I think this is kind of a backwards way of looking at things. All scientists go through a period of apprenticeship and very little of what it is to "do science" is written down. Textbooks contain descriptions of phenomena and experiments. There are protocols for performing common tasks. But there really isn't an extensive literature on "how to be a scientist." But I don't see why we should expect it to be communicable anyway. Why should we be able to provide a casual description of what people do? I think this expectation relies on the fallacy that explicit language is a mere translation of some internal "mentalese." Yet there's no reason to expect that language can capture thought or even behavior on anything but a technical level (i.e., a level not immediately useful to communicating practice). And even if we could express thought and behavior appropriately there's no reason to expect us to be adept at turning verbal descriptions back into thought and behavior. If the cognitive and behavioral sciences do manage to inform pedagogy I expect it will be in the form of providing better hands-on experiences and better apprenticeships rather than finding ways to express these ideas in textbooks.