In the Twilight of Social Structures

2 downloads 246 Views 1MB Size Report
whether consumption activities in everyday life can be explained by structural ...... characteristics of postmodernism are a critical attitude towards science and.
TURUN YLIOPISTON JULKAISUJA ANNALES UNIVERSITASIS TURKUENSIS

SARJA - SER. B OSA TOM. 263 HUMANIORA

In the Twilight of Social Structures: A Mechanism-Based Study of Contemporary Consumer Behaviour

By Pekka Räsänen

UNIVERSITY OF TURKU Turku 2003

From the Department of Sociology University of Turku, Finland

Supervised by: Professor Seppo Pöntinen Faculty of Social Sciences Department of Sociology University of Turku

Examined by: Professor Visa Heinonen Faculty of Agriculture and Forestry Department of Economics and Management University of Helsinki and Professor Timo Toivonen Department of Marketing Turku School of Economics and Business Administration

ISBN 951-29-2507-9 ISSN 0082-6987 Painosalama Oy – Turku, Finland 2003

…. and did you exchange a walk-on part in the war for a lead role in a cage? (R. Waters & D. Gilmour)

ABSTRACT

This study examines contemporary consumer behaviour from theoretical and empirical perspectives, and analyses the effects of economic and socio-demographic factors on the consumption patterns of individuals. The role of consumption in the advanced societies is discussed in terms of theoretical notions of postmodernism and the consumer society. In the empirical part of the study, the focus is on testing the arguments that have been presented in postmodern sociology. The central question is whether consumption activities in everyday life can be explained by structural factors. Consumption is measured as household expenditure (the actualised level of consumption) and as individual consumption desires (the imaginary level of consumption). The empirical analyses are based on cross-sectional data from the Finnish Expenditure Survey of 1998 and from the Finland 1999 Survey. The findings show that household expenditure and individual consumption desires are clearly influenced by economic and socio-demographic factors. However, the factors selected explain expenditure better than they do consumption desires. The study indicates that consumption is affected by linkages between a number of different conditions, which can be approached rather similarly at both levels of analysis. Throughout the study, interpretations of explanatory social mechanisms are offered in describing how individual behaviour is connected with economic and sociodemographic factors. Keywords: Consumption, postmodernism, consumer society, structural factors, explanatory social mechanisms.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS LIST OF FIGURES AND TABLES 1 INTRODUCTION 19 1.1 The many faces of consumption 20 1.2 Problem area and objectives 27 1.3 Structure of this work 29 2 POSTMODERNISM AND CONSUMER SOCIETY 33 2.1 From modern to postmodern society 33 2.1.1 Weakening of modern social structures 36 2.1.2 Living conditions after modernism 43 2.2 Postmodern society as consumer society 48 2.2.1 Rise in the level of consumption 49 2.2.2 The market as the main exchange mechanism of goods and services 53 2.2.3 Consumption as appropriate daily activity 58 2.2.4 Formation of consumer identities 62 2.3 Summary: Postmodernism and consumerism 67 3 UNDERSTANDING CONSUMPTION FROM A STRUCTURAL PERSPECTIVE 73 3.1 Economic resources 75 3.2 Social class 81 3.2.1 Class as cultural structure 83 3.2.2 Class as subjective experience 89 3.3 Age, gender and the type of household 94 3.4 Place of residence 99 3.5 Lifestyle and other relevant consumer groupings 103

3.6 Summary: Understanding consumption in contemporary society 110 3.6.1 Methodological difficulties related to postmodern notions 111 3.6.2 Approaching consumption from a structural perspective 113 4 EMPIRICAL STUDY: DATA AND METHODS 121 4.1 Empirical research design 121 4.1.1 Research questions 124 4.1.2 Social change in cross-sectional studies 126 4.2 Data sets and their validity 129 4.2.1 Selection of explanatory variables from the Finnish Expenditure Survey 1998 131 4.2.2 Selection of explanatory variables from the Finland 1999 Survey 134 4.3 Comparability of data sets and methods of analysis 136 5 EXAMINATION OF HOUSEHOLD PATTERNS OF CONSUMPTION EXPENDITURE 143 5.1 Expenditure on consumption 143 5.2 Exploring household expenditure on the selected categories 149 5.2.1 Expenditure on everyday life 154 5.2.2 Expenditure on culture and luxury 159 5.2.3 Expenditure on hedonistic consumption 165 5.2.4 Expenditure on home equipment 170 5.3 Summary of findings: Structural conditions and consumption expenditure 175 6 EXAMINATION OF INDIVIDUAL CONSUMPTION DESIRES 183 6.1 Subjective consumption experience measured as consumption desire 183 6.2 Exploring subjective consumption experiences 187 6.2.1 Differences in hedonistic consumption 192 6.2.2 Differences in culturally distinctive consumption 198 6.2.3 Differences in future-oriented consumption 202 6.2.4 Differences in change of dwelling or car 207 6.3 Summary of findings: Structural conditions and subjective consumption experience 212

7 CONCLUSIONS 219 7.1 Socio-demographic characteristics of consumption 220 7.1.1 Conditions based on economic resources 222 7.1.2 Conditions based on class 224 7.1.3 Conditions based on basic social categories 228 7.1.4 Conditions based on place of residence 231 7.2 Theoretical implications 233 7.2.1 Specific characteristics of the measures 235 7.2.2 Importance of explanatory social mechanisms 238 7.2.3 Individualistic features and structural conditions of consumer behaviour 239 REFERENCES 251 APPENDIX A 269 APPENDIX B 287

ACKNOWLEDGEMENTS

Many people have helped me in my efforts to achieve this thesis. It must therefore be said that the acknowledgements cannot provide a comprehensive list of credits. Instead, the following acknowledgements mention only those individuals who have given me the most important support and advice in the course of my work on this thesis. I wish to express my gratitude to my supervisor, Professor Seppo Pöntinen (University of Turku) for his advice on various theoretical and empirical questions. This thesis would have never seen the light of day without encouragement from Professor Timo Toivonen (Turku School of Economics), who originally introduced me to the sociological study of consumption and who is one of the outside examiners of the thesis. The other outside examiner is Professor Visa Heinonen (University of Helsinki); I am grateful for his detailed comments and constructive suggestions regarding the manuscript. In addition, I would like to thank my colleagues Jani Erola, Juho Härkönen and Pasi Moisio, who gave me useful feedback and showed genuine interest in my work. I also appreciate the comments given by Dr. Hannu Ruonavaara (University of Turku) on the theoretical parts of the manuscript. Most of the ideas presented in this thesis originated during 1999 and 2000, when I was involved with two research projects funded by the Academy of Finland. The first project was headed by Professor Olli Kangas, the second by Professor Kari Immonen (both at the University of Turku). These rather different projects provided me with methodological inspiration and helped me to understand what academic research is all about. I have written the thesis while working at the Center for Knowledge and Innovation Research (CKIR) of the Helsinki School of Economics and at the Department of Sociology, University of Turku. I wish to thank my colleagues and other people at these institutions. In particular Dr. Pekka Isotalus (University of Helsinki) and my colleagues Sami Kortelainen, Antti Kouvo and Olli Pyyhtinen have made important contributions through various intensive discussions. I also would like to thank research assistant Suvi Mäntylä for her contribution. In addition, Professor Pekka Sulkunen (University of Helsinki), Professor Liisa Uusitalo (Helsinki School of Economics) and Dr. Terhi-Anna Wilska (Finnish Youth

Research Network) have pointed out important things about the empirical study of consumer behaviour. I am grateful to the Emil Aaltonen Foundation and the Finnish Economic Society for the financial support of my research. My English has been revised by Dr. Ellen Valle of the Department of English, University of Turku. Finally, I want to thank my parents and my closest friends for their informal support, as well as for making me occasionally forget that one day I would have to finish this work. Henna, once my fifteen-year-old girlfriend and now my wife, has continually shown me what is good in life and what is not.

Turku, August 2003 Pekka Räsänen

LIST OF FIGURES AND TABLES

Figures Figure 1 Constitutive elements of consumer behaviour 25 Figure 2 Logarithmic time series of GDP and private consumption in Finland, 1900-1999 50 Figure 3.1 Economic resources and their link to consumer behaviour 79 Figure 3.2 The composition of class and its linkage to consumer behaviour 88 Figure 3.3 The origins of class identity and its linkage to consumer behaviour 93 Figure 3.4 Age, gender and type of household and their links to consumer behaviour 98 Figure 3.5 Place of residence and its link to consumer behaviour 102 Figure 3.6 The origins of lifestyle and its linkage to consumer behaviour 109 Figure 7 Different approaches to consumption 243 Tables Table 5.1 Main expenditure groups and average household expenditure 145 Table 5.2 Selected expenditure categories and average household expenditure 147 Table 5.3 Varimax-rotated factor matrix of the selected expenditure categories 150 Table 5.4 Scales for different types of consumption expenditure 152 Table 5.5 Expenditure on everyday life by independent background variables 156-157 Table 5.6 Expenditure on culture and luxury by independent background variables 162-163 Table 5.7 Expenditure on hedonistic consumption by independent background variables 166-167 Table 5.8 Expenditure on home equipment by independent background variables 172-173

Table 5.9 Expenditure dimensions analysed and their determinants 176 Table 6.1 Consumption desires related to the presented items 186 Table 6.2 Varimax-rotated factor matrix of the selected consumption items 188 Table 6.3 Scales for different types of consumption desires 191 Table 6.4 Hedonistic consumption by independent background variables 194-195 Table 6.5 Culturally distinctive consumption by independent background variables 200-201 Table 6.6 Future-oriented consumption by independent background variables 204-205 Table 6.7 Dwelling or car change by independent background variables 210-211 Table 6.8 Consumption desires analysed and their determinants 213 Table 7 Consumption dimensions and their connections with structural conditions 234

1 INTRODUCTION

‘Consumption’ is among the most widely used words in everyday life. It is a relatively mundane concept in many ways. Especially when it involves the behaviour of private individuals, we understand almost without reflection what type of activities the term actually refers to. It is used for instance in connection with shopping, going to the movies or eating in restaurants. Private consumption has also long been of interest as a subject of academic research. Within economic history, for example, some aspects of individual and household consumption have been studied for more than two hundred years. According to many writers, however, there has been a relatively sudden expansion of interest in the topic during the last few decades (e.g. Fine & Leopold 1993: 39-41; Miller 1995). Today, consumer research is one of the most visible areas in social sciences. The field attracts many disciplines, from economics, psychology and sociology to marketing, business studies and so on. Because of this, there exist many compelling ambitions for the development of theoretical approaches. Different approaches usually draw from the theories of different disciplines. It is normal, for example, to classify models of consumer behaviour into economic, psychological and sociological models, following broad disciplinary divisions (Uusitalo 1998: 215; Miller 1995; Bocock 1993). It is reasonable to argue that any attempt to approach consumption from a general perspective is likely to fail. The point of departure of this study is in theoretical and empirical research associated with the sociological tradition. This means that only a very limited number of interesting questions within the field of consumption will be covered. However, such a frame provides still much too broad a basis for our research agenda. This is because the sociological study of consumption itself is a vast and highly multidimensional enterprise, which has been flourishing worldwide for decades. The language and concepts of consumer research have been widely adopted in different areas of theoretical and empirical research. Many contemporary writers have argued that a loose ‘consumerist perspective’ is now commonly employed to frame the

19

discussion in many areas within the discipline (e.g. Campbell 1995: 101; Miles 1998). It is true that nowadays almost all special fields seem to have their own consumer-oriented perspectives. In such fields as urban sociology, the sociology of health or the sociology of sport, research problems are typically approached in the light of varied consumption activities. The focus of analysis is of course often restricted to certain specific phenomena; nevertheless a view that is shared among many culturally oriented studies is that varied social activities can be approached in terms of consumption (Räsänen 2000: 229). This has resulted from the increased role that mass consumption is argued to have achieved in advanced societies. In the presence of many distinctive perspectives on consumption, the field of this study has to be carefully defined. That is the concern of this chapter. I first try to offer a satisfactory working definition of consumption. This is followed by a description of the problem area and objectives of the study. I then outline the structure of the work. The general purpose of the chapter is to present a feasible approach to the theoretical and empirical analysis of contemporary consumer behaviour. 1.1 The many faces of consumption Consumption is approached in this study from a sociologically relevant perspective. This is to say that the examination of the phenomenon is restricted to the level of the role of individuals as consumers. It must of course be recognised that many aspects of consumers’ roles are also directed by other factors, at other levels, than those that have been seen as interesting in sociological – or perhaps, more generally, cultural – terms. Factors at the physiological and psychological levels, for example, are often important. While the effects of these kinds of phenomena will not be detailed in this study, it is acknowledged that consumer behaviour is always affected by various elements at different levels of analysis. Before going into the objectives of this study more specifically, however, let us take a brief look at the term ‘consumption’ itself. What does it mean, and how is it used in this study? There are many ways to define what consumption is, depending on which aspects of the concept we wish to emphasise. According to dictionaries, it can be defined as the act of using up different goods and

20

services (Collins Cobuild 1990: 303; OED Online 2002). Under this definition, consumption can be summed up as an activity that involves both the purchase and the use of goods and services. This seems a rather valid description; the term is generally used to refer to this type of activity in our daily lives. Buying a new car or other durable, for instance, provides a good example. However, while it is useful to approach consumption from the viewpoint of purchasing and using products, this also results in an unsatisfactory definition. This is because the definition does not take into account the fact that consumption is not a compulsory activity; that is, that a consumer choice is based on the consumer’s free will (Douglas & Isherwood 1978: 56). This means that even if there are many objects that we have to consume, the concept should always include a degree of sovereign choice. In this sense, such activities as paying taxes or fines do not represent acts of consumption. The dictionary definition can also be seen as unsatisfactory, since it reduces the use of the term to actualised economic activities. This is why consumption is often understood as a far more complex concept than economic activity alone. Among sociologists, consumption is usually defined as a cultural phenomenon that consists of acts involving the selection, purchase, maintenance, repair and disposal of any product or service (e.g. Campbell 1995: 102; Boden & Williams 2002: 494-495; Lee 1993). This also draws attention to the preparation and consequences of the actual consuming process. Consumption is thus seen as consisting of all those processes and influences that involve both the consumer and the product before and after such a purchase (Miles 1998: 4). From this point of view, consumption can be understood as a wider process, which describes what happens to objects once they have reached the attention and the hands of purchasers. It is true that most sociological views assume that consumption is much more than a merely economic phenomenon. It involves hedonistic and creative pursuits, in which imagination and emotions may play an important part. In many cases the consumed objects and services satisfy some basic human need. Naturally, it is possible to approach various forms of consumer behaviour from the viewpoint of need satisfaction (e.g. Csikszentmihalyi 2000: 268-269; Antonides & van Raaij 1998). If for instance we start from the hierarchical taxonomies developed by

21

motivational theorists, such as that by Abraham Maslow (1968), it is possible to argue that human behaviour is always determined by certain needs. The most basic needs that can motivate consumer behaviour are physiological survival needs, such as the need to eat, drink and sleep. The need for sex is also considered as a basic need in Maslow’s taxonomy. As soon as these needs are satisfied, certain higher needs will attract the individual’s attention. According to this theory, survival needs are followed by safety needs, such as the need to live in a stable environment. The need for love and belonging come next in the scheme; these are ultimately based on the fear of social isolation. The need for self-esteem and selfactualisation represent the highest level of needs in the taxonomy; these needs are related more generally to personal growth and self-realisation. The implications of the existence of different sets of needs are visible through various consumer activities. This means that it is possible to say that individual consumer behaviour is actually directed to meet certain lower or higher needs. For example, eating and drinking satisfies the basic need for nourishment. A new car, on the other hand, may function as a sign of wealth and style as well as a means of conveyance, and may be used to meet the need for self-esteem. Nonetheless, while it is unquestionable that consumption is often driven by various needs, it can also serve certain other purposes. Particularly in current theoretical writing, the symbolic importance of consumption behaviour is often stressed (e.g. Campbell 1987; Lee 1993; Noro 1995; Uusitalo 1998). According to these views, the consumption experience itself has to be taken into account if we wish to see the whole picture of consumption. In short, it is considered that consumption is more than just a realised activity. In these considerations consumer behaviour is treated as a concept that effectively describes our daily lives. It involves planning, rational and emotional calculation, and many aesthetic considerations. In other words, consumption is not merely an activity of spending money; it is also an activity that provides an arena for various personal and social experiences.1* There are many aspects of consumption that can be highlighted in definitions. The preceding review has shown that definitions draw attention to different types of phenomena and lead to different research questions. In this study, consumption is defined as a complex behavioural pattern that is *

Superscript numerals used in the text refer to numbered notes at the end of each chapter.

22

oriented towards the acquisition of any product or service. This definition takes it for granted that consumption necessarily requires expenditure, either actually or at the imaginary level. At the same time, however, it should be stressed that the definition does not equate consumption with purchase in the conventional sense. The actualised level of consumption refers to realised expenditure. For example, payments for all goods and services constitute actualised consumption. Payments result from a variety of daily purchases, such as for example daily payments to the supermarket or bus fares. In addition, the many operating costs arising from the obtaining of these things also have to be included. The actualised expenditure on certain products thus includes more than one type of spending (see Ilmonen 1993; Antonides & van Raaij 1998). The imaginary level of spending is clearly a more complicated concept to define. In principle, it involves such things as the thoughts, aspirations and emotions related to consumption. A relatively similar framework has been utilised in previous analyses of certain consumption activities. Shopping, for example, has been approached vaguely from the perspective of a particular type of collective activity that does not necessarily require spending at all (see Lehtonen 1994; Lehtonen & Mäenpää 1997). At a more restricted level, however, the imaginary level of spending can refer to practices that aim at realisation, but which have not yet been realised. From this perspective, the definition of such activities as shopping should include a clear orientation towards spending. It is in this way that the imaginary level of consumption is analysed here. According to this interpretation, consumption is very much a personal experience of what an individual wishes to do or might do in the future. In this context, it does not exactly refer to the type of activity of obtaining pleasure simply for the imagination (see Campbell 1987, 77-78; Campbell 1995: 118-119; Boden & Williams 2002). Rather, it refers to an individual’s willingness to realise the consumption practices in question. Realised consumption and imaginary aspects related to it can thus be seen as the opposite sides of a same coin. The realised side shows how certain consumption desires have actually been put into practice. The imaginary side, on the other hand, reflects how certain consumption practices might be carried out at a convenient time.

23

This definition of consumption is relatively closely tied to the assumption of personal purchase. Buying a magazine, or thinking of doing so, is a typical act of consumption. The same also goes for instance for payments for housing or for train tickets. Despite this, the definition can also cover indirect forms of spending, such as investing in shares or servicing a debt. At the theoretical level, then, consumption is discussed in the context of all types of everyday activities that involve purchases and the spending of money. Empirically, the actualised level of consumption is measured as household expenditure. The imaginary level of spending, on the other hand, is approached through the individual’s subjective consumption desires. As noted above, it is generally believed that individual behaviour is prompted by certain needs, but there are also other factors that are considered to play an important role that can explain observed patterns of consumption. For example, while the need for shelter is perhaps the most fundamental factor that drives people to seek accommodation in the first place, this need does not tell us what kind of housing one has acquired or how much one is paying for living there. There are many situational factors, such as the availability of housing and the current price level in the market, which can probably provide better alternatives for explaining an individual’s housing choices and expenditure. Furthermore, such factors as the individual’s income level, family phase, marital status, age and residential history may predict housing choices much more effectively than any need-based explanation. At the most abstract level, there may even be certain collective social norms which lead consumers to consume in particular ways. Nowadays, for instance, it is highly fashionable in advanced societies to prefer ecologically produced products. It can be argued that in the context of everyday activities such factors provide relatively competent predictors for individual consumer behaviour. In any case, it is true that different people consume different items rather differently in their daily lives. From this perspective, it might be feasible to view consumption determinants in terms of the basic characteristics of individuals. A graphic representation of the elements that can be seen as fundamentally affecting individual consumption behaviour is shown in Figure 1.

24

Consumption determinants • Personality factors • Earlier consumption experiences • Situational factors • Economic and socio-demographic factors

Consumption preferences

Consumer behaviour

• Needs and aspirations • Values and beliefs

Figure 1 Constitutive elements of consumer behaviour The figure assumes that the elements which affect consumer behaviour can be divided into two distinct groups. To highlight the analytical difference between these groups, they are labelled here as ‘determinants’ and ‘preferences’. The first group, that of consumption determinants, consists of such factors as personality, earlier experiences, situational factors and socio-demographic characteristics. Factors of this type are understood as the basic factors that should be taken into account when we start to analyse individual consumer behaviour. Consumption is thus understood in this study as a multidimensional behavioural pattern that requires information about the consuming actor at the level of the sociocultural context in order to be understood. It follows that information on consumers’ characteristics may provide highly relevant explanatory factors at the levels of both actualised and imaginary consumption. Consumption determinants are usually connected to individual consumption somewhat indirectly. This is to say that their effect is often mediated via individual preferences. For example, similar personality traits or socio-demographic characteristics may be connected to very different sets of needs and aspirations, leading to different patterns of consumption. However, it is also possible for certain determinants to influence consumption directly. The amount of money available, for instance, restricts many consumers from purchasing expensive cars or travelling abroad many times a year. In addition, it is reasonable to believe that 25

consumption behaviour itself may affect some of these determinants, such as situational factors. The adaptation of a certain consumer habit often increases the possibility that some other habits will change as well. For example, if I start to play golf it is very likely that my spending on such activities as travelling will also increase; this is because golf, as an object of leisure-time consumption, often requires relatively extensive travel. The second group of elements in the figure, labelled ‘consumption preferences’, consists of such factors as the individual’s needs, aspirations, values and beliefs. They can be seen as connected to consumption behaviour relatively directly. For example, choosing a certain hobby over another may be based on differentiated value-systems or simply on different personal needs. At the same time, however, consumption practices can also affect individual preferences. Trying a certain new product, for example, can change one’s earlier beliefs or even generate a totally new type of need. But it is also true, of course, that many preferences can be seen as originating from factors that are labelled here as consumption determinants. This means that values and aspirations, and even some basic needs, are often connected to the shared situational factors and sociodemographic characteristics of individuals. For instance, buying a house in a ‘better’ neighbourhood may be a more important decision for a middleclass family than for a working-class one. The role of other determinants, such as personality factors, should not be ignored either. In many situations, personality characteristics create the basis for the emergence of certain consumption preferences. For example, we may assume that generally outward-oriented people (extroverts) are perhaps more likely to try new types of restaurants than generally inward-oriented people (introverts). This may lead extrovert people to change their attitudes and beliefs towards food more often than introverts. Overall, it can be assumed that there are always certain underlying principles which both constrain and enable many consumer activities. Figure 1 stresses that the constitutive elements of consumption provide plural rather than exclusive sources of explanation. The focus in this study, however, is on those factors that are specific to sociological models of consumption. It is assumed that behavioural differences become visible via the economic, social and cultural characteristics of individuals.

26

1.2 Problem area and objectives While consumption undoubtedly arises out of many different and interconnected decisive factors, this study concentrates on the significance of economic and socio-demographic factors. More generally, this means that other important determinants will not be considered in this study. My intention is to examine how individual consumption patterns might be affected by basic structural factors in society, and to offer plausible explanations for these observations. I ask whether systematic differences can be found in consumption behaviour between different social groups. I also consider which factors can be seen as the most important structural conditions in different fields of consumption. My approach to these research problems is two-fold. I first present a theoretical discussion of trends in contemporary consumption patterns in advanced societies, leaning on the debate about the consumer society. The purpose of this part of the study is to explore the central arguments that have been presented with regard to the changing character of private consumption. Secondly, certain aspects of consumption will be examined empirically, utilising two Finnish data sets. The empirical part of the study will concentrate on testing some of the central arguments presented in the literature. The data used here are derived from the Finnish Expenditure Survey 1998 and a nation-wide postal survey carried out in 1999 (see Chapter 4). By combining theoretical and empirical approaches, the study offers a contribution to the ongoing debate on the changing role of structural conditions in individual behaviour, providing a sociological interpretation of certain key aspects of contemporary consumption. At a more abstract level, the study addresses certain claims that have been presented as to the weakened effects of modern social structures. Particular attention is paid to arguments connected with the postmodern and with postmodernism. In the social sciences, a distinction is typically made between the two terms. According to many writers, ‘postmodern’ can be understood to refer to a set of historical conditions, ‘postmodernism’ to an emerging intellectual and cultural orientation (e.g. Lemert 1997: 21; Blau 1993). In this sense, they represent distinct dimensions of contemporary social and cultural reality. However, it is also true that certain common themes appear in the context of both the postmodern and postmodernism.

27

According to Steven Best and David Kellner (1997), for example, there are three different aspects that can be identified as most recognisable in the literature. These are 1) the conditions of social life, 2) the cultural products created within it, and 3) the practices by which we study and interpret social life and cultural products. Briefly, the first aspect refers to arguments suggesting that modern arrangements and institutions are being replaced with postmodern ones. The second aspect, on the other hand, is related to the growing importance of new forms of political participation and communication, and of new type of cultural products, such as multimedia applications. The third aspect can be understood in terms of a shift in scientific knowledge and in aesthetic considerations (Best & Kellner 1997: 124). In this study, the terms ‘postmodern’ and ‘postmodernism’ are used in a restricted sense, to refer to the conditions of social life, i.e. to the decline of modern social arrangements and institutions. This definition can naturally cover only a limited number of questions. For example, the second and third aspect in Best & Kellner’s categorisation will be excluded almost completely. Since the late 1960s, postmodernism has been developing into a multidisciplinary approach to almost everything related to contemporary life, attitudes and values. At the broadest possible level, the terms ‘postmodern’ and ‘postmodernism’ indicate a movement away from the mechanistic worldview of modern science, along with a rejection of Enlightenment optimism, faith in reason, and the possibility of an ‘original’ human nature and of values which transcend culture (e.g. Pescasolido & Rubin 2000: 57-58; Lemert 1997). Today, the terms seem to draw together varying interests from the social and cultural sciences, the arts, literature, economics and theology. It would be impossible even to list the basic use of the terms in all of these areas. Thus only some of the key notions that can be regarded as essential to sociology (and to the sociology of consumption in particular) will be referred to in this study.2 From this point of departure, the research problems of the study are related to the question whether modern social structures can still explain and predict individual behaviour. In general, it is assumed in this study that socio-demographic structures affect everyday life not only at the levels of consumption expenditure and purchases, but also at the level of consumption desires.

28

Throughout the study, I try to develop explicit assumptions about the role of structural determinants and the ways in which individuals act in response to their social and cultural environment. In this sense, I try to avoid the tag of ‘variable sociology’ (Esser 1996), which has been associated with quantitative studies which try to describe and explain social behaviour simply in terms of statistical associations between the variables examined. My aim is to establish a solid theoretical grounding for methodological choices and empirical results. 1.3 Structure of this work The book is structured in such a way that each chapter includes a comparatively self-contained discussion of the issues covered in the study. The work falls into two parts. The first part deals with theoretical issues concerning consumption. The second part concentrates on the empirical analysis of consumer behaviour. In all chapters, one aim is to demonstrate that consumption is often a goal-directed activity, which can be approached in terms of social and cultural determinants. In developing my arguments, I will evaluate theoretical claims concerning changing consumption patterns in a relatively critical manner. Chapter 2 examines the general debate about postmodernism and the consumer society. First, current theoretical implications related to social change will be approached. I ask to what degree it is reasonable to argue that the majority of people live in conditions that can be described as postmodern. Basically, the essential features of the postmodern condition are approached through claims as to the weakened role of traditional social structures, particularly that of social classes, and the decline of modern social arrangements. These arguments are discussed in the light of the conventional features of modern social life. Secondly, the chapter examines the increased role of consumption in the advanced world. The premises and consequences of this are approached by concentrating on certain main elements of consumer society. In this sense, the chapter aims at describing the way in which consumption activities have become a ubiquitous aspect of everyday life. Chapter 3 discusses the role of basic economic and social factors affecting consumer behaviour, examining the associations by way of which social class, income, age, gender and other characteristics can influence

29

individual consumption patterns. Particular attention is paid to the importance of mechanism-based approaches in explaining social phenomena. This means that in the examination of various structural determinants I discuss the ways in which different forms of resources may regulate people’s expenditure, and how they also influence the planning of future expenditure. The characteristics of structural factors will also be explored by discussing how they have been measured in empirical studies. In general, the aim of the chapter is to show that it is still relevant to refer to traditional sociological factors in the empirical analysis of consumption. In addition, the chapter considers certain theoretical and methodological problems proposed by postmodern arguments for empirical research. The empirical questions of the study will be explored in the following three chapters. Chapter 4 outlines the empirical research design of the study. Research questions will be summarised on the basis of theoretical oppositions between structural and individualistic approaches to consumption. This is followed by a description of the data sets utilised. The methods used, and issues related to the comparability of the data sets, will also be briefly introduced. The focus of the chapter is on problems of operationalisation and on the validity of data sets measuring consumption at different levels of analysis. Chapters 5 and 6 present the results of the empirical analyses. I start by asking whether some of the most important consumption activities can be approached as distinct consumption dimensions. In Chapter 5 I show how socio-demographic variables influence household patterns of expenditure. Chapter 6 presents the connections between subjective consumption desires and socio-demographic factors. The analysis of the empirical data demonstrates the impact of structural factors on the consumption behaviour of individuals at both the actualised and the imaginary levels. Chapter 7 presents and discusses the conclusions of the study. Basically, the chapter attempts to interpret the empirical results in the theoretical context of the work. The results of the empirical analyses of different data sets are first compared, in order to determine the actual power of selected background variables (class, age, gender, etc.) in explaining different consumption practises. After this, I re-examine the structural explanations of consumer behaviour and their postmodern

30

critique at a more abstract level. The conclusions draw mainly upon the empirical findings and their interpretations in the analyses, but a variety of other theoretical writings dealt with in this study will also be considered.

Notes 1

It is interesting that even though consumption is often defined as something which is much

more than expenditure on goods and services, it is not usually studied according to this definition. Most empirical studies of consumer behaviour continue to measure it simply in terms of expenditure (see Uusitalo 1979; Toivonen 1991; Wilska 1999; Fisher 1987). It is true, of course, that many qualitative studies of individual consumer behaviour have also been published (e.g. Lehtonen 1994; Chattoe & Gilbert 1999). However, qualitative studies often deal with relatively different issues compared to quantitative ones. For example, generalisations in quantitative studies often include typical sample-to-population generalisations, while in qualitative studies generalisations often consist of drawing broader theoretical constructions from a relatively small number of cases represented in the data (Räsänen 2001, 65-68). 2

Even given this obvious restriction, questions associated with the postmodern and

postmodernism are by no means analysed in detail in this study. For a detailed and systematic analysis of the concepts, theoretical models and key writers associated with these terms see Winquist & Taylor (2000).

31

2 POSTMODERNISM AND CONSUMER SOCIETY

The concepts of the postmodern and of postmodernism have been among the most widely discussed themes in the social sciences during the last two decades. These terms have been brought up in almost every imaginable connection, when dealing with something essential to contemporary social life or its distinctive characteristics. In a sociological context the terms refer broadly to the ongoing societal process and the emerging new social condition. This new condition or process is typically theorised as against the previous period of modernity. In this chapter, the central implications behind postmodernism are discussed in a structural frame of reference. In general, I consider whether there can be a postmodern society, distinct from a modern one. I examine the relationship between social structures and social life primarily by looking at changes in the role of production. In addition, however, I also approach postmodernism through a set of cultural beliefs and values that exist in society. The purpose is to shed light on the key arguments that discussion of the term has brought to sociological research. I also discuss how arguments concerning the increased significance of consumption are influenced by certain conditions of contemporary society. It is assumed here that one of the key arguments about postmodern society concerns the changing role of consumption. Consumption can be seen to play an increasingly important role in how we construct our social lives. But, while the majority of the world’s population still lives on the edge of adequate survival, is it feasible to argue that consumption simultaneously occupies more and more of the time and energy of those living in advanced societies? In the latter part of the chapter I discuss the ways in which consumption can be understood to as one of the central practices of our everyday lives. 2.1 From modern to postmodern society Any attempt to outline the process of societal change or a new social condition must be theorised against the previous and following stages. The views of modern society, for example, that were theorised in the writings of 33

Émile Durkheim, Karl Marx, Max Weber, Georg Simmel or Ferdinand Tönnies juxtaposed traditional and modern society. These writers tried to describe new patterns of work, religion, social interaction and culture. Basically, all of the classics of the discipline have discussed the basic forms of new social relationships and have considered their effects on individuals, including the development of personalities, social structures and belief systems under those conditions (Pescasolido & Rubin 2000: 53-54). Thus portrayals of modern society offer an analytical distillation of basic differences in societal forms and of the characteristics of certain historical eras. The concept of the modern has many dimensions. For example, we can speak of modernity in terms of fashion, attitudes or values. In general, however, the emergence of the modern concept of social life is contemporaneous with the emergence of the concept of modern society (e.g. Owen 1997: 8; Hautamäki 1996). The line between traditional and modern runs along the assumed differences in social interactions and institutions. On the one hand, we can see modernity as a historical process, which has transformed societies based on a traditional order into something new that we call by the name ‘modernity’ or ‘modern’ society (Wilenius 1997: 25). The birth of modern technology, the emergence of industrialisation and a belief in rationality and material growth can be seen among the key features in this respect. On the other hand, descriptions of modern society also offer interpretations about the modern social condition. This can be partly understood by the increasing possibilities that the industrialisation of material production has made available to ordinary people. Among the key features of modern social life are a rising standard of living, a greater element of choice and multiple group membership. Perhaps the single most important factor from this perspective, however, is the centrality of social class as a determinant of individual life experience (e.g. Esping-Andersen 1993: 7; Bauman 1998: 17-18). The importance of class goes back to the roots of classic sociology and political economy. Following the development of industrialised and meritocratic societies, inherited privileges and statuses began to lose their importance as sources of social inequality. The emergence of class was originally linked to the notion of the surplus value of material products, referring to that which is left over after the subsistence needs of the

34

members of a society are satisfied. For the first time, perhaps, it became possible to manifoldly exceed the subsistence level of society. As a consequence, inequalities of wealth and power began to be attached to positions in the capitalist production process. In structural terms, so to speak, people’s situation in modern society was seen in terms of class position, which was determined by their relationship to the means of production and work. Views of postmodern society do not differ with respect to juxtaposition of the previous and following stages. It is believed that many of the trends involved in the transition from traditional to modern society have reached their limits in advanced industrial societies, and that sociocultural change is therefore taking a new direction. While very little agreement can be found as to whether postmodernism is a higher state of modernism or a fundamental critique of it, certain characteristic themes have become clearly identifiable in the postmodernist discussion.1 Relevant themes with regard to the area of this study can be found among both sociological and philosophical ideas. From a philosophical point of view, the most widely mentioned characteristics of postmodernism are a critical attitude towards science and the notion of progress, a relative emphasis on the other and on difference in general, and an interest in the particular and the indigenous. A dislike of doctrines and of sweeping generalisations, such as various ‘grand’ narratives, is also a common characteristic. (Etürk 1999: 27; Lemert 1997.) Perhaps the strongest claim in this respect is the postmodern awareness of the breakdown of the grand narratives in the Western world and the legitimisation crisis brought about by this. Numerous writings can be found stressing this dimension of postmodernism, concerning all possible fields – from contemporary political and cultural phenomena to empirical research traditions (e.g. Lyotard 1985; Beck et. al. 2003; Harvey 1990; Bauman 1997). From a sociological point of view, on the other hand, two fundamental characteristics can be abstracted: changes on the one hand in the institutional structure of society, on the other in value and belief systems in contemporary Western culture (Inglehart 1997: 27-28; Uusitalo 1998). According to David Owen (1997), for example, the modern understanding of society in its general sense has become highly problematic in terms of

35

the relationship between state and society and the impact of globalisation. The central question at issue concerns whether the features of socioeconomic modernism, specified in terms of the structural stability of nation-state capitalism, the centrality of class and class-based politics, and sustainability through material production, are still valid in contemporary society. These questions have been contrasted from the point of view of the increasingly disorganised character of capitalism, the fragmentation of class allegiances, the increasing role of information technologies and the growth of a culture of consumerism (Owen 1997: 14-15). The emergence of new social movements, such as the environmental, lesbian and gay movements, and shifts in the circulation of knowledge, can also be understood as central elements of the postmodern social condition (e.g. Craib 1992; 189-192; Hautamäki 1996: 36-37; Kouvo 2000: 9-10). Since it is impossible to cover all the relevant notions related to modernism and postmodernism, only some of the most challenging ones will be discussed below. It can be argued that the most relevant topics for this study are the weakening force of modern social structures and certain changes in the living conditions of individuals. 2.1.1 Weakening of modern social structures The weakening of modern social structures, in particular of class, is perhaps the best-known claim of postmodernism in a sociological context. While there has never been a definitive answer as to what exactly class is and how it can best be conceptualised, much of the discipline of sociology is built around the idea of class. The study of social inequality, social mobility and social structure, for instance, has primarily been the study of class (Stewart et. al. 1980: 10-11; Owen 1997; Esping-Andersen 1993). In general, class has been seen as the true sociological building material in theoretical frameworks and one of the most important background variables in empirical analyses. According to some writers, the main achievement of the whole discipline has been in demonstrating how the underlying realities of class can structure behaviour, life opportunities, attitudes and lifestyles. For many decades, class theories dominated practically every sociological discussion and extended into new fields of research. But in recent times the impact of social class has been challenged in a number of ways.

36

According to Harriet Bradley (1996) the major challenges to class theory can be identified as three main strands. The first concerns the validity of class theories and analysis. It has been suggested that in the post-war period, and especially in the past three decades, the class structure has changed so rapidly that the old frameworks are simply no longer applicable. This is to say that new concepts of class and stratification are required. The second main strand is the critique that comes from gender, race and ethnicity theories. According to many writers, class theories are not adequate to explain the gender and ethnic differences found in contemporary societies. The mechanisms of these structures are not necessarily related to class in any way. The third challenge to class theory comes from perspectives focusing their attention on the diversity of social experience. This kind of critique endorses new forms of pluralism, grounded in the specific positions of different groups and in a view of society made up of almost atomised individuals. In the most extreme versions, all notions of collectivity are undermined. (Bradley 1996: 2-3, 2123.) Bradley (1996) suggests that the third form of critique can be labelled most evidently as the challenge of postmodernism. It is true that many postmodern views include notions about gender and ethnicity as well; these themes, however, are not usually referred to as the emergent features of society in a structural sense. In addition, one of the starting points of postmodern sociology in general is that modern society and its structural relationships no longer hold true. I therefore refer to the critics of modern class theory simply as ‘postmodern critics’. It is true that both the complex and shifting nature of class as a set of ongoing relationships and the difficulties related to defining them as a social category have led to a situation in which there are countless views of class. We can say that class is everywhere and it is nowhere (Bradley 1996: 45; Anthias 2001: 368-369). Naturally, this analytical vagueness is not a problem of contemporary writings alone. At least according to some writers, class theorising began and has continued more or less in confusion (e.g. Crompton 1993; Lee & Turner 1996). Since class has no physical signs or markers, it is hard to observe in real life. This fact has led to fundamental problems related to the sources of class, what kinds of classes there are and how class should be measured. At this point, however, I do

37

not go further into theoretical issues of the definition of class or how it can most plausibly be empirically operationalised. These questions will be discussed at a more detailed level in Chapter 3. Up to now, one can be content with a rather vague notion of class as the reflection of an individual’s position in production. This is to say that class can be understood in terms of professions, occupational hierarchies and the ownership of productive means. Perhaps this view is relatively near to a Marxist framework, but it is reasonable to believe that this kind of simple definition can roughly refer to all the most familiar traditional classtheoretical templates (see Waters 1997: 23-24; Bradley 1996: 14). In general at least, all dominant class theories – whether Marxist or Weberian – identify classes through material ownership and economic status and their connections to industrial capitalism. In the case of Weberian theory, of course, possession of professional skills and the significance of personal status, which is basically independent of class position, is also highlighted. In any case, the strongest basis of class remains in economic factors. The work of Daniel Bell (1977) is pioneering in challenging the validity of the modern class structure. Bell characterised post-war industrial society in terms of the dominance of service employment and the rise of new technical-professional groups, whose social position would splinter former class cleavages. He basically theorised a new class division in a society dominated by service employment, in which the ruling principle lay in the control of scientific knowledge (Esping-Andersen 1993: 9). Bell suggested that technological change would lead to a society in which a technical elite would emerge and the majority of the population would belong to the middle classes. According to Bell, post-industrial society, with its emerging new occupational structure, would affect not only the jobs and careers of individuals but also their life chances in a new way. In general, courses of actions carried out by people are no longer considered to be dependent on a class structure similar to that which was the determinant factor in industrialised society. People see themselves at least partly as reflexive individuals, and the importance of class structure, based on productive position, is understood to diminish. Bell (1977) saw the possession of scientific knowledge and technical know-how as the key factors in social inequality. Many similar views have been presented, stressing the role of culture, the control of professional

38

skills or other cultural qualifications as the key sources of social stratification (e.g. Gouldner 1979; see also Goldthorpe 1982). These views, amongst many others, claim that there have been so many changes in classbased explanations of modern society that the idea of classes in their traditional form has become highly problematic. One important notion included in all of these views is that the general prosperity brought about by post-industrial change has provided an increased standard of living and leisure opportunities for almost everyone in the industrialised world. A higher standard of living has also diminished economic inequality between classes. Likewise the traditional career, understood as the deployment of a single set of skills throughout the course of a working life, can be assumed to become more infrequent.2 On this basis, it is argued that contemporary societal processes are leading to a much more complex pattern of class relationships than simple forms of occupational polarisation. The structural inequalities of people’s life chances are now believed to be connected to class in certain indirect ways. Many critics stress the role of consumption in the formation of class positions (e.g. Bourdieu 1984; Saunders 1984; Gronow 1997). Accounts of a more equal society in terms of increased consumption possibilities have also been presented in this context (see Featherstone 1991; Strasser et. al. 1998). Broadly speaking, the relationship between social structure and behaviour has turned around, indicating that behaviour, such as consumption, has come to determine social class. The work of Pierre Bourdieu (1984) is of particular importance in this respect. According to Bourdieu, the variation between different social groups may nowadays be established rather than expressed by cultural activities, opinions and consumption lifestyles. Along with economic capital there are also cultural and social forms of capital, which determine individual lifestyle and consumption preferences. In Bourdieu’s view too, however, access to class-distinctive lifestyles still depends greatly on access to economic resources, even though he stresses the role of cultural capital as another dimension of social stratification. Thus class background continues to play a fundamental role in the life chances of individuals. Bourdieu’s work will be discussed in more detail in Section 2 of this chapter. In general, the various new models of class which have been presented are not been unproblematic. The existence of a somewhat undifferentiated

39

working class, for instance, appears to be a relatively far-fetched idea when we consider the contemporary occupational structure in advanced societies. A scrutiny of the skills required in present-day manual work suggests that the characteristics of the working class are somewhat unclear. For example, a notable part of typical working-class occupations in advanced societies actually consists of occupations requiring various specialised skills. There has been a general trend over the last few decades towards the professionalisation of many occupations (see Abbott 1988; Perkin 1996). Recently it has been noted that the class situation of information specialists in particular can be seen as paradoxical in terms of skills and professional status (e.g. Lewin & Orleans 2000; Blom et al. 2001). The growing power of information technology and new democratic styles of work among people of different ages and working experience have dislocated the position of traditional managerial authority in many institutions. At the same time, agriculture and forestry have been declining for decades in the national economies of the industrialised world in terms of both employment and productivity. It follows that service occupations are now seen as the key economic sector in advanced societies. During recent decades there has been a tremendous rise in the number of service occupations all over the Western world (e.g. Florida 2002: 73-74; 321-322; Goldthorpe 2000; Ferrera 1996; Nätti 1989). The growth of the service sector is regarded as an indicator of the decline of the working classes in contemporary middle class societies. This trend is supported by the fact that the proportion of traditional working class jobs of the labour force has slipped considerably. The main trend in the Finnish labour market, for instance, indicates the decline of traditional working class occupations and the growth of managerial groups (Blom et. al. 1998: 43; see also Erola & Moisio 2002). Due to these ongoing changes in working life, even some of the pioneers of class analysis have come to doubt the future possibilities of a class-based sociology (see Clark & Lipset 1991: 397-398; see also Bauman 1982). This means that structural differences between classes based on occupational characteristics are expected to melt away. In the view of Gøsta Esping-Andersen (2000: 69), the whole concept of social class needs radical rethinking if it is to remain the cornerstone institution of sociology.

40

These notions lead to more aggressive views that challenge the whole concept of social class as such, regardless of the class model in question. According to many social theorists, class analysis has become inadequate in recent decades because of the decline of the old modern hierarchies and the emergence of new plausible social divisions. On the one hand, this tendency has been characterised as the expansion and fragmentation of the middle classes. In other words, while everyday life in general has become more and more middle-class oriented, the existing ‘new middle-class lifestyles’ (Sulkunen 1992) have simultaneously become more heterogeneous. On the other hand, however, ethnicity, race, gender and age have been reinvented as the key discriminating structures in contemporary social life. Ethnic and racial categories as sources of inequality have been widely discussed in the literature during recent years (see McDowell 1999; Cronin 2000; Anthias 2001), and will not be examined here in any detail. In Chapter 3, however, I will return to gender, age and life-cycle, in connection with structural explanations of consumption behaviour. At the most radical level of criticism, all kinds of structuring elements have been challenged, not class structure alone. It is assumed that there has been a movement from collective expressions of values and lifestyles to individualism. According to Mike Featherstone (1991), for example, we are moving towards a society without fixed status groups based on occupational characteristics. He argues that in postmodern society the adoption for instance of consumer lifestyles has made the connection between a given lifestyle and a specific group practically irrelevant (Featherstone 1991: 83-84). The general suggestion behind these individualist arguments is that a society contains multiple social collectivities and groupings and that individuals can move relatively easily between these groups. There are of course many differences in the theoretical ideas growing out of the notion of individualism; their key arguments, however, include the notions of class fragmentation or even classlessness in one form or another. It is a common statement in the recent sociological literature that during the last decades of the twentieth century society has become increasingly fragmented. While certain ideas of social fragmentation have also been proposed in the classic writings of the discipline, and in the views

41

of virtually every modern social thinker, current postmodern theorists have carried the idea much further. It is assumed that while there might be differences in the values, attitudes and life practices of people according to age, class, education or other status indicator, these differences cannot plausibly be included under the broad conception of structure. A decline of the social significance of modern structures is predicted in part because of the greater role played by the media and other leisure activities in influencing people’s identities and their experience of daily life. One group of writers has argued that distinct groups, sharing common attitudes, values and tastes, will emerge. For some of these writers certain kinds of new middle classes are emerging, which will have a fragmenting influence on post-industrial class divisions (e.g. Sulkunen 1992; Douglas 1996). Such notions are in a sense paradoxical: class-based lifestyles do exist, but these new classes are not clearly structured according to the old hierarchies. This means that certain common conditions and styles of life are understood to represent the kinds of positions in society that are available as people line up in political and cultural positions. To others, social collectivities in contemporary society consist simply of consumption-based cleavages and various communities based on common interests (e.g. Maffesoli 1996; Hetherington 1998). The key idea in such views is the formation of new lifestyle enclaves, formed by people who share certain features of their private lives in common. Members of a certain enclave express their identities through shared patterns of appearance, leisure activities or consumption (Bellah et. al. 1985: 72-73). The majority of postmodern writers argue that class or occupational structures still play at least some role in the experience of everyday life, but their effect is relatively indirect and fragmented; in other words, they are connected to various structural conditions in multidimensional ways (Bradley 1996: 212-213). According to some writers, however, contemporary development is bringing an end to classes and to traditional sorts of inequality. In particular, the post-war changes in the institutions of European welfare societies are often credited as the key influence leading to a classless society. The expansion of the state apparatus together with mass culture and consumerism is eroding traditional class-based hierarchies. The change is not merely a matter of cultural democratisation in terms of the interaction of high and low cultures, but also a matter of the

42

democratisation of life chances. Ulrich Beck (1992) has taken this idea to an extreme level. He argues that Western societies are now undergoing a surge of total individualisation. Beck refers to classlessness as individuation, a process of transformation through which individuals will be liberated from all the social forms of industrial society. The effect of class, stratification, family and gender statuses, among others, will be wiped out from people’s everyday experience. According to Beck, in the future we will see an individualised society. This new situation can be characterised as one in which individual members of society do not express any clearly identifiable structural patterns in their behaviour. (Beck 1992: 87; 98-99.) The most extreme versions of postmodernism can of course offer only remote visions of a possible future when compared to present-day society. But there are also some welcome ideas behind postmodern views. Perhaps social life can be assumed to be more individualistic than before. That is to say that in contemporary advanced societies social structures may be understood to be quite loose and their effects indirect – in comparison of course to previous stages of societal development. Accordingly, the living conditions of individuals may be fundamentally different. 2.1.2 Living conditions after modernism As a socio-cultural philosophy, postmodernism undermines the traditional notion of culture as something fixed and hierarchical. Opposition to any notion of an artistic or cultural canon is thus fundamental to postmodernism. In particular fixed standards of high and popular culture are seen as irrelevant (e.g. McQuail 1994: 27-28; Fidler 1997; see also Peltonen 1996). For example, one can refer to traditional forms of mass entertainment, such as comics or television serials, as particular forms of art. In mass media terms, postmodernism thus sees fashion as more important than tradition. The key features of everyday existence in the postmodern era include the mingling of various life spheres and social distinctions, as well as the disruption of organisational structures (Uusitalo 1998: 215). Postmodernism can be seen as stressing forms of culture which are transient and experienced as appealing at the particular moment. The momentary nature of culture is believed to illustrate the experience of

43

everyday life as well. Anthony Giddens (1991), for example, refers to the contemporary cultural condition as post-traditional culture, based on the loss of traditions and customs. In general, the weakening effect of traditions is a consequence of individualisation, which began with the transition to industrial society and which intensified in the last decades of the twentieth century (Beck-Gernsheim 1996: 140). In the course of individualisation people begin to be released from traditional ties and relationships. The role of religion and family institutions, for instance, has been declining. In the post-industrial era, individualisation has led to a point at which the life of individuals has become in many respects relatively open to individual choice. According to Giddens (1991), recent socio-cultural changes after modernism affect our understanding about the way we live in diverse ways. On the one hand, the scope of life has expanded. There is much more freedom of choice, for example, in education, place of residence or marital partner. On the other hand, according to Giddens there are also new obligations and demands on the individual: we have to learn to conceive of ourselves as centres of activity, with regard to our abilities, orientations and social relationships. These decisions as to education, spouse or lifestyles have to be made, and relatively little can be decided automatically. Giddens argues that social life can lead to a condition of ontological insecurity, a sense in which people have little understanding of the processes affecting their daily lives and inner selves or their sense of who they are and to which social groups and categories they belong. Giddens suggests the term ‘life politics’ to describe contemporary problems with life planning and self-actualisation. Life politics can be understood as referring to the problems that arise in the realm of everyday life and that obligate people to act reflexively and more independently than before (Erola & Räsänen 1998: 155-156). In Giddens’ words, individuals must now “integrate information deriving from a diversity of mediated experiences with local involvements in such a way as to connect future projects with past experiences in a reasonably coherent fashion” (Giddens 1991: 215). A reflexively ordered narrative of understanding of the self provides the means of giving coherence to individuals who are living in a realm of continuously changing external circumstances.3 In other words, we have to adopt a certain kind of politics of life decisions because social

44

life is to a great extent emancipated from social and cultural norms and customs. Giddens (1991) himself does not use the terms ‘postmodern’ or ‘postmodernism’. Despite this, it is evident that his writings represent a highly typical form of postmodern argumentation. The key point in his text is that contemporary conditions require a new perspective concerning one’s own life. Many similar thoughts on the nature of contemporary living conditions have been presented during recent years (e.g. Lash 1995; Beck 1992; Beck-Gernsheim 1996; Beck et. al. 2003). In general, life is seen as a reflexive project, sustained through narratives of self-identity and personal life politics. To a great extent these claims touch upon notions of the nature of sociality in contemporary societies. Important social categories for the individual consist of multiple reference groups, some perhaps without even particular self-awareness, and of various short-term institutions of somewhat weaker stability in time and space. It is true that people are members of many groupings at different times in their lives, sometimes several during a single day. Discontinuous groups have become the central feature of people’s experience of everyday life (Shields 1996: ix). Thus social categories can provide only temporary identifications and sources of identity. In particular traditional solid identities based on structural categories may have given way to a variety of other sources. Lifestyles, values and beliefs have become merely selfconstructed categories, and the establishment of a career and a personal identity requires a great deal of effort. Generally speaking, it is thus believed that one’s life has become a planning project, a task to be carried out alone. This is to say that modern social structures are not necessarily very effective in explaining contemporary social phenomena and practices. In such a situation, it can be argued that the attitudes and behavioural patterns of individuals can most effectively be approached in a framework of free choice. In effect, boundaries between life conditions cease to be given and instead become choices (Beck et. al. 2003: 19). This tendency is reinforced in the domain of values and belief systems. There seems to be diminishing trust, for instance, in the sociality associated with labour contracts. These notions have also gained some empirical support. According to comparative analyses, for example, public confidence in

45

hierarchical institutions in general is eroding throughout the advanced societies (Inglehart 1997; Inglehart & Baker 2000). It is widely recognised in the recent postmodern discussion that success in society is less dependent on basic social institutions. Life chances are not necessarily allocated according to birth or social position in production. It can be suggested that the institutions of the labour market and the welfare state have now developed a network of rules and requirements imposed on individuals. While employment and the social arrangements related to it still constitute institutional standards in many Western societies, they are now based to a greater or lesser extent on various models that stress individual responsibility and initiative. In Finland, for example, unemployed job-seekers are expected to market themselves using personal Internet advertisements and portfolios. The educational system is also being developed according to the principles of individual freedom and capabilities. The national system of secondary education in Finland, in particular, is based as a whole on separate courses taken by individual students; it thus includes very little participation in collectively organised study modules. Naturally, all this constitutes only anecdotal evidence. At a more general level, however, life conditions after modernism can be conceptualised as post-structural in nature. Even in welfare societies, the safety-net of the state can no longer be seen as the institution with primary responsibility for ensuring individual survival (Beck et. al. 2003: 24-25; Giddens 1991). The state and civil society may provide certain safety-nets for individuals (although there are also many differences among countries), but in general people have to take responsibility for themselves. Postmodern views of contemporary living conditions imply that it is up to one’s personal aspirations, abilities and choices what kinds of activities become the basis of one’s lifestyle. Alternative lifestyles or ideologies – for instance vegetarianism, veganism or freeganism – or values supported by certain social movements may be absorbed as the key principles of one’s life. But one’s way of life and specific lifecycle to be followed is still a matter of individual choice. In the contemporary “era of individualised life politics” (Lemert 1997: 124) it might even be questioned whether any ‘typical’ lifecycle can be identified at all.

