Industrial Democracy Economic and

Economic and Industrial Democracy http://eid.sagepub.com

East—West Differences in Employment Relations, Organizational Justice and Trust: Possible Reasons and Consequences Thomas Rigotti, Kathleen Otto and Gisela Mohr Economic and Industrial Democracy 2007; 28; 212 DOI: 10.1177/0143831X07076111 The online version of this article can be found at: http://eid.sagepub.com/cgi/content/abstract/28/2/212

Published by: http://www.sagepublications.com

On behalf of: Department of Economic History, Uppsala University, Sweden

Additional services and information for Economic and Industrial Democracy can be found at: Email Alerts: http://eid.sagepub.com/cgi/alerts Subscriptions: http://eid.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations (this article cites 42 articles hosted on the SAGE Journals Online and HighWire Press platforms): http://eid.sagepub.com/cgi/content/refs/28/2/212

Downloaded from http://eid.sagepub.com at Universitaetsbibliothek Trier on September 20, 2007 © 2007 Department of Economic History, Uppsala University, Sweden. All rights reserved. Not for commercial use or unauthorized distribution.

East–West Differences in Employment Relations, Organizational Justice and Trust: Possible Reasons and Consequences Thomas Rigotti, Kathleen Otto and Gisela Mohr University of Leipzig

In a cross-sectional questionnaire study with 245 employees from West Germany and 357 employees from East Germany, the inferior labour market and economic situation in East Germany could be shown to be related, as expected, to the psychological contract, with a poorer delivery of the deal and lower levels of justice and trust. Affective commitment, on the other hand, was shown to be higher in the East than in the West German sample. Moderated multiple regression analyses revealed that the cognitive perception of psychological contract breaches is less related to emotional strain and distributive justice in the East German than in the West German sample. Results are discussed within the framework of economic and normative differences, and also with reference to social exchange theories.

Keywords: commitment, East Germany, justice, psychological contract, trust

Introduction On 31 August 1990, the Federal Republic of Germany (FRG) and the German Democratic Republic (GDR) signed the contract of unification. Two systems merged together, with enormously different economies and labour markets. After the reunification of the two German states the whole economic, corporate, legal and institutional system of West Germany was transported virtually unchanged to the East. The adaptation process to the new demands turned the economy of East Germany completely on its head. In Economic and Industrial Democracy & 2007 Uppsala University, Sweden, Vol. 28(2): 212–238. DOI: 10.1177/0143831X07076111


Rigotti et al.: Organizational Justice in East and West Germany

213

addition to these massive economic changes, the forced implementation of the entire West German system continues to be perceived by many as being unfair. East Germans did not have a voice, i.e. they were not asked whether they accepted or rejected at least some of the West German institutions. On the contrary, even those structures established in the former GDR that could be seen as being more progressive (e.g. shorter education to get a university entrance diploma) were simply abolished. Thus, it is questionable whether or not East Germans can develop trust and perceive justice in the new economic, corporate, legal and institutional system. This study consequently aimed at exploring and comparing the perceptions of East and West German employees regarding the exchange of promises with employers (psychological contract), organizational justice and trust, and the impact on work-related outcomes. We assume that differences in perceptions between East and West Germany are related to different conditions on the labour market, as well as differences in values and norms.

East Germany vs West Germany: Are There Still Differences? Shortly after unification, tremendous economic modifications began in East Germany due to the fact that its industry still relied on mass production and its service sector was underdeveloped. Nonetheless, nearly 15 years after the unification have passed and differences between the eastern and western parts are still glaringly apparent. Today, unemployment rates in the new federal states are about 10 percent higher than in the old federal states. Looking only at the German territorial states, we find unemployment rates between 6.2 and 10.2 percent in the western and 16.7 and 20.5 percent in the eastern parts of Germany (Statistische A¨mter des Bundes und der La¨nder, 2005). The discrepancies in these rates surely speak for themselves. Furthermore, approximately 44 percent of the East German unemployed are long-term unemployed compared to (only) 35 percent of the West German unemployed (Statistisches Bundesamt Deutschland, 2005). The situation is not only more critical for trained East Germans but also for school students and trainees from the new federal states who are planning to start vocational training or to enter the labour market. Only 30 percent of the East German trainees can assume that they will get a permanent contract


