Types of IQ test
The label covers instruments that have very little in common: 90-minute supervised batteries, single-domain matrix tests, society entry exams and ten-question web quizzes. What separates them is norming, breadth and supervision.
The main kinds
WAIS / WISC
Supervised clinical batteryTen core subtests across verbal comprehension, perceptual reasoning, working memory and processing speed, administered one-to-one by a psychologist over roughly 90 minutes.
The reference standard. Produces a full-scale score plus index scores, and is the only kind of result that carries weight for diagnosis, accommodation or placement. Costs several hundred to a few thousand, with a wait.
Stanford-Binet
Supervised clinical batteryThe other major individually administered battery, with a similar structure and a longer history. Often preferred at the extremes of the range, where it has more usable items.
Equivalent standing to the WAIS. Same cost and access constraints.
Raven's Progressive Matrices
Single-domain instrumentMatrix reasoning only: a grid of figures with one cell missing. No language, no arithmetic, no cultural content.
Excellent at what it does, and deliberately narrow. Gives a clean read on fluid reasoning and says nothing about verbal ability, memory or speed. Widely used in research and in hiring.
Mensa entry tests
Screening examA supervised sitting designed to answer one question: are you above the 98th percentile. Some national branches also accept prior evidence from a clinical battery.
A pass/fail threshold rather than a measurement. It is not built to tell you your score, only which side of a line you are on.
Free online quizzes
EntertainmentTypically 10 to 25 items, no published norming, no item calibration, and a score generated by a formula the site does not disclose.
The scores are frequently inflated on purpose, because a flattering number gets shared. Treat as a game. The tell is a result with no confidence interval and no account of how it was normed.
The three things that actually differ
Norming. A score means nothing without a reference population and a documented conversion. Most online tests skip this entirely and invent a plausible-looking number instead.
Breadth. A matrix-only test measures one broad ability. A full battery samples several, which is what makes a profile of strengths and weaknesses possible rather than a single figure.
Supervision. An unproctored test cannot stop you pausing, looking things up or retaking it. That does not make the items worthless, but it does mean the result is self-reported, and it is the reason no online score should be used for any decision that matters.
Where this test sits
This one is an unsupervised online battery, but it is built like a clinical instrument rather than a quiz: 70 items across six domains from the Cattell-Horn-Carroll taxonomy, calibrated item parameters, item response theory scoring, and a reported confidence interval. It cannot substitute for a supervised assessment, and the methodology page is explicit about where its limits are.
If you want to see the kind of problems involved before starting, there are worked examples from each domain.