A Short History of the Progressive Matrices
John C. Raven developed the test while working with Charles Spearman, whose two-factor theory proposed that a general reasoning ability underpins performance on many different tasks. Spearman had a practical problem: verbal tests could not be used to study reasoning across language groups. Raven's solution — matrices made entirely of abstract figures — produced one of the first genuinely culture-reduced reasoning tests.
The Standard Progressive Matrices was published in 1938 and has since been adapted into the Coloured Progressive Matrices for young children and the Advanced Progressive Matrices for adults of high ability. The 60-item SPM remains the version most often used in schools, in occupational testing and in large cross-cultural studies.
What Each of the Five Groups Measures
The five groups are not five separate skills. They are five levels of the same underlying capability: the ability to work out a rule from incomplete information and apply it consistently. Difficulty rises steadily from A to E because the number of rules operating at once increases.
| Group | Questions | Cognitive demand | Typical pattern |
|---|---|---|---|
| A | 1 – 12 | Perceptual discrimination | A single figure is missing; the pattern continues without change |
| B | 13 – 24 | Analogy and comparison | Two figures relate, and the same relationship must be matched elsewhere |
| C | 25 – 36 | Comparative reasoning | Elements accumulate or cancel across rows and columns |
| D | 37 – 48 | Serial relationships | Figures rotate, move or grow in a consistent sequence |
| E | 49 – 60 | Abstract reasoning | Two or more rules operate simultaneously, including rules about rules |
Group boundaries follow the standard 60-item booklet arrangement, 12 items per group.
Grade Reference Table
Raven's original scoring divides the population into five grades. Your grade describes where your score falls in the reference distribution, not an absolute measure of ability.
| Percentile | Grade | Level | What it means |
|---|---|---|---|
| ≥ 95 | I | Intellectually Superior | Top 5% of the reference group |
| 75 – 94 | II | Definitely Above Average | Clearly above the middle of the population |
| 25 – 74 | III | Intellectually Average | The band containing the majority of people |
| 5 – 24 | IV | Definitely Below Average | Slightly below the middle of the population |
| < 5 | V | Intellectually Defective | Bottom 5% — retest when rested before drawing conclusions |
Grade labels follow Raven's 1938 original terminology for historical accuracy. This online adaptation is intended for informal self-assessment only.
Raw Score to IQ Conversion
When the raw score is mapped onto the IQ scale, the standardisation is deliberately set so that the population mean is 100 and the standard deviation is 15. That single convention is what allows different reasoning tests to be compared against one another.
| Raw score (of 60) | Percentile | Approx. IQ | Grade |
|---|---|---|---|
| 55 – 60 | 99 | 135+ | I — Superior |
| 50 – 54 | 95 – 98 | 124 – 134 | I — Superior |
| 45 – 49 | 85 – 94 | 116 – 123 | II — Above Average |
| 40 – 44 | 72 – 84 | 108 – 115 | II — Above Average |
| 33 – 39 | 45 – 71 | 94 – 107 | III — Average |
| 26 – 32 | 22 – 44 | 84 – 93 | III — Average |
| 20 – 25 | 9 – 21 | 77 – 83 | IV — Below Average |
| 12 – 19 | 2 – 8 | 70 – 76 | IV — Below Average |
| 0 – 11 | < 2 | < 70 | V — Lower |
Approximate mapping for adult reference groups. IQ values are indicative only and assume a mean of 100 with a standard deviation of 15.
SPM Standard Score Conversion Table
Raw scores on the Standard Progressive Matrices are converted to percentile ranks using age norms, because reasoning ability develops rapidly through childhood. Find your age band in the first column, then read across to the percentile that matches your raw score of 60.
| Age | Standard Score (%) | |||||||
|---|---|---|---|---|---|---|---|---|
| 95 | 90 | 75 | 50 | 25 | 10 | 5 | 1 | |
| 5.5 | 34 | 29 | 25 | 16 | 13 | 12 | 9 | 7 |
| 6 | 36 | 31 | 25 | 17 | 13 | 12 | 9 | 7 |
| 6.5 | 37 | 31 | 25 | 18 | 13 | 12 | 10 | 8 |
| 7 | 43 | 36 | 25 | 19 | 13 | 12 | 10 | 8 |
| 7.5 | 44 | 38 | 31 | 21 | 13 | 12 | 10 | 8 |
| 8 | 44 | 39 | 31 | 23 | 15 | 13 | 10 | 8 |
| 8.5 | 45 | 40 | 33 | 23 | 20 | 14 | 12 | 8 |
| 9 | 47 | 43 | 37 | 33 | 25 | 14 | 12 | 9 |
| 9.5 | 50 | 47 | 39 | 35 | 27 | 17 | 13 | 9 |
| 10 | 50 | 48 | 42 | 35 | 27 | 17 | 13 | 9 |
| 10.5 | 50 | 49 | 42 | 39 | 32 | 25 | 18 | 9 |
| 11 | 52 | 50 | 43 | 39 | 33 | 25 | 19 | 9 |
| 11.5 | 53 | 50 | 45 | 42 | 35 | 25 | 19 | 9 |
| 12 | 53 | 50 | 46 | 42 | 37 | 27 | 21 | 9 |
| 12.5 | 53 | 52 | 50 | 45 | 40 | 33 | 28 | 9 |
| 13 | 53 | 52 | 50 | 45 | 40 | 35 | 30 | 9 |
| 13.5 | 54 | 52 | 50 | 46 | 42 | 35 | 32 | 9 |
| 14 | 55 | 52 | 50 | 48 | 43 | 36 | 34 | 9 |
| 14.5 | 55 | 53 | 51 | 48 | 43 | 36 | 34 | 9 |
| 15 | 57 | 54 | 51 | 48 | 43 | 36 | 34 | 9 |
| 15.5 | 57 | 55 | 52 | 49 | 43 | 41 | 34 | 9 |
| 16 | 57 | 56 | 53 | 49 | 44 | 41 | 36 | 9 |
| 16.5 | 57 | 56 | 53 | 49 | 45 | 41 | 37 | 9 |
| 17 | 58 | 57 | 55 | 52 | 47 | 40 | 37 | 9 |
| 20 | 57 | 56 | 54 | 50 | 44 | 38 | 33 | 9 |
| 30 | 57 | 55 | 52 | 48 | 44 | 37 | 28 | 9 |
| 40 | 57 | 54 | 50 | 47 | 41 | 31 | 28 | 9 |
| 50 | 54 | 52 | 48 | 42 | 34 | 24 | 21 | 9 |
| 60 | 54 | 52 | 46 | 37 | 30 | 22 | 19 | 9 |
| 70 | 52 | 49 | 44 | 33 | 26 | 18 | 17 | 9 |
Normative reference values for the Standard Progressive Matrices. Read across from your age to find the raw score required to reach each percentile. Percentile bands are the conventional 95th, 90th, 75th, 50th, 25th, 10th, 5th and 1st cut-offs.