214

Economic and Industrial Democracy 28(2)

after finishing their vocational training (Gru¨nert and Lutz, 2003), and for only half of the East Germans of the same age group do jobs exist at all (Lutz, 2001). To sum up, the cited statistics illustrate that the economic and labour market situation in East Germany seems to be much worse than in the West. In addition to the aforementioned structural distinctions, differences can be expected on the individual level with respect to personal history and attitudes towards work and family resulting from the differences in the socialization processes: ‘Although linguistically and historically one people, 50 years of very different economic systems led people to perceive the nature of work quite differently’ (Kirkcaldy et al., 1999: 121). For example, East Germans in general perceive higher job insecurity (Otto and Dalbert, forthcoming) than West Germans, and this independent of their professional biography and actual work situation (e.g. current position, frequency of unemployment). Consequently, they value job security much higher than West Germans (Heyn et al., 1997), and prefer secure jobs over jobs that demand, among other things, a lot of responsibility or autonomy at work (Maier et al., 1994). Moreover, it was found that East Germans show higher uncertainty avoidance (Danne, 1996). The economic transitions in East Germany were also connected to avoidance of individual responsibility (Frese et al., 1997). It has been reported that workforces in East Germany show high organizational commitment (Mesner-Andolsek and Stebe, 2004) and possess an increased readiness to agree to concessions (Brixy and Christensen, 2002). In an East–West comparison, they were more ready to practise a job that was below their professional qualification or to accept changing working times, longer journeys to get to work and less satisfying working conditions. However, they were less ready to tolerate financial penalties. This should be seen in light of the fact that there is still a wage gap between East and West Germany, which is more accentuated for women, who earn 75 percent of the men’s wages in the western parts and 95 percent of the men’s wages in the eastern parts (Lu¨hrig, 2002). Although a general trend towards individualism can be observed all across Europe (see Guest, 2004), there is some evidence to suggest that East and West Germans still differ on the collectivisticindividualist dimension, with higher scores on a collectivistic orientation in the eastern part, even among students who had only spent their very early childhood years in the former GDR (Spieß and



215

Bru¨ch, 2002). Overall, it can be assumed that people in the eastern part of Germany are likely to have different perceptions of employment relations due to the diverse socialization and the different labour market conditions they experienced (see Borg and Braun, 1996; Frese et al., 1997).

Psychological Contracts Psychological contracts have become a prominent theoretical framework for the description of individually perceived employment relationships. The psychological contract (PC) is defined as an ‘individual’s belief about mutual obligations in the context of the relationship between employer and employee’ (Rousseau, 1990: 391). Those obligations implied in the employment relationship may also include the perception of promises and their fulfilment, which we call the ‘delivery of the deal’ (see Guest, 2004). Albeit, we agree with a social exchange approach of psychological contracts, including mutuality, in this article we limit our scope to organizational inducements, i.e. employers’ promises and obligations as they are perceived by employees. Both positive effects of contract fulfilment on work-related attitudes and citizenship behaviour, as well as negative outcomes of contract violation have formed the core of empirical studies over the past few years (see Lester et al., 2002). Recent empirical studies have shown that individual traits influence the perceptions of PCs, as well as moderating the effects of a perceived breach on diverse outcomes (Ho et al., 2004; Raja et al., 2004; Rigotti, 2005). When comparing samples from different nations, cultures or, in our case, federal states, the underlying assumption is that contextual factors have an impact on collectively shared norms and values within one group, which may then explain differences between two groups. Our behaviour is based upon cultural values and norms (Schein, 1992), which form individual perceptions about how things ought to be or how one should behave. Thomas et al. (2003) assume that cultural values may have an impact on the formation of PCs. According to these authors, culture may influence the perception of PCs via the shaping of cognitions and motives. Rousseau and Schalk (2000) postulate that the societal framework may influence