46

The pluralisation of life spheres can be seen as an important assumption which is part of postmodern reality. This is not merely a matter of new technology and its differentiating effects on social and cultural interest (see Lievrouw 2001: 15-16; Castells 1998). Various factors can affect the individual’s experience of daily life and what is considered to be important in it. Previously people judged themselves and others primarily according to positions in life determined by certain relatively strict social categories, such as class, age or gender. According to postmodern writers, this is no longer as evident a process as before. Social reality is more heterogeneous and people are understood to act outside the direct influence of modern social structures. It is easy to see that this has certain implications for social contexts and the experience of social life. In short, people constitute their lives in everyday practice; by their own activity, they deconstruct social structures and reconstruct new ones. This is to say that there has been a shift in the conceptualisation of what is important, what matters. Patterns of social interaction based on individuals’ own decisions can create new social life-worlds. The adoption of a certain ideology or belief, identification with a given reference group or following a certain lifestyle, for instance, may act as a basis for new kinds of social groupings. The practices that are seen and experienced as important can be almost anything – from the practising of a New-Age religion to hobbies and time spent in an Internet chatroom. Important social practices can also vary greatly from one individual to another, not necessarily following the requirements of any modern social structure at all. Nonetheless, it is argued that there are certain combining elements even in this individualistic society, which can operate as integrative forces in our social system. In particular the importance of consumer goods and services has been stressed in this respect. Consumption plays a significant and in many ways the most active role in the construction of people’s everyday lives (e.g. Bauman 1998: 2425; Gronow 1997; Featherstone 1991). It has been suggested that consumption patterns are an indicator of both individual self-identity and the actors’ positions in social hierarchies. The increased consumption potential of people living in welfare societies makes it possible to imitate the consumer lifestyles of different groups. Thus consumption is assumed

47

to break down earlier social communities and introduce novel criteria for new ones (see Baudrillard 1998; Sulkunen 1992). Consumption can also organise and maintain individual identities, i.e. how people perceive themselves and their place in society. These ideas are well represented in the writing on the consumer society. 2.2 Postmodern society as consumer society The transition from a traditional society to a developed industrial society can be seen as a transition from a shortage of material goods to affluence. This process subsequently induces a new kind of culture, which can be perceived as a distinct feature at the societal level (e.g. Antonides & van Raaij 1998: 64-65; Miles 1998; Strasser et. al. 1998). This new culture is characterised by particular elements of consumerism. Before turning to specific aspects of the consumer culture, however, it should be noted that the concept of consumerism is not completely synonymous with consumption. According to Steven Miles (1998: 4), consumerism should be seen as an attempt to come to terms with the complexities that lie behind the act of consumption.4 Thus the term can be understood as a cultural expression of the apparent ubiquity of acts of consumption in our everyday life. In other words, while consumption may be seen as an act, consumerism represents something that is a social condition or way of life. However, consumerism has become an important social factor only because we are indeed surrounded by consumer goods and services. As defined in the first pages of this work, consumption is understood as consisting of both material and immaterial consumption by private households or individuals. Thus the economic concept is seen as a necessary but not sufficient definition of consumption. Consumerism as a social condition is approached below through the literature on consumer society. It seems likely that all relevant aspects of consumerism can be assessed in terms of the concept of the consumer society because it is basically a rather expansive and vague term. Some of the societal demands of the consumer society have been discussed above. And of course the concept is so widely used in the present-day social sciences that a comprehensive overview would be impossible. Consumer society includes new social movements, new organisations and changes in

48

the ideological elements of the prevailing world view. In Finland, for example, the labour movement, advisory bodies originating in the activity of women’s organisations, and modern public consumer policy have been highly important in this respect (see Heinonen 1998: 30-32; 277-278; 297298). My purpose here is to focus on the level of everyday experience and on the cultural and structural conditions that enable the experience of the consumer society. Thus we can refer simply to a social condition which is dominated by the consumer culture in the form of consumer goods and services. The central elements of consumer society (and consumerism) can be summed up as the following four multidimensional features: 1) A rise in the level of consumption; 2) The market as the main exchange mechanism of goods and services; 3) Consumption as an appropriate daily activity; 4) The formation of consumer identities.

In the following, I discuss these features, their chief premises and their primary consequences for everyday life. For the sake of simplicity each feature is examined separately, even though it is evident that they form a single, integrated whole. While these four central elements of consumer society will be approached here from a contemporary angle, it should be stressed that in particular the first two elements can also be seen as distinctive features of modernism. In other words, while all four features can be argued to highlight consumer society, they should not be regarded as unproblematic features of postmodernism. To sum up: it is not my purpose to argue that consumer society should be equated to postmodern society. 2.2.1 Rise in the level of consumption The first feature of consumer society indicates that the level of consumption has risen. In its simplest sense, this refers to the increase in disposable income and to the fact that wealth can be used to increase expenditure. This tendency can easily be verified by referring to the official statistics. It is evident that the gross domestic product (GDP) of all industrialised countries has risen dramatically since the Second World War (e.g. Fine & Leopold 1993: 66-67; Laurila 1987). In most European countries this trend follows rather clearly the development of

49

industrialisation and – during the last few decades – the expansion of the service sector. In particular health care and the educational sector have expanded from the 1960s onward. In Finland, like in other developed economies, the GDP has risen steadily since the Second World War. As Figure 2 shows, the rising trend of the GPD has been relatively strong, excluding only the short-term effect of the economic depression of the early 1990s. Between 1945 and 1990, the volume of the Finnish GDP increased annually by a mean of 3.6 percent. This phase of rapid and sustained economic growth has been characterised as exceptional in Finland, in the sense that the ‘golden years’ did not end with the oil crisis of the 1970s (Heikkinen & Kuusterä 2001: 27). On the other hand, the depression of the early 1990s was deeper than any other crisis (excluding war) in Finland. However, recovery began soon and since then the GDP has again grown vigorously. In 2002, the Finnish GDP reached 139.7 billion Euro (Suomen Pankki 2003).

* 100 million € (Indexed, 1926 = 100) 10000

GDP

1000

Private Consumption

100

10

1900

1914

1928

1942

1956

1970

1984

1998

(Source: Tilastokeskus 2002)

Figure 2 Logarithmic time series of GDP and private consumption in Finland, 1900-1999

50

Figure 2 shows that the volume of private consumption follows growth in the GDP quite closely. In 2002, for example, private consumption reached a peak with 70.3 billion Euro. The peak of the national GPD volume was also witnessed that year (Suomen Pankki 2003). It is thus possible to argue that the rising GDP figures in Finland, as elsewhere in the Western world, sustained the rising standard of living which is manifested as increasing consumption expenditure. The industrial structure of the country has expanded since the Second World War, and many export industries have also begun to develop. Thus the modern massproduction system can be understood to be the primary force behind the rising GDP figures. The volume of private consumption in 1980 was almost twenty times that of 1880; this indeed indicates a dramatic rise in consumption during the last one hundred years (Laurila 1987: 10-11). This suggests that during this period the level of consumption has not merely risen in quantitative terms, but also now considerably exceeds the subsistence level. Of course, consumption in the Western societies has been above the subsistence level for a long time. In this respect, the late 20th century difference is one of scale. This means that since mass-produced goods have penetrated the national and international markets, people no longer have to struggle for life in terms of everyday necessities. It is feasible to say, for instance, that all Finns can afford to buy many times as many potatoes as they can eat. The welfare state is expected to ensure a basic livelihood for all members of society. It is true that food, heat and water can nowadays be taken for granted in basically every household, and scarcity is experienced only with regard to various luxury goods. During the last one hundred years the proportion of total household expenditure spent on food has dropped from 50 percent to less than 15 percent. This can be seen as a clear expression of the rising standard of living (Heinonen 2000: 15). To a great extent it is a consequence of the industrial revolution. Increased production capacity and improved technology have made it possible for goods to be produced more cheaply in greater quantities. This has resulted in lower prices, which is naturally followed by extended possibilities of consumption. A central development in the emergence of the consumer society was the increase in purchasing power among ordinary workers. Particularly

51

following the reform in working hours legislation in 1917, which made more leisure time available to workers, a clear trend of increasing recreational consumption could be observed in Finland (Toivonen 1986: 53-54). Elsewhere in Europe and North America similar reforms were passed during the late nineteenth and early twentieth century. Consequently, in the industrial mass-consumption system individual consumers were provided both with the necessary surplus of consumer goods and with sufficient time outside paid work. Money was increasingly deployed not only on housing and food, but on leisure and other activities. This is an important turning point with regard to the initial formation and consolidation of the consumer society. Later, in the 1950s and 60s, the accessibility of consumer goods began to transcend social hierarchies, and the status-conferring qualities of many daily durables become less important. With economic development, postwar workers, like almost all employed people, gained access to material wealth; this meant that they were able to satisfy more of the basic needs for their households, as well as considering purchasing new consumer objects such as televisions or cars (Miles 1998: 8-9). This can be understood as a relatively universal trend, occurring at about the same time in Europe and Japan, lagging a decade or so behind the United States. In Finland, the decontrolling of post-war rationing begun to have a clear effect in terms of household expenditure from the 50s onwards (Heinonen 2000: 14-15). In particular the retail network was broadened expeditiously. Following this, the 60s have been referred to in Finland as a decade of structural change, during which many people moved to cities, became prosperous and purchased cars, televisions and other modern domestic appliances (e.g. Toivonen 1996: 35-36; see also Ilmonen 1993). The diffusion of consumer durables made it possible for traditional attitudes towards consumption expenditure to change conspicuously. Basically, the earlier restrictive and disciplined attitudes towards spending started to give way to more hedonistic attitudes (Heinonen 2001: 89-91; Heinonen 1998). To understand some of the key features of consumer society at their simplest, then is to reduce them to a matter of growth in the consumption of marketed goods. Thus the consumer society can be perceived as a condition in which there is a tendency for everyone to want roughly the things that automated mass production is able to provide (Fine & Leopold 1993: 65;

52

see also Miles 1998). This has resulted in a relatively dramatic increase in the level of individual consumption throughout the industrial societies. However, the growth of consumption cannot be held as a sufficient explanation for the development of contemporary consumer society. 2.2.2 The market as the main exchange mechanism of goods and services The second central feature of consumer society relates to the market economy. Most goods and services are acquired in the first place by acquisition rather than by household production. This of course is definitely not merely a recent trend. Barter and trade are among the oldest activities in the world, and money established itself as an instrument of exchange many centuries ago. Nowadays, however, almost all goods and services are purchased from enterprises in the private sector. This can be seen not only in the case of foodstuffs, which nowadays increasingly take the form of convenience food, but also in the case of educational and health services (Fulton 1994: 226-227). At the same time the services of the public sector have also expanded enormously. This has led to a new kind of market situation – ‘quasi markets’ as they are called, in which private enterprises and public sectors interact and co-operate in order to provide services for citizens (see Harrinvirta 2000). In other words, both sectors act primarily as service providers. Overall, many activities in contemporary society can be seen as consumption: in other words, things are bought and sold in the market, and these activities are controlled by the principles of the market economy. According to some writers, people are increasingly treating even religion as a commodity, offering a range of options: the individual can choose those which best suit his or her consumer requirements (Heelas 1994: 102). It is possible to argue that an approach using acquisition and sale as a metaphor is becoming an increasingly plausible view of social life in general (Räsänen 2000: 229). This can be seen as most conspicuously highlighted in current leisure-time activities. Let us consider for instance sport. This is one of the most visible areas of social life, which has profoundly altered its character in recent years. Not so many decades ago sport was merely a leisure pursuit or hobby, and was not particularly closely linked with buying and selling. Here I am not referring primarily to gym membership fees or the price of sports equipment. Rather, I am arguing that earlier there

53

were very few professional athletes; in the contemporary consumer culture this has altered radically. Professionalisation has had a tremendous impact on sport in the form of commercialisation. During the past three decades or so, a long-term trend can be seen whereby sport as a participatory experience has been replaced by sport as a spectator event (see Rowe 1995; Miles 1998). This trend has been supported by the media in general and by television in particular. Many sports have been forced to professionalise because of the fear of being excluded from the sports marketplace. These claims are mostly meant to be provocative, but at the same time the underlying criticism is understandable. Sports events, such as the Olympic Games, can actually be seen as marketing opportunities for companies (Miles 1998: 127-128; 142). But what effect might commercialisation have on the daily lives of ordinary citizens? The first effect is, of course, that sport is nowadays entertainment, and is mainly consumed via television broadcasting. But secondly, those citizens who take an interest in sports probably evaluate it more and more in terms of ordinary consumption. In other words, commercialisation and professionalisation has legitimised a consumer ideology in relation to sport. This can be seen in the increased level of advertising and the statuses conferred on certain brands and trademarks. At the psychological level, going in for a certain sport often entails the purchase and use of appropriate items of clothing and other equipment. The personal commitment invested in hobbies is probably best understood sociologically as a consumption investment. In other words, individuals can decide themselves what kind of hobby they are interested in, but a central part of many activities involves the acquisition of equipment and the paying of membership fees. Parallel changes have occurred in many other areas. In this sense, consumption has become a powerful metaphor and a frame for the interpretation of daily activities. Fashion, education and almost every area of social life can equally be understood as consumption. These areas of social life consist of goods and services that can be bought and sold. Moreover, acquisition has gradually led to the effective target-group marketing of all identifiable goods and services. Likewise advertising and marketing has come to characterise practically all areas of life (Heinonen

54

2000: 19-20). Nowadays there are insurance offers for people under and over certain age limits, and universities and other educational institutions sell fully customised study modules for certain target groups. Workshopbased MBA-programs for business executives are good anecdotal examples of this. The influential role of acquisition, however, can most clearly be observed in the case of material goods, such as food and clothing. Clothing in different styles and sizes has been available for quite a long time, and there are more convenience food alternatives available with each passing day. This is partly the result of increased production capacity, which makes it possible to develop products and services that can be better adjusted to the needs of the consumer. In most cases it is easier and cheaper to buy a consumer item ready-made than to fabricate or produce it oneself. The supply of goods has become more versatile, while the distribution and organisation of labour has become more effective. In a way, then, one of the main prerequisites of the consumer society can be seen as a clear consequence of the industrial revolution. Nowadays, of course, new applications of information technology provide more sophisticated possibilities for manufacturing and retailing goods and for matching supply to demand. Just as much as it is a result of industrialism and capitalism, however, the consumer society is also a result of urbanism. In particular a relatively recent phenomenon, which can be described as post-industrial urban development, has accelerated the formation of the consumer society. Since the turn of the nineteenth and twentieth century, urbanisation in the Western world has been rooted in a manufacturing economy. During the early part of the period manufacturing developed within industrial districts and required a densely concentrated labour force. At the same time, retail and entertainment activities also began to be established and to flourish. The ultimate outcome of this development was the rise of cities, which was witnessed in the twentieth century throughout the developed world (Fainstein & Campbell 1996: 4-5). Later, processes of suburbanisation and deurbanisation emerged as the main trends in the largest metropolitan areas of many countries. This means that people started to move away from the dense streets to physically separate areas outside large cities. Changes in transportation technology made it possible for workers to live further away from their place of employment, and industrial and retail companies were able to spread out to cheaper and more

55

diffuse sites. While it is not necessarily sufficient to discuss these processes in a Finnish context, the more recent trends of post-industrial urban development can be seen as relevant here. As is well known, technical innovations have made it possible to simultaneously increase productivity and reduce the work-force. In the case of industry, this means that mature manufacturing has been displaced by other economic activities. This change was made possible in particular by the development of logistics and distribution channels. The departure of industry from city centres has generally been observable in the developed countries. Industry has moved out to suburban areas and to the periphery, and above all abroad. In some countries, such as the United States or Great Britain, this trend was obvious already in the mid-twentieth century. In certain other countries, the changes have started to happen far more recently. Nevertheless, the consequences of the processes of postindustrialism for city life are very much parallel. Nowadays the service sector is the only branch of the economy that continues to grow in urban areas. Old industrial buildings and blocks of flats are continually giving way to shops, supermarkets and shopping centres. Many theorists have also discussed the impact on city life of the ‘culture industries’, such as publishing or tourism (e.g. Jaeger 1986: 86-87; Soja 2000). In general, cities offer vast possibilities for private enterprises that can sell either goods or services. A number of institutional innovations, such as self-service supermarkets and department stores, have come to reflect city life. In Finland, it has been argued that for instance the huge shopping malls have significantly affected the public life of cities as a whole, partly because Finnish cities are relatively small (Lehtonen & Mäenpää 1997: 139-141). From the individual’s point of view, some of the basic characteristics of contemporary cities consist of advertising, shopping and entertainment; these are incorporated into every aspect of social life. In many respects, the urban landscape is an ideal setting for expenditure in terms of shopping, eating out and entertainment. From this perspective, cities are now frequently regarded as consumption centres (e.g. Hannigan 1998; Falk 1997). It is true, of course, that consumption in the city is not as recent a phenomenon as some postmodern theorists would have us believe; nevertheless, consumption has become a major urban preoccupation

56

(Hannigan 1998: 65). Nowhere else is the acquisition of goods and services as essential an activity as it is in contemporary cities. According to Anthony Giddens (1981), the city should not be regarded as incidental to social theory. Giddens suggests that on the contrary it belongs at its very core, because it stands at the heart of some of the most fundamental problems of general sociological interest. What urban life and the city are taken to mean is very much identical with that which we understand by contemporary late-modern social life. Typical features which are stressed in this context include for instance complexity and ambivalence. Giddens has suggested that social life in fact consists of regularised social practices.5 This means that everyday life is experienced as a continuity of day-to-day existence, in a context of the routinisation of practices and in given environment (see Giddens 1981: 150-151). While locale is a physical setting, it is simultaneously also a socially defined one. According to Giddens, human agents have a clear understanding of their social and spatial environment and they actively use this environment in the creation of their own lives and the social systems to which they belong. This is a relevant point also with regard to the role of consumption in present-day society. Firstly, if consumption is regarded as one of the major activities practiced in urban settings, it should be relatively important to almost everyone. This can be argued because most of the population in the advanced societies live in urban communities. In 1900 only one tenth of world’s population were city-dwellers; by 2006, in contrast, half of the world’s population will be living in urban areas – a twenty-fold increase in just 100 years (O´Meara 2001: 338-338). North America, Europe and Japan, for example, are already very highly urbanised and this trend seems by no means to be decreasing. In Finland, the vast majority of the population is concentrated in the southern part of the country, and many Finnish urban centres are located along the southern coastline. Over 20 percent of the population live in the Helsinki metropolitan area (Helsinki, Espoo, and Vantaa). At a more general level, more than a third of the population live in the ten largest cities of Finland (Tilastokeskus 2002). It is evident that a major impetus for growth and development in contemporary cities comes from the presence of business centres with diversified activities. More generally, a major role in this development has been played

57

by the growth of service sector jobs. Particularly in recent years, the growth of urban ‘high-tech centres’ has effectively contributed to urbanisation. Cities and their networks form a strong communication infrastructure, which is connected with services and educational institutions. This is also a highly visible feature of urban development in Finland, where the biggest cities tend to be university cities.6 Secondly, one can plausibly argue that in advanced societies portrayals of traditional urban and rural environment have lost their relevance. It is true that human-created environments are no longer distinct features of urban areas only, but are to a great extent universal. While the city can undoubtedly be described as the concentration of a large mass of individuals within a relatively restricted spatial area, the cityscape as such is not very easily defined according in traditional terms of the absence or opposite of untouched landscape and wildlife. There are telephone lines, paved roads, service stations and shops almost everywhere in the countryside too. These things remind us of our membership in the consumer society and of our possibilities to engage in its preoccupations. In this sense, it may be noted that to live in a created environment is also to live near streets and other public places filled with possibilities of continuous consumption. In other words, consumption has become one of the basic routines of the everyday life of citizens. These general notions regarding the market economy, urbanism and the created environment lead us to the third and perhaps the most important element of the consumer society: consumption, in terms of buying and spending on goods and services, has become an appropriate citizen activity. 2.2.3 Consumption as appropriate daily activity It has been suggested that contemporary society binds its members primarily as consumers. According to Zygmunt Bauman (1998), ours is a consumer society in a fundamental sense, just as modern society in its industrial phase was a producer society. In other words, in the present latemodern or postmodern stage the first and foremost role for the citizens is the role of consumer. In Bauman’s words: Ours is a ‘consumer society’ in a similarly profound and fundamental sense in which of our predecessors’ (modern society in its industrial phase .…) used to deserve the name of a ‘producer society’ …. The reason for calling that older type 58

of modern society a ‘producer society’ was that society shaped up its members primarily as producers; the way in which that society shaped up its members was dictated by the need to play their role and the norm that society held up its members was the ability and the willingness to play it. In its present late-modern, second-modern or post-modern stage, society engages its members – again primarily – in their capacity as consumers. (Bauman 1998: 24; emphasis in original.)

Basically, what Bauman is saying is that even though society earlier too has shaped its members as consumers, the all-inclusive consumer society has emerged only recently. At the level of everyday life, the passage from the producer to the consumer society entails many profound changes. The most obvious changes are the changing role of consumption and the new importance given to consumption activities (see also Baudrillard 1998). On the one hand, people can buy goods and services on a routine basis during their daily life, without even considering alternatives. When you feel hungry, you may want go to a restaurant; if you feel bored, you buy a magazine or go to see a movie or watch television. Nevertheless, acquisition is the principal strategy that crosses our mind when we require goods or services. To live a meaningful life, then, we all require our daily visits to the marketplace. On the other hand, however, consumption activities have become one of the main categories of time-use; time spent in various consumption activities has increased considerably in all advanced societies during the last thirty years or so. On a daily or weekly basis, the activities of individuals can be roughly divided into four main categories. These are 1) paid work, consisting of hours devoted to work and work-related activities; 2) unpaid work, i.e. time spent on household maintenance; 3) personal care; and 4) leisure. Personal care consists of time devoted to sleep, meals and personal hygiene. Leisure can be defined as residual time: that time which is not spent on earning money or on housework or personal care. It thus consists of all kinds of activities that we engage in outside working hours. The changing division between these four categories has been subject to lively theoretical debate by sociologists in the last few years. One of the most convincing interpretations comes from arguments relating to increases in the amount and distribution of leisure. It is frequently suggested in the literature that the fruits of affluence now take

59

the form of time spent in distinctive leisure-time activities (e.g. Featherstone 1991; Maffesoli 1996; see also Bourdieu 1984). In other words, the importance of paid work has decreased and that of leisure-time consumption has increased. It is also believed that leisure activities have become more important in expressing the individual’s place in the social hierarchies and in the construction of social identities. I return to this theme for a closer look in the following section of this chapter. In general, claims as to the increase in leisure enjoy some empirical support from simple measures of hours worked per week. At least since the 1960s and 1970s the most developed economies have in fact been reducing their paid work-time (Sullivan & Gershuny 2001: 333-334; Gershuny 2000: 28-29). Empirically, it is of course the case that the average hours worked by men have declined, while hours worked by women have increased slightly. Following the general notion of an overall increase in leisure-time, however, we can conclude that the centrality of consumption processes has also grown. Particularly according to American studies it is evident that in the 1970s people began to spend more time shopping and travelling (Robinson 1975: 847; Robinson et. al. 1998). The fact is that if more time is available outside working hours, then that time will be filled by other activities. It follows logically that unpaid work and leisure will take up more time. There is no reason to believe that certain consumption activities will not be among those categories which will occupy an increasing amount of time. According to comparative analyses (e.g. Niemi 1995; Sullivan & Gershuny 2001), it is evident that shopping, travel and other such activities take up increasing amounts of time for both women and men. This is very much an international trend, observable in all the most advanced societies. Moreover, the rate at which these changes take place has increased consistently for both sexes, as well as for different kinds of households (Sullivan & Gershuny 2001: 341). Thus it is possible to argue that consumption-oriented activities have in general become to take up more time. An overall increase in leisure time can be seen as a result of economic growth and technical efficiency. However, it is also important to note that during this period there has also been an increase in the use of cars, public transportation and telecommunications. Thus the increase in household time devoted to consumption cannot be a matter of technical efficiency

60

alone. Many leisure activities, as well as becoming more various, seem at the same time to be becoming more intensive, involving greater amounts of effort and expenditure. It has become mandatory to accept the increasing variety of consumer goods and to spend more and more time shopping and consuming. Thus theoretical discussions on the role consumption as a way of life (e.g. Bocock 1993; Miles 1998) appear fairly plausible. The importance of consumption has grown not only at in everyday life, but also at the societal level. To understand the relationship between production and consumption in modern societies, we have learned to say that production is justified by consumption. It has generally been believed that the economy produces goods to satisfy the needs of consumers. In reality, however, this is evidently no longer the case. The situation is in fact exactly the opposite: it is the imperative to produce that is dictating the need to consume (Csikszentmihalyi 2000: 271). In Finland too, economic forecasts are calculated taking into account the increasing demands for individuals to consume as much as possible. Shopping has come to be viewed as a patriotic act, and refraining from consumption as antisocial behaviour. People are expected to believe that unless they spend more money on cars, clothing and other consumer goods and services, the economy will suffer. According to this line of argument, non-participation in consumer activities is seen as a threat to the community’s economic functioning. Drawing a parallel between consumption and way of life, certain fundamental changes in the primary role of individuals in society can be perceived. In a way, consumption has come to provide a key arena for individual competitiveness in contemporary society. The role of the consumer can be regarded as the primary role of the citizen as well. It follows that the competent citizen is expected to be able to play that role. It is possible to say that when the abundance of supply is massive, consumption actually involves a great deal of effort. Consumer choices have become not only recreational or leisure activities, but also an obligation that must be satisfied. Acts of consumption are now seen as important building materials of everyday life. This has given rise to processes of self-construction and identity-formation through consumption.

61

2.2.4 Formation of consumer identities The fourth key feature of consumer society relates to its impact on individual social identities. The concept of ‘consumer identity’ refers to the fact that people have become inclined to evaluate both others and themselves in terms of their consumption expenditure level and consumption styles. This means roughly seeing oneself and the others as consumer citizens. While the increasing centrality of consumption processes is seen as a feature of recent social change, the idea was originally broached a long time ago; Thorstein Veblen introduced the idea in initial form in his best-known book, The Theory of the Leisure Class, first published in 1899 (Veblen 1994). Veblen developed a theory of the existence of a leisure class among the most affluent American social groups;7 according to this theory, upper class individuals tended to maintain a distinctive lifestyle by means of consumption. Veblen referred to this specialised consumption of goods as ‘conspicuous consumption’; the term reflected the waste of valuable goods in order to gain social prestige. This was seen as an obligation or even an occupation of the non-labouring leisure class (Veblen 1994: 43-44). In general, Veblen argued that the highest social groups continually updated their consumption habits in their efforts to stay one step ahead of the emerging groups of noveaux riches. The point was that those who were not under the necessity of earning a livelihood were able to express their social position and affluence in the consumption of food, clothing, dwellings and furniture. Veblen can thus be seen as drawing attention to the symbolic value of consumption, viewing consumer goods and activities as possible instruments to gain social prestige and status. Veblen’s analysis is one of the most significant contributions to a sociological understanding of consumption. He introduced the idea that markers of consumption might express a person’s place in the social hierarchies. It should be noted that Veblen focused on the role of consumption in the ways that members of the upper class associate with each other. He was also centrally concerned with the distribution of leisure and status among classes. In early twentieth century Europe, Georg Simmel wrote in some detail about the same processes: the achievement or maintenance of a certain social position through the consumption of fashion (Simmel 1957). Simmel argued that fashion can provide the only

62

apparent means of recovering and stabilising the diffusion of tastes and styles throughout society. According to him, the dynamics of fashion were largely determined by mobility towards higher social positions. People in lower social positions followed fashion for the sake of acceptability and higher status. Those at the very top, however, were just as active in protecting their privileged status as those in the strata below. The upper social strata thus were constantly inventing new fashions which differentiated them from the masses (Simmel 1957: 543-544). This idea, attributed in its original form to Simmel, has since then become known as the ‘trickle-down theory’, referring to the upward pressure to consume fashion that allows tastes to trickle down (e.g. Fine & Leopold 1993: 138; Davis 1992). At the beginning of the twentieth century only the highest social strata were able to engage themselves with distinctive consumption patterns. In contemporary society, however, this has probably become a more common condition. The symbolic value of consumption is at least in some cases one of the chief reasons why certain products are chosen for consumption in the first place. Let us consider the symbolic value of such brand-names as Levi Strauss, Lee or Diesel: at the moment, these labels represent some of the best-known jeans in the world. There are of course many other kinds of jeans available, made of exactly the same fabric and cut to the same pattern. Because of their brand-name, however, these products can be sold at a higher price than other similar garments. It is probably not necessary to go into detail as to the means by which symbolic value is attached to many goods.8 In general, everyone understands that it is a very different thing to drive a Mercedes than a Fiat. The former is generally perceived as a luxury car while the latter is not. The same thing can be said about certain clothing labels, alcoholic beverages, soft drinks, foodstuffs etc. Many related phenomena have emerged, ranging from the fan clubs of a certain make of car to simple brand loyalty in food consumption. Thus there is clearly a communicative role related to consumption. We do not even have to scrutinise brands and trademarks in support of this claim. Information concerning the ways various kinds of restaurant services or tourism package tours are usually consumed may reveal something very valuable about individuals’ general preferences and attitudes. In many situations, information as to

63

consumption preferences may thus tell us more about people’s lifestyles and identities than they seem to reveal at first look. Pierre Bourdieu (1984) has systematically addressed questions of social comparison through consumption. What is important in Bourdieu’s work is that it is based on the findings of systematic empirical research in France during the 1960s and 1970s. Broadly, his work has much common with that of Veblen; Bourdieu sees consumption behaviour as expressing variation between different social hierarchies. In addition, contrary to Veblen – whose ideas on conspicuous consumption were based on the rational selection of certain goods – Bourdieu argues that distinctive conduct has nothing to do with rationality; rather, it is the product of a person’s life history and passive learning (e.g. Campbell 1995: 104; Wilska 1999: 47; Gronow 1997). Bourdieu explains this by the concept of habitus, which he defines as a system of durable dispositions of everyday knowledge or cultural capital that reflects the routine experience of appropriate behaviour in particular cultures and contexts (Bourdieu 1984: 170-172). In a way, habitus can be seen as a group-distinctive frame for the interpretation of tastes, attitudes and social conventions, reproduced across generations. According to Bourdieu, the dominant classes demonstrate their superiority through access to high culture and high consumption; that is, they differentiate themselves from other groups by distinctive consumption patterns (Bourdieu 1984: 169-175). Social class of course remains the fundamental determinant of consumer lifestyle in Bourdieu’s analysis. However, what is important is that consumption is both the instrument of distinction and the field where one can stand out. Consumption necessarily involves signs and values, which are or can be used as means of distinguishing one group from another. Thus it is important that consumption does not merely reflect people’s identities; it can also create them. Products and services which serve to highlight certain symbolic values can be used as building blocks for the construction of social identities. In other words, consumption allows us to experiment with new identities. Consumption patterns might be used to distance oneself from one social group and to reinforce membership in another. But this does not necessarily tell us anything about individuals’ personal and social

64

identities.9 Broadly speaking, ‘personal identity’ can be seen as an individually unique self-portrayal that one is able to exhibit to others, and which the individual recognises as the image of him or herself. Personal identity is considered to develop during childhood and early adulthood (see e.g. Erikson 1980; Lunt & Livingstone 1992). It is basically experienced as a relatively stable conceptualisation of one’s primary being or personal character. ‘Social identity’ can refer either to the same thing with only a slight social emphasis, or it may refer to something that is more robust and resilient to change over time (Jenkins 1996: 20-21). The latter usage may refer vaguely to social roles which the individual displays in everyday life, but which are not experienced as part of one’s character. The important aspect of social identity is that it includes the notions of both separation from and identification with other people. Social identity may therefore emphasise similarity to others, i.e. what individuals are believed to have in common. But the bare outlines of both personal and social identity can be drawn by understanding identity as a synthesis of one’s self-definition and the definition of one’s social role and positions in society. Personal identity is a necessary prerequisite for social life. At the same time, however, that identity is not meaningful in isolation from the social world. When identities are discussed at the collective level, people must have something socially significant in common before it is possible to talk about their social identity. In the social sciences identity is usually treated merely as a feature of the long-term groups to which the individual belongs. Some of these features (such as sex, race, age or class position) are of course seen as more primary than others (such as friendship or recreational roles). Traditionally, sociological analysis has emphasised the importance of certain structural elements in identity dynamics and lifestyles. Classconsciousness in terms of working-class and middle-class identities has been the starting point of many analyses of everyday life. Lifestyle activities and daily routines, for example, have been analysed against a framework of class cultures, class-based traditions of solidarity and dignity related to one’s position in the production process (e.g. Sennett & Cobb 1972; Willis 1977; Lefebvre 1971). In the field of consumption too, it has been argued that class-based ideologies characterise both actual consumption activities and the images people have concerning their

65

consumer behaviour. The debate on postmodernism, however, has questioned this argument. It is particularly the importance of social classes that has been claimed to be weakening. In earlier stages of modernity, work – more precisely paid work – formed the basis for many people’s sense of identity. In current post-industrial society, however, one’s position in the production process is not the single most important factor of social life, as it probably was in industrial society. Still, it is reasonable to assume that people define themselves through the social practices they perform in their daily lives (Giddens 1991: 214-215). In this sense, consumer goods and household patterns of consumption can play an important part in the social construction of identities of men and women. According to many theoretical perspectives, the selection and purchase of goods and services should be envisaged as the central activity in everyday life. Thus one’s consumption preferences can also be regarded as a form self-expression and self-understanding. Some theorists, Robert Bocock (1993) for example, go so far as to argue that many people’s sense of identity is now bound up with their patterns of consumption rather than their occupational or gender roles. Questions as to what people think they are, or how they live, are at least as likely to be answered in terms of the kind of consumer lifestyle they aim at, as the kind of occupation they practice or seek (Bocock 1993: 109). Because of the wide range and character of goods and services on offer in the marketplace, individuals are regarded as free to select and maintain consumer identities. There is considered to be considerable freedom of choice in changing one’s consumption habits in order to adopt a new identity and to show it to others (Campbell 1995: 112). Some writers even speak of postmodern lifestyles and status group divisions that are experienced as stable categories only because of uniform set of consumption patterns (e.g. Maffesoli 1996; Featherstone 1991). A variety of typologies have been presented in the literature to illustrate the meanings that consumption may have for the individual. A typical example comes from those collected by Arto Noro (1995). Noro’s purpose is to reconstruct some of the most interesting discussions that have flourished in the field of sociology of consumption by extracting a total of five theoretical figures of the consumer. While Noro’s figures look at the

66

consumer from slightly different angles, all of them in fact contain the same idea of a consumer as a self-constructing character (Noro 1995: 911.) The term ‘lifestyle enclave’ is also used in a somewhat similar sense, expressing the significance of shared lifestyles based on individual choice (see Bellah et. al. 1985; Gabriel & Lang 1995). In these views, social factors are basically reduced to the pluralities of individuals who see themselves as similar to each other, or who share common behaviour in terms of consumption patterns. Thus views concerning the determining role of social structures can be seen as very different from traditional sociological views. I return to a closer look at the substance of lifestylebased notions in Chapter 3. In sum, then, it can be assumed that individuals’ lifestyles and social and personal identities have become more clearly defined in terms of consumption. There has been a dramatic increase during the last century in the consumption of market goods. In this sense, consumption can no longer be regarded as a secondary diversion. Likewise the results of more recent time-use studies suggest that consumption has shifted towards a central position in daily life. Perhaps, then, we can conclude that the cycle of use and re-use of goods and services, and the meanings related to these activities, provide a greater understanding of the dynamics of social life. The question, however, remains; is it feasible to approach a society of consumers as a postmodern society? 2.3 Summary: Postmodernism and consumerism The four features examined above – the rise in the level of consumption, the market as the main exchange mechanism of goods and services, consumption as an appropriate daily activity, and the formation of consumer identities – can be regarded as the key elements of the consumer culture. Some of them are relatively old, some of them quite recent. The revolution of the consumer culture is to a great extent a product of sophisticated exchange mechanisms and of industrialised modes of mass production. In addition, however, there have been several more recent changes contributing to the development of consumer society. Among the most important changes during the last few decades have been an increase in consumers’ capacity for consumption and a general democratisation process in advanced societies. Together, these factors represent the series

67

of trends that have brought us to a condition in which consumption is experienced as the core of social reality. Continuous consumption is nowadays basically synonymous with normal life. According to certain social theorists, if postmodernism means anything it surely means the establishment of the consumer society (e.g. Bauman 1998: 23-24; Hannigan 1998). It is believed that contemporary consumption-based lifestyles are at least partly replacing the effects of modern social structures. Following this, individual choices have increased and people are able to control more freely the course of their life. If the primary role of the citizen can really be understood as that of a consumer, it is surely true that what we see around us consists of potential objects of consumption. Thus it is not very peculiar that so many writers have recently approached different phenomena either as consumption or in a general framework of consumption. As noted above, in particular post-industrial cities have been regarded as shopping centres and entertainment hubs, and changing one’s personal consumption patterns has been seen to offer possibilities of experimenting with new social identities. Of course, incorporating consumption into urban life or social relations is not in itself a recent phenomenon. But what is understood to be relatively new is the fact that consumption has begun to affect other areas of social life. Expanded consumption possibilities are considered to have an effect for example on changes in the allocation of time between work and leisure. Consumption is also influencing general societal development. It has been suggested that economic prospects in society have become increasingly dependent on consumer demand and confidence. Some writers even argue that sociality is nowadays possible only through collective consumption activities. Consumption may cover some of the central characteristics of the postmodern condition. At least with respect to everyday life, the significance of consumption has been acknowledged to include something more fundamental than consumption behaviour alone. In this respect, it is plausible to argue that consumption choices provide the central arena for individual competitiveness in advanced societies. The possibility of selecting certain objects instead of others can be regarded not only as a reflection of personal freedom but also as a civic right. There is, however, also an opposite point to be made with regard to this question. It is

68

reasonable to consider that consumption not only constructs but also constrains social life. The increased distribution of consumption possibilities for one individual simultaneously means an increased distribution of impossibilities for another. If the road to understanding one’s place in society is through one’s consumer behaviour, then existing social inequalities should also be visible through differences in consumption patterns. This means that some of the key arguments that have been presented regarding the new social condition of postmodernism can be addressed by means of analysis of consumer behaviour. In the next chapter, the basic economic and socio-demographic factors that affect individual consumption patterns will be discussed. I discuss whether it is still reasonable to posit effects of class and other social factors in the empirical analysis of consumption. In particular the social mechanisms according to which the effects of these factors might be interpreted will be scrutinised. The criteria of consumer classification are examined in a framework of people’s life situations and their membership in different social categories. In addition, I note certain theoretical and methodological problems that arise in empirical consumer research based on a postmodern analysis of consumption and everyday life.

Notes 1

The overlapping concepts of postmodern and late-modern, for example, represent difficulties

in drawing distinctions between different types of societies and their social arrangements. In this study, however, the distinct aspects of the postmodern condition are constructed with a relatively broad conception of the postmodern. It follows that our criteria for postmodern include several assumptions that might equally well be attributed to various alternative terms, such as ‘second’, ‘reflexive’ or ‘late-modern’ (see Beck et. al. 2003; Giddens 1991; Harvey 1990). 2

Some empirical support can be found for this argument. It has been estimated that average

Americans, now in their late twenties and with at least two years of college, can expect to change jobs at least eleven times during the course of their working life (Sennett 1999: 22). It is reasonable to believe that the situation is relatively similar elsewhere in the Western world. In Finland, for instance, a notable proportion of new jobs from the 1990s onwards have consisted

69

of temporary positions (Kauhanen 2000). In addition, part-time jobs have become increasingly frequent in the last few years (Tilastokeskus 2003). More generally, the relative proportion of the labour force, i.e. the population consisting of persons of working age either in full-time or part-time work or unemployed, has diminished during the last decades. In 1970, over 72 percent of the Finnish population belonged to labour force, in 2000 only 66 percent. Especially among men the proportion of labour force has decreased strikingly from the early 1970s onwards. During this period women have established their position in the labour market. (Tilastokeskus 2002.) Similar changes have been observed during the last few decades throughout Western Europe (OECD 2001). 3

‘Reflexivity’ can be understood here in the simplest sense of defining the nature of human

action, which in essence involves continuous monitoring of behaviour and contexts. Thus social life consists of the fact that social practices are constantly examined and revised in the light of information about those practices (see Giddens 1991: 38-39). 4

Recent sociological writings have offered a broader definition for consumerism than has

generally been offered in the literature (e.g. Bocock 1993; Miles 1998). Earlier definitions of consumerism have tended to describe it in the context of social movements aimed at protecting consumer interests. In the contemporary sociology of consumption, however, it is used more widely to address the social and psychological impact of the consumption experience. 5

This refers to the duality of structure in Giddens’ structuration theory, which is presented in its

propositional and slightly matured form throughout Giddens’s production from the late 1970s onwards. At a very abstract level, the idea of duality of structure aims at connecting the production of social interaction to the reproduction of social systems across time-space. Basically, structuration theory argues that social systems, including entities such as social groups, are produced and reproduced in social actors’ interactive use of social structures that consist of rules and resources (Giddens 1984: 90-91; see also Ilmonen 1994). 6

This has been the case, however, only in a few cities, leading to an unbalanced spatial

population distribution. During the last decade population growth has been particularly rapid in the Helsinki area and in Salo, Tampere and Oulu, while many regions in the northern part of the country experienced an opposite development. See Kultalahti (2001) for a detailed study of regional development in Finland during the 1990s.

70

7

In this book, Veblen presented any novel ideas on American culture and on the origins and

development of ownership and property. The present discussion, however, is restricted to Veblen’s ideas on the origins of consumer identities. 8

Jean Baudrillard (1983; 1998) and Fredric Jameson (1991) are generally recognised as among

the important sociological theorists in this respect. The symbolic value of brands and trademarks has also been under intense discussion in recent ordinary discourse. Many sociological and psychological ideas, for example, have lately been popularised by Naomi Klein in her controversial bestseller No Logo (Klein 2001). 9

In spite of differences among various theories of the formation of personal and social

identities, certain principles are more or less common to all of them. First, identity can be understood as an answer to the question: who am I? Secondly, identity presumes some sort of consistency or continuity over time. Basically, in order to answer to the question of who one is, the individual must see him- or herself as a social being, with certain roles and certain positions in the community. Thus identity consists of experiences and beliefs concerning both oneself and one’s positions in relationships with other individuals. Thus identity refers to something that is experienced as integrated and coherent. The Latin root of the English word identity is idem, ‘the same’ (Figueora-Sarriera 1999: 130).

71

3 UNDERSTANDING CONSUMPTION FROM A STRUCTURAL PERSPECTIVE Consumption is a multidimensional behavioural pattern that cannot be reduced to the act of purchasing a certain product or service. It is defined very differently depending on the researcher’s theoretical framework. Some researchers, for example, prefer to concentrate on economic aspects of consumption, while others may emphasise the emotional and illusory dimensions related to it. In any case, however, it is agreed that people’s consumption preferences and styles in contemporary market-economy societies vary greatly. Accordingly, an exploration of the various psychological, cultural and social factors that affect or may affect consumption patterns is considered to be the best way to understand consumption. This chapter explores the factors influencing consumption patterns of individuals and households. It is assumed that there are many things that play a significant role in people’s consumption behaviour. Sometimes the most likely predictors of one’s consumption choices may vary from the individual’s emotional mood or personal characteristics to the outdoor temperature or the time of day. It is true of course that personal traits and seasonal changes are useful in explaining and predicting some aspects of consumption; human activity can never be explained by one single factor. But psychological concepts and explanations based on the time of day lie outside the scope of conventional sociological approaches. In the sociology of consumption, consumer behaviour has been understood as determined to a great extent by surrounding economic, social and cultural factors. In other words, the consumption patterns of individuals are heavily influenced by various collective expressions. Traditionally, other types of explanations are considered to be adequate only when no economic, social or cultural explanation can be found, or when explanations based on these factors seem to be theoretically insufficient.1 As an empirical phenomenon, however, consumption is a complex behavioural pattern that cannot be explained sufficiently by social and cultural factors. This means that other explanatory mechanisms may be acceptable if the researcher is able to specify the structural conditions in 73

which the assumed explanations are adequate. In other words, to avoid the contingency of interpretations, possible explanations of social phenomena should be connected to the surrounding social context. To provide a basis for a sociological understanding of the determination of consumption behaviour, different socio-cultural and economic factors will be discussed below. The discussion starts from the notions of economic resources and class position, and goes on to gender, age, type of household and place of residence. Finally, lifestyle and certain more recent ways of identifying consumer groups will be examined. These structural factors will be explored with regard to how they have been measured and assumptions as to how they can affect the formation of consumer practices. This means that theories as to the role of these factors in consumption may not be explored in detail. It should also be noted that all the factors are approached from a structural perspective, even when it is clear that the factors in question do not represent any structural element in a conventional sense. The main purpose of the discussion is to identify ‘social mechanisms’, as described in the sociological literature (e.g. Hedström & Swedberg 1998: 2-4; Hedström & Swedberg 1996; Esser 1996; Räsänen 2003). I contrast this approach with solutions which attempt to interpret social behaviour by reference to correlations between variables, and suggest that the latter are methodologically flawed. The concept of the social mechanism refers to the problem of identifying explanatory social processes that can be used in the interpretation of empirical results. In mechanism-based explanations, it is taken for granted that an appropriate explanation will explain the particular by the general, and that the explanatory mechanisms will operate according to the same logical principles from one occasion to another. This means that forms of descriptive approaches that aim at locating the chains of events that lead from one observation to another can be seen as analytically inadequate in an explanatory sense. At the same time, however, mechanism-based explanations typically establish explanations specially tailored to a limited range of phenomena. In other words, it is considered that social theory-formation should not try to establish universal social laws. Robert K. Merton is often acknowledged to be one of the first sociologists stressing this point. Basically, Merton suggested that the primary task of sociological analysis is to locate a ‘middle ground’ between

74

social laws and simple descriptions (Merton 1963: 46-50; 370-371). More recently, mechanisms have been referred in various formulations of explanatory models referring directly to the mechanisms which generate observations. In short, the concept of social mechanisms refers to a detailed set of explanatory descriptions, according to which structural factors produce effects at the individual level. It can be argued that the interpretation of these explanatory processes is of crucial importance for empirical analysis, since empirical results can provide only descriptive evidence as to the phenomena measured (Esser 1996: 160). In other words, empirical research can operate – even in its most advanced forms – only with measures of particular observations. It is essential in the social sciences to understand that it is people, not variables, that do the acting (Hedström & Swedberg 1996: 301). In this sense, mechanism-based approaches are concerned with providing theoretically plausible interpretations as to why a given factor systematically affects another. Accordingly, the purpose of this chapter is to identify and explicate the mechanisms that can be understood to produce the associations observed between consumption behaviour and the factors what are believed to explain it. 3.1 Economic resources The key variables in the analysis of all kinds of consumption are economic resources. The reasons for this are plain to see. If we exclude nature-based activities – such as picking wild mushrooms or berries – from our focus of interest, actualised consumption always involves spending money. Different people drive different kinds of cars, for instance, mainly because it is easier for a rich person to buy an expensive car than it is for a poor person. The same goes for many other daily practices which can be approached as consumption activities. In the simplest sense, the consumer market operates with commodities that have prices, and with money that people spend on these commodities. Thus it is surely crucial to take economic questions into account in order to understand people’s consumer behaviour. Consumer theory has been concerned with economic questions related to individual consumption especially in economics. Neoclassical economics, which has had a considerable impact on sociology, reduces the

75

consumer to homo oeconomicus, an abstract economic agent, who seeks to obtain maximum utility through purchase. It is assumed that the consumer is subject to financial boundary conditions only. Overall, economics has been confined to a motivational calculus of individual utility optimisation, which is generalised across all commodities and services (e.g. Smelser 1997, 3-4; Abell 1996). This is to say that the consumer behaviour of individuals is determined by prices, costs and benefits. Consumption is based directly on the utility preferences of individuals and their budget constraints. For example, the specific nature of what is consumed or the problems of changing consumption preferences usually receive little attention. In a sense, price elasticity or shadow prices remain the only important factor affecting people’s changing consumption preferences (Blundell 1988: 17-18; Pantzar 1996). Included in this view is the notion that the consumer has perfect information about products and services available in the market, time allocation and personal budget constraints (e.g. Fine & Leopold 1993: 46-47; Uusitalo 1998: 216). It should be noted, of course, that this narrow view of consumption determinants is usually sufficient for economic research, which is rather different from the typical sociological study. In addition, recent frameworks in economics have acknowledged the role of the social environment in the analysis of consumer behaviour (Becker & Murphy 2000: 5-7). But the idea that consumption is predominantly determined by the optimisation of utility within the prevailing budget and price constraints has been much criticised even as an theoretical notion (e.g. Campbell 1995; Uusitalo 1998). It is true that when the consumer stands in front of a looking glass, what is reflected back is surely more than just the image of an actor who is purposefully optimising the production of utility subject to the constraints of price and income. Despite this, one of the basic criteria for categorising consumers in all the social sciences is based on the economic position of the consumer. Economic resources naturally restrict the volume of consumption, but they also have an impact on how future expenditure will be planned. There are many determinants of the economic position of an individual, and they can affect consumption through various mechanisms. In empirical studies, however, the consumer’s economic position is usually measured in terms of household income from paid work. Earned income is

76

the easiest way to estimate the wealth and economic resources of individuals across different types of data. Information on property, such as holdings and ownerships, or other sources of income, may not be available at all. Mainstream research typically relies on the use of tax roll information, but subjective self-reporting are also sometimes used as income indicators (Heikkilä & Sihvo 1997, 120-121). In order to control the effects of taxation, disposable income is often preferred over gross earnings. Several reasons can be given for using household income rather than personal income. For example, it is often the case that the other members of the family are also important in determining an individual’s consumption possibilities. The household is often the economic unit. While it may well be the case that the main income-earner usually has the most to say concerning the budget, a partner or children may also greatly influence how money is spent. The effects of other adult members and children in the household have to be taken into account because households of different sizes have to budget their expenditure differently. This is why consumption unit equivalence scales has been introduced and utilised in many empirical analyses. These scales use different weightings in order to adjust household size in comparisons. The most widely used scale is the OECD consumption unit scale.2 Some alternative scales, such as the square-root scale, have also been introduced, and many modifications have been proposed to the traditional OECD scale (e.g. Atkinson et al. 1995; Ruggles 1990; Uusitalo 1993). However, the basic idea of equalising household size according to the number of persons in the household is basically the same across these scales. In short, income can be seen as the most important resource available to households, according to which consumption activities can be planned and carried out. In earlier times, the individual or household budget greatly restricted consumer spending. Goods were purchased only when there was money available. Especially before the industrial revolution and during early industrialisation, everyday life in many households was, if not a struggle for life, at least characterised by a scarcity of necessary goods. Food or clothing, for example, was scarce and people of lower social strata could not even think of spending on luxury goods. In this situation, income could predict the consumption of food, clothing and housing quite

77

accurately because there were few opportunities for households to deviate from the usual consumption pattern (Antonides & van Raaij 1998: 62; Heinonen 2000). The situation has naturally changed following the emergence of modern consumer society. All kinds of goods, from washing machines to mobile phones, have now become available to almost everyone. Obviously, income can still be seen as having explanatory power, although increased wealth has made it possible to deviate from the basic consumption patterns of necessary goods. Empirical studies show that in certain branches of total consumption the discriminating power of household income variables is still relatively strong. Scarcity and difficulty in satisfying basic material needs continue to trouble the poor even in advanced societies. In Finland, for example, it has been estimated that there are several thousand people without a home or other permanent shelter. As a more general notion, throughout the 1990s the proportion of households in Finland below the official poverty line – i.e. below 50 or 60 percent of the median income – has been nearly ten percent (Kautto & Moisio 2002: 330-332; Heikkilä & Sihvo 1997; see also Kangas & Ritakallio 1996). On a larger scale, too, household income is strongly related to the consumption possibilities of many goods. Let us for instance consider briefly consumers’ access to modern credit facilities. Even though there has been an expansion of credit-based forms of payment in recent decades, it is still the case that income is one of the key factors determining how much credit one can have. At the individual level, the greatest factor contributing to disruption of payments and running into debt has been a drop in income (e.g. Muttilainen & Tala 1998; Erola 2001: 180181). Thus income can be seen as a considerable predictor of daily purchases. The classic Engel curve assumes that as income increases, relatively more is spent on luxury goods and less on inferior goods.3 The distinction between luxury and inferior goods can be drawn on the basis of cultural statuses and tastes associated for instance with certain foods or consumer durables. Potatoes, for example, probably represent inferior food to many Finns, while caviar might be regarded as a luxury food. Underlying the choice of food are social conventions, values and interactions. It has been noted that in affluent societies food items become valuable and good only when they are conceived of as such (Pantzar 1996: 12). Following the

78

original idea of the Engel curve, it can therefore be proposed that the consumption of the kinds of goods that can be classified as luxurious might be explained quite well by economic factors. Luxury goods, which are usually more expensive than inferior goods, are simply easier to obtain in the highest income groups. Thus it is not surprising that in particular many leisure activities, such as travel, are explained relatively well by income; empirical evidence for Finland shows that households belonging to the highest income groups spend the most on tourism (e.g. Toivonen 1994; Wilska 1999: 38-39; Räsänen 2000). In general, economic status obviously has the greatest impact on the ownership of goods, and income is usually the most important factor affecting the economic status of an individual or a household. Thus we might call the explanatory mechanism according to which income level affects consumer behaviour the Mechanism of economic restriction. This mechanism can be understood in terms of a simple idea of how the available economic resources control one’s expenditure and how economic resources are connected to the planning of future expenditure. Figure 3.1 illustrates graphically how economic resources can be understood to affect consumer behaviour of people.

