What is IQ?
IQ, short for intelligence quotient, is a standardised score describing how efficiently a person reasons compared with a large reference population of the same age. It is not a count of correct answers and it is not a measure of stored knowledge. Raw test performance is converted onto a scale with a mean of 100 and a standard deviation of 15, so a score of 115 sits one standard deviation above the average of the reference group and a score of 85 sits one standard deviation below it.
The word quotient is a historical leftover. Early intelligence testing divided a child's estimated mental age by their chronological age and multiplied by 100. Modern testing dropped that arithmetic entirely and replaced it with norm-referenced scoring: your answers are compared against the distribution of answers from a normative sample, and your position in that distribution becomes your score.
Because the number is relative, an IQ score only means something in the company of three other facts: which test produced it, which reference sample it was normed against, and how wide the confidence interval around it is. Any online IQ test that reports a bare number with none of that context is asking you to trust a figure it has not justified.
What does an intelligence test actually measure?
An intelligence test samples reasoning processes rather than learned facts. The items are designed so that prior schooling, vocabulary size and cultural background matter as little as possible, while the ability to find structure, hold information in mind and apply rules matters as much as possible.
No single item type captures reasoning on its own, so a credible battery samples several distinct abilities and pools them into a general score. Testora IQ measures five:
- Logical reasoning — drawing valid conclusions from rules, conditionals and ordered sequences.
- Pattern recognition — detecting hidden regularities in visual and numeric series.
- Spatial reasoning — mentally rotating, folding and comparing shapes in two and three dimensions.
- Numerical ability — reasoning about quantity, proportion, rate and relative magnitude.
- Working memory — holding several items in mind and manipulating them while solving a problem.
How is an IQ score calculated?
Scoring happens in four steps. First, each response is weighted by the calibrated difficulty of the item it answers, so a correct answer on a hard item contributes more than a correct answer on an easy one. Second, weighted responses are pooled within each domain to produce five domain scores. Third, the domain scores are combined using weights that reflect how strongly each one loads onto general reasoning ability. Fourth, the combined figure is mapped onto the 100-mean, 15-standard-deviation scale using the normative sample.
The output of that mapping is your IQ score plus a percentile — the share of the reference group scoring at or below you. Percentiles are usually the more intuitive of the two: hearing that you scored at the 84th percentile tells you more about your standing than hearing the number 115.
One consequence of the normal distribution is that small differences near the middle of the scale move percentiles a lot, while the same difference at the extremes barely moves them. Two people three points apart around 100 can be several percentile points apart; two people three points apart around 140 are almost indistinguishable in rank.
Why confidence intervals matter more than the headline number
Every psychometric measurement carries error. The standard error of measurement quantifies it, and it is the reason serious reports present a range rather than a point. A test with a standard error of about 4.5 points gives a 90% confidence interval of roughly plus or minus 7 points, which means a reported 120 is best read as very likely falling between 113 and 127.
This is not a weakness in the test; it is honesty about what a 30-minute sample of behaviour can establish. It also explains why a small difference between two sittings of the same test is normal rather than meaningful, and why chasing a one or two point improvement is not a sensible goal.
What IQ does not tell you
IQ is silent on character, creativity, motivation, ethics, emotional skill and worth. It predicts some academic and occupational outcomes at a group level with modest strength, and predicts far less about any individual person than popular writing implies.
Scores are also sensitive to conditions. Sleep, stress, illness, caffeine, noise, screen size and simple familiarity with the item formats all move results by a few points. That variability is exactly why a single number should never be treated as a fixed property of a person.
The practical use of a score is direction, not judgement. A profile showing strong spatial reasoning and weaker working memory suggests different study tactics than the reverse — and that guidance is what a good report should give you.
Is a free online IQ test worth taking?
Cost and psychometric quality are unrelated. What determines quality is whether items were calibrated on a real sample, whether the normative group is large enough, whether reliability is reported, and whether the results page explains its own uncertainty.
A free online IQ test that meets those conditions is a useful screening instrument. What it is not, at any price, is a clinical assessment: diagnosis of learning differences or cognitive impairment requires an in-person evaluation by a qualified psychologist using an individually administered battery.
Frequently asked questions
- What is a normal IQ score?
- Scores between 90 and 109 are considered average and cover roughly half of the population, because the scale is built with a mean of 100 and a standard deviation of 15.
- Does IQ change over time?
- Relative standing is fairly stable in adulthood, but measured scores move with sleep, stress, practice and testing conditions, so small differences between sittings are expected.
- Is IQ the same as intelligence?
- No. IQ estimates reasoning efficiency on a specific set of tasks. Intelligence in everyday use includes creativity, judgement and expertise that no timed test samples.