Revisit the causal inference between organizational commitment and job satisfaction: A meta-analysis disentangling its sources of inconsistencies.

Existing theories and studies have exclusively focused on the direct temporal ordering of organizational commitment (OC) and job satisfaction (JS). However, their ordering varies highly across empirical studies. It is unclear whether such high variation is caused by statistical artifacts (i.e., spurious variation) or substantive moderators (i.e., true variation). Therefore, to disentangle artificial and substantial sources of this variation, we adopted traditional and full information meta-analytic structural equation modeling based on 71 independent samples (N = 16,698) with panel designs. After correcting for statistical artifacts, there was moderate heterogeneity in the lagged effects of OC and JS, suggesting that their ordering was not direct or simple but moderated by other variables. Further, the conceptualizations and/or measurements of OC and JS, time lag, and the timing of the investigation (e.g., newcomers vs. nonnewcomers) moderated their ordering. Finally, different specifications of the ordering between OC and JS resulted in differential cross-lagged relationships among JS, OC, and turnover intentions (k = 31, N = 6,876). Our meta-analysis provides new insights into the conceptualizations and/or measurements of OC and JS, adds some theoretical clarification to the dissenting theories by incorporating the time element, nuanced differences in the OC conceptualizations and the timing of the investigation, and raises concern over theory and research informed by studies overlooking alternative orderings of JS and OC. Looking forward, we suggest several promising future directions for explaining the heterogeneous cross-lagged relationships between OC and JS. (PsycInfo Database Record (c) 2023 APA, all rights reserved).