216


the ‘zone of negotiability’ of PCs. As we have the same legal employment regulations in all of Germany, we assume potential differences between East and West Germany to be caused by structural differences (like unemployment rates, union density, economic growth rates) and/or socialization. These possible determinants (culture and economy) may to some extent lead to different hypotheses when it comes to comparing East with West Germany. Cross-national research on PCs has so far focused mainly on the description of employment relations in different societal contexts (e.g. Rousseau and Schalk, 2000) and failed to establish theorybased comparative studies. First, comparative studies, however, point towards the cultural sensitiveness of the PC concept. Kickul et al. (2004), for example, report differences in the perceptions of PCs of American and Hong Kong Chinese, where Americans placed higher importance on PC outcomes but also perceived fewer breaches. Union density, how easy it is to fill a vacancy, the regional unemployment rate, the qualification structure of potential employees are all factors that may be connected to the establishment of a PC – for both partners. Employers who are not depending on the longterm commitment of their employees, because open vacancies can easily be filled, may give fewer pledges. If we follow this line of argumentation with respect to the inferior labour market conditions in East Germany as compared to West Germany, we can expect employers to need to give fewer promises in order to find and keep a qualified workforce, and thus we can expect employees in East Germany to perceive a lower delivery of the deal regarding obligations and promises of their employers: H1: Employees in East Germany report a lower delivery of the deal of employers’ obligations and promises compared to West German employees. Perceived breaches of PCs should have greater effects in situations where employees have high expectations towards organizations. It is possible that employees perceive a breach of promises without experiencing a strong emotional response to this event (see Morrison and Robinson, 1997). As discussed in the introduction, people in the East were found to possess an increased readiness to agree to concessions (Brixy and Christensen, 2002) and should therefore react to breaches with lower levels of strain than West Germans. This



217

would also be in line with the proposition that collectivists will have a ‘higher overall threshold for the perception of a PC violation than will individualists’ (Thomas et al., 2003: 461). Low values in the delivery of the deal can also be interpreted as a breach of perceived promises. We therefore assume: H2: East Germans experience lower emotional strain when perceiving a low delivery of the deal of the psychological contract than do West Germans. The perception of mutual promises and obligations is closely linked to perceived justice and trust in the social exchange process between employers and employees (see Kickul et al., 2001; Robinson, 1996). Hence, we also expected to find differences in perceptions of organizational justice and trust between East and West German employees. Therefore, in the following we would like to give brief definitions of these constructs and report some theoretical and empirical grounds that led us to our hypotheses.

Organizational Justice and Trust Organizational justice, in the psychological connotation, is an individual’s perception concerning the allocation of resources (distributive justice), the implementation of changes and procedures (procedural justice) and the social interaction (interactional justice) of people in the workforce (see Folger and Cropanzano, 1998). ‘Justice and fairness are properties that exist largely in the eye (and the mind) of the beholder’ (Morris and Leung, 2000: 101). Numerous investigations have shown that distributive and procedural justice cognitions are relevant for evaluating work-related attitudes (e.g. Tremblay and Roussel, 2001; Younts and Mueller, 2001). Studies focusing on distributive justice measured how jobs, wages or career advancements should be distributed if they are to be perceived as being fair, while studies from the procedural justice perspective looked more at the effects of the level of employees’ participation in the decision-making within the organizations (for a review, see Colquitt et al., 2001). Distributive justice perceptions are related more to cognitive, emotional and behavioural reactions to particular outcomes, whereas procedural justice appears to be a stronger predictor for reactions towards the organization as a


