What Are the Predictors of Math Ability in the World of Machine Learning?

• In Event: AERA Virtual Poster Session
  In Virtual Poster Session: Division C - Learning and Instruction Virtual Poster Session

Abstract

Previous studies had debates on the predictors of mathematics performance such as gender, self-concepts, spatial abilities, or home resources. (Else-Quest et al.,2010; Lee et al, 2018; Gilligan et al., 2017; Yoshino, 2012; Wittwer et al., 2008).However, none of these results can explain all possible reasons for improving mathematical ability.
The goal of the following report is to find out the correlation between various factors and academic performance by analyzing the eighth-grade student data of Taiwan and the United States from TIMSS 2015 and 2019 with logistic regression in machine learning. According to the result, self-confidence and the number of books at home are highly predictive to math performance.

Authors

• Yu-Jou Chou, Pennsylvania State University
  Presenting Author

• Yi-Ling Cheng, Kaohsiung Medical University
  Non-Presenting Author