How Learning About One's Ability Affects Educational Investments: Evidence from the Advanced Placement Program
Working Paper 52
Publisher: Oakland, CA: Mathematica Policy Research
This paper studies how receiving a customized signal of ability affects individuals’ educational investments. In 2013, students who took the Preliminary SAT received a message in their results report that signaled their potential to succeed in Advanced Placement (AP) coursework based on their test scores. Survey data from students in the Oakland Unified School District revealed that the signal had informational value, leading students to revise their self-assessed academic ability and plans for AP enrollment, consistent with Bayesian learning. Using a regression discontinuity design, I found that the signal increased the probability of participating in AP classes by 49 percentage points among surveyed students, leading them to enroll in and pass about one more AP course the following year compared to students who did not receive the signal. However, students who were not surveyed for the study did not respond to the signal. The results suggest that ability signals can be a cost-effective way to increase educational investments, but may only work when they are made salient.
- Students who received a personalized message about their potential to succeed in Advanced Placement coursework were 49 percentage points more likely to participate in AP classes than students who did not receive the message.
- Students who received the message enrolled in and passed about one more AP course in the following academic year.
- At the end of the academic year, these students also became substantially more likely to take an AP exam and passed a higher number of exams, increasing their eligibility for college credit.
- However, students who were not surveyed for the study did not respond to the signal. The results suggest that ability signals can be a cost-effective way to increase educational investments, but may only work when they are made salient.
Additional Working Papers