Abstract
Burnout and exhaustion has been extensively studied in organizational, work, and health psychology. Studies using the cross-lagged panel models have tended to conclude, explicitly or implicitly (e.g., in the form of policy recommendations), causal prospective effects of, for example, organizational demands, job insecurity, and depression on burnout and exhaustion. However, it is well established that effects in the cross-lagged panel model may be artifactual, e.g., due to correlations with residuals and regression to the mean. Here, we scrutinized 23 previously reported prospective effects on burnout/exhaustion by fitting complementary models to data that were simulated to resemble data in the evaluated studies. With one possible exception, the previously reported prospective effects did not withstand scrutiny, i.e., they appeared to be artifactual. It is important for researchers to bear in mind that correlations, including effects in cross-lagged panel models, do not prove causality in order not to overinterpret findings. We recommend researchers to scrutinize findings from cross-lagged panel models by fitting complementary models to their data. If findings from complementary models converge, conclusions are corroborated. If, on the other hand, findings diverge, caution is advised and claims of causality, explicit or implicit, should probably be avoided.