In short
An IQ test does not measure a quantity of intelligence. It measures how your performance on a fixed set of standardized tasks compares with the performance of other people your age, and reports that comparison on a scale built so the average is 100. A professional battery samples five domains, verbal comprehension, visual-spatial ability, fluid reasoning, working memory and processing speed; those subtest scores correlate with each other, which is the observation behind the general factor, and the total predicts group-level outcomes in education, work and health while saying much less about any one person. Psychology.com's IQ Test is a free reasoning quiz, not a clinical assessment, and this page is clear about the difference.
- What it actually measures
- Your performance against an age-matched comparison group
- The scale
- Average set at 100, standard deviation 15
- Domains sampled
- Verbal, visual-spatial, fluid reasoning, working memory, speed
- How it is given
- One to one by a qualified examiner, one to two hours
- What it does not sample
- Motivation, creativity, judgment, skill, character
- Test
- IQ Test: Free Reasoning Quiz
What an IQ score actually is
An IQ score is a rank expressed as a number. You sit a set of tasks, your answers are converted into points, and those points are compared with the points earned by a large sample of people your own age who sat the same tasks. The comparison is then rescaled so that the middle of the sample is 100 and a step of 15 points is one standard deviation. Nothing about the number describes a substance inside you. It describes where your performance fell in a line of other people's performances.
That is why the question "what does an IQ test consist of" has a more useful answer than "what is intelligence". The consisting-of part is concrete: a specified set of short tasks, given in a specified order, under specified conditions, scored against published reference data. Change the sample the score is compared against and the number changes without anything about you changing at all. This is also why a score has a date and an edition attached to it in any competent report.
The practical consequence is that a score is a measurement with a margin, not a reading off a dial. A professional report gives each score alongside a percentile and a confidence interval, because the same person tested twice will not produce the identical number. Treating the figure as exact is the single most common error readers make with their own results.
How the test is actually given
A clinical intelligence test is administered one to one. An examiner sits across from you, reads standardized instructions, times some tasks and not others, records your exact answers, and follows rules about when to stop a subtest and when to move on. That procedure is part of the measurement. Two people who answer identically but are prompted differently have not taken the same test, which is why the script matters as much as the questions.
The session is longer than most people expect. Our own result page tells readers to allow one to two hours for the core appointment, with the test battery itself sometimes shorter and a broader evaluation longer. The current Wechsler adult scale, published in 2024 for ages 16 through 90, is listed by its publisher at about 45 minutes for the seven subtests behind the full-scale score and about 60 minutes for the ten primary subtests.
Most batteries also use adaptive routing. Rather than giving every item to everyone, the examiner starts at a point chosen by age or by early performance, works outward, and stops after a set number of consecutive failures. That keeps the session short and keeps people out of long stretches of items that are far too hard or far too easy, and it is one of the features an online quiz cannot copy.
The five domains a modern battery samples
Current batteries group their subtests into five domains, each reported as its own index score alongside the total. Labels differ a little between publishers and the grouping has shifted across editions, but the five areas below are what the major instruments sample now. The tasks described here are generic examples of what each domain asks for, not items from any published test.
Verbal comprehension asks what you know and how precisely you can say it. The typical tasks are defining words, explaining how two things are alike, and answering general-knowledge questions. This domain draws heavily on what you have learned and read, which makes it sensitive to schooling and to whether the test is in your strongest language.
Visual-spatial ability asks you to work with shape and space: building a pattern from blocks to match a model, or picking which pieces would assemble into a given figure. It is the domain closest to what people mean by seeing the answer, and it is the one most people find either obvious or baffling.
Fluid reasoning asks you to find a rule you have never been taught. A matrix with one cell missing, a series that continues by a pattern you have to infer, a set of pictures you have to sort by a principle nobody states. Nothing here depends on vocabulary or schooling, which is why this domain is often described as reasoning on the spot.
Working memory asks how much you can hold and manipulate at once: repeating a string of digits backward, rearranging a jumbled sequence into order, holding a running total while new information arrives. It is the domain most affected by sleep, anxiety and distraction on the day, which is one reason a single session is never read as a final word.