Determinants of economic resources

Available resources

Mechanism of economic restriction

Consumer behaviour

• Payment for work • Property • Other sources of income

Figure 3.1 Economic resources and their link to consumer behaviour Figure 3.1 summarises the explanatory process in a relatively formal and technical manner. It is true, of course, that economic resources affect consumer behaviour also in other ways. For example, household income affects the values and attitudes of all family members. This is because individuals living in different economic conditions have become accustomed to certain patterns of spending. They also make their future 79

plans on the basis of their resources and earlier consumption routines and customs. It follows that economic factors can affect people’s choices and behaviour rather profoundly. However, it is reasonable to believe that the primary mechanism of income is based one way or another on restriction. It is the lack of economic resources that makes consumption difficult, not their availability. It is frequently suggested that recent decades have witnessed a growth of economic inequality within the advanced societies. According to one interpretation, the composition of income distribution during the 1980s and 1990s has been characterised by the diminishing significance throughout the industrialised world of between-nation income differences and the growing prominence of within-nation inequalities (Goesling 2001: 745746). This is especially true when we compare estimated average national incomes with common measures of inequality, such as the Gini index, in cross-sections of national data. However, claims as to the diminishing importance of between-nation income differences have also been challenged using alternative measures (see Korzeniewicz & Moran 2000; Firebaugh 1999). This indicates the existence of many contradictory views in the field of welfare research. Nevertheless, it is reasonable to assume that unequal income distribution within a society continues to be reflected in the consumption patterns of people belonging to different income brackets. Still, it should be noted that many features of contemporary consumption have become too complex and varied to be described by simple economic functions and curves. The consumption of food, clothing or housing, for example, is associated not only with the satisfaction of hunger or shelter but also with other needs and values. Certain garments might be purchased only because of their status or symbolic value, while some foods might not be eaten because of religious beliefs or lifestylebased notions. This also goes for most other consumption activities. This means that there are other basic factors in addition to economic resources that can play a significant part in the determination of consumer behaviour. Social class has often been proposed as the most noteworthy factor affecting people’s consumption preferences.

80

3.2 Social class The most influential theoretical approaches to social class have been those developed by Karl Marx and Max Weber; most subsequent theories are based on the ideas of these two classics (Giddens 1973: 16-17). In all theories which have been proposed, class is broadly an outline of the ranking of individuals in a given society, based on property ownership, occupation and education. Classes are often correlated with income, but income as such is not seen as a determinant of a social class. Four to seven basic class positions are usually posited in the sociological writings (e.g. Giddens 1973; Bourdieu 1984; Bauman 1982). These positions run more or less hierarchically from the upper and upper middle classes through the middle classes to the skilled and unskilled working classes. The peasantry and a class of the excluded or underclass are sometimes proposed as additional positions in a basic class categorisation. In general, classes refer to structures that distinguish groups of people from each other at the societal level. As already noted in Chapter 2, class is understood to be the most important source of structural inequalities in modern market-economy societies. This is why at least among sociologists it has been obvious that class should also be seen as one of the key determinants of the varying consumption patterns of individuals. It should be noted that there has never been agreement as to what exactly class is and how it can best be conceptualised (e.g. Crompton 1993; Marshall 1997; Rose & Pevalin et al. 2001). Controversy over the theoretical definition and empirical measurement of class has been, in a sense, at the heart of social theory and research over the years. The origin of this centrality probably lies in the fact that classes have no physical signs or markers, and are therefore hard to observe empirically. Basically however, they are considered to be quite permanent and homogeneous categories of real life. Social class is identified as an abstract positional structure, based to a great extent on a hierarchy of economic categories. The term refers to the intersection between the social and economic dimensions of life. According to Malcolm Waters (1997), the ways in which sociologists use the concept of class can be summarised in four basic propositions:

81

1) The proposition of economism; 2) The proposition of groupness; 3) The proposition of behavioural and cultural linkage; 4) The proposition of transformational capacity.

According to the first proposition, social class is viewed as basically a socio-economic phenomenon, referring to differential ownership of property and differential market capacity. This means that ownership and market capacity are held to be fundamental organising principles in social arrangements. The proposition of groupness assumes that classes are believed to be more than statistical aggregates or taxonomic categories. In other words, classes are seen as real features of social structure, with detectable boundaries, which establish the main lines of cleavage in society. According to the proposition of behavioural and cultural linkage, class membership is causally connected to the individual’s consciousness, identity and behaviour also outside the sphere of economic production. This is to say that class position can determine values and norms, political preferences, lifestyle choices, access to educational opportunity and so on. According to the fourth proposition, classes are seen as important collective actors in the economic and political fields of society. In other words, as social actors classes have latent access to resources that can affect the whole of society by transforming the general set of social arrangements or by inducing a conscious struggle between members of different classes. (Waters 1997, 23-24.) Class can be understood as acting both as an explanatory structure and as a descriptive category in social life (Giddens 1973: 99-100). Different definitions of class of course emphasise different propositions. It is impossible to examine all the commonly applied class definitions and their assumptions here. I will therefore summarise distinctive notions related to class in terms of how class can constrain or structure individual actions and how it can be understood to affect the subjective experience and consciousness of individuals. My purpose is to address the primary mechanisms whereby class can influence consumption. In general, class is considered to consist of particular determinants and to influence consumption through particular processes. These determinants and their processes will be discussed below. However, I should perhaps

82

make it clear that the explanatory mechanisms of classes are approached in an oversimplified and a stable framework. While we speak of classes as categories, it is seldom possible to draw a clear-cut line between a group that possesses all the properties claimed to be typical of a particular class and a group that has none of these properties. Also, classes are viewed as more or less permanent categories, even though it is evident that – due for instance to less restricted access to education in modern societies – class positions are not necessarily determined by birth. In this context, however, I do not examine phenomena of social mobility, i.e. movement upward and downward in the class structure, or other features related to changes in the individual’s social position over time. 3.2.1 Class as cultural structure As we all know there are differences, for instance, in the living conditions and attitudes of people. It is also believed that there are some underlying realities that can explain these differences. In sociology, structural mechanisms are considered to be the common explanatory factors. In the most typical sense the effect of a social class is approached exactly in this way. This means that class is held as an important mechanism through which people’s life chances are distributed. It is assumed that classes are sets of structural positions, which exist through social relationships within market, especially within labour market relationships (e.g. Marshall 1997, 26-27; Giddens 1973). Essentially then, class is seen as a generative or determining factor in social life, which is an objective condition of human existence. As noted earlier, property ownership, occupation and education are the explicit determinants of class positions. The problematic question addressed is how classes can be objectively defined by using these positions. Several systems to measure class exist, but in general, objective class positions have been operationalised empirically on the basis of socioeconomic classifications of occupations.4 This type of measurement of class positions can be seen as contradicting particularly the traditional Marxist views on class. However, at least one considerable justification for using occupation as the basic frame of class-analysis can be presented. It has resulted from the fact that property ownership has changed relatively much since the early industrial era. In contemporary societies, property

83

ownership can be recognised by the formation of shareholder corporations, which tend to separate company ownership from the control and to dispersal the ownership among the shareholders. It follows that the break between ownership and non-ownership has become less sharp and classes have become, in fact, groups of similar occupations (Waters 1997, 31). Simultaneously however, it is still evident that strong correlations between income levels and occupational structures can be observed all over the advanced societies (e.g. Pöntinen 1983: 12-13; Savage 2000: 68-69). Almost all theories of social class share the assumption of the family as the basic unit of stratification (Sørensen 1994: 27). This means that members of the same household are assumed to occupy a single position in the occupational hierarchy. Accordingly, the household level of analysis is often preferred in empirical measurements of occupations because it is believed that the characteristics of all adults in the family are important in determining the living conditions of every individual in the household. Despite this, however, the socio-economic position of a household is generally based on the status of the head of the household, who is typically the husband. The main reason for adopting this strategy has been the fact that women’s attachment to the labour force has traditionally been much less common than men’s. Proposals to include information about both adults in measures of the class position of households have also been rejected on methodological grounds. It has been shown, for example, that a household’s class position may change every time one of the spouses changes jobs or moves in or out of the labour force (Erikson & Goldthorpe 1992: 238). The conventional strategy for measuring class has provoked a lively debate as to how women’s new economic roles should be perceived in the study of class positions. The question of whether individuals or households should be considered as the basic units of class analysis continues to be an important aspect of this debate (see Erikson 1984; Leiulfsrud 1991; Martin & Barton 1996). Several similar international measures have been developed for the structuring and discrimination of different types of occupational positions. The labour market situation and typical tasks are used as the primary characteristics in these measures (e.g. Erikson & Goldthorpe 1992; Wright 1985; Rose & Pevalin et. al. 2001). At a general level, the classifications cover that part of the population which is in formal employment. While

84

occupations actually relate to the technical classification of labour, occupational classifications are based both on technical classifications and on employment status. A primary distinction is accordingly usually made between employers, self-employed persons and employees. The unemployed are categorised according to their last occupation or the occupation associated with their educational background. In addition to this basic categorisation, occupations are classified in more detailed level applying managerial, administrative and professional measures and for instance characterisations based on the skill levels required for occupations. The principles described above are followed for example in the standard socio-economic classification by Statistics Finland, which is based primarily on occupation, occupational status, and type of work (see Tilastokeskus 1989). The socio-economic classification used in the empirical part of the present study will be presented in Chapter 4. To put it briefly, socio-economic differentiation in occupations is understood to reflect different class positions. It should be noted, however, that the principles of occupational classification are not in clear accordance with any given theory of class (see Pöntinen 1983: 40-41). In addition, it is also often impossible to incorporate data that is classified in a particular way within other categories (Erikson & Goldthorpe 1992: 35-37). At a coarse level, higher managerial, administrative and professional occupations for instance are usually understood to represent upper middleclass positions. Similarly, various white-collar occupations represent middle-class positions, and blue-collar occupations working-class positions. There are also certain separate occupational categories, such as farmers or members of the underclass. In theoretical interpretations, members of the underclass have poor conditions of work and living. This is why the long-term unemployed are usually operationalised as representatives of the underclass (e.g. Erola & Moisio 2002; Erikson & Goldthorpe 1992). In empirical studies, information regarding socio-economic position is often treated directly as the class position of an individual or household (e.g. Pöntinen 1983; Toivonen 1991; Inglehart 1997). There are several problematic issues related to such an approach. First of all, such positions as entrepreneurs and farmers become treated as separate categories outside the hierarchical class positions of the upper, middle and working classes.

85

The situation is same in the case of students, the unemployed, the underaged and pensioners. Secondly, some features of the present-day labour market, such as part-time employment, temporary contracts and continuously changing occupational titles, are not easily included in the classifications. It is thus reasonable to argue that the use of socio-economic positions indicate several inadequacies in terms of distinguishing groups of people from each other at the societal level. However, since the general problems related to class-based views in this respect were already discussed at some length in Chapter 2, I do not return to them here.5 In order to understand the explanatory capacity of class positions, we now turn to the kind of possibilities and constraints occupations might generate in different fields of consumption. The implicit idea in sociology is that the presence of common socioeconomic resources strongly affects the types of lifestyle people are able to engage in. In the analysis of modern consumer behaviour, for instance, it has been assumed that socio-economic factors may to a great degree structure the values, attitudes and behaviour of individuals (Marshall 1997, 33-34; Savage 2000; Kraaykamp 2002). While consumers from lower income groups might be able to follow many of the higher groups’ consumption patterns in the form of cheaper mass products, certain factors may still play a significant role in the formation of consumption patterns. But what actually are the explanatory mechanisms according to which class may affect the behaviour of individuals? One objectively constraining and enabling mechanism will be summarised below. Its starting point is in the core assumption that people’s preferences are not fixed but differ across different social classes. Customs, learning and social conventions, rather than certain stable preferences, are seen as bringing meaning to different life patterns. The general aspect of how class is constructed and how class position may affect individual behaviour can be described by referring to common cultural factors that can restrict individuals’ access to certain goods or services. Let us explicate this notion with a simple example. Perhaps we can all agree that members of the upper and middle class generally go to the opera and the theatre more often than people from lower social groups. One possible explanation for this might be that making sense of an opera or a theatre performance requires certain knowledge-based resources that are

86

lacking to people belonging to the lower social groups. It could also be interpreted according to a general notion that the upper social groups simply appreciate high culture more than lower social groups. In any case, the key assumption in both explanations is the notion of a certain shared cultural factor that the upper social groups possess and the lower groups are lacking. Cultural factors are something that can be understood as shared and common resources that the members of a particular culture recognise and are familiar with (e.g. Grey & McGuigan 1993: viii-ix; Lefebvre 1971; Therborn 1991). In class-based explanations, culture is interpreted in relation to the underlying system of economic production. Conceptually, cultural factors are understood in terms of cultural capital and by means of learned codes of preferences and liking. ‘Cultural capital’ refers to everyday knowledge that reflects routine experience and appropriate behaviour in particular contexts (Bourdieu 1984: 172-173; Kraaykamp 2002). Differences in individual values, attitudes and tastes are connected with perceived differences in everyday knowledge. According to one of the most basic assumptions in sociology, cultural factors are internalised as social norms and values and are thus manifested as collective, rule-following behaviour (e.g. Parsons 1951: 11-12; 36-38; Wrong 1961: 184-185; see also Therborn 1991: 182-184). Psychologically, the internationalisation of norms and values can be understood as taking place broadly as a learning process, or as a process of habit-formation in the simplest sense. It is possible to understand the socio-economic variation observed in social activities in the following way: since there are common class-specific meanings, prohibitions and institutions, which create unequal possibilities for people with different backgrounds, it follows that people behave according to the class-specific resources they have adopted and internalised. The second part of the proposition can be logically concluded from the first if we are ready to accept the internationalisation process as a feasible sociological explanatory assumption and if the observed structural differences are relatively systematic between different class positions. The logic that can be used to explain class effects in the field of consumption is that people with greater resources are more easily able to adopt certain consumption patterns. We can refer to this mechanism, which is schematically illustrated below, as the Cultural condition of class. It

87

provides one mode of explanation of how individual consumption activities can be connected with the class structure of society. In the simplest sense this means that there are discriminating cultural differences between social classes, for example in terms of avoiding risks, taking advantage of available information, learning by utilising familiar customs and manners etc. These cultural differences are developed and continue to be regenerated through educational attainment, occupation, work content and other relevant aspects of work. This description of class belonging is basically an explication of individual-level phenomena in different living conditions. What is important is that underlying different cultural conditions are both processes based on rational activity and ones based on passively adopted habits. An individual may actively seek to obtain class-specific knowledge or adopt it passively throughout his or her life. The logic of how class is constructed as a cultural structure and how class position can affect people’s consumer behaviour is illustrated graphically in Figure 3.2.

Determinants of social class

Class positions

Cultural condition of class

Consumer behaviour

• Property ownership • Profession • Education

Figure 3.2 The composition of class and its linkage to consumer behaviour Figure 3.2 offers a somewhat formal model for understanding how class affects behaviour. The effects of cultural conditions can be seen as constraining individuals’ possibilities of performing certain activities on the basis of varying life conditions. For example, in most European countries the children of upper and middle class families are probably likely to hear more about French literature than working class children. Matters related to literature are probably not a typical subject of conversation in working class households; upper and middle-class adolescents may for instance be more interested in French literature than 88

working-class adolescents. Thus cultural boundaries between class positions can also enable people to choose a familiar or appropriate product or mode of operation for instance in the consumer market. Now, while the role of tradition has arguably been diminishing in latemodern societies (e.g. Inglehart 1997; Giddens 1991), certain groupspecific cultural meanings have probably become more distinctive. It is perhaps increasingly difficult to argue that socio-economic position is the single most important determinant of contemporary consumer lifestyles in general. It is also true that some of the differences observed between classes can usually be reduced to differences in household income levels. Nevertheless, employment relationships and working conditions can presumably affect what kind of goods and services one values and prefers. It is thus reasonable to assume that class cultures can still create considerable cultural conditions, which have either restrictive or constitutive effects on consumption patterns of people. Broadly, these effects can be understood in terms of various professional and educational qualifications that differ across occupational positions. However, class can also be seen as affecting consumption activities through another explanatory mechanism. 3.2.2 Class as subjective experience It is generally argued that occupational positions which are relatively similar in terms of activity, work conditions, and perhaps income, suggest a common identity and lifestyle. Thus the cultural conditions related to socioeconomic differences serve to preserve cultural boundaries for classdistinguished modes of life. The notion of a common identity, however, also leads us to examine the expressive mechanisms of class. There exist certain class-based experiences, or collective forms of consciousness, which can determine how people associate with each other. In other words, we can speak of the effects of class-consciousness or class identity; how social position is experienced and expressed in everyday life. The use of the terms ‘class identity’ and ‘class consciousness’ naturally varies greatly among different sociological writers. The nature of class experience, however, can be understood by referring to the beliefs that people hold about their social position and how it is related to the positions of the others. According to the traditional Weberian view in

89

particular, the subjective evaluations of social differences have been approached as status distinctions. While status might be understood as varying independently of class divisions, privileged status groupings have been seen as consisting of the wealthy members of the upper classes (e.g. Parkin 1979; Waters 1997). This status, however, concerns a relatively small number of people and families. Of course, also middle-class and working-class statuses provide important information about images of class structure and status. The significance of class-consciousness thus lies in subjective class identities, i.e. in personal experiences as to which social category one belongs to and in which categories other people would be placed. A basic strategy in investigating class identification is simply to ask people in a questionnaire which class they think they belong to.6 There are various studies that have used this method in different contexts over the years (see Centers 1949; Jackman & Jackman 1983; Räsänen 2000; Erola 2000; Räsänen 2002). Some cultural differences can be found, but in general people identify relatively easily with the basic class positions of Western societies. According to Mary and Robert Jackman (1983), for example, in the early 1980s all but 3 percent of Americans identified themselves with one of the five choices offered; the options were upper class, upper middle class, middle class, working class and the poor. Only one percent identified themselves as upper class and about 80 percent saw themselves as either working class or middle class. To a certain degree, similar findings have been reported among Finns, using the class positions of upper class, upper middle class, lower middle class and working class (e.g. Räsänen 2000; Erola 2000). Also Mike Savage (2000: 34-41) discusses the measurement of class identity, referring to various studies conducted in a British context. It has been noted that there is usually a high level of agreement as to which occupations and occupational positions become associated with particular class positions. In American studies, for instance, physicians, lawyers and business executives have been almost without exception placed in the upper or upper-middle class (Jackman & Jackman 1983: 2024). However, even if people generally seem to identify themselves as either middle or working class, a considerable proportion does not necessarily identify with any of the basic class positions, at least when the

90

alternative ‘no particular class’ is offered. In Finnish data from the late 1990s, for example, 16 percent of the respondents chose this option (Räsänen 2000: 234-235). This finding indicates that in advanced societies some people clearly have difficulties in conceiving their social position in terms of class. The subjective measurement of class positions used in this study will be presented in Chapter 4. It should be noted that the subjective method of measuring social class in fact merges views on the one hand of class structure as such, on the other how class structure is experienced (Giddens 1989: 222). There are two main reasons why the two should be kept separate. Firstly, all class differences are not necessarily perceived in the same way by different people, even people who consider themselves as belonging to the same class. Secondly, identification with a certain class does not have to be based on occupation. Class identification may be based for instance on education, family background or lifestyle. Whatever the reason, it is plausible to argue that identification with a social class and personal awareness of this may significantly affect the pursuit of particular consumption patterns. Let us explore this notion a bit further with a simple example. We can assume that it is usually more important for a business director to think about the visual image of his or her company than it is for a plumber. One reason for this could be that in modern business circles an imposing logo for instance can function as a visible sign of competence and prestige. The competence and prestige of manual workers, on the other hand, is usually measured more directly in terms of their actual craft and skills. While it is reasonable to admit that this is a far more complex thing in real life, the crucial point is in the interplay between identities and consumption patterns. From this perspective, the business executive may feel that he or she has an obligation to have an impressive visiting card because of his or her position in the workplace. The plumber may have no such obligation. This example suggests that certain behavioural patterns are expected from people with certain roles (in this case the roles of manager and blue-collar worker). Accordingly, identification with a particular class can partly guide one’s consumption behaviour because it is generally known that particular groups of people have particular patterns of consumption. Thus class identities provide possibilities of distancing

91

oneself from certain groups of people in terms of responding to the roleexpectations of certain generally known social roles. The fact that behaviour is often linked with many social and cultural categories, and that some of these categories construct identities more than others, is important in understanding class identities and their behavioural effects. Particularly those categories that are believed to be relatively stable and that can bring together people in terms of life changes, attitudes and values, may be highly important. It is therefore possible to approach one mechanism of class by referring to life conditions that are determined by subjectively experienced social position. This mechanism can be labelled here as the Mechanism of class identity. In a sense, it provides a reference for the consumer in comparing his or her consumption patterns to those of other people. It is possible to understand the explanatory power of class identity by referring to the surrounding cultural expectations. These expectations are directed towards the members of a given class and concern their personal identification with this class position. It is also important to address class identity in the context of people’s own experience. Briefly, this means that people who feel that they belong to a certain class also have a need to express it in their consumer behaviour. In this way certain class-based roles, consumption practices or the possession of certain kinds of goods may be used as a means of membership in one social group and of marking off certain groups from others. Class identity can therefore be considered to have an effect on consumption, since people who feel that they belong to a certain group often construct and maintain certain routines and act according to habits they have established. Basically, the mechanism of class identity can be understood rather similarly to the mechanism of cultural condition of class. Both mechanisms are manifested as collective behaviour that can be approached in the light of the underlying structural variation. But to a certain degree at least, identification with a class may be detached from occupational structures. While there may generally be a high level of agreement as to which occupations become associated with particular classes (see Jackman & Jackman 1983; Savage 2000), these associations do not necessarily have to be the case at the individual level. This means that while occupational position may create restrictive and regulative cultural conditions, class

92

identity is a clearly more subjective measure. To be ‘middle class’ does not mean that one has to work in an office, own a new Volvo or live in a private home in the suburb of a large city. Take the example of an unemployed university graduate. Such a person may feel that he or she belongs to the middle class simply because of his or her education, previous work experience or the jobs that he or she is seeking. Occupation probably represents the basic source and core of class identity, but it is not the only source. What is important is how one’s position is conceived and experienced in one’s mind. On this basis the effects of class identity should also be approached slightly differently from the effects of socio-economic position. This is depicted in Figure 3.3, which illustrates the mechanism of class identity.

Determinants of social class

Class identity

Mechanism of class identity

Consumer behaviour

• Property ownership • Profession • Education

Lifestyle, future expectations, identification with social groups, etc.

Figure 3.3 The origins of class identity and its linkage to consumer behaviour Figure 3.3 shows that distinctive identities can be derived from socioeconomic positions and from various social categories or group memberships. While the figure again offers a somewhat technical illustration, we can see that the content of an individual’s class identity consists of the groups and categories to which he or she belongs or wishes to belong. Consumer activities – expenditure on particular products and services – can also contribute to the formation of identity. It is reasonable to assume that certain brands, for instance, are important in defining who people feel that they are and how they wish to be seen. Class identity is a purely subjective experience as to which social category one belongs to. What is noteworthy in this respect is the interplay 93

between consumer behaviour, lifestyle and identification with certain social groups. By no means, however, should this suggest that exploration of the possible structural boundary conditions for identification with a class is irrelevant. On the contrary: it is of great importance to question what kinds of factors actually form the basis of one’s class identity. In many cases, for instance, crucial information may be provided by income, educational background or personal work history. If possible, controlling the effects of certain other dimensions of class, such as education, is desirable in the analysis. People’s class identities are influenced by the class position of other members of the household. This is also true of other relevant social and cultural factors that may create boundary conditions for the identification process. In various empirical studies, it has been argued that the economic situation of the consumer is a key factor that creates restrictions on the expression of class identities through consumption (e.g. Räsänen 2000; Räsänen 2002: 167-168; Wilska 2002). If we take the requirement of money into account, an investigation of economic factors thus naturally becomes unavoidable. Along with class, there are various other structural elements that can be of help in the sociological categorisation of consumers. Some factors can create differences in the life conditions of individuals in all class positions and income brackets. For instance, we can say that men and women consume certain things very differently from each other. So do people of different ages and people who live in different types of households. Next, we take a look at the mechanisms of these basic factors. 3.3 Age, gender and the type of household Social categories related to age and gender can be seen as constructing the basic outlines for everyday experience. Gender and age socialisation represent the core social processes, during which learning from adults and peer-group influences are integrated with information derived from books, stories, television and so on. All normal children go through a stage after which they are able to label themselves as boys or girls. This distinction is experienced as fundamental, relating not only to anatomical differences between males and females but also to psychological, social and cultural differences related to gender roles. Age is a very similar aspect of human

94

existence; the meanings of chronological age, physiological age and age as a cultural fact are also internalised relatively early in life. In many ways, age and gender can be regarded as the most basic and perhaps most powerful categories of social life. The typical course of modern life, if such exists, is built to a large extent around age and gender. The role of sexuality in the division of labour or advertising, for example, is extremely prominent. Gender and age relations are generally seen as representing particular activities and their meanings. This is why they also have an effect on consumption. A retired individual who lives alone has completely different consumption routines from an individual of working age with two children. Men and women have also been found to have relatively different consumption patterns at least in certain fields of activities. These questions have been explored profoundly in the sociological literature over the years (e.g. Lunt & Livingstone 1992; Wilska 1999; Reekie 1993; Chattoe & Gilbert 1999). In general, both the age and the gender of an individual should be regarded as highly important in most social and cultural contexts. Before going into the mechanisms through which age and gender potentially affect people’s consumption activities, I glance briefly at the empirical measurement of these variables, and at certain problems encountered in this respect. Basically, the measurement of a subject’s age and gender should be simple enough if the data are gathered at the individual level. Respondents typically mark their chronological age (typically the year of birth) and sex in the questionnaire or tell it to the interviewer. If the year of birth is not used as an indicator, there is often a margin of error of one to five years. It is generally considered that there is virtually no reason why subjects should misreport their age and sex in the typical conditions of social-scientific research. Subjective measures can therefore be relied on as valid indicators of the respondents’ age and gender. In the statistical analyses, age might be treated either as a categorical or a continuous variable, and gender as a nominal variable. However, determination of the age and gender of a respondent is greatly complicated when the research unit of the study is the household. As we know, households usually consist of men and women of different ages; thus the actual ‘age’ or ‘gender’ of a household is impossible to determine unambiguously. Certain technical principles have therefore been adopted to overcome this problem.

95

One widely used option is to draw inferences from information based on a single individual. In other words, age and gender are only known for the head of the household or the reference person in question. Such solutions have been common in the sociological analysis of consumption expenditure. But this may be problematic; the information on the reference person does not necessarily provide enough information on the composition of the household (Härkönen & Kosonen 2003). There may for example be more than two adults in the household, who as consumers may be relatively independent of the reference person. Power relations within households are important in this respect. It is not always the case that all family members make similar decisions. Rather, some members of the family are usually able to carry out their decisions even when opposed by others. Common sources of disutility reported in the literature include male oppression of women, the higher earnings of one member of the household, and various implicit contracts during cohabitation (see Blossfeld & Drobnič 2001: 2429). Also children can sometimes dominate decision-making in the families. Most people live in households, excluding of course people living in prisons or hospitals and people without a permanent residence. It is thus often more straightforward and more feasible to refer to the type of household in considering the effects of the basic characteristics of households (e.g. Raijas 2000: 50; Räsänen 2002: 164-165; see also Antonides & van Raaij 1998: 27). The classic concept of the family lifecycle (Glick 1947; Waite 1980) is useful in this respect. Households can be classified relatively easily according to the number of people living in the household and their age. The classification applied by Statistics Finland, for example, defines household type on the basis of the number of people in the household and their age (see Tilastokeskus 2001). Thus such households as adolescent families, middle-aged and elderly families can be easily distinguished. Households with and without children can be also kept distinct from each other. In practice, of course, the nature of the data utilised often imposes certain limitations on sophisticated categorisations. Information as to the gender of households, among other things, may be lost when household classifications are used. In general, however, it is reasonable to assume that the information available as to type of household will include the necessary information as

96

to the people living in the households in question. The type of household, as a variable roughly describing the life-cycle stage of the people living in the household, can be regarded as a basic category similar to age at the individual level. The age categorisations and the classification of household types used in the empirical part of this study will be described in Chapter 4. Since age, gender and household type are basic categories of social life, their mechanisms can also be approached from relatively similar frameworks. Both gender and age might be approached as factors that can affect consumption, such as structural conditions and age and gender identities. In other words, there are boundary conditions on both factors in analysing certain consumption activities. Individuals below a certain age, for instance, cannot legally purchase cigarettes or alcoholic beverages at all. It is often impossible for very old people to take part in a marathon or even to drive a car. In addition, however, the consumption patterns of men and women and of people of different ages may also be influenced by general role expectations as well as by generational differences connected with consumption. For instance, styles of dress in western societies are quite strongly connected to the individual’s age and gender. Naturally, both men and women may dress in jeans and tee-shirts; but a 70-year-old man wearing a skirt in public may easily become an object of astonishment under normal circumstances. Fashion appears to break down these cultural distinctions rather slowly. To put it more generally: in particular the consumption of products and services that are socially relatively visible, such as clothes, make-up and certain leisure pursuits, are culturally relatively fixed consumption activities. For example, in Western societies children are not expected to be very keen on golf, and the average man might not be as interested as a woman in such activities as buying clothes or furniture. It is often the case that men have to demonstrate their masculinity and women their femininity. Gender-specific associations, for instance, are deeply embedded in the institutional structures of work, leisure and household maintenance. As individuals become engaged in role behaviours, they absorb identities linked to those roles (Blossfeld & Drobnič 2001: 32). In this respect, it is meaningful on many occasions to refer to gender roles, while it is evident that people may consume many items and services rather differently in different phases of the life-cycle.

97

There are some noteworthy problems related to the analysis of age, such as whether the phenomena discovered are related to certain life-cycle stage or whether they are actually typical of broader groupings, such as generations. In the literature, these problems are typically referred to as the period, generational and age fallacies.7 In general, however, it can be argued that many consumption activities can be understood in terms of the individual’s age or life-cycle phase. Age has explanatory mechanisms similar to those of gender, which can both regulate an individual’s actions more or less directly and also lead people to act according to certain culturally defined expectations. In an empirical analysis where the unit is a household, on the other hand, these notions might be approached from a more demographically oriented point of view by referring to the type of household. The mechanisms according to which the age and gender of individuals can be seen as influencing consumer behaviour are labelled here as the Mechanisms of basic social categories. These mechanisms are illustrated graphically in Figure 3.4.

Basic social categories

Age, gender and the type of household

Mechanisms of basic social categories

Consumer behaviour

• Age • Gender

Cultural expectations, personal experiences, social group memberships, etc.

Figure 3.4 Age, gender and type of household and their links to consumer behaviour Consumption patterns that are based on age and gender differences can have many origins. As shown in Figure 3.4, various subjective factors, such as personal experience or social group identity, can affect the consumption activities expressed. In reality, of course, the mechanisms for age, gender and type of household may be somewhat different, but their origins lie in the same phenomena. This means that distinctions based on age and gender are profound distinctions in culture and in daily practices, which can 98

influence consumer behaviour by creating constraining or enabling structural differences and by promoting expectations related to these differences. While the impact of gender roles should not be assumed to be irrelevant in the determination of daily consumption patterns, it is reasonable to postulate that the restricting and enabling effects of ‘objective’ basic social categories on consumption is often of more importance. Many consumption patterns can be approached from a perspective of typical life situations, primarily defined by constraints and variations related to them. With regard to age and gender, for instance, the consumption of certain durable goods and services may simply not fit the lifestyle of seniors or of men. In this respect, men and women and people of different ages are structurally in different positions in many respects. In addition to the basic social categories of age and gender, there are also other factors that may structure the consumption behaviour of individuals. For example, different surroundings enable diverse activities and produce different forms of interaction for people living in different places. The role of the place of residence should therefore be considered relevant in the analysis of consumer behaviour. 3.4 Place of residence There are various reasons why place of residence can be viewed as one probable element affecting people’s consumption patterns. On the one hand, consumption possibilities often vary between different places. Prices may differ, and there are varying opportunities for shopping and consumption. A typical example has to do with housing type. Home ownership or rental are much more expensive in urban centres than outside cities. The supply of dwellings may also be different. Detached houses, for example, are more abundantly available outside city centres. In large urban communities there might be much less housing choice available in than in smaller ones (Ruonavaara 1993: 44-45). In addition, taking into account the differences between city centres and areas outside the centre, it is evident for instance that the former offer far more locales for consumption expenditure (shops, malls, restaurants etc.) than the latter. On the other hand, there may be different kinds of people living in different parts of both rural and urban environments. Particular groups, such as ethnic minorities,

99

professionals or artistic avant-gardes, come to live in certain areas and their lifestyles come to typify these areas and neighbourhoods. It can be argued that some residential settings might be prompted by different expressions and ideologies of consumption. An examination of the role of place of residence in consumer activities is of course a very complex task. For example, movement between different places is nowadays rapid and easy. It follows that people living in a certain town or area do not necessarily spend most of their time there. Many people have a job in another city, and many families spend all their weekends at their summer cottages. The integration of information technology and consumption, in the form for instance of online shopping, makes physical location a less and less important determinant of consumption behaviour. However, it should arguably also be possible to examine certain themes related to residential area by focusing on particular core questions. The distinction between urban and rural areas raises one question of considerable importance which is reflected in at least some areas of consumption. The advanced societies are highly urbanised. In spite of this, a considerable proportion of the population still live outside large cities. It can be assumed that residential area can be linked with variation in individual consumption patterns, but how exactly? As noted earlier, traditional urban and rural environments have by no means lost their relevance in contemporary suburbanised cities. So how is the decision made in the first place to label one area as urban and another as non-urban? Basically, the determination of the urban-rural distinction should be seen as an important empirical question. While there may be certain problems depending on the methods of data collection, the degree of urbanism can still be determined relatively easily using postal code information and statistics of population density. Official statistics, for example, typically combine these criteria in the formation of their residential classifications (see Tilastokeskus 2001). Another method used in empirical studies involves asking respondents to rate their place of residence – for instance using the Likert scales, anchored between urban and rural, or simply giving respondents alternatives to choose from (e.g. Kraaykamp & Nieuwbeerta 2000; Räsänen 2000). Measurements relying on register information and on subjective evaluations may yield slightly

100

different results, but both methods are utilised in contemporary studies. The classification of place of residence used in the empirical part of this study will be presented in Chapter 4. According to many sociological writers the geographical impact of the consumer society is primarily an urban one. It has been argued, for instance, that contemporary urban life is based to a considerable degree on networks of mass communication and entertainment, which disseminate the wares of a consumer culture (e.g. Thrift 1993; Miles 1998: 52-53). Thus it has been noted that consumption can be seen to a large degree as constructing urban life in everyday contexts. City centres are perceived as consumption havens; the adjacent suburban areas, on the other hand, emerge as exporters and importers of people and services to and from city centres. In this respect, car ownership for example has been viewed as an especially important item of symbolic display (Chaney 1990: 52-54). In rural settings, on the other hand, car ownership may still represent a mean of conveyance in its purest form, while in city centres a car might not be seen as necessary at all. But how, in fact, do certain consumption practices differ systematically across residential areas? Since the days of the Chicago School, and Louis Wirth in particular, urban and rural life have traditionally been distinguished as different psychological conditions.8 The urban milieu has been understood as the antithesis of a rural environment. The latter is often imagined as a small community, where relationships between people are quite personalised and intimate. Cities, on the other hand, are seen as large and heterogeneously populated, impersonal and anonymous places. According to such descriptions, city life is characterised by isolation and social disorganization, while rural life is characterized by a safe and coherent culture. While in the real world urban-rural differences are always a matter of degree, they can still affect modes of life in different areas. It is believed that city life-ways differ from life-ways in smaller places because larger places develop more specialised subcultures and more varied daily routines than less populous ones (Fischer 1995: 544-545). Accordingly, it can be concluded that in urban settings different consumer durables and services may provide rather different frameworks for lifestyle enactment than in rural settings. Shopping, for example, has been offered as a generic

101

description of the actual activities of what people are doing, thinking and experiencing while moving around in urban public places (e.g. Stone 1954: 36-37; Falk 1997; Lehtonen & Mäenpää 1997, 143-144). Shopping can thus be viewed as a leisure-time activity and an enjoyable use of time even aside from the purchase of goods or services. No analogue of this characterisation of the modern flâneur/flâneuse can be found in descriptions of public life outside city centres. To put it briefly, it can be assumed that consumption activities and aspirations may vary relatively much between different residential areas. Such activities as expenditure on vehicles and restaurant services or frequency of recreational travelling, for instance, may depend heavily on the individual’s place of residence. Whether or not systematic consumption differences can be found between people living in different types of residential areas, it is possible to approach the urban effect on consumption behaviour. This refers to the idea that individual consumption patterns may be structured by certain environmental features. The mechanism according to which place of residence may exert such an influence is labelled here as the Mechanism of residential area. Characteristics of the residential area can be understood as contributing to differences in the supply and demand of goods and services; or, to put it more simply, different living environments may give rise to different kinds of consumer needs. The effects of residential area are displayed in Figure 3.5.

Characteristics of residence

Place of residence

Mechanism of residential area

Consumer behaviour

• Physical location • Services, distances etc.

Figure 3.5 Place of residence and its link to consumer behaviour It is reasonable to assume that the variation in consumption patterns between places of residence can be scrutinised in the light of the physical characteristics of the surroundings. This means that we can assume that certain objective features of the environment, such as the number of shops 102

and malls or distances between places, are reflected in individuals’ needs, social activities, attitudes and spending. The origins of varying consumption preferences and spending can be attributed to the characteristics of the residential area, as illustrated in Figure 3.5. In a way, then, the mechanism of residential area can be approached without serious problems from the perspective of a restrictive structure. In other words, different consumption patterns in different places can be seen as originating from the different needs, requirements and preferences, which are constrained by the physical characteristics of residential areas. But it is also possible to assume that certain residential characteristics have also an enabling effect. For example, the presence of shops and restaurants in urban settings may lead people to discover various new ‘needs’, which are manifested in preferences and patterns of spending. In this respect, it should be noted that some differences between places can be purely incidental. This means that certain patterns of consumption may differ between urban and rural areas because, and only because, their residents differ. If possible, the analysis should therefore control for the effects of other possible sources of variation, such as individual or household income or socio-economic position. While economic factors, basic demographic features, reference groups and residential areas are often seen as the most obvious sociological predictors of consumer behaviour, the role of other factors may also be addressed. Particularly during the last two decades, discussion of the significance of culturally determined new consumer groupings has become popular within the sociology of consumption. Let us also view them from a mechanism-based angle. 3.5 Lifestyle and other relevant consumer groupings Numerous writings can be found in the recent literature, suggesting that the idea of a strong connection between traditional socio-demographic structures and consumption patterns is a misleading one (e.g. Tomlinson 1990; Featherstone 1991; Maffesoli 1996). Particularly among the sociologists of postmodernism it is believed that the effects of social structures in daily life have weakened and that the importance of individual choice has increased. As a result, behaviour patterns in general have become more variable. In this situation, many concepts related to ethnicity,

103

sexual orientation and cultural background have been the subject of lively debate in recent studies of consumer behaviour. In the few past decades, the advanced societies have witnessed a proliferation of alternative modes of life. A preference for ethnic reidentification over assimilation and the increased visibility of homosexual relationships, communal living and various subcultures, among many other changes, have been responsible for the emergence of new social distinctions. The living conditions of ethnic minorities, various youth subcultures and special interest groups, for instance, have been approached in various studies (see Zablocki & Kanter 1976; Hebdige 1979; Lachmann 1988; Kotzinets 2001). Many of these emerging phenomena are interesting mainly in an ethnographic sense, but some of them also have a considerable impact on the formation of consumption preferences. From this perspective, ethnicity and the newly recognised cultural factors might be regarded as providing relevant consumer groupings for a limited range of activities. However, the effect of these factors is somewhat difficult to evaluate because the necessary information – for instance on ethnicity – is always available. The concept of lifestyle is seen more generally in culturally oriented approaches as a profound source of consumption patterns differentiation. Lifestyle is seen as a complex phenomenon, composed of many different elements. It is regarded as an important and distinctive concept, and is believed to describe consumers’ behaviour and their preferences, choices and aspirations. In fact, lifestyle has been one of the most visible concepts in the literature of the social sciences since the late 1970s. Naturally, various alternative definitions of the concept have been presented (see Mitchell 1983; Roos 1986; Allardt 1986; Eskola 1985). Basically, the underlying idea of the concept of lifestyle is to combine descriptions of the actual behaviour of individuals with a holistic view of life situations. In the simplest sense the concept aims at describing the way people actually live. It is assumed that the common meaning of lifestyle can be understood almost instinctively; a more profound definition, however, is much more difficult to formulate. The concept has typically been used in the broadest possible sense in order to distinguish different groups or segments of people.9 Lifestyle is defined as the entire set of values, interests, attitudes, opinions and activities of people insofar as they

104

influence observable behavioural patterns (Antonides & van Raaij 1998: 376-377; Lee 1993). It is often added that while lifestyle categorisations may not be the same across all domains of life, they are stable to some extent and can therefore organise individuals’ activities (e.g. Eskola 1985: 167; Roos 1986: 37-38). These definitions indicate that lifestyle can be perceived from the viewpoint of the life practices of an individual or a group. Lifestyles have been conceived as relatively lasting structural characteristics and disposition of a given social group, gender or generation. The bond experienced among those who practice a similar lifestyle has been a central idea in many theoretical models. It is assumed that previous life experience can provide a common ground for the beliefs, values and habits that structure daily lives. Most conventional approaches have anchored lifestyles in the corresponding class categories (Roos 1986: 43-44), in the attempt to capture the distinctiveness of class-specific lifestyles (see Havinghurst & Feigenbaum 1959; Uusitalo 1979). The constellation of values and attitudes embodied in class experience has been regarded as crucial in the implementation of many activities. Distinctions between middle-class and working-class lifestyles, for example, have been discussed in many studies. A classic example is the work of Pierre Bourdieu (1984). Bourdieu has carefully examined the social factors which play a significant part in people’s everyday choices, ranging from clothing and food to leisure activities and other broader matters of taste. He argues that social snobbery, in terms of intellectual and aesthetic considerations, has been a common theme in the bourgeois world in the modern era. According to Bourdieu, snobbery remains pervasive almost everywhere in middleclass and upper-class life. Semiskilled and unskilled workers in turn can be seen as countering this snobbery by evaluating cultural objects and social practices according to their material worth, which often represents the ‘true value’ of things for the working classes. This basically indicates that the value placed on education, knowledge and high culture varies in different social positions. (Bourdieu 1984: 260-261; 372-375.) Bourdieu’s ideas have led to many later studies on the experience of consumption, in different parts of the world (see Miller et. al. 1998; Katz-Gerro & Shavit 1998; Räsänen 2000; Kraaykamp 2002).

105

Another conventional domain of lifestyle research has viewed class expression especially in the context of industrialised society. The processes related to industrialisation, such as the drift into the cities, factory work, and the prolongation of school attendance have been perhaps the most obvious applications of lifestyle analysis in Finland (see Kortteinen 1982; Ahponen & Järvelä 1983; Roos 1987). Finnish studies have focused on the changing life conditions of the working classes and problems emerging in the new situation. The analysis of the class experience of the working classes has been a visible theme in some of the most interesting studies of everyday life in a global perspective as well (see Sennett & Cobb 1972; Willis 1977). People’s lifestyles have been approached sociologically from the perspectives of labour participation, educational background, and the values and interests characteristic of different generations. More recent views, however, stress the fact that there exist various kinds of sport, hobby and other communities, which can function as specific sources of lifestyle with their own styles. The abstract and theoretical figures of the contemporary consumer, such as the ones collected by Arto Noro (1995), were briefly discussed in this respect in the previous chapter. Let us now examine the notion of self-creation through consumption especially in the context of lifestyle. Some of these ideas suggest that lifestyles are primarily constituted through an interest in shared activities and preferences. Such communities are typically quite loosely organised, but it is considered that they can be intensive in nature and experienced as important by the participating members. The idea of ‘lifestyle enclaves’ is perhaps the most prominent. According to Robert N. Bellah and his colleagues (1985), a lifestyle enclave is formed by people who share some features of their private life. Members of a given enclave express their identity through shared lifestyle, in terms for instance of leisure activities, appearance or consumption patterns. What is important in lifestyle enclaves is that the members are not interdependent in any way and do not share a history. Enclaves are relatively independent of ethnic and class background: they simply consist of groups of people who share certain practices. They are thought to exist most obviously in large cities, where groups of people have little in common except for the way they spend their leisure time. (Bellah et. al. 1985: 73-75; 335.)

106

Some writers argue that lifestyle in general is becoming far more individualistic than it used to be (see Beck 1992; Featherstone 1991; BeckGernsheim 1996; Beck et. al. 2003). This means on the one hand that characterisations of class-based activities or other homogeneous consumer cultures are no longer an obvious way to approach contemporary lifestyles; it is assumed that there are too many social groupings in society, and individuals are expected to move relatively easily between these groups (Bocock 1993: 181). On the other hand, middle-class values and attitudes are assumed to expand; thus the ambiguous ways in which people live cannot be located on a structural basis of conventional class positions (e.g. Sulkunen 1992; Maffesoli 1996). The vast majority of people share a common standard of living, and lifestyle differences between social groups are small. On a broader level, too, according to many individualistic ideas the roles of the central institutions of modern society, such as class and the family, have diminished. This is expected to lead to a situation in which the lifestyle one practices increasingly becomes the basis of one’s social identity, displacing class, age and gender as the central organising principles of social life. What opposes these views to basic structural views is their perception of lifestyle. In recent writing on lifestyle, it has been regarded merely as a personal project that shifts according to the social context. In its extreme form, this means that unity with other people may be experienced only because of certain shared practices, but that any durable lifestyle communities hardly exist (Featherstone 1991: 83-84; BeckGernsheim 1996). As is easily observed, there is a clear contradiction between conventional (structural) views and more recent (individualistic) views of lifestyle determination. It is nevertheless generally considered that the expression of one’s social position in contemporary society has become more closely attached to consumer lifestyles. Following the central premises and consequences of consumer society, such as the increase in the level of consumption and the formation of consumer identities, people’s life practices have become relatively consumption-oriented. In other words, consumer lifestyles are due to the ubiquity of consumption in present-day life. People who can be understood to live in a generally similar fashion can also be understood to have similar consumption patterns, and vice versa. But how can individual lifestyle be measured empirically?

107

The measurement of lifestyle involves numerous problems. In the first place, the concept of lifestyle consists of types of activities and the ways in which individuals practice them. In this sense lifestyle should be seen as a behavioural concept, comprising a number of factors. It is therefore a descriptive concept, with no explanatory power in the conventional sense. As noted earlier, however, values, interests and opinions can be regarded as the basic characteristics of one’s lifestyle. This indicates that lifestyle can be measured only indirectly. The usual way of approaching lifestyles systematically is by measuring values, attitudes and personality characteristics. These measures can offer, at least to a certain degree, theoretically grounded approaches to the identification of types of people according to personal characteristics and possible behaviour patterns. Lifestyles are believed to originate from differences in these factors. The VALS (Values and Lifestyles) typology, based on the acceptance of values (see Mitchell 1983: 243-272), is perhaps the most widely known lifestyle typology, but several other measures have been developed over the years.10 The idea of lifestyle typologies is that people fall into a given category on the basis of a lifestyle defined in terms of values or attitudes. It is reasonable to assume that some typologies can be useful in distinguishing different consumer and lifestyle types. Nevertheless they are not very often cited in the social science literature. This indicates that value-based measures in themselves are seen as somewhat insufficient for academic purposes. Lifestyle is in fact co-determined by many demographic and socio-economic characteristics; without information on these, lifestyle typologies remain merely descriptive names for certain types of activities. After all, as noted earlier, lifestyle is in the first place a behavioural concept. Thus simple segmentations of values, opinions or attitudes cannot provide a solid basis for a comprehensive understanding of individual consumer lifestyles and behaviour. Despite this, Figure 3.6 aims at illustrating the logic of lifestyle determination and its effects on consumer behaviour. According to Figure 3.6, the Mechanism of lifestyle can be seen as originating in life practices and appearances. In principle, the lifestyle patterns of individuals are seen to be relatively independent of traditional structural backgrounds. This indicates that the origins and expressions of lifestyle can be understood in rather individualistic terms. They can also be

108

seen as cumulative phenomena. For example, it is assumed that the practice of one consumer activity, such as reading, will lead to other related activities that together form a lifestyle. Among sociologically oriented researchers, however, lifestyle is seen as a combination of individual choice and normative action based on various structural restrictions. In this way, consumer lifestyles have been operationalised as patterns of consumption, consumption preferences and participation in various leisure activities. Common lifestyle indicators used in empirical studies include for example hobbies and leisure pursuits, holiday choices, and tastes in clothing, music and reading (e.g. Bourdieu 1984: 167-175; Katz-Gerro & Shavit 1998; Toivonen 1994). In the selection of indicators certain common dimensions of lifestyle differences are usually represented, such as distinctions between high and popular culture or between home-centred and community-centred activities. In practice, of course, almost any activity can be understood as measuring an aspect of individual lifestyle.

Lifestyle

Mechanism of lifestyle

Consumer behaviour

Daily activities, values and attitudes, social group memberships, etc.

Figure 3.6 The origins of lifestyle and its linkage to consumer behaviour It can be argued that even if information on a person’s life practices can be seen as the determining factors of his or her consumer behaviour, it will still be hard to answer the question whether these practices actually explain the differences observed. It is possible, for example, that both the values and the behaviour of an individual are affected by economic and demographic factors. If this is the case, then the value-behaviour relationship may be the result of the combined effect between demographics and values. This suggestion makes a value-based view of 109

lifestyle somewhat illusory. In addition, this indicates that the effects of the other possible determinants, such as income, class position and age, should be controlled in the analysis. Now, it is reasonable to argue that the effects of traditional economic and socio-demographic factors should be examined carefully before we are able to approach the mechanisms of lifestyle from a feasible sociological angle. Lifestyles will therefore be evaluated in this study by treating consumption activities as indicators of lifestyle patterns. In other words, differences in people’s consumption patterns will be addressed empirically by means of the economic and socio-cultural factors that have been discussed in this chapter from the point of view of explanatory structural mechanisms. 3.6 Summary: Understanding consumption in contemporary society One of the most distinctive features of the consumer society is that everyone can be seen as wanting the things that automated mass production is able to provide. Due to massive economic growth during the last few decades, many consumer durables have become available to practically everyone. This is to say that the standard of living has risen dramatically across all social strata. Some sociologists of consumption have accordingly argued that knowing for instance a person’s class position or occupation predicts almost nothing about their consumption patterns. To put it more generally: class is seen to be merging with many other aspects of social life, such as ethnicity and lifestyle, and society itself is expected to grow more fragmented and fluid. In this situation, it is believed that any systematic invariance between basic demographics and consumption hardly exists. However, it is possible to argue that this vision of contemporary society and social life is an oversimplification. According to this chapter, it is reasonable to assume that the basic socio-demographic categories can still provide fairly efficient approximations of preferences in daily expenditure and consumption choices. Obviously, the figures and mechanism-based descriptions presented above offer a relatively restricted view of the determination of social behaviour. These interpretations can offer only one kind of description of the nature and development of consumer choice differences. As already noted, consumption is surely too complex an empirical

110

phenomenon to be explained exhaustively by economic, social and cultural factors alone. In many cases, factors related to the context of everyday life, such as the situation and personal mood of the consumer, may help us to understand the ultimate causes for choosing one product or service over another. These factors, however, are outside the scope of this study. Even if the basic socio-demographic determinants can help us to understand consumer behaviour only on a limited scale, they can still reveal important information about the nature of people’s consumption patterns. It is from this perspective that the empirical research questions of this study will be addressed. Before going more deeply into the framework of the empirical part of this study, however, let me comment briefly on the analytical difficulties related to the individualistic postmodern notions presented in contemporary sociological interpretations. 3.6.1 Methodological difficulties related to postmodern notions As pointed out earlier, many postmodern notions assume that the importance of individual choice has expanded. This expansion of choice is followed by new demands and an increased sense of insecurity related to self-actualisation and personal life planning. Social structures and divisions based on position in the production process are believed to have relatively little or no effect at the level of everyday life activities. In other words, postmodern sociology commonly stresses the changing character of life. It is not always clear, however, what it is that has actually changed. It has been suggested that perhaps the greatest difficulties of postmodern critique have to do with the nature of the actual arguments (e.g. Bradley 1996: 68-69; Lemert 1997; Goldthorpe 2000). First, postmodern arguments are usually connected with rapid social change, which is described as taking place on a large scale and as including various political and cultural processes identified at different levels of society. These changes are understood to be manifested in the personal experience of individuals and in basic social institutions and arrangements alike. This means that it is not often easy to identify the actual focus of the argument in question. An article might be aimed, say, at the role played by individuals in the shaping of new forms of action, but the discussion will actually deal with the effect of postmodern changes on people’s lives. It is

111

thus not always easy to tell whether writers are saying something about the premises of social change or about its consequences. Secondly, the arguments usually indicate that the effects of modern social structures are weakening in one way or another, but they do not indicate what exactly the nature of current social life is. The arguments often fail to provide convincing interpretations of what will follow the current postmodern transition in terms of new social structures, with their concurrent opportunities and limitations. To put it another way, postmodernism has not come up with any new descriptions of what contemporary social arrangements are; it merely says what they are not (Pescasolido & Rubin 2000: 60). Thirdly, it is not always clear whether postmodern writers are presenting their arguments as general theories of change or merely as diagnostic narratives. This may place critics in an awkward position, since the same criticism cannot be levelled at all types of social theorising (Savage 2000: 104-105). It can be argued, for example, that some kinds of theorising do not have to seek empirical support.11 These weaknesses naturally also concern the arguments that have been presented regarding the changing character of consumer behaviour. Let us take for instance the common claim that socio-economic factors are very poor predictors of individual consumption patterns. Many writers have argued strongly for this idea during last few years (see Bocock 1993: 8081; Featherstone 1991; Beck 1992; Noro 1995; Sulkunen 1992). It is not always clear, however, whether they mean that the connection is now relatively insignificant compared to earlier times or that there is now no connection at all. Ulrich Beck (1992: 102), for example, stresses in one part of his book that class membership can still be very evident in everyday consumption and that class will remain an object of consciousness and identification also in the future. He says that the concept of lifestyle, which has become highly inclusive, is ultimately a category based on class memberships. Only a few pages earlier, however, Beck claims just the opposite, carrying the basic idea of postmodernism almost to an extreme by arguing that Western societies are undergoing a surge of individuation; this is characterised as a process in the course of which “people will be set free from the social forms of industrial society – class, stratification, family, gender status” (Beck 1992: 87). Thus it is not always easy to test particular ideas empirically in an appropriate way.

