Feb 1, 1996 - human resource (HR) practitioners is to create a work environment in which ..... style was the most significant predictor of organizational ...
The International Journal of Human Resource Management 7:1 February 1996
Employee acceptance of organizational change: the role of organizational commitment
Roderick D. Iverson Abstract This study tests a causal model that predicts the acceptance of organizational change using a sample of 761 employees from a large public hospital in the state of Victoria, Australia. The LISREL results indicate that employee acceptance of organizational change is increased by organizational commitment, a harmonious industrial relations (IR) climate, education, job motivation, job satisfaction, job security and positive affectivity, and is decreased by union membership, role conflict, tenure and environmental opportunity. Organizational commitment was found to act as both a determinant and mediator in the change process. Implications for the management of organizational change using human resource (HR) strategies and policies are discussed. Keywords Organizational change, flexibility, organizational commitment, industrial relations climate, hospital employees. Introduction
The French proverb, the more things change the more they remain the same (Goldstein, 1988) reflects the inevitable dilemma experienced by employees when confronting organizational change (Gardner et al., 1987; Watson, 1971). Management within organizations are experiencing a myriad of intemal and external pressures for organizational change. These pressures include production and process technology, govemment laws and regulations, major political and social events, the intemationalization of business, increases in the size and complexity of organizations and economies of scale (Pfeffer, 1994). In this situation, the one thing that has remained the same is the requirement of employees to adapt to organizational change. The concept, organizational change, implies an attempt or series of attempts to modify an organization's structure, goals, technology or work tasks (Camall, 1986). Moreover, concepts such as innovation, which is the development, acceptance and implementation of new ideas, processes, products or services (Kanter, 1983; Van de Ven, 1986), and functional flexibility, which is the capacity of employees to multi-skill (Cordery et al., 1993), are all part of the general rubric of organizational change. Hence, it is not surprising that employees display differences in receptiveness to the various forms of change through such responses as lack of trust (Gardner et al., 1987), frozen attitudes (Schein, 1979), non-participation (Coch and French, 1948) and fear of the unknown (Huse, 1980). As Dawson (1994: 14) notes, resistance to organizational change can result from one or a combination of the following factors: 'substantive change in job (skill requirements); reduction in economic security or job displacement; psychological 0985-5192
© Routledge 1996
Employee acceptance of organizational change 123 threats (whether perceived or actual); disruption of social arrangements; and lowering of status (redefines authority relationships).' Thus, the challenge to human resource (HR) practitioners is to create a work environment in which employees accept rather than resist change. Examining the change literature it becomes apparent that there is a plethora of both theories of change and strategies for the management of change within organizations. Although there is much debate surrounding the various theoretical perspectives of change, such as the typology of determinism to voluntarism (e.g., Astley and Van de Ven, 1983; Crozier and Friedberg, 1980; Hrebiniak and Joyce, 1985), Wilson (1992) notes that these perspectives can be divided into two broad conceptual schema. First, the systemic conflict framework views organizations as social systems which are characterized by conflict, politicking and inherent tensions. The impetus for change stems from attempting to resolve the conflicts and tensions within organizations (e.g., between employees). Theories which derive from this framework include contextualism, population ecology, organization life cycles, power in organizations, political models of change and social action theories (e.g.. Greenwood and Hinings, 1988; Hannan and Freeman, 1977; Kimberly and Miles, 1980; Pettigrew, 1985, 1990). Contextualism (Pettigrew, 1985, 1990), for example, focuses on the inner context of the organization (culture, structure and politics) and the outer context from which the organization operates (business, economic, political and societal). The process by which change unfolds or emerges is fundamental to contextualism's analytical framework. Contextualism, however, has been argued to be limited in regard to its abihty to operationalize the various propositions for use by managers (Dawson, 1994). Second, the strategic choice framework also views organizations as social systems, but emphasizes the notion of strategic choice (Child, 1972). The impetus for change stems from managers scanning the environment for possible solutions to organizational problems. Theories which derive from this framework include organizational development (OD), planned incrementalism, entrepreneuralism and intrapreneurialism and best practice (e.g., Japanization) (e.g., Lewin, 1951; Peters and Waterman, 1982; Quinn, 1980; Stace and Dunphy, 1991). One of the most prominent theories, the OD paradigm (e.g., Dunphy and Stace, 1988; Stace and Dunphy, 1991), for example, is founded on the premise of incrementalism, where the participation of individuals is integral to the change process. Lewin's (1951) three-phase model of change (unfreezing, changing and refreezing) provided much of the theoretical foundation for OD today. OD, however, is argued to be too normative and neglects to examine the processes by which change arises (Wilson, 1992). Two other theories that do not fit well within either framework are change which is initiated through mergers, joint ventures and acquisitions (Harrigan, 1987), and contingency theories which deal with organization-environmental fit (Bums and Stalker, 1961; Lawrence and Lorsch, 1967). The reason for their exclusion is that both theories can be classified as systemic conflict and strategic choice in nature. Turning to the strategies for the management of change, the main thrust is on delineating the steps involved in the change process, and attempting to build generalizable templates that can be successfully applied across work settings (e.g., Kanter, 1983, 1989; Morgan, 1988; Pfeffer, 1994). Both Kanter (1989) and Morgan (1988), for example, argue that what are required in the new global
124
Roderick D. Iverson