Processing speed asks how quickly you can do something easy and accurate. Matching symbols against a key, scanning rows for a target, copying codes under a time limit. The tasks are deliberately simple so that the score reflects pace rather than difficulty, and it is the domain that drops first with fatigue, medication and age.
Why the subtest scores hang together
People who do well on one of those five domains tend to do somewhat better than average on the other four. That is the observation, repeated for more than a century, behind the general factor usually written as g. John Carroll's 1993 re-analysis of hundreds of data sets found the same broad shape each time: a large number of narrow abilities, a smaller number of broad ones, and a single general factor sitting above them all.
What g is remains contested. Some researchers read it as one underlying capacity that feeds every task; others read it as a statistical summary of many partly independent processes that happen to develop together. The modern framework used by test publishers, the Cattell-Horn-Carroll theory, is explicitly a description of that layered structure rather than a claim about a single cause.
The practical reading matters more than the theoretical one. Because the subtests correlate, a full-scale score is a reasonable summary when all five indexes are close together. When they are far apart, the summary hides the thing worth knowing. A report showing strong verbal comprehension and weak processing speed is telling a story the single number erases, which is why index scores, not the total, drive most clinical decisions.
What an IQ score predicts, and for whom
At the level of large groups, the correlations are real and have been replicated. Tarmo Strenze's 2007 meta-analysis of longitudinal studies found corrected correlations between measured intelligence and later educational attainment of .56 across 59 samples and 84,828 people, .43 with occupational attainment, and .20 with income. Deary and colleagues, following more than 70,000 English schoolchildren over five years, estimated the correlation between general ability at 11 and national exam achievement at 16 at .81 at the latent level.
The same meta-analysis contains the correction that most popular accounts leave out. Parental socioeconomic status predicted educational attainment at .55 and school grades predicted it at .53, both close to the figure for intelligence. Strenze's own conclusion was that intelligence is a powerful predictor of success but, on the whole, not an overwhelmingly better predictor than parental background or grades.
A correlation of .5 across tens of thousands of people also leaves enormous room for individual difference. Two people ten points apart will frequently end up in the reverse order on anything you actually care about, because effort, health, money, teaching, timing and luck are all in the mix. A score describes a tendency in a population. It does not forecast a life, and it should never be used to decide what someone is capable of attempting.
What an IQ test does not measure
The list of things a battery leaves out is longer than the list it includes. It does not sample motivation, persistence or conscientiousness, which are the traits that decide whether ability is ever used. It does not sample creativity in any serious sense, because the tasks all have a correct answer and creative work mostly does not. It does not sample practical judgment, the ability to read a room, or the willingness to change your mind.
It also does not measure skill. A score is not a credential and it does not certify competence at anything, which is why no reputable publisher presents one as a qualification. It does not measure character, kindness or wisdom. And it does not measure potential: a result describes performance on a particular afternoon, not a ceiling you are stuck under.
Low scores carry the same restriction, in the direction that matters more. A single low result establishes nothing on its own. The American Association on Intellectual and Developmental Disabilities sets out three criteria together, significant limitation in intellectual functioning, significant limitation in adaptive behavior across conceptual, social and practical skills, and onset during the developmental period before age 22, and states that a clinician can only reach a determination after a comprehensive evaluation. A number is one input to that work, never the conclusion.
Who can give a clinical test, and why that matters
The major intelligence batteries are restricted products. Pearson, which publishes the Wechsler scales, states that all its professional assessment products are assigned a qualification level of A, B or C, with C being the highest, and an individually administered intelligence battery sits at the top of that scale. In practice that means a licensed psychologist or someone with equivalent graduate training and supervised experience in individual cognitive assessment.
The restriction is not a guild protecting its territory. The scoring rules, the stopping rules, the norms and the interpretation all depend on training, and an unqualified administration produces a number that looks like a score and is not one. The Standards for Educational and Psychological Testing, published jointly by the American Educational Research Association, the American Psychological Association and the National Council on Measurement in Education, set the professional expectations for how such a test is chosen, given and reported.
