Two hundred and forty-three female monozygotic (MZ) and 164 female dizygotic (DZ) twin pairs, aged 11 and 12 years, who participated in the ongoing Minnesota Twin Family Study, completed six specific scales of the Piers–Harris Children's Self-Concept Scale (P–H). Model-fitting analyses yielded three major conclusions. First, approximately 30% of the variance in specific self-concepts in female preadolescents was due to genetic factors, with the remaining variance being accounted for primarily by nonshared environmental factors and measurement error. Second, the underlying common genetic, shared environmental, and nonshared environmental factors influenced specific facets of self-concept directly and independently, rather than through an intervening phenotypic general self-concept. Finally, whereas genetic, shared environmental, and nonshared environmental factors were necessary to explain the commonality among the specific self-concept scales, only genetic and nonshared environmental factors were sufficient to explain the specificity of those scales.