Policy

Editorial policy

How the content on this site is written, checked, sourced and corrected. Short enough to read.

Sourcing

Every factual claim on this site falls into one of three categories, and each is handled differently.

Generated from the scale. Percentiles, rarity figures and classification shares are computed from the normal distribution by the same code that renders the page. They can be checked with any statistics package and they cannot drift out of agreement with the text.

Established findings. Correlations with outcomes, reliability figures, the structure of the CHC model, the Flynn effect. We state these at the level of confidence the literature supports and flag where estimates are disputed — the recent re-analysis of range restriction in job-performance meta-analyses is one example we mark on every page that touches it.

Outside datasets. The national IQ figures are the only one. That page carries its caveats before the table rather than after it, because a table read without the caveats is the problem we are trying to avoid.

Contested research

Where the field disagrees, we say so and describe the disagreement rather than choosing the version that makes a tidier page. The structure of g, the size of job-performance correlations after range-restriction correction, the causes of the Flynn reversal and the interpretation of national IQ data are all live arguments and are presented as such.

Where a popular claim is not supported — the Mozart effect, brain-training transfer, multiple intelligences as measured constructs, Einstein's IQ — we say that plainly and say what the evidence actually shows.

Test items

Every item is generated from an explicit rule, solved independently against that rule, and checked to confirm no distractor also satisfies it. Spatial items are verified as chiral by enumeration. The generating rule for every item is published on the test page, which is what makes it possible for a reader to tell us we are wrong. The full methodology.

Disclosure

We link to iq-test.io and recommend it as the longer free test. Other sites discussed in the reviews section are named and described but not linked.

This is stated on the reviews hub, on every individual comparison and on the about page. Each comparison also names what the competing test does better, because a comparison that finds nothing good to say about the alternative is not a comparison.

Corrections

If something here is wrong, tell us and we will change it. Substantive corrections are made to the page itself rather than appended as a note, because a reference page should be right rather than annotated.

We are particularly interested in hearing about test items with more than one defensible answer, explanations that do not follow, and reference figures that disagree with a source you can name. Contact.

What we are not

Not psychologists, not a clinical service, and not a substitute for assessment. Nothing on this site is a diagnosis or an official result, and pages that touch on clinical questions say what kind of professional to ask instead.