218


whole. Interactional justice is something that seems to affect the perception of the direct supervisor or source of justice (see CohenCharash and Spector, 2001). Others found justice to act as a mediator between leadership behaviour and trust (e.g. Pillai et al., 1999). And trust, to continue this path, has again been found to mediate between justice and work-related outcomes (e.g. job satisfaction or commitment, Aryee et al., 2002). Fischer and Smith (2004) could find greater perceived justice among British employees than among East German ones, and postulated that individual values may play a moderating role: ‘Thus, the perception of justice is an interaction between the actions of authorities and individuals’ values’ (Fischer and Smith, 2004: 672; see also Fischer, 2004). Several authors reported cultural differences in effects of perceived organizational justice (e.g. Brockner et al., 2001; Tyler et al., 2000). Regarding these earlier empirical findings, and also the described inferior labour market situation in the eastern part of Germany (higher unemployment rates, lower career prospects, higher job insecurity and lower wages), we expected to find lower values of perceived justice in the East as compared to the West: H3: Employees in East Germany report lower levels of justice than do West German employees. Whether unmet promises result in attitudinal and behavioural reactions may strongly rely on the justice judgement (see Morrison and Robinson, 1997). We expect that the relation between cognitively perceived breach of contract and perceptions of justice should be stronger in the West German sample than in the East German sample. H4: Delivery of the deal of psychological contracts is more strongly related to justice perceptions in West than in East Germany. A further important construct, when looking at the quality of employment relationships, is trust (e.g. Guest, 2004; Robinson, 1996). In line with the previous arguments, we also expected to find differences in perceptions of trust between the East and West German samples. Even though there is less empirical evidence for this relation than for organizational justice, both constructs are closely linked, i.e. they are based on one another, in the sense that



219

trust cannot be developed without perceiving justice and, in turn, that trusting other people in exchange relationships itself strengthens the perception that justice will be delivered in the long term (for a review, see Lewicki et al., 2005). Thus, we derive our hypothesis from the findings of organizational justice research: H5: Employees in East Germany report lower levels of trust than West German employees. While we assume that delivery of the deal is more strongly linked to perceptions of justice in the West as compared to the East German sample (H4), we do not expect a moderating role of being East German in the relationship between delivery of the deal and trust. Even though trust is closely linked to justice it is clearly a different construct. Besides the justice component, many other factors enhance the development of trust, like openness (the willingness to be open and transparent in one’s actions and intentions), integrity (adherence to principles and willingness to follow through on commitments), benevolence (the desire to do good towards the other) or competency (having skills and capabilities in a key relationship context) and so on (for a review, see Lewicki et al., 2005). Looking at the components of trust, there is no reason to expect that a perceived contract breach should have a more negative meaning for employees in West Germany compared to those in East Germany

Differences in Attitudes between East and West Despite the disadvantaged position of East Germany, organizational commitment is reported to be higher in East than in West Germany (Mesner-Andolsek and Stebe, 2004). Wasti (2003), who employed a Turkish sample and focused on an individual level, reports a substantial positive association of horizontal collectivism and affective commitment (N ¼ 801–890, r ¼ :29, p < :001). A collectivistic cultural background should therefore predict higher organizational commitment: H6: On average, East German participants report higher affective commitment than West Germans (which should be true even when background variables are controlled for).


220


There is no theoretical or empirical evidence, however, to support the idea that perceived contract breaches, organizational justice perceptions or trust, and their association to work-related attitudes, should generally differ between distinct groups (e.g. cultures, nations). Favourable correlates of perceived justice, such as trust or commitment, have been found across different country samples (see Morris and Leung, 2000).

Method Sample The sample is part of the research project PSYCONES. Data were gathered in 33 organizations from the educational, retail and food sectors across Germany in 2004. In these three sectors we have (1) private companies as well as public organizations, (2) a broad variety of educational and skill levels and (3) manufacturing as well as service orientations. Thus, central features across the sectors match, leading to comparable samples of employees working in the 33 organizations. Overall, we gathered data from 643 employees. As we were interested in distinguishing East and West German employees, we examined our East and West German participants further. In West Germany, 262 employees from 14 organizations, and in East Germany, 381 employees from 20 organizations, participated. However, we found 24 persons now working in East Germany who did not grow up in the former GDR, and 17 persons who grew up in the former GDR but were now working in West Germany. In accordance with our aim to compare the East and West German samples (based on both economic and valuedifferences), we focused only on those individuals who had not moved to the respective other part of Germany. Consequently, we excluded the movers from our sample, which then resulted in N ¼ 245 employees from West Germany and N ¼ 357 employees from East Germany. Almost half of the respondents (45 percent) had a non-permanent employment contract. The percentage of females in the West German sample, 35 percent, was far below the share in the East German sample, which was 62 percent (2 ð1Þ ¼ 43:40, p < :01). There were more bluecollar workers in the West German sample (38 percent as compared to 16 percent, 2 ð1Þ ¼ 38:80, p < :01), and the average age was lower



