Should organizations be concerned with “faking” when using the PI Behavioral Assessment?
The increased use of personality assessments in the workplace has brought renewed concerns about applicant response distortion (e.g., “faking”). Based on over 60 years of experience and a body of empirical research performed by organizational researchers, The Predictive Index does not believe faking is an issue when using well-designed personality assessments such as the PI Behavioral Assessment.
What is faking?
Faking occurs when applicants purposefully distort their responses to personality-based items to make themselves look more favorable. Research shows that applicants can alter their responses on tests when specifically instructed to do so. For example, in various studies conducted in carefully controlled laboratory settings, people instructed to “fake good” received higher or different scores than those told to respond honestly.
“Faking” is often regarded as deceitful behavior because it is perceived as manipulative or purposeful “gaming” of the assessment. Research suggests that most faking is due to socially desirable responding or impression management. “Socially desirable responding” simply means that applicants are uncomfortable admitting to faults or unfavorable self-descriptions. Impression management, on the other hand, involves applicants actively monitoring the situation (in this case, the job for which they are applying) and trying to put their best self forward. Both socially desirable responding and impression management are normal and occur during almost every job interview and social interaction. They are adaptive behaviors—not malicious ones—that help people get along or get ahead in the workplace.
What is the major concern with faking?
If an applicant fakes a personality assessment, it is likely that the results do not accurately describe the behavioral tendencies of the assessment-taker. The biggest concern from organizations is the belief that faking is so pervasive that the majority of applicants’ results are invalid. In other words, the concern is that the use of personality assessments may not be an effective tool in the hiring process.
Do most applicants fake on personality assessments, and how does this affect validity?
The consensus within the industrial-organizational psychology community is that faking is typically not a wide-spread problem. In fact, a March 2007 Journal of Applied Psychology research report indicated that less than 4% of job applicants distorted their responses to an extreme degree. In “real-world” hiring situations, most job applicants do not consciously manipulate their responses to personality assessments, which means that assessment tools serve their purpose as valid predictors of job performance.
Most of the evidence reported in scholarly journals of psychology indicates that faking has a slight (if any) negative impact on predictive validity, and for some jobs which require impression management skills (e.g., sales), faking may even increase validity. Thus, unless one has reason to suspect that most applicants within a job category are faking substantially, faking should not be considered overly problematic.
What steps does The Predictive Index take to reduce the likelihood of faking?
The PI Behavioral Assessment is a free-choice instrument in which there are no obvious “right” or “wrong” answers. It is extremely difficult, if not impossible, for people to accurately and consciously map an individual item on the assessment back to the underlying behavioral factor it was intended to measure.
The assessment uses an adjective checklist approach that provides less clarity around which responses are considered more “desirable” as opposed to framing questions as full sentences in which the “right” answer is more obvious (e.g., “I follow the rules every day”). This makes the PI Behavioral Assessment more challenging to manipulate. In addition, the PI Behavioral Assessment is periodically updated to remove adjectives that are clearly socially desirable or undesirable (i.e., those adjectives that are very highly endorsed or almost never endorsed).
Because PI Behavioral Assessment scores are reported and interpreted using a within-person format, the ability to manipulate the assessment is even further reduced. A test-taker who tries to distort their responses must not only choose the “right” words, but must also correctly identify the right combination of words (e.g., more A and B words, fewer C words than D) in order to arrive at an ideal pattern for a job. This would be extremely challenging even for a test-taker intimately familiar with the scoring mechanisms of the assessment.
The PI Behavioral Assessment is widely perceived by test-takers to be fair, non-invasive, easy to understand, and job-related. Research indicates that assessments perceived in such a manner do not typically suffer from faking.
The PI Behavioral Assessment is administered in a casual, non-threatening manner, and when administering the assessment to current employees, many organizations agree to review results with test-takers as soon as possible. These practices all reduce the likelihood of faking.
To date, no PI client has ever indicated that faking was a serious problem or concern whether among job applicants or current employees. Hundreds of jobs, organizations, and criterion-related validity studies covering 30 years consistently demonstrate that PI Behavioral Assessment patterns are related to objective job performance. If the assessment were truly impacted by faking, we would not have this overwhelming amount of evidence of the assessment’s ability to predict and improve job performance.
Assessment results should only be considered as a single data point in a hiring decision, with ample opportunities during interviews, reference checks, and other processes to indirectly validate the results of the assessment. In fact, social desirability and impression management are likely to be of bigger concern during interviews than during completion of the PI Behavioral Assessment.
Based on years of peer-reviewed academic research and our own experiences, there is no reason to believe that applicant faking systematically impacts the utility and validity of the PI Behavioral Assessment in any substantial manner.