In this Charles Murray article (alt. WSJ subs. link) returning yet again to his favourite theme -- the consequences of innate (and, according to him, inheritable) differences in intelligence -- he never cites a single specific source for his claim that a dominant determinant of educational outcomes is a one-dimensional, rankable individual characteristic called 'g.' He also wilfully confuses the propensity for academically tracked secondary education, which does seem to vary across students, with the propensity to acquire basic reading and numeracy skills -- which he seems to be arguing may not exist among 40 percent of the primary student population.
UPDATE: More via Brad DeLong. And [2nd update], this Niall Ferguson tirade from the Sunday Telegraph has points of interest in between the polemics; note in particular his observation that low measured educational attainment seems be an English-speaking country phenomenon. Note clear what Murray would make of that one.