The goal of this project is to uncover hidden biases in the standardized testing process and identify all factors that contribute to student performance on these exams. A secondary motivation for this analysis could be to find post‑secondary institutions with the most severe selection biases. Although once thought to be useful for classifying student competence and preparing students for admission to college, in recent years, the merits of standardized testing have been debated due to multiple problems preventing results from being an ideal metric for measuring students’ academic performance. Cheating scandals, test center closures during the pandemic, and, most pertinent to this project, inequity in test preparation resources based on socioeconomic factors, have all contributed to questions about the validity of standardized tests and their incorporation in college admissions decisions.
This project aims to investigate the impacts of socioeconomic factors and standardized testing on college admissions practices. We do so by analyzing institutional-level data about students’ race/ethnicity and economic status, as well as data about the institutions’ sizes, admissions practices, and aggregate secondary school performance to determine if these factors play a statistically significant role in test scores and admissions, and to what degree these factors are significant. The goal of this project is to uncover hidden biases in the admissions process and identify all factors that could contribute to an institution's admissions statistics, especially in the context of testing. This is accomplished through a random forest regression with feature analysis, an investigation into the effect of Pell Grant reception with a single component regression, and principal component analysis with K-Means clustering.