What is the relationship between high school grades, test scores, and academic success in college?

• In Event: Test-Optional Admissions Policies and Other Trends in Selective College Admissions

Abstract

Using rich administrative data from a selective public flagship university from 2015 through 2023, this study explores the role standardized tests play in predicting academic success in college. On a wide range of measures including first-year grades, first-year retention, and degree completion, this study finds standardized tests (both before and after the COVID-19 pandemic) tend to explain less variation and have smaller standardized coefficients than high school grades and other factors considered in holistic review.

Findings from this study are most similar to those found in “Crossing the Finish Line” (Bowen, Chingos, & McPherson, 2011), which focused on public selective and flagship institutions. In their work, which predates the recent expansion of test-optional admissions since the COVID-19 pandemic, they found similar results – students’ high school grades tend to provide more explanatory power and stronger relationship with various student success outcomes including academic performance and degree completion. One exception is first-year grades, where test scores tend to have a statistically significant and positive relationship.

To test the robustness of these results, the study explores whether the relationship between tests and academic outcomes at this institution changed after adopting test-optional admissions in 2020. It also explores whether students who withheld their test scores (“non-submitters”) had poorer academic outcomes than test submitters. Lower-scoring students tend to withhold their scores; however, preliminary results find little evidence that non-submitters have poorer academic outcomes than submitters. Aside from first-year grades, tests are not strongly associated with retention or degree completion. An additional robustness check explores whether these patterns hold for admitted students who ultimately enroll elsewhere (i.e., at less selective public institutions in the state). Preliminary findings suggest the patterns hold even among students who enroll elsewhere – the relationship between tests (and decisions to submit scores) are only weakly correlated with academic outcomes in these robustness checks.

Recent media attention on test-optional admissions has focused almost exclusively on the nation’s most highly-selective private institutions – largely drawn from the Ivy League – that admit fewer than 10 percent of applicants. Public flagship institutions like the one in this study that admit around 50 percent of applicants are often caught in between these major debates. On one hand, tests may provide useful information for differentiating applicants from one another. On the other hand, tests may provide redundant information that is already known about the applicant through other measures.

Ultimately, context is critical for determining whether and under what conditions test scores add valuable information for admissions review processes and enrollment management goals. As explained in “Crossing the Finish Line,” the question for public flagship institutions is less about “should we test” and more about “in what settings can we expect various kinds of tests, and other measures, to be especially helpful?.” For the institution studied in this analysis, preliminary findings suggest standardized tests can have some predictive power for first-year grades but provide little information beyond what is already collected from students in the application and review process.  

Author

• Nicholas W Hillman, University of Wisconsin-Madison
  Presenting Author