Predicting College Success

How Do Different High School Assessments Measure Up?
Predicting College Success
Michal Kurlaender
University of California, Davis
Kramer A. Dykeman
University of California, Davis


The Smarter Balanced Assessment (SBAC), implemented in California in 2014–15 as part of the California Assessment of Student Performance and Progress, is designed to evaluate students’ levels of college and career readiness. Student scores on the Smarter Balanced Assessment are currently used for both accountability and school improvement purposes. Aligned with Common Core State Standards for college readiness, student performance on the Smarter Balanced Assessment may also predict students’ success in college in a manner similar to other commonly used assessments for predicting college success.

In this report, we investigate how well the Smarter Balanced Assessment, high school grade point average (HSGPA), and SAT predict first-year college outcomes for students enrolled in the California State University (CSU) system and at the University of California (UC) campuses. We also report how the relationship between these assessments and early college outcomes differs by key student subgroups (race/ethnicity and socioeconomic disadvantage). Specifically, we find:

  • The Smarter Balanced Assessment does as well as the SAT in predicting college outcomes, specifically first-year college GPA and second-year persistence rates for students enrolled in CSU and UC.
  • HSGPA is a stronger predictor of first-year college GPA and second-year persistence than either SBAC or SAT at CSU.
  • None of the three assessments (HSGPA, SBAC, SAT) are strong predictors of second-year persistence at California State University or University of California campuses.
  • Patterns for predicting early college outcomes with HSGPA, SAT, and SBAC persist across all racial/ethnic groups, for students identified as socioeconomically disadvantaged, and across high schools of varying quality.
  • There are observable differences in the magnitude of the variation in performance explained by the assessments when comparing student subgroups. In particular, the three assessments are less predictive of first-year college GPA at CSU and UC for low income students than for more advantaged students.
  • High school GPA as a predictor of college success results in a much higher representation of low income and underrepresented minority students in the top of the UC applicant pool, than do SAT or SBAC test scores. Using high school GPA with SBAC versus high school GPA with SAT also results in a more socioeconomically and racial/ethnically diverse applicant pool.

Note: The March 2019 report updates data and analysis from a March 2018 version.

Suggested citationKurlaender, M., & Cohen, K. (2019, March). Predicting college success: How do different high school assessments measure up? [Report]. Policy Analysis for California Education.