The purpose of this study was to establish the construct validity of the NEO PI-R personality measure when used for high stakes employee selection purposes. Based on extant research from industrial/organisational psychology, social psychology, and behavioural economics it is argued that deliberate impression management, or faking good, by job candidates in high stakes selection contexts can occur. This can be regarded as a form of moral hypocrisy. Theoretical research showed that moral hypocrisy occurs in ambiguous contexts in the absence of reminders of moral standards. It was hypothesised that the use of a formal warning would eliminate or minimise faking good by participants in a field study of job applicants in a high stakes contexts, thereby allowing construct valid inferences to be made about the participant’s personality traits. To test this hypothesis a formal verbal warning about measures included in the assessment to detect deliberate impression management was given to the participants. They completed the NEO PI-R as part of the battery of tests used in the selection process for middle and senior management positions in a range of organisations. A bespoke impression management measure, based on a widely used measure used to detect deliberate impression management, was included in the battery of tests. A second field study sample was used to validate the findings of the managerial field study. Using confirmatory factor analysis the results showed that faking good was minimised, but not eliminated. Monte Carlo simulations showed that it was still possible that participants, who faked good in spite of the warning, could be selected from a short list of job applicants. The use of the bespoke impression management measure was shown to be of benefit in minimising bias and unfairness arising from the use of the personality measure when selecting a candidate from a short list.