Abstract:
The selection of students for higher education has been a burning issue on the agenda of South African institutions of higher education for the past decade. Institutions for higher education are experiencing pressure from both their clients and the government to broaden access, but at the same time financial realities force these institutions to admit only those candidates with the potential to be successful in their chosen course of study. The main aim of this study was the identification of variables which relate to academic success amongst Engineering Technology students at Technikon Pretoria, and to incorporate them into a selection battery which would be both valid and reliable. A non-experimental, correlational design was selected, as this research technique is considered the best controlled and most accurate of all non-¬experimental designs. Since a quantitative technique was selected for data gathering, the necessity for a statistical method in the data analysing process was obvious. The sample for this study consisted of a total of 732 Engineering Technology students at Technikon Pretoria. From these, 512 were Civil Engineering Technology students and the remaining 220 were Mechanical Engineering Technology students. These subjects were the total number of students from these two academic departments, enrolled from 1997 to 1999, of whom both psychometric and academic data were available. The sample consisted of 14.75% female and 85.25% male respondents and was representative of the cultural diversity of the Technikon campus. The competencies indicated by academic staff involved with the training of Engineering Technology students at Technikon Pretoria were hypothesised to be indicative of a potentially successful student. After the identification of these predictor variables the assessment battery to be used in this study was compiled. This was then included in a comprehensive set of data regarding each applicant, together with the required school performance. A forward stepwise multiple regression analysis was performed on the data in order to establish the predictive validity of the assessment battery. The expansion of the traditional selection procedure to include the potential assessment phase proved valuable, as the validity of all prediction models improved with the addition of the indices from the Potential Index Batteries. The prediction models were found to be unbiased against students from the previously disadvantaged school systems and can thus be said to be culture fair.
