Can Personality Tests Be Faked?

Can personality tests be faked? Yes. Someone who wants the job answers differently than the same person filling out the identical questionnaire with nothing at stake, and a 2025 meta-analysis of 80 paired samples and 14,041 participants, published in the International Journal of Selection and Assessment, found that the shift is large enough to measurably weaken how well the test then predicts job performance, not just how the score looks on paper.

That distinction is the whole story. A faked test can still look internally consistent, since the person is answering steadily toward the image they want to project. What breaks is the link between the score and the thing the test was built to predict: what that person actually does on the job.

Can personality tests be faked enough to matter

Personality tests already predict job performance only modestly under honest conditions, and the same meta-analysis put the honest-condition correlation at r=0.21 across the five major personality traits, against r=0.14 once respondents were motivated to make a good impression, a drop to roughly two thirds of the original strength.

The gap widens further in the samples that come closest to a real hiring decision: matched groups who answered the same test once with nothing at stake and once as an actual job applicant. There, the same research found validity fell from r=0.18 to r=0.09, a correlation weak enough that it becomes hard to say the test is still measuring anything a hiring decision should rely on.

Why the format of the test changes how much faking costs

Not every personality test loses ground equally. A test that asks someone to rate single statements one at a time, "I stay calm under pressure," strongly agree to strongly disagree, gives a motivated respondent an easy path: rate every desirable statement higher. A test that instead forces a choice between several equally flattering-sounding statements removes that shortcut, since picking one does not let every trait score rise at once.

The same meta-analysis measured that difference directly: single-statement formats lost 0.09 points of validity to faking, against 0.02 for forced-choice formats, a gap large enough to matter even though fewer forced-choice studies were available to confirm it precisely.

What this means for a hiring decision

This is a different weakness than the one covered in how accurate self-assessment is: that post is about people misjudging their own ability even when answering honestly. Faking is deliberate, and it gets worse exactly when the stakes are highest, which is the opposite of what a hiring decision needs from a test.

It is also why the field has moved toward formats that do not ask a candidate to self-report a trait at all. A situational judgement test asks what someone would do in a specific scenario and scores the answer against agreed criteria. A behaviorally anchored rating scale ties every point on a scale to an observed behavior instead of a self-description. Neither format asks a candidate to rate their own conscientiousness or emotional stability, so neither depends on that candidate answering honestly about it. The six-domain check behind this site follows the same logic: a scored response to a real scenario, not a self-rated trait a motivated applicant could adjust upward.