This article examines how the literatures of dynamic performance and the performance–turnover relationship inform each other. The nonrandom performance–turnover relationship suggests that dynamic performance studies may be biased by their elimination of participants who do not remain for the entire study period. The authors demonstrated that the performance slopes of those who leave an organization differ from the performance slopes of those who remain. This finding suggests that studies of the performance–turnover relationship need to consider employee performance trends when predicting turnover. Replicating and extending the research of D. A. Harrison, M. Virick, and S. William (1996), the authors found that performance changes from the previous month and performance trends measured over a longer time period explained variance in voluntary turnover beyond current performance. Finally, the authors showed that performance trends interacted with current performance in the prediction of voluntary turnover.