Non-verbal reasoning test

What non-verbal reasoning is

The APA Dictionary defines a nonverbal test as one in which neither the problems nor the answers are given in words (APA Dictionary: nonverbal test). Non-verbal reasoning is the skill those tests target: working out rules from shapes, positions, counts and turns. It is the counterpart of verbal intelligence, which runs on vocabulary and language; that page covers the verbal side, and this one is where you take the shape-only test and practice it.

Removing words does not change what is being measured as much as people expect. The 1994 “Mainstream Science on Intelligence” statement said that while there are different types of intelligence tests, they all measure the same intelligence, some using words or numbers and others using shapes or designs, and that people who do not understand English well can be given a nonverbal test (Gottfredson, 1997). Most nonverbal items are tests of fluid reasoning, which Raymond Cattell described as reasoning with novel material rather than acquired knowledge (Wikipedia: Fluid and crystallized intelligence). The best-known example, Raven's Progressive Matrices, first published in 1938, is widely used as a measure of fluid, culture-reduced reasoning (Wikipedia: Raven's Progressive Matrices).

“Culture-reduced” is the honest word. The APA Dictionary describes culture-fair tests as using nonverbal, nonacademic items such as choosing a design that completes a series, and cautions that every assessment reflects some cultural norms, so no test is fully culture-free (APA Dictionary: culture-fair test). David Lohman goes further and argues that most nonverbal tests still rely on verbally mediated thinking and are neither culture free nor culture fair (Lohman, 2005). You will notice this yourself: on a hard grid, most people talk themselves through the rules (“one more each time, solid then outlined”).

So why test without words at all? Because for some people, words get in the way of seeing their reasoning. The publisher of the Leiter-3 says nonverbal testing is essential for people with autism or speech and language disorders, because verbal deficits can hide their cognitive strengths and weaknesses (Stoelting: Leiter-3). The same logic applies to anyone tested outside their first language.

How non-verbal reasoning is measured

General IQ batteries split out a nonverbal score. The Stanford-Binet 5 is organized into two domains, Verbal and Nonverbal (PRO-ED: SB5). The WAIS-5 includes a Nonverbal Index among its ancillary scores (Pearson: WAIS-5 overview), and the WISC-V reports Visual Spatial and Fluid Reasoning indexes alongside Verbal Comprehension (Pearson: WISC-V). Short batteries do the same: the KBIT-2 reports a Fluid (Nonverbal) standard score (Pearson: KBIT-2), and the RIAS-2 a Nonverbal Intelligence Index (PRO-ED: RIAS-2).

Dedicated nonverbal batteries go all the way. The UNIT2 is designed for people with speech, language or hearing impairments, different language backgrounds, or who are verbally uncommunicative, and has six subtests including Analogic Reasoning, Numerical Series and Cube Design; its coefficient alpha is .88 to .96 for subtests (PRO-ED: UNIT2). The Leiter-3 uses language-free administration for ages 3 to 75 and older (Stoelting: Leiter-3). The TONI-4 lets people answer by pointing, nodding or blinking, and its 60 abstract figural items vary on eight features (shape, position, direction, rotation, contiguity, shading, size and movement), with harder items combining more of them (PRO-ED: TONI-4). The CTONI-2 measures analogical reasoning, categorical classification and sequential reasoning with pictures and geometric designs (PRO-ED: CTONI-2). Those three kinds of reasoning map closely onto the analogy, odd-one-out and sequence items in the test above.

Schools meet nonverbal tests mostly through gifted programs. The NNAT3 is a group test for kindergarten through grade 12 with 48 items in 30 minutes (Pearson: NNAT3), and its publisher describes it as a nonverbal measure of general ability that also serves gifted identification, especially for underrepresented groups (Buros NNAT3 listing). Each item is a pattern with a missing section and five options, and Pearson calls it a power test: students given extra time on the previous edition did not score higher (NNAT3 Manual A to D). It uses only yellow and blue so that students with minimal color vision impairments can take it (Pearson NNAT3 brochure), and Naglieri describes its aim as measuring how well students think without that being confounded by how much they know (Pearson NNAT3 infographic). The CogAT Nonverbal Battery has three subtests: Figure Matrices, Paper Folding and Figure Classification (Riverside Nonverbal Battery help). The OLSAT 8 nonverbal side has Pictorial Reasoning and Figural Reasoning clusters, the latter with Figural Classification, Figural Analogies, Pattern Matrix and Figural Series items (Pearson OLSAT 8 brochure).

These tests are useful, not interchangeable. Lohman, Korb and Lakin gave the Raven, the NNAT and CogAT Form 6 to 1,198 elementary children, about 40 percent of them English learners (Lohman, Korb and Lakin 2008). U.S. national norms for the Raven substantially overestimated the number of high scorers, norming errors led the NNAT to overestimate both high and low scorers, its standard error of measurement was twice that of the Raven or CogAT, and none of the three predicted achievement very well for English learners. The study predates the 2016 NNAT3, so it does not directly describe today's editions. A 2010 study co-authored by Naglieri found the NNAT and the Raven's Colored Progressive Matrices had moderate validity for math and reading in Italian third and fifth graders, and that their predictive validity did not depend on students' sociocultural level (Balboni, Naglieri and Cubelli 2010). For a side-by-side view, see types of IQ tests.

