Are personality tests valid for hiring? What the evidence shows

The short answer

Some personality tests predict how people do at work, but weakly. The strongest trait, conscientiousness, correlates about .19 with job performance ratings in a 2022 re-analysis — roughly 3–4% of the variation in those ratings. The MBTI is a separate case: the Myers & Briggs Foundation's code of ethics calls using its results to screen job applicants unethical, says taking it should be voluntary, and says results go only to the person who took it unless they agree otherwise. That is the position of the instrument's owner, not a validity study, and it covers only the MBTI.

What meta-analyses say about validity

In 1998 Schmidt and Hunter summarised 85 years of research and gave figures that are still widely quoted: .51 for general mental ability tests, .51 for structured interviews, .31 for conscientiousness. In 2022 Sackett, Zhang, Berry and Lievens re-analysed the published meta-analyses and concluded that earlier estimates were inflated by over-correcting for range restriction; average validities came out .10–.20 lower.

Their revised figures: structured interviews .42; job knowledge .40; empirical biodata .38; work samples .33; cognitive ability tests .31 (previously .51); integrity tests .31; assessment centres .29; unstructured interviews .19 (previously .38). For personality, conscientiousness on general scales is .19 (previously .31), and .25 when the questions are framed in a work context ("at work, I…"). The other Big Five traits on general scales: emotional stability .09, extraversion .10, agreeableness .10, openness .05. For comparison, Schmidt and Hunter listed graphology at .02.

An earlier meta-analysis by Barrick and Mount (1991, 117 studies) had already found conscientiousness to be the only trait consistently linked with success across occupational groups.

Limit: these are averages across many studies and jobs, mostly against supervisor ratings, and say nothing about any specific product. The spread is wide — for conscientiousness the lower 80% bound is .02, and for the other traits it is around zero or below — so for many jobs there may be no link at all. The 2022 re-analysis has been disputed in published replies; the 1998 figures are best read as a historical point.

Fairness between groups

In the same 2022 summary, Black–White score differences (d) are close to zero for personality scales — conscientiousness −.07, emotional stability .09 — compared with .79 for cognitive tests, .67 for work samples, and .23 for structured interviews. Under the US Uniform Guidelines, a selection rate for any group below four-fifths (80%) of the rate for the most successful group is generally treated as evidence of adverse impact, and the employer is then expected to show validity.

Limit: the d figures cover only the Black–White comparison in US data; small gaps do not show an absence of discrimination by disability, age, or language. The four-fifths rule is a rule of thumb, not a legal threshold, and applies only in the United States. This is not legal advice.

Mental health and the law (United States)

In the United States, according to EEOC guidance on the Americans with Disabilities Act, employers may not require medical examinations before a conditional job offer. EEOC guidance treats a psychological test as medical if it identifies a mental disorder; under that guidance, tests of ordinary personality traits generally are not. In Karraker v. Rent-A-Center (7th Circuit, 2005), an employer's use of the MMPI was held to be a medical examination under the ADA.

When a hiring decision tool is automated, New York City's Local Law 144 requires a yearly independent bias audit, publication of its results, and notice to candidates 10 business days in advance. A 2025 New York State Comptroller audit found enforcement weak.

Limit: Karraker is one ruling from one circuit and concerns a clinical test, not Big Five questionnaires. Local Law 144 applies only in New York City and only to automated tools, and its audit looks at selection rates, not validity. This is not legal advice.

What candidates do and feel

A 2006 meta-analysis of 33 studies found that job applicants score higher than non-applicants on emotional stability (d = .44) and conscientiousness (d = .45), less so on extraversion (.11) and openness (.13); direct Big Five scales were distorted more. A 2004 meta-analysis of 86 samples (48,750 people) found that candidates rate personality tests less favourably than interviews and work samples, and that how job-related a test seems is linked with how fair it feels and how people see the employer.

Limit: the first finding compares groups, so part of the gap may be self-selection rather than faking, and its effect on validity is not established. The second measures perception, not validity, and draws heavily on student and hypothetical samples.

What to ask a test vendor

The professional standard for validating selection procedures is the SIOP Principles (5th edition, 2018), adopted by the American Psychological Association as official policy. In the United States, the Uniform Guidelines expect evidence of validity when a procedure has adverse impact. So the document to ask a vendor for is the validation file. NOOS will sell diagnostics only when that file exists and the candidate can see the kind of score being produced.

Limit: the SIOP Principles are a professional standard, not law. NOOS has not yet audited any commercial instrument itself.