IQ test questions fall into six families: verbal, fluid, spatial, working memory, processing speed, and quantitative. See what each format looks like.
Dr. Russell T. WarneChief Scientist
Share
Real IQ test questions fall into six families: verbal reasoning, fluid reasoning, spatial ability, working memory, processing speed, and quantitative reasoning. The actual items on professional tests are secure and unpublished, so anything posted online is practice material or a leak, never the real test. This article walks through each family's format, the ability it measures, which real tests use it, and what separates an IQ question from a brain teaser.
One note first: every example item below was written for this article to illustrate a format. None comes from any published test. Modern batteries follow the "Cattell-Horn-Carroll" model, which places a general ability factor above several broad abilities, and each question family samples one broad ability with many short items.
Verbal reasoning questions
Verbal items measure "comprehension-knowledge," the store of words and facts you have acquired and your ability to reason with it. The common formats are vocabulary, general information, similarities (how are two things alike?), and verbal analogies.
• Vocabulary: "Candid" most nearly means: (A) hidden (B) frank (C) clever (D) cautious. Answer: B.
• Similarities: In what way are a thermometer and a ruler alike? Full-credit answers name the category: both are measuring instruments.
• Analogy: Sapling is to tree as cub is to: (A) den (B) bear (C) fur (D) forest. Answer: B.
On the WAIS-5 this family appears as the Vocabulary and Similarities subtests, the Stanford-Binet 5 covers it under its Knowledge factor, and the RIOT IQ test's Verbal Reasoning index uses Vocabulary, Information, and Analogies.
Fluid reasoning questions
Fluid reasoning items pose novel problems that prior knowledge cannot solve. The best-known format is "matrix reasoning": an incomplete grid of abstract shapes, completed by working out the relationships among them. Number series and balance-scale logic belong to the same family, which sits closest to the general factor IQ tests estimate.
• Number series: What comes next: 2, 6, 18, 54, __? Answer: 162 (multiply by three).
• Matrix: Row one: one, two, three dots. Row two: one, two, three squares. Row three: one triangle, two triangles, and a blank. The missing cell is three triangles.
• Figure weights: A circle balances two triangles, and a triangle balances three squares. How many squares balance one circle? Answer: six.
The WAIS-5 measures this family with Matrix Reasoning, Figure Weights, and Set Relations; the RIOT IQ test's Fluid Reasoning index uses Matrix Reasoning, Visual Puzzles, and Figure Weights. Raven's Progressive Matrices is built from the matrix format alone, and Pearson describes the current edition as minimizing "the impacts of language skills and cultural differences." Our guide to inductive reasoning tests goes deeper on this family.
Spatial questions
Spatial items ask you to generate, hold, and transform mental images, the ability testers call visual-spatial processing. Typical formats include mental rotation, paper folding, and block problems.
• Mental rotation: Rotate a capital "M" by 180 degrees and it most resembles a "W."
• Block counting: A 3×3×3 cube built from 27 small blocks is painted on every outside face. Exactly one, the center block, has no paint.
The WAIS-5 gives this family its own Visual Spatial index through Block Design and Visual Puzzles. The Stanford-Binet 5 tests it as Visual-Spatial Processing, and the RIOT IQ test's Spatial Ability index uses three subtests, including object rotation and spatial orientation tasks.
Working memory questions
Working memory items measure holding information in mind while actively working on it, storage plus processing rather than rote recall. The classic format is the "digit span" task: repeat a string of digits forward, backward, or reordered.
• Digits reversed: Hear 4, 9, 1, 7 once, then repeat the digits in reverse: 7, 1, 9, 4.
• Computation span: Solve 5+2, 9−3, and 4+4 aloud, then recall your three answers in order: 7, 6, 8.
On the WAIS-5 the family runs through Digit Sequencing and the new Running Digits (the "Digit Span" subtest name belongs to earlier editions), plus the visual Symbol Span and Spatial Addition. The Stanford-Binet 5 tests it verbally and nonverbally, and the RIOT IQ test's Working Memory index uses Visual Reversal, Computation Span, and Exposure Memory.
Processing speed questions
Processing speed items are individually trivial. The score is how many you finish quickly and accurately, often in about two minutes. Formats include symbol search, coding (fill in symbols under a row of digits using a key), and rapid comparison.
• Symbol search: Target: a star. Row: diamond, cross, star, triangle, circle. The target is present.
• Comparison: Identical or different? 58371294 versus 58372194. Different (two digits are swapped).
The WAIS-5 measures this family with Symbol Search, Coding, and Naming Speed Quantity. The RIOT IQ test's Processing Speed index uses Symbol Search and Abstract Matching.
Quantitative questions
Quantitative items test reasoning with numbers, usually mental arithmetic word problems done without a calculator or scratch paper, so they lean on working memory too.
• Arithmetic: A jacket costs $40 after a 20 percent discount. The original price was $50.
• Comparison: Which is larger, 3/7 or 0.45? The decimal: 3/7 is about 0.43.
The WAIS-5 tests this family with its Arithmetic subtest, and the Stanford-Binet 5 gives quantitative reasoning its own factor. On the RIOT IQ test, quantitative content lives inside subtests such as Computation Span and Figure Weights rather than in a separate index.
