Investigated the effect of three different variables on agreement effects in p-o-x triads. The subjects were 740 male and female undergraduates at the University of North Carolina. Experiment 1 found that assumed similarity modified the agreement effect—but only on cognitive rating scales (expectancy, consistency, stability). Experiment 2 found that assumed knowledgeability of o regarding × modified the agreement effect on both affective (pleasantness, harmony) and cognitive (expectancy, consistency, stability) scales. Experiment 3 found that a manipulation of o to p liking that was either consistent or inconsistent with p and o agreement modified the agreement effect primarily on the affective scales.