This study explored how self-concept in arts-related areas could be integrated into Shavelson and Marsh's Hierarchical Model of Self-Concept. Undergraduate and graduate students (N=362) from a large midwestern university completed the Self Description Questionnaire III and a questionnaire assessing self-concept in music, art, dance, and drama. Confirmatory factor analyses and hierarchical confirmatory factor analyses were used to test a series of competing models based on previous studies by Marsh and Shavelson and other hypotheses. The results provided some support for Shavelson and Marsh's Hierarchical Theory of Self-Concept, but also highlighted weaknesses in the hierarchy. Consistent with prior research, most of the covariance among first-order factors was explained by the final hierarchy, and a third-order general self-concept factor was well-defined. Evidence also was obtained supporting prior findings that the second-order academic self-concept factor should be replaced by separate mathematics and verbal factors, and the second-order nonacademic self-concept factor should be replaced with separate physical/social and moral factors. An important new finding was that self-concept in the arts was best represented by a second-order factor on equal footing but distinct from academic and nonacademic self-concepts. One-half of the variance in the majority of first- and second-order factors was unexplained by higher-order factors in the model, suggesting that researchers and practitioners should continue to assess self-concept separately in different content areas. (Author/NB)