MBTI For Hiring: What The Research Actually Says (And Why The Test's Own Publisher Warns Against It)

Six things to know before reading further:

The Myers-Briggs Foundation publishes Ethical Use Guidelines that include explicit positions on what MBTI is and is not designed for. The relevant clause: "The MBTI instrument is intended for use in development; it is not intended for use in selection or hiring decisions, and its use for those purposes is unethical." This position is not a contested claim — it is the publisher's own framing, repeated across the Foundation's documentation, the CPP commercial license materials, and the certified MBTI practitioner training curriculum.

The Foundation's reasoning, paraphrased: MBTI is designed to help people understand their own preferences and how those preferences interact with others. It was not designed to predict job performance, and it lacks the predictive-validity evidence required for selection use. Using MBTI in hiring also creates a high risk of legal challenge under U.S. employment law (more on this below), because the framework can be argued to have disparate impact across protected demographic groups depending on how the test is administered and interpreted.

This is an unusual situation in commercial assessment markets. Most personality test publishers position their tests for the broadest possible market including hiring. The Myers-Briggs Foundation's explicit narrowing of legitimate use to development contexts is a notable artifact of how the framework was originally positioned — Isabel Myers framed MBTI as a self-development tool, not as a selection instrument, and the Foundation has held that line for decades despite obvious commercial incentives to relax it.

Per the Foundation's published guidelines, the legitimate workplace uses of MBTI are: team-building and team-effectiveness conversations, leadership development programs, conflict-resolution work, communication-style awareness, and individual coaching. The illegitimate uses are: pre-employment screening, hiring decisions, promotion decisions, performance evaluation, and any context where the assessment determines whether a person obtains or retains employment.

John Pittenger's 2005 review "Cautionary comments regarding the Myers-Briggs Type Indicator" (Consulting Psychology Journal: Practice and Research, 57(3), pp. 210-221, DOI 10.1037/1065-9293.57.3.210) is the most-cited academic critique of MBTI's measurement properties. Pittenger reviews the test-retest reliability literature and finds approximate per-dimension reliability of 0.5-0.6 — meaning that across the four MBTI dichotomies, somewhere between 25% and 50% of the variance in scores is reproducible between sessions. By comparison, Big Five trait measures typically show per-dimension test-retest reliability in the 0.7-0.9 range.

The implication for hiring: when a personality measure has per-dimension reliability of 0.5-0.6, the categorical type code (a binary cutoff on each dimension) flips on retest in approximately 50% of cases. Pittenger documents that across multiple studies, around 50% of MBTI test-takers receive a different four-letter type code within five weeks of their first test. A measurement instrument that flips its categorical output 50% of the time on retest does not have the consistency properties required to underpin hiring decisions.

Pittenger also reviews the predictive-validity literature (the question of whether MBTI score predicts subsequent job performance). The conclusion: there is no accumulated body of evidence demonstrating that MBTI score predicts job performance with the magnitude or consistency required for selection use. Some individual studies show small associations between MBTI dimensions and specific work outcomes, but the associations do not reproduce reliably across studies, and they are smaller than the associations Big Five Conscientiousness has with the same outcomes.

The honest read: MBTI's measurement properties are modest by psychometric standards. They are not zero — the dimension scores carry some signal about underlying preferences — but they are not strong enough to underpin a decision about who gets hired. The publisher's position and the academic-research consensus point in the same direction here, which is unusual; usually publishers and academic critics disagree.

Murray Barrick and Michael Mount's 1991 meta-analysis "The Big Five personality dimensions and job performance" (Personnel Psychology, 44(1), pp. 1-26, DOI 10.1111/j.1744-6570.1991.tb00688.x) is the canonical reference for the relationship between personality traits and work outcomes. Barrick and Mount aggregated 117 individual studies covering five occupational groups (professionals, police, managers, sales, and skilled/semi-skilled workers) and four job-performance criteria (job proficiency, training proficiency, personnel data, and overall performance).

