TIMSS Student and Teacher Predictors for Interest in Learning Mathematics Explored via Machine Learning, Mnet

• In Event: Machine Learning Applications in Education

Abstract

Despite having high academic achievement in mathematics, Korean students have low interest in learning mathematics. This study aims to investigate important student and teacher predictors for interest in learning mathematics. To explore important predictors, the authors employed penalized regression, specifically Mnet, a machine learning technique, to handle multicollinearity problems. Of the 306 TIMSS variables explored, Mnet selected 25 student and 3 teacher variables as important predictors for interest in learning mathematics. Newly found variables of interest in learning mathematics included students’ tiredness, students’ confidence, value recognition in mathematics, attitude toward mathematics, and school focus on academic achievements. The study results can shed light on how to improve educational policies in order to increase student interest in learning mathematics.

Authors

• Miryeong Koo, Korea National University of Education
  Presenting Author

• Jin Eun Yoo, Korea National University of Education
  Non-Presenting Author