An investigation of the fakability of an empirically scored biographical inventory with variations in the subtleness and specificity of subject response set

The present research was an attempt to evaluate the effects of subtlety and specificity of subject response sets upon the fakability of an empirically keyed biographical inventory. The subject sample included 148 undergraduates from psychology courses at the University of Houston, of which 91 were male and 57 female. The first contact with the students involved the administration of a 57 item biodata questionnaire, along with the random distribution of four sets of written instructions to create the "fake" conditions. The second dispersion of the background form was done with a request for honest responses. Faking was defined as observed differences in the fake and honest means when responses were scored with a "salesman" key and a "manager" key. Contrary to the author's expectations, the results demonstrated that it was possible for the subjects to fake both keys under nearly all experimental conditions. In addition, faking under unsubtle conditions proved to be significantly more effective than faking under subtle conditions. Scaling difficulties made it impossible to evaluate the relative ease of faking for a more specific job as opposed to a more general position. Previous research had indicated that faking was of a more focused, rather than a generalized nature. Therefore differences were anticipated in the scores of individuals applying for a sales versus a management position when both sets of responses were scored with either the salesman or manager key. That no differences were found was attributed to the overlap between the two keys and the detrimental effects of an inappropriate sales stereotype. A comparison was made between the norms of the experimental subjects and those of a nationwide sample of applicants for the dealer sales representative position. There was no significant difference between the mean of the applicants and the mean of the most successful faking condition. However, the variability within the faking sets was significantly higher than the variability of the applicants. While not ruling out faking by the applicants, this led to the conclusion that there was evidence to indicate greater faking on the part of the students. Conclusions and suggestions concerning future research on biodata items were presented.