What Is Raven’s Matrices? Why Figure Problems Best Capture Fluid Intelligence

You have certainly seen this kind of problem: a 3×3 grid with the bottom-right cell missing, and five options to choose what belongs there. It is called Raven's Standard Progressive Matrices (SPM), designed by British psychologist John C. Raven in 1938 and still one of the most widely used cognitive ability tests in the world. Its most remarkable feature: the entire test contains not a single word.

Why no words? Because it targets "fluid intelligence"

Psychologist Raymond Cattell split intelligence into two kinds in 1963. Crystallized intelligence is the knowledge, vocabulary and facts you accumulate over a lifetime — it grows with age and education. Fluid intelligence is your ability to face a brand-new problem and find the pattern through pure induction and reasoning, without relying on prior knowledge. It peaks around age 20 and slowly declines afterward. Raven deliberately uses abstract shapes to minimize the "how much have you learned" variable. Even if you never went to school and cannot read, you can score as long as you can spot the transformation rule between figures. That is why it is so widely used in cross-cultural research, multilingual hiring assessments and child cognitive evaluation.

Crystallized intelligence is the screwdriver in your toolbox; fluid intelligence is the knack of turning a screw by hand in an unfamiliar situation.

What do matrix puzzles actually measure?

On the surface it is just "find the pattern and fill the gap", but underneath it tests three things at once:

• Inductive reasoning: extracting a shared rule from multiple figures (sides increasing, rotation, overlap cancellation, etc.)
• Working memory: holding both the "row rule" and the "column rule" in mind and checking whether they are consistent
• Inhibitory control: rejecting distractor options that look right but do not actually satisfy the rule

Three classic matrix rules (master these and you crack 80% of items)

The vast majority of 3×3 figure puzzles fall into one of these three transformations — which is exactly why people who have practiced score noticeably higher than first-timers.

The essence of a matrix puzzle is "satisfy both the row rule and the column rule at once" — find that intersection and the answer is unique.

Does a high score mean you are smart? Three limits to keep in mind

Raven, like every cognitive test, measures "a specific ability in a specific situation", not life outcomes. Three common misconceptions:

• It is not overall intelligence: it measures only fluid reasoning, not creativity, emotional intelligence, practical wisdom or social skill — all equally important in real life
• Practice effects are strong: first-timers usually score 10-15 points lower than practiced solvers, so a low "cold" score does not equal low ability
• State matters a lot: poor sleep, anxiety and time pressure all drag the score down. A single result is a reference, not a label

Can fluid intelligence be trained?

Yes, but within bounds. A 2008 study by Jaeggi et al. on "n-back training" claimed to boost fluid intelligence, but later replications have been mixed. The current consensus: practice on matrix puzzles themselves reliably raises scores (because you learn the templates), but whether that transfers to "becoming smarter overall" is still debated. What has been proven to help: sleep, aerobic exercise and lifelong learning of new skills — these put your brain in a better state whenever fluid intelligence is called on.

Seven minutes to find your fluid-reasoning tier

Drawing on the design logic of Raven, we created 16 original reasoning items — four each in letter/number series, verbal logic, 3×3 matrices and 3D spatial rotation. None use official items. After finishing you get a reasoning index, a norm percentile and a four-axis radar chart — not a clinical diagnosis, but an entertaining exercise in self-insight.