business world are leaders with strategic vision who are strong, competent and lateral thinkers. Nevertheless, a major problem in the literature, again, is the emphasis on normative approaches to the management of organizational change. Moreover, the paucity of empirical information on the success of such change programmes raises issues over both their utility and link with change models (Wilson, 1992). None the less, with the increase in market emphasis prompted by the current economic environment, HR managers are recognizing that to achieve a competitive advantage they need to be able to introduce and manage change successfully within organizations (Aryee, 1991; Bamber, 1990; Dawson, 1991; Pfeffer, 1994). This is particularly the case within the Australian economy, which is continuing to open up to external market forces in formerly protected domestic markets. Moreover, for public-sector organizations there is considerable pressure to 'improve efficiency, by reducing costs, improving qudity, and increasing client responsiveness' (Curtain, 1993: 31). When this pressure is combined with rapid changes in technology and market demands (Piore and Sabel, 1984), organizations are increasingly accepting that work-force flexibility (e.g., functional, numerical or temporal flexibility) must be a normal part of their business practices (Bamber, 1990). Pfeffer (1994) notes that the traditional sources of success such as product and process technology provide less of a competitive advantage today than in the past, leaving how the work-force is managed as comparatively more vital. Thus, for organizations to improve their bottom line they must have adaptable work-forces which are positively disposed to change (Guest, 1992). The purpose of this paper is to develop a causal model which is capable of predicting employee acceptance of organizational change. Deriving from the various theories and strategies for the management of change, the model includes HR policy variables, as well as individual and environmental factors that enhance our understanding of the change process. However, in contrast to many of these theories and strategies, pivotal to this change model is the mediating role of organizational commitment, which is argued to act both as a determinant and an intervening variable in the change process. It is via organizational commitment that appropriate HR policies may be employed to create a work-force receptive to change (Stace and Dunphy, 1991). The various HR policies enable employees to reduce uncertainty and gain greater predictability and control in their work setting (Tushman and Romanelli, 1985). Thus, HR policies may induce the commitment of employees to the organization, where committed employees are assumed to be more productive and more adaptable to change (Guest, 1987). As a public hospital is the site for this study, a brief discussion of the changes affecting the hospital is provided. Background to change process
In recent years the formula by which Federal and State governments fund public hospitals in Australia has undergone several changes. Previously, each public hospital spent whatever was required to treat public patients and the government met the costs of the treatment (i.e., the bulk of the operational funds were provided by the government). If public hospitals treated patients with private health insurance then the insurance company paid those costs, although those patients with private insurance have been in the minority. Hence, the major source of
Employee acceptance of organizational change 125 income has been by public funding allocation. Nevertheless, as governments have found it increasingly more difficult to cover the mounting costs of funding hospitals, they have emphasized the need for hospitals to become more efficient and more productive. In recent years public hospitals negotiated with the government, signing the Health Services Agreement, which set specific performance targets to be met in return for an agreed amount of money. According to this arrangement, hospitals had specific amounts to spend, whereby the government would not cover any deficiencies if the hospitals had overspent. On 1 July 1993, the Victorian state government introduced casemix funding which aimed to promote a 'flexible hospital system, better positioned to meet the constant demands for change and improvement' (Victorian Government Department of Health and Community Services, 1993: iii). A casemix measure known as Diagnosis Related Groups (DRGs) allows the units of hospital output to be recorded, valued and charged to the state government by the hospital. T'he funding formula is based on the number and type of patients treated by the hospital. A price schedule determines each type of treatment given. For example, the DRG cost weight schedule is higher for a liver transplant compared to that for an arthroscopy. Moreover, hospitals are grouped together in terms of si2K and cost overheads, and receive the same amount of funding for providing the same type of service. If a hospital within a group spends more than others to dehver the same service, then it is deemed to be losing money and will be financially penalized by the government. The public hospital in the present study responded to government funding changes by undertaking an Operational Efficiency Review (OER). The executive director of the hospital outlined in meetings and newsletters to the staff that 'we have outdated buildings that need to be replaced, we need to reduce our lists of patients waiting for treatment, we have to keep up with improved medical technologies, and we need to meet increased costs of employing staff - covering Workcare and superannuation as well as salary increases.' The executive director emphasized that the OER focused on restructuring jobs, increasing the use of technology and eliminating activities that were redundant. A group of consultants were contracted to undertake the OER, which was carried out in two phases. Phase 1, the diagnostic stage, sought to identify the opportunities for improvement in the way the hospital carried out its business, and to estimate the benefits to the hospital if those improvements were implemented. This phase identified potential savings of approximately $8.5 million, which was 5 per cent of the hospital's annual operating budget. Apart from the savings factor, the consultants identified areas where patient care could be improved through streamlining systems or changing work practices. Other recommendations included the investment in new equipment, multi-skilling employees, merging and restructuring some departments, strategies for reducing the waiting time for outpatient and X-ray services, increasing the usage of operating theatres, reducing unnecessary bed days (e.g., pre-operative stays, discharge planning processes and assessment and arrangements for transfer to other care facilities) and reducing staff levels in some departments by using automated technology. A change committee was set up to review the OER data as it was collected so that staff were kept abreast of the recommendations. The committee was consulted every two weeks and had the opportunity to question the findings of the OER and to offer alternative information about aspects of the report. Phase 2 comprised the establishment of four task forces covering food services.
