The authors examined how certain ability-related academic self-perceptions are organized in the self-systems of 5th graders. One hundred ninety children completed questionnaires concerning their perceived ability, expectancies for the near future, expectancies for the distant future, and causal attributions to ability. Each of these 4 self-perceptions was assessed separately for math, reading, and general school. Confirmatory factor analyses revealed that children differentiated their ability-related self-perceptions in ways that are not accounted for by current models of academic self-concept. That is, in addition to differentiating the 3 domains of math, reading, and general school, children also differentiated the 4 selfperceptions within academic domains. These findings both support and extend current theories of academic self-concept.Children's ability-related self-perceptions are important predictors of their task engagement and performance. For example, children who perceive themselves as highly competent are more likely to persist when they confront difficulties and are more likely to use the abilities and strategies they possess (Harter, 1990;Pintrich & Schrauben, 1992). However, ability-related self-perceptions encompass a wide range of constructs, such as self-concept, perceived ability, and performance expectancies. Although these constructs may be highly related, there are potentially important distinctions between them that are often ignored.The purpose of the present study was to examine how certain ability-related academic self-perceptions are organized in the self-systems of late elementary school children. Specifically, this study focused on the interrelationships among four ability-related self-perceptions in the academic domains of math, reading, and general school. The four self-perceptions were (a) expectancies for the near future (i.e., expectancies for their next report card), (b) expectancies for the distant future (i.e., during the next academic year), (c) perceived ability (i.e., perceived "smartness"), and (d) causal attributions to ability (i.e., success attributions to high ability and failure attributions to factors other than low ability).The present study was influenced by the work of Marsh and colleagues on the organization of academic self-concept