Using Hierarchical Classes to Analyze Organization of the Self-Concept

Abstract

This study evaluated in several ways the concurrent and discriminant validity of measuring self-concept organization via Hierarchical Classes (HICLAS) analysis of interview-derived self-descriptions. College students (N = 85) listed and then described using their own words various self-aspects, and they completed standardized measures of depressive symptoms, self-esteem, and verbal ability. As expected, self-views were multifaceted and hierarchically organized. The proportion of negative self-descriptors provided was positively correlated with depressive symptoms and negatively correlated with self-esteem. Positive and negative self-descriptors were compartmentalized across self-aspects. These results supported the concurrent validity of the information yielded by HICLAS analysis. Supporting the discriminant validity of the method, indices of the elaboration of the self-concept were not correlated with depressive symptoms. No HICLAS variables were significantly confounded by individual differences in verbal ability. We conclude that HICLAS analysis of self-generated descriptors is a promising, flexible method of idiographic assessment of self-concept organization.