A critical appraisal of selection criteria and academic progression of first and second year medical students at the University of the Free State

Abstract

English: The changing of the evaluation systems used for Grade 12 scholars in South African schools and the transformation principles of the Department of Education, compelled the University of the Free State (UFS) to start looking into alternative criteria for the selection process of medical students. One of the alternative criteria explored is the Health Science Placement Tests (HSPTs). The overall aim of this study was to assess the relationship between the HSPTs, school performance and other factors and academic performance during the first two years of study at the UFS. The specific objectives of the study were to conceptualise and contextualise the problem of selection of medical students at the UFS and to identify factors in different regions of the world that play a role in the selection of medical students by means of a thorough literature survey, but also to assess the influence of the current selection criteria and additional criteria on the performance of first and second year medical students at the UFS. A quantitative research approach was followed. The study population comprised of the first year medical students of 2004 and 2005 and second year medical students during 2005 and 2006 at the UFS. The demographic information of the students, their HSPTs results, school performance and academic performance results during first two years of study were statistically analysed to detect associations. Data for the study was obtained from the several databases of the University of the Free State and was collated by the researcher. The data management and analysis in this study was conducted by the staff of Statistical Consulting Service, Department of Statistical Sciences, University of Cape Town, using a variety of available statistical techniques. The correlation between all the numeric and categorical variables and the outcome variable were checked. These results showed the degree to which the variables changed together and allowed the researcher to indicate those with a predictive relationship. Strong to moderate correlations were found to be present between the averages of the first two years of study and English, Mathematics, Science and Biology of the Grade 12 marks, the PTEEP, MACH, MCOM and SRT of the HSPTs and the M-score. A weak negative correlation was found between the age of the student and whether or not they had any tertiary education and both the first and second year averages.By using the simple linear regression technique of analysis, the researcher evaluated the effect that each of the individual variables had on the first and second year averages. The following variables had a significant influence on the first two year's average marks: English, Mathematics, Science and Biology average mark, School Poverty Ouintile Index, M-score and the HSPTs average.By using a multiple regression analysis, the predictors of dependent variables upon the outcome variable were tested, while the independent variables were held fixed. After following a step-wise regression analysis, the best fit model was the model evaluating the relationship between the first and second year average marks independently and the age of the student, the English, Mathematics, Science and Biology scores of Grade 12 and the PTEEP, MACH, MCOM and SRT tests of the HSPTs and the School Poverty Ouintile Index. This model explained 50% variance of score in the first year and 70% variation of score in the second year as a result of the combination of these variables. Although some of the variables were not statistically significant, they were still of conceptual significance. From this analysis it was clear that the more variables that were included, the more reliable or predictive the model was to determine how a student would perform academically at the end of the first two years of study. The conclusion of this study was that the application of different statistical approaches presents a case for the complimentarity of data for use in selection models andapproaches. Through the exploration of different models of regression and association, a particular model was found acceptable as an indicator for good performance during the first two years of study. This choice was based on the fact that the multiple regression model was able to predict the effect that a variable would have on the outcome and the size of the effect. It was able to explain 50% variance of score in the first year and 70% variation of score in the second year and also took into account the effects of other confounding variables. This study and similar future studies will identify reliable and valid selection criteria for medical students who will perform well academically within the M.B.,Ch.B. tertiary education programme.

