The Hierarchical Structure of Self-Concept and the Application of Hierarchical Confirmatory Factor Analysis


  • The author would like to acknowledge the helpful comments by Dennis Hocevar, Jenifer Barnes, and Raymond Debus on earlier drafts of this article, and to thank Gary Richards and the Outward Bound program for assistance in the collection of the data and financial support.

HERBERT W. MARSH, Reader, Educational Psychology, Faculty of Education, University of Sydney, Sydney NSW 2006, Australia. Degrees: BA, Indiana University; MA, PhD, University of California at Los Angeles. Specializations: self-concept, measurement and evaluation.


This investigation (a) tested the ability of an a priori hierarchical structure of self-concept derived from the Shavelson model to explain responses to the Self Description Questionnaire (SDQ) III, and (b) demonstrated the application and problems with the use of hierarchical confirmatory factor analysis (HCFA). A first-order factor analysis clearly identified all 13 facets of self-concept that the SDQ 111 is designed to measure. A series of HCFA models supported the separation of the 13 SDQ III facets of self-concept into academic and nonacademic components, and the academic facets into math/academic and verbal/academic components. However, support for separation of nonacademic facets into the physical, social, and moral second-order factors was not strong. Third-order HCFA models resulted in a clearly defined hierarchical general self-concept that was substantially related to Esteem and to physical, social, and emotional components of self-concept, but not to the academic and moral components. However, the hierarchy was so weak that first-order facets could not be accurately inferred from the higher order facets.