Predicting academic success : towards an admission and placement model

The study aimed to develop a model predictive of academic success based on variables assessed during the admission process and the relationships of the variables with academic progress and academic success at a South African university. To do this, the study investigated if school exiting results and admission tests were significant predictors of first-year grade point average and of final academic grade point average over a period of six years, taking into account different race and gender groups, as well as different school examination authorities. The relationship between first-year and final grade point average was also assessed in order to develop a comprehensive and integrated model predictive of academic success. The disaggregation of the dataset (N=3418) into different race and gender subgroups and different school examination authorities revealed to be useful and necessary as differences in mean scores of the predictor and criterion variables were observed and vastly different predictive models were presented, indicating that an overall model to predict academic success for all students is not appropriate. The results of this study highlighted the reality of significant inequalities in university outcomes for students of different race and gender subgroups in particular. The results of this study should be interpreted in the context of guiding admission decisions, and developing policies and rules that are fair, equitable, reliable and justifiable in terms of the ability and probability of students to succeed, bearing in mind individual differences in the prediction model with regard to race, gender and different school examination authorities.Thesis (PhD)--University of Pretoria, 2019.TM2019Human Resource ManagementPhDUnrestricte