Within the school domain, studies have suggested that prior academic achievement may be an important influence on an adolescent's academic self-concept. For instance, Marsh & Yeung (1997) found that not only can adolescents' levels of academic self-concept affect their later performance in school, their self-concepts are also influenced by their prior academic achievement, as indicated by their grades and their test scores. Hence, the relationship between academic self-concept and academic achievement seems to be reciprocal in nature, with each affecting the other. The sample in this study, composed of 603 mostly white, Catholic boys from Australia, makes it difficult to draw generalizations about the extent to which this pattern would hold for all adolescents. Yet, a second longitudinal study (Marsh, 1994) examining a national sample of adolescents in the U.S. (the National Education Longitudinal Study of 1988) also found a link between students' test scores and grades, and their levels of academic self-concept, lending support that these findings are probably not specific to the sample in the Marsh &Yeung (1997) study. Additionally, this second study found that the path from students' earlier test scores in a given domain to their later levels of academic self-concept in that domain occurs mostly through the grades they receive within that subject - in other words, students who score well on tests tend to receive higher grades in school, which in turn leads to their having higher levels of academic self-concept.

In addition to the reciprocal relationship between achievement and academic self-concept, it also appears that adolescents' academic self-concepts in a given domain can be affected by their achievement within another academic domain. For instance, a study following a national sample of adolescents for a three-year period found that adolescents' who received a high grade in one academic domain (e.g., math) demonstrated lower self-concepts within another academic domain (e.g., English) at a later point in time. This relationship was found in the presence of controls for earlier levels of English self-concepts and grades in English class, suggesting that the result is not due to differences in achievement or self-concepts at the start of the study. The authors of this study, along with other researchers interested in academic self-concept, suggest that adolescents form their academic self-concepts based on both internal and external frames of reference. In other words, the self-concept in a given domain is based not only on "external" indicators of competence in that area (such as grades or test scores), but also on relative performance in one domain compared to other domains. Hence, the better an individual performs in math, the lower his or her English self-concept will likely be and vice versa.