Log in

Q&A: Online IQ tests

Q: Where can I find a valid online IQ test?

It’s not really possible to get a valid IQ test online. They have to be administered and scored by humans because there are a lot of judgment calls involved, and a multiple-choice format simply doesn’t lend itself to that kind of thing.

I’m from the US, and I’ve studied IQ a lot because I’m fascinated with statistics and with tests and measures (I’ve got a psychology degree, plus I’m autistic, which makes a very obsessive type of researcher!). And take it from me: People in the US overvalue IQ. It means much less than it seems, and says less about intelligence than people think it does.

IQ tests break down whenever someone who's atypical in some way gets tested. If your neurology is unusual, your communication style is unusual, even your culture is different, IQ tests start to say less and less until in the end, they say nothing.

The tests aren't utterly useless. Generally, we can tell if somebody’s outright gifted or outright learning-disabled from an IQ test, if it’s administered carefully on a good day with no cultural barriers, but the precise numbers themselves are really very deceptive. The idea that somebody with an IQ of 112 is smarter than someone with an IQ of 110 is just ridiculous. It just isn’t that precise. Only once you get to two or more standard deviations worth of difference do I feel that the differences are worth making a note of—and since the IQ test has a standard deviation of 15 or 16, that’s a big difference, the difference between average and gifted or average and intellectually disabled.

That’s not so surprising, considering that the original IQ tests were meant to identify students who needed extra help. They still fulfill that function reasonably well. But they were never meant to rank people by intelligence.

If face-to-face IQ tests are of so little worth, fail so often and say so little about us, you really can’t expect online IQ tests to be worth much at all.

Instead of worrying about IQs, we should focus on what we’re good at doing, what we’ve worked hard on, what we enjoy learning. That’s what really matters.

To those who want to learn more about IQ, I recommend the book "The Mismeasure of Man", by Stephen Jay Gould. It’s old—written in 1981—but it addresses a lot of the issues with testing intelligence and cognition, and explains why it’s hard to do and why it's much less applicable to daily life than you’d think.


More on overvaluing IQ

Another related observation:

I've read that even if you buy into IQ tests, high values aren't needed to do excellent "Brain work". Apparently 110 is perfectly adequate for university, and 120 for nobel-quality research. Note both are within a standard deviation of each other.

Plus, grit (stick-to-it character) can be more important than IQ.
...1981 is old..? jeez, then I must be old..!

I think that IQ test's have a limited validity and scope... and grit has been accounted for more success than high IQ
I don't get why people assume that because IQ isn't ridiculously precise means it's not really a valuable measurement. You don't even need that much precision, anyway.

I'm going to use the MA/CA model here, even though it's not exactly accurate, but it helps illustrate the basic idea. Let's say you have a 10 year old (just had his birthday) who was given a magically precise IQ test and scored 110. His twin brother took the same IQ test and got 112. What would that mean for MA?

The first boy would have an MA of 11 years. The second boy would have an MA of 11 years and 2.5 months. In other words, the difference in their cognitive abilities would be the same as the difference between two 11 year olds with IQs of 100 who are 2.5 months apart. Which is not really meaningful at all. So why would it matter if an IQ test isn't accurate to that degree?

On the other hand, a 10 year old with an IQ of 115 would have an MA of 11 and a half, making that kid, in theory, a year and a half ahead of a kid the same age with an IQ of 100. While the difference between a kid who's just turned 10 and one who's 11 and a half is slight, it's noticeable, particularly if you're looking at their academic scores. If you put an 11 and a half year old kid in a grade 5 class, assuming their IQ is roughly 100 and they don't have specific learning disabilities or educational neglect, they'll probably get better than average marks. Which is also true of kids with an IQ of 115.
Your argument makes the assumption that IQ is an interval scale. It's not; it's ordinal. The difference between 100 and 110 has nothing to do with the difference between 110 and 120, for example.