How much do we need college admission tests?

Abstract

Scrutiny about and controversy regarding the use of college admissions tests have increased since COVID-19 eliminated mass face-to-face testing. At that time, many selective colleges chose to make their admissions decisions without using such tests. More recently, the controversy has been sustained as some selective colleges and universities have reinstated requiring test scores. These tests, primarily the ACT and SAT, have received empirical support but also have faced substantial criticism from test critics and researchers. This paper examines contributions of admissions tests for predicting college outcomes in the presence of additional information available for making college decisions. We investigate relations of ACT tests with academic outcomes for four successive (2011–2014) first-year cohorts at a large, moderately selective Midwestern Research 1 university. Analyses examine how strongly ACT college admissions tests are related to short- and long-term college success for students controlling for their backgrounds (citizenship, race/ethnicity, first-generation, sex, socio-economic status). ACT tests contributed modestly to the prediction of first-year performance. However, with first year college GPA included, ACT tests failed to add to the prediction of GPA in years 2, 3, and 4, and to prediction of dropout and graduation. These findings inform conversations about the use of college admissions tests post-COVID.