221

TABLE 1 Sample Characteristics after Exclusion of People Who Had Moved to the Respective Other Part of Germany West

East

N N N N Organizations Employees Organizations Employees Food Retail Education Total Percentage blue-collar Percentage females Age Educational level a Lower secondary level Upper secondary level Post-secondary level First stage of tertiary education Second stage of tertiary education a

7 2 5

142 20 83

2 9 9

245 38% 35% M ¼ 35:15 ðSD ¼ 11:69Þ

20

14

51 171 135

357 16% 62% M ¼ 38:93 ðSD ¼ 11:89Þ

22% 42% 5% 31%

7% 28% 6% 50%

0%

5%

Levels according to the international educational classification scheme ISCED (OECD, 1999).

in the West (M ¼ 35:15, SD ¼ 11:69), than in the East (M ¼ 38:93, SD ¼ 11:89; t ¼ 3:83, p < :01; see Table 1 for details). Moreover, the educational background was, on average, higher in the East German sample (M ¼ 4:16, SD ¼ 1:14) than in the West (M ¼ 3:45, SD ¼ 1:15, t ¼ 7:47, p < :01), measured against the international classification of educational levels, ISCED (OECD, 1999). Therefore, we control for these differences in all multivariate analyses.

Instruments Controls. Age was assessed as a continuous variable and concatenated in five age categories for multiple classification analyses. Gender is a dichotomous variable with 0 for females and 1 for


222


males. Type of contract is a dummy, with 0 representing temporary and 1 permanent contracts. Position, albeit assessed on six levels, is also used as a dummy, contrasting blue-collar workers (1) with others (0). Educational level was measured by using a combination of school and vocational education and transformed to the six levels of the international standard scheme, ISCED (OECD, 1999). As usual, we considered this measure to be metric. The regional unemployment rate connected to the participating organizations was gathered in October 2004 from the official websites of the respective towns. The unemployment rate ranged from 3.80 percent to 19.70 percent. Delivery of the deal. Respondents were asked to indicate whether the organization had committed itself to a diverse range of 15 promises. Seven topics were covered: content of the job, employment relationship, compensation, advancement, participation, social relationship and work–family balance. Our subjects were requested to answer the 15 items after having been given the instruction ‘Has your organization promised or committed itself to . . .’, e.g. ‘. . . provide you with a reasonably secure job?’ Answers could range from 0 ¼ ‘no’, in cases where employees perceived no promise, and from 1 ¼ ‘yes, a promise was given but not at all kept’ to 5 ¼ ‘yes, promise given and fully kept’ to rate the fulfilment of the promise (see Rigotti and Mohr, 2004). Delivery of the deal is operationalized by computing the mean of promise keeping for every individual, with values ranging from 1 to 5. We performed factor analyses based on tetrachoric correlations (see, for example, McLeod et al., 2001) using the programme Testfact (Bock et al., 2003) on the dichotomous answers. Although there was some evidence for a two-factor solution (Eigenvalues: 8.86, 1.46, 0.93), a single factor explains 56 percent of the variance. The ratios of chi-square and degrees of freedom dropped only slightly from 2.52 (2 ¼ 1544:42= d:f: ¼ 612 for one factor), to 2.15 (2 ¼ 1282:73=d:f: ¼ 598 for two factors), applying 100 iterations. Nevertheless, even after rotation of factor loadings, there were many cross-loadings of items. For the one-factor solution all loadings were greater than .60. Hence, for the purpose of this study the use of the measure as only one underlying latent construct seems to be justified. Analyses of internal consistencies (using codings from 0 to 5) yielded alphas of .87 (East) and .92 (West).