Afrikaans:Die veranderinge van die evalueringssisteme wat gebruik word vir graad 12-skoliere in Suid-Afrika en die transformasiebeginsels van die Departement van Onderwys, het die Universiteit van die Vrystaat (UV) genoodsaak om te begin kyk na ander kriteria vir die keuring van mediese studente soos die Gesondheidswetenskappeplasingstoetse ("HSPTs"). Die oorhoofse doel van die studie was om die verband tussen die "HSPTs", skoolprestasie en ander relevante faktore en die akademiese prestasie gedurende die eerste twee jare van studie by die UV te ondersoek. Die spesifieke doelwitte van die studie was om die probleem aangaande die keuring van mediese student by die UV en verskeie streke in die wêreld te koseptualiseer en te kontekstualiseer deur 'n deeglike literatuurstudie, asook om die invloed van die huidige keuringskriteria en addisionele kriteria van die prestasie van die eerste twee jare van studie te ondersoek. Die navorsingsontwerp was op 'n kwantitatiewe aanslag geskoei. Die studiepopulasie was die eerstejaar-mediese studente van 2004 en 2005 en die tweedejaarstudente van 2005 en 2006 aan die UV. Die demografiese inligting van die studente, hulle "HSPTs", skoolprestasie en akademieseprestasieresultate van die eerste twee studiejare is statisties geanaliseer om verbande vas te stel. Data van die studie is verkry vanaf verskeie databasisse van die UV en is deur die navorser geamalgameer. Data-ontleding is hanteer deur die personeel van Statistiese Konsultasiedienste, Departement Statistiese Wetenskappe van die Universiteit van Kaapstad wat "n verskeidenheid van analitiese tegnieke gebruik het. Die korrelasie tussen al die kategoriese en numeriese veranderlikes en die uitkomsveranderlikes is gekontrolleer. Hierdie resultate het die graad waarmee die veranderlikes saam verander het aangedui en het die navorser in staat gestelom "n voorspellende verhouding vas te stel. Sterk tot matige korrelasies is gevind tussen die gemiddelde van die eerste twee jare van studie en die Engels-, Wiskunde-, Wetenskap- en Biologie-resultate van die graad 12 punte, die PTEEP, MACH, MCOM en SRT van die HSPTs en die M-telling. "n Swak negatiewe korrelasie is gevind tussen die ouderdom van die student en of hy/sy tersiêre studies voltooi het en die gemiddelde van die eerste- en tweedejaar gemiddelde punte onderskeidelik. Deur gebruik te maak van die eenvoudige liniêre regressiewe tegniek van analise, het die navorser die effek wat elk van die veranderlikes op die eerste- en tweedejaarpunte gehad het, geëvalueer. Die volgende veranderlikes het "n betekenisvolle invloed op die eerste twee jare se gemiddelde punte gehad: die Engels-, Wiskunde-, Wetenskap- en Biologie-gemiddelde punt, die Skool Armoede Indekstelling, die M-telling en die gemiddelde HSPTs. Deur die uitvoering van die veelvuldige regressiewe tegniek van analise, kon die effek van die afhanklike veranderlikes op die uitkomsveranderlike getoets word, terwyl die onafhanlike veranderlikes gefikseer is. Na die uitvoer van "n stapsgewyse regressieanalise, is gevind dat die beste model die een was wat die verhouding tussen die eerste twee jare se gemiddelde en die ouderdom van die student, die Engels-, Wiskunde-, Wetenskap- en Biologiepunte van graad 12, die PTEEP, MACH, MCOM en SRT toetse van die HSPTs en die Skool Armoede Indeks as model getoets het. Hierdie veranderlikes as deel van "n model kon 50% van die verandering in die eerstejaarpunte en 70% van die verandering in die tweedejaarpunte verduidelik. Alhoewel sommige van die spesifieke veranderlikes in die model nie statisties betekenisvol was nie, was hulle steeds belangrik as deel van die model en die invloed van die modelop die punte. Uit die analise kon afgelei word dat hoe meer veranderlikes in die model ingesluit is, hoe meer betroubaar en voorspellend die model was om te bepaal hoe suksesvol die student sou wees aan die einde van die eerste twee jaar van studie. Die gevolgtrekking van die studie was bevestig deur verskillende tegnieke van analise. Deur die ondersoek van verskillende modelle van regressie en assosiasie, is 'n spesifieke model aanvaarbaar gevind as 'n indikator vir prestasie in die eerste twee jare van studie. Hierdie keuse is gebaseer op die feit dat die veelvuldige regressiemodel in staat was om die effek van die afsonderlike veranderlike op die uitkomste en die grootte van die effek te bepaal. Die model was verder in staat om 50% van verandering in die eerstejaar-gemiddelde punte en 70% in die tweedejaargemiddelde punte te voorspel en terselfdetyd die effek van die ander veranderlikes in ag te neem. Hierdie studie en soortgelyke studies sal keuringskriteria vir mediese studente daar stel in die toekoms wat betroubaar en geldig is vir mediese studente wat akademies goed sal vaar in die M.B.,Ch.B.-tersiêre-onderwysprogram.