Strikes me that that there's a serious specification issue and
potential tautology involved in asserting that the «top one - third of students» get the «best one - third of teachers,» especially when the gauge of teacher quality is student performance.