Their headline finding: Big Five Conscientiousness predicts job performance across all five occupational groups and all four performance criteria. The estimated true-score correlation between Conscientiousness and overall job performance was approximately 0.22 — modest in absolute terms but consistent across job types, which makes it the most generalizable personality predictor of job performance ever documented in meta-analytic work.

The other Big Five traits (Openness, Extraversion, Agreeableness, Neuroticism) showed job-specific predictive patterns. Extraversion predicts performance in sales and management roles where social initiative matters. Agreeableness predicts performance in customer-service and team-collaboration roles. But none of the other four show the cross-occupational generalizability that Conscientiousness does. If a hiring process can use only one personality measure, the research evidence supports Conscientiousness.

Translating the Barrick & Mount finding to MBTI terms via McCrae & Costa's 1989 mapping (DOI 10.1111/j.1467-6494.1989.tb00759.x): the MBTI J/P dichotomy correlates partially with Big Five Conscientiousness, with J-types tending higher on Conscientiousness than P-types. So in principle, MBTI's J/P axis carries some predictive signal. In practice, the partial correlation (~0.4) and MBTI's modest test-retest reliability mean that J/P alone is a much noisier predictor than direct Big Five Conscientiousness measurement. Why use the noisier proxy when the direct measurement is available?

John Hunter and Ronda Hunter's 1984 meta-analysis "Validity and utility of alternative predictors of job performance" (Psychological Bulletin, 96(1), pp. 72-98, DOI 10.1037/0033-2909.96.1.72) is the foundational reference for ranking selection methods. Hunter and Hunter compiled the predictive validity (correlation with subsequent job performance) of every commonly-used selection method available at the time.

Their ranking, with the predictive validity coefficient (higher = better predictor):

MBTI does not appear in the Hunter & Hunter list because it was not commonly studied as a selection predictor at the time of the 1984 meta-analysis. Subsequent studies that have tried to estimate MBTI's predictive validity for job performance generally find correlations in the 0.05-0.15 range — putting MBTI between graphology (essentially zero) and unstructured interviews (low signal) on this scale, well below structured interviews and cognitive-ability tests.

The practical implication: a hiring process that uses MBTI is using a selection method whose predictive validity is roughly an order of magnitude weaker than the standard high-validity selection methods (work samples, structured interviews, cognitive ability). It is also weaker than direct Big Five Conscientiousness measurement, which itself is a moderate predictor. There is no level of analysis at which MBTI is the best available selection method.

This is the cumulative case against MBTI in hiring: (a) the publisher explicitly says don't use it for hiring, (b) the test-retest reliability is modest, (c) the empirical predictive validity for job performance is weak in absolute terms, (d) every standard high-validity selection method outperforms it, and (e) better personality-based predictors exist if the goal is to add a personality dimension to the hiring process. The absence of any single overwhelming argument against MBTI in hiring is misleading; the cumulative case is what makes the publisher's position align with the research consensus.

If the publisher says don't, the research says don't, the law-risk says don't, and better alternatives exist, why does MBTI persist in hiring contexts? Three reasons that show up consistently in HR-practice surveys.

**First, recognition heuristic.** MBTI is well-known and widely-recognized by both candidates and managers. A hiring process that includes MBTI feels more thorough than one that doesn't, even when the included assessment doesn't add predictive validity. Hiring managers who don't have a strong psychometric background often assume that a recognized assessment must be useful — the recognition substitutes for the validity check that should be happening.

**Second, the labels are conversationally portable.** "INTJ" travels in a job interview debrief in a way that "73rd percentile Conscientiousness, 41st percentile Openness" does not. Hiring managers can talk about "the candidate's INTJ profile" with each other; they can't easily talk about a Big Five percentile spread. This portability is real value, but it is value as conversational vocabulary, not as predictive validity. The mistake is to confuse the two.

