In a review of two areas of turnover researchuindividual motivated choice behavior and organizational consequences--five alternative turnover measures are defined: reasons, voluntariness, avoidability, functionality, and utility. Turnover data for one year (1987) were gathered from 325 former employees, 568 supervisors, 418 replacement employees, and the personnel files of a university. Analyses indicated that organizational records are deficient as a source of information, especially because of the usual practice of recording a single reason for turnover. Voluntariness may result in a classification system that is too gross for validating motivational models. Avoidability, functionality, and utility each measure unique aspects of organizational consequences, but each has limitations. Turnover measures should be viewed as continua rather than as dichotomies. Recommendations for future research are provided.