Temperament and objective and subjective measures of academic competence were interrelated in order to test two alternative models: (1) a direct effects model stressing intraorganismic, noncontextually mediated links; and (2) a developmental contextual model emphasizing social interactional processes between students and teachers. Data from the Pennsylvania Early Adolescent Transitions Study were also used in the test. Results of LISREL analyses, both at the end of grade 6 and the end of grade 7, supported the developmental contextual model, and indicated that there were significant paths between second-order temperament factors and teachers' ratings of students' academic competence. In turn, these ratings were related to students' self-rated competence, to grade 6 and 7 grade point averages, and to grade 6 scores on a standardized achievement test. In sum, individual differences in developmental change occurred by means of person-context relations evoked by a person's characteristics of individuality. Findings are discussed in regard to the possible role of temperament in moderating early adolescent coping. (RH)