“It’s not about what the instructor covers, it’s about what the students uncover” (Dan Levy)
I was fortunate to learn from many great teachers as a student and a teaching assistant not too long ago. No one impacted my teaching more than Dan Levy, a senior lecturer at Harvard Kennedy School who I TA’d for starting in fall 2019.
One of Dan’s pedagogical adages that has always stuck with me is his caution to lecturers:
It’s not about what the instructor covers, it’s about what the students uncover
In statistics, there is always a lot I want to teach. For example, in a class on interpreting regression coefficients, I would think there are so many important topics to cover: e.g., categorical predictors, binary outcome variables, interactions, logit vs. linear regression, control variables, fixed effects — all of these are important! Not to mention the many examples I could use.
But what this provocative claim suggests is that ultimately, it does not really matter if you covered 5 things vs. 10 things vs. 20 things — no matter their objective importance. What matters is what the students retain after the class ends, and furthermore, people retain more information when they uncover it themselves.
For example, say I want to teach that randomization of treatment assignment ensures that potential outcomes are balanced in expectation. Having students uncover this key concept in causal inference themselves is a different matter. Here is one exercise I tried: Randomly assign the students into groups A vs. B live, and then calculate the proportion of seniors in each group. Hopefully, the proportions are similar across groups. Then you can use that exercise to drive home the intuition that any prior potential outcome should be balanced, too.
I also teach introduction to Congress at Yale. In a non-statistics class like this, using simulations and games for teaching congressional negotiations can be quite helpful, because students uncover notions of agenda power, logrolling, time management, etc., themselves. (Frances Lee and Robert Oldham developed such a course at Princeton, and Keith Hennessey teaches a wildly popular MBA class at Stanford with this, called “I’m Just a Bill”).
This does not mean, though, that every lecture has to be some kind of group activity (those who know me personally probably can tell I wouldn’t be a group activity die-hard myself if I were a student). A traditional combination of pure lecture and homework later also can lead students to “uncover”. Doing the homework allows students to uncover the material. There is a particular type of joy when you are solving a problem after class (even during the final exam!) and you realize you can now solve something that was incoherent before. I take that to be a part of uncovering too.
On net, this adage can make class fun, but it is also sobering — just because I managed to pack a lot of important content into a lecture, or delivered it correctly, none of that matters if my audience does not retain it. It’s not an exaggeration to say that it has changed how I see course design 180 degrees.
I’m grateful to Dan for instilling this concept in me. Many of the things Tina or I post here are ideas I borrowed from being in his class.
[Note: Dan told me that he did not come up with the exact quote. He had heard the expression before but cannot remember who said it.]
P.S. Another TA who I served with in Dan’s class took extensive notes and later wrote a book about his pedagogy, titled Invisible Learning — a very good book I learned from. Unlike other pedagogy books, this one is written by an observer in the classroom, not the popular professor themselves, which I think helps a lot (and is why I wanted to include an observer’s notes, like Tina’s, in this blog).