Categories: GSI Online Library, Teaching Effectiveness Award Essays
By Zhixiang Su, Sociology
Teaching Effectiveness Award Essay, 2025
Teaching quantitative methods to students primarily trained in qualitative research is never easy—especially as a first-time Graduate Student Instructor. Last year, I began teaching a two-semester statistics sequence in the sociology department to PhD students. Many students took these classes primarily to fulfill mandatory program requirements and approached statistics with limited initial interest. What makes the learning environment even more challenging was the imbalance in the classroom, which is unfortunately the case for every incoming cohort. Most students without prior training in math or programming struggled to keep pace, while others were already comfortable with the material. Having been a student in the same class myself, I am fully aware how such a gap can deepen math anxiety and self-doubt, especially among students from disadvantaged backgrounds.
After several unsatisfying lab sessions, I realized that the problem lay in what Paulo Freire described as the “banking model of education”—a structure that statistical class typically follows, where instructors “deposit” dense blackboards of mathematical notation and expect students to passively memorize and reproduce them for exams. To break this cycle, I began experimenting with games and critical dialogue as tools to support statistical learning. First, I designed interactive games to introduce new statistical methods, transforming the classroom into a kind of mini-theater to help students overcome math anxiety. For example, when teaching instrumental variables (IV), an econometric method, I started the session with a scenario: students ran a currency exchange shop, converting dollars to euros and then to Japanese yen. I then drew parallels between this process and IV, treating variables as different currencies and regression coefficients as exchange rates. This helped students build intuition for calculating effect sizes in the model. When teaching multivariate regression, I invited students to participate in a physical demonstration where we projected data points onto a “regression plane” using balls, flashlights, and cardboard—visually illustrating concepts such as dimensionality, best fit, and residuals. Simulations and gambling games proved invaluable for exploring probability theory. By leveraging the power of play, I helped students approach complicated statistical models not as intimidating math, but as tangible, interactive “toy models” they could pay with and reflect upon, laying the groundwork for deeper critical thinking.
Second, I incorporated critical dialogue—an approach commonly used in humanities and theory courses—into the statistics classroom. Together with the professor, we introduced case studies that applied statistical methods to contentious social issues, prompting students to draw on their lived experiences to question the assumptions, results, and implications of statistical analysis. For instance, we analyzed statistical reports used in court filings on racial bias in Harvard’s admissions process, encouraging students to interrogate how the implicit assumptions of different models could be mobilized to support or undermine either side. In another session on causal inferences, we explored why qualitative studies using the same dataset reached conflicting conclusions about whether children of same-sex couples experience better health outcomes, and reflected on how such publications shape political discourse and struggles. Through these dialogues, statistical learning became a form of emancipatory inquiry. The instructor was no longer the sole authority but became part of the collective learning process with the class. When students debate and adjudicate statistical evidence through their own experiences and political standpoints, they begin to demystify the seeming objectivity of statistical methods and come to see it as a powerful tool in the struggle of social change. More importantly, by learning to use statistics not as passive recipients, but as critical researchers and reviewers, students became more attuned to how the misuse of statistics can reinforce inequality and domination.
These efforts ultimately paid off. By the end of the series, students reported feeling significantly more confident using statistics. Several students who initially came to my office hours questioning whether they would ever use statistics in their research, later became enthusiastic about turning their final papers into publications concerning pressing social problems. The impact of this critical pedagogy became even more evident when I saw my students asking critical methodological questions at departmental job talks, demonstrating both confidence and the ability to engage with cutting-edge research. I have therefore managed to combine Paulo Freire’s pedagogy of the oppressed to statistical learning, a model that I wish to replicate when teaching statistics to social science students.