The Psychopathic Personality Inventory (PPI; S. O. Lilienfeld & B. P Andrews, 1996) is a self-report test that has shown considerable promise as a screening measure for psychopathy. A current limitation of the PPI is that no data exist regarding the impact of response sets such as positive impression management. Although the PPI contains a validity scale (Unlikely Virtues) designed to identify response biases such as "faking good," its utility has not yet been assessed. In this study a repeated measures analogue design was employed in which 186 respondents completed the PPI both under standard conditions and with specific instructions to create a favorable impression of themselves. In the "fake good" condition, participants were able to appear significantly less psychopathic, with those who obtained higher scores in the standard instruction condition showing the largest decreases in their PPI scores. Receiver Operating Characteristic analyses indicated that, although the Unlikely Virtues scale significantly differentiated between "fake good" and honest protocols (area under the curve = .73), a considerable number of misclassifications occurred. The clinical and forensic implications of these findings are discussed.