Machine Learning to Identify Important Predictors of School Science Literacy Achievement: Evidence From the United States

• In Event: New Pathways: Breaking Ground on Pioneering Data Use Practices for Education

Abstract

Many studies have examined the relationships between school science literacy achievement and various school level factors such as school resources or teacher quality. Different previous studies, the main purpose of this study was to examine student, teacher, as well as principal reported variables at school level in their predictive power of school science literacy achievement using the machine learning approach. Based on data from 266 U.S. schools collected via the Program for International Student Assessment 2015, we specifically examined relative importance of 40 predictors in predicting school science literacy achievement. The results indicated 10 relative important predictors, with seven being students’ science learning dispositions. The results were discussed with respect to the implications for school leaders in setting program priorities.

Author

• Cody Ding, University of Missouri - St. Louis