112

Still, it is possible to ask questions about the certain general ideas behind postmodern notions. For instance, all the arguments presented challenge the significance of modern social structures in one way or another. The argument which is common to these concepts of postmodern, late-modern and reflexive modern society questions normal life histories and the structural conditions determining them (see Beck et. al. 2003: 4-5; Harvey 1990; Beck 1992; Lash 1995; Giddens 1991). The relatively closed social milieus and life-worlds based on classes and other structures are assumed to be giving way to individual choices. Against this new background, it is no longer possible to predict life patterns from a framework of stable conditions. In other words, individual life practices have become unpredictable and can be represented only individually. This notion leads us to several new questions. For example, if there exist considerable connections between socio-demographics and consumption behaviour, then it becomes possible to evaluate postmodern ideas in a new light. Secondly, while many writings are concerned with assumptions that exclude the consideration of empirical evidence, they still aim at offering a vision of contemporary times or a possible future. This allows us to consider whether there is cause to believe that current life may actually be leading to the condition depicted in postmodern writings. It is thus reasonable to assume that at least some aspects of the individualistic critique can be addressed empirically. 3.6.2 Approaching consumption from a structural perspective When certain consumer goods are distributed in the general market, there are structural conditions that determine the rate of their adoption among different population groups. The ‘diffusion process’ (Fidler 1997, 12) whereby things are adopted within a society naturally depends on many factors, but economic and cultural resources can be understood to be among the important ones. In this sense, it is possible to identify different categories of consumers, such as early adopters, the majority, and late adopters. Such categorisations are typically referred to in analysing the distribution of technological innovations in society (e.g. Rogers 1983; Raijas 2001). In general, the model assumes that those who have better resources perceive the relative advantages of certain items first; thus there are usually strong status distinctions when an innovation is first introduced.

113

Later, as the new products or services become more commonplace, the links between social or cultural resources and consumer behaviour begin to diminish. However, it has also been noted that the proliferation of modern consumer goods does not necessary weaken the connections between consumption behaviour and social background (e.g. Bourdieu 1984; Toivonen 1997; Kraaykamp 2002). Certain patterns of consumption may remain distinctive for certain groups of people for relatively long periods of time. Such culturally determined differences, which seem to reproduce over time, may be related to many activities. Thus an examination of the differences between class positions and other socio-demographic factors can be seen as relevant, despite the recent theoretical assumptions concerning the blurring of structures which would seem to argue the opposite. In the analysis of consumer behaviour, we can assume that there are always certain underlying principles which both constrain and enable many activities. These principles are usually structurally determined and become visible through various economic, social and cultural characteristics. Many factors, ranging from disposable income to educational qualifications, may therefore lead to systematic differences among people. In other words, certain structural conditions for consumer behaviour can still be pointed out. The need to pay money for things to be consumed is surely a very prominent fact. While continuous consumption can be regarded as normal life, without sufficient money buying things and services is just as impossible as it always has been. Zygmunt Bauman (1998) has argued that certain social groups have come into being which are permanently displaced from society because they cannot engage in the role of consumers. The conditions of defective consumption are not characterised merely as lack of money for the necessary payments. An individual’s inadequacy as a consumer prevents long-term life planning, because without money there are few activities and possibilities available in present-day society. Being poor in the consumer society means simply being ‘a flawed consumer’ (Bauman 1998: 38). From this perspective, it is of course naïve to postulate that someone from the lowest social group might have the same consumption pattern as someone in the highest income bracket.

114

But not even money is merely a matter of a constraining structure. As Bourdieu (1984: 374) puts it, “having a million does not in itself make one able to live like a millionaire”. In other words, it generally takes a long time for people to learn new expenditure patterns. Gender roles and class roots, for instance, continue to affect behavioural patterns even under new and changing social conditions. In this respect it can be suggested that while wealth probably creates the most central and primary conditions, there are other factors as well that affect individual consumption behaviour. In this chapter, the mechanisms in particular of income, social class, gender, age, type of household and place of residence were discussed. The possible effects of these factors were evaluated, drawing on the idea of explanatory processes that can explicate the source of the relationships observed between consumption behaviour and the assumed explanatory factor. By describing these social mechanisms I have aimed at offering a theoretical account of the links between independent and dependent variables in typical empirical designs. All of the factors examined, which were offered a plausible explanatory mechanism, can be understood as having either a constraining or an enabling effect on individual consumer behaviour. In the case of income, however, the effect is more likely to result from the restriction of resources. The difference between a constraining and an enabling effect lies in the fact that some factors – such as lack of money – make certain activities – such as shopping – simply impossible. On the other hand, certain other factors, such as free time, can be understood to make the same activities possible. Distinct cultural practices, for instance, can offer typical leisure activities for the upper and middle classes. Participation in these practices by the working classes, however, may be extremely rare. This means that differences can be observed in various consumption patterns, and that these differences can be attributed for instance to the mechanisms of cultural condition or class identity. The purpose of this chapter has been to demonstrate that the basic idea of mechanism-based assumptions implies that economic and sociodemographic factors can explain consumer behaviour – depending of course on the type of consumption to be explained. The empirical part of this study focuses on the question whether economic and sociodemographic factors affect consumption patterns at the household and

115

individual level, including an evaluation of which factors appear to be the best determinants in different fields of consumption. In addition, the study examines the degree to which the effects of the explanatory factors selected can be interpreted on the basis of similar social mechanisms. In particular consumption in the fields of personal goods, everyday life and leisure activities are analysed empirically. The next chapter takes a closer look at the empirical research design. The research questions will be enumerated following the structural framework outlined in this chapter. In order to cover the levels of both realised consumption and consumption desires, consumer behaviour will be approached through two different data sets. These data sets, the choice of explanatory variables, and the methods of analysis applied are discussed in the next chapter.

Notes 1

The theories of rational choice can be seen as an exception in this respect. Explanations based

on rational choice have become popular in current theoretical and empirical sociology (Zafirovski 1999: 47-48). In particular John Goldthorpe has argued strongly for explanations based on rational choice in empirical research (e.g. Goldthorpe 2000). In general, rational choice approaches are based on rather idealised assumptions of the rationality and utilitarian nature of human activities. These approaches attempt to explain individual behaviour on the basis of stable preferences. Thus the starting point is very different from traditional sociological approaches, where the basic underpinning of the analysis is in social structures, norms and institutions. It should be noted that while all approaches based on rational choice share certain features, such as the assumption of the actor’s rationality or methodological individualism, they also differ in many ways. Nevertheless, the assumption of a certain static foundation to which individual choices refer is essential in any specification of rational choice. 2

On this scale the first adult of the household is given a weight of 1.0 and all other adults a

weight of 0.7. Children get a weight of 0.5. Basically, the OECD consumption unit scale implies that the first adult may use 100 percent of the budget, while each other adult requires an additional 70 percent and each child 50 percent to the budget (OECD 1982). The income of a household consisting of two parents and two children, for example, would accordingly be divided by 2.7.

116

3

The Engel curve describes expenditure on basic needs, such as food or housing, as a

marginally concave decreasing function of household income. With increasing income, the share of this expenditure decreases. This function is based on the distinction between necessary goods, inferior goods and luxury goods. Necessities take up a large part of the budget of the lower income groups but a smaller part of the budget of higher income groups. The curve and the concept are named after the economist Ernst Engel, and are sometimes referred to as ‘Engel’s Law’ (see Antonides & van Raaij 1998; Fine & Leopold 1993). 4

According to David Rose and David Pevalin and their colleagues (2001), empirically-oriented

social scientists can be divided between researchers who prefer continuous measures and those who favour categorical approaches to socio-economic classification. Some researchers prefer measures that are based on indexes with occupational, educational and income measures. In this way, an individual’s socio-economic position becomes understood through combined information on his or her occupation, education and income. Others, on the other hand, favour to divide the population into a discrete number of categories or social positions. Only categorical approaches to socio-economic classification are discussed here, because in this way it is possible to discuss the effects of occupational position to consumption, without its connection to income. See Alexander Stewart et al. (1980), for example, for comprehensive presentation of the continuous measurement of occupational positions. 5

See Rose & Pevalin et. al. (2001) or Erikson & Goldthorpe (1992) for a concrete account of

the development of comparable occupational classifications. 6

There are naturally various alternative methods for studying subjective class experience.

Qualitative research techniques, such as in-depth interviews, offer many possibilities for asking people what kind of images they have about their own class position and about class structure in general. However, lengthy interviews can only be effectively applied in the study of small groups of people (e.g. Newman & Benz 1998; Luomanen & Räsänen 2000; Räsänen 2001). In addition, in qualitative research designs it is often difficult to conduct comparative analyses which control for the effects of other possible independent variables. For these reasons, research strategies that use unstructured data will not be explored here. 7

The effect of age is often a problematic issue in empirical analyses of consumption or any

other activity. Sometimes the differences observed in age-bound behaviour might be explained as period and cohort differences rather than as age differences (e.g. Riley 1973: 35-36; Siurala 1987, 10; Lunt & Livingstone 1992). It is most likely, however, that age and cohort effects

117

occur simultaneously. Thus a systematic analysis that takes into account both variations in lifecycle patterns and historical correlations is very difficult. 8

Wirth’s essay, Urbanism as a Way of Life (Wirth 1938), and its influence on urban sociology

should be stressed in this respect. Wirth’s essay is often claimed to be one of the most influential sociological articles ever written (e.g. Savage & Warde 1993: 97). In the article, cities are seen as having distinct social effects on people’s behaviour. What Wirth was basically claiming was that in the new cities a new urban personality had emerged. This distinct type of personality could be distinguished in terms of competition, aggrandisement, mutual exploitation and separation from nature in particular. Thus Wirth can be said to be arguing that the cities themselves were actively generating a new way of life (Dickens 1990: 45; 48). 9

It is sometimes stressed that the term ‘lifestyle’ has two different meanings: it may refer to

particular practices or to systems of different elements. Particularly in the Finnish literature, ‘style of life’ (elämäntyyli) and ‘way of life’ (elämäntapa) have been distinguished (see Roos 1986: 38; Eskola 1985); this analytical distinction, however, is not found elsewhere (see Miles 1998: 36-37; Maffesoli 1996: 127-128; Mitchell 1983). 10

It should be noted that most of these typologies have been developed for marketing and other

business purposes. They are usually developed through psychometrics, involving factor analysis of various attitudes, psychological characteristics, opinions and certain selected activities into clusters defining different segments of the population. Some typologies utilising information on social activities have also been developed. In general, however, it has been pointed out that most lifestyle typologies reveal rather little about individual behaviour outside the given marketplace and the purchase activities examined. Most researchers consider that typologies are usually based on relatively arbitrary and unstable categorisations (e.g. Puohiniemi 1991: 50; Antonides & van Raaij 1998: 380). 11

An interesting comment related to this issue can be made. A strange article was published in

Sosiologia, the Finnish sociology journal, in the late 2000 (Noro 2000). In the article, Zeitdiagnose was proposed as a third possible genre of sociological theory, in addition to research theory and general theory. It was presented that this Zeitdiagnose can be evaluated in the light of its sociological components, but it cannot be totally incorporated into the scientific debate. Thus a diagnosis (unlike the typical research theory) cannot be used as a theory for the interpretation of empirical evidence. While this is a valid point related to differences in the nature of scientific argument, it is also somewhat problematic as a normative argument. The

118

article may actually have been designed to defend the kind of social theorising that cannot be tested empirically. The problem arises particularly because some of the ideas put forward by Ulrich Beck, Anthony Giddens and Zygmunt Bauman were named as examples of this genre. In addition, Noro’s article says nothing about whether is it acceptable to label all writings that do not present relatively unambiguous statements as Zeitdiagnose. Similar comments on Giddens’ writings have also been presented earlier (see Gregson 1991).

119

4 EMPIRICAL STUDY: DATA AND METHODS

In this chapter I outline the empirical research design of the study. I first summarise the research questions on the basis of what has been said about the role of social structures in everyday life. In general, the aim of the study is to sum up the key differences between structural and individualistic approaches, the latter representing the postmodern critique of the former. I also discuss some of the problems involved in the empirical examination of social change by means of cross-sectional data. After this, I examine the empirical data and their validity. As already mentioned, the empirical part of this study consists of the analysis of two different data sets. The data sets and their utilisation are described in this chapter. Basically, what is presented here is the selection of background variables; dependent measures are discussed in Chapters 5 and 6. Problems related to the methodological operationalisation of sociological concepts are not dealt with here, since these questions were examined at an analytical level in the previous chapter. Rather, I here discuss how explanatory factors are measured in the data as operationalised categories. The empirical testing procedure is described at the end of the chapter. In addition to a short presentation of statistical methods, the chapter also discusses questions related to the comparison of the data sets utilised. 4.1 Empirical research design In this study, the characteristics of contemporary consumer behaviour are approached from a structural perspective. The general question to be answered is whether traditional structural determinants are important in the interpretation of people’s consumption patterns. It should be acknowledged, however, that it is not easy to test the validity of the individual ideas that have been presented in various culturally oriented approaches to consumption (see Woodward et. al. 2000: 340-342; Wilska 1999; Räsänen 2000). Several reasons for that exist. First of all, it is often impossible to draw testable hypotheses from contemporary writings. As noted at the end of the previous chapter, this is 121

the case mainly because of the nature of postmodern social analysis. Most culturally oriented studies are not primarily concerned with the social determination of consumption. Their focus is not on the differences in attitudes or tastes between various groups in society (e.g. Gronow 1997: ixx; Boden & Williams; see also Bauman 1997). Also, it is often relatively unclear what the core arguments within theoretical treatments actually are. For example, frequent reference is made to the idea that we are living in ‘new times’ or in an ‘interregnum’ (Smart 1997: 399), but it is not easy to say what is actually meant by this. One reason for this vagueness may be that the philosophical background of postmodernism is in the conscious effort to resist rigorous empirical testing. In other words, postmodernists try to avoid the ‘positivist’ doctrine of science which upholds the primacy of empirical evidence as the basis of knowledge (e.g. Etürk 1999; Pescasolido & Rubin 2000). It follows that many writers accept theoretical ideas as true instead of testing them. Secondly, many writers identified with postmodernism treat their issues in a relatively unsystematic and paradoxical way. One may argue, for instance, that the current social condition is something which comes after the modern, but at the same time it is undoubtedly part of the modern. This also means that it is difficult to draw testable set of hypotheses from a certain kind of theoretical discussion. While certain ideas related to postmodernism and consumption will be addressed in this study, the contributions of individual writers will not be evaluated as such. Research questions will be drawn generally from the body of individualistic notions that aim at describing the nature of people’s consumption behaviour. According to these notions, consumption has become heavily influenced by individualistic and personal meanings. Some postmodern notions also include assumptions about new types of social linkage, not connected to socio-demographic structures. It was concluded in Chapter 2 that the quantity and variety of different interpretations is enormous. In a broad sense, however, we can argue that individualistic and other postmodern conceptions include the notion that the collective expression of class and other earlier socio-demographic variables has become more or less irrelevant in the analysis of individuals’ consumer behaviour. In other words, it is argued that the activities of the average consuming individual are not determined or constrained by

122

traditional social and cultural factors. The use of earlier socio-demographic models is therefore inadequate for understanding or predicting phenomena related to consumer behaviour. The key to understanding and explaining consumption is considered to lie in various personal and situational factors. For example such terms as mindset, need state, viewpoint and more general conceptions of lifestyle are often mentioned as factors explaining the consumer behaviour of individuals (e.g. Spencer & Wells 2000: 235; Falk & Campbell 1997; Miles 1998). It is believed that several types of fragmented groupings may exist in society, but that they are not based on traditional criteria such as class position. There are many definitions of who consumers are, how they behave and what drives them. A consumer may be regarded for instance as an identity-seeker trying to find a real self in the objects which he or she consumes, or an explorer looking for new experiences (e.g. Campbell 1987: 77-78; 90-91; Bocock 1993: 80-81; Ilmonen 1993). In many portrayals consumers are described in general terms as either being interested in getting value for money or as hedonistic pleasure-seekers. This means that people are regarded as free problem-solvers; the problem they are mainly faced with is what to consume and where. At the same time, however, each particular attempt to frame the consumer fails to capture certain features that have been stressed in other views (Gabriel & Lang 1995: 188-189). In this sense, most contemporary images of consumers are both restricted and ill-defined. In real life, consumers are not uniform entities and their behaviour can be inconsistent and even contradictory. In general, however, while traditional structural perspectives assume that there exist both constraining and enabling social structures that influence individual thinking and behaviour, individualistic notions claim that one’s thought and behaviour are to large degree a matter of choice. This point can be seen as the key difference distinguishing (traditional) structural and (postmodern) individualistic ways of thinking. It follows from individualistic ideas that contemporary consumption patterns cannot be explained by socio-demographic structures. It is assumed that the existing differences in the consumer market are highly fragmented and cannot be dealt with by structural models. In other words, there are no clearly recognisable structural differences in terms of

123

traditional cultural or demographic factors. It is reasonable to question whether this view will be verified by empirical analysis. In particular the examination of various types of consumption activity and the evaluation of possible differences between them is relevant. The empirical research questions of this study are outlined in the next section. 4.1.1 Research questions The first empirical objective of this study is to identify the structural dimensions of contemporary consumer behaviour by analysing some of the most important consumption activities of everyday life. These activities will be selected on the basis of their availability in the data sets utilised. It follows that some key categories of daily consumption are not included in the analysis. This question and its implications for the interpretation of the results are discussed further in Chapters 5 and 6. The second objective is to determine the socio-demographic factors that can help us to understand the observed differences most effectively. People’s consumption activities are examined by exploring possible relationships between consumption and demographic factors, and by determining the strength of these relationships. Social class is treated as the factor of key interest throughout the analysis. This is because it has been argued in contemporary theoretical writings that in particular the effect of class has become insignificant in explaining consumer behaviour. Class has also been one of the most important theoretical concepts and background variables in empirical studies in sociology. It is generally argued in the sociology of consumption that consumption should be regarded as an almost ubiquitous process in our society. It incorporates many things besides the actual spending on goods and services. On the one hand it is a matter of satisfying basic needs and of gratifying various desires. On the other hand it is also a form of communication, whereby signals concerning wealth and social status can be expressed. Thus not only consumption expenditure itself but also the subjective experiences related to consumer behaviour can be seen as providing an interesting area for empirical exploration. It can be argued that consumption includes many other things besides the actual spending on goods and purchases. For example, one characteristic of consumption can be approached by measuring the individual’s willingness to spend on

124

certain activities. Such subjective aspects allow us to draw conclusions as to how people perceive their own positions as consumers. Consumption is therefore measured in this study by means of both subjective experiences and expenditure. This dual approach is expected to lead to a better understanding of the characteristics of contemporary consumer behaviour. The empirical objectives of the study can be divided into two sections. The first objective is to analyse the dimensions of consumption, the second to analyse the socio-demographic features of consumption. On the basis of these tasks, a total of five research questions can be formulated: A. Dimensions of consumption 1) Are there systematic associations between the consumption activities selected, such that they can be approached as distinct dimensions? 2) What kind of structural similarities and differences can be observed when different types of consumption measures in the data sets are dimensioned? B. Socio-demographic features of consumption 3) Do patterns of consumption differ significantly according to sociodemographic group? 4) What type of consumption can be explained most effectively by sociodemographic factors? 5) What kind of socio-demographic similarities and differences are found employing various consumption measures?

The dimensions of consumption are approached by analysing a variety of different consumption measures. The analysis focuses on some of the most prominent questions in the current debate on structural and individualistic features of consumption. In order to analyse many different consumption activities, one question asked concerns the possible presence in the measures of certain general associations that can be described as consumption dimensions. The first task of the empirical investigation is thus to explore possible correlations between the consumption measures chosen, and the differentiation of the main dimensions of the existing correlations. The similarities and differences between the associations of

125

different data sets are also evaluated. The basic subject thus is the degree to which different types of measures provide different results. The second task is to explore the variation in consumption patterns between different social groups. The aim is to provide new information on the determination of individuals’ consumer behaviour. I first ask whether socio-demographic factors can function as significant predictors of consumption behaviour. Secondly, I determine which factors provide the best means of distinguishing between patterns of consumption. This is carried out by comparing the effect of the independent variables on the dependent measures. The final question concerns the socio-demographic differences between different types of consumption measures. The five questions listed above will be examined by using crosssectional data. Briefly, this means that the data have been collected at approximately the same time. The data include information about people with different economic and social backgrounds. Can we account for different consumption patterns from a structural perspective, and what type of consumption activities can most feasibly be attributed to structural factors? Thus the purpose is to evaluate the tenability of certain general postmodern arguments. Consumer behaviour is approached primarily in a framework of daily life. This means that certain basic consumption activities are used as dependent measures, while basic socio-demographic measures are used as independent measures. Before going into the selection of independent background variables, however, let me briefly discuss the difficulty of interpreting social change by means of a cross-sectional research design. 4.1.2 Social change in cross-sectional studies It is well known that researchers who wish to understand the dynamics underlying processes of social change are often constrained by the limits of cross-sectional empirical data. One of the ways most commonly relied on to overcome this problem is to use repeated cross-sectional data sets, which in many countries are provided by official statistics. In Finland, for example, Statistics Finland has collected Expenditure Surveys regularly since the late 1960s (see Tilastokeskus 2001: 5). Official statistics, however, are not necessarily suitable for describing the kind of phenomena that the researcher would like to study. The Finnish

126

Expenditure Surveys, for example, do not include any information about subjective consumption experience. It has also been argued that – while such data are eminently suited to portraying structural transformations from one point in time to another – they do not actually permit researchers to study the causal mechanisms underlying these processes (Djurfeldt & Gooch 2002: 75). It may well be, for example, that the differences observed are due to the characteristics of the respondents and not to time-related factors. In other words, the results cannot necessarily be explained in a satisfactory way. It has therefore been suggested that explaining social change is possible only by using longitudinal (panel) data from the same respondents. Longitudinal research can be defined as a type of research in which data are collected for each item from at least two distinct periods in time, and in which the subjects or cases analysed are the same or at least comparable from one period to the next (Menard 1991: 4; Blossfeld & Rohwer 2002). With such a design, it can be argued that the effects of respondents’ characteristics are quite effectively controlled. There may of course remain other factors which affect the results, and statistical controls for other independent variables are often the only way to control alternative explanations (e.g. Achen 1986: 12-13; Stevens 1992). Thus causal explanation for any social phenomenon, at least in terms of ultimate causes, is almost impossible, regardless of the kind of data available.1 The methodological difficulties of longitudinal analysis have been well documented in the methodological literature (see Biljeveld et. al. 1998; Menard 1991), and I do not discuss them more closely here. The empirical part of this study is in any case based purely on cross-sectional data analysis. Naturally, a number of difficulties related to cross-sectional research designs have been recognised in the literature (see Babbie 1989; Biljeveld et. al. 1998). In particular the two key restrictions of cross-sectional designs have to be taken into account. First, by definition cross-sectional research designs do not involve comparative data analysis between different periods of time; the data are collected at exactly or at least approximately the same time. In other words, there is no follow-up of subjects or cases over time. It necessarily follows that comparisons over time are not possible. The objective of cross-sectional studies is therefore limited to describing a

127

research population at a particular time. Secondly, these studies are highly sensitive to information bias due to the effect of that particular point of time. This problem is often controlled by including in the data subjects of different age. This, however, does not eliminate all possible biases. The weakness is that the subjects, in addition to being different ages, were also born in different years, and their behavioural patterns may thus be influenced by generational effects rather than by age and other sociocultural differences. In addition, the time of research may be exceptional in many ways; this should be explicated when generalising from the results. Now, it is important to be clear that we cannot conclude very much about the nature of change on the basis of our data. This means that to answer the question for instance of the extent to which the importance of class has actually weakened over time in a particular society would require a different research design. On the other hand, interpretations from our data have to be highly time-restricted as well. For example, we cannot argue anything about changes in consumption patterns in the early 21st century. What we can do with cross-sectional data, however, is to address the question whether certain ideas proposed as to the nature of contemporary society seem reasonable or not. Possible similarities between different types of measures provide grounds for more accurate claims as to the nature of the current situation. We might also be able to argue something about future trends in individual consumption patterns. This is of course only possible on the basis of theoretical discussion, not directly on the basis of the data analysis. In any case, it can be assumed that exploring the basic characteristics of consumption patterns at a given time can be of substantial importance in interpreting the nature of contemporary patterns of consumption. The key methodological issues of this study are naturally the selection of a representative sample and of reliable measures, and controlling as far as possible the effects of confounding variables. The purpose of the empirical part of this study is to give a coherent picture of the patterns of selected consumption activities in contemporary Finland. Towards this end, consumption behaviour is approached from more than one angle. I analyse both consumption expenditure and individuals’ willingness to increase their consumption. This will arguably allow more effective evaluation of the validity of theoretical ideas than by using just one type of measures. The study utilises one wave of the Finnish

128

Expenditure Survey and a nation-wide postal survey. In the next section, these data sets are briefly described. 4.2 Data sets and their validity The data sets utilised in this study consist of the Finnish Expenditure Survey for 1998 (FES98) and the Finland 1999 Survey (FIN99), a nationwide postal survey conducted in 1999. These two sets of data provide the means to approach the nature of contemporary consumption patterns from the angle of both individuals and households. The data sets, however, approach consumption rather differently. In the FES98 data consumption is measured as household expenditure. The variables in the data report amounts of household spending. The data cover all the acquisition of goods and services during the time of the research. In these data, consumption is therefore defined simply as expenditure. They thus allow the analysis of actual consumption that has occurred in households. The selection of indicators is described in Chapter 5. In the FIN99 data, on the other hand, consumption is reported in the form of subjective evaluations. Basically, the data consist of the respondents’ evaluations of how much they would increase different types of consumption activities if they were able to do so. This represents analysis beyond actualised activity – the analysis of consumption at the imaginary level. Consumption is interpreted in terms of the desire to increase one’s current consumption. These measures are described more closely in Chapter 6. The FES98 survey is one wave in the data which have been collected by Statistics Finland approximately every five years since 1966. It should be noted, however, that these expenditure surveys do not represent longitudinal (panel) data. The sample size in 1998 was 6870 households and the response rate was approximately 63 percent (N = 4359). The sampling frame in the data consisted of persons over 15 years of age who had a permanent home address according to the Central Register of Population. Most of the data were collected in face-to-face interviews and by the keeping of diaries of household expenditure over two-week periods. In addition, various official sources, such as the tax registers and the Social Insurance Institution, provided information concerning income, pension and insurance. Household expenditure is reported in the data in the form of

129

annual estimates. However, in particular households in the lowest income groups and single households are underrepresented in the data. Thus it is necessary to use a weight coefficient in the statistical analyses of this data to weight the sample. Different household types are assigned different weight coefficients, to correct the bias in the sample.2 A detailed description of data collection techniques, the sample of respondents and the representativeness of data is given elsewhere (see Tilastokeskus 2001). The FIN99 survey was carried out at the Department of Sociology, University of Turku, during spring and summer 1999. The data were collected for the research projects of Turku Center for Welfare Research (TCWR). The data represent the total Finnish-speaking population 18-74 years of age. Swedish-speakers, the second official language group in Finland, are not represented in the data.3 The questionnaires were sent to 4001 people who according to the Central Register of Population were resident in Finland. A total of 43 people in the original random sample were not reached. Three reminders were sent to those who had not answered. The second reminder came with a new question form. The effective response rate was approximately 61 percent (N = 2417). There was no systematic bias in the response and no weights were applied in the analysis of the data. A detailed description of data collection, sampling procedure, and the representativeness of the sample is available elsewhere (see Erola & Räsänen 2000). The research questionnaire is given in Appendix A. These two data sets are used here to determine to what degree the sociological background variables selected explain patterns of consumption at both the household and the individual level. In other words, the question is whether economic and socio-demographic information can help us to understand people’s consumer patterns. The first data set (FES98) makes it possible to study consumption expenditure at the household level; the second data set (FIN99), makes it possible to study people’s consumption desires at the individual level. Consumer behaviour is approached through relatively different activities represented in the data sets. It can be argued that the measurement of consumption as expenditure follows a more clearly traditional and empirical form of consumption study, and represents a more objective indicator of actual consumption than individuals’ subjective evaluations.

130

However, since consumption is approached in this study by means of both expenditure and subjective experiences, perhaps a more coherent and complete picture emerges of the nature of individuals’ consumption behaviour. In addition, combining the analysis at the individual and household level can reveal whether certain patterns of consumption occur at both levels of analysis. The selection of independent measures from the data sets is presented in the following two sections. 4.2.1 Selection of explanatory variables from the Finnish Expenditure Survey 1998 FES98 survey measures consumption in terms of household expenditure. It follows that if we wish to approach the effect on expenditure for instance of class or gender, we have considerable problems because class position and gender are actually individual measures. As shown in Chapter 3, the methods available for defining household-level measures for individuallevel concepts are somewhat inadequate. Basically, we need measures that express the categories of only one individual in the household. We then have to assume, however, that the reference person can influence decisions as to money allocation more than the other members of the household. It is of course by no means always the case that the same person in the household has the most to say over each decision, but at least sometimes this may be a correct assumption. In any case, for some measures we have to rely on the characteristics of the reference person. In this sense it must be accepted that the available data set restricts us in some ways. In general terms, similar background variables were selected from this data that could also be selected from the FIN99 data. However, the differences of the measures on the household level were kept in mind. It should also be noted that only the general characteristics of the selected variables are discussed here. This means that only modifications that were made for the original variables are described; no detailed characterisations are given of the original variables. Detailed information on the creation and composition of the background variables can be found elsewhere (see Tilastokeskus 2001: 10-17). The following five variables were selected as independent background variables:

131

1) Socio-economic status of head of household; 2) OECD-equalised household income; 3) Type of household; 4) Gender of head of household; 5) Place of residence.

The socio-economic status of the head of household (HOH) is measured as the socio-economic position of the reference person of the household. In practice, this variable indicates the reference person’s occupational position. The reference person is defined as the person who at the end of 1997 had the highest personal earnings in the household during the preceding 12 months. The principles according to which different specific occupations are classified into more general occupational positions can be found in the Statistics Finland manual (see Tilastokeskus 1989). Since education is one of the criteria used in the socio-economic classification of occupations, there is in practice no need to control its effect in the analysis. In addition, our interest is in class effects in general. Originally, there were nine socio-economic categories in the data for the occupational classification: 1) farmer, 2) other entrepreneur, 3) manager or professional, 4) intermediate non-manual employee, 5) manual worker, 6) student, 7) pensioners and elderly, 8) long-term unemployed, and 9) other economically inactive. Some categories of the socio-economic status of HOH were combined in order to make socio-economic status suitable for the measurement of social class positions. However, none of the categories were excluded from the further analysis. Pensioners and elderly, long-term unemployed, other economically inactive and students were placed in a category labelled ‘other households’. Thus six categories remained. The category of manual workers is used as a reference group. It can be assumed that this kind of rough socio-economic categorisation will capture some of the aspects related to the educational dimensions of different class positions, such as white-collar and blue-collar jobs. The socio-economic status classification of Statistics Finland, however, is designed primarily for sorting occupations according to type of work and the level of expertise required in different occupations. Income measurement is based on the disposable incomes of all members of the household. It reports income from work plus property

132

income and income transfers, minus transfer payments from 1997. The information is obtained from the tax registers, and the variable is calculated using the OECD consumption unit equivalence scale. As described in Chapter 3, the OECD scale uses different weights in order to make households of different size equivalent for comparisons. The original measures were in Finnish marks, but marks were converted to euros using the official exchange rate (1 € = 5,94573 FIM).4 The income variable was also categorised in order to make the presentation of the results consistent with the other independent variables. A total of six categories for disposable household income appear in the analyses. The lowest group consists of households with an income of less than 6500 € a year, the highest of those with over 20 000 € a year. The median income in the data is around 10 500 € a year. Thus the lowest group consists of households with approximately 60 percent or less of the median income level. The highest income category is treated as a reference group. The type of household is treated in the analysis of this data as a rough measure of the ‘age’ of the household. This variable was preferred over the age of the reference person simply because it provides more information on the composition of households. The type of household is defined in terms of the number and age of the persons in the household. The original composition of the variable consisted of six categories: 1) single person, 2) couple without children, 3) single parent with children, 4) couple with children, 5) elderly household, and 6) other type of household. An elderly household consists of persons over 65 years of age, living either alone or together. The category of other households includes for example such households as student communes and other mixed types of household. The category of single person is used as a reference, but otherwise no modifications were made. Gender of the head of household (HOH) simply reports the reference person’s sex. The reference person is again the person with the highest personal income in the household. Use of this variable has its difficulties when there is more than one adult living in the household. In spite of this, gender of HOH can be assumed to have an great impact on certain types of consumption expenditure, and should therefore be included in the analysis. The category of households headed by women provides the reference group for the statistical analyses.

133

Place of residence is the final background variable selected from the data. It is measured by the population area variable, which is defined by the population density and size of the municipality (see Tilastokeskus 2001: 13-14). The variable is based on official register information that was available at the time of research. Basically, the variable can be seen as measuring the degree of urbanisation of the households’ location. The variable has three categories: 1) rural, 2) densely populated and 3) urban. The category of rural is used as a reference. 4.2.2 Selection of explanatory variables from the Finland 1999 Survey The FIN99 Survey measures subjective consumption experience at the individual level. The unit of analysis for most of the background variables is also at the individual level. Thus there is no problem of operationalising certain sociological conceptions, such as age or gender. This means that most of the concepts discussed earlier can be defined rather painlessly by means of the variables in the data. The following six variables were selected as independent background variables: 1) Class identity; 2) Vocational education; 3) OECD-equalised household income; 4) Gender; 5) Age; 6) Place of residence.

Class identity is treated in the analysis of this data as the respondent’s class position. While it is true that class identity is a purely subjective measure, it gives relevant information on the respondent’s social rank. It was measured in the questionnaire simply by asking which social class the respondents identified with.5 There were originally six response options to choose from: 1) upper class, 2) upper middle class, 3) lower middle class, 4) working class, 5) none of these classes, and 6) other (with an empty line for filling in). The answers giving the sixth option (N = 78), however, were excluded from the analysis because it would have been impossible to create a label combining the answers for that option. Moreover, the sixth option would have produced obvious difficulties for the further analysis in the

134

sense of identification with the basic model of class structure. The other five categories were included in the classification; working class was used as a reference group. This kind of hierarchical classification can be treated relatively similarly to the socio-economic classification presented above. Vocational education is used as a control variable for class identity. It was selected because there was no way to construct a reliable occupational classification from the data. Vocational education was measured by means of seven options: 1) unskilled, 2) vocational school or course, 3) intermediate level, 4) higher intermediate level degree, 5) academic degree, 6) post-graduate academic degree, and 7) other (with an empty line for filling in). Subjective reporting offered the only possibility of measuring the respondents’ educational background. Some modifications were made in order to make the original variables more suitable. First, the answers giving the seventh option (N = 177) were excluded from the analysis. This was done because it would have been impossible to label this category unambiguously. Secondly, categories three and four and categories five and six were combined. This resulted in four categories for the classification of vocational education. It is reasonable to assume that this variable can still differentiate the respondents’ educational background both by skill level and by qualification. The category of academic degree is used as a reference group. Income is measured as the OECD-equalised income of the household. Rather than the official tax registers, subjective report was used as income indicators. Respondents recorded their household’s disposable income as an answer to an open-ended question (in FIM per month). Finnish marks were converted to euros using the official exchange rate. A household-level measure was preferred over an individual measure because it is reasonable to assume that the economic situations of other adult household members are important in determining one’s economic position. Household income was also categorised. A total of five income groups equalised by consumption-unit were created. The lowest group consisted of cases that had less than 450 € a month, the highest of those with more than 1650 € a month. The lowest group, however, was not calculated on the basis of 60 percent of the median income because there would have been more than 400 cases below that point. The median income in the data was around 1000 € a month. Thus the lowest income group consists of respondents who

135

live in households that actually earn less than 50 percent of the median income level. Again, the highest income group is used as a reference in the analyses. The respondent’s gender and age are based on subjective reporting. Female respondents represent a reference for gender in the analyses. Age was specified in the questionnaire as the year of birth; thus it provided a continuous measure. However, age was categorised into four groups in order to allow parallel comparisons with the other independent measures. The first two age groups were 18-30 and 31-45 years. Respondents aged 45-60 years were placed in the third group, and those over 60 in the fourth group. This categorisation can be seen as reflecting a broad classification of the phases distinguished in adult life. The years from 18 to 30 are often referred to as early adulthood. The next two age groups can be defined as early and late middle age. Finally, people over 60 years of age are often characterised as the elderly because this is usually a period entitling the person to a pension. The oldest age group is treated as a reference group. The last background variable selected is place of residence. This was measured simply by asking the respondents to choose their type of residential area. The questionnaire offered just two options to choose from: 1) urban and 2) rural. While a dichotomised option might be seen as a somewhat coarse method of measurement, it can also be understood to provide a relatively reliable measure; it is probably more difficult for the average respondent to rate his or her residential area according to several different options, based for instance on the classifications of the municipal registers. On the other hand, it is presumably fairly easy to say whether one’s residence is located either in an urban setting or in a rural setting. In this sense this measure should be considered as an unambiguous measure. The option rural is used as a reference category. 4.3 Comparability of data sets and methods of analysis The purpose of the empirical analysis is to provide a multidimensional look at people’s consumer behaviour. This is carried out by studying both the subjective experience of individuals and the expenditure of households. In general, similar procedures are followed in the analysis of both datasets. Despite this, data from the FIN99 survey and from the FES98 should not be treated as comparable. The first problem is of course the different levels of

136

analysis. The first measures consumption at the household level, the second at the individual level. It is also true that while both data were collected in Finland in the late 1990s, the FIN99 represents a smaller share of the total population than the FES98; the FIN99 data represent only Finnishspeakers, while the FES98 represents the entire population. The ways of collecting data are also likely to influence the reliability of the data sets. The second data set was obtained through postal survey, while the first was based on diary-keeping, face-to-face interviews, and official register data. It follows that the lack of official information for instance on household income may have led to incorrect reporting in the FIN99 data. Information on income is thus likely to be more reliable in the first data set. Likewise the reliability of other independent variables should be kept in mind in interpreting the results. All in all, however, we can assume that the greatest difficulties stem from the nature of the items measured. Subjective evaluations by individuals and objectively measured household expenditure can never be fully comparable or compatible. Actualised and imaginary levels of consumption provide information about different aspects of behaviour. The imaginary level measures thoughts, emotions and aspirations related to consumption, while the actualised level simply measures spending. In addition, the actualised level of consumption is measured at the household level; it follows that social relations within families are likely to affect the results considerably. The imaginary level of consumption, on the other hand, is not necessarily influenced by the family context at all. Thus it can be assumed that observations on these two levels offer tools for understanding the role of the family or household in the formation of consumption patterns more profoundly than by using one type of observation only. It is also evident that the FES98 data set includes more detailed information on consumption practices and more background variables than the FIN99 data set. Expenditure measures are more or less specific in the FES98 data, while consumption desires are measured as relatively broad categories in the FIN99 data. However, it should be stressed here that the aim of the analysis is not to offer comparable results as such. This would be impossible due to the obvious differences between the data sets. Rather, the

137

aim is to find out whether similar socio-demographic features can be found underlying consumption on the basis of different types of data. The empirical analysis of both sets of data starts from the selection of suitable consumption measures. After that, the structure of the dependent variables is explored. The purpose of the study is to find out whether certain dimensions can adequately characterise the consumption activities selected. It is by means of these dimensions that socio-economic variation in expenditure and subjective consumption is explored. While we are examining consumption at different levels (household and individual) and using different measures (expenditure and consumption desires), the categories are treated as similarly to each other as possible. For example, the same independent variables are used in each analysis. The nature of the consumption indicators and their restrictions will be further discussed in connection with the selection of dependent variables. Nevertheless, only items of considerable importance and interest will be selected, since it is not possible to cover all areas of consumption in this study. Consumption is defined here as consisting of total spending and the allocation of money, and all possible decisions related to these. From this perspective, total consumption is approached above all in a framework of economic resources. It follows that income and wealth are seen as the most important factors influencing the total consumption of a household or an individual. However, economic factors may not always be the best predictors when consumption is explored at a more itemised level. Such factors as class, gender or type of household may have the greatest impact in certain fields of expenditure. Moreover, analysing individual consumption in terms of subjective desires also has consequences. For example, it would not be reasonable to ask subjects to say how much they would increase their total consumption if the term ‘total consumption’ is not unambiguously understood by all respondents. This argument provides perhaps the best justification for analysing consumption by means of the selected consumption areas or categories rather than as a totality. It should also be stressed that the data sets utilised restrict the consumption categories that can be covered in this study. The two basic techniques primarily employed in the study are principal component factor analysis (PCA) and analysis of variance (ANOVA). Factor analysis is first applied to the data, to reduce the number

138

of variables to more general dimensions and to detect structures in the relationship between them. In general, it can be argued that factor analysis provides a feasible method of combining variables that are highly intercorrelated. Variance analysis is used as a method of exploring the effects of independent variables on dependent ones. Here it is used to distinguish socio-demographic differences in consumption activities. Analysis of variance can be argued to provide a basic tool of examining mean scores and their deviation from the overall mean of the different groups in the data. These statistical methods were selected on the basis of their simplicity and suitability for the purposes of this study. In factor analysis it is in principle possible for instance to use the similar criteria to extract factors from different data sets. Variance analysis, on the other hand, can be seen as an efficient way of analysing data and interpreting the effects of various independents. Tests can be modelled for instance by entering the selected background variables into the equations in same order. The analysis procedure is presented as clearly and as accessibly as possible, avoiding the camouflaging of the actual results behind sophisticated mathematical concepts or formulae. A more detailed description of the methods used will be given in connection with each analysis. Throughout the empirical analysis, SPSS version 10.1 is utilised.

Notes 1

Causal explanations are understood in this context as sufficient answers to why-questions. A

causal explanation can be roughly defined as a logical operation, in which a proposition or phenomenon to be explained is a deduction from a certain scientific law together with the relevant initial conditions (e.g. Elster 1989: 3; Ruonavaara 1990: 129). This tradition of scientific explanation is of course somewhat problematic when we are dealing with objects that are capable of consciousness about their own existence and about their situation. In this study, however, ‘explanation’ is used in the sense of ‘social mechanism’, as discussed in Chapter 3. 2

The weight coefficient available in the data is calculated to show how many households out of

the total population are represented by a particular household (Tilastokeskus 2001: 23-25). In other words, the coefficient indicates how many cases are represented by a particular case in Finland. Using a weight that grosses up the sample, however, is not necessarily sensible when

139

using certain statistical methods. This is because the coefficient exaggerates the statistical significances of tests. In the case of the data in question, this is to say that the original N of cases actually means thousands of cases. The original weight, grossed up to the population level was therefore divided by the sample. This modified weight is applied to each household in the analyses. This allows us to benefit from the original weight, but does not exaggerate the statistical significances. 3

Approximately six percent of the population of Finland is Swedish-speaking; this group is

mainly concentrated along the southern and southwestern coast of the country. 4

The euro became the official currency of the European Monetary Union (EMU) in 2002 and

replaced national currencies in the euro area. 5

The following question was used: ”Which of the following social class do you predominantly

identify with?”. It can be argued that this question measures class identification in the clearest possible sense. We can also assume that the question would be understood very similarly by each respondent.

140

5 EXAMINATION OF HOUSEHOLD PATTERNS OF CONSUMPTION EXPENDITURE An examination of household expenditure can be considered for a number of reasons to add to our understanding of the roles played by social structures in the formation of individual consumption patterns. First of all, in daily life consumption is often perceived as spending on goods and services. This means that what people are buying is perhaps the most unambiguous indicator of their consumer behaviour. Secondly, the role of the household in the formation of consumption patterns should be seen as relevant because decisions concerning money allocation often take place in families. The fact is that most people in our society live together with others, and many consumer goods come as durables that are used by more than one person. Finally, in empirical research consumption has typically been measured as household expenditure. In this chapter the central dimensions of consumption expenditure by households are surveyed, using data from the FES98 survey. I first take a look at the basic structure of household expenditure in Finland. This is followed by an analysis of expenditure in certain subcategories of the general household consumption patterns. In particular expenditure on culture, hobbies and other common activities is examined. Expenditure on housing and driving will also be analysed. Consumption expenditure variables are selected for their expressiveness and for their general comparability with the variables in the FIN99 data. However, while the aspect of comparability of the data sets is kept in mind in the empirical analyses of this study, it should be stressed that my aim is merely to draw certain rough comparisons between patterns of expenditure and of imaginary consumption. 5.1 Expenditure on consumption If we wish to explore the nature of everyday consumption and consumption determinants, it may be sensible to start from the basic pattern of consumption. At first sight it may seem quite unproblematic to piece together what we understand by the basic pattern of everyday consumption. As described earlier, our consumption activities can be approached to a 143

large degree from the viewpoint of need satisfaction. For example, we have to eat, drink, sleep, and wear clothes. We also need to live somewhere, so we have to pay rent, maintenance charges or a mortgage for an apartment or house; and perhaps we need to own a car. Ordinary members of contemporary society have to continually buy things and services in order to satisfy their needs. At the same time, however, the viewpoint of need satisfaction is in many ways insufficient. Many of our consumption activities cannot be regarded as needs that require satisfaction. For example, it would be very unconvincing to claim that there exists a basic need to buy a new car, and that this need has to be satisfied every now and then even if the old car is in excellent condition. In a way, then, it is more accurate to say that the need for continual consumption arises out of our living conditions. It might therefore be adequate to approach the nature of consumption in terms of the general household pattern of expenditure. The general household pattern can be derived from the total house expenditure. It includes expenditure on all possible goods and services, including loans and other financial expenditures. The FES98 data set contains twelve main expenditure groups measuring total household expenditure. These main groups include several subcategories and about 800 items. The average amounts of euros (€) that households allocate to the main categories, the relative percentages of total consumption expenditure and the standard deviations can be seen in Table 5.1. The figures indicate absolute levels of household expenditure (all the subcategories belonging to these main categories are listed in Appendix B). As Table 5.1 shows, the general pattern of household expenditure is measured by means of relatively broad main categories. The table also indicates that the proportions of some expenditure categories are quite large, some other rather small. Housing, food, transportation, and recreation and culture stand out as the largest single expenditure groups. Expenditure on education is clearly the smallest main group in the data. It can be argued that some of the groups seem considerably too broad to represent single expenditure categories. For example, it is not easy to approach recreation and culture, or furnishings, household equipment and routine home maintenance, as single expenditure groups. In other words, expenditure on some items under a given main category may illustrate very different type of spending than others. In addition, there are also various

144

types of goods and services that are not included in the main groups. More than ten percent of total consumption expenditure is labelled as miscellaneous goods and services. Since it is not the purpose of this study to analyse expenditure as whole, I focus below on the central features of the expenditure dimensions selected. Table 5.1 Main expenditure groups and average household expenditure Item

%

€ / household / year

Std. Deviation

13,6 2,8 4,5 27,3

3017,3 629,4 990,6 6032,7

1932,3 1008,7 2154,6 2947,1

1 Food and non-alcoholic beverages 2 Alcoholic beverages and tobacco 3 Clothing and footwear 4 Housing, water, electricity, gas and other fuels 5 Furnishings, household equipment and routine maintenance of the house 6 Health 7 Transportation 8 Communication 9 Recreation and culture 10 Education 11 Restaurants and hotels 12 Miscellaneous goods and services

4,5

1003,4

1477,4

3,6 16,2 2,7 10,1 0,2 4,0 10,5

795,5 3593,8 598,4 2239,8 38,7 877,3 2317,5

1146,3 4998,7 539,8 2357,5 282,1 1186,8 2247,8

Total consumption expenditure

100

22 234,5

Some of the most interesting features of consumption would be difficult to capture by concentrating on the main expenditure groups alone. For example, we would miss the connections between certain goods and activities, simply because they are entered under different main categories. It would also be impossible to draw any comparative interpretations with the FIN99 data even at the most general level. The FIN99 data do not include any items concerning health care, communication, or the consumption of groceries, tobacco or alcohol (see Chapter 6 for a closer description of the measures in the FIN99). After these restrictions, a total of only 18 different variables were selected from the data. Five of the twelve main expenditure categories as a whole were excluded from the analysis. It was concluded that they did not contain feasible items for our purposes. The excluded categories were 1) food and non-alcoholic beverages, 2) alcoholic beverages and tobacco, 3) 145

health, 4) communication, and 5) education. Some more itemised sub-items (e.g. taxes, fines, membership payments, loans, consumption credit charges) were excluded as a rule because money allocation to such items cannot actually be regarded as consumption expenditure. In addition, the effects of some of these will also be partly represented through active consumption expenditure (e.g. vehicles and household equipment are usually bought on hire purchase). Monthly mortgage payments, however, are included in housing costs similarly to rent and routine maintenance charges. Basically, seven of the main expenditure groups are represented in the selected categories. Some of the main categories were picked up in their original form, but generally little more itemised subcategories were selected. It can be argued that the selected variables represent consumption expenditure at least in a slightly more detailed level than the original main expenditure groups in the data. In spite of this, all of the selected variables represent expenditure categories, that is, they do not measure spending on individual items. The average amounts that households allocate to the selected categories, their relative percentages of total consumption expenditure, and standard deviations can be seen in Table 5.2 (see Appendix B for all subcategories in FES98 data). The largest expenditure categories appear first in the list. The variables selected account for 61 percent of total household expenditure in the data. Expenditure on housing (housing, water, electricity, gas and other fuels) is clearly the largest single category, accounting for more than 44 percent of the selected total expenditure. This main category is structured problematically in the FES98 data set and cannot easily be divided into more itemised subgroups (see Appendix B); it is therefore used in its original form. The relative size of this particular category naturally has to be taken into account in the further analyses. It should also be noted that some of the most important consumption categories of everyday life are not included in the list. Perhaps most visible is the absence of food and non-alcoholic beverages, which accounts for almost 14 percent of total household expenditure. The absence of this category, and the potential effect on the analysis, should be kept in mind when interpreting the results.

