Cognitive | Fluid Intelligence
Raven's Progressive Matrices
A 3×3 grid of patterns — the bottom-right cell is missing. Pick which of 6 options completes the pattern. 12 problems. 10-minute timer. The gold standard for culture-fair fluid IQ.
Save your fluid intelligence score
Track your Raven's Matrices performance over time with a free account.
Raven's Progressive Matrices
Raven's Progressive Matrices were developed by John C. Raven in 1936 at University College London. Raven designed the test to measure eductive ability — the ability to extract meaning from complex information and derive novel insights — which he identified as the core of general intelligence. The matrices were specifically designed to minimize the influence of education, language, and cultural knowledge, making them the most widely used culture-fair intelligence measure in the world.
The original SPM (Standard Progressive Matrices) contains 60 items arranged in five sets of 12, increasing in difficulty. Variants include the Coloured Progressive Matrices (CPM, for children and elderly) and Advanced Progressive Matrices (APM, for high-ability adults). The test is used clinically, educationally, and in occupational settings across 180+ countries and has been translated into over 40 languages — though translation is largely irrelevant since the test contains no words at all.
The test has one of the highest g-loadings of any psychometric instrument — meaning it correlates more strongly with general intelligence than almost any other single test. This is why it is the primary reference test used in research on the Flynn Effect — rising IQ scores over generations.
Why Raven's Is Culture-Fair
Most intelligence tests require language comprehension, cultural knowledge, or familiarity with specific academic content — meaning they measure a combination of intelligence and educational opportunity. Raven's matrices avoid this by using only abstract geometric patterns with no labels, no words, and no culturally specific content. All you need is the ability to recognize and reason about visual relationships.
Zero text — works identically for all speakers
Requires only abstract pattern reasoning
Simple choice format learned in seconds
Even with these advantages, some cultural and socioeconomic factors still influence Raven's scores — primarily through nutrition, stress, test anxiety, and access to cognitively stimulating environments in childhood. "Culture-fair" means less biased, not unbiased. For complementary assessment, the Cognitive Reflection Test measures analytical thinking disposition, which is a distinct dimension of cognitive ability.
Score Distribution
Accuracy on 12-problem version
| Score (/12) | Percentile | IQ Estimate | Classification |
|---|---|---|---|
| 12 | Top 5% | 130+ | Very Superior |
| 10–11 | Top 20% | 115–129 | Superior |
| 7–9 | 25th–75th | 90–114 | Average |
| 5–6 | 75th–90th | 80–89 | Low Average |
| <5 | Bottom 10% | <80 | Below Average |
Fluid vs Crystallized Intelligence
Raymond Cattell's influential distinction between fluid intelligence (Gf) and crystallized intelligence (Gc) — developed in the 1940s and formalized with John Horn in the 1960s — describes two broad dimensions of cognitive ability that follow different developmental trajectories.
Fluid Intelligence (Gf)
The ability to reason with novel problems, identify patterns, and draw inferences — independent of accumulated knowledge. Peaks in the mid-20s, then slowly declines with age. Raven's matrices are the defining Gf measure. Also tapped by our Mental Rotation Test.
Crystallized Intelligence (Gc)
Knowledge, vocabulary, and expertise accumulated through experience and education. Continues growing well into the 60s and 70s in intellectually active individuals. Measured by vocabulary tests, general knowledge, and verbal analogies.
As people age, the typical pattern is declining fluid intelligence offset by increasing crystallized intelligence. This is why older adults often outperform younger people on domain-specific expertise tasks while performing below their younger selves on pure pattern reasoning — exactly the skill measured here.
How to Improve Pattern Recognition
Systematic practice of matrix problems
Regular practice with progressive matrix problems — starting with simpler patterns and advancing to more complex ones — demonstrably improves scores on both Raven's and transferred pattern recognition tasks. The key is learning to systematically check multiple dimensions (shape, size, fill, rotation, count) rather than relying on intuition. 20 minutes of practice three times per week shows measurable improvement within 6 weeks.
Aerobic exercise and sleep
Fluid intelligence is more sensitive to physiological condition than crystallized intelligence. Aerobic exercise at 60–70% of VO2max for 30+ minutes raises BDNF levels and improves performance on Gf tasks for 1–2 hours post-exercise. Adequate sleep is equally critical: REM sleep consolidates the abstract pattern representations that underlie matrix reasoning. Just one night of poor sleep reduces Raven's scores by 10–15%.
Learn to work systematically
Most errors on Raven's come from checking too few dimensions. Expert test-takers systematically check: (1) what changes across each row, (2) what changes down each column, and (3) whether there is a diagonal rule. Then they verify their answer by checking that it satisfies all identified rules. This explicit strategy transfers to other matrix problems and reduces errors by 25–40% compared to purely intuitive approaches.
Engage in varied abstract reasoning
Chess, programming, mathematical problem-solving, logic puzzles, and strategy games all exercise the abstract relational reasoning circuits that underlie Raven's performance. Near-transfer from abstract puzzle training is consistently observed in the cognitive training literature. The key is novelty — problems that require building new mental representations rather than applying memorized procedures.
Track Your Fluid Intelligence
Create a free account to save your Raven's score and monitor changes in pattern recognition ability over time.