How to Reduce Job Applicant Faking on Personality Tests

woman working on computer
Topic(s): assessment, personality, selection
Publication: International Journal of Selection and Assessment
Article: Comparing personality test formats and warnings: Effects on criterion-related validity and test-taker reactions
Authors: P.D. Converse, F.L. Oswald, A. Imus, C. Hendricks, R. Roy, H. Butera
Reviewed by: Benjamin Granger

Although personality testing in employee selection settings is a common practice, it hasn’t gone without critique. The reason for this is simple: personality tests can be faked. A quick glance at many of the commonly used personality test items will corroborate this concern, for example, “I am always prepared,” or “I make friends easily.” Most job applicants would not even consider disagreeing with such statements in a high stakes hiring situation, and would instead provide the answer they believe will help them land the job, even if it is untruthful. Despite this problem, personality tests have been shown to predict employee job performance. Thus, there is a major dilemma over what to do about personality testing in employee selection settings.


In general, most personality measures allow job applicants to read a series of items and indicate the extent to which they agree or disagree with each statement. Due to the transparency of such items, several methods are currently being investigated which aim to ameliorate the faking issue. Two of the most promising methods include using (1) forced choice (FC) items (i.e., force applicants to choose one personality trait characteristic over another) and (2) warnings to test takers about faking (i.e., inform applicants of the consequences of faking). Importantly, these two options effect faking in different ways. FC methods reduce an applicant’s ability to misrepresent him or herself while warnings effect an applicant’s motivation to fake his or her personality.

Although both FC and warning methods have been shown to reduce faking, relatively little research has looked at these two methods simultaneously, until now. In this article, researchers (Converse et al., 2008) were interested in the combined effects of FC and warnings on two important outcomes: (1) criterion-related validity (how well personality measures predict job performance) and (2) test-taker reactions.


The researchers found that FC personality tests not only predicted important work-related criteria (e.g., leadership, perseverance), but they did so above and beyond cognitive ability. However, their findings suggest that in some situations, job applicants may have more negative reactions toward FC formats compared to more traditional formats (e.g., Likert scale). On the other hand, it was not completely clear whether warning applicants of the consequences of faking showed acceptable criterion-related validity. Thus the authors warned against drawing firm conclusions.  Moreover, when warnings were used and were framed negatively, applicants reported high levels of test anxiety. However, positively framed warnings did not have such an effect on applicants.


So what do organizations need to know about these methods that aim to curb faking? First, both FC and warnings reduce the occurrence of faking. Second, FC methods, although more time-consuming than traditional personality tests, predict workplace criteria over and above cognitive ability. Third, if organizations choose to use warnings against faking, they may consider wording the warnings in a positive way because negatively worded warnings may cause test-taker anxiety. For example, organizations can say, “If you respond honestly, it is more likely that you will be placed in a job that suits you well.” This may be more beneficial than saying, “We are going to catch you and you will no longer be considered for the job.”

Converse, P, D., Oswald, F. L., Imus, A., Hedricks, C., Roy, R., & Butera, H. (2008).  Comparing personality test formats and warnings: Effects on criterion-related validity and test-taker reactions. International Journal of Selection and Assessment, 16(2), 155-169.