Statistical identification of predictors of dropout in secondary education: evidence from Malaysia

Abstract

The issue of school dropout has been long rooted in Malaysia. To address this issue effectively, further insight into the predictors of the school dropout and the early statistical identification of dropouts are crucial. Nevertheless, uncertainty has been persistent on these two aspects. This paper aims to identify the predictors and assess the feasibility of early identification of school dropouts. Using a school administrative database that comprised data on more than 100,000 students, the school dropout predictors, such as gender, academic achievement, and household income were identified. The estimated model correctly identified early on the 20.83% of the school dropouts. This is substantially higher than the overall sample mean of school dropouts (5%). Thus, it is feasible to identify early on the school dropout statistically. The findings have provided insightful inputs that could strengthen our intervention strategies and policies to help alleviate the problem of school dropout in Malaysia.