146

Table 5.2 Selected expenditure categories and average household expenditure Item

%

€ / household / year

Std. Deviation

1 Housing, water, electricity, gas and other fuels 2 Acquisition of vehicles 3 Operating expenses on vehicles 4 Clothing and footwear 5 Restaurants and cafés 6 Books and magazines 7 Recreation and leisure equipment 8 Cosmetics and beauty care 9 Insurance 10 Furnishings, art objects and carpeting 11 Audiovisual equipment and computers 12 Package tours 13 Cultural services 14 Household equipment 15 Sport and leisure services 16 Hotels and hostels 17 Personal goods 18 Banking and other financial services

44,4

6032,7

2947,1

9,7 8,8 7,3 4,3 3,3 3,2 3,1 2,9 2,9 2,8 2,1 2,0 1,6 0,9 0,5 0,1 0,0

1325,2 1195,2 990,6 587,9 446,9 417,5 415,4 379,5 387,6 380,3 284,7 263,7 202,2 124,2 72,3 87,1 2,7

4122,0 1959,8 2154,6 951,0 546,7 772,3 605,6 380,7 857,2 740,9 663,7 399,2 513,6 509,2 261,3 431,9 40,9

Selected total consumption expenditure Proportion of total household consumption expenditure

100 .61

13 596,2

The following analytical procedure may appear somewhat problematic in the framework of total household expenditure. We should remember, however, that food and other excluded categories were removed from the analysis because they are not included in the FIN99 survey. In addition, we can argue that health and education represent relatively different consumption fields compared to the measures selected. Expenditure on health (doctors’ fees, medicine etc.) cannot be regarded as an everyday field of consumption at all. Rather, it is usually constrained by illness and conditions related to ageing. Likewise education is a problematic indicator of consumption. The amount of money spent on education in Finland is strongly associated with particular age groups: in the 1990s, for instance, when the FES98 study was carried out, the baby-boom generation was the primary group of private consumers of various commercial educational

147

packages in adult education centres and folk high schools (see Rinne et. al. 1992). Study at a university or college, on the other hand, does not require tuition fees from regular Finnish students. Overall, it is reasonable to assume that the 18 variables selected measure at least partly similar consumption activities; these will be analysed in the next chapter using subjective measures.1 But should we assume that the expenditure variables selected are mutually connected in some way? In other words, can they be analysed in terms of distinctive expenditure dimensions? According to basic methodological principles, a large number of variables cannot be reduced to a smaller number if the correlations between the variables are very weak. This would be problematic, not only in technical terms but also for interpretations concerning the structure of household consumption. The next step in the analysis is therefore to apply factor analysis to the data, in order to explore the possible dimensions in which the consumption categories can be combined. Factor analysis can be seen as a feasible multivariate technique for uncovering various dimensions in data containing many variables: a larger set of variables is compressed into a smaller set, which is included in the joint variance of variables.2 In our case the variables are the various specific expenditure categories, and the factors can be seen as describing more general areas of household consumption. However, it should be noted that during recent decades the use of factor analysis in the social sciences has been subject to considerable criticism. According to some writers, its use can quite easily lead to over-interpretation and perhaps even distortion of the results (see Tabachnick & Fidell 2001: 283-294; Graham & Collins 1991). There is naturally a risk of over-simplification in the use of factor analysis. Despite this risk, however, the method has been commonly used in lifestyle-oriented consumer research over the years (see Wilska 2002; Wilska 1999; Katz-Gerro & Shavit 1998; Hendry et. al. 1993; Uusitalo 1979). Furthermore, factor analysis is used here as a method of data reduction, rather than for final explanatory purposes. The method is applied in order to allow the combination of correlated variables in factors, which can then be used in the further analysis.

148

5.2 Exploring household expenditure on the selected categories The extraction technique used in the analysis is principal component analysis (PCA); the rotation technique used is Varimax. The PCA technique was selected because it forms factor components from the original variables without separating the joint variance from the variance of single variables (Harman 1976: 140-142). This can be regarded as a feasible solution in the interpretative sense, since the basic goal in PCA is to reduce the size of the data by providing independent components. Varimax rotation, on the other hand, yields the highest possible loading of a variable on one factor and minimises it on other factors. The number of factors in the analysis is defined according to Eigenvalues and the proportion of total variance reduced by each factor. An Eigenvalue indicates how much variation in the original group of variables is accounted for by a single factor. The criteria for Eigenvalues were set as ≥ 1. The proportion of total variance reduced is set at ≥ 5 %, which is the common recommendation in the literature (see Alkula et. al. 1994: 276; Toivonen 1999: 338-339). The interpretation of the factors is based on the highest loadings for 2 each factor. Communalities (h ), i.e. the sums of the squared factor loadings for each variable, are used as estimations of the explained shares of single variables. The first procedure, where only the Eigenvalue criterion was used, yielded a total of eight factors. Such a large number of factors, however, proved impossible to interpret; the first three factors explained more than 30 percent of the variance, all the other factors combined only five percent. The eight-factor solution was seen as problematic because the aim of the analysis was to find distinct consumption dimensions that could be examined separately. The next step, therefore, was to reduce the number of factors. However, even solutions with six or five factors seemed too complex to lend themselves to interpretation; there were some high loadings on the first factors, but in general the loadings on the other factors were relatively weak. Several analyses were therefore conducted; it was only when the number of factors was limited that at least relatively consistent factor interpretations emerged. Eventually, a four-factor solution was found, which explained 37 percent of the total variance of the selected expenditure variables; this solution was relatively easy to interpret. This was considered

149

to be the best solution, even though a three-factor solution, which lent itself relatively easily to interpretation, was also constructed. The four-factor 2 solution with the extraction communalities (h ) is shown in Table 5.3. The factor matrices show only the Varimax-rotated factors. The highest positive or negative loading of each item is shown in boldface in the table. Table 5.3 Varimax-rotated factor matrix of the selected expenditure categories Item

I

II

III

IV

h

2

Housing, water, electricity, gas and other fuels Acquisition of vehicles Operating expenses of vehicles Recreation and leisure equipment Package tours Hotels and hostels Insurance Audiovisual equipment and computers Sport and leisure services Cultural services Restaurants and cafés Clothing and footwear Books and magazines Cosmetics and beauty care Personal goods Banking and other financial services Furnishings, art objects and carpeting Household equipment

0,63

0,14

0,11

0,27

0,49

0,57 0,57 0,42 0,40 0,38 0,77 0,12 0,07 0,06 0,15 0,17 0,17 0,11 -0,01 -0,02 0,13 0,09

-0,13 0,07 0,20 0,31 0,17 0,12 0,37 0,39 0,69 0,61 0,22 0,20 0,37 -0,02 0,34 0,09 -0,02

0,04 0,12 0,24 -0,12 0,05 0,10 0,21 0,17 0,00 0,15 0,50 0,44 0,43 0,69 -0,37 0,11 -0,03

0,13 0,01 0,05 -0,03 -0,24 0,19 0,34 -0,03 0,07 -0,02 0,07 0,17 0,16 -0,04 0,12 0,67 0,75

0,35 0,34 0,27 0,27 0,23 0,66 0,31 0,19 0,40 0,42 0,33 0,29 0,36 0,47 0,28 0,48 0,57

Eigenvalue Total variance explained

2,26 .12

1,63 .09

1,45 .08

1,39 .08

.37

Note: Extraction method: Principal Component Analysis; rotation method: Varimax with Kaiser Normalization.

In general, the explanation percentages for the factor components are relatively small. The first factor explains almost 12 percent, the second nine, and the next two factors eight percent of the total variance. It should also be noted that the cumulative proportion of the total variance explained (37 percent) is quite modest. However, the proportions explained can be regarded as satisfactory if we take into consideration the nature of the items selected. Indicators of the same type of consumer behaviour do not

150

necessarily correlate very strongly with each other. This is because the variables selected consist of relatively broad categories that include many types of items. It can be assumed, for example, that the allocation of money to one type of consumption, such as cultural services, displaces them from other related activities, such as reading books and magazines. This is because the household budget is always restricted to at least some degree, and consumption expenditure on various activities is the end result of choosing a given alternative over the others available. Even though the four factors explain less than 40 percent of total variance, we can argue that they offer a fairly clear model for the interpretation of basic patterns of household expenditure. The structural dimensions of the factors become apparent when we examine their highest loadings on each dimension and assign interpretations to them. None of the factors stands out clearly as the most important consumption dimension of daily life, in the sense of a general factor connected with all types of activities. Rather, it seems that the factors consist of correlations of only a few items, which can be distinguished relatively easily. Despite this, all the items appear to have considerable loadings on at least one factor (0.37 is the weakest factor score in this sense). As shown in Table 5.3, the first factor loads high on housing, the acquisition and operating costs of vehicles, recreation and leisure equipment, package tours, hotels and hostels, and insurance. This factor can be labelled Expenditure on everyday life. While this factor is labelled as referring to the basic activities of daily life, we should remember that our list of variables does not include food. However, in preliminary analyses of the FES98 data (not shown here) food was found to load high on the same factor as housing and other general items related to daily expenditure. In this sense the factor label may not be seriously misleading. In any case, it should be kept in mind that the exclusion of food probably has some effect on the further analysis procedure. The second factor, on the other hand, loads high on cultural services, audiovisual equipment and computers, sport and leisure services, and restaurants and cafés. This factor can be interpreted as Expenditure on culture and luxury. The third factor seems to be connected with many hedonistic consumption activities. It loads high on clothing and footwear, books and magazines, cosmetics and beauty care. Banking and other

151

financial services load negatively high on this factor. The factor is labelled here as Expenditure on hedonistic consumption. The last factor has a high loading for only two items: furnishings, art objects and carpeting, and household equipment. Simply on the basis of the nature of these variables, the factor can be labelled as Expenditure on home equipment. The original idea of running the factor analysis was that the 18 consumption expenditure variables could be reduced to a smaller number. The analysis produced four dimensions. The interpretation of these expenditure dimensions was found to be relatively unproblematic. It should of course be noted that the first factor (expenditure on everyday life) is dominated by housing expenditure. But it can be argued that all the variables with the highest loading on the same factor can be dealt with together. Accordingly, the variables with the highest loadings for the same factors were summated and labelled for further analysis as new intervalscale variables.3 A total of four summated scales were constructed. The first scale consists of seven variables, the second of four variables, the third of five variables and the fourth of two variables. The scales and their relevant parameters are given in Table 5.4. All the variables included in the scales (or sum variables) can be seen in Table 5.3. Table 5.4 Scales for different types of consumption expenditure Sum variable 1 Expenditure on everyday life 2 Expenditure on culture and luxury 3 Expenditure on hedonistic consumption 4 Expenditure on home equipment

N

€/ Minimum

€/ Maximum

€/ Mean

Std. Deviation

4359

0,0

86960,2

9707,3

7030,1

4359

0,0

23245,9

1356,2

1642,7

4359

0,0

44024,8

1942,8

2647,9

4359

0,0

16711,2

589,8

1098,6

The first point to be noted in Table 5.4 is the considerable variation in the scope of the scales. For example, average household expenditure on everyday life is almost 10 000 € a year, while average expenditure on home equipment is less than 600 € a year. Expenditure within the scales also 152

seems to vary considerably. At one end there are households that report spending very much on the selected items, while at the other end there are households that do not report spending at all. The number of households that report not spending on any item on a particular scale, however, is very low. Moreover, what is important here is the average amount of expenditure, not minimum or maximum values. There is thus no reason to exclude any deviant cases from the further analysis. No transformations to the raw scores of the households were employed either.4 After inspecting the scatter plots, it was concluded that the constructed expenditure variables appeared relatively symmetric and approximately normally distributed. It is well known that different groups of people have different expenditure styles. The next question to be addressed in the analysis is whether variation in the expenditure dimensions indicated by the summated scales can be explained by the selected independent background variables. The statistical method used for this purpose is analysis of variance (ANOVA). ANOVA was originally developed as a method of experimental science, but it widely used in all areas of social research. It is basically a tool for determining the statistical differences between the means of two or more different populations (McNeil et. al. 1996: 3-4; Toivonen 1999). An important assumption of the method is the homogeneity of variance: in other words, it is assumed that the data within each cell come from normally distributed populations with equal variances. In an ANOVA analysis, the relationship between the mean and the variance of a given group provides the information needed to determine if the difference between the two is significant.5 Main-effects ANOVA models are used here to compare the significance of the effects of background variables for the summated expenditure scales. The effects of other independent background variables will also be controlled in the analyses. Thus the aim is to show the extent to which the selected independent variables explain the variances in the dependent variables both together and separately. Some interpretations of possible interactions between the independents are also made. In general, an interaction effect indicates a change in the simple main-effect of one variable over levels of the second (see McNeil et. al. 1996; Achen 1986).

153

The analysis concentrates on the exploration of two-way interactions. The analysis starts from the sum variable labelled Expenditure on everyday life. 5.2.1 Expenditure on everyday life The aim of the ANOVA analysis is to uncover possible socio-demographic differences in household patterns of consumption expenditure. The analysis procedure was carried out in two parts. First, the unadjusted effect of each background variable was tested; after that the effects of other variables were explored by entering a new variable one at a time into the model. Table 5.5 shows the results of the main-effect tests for expenditure on everyday life. The socio-economic status of the head of household (HOH) is entered first because it represents a class measure in this data. Class can be seen as a general factor affecting the distribution of life changes. Statistical significance is indicated in the table by the value F, the parameter estimates ( β ) and also by the variance explained for all 2 independent variables (Adj. R ).6 The unadjusted main-effects shown in the first column of the table reveal that all background variables are significantly associated with the expenditure on everyday life variable. The most notable effects seem to be those of type of household, gender of head of household (HOH) and the socio-economic status of HOH. The effect of OECD-equalised household income is also considerable. On the other hand, the significance of place of residence is clearly the weakest. In terms of euros allocated, the average differences are largest between income groups: expenditure decreases steadily from the highest to the lowest income group. It should be noted, however, that the F statistic does not report the power of the independent variable as such. Rather, it is a value of statistical significance, which should be interpreted in relation to the number of categories in the independent variable. In general, the differences observed seem to be more or less as expected. For example, it can be regarded as fairly predictable to find that the households of entrepreneurs and professionals or managers allocate clearly more money to this type of consumption than the households of manual workers. The same can be said of the differences between types of households. Couples without children and couples with children spend the most. Households categorised under the label of ‘other’ come next,

154

followed by the elderly and the households of single persons. It may also be noted with regard to the gender variable that the head of the household in families that spend more on everyday life is more typically male than female. Perhaps the only surprising result is that households located in densely populated (non-urban) areas spend more on everyday life than households in urban areas. However, this is probably due to the fact that most households in urban areas live in apartments and flats rather than detached or semi-detached houses. This means that energy costs, which account for a great deal of expenditure, are lower in urban settings than in areas where other forms of housing are more common. In any case, the differences between residential groups are not very great. It may be assumed that the effect of the place of residence will weaken considerably when the effects of other variables are controlled. 2 Model 1 shows that the part of the variance accounted (Adj. R ) for by socio-economic status of HOH is 15 percent. This can be seen as a major share. It appears, however, that available income also has a strong affect on expenditure. In the second model, it is not surprising that differences between socio-economic groups decline when the effect of income is controlled. Differences between income brackets also level out. In this sense, socio-economic status and income seem to be interconnected. A total of 21 percent of variance can be explained by these two variables. However, even more significant changes occur when type of household is added to the equation in Model 3. This increases the explained variance by another 15 percent, and reduces the differences between socioeconomic groups considerably. In this model it is interesting to note that the parameter estimates of farmer and intermediate non-manual households change relatively much. It may well be that some of the differences explained by the socio-economic status of the HOH can be attributed to the type of household. Moreover, the parameter estimates suggest that of these three, it is type of household that is the most important predictor of expenditure.

155

271 1013 1274 902 543 587

Household income Less than 6500 €/year 6500-10 000 €/year 10 001-13 000 €/year 13 001-17 000 €/year 17 001-20 000 €/year Over 20 000 €/year

(cont.)

86 249 617 748 1875 1015

N

Socio-economic status of HOH Farmer Other entrepreneur Professional, manager Intermediate non-manual Other household Manual worker

Source of variance

146,28*** -8770,0 -7337,0 -5441,2 -3655,4 -1779,4 (a)

2732,5 3001,5 3252,2 86,3 -3833,6 (a)

160,24***

Unadjusted main effects

2732,5 3001,5 3252,2 86,3 -3833,6 (a)

160,24***

Model 1

66,52*** -6299,8 -5161,2 -3705,8 -2523,5 -1169,7 (a)

2500,8 2601,6 1905,7 -182,2 -2957,7 (a)

79,42***

Model 2

113,56*** -7122,3 -6767,1 -5219,9 -3660,3 -1863,7 (a)

1234,9 2242,6 1167,1 728,3 24,5 (a)

9,00***

Model 3

110,84*** -7013,2 -6677,3 -5131,7 -3588,1 -1814,3 (a)

1185,2 2272,7 1300,8 1041,0 128,8 (a)

9,99***

Model 4

110,24** -7003,7 -6683,4 -5112,7 -3579,0 -1785,7 (a)

1145,5 2260,2 1347,7 1075,9 153,4 (a)

10,11***

Model 5

Table 5.5 Expenditure on everyday life by independent background variables

2755 1835 2874 657 1059

Gender of HOH Male Female

Place of residence Urban Densely populated Rural

8,89*** 206,5 1365,3 (a)

206,04*** 2953,1 (a)

286,06*** 5508,6 2628,0 8479,3 497,0 5354,3 (a)

.15

.21

.35

215,53*** 4333,3 4121,2 8364,3 1551,9 5235,0 (a)

Note: *** = p < 0.001; ** = p < 0.01; * = p < 0.05; (nfs) = p > 0.05; (a) = redundant parameter.

Adj. R

2

898 221 1088 923 395 1065

Type of household Couple Single parent Couple + children Elderly Other household Single person

.36

24,35*** 920,8 (a)

197,48*** 4203,5 4440,2 8118,2 1672,8 5230,8 (a)

.36

3,91* -72,3 606,5 (a)

24,24*** 918,4 (a)

193,33** 4165,0 4420,2 8058,7 1615,9 5193,7 (a)

It should be remembered, of course, that expenditure on housing is the largest share in this dimension. However, in additional analyses (not reported here) the type of household was found to be the best predictor of some of the individual items included in the scale as well. Operating expenses of vehicles provided a good example of this. In this respect, there is no need to question whether the observed power of this independent variable depends only on the selection of housing variable. There are no surprising changes in the following models. It appears that the effects of gender and place of residence are actually rather weak compared to the effects of the other three background variables. In any case, the effects of both remain statistically significant. The explanation proportions of the fourth and the fifth models are the same, i.e. 36 percent. On the basis of the main-effect tests, it can thus be concluded that the best predictors of expenditure on everyday life are disposable income and type of household. Socio-economic status and the gender HOH come next. The examination of the interactions showed that the socio-economic status of HOH had a significant interaction effect with all other background variables (p < 0.01). In addition, household income was found to interact significantly with type of household and place of residence. A closer look, however, indicated that these effects were not very interesting. The effects of socio-economic status of HOH, for instance, were found to be significant mainly due to the effect of households of farmers, entrepreneurs and other households. These socio-economic groups seem to spend on everyday life rather differently according to income level, type of household, gender of HOH and place of residence. This may result from possible differences between the income composition in these socio-economic groups. The significant interaction between income and type of household, on the other hand, revealed that within higher income groups, households of single parents spend clearly less than others. Interaction between income and place of residence showed that urban people in the lowest income groups spend significantly more on everyday life than other residential groups. These results might be due to the fact that in many cases households with a low income must allocate almost all their money to the basic goods of everyday life. It should be noted, however, that the measurement of basic goods is based mainly on such items as housing and

158

motoring expenses. Variation in these categories may be somewhat different from variation in expenditure for instance on food or on alcohol. While these interactions revealed some important connections, we do not have to modify our interpretations greatly. The observed differences can be attributed mainly to the main-effects appearing in Table 5.5. We can therefore conclude that the largest differences between socio-economic groups were found between professionals or managers and manual workers. Couples with children were found to spend the most among types of households. Households with a man as the head spend more on everyday life than households headed by a woman. In addition, expenditure seems to decrease steadily from the highest income group to the lowest. The effect of place of residence, for its part, turned out to be relatively weak compared to other variables. In general, then, it is unproblematic to approach this kind of expenditure in modern life: income level and the basic positions of life can be seen as determining rather strongly how much money households spend. It may also be stressed that all the independent variables combined explain a considerable proportion of the variance. In this respect, some of the structural mechanisms discussed earlier are preferable in the theoretical interpretation of these findings. Next, however, I turn to socio-demographic variation in expenditure on culture and luxury. 5.2.2 Expenditure on culture and luxury The statistical procedure for analysis of variance was conducted in a similar way as described earlier. Table 5.6 shows the results of analysis for the dependent variable. As the table indicates, each independent variable has a significant effect on expenditure on culture and luxury. Socio-economic status of HOH, type of household and gender of HOH appear to have the strongest unadjusted effects according to the parameter estimates ( β ). However, it also appears that variation is relatively strong between different categories of income and place of residence. Professionals or managers seem to be the socio-economic group that spend most on culture and luxury. Intermediate non-manual workers and other entrepreneurs than farmers come next. On the other hand, households labelled as other spend clearly the least. Despite this, in particular manual workers and farmers should be seen here as reference groups for the groups

159

that spend most. There are clear differences in relation to these categories. Their expenditure increases steadily from the lowest to the highest bracket; the lowest income bracket itself, however, forms an exception. Couples, whether with or without children, are again the type of households that spend the most. The elderly, on the other hand, spend clearly the least of all types of households. Gender differences also appear to be as expected. In other words, households headed by a man report on average more abundant expenditure than those headed by a woman. Expenditure on culture and luxury is also greater in urban and densely populated areas than in rural settings. Overall, it can be argued that the results are not surprising. For example the socio-economic variation is easy to understand against the framework of any theoretical model of social class. The items included in the scale measure expenditure on cultural services, restaurants and cafés, home electronics, and sport and leisure services. Most of these items can be regarded as representing ‘high’ culture in one dimension, that is, roughly in typical lifestyle orientations. This is true in particular of the differences observed between on the one hand the middle-class groups of professionals or managers and intermediate non-manual workers, on the other the working-class manual workers. In this sense, the results characterise occupational differences in terms of the basic class hierarchy. There is likewise nothing of particular interest in the results concerning type of household and gender of HOH. Couples, with or without children, are probably the primary consumers of selected types of expenditure items for several reasons. Most cultural, sport and leisure services, for instance, represent typical objects of money allocation among families with children and among couples. Old people, on the other hand, do not necessarily spend much on cultural activities at all. The gender effect might be approached in a highly similar fashion. The difference between male and female HOHs, however, is not very strong. In addition, the effect of income level is not yet controlled in these main-effect results. The proportion of variance explained by socio-economic status of HOH can be seen in Model 1. It might be said that this variable explains a relatively high proportion (15 percent) of the variation in expenditure on culture and luxury. The proportion increases by only one percent in the following model, in which the OECD-equalised income of the household is

160

added. Not many changes, however, occur in comparison with the unadjusted effects. Basically only the differences between income brackets appear to diminish. The same can be said of Model 3, which adds the type of household to the equation. In this model, some changes can be observed in the socio-economic status of the HOH. In general, the differences between occupational categories diminish only slightly, while the parameter estimate of ‘other household’ diminishes considerably. It can thus be assumed that controlling the effect of the type of household levels out the effect of the ‘other household’ category, which in fact consists of HOHs of heterogeneous occupations. A total of 19 percent of the variance can be accounted for by these three independents. Controlling the effects of gender of HOH and place of residence does 2 not increase the adjusted R much in the next model. Despite this, all variables are very significant also in the final model, which explains no less than 22 percent of the total variance. The examination of interaction effects, however, revealed a few interesting connections between the independent variables. First of all, the interaction of socio-economic status of HOH with income, with type of household and with place of residence were significant (at p < 0.01 level). Similarly, the effects of income with type of household, with gender of HOH and with place of residence were found to be significant. The interaction between socio-economic status of HOH and income revealed that professionals and managers who are in the lowest income bracket spend more on culture and luxury than other groups. This indicates that cultural conditions related to different class positions may also affect consumption, independent of the economic resources of the household. A closer examination between the interaction of income and type of household, however, showed that this interaction was due to the households of couples with and without children. It can thus be approached mainly in terms of an ‘age effect’, indicating that the households of young and working-aged people generally spend more on culture and luxury than older people. Those households labelled socio-economically as ‘other households’ also showed a significant interaction with the different categories of the type of household variable. These differences will not be elaborated here in more detail, since households labelled as ‘other’ in our socio-economic classification do not form a homogenous category.

161

271 1013 1274 902 543 587

Household income Less than 6500 €/year 6500-10 000 €/year 10 001-13 000 €/year 13 001-17 000 €/year 17 001-20 000 €/year Over 20 000 €/year

(cont.)

86 249 617 748 1875 1015

N

Socio-economic status of HOH Farmer Other entrepreneur Professional, manager Intermediate non-manual Other household Manual worker

Source of variance

52,67*** -860,0 -1068,7 -790,5 -469,2 -72,5 (a)

-299,6 333,3 1044,6 306,1 -790,2 (a)

166,81***

Unadjusted main effects

-299,6 333,3 1044,6 306,1 -790,2 (a)

166,81***

Model 1

9,71*** -67,2 -396,0 -283,6 -152,3 84,6 (a)

-334,1 282,2 938,5 285,6 -738,2 (a)

118,68***

Model 2

14,48*** -338,7 -576,6 -415,4 -269,2 10,6 (a)

-373,9 283,8 846,8 354,6 -214,0 (a)

37,47***

Model 3

13,61*** -306,2 -549,8 -389,1 -247,6 25,3 (a)

-388,7 292,8 886,6 447,8 -182,9 (a)

40,80***

Model 4

10,43*** -270,6 -469,4 -363,8 -424,4 20,4 (a)

-133,6 328,5 812,9 403,4 -183,8 (a)

33,96***

Model 5

Table 5.6 Expenditure on culture and luxury by independent background variables

2755 1835 2874 657 1059

Gender of HOH Male Female

Place of residence Urban Densely populated Rural

82,5*** 729,5 380,4 (a)

124,9*** 542,5 (a)

131,45*** 398,1 13,8 853,5 -839,8 -53,6 (a)

.15

.16

.19

37,40*** 222,1 102,3 702,8 -439,1 65,0 (a)

Note: *** = p < 0.001; ** = p < 0.01; * = p < 0.05; (nfs) = p > 0.05; (a) = redundant parameter.

Adj. R

2

898 221 1088 923 395 1065

Type of household Couple Single parent Couple + children Elderly Other household Single person

.20

31,60*** 274,4 (a)

28,82*** 183,4 197,4 629,5 -403,1 63,7 (a)

.22

59,03*** 571,2 253,4 (a)

36,38*** 290,9 (a)

31,57*** 232,5 199,6 697,6 -343,6 123,0 (a)

Socio-economic variation by place of residence showed that two groups – ‘other entrepreneurs than farmers’ and ‘intermediate non-manual workers’ – report clearly greater spending in urban areas than in other residential areas. The significant interaction between income level and type of household in turn revealed that families with children in the second and third lowest and in the second highest income group spend significantly less on this type of activity than the same household type in other income groups. It can be assumed that these differences depend at least partly on the variation that exists within the group of households labelled as ‘other’ in terms of socio-economic status of HOH. There was also a significant effect for type of household by gender of HOH and by place of residence. Three household types – couples without children, couples with children, and other types of households – reported spending less when the HOH was male. This interaction is quite interesting, as is the finding that elderly households in urban areas spend less on culture and luxury than elderly households in other areas. A more general interpretation of these effects, however, is problematic; households labelled as ‘other’ is a heterogeneous category. While there are some noteworthy interactions between the independent variables, the main-effects presumably explain the variance more or less adequately. The proportion accounted for out of the total effect (22 percent) can be regarded as moderately high. Thus a relatively clear model is again observed for the interpretation of expenditure on culture and luxury. Basically, extensive expenditure seems to be typical of three types of household: managerial or professional, intermediate non-manual and entrepreneurial. In other words, upper and middle class households spend more on these activities than other socio-economic groups. While a clear main-effect for gender of HOH exists, the examination of interactions also revealed differences in this respect among various types of household. This probably reflects one of the key features of the Finnish culture, which cannot be regarded as a typical male breadwinner culture. At the same time, however, place of residence and type of household also seem to affect this dimension of household expenditure considerably. In general, expenditure on culture and luxury can be said to represent a more urban dimension of consumption, which is more typical of families with

164

children. The mechanisms which might be understood as relevant here are those of class, of basic social structures and of residential area. 5.2.3 Expenditure on hedonistic consumption This section presents an analysis of the structural features of expenditure on hedonistic consumption. The results of the main-effect ANOVA tests for expenditure on hedonistic consumption can be seen in Table 5.7. Again, the unadjusted main-effect for each variable is significant according to the value of F. Moreover, the structure of variation seems to be relatively similar to the variation in expenditure on culture and luxury: households with a head in the categories of manager or professional, other entrepreneur than farmer, and intermediate non-manual worker spend the most. The results according to household income are also similar. Expenditure increases stepwise with rising income. The two lowest income brackets, however, again form an exception. The differences between different types of household, not surprisingly, are quite great; couples with children spend the most and elderly households the least. In this respect, it should be noted that some of the items included in hedonistic consumption, such as expenditure on clothing and footwear, represent daily necessities as well as hedonistic forms of consumption. The gender difference revealed that households with a man as their head spend a little more than households headed by women. Differences by place of residence indicate that urban dwellers spend the most and those who live in rural settings the least. Considering the analysis of the previous expenditure category, the results of the unadjusted main-effect test thus revealed pretty much what might have been expected. But does controlling other variables change these findings? Model 1 shows that the socio-economic status of the HOH can explain 10 percent of variance. This is a fairly high share for a single variable. It may be noted that the explanatory proportion of the model does not increase when the effect of equalised household income is added in the following model. In other words, income and socio-economic status together do not predict this type of expenditure better than socio-economic status or income alone. Controlling the effect of income naturally reduces the differences between socio-economic groups, but, perhaps surprisingly, not very much.

165

271 1013 1274 902 543 587

Household income Less than 6500 €/year 6500-10 000 €/year 10 001-13 000 €/year 13 001-17 000 €/year 17 001-20 000 €/year Over 20 000 €/year

(cont.)

86 249 617 748 1875 1015

N

Socio-economic status of HOH Farmer Other entrepreneur Professional, manager Intermediate non-manual Other household Manual worker

Source of variance

36,47*** -1473,4 -1629,4 -1143,0 -836,6 -472,6 (a)

221,6 990,6 1602,6 498,2 -808,3 (a)

100,43***

Unadjusted main effects

221,6 990,6 1602,6 498,2 -808,3 (a)

100,43***

Model 1

6,13*** -469,4 -731,4 -432,5 -357,1 -209,9 (a)

168,4 918,5 1432,9 431,7 -725,3 (a)

67,96***

Model 2

18,74*** -852,0 -1280,6 -933,1 -716,2 -402,8 (a)

-105,2 829,2 1171,5 613,0 26,2 (a)

21,76***

Model 3

18,76*** -855,6 -1283,7 -936,1 -718,6 -404,6 (a)

-103,5 828,2 1167,0 602,1 22,7 (a)

21,27***

Model 4

17,17*** -833,9 -1234,1 -920,9 -715,6 -408,1 (a)

54,11 850,4 1120,8 574,2 21,8 (a)

19,50***

Model 5

Table 5.7 Expenditure on hedonistic consumption by independent background variables

2755 1835 2874 657 1059

Gender of HOH Male Female

Place of residence Urban Densely populated Rural

13,77*** 501,8 357,2 (a)

32,01*** 453,8 (a)

116,08*** 790,3 777,7 2180,7 -194,6 503,0 (a)

.10

.10

.17

71,61*** 558,6 1010,7 2076,0 256,8 629,0 (a)

Note: *** = p < 0.001; ** = p < 0.01; * = p < 0.05; (nfs) = p > 0.05; (a) = redundant parameter.

Adj. R

2

898 221 1088 923 395 1065

Type of household Couple Single parent Couple + children Elderly Other household Single person

.17

(nfs) (a)

69,73*** 563,0 1000,0 2084,3 252,8 629,2 (a)

.17

8,04*** 352,4 145,0 (a)

(nfs) (a)

71,70*** 593,9 1001,7 2127,2 290,4 666,3 (a)

Model 3, however, shows that the type of household appears to have a very significant impact on expenditure on hedonistic consumption. This further reduces the differences between socio-economic categories. Particularly the parameter estimate of ‘managerial or professional household’ decreases. On the other hand, the effect of income, interestingly, appears to be little stronger in the third than in the second model. Controlling the effects of type of household thus makes the existing differences between income brackets more visible. At the general level, however, it can be considered that there are no striking differences in socio-economic variation between the models. This means that there is still a socio-economic pattern indicating that the household categories of managerial or professional, entrepreneurial other than farmers, and intermediate non-manual households spend the most on hedonistic activities. Couples with children are clearly the type of household that spends the most. It is interesting to note that the households of single persons do not spend the most on hedonistic consumption. It is true, of course, that spending for instance on clothing and personal goods is typically higher among families with children. On the other hand, the finding may result simply from the fact that singles always have a certain disadvantage in terms of collective goods and services compared to households of many persons. It is often the case, for example, that a single needs proportionally more money to cover food costs than a couple living together. A total of 17 percent of the variance can be accounted for by variables appearing in the third model. In the fourth and the fifth model the explanation proportions do not change at all. Moreover, the effect of gender of HOH is insignificant in both the fourth and the fifth model. Place of residence has a significant effect, indicating that households in urban areas spend slightly more on average than households that are located elsewhere. To summarise the results of the main-effect tests: one might conclude that the findings based on unadjusted and adjusted effects are not very different. The greatest difference is of course that the effect of gender of HOH becomes insignificant when the parameter estimates ( β ) are adjusted for the independents. This is probably due to the fact that various hedonistic items that are equally typical of both sexes, such as spending on books and magazines and on clothing, were included in the sum variable.

168

The examination of interactions between the independent variables showed that socio-economic status of HOH and type of household had a significant effect. A closer examination revealed that couples with children with the socio-economic status of ‘other entrepreneur’ as their head spend clearly the most on hedonistic consumption. This finding is not necessarily particularly remarkable if we consider the type of activity under analysis, but it provides more information on the socio-economic impact on this type of consumption expenditure. The interactions between other independents were found to be non-existent at all significance levels (p > 0.05). It is possible to approach the effects of the background variables by means of main-effects appearing in Table 5.7. The share accounted for out of the effects of all variables (17 percent) in the final model is certainly a moderate one. It should be noted, however, that type of household and socio-economic status of HOH are clearly the single most important sources of variation according to the comparison of parameter estimates. In this sense, it can be suggested that the cultural characteristics of households seems have a great impact on the pattern of this type of expenditure. The items included in the scale included for example expenditure on personal goods, on books and magazines, and on clothing and footwear. It is perhaps plausible that conventional class-based tastes in fashion and style, and the habits generated in different cultural environments in general, may play an important role here, just as they can be assumed to play a similar role in expenditure on culture and luxury. Still, while it may be feasible in this context to refer to the mechanism of class, we should also remember that type of household again appears to be the most significant predictor of hedonistic expenditure. This observation does not have to contradict the suggested cultural interpretation. Moreover, the models presented in Table 5.7 indicate that socio-economic status of HOH and type of household clarify the hierarchical structure of the other’s effects. In other words, the socioeconomic groups that spend the most on hedonistic activities are easily identified on the basis of models in which the effect of the type of household is adjusted. While these two variables are clearly the best predictors of hedonistic consumption, the role of disposable household income should also be seen as relevant.

169

The last scale constructed on the basis of the factor analysis of FES98 data was that for expenditure on home equipment. This consumption dimension is analysed in relation to the background variables in the following section. 5.2.4 Expenditure on home equipment The question addressed in the final analysis procedure of FES98 data was whether expenditure on home equipment could be explained by the same socio-demographic variables as the other consumption expenditure dimensions. Table 5.8 shows the results of the main-effect ANOVA tests. The unadjusted effects are again significant for all variables. At a general level, the structure of variation is relatively similar for this type of expenditure as it was for hedonistic consumption. Certain considerable differences, however, can be observed. First, it appears that the variation according to the socio-economic status of the HOH is relatively different. Households headed by managers or professionals and by farmers seem to spend the most on home equipment. On the other hand, households with an intermediate non-manual head report spending slightly more cautiously than manual workers. The differences between types of household also seem rather interesting. Again, couples with and without children report spending the most; interestingly, however, the difference between these two is definitely not a large one. In addition, it might be observed that the relative variation by type of household is not very massive compared to the variances of most of the other dimensions scrutinised here. With regard to the other independent variables, the effect of the OECD-equalised household income appears to be as might be expected. The general trend thus shows that households belonging to the lowest income groups spend clearly less than households belonging to the highest income groups. The same can be said of differences in gender of HOH and place of residence. It is again the case that male-headed households spend more than female-headed households. Also, households located in urban and densely populated areas appear to spend slightly more than rural households. However, the unadjusted effect of place of residence is by no means particularly strong. Model 1 indicates that the socio-economic status of the HOH explains only five percent of variation. This is the lowest share of the expenditure

170

dimensions examined; nevertheless, it cannot be regarded as insignificant. In the next model, income is added into the equation. This has only a marginally weakening effect on the differences between socio-economic groups. It is interesting to note, however, that the differences between income groups decrease considerably. The proportion of variance accounted for by these two variables is six percent, which is low. Basically then, socio-economic status and income together do not explain variation more than either variable alone. Model 3 represents the adjusted effects for socio-economic status of HOH, household income and household type. The variance accounted for by these three variables rises to nine percent. It also appears that the parameter estimates ( β ) of this model are rather different from the unadjusted estimates. This means that controlling the effect of type of household thus seems to weaken the differences between some socio-economic and income groups, but not between all of them. Perhaps the most interesting result is that households with farmers as heads continue to spend the most. In addition, the variation according to type of household changes in structure if compared to the results of the unadjusted test. The spending of single parents and elderly households increases in the third model, indicating that there may be an interaction between these independents. In the last two models no significant changes are observed. The effect of gender of HOH becomes insignificant in the fourth model, but place of residence continues to have an effect in the final model. The effect indicates that urban households and household located in densely populated areas spend slightly more on home equipment than households in rural settings. Model 5 explains only a total of 10 percent of the variance. This is clearly the lowest share of all expenditure dimensions analysed in this chapter. The scrutiny of interaction effects (at the p < 0.01 significance level) revealed that socio-economic status of HOH was significant for all other background variables, except for gender of HOH. An interaction was also found between income and place of residence. The examination of these effects indicates that socio-economic status of HOH interacts with income in two ways.

171

271 1013 1274 902 543 587

Household income Less than 6500 €/year 6500-10 000 €/year 10 001-13 000 €/year 13 001-17 000 €/year 17 001-20 000 €/year Over 20 000 €/year

(cont.)

86 249 617 748 1875 1015

N

Socio-economic status of HOH Farmer Other entrepreneur Professional, manager Intermediate non-manual Other household Manual worker

Source of variance

34,75*** -625,2 -642,1 -506,8 -369,6 -194,7 (a)

336,5 275,1 355,2 -49,5 -334,9 (a)

50,52***

Unadjusted main effects

336,5 275,1 355,2 -49,5 -334,9 (a)

50,52***

Model 1

12,04*** -371,9 -415,4 -319,5 -241,3 -123,1 (a)

302,2 228,9 246,2 -67,7 -281,1 (a)

27,43***

Model 2

16,31*** -441,0 -505,8 -399,7 -303,1 -169,0 (a)

255,8 219,4 198,4 -2,0 -17,6 (a)

5,07***

Model 3

15,84*** -433,2 -499,4 -393,4 -297,9 -165,5 (a)

252,2 221,6 207,9 20,3 -10,2 (a)

4,97***

Model 4

14,82*** -426,6 -484,0 -383,7 -297,0 -166,7 (a)

301,0 228,5 193,5 11,5 -10,5 (a)

5,10***

Model 5

Table 5.8 Expenditure on home equipment by independent background variables

2755 1835 2874 657 1059

Gender of HOH Male Female

Place of residence Urban Densely populated Rural

5,13** 127,0 98,7 (a)

59,9*** 255,8 (a)

62,95*** 489,6 124,6 613,5 -69,2 246,1 (a)

.05

.06

.09

32,35*** 397,0 238,0 582,2 30,7 236,6 (a)

Note: *** = p < 0.001; ** = p < 0.01; * = p < 0.05; (nfs) = p > 0.05; (a) = redundant parameter.

Adj. R

2

898 221 1088 923 395 1065

Type of household Couple Single parent Couple + children Elderly Other household Single person

.09

(nfs) (a)

28,8*** 387,7 260,8 564,6 39,3 236,3 (a)

.10

4,16* 109,1 43,0 (a)

(nfs) (a)

29,75*** 397,4 261,4 578,1 51,1 247,8 (a)

First, households in the lowest income groups with professionals or managers as heads reported more intensive spending than the other socioeconomic groups. This finding suggests that there may be a relatively clear cultural pattern for this type of expenditure in these socio-economic positions. Secondly, households in the second highest income group that are labelled socio-economically as ‘other households’ indicated the highest expenditure of all households belonging to this income group. Socioeconomic status of HOH likewise interacted with type of household in two ways. Farmer households consisting of couples spent more than other farmer households. On the other hand, manual workers living alone also reported relatively intensive spending compared to the average pattern of expenditure by this socio-economic group in general. No clear patterns of two-way interactions between OECD-equalised income groups and other background variables, however, could be observed. Thus interactions between income and other independent variables will not be interpreted here in more detail. In this respect, it need only be mentioned that household income has significant interaction effects with other independent variables; the nature of these effects, however, cannot be interpreted by means of simple models. To sum up the results, then: no particularly clear socio-demographic pattern can be observed in expenditure on home equipment. Of course, this type of expenditure might be explained by the socio-economic status of HOH, by household income and by the type of household, the three most significant individual determinants. The effects of other variables are somewhat weaker. At the same time, however, many statistically significant interactions were found between the background variables. The results of the main-effect tests are thus somewhat problematic. It might be concluded that the impact of socio-demographic variables on expenditure on home equipment is clearly weaker than was observed in the case of the other expenditure dimensions studied. This is naturally affected also by the fact that average household expenditure on home equipment is less than 600 € a year. But in a way, it also appears that spending on home equipment varies rather irregularly between different households; the maineffects of the independent variables are relatively weak and many interactions are found.

174

In general, what the results tell us is that similar specific cultural (or other socio-demographically interesting) characteristics as in the previous expenditure dimensions might not be elaborated here as easily. Expenditure on home equipment can therefore be regarded as a typical example of a consumption activity that is connected to basic social structures, but in a rather multidimensional way. Clearly identifiable mechanisms which can be understood as determining this type of expenditure cannot be distinguished without analytical problems. In the next section, the key findings related to all the expenditure dimensions studied are summarised. 5.3 Summary of findings: Structural conditions and consumption expenditure The results presented in this chapter indicate that a large number of the selected measures of household expenditure can be reduced to a few relatively clearly interpreted dimensions. A total of four sum variables or scales were constructed, in order to determine whether they can be explained by socio-demographic factors. The independent background variables used were socio-economic status of HOH, OECD-equalised household income, type of household, gender of HOH, and place of residence. The ANOVA found considerable effects by these variables on the variances in the scales constructed. The analysis revealed that in particular expenditure on everyday life (measured as housing, vehicle acquisition and operating expenses, recreation and leisure equipment, package tours, hotels and hostels, and insurance) and expenditure on culture and luxury (measured as audiovisual equipment and computers, sport, leisure and cultural services, and restaurants and cafés) could be explained effectively by socio-demographic models. Likewise expenditure on hedonistic consumption (measured as clothing and footwear, books and magazines, cosmetics, beauty care and other small personal goods, and banking and other financial services) were found to be quite strongly connected to socio-demographic factors. At the general level, a variety of statistically significant associations were found, and some relatively clear models for the interpretation of differences could be presented. At the same time, however, analysing the structural determination of at least one of the dimensions was concluded to

175

be rather difficult. Expenditure on home equipment (measured as furnishings, art objects and carpeting, and household equipment) was not very clearly connected to socio-demographic conditions; the proportions explained were not very high, and interactions between the independents were found. A summary of the main-effect ANOVA tests for the expenditure dimensions examined is given in Table 5.9. The table shows the 2 explanatory powers of all significant background variables (Adj. R ) in the final models. The variances accounted for in these models by each 2 independent variable are reported by the Partial Eta values, which allow us to compare the proportions of total variance attributed to different 2 explanatory factors.7 The best predictors on the basis of the Partial Eta values appear first in the lists. Table 5.9 Expenditure dimensions analysed and their determinants Expenditure on everyday life 2

Partial Eta values of the determinants

2

Adj. R

Expenditure on culture and luxury

Expenditure on hedonistic consumption

Type of Socio-economic Type of household status of HOH household 0.07 0.04 0.18 SocioType of Household economic household income status of HOH 0.03 0.11 0.02 Place of SocioHousehold residence economic income 0.03 status of HOH 0.02 Household 0.01 Place of income Gender of residence 0.01 HOH 0.00 Gender of HOH 0.01 0.01 Place of residence 0.00 0.36

0.22

0.17

Expenditure on home equipment Type of household 0.03 Household income 0.02 Socio-economic status of HOH 0.01 Place of residence 0.00

0.10

In general, type of household and socio-economic status of HOH provided the best determinants for all dependent variables. In addition, however, available income of household and the place of residence were 176

also good predictors. As Table 5.9 shows, expenditure on everyday life, and expenditure on culture and luxury are explained best by the selected background variables. The shares accounted for by these variables are 36 and 22 percent. The lowest explanation proportion, at only ten percent, is for expenditure on home equipment. This dependent variable can be regarded as the most difficult dimension to explain by the sociodemographic factors used. However, it would surely be incorrect to assume on the basis of the statistical associations observed between the independent and dependent variables that these are necessarily cause-and-effect relationships. That can be ascertained only from ANOVA designs applied to experimental data. In real-life survey research, it is not easy to say which phenomena are more highly accounted for than others. At least to a certain degree, this can be seen as resulting from the nature of social life. For example, people make different choices from one time to another, and may behave totally differently even under the same conditions. It has also been noted that few 2 studies in the social science literature report R values higher than 20 or 30 percent (e.g. McNeil et. al. 1996: 297-298; Stevens 1992). Nevertheless, it is reasonable to assume that the use of several household expenditure measures makes it at least slightly easier to estimate, from the share of variance accounted for, which dimension is more structurally determined than others. It is clear that it is relatively valid to approach some forms of money allocation in terms of the mechanisms of social structures. In general, type of household and socio-economic status of HOH appear to be the most important determinants. In particular expenditure on everyday life, expenditure on culture and luxury, and expenditure on hedonistic consumption seem to be connected with these factors. Families with children with a socio-economic status associated with the upper or the middle class position spend the most on these types of goods and activities. The effects of place of residence and gender of HOH are often of lesser significance. The mechanisms of basic social structures and cultural condition of class can thus be seen as adequate. On this basis, we can say that socio-demographic factors can provide fairly valid approximations in evaluating differences in consumption expenditure. At a general level, many fields of spending are evidently structured by the socio-economic

177

hierarchy, by income level, and by family type. Households belonging to upper class positions and the highest income groups spend more on the consumption dimensions selected. The consumption expenses of families with children and couples are also typically much higher than the expenses of singles or the elderly. In the interpretation of these results, it seems in fact possible to apply the principles of several explanatory mechanisms discussed in Chapter 3. It appears that the general condition of life in economic and sociodemographic terms determines to a considerable extent the expenditure of Finnish households. Some aspects of the findings may of course raise speculative questions as to the possibility of structural interpretations on the basis of the analysis. For example, it would be impossible to evaluate comprehensively the validity of postmodern arguments merely on the basis of this analysis. It is often argued in the literature that the power of class and other modern structures is less significant in everyday life than might appear from interpretations of one type of social statistics (e.g. Featherstone 1991; Maffesoli 1996; Gabriel & Lang 1995). The role of social structures in the formation of an individual’s identity, for example, is argued to be more or less insignificant. In other words, it is assumed that individualistic notions are expressed far more visibly in subjective experiences related to consumption and other aspects of daily life. In the next chapter, I therefore explore the question whether similar patterns of socio-demographic variation can be observed utilising measures of individual consumption desires. Is it true that in Finland people’s attitudes and views about consumption are too heterogeneous to be approached on the basis of class and other structurally organised life positions?

Notes 1

None of the selected items are of course directly comparable to or show a one-to-one

correspondence with the items in the other data (see Chapter 6). This, however, is not a problem, since what is addressed in this study is whether similar socio-demographic variation can be observed between household expenditure and the consumption desires of individuals. This means that we are primarily interested in patterns of variation in different consumption fields. The different data sets are not treated as fully compatible.

178

2

In this sense, factor analysis is a technique for describing the interrelationships among many

variables in terms of a few underlying random variables or factors. There are numerous alternative procedures in conducting factor analysis. This is the case even when particular statistical software packages and extraction techniques are used. Thus I do not present any general account of model construction in factor analysis. For a comprehensive survey of different factor analysis techniques, see Harman (1976) or Tabachnick & Fidell (2001). 3

The factor scores are not directly used as new latent variables in the further analysis. This is

because it is often easier to make interpretations on the scales that are summated from the variables with the highest loadings on the same factor. If the factor scores for the constructed factors were used as new variables, then the loadings of all other variables would also be included in the scales. In addition, by using factor scores we would also miss out the possibility of examining variation in terms of average household money allocation. 4

For example, transformations based on normal logarithm functions ( e ln( x ) ) were not employed,

since our purpose is to interpret differences between household groups in terms of money allocation. Also, using logarithmic functions instead of real numbers would have levelled out the existing variation within the household groups. 5

Thus ANOVA basically assumes that there is a cause-effect relationship between variables.

The statistical procedure on which the ANOVA is used describes the relationship between cause and effect through the following formula: Yij = µ + ai + ε ij , in which µ = the grand mean, ai = the effect of i category of the independent variable, and ε ij = error. An ANOVA procedure tests whether the variation between groups is greater than variation within groups. 6

The F statistic indicates whether the effect is statistically significant for the overall model. The

significance of the F-value indicates whether the model explains a significant portion of the variation in the data; probabilities are referred to as p-values. Parameter estimates ( β ), or betas, are the original estimates of the effects of independent variables in the constructed models. R squared, i.e. the coefficient of determination, is the percent of the total Sum of Squares which is indicated by deviation and divided by total deviation. Each new variable added into the equation will increase the numerator at least slightly, resulting in a higher value of R squared, even when the new variable reduces the efficiency of the equation. In theory, then, using an infinite number of independent variables to explain the change in a dependent variable would result in an R 2

squared of one. For this reason the adjusted R squared (Adj. R ) value is preferred, since it

179

attempts to correct this shortcoming by adjusting both the numerator and the denominator by their respective degrees of freedom. 7

The Partial Eta squared ( η p2 ) is an estimate of effect size that can be assigned to a value for

each effect in an ANOVA model. The Partial Eta squared is the proportion of the effect plus the error variance that is attributable to the effect. It is therefore a value that describes the proportion of total variability (or effect size) of an explanatory factor. In a way, it is analogous to the Eta squared statistic, but it applies to single factors only. Eta-squared values describe the amount of variance accounted for in the sample, not in the population. In general, there are no commonly accepted critical values for ‘small’ or ‘large’ effect sizes. Effect sizes less than 0.10, however, are sometimes considered to be relatively small for typical independent variables.

180

6 EXAMINATION OF INDIVIDUAL CONSUMPTION DESIRES Many essential activities in our everyday lives are based on consumption practices. The fact is that we all have to spend money and make purchases in order to survive. Continuous or ubiquitous consumption can thus be seen as a powerful metaphor of our daily existence. However, it is also true that some people consume more on certain goods or services, some less. There are those who prefer for example to spend a great deal on travel, while others prefer other ways to allocate their leisure time expenditure. Such differences illuminate the central dimensions of our consumption activities and life as consumers. In the previous chapter, it was shown that considerable differences can be found between household patterns of consumption according to many basic socio-demographic characteristics. In particular variation by type of household, by socio-economic status of head of household (HOH) and by household income was found to be strong. It can be argued that imaginary aspects related to consumption will vary in a similar fashion. People belonging to the upper social strata, for instance, will probably wish to spend more on consumption activities that can be generally regarded as upper-class activities. In this chapter, the question explored is the existence of systematic differences in the ways different socio-demographic groups wish to allocate their consumption expenses. Another question has to do with the differences between various consumer activities that individuals are evaluating. The data analysed in this chapter are derived from the FIN99 survey. It is assumed that by studying individual consumption desires it is possible to address more profoundly questions related to economic and socio-demographic characteristics and the structural conditions of consumption. 6.1 Subjective consumption experience measured as consumption desire It has been suggested that subjective evaluations of people’s attitudes allow a more profound understanding of individual consumption patterns than expenditure measures alone (Räsänen 2002: 164-166; Wilska 2002). It can 183

be argued that this type of analysis can produce more realistic information about the interconnections between structural settings and consumer preferences. This means that the household level of analysis, concentrating on the measurement of actualised consumption, may fail altogether to capture certain aspects of consumer behaviour. For example, we might conclude that household budget constraints affect the formation of consumption patterns more than they actually report as expenditure. If we are primarily interested in understanding the true nature of the economic and socio-demographic conditions of consumption, one reasonable possibility is to approach consumption aspirations, whose effect might transcend realistic patterns of spending. Such an approach will yield more profound answers to such questions as which materialistic aspirations appear to be structurally determined and which are not. In order to evaluate consumption desires, we need completely different types of consumption indicators from those offered by measures of household or personal expenditure. Subjective consumption experiences are approached in this study by way of consumption desires. Respondents were presented with a list of forms of consumption, and were asked which ones they would like to increase if this were possible. It can be argued that by measuring consumer desire we can basically find out what are popular spending activities and what are not. As discussed earlier, desires indicate a willingness to actualise certain behavioural patterns. It is, however, another question whether the evaluation of a large number of items actually measures what type of consumption one would like to increase most. Expensive purchases usually prevent individuals from buying other expensive things. This is not the case for consumption desires, which are not necessarily constrained by other desires at all. In other words, there may be several different strategies for the average respondent to evaluate the desirability of the items listed in the questionnaire. Basically, however, consumption desire measures enable us to make general interpretations as to how people perceive their own abilities as consumers in contemporary society.1 The survey questionnaire consisted of a total of fifteen items that can be seen as serving the common indicators of desires related to personal consumption (see Appendix A). The list included items of leisure activities, ‘high’ culture, donations and investments, and more general indicators of

184

economic behaviour. Consumption desires were measured by means of the question, “Would you increase your personal consumption of the following goods or services if you were able to do so?” Items were evaluated using five-point Likert-type scales, anchored between 1 = ‘very much’ and 5 = ‘not at all’. A total of 13 items were selected for further analysis;2 the purpose was to capture some of the common features related to daily consumption activities. Items related to food consumption, however, were not included. While it has often been theorised in the sociological literature that food, drink and eating habits may be effective descriptors of taste and of social distinctions, it is not easy to find unproblematic measures of personal desires related to food. One option would have been to measure particular foodstuffs or product groups one by one. Many empirical studies of this kind have been reported (e.g. Tuorila 1986; Goody 1984; Warde 1997; Toivonen 1997; Alasuutari 1997). Selecting specific foodstuffs, however, would have made the analysis much more detailed and narrow than was originally planned. Furthermore, it is extremely hard to say what kind of meal reflects highly stylised culinary behaviour and which does not. Alan Warde (1997: 188-190), for example, has argued that it is impossible to address consumer tastes simply on the basis of foodstuffs purchased. Expressing style with food is more visibly linked with such activities as going to restaurants rather than selecting different foodstuffs in grocery shops and supermarkets. For these reasons, general items regarding food consumption were not included in the FIN99 questionnaire (Erola & Räsänen 2000). The items selected from the data, the number of valid cases, and their mean scores and standard deviations are shown in Table 6.1. Ascending scores in the table indicate decreasing consumption desires. Thus the items that respondents would like to increase the most appear first in the list. Holiday trips appear to be the most common item that respondents would like to increase if they were able to do so. Such activities as selfpampering and helping family members come next. It is of some interest that perhaps the most typical consumption activity, shopping, and the common economic investments, such as change of dwelling or car, come only after these. It is also interesting that art and antiques, credit cards and charge accounts are the activities that people wish to increase the least.