223

Organizational justice. Three dimensions of organizational justice were assessed by single or two-item measures (items were taken from Guest and Conway [2002], who considered these items to reflect only one latent construct, ‘fairness’). Distributive justice was reflected by the two items ‘Overall, do you feel you are rewarded fairly for the amount of effort you put into your job?’ and ‘Do you feel you are paid fairly for the work you do?’, which showed a correlation of r ¼ :80 (p < :01), and yielded alphas of .88 (West) and .89 (East). Procedural justice was assessed with the item ‘Do you feel that organizational changes are implemented fairly in your organization?’; and interactional justice with the item, ‘Do you feel fairly treated by managers and supervisors?’ The response format ranged from 1 ¼ ‘not at all’ to 5 ¼ ‘totally’. The correlations between the three dimensions of justice (rdp ¼ :40, rdi ¼ :41, rpi ¼ :46, p < :001, N ¼ 590–599) are in line with findings of Cohen-Charash and Spector’s (2001) meta-analysis and underpin the validity of the one- and two-item measures respectively (see Table 2). Trust. To measure trust, we used a subset of three items from Guest and Conway (2002; e.g. ‘In general, how much do you trust your organization to keep its promises or commitments to you and other employees?’). Answers could be given on a five-level scale ranging from 1 = ‘not at all’ to 5 = ‘totally’. The scale yielded alphas of .79 in both subgroups. Commitment. A four-item measure was used to assess affective commitment (Cook and Wall, 1980; e.g. ‘I am quite proud to be able to tell people who it is I work for’). Answer format ranged from 1 ¼ ‘strongly disagree’ to 5 ¼ ‘strongly agree’. Cronbach’s alphas of the scale in our samples are .70 (West) and .65 (East). Emotional irritation. This variable was measured using a subset of the Irritation Scale (Mohr et al., 2005a, 2005b). Irritation is defined as a state of psychological impairment, caused by perceived goal discrepancy, and includes rumination about problems at work (¼ cognitive irritation) and irritability (¼ emotional irritation; see Mohr et al., 2006; Mu¨ller et al., 2004). We chose to focus on emotional irritation as this concept has been proven to be a precursor of further impairments, such as psychosomatic complaints or depression (e.g. Dormann and Zapf, 1999). The scale consisted of


.10

.18

.24

.03NS .09

East vs West

.31

.14

.27

.36

.59

.72

.58

.55

5

.14

.26

.39

.39

.31

.39

.47

6

.03NS

.19

.09NS

.14

.04NS

.12

.19

7 (4.10 (4.10) (3.32 (3.35) (3.31 (3.24) (4.08 (4.03) (3.56 (3.46) (3.76 (3.76) (2.30 (2.37)

M

.70 .85

1.16

.79

–

–

.88

.92

a

0.76

0.91

0.97

1.16

1.11

0.61

SD (3.97 (3.97) (3.24 (3.19) (3.10 (3.17) (3.89 (3.92) (3.28 (3.35) (3.95 (3.95) (2.39 (2.34)

M

1.12

0.65

0.96

1.04

1.13

1.13

0.72

SD

East

.85

.65

.79

–

–

.89

.87

a

0.64

4.77**

3.33**

2.40*

2.14*

0.81

2.38*

t

.08

.27

.30

.19

.18

.07

.19

d0

.01

.13**

.06**

.05*

.03*

.07

.10*

b

Notes: Upper diagonal: East German sample, N ¼ 314–357; lower diagonal: West German sample, N ¼ 228–245. All Pearson correlations are significant on a 1% alpha-level, except those indicated with NS, ¼ Cronbach’s alpha, t ¼ t-value of simple t-test for independent samples, d 0 ¼ referring effect size, ¼ partial correlations from MCA. * p < :05, ** p < :01 for t-tests and partial correlation coefficients of MCA. Means in parentheses after control of sex, age, education, blue collars, fixed-term contracts and sector.

