Stress and well-being in the workplace: a longitudinal cross-lagged structural equation modelling investigation

The current research expands upon previous knowledge by further investigating the causal relationship between self-ratings of occupational stress, psychological well-being, personal control and work performance. The work has also been designed to address the methodological pitfalls and deficiencies apparent in longitudinal research by incorporating the methodological and statistical rigor of structural equation modeling (SEM). A review of the occupational health literature indicates a broad range of inconsistencies regarding the causal pathways between variables. Based upon these inconsistencies, the aims of this research are to address three main hypotheses: the relationship between stress and wellbeing (H1), the relationship between stress, control and well-being (H2) and the association between stress, work performance and well-being (H3). All three hypotheses measure variables across work, non-work and context-free domains. Three samples of data were incorporated within the study in order to cross-validate findings. SEM techniques were conducted to analyse data to examine the intricate one-way, reverse and reciprocal relationship between variables. In relation to H1, results support a best fitting reciprocal cross-lagged model where both sources of stress and psychological well-being simultaneously influence one another across work, non-work and context-free life domains. A good fitting reciprocal cross-lagged model revealing that sources of stress and control across domains simultaneously effect each other was also produced in regards to H2. In relation to H3, again findings support a best fitting reciprocal cross-lagged model where both sources of work/non-work stress and work performance simultaneously effect one another. Overall the results theoretically build upon previous work conducted by further emphasising the complex causal relationship between organizational health factors including one-way, reverse and reciprocal associations. This research suggests that stress and well-being models should not be designed in the future without considering the strong causal influence of factors outside of work. This study also addresses all seven longitudinal methodological and statistical recommendations put forward by Zapf, Dormann & Frese (1996). The practical implications of the current study and recommendations for future research are also put forward.