Doctor of Philosophy (Ph.D.), Bowling Green State University, 2019, Psychology/Industrial-Organizational

This study's main argument was that faking on personality tests is a function of intelligence, personality (adaptive and maladaptive), and values. I expected that highly intelligent fakers would successfully fake the job-relevant traits for the job they were either already holding (incumbents) or have been applying for (applicants). I used archival data provided by Hogan Assessment Systems, and respondents were either applicants (N=1073) or incumbents (N=793) within the managerial job family. All respondents had taken the Hogan Personality Inventory (HPI), the Hogan Development Survey (HDS), the Motives, Values, and Preferences Inventory (MVPI), and the Hogan Business Reasoning Inventory (HBRI). I used the method of mixedmodel item response theory with covariates (MMI-IRT-C) and the statistical software Latent Gold to estimate latent classes of respondents on the HPI items. To isolate the latent classes, I entered HBRI's raw scores and the assessment reason (applicants or incumbent) as covariates and used the items of each HPI's main scale as observed indicators of these classes. I extracted between two and four latent classes of respondents for each of the seven HPI main scales. Using various indicators of faking such as significantly more applicants than incumbents within a latent class, higher HPI scale scores for applicants, decreased internal reliabilities, and very low difficulty parameters, I identified and categorized three groups of respondents – honest, standard faking, and a third group which I tentatively called deceptive faking. Although the deceptive faking group represented the most intelligent respondents, their HPI main scale scores were not significantly different than honest respondents' scores, whereas standard faking respondents' scores were significantly higher than honest respondents' scores except for the Sociability scale. In terms of maladaptive personality, which was explored through the HDS, contrary to hypothesized, (open full item for complete abstract)

Committee: Michael Zickar Ph.D. (Advisor); Kenneth Borland Ph.D. (Other); Eric Dubow Ph.D. (Committee Member); Samuel McAbee Ph.D. (Committee Member) Subjects: Personality Psychology; Psychological Tests; Psychology

Doctor of Philosophy, University of Akron, 2005, Psychology-Industrial/Organizational

Previous research (Dwight & Donovan, 2003) suggests that warnings or instructions are useful in reducing applicant faking. This study hypothesized that instructions which promote accuracy would be more effective than those which discourage faking. Likewise, it was hypothesized that instructions which emphasize social verification would be more effective than methods that emphasize embedded social desirability items. This study compared the effectiveness of varying the focus of consequences and the method of identification of instructions on mean applicant scores, honesty of scores and applicant reactions. This study also sought to understand the processes by which warnings may influence applicant faking. The results of this study suggest that, contrary to hypotheses, instructions that emphasized the use of embedded items were more effective than those that emphasized the use of social verification in lowering mean applicant scores. Interestingly, the use of embedded items was also associated with lower levels of expectancy of producing accurate ratings. It was also found that warnings led to more honest applicant ratings. When analyses were conducted only on response elevators, it was found that the focus of consequences and method of identification interacted to influence applicant ratings and the credibility of the method of identification. Specifically, embedded items were more effective and credible when paired with promote accuracy instructions. It is suggested that warnings may impact response elevators and non-elevators differentially.

Committee: Andrea Snell (Advisor) Subjects: Psychology, Industrial