Does Personality Predict Job Performance?
By PsyCove Editorial · 12 min read · Updated 2026-10-27 · Checked against the sources listed below
- One trait generalises and the rest are conditional. Conscientiousness showed consistent relations with performance across all criteria and all five occupational groups studied; other traits were job-specific (Barrick & Mount, 1991).
- The numbers were too high. A 2022 reanalysis found that range-restriction corrections had been overcorrected, and revised mean validities downward by .10 to .20 points (Sackett et al., 2022).
- Structured interviews came out on top. After the correction, most procedures kept their ranking, but the interview led and the whole field sat lower than previously reported.
- Type adds nothing here. Two decades of management research using the type indicator produced no consistent evidence that type predicts performance (Gardner & Martinko, 1996).
What Did the First Meta-Analysis Actually Find?
The 1991 meta-analysis found one trait that travels and four that do not, which is a narrower result than the way it is usually summarized. That reading would be wrong if the five occupational groups had produced five different rankings; the consistency across criteria is the finding.
Barrick and Mount pooled the studies available by 1991 across five occupational groups — professionals, police, managers, sales, and skilled or semi-skilled work — against three criteria. Conscientiousness showed consistent relations with every criterion in every group. Extraversion predicted performance for the two occupations built on social interaction, managers and sales, and both extraversion and openness predicted how well people did in training (Barrick & Mount, 1991).
Everything else was small. Where the remaining traits did predict, the estimated true correlations sat below .10, which is a signal you would need a very large sample to see and would never see in one applicant.
How Much Did the Numbers Change?
The correction was large enough to matter in practice. The figures that circulated for thirty years were, in the technical sense, too generous.
The problem was not the underlying studies but the arithmetic applied to them. Meta-analyses of selection tools correct observed correlations for range restriction, because applicants are a narrower group than the general population. A 2022 review of five common ways of building that correction found each of them flawed in ways that overcorrected, and therefore overstated how well selection procedures predict performance (Sackett et al., 2022).
Two findings from the reanalysis are worth keeping. Most procedures kept their relative ranking, so the ordering you may have seen was roughly right. But mean validities dropped by .10 to .20 points, and the structured interview became the top-ranked procedure (Sackett et al., 2022).
- Ranking survived. Procedures that were near the top before are near the top now, so the ordering in older summaries is not useless (Sackett et al., 2022).
- Size did not. Every mean validity fell by .10 to .20, which is a large change on a scale where the whole field spans about .3.
- One position moved. The structured interview took the top spot, ahead of the ability tests that had led for decades.
Which Traits Matter for Which Jobs?
Conscientiousness is the generalist among the five traits. Everything else is a specialist, and the specialty has to be named.
| Trait | Predicts overall performance? | Where it does the work |
|---|---|---|
| Conscientiousness | Yes, across occupations and criteria | Every job studied. The effect is strongest for the achievement and order facets |
| Emotional stability | Yes, for overall performance | Generalisable in the larger review, though less consistent across specific criteria (Barrick et al., 2001) |
| Extraversion | No | Managers and sales, plus training proficiency. The advantage is specific to jobs built on interaction |
| Openness | No | Training and learning speed rather than on-the-job output |
| Agreeableness | No | Team and customer settings. It can cut against output in individually scored, competitive roles |
The pattern is consistent across reviews. A later summary of fifteen prior meta-analyses reached the same conclusion and then asked the field to stop running this kind of meta-analysis, on the grounds that the question had been answered (Barrick et al., 2001).
How Does That Compare With Other Selection Tools?
Personality measures do not compete against nothing in a hiring decision. They compete against interviews, work samples and tests of general mental ability, and the comparison is unflattering in one respect and useful in another.
The influential summary of the field put general mental ability and structured interviews at the top of the ranking, with personality measures below them (Schmidt & Hunter, 1998). The 2022 correction kept that shape while lowering every number, and it put the structured interview first (Sackett et al., 2022).
The useful part is that these tools measure different things. A personality score does not predict what an ability test predicts, which is why adding one to the other raises prediction slightly rather than duplicating it. The unflattering part is that a personality score alone is a weak instrument for choosing a person.
That additivity is easy to overstate. The gain from adding a personality measure to an interview comes from the part of performance the interview did not capture, and if the criterion is a single supervisor rating, that part is small. Where the criterion is broader — retention, counterproductive behavior, teamwork — the added information is worth more, which is why job-specific criteria matter more than a global performance score.
Does Your Type Predict Anything at Work?
The evidence for the four-letter code in workplace settings is much thinner than its use, and the reviews are blunt about it.
