How IQ Tests Work: Question Types and Scoring

An IQ test looks like a set of puzzles, but each puzzle is there for a reason. Every question type targets a specific mental ability, and the way answers are turned into a score follows well-established statistical methods. This guide explains what kinds of questions IQ tests use, how scoring works, and how to think about the results of an online test.

What IQ tests try to measure

Psychologists have found that people who do well on one type of reasoning task tend to do well on others. This shared factor is often called general intelligence. Most modern tests estimate it by combining several narrower abilities, typically:

• Fluid reasoning: solving new problems and spotting rules you have never seen before.
• Visual-spatial thinking: picturing and manipulating shapes in your mind.
• Working memory: holding information and working with it at the same time.
• Quantitative reasoning: applying numerical thinking to a problem.
• Verbal reasoning: understanding relationships between words and ideas.

Testing several abilities gives a more balanced estimate than any single puzzle type, and it also shows where your relative strengths lie.

Common question types

Matrix reasoning. You see a grid of shapes, usually 3 by 3, with one cell missing. The shapes change across rows and columns according to hidden rules: a shape might rotate, a count might increase, or features might combine. You pick the piece that completes the pattern. Matrices are a classic measure of fluid reasoning because they rely very little on language or prior knowledge.

Number and letter series. You are given a sequence such as 3, 6, 12, 24 and asked for the next item. The rule might be adding, multiplying, alternating between two patterns, or moving through the alphabet in steps. These test your ability to find and extend a rule.

Spatial rotation. You compare a shape with several rotated or mirrored versions and decide which one is the same object turned, and which ones are mirror images. Three-dimensional versions use figures made of cubes, which you have to rotate mentally.

Paper folding. A sheet of paper is folded, a hole is punched through it, and you choose how it looks when unfolded. You have to track each fold in reverse.

Cube nets. You see a flat pattern of six connected squares and decide which cube it folds into, or which faces end up opposite each other.

Verbal analogies. "Hand is to glove as foot is to ...". You identify the relationship between the first pair and apply it to the second. Because these depend on vocabulary, they must be written separately for each language.

Working memory tasks. You might see a sequence of digits and repeat it forward or backward, or watch blocks light up in an order and tap them back. Sequences grow longer until you reach your limit.

How scoring works

Scoring happens in three steps.

1. Measure performance. The simplest method counts correct answers. Better methods weigh questions by what they reveal. Many modern assessments use item response theory (IRT), in which every question is calibrated in advance using responses from many people. Each question gets a known difficulty, and the model estimates the ability level that best explains your pattern of correct and incorrect answers. Getting a hard question right says more than getting an easy one right, and a correct answer on a question that is easy to guess counts for less.

2. Compare with a norm group. Your ability estimate is then compared with a reference group, usually people of a similar age. This tells you how far above or below the group average you are, measured in standard deviations.

3. Convert to the IQ scale. That distance is placed on a scale with an average of 100 and a standard deviation of 15. One standard deviation above average becomes 115, two becomes 130, and so on. A good test also reports a confidence range, because every measurement has some error. Our guide to IQ percentiles explains how that range is calculated.

How IQTestScale builds and scores its test

The IQTestScale test has 20 questions across four abilities: fluid reasoning, spatial thinking, working memory and quantitative reasoning. Each question is a different type of puzzle, and the test takes about 20 minutes. A few design choices shape the result:

• The same difficulty for everyone. Every test follows the same order of difficulty: ten easier questions, then ten harder ones. Nobody gets an easier test than anyone else.
• New questions every time. Puzzles are generated from rule sets at ten difficulty levels, so two people never see the same test and answers can't be shared or memorized.
• IRT scoring. Harder questions count for more, and lucky guesses count for less.
• An honest range. Your score is reported with a likely range around it, for example 118 (112 to 124), rather than as a single number pretending to be exact.
• Every answer explained. The full report walks through each question, the correct answer and the rule behind it.

Why online tests are self-assessments

Professionally administered tests are given in person by a qualified psychologist, under controlled conditions, with norms built from carefully sampled populations. Those conditions make clinical and educational decisions possible.

An online test works differently. You take it on your own device, in your own environment, with whatever distractions are around you. The comparison group and the testing conditions are less controlled. Even a carefully built online test is therefore best read as a self-assessment: a well-informed estimate of how you reason, useful for personal insight, but not a diagnosis and not a substitute for a clinical evaluation.

When choosing an online test, look for clear signs of care: an explanation of how scores are calculated, a range instead of a single number, and an honest statement of what the result can and cannot tell you. For context on what your number means, see what is a good IQ score and the IQ score chart.

Try it for yourself

The clearest way to understand how an IQ test works is to take one. The IQTestScale test needs no account, takes about 20 minutes, and your results show your score, its likely range and your percentile.

Take the IQ test