Factors influencing two types of congruence in multirater judgments

In two studies of 360° judgments, it was shown that congruence measured in terms of intercorrelations between ratings by different people (congruence‐r) and in terms of the discrepancy between those ratings (congruence‐d) were empirically independent of each other. Congruence‐r was greatest, and congruence‐d was smallest, between judgments made by a target person and his or her supervisor, rather than by subordinates or peers; the differential observability of rated behaviors predicted congruence‐r but not congruence‐d. Variations in both forms of congruence were found to arise from the source of a judgment, and from a target person's age, gender, cognitive ability, and certain personality attributes. Over‐ and under‐rating associated with a personality or another third factor was shown to derive from substantially different correlations between such a factor and the two separate judgments under consideration.