Research investigated: 1) the effect of instructional mode on student achievement; 2) the relationship between a) the student factors of aptitude, sex, previous achievement, educational level, and academic major and b) achievement; and 3) the possibility of a homogeneity of regression equations for achievement of students in varying instructional modes with respect to student traits. Students in an engineering course received either conventional or computer-assisted instruction (CAI), and curvilinear regression procedures were used to construct models of achievement. Findings showed the CAI students achieved better in the course, although their overall grade point averages for the semester were not significantly better. Significant relations were found between instructional modes and student factors in terms of achievement, but homogeneity of regression equations was not manifested. It was concluded that: 1) CAI was an effective instructional mode; 2) models of the student could be developed to predict achievement with a particular instructional mode, but their predictive value was limited; and 3) it was not possible to construct a predictive model of achievement which did not take into account the instructional mode. Additional research was recommended on curvilinear analysis and other student variables. (LB)