Electronic Thesis and Dissertation Repository

Degree

Doctor of Philosophy

Program

Psychology

Supervisor

Goffin, R.D.

Abstract

This dissertation examined the measurement and prevention of applicant faking on personality tests. Study 1 compared how 12 different faking indices differentiated between the same people’s personality scores during a job application and non-applicant condition, and how these faking indices discriminated between separate groups of job applicants and non-applicants. We found that researchers and practitioners should assess applicant faking using multiple faking indices, including Idiosyncratic Item Responding, Blatant Extreme Responding, and Communal Impression Management, and when possible, Residualized Change Scores. Interestingly, Bogus Items—a common faking measure in the literature—were ineffective. Similarly, two covariance indices, were also ineffective and failed to predict applicant faking or relate to any other measure of applicant faking. The findings also suggest that we can disregard a few measures from future analysis because they are inferior to existing options, and offer no utility, which include Percent Agreement, Individual Change Scores, Within-Subject Correlations, and Within-Subject Variances of the Differences. Studies 2 and 3 examined the efficacy of three new faking dissuasion messages in reducing applicant faking compared to a traditional faking warning and a no-warning control group. In Study 2, we found no evidence for any of the dissuasion messages, including one adapted from the existing literature. In Study 3, we tested military recruits from the Canadian Armed Forces. We found some evidence that an Immediate Authentication Warning, which informed recruits that faking could be identified using reference checks and internal integrity checks, helped reduce faking. However, none of the other faking dissuasion messages were effective relative to the control group. We discuss several potential explanations for these conflicted findings.

Download

Share

COinS