Modelling employee withdrawal behaviour over time: A study of turnover using survival analysis

Survival analysis techniques were used to test a model of turnover with a sample of 244 staff nurses. Estimates of survival and hazard functions indicated that withdrawal was not uniform over time, but rather occurred in distinct waves. Formal hypotheses were tested with a regression analogue of survival analysis, proportional hazards regression, and provided little support for a reasoned action model of turnover. Specifically, job satisfaction emerged as predictive of turnover while job search behaviour did not. Results from OLS and logistic regressions were consistent with prior research findings in that job search behaviour was a strong influence of employee turnover. Implications of these findings and directions for future research are discussed.