Socially desirable responding (SDR) on personality questionnaires is a phenomenon which can greatly complicate the selection process and stand between the organization and the recruitment of the best available candidates for the job. Of the many methods of reducing SDR, verbal and/or textual warnings of measuring SDR are certainly one of the most popular ones. Among those, warnings mentioning computational and statistical methods of measuring SDR are an exciting new prospect. In this study, we attempted to examine the effectiveness of one such warning, as well as compare participants' reactions to that warning with the reactions to one mentioning the use of individual items. The study involved 203 participants, which were randomly assigned to one of four situations where they filled out a personality questionnaire. In three experimental situations, we encouraged participants to respond desirably by simulating a job selection process using a short text, while in two of them participants were also warned that SDR was being measured by computational and statistical methods and by items respectively, then penalized. The group which was administered the computational-statistical warning, contrary to the hypotheses, did not give less desirable assessments of personality nor did they achieve lower results on SDR scales when compared to the other two groups with simulated selection. On the other hand, a hypothesis predicting equal reactions across groups was confirmed. However, the absence of a statistically significant difference between control and selection groups indicates that the simulated selection was not particularly effective. Given that the research hypotheses depended on a successfully simulated selection, the findings of the study are inconclusive. Nevertheless, we believe that the warning mentioning computational-statistical measurement methods still holds potential and warrants further research.