How Raven Scores Compare Around the World
Because the Progressive Matrices were designed to be independent of language and schooling, they are the most widely used reasoning test in cross-cultural research. The table below summarises average IQ estimates reported for world regions. Treat these figures as rough population summaries rather than measurements of any individual.
| Region | Reported average IQ | Representative studies |
|---|---|---|
| East Asia | 105 – 108 | Japan, South Korea, China, Singapore, Taiwan |
| Europe | 97 – 102 | Western and Northern Europe cluster near 100; Southern and Eastern Europe slightly lower |
| North America | 98 – 100 | United States, Canada |
| Oceania | 98 – 100 | Australia, New Zealand |
| South America | 87 – 93 | Brazil, Argentina, Chile, Colombia |
| South & Southeast Asia | 85 – 92 | India, Indonesia, Philippines, Thailand, Vietnam |
| Middle East & North Africa | 84 – 91 | Turkey, Iran, Egypt, Saudi Arabia |
| Sub-Saharan Africa | 71 – 82 | Widely varying national estimates |
Averages for a whole region hide enormous within-region variation and are strongly affected by schooling, nutrition, health and test familiarity. The Flynn effect — roughly a 3-point rise per decade across the 20th century — means older norms understate present-day performance.
Country-Level Estimates
Individual country estimates are frequently quoted in the literature. The figures below are representative values from large meta-analyses; different sources vary by several points, so read them as a ranking aid rather than a precise number.
| Country | Est. IQ | Country | Est. IQ |
|---|---|---|---|
| Singapore | 108 | Poland | 99 |
| Hong Kong SAR | 108 | Canada | 99 |
| South Korea | 106 | Australia | 99 |
| Japan | 105 | United Kingdom | 99 |
| China | 105 | Germany | 99 |
| Taiwan | 106 | France | 98 |
| Italy | 102 | United States | 98 |
| Switzerland | 101 | Russia | 97 |
| Netherlands | 100 | Malaysia | 92 |
| Sweden | 100 | Mexico | 88 |
| New Zealand | 99 | India | 85 |
| Spain | 99 | Brazil | 87 |
Compiled from published national IQ estimates. The concept of a single national IQ figure is contested — sampling, urbanisation and education differ far more than the numbers suggest.
Answer Patterns
An often-overlooked feature of the Progressive Matrices is that the correct answer tends to concentrate in the middle of the option row for the easier groups and spread out for the harder ones. This is a deliberate design choice: it discourages guessing strategies based on position.
| Group | Difficulty | Median completion time | Cumulative time |
|---|---|---|---|
| A | Very easy | about 4 min | 4 min |
| B | Easy | about 5 min | 9 min |
| C | Moderate | about 7 min | 16 min |
| D | Hard | about 10 min | 26 min |
| E | Very hard | about 12 min | 38 min |
Typical timings for untimed group administration. Individual pacing varies widely and there is no time limit on this online version.
Strengths and Limitations
| Strengths | Limitations | |
|---|---|---|
| Design | Non-verbal, so language and schooling matter less | Measures only fluid reasoning, not knowledge or creativity |
| Use | Comparable across cultures and age groups | Cannot diagnose learning difficulties on its own |
| Administration | Short, cheap, and suitable for group testing | Practice effects are real, so repeated retakes inflate scores |
| Interpretation | Well-established norms spanning more than 80 years | Old norms require adjustment for the Flynn effect |
Frequently Asked Questions
Is this the official Raven's test?
No. This is a free online adaptation of the Progressive Matrices format. The official SPM is a copyrighted, controlled instrument that must be administered by a qualified professional.
Why is there no time limit?
Raven designed the test to measure power rather than speed. Untimed administration gives a purer reading of reasoning ability, which is also why it is preferred in cross-cultural research.
Can I practise to improve my score?
Familiarity with matrix puzzles does raise scores, mainly by teaching you to scan rows before columns and to eliminate options that break a single obvious rule. The gain is typically modest and does not generalise indefinitely.
Does a low score mean anything bad?
It means your performance on this particular task, on this particular day, fell below the middle of the reference group. Fatigue, distraction and an unfamiliar format all move the number. A retake when rested is far more informative than a single sitting.
Where to Go Next
If matrix reasoning suited you, the 72-question combined Raven and CRT test extends the same format with a larger item pool, and the Baron EQ assessment covers the emotional side of intelligence that matrix tests cannot reach.