126 Roderick D. Iverson cleaning, operating theatres and anaesthesia, and radiology to investigate the savings opportunities identified in phase 1 of the OER. The role of each task force was to review, trial and monitor the recommendations made in phase 1. The task forces enabled management and staff to reach consensus over the means of achieving savings through increased efficiency. Progress reports from each task force were also made available to the change committee for consideration and comment, and to the steering committee, which comprised members of the hospital's board, senior management, medical consultants and union representatives, before they went to the full board of the hospital. All staff were kept informed in regard to the task forces' recommendations and findings through the OER newsletters and regular meetings. The following section develops the causal model based on the acceptance of the OER by employees in the hospital. Theoretical framework
The causal model as shown in Figure 1 integrates personal, job related, environmental and employee orientations which were selected from the economic, psychological and sociological perspectives from the commitment literature (see Iverson and Roy (1994) for more detail on the perspectives). The variables can be categorized as: (a) personal variables, comprising the individual characteristics of age, tenure, education, occupation, gender and union membership, as well as the predispositions of positive and negative affectivity which employees bring to organizations; (b) job-related variables (which Morris, Lydka and O'Creevy (1993) equate to HR policies), including autonomy, job security, promotional opportunity, distributive justice, role ambiguity, role conflict, role overload, supervisory support, co-worker support, job satisfaction and job motivation; (c) environmental variables, relating to the non-work setting, consisting of industrial relations (IR) climate, environmental opportunity and kinship responsibility; and (d) employee orientation, dealing with the affective responses such as organizational commitment, which is produced by the personal, job-related and environmental variables. The inclusion of" this variable provides a better explanation of the change process. The path diagrammatic representation of the model in Figure 1 contains positive (+) and negative (-) signs to indicate the hypothesized causal relationships between the variables. If high levels of a variable are hypothesized to lead to high levels of a second variable, a positive sign appears. Similarly, if high levels of a variable are hypothesized to lead to low levels of a second variable, a negative sign appears. For example, it is hypothesized that employees who experience a high level of job security will also experience a high level of organizational commitment. Definitions of the variables in the model are contained in Table 1. Employee orientation Although there is some disagreement over the definition of organizational commitment (e.g., Coopey and Hartley, 1991; Farrell and Rusbult, 1981; Guest,
Employee acceptance of organizational change 127 Personal variables Age Tenure Education Occupation Gender — Union membership Positive affectivity Negative affectivity Organizational commitment Job related variables Autonomy Job security Promotional opportunity Distributive justice Role ambiguity Role conflict Role overload Supervisory support — Co-worker support Job satisfaction Job motivation Organizational change Environmental variables IR climate Environmental opportunity Kinship responsibility - —
Figure 1 Hypothesized model of organizational change.
1992; Iverson and Roy, 1994; Kuruvilla and Iverson, 1993; Meyer and Allen, 1984; Morris et al, 1993; Salancik, 1977), researchers generally agree that the term 'commitment' can be employed to describe two distinct but related concepts, attitudinal and behavioural commitment. Attitudinal commitment, also referred to as affective organizational commitment, represents the degree of
128 Roderick D. Iverson Table 1 Definitions of variables Variable Personal variable^ Union membership Positive affectivity Negative affectivity
Definition Degree of affiliation with union. Extent to which an individual feels enthusiastic over time and across situations (Watson, Pennebaker and Folger, 1987). Extent to which an individual experiences aversive emotional states over time and across situations (Watson et al., 1987).
Job-related variables
Autonomy Job security Promotional opportunity Distributive justice Role ambiguity Role conflict Role overload Supervisory support Co-worker support Job satisfaction Job motivation
Degree to which an individual has influence over his/her job. Extent to which an organization provides stable employment for employees (Herzberg, 1968). Degree of movement between different status levels in an organization (Martin, 1979). Degree to which an organization treats employees fairly. Degree to which role exjjectations are unclear (based on Kahn et al, 1964). Degree to which employee role expectations are incompatible (Kahn et al., 1964). Extent to which the job performance required in a job is excessive. Degree of consideration expressed by the immediate supervisor for the subordinates (Michaels and Spector, 1982). Degree of consideration expressed by co-workers (Blau, 1960). Overall degree to which an individual likes his/her job (Price and Mueller, 1981). Degree to which the individual psychologically identifies with the job (Kanungo, 1982).
