We estimate the causal impacts of a new Advanced Placement (AP) course on participation in CS education. AP offers college-level courses, exams, and potentially college credit to U.S. secondary students. Historically, female, Black, and Hispanic students have been significantly underrepresented in AP Computer Science [1]. A new course, CS Principles, was launched in 2016-17 to attract a more diverse group of students. Compared to the Java-focused preexisting course, CS A, CS Principles features a broader framing of CS, a creative project component in assessment, and flexibility for teachers in choosing programming languages. We assemble a dataset of annual AP exam-taking and course offerings at 294 schools in Massachusetts over fifteen years, using publicly available data from the Massachusetts education department and the College Board. Using an interactive fixed effects counterfactual estimator, we estimate that offering CS Principles increased a school’s chance of having any students take AP CS exams by 25 percentage points and its number of CS exams by 24 [3]. Gains were larger for Black and Hispanic and female students. We find some substitution effects on CSA, as suggested by previous research, but none on other AP STEM exams [2]. Our preliminary results suggest that elements of the CS Principles launch, including course design, curriculum development, marketing, and teacher training, can inspire other large-scale initiatives promoting engagement and diversity in STEM education. In this Lightning Talk, we hope to outline our work, gather feedback, and connect with potential collaborators interested in quantitative policy analysis in CS education.