Abstract
This study proposes an two-step methodology for psychosocial risk assessment. In the first step for each psychosocial factor, latent classes are estimated as a function of the level of exposure. In the second step, exposure to the psychosocial factors is related to harm by studying the criterion validity of the estimated classes. To test this approach we first adaptation of the Short Inventory to Monitor Psychosocial Hazards – 5A (SAMU) to the Norwegian context. To this aim, a sample of 1,002 Norwegian employees from various sectors was drawn. The confirmatory factor analysis in Mplus 8.4 supported the factorial validity. With Latent Gold 6.0 we identified different levels of exposure to each psychosocial hazard. The criterion validity showed significant and substantive effect sizes, allowing us to distinguish between varying levels of risk for burnout and engagement. Compared to no high or very high exposure to psychosocial demands, a greater risk of burnout was observed. A lack of or very low job resources was associated with little or no job engagement. This mixture approach offers an alternative to traditional correlational and covariance methods used for risk assessment. This configural approach to risk assessment potentially aligns more effectively with the complementary view that contemporary occupational stress and health theories offer on why employees experience stress or demotivation. However, the study acknowledges certain limitations, including common method bias, reliance on self-reported data, and issues with generalizability. This highlights the need for future research to include invariance testing and longitudinal and registry-based studies.