Subjects, 432 convenience store cashiers, were divided into long-tenure and short-tenure groups. Chi-square analysis of application blank information for a weighting sample drawn from both groups revealed two items which significantly (p < .05) differentiated between the long tenure and short-tenure groups: number of previous jobs and full-time/part-time preference. Response weights were computed for these two items and used to calculate composite scores for the remaining holdout sample. A significant reduction in turnover would have occurred at the highest composite score level, if used as a hiring cut off. Results were tempered by several considerations, including a high percentage of false negatives and an insignificant linear relationship between composite scores and tenure.

In recent years, organizations have invested increasing financial and labor-related resources on employee training. The assumption is that training will benefit the organization through improved performance which will result in greater efficiency, greater customer satisfaction and, ultimately, increased revenue and profits. Further, employees are assumed to benefit because their improved performance should lead to career advancement and increased compensation. However, measuring the effect of training on employee performance has been problematic due to the difficulty of isolating the effect of training from other human resource management practices and environmental and organizational influences. Hierarchical multiple regression analyses were used to test a model for predicting merit pay increase, job promotion and performance ratings from measures of general and finance training, as well as employee tenure, gender, educational level and organizational level. It was found that while significant contributions (i.e., betas) were made by finance and general training for performance ratings, promotion and merit pay increase, they did not increase the variance accounted for by tenure, organizational level and gender.