Every personality product on the internet describes itself as science-backed. The phrase is doing very different amounts of work in each case. Here's how to tell.
Three questions that sort the field
1. If I retake it in five weeks, do I get the same answer? This is test-retest reliability, and it's where the famous four-letter type instruments quietly collapse — studies find that roughly half of people get a different type on retest within weeks. A measurement that changes with your mood isn't measuring your personality.
2. Do the scores predict anything? The Big Five — the five-trait model that emerged from decades of research — has meaningful, replicated correlations with real outcomes: how people perform, lead, and collaborate. It's the consensus model in academic psychology not because it's fashionable, but because it keeps replicating.
3. Can two people's scores be honestly compared? Some popular instruments use "ipsative" scoring — they rank traits within you, so scores can't be compared between people. For an individual reflecting alone, fine. For anything team-shaped, comparison is the whole point, and ipsative scores quietly can't support it.
Where that leaves the familiar names
The four-letter tests are beloved and useful as conversation starters — and weak as measurement. The strengths-branded tools have real research behind them but ipsative scoring and vendor lock-in. DISC is best understood as a friendly language sitting on top of a subset of the Big Five: excellent for communicating, not what you'd choose to measure with.
And the frontier stuff — inferring personality from your voice, your face, your game-play? The accuracy is modest (about as good as one acquaintance's guess), the bias problems are documented, and workplace emotion-inference now sits inside the EU AI Act's restricted territory. Exciting demos; poor foundations.
The stack that survives scrutiny
Measure with the boring, bulletproof thing: validated Big Five instruments from the public-domain research item pool, scored normatively so colleagues can be compared honestly. Then present with the friendly thing: plain working-style language people actually enjoy using.
Rigorous underneath, human on the surface — in that order, and never the reverse.
Ask any personality vendor those three questions. The ones with good answers will be delighted you asked. The other kind will change the subject to how fun the results are at offsites.
