The secret behind Singapore’s success: identifying the top predictors of student reading achievement using machine learning / El secreto detrás del éxito de Singapur: identificación de los principales predictores del rendimiento lector de los estudiantes mediante el aprendizaje automático

Abstract

Despite the consistently remarkable success of Singapore students in international assessments, little is known about the critical factors that drive their reading achievement. Much of the existing research has narrowly focused on a few relevant factors. However, since reading achievement is a complex phenomenon simultaneously determined by numerous different factors, a more integrative lens is needed. This paper aims to demonstrate the application of machine learning to determine the most critical factors that could predict Singapore students’ reading performance in the Programme for International Student Assessment (PISA). Based on the PISA framework, the variables were categorized into ‘student background’, ‘schooling’ and ‘non-cognitive/metacognitive’ constructs. The results indicated that the variables associated with the ‘non-cognitive/metacognitive’ constructs (e.g., metacognition and joy of reading) were deemed as the key predictors of achievement. Our study can provide valuable insights for policymakers and educators, aiding them in prioritizing factors to address in their endeavours to improve learning outcomes.