presented in the literature as sufficiently robust to
warrant transfer to practice is exclusively of the
variance kind; specifically, it concerns knowledge
of "what works," usually based on comparisons of
performance in large samples or on controlled lab
experiments.