Using Multilevel Logistic Regression Model to Investigate the Factors Influencing 15-Year-Old Students' Mathematics Achievement

• In Event: Poster Session 9
  In Poster Session: Using Multilevel Modeling in Substantive Research

Abstract

The purpose of this study was to investigate the factors affecting the mathematics performance of 15-year-old students. The sample was obtained from students who participated in Programme for International Student Assessment (PISA) 2012 in the United States. The two-level logistic regression model, as a special case of hierarchical generalized liner models, was applied to examine the factors to influence mathematics performance for 4,798 15-year-old students nested within 162 schools. Results from this study indicate that gender, parental education level, family socioeconomic status, learning time on mathematics, and student-teacher relations have positive effects at the student level. Teacher morale was found to have positive effects on student mathematics performance while availability of computer has negative effects on student mathematics performance.

Authors

• Ruixue Liu, University of Kentucky

• Kelly D. Bradley, University of Kentucky

• Meng Fan, University of Delaware