eric.ed.gov har udgivet:
This paper asks whether placement recommendations for a developmental math course at an elite public institution impact students’ future academic performance, course-taking, and college outcomes. Researchers use these specific outcomes to measure whether developmental courses help students develop the skills necessary to succeed in college, inspire them to take different courses, and help them graduate or persist in college. The study examines the ways in which instructor characteristics can drive these outcomes, and whether instruction at this university in a program for low-achieving students and particularly underprepared low-income, first-generation, and underrepresented minority students achieves its goal of reducing achievement gaps. This informs specific course and instructor policies to help underprepared students in their first semesters in college. The research setting is an elite public institution with a large number of students, with an undergraduate enrollment of over 20,000. Researchers gathered the administrative formulas the Office of the Registrar used to calculate student recommendations, and accessed student and instructor administrative data. A regression-discontinuity (RD) framework was used to overcome the hurdle of making estimations since students who take developmental courses are different from students who do not. With the RD strategy, researchers can measure the causal impact of receiving a recommendation for developmental courses. A linear regression was used on the selected sample to estimate the causal impact of the recommendation. Researchers included the index distance above and below the cutoff as additional controls. Using the linear regression and RD framework, results show that students who receive a Definite Recommendation to take Pre-Calculus are more likely to take Pre-Calculus compared to students who receive only a Tentative Recommendation. This shows these recommendations have a significant effect on students’ course-taking decisions. Figures are appended.