This paper aims to provide a parsimonious but rigorous model to assist decision-makers to determine critical factors which can lead to higher graduation rates amongst higher education institution (HEI) participants. It predicts the odds of dropout amongst university students, using HEI data from a developing country. This is used as a basis for a Student Retention Predictive (SRP) Model to inform HEI administrators about predicted risks of attrition amongst cohorts.
A classification tool, the Logistic Regression Model, is fitted to the data set for a particular HEI in a developing country. The model is used to predict significant factors for student dropout and to create a base model for predicted risks by various student demographic variables.
To reduce dropout and to ensure higher graduation rates, the model suggests that variables such as age group, faculty, academic standing and cumulative GPA are significant. These indicative results can drive intervention strategies to improve student retention in HEIs and lessen the gap between graduates and non-graduates, with the goal of reducing socio-economic inequalities in society.
This research employs risk bands (low, medium and high) to classify students at risk of attrition or drop out. This provides invaluable insights to HEI administrators in the development of intervention strategies to reduce dropout and increase graduation rates to impact the wider public policy issue of socio-economic inequities.
