Some are, many aren't — and telling them apart doesn't take a PhD. The honest checklist.
In everyday language, reliable means “I can trust it”. In psychometrics it splits into two separate questions: reliability — does the test measure precisely, without too much noise (would it give you a similar answer next week)? — and validity — does it measure what it claims to, and does that relate to anything real? A test can be exquisitely precise at measuring nonsense, and a good idea can be measured with a noisy instrument. You need both.
The third leg almost nobody looks at: norms. A “73” means nothing on its own; it means something compared to a reference group. Which group? From what country, age and era? A serious test tells you; an entertainment one never even asks itself the question.
One: does it measure continuous traits, or does it put you in a box? People are distributed along gradients; types are a rounding. The rounding is useful for conversation and treacherous for decisions — and cutting a continuum down the middle turns two nearly identical people into “opposite types”.
Two: does it tell you its margin of error? All measurement carries noise. Serious instruments acknowledge it (bands, intervals, firmness); magazine ones hand you a verdict with the confidence of a horoscope.
Three: does the result always flatter you? If everything it says sounds good, it isn't describing you: it's retaining you. The Forer effect — vague, favourable descriptions anyone feels are their own — is the engine of half this sector.
Four: where do the items and the norms come from? Documented public domain, described samples, limitations stated. If the answer is “just trust us”, you already have your answer.
We measure continuous traits with public-domain items used in research, we show bands and firmness instead of faking infinite precision, we write every trait with its gift and its price, and we tell you where the norms come from — including the uncomfortable part: our own base is still small and growing with every real response. The deep profile additionally leans on published norms built from hundreds of thousands of cases.
And the limit we won't dress up: this is self-report. It measures how you see yourself and how you answer today, not a truth carved in stone. No honest test can promise you more.
The full checklist of how we measure, with its limits stated, is published.
KEEP READING