Environmental variables
IR climate
Degree of harmony between management and the unions (Dastmalchian, Blyton and Adamson, 1991). Environmental opportunity Availability of alternative jobs outside the organization. Kinship responsibility Degree of an individual's obligation to immediate relatives in the community (Iverson, 1992). Employee orientation Organizational commitment Degree of loyalty to the organization (Porter et al., 1974). Age, tenure education, occupation and gender have accepted definitions in the literature.
Employee acceptance of organizational change 129 loyalty an individual has for an organization. Specifically, it emphasizes an individual s identification and involvement in the organization (Porter et al, 1974). Conversely, behavioural commitment reflects the process by which individuals hnk themselves to an organization and focuses on the actions of the individuals. Kuruvilla and Iverson (1993) note that, although there are other forms of commitment such as continuance or calculative commitment (i.e., the costs associated with employees leaving the organization) and normative commitment (i.e., the right or moral thing to do (Weiner, 1982)), all forms of commitment reflect a bond between the individual and the organization. Guest (1987) recommends researchers use the attitudinal component of organizational commitment (Mowday, Porter and Steers, 1982) due to it being more applicable to HR policy formation than the more specific types of conmiitment. Accordingly this component (from this point on known as organizational commitment) is employed in this study. Guest (1987, 1992, 1995) and Coopey and Hartley (1991) further argue that HR policies have been designed to produce HRM outcomes, such as eliciting high employee commitment. This commitment then leads to organizational outcomes, such as successful organizational change. Guest's model asserts that commitment is best considered as an intervening variable between antecedents and behavioural consequences, where commitment will result in a 'willingness to accept change' (1987: 514). Therefore, it is expected that organizational commitment will mediate the effects of the job-related and environmental variables on organizational change. A discussion of the personal, job-related and environmental variables and their relationship with organizational commitment and acceptance of organizational change follows. Personal variables There is evidence linking both age and tenure to organizational change. Age, for example, has been found to have a negative impact on change (Cordery et al, 1991; Cordery et al, 1993; Ellis and Child, 1973). That is, younger employees are more likely to accept change than older employees. Younger employees tend not to be as 'set in their ways' as older employees (Cordery et al, 1991), and as such are less resistant to change. Similarly, Broadwell (1985) argues that the less time employees have spent within an organization, the more likely they are to accept change. This primarily stems from employees having relatively few 'preconceived notions' about how the organization should do things and is also due to these employees having less established routines than employees with longer tenure. Therefore, it is hypothesized that age and tenure will have direct negative impacts on the acceptance of change. Education is expected to have a positive impact on the acceptance of organizational change, as employees with higher education have increased opportunities for skill utilization (Cordery et al, 1993). This increased skill utilization enables employees to better meet the new challenges of their job. In relation to occupation, there is support in the literature for white-collar employees being more accepting of change than blue-collar employees. Sproul (1981) reported that attitudes towards change were related to hierarchical position in the organization. Specifically, employees at higher levels in the hierarchy expressed greater belief in the change process when alerted to the problems facing the organization.
130 Roderick D. Iverson Research on gender and its relationship with organizational change has had inconsistent findings (Cordery et al, 1991, 1993). Although Cordery et al. (1993) found no relationship, an earlier study by Cordery et al. (1991) reported that men were resistant to change when they perceived they were gaining skills that were traditionally female (e.g., typing). It is therefore hypothesized that women would be more accepting of change than men. Union membership is hypothesized to have a negative impact on both organizational commitment and the acceptance of organizational change. This derives in part from the research surrounding the 'exit-voice' (Freeman and Medoff, 1984; Hirschman, 1970) model, where unionization is associated with 'lower autonomy, task complexity, task significance, and feedback' (Barling, FuUagar and Kelloway, 1992: 185), which is expected to reduce organizational commitment. That is, lower loyalty to the organization is attributable to the union making members more aware of their working conditions and also providing an outlet for members to voice their attitudes. In addition, union members are argued to be resistant to change due to management potentially reacting to the 'increased costs associated with unionisation by replacing labor with capital' (Barling et al, 1992: 175), and by the emphasis on increased efficiency in work practices leading to the perceived deterioration of wages and conditions of employees. However, it is also hypothesized that the impact of union membership on both organizational commitment and change would differ by the degree of harmony in the IR climate. That is, union membership would have a positive effect on commitment and change when the IR climate is harmonious (Deery, Iverson and Erwin, 1994). The individual traits of positive and negative affectivity respectively describe the dispositions of individuals to perceive events to be generally positive and to be enthusiastic (George, 1989; Judge, 1993) or negative (Watson and Clarke, 1984) over time. Such