ABSTRACT
Bias is easily alleged than demonstrated; it is easier to imagine the various kinds of third variables that may bias scores than to show their influence. Cultural bias can be expected across countries in multinational organizations, but it is less certain for subcultures, within a single national experience where the same media of mass communication, give subcultures much in common despite some profound differences. Unfair discrimination exists "when persons with equal probabilities of success on the job have unequal probabilities of being hired for the job". Analysis of bias tries to ferret out instances of unfair discrimination. Distributional differences may stem from true differences or from systematic sources of measurement error related to group membership. In criterion-related validation, the criterion is reliable, valid, and free from third variable biases. Evidence of valid measurement of the intended criterion construct is the sort of evidence most appropriate; a major question in psychometric validation is whether extraneous sources of variance influence the measures.
