Predictors of student success in Higher Education: Secondary school internal scores versus national exams

Abstract

In many countries entrance to Higher Education is determined by the performance of students in secondary school and/or the scores obtained in national exams. The relative weight of these two scores on the admission decision is a relevant policy topic, given its implication on who is admitted to university. The purpose of this paper is to investigate the relative predictive power of these two measures on the academic performance of students in Higher Education. It makes use of a dataset of Bachelor students from Portuguese Higher Education institutions with detailed information about their characteristics and past achievement results. The measure of academic achievement considered is the Bachelor's final average score. The main finding is that the scores given by teachers in secondary school are better predictors of subsequent performance than the access exam scores. The relevance of factors like working status, social support and gender vary with the reputation of the degree and the type of Higher Education institution, university versus polytechnic. A noteworthy result is that the added predictive contribution of parents' education is always negligible when past success measures are already taken into account.

In search of valid non-cognitive student selection criteria

“…For personnel selection, the predictive validity is lower, at 75%. That grades are more often a better predictor of student performance than standardized tests is further supported by numerous studies on grades and standardized tests, which display the positive effects of both but confirm grades as generally having a higher positive predictive validity (Rhodes, Bullough, and Fulton 1994;Hoffman and Lowitzki 2005;Cerdeira et al 2018). One explanation for this is that grades reflect students' self-regulatory competencies better than standardized tests, and these are needed to be a successful student (Galla et al 2019).…”

Benchmarking of secondary schools based on Students’ results in higher education

“…The predictive power of the variables considered (gender, entry grades and type of secondary school attended) is very low as these variables can only explain about 6% of the variability around the mean of scores on exit (corresponding to the final year of the degree attended by the students). Cerdeira et al (2018) investigated the predictive power of internal grades and exam scores in the success of university students in Portugal. They used a sample of more than 20 thousand students and a regression model to predict their success, measured as the final grade at the end of the higher education degree.…”

The Effectiveness of Remedial Courses: The Case of Undergraduate Students in Industrial Engineering

“…Significant predictors of academic performance typically correlate with field-specific prior knowledge, with the regularity of the earlier schooling career, and with the socio-economic background of candidates (Perlberg, 1967;Hagedorn et al, 1999;Kreysa, 2006;Ehrenberg, 2010;Kokkelenberg and Sinha, 2010;Rask, 2010;Faulkner et al, 2014;Cerdeira et al, 2018;Derr et al, 2018;Windle et al, 2018). Prior knowledge is usually measured by high school outcomes (grade point average, GPA, or score achieved at high school final exams), by the coherence between high school type and tertiary education field of studies, and by performance in the admission test 4 .…”

“…Söderlind and Geschwind (2017) compare the predictive power of an aptitude test with high school GPA and find that the former predicts retention rate, whereas GPA predicts credits earning. 5 Cerdeira et al (2018) (Kokkelenberg and Sinha, 2010). In addition, domain-related prior knowledge from secondary school plays a more crucial role (Derr et al, 2018), possibly due to the cumulativeness of mathematical knowledge across the subsequent courses of engineering undergraduate programmes (Rask, 2010;Derr et al, 2018).…”