You’ve probably, at some point, taken an online personality test. Possibly Myers-Briggs, which gives you a four-letter type: ENFP, INTJ, and so on. Possibly something more playful — which Hogwarts house, which Disney princess, which type of bread. You probably, also, took the result at least a little seriously. The label felt descriptive, perhaps even revealing. It gave you vocabulary for something you already half-knew about yourself.
Here’s something worth understanding. Most of the personality tests most people encounter have, at best, weak scientific support. The Myers-Briggs Type Indicator — despite being the most widely-used personality test in corporate and educational settings globally — has performed poorly in reliability and validity studies for decades. The same person takes it twice and often gets different letters. The types don’t correspond to anything that shows up in other psychological research. The theory underneath it, drawn from Carl Jung via a mother-daughter pair of amateur enthusiasts in the 1940s, has never been empirically supported.
And yet Myers-Briggs is everywhere, while a different framework — one with genuine scientific backing, with cross-cultural replication, with predictive power for real-life outcomes — remains almost unknown outside academic psychology. The gap between what the research supports and what the public uses is, in the personality field, unusually stark.
The well-supported framework is called the Big Five, and understanding it tells you something useful both about your own personality and about the strange gap between pop psychology and the science it usually ignores.
The framework itself
The Big Five describes personality along five broad dimensions, each of which is itself a composite of more specific traits. Most people fall somewhere along a continuum on each dimension, not at an extreme. The five are:
Openness to experience: how curious, imaginative, and intellectually adventurous you are. High openness predicts enjoyment of novelty, abstract ideas, and aesthetic experiences. Low openness predicts preference for familiar patterns and practical concerns.
Conscientiousness: how organised, disciplined, and goal-directed you are. High conscientiousness predicts academic and professional success, health outcomes, and life satisfaction — it’s one of the strongest single predictors of life outcomes in the entire personality research literature. Low conscientiousness predicts flexibility but also higher rates of missed commitments.
Extraversion: how much energy you draw from social interaction and external stimulation. High extraversion predicts larger social networks, higher reported positive emotions, and more leadership positions. Low extraversion — what’s casually called introversion — predicts preference for solitary activities and smaller but often deeper social connections.
Agreeableness: how cooperative, trusting, and sympathetic you are. High agreeableness predicts better interpersonal relationships and more prosocial behaviour. Low agreeableness predicts more competitiveness and more willingness to prioritise self-interest over group harmony.
Neuroticism: how prone you are to negative emotions — anxiety, sadness, worry, emotional volatility. High neuroticism predicts higher rates of mental health problems, but also higher creativity in some studies. Low neuroticism predicts emotional stability and lower stress reactivity.
Each of these dimensions is continuous — people aren’t open or closed, extraverted or introverted, in absolute terms. They’re somewhere along each scale. And the five dimensions are relatively independent — someone can be high on conscientiousness and low on agreeableness, or high on openness and high on neuroticism, in any combination.
Where the framework came from
Unusually for a psychological model, the Big Five wasn’t developed from a single theorist’s intuition and then tested. It emerged from decades of attempts to find structure in the words people use to describe themselves and others.
The approach, called lexical analysis, started from a straightforward premise. If a personality characteristic is important enough that humans need to talk about it, it will eventually get a word. If you collect all the words a language uses to describe personality, and you apply statistical techniques to see which words tend to cluster — which descriptive words tend to apply to the same people — you should be able to identify the underlying dimensions of personality that human language has evolved to track.
This work was done across several decades, by researchers including Gordon Allport at Harvard in the 1930s, Raymond Cattell in the 1940s, and then refined by Lewis Goldberg and others from the 1970s onward. The Big Five emerged repeatedly from these analyses, in English first, and then in language after language. The same five dimensions kept reappearing, whether researchers started from English, German, Dutch, Chinese, or Turkish descriptive vocabulary. Human cultures, it seems, have all developed ways of noticing the same five rough dimensions of how people differ, presumably because these differences have been socially important across human history.
The Big Five was then refined further through a different approach — factor analysis of personality questionnaires, extensively developed by Paul Costa and Robert McCrae at the National Institute on Aging, whose NEO Personality Inventory became the standard research tool for measuring the five dimensions. Costa and McCrae’s work extended the Big Five into detailed sub-facets (six for each major dimension) and documented that the structure holds up with impressive consistency across ages, genders, and cultures.
This is significant. Most personality frameworks don’t replicate cleanly across cultures, or change shape with age, or give different results when measured with different instruments. The Big Five does none of these. It’s among the most robust structural findings in all of personality psychology.
What the Big Five predicts
The practical value of a personality framework depends on what it can predict. Here the Big Five performs substantially better than its competitors.
Conscientiousness, in particular, predicts real-life outcomes with consistency. The University of Illinois researchers Brent Roberts and Nathan Kuncel, in a series of comprehensive reviews, have documented that conscientiousness predicts academic performance, job performance, health outcomes, relationship stability, and longevity — with effects comparable to or larger than those of IQ. A person higher on conscientiousness, other things equal, tends to do better across almost every life domain where reliable effort matters.
Neuroticism predicts mental health risk. High neuroticism is one of the strongest predictors of later development of anxiety and depression disorders, and it predicts responses to stress, relationship satisfaction, and general life satisfaction in ways that have replicated extensively.
Extraversion predicts sociability-related outcomes. Openness predicts engagement with intellectual and creative pursuits. Agreeableness predicts prosocial behaviour and interpersonal satisfaction. None of these effects are individually huge — personality is one variable among many in shaping a life — but they’re real, they’re measurable, and they hold up across populations.
