A Study of predictive validity of O-level, A-level and an aptitude test in relation to the performance at the University of Nairobi

Abstract

This study is an empirical investigation to determine the predictive validity of A-level, O-level and an aptitude test in relation to performance at the University of Nairobi. In this prediction study a sample of about 260 subjects was used. The subjects were admitted to the University in academic year 1972/73 and were students in the following faculties:- Agriculture, Arts, Architecture, Education and Medicine. O-level and A-level are achievement tests done by candidates after completing four and six years of secondary school curriculum respectively. The candidates sitting for the O-level do papers ranging from Geography t H.istory, English language etc. (social sciences) to Biology, Mathematicst Chemistry etc. (Pure Sciences). The candidates who proceed on for A-level can either offer Arts.(social science) or Science (Pure Science). The A-level is done after two years of study after completion of O-level. Selection into A-level classes is based solely on O-level performance. The O-level and A-level grades and c~ores of an aptitude test were used as predictor variables and the standardized scores obtained~n University examinations were used as criterion variables. The O-level grades and A-level grades were weighted to make analysis possi~le. For O-level the weigh~ing sy&tem of 9,8, ••••t2,1 was used for the grades 1, .»: 2, •••• ,8,9 respectively i.e. grade 1 (the best O-level grade) was giv7n a weight of 9 etc. For A-level a weighting system of 6,5, ••••,1,0 was used for the grades A, •••• ,E,O and F -2- respectively. A similar kind of weighting system of A-level is used by the selection Board of the University for their selection purposes. The Board uses the A-level grades only. The analyses based on the intercorrelations between variables were carried out namely, (i) Multiple regression analysis and (ii) Canonical correlation analysis It is usually felt that estimates of previous school performance and maximum performance tests are not the only categories of possible predictors. There are two other possible types (Drenth, 1975):- 1. Personality variables .• 2. Miscellaneous category of biographical information, antecedents, physical qualities, socio-economic factors, linguistic, racial and tribal background variables etc. In this study, however, attention has been paid only to previous school performance and maximum performance test (aptitude test) in predicting performance at the University of Nairobi. The multiple regression analysis showed A-level to have an appreciable redictive efficiency for university examinations in these faculties; (i)~ Architecture during second and third year (ii) Medicine in both first and second year • ..,;' O-level has pred~c~ive efficiency in medicine during first year. The aptitude test score bas a predictive efficiency in - 3 - faculties of:- (i) Agriculture during first and third year and (ii) Education in first year. For the faculty of Medicine the study showed that the set consisting of O-level, A-level and English language is more efficient as a predictor of first and second year university examination than best single predictor, A-level. For Architecture the optimum prediction can be achieved by considering A-level and aptitude test when we take second year uni versi ty examination as a cri terion. The optimum prediction can be obtained by considering ..aptitude test, O-level and Mathematics in faculty of Education when first year performance is taken as a criterion. The canonical correlation analysis showed it is only in faculty of Medicine in which the set of five predictors (i.e. English language grade, Mathematics grade, O-level grade aggragate, A-level grade aggr~gate and aptitude test score) can predict the university success at level of significant less than 0.05. The university success (a linear combination of either two or three criterion variables) can be predicted w~h greater accuracy for medical students followed by Architecture, Arts, Education and lastly Agriculture students. The study has shown that A-level does exhibit ....-" satisfactory preaictive validity but not in all faculties consid~red, hence the evidence in the study does not lend strong support to use of A-level for selection to all faculties in the University.