dispositional differences will affect the perception of organizational change. Research indicates that individuals with positive affectivity are more able to control their environment (George, 1989; Judge, 1993), and as such would embrace change more readily. In contrast, individuals with negative affectivity are more likely to judge change as a stressful and demanding event because of their lack of coping strategies in dealing with change (Parkes, 1990). Therefore, it is proposed that positive and negative affectivity will have direct impacts on the acceptance of organizational change. Job-related variables Autonomy refers to the control employees have over their job. It has long been argued that, with increased autonomy, employees are able to buffer job stress and burnout (Miller et al, 1990; Ray and Miller, 1991). As change is generally associated with uncertainty and stress, it is expected that autonomy will predict employees' acceptance of change. Mathieu and Zajac (1990), in a meta-analysis, also found autonomy to be positively associated with organizational commitment. That is, the greater the autonomy, the greater the loyalty to the organization. Hence, it is posited that autonomy will have a positive impact on the attitudes to change via organizational commitment. The importance oi job security in the organizational change literature is somewhat understated, with a small but consistent body of research finding a positive
Employee acceptance of organizational change 131 relationship (Cordery et al, 1991, 1993). The impact of job security on organizational commitment is thought to be influenced by labour market conditions (Morris et al, 1993). That is, job security would have a strong positive relationship with an employee's loyalty to the organization when labour market conditions are contracted. This has been empirically supported by Armknecht and Early (1972) who found that an employee's level of job security is contingent on the behaviour of the labour market. As the present study was undertaken during the recessionary economic conditions of the early 1990s, it was posited that job security would have a positive impact on organizational commitment, which in tum, would impact on the acceptance of organizational change. Promotional opportunity has been found to be an important determinant of organizational commitment (Iverson and Roy, 1994; Mathieu and Zajac, 1990; Morris et al, 1993; Price and Mueller, 1986). That is, the greater the opportunities for career advancement the greater the loyalty to the organization. As employees move up the organizational hierarchy they are able to realize goals such as increased responsibility, status and money rewards (Morris et al, 1993). Realizing goals such as these serves to commit the employee to the organization. Moreover, employees with greater career prospects tend to display greater skill utilization and as such are more receptive to change (Sproul, 1981). Distributive Justice focuses on the fairness to which employees are treated in terms of their effort, responsibility and education and training. It is expected that employees with high distributive justice will display high commitment to the organization (Iverson and Roy, 1994; Price and Mueller, 1986). Cordery et al. (1993) reported that employees who exhibited positive attitudes to change were higher on distributive justice than those employees who exhibited negative attitudes to change. Current research considers job stress to be composed of at least three basic factors: role ambiguity, role conflict and role overload (Iverson, Deery and Erwin, 1994, 1995). Working conditions are perceived as stressful when the demands of the job exceed the abilities of the individual. Job stress has been found to be an aetiological cause of physiological, psychological and behavioural problems. Many studies document the fact that hospital employees work in a stressful environment (Iverson, Deery and Erwin, 1994, 1995; Price and Mueller, 1986). Role ambiguity relates to task-related information and is usually found under labels such as 'feedback' or 'instrumental information'. An example of role ambiguity would be the superordinate member of a hospital, such as a charge nurse, not conveying to a sister what is to be done, how the patient is supposed to be treated and any other information that is task related. The second component of job stress, role confiict, refers to the dimensions of congruency-incongruency or the compatibility-incompatibility in the requirements of the job role. In contrast to role ambiguity, role conflict is not concemed with the clarity of an employee's role, but with the inconsistent demands of it. Role overload relates to undue performance required on a job. The greater the amount of job performance required, the greater the workload. There is consistent support (Iverson, Deery and Erwin, 1994, 1995; Mathieu and Zajac, 1990; Price and Mueller, 1986) for the negative impact of role ambiguity, role conflict and role overload on organizational commitment. Social support is the degree of consideration individuals receive from members of their social network. Social support is important as it is thought to buffer job