.10

.29

.27

.38

6 Affective .36 commitment 7 Emotional irritation .33

.60

.44

.49

5 Trust

.67

.40

.46

.42

.54

4

4 Interactional justice .45

.45

.40

.46

3

.39

.43

2

.39

3 Procedural justice

2 Distributive justice .52

1 Delivery of the deal

1

West

TABLE 2 Pearson Correlations, Means (with and without control), Standard Deviations, Cronbach’s Alpha Separately for the East and West German Samples, Results of t-Tests for Independent Samples and MCA

224 Economic and Industrial Democracy 28(2)



225

five items (e.g. ‘I get in a bad mood when I am disturbed by others’), which could be rated on seven levels with 1 ¼ ‘strongly disagree’ to 7 ¼ ‘strongly agree’. With an alpha of .85 in both subgroups, the measure can be considered highly reliable.

Analyses Multiple classification analysis (MCA). MCA, as well as dummy variable regression (DVR), estimates additive models and assumes interactions between independent variables to be non-apparent (Jacobsen, 1978). Concerning the prediction of the overall model, MCA and DVR are interchangeable. The difference between the two approaches is, however, that MCA measures the association of the factors as a whole with the criteria, whereas DVR tests the ‘pattern of effects in relation to a reference category for each factor’ (Jacobsen, 1978: 89). In contrast to DVR, the criteria may be dichotomous in MCA (Jensen, 1978). As an additional result, we can see the observed mean values of each subgroup in the dependent variable in an MCA, as well as the estimation of the mean, controlling for all other predictors. The beta-values are analogous to beta-coefficients in multiple regressions, and can be interpreted as part correlation coefficients. For our main purpose of comparing the East and West German sample, we chose MCA. Hence, some results will be cross-validated with multiple regression analyses, which are used to test for moderating effects. Moderated multiple regression analysis (MMR). In order to test hypotheses concerning the difference in strength of an association between two variables across different groups, we performed MMR (see, for example, Aguinis, 2004; Aiken and West, 1991). In the first step, the predictor variable and the assumed moderator variable are entered and, in the second step, the product term of both variables is entered. A variable can be considered a moderator when the interaction term yields a significant beta-weight in the multiple regression model and the change in R2 (variance explained) is significant, compared to when the independent and moderator variables are only included separately (see Jaccard and Turrisi, 2003). From the R2 of the first and second step, effect sizes can be calculated (Aiken and West, 1991). Although Cohen et al. (2003) proposed labelling effect sizes of .02, .15 and .35 as small, medium


226


or large, Aguinis et al. (2005) found a mean effect size of f 2 ¼ :009 across 636 published MMR analyses. An often ignored problem in MMR with categorical moderator variables is the assumption of homogeneity of error variances across categories of the moderator variable, which may lead to an increase in Type I errors if sample sizes are unequal across subgroups, and an increase in Type II errors if the subgroup with larger sample size presents the larger error variance (Aguinis, 2004). In their meta-analytic review, Aguinis et al. (1999) could show that the assumption of homogeneity of error variances was violated in 39 percent of analyses reviewed. We therefore also present Bartlett’s (1937) test of homogeneous error-variances.

Results Mean Differences between East and West German Employees We hypothesized that participants from East Germany would give lower ratings on factors associated with the employment relationship, namely delivery of the deal (H1), justice perceptions (H3) and trust (H5). In Table 2, results from simple t-tests for independent samples, as well as from MCA controlling for important sample differences in age, sex (0 ¼ female, 1 ¼ male), education (ISCED classification), type of contract (0 ¼ temporary, 1 ¼ permanent), blue collars (1 = blue-collar, 0 = others) and sectors (sector_dummy1 ¼ 1 for food industry, sector_dummy2 ¼ 1 for retail, reference group educational sector). Except for distributive justice, the null hypotheses have to be rejected for all employment relationship measures, indicating that East Germans gave lower ratings on delivery of the deal (t ¼ 2:38, p < :05, d 0 ¼ :19), procedural justice (t ¼ 2:14, p < :05, d 0 ¼ :18), interactional justice (t ¼ 2:40, p < :05, d 0 ¼ :19) and trust (t ¼ 3:33, p < :01, d 0 ¼ :30). These mean differences do not vanish when controlling for sample characteristics in MCA, although effect sizes point towards small effects. Furthermore, we hypothesized that we would find higher ratings in commitment (H6) in the East German than in the West German sample. Differences in affective commitment turned out to be stronger than for the employment relationship measures. East Germans gave higher ratings on organizational commitment



