6. Conclusion
The use of overall grade point averages is ubiquitous for determining access to college. Guaranteed admission policies explicitly use class rankings and individual colleges and universities use GPAs at their own discretion. Further, many colleges, universities, and scholarship organizations have eligibility rules based on overall GPAs. However, the empirical evidence suggests that performance in later grades is more predictive of success in college and the labor market than performance in 9th and 10th grades. Students who exhibit their best performances as high school juniors and seniors are less likely to drop out and are more likely to graduate on time than students with identical GPAs but whose best performances occurred as freshmen or sophomores. Examining the predictive nature of trajectories and non-academic courses suggests that persistent changes in effort provide at least a partial explanation for this phenomenon. The results have important implications for policy. Distributing admissions slots and financial support in a way that places greater weight on performance in later grades would result in the selection of students who are more likely to succeed. Further, abstracting from competition across colleges, changing deadlines to allow the consideration of more grades from students’ senior years would result in the selection of stronger candidates. Of note is that for the population examined, disaggregated GPAs by grade level are the dominant predictor of college completion and labor market success, exceeding the contribution of entrance exam scores. Further, switching to a system that places greater weight on performance in later years does not appear to negatively affect diversity, as minority students do not exhibit disproportionately negative performance trends during high school.
