Comprehensive Predictors of Engineering Success Using Cognitive and Non-Cognitive Variables

This study investigates predictors of student success in an undergraduate civil engineering program.Building upon previous work, the scope encompasses a broader population of engineering students and a more comprehensive set of predictive variables.Anonymized academic records from the Virginia Military Institute (VMI) were analyzed using a multi-faceted statistical framework to examine the relationship between student success, defined by graduation GPA and on-time graduation, and a wide range of pre-college metrics.The methodology used in this study includes Principal Component Analysis (PCA) to address potential collinearity among variables and to support qualitative assessment of the most influential factors.A multilinear framework was employed to quantify relationships between precollege metrics and graduation GPA, while a probabilistic classification approach was used to estimate the likelihood of on-time graduation.Independent variables include standardized test scores (SAT/ACT), high school GPA, incoming college credits, mathematics placement exam scores, and participation in a summer transition program.Initial findings suggest that a student's academic foundation, as reflected by high school GPA and placement scores, demonstrates a stronger correlation with long-term success than performance on a single standardized test.This research aims to provide a holistic and interpretable framework for identifying critical factors in engineering student success by combining probabilistic modeling with qualitative assessment.It also offers actionable insights to inform institutional recruitment, retention, and student support strategies.