We taught a class using Zoom yesterday. Here’s what we learned.

Like many schools, Columbia is moving to online teaching for awhile to minimize the potential for virus transmission, so Merlin, Pablo, and I used Zoom for our class on applied regression and causal inference. Yesterday’s topic was chapter 11, Assumptions, diagnostics, and model evaluation.

Some things about the class went well, some things could be improved.

I’m sharing our story in the hope that some of you can learn from our experiences when holding videoconferences or remote classroom teaching.

– To get in the right frame of mind, I dressed for work and went to my office as usual and set up the videolink from there.

– Getting Zoom working wasn’t too difficult. But the first Zoom link didn’t work, so we had to try again and then email all the students to let them know about it.

– Turnout was excellent. The students had the time already available, of course, and they’re all online. Just about all of them joined in.

– Getting started was difficult. In retrospect, I wish I’d prepared a start-of-class activity to avoid the initial awkwardness of the online interaction, in the same way that I prepare first-day-of-class activities to avoid the waste of time that often occurs in the first meeting of the semester.

– After a few minutes, things went more smoothly. But I spent just about all the time talking. I think I must have been shouting at my computer, because when the hour and a half was over, I had a bit of a sore throat.

– What did we actually do? This wasn’t too hard, because much of the course already exists online. Students submit their homeworks online, so they didn’t need to come to class to hand anything in. Also, before each class all students are required to contribute to a Google doc which has space for questions about the homeworks and the readings. Students can type in their questions on this document or respond to other students’ questions. There are about 20 or 25 students in the class, which is enough that we have lots to talk about but we’re not overwhelmed. I guess in a larger class you’d want to do this sort of thing separately for each section.

– Anyway, the remote class went pretty smoothly because I just answered the questions one by one. We set up a separate shared Google doc which served as a scratchpad for the discussion. There I typed formulas, prototype R code, etc., and students put in their questions from the discussion.

– The main difficulty was that it was mostly me talking. Me talking is better than nothing, but I don’t think it’s the best way for the kids in the class to learn. They learn by doing. In a live class I can walk around the room and look and listen to what they’re doing as they work in groups. Online is tougher.

What will we do next?

– For next class I plan to prepare some activities ahead of time—I do this anyway for the course—but this time I will type the activities into a shared sheet, rather than saying them.

– How can students work in pairs? They’ve all networked, so we paired them up and asked each pair to set up their own shared google doc, which they can use as a scratchpad when working together. Then we can screenshare some of their solutions.

– I’ll also ask Merlin and Pablo to extract some relevant pieces from their homework solutions and then we can screenshare these and discuss them in class. I feel like we’ve been spending too much time lately discussing the readings and not enough time on the homework. And on drills.

I’ve probably missed a few things. So far I’ve been happy with the remote teaching experience. The challenge will be keeping the students engaged. I have a horrible feeling that half of them are texting or reading the news on the web while half-listening to the class. I appreciate students’ patients with our technology struggles, but going forward I want them to be even more engaged. I don’t want to be wasting their time and attention. Any suggestions?