Homogeneity of Personality in Occupations and Organizations: A Comparison of Alternative Statistical Tests

Purpose The attraction-selection-attrition (ASA) model has served as the foundation for numerous investigations. However, the generally supportive evidence for ASA's homogeneity hypothesis has often been based on statistical tests (e.g., MANOVA) that rely on between-group differences to evaluate within-group agreement. The primary purpose of this article was to discuss advantages of direct statistical tests of homogeneity—average deviation (AD) and r wg —when testing ASA's homogeneity hypothesis, and advantages of other statistical tests for testing other aspects of ASA theory. A secondary goal was to evaluate the extent to which occupational homogeneity is distinct from organizational homogeneity. Design/Methodology/Approach Data were obtained from the Center for Creative Leadership (CCL) and included scores on the Myers-Briggs Type Indicator ® (MBTI) personality measure for 1,103 managers from 25 organizations and 17 occupations. Findings Results were generally supportive of the homogeneity hypothesis. AD values showed that most groups were homogeneous on most assessed personality dimensions. A comparison analysis using traditional statistical tests (i.e., MANOVA) indirectly suggested within-group homogeneity by revealing a significant between-groups effect. In addition, results suggested possible boundary conditions to ASA; notably, meaningful heterogeneity was observed for the S-N (sensing-intuition) MBTI ® dimension. Implications The current study provides direct support for ASA's homogeneity hypothesis for both organizations and occupations and offers guidance for future research on ASA theory and its possible boundary conditions. Originality/Value This is one of the first studies to test the predictions of ASA in both organizations and occupations using a direct index of agreement.