Role theory (Kahn, Wolfe, Quinn, Snoek & Rosenthal, 1964) provided a framework for an investigation of moderators of MSA agreement. Specifically, this study investigated moderating effects of different types of actual role conflict (person-role conflict, role sender dyad conflict, and inter-sender conflict), frequency of behavior observation, and feedback seeking behavior (feedback inquiry and feedback monitoring) on agreement between self, peer, superior and subordinate assessments within a 360 degree assessment. Data from 350 assessors of 27 managers and 124 professional employees produced 839 assessments (86% return rate). An exploratory factor analyses of results (prior to full analyses) indicated five behavioral categories of competence would best represent the data. These factors were used in Moderated Multiple Regression analyses. Results indicated weak to moderate support for moderating effects. Results replicated trends in past research regarding self-peer, self-superior, and peer-superior dyads.