A Multilevel Analysis of the Occupational Stressor-Strain Process: Contextual Variables as Additional Moderators in the Job Demand-Control Model

Abstract Poor working conditions and other environmental stressors have been linked to detrimental effects on employee well-being (Ganster & Schaubroeck, 1991; Jex & Crossley, 2005; Siegrist, 2008; Segerstrom & Miller, 2004). A major objective of research in this area is to identify those factors that mitigate or lessen the negative impact of work stressors. Karasek's (1979) Job Demand-Control Model (JD-CM) endeavours to explain the stressor-strain process from a job design perspective, whereby job control is proposed to mitigate the negative effects of job demands on strain. However, the evidence in regards to the interactive effects of job demand and job control remain mixed, and it has been hypothesised that there may be additional moderating variables involved in the stressor-strain process that better explain how strain may be reduced. Although a number of additional individual-level moderators have been examined, such as self-efficacy, and social support, the role of broader contextual team and organisational factors, and the transaction between the individual and the environment have been largely overlooked. It is, therefore, important to investigate cross-level effects between individual-and group-level factors to gain a better understanding of the stressor-strain process. One such potential contextual factor is collective efficacy (Bandura, 1997) – the shared beliefs of a group that they have the resources and skills to manage a problem. It is hypothesised that group-level collective efficacy buffers the negative effects of stressors by providing the group with the confidence that they can cope in a productive manner, by using personal control to manage stressors. A second potential group-level moderator is group identification (Tafjel & Turner, 1979). Groups high in identification are characterised by shared values, goals, attitudes, and mutual respect. It has been argued that a strong sense of group identification buffers individuals and workgroups from the negative effects of stressors, by reducing uncertainty and providing a socially supportive environment that allows individuals to feel confident exercising personal control to deal with stressors. The objective of this thesis is to examine the role of these group-level factors as additional moderators of the JD-CM. The individual-, group-, and cross-level interactive effects of job design and contextual variables on strain will be investigated to develop a better understanding of how the broader social context in which one works impacts on individual factors and how strain is experienced at both the individual- and group-levels. The first study employed a cross-sectional survey design, whereas Study 2 was a longitudinal field study. A broad overview of each study is presented below. In Study 1, data was collected from 544 employees nested within 23 workgroups. Perceptions of job demands, job control, collective efficacy, workgroup identification, job satisfaction, and strain were measured. Variance components analyses found that significant amounts of variance in both anxiety and job satisfaction were shared collectively at the group-level. The major objective for this study was to evaluate whether collective efficacy and identification acted as group-level moderators of the individual-level JD-CM variables. Significant 3-way interactions were found among group-level collective efficacy, individual-level job control, and individual-level job demand on anxiety and job satisfaction, as well as among group-level identification, individual-level job control, and individual-level job demand on anxiety and job satisfaction. These interactions revealed that when collective efficacy was high, individuals perceiving high levels of job control were buffered from the negative effects of high job demands. Moreover, when group-level collective efficacy was low, high job control exacerbated the negative effects of job demands on both anxiety and job satisfaction. When group identification was high, the positive effects of high job control on job satisfaction were more marked. Conversely, when identification was low, high levels of perceived individual control were related to higher strain. These findings suggest that contextual factors may influence whether job control is perceived as a buffering or exacerbating factor. Study 2 examined the effects of these two contextual variables in the JD-CM over time. Using a sample of 120 participants nested within 8 workgroups, two surveys were conducted in a partial two-wave panel design. The results revealed significant 3-way interactions among group-level collective efficacy, individual-level job control, and individual-level job demands for depression, but not for anxiety, stress, or job satisfaction. The 3-way interaction among group-level workgroup identification, individual-level job control, and individual-level job demands was significant in the prediction of job satisfaction only. Both interactions followed similar patterns as those found in Study 1, providing additional support, albeit less consistently, for the notion that contextual factors need to be in place in order for job control to operate as a stress-buffer. Overall, the findings from both studies have several important implications for the JD-CM and for stress research in general. First, the findings suggest that aspects of strain are shared experiences among employees nested within workgroups, and this provides support for the notion that individuals are not independent of their contexts. Additionally, the findings from both studies suggest that the expansion of the JD-CM is warranted and that group-level contextual factors may act in concert with individual-level variables in buffering the effects of stressors in the work environment