Event History Turnover Analysis

Event history turnover analysis models not just whether employees leave but when they leave, treating tenure as a duration and the act of quitting as an event whose timing carries information. Paul Allison's 1984 monograph brought event history methods — survival and hazard models — into the social sciences with a regression-oriented treatment that handles the censoring inherent in longitudinal data. Morita, Lee, and Mowday's 1993 Academy of Management Journal paper applied these techniques to turnover research, showing organizational scholars how to model the hazard of leaving and why time-to-event methods are superior to simple stayed-versus-left comparisons. The core object is the hazard function, the instantaneous risk of quitting given that one has stayed so far, which can depend on tenure and on employee and job characteristics. Because some employees are still present when the study ends, the analysis must correctly handle censored observations rather than discarding or mis-coding them. The result is a model that explains and predicts the timing of turnover.