Raven’s Progressive Matrices Explained: How They Work and What They Measure

What Are Raven’s Progressive Matrices?

Raven’s Progressive Matrices are nonverbal reasoning assessments built around visual patterns with a missing piece. Instead of asking for a vocabulary definition or a calculation, a question asks you to inspect the arrangement and choose the option that completes it. This makes matrix problems a useful way to explore abstract and relational reasoning. It does not make them a culture-free window into someone’s total intelligence.

You may encounter Raven’s-style puzzles on learning sites, in preparation materials or in discussions about IQ. However, a puzzle that looks like a Raven item is not necessarily the officially standardized test. The edition, administration method and score interpretation matter. This guide explains the original assessment, common forms, how to approach visual matrices and what a result can—and cannot—tell you.

History: John C. Raven and the 1938 Assessment

British psychologist John C. Raven developed the original Progressive Matrices during the 1930s. The Smithsonian’s National Museum of American History documents the 1938 published test, which consisted of 60 visual problems in five sets of 12, labelled A through E. Historical materials also included instructions, an answer key and early norm information. This is a more defensible description than assigning the date of the standardized publication to 1936.

The Wellcome Collection archive records later revisions and norm editions. Over time, the Raven family of assessments expanded beyond its initial format. When comparing scores or articles about the test, it is therefore useful to identify which version the writer means, instead of assuming every “Raven test” is the same instrument.

How the Missing-Piece Pattern Works

A typical item presents a visual arrangement with a blank area and several candidate pieces. You need to infer a rule that links the surrounding figures. Many familiar practice puzzles are arranged in a 3 × 3 grid, but that is not a universal format: the historical Standard Progressive Matrices also used simpler arrangements, including 2 × 2 patterns.

Useful relationships to investigate include shape count, direction, rotation, reflection, position, shading and combinations of elements. For instance, in a made-up practice example, the shapes in each row might rotate a quarter-turn between columns. The missing shape should then continue that row’s rule while also remaining consistent with the other rows. The point is not to memorize an answer but to test a proposed pattern against the entire arrangement.

As items become more complex, a single attractive pattern may be misleading. An explanation should account for the row and column relationships that matter, rather than relying on visual resemblance. Correct answers are determined by the test’s actual item rules and scoring key—not by a universal trick that works for every matrix.

Different Raven Test Versions

Published versions vary in form and difficulty. The terms below describe recognized members of the assessment family; they are not levels in one free online game.

For edition-specific details, see Pearson’s Raven’s 2 product documentation and the SAGE Encyclopedia of Measurement and Statistics overview.

What Does Raven’s Actually Measure?

Raven tasks are often described as measures of nonverbal or fluid reasoning: identifying relationships, keeping several visual rules in mind and applying them to unfamiliar patterns. The historical theory of general cognitive ability, often discussed using g, influenced the design of the instrument. A matrix score can be associated with broader cognitive performance, but that does not mean it is a complete measurement of vocabulary, memory, creativity, social understanding or every other ability.

Some articles claim that Raven scores predict academic or job performance with a fixed correlation such as r = 0.5 or r = 0.4. A single context-free number is misleading: relationships vary with the population, version, outcome measures and study design. This guide therefore does not present one made-up coefficient as the guaranteed predictive power of all Raven forms.

To understand why different abilities are not identical, read our guide to Fluid vs Crystallized Intelligence. It explains how reasoning with new information differs from knowledge developed through prior learning.

Nonverbal Does Not Mean Culture-Free

Matrix puzzles need fewer spoken or written language responses than many vocabulary-heavy tests. This may be useful in some multilingual or communication-related assessment settings. But less verbal does not mean unaffected by culture or experience. Familiarity with diagrams, school tasks, visual conventions, prior testing, attention and the quality of instructions may influence what a score represents.

