How IQ Tests Work: Questions, Norms and Scoring

An IQ test does not simply count correct answers and label the total “intelligence.” It samples performance across selected cognitive tasks, compares the result with an appropriate norm group and converts that comparison into standard scores.

What does an IQ test try to measure?

Intelligence tests aim to estimate general cognitive ability: the capacity to reason, learn, solve unfamiliar problems and work with complex information. Psychometric research finds that performance across many cognitive tasks tends to be positively related. This shared pattern is often described as general intelligence, or g.

That does not mean every task measures the same thing. Vocabulary, spatial reasoning, working memory and processing speed each contribute distinct information. Modern batteries therefore produce both an overall score and a profile across narrower domains.

How are IQ test questions selected?

Developers usually begin with a larger pool of candidate items than the final test will contain. Those items are piloted with people from the intended population. Weak, ambiguous or unfair questions are revised or removed.

Useful items vary in difficulty and distinguish meaningfully between different levels of performance. A test made only of very easy questions cannot differentiate well among average and high performers. A test made only of extremely difficult questions provides little information about most users.

Professional test development also examines whether items behave differently across groups for reasons unrelated to the intended construct. Fairness cannot be guaranteed by using only pictures or abstract shapes; nonverbal tasks can still depend on familiarity, education and strategy.

How standardisation and norms work

A raw total has little meaning on its own. Fifteen correct answers might be excellent on one test and ordinary on another. To interpret performance, the test is administered to a norm sample and raw scores are converted into standard scores.

Many IQ scales use a mean of 100 and a standard deviation of 15. The score therefore indicates where a result falls relative to the norm group. Age-based norms are particularly important because cognitive performance changes across childhood, adolescence and later adulthood.

From raw answers to a full-scale IQ score

1. The test records correct responses, response times or both.
2. Raw results are converted using the relevant norm tables or statistical model.
3. Scores from related tasks are combined into domain or index scores.
4. Selected domain scores are combined into an overall estimate, often called full-scale IQ.
5. The report should communicate uncertainty and any reasons the result may be less interpretable.

The precise procedure differs between tests. Some tasks are timed because speed is part of the intended construct. Others allow more time so the score reflects accuracy rather than rapid responding.

Why current professional tests use multiple domains

Contemporary instruments such as the WAIS-5 use several cognitive domains and multiple subtests. This broader structure reduces the risk that one unusual strength or weakness dominates the entire interpretation. It also gives clinicians information about how a person arrived at the overall score.

Consumer online tests are usually shorter and narrower. They may focus mainly on fluid reasoning and pattern recognition because these tasks are practical to deliver on a screen. That can still be informative, but it should not be described as equivalent to a full clinical battery.

What an IQ score can—and cannot—tell you

An IQ score can describe performance on a particular sample of cognitive tasks relative to a reference group. It can be useful when asking about general reasoning, learning demands or a pattern of cognitive strengths and weaknesses.

It does not directly measure creativity, motivation, personality, emotional intelligence, practical wisdom or moral worth. It also does not explain why a person obtained a particular result. Formal interpretation may require developmental, educational, linguistic and medical context.