This is also why a clinical assessment costs real money and takes real time, and why there is no legitimate way to shortcut it. If a decision is riding on the result, placement, accommodations, a disability determination, an evaluation of a change in your thinking, the only useful route is a qualified assessor and a written report. Nothing free and online substitutes for that, including ours.
How a free reasoning quiz differs from a battery
Psychology.com's IQ Test is a reasoning quiz: 30 original text-only questions for adults, untimed, with no speed bonus, taken on your own and scored instantly into an estimated range with a five-category breakdown. It is self-administered, unproctored and unnormed. We say that on the result screen and we will say it here: it is not an IQ score and it is not a clinical assessment.
The differences are structural, not a matter of our quiz being a slightly worse version of the real thing. It is text only, so it cannot sample visual-spatial ability with blocks or pictures. It is untimed, so it cannot sample processing speed. Nobody is watching, so there is no way to know whether the conditions were quiet, whether a break was taken, or whether a search engine was open. And it has not been given to a representative sample of adults, so there is no reference group behind the number.
What it is good for is the format. If you want to know what verbal, numerical, abstract and logical reasoning items feel like before you meet them somewhere that counts, a quiz is a reasonable way to find out, and the category breakdown can point you at a kind of task you would enjoy getting better at. If you want to know where you stand against other adults, that is a different instrument and a different appointment.
What the research says
This is one of the better-evidenced areas in psychology, and saying so plainly is part of being honest about it. The layered structure of cognitive abilities has been found again and again across decades of factor-analytic work, the modern batteries are highly reliable instruments by any psychometric standard, and the group-level links to education, occupation and health outcomes have held up in large prospective studies, including a 68-year follow-up of a Scottish birth cohort published in the BMJ in 2017.
The limits are equally well established. Average scores on the same tests rose through most of the twentieth century and have stalled or reversed in several countries since, which no theory of a fixed quantity explains comfortably.
So the fair summary is narrow and useful. These tests measure something real, stable enough to be worth measuring, and informative about groups. They are also a sample of a few hours of a person's thinking, scored against a reference group, carrying a margin of error, and silent on most of what makes someone good at a life.
The rest of the intelligence pages here
If you have a number and want to read it properly, the IQ scale page explains how the scale is built, why the range labels are conventions rather than natural categories, and how to use a confidence interval. Average IQ score answers the related question of what sitting near 100 does and does not tell you. Who created the IQ test covers the history, which is a chain of people rather than one name, including the uses the early instruments were put to.
For the mechanics, Cattell-Horn-Carroll theory sets out the three-level model that shaped the subtests in every modern battery, and Stanford-Binet describes one of the two main clinical instruments in detail. For the setting where most adults actually meet a test, cognitive ability tests explains the shorter screens employers use and what preparation legitimately helps. Gardner's multiple intelligences covers the best-known alternative to a single score, and it sits under our Multiple Intelligences Test rather than this one.
| Feature | Clinical battery | Our free reasoning quiz |
|---|---|---|
| Who gives it | A qualified examiner, one to one | You, on your own device |
| Length | One to two hours for the appointment | 30 questions, untimed |
| Domains covered | Verbal, visual-spatial, fluid, memory, speed | Text-based reasoning in five categories |
| Reference group | A representative normative sample | None, so no percentile is possible |
| What you get | Index scores, percentiles, confidence intervals, a report | An estimated range and a category breakdown |
| Conditions controlled | Standardized script, timing, observation | Unproctored, breaks and notes allowed |
| Can be used for decisions | Yes, by the professional who gave it | No, for curiosity and format practice only |
| Cost | A paid professional appointment | Free |
Key takeaways
- An IQ score is a comparison with an age-matched reference group, rescaled so the average is 100 and one standard deviation is 15 points, not a quantity of anything.
- A professional battery samples five domains: verbal comprehension, visual-spatial ability, fluid reasoning, working memory and processing speed, given one to one over one to two hours.