185

Table 6.1 Consumption desires related to the presented items Item

N

Mean

1 Holiday trips 2 Self-pampering 3 Helping family members 4 Dwelling change 5 Car change 6 Investments in stocks 7 Donations to charities 8 Cultural services (e.g. theatre, opera) 9 Impulse purchases 10 ‘Shopping’ 11 Art and antiques 12 Credit cards 13 Charge accounts

2341 2305 2328 2310 2286 2287 2308 2301 2281 2282 2287 2282 2276

2,15 2,44 2,61 3,00 3,11 3,24 3,25 3,31 3,41 3,58 3,73 4,73 4,75

Valid N (listwise)

2164

Std. Deviation 1,22 1,14 1,17 1,61 1,44 1,44 1,26 1,34 1,29 1,31 1,34 0,69 0,64

It is of course possible that the respondents may have understood the items differently. In particular the item labelled ‘self-pampering’ can be seen as ambiguous; what kind of activities people see as self-pampering vary from one individual to another. Still, it can reasonably be assumed that all the items were understood similarly enough. ‘Shopping’, for example, was presented in the questionnaire in quotation marks, referring to leisure or recreational shopping rather than the purchase of food and other necessary goods. In other words, there should be no fundamental problems in the drawing of valid comparisons among these items. At the same time, however, it should be recognised that issues of ambiguity often remain at least partly unsolved in survey studies (e.g. Haller 2002: 152-153; Räsänen 2000: 239). It is evident that the list consists of relatively different consumption categories compared to the variables selected from the FES98 data. The latter included for instance such items as total expenditure on housing, operating expenses of vehicles, and expenditure on recreation and leisure equipment, which were regarded as typical variables describing some of the basic dimensions of household expenditure. On the other hand, there were no single items in the FES98 data that measured helping family members or any form of donation. In this respect, the lack of overlap between the

186

data sets can be seen as a clear weakness of the empirical analysis from a comparative perspective. It can also be argued, however, that some of the 18 variables selected from the FES98 data set are at least partly similar to the items listed in Table 6.1. For example cultural services, travelling tours, acquisition of vehicles, furnishing, art objects and carpeting can be seen as more or less comparable items. Likewise expenditures on banking and other financial services may be seen as comparable to the credit cards and charge accounts included in the list of subjective evaluations. Moreover, such items as household equipment, books and magazines, and audiovisual equipment may be regarded as typical objects of impulse purchase and shopping. Some of the items, such as cosmetics and beauty care and personal items, can be considered to represent typical forms of self-pampering activities. Some of the items can thus be seen as similar in content. Moreover, the purpose of the study is to explore structural similarities at the actualised and imaginary level of consumption, not to provide a fully comparable analysis as such. Table 6.1 indicates that there are differences in the extent to which respondents wish to increase the consumption of the items listed. Some activities are rated relatively similarly, others rather differently. But can it be argued that certain categories of consumption desire are actually interconnected? The next step in the analysis is to explore the possible dimensions in which the consumption items selected may combine. For this purpose, a factor analysis was carried out for the 13 measures selected. 6.2 Exploring subjective consumption experiences In our exploratory search for a model, principal component analysis (PCA) was again selected as the appropriate extraction technique. The Varimax criterion was applied in the rotation procedure, to produce factors that would differ from each other. The same recommendations and criteria that were described in Chapter 5 were thus followed in the analysis of consumption desires. The number of factors in the analysis was defined on the basis of their Eigenvalue and the proportion of total variance reduced to each variable. The interpretations of the results are based on the rotated factors.

187

The consumption desire items were analysed using the abovementioned criteria for the Eigenvalues and the proportion of total variance explained. The analysis first yielded a six-factor solution. The number of factors was reduced, however, because the first factor had the highest loadings for almost every variable entered. This was not a rotated solution, but other factors did not load very high on any particular item in this extraction. In addition, the number of factors was reduced because after the first two factors the Eigenvalues began to drop heavily. In addition, the factor structures were difficult to interpret on the basis of the highest loadings. Finally, a three-factor solution that explained 52 percent of the cumulative variance was decided as the best. The extracted factors of this 2 solution, with their extraction communalities (h ), are shown in Table 6.2. The factor matrices in the table show only the rotated factors, since the interpretation will be based on them. The highest loading of each item is shown in boldface. Table 6.2 Varimax-rotated factor matrix of the selected consumption items Item

2

I

II

III

h

Holiday trips Dwelling change Car change ‘Shopping’ Impulse purchases Self-pampering Cultural services (e.g. theatre, opera) Donations to charities Contributions to fellows Art and antiques Investments in stocks Credit cards Charge accounts

0,52 0,44 0,37 0,82 0,84 0,69 0,18 0,08 0,06 0,04 0,06 0,11 0,14

0,40 0,03 -0,08 0,02 0,03 0,25 0,71 0,79 0,70 0,63 0,24 0,02 0,01

0,14 0,30 0,36 0,12 0,13 -0,01 0,01 0,07 -0,10 0,24 0,51 0,87 0,85

0,43 0,28 0,27 0,69 0,66 0,54 0,53 0,62 0,49 0,46 0,32 0,77 0,74

Eigenvalue Total variance explained

2,46 .19

2,28 .18

2,04 .16

.52

Note: Extraction method: Principal Component Analysis; rotation method: Varimax with Kaiser Normalization.

As Table 6.2 shows, the first factor loads high on popular leisure activities, such as holiday trips, shopping and impulse purchases, and also 188

on self-pampering. In addition, it loads relatively high on dwelling or car change. The second factor loads high on cultural services, contributions to fellows, donations, and art and antiques. The third factor relates fairly clearly to economic activities. It loads high on credit cards, charge accounts and investments in stocks. These themes were followed in the labelling of the factors. However, not all factors are easy to label; some of the items with high loadings seem to have very little in common. At first glance, the second and the third factor seem quite easy to label. The second factor can be labelled Culturally distinctive consumption; it loads high on activities and services that include cultural, artistic and charitable interests and pursuits. The third factor loads high on activities related to future-oriented financial activities; the factor can be labelled Future-oriented consumption. The first factor seems odd in terms of content; it loads high on two different types of activities that are not substantially related to each other. On the one hand, the factor seems to be characterised by hedonistic activities, such as holiday trips, shopping etc; however, dwelling and change of car also have their highest loadings on this factor. Still, the loadings of the latter two activities are clearly weaker than those of the first type. Thus the first factor can be labelled Hedonistic consumption. It can be argued that these three underlying dimensions of consumption desire may help us to understand how consumer experiences vary between individuals. Nevertheless, further analysis had to be performed, to find out whether any of these dimensions reflect differentiation in income, class or other socio-demographic structures. On the basis of factor analysis, three sum variables were constructed. The following variables were selected for the first summated scale: 1) holiday trips, 2) shopping, 3) impulse purchases, and 4) self-pampering. Items related to dwelling and change of car were excluded from the scale, since they cannot be considered to describe similar activities as the other items in the first factor. Furthermore, the excluded items do not load notably high on the factor. Their loadings are lower than 0.50, which can be regarded as a feasible cut-off value. The second scale was constructed on the basis of the second factor; the variables included were 1) cultural services, 2) donations, 3) helping family members, and 4) art and antiques. The following variables were selected for the third summated scale: 1) investments in stocks, 2) credit cards, and 3) charge accounts.

189

Since the summated scales represent an assembly of interrelated items that are intended to measure underlying constructs, it is important to know whether this is really the case. The reliability of the scales was therefore tested with Cronbach’s Alpha ( α ). This is a coefficient that describes interitem consistency, i.e. how well a group of items focuses on a single idea or construct.3 The Alpha indicated a reliability of 0.73 for the first, 0.70 for the second, and 0.67 for the third sum variable. It is often hard to determine what cut-off value for Alpha should be considered acceptable. Considering typical suggestions to reliability, however, the reliability of the scales can be seen as good. In addition, it should be noted that some of the items included in the scales were rather vague in nature. It should also be noted that the constructed scales probably have weaker connections with some of the independent variables than is indicated by the individual items. On the other hand, multi-item scales can be argued to achieve greater validity by combining the reliable information that all the items selected have in common. Thus it may also be easier to describe aspects of consumption desire by an index of several items combined rather than many single items individually. In any case, the general idea of relatively high reliability is that it is reasonable to analyse the summated variables as indexes. On the basis of the factor analysis, two variables were omitted from the scales. Still, items measuring the desire for change of dwelling and car contain valuable information on the basic consumption desires of individuals. It was therefore important to preserve these items in the analysis procedure. It can be assumed that the desire for change of dwelling or car represent similar desires in terms of improving one’s material wellbeing and quality of life. The two items were correlated in order to find out whether they were connected, and what the strength of the relationship was.4 The correlation coefficient between the variables was found to be moderately strong (r = 0.50). On this basis it was possible to construct a sum variable combining the items of dwelling and car change. This variable was labelled simply Dwelling or car change; it reports the desire to change one’s dwelling or car, or both. All the scales (or sum variables) and their relevant parameters – minimum and maximum values, mean scores, and standard deviations – are shown in Table 6.3.

190

Table 6.3 Scales for different types of consumption desires Sum variable 1 Hedonistic consumption 2 Culturally distinctive consumption 3 Future-oriented consumption 4 Dwelling or car change

N

Minimum

Maximum

Mean

Std. Deviation

2231

4

20

11,59

3,68

2249

4

20

12,98

3,69

2262

3

15

12,73

2,11

2260

2

10

6,14

2,43

Some points about the shape of the constructed scales can be noted here. In general, means and standard deviations indicate that the average scores are more concentrated in the centre than in the tails. According to the graphic representation of the scatterplots too, the scales appeared to be relatively symmetrical. Thus the scales can be seen as approximately normally distributed. The third scale, however, is an exception, in that it shows some skewness. Despite this, transformations from the raw scores were not employed on any of the scales. It was concluded that the third variable too was only moderately skewed. The purpose of the factor analysis was to examine whether similar dimensions would emerge from the FIN99 measures as were found for FES98 data. Now, it is obvious that the dimensions described above are not identical with the household expenditure dimensions. Especially in the respect of future-oriented consumption, it is clear that certain specific items (e.g. investments in stocks, credit cards) load high on it. These items were not measured at the itemised level in the expenditure data. The factor structures of culturally distinctive consumption and expenditure on culture are also relatively different from each other. However, there are also some similarities. In particular a dimension similar to hedonistic consumption expenditure can be observed. The two measures connected with basic activities in everyday life (e.g. housing and car expenses) can also be seen here, even though the variable of dwelling or car change was not formed directly on the basis of factor loadings. Other similarities are less evident, even though most expenditure variables seem to load on factors of a type

191

similar to the items of subjective consumption desire (see Tables 5.3 and 6.2). Overall, the factors generated by these data are relatively different from those encountered in the analysis of expenditure data. However, this does not prevent us from exploring the similarities that reflect underlying dimensions of consumer behaviour. In other words, even if the similarities between the factors constructed here and in the previous chapter appear somewhat far-fetched, this is not necessarily a problem because the aim of the study is not to compare different measures as such. The next problem in the analysis is whether the consumption dimensions indicated by the summated variables can be explained by sociodemographic factors. This question is explored in the following sections of the chapter. The purpose is also to evaluate possible similarities with the structure of variation in the household expenditure data. The constructed consumption scales will be used as dependent variables and the selected socio-demographic background variables as independent ones. The scales are treated as interval scales and the statistical method used is analysis of variance (ANOVA) (see Chapter 5 for a brief methodological discussion of this method). Main-effect ANOVA models are used to compare the significance of the effects of different socio-demographic variables for the consumption desire scales. Possible two-way interactions between the independent variables will also be discussed. I first explore the question whether certain socio-economic groups indicate a stronger desire than others to increase their hedonistic consumption. 6.2.1 Differences in hedonistic consumption The statistical procedure was carried out in two parts. First, the unadjusted effect of each background variable was tested. The effects of other variables were then controlled by entering other variables into models one at a time. Class identity was selected as the first variable for all models. This is because class identity represents the class measure in this data. Class can also be regarded as a general sociological factor through which life changes of individuals can be described. The results of the main-effect ANOVA tests for hedonistic consumption are given in Table 6.4. The statistical significance of the models is reported in the form of F values. The table also shows the parameter estimates ( β ) and the proportion of

192

2

variance explained for all independent variables (Adj. R ) in the models. It should be noted that the scales of the dependent variable are inverted in the tables for purposes of interpretation; in other words, higher parameter estimates indicate a stronger consumption desire. The table shows both unadjusted and adjusted parameter estimates for all groups of independent variables. On the basis of the unadjusted maineffects shown in the first column, the first conclusion we are able to draw is that class identity has no significant effect on hedonistic consumption desires. All other variables, however, seem to have significant effects. According to the parameter estimates, the highest explanatory power seems to be that of age; this is followed by gender and place of residence, but these independents are less powerful than the effect of age. Vocational education and OECD-equalised income of household seem to have rather marginal effects. In general, most of these results seem relatively unsurprising. For example, it is not odd that urban dwellers wish to increase their hedonistic consumption more than people living in rural areas. Similarly, women and younger age groups wish to increase their consumption more than men and older age groups. Differences between income groups indicate that the desire to increase one’s consumption is greatest in the lowest income group. This is unsurprising, considering the different budget constraints of households belonging to different income brackets. However, interesting conclusions can be drawn from the differences between educational categories. People with a lower educational background, in particular people with secondary education, wish to increase their hedonistic consumption more than university graduates. This can be seen on the one hand as a typical finding when no cultural connotations can be attached to the activity analysed. It is true that most activities included in the summated scale, such as impulse purchases or shopping, do not represent the kind of activities that are valued more highly by educated people. In particularly impulse purchases are probably greatly determined by situational factors, regardless of the kind of goods or services they consist of. On the other hand, it is interesting that people with secondary education actually express slightly stronger desires than those without any vocational training.

193

(cont.)

4,96** 0,53 0,53 1,05 (a)

560 222

(nfs) (a)

Unadjusted main effects

519 716

25 466 634 352 659

Class identity Upper class Higher middle class Lower middle class Other Working class

Vocational education Unskilled Vocational school or course Intermediate level Academic degree

N

Source of variance (nfs) (a)

Model 1

1,23 (a)

5,49** 0,82 0,86

(nfs) (a)

Model 2

1,23 (a)

5,41** 0,70 0,73

2,73* 1,73 0,61 0,02 0,46 (a)

Model 3

1,18 (a)

4,87** 0,79 0,93

3,49** 1,99 0,67 0,03 0,49 (a)

Model 4

0,94 (a)

3,92** 1,00 0,88

(nfs) (a)

Model 5

Table 6.4 Hedonistic consumption by independent background variables

0,98 (a)

4,45** 1,09 0,96

(nfs) (a)

Model 6

730 708 559 225 1712 521

Age 18-30 years 31-45 years 46-60 years Over 60 years

Place of residence Urban Rural

31,23*** 1,03 (a)

119,80*** 4,08 2,72 1,41 (a)

89,5*** -1,47 (a)

5,60*** 1,47 1,05 1,22 0,53 (a)

0.00

0.01

0.02

4,54** 1,40 1,10 1,17 0,44 (a)

Note: *** = p < 0.001; ** = p < 0.01; * = p < 0.05; (nfs) = p > 0.05; (a) = redundant parameter.

2

1015 1184

Gender Male Female

Adj. R

186 652 574 220 174

Household income Less than 450 €/month 450-850 €/month 851-1250 €/month 1251-1650 €/month Over 1650 €/month

0.06

60,22*** -1,41 (a)

3,41** 1,21 0,98 1,03 0,43 (a)

0.16

62,68*** 3,88 2,33 1,45 (a)

55,61*** -1,29 (a)

(nfs) (a)

0.17

9,03** 0,64 (a)

60,05*** 3,84 2,24 1,46 (a)

52,13*** -1,25 (a)

(nfs) (a)

In other words, university graduates wish to increase their hedonistic consumption less than any other group, but this does not seem to apply to those with secondary education. In any case, the overall impact of vocational education is fairly weak and may be even weaker when the effects of other variables are controlled. 2 As Model 1 shows, the part of the variance explained (Adj. R ) by class identity is non-existent. The overall effect of class identity is also insignificant on the basis of the value F. In the second and the third model, where vocational education and household income are entered, the explanation proportions of the models do not change very much. The variance explained in the third model is still only two percent. The only notable change is that in the third model the effect of class identity is significant. This may be due to a connection between class and income. Thus it can be assumed that certain aspects of class can be reduced to household economic resources, and vice versa. The finding is not necessarily very exceptional in the analysis of this type of items. In general, hedonistic consumption desires appear to be a little stronger among those who identify with upper social positions. Even in Model 4, however, which includes the effect of gender, the explanation proportion of all variables remains at six percent. This must be seen as a rather small share for a total of four independent variables. The relatively weak effects of class identity, vocational education and income level become even more evident in Model 5, when age is added to the equation. The effects of class identity and income are not statistically 2 significant in this model, while the adjusted R accounted for by all independents rises to 16 percent. Vocational education still has a statistically significant impact. But more generally, an examination of the fifth model indicates that age has the highest explanatory power for the variance observed. Model 6 does not change this interpretation, even though the proportion explained rises slightly. While place of residence has a significant effect in the final model, it is easy to see that hedonistic consumption desires is most effectively explained by the respondents’ age and – to a lesser degree – by gender. It is also worth noting that there were no statistically significant interactions (at a level of p > 0.05) between the independent variables.

196

To summarise the results of the analysis of the first scale: the most important finding was the absence of any actual effect of class identity. This may be due to the fact that our measure of class identity was based purely on the subjective evaluation of one’s social rank. The effects of place of residence and vocational education can be seen as noteworthy, but basically there are no strong explanatory factors for consumption desires except for age and gender. The effect of household income was found to be relatively weak. However, findings related to household income may depend greatly on the original formulation in the questionnaire of the question about consumption desires.5 In this respect we can also assume that a general wish to increase consumption is not necessarily as important for rich people as it is for poor people. This may affect the results: the differences are smaller than would be the case if some other type of measurement were used. It is also worth noting that the income measure in the FIN99 data is based on the respondents’ subjective evaluations. In the light of these findings, it is not easy to suggest any direct interpretations of the structural determination of individuals’ hedonistic consumption desires. Still, some conclusions can be drawn. One alternative is to consider why hedonistic consumption desires seem to be mainly connected to the life practices of women and adolescents, as compared to men and older age groups. Other possible points of reference are offered by the effect of living in an urban area and of not being a university graduate. The other alternative, of course, would be to see these findings as strong evidence for the validity of the arguments related to postmodernism. As noted earlier, postmodern consumers are believed to live according to their own individualistic experience, rather than according to the habits of a specific social group, and of a class-based group in particular. Still, the considerable impact of age on hedonistic consumption desires seems to contradict the likelihood of postmodern interpretations. These questions will be discussed in slightly more detail at the end of the chapter. Broader theoretical implications will be considered in the concluding chapter of the study. I next take a look at the possibility of explaining culturally distinctive consumption desires by socio-demographic variables.

197

6.2.2 Differences in culturally distinctive consumption The analysis of culturally distinguished consumption desires was carried out similarly to the analysis described above. The unadjusted main-effects for all independent background variables were first analysed separately, followed by an analysis of controlled effects. Main-effect ANOVA models were employed. Class identity is used as the first variable in all models. The results of the ANOVA tests are shown in Table 6.5. As the table indicates, each independent variable has a significant effect on culturally distinguished consumption desires. Income again appears to have a quite minor effect. Variation between the different categories of class identity and vocational education seems to be relatively great. According to the parameter estimates ( β ), however, the best predictor is evidently gender. In general, the results are in many ways fairly intelligible. For example, it is easy to understand why people with a higher educational background wish to increase their culturally distinctive consumption more than less educated people. According to many interpretations, such as the well-known one of Bourdieu (1984), activities that can be valued in terms of cultural sophistication have always been associated with educated members of the middle and the upper classes. It is true that some of the items included in the summated scale, such as cultural services or art and antiques, can be seen to measure a kind of ‘high’ culture that is generally valued among educated people. The same effect is probably reflected in the differences observed between different class identities. Those who identify with the upper or middle classes evidently express a significantly stronger desire for the form of consumption in question. The effect of place of residence is likewise somewhat unsurprising in the light of everyday knowledge. It is easy to understand that cultural activities are common leisure activities for city dwellers rather than for people outside cities. It is also a relatively normal finding that women indicate a stronger desire for culturally distinguished consumption activities (e.g. Lunt & Livingstone 1992: 94-95; Falk & Campbell 1997). However, the gender effect can be seen as quite prominent. This finding perhaps needs explanation. The finding may be influenced for example by items measuring altruistic willingness, i.e. donations and helping family members; these items might be regarded as generally more important to

198

women than to men. Some interesting suggestions about the effects of the other independents can also be made. Those in the youngest age groups have no wish to increase their consumption, but the other two age groups wish to increase their consumption more than the oldest. This seems somewhat curious. It is probably due, however, to the fact that the oldest age group consists to a large degree of people in retirement, who may not express a particularly strong desire for any type of consumption activities. It should perhaps be noted that the effect of age does not seem particularly strong either. Residential differences indicate that urban dwellers are slightly more eager to increase this type of consumption than those who live in rural settings. The explained share attributable to class identity can be seen in Model 1. It is basically a modest share (only two percent). Furthermore, the effect of class identity diminishes relatively much in the second model, when vocational education is added to the equation. We can conclude that education partly controls the effects of class identity. The OECD-equalised income of the household is added in Model 3. This does not greatly change the earlier interpretation, but the differences between class identity positions become a little stronger. Household income is no longer statistically significant in this model. The variance accounted for by these independents remains at just four percent. Gender is added in the fourth model. This weakens the effect of vocational education a little more, while at the same time the effect of class identity becomes stronger. In particular the parameter estimate of upper class identity increases in relation to working class identity. In addition, as one would expect, the total variance explained is much higher for this model compared to the previous one. Nevertheless it is just eight percent; this is not much considering that it is accounted for by a total of four different variables. Age is added in the equation in Model 5. The effects of class identity and education are still significant, and the effect of class identity once again appears to be slightly stronger. More generally, age seems to be a relatively strong source of variation, although gender continues to have the highest explanatory power. The strength of the effects encountered in the fifth model do not decrease considerably in the final model, which adds the effect of place of residence. In this final model all independent variables are significant, except for household income.

199

(cont.)

19,24*** -1,50 -1,48 0,32 (a)

559 222

11,68*** 1,11 1,44 0,94 0,73 (a)

Unadjusted main effects

510 722

25 462 627 350 666

Class identity Upper class Higher middle class Lower middle class Other Working class

Vocational education Unskilled Vocational school or course Intermediate level Academic degree

N

Source of variance 11,68*** 1,11 1,44 0,94 0,73 (a)

Model 1

-0,11 (a)

7,58*** -1,05 -0,95

4,17** 0,78 1,02 0,69 0,57 (a)

Model 2

-0,06 (a)

5,59** -0,99 -0,85

3,45** 1,08 1,08 0,52 0,53 (a)

Model 3

-0,11 (a)

3,40* -0,86 -0,63

4,22** 1,39 1,17 0,47 0,49 (a)

Model 4

-0,08 (a)

4,35** -0,96 -0,64

4,71** 1,46 1,22 0,41 0,55 (a)

Model 5

Table 6.5 Culturally distinctive consumption by independent background variables

-0,06 (a)

4,16** -0,94 -0,60

4,52** 1,12 1,22 0,41 0,58 (a)

Model 6

731 710 549 224 1709 515

Age 18-30 years 31-45 years 46-60 years Over 60 years

Place of residence Urban Rural

11,72** 0,63 (a)

7,77*** -0,18 0,34 0,79 (a)

133,34*** -1,77 (a)

2,42* -0,63 -0,74 -0,39 0,06 (a)

0.02

0.03

0.04

(nfs) (a)

Note: *** = p < 0.001; ** = p < 0.01; * = p < 0.05; (nfs) = p > 0.05; (a) = redundant parameter.

2

1014 1180

Gender Male Female

Adj. R

186 654 578 221 168

Household income Less than 450 €/month 450-850 €/month 851-1250 €/month 1251-1650 €/month Over 1650 €/month

0.08

75,72*** -1,60 (a)

(nfs) (a)

0.11

16,13*** -0,77 0,07 0,94 (a)

82,97*** -1,67 (a)

(nfs) (a)

0.11

4,90* 0,50 (a)

16,40*** -0,83 0,06 0,91 (a)

80,47*** -1,65 (a)

(nfs) (a)

A test was also conducted to determine whether any interaction effects among the independents entered into the equations. No significant interactions were discovered (p > 0.05). The final conclusion was that the variation observed in consumption desires could be attributed to the maineffects of the selected independent variables. To sum up: culturally distinctive consumption desires are perhaps not particularly well explained by the selected background variables. The share accounted for is no more than 11 percent. It might be concluded from this that socio-demographic factors play a relatively small role in the determination of these kinds of consumption desires. We must remember, however, that statistically significant effects of background variables certainly do exist, and that a relatively clear model for variance can be observed. While it is true that gender has the highest impact on the consumption desires scrutinised in this section, the effects of age, class identity and vocational education should also be recognised. In general, distinctive consumption desires seem to be at least partly structured around the basic factors of daily life, from age (or stage of lifecycle) to gender roles, class and educational background. It can thus be argued that the mechanisms of class identity and those of basic social categories can be seen as influencing the determination of culturally distinctive consumption. It should also be noted, however, that the effect of class identity on the formation of culturally distinctive consumption desires is weaker than that of certain other background factors. These issues are evaluated in more detail in the concluding chapter of the study. The next section explores possible explanatory relationships between the background variables and patterns of future-oriented consumption. 6.2.3 Differences in future-oriented consumption The analysis of future-oriented consumption desires followed the same procedure as the earlier analyses. The results of the main-effects of ANOVA tests can be seen in Table 6.6. The first conclusion we can draw from the table is that all the background variables appear to have a significant impact on the consumption desire scale; their strength, however, varies relatively greatly. According to the differences in the parameter estimates, the most powerful unadjusted main-effects are those of gender

202

and class identity; the weakest effect, on the other hand, is that of vocational education. The table indicates that those who identify with the upper class are most eager to increase their future-oriented consumption. People identifying with the upper middle classes come next. Those identifying with the working class and those who do not identify with any class position are the least eager to increase their consumption. These results are easily understood in the light of writings on class-based lifestyles. It may be typical of people with upper-class values and attitudes to consider investment and related activities as highly important. This is one reason why surprise is sometimes expressed in the sociological literature that economic inequalities are usually separated from the power of class (Savage 2000: 68-69). These of course are primarily methodological questions, but it is generally acknowledged that class can also shape economic inequalities. From this point of view, our results can be said to reflect differences in economic life planning strategies between different class positions. The effect of income of the household seems more or less predictable. Those in lower income groups wish to increase their futureoriented consumption less than people in the higher income groups; the two highest groups, on the other hand, express an almost equal desire to do so. Income, however, should not be regarded as a particularly strong predictor of this type of consumption desires. Variation between different educational categories was found to be weak. In principle, educated people tended to express a slightly stronger desire to increase their consumption. Similarly, the differences between age groups are relatively small. The mean scores are almost the same from the youngest group to the second oldest one; all the younger age groups apparently desire to increase their future-oriented consumption more than the oldest one. This makes the age-effect significant and noteworthy, even though the differences between the age groups seem relatively unsurprising. For example, it is relatively easy to assume that people in their sixties and over do not necessarily wish for more credit cards or charge accounts. They are probably not the most eager to participate in the stock market either. Their future-oriented activities are probably directed towards possible offspring, rather than themselves.

203

(cont.)

3,82* -0,43 -0,33 -0,07 (a)

563 225

13,37*** 1,79 0,66 0,13 -0,16 (a)

Unadjusted main effects

520 721

26 464 639 352 661

Class identity Upper class Higher middle class Lower middle class Other Working class

Vocational education Unskilled Vocational school or course Intermediate level Academic degree

N

Source of variance 13,37*** 1,79 0,66 0,13 -0,16 (a)

Model 1

(a)

(nfs) -

11,96*** 2,06 0,75 0,14 -0,07 (a)

Model 2

(a)

(nfs) -

8,04*** 1,89 0,68 0,17 -0,01 (a)

Model 3

(a)

(nfs) -

6,91*** 1,75 0,64 0,21 0,00 (a)

Model 4

(a)

(nfs) -

6,93*** 1,73 0,63 0,21 -0,04 (a)

Model 5

Table 6.6 Future-oriented consumption by independent background variables

(a)

(nfs) -

6,67*** 1,74 0,64 0,21 -0,01 (a)

Model 6

513 682 657 375 1713 525

Age 18-30 years 31-45 years 46-60 years Over 60 years

Place of residence Urban Rural

8,41** 0,31 (a)

4,43** 0,46 0,37 0,45 (a)

66,26*** 0,73 (a)

5,58*** -0,63 -0,43 -0,12 0,07 (a)

0.03

0.03

0.03

(a)

(a)

0.05

39,15*** 0,66 (a)

(nfs) -

(nfs) -

Note: *** = p < 0.001; ** = p < 0.01; * = p < 0.05; (nfs) = p > 0.05; (a) = redundant parameter.

2

1020 1181

Gender Male Female

Adj. R

188 650 581 221 172

Household income Less than 450 €/month 450-850 €/month 851-1250 €/month 1251-1650 €/month Over 1650 €/month

0.05

3,54* 0,57 0,58 0,67 (a)

38,84*** 0,66 (a)

(a)

(nfs) -

0.06

5,33* 0,30 (a)

3,48* 0,54 0,58 0,67 (a)

39,80*** 0,67 (a)

(a)

(nfs) -

Also perhaps as expected, men express a stronger wish to increase their consumption than women. This gender difference is the reverse of that found for the previous two consumption desire dimensions. This finding can be viewed from the perspective of typical gender roles in Western societies. The items included in the summated scale in fact measure economic activities that can be seen as more typical of men than of women; it is thus not surprising that a significant effect was found. There is also an ‘urban effect’ on future-oriented consumption; consumption desires are somewhat stronger in urban areas, but the effect is relatively weak. But do these interpretations also apply to the results of the adjusted tests? According to Model 1, the variance explained by class identity is three percent. In fact, this cannot be regarded as a considerable share even though it is by no means a negligible one. Moreover, the difference in parameter estimates between middle-class and working-class identities is relatively small; the difference between upper-class and working-class identities, on the other hand, is a notable one. Similar differences can also be observed in the second model. The effect of education, however, is insignificant when it is added to the same equation with class identity. This also holds true of income when it is added to Model 3. Class identity continues to have a significant effect, even though the variation between the different categories of class identity diminishes a little. The explained share of total variance increases for the first time in the fourth model, with 2 the addition of gender; the adjusted R , however, is still only five percent. Age and residence do not really increase the share in the fifth and the sixth models. Despite this, the effects of class identity, gender, age, and place of residence are statistically significant in the final model. Does this mean that relatively much of the variation can be reduced to the effect of classidentity? This is a possible and perhaps also a plausible interpretation. In any case, it should be noted that the single best predictor of patterns of future-oriented consumption desires is offered by gender. In the analysis of interactions, three statistically significant effects were found. This indicates that respondents belonging to different categories of one independent variable respond differently at the level of other independents. Interaction between class identity and education was revealed, along with interaction between age and gender and between gender and place of residence. The interactions between class identity and

206

education and between age and gender, however, were comparatively weak (p > 0.01); they can be considered to be of only little account for the interpretations of results here. The interaction between gender and residence, on the other hand, is more important. The test revealed that men living in urban settings actually indicate fewer future-oriented desires compared to urban women. This is an interesting addition to the results of the main-effect tests, although rather difficult to interpret unequivocally. To sum up the results of this analysis: the variance of future-oriented consumption desires could not be satisfactorily explained by the selected independent variables. The variation observed cannot be reduced to maineffect models. The first finding was that the unadjusted main-effects were significant for all independents, but in the further analyses the unadjusted main-effects were weakened by the effects of other variables. Class identity and gender have the most powerful impact, but even these are not particularly strong. In other words, it appears that there exist certain structural patterns connected with gender and class identity, but these patterns are not very visible in terms of statistical regularities. The need to explain desires related to this kind of consumption constitutes an obvious problem for arguments based on the assumption of structural effects. All in all, future-oriented consumption desires seem to be connected to some of the background factors, but these connections are to some extent multidimensional. In other words, desires differ considerably between individuals belonging to the same socio-cultural position. This conclusion is discussed in more detail in the final chapter. First, however, let us take a look at desires related to change of dwelling or car in different sociodemographic groups. 6.2.4 Differences in change of dwelling or car The variable measuring the desire for a change of dwelling and/or car was constructed from the items that were excluded from the first summated scale. A ‘dwelling or car change’ scale was formed because for most people these items are related to some of the most basic economic activities. In addition, the correlation between these two items was found to be quite strong. The analysis between the independent background variables and dwelling or car change followed the same procedure as

207

described earlier. The results of the main-effect ANOVA tests are displayed in Table 6.7. With regard to the unadjusted main-effects, the first conclusion to be drawn is that class identity is not statistically significant, while all other background variables are. According to the parameter estimates, the effects of age and place of residence seem to provide the strongest single maineffects. Men would evidently like to change their dwelling or car more than women, and so would do city dwellers compared to those living outside urban areas. Younger respondents also indicate a stronger desire for dwelling or car change than older ones. These findings are in no way unexpected. The differences observed can be understood in the light of typical gender roles and life-cycle phases. The urban effect is probably due to the fact that one’s car and dwelling represent more visible symbols of material property in urban settings than elsewhere. At the same time, we should keep in mind that rural residents nowadays are by no means a homogeneous group. Besides farmers, many other socio-economic groups also live outside urban areas. It is reasonable to assume, however, that contemporary modes of life share some common features in rural settings as they do in urban ones. There is probably a more fixed relationship towards land and dwelling. From this perspective, the differences observed in relation to the desire for dwelling or car change are perhaps completely understandable. The effect of the OECD-equalised income of the household and of vocational education appear to be interesting. First, respondents without vocational education are the group least eager to change their dwelling or car. University graduates come next. People with secondary education and with vocational qualifications are the groups with the strongest desire for a change. In terms of income group, people in the lowest income group are the least eager for a change, while variation between all other groups is almost nonexistent. It is interesting that in the other dimensions of consumption desires analysed, consumption desires did not vary particularly significantly by income. In this sense, this scale is relatively different from all other constructed scales. More generally, however, the effects of income and education according to the unadjusted main-effects appear to be quite small.

208

Model 1 shows that class identity cannot explain eagerness for a change of dwelling or car at all. The share of the total explained variance 2 (Adj. R ) does not increase in the following three models either. Despite this, the effects of education, income and gender continue to be significant. Age is entered into the equation in Model 5. In this model the effect of 2 vocational education is no longer significant, but the adjusted R of the model rises to eight percent. This in itself may not be regarded as a large share, but the strongest main-effect of all independent variables on dwelling or car change is clearly that of age. In addition, it should be noted that the proportion explained rose by seven percent (from one to eight) after age was entered into the equation. The effect of age does not decrease much in the sixth model, where place of residence is added. Residential area can also be regarded as having at least some explanatory power in the model. The effects of income, gender, age and residence are statistically significant and the proportion explained by the final model is 10 percent. This can be seen as a moderate share; gender and household income combined did not add to explanation of desires related to dwelling or car change more than one percent. In other words, the observed variation can be accounted mainly to two factors, that is, age and the place of residence. The examination of interaction effects did not reveal any significant effects (p > 0.05) for the independent background variables. To sum up: dwelling or car change appears to be a consumption dimension that does not attach itself very clearly to any cultural distinctions. The most accurate predictors of this consumption desire are age and place of residence. It is of course possible that age differences reflect generational shifts, which could be approached as culturally based conditions. Our cross-sectional data, however, do not allow the testing of this assumption at a more detailed level. In general, then, the most plausible interpretation is that younger and urban people express a stronger desire to change their dwelling or car, and men are slightly more eager to do so than women. The overall differences between income groups are quite small, but those in the lowest income group indicate a weaker desire than any of the other groups. However, none of these variables can be regarded as important sources of variance according to the shares of variance explained.

209

(cont.)

6,20*** -0,12 0,38 0,41 (a)

563 221

(nfs) (a)

Unadjusted main effects

520 716

25 461 638 348 666

Class identity Upper class Higher middle class Lower middle class Other Working class

Vocational education Unskilled Vocational school or course Intermediate level Academic degree

N

Source of variance (nfs) (a)

Model 1

0,41 (a)

6,54*** -0,16 0,39

(nfs) (a)

Model 2

0,35 (a)

2,71* -0,01 0,37

(nfs) (a)

Model 3

0,34 (a)

2,72* -0,06 0,34

(nfs) (a)

Model 4

(a)

(nfs) -

(nfs) (a)

Model 5

Table 6.7 Dwelling or car change by independent background variables

(a)

(nfs) -

(nfs) (a)

Model 6

511 681 659 374 1712 525

Age 18-30 years 31-45 years 46-60 years Over 60 years

Place of residence Urban Rural

46,66*** 0,82 (a)

48,18*** 1,76 1,10 0,52 (a)

15,69*** 0,41 (a)

4,24** -0,48 0,07 0,04 -0,02 (a)

0.00

0.01

0.01

2,65* -0,45 0,03 0,28 0,01 (a)

Note: *** = p < 0.001; ** = p < 0.01; * = p < 0.05; (nfs) = p > 0.05; (a) = redundant parameter.

2

1018 1186

Gender Male Female

Adj. R

184 654 579 219 174

Household income Less than 450 €/month 450-850 €/month 851-1250 €/month 1251-1650 €/month Over 1650 €/month

0.01

4,45* 0,27 (a)

2,36* -0,38 -0,11 -0,31 0,06 (a)

0.08

37,63*** 2,05 1,10 0,68 (a)

6,44* 0,31 (a)

4,13** -0,81 -0,13 0,11 -0,01 (a)

0.10

24,60*** 0,74 (a)

34,20*** 1,99 1,10 0,70 (a)

7,96** 0,35 (a)

3,43** -0,75 -0,06 0,08 -0,08 (a)

In the following section, the findings related to all types of consumption desire are summarised in reference to the original research questions of this study. 6.3 Summary of findings: Structural conditions and subjective consumption experience In this chapter, a factor analysis of the items measuring consumption desires in the FIN99 data has been presented. The analysis yielded certain relatively clear dimensions. A total of four sum variables (or scales) were constructed and labelled in order to explore the nature of their sociodemographic determination. However, interpreting the determination of consumption desires by the selected background variables was found to be relatively difficult. Many statistically significant associations were found, but the shares of variance explained were generally quite low. This suggests that economic and socio-demographic variables by themselves may not be adequate determinants for the proper explanation of individual consumption desires. On the other hand, these factors should be seen by as no means irrelevant to the determination of contemporary consumption experiences. The analysis indicated that particularly the hedonistic desires of individuals (measured as holiday trips, ‘shopping’, impulse purchases and self-pampering) were linked with socio-demographic factors. Likewise culturally distinctive consumption desires (measured as cultural services, donations, helping family members, and art and antiques) and change of dwelling or car could be partially explained by economic and sociodemographic variables, although to a lesser degree. Future-oriented consumption desires (measured as investments in stock, credit cards and charge accounts) were not very clearly influenced by socio-demographic variables. A summary of the ANOVA tests is shown in Table 6.8. The table 2 shows the explanatory powers of all background variables (Adj. R ) in the final models. The table also lists the relative powers accounted for by each 2 significant determinant (Partial Eta ) in these models. The background variables are listed in descending order, starting from those with the highest 2 explanatory power according to the Partial Eta statistics.

212

Table 6.8 Consumption desires analysed and their determinants Hedonistic consumption

Culturally distinctive consumption

Futureoriented consumption

Dwelling or car change

Partial Eta values of determinants

Age 0.11 Gender 0.04 Vocational education 0.01 Place of residence 0.01

Gender 0.05 Age 0.03 Class identity 0.01 Vocational education 0.01 Place of residence 0.00

Gender 0.03 Class identity 0.02 Place of residence 0.00 Age 0.00

Age 0.06 Place of residence 0.02 Household income 0.01 Gender 0.01

2

0.17

0.11

0.06

0.10

2

Adj. R

In general terms, age and gender can be considered to provide the best determinants of the consumption desires explored. The effects of place of residence, class identity and vocational education were also found to have relevant effects on certain types of desires. According to the parameter estimates, however, the impact of socio-demographic factors was found to be considerably weaker than in the analysis of household expenditure patterns. Table 6.8 summarises the finding that hedonistic consumption desires are best explained by age and gender. The share of variance explained by the independent background variables is 17 percent, which is the highest proportion of all scales examined here. Culturally distinctive consumption desires, on the other hand, are best explained by gender, age and class identity. The share explained by the independent variables is 11 percent. As already mentioned, the variance within future-oriented consumption desires is the most difficult to explain by the background variables. The most significant single impact is that of gender, but the overall share accounted for by the independent variables is only six percent. Desires related to dwelling or car change are best explained by the 2 respondents’ age and place of residence. The Adj. R is no more than 10 percent, but this may be seen as an adequate figure in relation to phenomena such as these. 213

Again, it would be incorrect to conclude from the occurrence of statistical associations that these represent cause and effect relationships between the independent and dependent variables. As noted at the end of the previous chapter, it is very hard to say – outside controlled experiments – what kinds of phenomena are more easily accounted for by certain independent variables than others. However, one reason why the shares of variance explained remain relatively low has to do with the nature of the summated scales. As described earlier, the reliability of the scales measured by Cronbach’s Alpha coefficient ( α ) were at least moderately reliable. The correlation coefficient (r) between the items measuring the wish to change one’s dwelling and/or car was also quite strong. But the respondents did not answer all the items in the questionnaire in exactly the same way. It follows that both the statistical significance levels and the amount of variance accounted for by the scales are lower than they are for certain individual items. On the other hand, the use of multi-item measures can be argued to possess certain advantages over scores yielded by single items: they can be considered to represent the concepts measured with greater depth. In our case this means that different aspects of consumption desires are described fairly comprehensively by multi-item scales. From the analysis described in this chapter, we can perhaps reasonably say that the background variables used can provide at least some valid though approximate conclusions explaining consumption. Relatively clear models for differences in consumption desires were found, even though the variables did not always increase the explained shares of variance very much. However, we need to return briefly to certain theoretical problems related to the analysis of consumption desires. The purpose of this study is to address the question whether it is feasible to refer to the idea of explanatory social mechanisms in understanding and explaining contemporary consumption patterns. It can be argued that the results of the analysis of subjective consumption desires give us some hints about the nature of certain of these mechanisms. It may well be that basic economic and socio-demographic variables are inadequate as the only determinants, but they appear to have a definite impact on the determination of consumption desires. Consumption desires are evidently organised at least to some degree along the basic social categories of everyday life. This is to say that in particular such categories

214

as gender, age or place of residence influence what kind of consumption individuals will wish to increase. It may also be possible to refer to typical modern (or postmodern) life situations and their cultural conditions and routines, which greatly influence our experience of consumption possibilities. At the same time, however, we have to be very cautious about overly broad interpretations. The fact is that the effects of many of the factors examined were found to be relatively weak. The consumption experiences examined in this chapter consisted of consumption desires. This means that actual consumption styles, whether in terms of individual or household expenditure, were not under investigation. Rather, the question addressed was whether patterns of subjective consumption desires reflect similar structural conditions to those observed in the analysis of household expenditure. There were certain substantial difficulties in achieving compatible comparisons between the two data sets in the first place. The data sets were found to be rather different in nature: the first was concerned with measures of household expenditure, the second with individual consumption desires. Thus no compatible consumption measures or identical background variables were available. In addition to these basic problems, the process of data reduction, which formed the first step in the analysis in both cases, produced somewhat different results; the PCA factor analysis of the variables selected yielded a smaller number of consumption dimensions compared to the FES98 measures. The dependent measures were also rather differently composed in the two data sets. The analysis of the second set of data provided three factors, labelled hedonistic consumption, culturally distinctive consumption and futureoriented consumption. A scale measuring dwelling or car change was also constructed. The principal components of the data analysed in this chapter provided rather different types of expenditure dimensions. While some factors loaded high on certain measures that could be seen as relatively comparable, many exceptions were also found. Thus it may not be feasible to draw direct comparisons between the scales constructed in Chapter 5 and the scales constructed here. At best this would merely lead to overinterpretation, and at worst it could obscure the basic idea of this study. It is still important to keep the two levels of analysis (household and individual) and the two types of data (expenditure and consumption

215

desires) separate. Thus there is no need to go back to the original measures in the data, as there is no need to present any (probably far-fetched) ideas as to how the different variables might be treated as similar consumption dimensions. What we need to do is to continue the discussion of the importance of structural elements in everyday life at a broader level. In other words, what is important in relation to the purpose of this study is to ask questions about the role of social structures in everyday life. We may ask, for example, whether it is still feasible to refer to typical life situations, or whether such notions should be viewed as totally obsolete. By this means, it may be possible to draw out some implications that would contribute to both theoretical and empirical research on contemporary consumer lifestyles. In the following chapter, I discuss ways in which socio-demographic factors can help to explain the determination of consumer behaviour, summarising the most significant empirical findings concerning individual consumption desires and household expenditure, and offering provisional answers to the original questions addressed in this study. The concluding discussion draws both on the theoretical ideas presented in the first chapters and on the empirical findings from the analysis of the FES98 and FIN99 data.

Notes 1

Included in the questionnaire was a set of questions in which respondents were asked to

compare their spending on certain common goods and services with the ‘average consumer’. However, these measures are not used in this study, since it can be assumed that respondents restrict their comparisons with the average consumer to their closest peers, in terms of such factors as gender, class, age, and other socio-demographics. Similar interpretations have been reported in earlier studies using the same survey data (see Räsänen 2000: 239). It has also been reported that the respondents generally answer that they spend less than average on almost everything (Wilska 2002: 199-200). It was therefore concluded that measures of consumer desires illuminate the subjective aspirations of individuals more effectively than measures based on comparisons with the average consumer. 2

Two variables were thus omitted from the analysis. The first was an open-ended question that

cannot be used here, while the second measured the wish to save money. Saving is of course an

216

essential part of daily life for most people. However, this variable was omitted since it does not measure consumption as a money allocation activity similar to the other items on the list (see Appendix A, question 36). 3

2 The Alpha coefficient can be specified by computing the formula (k / (k − 1))* (1 − ∑ (s12 / ssum )) , where

s12 denotes the variances for the

2 k individual items and ssum denotes the variance for the sum of

all items. The coefficient ranges in value from zero to one, and can be used to describe the reliability of factors extracted from both dichotomous and multi-point questionnaires or scales. The basic idea of the coefficient is that the higher the score, the more reliable the generated scale. A value of 0.7 or over has generally been considered an acceptable sign of a reliable scale (e.g. Nunnaly 1978; Alkula et. al. 1994), but lower thresholds are also used. 4

Pearson’s product moment correlation or Correlation coefficient (r) is a common measure of

the degree of linear relationship between two variables. It may take on any value between plus and minus one. The sign of the correlation coefficient (+ or -) defines the direction of the relationship, positive or negative. A positive correlation coefficient means that as the value of one variable increases the value of the other increases and vice versa. A negative coefficient indicates that as one variable increases, the other decreases. 5

Respondents were asked to evaluate how much they would increase their personal

consumption if they were able to do so (see Appendix A). It can reasonably be assumed that many respondents understood the question as meaning “if they could afford doing so”. This may explain why the straightforward effect of household income is as weak as it appears to be. The same point naturally also applies to the other consumption dimensions analysed in this chapter.

217

7 CONCLUSIONS

The purpose of the empirical analyses presented in the previous two chapters was to find answers to the following questions: first of all, whether certain important consumer activities, both actualised and imaginary, can be approached as dimensions which are reasonably easy to interpret; secondly, whether variation in these dimensions can be explained by economic and socio-demographic factors; and thirdly, whether similar socio-demographic variation is observed when consumption is approached as expenditure and as consumption desire. Patterns of consumption expenditure were first analysed using data from the FES98 survey, which measured expenditure at the household level. In addition to the analysis of actualised spending, however, consumption was also approached at the imaginary level, describing individuals’ consumption desires on the basis of data from the FIN99 survey. The results indicated the selected consumer activities and desires can indeed be approached as dimensions, and that some relatively clear socio-demographic patterns of variation can be observed in these dimensions. Comparisons between expenditure and consumption desires, however, were difficult, mainly because the data sets utilised in this study differed considerably in nature. The first data (FES98) represented the household level, the second (FIN99) the individual level. Neither consumption measures nor independent background variables were identical in the two data sets. Despite this, certain rather similar structures of variation were found between the reduced consumption dimensions and the selected background variables. In this chapter, the role of social structures in the determination of contemporary consumption patterns is discussed at a more focused level. I first summarise the results of the empirical parts of the study, examining the socio-demographic influences underlying consumption dimensions and describing the structural criteria by means of which consumer activities can best be interpreted. The validity of the proposed explanatory social mechanisms is also discussed. I then discuss the theoretical issues that have been explored in this study. Can the empirical findings reported here 219

contribute to the current sociological debate over the consumer society? Are contemporary patterns of consumption structured in accordance with modern social categories or do they tend to be socially fragmented? The aim of this concluding chapter is thus to re-examine the structural explanation of consumer behaviour and its postmodern critique, presented in the early chapters of this study. 7.1 Socio-demographic characteristics of consumption The empirical interest of this study was in the structural characteristics of consumption activities. The presumption was that since consumption is such a complex field, some of its most interesting aspects could be approached more effectively by other means than information on actual spending or budgetary decisions. It was therefore considered necessary to explore the subjective evaluation of consumer behaviour more directly. Simultaneously however, it was assumed that an examination of variation in the desire to consume different items would not necessarily illuminate variation in actual consumption patterns. On the basis of one type of analysis, we cannot draw firm conclusions about connections between patterns of money allocation and consumption desires; we can only form tentative and speculative assumptions. In some cases these connections may be relatively strong, in others they may be nonexistent. It is thus reasonable to argue that concentrating solely on one level of consumption may leave many questions linked with postmodern critique completely unanswered. The analysis of two data sets, on the other hand, allows certain rough comparisons between measures representing different aspects of consumer behaviour. In general terms, the dependent variables selected from the data sets were those measuring housing, transport, household maintenance and various cultural activities. It was considered that variations in household expenditure and individual consumption desires cannot be based on comparable measures. Despite this, independent background variables which were to some extent similar were employed in the analysis of expenditure and desires. I also attempted to select dependent measures which would be as compatible as possible. While there were certain obvious differences between the data, the analysis revealed many similar structures in the socio-demographic variation of household expenditure and

220

individual consumption desires. This suggests that socio-demographic factors may create at least partly similar structural conditions for people at the levels both of actualised spending and of imaginary consumption. It can thus be argued that structural factors are adequate in the analysis of contemporary consumer behaviour. In particular the effects of class and of the basic social categories are apparent in interpreting the results. Place of residence also appears to have an impact on certain aspects of consumption behaviour. In addition, household expenditure is restricted by economic resources; no such effect, on the other hand, was found in the case of individual consumption desires. The structure of variation by the independent variables was described at a detailed level in Chapters 5 and 6. I now discuss the structural determination of consumption more generally. I first evaluate the impact of economic resources and social class, followed by an examination of age, gender and type of household. After that, the effect of place of residence is discussed. Basically, my purpose is to clarify the possible influence of different social factors on the selected consumption activities or measures. Of particular interest are possible similarities between patterns of variation at the household and individual level. It should be noted that in the interpretation of the empirical results, the impact of structural factors is represented in the most accountable sense. This means that variation in the consumption dimensions analysed is attributed only to one or two primary mechanisms, which can be seen as providing the best explanatory source(s) for the dimensions of consumption in question. It should also be stressed that even though several significant two-way interactions of the independents in the data sets were observed, only the results of the main-effect models are discussed here.1 My aim is thus to present certain abstract examples of different types of structural conditions, and to illustrate how they can be understood to affect the consumer behaviour of individuals. The discussion of explanatory mechanisms is also relatively descriptive by nature. The methodological principles of different social mechanisms were discussed in more detail in Chapter 3.

