Personality measures and assessments are commonly used for selection purposes in many organizations. However, there are some concerns surrounding their use. Specifically, many individuals believe that these types of assessments can be easily faked. In other words, applicants applying for a job may distort their responses to be more in line with how they believe the organization wants them to respond. Research has demonstrated that some applicants fake and faking is linked to decreases in test validity.
There has been a large effort to tackle the issue of faking in personality assessments by developing a variety of methods to detect individuals who are attempting to improve their scores. A recent study conducted at Select International used four faking indicators as an attempt to hone in on fakers. The four indicators used were:
- Covariance Index (CVI)
- Bogus Items
- Social Desirability
- Blatant Extreme Responding (BER)
The covariance index is computed by identifying pairs of unrelated items from a personality assessment using a sample that has low motivation to fake, in this case, current job incumbents. The CVI is then applied to the sample of interest, job applicants, and fakers are identified.
Bogus items are used to identify people who are purposefully lying about their experience. Specifically, these items are fictional and do not represent true components of the job. Therefore, if an individual reports that they have experience with these items, they must be lying. Individuals who endorse a high amount of bogus items are identified as fakers.
Social desirability is computed using scores from a set of socially acceptable responses. Individuals who respond to this scale at an extremely high rate are flagged as fakers.
Finally, BER is a relatively new technique for detecting fakers by flagging individuals who endorse the desirable extreme points (e.g., 1s and 5s) on the scale. Individuals who endorse the desirable extreme points at a high rate are identified as fakers because it is unlikely that an individual would score at the extreme across all personality traits.
The Select International study used these four indicators individually as well as in conjunction as an attempt to triangulate them to more accurately identify individuals as fakers. Therefore, various combinations of the four indicators were also computed. A second goal of the study was to determine whether or not using these triangulation methods to identify and remove fakers actually led to improved test validity.
Results revealed that relationships between the personality assessment and actual performance on the job were stronger when using combinations of the faking indicators to filter out individuals. This supports the notion that we can in fact combine multiple faking indicators to triangulate fakers and that excluding those fakers from the sample improves test validity. These results have two main practical implications for organizations. First, identifying fakers using triangulation methods and removing them increases test validity. Second, individuals flagged as fakers have lower job performance.