Rater and technique contamination in criterion ratings

Abstract

Rater contamination may occur when the same rater makes both prediction and criterion ratings; technique contamination may occur when both the prediction and criterion ratings are on the same form. 400 Army officers provided ratings on a graphic scale and two versions of a forced-choice scale. Each officer rated 20 of his fellow officers on the graphic scale, and eight days later rerated two of them on the graphic scale and one of the forced-choice scales. Intercorrelations for the same raters on the same (graphic) form were .69-.82. Inter-correlations for the same rater on different forms were .52-.57. For different raters, inter-correlations were below .3, for same or different forms. It is concluded that the rater is the principal source of contamination. (PsycInfo Database Record (c) 2025 APA, all rights reserved)