Great article.

The history of class size reduction is a great lesson in why science cannot be the end-all-be-all of policymaking.

Why did America go all in on class-size reduction? Because of a famous study known as Project STAR that used genuine randomization to show that students in smaller class sizes did better than students in large class sizes. For its time, it was considered a gold standard in scientific evidence because there was a true randomized control group.

However:

• Just because class size reduction was demonstrated to help, that doesn't mean it's the best use of resources. Teacher quality wasn't really tested because we don't have good measures of teacher quality. Just because something is difficult to measure scientifically doesn't mean it isn't valuable. And that's the huge downside of data-driven policymaking, which is that it often ignores hard to measure factors that are potentially hugely important.
• Effective class size reduction wasn't scalable. Project STAR was done on a small scale and with a relatively homogenous student population. To try and scale it up would require hiring a lot more teachers, which means digging deeper into the pool of less qualified, less experienced potential teachers.

I basically agree with the author that teaching needs to become a prestigious profession again. It should be at least on par with doctors and lawyers in terms of desirability, if not pay.