Good question. The government seems to think that random assignment to experimental groups is required and we certainly need more of those types of studies, though it should be clear that education studies cannot be done in situ and meet this requirement very often. That said, I think there should be more willingness on the part of educational administrators (and parents) to cooperate in studies.
Many large "studies" funded by the government are non-experimental. Such studies have their place in finding relationships among among variables and constructs of interest. But, I think that too much emphasis is placed on such findings, especially in suggesting causal relationships and because educators, politicians, and the public tend to assume that educational interventions can change any variable or construct. For example, most people associated with education believe that "motivation" is important for a student achievement, but they seldom stop to ask if we can actually change motivation in a way that can have clear and lasting affects on a student's formal learning.
Many people associated with education or the training of teachers are either suspicious of scientific investigation of instruction or claim outright that science doesn't work for instruction. Others claim that only educational practice can "inform" educational practice. These view stem from the non-scientific background of many in education. They have little understanding of scientific method and scientific reasoning and instead rely on philosophical attacks on science to justify their distrust of scientific analysis and support their manly ideological views about how learning works and education should work..
I would argue that science should be informing educational practice. A major outcome of scientific method is that it "works" for achieving specific goals. Controlled scientific investigation results in real understanding and progress because scientific method seeks to control the intuitive biases we have about how learning works. Philosophical arguments rarely lead anywhere useful when compared to scientific investigation. I suggest that in addition to seeing teachers as nurturers (stand-ins for parents) that we also much see them more as engineers--professionals applying science in practical and useful ways.