From the point of view of variance reduction, the adjustment of the 's solely by seems very little different to that obtained by adjusting fully by and . We can examine this suggestion, and the determine the effects upon expectation, by considering separately the effects of adjusting by the two-hour measurement, , in addition to .