Approaches to learning does matter to predict academic achievement

Abstract

Meta-analyses show that the correlations of approaches to learning with academic achievement are low. However, only one study has controlled these correlations in the presence of intelligence. Our study investigates the incremental validity of the surface and deep approaches taking cognitive abilities as control and applied the multiple linear regression and CART regression tree method in a diversified sample of high school students from Minas Gerais, Brazil. We found a prediction of 20.37% and 35.20% of the outcome variance, through the linear regression and CART, respectively. Our results show that although cognitive abilities are more important than students’ approaches, the approaches have incremental validity and important roles as predictors of academic achievement.

Keywords

• Students’ approaches to learning
• Achievement
• Intelligence
• Incremental validity
• CART algorithm