Working without a net: Time, performance, and turnover under maximally contingent rewards.

Researchers examined the performance-turnover relationship under conditions of varying reward contingency with 189 sales representatives who earned base pay plus commission (moderately contingent rewards) during their first 2 months of work and commission-only pay (maximally contingent rewards) thereafter. As predicted by an expanded version of the contingent rewards hypothesis, the performance-turnover relationship was stronger under maximally contingent rewards, higher than any reported in the literature. Using event history analyses, the researchers also showed that current (time-dependent) performance affords a better prediction of turnover than average (time-stationary) performance. Performance velocity (slope overtime) had a further, unique effect on turnover risk. The implications of these findings for future studies of the employee turnover process and pay systems were discussed.