Practitioners used different strategies to prevent faking on self-report inventories such as forced-choice formats, subtle versus obvious items, respondent warnings, corrections based on social desirability scales. The paper presents an indirect strategy to reduce faking in self-report personality testing (the strategy was called constrained ethical commitment because applicants were indirectly forced to sign a declaration whose content highlights the importance of honest answers). The sample consists in 91 male students who applied for Police Academy and completed different questionnaires for measuring the Five Factor Model of personality and response validity. The subjects formed two groups: (a) the control group filled in the questionnaires without special instruction and (b) the experimental group signed the ethical commitment form before answering. The main idea was that the experimental group produces significant lower feigning indexes than the normal applicants do and four indicators have been analysed to test this hypothesis: the scores of the validity scales, the scores of the FFM, the percentage of invalid profiles in both groups, the relation between validity scales and the FFM. The results showed that constrained ethical commitment has no effect on the scores of validity scales or the means of the FFM, but it reduces the percentage of invalid profiles (Cohen's d=1.54) and also the relation between dissimulation and neuroticism (from .50 to .01). This strategy might be useful especially in settings where ethics plays a great role and the candidates should provide evidence for such behaviour.