The classic shape-only tests are still around. Raven's Progressive Matrices come in three forms, from an easy level for young children to an advanced level that spreads the scores of the top 20 percent, and are untimed in Pearson's classic line (Pearson: Raven's brochure). The Cattell Culture Fair test was designed to measure intelligence with minimal influence from verbal fluency and education (Buros: CFIT), using four subtests: series, classification, matrices and typology (Zahra et al., 2022). At work, Criteria's UCAT leaves out verbal ability, which makes it easy to translate (Criteria: UCAT), and the CCAT includes spatial reasoning questions (Criteria: CCAT prep). British Mensa's online test includes a non-verbal diagrammatic fluid reasoning test (British Mensa: online IQ test).

The ten questions above sample five formats from the full IQ Octopus test: two shape sequences, three 3x3 matrices, two mental rotation items, two odd-one-out sets and one figure analogy, ordered from easy to hard. Sequences and matrices are series and matrix reasoning, formats that recur across the batteries above. Rotation is spatial reasoning, odd one out is classification, and the analogy is analogical reasoning.

Worked examples: easy, medium and hard

The easy example is mental rotation. One option is the original figure turned, and every other option is a mirror image. Pick a landmark, such as the marked cell or the end of the longest arm, turn it in your head, then check that the part next to it is still on the same side.

The medium example is a matrix with no progression. Each shape and each fill appears once in every row and once in every column, in shuffled order. Find which shape and which fill the bottom row is missing, then confirm with the right-hand column. Also note the attribute that never changes.

The hard example is a sequence with three changes running at once: the arrow turns, the fill cycles, and a small dot moves around the edge. Track each change on its own across the whole row, then apply all three to the last tile. Most wrong options get two of the three right.

Common mistakes on non-verbal items

• Not resetting between formats. After two sequences your eyes compare neighbors; a matrix needs rows and columns, and a rotation item needs one landmark. Read the prompt above each puzzle to know which method applies.
• Reacting to surface differences. Size, tilt and fill are often there to distract. On odd-one-out sets, the figure that looks most different is rarely the answer; the rule is a side count or where a dot sits.
• Choosing a mirror image. On rotation items every wrong option looks nearly identical. Ask whether a turn can reach it, not whether it resembles the original.
• Stopping one rule short. Harder sequences, matrices and analogies stack two or three changes. Wrong options satisfy all but one, so stopping at the first rule leaves you choosing between near twins.
• Getting direction backward. Clockwise versus counterclockwise, and a turn versus a flip, are the easiest things to reverse. Trace a single point with your eye to settle it.
• Forgetting the constants. If every tile in a grid is large, the answer is large too. What does not change is part of the rule.
• Scanning the options first. They are designed to look reasonable. Describe the answer before you look, then match it.

How to practice non-verbal reasoning

Because this test mixes formats, practice them one at a time and then together:

1. Sequences. Start with the easy sequence puzzles, naming every change between neighboring tiles, then move to the hard sequences with three changes at once. The pattern recognition guide explains the method.
2. Matrices. Work through the easy, medium and hard matrix puzzles with the matrix reasoning guide open.
3. Rotation and odd one out. The rotation library and the odd-one-out library train the landmark method and the say-the-rule habit.
4. Analogies. A few figure analogies teach you to tell flips from turns and to watch for two changes at once.
5. Mix and time it. Retake this test, then try the culture fair test practice and the spatial reasoning test with the 30-second clock.
6. Read about the limits. The guide to culture-fair tests explains what shape-only testing can and cannot promise.

Familiarity helps more than people think, which is also a fairness issue: someone who has seen these formats before starts ahead. Retaking a cognitive ability test raises scores by about a quarter of a standard deviation on average, more on the same form (Hausknecht et al., 2007), and a meta-analysis of 174 samples found no further gains after the third sitting (Scharfen, Peters & Holling, 2018). A practice set like this closes that gap cheaply before a test that counts.

Beyond the format, the evidence splits. Spatial training works and spreads within its own area: a meta-analysis of 217 studies found an average gain of g = 0.47 that lasted and transferred to untrained spatial tasks (Uttal et al., 2013). Broad brain training is weaker; reviews of chess, music and working memory training find effects shrink as study quality rises and far transfer rarely occurs (Sala & Gobet, 2017). Practice makes you better at these puzzles. It does not make you smarter overall.

Non-verbal reasoning on the full IQ Octopus test

The full test has 25 items in five sections: Pattern Recognition (5), Abstract Reasoning (9), Spatial Reasoning (5), Quantitative Reasoning (3) and Logical Reasoning (3). Apart from the three number series in Quantitative Reasoning, every item is a shape puzzle: sequences, matrices, rotation and odd one out, and figure analogies, the same five formats you just met. Items are generated fresh for every attempt.

Your score above is a count out of ten, not an IQ, and it cannot be used for any school, job or membership application. Real IQ scores compare you with people your own age on a scale with a mean of 100 and a standard deviation of 15 (Wikipedia: Intelligence quotient). The full test scores all 25 answers with a Rasch model and reports an estimated IQ with a confidence range. When you are ready, take the full IQ test.