- Subtest scores correlate, which is the basis for a general factor, but when the five indexes are far apart the single total hides the useful information.
- Group-level prediction is real and modest: Strenze's meta-analysis found .56 with educational attainment, .43 with occupation and .20 with income, with parental background and school grades predicting education almost as well.
- The test does not sample motivation, creativity, judgment, skill or character, and no single score establishes anything about one person on its own.
- Our free reasoning quiz is untimed, text only, unproctored and unnormed: useful for format practice, not a clinical assessment.
Curious how you handle this kind of reasoning?
Our free IQ Test is 30 untimed text questions with an instant estimated range and a five-category breakdown. It is a reasoning quiz, not a clinical assessment, and the result says so.
Cite this source
Fontane Pennock, S. (2026, November 8). What an IQ Test Measures and What It Includes. Psychology.com. https://psychology.com/types/what-an-iq-test-measures
Frequently asked questions
What does an IQ test consist of?
A professional battery is about ten to thirteen short subtests grouped into five domains: verbal comprehension, visual-spatial ability, fluid reasoning, working memory and processing speed. Tasks include defining words, building patterns to match a model, completing a matrix by inferring a rule, repeating digit strings in reverse, and matching symbols against a key under time pressure. An examiner gives them one to one from a standardized script.
How is an IQ test done, step by step?
You meet an examiner, answer background questions, then work through the subtests in a set order. Each one starts at a point chosen by your age or early performance and stops after a run of items you miss, so nobody sits the whole set. Some tasks are timed. Your raw points are converted against published reference data for your age group, and the report gives index scores with percentiles and confidence intervals.
How long does an IQ test take?
Allow one to two hours for the core appointment. The test battery alone can be shorter, and a broader evaluation that includes memory, attention or academic testing can run across more than one session. The current Wechsler adult scale is listed by its publisher at roughly 45 minutes for the seven subtests behind the full-scale score and roughly 60 minutes for the ten primary subtests.
Who can give a real IQ test?
The main batteries are sold only to qualified purchasers. Pearson assigns its professional assessments a qualification level of A, B or C, with C the highest, and an individually administered intelligence test sits at that top level. In practice the administrator is a licensed psychologist or someone with equivalent graduate training and supervised experience in individual cognitive assessment.
Are online IQ tests accurate?
They are accurate about one thing: how you did on those particular questions. They cannot produce a percentile, because that requires a representative reference sample, and they cannot control the conditions, the timing or whether you looked anything up. Our own quiz says so on the result screen. Use one for curiosity or format practice, and a qualified assessor if a decision depends on the answer.
What does an IQ test not measure?
Motivation, persistence, creativity, practical judgment, social skill, specific knowledge, character, and anything resembling potential. A score also carries no information about competence at a job or a course, which is why reputable publishers state that their instruments are not credentials. The five domains sampled are a narrow slice of thinking, deliberately chosen because they can be measured reliably.
Can an IQ score change?
The score can move, both because people change and because measurement is imperfect. Any report should include a confidence interval for that reason. Processing speed and working memory are especially sensitive to sleep, mood, medication and distraction on the day. Across whole populations average raw performance also rose through the twentieth century, which is why tests are re-normed every decade or two.
Related types
References
- Carroll JB. Human Cognitive Abilities: A Survey of Factor-Analytic Studies. Cambridge University Press; 1993. doi.org
- Strenze T. Intelligence and socioeconomic success: A meta-analytic review of longitudinal research. Intelligence. 2007;35(5):401-426. doi.org
- Deary IJ, Strand S, Smith P, Fernandes C. Intelligence and educational achievement. Intelligence. 2007;35(1):13-21. doi.org
- Calvin CM, Batty GD, Der G, et al. Childhood intelligence in relation to major causes of death in 68 year follow-up: prospective population study. BMJ. 2017;357:j2708. doi.org
- AERA, APA, NCME. Standards for Educational and Psychological Testing. 2014 edition. testingstandards.net
- American Association on Intellectual and Developmental Disabilities. Defining criteria for intellectual disability. aaidd.org