The RIOT IQ test (the Reasoning and Intelligence Online Test) adds a sixth index most paper-era batteries omit: Reaction Time, simple and choice reaction tasks timed to the millisecond, which online administration makes practical. All six indices and fifteen subtests are documented on the RIOT IQ assessments page, and you can try the format with the free IQ test.
What makes a good IQ test question
Test developers judge items with statistics. An item's "difficulty" is the percentage of examinees who answer it correctly, and a battery needs a spread of difficulties to measure the whole ability range. Its "discrimination" is how well it separates higher scorers from lower scorers; common rules of thumb treat values at or below 0.20 as poor and 0.35 or above as excellent. A clever riddle that able people tend to miss is, by that standard, a bad item. The American Psychological Association's ethics code requires developers to use "appropriate psychometric procedures and current scientific or professional knowledge" in test design, standardization, and validation.
The same code explains the secrecy: Standard 9.11 directs psychologists to make reasonable efforts to maintain the security of test questions and materials, the International Test Commission publishes dedicated test-security guidelines, and published nonverbal tests are typically protected by copyright. Secrecy protects the meaning of the score. A meta-analysis of 107 samples and 134,436 test takers found that retesting raises cognitive scores by about a quarter of a standard deviation, with larger gains from coaching or identical forms. A score inflated by familiar items no longer means what the test's norms say it means.
Why all six families correlate
In 1904 Charles Spearman published the observation that scores on very different mental tasks correlate positively, and explained the pattern with a single general factor, "g." A vocabulary question and a matrix puzzle look nothing alike, yet performance on one predicts performance on the other. A 2010 review in Nature Reviews Neuroscience grounds g in biology, tying general intelligence to additive genetic influences and the efficiency of parieto-frontal brain pathways.
That is why a real IQ score aggregates many items across families. The WAIS-5 reports five index scores, the RIOT IQ test reports six, and no single question type is "the" IQ question. It is also why a quick IQ test built from one or two families can rank you roughly but cannot map your cognitive profile.
Common misconceptions about IQ test questions
• Riddles and brain teasers are IQ questions: No professional battery uses trick riddles; real items are standardized formats with known difficulty and discrimination statistics.
• An online quiz can give you an IQ: Without norms and a reported margin of error, a score is a raw count of correct answers. A genuine IQ is a comparison against a normed reference group.
• You can find real IQ test questions online: Professional ethics and copyright keep real items secure. Freely posted "real" items are fakes, practice clones, or leaks, and practicing leaked items inflates a score without raising ability.
• One question can reveal your IQ: A single item is a small, noisy sample. Reliable measurement comes from many items across all six families.
The way to see professionally built, normed questions is a real test. Our guide to IQ tests for adults explains how to get a score you can trust.
Frequently asked questions
Are online IQ test questions real?
Almost never. Real items are protected by test-security standards and copyright; what circulates online is practice material in the same style, and a quiz without norms cannot produce a genuine IQ.
Can you practice IQ test questions?
You can practice the formats, and it moves scores: retesting gains average about a quarter of a standard deviation, more with coaching or identical items. The gain is familiarity rather than higher intelligence.
Why do IQ tests use matrix puzzles?
Matrix items measure fluid reasoning with almost no language, which is why Raven's Progressive Matrices relies on them exclusively and can be used with nonverbal test takers.
How many questions are on an IQ test?
It varies, and publishers count subtests rather than questions. The WAIS-5 yields a full-scale IQ from seven subtests in about 45 minutes; the full RIOT IQ test runs fifteen subtests in about 52 minutes. Each subtest holds many short items.
References
1. American Psychological Association. (2017). Ethical principles of psychologists and code of conduct (2002, amended 2010, 2016). apa.org
2. Chierchia, G., et al. (2019). The matrix reasoning item bank (MaRs-IB): Novel, open-access abstract reasoning items for adolescents and adults. Royal Society Open Science, 6(10), 190232. doi.org/10.1098/rsos.190232
3. Deary, I. J., Penke, L., & Johnson, W. (2010). The neuroscience of human intelligence differences. Nature Reviews Neuroscience, 11(3), 201-211. doi.org/10.1038/nrn2793
4. Hausknecht, J. P., Halpert, J. A., Di Paolo, N. T., & Moriarty Gerrard, M. O. (2007). Retesting in selection: A meta-analysis of coaching and practice effects for tests of cognitive ability. Journal of Applied Psychology, 92(2), 373-385. doi.org/10.1037/0021-9010.92.2.373
5. International Test Commission. (2016). The ITC guidelines on the security of tests, examinations, and other assessments. International Journal of Testing, 16(3), 181-204. doi.org/10.1080/15305058.2015.1111221
9. Rezigalla, A. A., et al. (2024). Item analysis: The impact of distractor efficiency on the difficulty index and discrimination power of multiple-choice items. BMC Medical Education, 24, 445. pmc.ncbi.nlm.nih.gov
10. Spearman, C. (1904). "General intelligence," objectively determined and measured. The American Journal of Psychology, 15(2), 201-292. doi.org/10.2307/1412107
Take our professional IQ test
Want to know your IQ? Try the first ever professional online IQ test.
