Teaching Statistics: A Quasi-Experimental Analysis of Ensuring Undergraduate Student Success by Teaching and Exam Format Choice

Statistics is often viewed as a challenging and unpopular subject among political science students, yet its significance is growing with the increasing relevance of big data and digital social sciences. This study examines the impact of course design in statistics education on student behavior and academic success by comparing two similar social science programs within the same institute at a large German research university. Program A employs a traditional lecture-based approach with a final written exam, while Program B integrates lectures with project-based learning in additional seminars and assesses students through a project report as exam format. Utilizing administrative data from two student cohorts (n = 282 for Program A and n = 328 for Program B), we assess exam registration, participation, and grades as key indicators of student success. Our findings reveal that students in Program B register and participate in exams earlier and achieve higher grades than their counterparts in Program A. These results underscore the value of an engaging, project-based course design in minimizing procrastination and enhancing the effectiveness of statistics education. However, implementing such active learning formats may incur higher financial costs or more internal coordination. Our study emphasizes the need for educational institutions to reconsider traditional teaching methods and highlights the benefits of fostering an interactive, student-centered learning environment in statistics education.