What this test measures

Where the items come from

The item types come from psychometrics, the field that measures reasoning. The same families sit in published research batteries, which is why they can be built to rules instead of invented.

We use no material from Raven’s Matrices or any other protected test. Every item is generated from named rules, following published work on item generation.

The six kinds of item

• What completes a grid
• What continues a series
• What does not belong
• What stays the same when a shape is stretched
• What an object looks like turned around
• What number comes next

Six subtests, 36 scored items in six blocks, one practice item before every block, and a timer on each block

What we guarantee

Nothing to remember and nothing to look up: it is not knowledge, vocabulary, schooling or memory

• Exactly one answer is correct: a solver proves it for every item before the item enters the bank
• Wrong options are named mistakes, not random shapes
• Six options in every block, so a guess is worth the same everywhere
• No vocabulary: “?” is the only character you will see in an item
• No colour: items are ink on paper, so colour vision does not change your score
• Three forms, matched position by position, so a retest compares with your first run
• Difficulty is a tier and a count of rules, not an opinion

Each line above is a test in our build. A bank that fails one of them does not ship.

What your score is

There is no unit of intelligence, so a score is a position, not an amount. 100 is the middle of the group you are compared with, and that group is named on your result.

• A range, not a single number. The range is the measurement; a bare number is a claim
• A percentile among people who finished the same test, and the group you are compared with is always named on screen
• Norms carry a version, and so does the item bank; your result records both

How the comparison grows

At first you are compared with everyone who finished. Age groups open once enough people of your age have taken the test.

What we do not claim

• This is not a clinical or diagnostic assessment
• Our test has not been validated against WAIS, Stanford–Binet or any other clinical instrument
• Research findings about intelligence describe groups, never the person reading them
• We do not tell you what your score predicts about your income or your work
• One sitting gives one overall score, not a map of your strong and weak sides; that needs more items than a single session holds

What we build on

Open, published work on matrix item generation and on short public intelligence batteries: ICAR and the Sandia matrix generator