Test Format
Raven's Progressive Matrices is one of the most widely used non-verbal reasoning tests in psychology, built entirely from visual pattern puzzles rather than words or numbers. This guide explains how the test works, what the three official versions measure, how scoring and percentiles are calculated, and lets you try an interactive sample item built on the same rule families the real test uses.
IQMog's own assessment is built in the Raven tradition — a fixed-form, non-verbal matrix reasoning test — though it is a shorter 20-item form and is not a reproduction of the clinical SPM, CPM, or APM item banks.
John C. Raven introduced Progressive Matrices in 1936 as a way to measure reasoning ability without relying on language, cultural knowledge, or formal schooling. Each item presents a 3×3 (or smaller) grid of related abstract patterns with one cell left blank. The test-taker must infer the underlying rule — a rotation, a size progression, a logical combination of shapes — and select the option that correctly completes it from a set of alternatives.
Because the format uses only abstract shapes rather than words, numbers, or culturally specific content, RPM is often described as one of the more culture-fair intelligence measures available, though researchers note that no test is fully free of cultural or educational influence. RPM specifically targets fluid intelligence (Gf) — the capacity to reason and solve unfamiliar problems — as distinct from crystallised intelligence (Gc), which reflects accumulated knowledge and vocabulary.
RPM scores correlate strongly with Spearman's g — the general intelligence factor that underlies performance across nearly all cognitive tasks. Because a single well-designed matrix-reasoning battery loads heavily on g, researchers frequently use RPM as an efficient single-construct proxy for general cognitive ability in studies where administering a full battery like the WAIS-IV is impractical. For a deeper comparison of matrix-based tests against comprehensive batteries, see the Wechsler IQ test guide.
The demo below cycles through three of the most common rule families used in matrix reasoning tests. Select an answer to see whether you identified the underlying rule correctly, then move to the next pattern.
Try a Sample Item — Rotation
Pattern 1 of 3Which option replaces the question mark?
Nearly every matrix-reasoning item — across RPM and Raven-style tests generally — is built from a small set of underlying rule types. Recognising which family an item belongs to is the single highest-leverage skill for solving them quickly.
| Rule family | How it works | Typical difficulty |
|---|---|---|
| Progressive rotation | A shape rotates by a fixed angle from one cell to the next, continuing consistently across rows and columns. | Low to medium |
| Size or quantity progression | A shape, count, or line thickness increases or decreases by a fixed increment across each row or column. | Low to medium |
| Logical overlay (AND / OR / XOR) | The missing cell is derived by combining the shapes in the other two cells of its row using a consistent logical rule. | Medium to high |
| Distribution of three | Three distinct values (shapes, colours, or orientations) are distributed once across each row and column, similar to a visual Sudoku constraint. | Medium to high |
| Figure addition or subtraction | Elements are added to or removed from a base figure according to a rule that must be inferred from earlier cells. | High |
Raven's Progressive Matrices is not a single fixed test — it is a family of three instruments, each targeting a different population and difficulty range.
Children ages 5–11, older adults, some clinical populations
Uses colour to reduce fatigue and maintain engagement for younger or vulnerable test-takers.
General population, ages 6 through adulthood
The original and most widely used version. Forms the basis for most population norms and research applications.
Above-average adults, gifted assessment contexts
Designed to differentiate among high performers where the SPM produces ceiling effects.
In its original design, the SPM is a power test administered without a strict time limit — the goal is to see how far into increasing difficulty a person can progress, not how quickly they respond. Most modern administrations, including group testing and online formats, apply a time limit of roughly 40 to 60 minutes for the full 60-item SPM to keep conditions consistent across test-takers. Items are not negatively marked, so there is no penalty for an incorrect guess.
Raw RPM scores are the simple count of items answered correctly. That raw score is then converted to a percentile using age-based norm tables built from large reference samples — the same conversion logic used across most standardised cognitive tests. A raw score of 48 out of 60 might correspond to the 75th percentile for one age band and the 60th for another, since raw performance naturally shifts with age.
Percentiles from a norm table can then be mapped onto a standard IQ-style scale (mean 100, SD 15) for easier interpretation alongside other cognitive measures. For the complete percentile-to-score mapping, see the IQ score chart and IQ score ranges guide.
Because percentiles are relative to a specific reference sample, the same raw score can map to different percentiles depending on when and where the norms were collected. This is the same dynamic behind the Flynn Effect, the well-documented rise in raw test scores across the 20th century. Fluid-reasoning tasks like matrix items showed some of the largest Flynn Effect gains of any test category, which is why current norm tables are re-standardised periodically.
1936
Year first published
Developed by psychologist John C. Raven as a non-verbal alternative to language-dependent reasoning tests.
60
Items in the Standard form
Arranged across five sets (A–E) of 12 items each, with difficulty increasing within and across sets.
5
Core rule families
Nearly every matrix item is built from rotation, progression, overlay, distribution, or addition/subtraction rules.
Raven's Progressive Matrices (RPM) is a non-verbal reasoning test developed by psychologist John C. Raven in 1936. Each item shows a grid of related visual patterns with one piece missing, and the test-taker selects which of several options correctly completes the pattern. It is designed to measure fluid intelligence — the ability to identify relationships and solve novel problems without relying on prior knowledge, vocabulary, or formal education.
RPM measures a specific component of general cognitive ability — fluid reasoning — rather than the full construct captured by comprehensive batteries like the WAIS-IV, which also assess verbal comprehension, working memory, and processing speed. RPM scores correlate strongly with general intelligence (g) and are frequently used as a proxy for IQ in research settings, but a psychologist administering a full clinical assessment would typically treat it as one component measure rather than a complete standalone IQ test.
It depends on the version. The Standard Progressive Matrices (SPM) has 60 items across five sets (A through E) of 12 items each, with difficulty increasing within and across sets. The Coloured Progressive Matrices (CPM), designed for young children and older adults, has 36 items across three sets. The Advanced Progressive Matrices (APM), designed to differentiate among above-average adults, has 48 items across two sets. IQMog's assessment uses a fixed 20-item Raven-style form rather than reproducing the full clinical item bank.
In its original clinical design, the SPM is untimed and administered as a power test, meaning the goal is to measure how far a person can progress through increasingly difficult items rather than how fast they respond. In practice, most modern administrations — including group and online settings — apply a time limit, typically 40 to 60 minutes for the full SPM, to standardise conditions across test-takers.
Familiarity with the item format produces measurable score gains, particularly for first-time test-takers. Research on practice effects shows repeated exposure to matrix-style reasoning tasks improves scores by recognising common rule families — rotation, size progression, and logical overlay being the three most common — faster under time pressure. This reflects improved pattern-recognition strategy and pacing rather than a change in underlying fluid reasoning capacity. See the full Raven's matrices strategy guide for a breakdown of every common rule type.
IQMog's assessment is built in the Raven tradition — a fixed-form, non-verbal matrix reasoning test with a consistent scoring model. Free to start, with your score and percentile band available immediately after the final item.