The assertion by Landy et al. (1980) that purer criteria for test validation would result from statistically controlling for halo in dimension ratings by partialing out a general impression rating was investigated by Barry R. Nathan and Nancy Tippins.