APS Associate Superintendent Fails to Grasp Statistics

In today’s Journal, I read with an article by Diane Kerschen, an Associate Superintendent for Elementary Education with APS. There has been a great deal of discussion recently about Reid Nunn’s all-boy class at McCollum Elementary and whether single-sex education works. For the record, I went to an all-boy high school and did not find the single-sex aspect to improve things, but that was only my experience.

This brings me to Kerschen’s assertion that because the all-boy class had only 15 students in it (as opposed to 15), that the smaller class size is what caused the improved results. That is plain silly. The fact is that the connection between class size and educational achievement is not well established.

More importantly, extrapolating data from one class — one in which same-sex teaching was also implemented — defies any standard of statistical analysis. It’s called “sample size.” The fact is that you need to allow many different techniques to be tried in a controlled environment before jumping to such conclusions.

I’d like to see a great deal of experimentation — including smaller (and larger) class sizes (and a whole host of other schooling techniques) — and have relevant data collected and made available for analysis. Unfortunately, this is the very kind of thing that the teachers’ unions have attempted to stop in the past.