Karasek's activation hypothesis: A longitudinal test of within-person relationships

Karasek's job demand-control model has received considerable research attention. However, only a few studies have examined its activation hypothesis and two problems have emerged. First, the only consistent finding concerns the activating role of job control. Findings related to job demands and the interaction between control and demands are inconsistent. Second, a dynamic approach focusing on intrapersonal dynamics is lacking. This study proposes and tests a dynamic activation hypothesis using theory elaboration, combining construct specification (clarifying the nature of activation as fluctuations in self-efficacy), vertical contrasting (distinguishing between- and within-person effects), and structuring (proposing recursive relationships). Hypotheses were tested in three-wave data (1755 teachers), analyzed using random intercept cross-lagged panel analyses. The results differed at the between- and within-person level. A baseline level of autonomy appeared essential (between-level), but increases in autonomy over time do not predict increased self-efficacy (within-level). High baseline levels of workload are negatively related to self-efficacy (between-level), but modest increases in workload backed-up with added autonomy are related to increased self-efficacy (within-level). In turn, increases in self-efficacy stimulate workload. This hints at a feedback loop, which is balanced by a negative feedback process. Too strong increases in workload or enduringly high workload negatively affects self-efficacy.