221

7.1.1 Conditions based on economic resources As has been discussed previously in this study, economic resources can be seen as the basic factor influencing almost all forms of consumption. This is because consumption generally means that people have to spend on certain items and services if they wish to consume. Consumption activities were approached from this perspective also in this study. The economic resources of consumers were measured in both data sets as the OECDequalised income of the household. Tax-roll information was utilised to construct the income variable in the expenditure data, while subjective reports of household earnings were relied on in consumption desires data. While income cannot be regarded as the strongest source of variation for any consumption dimension, its effects were important in the analysis of the first data. Household income was found to have a very notable effect on expenditure on everyday life and on home equipment. The effect of income was in fact statistically significant for all other expenditure dimensions as well, even though the effects did not appear to be as strong as they did for one or two other independent variables. In general, the results reveal a trend according to which households in higher income brackets spend more than households in lower ones on the selected types of activity included in the dimension. Contrary to these findings, however, none of the individual consumption desire dimensions in the second data were greatly influenced by income. Only one statistically significant effect adjusted for the other independents was found: this was the desire for a change of dwelling or car. Households belonging to the lowest income bracket indicated a slightly weaker desire for such changes. This finding was considered to be relatively self-evident. However, some general interpretations can be suggested on the basis of our findings. Explanatory mechanisms related to income can be approached primarily as restrictive mechanisms. As discussed at some length earlier, this means that economic resources can be seen as creating conditions that restrict individuals from accomplishing consumption activities. For example, individuals and households are always subject to constraints of price and assets. Resources differ between income brackets and these differences affect both the actual spending and the desires of individuals. With this characterisation of the explanatory mechanism in mind, it is understandable that the effects of income are most visible in particular

222

dimensions of expenditure and desire. From the perspective of economic restrictions, it is easy to imagine that household income greatly affects how much is spent on everyday life. A major part of the dimension of expenditure on everyday life consists of housing costs. The average income level can be seen as correlating with housing costs, for example in terms of monthly rent and mortgage payments. At the level of consumption desires, on the other hand, it may be that households in the lowest income brackets indicate weaker desires for housing and car change because they know that in their current economic situation it is not very realistic for them to think about new housing or a car; these desires are out of reach for them. Despite this, the effect of income on consumption desires was by no means weak. It has been suggested that the wide range available of consumer goods and services can no longer be evaluated in the context of Engel’s law (e.g. Pantzar 1996; Uusitalo 1998; Warde 1997). This means that patterns of individual and household consumer spending cannot be reduced to financial boundary conditions. It is true, for instance, that the differences between income categories with respect to the consumption desires measured are not very conspicuous in terms of statistical significances. At the same time, however, it is still the case that the need for shelter and food takes up the largest part of budget in the average households. In this sense we can argue that the mechanism of economic restriction provides the most obvious predictor for consumer behaviour related to daily routines and necessities which require considerable economic investment. In our analyses particularly expenditure on everyday life was found to be a dimension of this type. Expenditure on home equipment represents a similar dimension of spending, since it requires continuous monetary allocation by households. New kitchen equipment, for instance, is acquired every now and then. While there are some problems related to the relative power of income as an explanatory variable, we can assume that the effect of income is an important factor in theoretical interpretations of consumer behaviour. It appears to be more important in the interpretation of actualised patterns of consumption. Income was significant for all expenditure dimensions analysed. The effects of income were also found to be similar to those of class. Upper-class families and families in high income brackets often spend the most. In this sense, it can be assumed that some features of

223

income co-occur with the socio-economic characteristics of households. This is to suggest that wealth is a resource that is connected to class and that it may be manifested in more than one way (see Savage 2000; Marshall 1997). It should also be kept in mind that income appeared to have a solitary main effect for each consumption desire dimension. This allows us to conclude that economic conditions may generally affect the distribution of life chances, even if the visible effects related to these positions appear to be relatively weak. Despite these important points, however, the effects of certain other factors were found to be far more powerful sources of variation than household income. 7.1.2 Conditions based on class According to the empirical analyses, class seems to have a considerable impact in various fields of consumption. However, not all statistically significant effects should be considered as of equal importance. While it is true that the effect of the socio-economic position of the head of household (HOH) was significant for all extracted expenditure dimensions, we may consider class to be primarily responsible for differences in expenditure on culture and luxury and on hedonistic consumption. This is because the effect of socio-economic status is relatively strong for these dimensions compared to the effects of other independent variables. Socio-economic differences in expenditure on culture and luxury and on hedonistic consumption can also be interpreted relatively unequivocally. Thus the above-mentioned dimensions of expenditure can be approached most effectively in a framework of classes. The same holds true of the cultural dimension of consumption desires: variation in culturally distinctive desires can evidently be seen as affected to a considerable extent by the respondent’s class identity. It should of course be kept in mind that class identity is not a measure of class position similar to socio-economic status. Class identity measures membership in a particular class and thus represents the purely subjective experience of a respondent. This is why vocational education was used as a control variable in the empirical analyses of consumption desires. The results indicated that class identity and education are not actually the strongest sources of variation, although they are prominent in this respect. The role of the class effect is not necessarily as obvious for the other consumption desire

224

dimensions analysed. While class identity has a significant effect on futureoriented consumption desires as well, this dimension is less clearly determined by class; in fact, future-oriented desires cannot easily be accounted for by economic and socio-demographic factors at all. The analysis of the first set of data showed that class explains expenditure on culture and luxury and expenditure on hedonistic consumption fairly similarly. Households headed by managers and professionals or by other entrepreneurs spend clearly more on these activities than households headed by manual workers. Likewise the households of intermediate non-manual workers spend more than manual workers on the activities included in these two expenditure dimensions. But there were also some problematic, yet interesting features. Especially farmers are a typical group in this respect; they spend less than manual workers on culture and luxury, but more on hedonistic activities. On the other hand, farmer households spend clearly less than managers and professionals, or than other entrepreneurs, on both culture and luxury and hedonistic consumption. The structure of expenditure thus seems to be clearly connected with the hierarchy of modern class positions. While the effect of type of household was found to be very strong for both of the mentioned consumption dimensions, it can be suggested that the differentiation of these activities is clearly based on the cultural conditions of class. Consumer choices are often influenced greatly by earlier choice and routines. Thus patterns of expenditure can be seen as being at least partly dependent on certain class-based tastes and preferences. The dimension of hedonistic consumption was measured by such activities as expenditure on clothing, books and magazines, and beauty care. The other dimension was measured as expenditure on such activities as sport, leisure and cultural services, restaurants and coffee bars, and audiovisual equipment. Interestingly, household income was not a very strong predictor of these dimensions. Thus it may well be that preferences for these types of activities are relatively highly controlled by the cultural condition of class, the influence of which may transcend budgetary restrictions. It may be concluded that there exist certain style-related differences among households in different class positions. Such an interpretation is consistent with many general ideas in the current sociology of consumption (e.g. Woodward et. al. 2000: 350-351; Kraaykamp 2002; Warde 1997). It

225

can be said that personal tastes and preferences related to cultural and leisure-time activities are deeply embedded in class cultures and can thus be seen as highly constrained or regulated by class positions. These capacities are usually automatic in their nature rather than systematically planned. This means that certain hierarchical structures are generated by learned set of dispositions that support and generate cultural codes or judgements in different social situations. The cultural condition of class thus seems a valid mechanism in explaining variation in expenditure on culture and luxury and on hedonistic consumption. On the other hand, both expenditure dimensions can also be understood as reflecting distinctive consumer lifestyles more expressively, in the classic sense as used by Bourdieu (1984) or Veblen (1994). This tradition of explanatory constructions shares many broader assumptions with the above interpretation of class cultures. It is assumed, for example, that particular cultural conditions do exist and that they are distributed more or less systematically between class positions. Structural differences between class positions also display forms of cultural capital that are in many ways unconscious and may be manifested in thought and action. In addition, however, interpretations based on distinctive behaviour also include the idea that people in different class conditions are fairly actively seeking ways to differentiate themselves from people belonging to other classes. Such descriptive terms as ‘distinction’ and ‘conspicuous consumption’ quite effectively characterise these ambitions at selfdifferentiation from others. It is true that differences are apparent especially between lower socioeconomic positions (manual workers and farmers) and upper ones (managers and professionals, intermediate non-manual workers and other entrepreneurs). On these grounds, it may be assumed that there occur certain attempts at distinctive consumer behaviour among the upper class positions in the fields of culture and hedonistic activities (see also Räsänen 2002: 173-174). The concept of distinctive consumer behaviour can also be seen as a valid explanation because the results of the analysis of consumption desires show similar differences between working-class and upper-class positions. People who identify with upper-class positions wish to increase their culturally distinctive consumption activities clearly more than those identifying with lower class positions.2 The distinctive

226

difference appears to be strong between the upper and the upper middle classes on the one hand, the working class on the other. The effect of income was controlled in the analyses, which means that economic constraints do not provide a plausible explanation of consumer behaviour here. While it is true that gender is clearly the most important source of variation for culturally distinctive consumption desires, it can be argued that at least some aspects can be attributed to class. This is because class identity offers quite a clear model for the interpretation of variation in this dimension of desires. Differences between class identifications are structured fairly hierarchically. Variation by vocational education also supports this interpretation. People with a secondary or higher education indicate stronger desires to increase their culturally distinctive consumption. On this basis, it may be concluded that educational qualifications and their linkages with class distinctions provide an important factor in explaining cultural consumption patterns beyond actual spending. Thus the mechanism of class identity can also be seen as a relevant source of explanation in the field of cultural activities. The dimension of culturally distinctive consumption desires consisted of such activities as cultural services, donations, and art and antiques. In this sense, it was interesting to find that the dimension of hedonistic consumption desires is not connected to class identity at all. The effect of vocational education is significant but relatively weak, compared for instance to the effects of age or gender. Thus only culturally distinctive consumption desires can be interpreted against class-based expressions. The finding that the dimension of hedonistic expenditure is connected to class while the dimension of hedonistic consumption desires is not is likely to result from the fact that relatively different items were included in these dimensions. For example, such activities as shopping or impulse purchases at a general level may not reflect similar cultural expressions as books and magazines or clothing at a more specific level. Despite this, cultural activities in general can be seen as representing the most important classbased activities among the consumption measures used in this study. Overall, the effect of social class measured at either the household or the individual level is statistically significant for many consumption dimensions. However, it cannot be concluded that class is a very good

227

predictor for all the consumption dimensions analysed in this study. This is because the variation observed is sometimes explained more effectively using other background variables. The interpretation of class effects is not always unproblematic either. For example, the results of the analysis of expenditure dimensions cannot be understood in terms of hierarchical class structure. The existence of a hierarchical structure is a typical assumption in all modern views of class (see Marshall 1997: 31-32; 43-33; Erikson & Goldthorpe 1992; Waters 1997). The basic finding was one of an apparent difference between the structure of socio-economic variation between the dimensions of expenditure on everyday life and expenditure on culture and luxury. We can thus conclude that class does not always provide the best means of understanding all structural aspects of current contemporary consumer behaviour. It is therefore also important to seek explanations from the basic social categories. 7.1.3 Conditions based on basic social categories According to our empirical analyses, the basic social structures – measured as age, type of household, and gender – represent the most visible factors influencing many consumption dimensions. There is reason to believe that the effects of an individual’s gender and the gender of the head of household (HOH) are connected to each other in that they represent similar types of influence. The effects of both of these factors can be approached from the perspective of gender restrictions and gender roles. Accordingly, it can be assumed that an individual’s age can affect consumer behaviour in a similar way to the type of household; both of these represent coarse measures of life-cycle phase. Accordingly, the type of household was also treated in the analyses as a measure of the ‘age’ of the household. In particular expenditure on everyday life can be explained by the type of household. Gender too is a relatively strong predictor for this expenditure dimension. In addition, the type of household is also significant for expenditure on home equipment. Variation in spending according to type of household can evidently be interpreted rather similarly in both of these dimensions. Families with children spend the most on everyday life and home equipment, while single persons and elderly households spend the least. This is not very surprising if we remember of what the dimension is actually composed. Expenditure on everyday life

228

consists for instance of housing costs, acquisition and operating expenses of vehicles, and recreation and leisure equipment. Expenditure on home equipment, on the other hand, consists of such things as furniture and other household equipment. Results consistent with this finding have been reported using a relatively similar set of consumption items from earlier Finnish Expenditure Survey data sets (see Wilska 1999: 154-158; Raijas 2000: 53-54). It would in fact be surprising if the results had been different. Basically, the type of household can be seen as affecting consumption activities as both a constraining and an enabling factor. In other words, we can assume that household spending is based on budgetary decisions that take into account both the structure of the household and the available economic resources. Presumably, for example, expenditure on everyday life is much more extensive in families with children than in any other type of households mainly because children need more space. It follows that in families with children housing costs are on average more substantial than in any other type of household. This of course is merely an oversimplified interpretation of the results. It does not take into account the effects of income, residential area and other relevant explanatory factors. Nevertheless, the type of household is the single most important determinant of expenditure on everyday life, and this simple interpretation may offer the most plausible explanation of the results. On the other hand, households with many adults can take advantage of certain consumption activities, such as acquisition of consumer durables. This is probably one of the reasons why so many significant interactions between type of household and disposable income household were found in the analyses. Gender is also an obvious predictor of expenditure on everyday life. It is somewhat surprising that male-headed households indicate clearly greater spending on everyday life than households headed by a woman. This is also the case in those models where the effect of income is controlled. However, it is true that expenditure on housing and vehicles account for the two largest shares of this expenditure dimension. The results are probably affected in particular by the fact that the dimension includes expenditure on vehicles. It is feasible to approach this difference from a perspective of gender roles. Spending on cars is often more intensive in those households where a man has the most to say on the allocation of money. This argument remains

229

consistent when we look at the gender effect in the analysis of the desire for a change of dwelling or car. It appears that gender is significant: men indicated more often than women that they would change their dwelling or/and car if they were able to do so. Either way, then, the explanatory structural mechanism of gender can be argued to be relatively simple. Certain activities are more typical of one gender than of the other. On the basis of the results, gender differences seem to have a considerable impact at the levels both of household expenditure and of individual consumption desires. The variation in consumption desires related to dwelling or car change is best explained by age. Younger people indicate a clearly stronger desire to change their dwelling or car than older people. But this finding does not have to contradict the previous interpretations. This finding can be interpreted in the light of the fact that older people have often already acquired satisfactory homes and vehicles; thus these desires will probably be less powerful for them than for younger people. The analysis indicated that the difference is particularly clear between the oldest and youngest age groups. Obviously, this supports the above simple interpretation concerning the role of age and life phase in consumer behaviour. Hedonistic consumption desires, which are accounted for best by age, can be interpreted in terms of a similar notion of age. The results indicated that the younger the respondent, the stronger the desire. This dimension was measured by means of rather abstract items, such as the desire for impulse purchases, holiday trips and shopping; many of the items included in the dimension are more likely activities for younger persons in the first place. It is thus not at all surprising that the age effect was found to be quite strong. This differentiation of daily consumption routines by age has also been recognised in previous studies (see Lunt & Livingstone 1992: 120121; Wilska 2002: 201; Chattoe & Gilbert 1999). A significant gender effect was also observed in the analysis of hedonistic consumption desires. Women would like to increase their consumption more than men, again not surprisingly considering the items included in the dimension. Shopping, for example, is much more likely to be associated with women than men. Thus this result too is rather easily understood in the light of the findings from the expenditure data.

230

We can conclude, then, that structural conditions based on basic social categories can be approached quite well from a general perspective. This means that the mechanisms of the basic social categories may have many origins, ranging from social group identities and gendered practices to legal restrictions. In any event, age, gender and type of household reflect some of the most basic factors influencing everyday life. Their effects can be approached in terms of specific social and cultural conditions. From our empirical analysis, it appears that especially consumption activities related to daily routines and the necessities of life are strongly dependent on conditions generated and maintained by the basic social categories. Thus the mechanism of basic social categories can be understood as an explanation that is based on ordinary conditions associated with the typical (modern) way of life. In addition to economic resources, class, age, gender and type of household, place of residence was also found to be significant in the empirical analyses. I next summarise the effect of place of residence on the consumption activities analysed in this study. 7.1.4 Conditions based on place of residence There were no consumption dimensions in the data sets that could be explained primarily by place of residence. Regardless of this, place of residence should not be ignored in interpreting the results. Residential area can be considered of great importance in determining on the one hand expenditure on culture and luxury, on the other consumption desires related to change of dwelling or car. At a general level, expenditure was higher in urban than in rural settings, except in the case of expenditure on everyday life. The results indicated clearly that people living in densely populated areas spend the most on culture and luxury. Consumption desires for change of dwelling or car were likewise greater in urban settings. This urban effect is also noteworthy for hedonistic consumption desires. Otherwise the effect of place of residence was relatively weak for the dimensions analysed. Place of residence was measured using compatible categorical measures in both sets of data. Thus the interpretation of the results is more or less unproblematic. Basically, the effect of place of residence can be approached in terms of restrictive and enabling factors. This means that

231

one’s place of residence area creates conditions that can either restrain individuals from carrying on certain activities or enable them to do so. Supply and demand may also differ between residential areas; this is reflected at the actualised and imaginary levels of consumption. The effect of the place of residence may be significant specifically for expenditure on culture and luxury because the supply of this kind of activity is more diversified in urban areas. The measure included such activities as expenditure on cultural services, restaurants and coffee bars. It follows that people in urban environments are able to adopt and maintain different consumption patterns than elsewhere. This suggests that the key to the mechanism of residential area is in the qualitative differences connected with one’s surroundings. This interpretation is consistent with many theoretical writings related to the special character of urban consumer activities (see Fischer 1995: 568-569; Hannigan 1998). It has been suggested for example that explanations of urban-rural differences can be attributed to the fact that urbanism promotes more diverse life practices within communities. In contemporary cities, daily activities are often described as being consumption-oriented (e.g. Lehtonen & Mäenpää 1997; Miles 1998). This urban effect is also reflected, though in slightly weaker form, in the other expenditure dimensions. On average, urban dwellers spend more than rural dwellers on cultural and hedonistic practices. Urban dwellers also indicate a slightly stronger desire for hedonistic and culturally distinctive consumption activities. At the level of consumption desires, however, the difference between urban and rural areas is most obvious when it comes to change of dwelling or car. This can be considered an interesting finding; unlike rural areas, in the city possession of a car cannot be regarded as a necessity. This means that in urban settings car ownership can be seen as an “important piece of symbolic display” (Chaney 1990: 52), whereas in rural settings a car may still represent a means of conveyance in its purest sense. Perhaps an analogous interpretation of symbolic display can also be applied to consumption desires related to change of dwelling. It may well be that the home has a greater symbolic value in urban areas than elsewhere; in rural area home is perhaps more clearly associated with one’s ‘own plot’ and ‘own land’. From this perspective, both the car and the dwelling represent means of displaying material wealth or personal style. This interpretation

232

offers at least one possible explanation of the finding related to dwelling or car change. It should also fit in relatively well with our broader theoretical construction, the mechanism of residential area. Place of residence seems to influence many culturally oriented consumption patterns with a symbolic value. These activities are more typically urban than rural in nature. This becomes most evident when we look at the variation in expenditure on culture and luxury and in the desire for a change of dwelling or car. These two dimensions represent relatively different activities, but both involve a concept of symbolic value one way or another. In this sense, differences between residential areas can be seen as providing a plausible explanatory pattern in a mechanism-based framework. The findings can be seen as resulting primarily from differences in level between the measures in the data sets. As already mentioned, while the differences between the measures in the FES98 and FIN99 data sets prevent us from making detailed comparisons, they also improve our ability to draw generalisations from the results. In the following section of this chapter, the discussion moves towards a more abstract and conceptual level. What kind of theoretical conclusions is this study able to offer on the basis of the theoretical exploration and the empirical analysis? 7.2 Theoretical implications The research problems of this study were addressed following structural theories of social action. From the conceptual discussion and the empirical results, it is clear that most consumption activities are influenced by the structural basis of modern social factors. The empirical findings allow us to put forward certain theoretical arguments that stress the role of structural determinants in the analysis of consumer behaviour. Hence, we may argue that if the structural bases of society continue to provide “a link between the organisation of society and the position and behaviour of individuals” (Breen & Rottman 1995: 455), then the importance of structural factors also needs to be taken into account in theoretical considerations. This assumption is elaborated in the following sections. An overview of structural conditions and the consumption dimensions analysed in this study is shown in Table 7. Each dimension in the table is attributed to the best explanatory condition, which is marked with X. The

233

comparison of the relative strengths of the explanatory factors is based on 2 Partial Eta statistics. However, some dimensions are attributed to more than one structural condition. This is because strong effects of many variables were found for each dimension. Table 7 Consumption dimensions and their connections with structural conditions Income

Class

Basic social categories

X

(X)

X

(X)

X

(X)

(X)

X

X

X

(X)

X

Place of residence

Dimension Everyday life (expenditure) Culture and luxury (expenditure) Hedonistic consumption (expenditure) Home equipment (expenditure) Hedonistic consumption (desires) Culturally distinctive consumption (desires) Future-oriented consumption (desires) Dwelling or car change (desires)

X

(X)

X

X

(X)

(X) X

(X)

(X)

X

Note: X = Primary and/or strong explanatory effect; (X) = Secondary and/or weaker explanatory effect.

The most important secondary effects are also presented in the table, marked with (X). Secondary effects can be understood as having quite considerable power in explaining variance in the consumption dimension in question, but they are not considered to be the primary sources of variation. Income, for instance, was interpreted as having a notable effect on expenditure on culture and luxury, but the variation in this dimension was more clearly based on class. On the other hand, gender and class identity were found to significantly affect future-oriented consumption desires, but this dimension is not primarily accounted for by any structural factors at all. This is because each background variable was a relatively weak source of variation in comparison with its effect on the other dimensions analysed. 234

Comparisons of the proportions of total variability attributable to different independent variables were presented in Chapters 5 and 6 (see Tables 5.9 and 6.8). Perhaps the most prominent observation summarised in Table 7 is that income does not have an explanatory effect on every dimension. This results, of course, from the fact that statistically income did not always have a high explanatory power. With the exception of dwelling or car change, the effect of income was not even statistically significant in the final models of consumption desires. This, of course, does not mean that there are many consumption activities in our daily lives that are not affected by economic factors. The findings from the empirical analyses are more likely to suggest that there exist certain hierarchies in our consumption: certain activities are more directly influenced by income, others more indirectly. The table illustrates quite consistently that there are varying conditions related to consumer behaviour and that these conditions are based on different (primary and secondary) structural factors. In the following, certain theoretical implications of this study will be presented. First, I briefly discuss the usefulness of the measures selected and the idea of social mechanisms in the explanation of consumption activities. After this, I make certain suggestions as to how the role of structural determinants should be taken into account in current theoretical debates. I should perhaps also point out that my purpose is not to construct descriptive models for application in the theoretical or empirical investigation of consumption. Rather, my aim is to explicate certain enabling and constraining principles for individual consumer behaviour that continue to be created by the structural conditions of society. Also, while at least some aspects of current approaches stressing the expanded role of consumer society will be questioned, the discussion is primarily confined to the existing economic restraints and to the conditions of sociocultural regulation of consumption. 7.2.1 Specific characteristics of the measures In many contemporary views it is stressed that general propositions about consumption practices are impossible because different areas of consumer culture and different commodities do not operate according to a single rationale (e.g. Fine & Leopold 1993: 7-11; Warde 1997; Miles 1998). In

235

other words, consumption behaviour is not a coherent field; thus its determination cannot be explained by means of only a limited number of explanatory factors. It is true that what we are able to say about one area does not necessarily apply to another. For example, our empirical analysis indicated that household expenditure on items connected with domestic maintenance and daily necessities was best explained by the basic social categories, such as type of household. Expenditure on various cultural activities, on the other hand, seems to be more clearly influenced by socioeconomic differences. Moreover, sometimes consumption desires related to similar types of items vary relatively independently of the patterns of household expenditure. These findings raise doubts about the potential usefulness of the measures in the analysis of consumption in broader terms. How can we reply to these doubts? The differences in the explanatory models of the dimensions analysed indicate that a single type of consumer activity can be connected to sociodemographic factors through various mechanisms. For example, all basic social categories can be seen as providing many explanatory processes, depending on the type of consumer activity in question. In the case of certain daily activities related to such items as housing and household maintenance, for example, this mechanism can be approached as the lifephase effect. In certain other types of activity, such as hedonistic or cultural activities, the mechanism of basic social categories may be interpreted against more general notions of class-based lifestyles and their connections with age and gender roles. What is important is that the interpretation of explanatory principles should always take place in the context of the phenomena in question. In certain respects, however, all the consumption dimensions constructed in this study – expenditure on everyday life, expenditure on culture and luxury, expenditure on hedonistic consumption and expenditure on home equipment, as well as hedonistic consumption desires, culturally distinctive consumption desires, future-oriented consumption desires and desires related to change of dwelling or car – can be understood rather similarly. First, they represent fields of consumption that are presumably constrained by many earlier decisions, acquired customs and habits. This means that amounts of household spending or an individual’s desires are not likely to change very rapidly. For example, if a person has become

236

accustomed to spending relatively much on travel, clothing, or dining out, that person will presumably not change his or her habits overnight. The destination of holiday trips or favourite restaurants may of course change relatively often for most people, but not the habit of travelling and eating out as such. This suggests that the consumption activities analysed in this study represent relatively important objects for the respondents. In effect, while it is true that such activities as for instance housing or beauty care can be assumed to be mundane, they are also integrally joined with many other activities. Our daily existence is constructed on the basis of cumulative styles of doing things, in which involvement in a certain type of activity often leads to several related activities. Accordingly, the consumption measures used in this study can be seen as representing certain meaningful aspects of life more generally than any one item treated separately. Secondly, in the case of most of the measures, if a person or a household has made a decision to spend on one item this also constrains their possibility of spending on something else. Buying a new car, for example, normally prevents one from making some other expensive purchase. Consumption desires may of course not operate exactly in accordance with this logic. But it is also true that evaluating many items results in personal rankings of the type of consumption one would like to increase the most. It can be assumed that these lists operate at least partly in a similar fashion: evaluation of the desirability of one option affects the evaluation of others. The most desirable items are often grouped together and evaluated from the broader perspective of personal values or attitudes. This also indicates that the consumption dimensions constructed here reflected some quite meaningful aspects of life. In principle, then, the empirical investigations reported here allow us to formulate certain general propositions. This is because the study did not concentrate solely on one or two types of activity, but used many different items to measure consumption behaviour. This enables somewhat broader options in discussing the general conditions related to consumption. In addition, consumption was approached by means of both household expenditure and subjective desires. This makes it possible to determine the consumption preferences shared by particular socio-demographic groups in other terms as well as merely patterns of money allocation. At the same

237

time, however, it should be kept in mind that the theoretical propositions presented here are perhaps best applied specifically to the analysis of those consumer objects that were measured empirically in this study.3 7.2.2 Importance of explanatory social mechanisms The interpretation of the results was based on the idea of social mechanisms, that is, the explication of what actually produces the relationship between the observed phenomenon and what is believed to explain it (e.g. Hedstöm & Swedberg 1996; Esser 1996; Hedström & Swedberg 1998). In this study, the explication of mechanisms was held to be of crucial importance, because empirical results can provide only descriptive evidence about the phenomena in question. But how in fact are mechanism-based interpretations considered to contribute to our understanding of the nature of consumer behaviour? The common assumption in social science is that the significance of theoretical approaches should be evaluated in terms of their empirical relevance. The effects of the background variables were explored by following structural assumptions as to how economic and different sociocultural conditions may affect individual behaviour. Such attributes as age, gender, occupational status, place of residence etc. were found to explain relatively much of the variance in many fields of consumption. These interpretations were based in the first place on empirical results. Since, however, certain factors were of more importance than others, the interpretation of the best explanatory factors made use, in addition to statistical tests and parameters, also of ideas derived from theoretical ideas of structural determination. The notion of systematic associations between independent and dependent variables did not seem adequate, and the focus was directed to the possible processes causally underlying these effects. What are the reasons for these differences in the patterns of household expenditure and individual consumption desires? The primary concern was not with establishing general explanations for consumption as such, but rather with finding explanations for a limited and specific range of consumer activities. As has long been known, any small element of individual action is determined by the total make-up of the person at the moment as well as the total situation in which he or she finds himself (Lazarfeld & Rosenberg

238

1955: 393). With this in mind, it can be argued that the mechanism-based interpretations referred to in this study can explain the existing associations between economic and socio-demographic factors and certain customary patterns of consumption. Thus our interpretations may not provide finegrained explanations for instance of how a consumer decides between two similar products in the supermarket. Satisfactory descriptions of such phenomena would require explanations deriving from different theoretical assumptions, such as psychological treatments of rationality and individual cognitive abilities. In other words, while our mechanisms are of some generality, they definitely do not offer universal explanations of individual consumer behaviour. By focusing on why-questions, the intention was to offer feasible comparisons of the effects of different socio-demographic factors. The purpose of these comparisons was to estimate the possible influence on consumption of the variables representing structural determinants. Although the mechanisms of explanatory variables were actually observable only in their effects, it was possible to arrive at a number of conclusions as to possible connections between habits, intentions, likings and preferences on the one hand and particular socio-demographic background variables on the other. Likewise the influence of economic factors was approached in the framework of a restrictive mechanism. Overall, the idea of social mechanism was highly useful in making it possible to integrate empirical results with theoretical assumptions. This method of ‘understanding’ statistical regularities explained why different social groups seem to have systematically varying patterns of expenditure; they also report varying desires to consume. It was thus possible to offer certain meaningful explanations for the empirical observations. Now let us turn to a discussion of how structural conditions in general should be taken into account in theoretical approaches to consumer behaviour. 7.2.3 Individualistic features and structural conditions of consumer behaviour According to almost all ideas related to consumer society, consumption preferences can be seen as reflecting our lives as both consumers and ordinary citizens. This means that consumption is described as one of the most visible aspects of daily life. Further, as many contemporary writers

239

have argued, consumption has become a powerful metaphor of our entire existence (e.g. Campbell 1995; Bauman 1998; Lunt & Livingstone 1992). It is assumed that people increasingly evaluate various social phenomena according to the principles of the market economy. The life aspirations of the average citizen consist merely of practices which are oriented towards buying commodities. According to this slight exaggeration, shopping for example has been conceptualised in terms of a plausible characterisation of how people generally spend their time (Lehtonen 1994; Falk & Campbell 1997: 2-3). Consumption activities are considered to reflect much more than just spending money: they play an essential role in the maintenance of daily routines and lifestyles. According to a conventional structural view, the connections between socio-demographic factors and personal consumption patterns continue to provide the most important research tasks for the sociology of consumption (e.g. Toivonen 1997; Räsänen 2000; Woodward et. al. 2000; Kraaykamp 2002). In particular the study of the relative powers of class and other traditional sociological variables has been seen as highly relevant and interesting. Structural views share the assumption that many structural constraints still persist in contemporary late-modern market economy societies. It follows that many basic social differences affect individuals’ life chances in terms of consumption possibilities as well. The postmodern critique, however, has questioned the validity of these assumptions. Perhaps the most distinct characteristic of postmodern theory is its lack of explicit references to socio-demographic structures. Most theoretical writings are connected with general discussions about ongoing socio-cultural changes in society. The key tendency for many theorists is individualisation; Anthony Giddens and Ulrich Beck are probably the most prominent and explicit writers in this respect (Warde 1997: 14-15; 181182; Savage 2000: 101-102). According to these writers, contemporary culture in advanced societies has broken away from the collective cultures of the modern age (Giddens 1991; Beck 1992; see also Beck-Gernsheim 1996; Beck et. al. 2003). Briefly, Beck and Giddens insist that social classes no longer hold the place of status groups, or the family as the stable frame of one’s life world. Life choices have become progressively independent of income, and thus even of labour force participation (Beck et. al. 2003: 7). Individuals need to reflexively construct their own

240

identities in a fragmented world. It follows that individuals themselves become reproductive units and that the surrounding social structures lose their integrative influence on behaviour. Many contemporary writers on consumer behaviour share these assumptions concerning the new distribution of collective factors (e.g. Featherstone 1991; Hetherington 1998). It is believed that consumer behaviour represents the freedom to choose a new social identity and a new self. Some writers argue that most people are less restricted in the field of consumer activities than in any other fields of their lives (see Bocock 1993; Noro 1995). Thus in the field of consumption individualisation is interpreted as meaning that the reduced importance of modern social categories either obliges or enables people to choose different patterns of consumption. It follows that consumption styles are becoming increasingly expressive and individualised. Consumption consists of a field of unpredictable activities; it is determined individually with reference to changing attitudes and personal interests. This means that consumers have been released from the pressure towards uniformity within large social groups. In particular the normative regulation of the class-based consumer cultures is believed to be declining. On this basis, the exploration of the structural characteristics of consumer behaviour may not be seen as either relevant or interesting. Other postmodern theorists stress the role of new group formations which are disengaged from traditional social networks and groupings, again in particular from classes. Perhaps the most consistent view with respect to new kinds of social communities has been presented by Robert. N. Bellah and his colleagues. Basically, Bellah et. al. (1985) argue that many activities in contemporary life are quite loosely organised by many simultaneous social groupings, but that the organising forces of these may be experienced as important by the participating individuals. Typical forms of such groupings are ephemeral communities, so-called ‘lifestyle enclaves’. They are considered to exist only because of shared (leisuretime) activities; participants have little in common in terms of their sociodemographic conditions. In the sociology of consumption, many concepts have been proposed similar to these ‘lifestyle enclaves’ (e.g. Maffesoli 1996; Sulkunen 1992; Tomlinson 1990; Bauman 1997). Contemporary consumer life is

241

considered to be marked by membership in overlapping groups; identification with these is relatively flexible. Fashion subcultures and other interest-based groupings, for example, are seen as typical forms of new social groupings. The collective expressions of such social groupings are considered to be temporary and fleeting. In these arguments it is suggested that consumer patterns are inconsistent and fluid, and cannot easily be mapped against typical structural positions. In theoretical terms, then, there are actually three ways to approach the nature of consumer activities. The first option is of course to approach consumption as an activity that can be sufficiently interpreted and explained in terms of socio-demographic categories. This position is labelled here as Structurally patterned consumption, and it can be seen as representing the traditional sociological approach. According to this position, contemporary consumption patterns continue to be highly influenced by socio-economic and other demographic conditions. The second option is that there is great diversity and that individual consumer choices are not constrained by socio-demographic factors. This approach follows individualistic views and can be labelled the position of Fragmented consumption. According to the fragmented position, current individual consumption patterns are highly changeable and unpredictable in a collective sense. According to the third option, consumer behaviour has become detached from socio-demographic structures but this has not led to a patternless individualisation. Instead, various other groupings and lifestyle communities are considered to influence the formation of individual consumption patterns. This position can be referred to as Segmented consumption; it takes into account the growing importance of new forms of groupings and social networks. These three positions and their interrelations are represented in Figure 7. The different positions are placed in the figure along vertical and horizontal trajectories. Basically, the vertical trajectory illustrates the continuum between individualistic and structural consumption patterns. At one end of the continuum it is assumed that consumption can be approached in a framework of clearly distinguished patterns, at the other end that heterogeneity and unpredictable consumer styles offer a better explanation. The horizontal axis, on the other hand, illustrates the dichotomy between (modern) socio-demographic explanations and their

242

(postmodern) critique. One end assumes that consumption is strongly connected to socio-demographic categories, the other that these connections have become insignificant. Thus both the fragmented and the segmented approaches to consumption can be understood as entailing the weakened power of modern social structures.

Structurally patterned consumption

Segmented consumption

Individualistic-Structural

Fragmented consumption

Modern-Postmodern

Figure 7 Different approaches to consumption The empirical relevance of all three positions in the figure cannot be evaluated effectively here. Postmodern considerations are puzzling in many ways, making it very difficult to test individual ideas systematically. The second problem is that most postmodern writers do not actually seek empirical support for their arguments. Preferably, they claim that this type of theory-construction is not the kind that needs adjudication from empirical research in the first place (see Gregson 1991; Noro 2000). These problems were pointed out in Chapter 3. In addition, the background variables in the data sets used here did not allow testing for the possible effects of new social groupings and communities on individual consumption patterns. Thus the discussion has to be restricted to the significance of socio-demographic factors. In other words, we are only able to evaluate whether claims as to social fragmentation and segmentation, and the weakened effect of modern structures, seem valid or not. The empirical analyses of this study do not yield information on other aspects of postmodern approaches. The empirical findings indicated that most consumer activities can be approached rather painlessly from a traditional structural approach. It 243

appears that different kinds of resources that accumulate over the course of the lives of individuals can be observed from their patterns of consumption. In other words, many activities can be argued to follow the ‘normal’ dispositions of modern life. This is to say that certain spending patterns can be seen as more typical of certain demographic groups than of some others. Typical spending patterns can thus be understood in reference to the fact that consumers remain influenced by socially and culturally absorbed habits and economic restrictions. It is also clear that particular structural resources obviously work better in some fields than in others. Despite this, one of the most important findings was that neither social class nor any other single background variable could be identified clearly as the most important explanatory variable. On the contrary: it was found that there is hardly any field of consumption that can be explained sufficiently in terms of one or a few socio-demographic factors. Adding many variables into the models resulted in higher explanation proportions. On this basis, it can be claimed that structural processes based on socio-demographic conditions are complementary rather than alternative sources of explanations.4 The intersections of these factors provide the best possibility of understanding structurally patterned conditions of consumption. These ideas set the stage for a final theoretical recapitulation. In general, it can be argued that in consumer research the role of sociodemographic factors should be taken into account as persistently influential background features. In exploring the nature of consumption experiences and preferences, we should also question how these might be connected to individual socio-demographic conditions. For example, contemporary writings on the formation and maintenance of consumer identities might be sociologically more convincing if identities were approached from a perspective stressing the constraints of various economic and socio-cultural restrictions. Perhaps we will also be able to accept theoretical notions on the increased importance of consumer identities more easily if they are not seen as categories based merely on individualised choices. In many ways, it makes sense to assume that the society we live in is very much a society of consumers. This means that the use of products and services for purposes other than the satisfaction of needs has become more commonplace. Consumption is seen as a visible metaphor of our existence, along with typical social roles in working life or the family. In this sense,

244

contemporary developments in the focus of life can be understood to reflect postmodern consumerism. However, accepting this shift does not mean that structural inequalities in society should be forgotten. It continues to be the case that those who possess better resources have more choice as to how to use them. The basic economic and socio-cultural factors can be seen as important resources that can be allocated to consumption activities. On both the theoretical and empirical grounds of this study, we can say that it is not common in real life to encounter many individuals that are able to choose their spending style and identity practically freely. At the current stage of the Finnish consumer society at least, this is not the case. We are more likely to find various economic and socio-cultural structures which allow individuals to spend more on some things and less on others. This claim can be made with regard to current notions on the adaptation of consumer lifestyles as well. Lifestyles arguably cannot be regarded as individualised habits, which can be altered simply by changing one’s consumer routines. It is structural factors that create dispositions, where a particular set of values, beliefs and attitudes tends to reinforce certain preconceived images of oneself as a consumer and as a member of social category. Thus the structural framework offered in this study can be considered to have great potential in understanding the nature of social relations. It is true that one major trend in current writings on consumer behaviour is to highlight the fragmentation of various lifestyle- and identity-sustaining processes (Lunt & Livingstone 1992: 25-25; Wilska 2002: 195-196). According to many postmodern writers, the symbolic meaning of one’s activities does not express any particular structural position in society. These meanings are created personally in the course of shopping activities and other daily routines (e.g. Lehtonen 1994: 194-195; Gabriel & Lang 1995; Falk & Campbell 1997). Basically, these arguments indicate that lifestyles and identities are continually constructed rather than influenced by membership in permanent social categories. However, Daniel Miller and his colleagues (1998) have noted that the the empirical analysis of typical shopping activities includes far more complex issues than a certain free plurality of agents choosing their consumer products and experiencing new social positions by doing so. According to Miller and his colleagues, there are complex processes related to confirming one’s public

245

standing by means of consumption, ranging from disposable money to adopted customs, the wish to meet people from a similar background, and so on. The constellation of values represented as the experience of social categories is powerfully expressed in the form of shopping as well as in the individual consciousness (Miller et. al 1998: 187-188). It makes sense to argue that the same is true of consumer lifestyles more generally: people cannot simply choose a lifestyle from every imaginable option offered in the media, in advertisements or in public discourse. It is true, of course, that lifestyles can be characterised as processes which consist of actions and routines engaged in over the course of time. But there exist certain boundary conditions for lifestyle practices, as there exist for every human activity. In the real world, lifestyles can be seen as consisting of practices that lead to a number of interrelated activities; together these provide a coherent basis for the organisation of daily activities. Organised activities naturally also provide one source of social identity, but less detached from its socio-cultural connections. What should be stressed is thus the idea that individuals’ lifestyle activities are structured at least to a certain degree and that these underlying structures can be identified by means of sociological variables. The most important factors that tend to determine distinct lifestyles in our daily lives are probably social class, age and gender. It should be noted that I am not actually assuming that lifestyle is the determinant of one’s consumer behaviour. Rather, the explanatory mechanisms for differences in consumer behaviour are sought among other relevant life conditions. The results from this study suggest that it is important to address the interplay of many structural factors, often related to economic resources and membership in various social groups. Of course, certain patterns of consumption may well be collective in other than socio-demographically identifiable ways, and certain other patterns may be primarily individualistic. Moreover, it can be assumed that many people move with relative ease between different structurally patterned practices. An individual may prefer, say, spending relatively much on opera and theatre and also on popular music. Nevertheless, we can argue that the general logic of consumer behaviour may be best understood as the interplay of various institutional forces. Moreover, it is reasonable to assume that the

246

influence of these institutional forces will continue to contribute to the stability of behavioural patterns in the near future. It has been suggested that extreme theoretical oppositions often fail to capture the complex interweaving of social structures and individual behaviour (Savage 2000: 101). The analytical contradictions between traditional collective views and new individualised ones can be seen as one example of this failure. This is the case particularly because individualised views are often connected with conceptions of postmodern and postmodernism, which have been held responsible for many of the ills besetting contemporary social science. However, we can also resist the use of these terms altogether. We can agree that consumption behaviour results from both structural influences and individualistic decisions. From this point of view, all three trends represented in Figure 7 probably co-exist, even though many activities can be approached effectively from a structural angle. In other words, we should not shut the door on the possibility of exploring alternative interpretations. It is ultimately a matter of definitions and measurements that determines how we categorise the main trends of current consumption. Under these circumstances, the most visible characteristics of consumption seem somewhat unsurprising from the perspective of modern sociology. In conclusion: this study lent support to the assumption that sociodemographic structures determine both patterns of spending and consumption desires. Structural conditions should thus be recognised in all theories that try to capture certain essential characteristics related to contemporary consumer practices. How fragmented and segmented consumption positions may be reflected in an individual’s behaviour, and how they may affect lifestyles and experiences of social identity, are altogether different empirical questions. However, it is sensible to assume that approaches based on fragmented and segmented positions may become sociologically more relevant if structural characteristics based on traditional sociological variables can no longer be identified.

Notes 1

It is necessary to note that the interaction effects of independent variables may also help us to

interpret the results. For example, in Chapter 5 the interaction between class and income

247

revealed that professional and manager families in the lowest income bracket spend more on culture and luxury than other professional or managerial income brackets. On the one hand, this finding suggests that there is a class effect that cannot be reduced to the economic resources of the household; on the other hand, income also interacted with many other independent variables. These findings make our theoretical interpretations much more complex than interpretations based on main-effect models. For this reason, the observed interactions of background variables will not be elaborated here in detail. 2

It has been assumed that strategies of class distinction are practically non-existent in Finland

(Mäkelä 1994: 261-262; see also Alasuutari 1997: 3-4). However, our analysis enables an opposing view, stressing the role of certain visible consumption patterns. Strategies of social distinction based on the selected types of consumer practices may be found among Finns as well. This opposing view can of course only be offered as an interpretation; it cannot be verified on the basis of the present empirical data. Thus there is no need to contradict the conclusions based on a single restricted interview study conducted in the 1980s (see Roos & Rahkonen 1985). 3

For example, the activities measured in this study might possibly be less open to the

individualistic tendencies of postmodern consumerism than certain others. Particularly the absence of measures related to food is important in this respect. In addition, the measures used in this study consist of only two types of quantitative measures (household expenditure and individual consumption desires). It may also be that the use of other types of data, such as lengthy qualitative interviews, would have indicated somewhat weaker connections between socio-demographic factors and individual consumption patterns than is suggested by the analyses presented here. On the other hand, the structural connections could also have appeared to be even stronger. 4

It is of course not surprising that numerous variables result in better explanation proportions in

analyses based on generalised linear models (GLM). Nevertheless, stronger arguments can be put forward on the basis of correlations between many factors than would have been possible if the analyses had focused on single variables alone. In this sense, my intention is to seek ways in which structural-based explanatory factors can be integrated with current sociological approaches to consumption.

248

REFERENCES

Abbott, A. (1988). The System of Professions: An Essay on the Division of Expert Labor. Chicago: University of Chicago Press. Abell, P. (1996). Sociological Theory and Rational Choice Theory. In Turner, B. S. (Ed.): Blackwell Companion to Social Theory. Oxford: Blackwell, 252-273. Achen, C. H. (1986). The Statistical Analysis of Quasi-Experiments. Berkeley & Los Angeles: University of California Press. Ahponen, P-L. & Järvelä, M. (1983). Maalta kaupunkiin, pientilalta tehtaaseen. Helsinki: WSOY. Alasuutari, P. (1997). Kulttuuripääoma summamuuttujan valossa. Sosiologia 34:1, 314. Alkula, T. & Pöntinen, S. & Ylöstalo, P. (1994). Sosiaalitutkimuksen kvantitatiiviset menetelmät. Helsinki: WSOY, 3-32. Allardt, E. (1986). Elämäntapa, harkinta ja muoti ihmisten valintojen perustana. In Heikkinen, K. (ed.): Kymmenen esseetä elämäntavasta. Lahti: Yleisradio, 1-33. Anthias, F. (2001). The Material and the Symbolic in Theorising Social Stratification: Issues of Gender, Ethnicity and Class. British Journal of Sociology 52:3, 367-390. Antonides, G. & van Raaij, W. F. (1998). Consumer Behaviour: A European Perspective. Chichester: John Wiley & Sons. Atkinson, A. & Rainwater, L. & Smeeding, T. (1995). Income Distribution in OECD Countries: Evidence from the Luxembourg Income Study. Paris: OECD. Babbie, E. (1989). The Practice of Social Research. Fifth Edition. Belemont: Wadsworth. Baudrillard, J. (1983). Simulations. New York: Semiotext. Baudrillard, J. (1998). The Consumer Society. Myths and Structures (translated by Chris Turner). London: Sage. Bauman, Z. (1982). Memories of Class: The Pre-History and After-Life of Class. London: Routledge & Kegan Paul. Bauman, Z. (1997). Postmodernity and Its Discontents. Cambridge: Polity Press. Bauman, Z. (1998). Work, Consumerism and the New Poor. Milton Keynes: Open University Press.

251

Beck, U. (1992). Risk Society: Towards a New Modernity (transl. Ritter, M.). London: Sage. Beck, U. & Bonss, W. & Lau, C. (2003). The Theory of Reflexive Modernization: Problematic, Hypotheses and Research Programme. Theory, Culture & Society 20:2, 133. Beck-Gernsheim, E. (1996). Life as a Planning Project (transl. Chalmers, M.). In Lash, S. & Szerszynski, B. & Wynne, B. (Eds): Risk, Environment and Modernity: Towards a New Ecology. London: Sage, 139-153. Becker, G. & Murphy, K. M. (2000). Social Economics. Market Behaviour in a Social Environment. Cambridge: Harvard University Press. Bell, D. (1977). The Coming of Post-Industrial Society. New York: Basic Books. Bellah, R. N. & Madsen, R. & Sullivan, W. M. & Swidler, A. & Tipton, S. M. (1985). Habits of the Heart. Individualism and Commitment in American Life. Berkeley: Haper & Row. Best, S. & Kellner, D. (1997). The Postmodern Turn. New York: Guilford Press. Biljeveld, C. C. J. H. & van deer Kamp, L. J. T., with Moijaart, A. & van der Kloot, W. A. & van der Leeden, G. & van der Burg, E. (1998). Longitudinal Data Analysis: Designs, Models and Methods. London: Sage. Blau, J. (1993). Social Contracts and Economic Markets. New York: Plenum. Blom, R. & Melin, H. & Nikula, J. (1998). Changes in the Work Situation: Proliferation of Good Working Life or Just the Same Old Pall? In Blom, R. & Melin, H. (Eds): Economic Crisis, Social Change and New Divisions in Finland. Publications of Department of Sociology and Social Psychology, Series A/29. Tampere: University of Tampere, 25-52. Blom, R. & Melin, H. & Pyöriä, P. (2001). Tietotyö ja työelämän muutos: Palkkatyön arki tietoyhteiskunnassa. Helsinki: Gaudeamus. Blossfeld, H-P. & Drobnič, S. (Eds) (2001). Careers and Couples in Contemporary Societies: From Male Breadwinner to Dual Earner Families. Oxford: Oxford University Press. Blossfeld, H-P. & Rohwer, G. (2002). Techniques of Event History Modeling: New Approaches to Causal Analysis. Second Edition. London: Lawrence Erlbaum. Blundell, R. (1988). Consumer Behaviour: Theory and Empirical Evidence – A Survey. Economic Journal 98:3, 16-65. Bocock, R. (1993). Consumption. London: Routledge.

252

Boden, S. & Williams, S. J. (2002). Consumption and Emotion: The Romantic Ethic Revisited. Sociology 36:3, 493-512. Bourdieu, P. (1984). Distinction. A Social Critique of the Judgement of Taste (translated by Nice, R.). London: Routledge & Kegan Paul. Bradley, H. (1996). Fractured Identities: Changing Patterns of Inequality. Cambridge: Polity Press. Breen, R. & Rottman, D. (1995). Class Analysis and Class Theory. Sociology 29:3, 453-473. Campbell, C. (1987). The Romantic Ethic and the Spirit of Modern Consumerism. Oxford: Blackwell. Campbell, C. (1995). The Sociology of Consumption. In Miller, Daniel (Ed.): Acknowledging Consumption: A Review of New Studies. London: Routledge, 96-126. Castells, M. (1998). The Information Age: Economy, Society and Culture. Volume III. End of Millennium. Oxford: Blackwell. Centers, R. (1949). The Psychology of Social Class. Princeton NJ: Princeton University Press. Chaney, D. (1990). Subtopia in Gateshead: The Metro Centre as Cultural Form. Theory, Culture, Society 7.4, 49-86. Chattoe, E. & Gilbert, N. (1999). Talking About Budgets: Time and Uncertainty in Households’ Decision Making. Sociology 33:1, 85-103. Clark, T. & Lipset, S. (1991). Are Classes Dying? International Sociology 6:4, 397-410. Collins Cobuild (1990). Collins Cobuild English Language Dictionary. London & Glasgow: Collins. Craib, I. (1992). Modern Social Theory: From Parsons to Habermas. Second Edition. New York: Harvester Wheatsheaf. Crompton, R. (1993). Class and Stratification. London: Blackwell. Cronin, A. M. (2000). Advertising and Consumer Citizenship: Gender, Images and Rights. London: Routledge. Csikszentmihalyi, M. (2000). The Costs and Benefits of Consuming. Journal of Consumer Research 27:3, 267-272. Davis, F. (1992). Fashion, Culture, and Identity. Chicago: Chicago University Press. Dickens, P. (1990). Urban Sociology: Society, Locality and Human Nature. London: Harvester Wheatsheaf.

253

Djurfeldt, G. & Gooch, P. (2002). Farm Crisis, Mobility and Structural Change in Swedish Agriculture, 1992-2000. Acta Sociologica 45:2, 75-88. Douglas, M. (1996). Thought Styles: Critical Essays on Good Taste. London: Sage. Douglas, M. & Isherwood, B. (1978). The World of Goods: Towards an Anthropology of Consumption. New York: Penguin Books. Elster, J. (1989). Nuts and Bolts for the Social Sciences. Cambridge: Cambridge University Press. Erikson, E. H. (1980). Identity and the Life Cycle. New York: Norton. Erikson, R. (1984). Social Class of Men, Women and Families. Sociology 18:4, 500514. Erikson, R. & Goldthorpe, J. H. (1992). The Constant Flux. Oxford: Clarendon. Erola, J. (2000). Häviävätkö luokkaerot? Kulutuspohjaisten luokkaerojen merkitys riskiyhteiskunnassa. Futura 19:3, 26-40. Erola, J. (2001). Maksuhäiriöisyys ja luottamusrakenteiden kriisi sosiaalisessa toiminnassa. Sosiologia 38:3, 174-191. Erola, J. & Moisio, P. (2002). Jähmettyikö Suomi? Sosiaalinen liikkuvuus ja pitkäaikaistyöttömyys Suomessa 1970-1995. Sosiologia 39:3, 185-199. Erola, J. & Räsänen, P. (1998). Riskirationaalisuus ja epävarmuuden elämänpolitiikka. Janus 6:2, 153-166. Erola, J. & Räsänen, P. (2000). Suomi 1999: Aineiston keruu ja tutkimusseloste. Teoksessa Erola, J.: Maksuhäiriöisyys laman jälkeisessä Suomessa. Sosiologista keskustelua B 36 Sociological discussions. Turku: Turun yliopisto, 75-88. Eskola, A. (1985). Persoonallisuustyypeistä elämäntapaan. tutkimuksen metodologisia opetuksia. Helsinki: WSOY.

Persoonallisuuden

Esping-Andersen, G. (1993). Post-Industrial Class Structures: An Analytical Framework. In Esping-Andersen, G. (Ed.): Changing Classes. Stratification and Mobility in Post-Industrial Societies. London: Sage, 7-31. Esping-Andersen, G. (2000). Two Societies, One Sociology, and No Theory. British Journal of Sociology 51:1, 59-77. Esser, H. (1996). What is Wrong with “Variable Sociology”? European Sociological Review 12:2, 159-166. Etürk, K. A. (1999). Marx, Postmodernity and the Transition of the Individual. Review of Radical Political Economics 31:2, 27-45.