The official publisher notes that Raven’s 2 seeks to minimize certain language and cultural influences; that is a narrower and more appropriate claim than stating these influences have been eliminated. Testing across language and cultural groups still requires careful study of interpretation and norms. It is not sound to infer “raw intelligence” or compare unrelated populations from a casual browser quiz.

Raw Scores, Percentiles and IQ Numbers

The most direct result of a matrix test is the number of items answered correctly under its defined scoring rules. To interpret that result as a standard score or percentile, a validated assessment uses an appropriate reference population and scoring method. The relevant manual matters, including considerations such as the edition, age group and administration conditions.

A percentile tells you about position relative to a specified comparison group; it does not say what percentage of problems you solved. Likewise, 80% correct on an informal matrix quiz is not automatically an IQ of 120. A publisher-supported Raven result and a hobby quiz’s accuracy percentage belong to different categories of evidence.

For a wider discussion of normed scores, see What Is a Good IQ Score? and Free IQ Test Online: What to Expect. Neither article should be read as a substitute for professional assessment guidance.

How to Approach Raven’s-Style Practice Puzzles

These steps can help you approach a new visual-reasoning puzzle systematically. They are practice strategies, not leaked official answer keys or a guarantee of a higher IQ:

1. Scan the whole arrangement. Identify what remains the same and what changes across the visible positions.
2. Check one relationship at a time. Look for shifts in direction, shape count, size, shading or orientation.
3. Compare rows with columns. See whether one consistent rule explains both directions, or whether different rules operate independently.
4. Test your rule on visible cells. A good hypothesis should explain more than the single missing position.
5. Eliminate incompatible options. Remove candidates that break a rule that holds elsewhere in the matrix.
6. Revisit your assumption if nothing fits. You may have missed a second rule, combined two features or interpreted the movement incorrectly.
7. Review explanations after practice. Learning why a solution works can improve familiarity with these puzzles, even when it does not establish a change in general intelligence.

For independent practice, start with simple transformations and then try multi-rule visual patterns. Avoid memorizing copies of protected standardized test questions; official assessment materials have intended administration procedures and test-security considerations.

Can Practice Make You Better at Matrix Problems?

It can help you understand a task format, become less surprised by common transformations and recognize mistakes in your own reasoning. Such improvements are meaningful for a learning activity. But retaking the same items can also create familiarity and practice effects, making later scores difficult to interpret as independent evidence of improved cognitive ability.

To track practice honestly, record the name of the activity, question set or version, date and whether you have seen the items before. Avoid equating a quick improvement on a familiar online game with a verified increase in general intelligence. If an assessment score will inform clinical, educational or another consequential decision, seek the appropriate qualified professional and standardized procedure.

Try Visual Reasoning on ReflexBenchmark

Our Pattern Recognition Test contains 20 visual logic puzzles involving patterns and transformations. It is a free, informal browser activity that lets you practise noticing visual rules. It is not an official Raven’s form, uses no Raven-provided scoring norms, and should not be described as a medical or professional IQ assessment.

For additional reasoning practice with different question types, see the 20-question Mini IQ-Style Reasoning Quiz. It gives a raw score out of 20 rather than a clinically standardized IQ result. Trying both can be a useful way to explore how visual puzzles differ from verbal, numerical and logical questions without overstating what they measure.

Research and Primary Sources

1. Smithsonian Institution, National Museum of American History. Psychological Test, Raven Progressive Matrices. Museum documentation of the 1938 original and its five sets.
2. Wellcome Collection. Raven, J. C. — Standard Progressive Matrices, Sets A–E. Archival record of revisions, editions and norms.
3. Raven, J. C. (1941). Standardization of Progressive Matrices, 1938. British Journal of Medical Psychology, 19(1), 137–150.
4. Pearson Clinical Assessment. Raven’s Progressive Matrices, Second Edition (Raven’s 2). Publisher description of forms, updated norms and score interpretation.
5. Alexopoulos, D. S. (2007). Raven’s Progressive Matrices. Encyclopedia of Measurement and Statistics, SAGE.