Author ORCID Identifier

Date of Award

Spring 5-13-2022

Degree Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Dr. Thomas Mroz

Second Advisor

Dr. Daniel Kreisman

Third Advisor

Dr. Melinda Pitts

Fourth Advisor

Dr. Jonathan Smith


Increasing participation in STEM is a major policy goal, yet not much is known about how exposure to math and science in high school influences future STEM attainment. To examine this connection, I develop a dynamic structural model of high school curriculum and college major choice using data from the Educational Longitudinal Study of 2002. I examine the effectiveness of high school STEM exposure on resolving student uncertainty about academic aptitude and derive the impact of raising graduation requirements in math and science on educational outcomes and student welfare. I find that, while students are initially uncertain about their abilities at the beginning of high school, much of this is resolved upon completion. Therefore, counterfactual simulations that increase the amount of information contained in high school STEM courses do little to affect students' ultimate educational choices. Raising graduation requirements in math and science primarily boosts post-secondary attainment in college non-STEM for median individuals. I also find suggestive evidence that higher requirements in mathematics increase high school dropout rates for men. These effects may vary across the ability distribution, with effects on dropout concentrated among lower ability individuals. Because uncertainty is resolved quickly in high school, these effects are driven by changes in course-taking patterns rather than changes in beliefs. Ultimately, my findings do not rule out only modest negative welfare effects of these policies, with (imprecise) point estimates of the reduction in expected lifetime utility from increasing graduation requirements in mathematics from 2 to 4 years to be around 4.8% for men and 4.0% for women.