Abstract
Prediction of academic success of first year students. The recent amalgamation of the various regional education systems in South Africa and as a result, the restructuring of the education dispensation, led to a number of problems which impacted negatively on the examination process. Against this backdrop, the present study attempted to find a valid alternative for matric marks as a predictor of academic success. Multiple regression analyses and canonical discriminant analyses were performed to construct a comprehensive model, which can be used to predict the academic achievement of students, and serve as an alternative for matric marks. Regression analyses with average exam mark as criterion were performed on the data of the three largest faculties. Multiple correlations of 0.59, 0,56 and 0,52 were obtained. Discriminant analysis were done against a promotion criterion (promoted/not promoted), for each of the faculties. According to the results of the discriminant analyses the percentage correct classifications varied from 62,7% to 81,1%.