To expand our knowledge of personality assessment, this study connects research and theory related to two common selection methods: assessment centers (ACs) and personality inventories. We examine the validity of personality-based AC ratings within a multi-method framework. Drawing from the self-other knowledge asymmetry model (Vazire, 2010), we propose that AC ratings are suited to capture personality traits that are observable in social interactions, whereas other methods (i.e., self-ratings) are useful to assess more internal traits. We obtained data from two personality-based ACs, self- and other-rated personality inventories, and supervisor ratings of job performance. Confirmatory factor analyses indicated that personality-based AC ratings reflected the Big Five traits. Consistent with the self-other knowledge asymmetry model, AC ratings of more observable personality traits (Extraversion, Agreeableness, and Intellect/Openness) were correlated with inventory-based measures of these traits. AC ratings demonstrated incremental validity in predicting job performance over inventory-based personality measures for some traits (including Agreeableness, and Intellect/Openness) but self-ratings also demonstrated incremental validity over AC ratings (for Conscientiousness). This implies that different personality measures capture unique information, thereby complementing each other. Yet, AC effect sizes were modest, suggesting that running personality-based ACs is advisable only under specific circumstances.