This is what a scientifically supported personality framework looks like. Not magical insight into the soul, but modest, honest, cross-culturally-replicated predictions about broad tendencies.
What Myers-Briggs claims, and what it delivers
Set the Big Five alongside Myers-Briggs, and the comparison is brutal.
Myers-Briggs sorts people into one of sixteen types defined by four either-or dimensions: extraverted or introverted, sensing or intuitive, thinking or feeling, judging or perceiving. Each type has an associated personality profile, often quite elaborate. The framework is widely used in corporate training, career counselling, and personal development.
The problems, from the research perspective, are severe. The test-retest reliability is poor — about 50 per cent of people who retake the test get different letters within a few weeks, even though personality doesn’t actually change on that timescale. The types don’t correspond to anything in other psychological research — there’s no evidence that the sixteen types are meaningfully distinct categories rather than arbitrary cut-offs on continuous distributions. The underlying theory, drawn from Carl Jung’s speculative writing about psychological functions, was never designed as a scientific model and has never been empirically tested successfully.
Most damagingly, the either-or structure of Myers-Briggs forces people into categories when the actual personality data shows smooth distributions. Being called “an ENFP” makes it sound like there’s a discrete type of person you are. The underlying scores, when examined, usually show that you’re near the middle on at least one or two of the dimensions — and small changes in your score would flip you into a different supposed type. The type is, in a statistical sense, often just noise around the middle of the distribution.
The question of why Myers-Briggs has become so popular despite being scientifically weak is itself interesting. Partly it’s that the types feel meaningful — the descriptions are vague enough to resonate with almost anyone (the Barnum effect). Partly it’s that the framework is more accessible than the continuous, less-tidy Big Five. Partly it’s that Myers-Briggs was heavily marketed and licensed to corporate users in ways that created a commercial infrastructure around it. The scientific support doesn’t drive the usage. The usage drives the usage.
Alternatives worth knowing
The Big Five has some honest competitors. The most significant is HEXACO, developed by Michael Ashton and Kibeom Lee, which adds a sixth dimension — honesty-humility — capturing traits like sincerity, fairness, and freedom from greed and manipulation. HEXACO research suggests that the sixth dimension captures variation that the Big Five partly misses, particularly at the negative end (people with dark-triad traits — narcissism, Machiavellianism, psychopathy — score low on honesty-humility in ways that the Big Five’s agreeableness doesn’t fully capture). The framework is less well-established than the Big Five but has a growing research base.
A different extension is the Dark Triad research tradition, associated with Delroy Paulhus and others, which studies three specific trait complexes linked to interpersonally costly behaviour: narcissism, Machiavellianism, and psychopathy. These aren’t personality types in the Big Five sense — they’re clusters of traits that tend to co-occur in ways that predict specific behavioural patterns. The Dark Triad research is interesting for what it adds to the Big Five rather than as a replacement.
The counter-thread worth hearing
Before accepting the Big Five uncritically, some honest caveats.
First, the Big Five describes personality at a broad statistical level but has limits when predicting specific individual behaviour. Two people with identical Big Five profiles can behave quite differently in specific situations. The framework is about average tendencies, not specific predictions.
Second, the cross-cultural replication of the Big Five, while strong, isn’t perfect. Some studies in non-Western cultures have found subtle differences in the factor structure, with some dimensions loading slightly differently than in Western populations. The broad picture replicates, but the details are still being worked out.
Third, personality is more stable than situationists once claimed but less stable than the Big Five sometimes implies. Meaningful change in Big Five traits does occur across the lifespan, and specific interventions can produce modest changes. Treating the Big Five as a fixed assessment of who you are misses the genuine (if slow) capacity for personality change.
What to do with this
If you want to understand your own personality with any scientific rigour, take a well-constructed Big Five test rather than Myers-Briggs or its cousins. Free and reasonably accurate versions are available (the IPIP-NEO is one well-validated option). The results will come as five scores on continuous scales, not as a type label. This is less satisfying than being told you’re an ENFP, but it’s more accurate.
The scores themselves are modestly useful as self-knowledge. Someone who scores very low on conscientiousness should probably build structural supports into their life rather than relying on willpower. Someone who scores high on neuroticism should probably take their emotional signals with some discount and cultivate calming practices. Someone who scores low on extraversion should probably build a life that doesn’t demand constant social performance. None of this is startling; the framework just gives you more accurate vocabulary for things you probably half-know.
What the Big Five doesn’t do is tell you who you are or who you should be. It describes, statistically, some broad dimensions of how you tend to respond to the world. That’s useful. It’s not everything.
The question that remains
The deepest thing the Big Five research teaches is, ironically, humility about how much can be said precisely about personality. After a hundred years of trying, the best available scientific framework gives you five rough scales and modest predictive power. Most of who you are, in any specific sense, isn’t captured by any personality framework. The frameworks catch broad tendencies. Your life is built from much more specific material than tendencies.
This suggests holding personality frameworks — even the well-supported ones — lightly. They’re useful. They’re not deep. The sense of recognition you got from a pop-psychology personality test was probably not, strictly, the test being accurate. It was probably you being generous to the description.
The question to carry, before the next time someone asks you your personality type:
Of the things you know about yourself that feel most important, how many could any personality test ever have told you?
Key research referenced: the lexical and factor-analytic research on the Big Five by Gordon Allport, Raymond Cattell, Lewis Goldberg, Paul Costa and Robert McCrae (extensive, spanning 1930s-present); Brent Roberts and Nathan Kuncel’s meta-analyses of Big Five predictive validity; critiques of the Myers-Briggs Type Indicator (e.g., Pittenger, 2005); Michael Ashton and Kibeom Lee on HEXACO; Delroy Paulhus on the Dark Triad.