According to previous studies based on the Social Relations Model (SRM), most of the variance in observer reports of personality is perceiver and relationship variance, and not much is target variance. However, most SRM studies have employed short adjective scales instead of personality questionnaires. Results based on the HEXACO-PI-R in family and work groups showed high levels of consensus (target variance) and self-other agreement for all traits and, except for Honesty-Humility and Openness to Experience, low levels of generalized rater bias (perceiver variance) and of assumed similarity. Additionally, intraclass correlations suggested a 'group personality' for some traits. The findings suggest that the use of personality questionnaires in Social Relations Analyses may promote higher estimates of consensus in personality judgments. Copyright © 2010 John Wiley & Sons, Ltd.