Personality Traits

Which Personality Test Should I Take?

PsyCove Editorial · 8 min read · Updated 2026-10-28

The answer depends on the question you are bringing. To measure and compare, take a Big Five inventory with published norms. For vocabulary about how you work with others, type language is fine if held lightly. For a hypothesis about motive, the Enneagram. For a career direction, RIASEC. To judge whether any test is worth taking, skip the brand and check four things: norms, retest stability, published validation, and whether the scoring is visible.

Key takeaways

There is no best personality test, only a best match between an instrument and a question.
Four properties separate a real scale from a quiz: published norms, retest stability, validation against outcomes, and visible scoring.
Free tests are usually free because they skip the norming — which is the part that makes a score mean anything.
A good test gives you roughly the same answer twice. If it does not, that is information about the test rather than about you.
No personality test diagnoses anything, and none should be used to decide whether someone else is suitable.

Start with the question, not the test

Most people arrive at this decision from the wrong end: they have seen a test name and want to know whether it is good. The instrument is the last thing to choose, not the first, because the same questionnaire can be excellent for one purpose and useless for another. Three questions settle it. What do you actually want to know? What will you do with the answer? Who else will see it? If the honest answer to the second is 'nothing in particular', any test will do and none will help. If the answer is 'decide whether to hire someone', no instrument in this category is the right one, and the conclusion it produces would be wrong even before it was unlawful.

The third question, who else will see it, is the one people skip. A result you keep to yourself costs nothing when it is wrong. A result that ends up in a personnel file, a dating profile, or a family argument carries consequences that an instrument of modest accuracy cannot support.

A short field guide

Six families cover almost everything people are looking for, and each has a legitimate use alongside a characteristic failure.

Big Five inventories: the measurement instruments. Continuous scores, published norms, decades of validation. Use them to compare yourself with a population and to track change. Failure mode: nothing about 'moderately high conscientiousness' is memorable, which is why people drift back to types.
Type systems, MBTI and relatives: vocabulary. Best for a shared language about working preferences, and their correspondence with the five factors is documented (McCrae & Costa, 1989). Failure mode: the letter discards magnitude, so people near the middle get a label that does not predict them.
The Enneagram: motive. Best used as a hypothesis generator about what you avoid. Failure mode: its most distinctive machinery — wings, arrows, levels — has almost no empirical support.
RIASEC: interests and direction, not personality. Best for choosing a field, with the caveat that it says nothing about capability, and that its performance prediction is modest (Nye et al., 2012).
Attachment measures: relationships. Best for understanding a pattern under intimacy, with the caveat that they describe a style rather than a condition.
Everything else — the four-question quizzes, the spirit animals, the color wheels: entertainment. Fine as entertainment, and worth being honest about which one you are doing.

Notice what the list is organized by: purpose, not quality ranking. A type system is a worse measurement instrument than a Big Five inventory and a better conversation-starter, and both of those sentences are true at the same time.

What makes a test worth taking

If you want a criterion that travels across instruments rather than a list of brand names, four properties do most of the work.

Norms: a score means nothing until you know the distribution it sits in. 'High conscientiousness' is a percentile claim, and a test without norms cannot make it.
Retest stability: take it twice, weeks apart. A decent instrument gives you roughly the same profile; a poor one moves for reasons that have nothing to do with you. Dichotomous scoring makes this worse — Pittenger's 1993 review reported a retest study in which about half of respondents received a different type.
Validation: has anyone published what this scale predicts, in samples the vendor did not choose? Conscientiousness has that record (Barrick & Mount, 1991); most popular instruments do not.
Visible scoring: can you see how your answers turned into the result? Opaque scoring is where cold reading lives.

None of this requires a psychology degree to check, and the first two you can test yourself in an afternoon. That is the practical use of the list: it turns 'is this test scientific' into four answerable questions.

The free-test problem

Free tests are not worse because they are free; they are usually worse because norming is the expensive part. Building a scale is cheap — write statements, ask people to agree on a five-point scale, average the answers. Building norms means collecting data from a large representative sample until you know what a score means, and then validating it against something outside the questionnaire. That step is where the money goes, and it is the step online quizzes skip.

There is a second cost that is less visible. A free test that returns a flattering result is a product designed to be shared, and the sharing is the business model. That does not make the result false, but it does mean the result has been selected for how it feels, which is the opposite of what a measurement is supposed to do.

What no test will tell you

Three limits apply to every instrument in the list, including the good ones, and knowing them in advance is what keeps a result in proportion.

Not a diagnosis. No personality test detects or rules out a condition. An instrument that appears to is probing mental health, which is a different activity with a different evidence standard.
Not a hiring tool. Scores are self-reported, modestly stable, and easy to manage when a job is on the line — quite apart from the legal constraints that apply to screening with personality measures.
Not an identity. A profile is a description at one moment under one set of instructions. It is not a ceiling on what you can do, and treating it as one converts a description into a permission slip.

A sequence that works

If you want to do this once and get something out of it, the order matters more than the choice of instrument.

Write the question first, in one sentence. 'I want to know why I go quiet in meetings' is a better starting point than 'I want to know my personality'.
Pick the instrument that answers that question, not the one with the best website.
Write down what you expect to be true about yourself before you see the result. It is the only way to distinguish a finding from a flattering mirror.
Take it a second time, a month apart. Consistency is the cheapest available test of whether the instrument is measuring anything stable, and it is information about the test rather than about you.

One last thing to expect: a good result is usually boring. 'Moderately high in conscientiousness, average in everything else, with one facet that stands out' is what real measurement sounds like. If a result feels like a revelation, the instrument is doing something other than measuring, and that something can be useful — as long as you know which one you bought.

The useful question is never 'what is my personality'. It is 'what am I trying to decide, and does this instrument have anything to do with it'.

Related types

🌙 INFP The Mediator

Related guides

Frequently asked questions

What is the most scientifically valid personality test?

For measuring traits, a Big Five inventory with published norms, such as one built on the NEO or IPIP item pools. 'Valid' is always relative to a purpose: a Big Five measure is the better instrument for measurement, and a type system is the better instrument for a conversation, which is a different job.

Are free online personality tests accurate?

Usually not, and the reason is structural rather than malicious. Norming a scale is the expensive part, and free quizzes skip it, so the score has no reference distribution behind it. The result can still be interesting as self-reflection; it is not a measurement.

How do I know if a test is any good?

Four checks: is there a published norm group, does it give you a similar result on retest, is there published evidence of what the scale predicts, and can you see how the scoring works? If the answer to all four is unclear, you are looking at a questionnaire rather than an instrument.

Should I take one test or several?

One, twice, beats several once. Instability between two different instruments is uninterpretable, because you cannot tell whether the difference is the scale or the construct. Instability on the same instrument is informative, because the only thing that changed is you or the conditions.

Can a personality test tell me what career to choose?

A RIASEC-style interest inventory can help with direction, and a Big Five profile can tell you which conditions you will tolerate. Neither predicts performance well enough to make the decision for you, and matching a personality profile to a job description tends to produce people who are well suited to work they do not want.

References

This guide summarizes published research. It is not a diagnosis, not therapy, and not a hiring, admission, or screening instrument. PsyCove Editorial is an in-house editorial team, not a clinical practice; if distress is persistent, contact a qualified professional or a local crisis line. PsyCove is not affiliated with or endorsed by the MBTI® trademark.