The purpose of this research was to estimate the educational production functions for students belonging to different subgroups in a principles of economics course. Three samples were used: (1) a large lecture class at a large state university (185 students), (2) several small sections taught by different instructors at a small private university (149 students), and (3) two sections taught by one instructor at a small private university (66 students). The three groups were disaggregated on the basis of sex of the student, age of the student, the student's score on a pretest (preTUCE), the student's pre-attitude toward economics, and the student's ability as measured by American College Test (A.C.T.) scores. Ordinary least square regression analysis was used to estimate the production function for each group and an F test was used to test for group differences. Results indicate that there was no difference in the set of coefficients for the production function when the sample was disaggregated by sex or age. However, a significant difference was indicated when the sample was disaggregated by preTUCE or A.C.T. scores. A test of coefficients of the disaggregated regressions showed that teachers have different effects on low, average, and high ability students. These results indicate that students with different levels of preknowledge or different ability levels utilize their individual attributes in different ways in order to learn. Findings suggest that students of different ability levels might learn more if separated and taught by different pedagogical methods and instructors. Over 25 data tables are provided. (Author/LH)