132 Roderick D. Iverson stress (House, 1981). There are two types of social support: supervisory and coworker support. Supervisory support is characterized by the human relations ability of superiors, and is displayed in terms of trust, respect, friendship and a deep concem for subordinates' needs (Iverson, 1992). In relation to co-worker support, it generally refers to quasi-primary relationships (Price, 1977), such as having close friends at work. Implicit in co-worker support is the notion of work group cohesion, whereby co-workers listen to job-related problems, are helpful in getting the job done and can be relied upon when things get difficult on the job. There is an abundance of literature linking supervisory and co-worker support to organizational commitment (Iverson, Deery and Erwin, 1994, 1995; Mathieu and Zajac, 1990; Price and Mueller, 1986). There is a growing body of literature supporting the positive relationship of job satisfaction and job motivation to employees' attitudes to change (Coopey and Hartley, 1991; Cordery et al, 1993; Gardner et al, 1987; Guest, 1987). In relation to job satisfaction, Cordery et al (1993) reported that low levels of extrinsic satisfaction were associated with unfavourable attitudes to change, while Gardner et al. (1987) observed significant interactions between satisfaction and motivation and job change. Moreover, job satisfaction and job motivation have been found to be strongly related to organizational commitment (Iverson and Roy, 1994; Mathieu and Zajac, 1990). Therefore, it is hypothesized that job satisfaction and job motivation will have indirect effects on change via organizational commitment. Environmental variables IR climate refers to the degree of harmony in the union-management relationship. This is reflected in the faimess of the interaction between management and unions and the degree of co-operation of both parties in solving common problems (Dastmalchian et al, 1991). Other writers such as Lorenz (1992) emphasize the role of tmst between unions and management, while Zeffane (1994) argues that IR climate may be best represented as a particular feature of management style. Much attention has focused on the relationship between a positive and cooperative IR climate and unilateral (organizational or union) and dual (organizational and union) commitment (e.g.. Angle and Perry, 1986; Beauvais, Scholl and Cooper, 1991; Deery et al, 1994; Gaertner and NoUen, 1989). In the case of unilateral organizational commitment, Deery et al (1994) found that a harmonious IR climate was positively and significantly related to organizational commitment. That is, the more positive the IR climate, the more likely employees were to be loyal to the organization. This is consistent with Zeffane (1994) who reported in a study of public-sector employees that participative and flexible management style was the most significant predictor of organizational commitment. It is also hypothesized that IR climate will have a direct and positive impact on the acceptance of organizational change. As a harmonious IR climate conveys co-operation, trust and participation in the solving of problems, it is reasonable to expect that the challenges facing the organization in the present study (e.g., achieving savings through increased efficiency) would be shared between management and unions. Hence, a congruence in the acceptance of change for the benefit of the organization would be predicted. Environmental opportunity refers to the availability of jobs that are extemal to
Employee acceptance of organizational change 133 the organization. Environmental opportunity has been found to have a negative impact on organizational commitment (Mueller et al, 1994; Price and Mueller, 1986). That is, when jobs are plentiful in the labour market, employees' loyalty to their organization is decreased. This may reflect a re-evaluation of current job attitudes of employees when there are altemative job opportunities (Steers and Mowday, 1981). Moreover, it is hypothesized that environmental opportunity will also have a negative effect on organizational change. Mowday et al (1982) assert that in times of high unemployment, the restricted availability of altemative jobs will influence employee behaviour for fear of losing one's job. Hence, when environmental opportunity is low, it would be anticipated that employees would be more accepting of organizational change. Marital status, sex, number of children or dependants, age of children, number of relatives in a community and/or commitments such as buying a home have been cited as indices of kinship responsibility by researchers (Blegen, Mueller and Price, 1988; Gupta and Beehr, 1979; Mobley, 1982; Price and Mueller, 1981). In addition to these measures, both Muchinsky and Morrow (1980) and Muchinsky and Tuttle (1979) suggest that an important consideration in kinship responsibility is whether an employee is the primary or secondary wage eamer. This study focuses on the number of dependants, marital status (i.e., single or married) and income eamer status (i.e., primary or secondary) as measuring kinship responsibility. Hence, an employee who is married with dependants, and who is also the primary income eamer, would be an example of high kinship responsibility. The relationship of kinship responsibility with organizational change is explained by the dependency of individuals on organizations. That is, a married person with dependants, who is also the primary income eamer, for example, relies on the organization as a means of fulfilling important kinship obligations (Iverson, 1992). Therefore, it is hypothesized that kinship responsibility will have a direct negative impact on organizational change due to the uncertainty that is created (Fox and Krausz, 1987). Methods Research setting and sample The site for this research was a large public hospital in the state of Victoria, Australia. The hospital employs approximately 2000 staff. The sample consisted of 761 (74 per cent female and 26 per cent male) staff of whom 10 per cent were from blue-collar occupations and 90 per cent were from white-collar occupations, where 65 per cent were union members. The average age, tenure and education of employees were 33.71 years (s.d. = 9.89), 6.14 years (s.d. = 5.17) and 13.56 years (s.d. = 2.30), respectively. Data collection A multiple-item survey was administered to a random sample of 1100 employees at the hospital. The response rate was 74 per cent. Following the listwise deletion of missing data, a total of 761 questionnaires were retained. Chisquare analysis was undertaken to evaluate the representativeness of the sample. Through data obtained from the personnel records and questionnaire data, no significant differences were found in the demographic characteristics for age
