Failure to consider errors of measurement when using partial correlation or analysis of covariance techniques can result in erroneous conclusions. Certain aspects of this problem are discussed and particular attention is given to issues raised in a recent article by Brewar, Campbell, and Crano. (Author)