Belmont University Research Symposium (BURS)

Predicting SAT Scores in Massachusetts Public Schools

Publication Date

Spring 4-2024

College

Sciences and Mathematics, College of

Department

Math and Computer Science, Department of

BURS Faculty Advisor

Dr. Christina Davis

Presentation Type

Poster Presentation

Abstract

The SAT is a standardized test that evaluates a high schooler's reading, writing, and math skills. These scores are used across the United States for college admissions. Based on previous studies, SAT scores are strongly predictive of college performance, student retention, and income. This research project analyzes public school data with a specific interest in predicting SAT scores. Various predictive models were used on data from 292 high schools in Massachusetts in 2017. These models aimed to predict which schools performed above average on the SAT. Python was used to run models including Random Forest, K Nearest Neighbors, and Logistic Regression, all with an accuracy score above 85%. Through feature importance analysis, it was observed that the most significant predictor of a school's SAT score being above average is the percentage of high-needs students. Additionally, schools with high percentages of students with high needs not only performed significantly worse on the SAT but graduated at lower rates and scored lower on Advanced Placement exams. This strong negative relationship between the percentage of high-needs students and academic performance illustrates how the public school system in Massachusetts should make an intentional effort to better serve students with high needs through support and aid so that they can achieve academic success.

This document is currently not available here.

Share

COinS