254

Fainstein, S. & Cambell, S. (1996). Introduction: Theories of Urban Development and their Implications for Policy and Planning. In Fainstein, S. & Cambell, S. (Eds): Readings in Urban Theory. Cambridge: Blackwell, 1-17. Falk, P. (1995). The Consuming Body. London: Sage. Falk, P. (1997). The Scopic Regimes of Shopping. In Falk, P. & Campbell, C. (Eds): The Shopping Experience. London: Sage, 177-185. Falk, P. & Campbell, C. (1997). Introduction. In Falk, P. & Campbell, C. (Eds): The Shopping Experience. London: Sage, 1-14. Featherstone, M. (1991). Consumer Culture and Postmodernism. London: Sage. Ferrera, M. (1996). The “Southern Model” of Social Welfare in Europe. Journal of European Social Policy 6:1, 17-37. Fidler, R. (1997). Mediamorphosis. Understanding New Media. Thousand Oaks: Pine Forge. Figueora-Sarriera, H. J. (1999). In and Out of the Digital Closet: The Self as Communication Network. In Gordo-Lopez, A. & Parker, I. (Eds): Cyberpsychology. London: Macmillan, 130-145. Fine, B. & Leopold, E. (1993). The World of Consumption. London: Routledge. Firebaugh, G. (1999). Empirics of the World Income Inequality. American Journal of Sociology 104:6, 1597-1630. Fischer, C. S. (1995). The Subcultural Theory of Urbanism: A Twentieth-Year Assessment. American Journal of Sociology 101:3, 543-577. Fisher, J. E. (1987). Social Class and Consumer Behaviour. The Relevance of Class and Status. Advances in Consumer Research 14, 492-496. Florida, R. (2002). The Rise of the Creative Class. New York: Basic Books. Fulton, O. (1994). Consuming Education. In Keat, R. & Whiteley, N. & Abercrombie, N. (Eds): The Authority of the Consumer. London: Routledge, 223-239. Gabriel, Y. & Lang, T. (1995). Unmanageable Consumer. Contemporary Consumpiton and its Fragmentations. London: Sage. Gershuny, J. (2000). Changing Times. Work and Leisure in Postindustrial Societies. Oxford: Oxford University Press. Giddens, A. (1973). The Class Structure of the Advanced Societies. London: Hutchinson & Co.

255

Giddens, A. (1981). A Contemporary Critique of Historical Materialism. Vol. 1 Power, Property and the State. Cambridge: Macmillan. Giddens, A. (1984). The Constitution of Society. Oxford: Polity Press. Giddens, A. (1989). Sociology. Cambridge: Polity Press. Giddens, A. (1991). Modernity and Self-Identity. Self and Society in the Late Modern Age. Cambridge: Polity Press. Glick, P. C. (1947). The Family Life Cycle. American Journal of Sociology 12 (April), 164-174. Goesling, B. (2001). Changing Income Inequalities Within and Between Nations: New Evidence. American Sociological Review 66:4, 745-761. Goldthorpe, J. (1982). On the Service Class, Its Formation and Future. In Giddens, A. & Mackenzie, G. (Eds): Social Class and the Division of Labour. Cambridge: Cambridge University Press, 161-185. Goldthorpe, J. (2000). On Sociology  Numbers, Narratives, and the Integration of Research and Theory. Oxford: Oxford University Press. Gouldner, A. (1979). The Future of Intellectuals and the Rise of the New Class. London: Macmillan. Goody, J. (1984). Cooking, Cuisine and Class. A Study of Comparative Sociology. Cambridge: Cambridge University Press. Graham, J. W., & Collins, N. L. (1991). Controlling Correlational Bias via Confirmatory Factor Analysis of MTMM Data. Multivariate Behavioral Research 26:4, 607-629. Gregson, N. (1991). On the (Ir)relevance of Structuration Theory. In Held, D. Thompson, C. (Eds): Structuration Theory. Cambridge: Polity Press. Grey, A. & McGuigan, J. (1993). Introduction. In Grey, A. & McGuigan, J. (Eds): Studying Culture. London: Edward Arnold. Gronow, J. (1997). The Sociology of Taste. London: Routledge. Haller, M. (2002). Theory and Method in the Comparative Study of Values. Critique and Alternative to Inglehart. European Sociological Review 18:2, 139-158. Hannigan, J. (1998). Fantasy City. Pleasure and Profit in the Postmodern Metropolis. London: Routledge. Harman, H. H. (1976). Modern Factor Analysis. Chicago: University of Chicago Press.

256

Harrinvirta, M. O. (2000). Strategies of Public Sector Reforms in the OECD countries. Helsinki: Finnish Society of Sciences and Letters. Harvey, D. (1990). The Condition of Postmodernity. Cambridge: Polity Press. Havinghurst, R. & Feigenbaum, K. (1959). Leisure and Lifestyle. American Journal of Sociology 5:4, 396-404. Hautamäki, A. (1996). Individualismi on humanismia. Teoksessa Hautamäki, A. & Lagerpets, E. & Sihvola, J. & Siltala, J. & Tarkki, J.: Yksilö modernin murroksessa. Helsinki: Gaudeamus, 13-44. Hebdige, D. (1979). Subculture: the Meaning of Style. London: Routledge. Heelas, P. (1994). The Limits of Consumption and the Postmodern Religion of New Age. In Keat, R. & Whiteley, N. & Abercrombie, N. (Eds): The Authority of the Consumer. London: Routledge, 102-115. Hedström, P. & Swedberg, R. (1996). Social Mechanisms. Acta Sociologica 39:3, 281308. Hedström, P. & Swedberg, R. (1998). Social Mechanisms. An Introductory Essay. In Hedström, P. & Swedberg, R. (Eds): Social Mechanisms. An Analytic Approach to Social Theory. Cambridge: Cambridge University Press, 1-31. Heikkilä, M. & Sihvo, T. (1997). Concepts of Poverty and Exclusion in Europe. Scandinavian Journal of Social Welfare 6:2, 119-126. Heikkinen, S. & Kuusterä, A. (2001). Finnish Economic Crises in the 20th Century. In Kalela, J. & Kiander, J. & Kivikuru, U. & Loikkanen, H. A. & Simpura, J. (Eds): Down From the Heavens, Up From the Ashes. The Finnish Economic Crisis of the 1990s in the Light of Economic and Social Research. VATT Publications 27:6. Helsinki: Valtion Taloudellinen tutkimuskeskus, 25-51. Heinonen, V. (1998). Talonpoikainen etiikka ja kulutuksen henki. Kotitalousneuvonnasta kuluttajapolitiikkaan 1900-luvun Suomessa. Bibliotheca Historica 33. Helsinki: Suomen Historiallinen Seura. Heinonen, V. (2000). Näin alkoi ”kulutusjuhla”. Suomalaisen kulutusyhteiskunnan rakentuminen. Teoksessa Hyvönen, K. & Juntto, A. & Laaksonen, P. & Timonen, P. (toim.): Hyvää elämää — 90-vuotta suomalaista kuluttajatutkimusta. Helsinki: Tilastokeskus & Kuluttajatutkimuskeskus, 8-22. Heinonen, V. (2001). Niukkuuden ja kulutuksen mentaliteeetit modernisoituvassa Suomessa. Teoksessa Näre, S. (toim.): Tunteiden sosiologiaa II. Historiaa ja säätelyä. Helsinki: Suomalaisen kirjallisuuden seura, 75-98. Hendry, L. & Schucksmith, J. & Lowe, J. (1993). Young Person’s Leisure and Lifestyle. London: Routledge.

257

Hetherington, K. (1998). Expressions of Identity. Space, Performance, Politics. London: Sage. Härkönen, J. & Kosonen, P. (2003). Kotitaloudet ja työmarkkinat Euroopan Unionissa – Eroavatko hyvinvointiregiimit toisistaan? Sosiologia, 40:1, 23-39. Ilmonen, K. (1993). Tavaroiden taikamaailma. Sosiologinen avaus kulutukseen. Tampere: Vastapaino. Ilmonen, K. (1994). Anthony Giddensin rakenteistumisteoria ja sen kritiikki. Teoksessa Heiskala, R. (toim.): Sosiologisen teorian nykysuuntauksia. Helsinki: Gaudeamus, 316347. Inglehart, R. (1997). Modernization and Postmodernization. Cultural, Economic, and Political Change in 43 Societies. Princeton NJ: Princeton University Press. Inglehart, R. & Baker, W. E. (2000). Modernization, Cultural Change, and the Persistence of Traditional Values. American Sociological Review 65:1, 19-51. Jackman, M. & Jackman, R. (1983). Class Awareness in the United States. Berkeley: University of California Press. Jaeger, M. (1986). Class Definition and the Esthetics of Gentrification: Victoriana in Melbourne. In Smith, N. & Williams, P. (Eds): Gentrification of the City. London: Allen & Unwin, 82-104. Jameson, F. (1991). Postmodernism, or the Cultural Logic of Late Capitalism. London: Verso. Jenkins, R. (1996). Social Identity. London: Routledge. Kauhanen, M. (2000). Määräaikaiset työsuhteet ja sosiaaliturvajärjestelmän kestävyys. Sosiaali- ja terveysministeriön selvityksiä 2000:9. Helsinki: Sosiaali- ja terveysministeriö. Kangas, O. & Ritakallio, V-M. (1996). Eri menetelmät – eri tulokset? Köyhyyden monimuotoisuus. Teoksessa Kangas, O. & Ritakallio, V-M. (toim.): Kuka on köyhä: Köyhyys 1990-luvun puolivälin Suomessa. Sosiaali- ja terveysalan tutkimus- ja kehittämiskeskuksen tutkimuksia 65. Helsinki: Stakes, 11-67. Kautto, M. & Moisio, P. (2002). Suomalainen köyhyys vertailevasta näkökulmasta. Teoksessa Heikkilä, M. & Kautto, M. (toim.) Suomalaisten hyvinvointi 2002. Helsinki: Stakes, 312-335. Katz-Gerro, T. & Shavit, Y. (1998). The Stratification of Leisure and Taste: Classes and Lifestyles in Israel. European Sociological Review 14:4, 369-388. Klein, N. (2001). No logo. Tähtäimessä brändivaltiaat (translated by Laaksonen, L. & Tillman, M.). Helsinki: WSOY.

258

Kortteinen, M. (1982). Lähiö: tutkimus elämäntapojen muutoksesta. Helsinki: Otava. Korzeniewicz, R. P. & Moran, T. P. (2000). Measuring World Income Inequalities. American Journal of Sociology 106:1, 209-214. Kotzinets, R. V. (2001). Utopian Enterprise: Articulating the Meanings of Star Trek’s Culture of Consumption. Journal of Consumer Research 28:2, 67-88. Kouvo, A. (2000). Kestävä kehitys ja käyttäytyminen Varsinais-Suomessa: Tutkimusraportti. Varsinais-Suomen Agendan julkaisuja. Turku: Varsinais-Suomen Agendatoimisto. Kraaykamp, G. (2002). Cumulative Advantages and Inequality in Lifestyle. A Dutch Description of Distinction in Taste. The Netherlands Journal of Social Sciences 38:2, 121-143. Kraaykamp, G. & Nieuwbeerta, P. (2000). Parental Background and Lifestyle Differentiation in Eastern Europe: Social, Political, and Cultural Intergenerational Transmission in Five Former Socialist Societies. Social Science Research 29:1, 92-122. Kultalahti, O. (2001). Internal Migration and Specialising Labour Markets in Finland. In Yearbook of Population Research in Finland 37, 103-125. Lachmann, R. (1988). Graffiti as Career and Ideology. American Journal of Sociology 94:2, 229-250. Laurila, E. H. (1987). Yksityinen kulutus Suomessa ajanjaksona 1880-1980. ETLAn tutkimuksia, Sarja B 52 Series. Helsinki: ETLA. Lash, S. (1995). Refleksiivisyys ja sen vastinparit: rakenne, estetiikka, yhteisö. Teoksessa Beck, U. & Giddens, A. & Scott, L. (1995): Nykyajan jäljillä. Refleksiivinen modernisaatio (translated by Lehto, L.). Tampere: Vastapaino, 153-253. Lazarfeld, P. & Rosenberg, M. (eds.) (1955). The Language of Social Research. New York: The Free Press. Lee, D. J. & Turner, B. S. (1996). Conflicts About Class. Debating Inequality in Late Industrialism. London: Longman. Lee, M. J. (1993). Consumer Culture Reborn. London: Routledge. Lefebvre, H. (1971). Everyday Life in the Modern World. New York: Harper & Row. Lehtonen, T-K. (1994). Shoppailu sosiaalisena muotona. Sosiologia 31:3, 192-203. Lehtonen, T-K. & Mäenpää, P. (1997). Shopping in the East Centre Mall. In Falk, P. & Cambell, C. (Eds): The Shopping Experience. London: Sage, 136-165. Leiulfsrud, H. (1991). Det familjärä klasssamhället: en teorisk och empirisk studie av blandklassfamiljer. Lund: Arkiv.

259

Lemert, C. (1997). Postmodernism Is Not What You Think. London: Blackwell. Lewin, C. & Orleans, M. (2000). The Class Situation of Information Specialists: A Case Analysis. Sociological Research Online 5:3. . (May 2002). Lievrouw, L. A. (2001). New Media and the ‘Pluralization of Life-Worlds’. A Role for Information in Social Differentiation. New Media & Society 3:1, 7-28. Lunt, P. & Livingstone, S. (1992). Mass Consumption and Personal Identity. Everyday Economic Perspective. Milton Keynes: Open University Press. Luomanen, J. & Räsänen, P. (2000). Systemaattisuudesta laadullisessa tutkimuksessa. Tieteessä tapahtuu 18:1, 55-58. Lyotard, J-F. (1985). The Condition of Postmodernity. London: Sage. Maffesoli, M. (1996). The Time of the Tribes. The Decline of Individualism in Mass Society (transl. Nice, R.). London: Sage. Marshall, G. (1997). Repositioning Class. Social Inequality in Industrial Societies. London: Sage. Martin, J. & Barton, J. (1996). The Effect of Changes in the Definition of the Household Reference Person. Survey Methodology Bulletin 38, 1-8 Maslow, A. (1968). Toward a Psychology of Being. New York: Van Nostrand. McDowell, L. (1999). Gender, Identity and Place: Understanding Feminist Geographies. Cambridge: Polity Press. McQuail, D. (1994). Mass Communication Theory. Third Edition. London: Sage. McNeil, K. & Newman, I. & Kelly, F. J. (1996). Testing Research Hypothesis with the General Linear Model. Carbondale: Southern Illinois University Press. Menard, S. (1991). Longitudinal Research. Sage University Paper Series on Quantitative Applications in the Social Sciences, 07-076. London: Sage. Merton, R. K. (1963). Social Theory and Social Structure. Revised and enlarged Edition. Glencoe: The Free Press. Miles, S. (1998). Consumerism  as a Way of Life. London: Sage. Miller, D. (1995). Consumption as the Vanguard History. In Miller, D. (Ed.): Acknowledging Consumption: A Review of New Studies. London & New York: Routledge, 1-57.

260

Miller, D. & Jackson, P. & Thrift, N. & Holbrook, B. & Rowlands, M. (1998). Shopping, Place and Identity. London: Routledge. Mitchell, A. (1983). The Nine American Life Styles. Who We Are and Where Are We Going? New York: Warner Books. Muttilainen, V. & Tala, J. (1998). Kuka vapautuu veloistaan. Velkajärjestelyn maksuohjelmat ja vuoden 1997 lakimuutos. Oikeuspoliittisen tutkimuslaitoksen julkaisuja 155. Helsinki: Oikeusministeriö. Mäkelä, J. (1994). Pierre Bourdieu − erottautumisen teoreetikko. Teoksessa Heiskala, R. (ed.): Sosiologisen teorian nykysuuntauksia. Helsinki: Gaudeamus, 243-269. Newman, I. & Benz, C. R. (1998). Qualitative-Quantitative Research Methodology. Exploring the Interactive Continuum. Carbondale: Southern Illinois University Press. Niemi, I. (1995). A General View of Time Use by Gender. In Niemi, I. (ed.): Time Use of Women in Europe and North America. New York: United Nations Economic Commission for Europe, 1-22. Nunnaly, J. (1978). Psychometric Theory. New York: McGraw-Hill. Noro, A. (1995). Uudemman kulutussosiologian mallit ja figuurit. Sosiologia 32:1, 111. Noro, A. (2000). Aikalaisanalyysi sosiologisen teorian kolmantena lajityyppinä. Sosiologia 37:4, 321-329. Nätti, J. (1989). Työmarkkinoiden lohkoutuminen ja segregaatioteoriat. Suomen työmarkkinat ja yritysten työvoimamarkkinat. Jyväskylä studies in education, psychology and social research 68. Jyväskylä: Jyväskylän yliopisto. O´Meara, M. (2001). Exploring a New Vision for Cities. In Freire, M. & Stren, R. (Eds): The Challenge of Urban Gowernment. Policies and Practices. Washington & Toronto: World Bank Institute & University of Toronto, Centre for Urban and Community Studies, 337-355. OECD (1982). The OECD List of Social Indicators. Paris: OECD. OECD (2001). Historical Statistics 1970-2000. Paris: OECD. OED Online (2002). Oxford English Dictionary. Second Edition. . (November 2002). Owen, D. (1997). The Postmodern Challenge to Sociology. In Owen, David (Ed.): Sociology After Postmodernism. London: Sage, 1-22. Pantzar, M. (1996). Rational Choice of Food: On the Domain of the Premises of the Consumer Choice Theory. Journal of Consumer Studies and Home Economics 20:1, 120.

261

Parkin, F. (1979). Marxism and Class Theory. London: Tavistock. Parsons, T. (1951). The Social System. Illinois: The Free Press Perkin, H. (1996). The Third Revolution. Professional Elites in the Modern World. London: Routledge. Peltonen, M. (toim.) (1996). Rillumarei ja valistus: Kulttuurikahakoita 1950-luvun Suomessa. Helsinki: Suomen historiallinen seura. Pescasolido, B. A. & Rubin, B. A. (2000). The Web of Group Affiliations Revisited: Social Life, Postmodernism, and Sociology. American Sociological Review 65:1, 52-76. Puohiniemi, M. (1991). Value-based Segmentation, Social Change and Consuming Orientations. In ESOMAR: Seminar on the Growing Individualisation of Consumer Lifestyles and Demand. How is Marketing Coping With It? Helsinki (Finland) 12th-14th June 1991. Amsterdam: The European Organisation for Opinion and Marketing Research, 49-59. Pöntinen, S. (1983). Social Mobility and Social Structure. A Comparison of Scandinavian Countries. Commentationes Scientiarum Socialium 20/1983. Helsinki: Societas Scientiarum Fennica. Raijas, A. (2000). Erilaisten kotitalouksien kulutusmuutokset 1990-luvulla kulutustukimuksen valossa. Teoksessa Hyvönen, K. & Juntto, A. & Laaksonen, P. & Timonen, P. (toim.): Hyvää elämää — 90-vuotta suomalaista kuluttajatutkimusta. Helsinki: Tilastokeskus & Kuluttajatutkimuskeskus, 48-56. Raijas, A. (2001). Kokemattomien käyttäjien näkemyksiä Interneteistä ja sen tarjoamista palveluista. LTT tutkimus Oy:n Elektronisen Kaupan Instituutin julkaisuja. Helsinki: Helsingin kauppakorkeakoulu. Reekie, G. (1993). Sex, Selling and the Department Store. London: Allen & Unwin. Riley, M. W. (1973). Ageing and Cohort Succession: Interpretations and Misinterpretations. The Public Opinion Quarterly 1973:1, 35-49. Rinne, R. & Kivinen, O. & Ahola, S. (1992). Aikuisten kouluttautuminen Suomessa. Osallistuminen, kasaantuminen ja preferenssit. Koulutussosiologian tutkimuskeskuksen raportteja 10. Turku: Turun yliopisto. Robinson, J. (1975). Time as an Indicator of Social Change and the Quality of Life. Advances in Consumer Research 2:1, 847-850. Robinson, J. & Landry, B. & Rooks, R. (1998). Time and the Melting Pot. American Demographics 20:6, 18-24. Rogers, E. M. (1983). Diffusion of Innnovations. Third Edition. New York: Free Press.

262

Roos, J. P. (1986). Elämäntapateoriat ja suomalainen elämäntapa. Teoksessa Heikkinen, K. (toim.): Kymmenen esseetä elämäntavasta. Lahti: Yleisradio, 34-76. Roos, J. P. (1987). Suomalainen elämä. Tutkimus tavallisten suomalaisten elämänkerroista. Helsinki: SKS. Roos, J. P. (1990). Suomalaisen elämäntavan muutokset. Teoksessa Riihinen, O. (toim.): Suomi 2017. Helsinki: Gummerus, 477-494. Roos, J. P. & Rahkonen, K. (1985). [Will to a Distinctive Life Style:] In Search of the New Finnish Middle Class. Acta Sociologica 28:3, 257-274. Rose, D. & Pevalin, D. with Elias, P. & Martin, J. (2001). Towards a European Socioeconomic Classification. Final Report to Eurostat of the Expert Group. Great Britain: Eurostat. Rowe, D. (1995). Popular Cultures: Rock Music, Sport and the Politics of Pleasure. London: Sage. Ruggles, P. (1990). Drawing the Line: Alternative Poverty Measures and Their Implications for Public Policy. Washington: Urban Institute Press. Ruonavaara, H. (1990). Four Models of Explaining the Growth of Home-Ownership. Scandinavian Housing & Planning Research 7, 129-142. Ruonavaara, H. (1993). Omat kodit ja vuokrahuoneet. Sosiologinen tutkimus asunnonhallinnan muodoista Suomen asutuskeskuksissa 1920-1950. Turun yliopiston julkaisuja, sarja C, osa 97. Turku: Turun yliopisto. Räsänen, P. (2000). Kulutusvalintojen postmodernit piirteet ja rakenteelliset reunaehdot. Sosiologia 37:3, 228-242. Räsänen, P. (2001). Computer Assistance and the Evolution of Qualitative Data Analysis. The Yearbook of Population Research in Finland 37, 63-82. Räsänen, P. (2002). Vaikuttaako kuluttajan tausta uuden teknologian kulutukseen? Teoksessa Uusitalo, L. (toim.): Kuluttaja virtuaalimarkkinoilla. Helsinki: Edita, 160176. Räsänen, P. (2003). The (In)adequacy of Explanatory Mechanisms in Empirical Research. Teoksessa Toivonen, T. (toim.): Individuaalisuus, rationaalisuus ja sosiaalinen rakenne – Individuality, Rationality, and Social Structure. Turun kauppakorkeakoulun julkaisuja, Sarja Keskustelua ja Raportteja 3:2003. Turku: Turun kauppakorkeakoulu, 91-115. Saunders, P. (1984). Beyond Housing Classes: The Sociological Significance of Private Property Rights in Means of Consumption. International Journal of Urban and Regional Research 8:2, 202-225.

263

Savage, M. (2000). Class Analysis and Social Transformation. Buckingham: Open University Press. Savage, M. & Warde, A. (1993). Urban Sociology, Capitalism and Modernity. New York: Continuum. Sennett, R. (1999). The Corrosion of Character: The Personal Consequences of Work in the New Capitalism. New York: Norton. Sennett, R. & Cobb, J. (1972). The Hidden Injuries of Class. Cambridge: Cambridge University Press. Shields, R. (1996). Foreword: Masses or Tribes. In Maffesoli, M.: The Time of the Tribes: The Decline of Individualism in Mass Society. London: Sage, ix-xii. Simmel, G. (1957). “Fashion”. American Journal of Sociology 62:4, 541-548. Siurala, L. (1987). Nuorten työasenteet: Muuttuvatko iän mukana vai jäävätkö sukupolvea leimaamaan. Helsingin kauppakorkeakoulun julkaisuja D-98. Helsinki: Helsingin kauppakorkeakoulu. Smart, B. (1997). Postmodern Social Theory. In Turner, B. S. (Ed.): Blackwell Companion to Social Theory. Oxford: Blackwell, 396-428. Smelser, N. J. (1997). The Rational and the Ambivalent in the Social Sciences. American Sociological Review 63:1, 1-16. Soja, E. W. (2000). Postmetropolis: Critical Studies of Cities and Regions. Oxford: Blackwell. Spencer, D. & Wells, S. (2000). Qualitative Research and Innovation. In Marks, L. (Ed.): Qualitative Research in Context. Admap: Henley-On-Thames, 233-252. Stevens, J. (1992). Applied Multivariate Statistics for the Social Sciences. Second Edition. London: Lawrence Erlbaum. Stewart, A. & Prandy, K. & Blackburn, M. (1980). Social Stratification and Occupations. New York: Holmes & Meier. Stone, G. P. (1954). City Shoppers and Urban Identification: Observations on the Social Psychology of City Life. American Journal of Sociology 60:1, 36-45. Strasser, S. & McGovern, C. & Judt, M. (1998). Getting and Spending: European and American Consumer Societies in the Twentieth Century. Cambridge: Cambridge University Press. Sulkunen, P. (1992). The New European Middle Class: Individuality and Tribalism in Mass Society. Aderhot: Avebury.

264

Sullivan, O. & Gershuny, J. (2001). Cross-National Changes in Time-Use: Some Sociological (Hi)Stories Re-Examined. British Journal of Sociology 52:2, 331-347. Suomen Pankki (2003). Suomen taloudelliset näkymät 2003-2005. Ennusteen taulukkoja kuvioliite. . (June 2003) Sørensen, A. (1994). Women, Family and Class. Annual Review of Sociology 20, 27-47. Tabachnick, B. G. & Fidell, L. S. (2001). Using Multivariate Statistics. Fourth Edition. London: Allyn & Bacon. Therborn, G. (1991). Cultural Belonging, Structural Location and Human Action. Acta Sociologica 34:3, 177-191. Thrift, N. (1993). An Urban Impasse. Theory, Culture & Society 10:2, 230-231 Tilastokeskus (1989). Sosioekonomisen aseman luokitus 1989. Käsikirjoja 17. Helsinki: Tilastokeskus. Tilastokeskus (2001). Kulutustutkimus 1998. Laatuselvitys. Tulot ja kulutus 2001:4. Helsinki: Tilastokeskus. Tilastokeskus (2002). Tilastotietoa aiheluokituksen perusteella. . (May 2002) Tilastokeskus (2003). Tilastotietoa aiheluokituksen perusteella. . (April 2003) Toivonen, T. (1986). Ainainen puute ja kurjuus? Kulutus ja yhteiskuntakerrostumat Suomessa 1928-1950. Turun kauppakorkeakoulun julkaisuja A 1:1986. Turku: Turun kauppakorkakoulu. Toivonen, T. (1991). Melting Away Class Differences: Consumption Differences Between Employee Groups in Finland 1995-1985. Social Indicators Research 26:3, 277-302. Toivonen, T. (1994). Does Consumption Determine Social Class? On the Changing Pattern of Consumption Determination. Journal of Consumer Studies and Home Economics 18:1, 45-63. Toivonen, T. (1996). Luokat ja kulutus: Episodi massakulutuksen läpimurrosta 19551971. Sosiologia 34:1, 35-48. Toivonen, T. (1997). Food and Social Class. Journal of Consumer Studies and Home Economics 21:4, 329-347. Toivonen, T. (1999). Empiirinen sosiaalitutkimus: Filosofia ja metodologia. Helsinki: WSOY.

265

Tomlinson, A. (1990). Introduction. In Tomlinson, A. (Ed.): Consumption, Identity and Style. London: Routledge & Kegan Paul, 1-21. Tuorila, H. (1986). Hedonic Responses to Colour, Sweetness, Saltiness and Fattiness in Selected Food as Related to Corresponding Attitudes and Other Behavioral Measures. Elintarvikekemian ja -teknologian laitoksen EKT-serie 742. Helsinki: University of Helsinki. Uusitalo, H. (1993). Ekvivalenssiskaalat: Tarpeellisia vai tarpeettomia tilastotyökaluja? Hyvinvointikatsaus 2/1993, 37-40. Uusitalo, L. (1979). Consumption and Way of Life: An Empirical Identification and Explanation of Consumption Style Dimensions. Acta Academie Oecomomicae Helsingiensis, Series A:27. Helsinki: Helsinki School of Economics. Uusitalo, L. (1998). Consumption and Postmodernism: Social Structuration and the Construction of the Self. In Bianchi, M. (Ed.): The Active Consumer. Novelty and Surprise in Consumer Choice. London: Routledge, 215-235. Veblen, T. (1994). The Theory of the Leisure Class. New York: Dover Publications. Waite, L. J. (1980). Working Wives and the Family Life Cycle. American Journal of Sociology 86:2, 272-294. Warde, A. (1997). Consumption, Food and Taste: Culinary Antinomies and Commodity Culture. London: Sage. Waters, M. (1997). Inequality after Class. In Owen, D. (Ed.): Sociology After Postmodernism. London: Sage, 23-39. Wilenius, M. (1997). Faust on Wheels. Conceptualising Modernization and Global Climate Change. Commentationes Scientiarum Socialum 52:1997. Helsinki: Finnish Society of Letters and Finnish Academy of Science and Letters. Wilska, T-A. (1999). Survival with Dignity? The Consumption of Young Adults during the Economic Repression: A Comparative Study of Finland and Britain 1990-1994. Publications of the Turku School of Economics, Series A-3:1999. Turku: Turku School of Business Administration. Wilska, T-A. (2002). Me − A Consumer? Consumption, Identities and Lifestyles in Today’s Finland. Acta Sociologica 45:3, 195-210. Willis, P. (1977). Learning to Labour. Aldershot: Avebury. Wirth, L. (1938). Urbanism as a Way of Life. American Journal of Sociology 44:3, 124. Winquist, C. E. & Taylor, V. E. (2000). The Routledge Critical Dictionary of Postmodern Thought. London: Routledge.

266

Woodward, I. & Emmison, M. & Smith, P. (2000). Consumerism and Postmodern Space: A Modest Test of An Immodest Theory. British Journal of Sociology 51:2, 339354. Wright, E. O. (1985). Classes. London: Verso. Wrong, D. H. (1961). The Oversocialized Conception of Man in Modern Sociology. American Sociological Review 26:2, 183-193. Zablocki, B. D. & Kanter, R. M. (1976). The Differentiation of Lifestyles. Annual Review of Sociology 2, 269-298. Zafirovski, M. (1999). What is Really Rational Choice? Beyond the Utilitarian Concept of Rationality. Current Sociology 47:1, 47-113.

267

APPENDIX A

This appendix presents an English translation of the Finland 1999 (FIN99) research questionnaire. The original Finnish version of the questionnaire is available elsewhere (see Erola & Räsänen 2000). BACKGROUND INFORMATION 1. Are you? 1=Male 2=Female 2. What is the year of your birth? (With an empty line for filling in) 3. What type of accommodation do you live in? 1=Owner-occupied flat 2=Parents flat 3=Rented flat 4=Right of residence apartment 5=Other accommodation 4. What type of residential area do you live in? 1=Urban 2=Rural 5. What is your marital status? 1=Single 2=Cohabitation without marriage 3=Married 4=Divorced 5=Widowed 6. What type of household does your household represent? 1=Single 2=Single parent 3=Couple without children 4=Couple with children 5=Living with parents 6=Other type of household 7. Number of persons living in your household? (With an empty line for filling in)

269

8. Do you have descendants living elsewhere? 1=No 2=Yes, the number of under aged descendents (With an empty line for filling in) 3=Yes, the number of adult descendents (With an empty line for filling in) EDUCATION, OCCUPATION AND LIVELIHOOD 9. What is your basic education? 1=Part of lower elementary school or less 2=Elementary school 3=Higher elementary school 4=Secondary school or A-levels 10. What is your vocational training? 1=Unskilled 2=Vocational school or course 3=Intermediate level 4=Higher intermediate level degree 5=Academic degree 6=Post-graduate academic degree 7=Other (With an empty line for filling in) 11. What is your educational domain? (With an empty line for filling in) 12. If you are not employed at the moment, which one of the following is the reason? 1=Student 2=Unemployed 3=Pensioner 4=Disability pension 5=Parental leave 6=Nursing leave 7=Military service 8=Housewife/househusband 9=Other (With an empty line for filling in) 13. In which sector do you work? (Note: if you are not employed at the moment, answer in reference of your last job) 1=Public sector 2=Private sector 14. What is your occupation? (Note: if you are not employed at the moment, answer in reference of your last job) (With an empty line for filling in)

270

15. If you are participating working life, what kind of attachment you have? (Note: if you are not employed at the moment, answer in reference of your last job) 1=Permanent 2=Periodical, 1-5 years 3=Periodical, less than a year 4=Periodical, less than six months 5=Jobbing, contemporary 6=I am an entrepreneur 16. Have you been unemployed during the last five years? (Note: if you are not employed at the moment, answer in reference of your last job) 1=Yes 2=No 17. How many times and for how long have you been unemployed? (Note: if you are not employed at the moment, answer in reference of your last job) 1=I haven’t been unemployed 2=Once, for less than six months at a time 3=Several times, for less than six months at a time 4=Once, for more than six months at a time 5=Several times, for more than six months at a time 18. What is the occupation of your spouse? (Note: if you do not have a spouse, go to the question number 22) (With an empty line for filling in) 19. If your spouse is not participating working life, what is the reason? (Note: if you do not have a spouse, go to the question number 22) 1=Student 2=Unemployed 3=Pensioner 4=Disability pension 5=Parental leave 6=Nursing leave 7=Military service 8=Housewife/househusband 9=Other (With an empty line for filling in) 20. Have your spouse been unemployed during the last five years? (Note: if you do not have a spouse, go to the question number 22) 1=Yes 2=No 21. How many time, and for how long? (Note: if you do not have a spouse, go to the question number 22) 1=She/he haven’t been unemployed 2=Once, for less than six months 3=Several times, for less than six months at a time 4=Once, for more than six months 5=Several times, for more than six months at a time

271

22. What is/was the primary occupation of your father? (With an empty line for filling in) 23. What is/was the primary occupation of your mother? (With an empty line for filling in) 24. What is the primary source of income in your household? 1=Paid work 2=Entrepreneurship 3=Capital income 4=Pensions 5=Unemployment benefits 6=Home care subsidy 7=Subsistence subsidy 8=Family help 9=Other (With an empty line for filling in) 25. How would you describe your current economic situation? 1=Very good 2=Good 3=Moderate 4=Poor 5=Very poor 26. How would you describe your current economic situation in relation to the situation five years ago? 1=Much better 2=Better 3=Similar 4=Worse 5=Much worse 27. How would you expect your economic situation to become in the next five years? 1=Much better 2=Better 3=Similar 4=Worse 5=Much worse 28. Has your total consumption increased during the last five years? 1=Increased very much 2=Increased a little 3=Neither increased nor decreased 4=Decreased a little 5=Decreased very much

272

CONSUMPTION 29. How would you compare your outlay to an ‘average consumer’? (Evaluated at a Likert scale from 1=Very much, 5=Not very much) a) Food b) Housing c) Health d) Clothes e) Household equipment f) Furniture, decoration g) Kids’ accessories, toys etc. h) Household services i) Traffic and communication j) Movies, videos, records k) Leisure-time electronics l) Leisure travel m) Alcohol n) Entertainment, social life o) Sports p) Studying q) Culture r) Beauty care s) Self-pampering t) Special hobby, which one (with an open line for marking) u) Other (with an open line for marking) 30. Which are the three consumer activities that you would most likely increase if you could afford doing so? (With an empty line for filling in) 31. If your economic situation is not very good, how much do you do the following things? (Evaluated at a Likert scale from 1=Very much, 5=Not very much) a) I try to get an extra job b) I save from daily purchases c) I keep away form taking any loans d) I take a loan from my friends e) I take a consumer credit (credit card, account card, etc.) f) I take a bank loan g) I try to cut down all my expenses h) I begin to gamble and do lotto i) I try to live self-sufficiently j) I try to buy discount price products k) I do something else, what (With an empty line for filling in) 32. Have the patterns of your consumption expenditure changed during the economic recession? If yes, in which ways? (With an empty line for filling in)

273

33. Do you feel that the lack of time prevents you from leisure-time consuming? 1=Very often 2=Often 3=Sometimes 4=Seldom 5=Never 34. Which of the following things are necessities for you? (Evaluated at a scale from 1=Necessity and useful, 2=Useful, but not necessity to 3=Not useful, and not necessity) a) Paid work b) Leisure-time c) Home or children d) Watching television e) Own car f) Credit card g) Holiday trip twice a year h) Cultural services (theatre etc.) i) Restaurant evenings j) Proper personal life k) Living clean l) Other (With an empty line for filling in) 35. Which of the following durables are necessities in your everyday life? (Evaluated at a scale from 1=Necessity and useful, 2=Useful, but not necessity to 3=Not useful, and not necessity) a) Fridge b) Television c) Cable television d) Daily Newspaper e) Microwave f) Stereo set g) Video recorder h) Digital camera for photos i) Personal computer j) Internet connection k) Television game (Playstation etc.) l) Telephone m) Mobile phone n) Kitchen machine o) Other (With an empty line for filling in)

274

36. Would you increase your personal consumption of the following goods or services if you were able to do so? (Evaluated at a Likert scale from 1=Very much, 5=Not at all) a) Holiday trips b) Cultural services (e.g. theatre, opera) c) Donations to charities d) Helping family members e) Art and antique f) Dwelling change g) Car change h) Investments in stocks i) Credit cards j) Charge accounts k) ‘Shopping’ l) Impulse purchases m) Money saving n) Self-pampering o) Other (With an empty line for filling in) ATTITUDES TOWARDS CONSUMPTION AND LIFESTYLE 37. Do you agree with the following propositions? (Evaluated at a Likert scale from 1=I agree completely, 5= I don’t agree at all) 1) I don’t care about fashion 2) I don’t care about what people think about me 3) I take good care of my appearance 4) I’m concerned about the environmental effects of my consumption 5) Quality is often more important to me than the price 6) I think that the younger generation is irresponsible 7) I think that the older generation consumes irresponsibly 8) I want pleasure from my consumption 9) I often eat out 10) I often go pubs and bars 11) If I had more money, there wouldn’t be so many problems 12) High culture is snobbery 13) One should always save for ‘bad days’ 14) Our family dines always together 15) It is the obligation of the citizens to increase their consumption during the economic boom 16) I often shop from flea markets 17) I am know the latest trend in popular music 18) I feel that I’m not able to spend as much as I want 19) Life is too materialistic and there are too many products in the market 20) I often drink vine when I eat 21) I live economically 22) I often read fashion and home magazines 23) I consume consciously in an environment-friendly way 24) I often do impulse shopping

275

25) I often listen to classical music 26) People should live so that they can leave a proper legacy for their children 27) The fear of the next recession prevents me already from consuming 28) I often shop from sales 29) I am concerned about the origin of the food we eat 30) Consumer credit makes a good life possible 31) Taking loans from the bank is a good way to finance one’s studying 32) It is possible to increase one’s living standard permanently with reasonable loans from banks 33) Living on interest rates should be condemned 34) I finance my purchases by saving beforehand 35) One should avoid taking loans from the bank 36) One is able to achieve a high standard of living by working hard 37) Public nuisances of economic crime are exaggerated 38) Banks has been blamed too much from the economic depression FICTIONAL ACTIVITIES IN EVERYDAY LIFE 38. Do you agree with the following propositions if you buy a consumer durable and you have to make a decision between different alternatives? (Evaluated at a Likert scale from 1=I agree completely, 5=I don’t agree at all) a) Before buying I have to decide whether it is right or wrong to buy a product b) It is hard to make a right decision c) I can trust what somebody else recommends as a best solution d) I can make my decision on previous knowledge e) I prepare that my decision can eventually turn out as a failure 39. Do you agree with the following propositions if you buy a consumer durable and you have to make a decision between credit card payment and cash? (Evaluated at a Likert scale from 1=I agree completely, 5=I don’t agree at all) a) I have to decide whether it is right or wrong to pay with credit b) It is hard to make a right decision c) I usually accept what other people recommend me d) I know beforehand what type of payment I’m going to use e) I prepare that my decision can eventually turn out as a failure 40. Do you agree with the following propositions if you need a loan (for a house or a car, for instance) and you have to make a decision between different banks? (Evaluated at a Likert scale from 1=I agree completely, 5=I don’t agree at all) a) I have to decide whether it is right or wrong to take loan b) It is hard to make a right decision c) I can make my decision on previous knowledge d) I prepare that my decision can eventually turn out as a failure

276

41. Do you agree with the following propositions if you have to make an important decision (move to another town, for instance)? (Evaluated at a Likert scale from 1=I agree completely, 5=I don’t agree at all) a) I have to decide whether it is right or wrong to make a decision b) It is impossible to evaluate that kind of decisions c) It is usually the case that what other people recommend is also a good decision d) I can make my decision on previous knowledge e) I prepare that my decision can eventually turn out as a failure 42. Do you agree with the following propositions if you have to make an important decision related to your tasks at work? (Evaluated at a Likert scale from 1=I agree completely, 5=I don’t agree at all) a) I have to decide whether it is right or wrong to make a decision b) Every decision has to be considered individually c) I can make my decision on previous knowledge d) I prepare that my decision can eventually turn out as a failure 43. Do you agree with the following propositions if you have to make an important decision related to your companion with your loved one (moving together, for instance)? (Evaluated at a Likert scale from 1=I agree completely, 5=I don’t agree at all) a) I have to decide whether it is right or wrong to make a decision b) It is impossible to evaluate that kind of decisions c) It is usually the case that what other people recommend is also a good decision d) I can make my decision on previous knowledge e) I prepare that my decision can eventually turn out as a failure ECONOMIC RECESSION 44. There was a severe economic recession in Finland in the early 1990’s. How much did the recession affect your life? (Evaluated at a Likert scale from 1=Very much, 5=Not at all) a) Did it have negative effects b) Did it have positive effects c) Did you have to cut down your consumption in terms of considerable purchases (house, car etc.) d) Did you have to cut down your daily consumption (food etc.) e) Did the recession degrease your living standard f) Did the recession have negative effects on your work g) Did the recession have negative effects on your career h) Did the recession have negative effects on your family life i) Did the recession have positive effects on your family life j) Did the recession have an effect on your health k) Does the recession have negative effects on your life today l) Does the recession have positive effects on your life today

277

45. Do you think that the economic recession is completely over in Finland? 1= Yes, completely 2=Yes, partly 3=Not completely 4= Not at all SOLIDARITY AND DISSIMILARITY 46. How much do feel of belongingness to the following social categories? (Evaluated at a Likert scale from 1=Very much, 5=Not at all) a) Family b) Circle of friends c) Staff at work/friends at school d) Neighborhood e) Part of the town f) Town or municipality g) Church h) Hobby or other leisure time group i) Finnish Society j) European Union 47. How similar do you think you are in terms of the following things? (Evaluated at a Likert scale from 1=Very much, 5=Not at all) a) Expenditure b) Using consumer loans c) Taking loans from the banks d) Saving e) Travelling f) Activities with friends g) Values h) Dressing i) Fashion j) Alcohol consumption k) Sex life GENERATIONS AND IDEOLOGIES 48. Do you think that there exists a conflict between the generations? (Evaluated at a Likert scale from 1=Very much, 5=Not at all)

278

49. Which one of the following terms characterises your generation? 1=Baby-boomer generation 2=Turning point generation 3=Generation X 4=Suburban generation 5=Computer generation 6=Generation of war and shortage 7=Welfare generation 8=Reconstruction generation 9=Generation of the recession 10=Weakling generation 11=Incapable generation 12=Other (With and empty line for filling in) 13=Unnamed generation 50. Which one of the following terms characterises your children’s generation? 1=Baby-boomer generation 2=Turning point generation 3=Generation X 4=Suburban generation 5=Computer generation 6=Generation of war and shortage 7=Welfare generation 8=Reconstruction generation 9=Generation of the recession 10=Weakling generation 11=Incapable generation 12=Other (With and empty line for filling in) 13=Unnamed generation 51. Which one of the following terms characterises your parents’ generation? 1=Baby-boomer generation 2=Turning point generation 3=Generation X 4=Suburban generation 5=Computer generation 6=Generation of war and shortage 7=Welfare generation 8=Reconstruction generation 9=Generation of the recession 10=Weakling generation 11=Incapable generation 12=Other (With and empty line for filling in) 13=Unnamed generation

279

52. Which of the following social class do you predominately identify with? 1=Upper class 2=Upper middle class 3=Lower middle class 4=Working class 5=None of the classes 6=Other (With an empty line for filling in) 53. How do you predominately vote in political elections? 1=I predominately vote a person regardless of the party 2=I predominately vote a party regardless of the persons 3= I do not vote very often 54. Which one of the following political parties is closest to you? 1=Social democrats 2=Coalition 3=Center 4=Left Wing 5=Swedish party 6=Greens 7=Christians 8=Young Finns party/Progressive Liberal party 9=Communists 10=Other (With an empty line for filling in) SECURITY, INSECURITY AND RISKS 55. How important the following things are for you? (Evaluated at a Likert scale from 1=Very important, 5=Not at all important) a) Self-respect b) Interesting life c) National security d) Love e) World peace f) Material well-being g) Good life h) Economic equality i) Other equality j) Effectuation k) Freedom l) Independence m) Health n) Work o) Leisure time

280

56. How safe do you think that the following things are? (Evaluated at a Likert scale from 1=Very safe, 5=Not at all safe) a) Your own life b) Family c) Personal relationships d) Residential area e) Downtown are of the city/municipality f) Finnish Society g) European intergration h) Internationalization i) World situation 57. How much the following things increase your social insecurity? (Evaluated at a Likert scale from 1=Increase very much, 5=Does not increase at all) a) Restriction of the social and health care resources b) Restriction of police resources c) Restriction of military resources d) Restriction of social benefits e) Immigration f) Restriction of schooling recources g) Restriction of the pensions h) Restriction of unemployment payments i) Cutting down work benefits j) Changes in the national alcohol policy k) Restriction of the institutional health and crime care l) Restriction of the family support m) Debts of the state n) Environmental pollution o) Other (With an empty line for filling in) 58. How much the following things increase your feeling of personal insecurity? (Evaluated at a Likert scale from 1=Increase very much, 5=Does not increase at all) a) Your own health b) Personal relationships c) Housing d) Studying e) Tasks at work f) Unemployment g) Incapacity for work h) Dependency upon someone else’s care i) Problems in family members lives j) Residential area k) Household economy l) National economy m) National situation n) World situation o) Environmental factors p) Contemporary values and attitudes g) Other (With an empty line for filling in)

281

59. To what degree do you consider the following as social problems? (Evaluated at a Likert scale from 1=Very much, 5=Not at all) a) Using nuclear power b) Drugs c) East criminality d) Unemployment e) Low voting rates f) Slight punishments for economic criminals g) The level of crime punishment h) Television and other entertainment violence i) Pornography j) Prostitution k) AIDS l) Gene technology m) Aging of the population n) Internationalization o) Global growth of population p) Other (With an empty line for filling in) 60. How much do you think that the following things include risks? (Evaluated at a Likert scale from 1=Very much, 5=Not at all) a) Part-time employment b) Investments in stock c) Taking loans d) Guaranteeing loans e) Consumer credits f) Saving in banks g) Over-consuming h) Gambling i) Impulse purchases j) Travelling k) Speeding with a car l) Flying m) Drugs for medication n) Alcohol o) Drugs abuse p) Staying out at night q) Speaking to total strangers r) Sleeping around s) Other (With an empty line for filling in)

282

INCOME, EXPENSES, SAVINGS AND LOANS 61. What is the amount of household income a month? a) Please, indicate cross amount in FIM (With an empty line for filling in) b) Please, indicate net amount in FIM (With an empty line for filling in) 62. What is the amount of your personal income a month? a) Please, indicate cross amount in FIM (With an empty line for filling in) b) Please indicate net amount in FIM (With an empty line for filling in) 63. Consider all costs related to your consumption; what is the amount of additional savings of your household? Please, indicate the amount in FIM (With an empty line for filling in) 64. How much your household expenses exceed the income? Please indicate the amount in FIM (With an empty line for filling in) 65. How much does your household have mortgage? Please indicate the amount in FIM (With an empty line for filling in) 66. How much does your household have study loan? Please indicate the amount in FIM (With an empty line for filling in) 67. How much does your household have loan for a car? Please indicate the mount in FIM (With an empty line for filling in) 68. How much does your household have other loans? Please indicate the amount in FIM (With an empty line for filling in) 69. Are you planning to take loan in the near future? If yes, please indicate the amount in FIM (With an empty line for filling in) 70. Do you have personal consumer credits? If yes, please indicate the amount in FIM (With an empty line for filling in) 71. How many credit or account cards do you have? a) (With an empty line for filling in) b) The limit of my account is (With an empty line for filling in) 72. How often do you make purchases by using credit or account card? 1=Daily 2=A couple times a week 3=Weekly 4=A couple time a month 5=Once a month 6=Less than once a month

283

73. How much do you have other types of loans? Please indicate the amount in FIM (With an empty line for filling in) 74. Do you have intentions to take more consumer credit in the near future? 1=Yes 2=No 75. Do you have problems with your payment records or credit status information? 1=No 2=I have had problems with the payment records or credit status information, but not any more. Please indicate also the number of cases (With an empty line for filling in) 3=I currently have problems with the payment records or credit status information. Please indicate also the number of cases (With an empty line for filling in) 4=I don’t know 76. Do other people living in the same household have problems with the payment records or credit status information? 1=No 2=They have had problems with my payment records or credit status information, but not any more. Please indicate also the number of cases (With an empty line for filling in) 3=They currently have problems with my payment records or credit status information. Please indicate also the number of cases (With an empty line for filling in) 4=I don’t know COMMENTS AND FEEDBACK (With empty lines for filling in)

284

APPENDIX B

This appendix presents a translated variable list of the main categories and subcategories in the Finnish Expenditure Survey 1998 (FES98) data. Categories are presented according to the original variable codes in the data (C01-C12). The list of original variables in Finnish is available elsewhere (see Tilastokeskus 2001). C01 FOOD AND NON-ALCOHOLIC BEVERAGES C011 Food C012 Non-alcoholic beverages C0121 Coffee, tea and cocoa C02 ALCOHOLIC BEVERAGES AND TOBACCO C021 Alcoholic beverages C022 Tobacco C03 CLOTHING AND FOOTWEAR C031 Clothing C032 Footwear C04 HOUSING, WATER, ELECTRICITY, GAS AND OTHER FUELS C041 Housing costs in all C042 Real housing costs C043 Maintenance of house and reparations C044 Miscellaneous costs related to housing C0441 Water C0442 Waste treatment C0444 Miscellaneous costs related housing C045 Electricity, gas and other fuels C05 FURNISHINGS, HOUSEHOLD EQUIPMENT AND ROUTINE MAINTENANCE OF THE HOUSE C051 Furniture, interior decorations, carpeting and other floor coatings C0511 Furniture and interior decorations C0512 Carpeting and other floor coatings C052 Household textile C053 Household equipment C054 Glassware, cutlery and household utilitary articles C055 Tools and equipment for maintenance of the house and gardening C056 Products and services for routine maintenance of the house C0561 Short-lived consumer durables C0562 Domestic help and other services for maintenance of the house

287

C06 HEALTH C061 Drugs and pharmaceutical instruments and equipment C0611 Drugs C0612 Other pharmaceutical products C0613 Therapeutical equipment C062 Fees for treatment of outpatients C0621 Doctors’ fees C0622 Dentists’ fees C0623 Other fees for treatment of outpatients C063 Hospital fees C07 TRANSPORT C071 Acquisition of vehicles C0711 Cars C0712 Motorcycles and snowmobiles C0713 Bicycles C072 Running costs of private vehicles C0722 Fuels and lubricants C0723 Service and repair costs of vehicles C0724 Miscellaneous costs of vehicles C0703 Transport services C0703 Train, tram and subway fares C0732 Bus and taxi fares C0733 Flights C0734 Boat trips C0735 Miscellaneous transport services C08 COMMUNICATION C081 Communication in all C0811 Postal services C0812 Communication equipment C0813 Communication services C09 RECREATION AND CULTURE C091 Audiovisual equipment, cameras and equipment for data-processing C0911 Equipment for audiovisual playback C0912 Cameras for photographing and filming and other optic equipment C0912 Computers, calculators and typewriters C0913 Audiovisual recorders C0915 Repair costs for audiovisual equipment, cameras and equipment for data processing C092 Other durables for leisure and recreation C0921 Durables for outdoor leisure and recreation C0922 Durables for indoor leisure and recreation C0923 Repair costs for other durables for leisure and recreation C093 Miscellaneous goods and equipment for recreation; gardening products and pets C0931 Games, toys and equipments for hobbies C0932 Sports and camping equipment C0933 Flowers and gardening products C0934 Pets and pet facilities

288

C0935 Medical and other services for pets C094 Services for recreation and culture C0941 Sport and recreational services C0942 Cultural services C0943 Pools and lottery C095 Newspapers, books and stationary C0951 Newspapers C0952 Newspapers and magazines C0953 Maps, calendars, cards and other printed papers C0954 Stationary and office supplies C096 Package tours C10 EDUCATION C11 RESTAURANTS AND HOTELS C111 Restaurant services C1111 Restaurants and café C1112 Lunchrooms and canteens C112 Accommodation services C12 MISCELLANEOUS GOODS AND SERVICES C121 Cosmetics and beauty care C1211 Hairdressers, barbers and other services for beauty care C1212 Hairdryers, shavers and other electric devices C1213 Miscellaneous goods for beauty care C123 Miscellaneous goods C1231 Jewelry and watches C1232 Other personal goods C124 Social security C125 Insurances C126 Bank and other financial services C127 Miscellaneous services

289