University of Cambridge Cognitive Ability Test

"The entire test will take about an hour to complete."

Doesn't that skew things? That is a lot of time a lot of people don't have.

No, not in any way. Ability to be present for a test does not have any bearing whatsoever on cognitive ability, nor does any test on the planet account for whether or not the taker is present.

I think the suggestion is that people who have less free time might be in more of a rush to complete the test, and that surely affects test results without indicating cognitive ability.

I'd bet they don't consider incomplete tests, and most people I think would just peace out in ten or less questions vs continuing on either uninterested or rushed for time

I’m going to guess this is an economically comfortable person with lots of free time imagining that people less well off have none. To an approximation, that trend is surely true - poor people have less free time - but almost everyone can find time to do things they want to do. Would it skew results? Hard to say.

> Ability to be present for a test does not have any bearing whatsoever on cognitive ability

The irony of making such a big mistake while discussing cognitive ability...

Nope, I’m not thinking about their test in any way. I’m explaining that it’s explicitly unrelated to your cognitive ability. Getting a more convenient job does not boost your brain.

I’m going to guess you’re obsessed with the plight of those you perceive to be “under” you. I mean, they’re poor - there’s NO WAY they’re doing well on the test right!?

Such an obnoxious point of view. Of COURSE your job has no causative effect on your natural cognitive ability. What a ridiculous way to try and look down on the poor.

Why would they rush when they have a specified time limit to complete it that everyone else suffers from? Does having less free time make you stupid? Is Jeff Bezos a brain-dented man as a result of being busy?

Please tell me how having a free period in your schedule causally makes you smarter.

A smarter approach would be to run an experiment and look at what the data says. Just because you cannot think of any doesn’t make it so.

It doesn't. The fact that the IQ test takes a lot of time means that generally only people who want validation of their high IQ will be motivated enough to take the test.

"Test your fitness with this 1 hour workout! ... Hmm our totally unbiased test shows that everyone is really fit."

See? Unbiased sampling is really hard but an hour long test means you're not even trying. (Which tbf they might not be.)

I think you took my comment the wrong way. I was suggesting the opposite.

To map the raw score out of 60 to a population (normalize) you would need to control for biases like free time. For example, with the cooperation of institutions, it could officially replace an hour of school or work. You can look up such normalization tables for this test.

continues to be ironic

“Test your mountain climbing ability by summiting K2!” Hey wow turns out 70% of the population can summit K2!

You’re right, I completely misread it, thank you for correcting me

That’s true of every test, right? Is there a single test where unmotivated people do well?

And an hour is a pretty reasonable ask of most people. How long do you think studies usually take?

Oh man you are lost lol

> And an hour is a pretty reasonable ask of most people.

lol no. Most people are not going to spend an hour doing an IQ test. 5 minutes? Sure. Look at how many people here are commenting about it - and HN has a very high concentration of people that love IQ tests.

> That’s true of every test, right?

You can reduce this bias by either making the test a lot shorter (5 minutes) or paying or forcing people to take it (e.g. tests in school don't suffer from this bias).

I don't think you're ever going to get anything that isn't totally useless out of a 5 minute IQ test, and I have no idea why you'd think HN has a high concentration of people that like IQ tests? I mean, maybe conceptually?

Anyways, asking for an hour of someone's time is fine. I really don't believe that dumb people would, for some reason, be particularly short on time. Maybe you're right that we won't get a fantastic sample of people working 4 jobs at once with 8 children at home.

> I don't think you're ever going to get anything that isn't totally useless out of a 5 minute IQ test

Depends what you're trying to learn, but yeah it would be difficult. That's not really relevant to whether hour-long tests are biased due to their length though.

> I have no idea why you'd think HN has a high concentration of people that like IQ tests?

Because it's full of programmers and geeks who value intelligence and generally think of themselves as more intelligent than the average person. This is really obvious IMO.

> I really don't believe that dumb people would, for some reason, be particularly short on time.

I never said they would. I said that they would be unlikely to invest that free time in an hour long IQ test.

There's no time limit on the test either, so people could take arbitrarily long.

But it seems from what's written at the start that what they're looking for is correlations between the different types of questions, rather than scores across people