**Third, organizational inertia.** Once an HR department has invested in MBTI training, certified practitioners, and assessment licenses, the cost of switching to a different assessment is non-trivial. Even when the HR director knows MBTI is not the right tool for hiring, the path of least resistance is to keep using it for whatever purposes have been established. This inertia is well-documented in HR-practice research and is one of the strongest forces keeping MBTI in hiring contexts despite its weak fit.

None of these three reasons are evidence that MBTI predicts job performance. They are evidence that MBTI persists in hiring for reasons unrelated to predictive validity. Recognizing this distinction is the first step toward replacing MBTI in hiring with measures that actually predict the outcome you care about.

U.S. employment law treats personality assessments used in hiring as employment tests subject to Title VII of the Civil Rights Act of 1964 and the Equal Employment Opportunity Commission's Uniform Guidelines on Employee Selection Procedures (29 CFR 1607). The relevant standard is the four-fifths rule: if a selection procedure produces a selection rate for any protected demographic group that is less than four-fifths (80%) of the rate for the highest-selecting group, the procedure may have disparate impact and the employer must demonstrate job-relatedness.

MBTI in hiring carries litigation risk on this standard for two reasons. First, MBTI types correlate weakly with demographics in some samples — for example, women self-identify as Feeling-types more often than men, and Sensing/Intuition has age-related self-identification differences. If a hiring process screens for or against specific MBTI types, it may produce selection-rate disparities across protected groups that fail the four-fifths test. Second, when the disparate-impact analysis happens, the employer must demonstrate the assessment's job-relatedness — and as the Pittenger 2005 review documents, MBTI does not have the predictive-validity evidence required to clear that bar in litigation.

The combination — possible disparate impact plus weak job-relatedness defense — is why employment-law attorneys generally advise against using MBTI as a hiring screen. Companies that use MBTI for development purposes (post-hire team-building, leadership coaching) face minimal legal exposure because the assessment is not determining employment outcomes. Companies that use MBTI as a pre-employment screen have a much harder defense if they get sued.

This is why the Myers-Briggs Foundation's Ethical Use Guidelines specifically call out hiring as an unethical use case. The Foundation's reasoning includes the legal exposure as well as the predictive-validity gap. The Foundation has institutional incentives to keep MBTI usable across many contexts, and it has still drawn the line at hiring — that institutional choice is itself evidence about the strength of the cumulative case.

If the underlying goal of including a personality assessment in your hiring process is to add information about how a candidate is likely to perform, the practical alternatives, ranked by predictive-validity evidence:

Three caveats worth naming explicitly so the case against MBTI in hiring isn't over-extended.

**Caveat 1: "MBTI doesn't predict job performance" is not the same as "personality doesn't predict job performance."** The Big Five evidence is robust: Conscientiousness predicts performance, and Big Five facets predict performance with job-specific patterns. The case is specifically against MBTI as the personality assessment used for hiring; it is not a case against personality assessment in hiring generally. Replace MBTI with Big Five Conscientiousness measurement and the predictive-validity problem largely resolves.

**Caveat 2: "Don't use MBTI for hiring" is not the same as "don't use MBTI."** MBTI has documented uses in development, team-building, and individual coaching contexts. These uses are endorsed by the publisher and supported by practitioner literature. The framework's weakness is specifically its predictive validity for selection, which is a narrow technical claim, not a wholesale rejection of the assessment.

**Caveat 3: The research case described here is the dominant academic position, not a unanimous one.** Some researchers and practitioners disagree with elements of Pittenger's 2005 critique or with the application of Barrick & Mount 1991 to MBTI displacement. The minority view typically argues that MBTI provides team-fit information that pure Big Five Conscientiousness does not. That argument is plausible for team-building use cases (where MBTI is endorsed) but is much weaker for selection (where MBTI's measurement properties don't support the use). For the long-form treatment of MBTI's measurement properties and the four primary citations, see /blog/mbti-common-misconceptions-and-data. For the framework comparison without the hiring-specific framing, see /blog/mbti-vs-big-five.