Evaluating a predictive model of student performance in introductory calculus-based physics

Abstract

Various factors are known to correlate with performance in physics and other STEM courses. In this study, we report on students’ performance in an introductory calculus-based physics course at the university level and correlate their performance (as measured by final grade) with a variety of measures of academic and cognitive ability. These measures include mathematical and English language ability (as measured by the SAT), incoming GPA, and spatial and scientific reasoning ability (as measured by the Mental Rotation Test and Classroom Test of Scientific Reasoning). We also consider the effect on student performance of factors such as employment status, and science and mathematics self-efficacy (as measured by a survey). In this study, we report the prediction efficiency of the ability of two multivariate linear models to predict student success in the course.