About
What this site is, what it is not, and how it pays for itself.
What this is
CogniScale is a free, 30-question cognitive ability test that runs entirely in your browser, together with a set of sourced guides on what such scores mean. It was built because most free IQ tests share two problems: they do not explain how they arrive at a number, and they present that number as far more precise than any short test can be.
The response to both is the same: publish everything. The methodology page describes the construct, the item construction rules, the scoring model, the measured precision and nine specific limitations. Every score comes with a confidence interval attached.
What this is not
This is not a clinical assessment and carries no diagnostic, educational or legal standing. It cannot diagnose a learning disability, support an application for accommodations or gifted programmes, or substitute for an assessment by a qualified psychologist. If you need a score that carries weight somewhere, a supervised battery administered by a licensed practitioner is the only route.
What it is based on
The test blueprint follows the item families used by the International Cognitive Ability Resource, a public research project set up to build validated cognitive measures outside the commercial test publishers. Its initial validation drew on nearly 97,000 participants, and a later study found its 16-item short form correlated r = .81 with WAIS-IV Full Scale IQ in a sample of 97 students.
The ICAR item bank itself is licensed for academic use only, so no ICAR item appears on this site. What is borrowed is the published methodology. Every question here is original, generated from explicit construction rules and machine-verified before release. The details are here.
Editorial standards
- Claims are sourced. Factual statements cite a named paper or textbook, or describe how this site itself works. No "studies show" without the study.
- Uncertainty is stated, not smoothed over. Where the evidence is contested, the pages say so. Where this test is weak - and its norms are genuinely its weak point - that is written plainly rather than buried.
- No score inflation. Reported scores are capped at 135 because a 30-item test cannot resolve beyond that. A flattering number would be better for sharing and worse as a measurement.
- Corrections are made and dated. If something here is wrong, report it and it gets fixed on the page with the date changed.
How this is funded
Display advertising through Google AdSense. That is the entire business model: there is no paid tier, no upsell, no report to purchase, no email list and no data sold. The test cannot be paywalled because there is nothing behind a wall — scoring happens on your device.
Advertising creates an obvious incentive to maximise page views, and the way that usually plays out on quiz sites is ads wedged between questions, a score held back behind a "continue" that loads another unit, or a result split across five pages. None of that happens here. The rules, which are enforced in code rather than left to good intentions:
| Where | Ads? |
|---|---|
| During the test, on any question | Never. The ad loader refuses to fill any slot while a question is on screen |
| Between questions | Never. There is no interstitial of any kind |
| Above your score on the result page | Never. The score card comes first |
| Below your score, once you have it | One unit |
| Home page, below the fold | One unit |
| Within a guide, at a paragraph break | One unit |
Additionally, and permanently: no pop-ups, no pop-unders, no sticky or floating units, no auto-playing video, no full-screen scroll-over, and no countdown before content. Those formats are Better Ads Standards violations, and sites using them get filtered by Chrome — but the more immediate reason is that they would make the test worse. Every ad container also reserves its height in advance, so a slow-loading ad can never shove the page around under your cursor.
Your test answers are never collected, shared or sold, because they never leave your browser in the first place. The privacy policy has the specifics, including what Google may set independently of us.
Who made it
CogniScale is an independent project, not affiliated with any test publisher, university or professional body. It is not endorsed by Pearson, Riverside, the ICAR project or any of the authors cited on the methodology page. Where their work is described, it is described from the published sources, which are all listed and linked.
Trademarks referenced — WAIS, Wechsler, Stanford-Binet, Raven's Progressive Matrices — belong to their respective owners and are used here only to describe those instruments.
Contact
Errors, ambiguous questions, broken pages, or a claim you think overstates the evidence: get in touch. Reports about specific test questions are especially welcome, since an item that reads as ambiguous to a careful reader is a real defect regardless of what the verifier says.
Take the test
Thirty questions. One score, with the uncertainty attached.
Start the test30 questions · about 30 minutes · no sign-up