Read on Twitter
I just finished teaching a 60-student asynchronous undergraduate course in Health Economics. Here are some thoughts about teaching, learning, and grading in the Brave New Online World
1. Requiring an empirical term paper is an effective way to engage and motivate students (1/n)
1. Requiring an empirical term paper is an effective way to engage and motivate students (1/n)
2. Hands-on via 1-on-1 office hours and with the data — whether in R, Stata, Excel, or Sheets — is a clear winner in several important dimensions. Students are more motivated. They understand how to make progress. They destroyed any grading curve I would have had. (2/n)
3. When creativity is on the table without a lot of guardrails, you get some pretty impressive buy-in from students. As you might imagine, term papers in health economics in Summer 2020 were heavily weighted toward topics of racial and socioeconomic inequality. (3/n)
But the amazing thing is that even though the Zeitgeist will push people to explore a set of topics, everyone does it differently, with their own skills and voice, which they work on. It's not as hard to grade as one might initially expect, because there are no guardrails (4/n)
4. But when there are no guardrails, a grading rubric that checks for specific concepts and skills is virtually impossible. Rather, a rubric in this situation more naturally outlines qualities of a good paper, like clarity and reasonable speculations based on the data. (5/n)
This ends up rewarding students for *relative* progress, not absolute standing. "An A for effort," if you like, but of course the efforts have to be successful. (Spoiler: they were.) This can create some potential challenges depending on how the course fits the curriculum (6/n)
I'm hoping that as an upper division elective, this course isn't heavily relied upon by department staff to determine GPA eligibility for honors, for example. Ironically, a course like this has prepared students for exactly what honors would require: independent work. (7/n)
5. In retrospect, I think I was overly concerned with how students would react to one-shot, timed online quizzes in this environment. I got what I wanted: no drama about the scored quizzes at all. But I also didn't get what I would have liked: some gradient in outcomes (8/n)
To be clear, what I did was put 50% grade weight on an empirical term paper in health economics, and 50% on 5 repeatable online quizzes in this short session. I wish I'd set the quizzes to be more unforgiving. Maybe that would have caused trouble, but it's my gut now. (9/n)
6. Ultimately I invested a lot of time directly into students with the term paper, 26 hours in Zoom 1-on-1's across 31 unique students. That paid off, and it also was a real cost. True, I spent no time trying to police cheating, but I spent a lot of time. Maybe there is (10/11)
A more cost-effective way to configure everything, but I'm happy with the learning outcomes. These students, who may be highly select, really impressed me with what they pulled off. Or was it a true treatment effect of my 26 office hours? Well, that probably helped. (11/11)
