Mapping the Reliability Landscape of the Utrecht Work Engagement Scale

Abstract: The Utrecht Work Engagement Scale (UWES) underpins much organizational psychology research, yet cumulative evidence on the reliability of UWES scores across contexts has not been synthesized using reliability generalization methods. We conducted a reliability generalization meta-analysis of internal consistency estimates for the UWES, identifying 161 independent samples. For the UWES total score, 132 samples provided usable Cronbach’s alpha estimates. Random-effects models (REML) using Fisher’s Z transformation yielded a high pooled alpha (α = .912, 95% CI [0.903, 0.920]) but substantial heterogeneity (I 2 = 98.85%). The 95% prediction interval [0.744, 0.972] indicates that reliability is typically strong but may be meaningfully lower in some settings. Moderator analyses showed that sample context (student vs. worker) did not explain variability, and risk of bias screening suggested uniformly low concern across included studies. These findings support the UWES as a generally reliable measure while underscoring the importance of reporting reliability in each new sample.