Two studies compared specific versus broad measures in linking personality with work behavior. In Study 1, 100 university students completed the 20 subscales of the Personality Research Form and an in-basket exercise scored on 16 distinct managerial behaviors. In Study 2, 335 market research field representatives completed the Hogan Personality Inventory, containing 41 specific trait scales organized into seven primary scales, and were rated by their supervisors on seven performance dimensions. In both studies, significant linkages between broad personality and criterion variables (e.g., factors) were explained by stronger relations among relatively few specific variables. Moreover, consistent with prior research, weak relations among broad measures obscured important linkages at the specific level, including several cases of cancellation (i.e., specific traits loading the same factor in the same direction correlated with criteria in opposite directions). Canonical correlations with appropriate shrinkage correction revealed notable improvements in criterion validity over inter-factor correlations and helped summarize the data while retaining the diagnostic advantages of specificity. Our findings are unique by demonstrating the value of specificity on both sides of the prediction equation. Implications for personnel selection are discussed. Copyright © 2003 John Wiley & Sons, Ltd.