A review of two decades of management research using the type indicator found no consistent evidence that type predicts managerial behavior or performance, and it noted that the instrument was being used well beyond what it was designed for (Gardner & Martinko, 1996). The broader cautionary literature makes the same point about claims that outrun the measure (Pittenger, 2005).
- A criterion that is not self-report. Ratings from supervisors on outcomes defined before the study starts.
- Samples large enough to detect a small effect. With a plausible validity around .1, a single team is hopeless.
- A comparison against conscientiousness. If the type adds nothing over a five-minute trait measure, the type is the expensive way to get the same information.
There is also a structural reason to expect little. The letters are cut from continuous scores at midpoints, so distance from the midpoint disappears and the categories behave like divisions rather than kinds (Bess & Harvey, 2002). If conscientiousness already predicts weakly, a coarse version of it cannot predict better.
Why Do Employers Still Use Type Tests?
Because they are cheap, because they produce a shared vocabulary, and because a workshop with four letters in it is easier to run than a conversation about validity.
- Cost. A type questionnaire is short, self-scored and needs no trained administrator.
- Legibility. “We need more Js on this team” is a sentence a manager can say in a meeting. A correlation coefficient is not.
- The feeling of insight. Type descriptions are written to be recognized, and recognition feels like evidence (Forer, 1949).
- Weak feedback loops. Nobody tracks whether the hires who scored a particular way performed better, so nothing ever contradicts the practice.
Applicants do adjust their answers when a job is at stake, and that is a real effect. It is also a smaller problem than it sounds for what the scores predict, which is why the selection literature has long treated social desirability as a red herring (Ones et al., 1996). None of that is a conspiracy. It is what happens when a cheap tool with a good story meets an organization with no way to check.
Is a Small Correlation Worth Hiring On?
For a single hiring decision, the answer is no. For a process that makes many decisions and cannot test everyone deeply, sometimes yes — and the difference is arithmetic rather than philosophy.
A correlation around .2 is a real signal across a large applicant pool and close to noise for one candidate. If you hire one person, the score adds little. If you hire two hundred a year into similar roles, a small edge compounds, and the same arithmetic that makes the effect weak for a person makes it visible for an organization. Over whole lives rather than single jobs, the effects of personality traits on outcomes are comparable to those of socioeconomic status and cognitive ability (Roberts et al., 2007) — which is a claim about populations, not about your next hire.
The 2022 review adds a constraint that the older summaries mostly left out: it reported subgroup differences alongside validity for each procedure, so the tradeoff between predictive accuracy and adverse impact is visible in the same table (Sackett et al., 2022). A tool that predicts slightly better and screens out one group more heavily is not a straightforward improvement.
And the criterion itself is less stable than the predictor. Performance varies by situation and by week for the same person, in the if-then patterns the personality literature has documented for decades (Mischel & Shoda, 1995). Hiring on a trait assumes a stability in the outcome that the outcome does not have.
What Should You Do With a Low Conscientiousness Score?
A low score is a description of a default, not a forecast about your life. The correlation that makes the trait valuable to employers was estimated across occupations and averaged over people, and it says nothing about what one person will do in one job in one week.
The mechanisms that produce the conscientiousness advantage are mostly substitutable. Tidiness, punctuality and follow-through are behaviors with external supports: a checklist, a calendar with alarms, a colleague who asks. Someone who scores low and uses those supports can produce the same output as someone who scores high and does not need them.
- Choose the setting, not the verdict. Work that rewards improvisation under time pressure makes different demands from work that rewards repeated accuracy. The score says more about fit than about worth.
- Name the failure mode. Low conscientiousness predicts missed deadlines more reliably than it predicts poor judgment. A named failure is one you can design around.
- Check whether the trait is the active ingredient. In roles where the work is supervised and paced by a system, the structure supplies what the trait would otherwise supply, and the trait has less room to matter.
None of this makes the score meaningless. It makes the score a description of what happens by default, and a default is the part of behavior that a change in the situation can move.
Can an Employer Require You to Take One?
In the United States, often yes, with a limit narrower than most candidates assume. The constraint is not the questionnaire. It is what the questionnaire was designed to reveal.
Three rules set the boundary, and they are not the same rule in each jurisdiction.
- The medical-examination line. The ADA bars most medical examinations before a job offer. EEOC guidance treats a test designed to identify a mental disorder or impairment as medical, and a test measuring traits such as honesty or habits as not medical. A Big Five questionnaire sits on the second side. An instrument built around clinical scales does not.
- The case that drew it. Karraker v. Rent-A-Center, 411 F.3d 831 (7th Cir. 2005), held that a management assessment containing MMPI items was a medical examination, even though it was scored with a vocational protocol and read by no psychologist. Intent did not save it. Effect decided it.
