A Machine Learning Approach to Predicting STEM College Major Choice

• In Event: What Matters for STEM Participation in College? A Longitudinal View of STEM Identities and STEM Pathways

Abstract

Despite the high demand for STEM talent, only a small portion of high school students aspires to STEM-related careers. To understand the reasons for the current low STEM participation, this study presents a more complete picture describing the factors relating to the U.S. 9th grade students in fall 2009 who eventually enrolled in STEM-related college majors. Using a machine learning technique, this study found that four factors – science identity, math achievement, total STEM credits, and GPA in AP/IB math courses – played important roles in predicting the likelihood of students enrolling in STEM college majors.

Authors

• Chi-Ning Chang, The University of Kansas

• Shuqiong Lin, American Institutes for Research

• Oiman Kwok, Texas A&M University - College Station

• Guan Saw, Claremont Graduate University