134 Roderick D. Iverson (^\'i) - 5.56, p > 0.05), tenure (x^(3) = 1.63, p > 0.05) and occupation (x^(l) = 0.21, p > 0.05). Measurement A 5-point Likert-type scale format was used to measure employees' perception to each item, ranging from strongly disagree (1) to strongly agree (5). Established scales were used where possible. The reliability of the multiple measures was computed by estimating Cronbach's Alpha (1951). The reliability, descriptive statistics and correlations (LISREL) among measures are contained in Table 2. Endogenous variables Employee orientations Organizational commitment was operationalized by the 9-item short form of Porter et al. (1974). The dependent variable of organizational change was developed by the researcher. The scale measured respondents' attitudes towards change based on the OER at the hospital and comprised three questions relating to the impact of budget cuts on the tendering and closure of some services and whether employees perceived the hospital as a better place to work since the OER. Exogenous variables Personal variables Occupation was coded as 1 if the respondent was whitecollar and 0 if they were blue-collar. Gender was coded as 1 if the respondent was female and 0 if they were male. Union membership was coded as 1 if the respondent was a union member and 0 if they were not. Positive and negative affectivity were operationalized by an adaptation of the Multidimensional Personality Index obtained from David Watson (see Agho, Price and Mueller, 1992, 1993). Job-related variables Autonomy was measured using Tetrick and LaRocco's (1987) scale. The variable of job security was assessed by a modification of the scale by Oldham et al (1986). Promotional opportunity and distributive justice were measured using by Price and Mueller's (1981, 1986) scale. The scale of stress measured the psychological symptoms of stress by focusing on the components of role ambiguity, role conflict and role overload from the scales of Kahn et al (1964), Rizzo, House and Litzman (1970), and Caplan et al (1975), respectively. Social support evaluated the consideration expressed by the supervisors and co-workers using a modification of House's (1981) measure. Job satisfaction was measured using a shortened version of the established scale of Brayfield and Rothe (1951). Job motivation was operationalized using Kanungo's (1982) scale. Environmental variables IR climate was assessed using the 10-item harmony component from Dastmalchian, Blyton and Adamson (1989). Environmental opportunity was operationalized using Price and Mueller's (1981, 1986) scale. Kinship responsibility was measured using a three-item composite index based on Blegen, Mueller and Price (1988). The first item, relating to the number of dependants (e.g., spouse, children, relatives) living with the respondent, was
Employee acceptance of organizational change
* I
if
O^ fN O
"1 ^
' r ' so CXJ
*r^ n on 'Ow^^ -^Os
Is T3 a
1 m f^ m m vo t^
as
•&>..§>
:i
il^l,
ll I
2 5. aa a g ROT
i sI
• •§ CO O >^ so r - oo CN
^O O
.5-5-
i o •
:w i
ig S fs S ^ S S
al
II 0^ u
135
136 Roderick D. Iverson coded as 1 if the respondent had one dependant living with them, 2 if more than one dependant lived with them, and 0 if no dependants lived with them. The second item, relating to whether respondent was a single parent, was coded as 1 if they were a single parent and 0 if they were not. The third item, relating to whether the respondent was the main income eamer, was coded as 1 if the respondent was the main income eamer and 0 if they were not. The possible range of the index was between 0 and 4. Analysis The statistical techniques of multiple regression. Chow test and Linear Structural Relations (LISREL) were utilized in the analysis. Multiple regression was employed to assess and support the assumptions of linearity, additivity, model specification, multicoUinearity and homoscedasticity (Berry and Feldman, 1985). The statistical technique known as the Chow test (Gujarati, 1978) was employed to test whether sample populations from the blue- and white-collar occupations could be combined to form an aggregate sample. The Chow test involved testing for differences between the regression coefficients of the two subgroups (occupations). No consistent pattem of differences was found, confirming that aggregation of the two subgroups was appropriate. The statistical technique of LISREL was used to estimate the causal model. LISREL VIII produces a stmctural equation model and a measurement model (Joreskog and S6rbom, 1993). The structural equation model specifies the hypothesized causal relationships among the concepts (latent constructs), whereas the measurement model specifies a confirmatory factor analysis of hypothesized relationships between the latent (unobserved) constmcts and the manifest (observed) variables. LISREL adds precision to the estimation of the causal model by correcting for attenuation in random measurement error of manifest variables and, as maximum likelihood method is employed, LISREL produces both a statistical measure of goodness-of-fit and explained variance (Rsquare) of the model. Moreover, as the coefficients can be interpreted as standardized regression coefficients, a path analysis (decomposed into direct, indirect and total effects) can also be undertaken (Alwin and Hauser, 1975). Applying the procedures as recommended by Bagozzi and Yi (1988), the measurement model was tested for convergent validity (i.e., the degree of association between measures of a construct) and discriminant validity (i.e., the degree to which measures of constructs are distinct). In relation to the convergent validity, the null model, followed by the one-factor and hypothesized model were estimated. The convergent validity of the model was affirmed, where the hypothesized model was found to fit the data significantly better than both the null (Ax^ (341) = 21304.85, p < .001) and one-factor (Ax^ (269) = 13727.51, p < .001) models. The discriminant validity of the model (tested by calculating the difference between one model which allowed the correlations between the constructs to be constrained to unity, i.e., perfectly correlated, and another model which allowed the correlations between the constructs to be free) was also supported. The measurement model was found to have a normed comparative fit index (CFI) (Bentler, 1990) of .90. As the CFI (which avoids the underestimation of fit) indicated an accepted fit of the measurement model, it was appropriate to estimate the stmctural model. The findings of the stmctural model are
