I'll toss some gas on this one. One of my college instructors- I'll not say which one, but it was in the education component of my undergrad degree- was fond of saying "the research tells us". Research is a funny thing- it can be twisted quite easily to support a preconcieved notion, idea, or hypothesis. In other words, the research can be made to tell whatever the researcher wants it to- if they do not report the facts fully. I also noticed that the education research I looked at used for the most part psychology models to prove the accuracy of their statistics- generally only using a sample of 1000 students or teachers, or whatever. In a country where the students number in the millions a sample of 1000 anything, or even 10,000 isn't significant. My opinion- I think a lot of educational research is very flawed, and done mostly to avoid having to enter a classroom and teach.