227

(t ¼ 4:77, p < :01, d 0 ¼ :27). Effect sizes indicate a medium effect and differences are still significant after controlling for sample characteristics. We also tested whether or not the regional unemployment rate, seen as an important indicator for the economic situation, may explain differences between the East and West German samples. We therefore performed partial correlation analyses. The results are shown in Table 3. It turned out that perceptions related to the employment relationship are mediated by the regional unemployment rate, as partial correlations fail to be significant. Differences in affective commitment cannot, however, be explained by the regional unemployment rate.

The Impact of Delivery of the Deal when Comparing East and West German Employees Our first moderating hypothesis was that the effects of perceived breaches of the PC may be less pronounced for East Germans than for West Germans (H2). We employed an MMR by first introducing possible explanatory control variables: age, sex (0 ¼ female, 1 ¼ male), education (ISCED classification), type of contract (0 ¼ temporary, 1 ¼ permanent), blue collars (1 ¼ blue-collar, 0 ¼ others) and sectors (sector_dummy1 ¼ 1 for food industry, sector_dummy2 ¼ 1 for retail, reference group educational sector). In the second step, we introduced the sample dummy (0 ¼ West, 1 ¼ East) and delivery of the deal as a predictor variable. In the final step, we introduced the interaction between sample and delivery of the deal. The interaction term yielded a significant beta, and could explain 1 percent of additional variance. Effect size was f 2 ¼ :009 (see Table 4). Bartlett’s test indicated homogeneous error variances (M ¼ 0:44, p ¼ :50) across subgroups. The interaction plot using 1 SD (see Aiken and West, 1991) indicates that, in line with our hypotheses, the association between delivery of the deal and emotional irritation is stronger for West Germans than for East Germans (see Figure 1). Along the same lines, we tested whether or not perceived breaches of the PC (or delivery of the deal) led to different judgements of justice in the East than in the West German sample (H4). We used the same set of steps as predictors of distributive, procedural and interactional justice separately in three MMRs. A moderating


* p < :05, ** p < :01:

East/West Unemployment rate Partial correlation rsample.unemployment

rate

.10* .10* .02*

Delivery of the Deal .03 .03 .02

Distributive Justice .09* .10* .01*

Procedural Justice .10** .11** .01**

Interactional Justice

TABLE 3 Influence of Regional Unemployment Rate on Sample Differences

.14** .15** .01**

Trust

.14** .11** .11**

Commitment

.03 .00 .06

Emotional Irritation

228 Economic and Industrial Democracy 28(2)



229

TABLE 4 Moderated Multiple Regression Analyses DV:

Step 1 Age Sex Educational level Type of contract Blue-collar Sector_dummy1 Sector_dummy2 R2 Step 2 Delivery of the deal Sample R2 Step 3 Sample delivery of the deal R2 Adj. R2 Bartlett’s M f2

Emotional Irritation

Distributive Justice

Procedural Justice

Interactional Justice

.02 .04 .03 .07 .07 .09 .05 .03

*.12* .00 .06 .08 .06 *.04* .11 *.03*

.03 .00 .05 .01 .05 .01 .13* .03

.04 .01 .01 **.12** .06 .07 .01 **.04**

**.36** *.59* **.06**

**.57** .45 **.21**

**.45** .08 **.18**

**.48** .15 **.25**

*.57*

*.49*

.07

.12

*.01*

*.01*

.00

.00

.07

.24

.19

.27

M ¼ 0:44 (p ¼ :50) .009

M ¼ 1:54 (p ¼ :21) .008

M ¼ 1:14 (p ¼ :29)