Employee acceptance of organizational change 137 contained in the results section. However, before proceeding to the results, it is pertinent to discuss the problem of common method variance. When using self-report material, such as regressing attitudes on attitudes, a problem of common method variance (Campbell and Fiske, 1959; Fiske, 1982) may occur. As common variance tends to inflate the relationship between the constructs (Podsakoff and Organ, 1986), an attempt to minimize the impact was undertaken by having positively and negatively worded items in the measures (Gordon and Ladd, 1990). Another way to overcome this problem is to regress objective data, such as change data collected from the organization's records on respondents' attitudes. However, when objective data are not available (as was the case in the present study), it is necessary to conduct further analysis, such as Harman's (1967) one-factor test. In this test, all variables, including exogenous and endogenous are entered into the factor analysis. The results of the unrotated factor solution are then examined to determine whether '(a) a single factor will emerge from the factor analysis, or (b) one "general" factor will account for the majority of the covariance' (Podsakoff and Organ, 1986: 536). The unrotated factor solutions using scree tests (Cattell, 1966; Cattell and Vogelmann, 1977) examined whether a single factor emerged. Seven meaningful factors were found, explaining 38.7 per cent of the variance. Following a similar procedure to Miceli et al (1991), the factor solution was rotated. Of the nineteen item loadings for the first factor, four were greater than or equal to .40 (displaying convergent validity [Kim and Mueller, 1978]), while fifteen were less than .40. These results indicate that a general factor did not originate, limiting the probability of common method variance problems. Results The following section presents two sets of results: first, the original hypothesized model and, second, the revised causal model. Hypothesized causal model Table 3 presents the LISREL VIII results used to estimate the hypothesized model, which is shown in Figure 1. Six variables as predicted by the model were found to have statistically significant net impacts on organizational commitment (see column 1 of Table 3). The personal variable of union membership, the jobrelated variables of job security, job satisfaction and job motivation, and the environmental variables of IR climate and environmental opportunity. Organizational commitment is increased when employees are not union members, have stable employment, like their jobs, psychologically identify with their job, where there is a harmonious IR climate and where there are few alternative jobs in the labour market. In testing for the moderating effect of IR climate, no interaction was found for union membership on organizational commitment. It can be seen from the results that 54 per cent of the variance in organizational commitment was explained by the variables in the model. Organizational change, the dependent variable, had 43 per cent of variance explained by the personal variables of tenure, education and union membership, the environmental variable of IR climate, and the employee orientation of organizational commitment (see column 3 of Table 3). As predicted by the model, the acceptance of organizational change is increased when employees have shorter
138
Roderick D. Iverson
Table 3 LISREL {standardized coefficients) for organizational change: hypothesized and revised causal models^
Determinants Personal variables Age Tenure Education Occupation Gender Union membershipf Positive affectivity Negative affectivity Job-related variables Autonomy Job security Promotional opportunity Distributive justice Role ambiguity Role conflict Role overload Supervisory support Co-worker support Job satisfaction Job motivation Environmental variables IR climatef Environmental opportunity Kinship responsibility Employee orientation Organizational commitment R^
Organizational commitment (2) (1)
-.11* —
-.10* .09*
-.04 .11* -.04 .05 -.04 .03 -.01 .03 -.03 .29* .28*
-.05 .09* -.04 .06 -.03 .03 -.01 .02 -.04 .26* .29*
.30* -.13*
.31* -.13*
.54
.55
Organizational change (3) W .03 -.10* .14* .03 .04 -.36* -.03 .01
.01 -.11* .13* .01 .03 -.36* -.03 .03
—
-.13*
.20* -.05 -.05
.19* .02 .03
.35* .43
.35* .47
» N = 761; the results of the hypothesized and revised models are contained in the odd and even numbered columns, respectively. t Interaction tests examining the relationship between union membership and IR climate with organizational commitment and organizational change were not significant. * p < .05, one tailed test.
tenure, have higher education, are not union members, where there is a harmonious IR climate and when employees are loyal to the hospital. Similar to organizational conmiitment, no interaction effect was found for IR climate and union membership on organizational change. Revised causal model
Table 3 also presents the LISREL VIII results used to estimate the revised causal model. The modification procedure used to fit the LISREL model is consistent
Employee acceptance of organizational change 139 with the technique employed by previous research (Iverson and Roy, 1994; Mueller et al., 1994). The procedure consisted of two steps. First, all originally hypothesized paths (see Figure 1) were retained, even if they were non-significant after analysis. Second, the hypothesized model was expanded by estimating additional causal paths that were indicated by the LISREL VIII modification indices and considered to be theoretically plausible. The goodness-of-fit index for the revised model (x^ (2232) = 4426.07, p < .001) compared to the hypothesized model (x^ (2234) = 4436.14, p
