The American Cancer Society recommends that everyone get a colonoscopy at age 50 and then every ten years after that. It is an effective way for for doctors to identify colo-rectal cancer in its early stages, but it is usually a painful procedure. Physicians insert a camera into the rectum and push it pretty far up into the colon looking for potential problem areas.

In one of my favorite academic studies, researchers randomly divided patients who were about to get colonoscopies into a control group and a treatment group. The control group got a standard colonoscopy. The treatment group got the same procedure, except they didn’t just extract the camera when they were done. Instead, they pulled it almost all the way out and then just left it there for an extra minute. This part of the procedure wasn’t pleasant, but it was clearly less painful than the rest of the process. According to Wikipedia, the real pain from the procedure comes from having to inflate the colon with air in order to do the inspection.

Objectively, the control group’s procedure was better–those patients didn’t have to deal with an extra minute of a camera in their bowel. Their perception of the experience by study participants was the opposite: Average pain levels reported by patients after the procedure were significantly lower for those in the treatement group. Their overall impression of the procedure was weighted heavily toward their experience at the end.

I see the same thing when I teach. The body of the lecture is really important–That’s where the bulk of the learning happens, but the end of the lecture is what often determines students’ impression of the whole. This can be terrific when you’ve covered some tough material or hit some rough patches in the middle and you can finish strong.

On Wednesday I walked into my class with a great lecture prepared. It started with interesting substantive questions and together we developed the tools of hypothesis testing to answer them. The first hour and five minutes were smooth, interactive, and I thought very well-paced. Then I realized I had ten minutes left to explain how testing in small samples worked, teach them the corresponding commands to do tests using Stata, and then preview the lab that was coming on Thursday or Friday.

Just when my students were starting to disengage and think about what they were going to do after class, I cranked up the pace, talking and flipping slides faster and faster. Even in the moment I knew it was wrong. My initial plan was too ambitious, but that’s going to happen from time to to time. The big mistake was not adjusting on the fly. I didn’t have to cover small samples right then. That could wait until next week. I didn’t have to cover the Stata commands–I could have emailed them or recorded a short video that they could watch before they got to lab.

Messing up the last ten minutes didn’t ruin the whole lecture. I’m convinced that my students learned a lot in the first hour. It did disproportionately affect their impression of the class, and that’s a shame. Lesson learned.