Abstract

Empirical and intuitive approaches to test construction were compared during an attempt to develop a scale that would discriminate between successful and unsuccessful athletes. Eight scales were developed using items from the Minnesota Multiphasic Personality Inventory item pool. Two scales were constructed using an empirical approach, five scales using an intuitive strategy, and one scale based on a random selection of MMPI items. Results of this investigation indicate that reliable scales were developed using the empirical and intuitive test construction methods, yet all scales were equally ineffective as valid measures of athletic potential. Post hoc analyses indicated that while overall classification rates were not significant, classification of successful athletes was made with greater precision than classification of unsuccessful athletes. Findings are interpreted in terms of characteristics inherent in the MMPI item pool and the domains assessed by the MMPI. Recommendations are made for future scale development in sport psychology. Verf.-Referat