A Comparative Analysis of Selected Studies in Student Performance Prediction

Predicting student success has long been an interest of institutions of higher education as well as organisations responsible for preparing high-stake, standardised tests administered at national and international levels. This study discusses how performance prediction studies have evolved from those that use demographic data and high school grades to predict success in college to those that utilise sophisticated data collected in non-traditional educational platforms to predict end-of-course performance and to those that show how student progress can be tracked in a continuous manner. A total of 56 studies published since the nineties are discussed. Views on strengths and weaknesses as well as observed opportunities for improvement are presented. The consistently high results reported in many of the studies shall convince the reader that automated solutions to the problem of predicting student progress and performance can either be tailored for specific settings or can be adopted from similar settings in which they have been utilized successfully. A recommendation on how to build upon recent success is provided.