Understanding Where IQ Numbers Actually Come From
IQ scores are standardized measurements derived from psychometric tests that compare your performance against a large representative sample. The numbers themselves aren't some fixed truth about your intelligence. They're relative rankings on specific cognitive tasks, normalized so the average lands at 100 with a standard deviation of 15 in most modern tests. The straightforward path is a professionally administered test. WAIS-IV or WAIS-V for adults, WISC-V for children under sixteen. You book an appointment with a licensed psychologist, show up, and take a series of subtests covering verbal comprehension, perceptual reasoning, working memory, and processing speed. The whole thing takes about 60 to 90 minutes. You get a full report back within a week or two, usually between $200 and $600 depending on your location and the practitioner. I went through this process myself a few years back when a neurology referral asked for baseline cognitive data. The psychologist I worked with was decent but clearly running behind schedule. She skipped a couple of the optional supplementary subtests without mentioning it until after we were done. The score she gave me was technically valid, but the omission of those extra indices meant the full profile wasn't as complete as it should have been. If you're paying for this, ask upfront which subtests are included and which are optional. Write down which ones she actually covers. You won't get that information back later.
Online tests exist. Most of them are useless for anything except entertainment. The ones that come close to legitimate are the ones licensed from actual psychometric instruments. Mensa offers a supervised diagnostic test. Some national Mensa chapters run these periodically. The scores from those carry more weight than anything you'll find on a random website, but they still don't replace a clinical evaluation if you need documented results for academic accommodations or medical purposes.
What the Numbers Actually Mean
A score of 100 means you performed at the exact median of the normative sample. 115 is one standard deviation above that, which puts you ahead of roughly 84 percent of people your age group. 85 is one standard deviation below, meaning you scored higher than about 16 percent. The distribution is bell-shaped by design. Most people cluster around 85 to 115. Scores above 130 or below 70 are genuinely rare and require careful interpretation. Here's something most people miss. Your IQ isn't a single number. It's a collection of index scores that can vary significantly from each other. Someone might score 130 on verbal comprehension and 95 on working memory. The full-scale IQ would be somewhere in the middle, maybe 112, but that single number hides a real cognitive profile. That discrepancy matters if you're looking at learning differences or career fit. A rounded score smooths over the interesting parts. Another counter-intuitive detail. Practice effects are real and they're substantial. If you take the same test twice within a few years, your second score will likely be higher by three to five points even if nothing about your actual cognitive ability changed. This is why professionals don't reuse the same test form at follow-up sessions. They switch to an equivalent but different version. Online tests don't do this. They just give you another number that's slightly inflated from the first time around.
Get the Full Details
What Online Tests Get Wrong
Free online IQ tests are generally calibrated on nowhere near enough participants to establish reliable norms. Some claim to use proprietary algorithms or AI scoring. That doesn't fix the fundamental problem of inadequate normative samples. A test built on fifty thousand self-selected internet users will have a very different distribution than one built on a stratified sample of tens of thousands representative of the general population. The timing element is another failure point. Many online tests don't properly enforce time limits or they make them arbitrarily short. Processing speed is a measured component on legitimate tests because it captures something real about how efficiently your brain retrieves and manipulates information. Remove the time pressure and you're measuring something else entirely. You're measuring how carefully someone wants to work, not how quickly they can work when speed matters. I tried three different free online tests last year just to check what they produced. The scores ranged from 98 to 127. All of them claimed high reliability. None of them had published normative data or any way to verify their calibration. The variation between them was larger than the typical measurement error on a real clinical test, which is usually around two to three points for a full-scale score. Those online results told me absolutely nothing useful beyond confirming that I could take a puzzle test.
When a Score Actually Matters
If you need an IQ score for a formal purpose, like gifted program placement, learning disability evaluation, or certain professional admissions, a paper-and-pencil test administered by a credentialed professional is the only option anyone will accept. Reports from unsupervised online tests get discarded immediately in these contexts. There's no negotiation about it. For personal curiosity, a single number is fine. But don't treat it as definitive. Intelligence is multi-dimensional. Working memory capacity predicts different outcomes than processing speed, which predicts different outcomes than fluid reasoning. The full-scale IQ conflates all of these into one digit. It's convenient but reductive. The best approach if you want something reasonably accurate without spending money is to find a university psychology department that offers low-cost testing through graduate training clinics. These programs are supervised by licensed psychologists and the tests used are the same ones used in private practice. Costs typically run between $50 and $150. The tradeoff is scheduling. You might wait six to eight weeks for an appointment and the session could run longer than usual because the trainee is still building competence.
If you do go the professional route, request the breakdown. Don't accept just the full-scale number. Ask for the index scores, the confidence intervals, and any notes about subtest scatter. A responsible examiner will have all of this written into the report. If they don't volunteer it, ask. That's the information that's actually useful.
