So You Want To Know Your IQ
I've spent years working with cognitive assessments across clinical and industrial settings, and the first thing most people get wrong is assuming an IQ score is a permanent, universal number. It isn't. It's a snapshot of how you performed on a specific test at a specific time under specific conditions. When someone asks me How To Work Out Their Iq, they're usually starting from the wrong premise, so let me walk through what actually happens in practice. The only reliable method is administering a standardized, norm-referenced intelligence test. The gold standard is the WAIS-IV (Wechsler Adult Intelligence Scale, fourth edition), which produces a Full Scale IQ score with a mean of 100 and a standard deviation of 15. There's also the Stanford-Binet Fifth Edition and the MATS (Mensah Intelligence Tests), but the WAIS is what most clinicians use because it breaks down into index scores — verbal comprehension, perceptual reasoning, working memory, and processing speed — which tells you more than a single number ever could. Here's the practical reality of taking one. A proper WAIS-IV administration takes about 60 to 90 minutes. It's not a series of multiple choice questions you can cram for. The subtests include things like Block Design where you recreate patterns with colored blocks under time pressure, Digit Span where you repeat sequences of numbers backward, Arithmetic mental calculations, Similarities where you explain how two concepts relate, and Vocabulary definitions. Each subtest is scaled to a mean of 10, and the sum of the core subtests gives you the Full Scale IQ.
The scoring process itself is mechanical once you have the raw scores. You look up the scaled score in the WAIS-IV table for your age group, sum the relevant subtest scaled scores into a composite, convert that composite to a standard score, and you're done. An IQ of 115 means you scored one standard deviation above the mean for your age cohort. That's it. Nothing mystical about it.
Online Tests: What Actually Works and What Doesn't
There are free online IQ tests everywhere, and the vast majority of them are garbage. They lack norms, they lack standardization, and many are designed to inflate scores so you feel good and share the results. A test that gives you an IQ of 135 because you answered five pattern-matching questions correctly is not measuring anything useful. The Mensa Norway or Mensa Denmark online practice tests come closer because they're derived from actual Raven's Progressive Matrices items, which measure fluid reasoning. But even those are estimates at best. A proper practice test will give you a rough ballpark within maybe 10 to 15 points of a full clinical assessment, and that's only if you take it under timed conditions without distractions. I've seen people score 140 on a free online test and then get a 112 on the WAIS. The discrepancy isn't rare. If you want a rough ballpark without spending money, the Mensa practice tests are the least unreliable option. Take them seriously — timed, no phone, no pausing. But understand that a 30-minute online test cannot replicate what a 90-minute clinician-administered battery does.
Get the Full Details
Self-Directed Calculation: The Math Behind the Score
Sometimes you don't need a psychologist. If you have access to a published test with normative tables — like a textbook version of the WAIS or even a well-constructed older edition like the WAIS-R — you can calculate the score yourself. Here's how that actually works. First, administer the test under standardized conditions. Read the instructions exactly as written. Time each subtest precisely. Record raw scores — correct answers, not percentages. Then locate the conversion tables in the manual. For the WAIS-IV, you convert raw scores to scaled scores using age-band tables. The age bands are 16-17, 18-22, 23-29, 30-34, 35-39, 40-44, 45-49, 50-54, 55-64, and 65-74. A raw score of 14 on Digit Span might be a scaled score of 10 for a 25-year-old but only a 7 for a 60-year-old, because the norms adjust for age. This is why you can't just use a single table — age matters significantly, especially at the extremes. Once you have scaled scores for the four index subtest clusters, you sum them, convert to a composite score using the manual's table, and apply the confidence interval. A Full Scale IQ of 105 isn't a single point — it's typically reported as a range of about 100 to 110 at the 95% confidence level. I always report it that way because treating it as a precise number is misleading. Your performance on any given day can fluctuate by several points due to sleep, anxiety, motivation, or even the temperature in the room.
A Problem I Actually Encountered
Not long ago I was scoring a WAIS-IV for someone whose Working Memory Index was dramatically lower than their Verbal Comprehension Index — a gap of over 20 points, which is clinically significant. On Digit Span backward, they were scoring near the bottom. But when I reviewed the raw data, I realized they had a mild untreated hearing deficit in the high frequencies. The digit strings contained consonant sounds that were hard for them to distinguish, which artificially depressed the score. This isn't something the test manual warns you about directly, but any experienced examiner learns to flag it. The workaround was straightforward: I noted the discrepancy in the report, recommended an audiological evaluation, and interpreted the Full Scale IQ with a caveat that the Working Memory component was likely suppressed. The true cognitive profile was much more balanced than the raw numbers suggested. This is exactly why self-scoring without clinical training is risky — you might take a suppressed score at face value and draw the wrong conclusion. The most important thing beginners don't understand is that IQ tests measure a narrow band of cognitive abilities. They don't measure creativity, practical problem-solving, emotional intelligence, or wisdom. Someone who scores in the 90s on a WAIS can be exceptionally capable in domains the test simply doesn't assess. I've seen people with IQs in the low average range run successful businesses and solve complex real-world problems that would stump someone with a higher score. The test is a tool, not a verdict. Another pitfall is the Flynn effect. IQ norms are periodically re-normed because average scores have been drifting upward across generations — roughly three points per decade. A test normed in 1995 will give systematically different scores than the same test normed in 2025. If you're comparing your score to someone else's, make sure the tests are from roughly the same era. An IQ of 110 on an old norm table might be closer to 100 on a current one.
Test anxiety is a third factor that people consistently underestimate. I've watched competent people perform 10 to 15 points below their typical level because they froze on the timed subtests. The clock pressure is real and it affects people differently. If you know you're prone to anxiety under time pressure, a single test session is not a reliable measure. Clinicians sometimes administer the test across two sessions to mitigate this.
What an IQ Score Actually Tells You
An IQ score is a ranked position within an age cohort, expressed on a normal distribution. About 68% of people score between 85 and 115. About 95% score between 70 and 130. Scores below 70 may indicate intellectual disability when accompanied by adaptive functioning deficits. Scores above 130 are considered gifted. These are descriptive categories, not destiny markers. The predictive validity of IQ for academic achievement and job performance is well established, with correlations in the 0.5 to 0.7 range depending on the outcome. But correlation is not identity. IQ explains a substantial portion of variance in certain domains, but it leaves a lot unexplained. Socioeconomic background, education quality, motivation, and opportunity all interact with cognitive ability in ways that a single number cannot capture.
Practical Steps If You Want to Proceed
If you want an accurate score, contact a licensed psychologist who administers the WAIS-IV or Stanford-Binet. Expect to pay between $500 and $2,000 depending on your location and whether the assessment includes additional subtests. The appointment will last 60 to 90 minutes. Bring ID and any previous test scores if you have them. Arrive well-rested. Don't try to game the test — it doesn't work, and pretending to perform differently will just produce an invalid score. If budget is a constraint, Mensa's national organization sometimes offers subsidized testing or can direct you to low-cost clinics affiliated with universities. Graduate psychology programs often provide reduced-rate assessments as part of their training. It's the same test, administered by a supervised trainee, and the results are equally valid. For a rough estimate, take the Mensa Norway or Denmark practice test under strict timed conditions. Treat it as an approximation, not a diagnosis. And whatever you do, don't treat any online result as definitive. The difference between a useful number and a vanity metric is often just whether someone sat down with normative data and followed the manual.