- Elsewhere. The Equality Act 2010 asks whether a selection criterion is a proportionate means of achieving a legitimate aim. The GDPR restricts decisions taken solely by automated means. "Personality test" is not the operative category in either.
In practice this means you can be asked to sit one, and you can be rejected for reasons you will never see. It does not mean a score is a lawful reason by itself, and it does not mean you can tell from the outside which side of the line a given test falls on. What follows is not a diagnosis of you; it is a description of what an employer may and may not ask. If you are handed a questionnaire, "which traits does this score, and what decision does the score change?" is a fair question.
What Should You Take From This?
Four conclusions, ordered by how confident you should be in them.
- Conscientiousness is the one robust finding. It predicts across occupations and criteria, and it survived the correction in rank if not in size (Barrick & Mount, 1991; Sackett et al., 2022).
- Every number you have seen is probably too high. If a validity figure is from before 2022 and was not revised, treat it as an upper bound.
- Personality is a supplement, not a substitute. Its value at work comes from adding information the other tools miss, not from replacing them.
- Do not hire on type. There is no body of evidence that four letters predict performance, and the instrument was never designed for that use.
Our test is built for self-description, and we do not offer a hiring report, because the effect sizes in this literature are too small to carry a decision about an individual. We keep the result page to four positions and no suitability score, for the same reason an employer should not use it as one.
Frequently asked questions
Which personality trait best predicts job performance?
Conscientiousness, and it is the only one that predicts consistently across occupations and criteria (Barrick & Mount, 1991). The effect is real and modest, and its size was revised downward in 2022 (Sackett et al., 2022).
Are personality tests worth using in hiring?
As a small supplement to better tools, sometimes. As a primary screen, no. Structured interviews and ability measures predict better, and a personality score's main value is the information it adds rather than the information it replaces.
Why were the old validity numbers too high?
Because the corrections applied for range restriction were themselves overstated. Five common ways of building the correction were each found to overcorrect, which inflated the resulting validities by .10 to .20 points (Sackett et al., 2022).
Does MBTI type predict job performance?
No consistent evidence supports it. A review of two decades of management research using the indicator found no reliable prediction of managerial behavior or performance (Gardner & Martinko, 1996).
If the effect is small, why do employers still use these tests?
Because they are cheap, produce a shared vocabulary, and rarely face a feedback loop that could contradict them. The feeling of insight comes from the description rather than from the prediction (Forer, 1949).
Further Reading (primary sources & authoritative references)
- Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A meta-analysis. Personnel Psychology, 44(1), 1–26. — DOI
- Barrick, M. R., Mount, M. K., & Judge, T. A. (2001). Personality and performance at the beginning of the new millennium: What do we know and where do we go next? International Journal of Selection and Assessment, 9(1–2), 9–30. — DOI
- Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2022). Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection. Journal of Applied Psychology, 107(11), 2040–2068. — DOI
- Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology. Psychological Bulletin, 124(2), 262–274. — DOI
- Gardner, W. L., & Martinko, M. J. (1996). Using the Myers-Briggs Type Indicator to study managers: A literature review and research agenda. Journal of Management, 22(1), 45–83. — DOI
- Pittenger, D. J. (2005). Cautionary comments regarding the Myers-Briggs Type Indicator. Consulting Psychology Journal: Practice and Research, 57(3), 210–221. — DOI
- Bess, T. L., & Harvey, R. J. (2002). Bimodal score distributions and the Myers-Briggs Type Indicator: Fact or artifact? Journal of Personality Assessment, 78(1), 176–186. — DOI
- Ones, D. S., Viswesvaran, C., & Reiss, A. D. (1996). Role of social desirability in personality testing for personnel selection: The red herring. Journal of Applied Psychology, 81(6), 660–679. — DOI
- Mischel, W., & Shoda, Y. (1995). A cognitive-affective system theory of personality. Psychological Review, 102(2), 246–268. — DOI
- Roberts, B. W., Kuncel, N. R., Shiner, R., Caspi, A., & Goldberg, L. R. (2007). The power of personality: The comparative validity of personality traits, socioeconomic status, and cognitive ability for predicting important life outcomes. Perspectives on Psychological Science, 2(4), 313–345. — DOI
- Forer, B. R. (1949). The fallacy of personal validation: A classroom demonstration of gullibility. Journal of Abnormal and Social Psychology, 44(1), 118–123. — DOI
Related reading
- Which systems have real evidence? — the same question asked of the frameworks rather than of the workplace.
- Do you answer honestly? — why faking is a smaller problem in hiring than employers assume (Ones et al., 1996)
- Where MBTI stops — the structural problems behind the type-prediction gap.
- Trait versus type — why a continuous score is the only version of this that could predict anything.
- The test page reports four positions for your own use. It is not a suitability score, and it is not built to be one.