Asking candidates the right questions is vital

How to spot fake answers to personality tests

Worried about potential employees faking the answers on a personality test? According to the experts, employers who use personality tests should be cautious — not only about the possibility of applicants faking the “right” answers, but about the methods testers use to detect such faking.

On The Wall Street Journal Online Erin White writes about a US psychology professor Richard Griffith, whose research shows that candidates do give fake answers to ensure they give the “right” answer for the job being sought.

He tested the same students twice — once when their instructor asked them to be totally honest and again, six weeks later, when they thought they were being tested by recruiters.

He found that about a third of the students changed their answers to achieve “significantly higher scores”.

White writes that such behaviour "substantially harms the ability of the test to predict" people's performance.

However, Griffith told White that such tests are still more reliable than a job interview alone. While employees can include questions to weed out people faking their answers they do not always work.

“One common question designed to trap fakers asks applicants whether they have or haven't done something common but frowned upon, such as ‘I have never looked at a dirty book or magazine.’ In theory, liars are more apt to say they never have, while honest people admit the behaviour.”

However, Griffith told White that applicants who are smart know that they should confess — effectively they know to fake the answer to the question designed to see if they are faking.

There is also the chance that the applicant has not looked at a dirty magazine and therefore wrongly gets labelled a liar.

Whatever the outcome, experts recommend that personality tests are not used as the only tool in hiring. Writing in the Success Performance Solutions newsletter, Ira Wolfe says such tests should contribute to about one-third of a candidate's total score, with the interview making up another third and “reference checks, experience, education, and technical skills” the rest.

Wolfe points out that there seems little point trying to fake a personality test when there is not really a right or wrong answer, yet people do, which is why good tests include questions designed to pick out fakers. Wolfe believes “lie detector” or “social desirability” scales (where applicants’ answers are used to plot them on a scale which shows how honest they are), easily detect people faking it.

Wolfe gives the example of one system where the social desirability scale is integrated into test reporting. If candidates score between three and eight on the scale, then employers know they can believe the predictive personality traits.

“Does that mean that candidates who score 1-2 or 9-10 are liars? No. It does however throw up a caution flag to hiring managers, who should investigate the responses further.

“Scoring above or below the 3-8 range could be a warning that the candidate tried to game the system. It could also just mean that the candidate is a ‘saint’ and strives to always act in a socially desirable way.”

Like Griffith, Lynn McFarland, an assistant professor at Clemson University, also studies "faking" and is also sceptical about the ability to detect fakers using the social desirability scale.

Quoted in an article by Kathleen Groll Connolly on the workplace assessment firm Performance Programs’ website, McFarland says some people may have unusually high scores on social desirability scales, but may “simply be trying to please, may have poor reading skills, or may have overly optimistic beliefs about themselves.”

Connolly writes that questionnaire wording is one of the best tools to catch fakers.

“People who give extreme answers to questions are often — but not always —overcompensating to hide their true intent [when compared with] those who do not answer with the extreme choices.”

Connolly also suggests giving people two tests because it is less likely someone will be able to deceive in both.

“Tell participants they will be discussing their answers at a later date. Dr McFarland says this strategy shows some promise. It holds an applicant responsible for consistent answers at a later date.”

Join the newsletter!

Error: Please check your email address.

Tags managementcandidatesAnswers

More about Wall Street

Show Comments

Market Place

[]