Personality tests are sold for hiring and for team workshops. Do they predict how people will perform at work, and should we use MBTI or DISC?

Personality traits do relate to job performance. In a 2003 study of 159 employees of a pharmaceutical company, published in a South African journal, three traits (emotional stability, openness and agreeableness) explained 28% of the variance in management performance. The links with task performance were weak: no single trait's correlation with it reached the authors' threshold of practical significance, which they suggest may reflect well-defined jobs with little autonomy. It was a cross-sectional survey in one company. (Read in library)

But the headline figure for the strongest trait is inflated. Conscientiousness is, in the words of one study, "a valued predictor in employment selection". A 2015 reanalysis of the largest collection of data on how well it predicts performance found that estimate "moderately overestimated (by around 30%; a correlation difference of about .06)", mainly because small effects tend to go unpublished. The authors add that the inflated estimates overstate the money value of selecting on personality "by millions of dollars". (Read in library)

Applicants can fake, and faking changes who ranks where. In a 2017 study, 243 applicants to a teacher-education college in Austria completed personality scales as part of their admission. The authors conclude that applicants' answers in a real selection were "quite stable", but that faking introduces a uniform bias that "detrimentally effects the stability of the rank-order" of candidates. In a 2021 laboratory study of 939 university students in Spain, conscientiousness remained the best predictor of performance even when participants were asked to fake, but reliability and validity "were attenuated under faking conditions".

A test introduced for selection predicted poorly at scale. UK medical schools added personality ("non-cognitive") tests to their admissions test. A 2016 study looked at the 6,294 students who graduated in 2013 from 30 of them, and analysed the 3,343 who had taken the new tests. A few traits predicted exit assessments, but "the predictive validity of these tests was limited, and the relationships revealed do not fit neatly with theoretical expectations". The authors' conclusion: "This study does not support their use in selection."

Other methods added to the questionnaire, or matched it. In a 2020 study, 223 employees took a structured job interview designed to measure the Big Five traits, and a personality questionnaire, in a simulated selection. The interview ratings predicted the helpful, cooperative behaviour their supervisors reported, and added to that prediction over and above the questionnaire and a verbal ability test. (Read in library) In a 2022 study of 528 employees in Spain, a brief job-related biodata scale, a short set of questions about a person's work history, predicted two kinds of job performance. It did as well as most personality traits, and added prediction over and above the Big Five and the "Dark Tetrad" traits. (Read in library)

On MBTI and DISC, we found almost no evidence, and what there is points to weak prediction. Our search of openly licensed, peer-reviewed research found no study testing whether MBTI or DISC predict job performance. The one study we found that tests MBTI against behaviour at all surveyed 529 business-administration students at Colombian universities, most of them graduate students, in 2023 (464 valid responses). Their MBTI results explained about one percent of the variance in their leadership practices, a relationship the authors call weak. It was a student sample using questionnaires, it did not test hiring or job performance, and the authors still see a use for MBTI in self-awareness. (Read in library) We found no openly licensed study on DISC. The studies above, the ones that do link personality to performance, measure personality traits, mostly the Big Five, not the types these tools report. A test used to decide who gets hired should be able to show what it predicts before it is used; we could not find that evidence for either.

How much applicants fake is disputed. A 2020 meta-analysis of 66 studies (115 samples, 46,265 people) found that, in selection situations, applicants with higher cognitive ability report higher personality scores, a pattern its authors attribute to their greater ability to fake; the pattern varied between laboratory and field studies. But when 128 firefighter applicants took a personality inventory during selection and again three months later, a 2021 study found "very little faking behavior on average", and higher cognitive ability went with less score inflation, not more.

A 2018 study goes further. Measuring faking ability directly in online samples, it found that ability largely explained by cognitive abilities, and concluded that research linking self-reported personality to job performance "is presumably confounded by cognitive abilities". If that holds, some of what personality questionnaires seem to predict may really be intelligence. (Read in library)

The size of the relationship is also contested. A single-company survey found traits explaining 28% of the variance in management performance; the 2015 reanalysis found the best-known estimate for conscientiousness inflated by publication bias. They measure different things in different samples, so neither cancels the other.

Don't hire on a personality questionnaire alone, and don't use MBTI or DISC to decide who gets hired unless the provider shows evidence that they predict performance in jobs like yours. Trait measures carry some real information, less than the headline figures claim, and self-reports can be shaded in ways that change the ranking. If personality matters for a role, assess it in a structured interview built around the traits the job needs, add job-related history, and treat any questionnaire as one input among several.

Confidence: moderate that self-report personality measures predict performance less well than their headline figures suggest and can be distorted in selection. On MBTI and DISC we found no study showing that they predict job performance, which is why we advise against using them to select people until a provider can show that evidence for jobs like yours.

1. Check what your tests are used for. If MBTI, DISC or any other questionnaire screens people out of a hiring process, pause that use until the provider shows validation for your roles.
2. Ask every provider four questions: what does it predict, in which jobs, how strongly, and in real applicants or in volunteers?
3. Put personality into the interview. Name the two or three traits the job needs, write questions for each, and score every answer on a fixed scale.
4. Add a short job-related history form. In the study of 528 employees in Spain, a brief biodata scale predicted performance about as well as most personality traits.
5. Never use a single personality score as a cut-off. Faking can change the order in which candidates rank.

None of the single studies here was set in the Gulf.

If you use a questionnaire in translation, or with applicants of many nationalities, none of these studies tells you how it performs. Before using any test for hiring, ask for evidence from applicants like yours, in the language they will take it in.

Ten of the eleven research sources below are open access under CC BY 4.0, with the licence verified per article through Crossref. The eleventh (1) carries a CC BY 4.0 notice on the published article, but Crossref holds no licence record for it. All are academic journal articles; none was published by a company that sells personality tests.

1. The big five personality dimensions and job performance — SA Journal of Industrial Psychology, 2003
2. The Validity of Conscientiousness Is Overestimated in the Prediction of Job Performance — PLOS ONE, 2015
3. The Psychometric Costs of Applicants' Faking: Examining Measurement Invariance and Retest Correlations Across Response Conditions — Journal of Personality Assessment, 2017
4. Effects of Faking on the Predictive Validity of a Quasi-Ipsative Forced-Choice Personality Inventory: Implications for Sustainable Personnel Selection — Sustainability, 2021
5. Do personality traits assessed on medical school admission predict exit performance? A UK-wide longitudinal cohort study — Advances in Health Sciences Education, 2016
6. Who will go the extra mile? Selecting organizational citizens with a personality-based structured job interview — Journal of Business and Psychology, 2020
7. Analysis of a brief biodata scale as a predictor of job performance and its incremental validity over the Big Five and Dark Tetrad personality traits — PLOS ONE, 2022
8. The relationship between cognitive ability and personality scores in selection situations: A meta-analysis — International Journal of Selection and Assessment, 2020
9. Predictors of faking behavior on personality inventories in selection: Do indicators of the ability and motivation to fake predict faking? — International Journal of Selection and Assessment, 2021
10. The "g" in Faking: Doublethink the Validity of Personality Self-Report Measures for Applicant Selection — Frontiers in Psychology, 2018
11. How good is the Myers-Briggs Type Indicator for predicting leadership-related behaviors? — Frontiers in Psychology, 2023