A Meta-Analytic Test of Multiplicative and Additive Models of Job Demands, Resources, and Stress

Multiplicative and additive theoretical models have been proposed to explain how job demands and job resources (e.g., job control, social support) relate to strain. However, there has been mixed support for the multiplicative model, and there are questions about the generalizability of both models to strains varying in severity and type, and to different types of demands. Thus, we conducted a meta-analysis of regression coefficients testing the multiplicative and additive models from 77 unique samples and over 141,505 participants, using raw data or correlation matrices supplied by study authors. Overall, we found weak support for the multiplicative model and some support for the additive model. We found that job demands and job control/social support were all weakly to moderately related to employee strain. However, neither the job demands-control nor job demands-social support interactions were meaningfully related to strain in almost all cases. The components of the additive model were more strongly related to primary and tertiary strain indicators than secondary indicators and were more strongly related to psychological strain than to physical or behavioral strain. Both the additive and multiplicative models were more strongly related to strain when demands are conceptualized as hindrances rather than challenges. Meta-analyses of the simple slopes showed that job control and social support generally buffered the effects of job demands on strain, but the effects were weak and not significantly different across different levels of control and support. Theoretical and practical implications, as well as future research directions, are discussed.

Automated summary

The multiplicative model has weak support, while the additive model has some support. Job demands and job control/social support are weakly to moderately related to employee strain, but the interactions between demands-control and demands-social support are almost unrelated.