The effect of socially desirable responding (SDR) on the consensual validity of personality traits was studied. SDR was operationalized as the sum of items weighted by their respective social desirability values (Social Desirability Index, SDI), which could be computed for both self‐ and peer‐reports. In addition, the Balanced Inventory of Desirable Responding (BIDR) was used as a measure of SDR. It was shown that both self‐peer and peer‐peer agreement rose significantly for most studied traits when SDI was controlled in both self‐ and peer‐reports. BIDR was a significant suppressor variable in only one of the analyses involving Neuroticism. The SDI detected faking on personality scales somewhat better than the BIDR scales. It is argued that the SDI is a measure of evaluativeness of a person description, and that people agree more on descriptive than on evaluative aspects of a target's personality traits. Copyright © 2006 John Wiley & Sons, Ltd.