| Rebecca Zwick
University of California, Santa Barbara
Do SAT Scores and High School Grades Predict College Grades and Degree Completion? A Study of Four Language/Ethnicity Groups
FINAL REPORT
We investigated the appropriateness of standard college admissions criteria for 4 student groups: Hispanic students whose first Ianguage was Spanish, and Hispanic, Black, and White students whose first language was English. The students were part of the High School and Beyond survey and were high school sophomores in 1980. We examined the degree to which SAT scores, high school grade-point average (HSGPA), and other factors predicted first-year college grade-point average (FGPA) and college graduation for these groups.
When linear regression analysis was performed to predict FGPA for all students combined, HSGPA and SAT score jointly explained 22% of the FGPA variance. As is typical in such analyses, HSGPA was a stronger predictor than SAT score. Adding language/ethnicity information to the regression equation raised the proportion of explained variance by only a small amount, to 22.5%. However, the effect of membership in the black/English group was statistically significant, and the effect of membership in the Hispanic/English group, though not significant, was sizeable. For these groups, applying the combined-group regression equation would lead to predicted FGPAs that were higher, on average, than the FGPAs actually attained. Overprediction was even more severe when only HSGPA (without SAT) was used to predict FGPA. In previous research, overprediction has often occurred for Black and Hispanic students. Surprisingly, the selectivity of students' colleges did not contribute substantially to the prediction of FGPA. Gender also proved not to be a statistically significant predictor, although supplementary analyses revealed a greater tendency toward overprediction errors for men than for women.
Survival analyses showed that the White/English group had a substantially higher graduation rate than the remaining groups. After five years of enrollment, 82% of the White/English group had graduated, with rates for the other groups ranging from 50% to 66%. HSGPA was found to have a statistically significant influence on graduation in the White/English group; SAT had a significant effect in the Hispanic/English and White/English Groups. Institutional selectivity and gender had little effect on graduation. Of particular interest in this study were differences between the Hispanic/Spanish and Hispanic/English groups. One difference was the existence of some overprediclion of FGPA for the Hispanic/English group, but not for the Hispanic/Spanish group. This result indicates that, compared to Hispanic/English group members with similar academic backgrounds, Hispanic/Spanish group members tend to receive higher college grades. In fact, although average SAT score was lower for the Hispanic/Spanish group than for the Hispanic/English group, the Hispanic/Spanish group had a higher average FGPA. A second difference in analysis results for the two Hispanic groups is that both FGPA and graduation were more strongly associated with HSGPA and SAT scores in the Hispanic/English group than in the Hispanic/Spanish group.
The disparities in achievement patterns between the two Hispanic groups demonstrate the value of taking language background into consideration in educational research. This factor is likely to be at least as important as ethnic background in facilitating the understanding of students' academic success in coIlege.
Back to Funded Research Grants Page |