[JustinSkycak]

I would agree there are a lot of problems with a lot of education research. Many purported findings do not replicate or are otherwise impossible to replicate.

However, there are also many findings that are actually legit. As you say, they're rare, but there are enough of them to paint a surprisingly complete picture when you pull them together.

Discussed at length a couple months ago here: https://news.ycombinator.com/item?id=40348986