Adopting Mixed Effects Random Forest to Predict Students' Reading Performance Using ICT–Related Variables

• In Event: AERA Virtual Poster Session 4
  In Poster Session: Adopting Mixed Effects Random Forest to Predict Students' Reading Performance Using ICT–Related Variables

Abstract

The aim of this study is to predict students’ reading performance using Information and Communication Technology (ICT)-related variables with the mixed effects random forest method. The study utilized Programme for International Student Assessment 2018 United States data, which included 2,592 students across 118 schools. In terms of prediction ability, the accuracy of the prediction model reached 93%. Regarding the variables’ degree of importance, our prediction model showed that the use of technology in schools and the duration of technology use by the students in class were more meaningful predictors for the students’ reading performance than other ICT-related factors. The findings indicate that ICT availability and use in classrooms and in schools can have a tremendous impact on students’ reading performance.

Authors

• Wenting Weng, Texas A&M University - College Station

• Seoyeon Park, Texas A&M University - College Station