Are IQ tests generally administered?

[Stranger_On_A_Train]

@DSYoungEsq

For example, the SAT has had over the years a high correlation with measured IQ. That’s why I’ve always told my Sophomore Geometry students not to get overly worried about “studying” for the SAT, because it’s not really that dependent upon information and more dependent upon so-called “intelligence”. As an example, I go through a few of the math problems and show that, in almost every case, the key to the answer is identifying some relationship, the “trick” to the question, rather than knowing specific things about Geometry or Algebra.

It is true that the SATs are a test of reasoning ability rather than knowledge; the same is true for the ACT, basic GRE, and many other non-discipline acceptance tests. This is intentional in what they are testing is not some specific body of knowledge but general literacy and the ability to apply it. As a metric for academic success, college acceptance tests are somewhat correlated versus IQ tests where there is actually a poor correlation; however, many organizations (like the afformentioned Mensa) accept some scaling of college acceptance tests as demonstrating a threshold of intelligence.

@Shagnasty

However, real IQ tests are not designed that way. It isn’t just people sitting around making up questions that they think someone should know. Instead, it is math (statistics) used to differentiate people along a normal (Bell) curve within their own group and all other groups as a whole. It doesn’t really matter what the question or task is. As long as it differentiates people along a normal curve and coorelates highly with questions on similar tests and real world measures, it is considered valid and reliable.

It is basically the same thing as if you tried to design a test of generic athletic ability. It would be something like a modern decathlon with some skills from various team sports added. It won’t tell you who is going to be an athletic superstar (even college sports careers can’t do that) but it will narrow the field to tell you the group you should be looking at.

Slight nitpick; IQ is distributed on a normal or Gaussian distribution; this is sometimes colloquially referred to as a “bell curve” because of its shape (which is roughly like a classical church bell) but it is not named after a person and shouldn’t be capitalized, except of course in the case of the controversial book by psychologist Richard J. Herrnstein and political scientist Charles Murray which was interpreted by some as making the case for intelligence as an innate genetic quality.

Without getting into the nature vs. nurture argument and any of the attendant discussions of culture, ‘race’, et cetera, two overarching criticisms of the intelligence quotient should be identified; one is that intelligence is capable of being measured as a single metric across all populations based upon a written test which requires language and cultural literacy, and the second is that intelligence (or whatever is being measured by an IQ test) is distributed in a Gaussian fashion (randomly about a central mean).

The first is obviously untrue; if I asked you to take an intelligence test written in Magyar and referencing Hungarian culture, you would excel if you were from Hungary and would likely fare poorly otherwise. IQ tests are innately biased toward the kind of intelligence and literacy that the developers are measuring against. Even in the case of non-verbal tests, the testing is biased toward specific reasoning methods used by the target population and may not reflect a universal cognitive mechanism.

The second is provisionally untrue; while IQ data is fit to a normal distribution, there is significant reason to believe that intelligence (however you measure it) is distributed in a log-normal fashion, just as many physical continuous characteristics such as height is. A log-normal distribution provides a very small number of people on the very low end, a clustering of people at somewhat less than the mean, and a long thin tail of people above the mean with an asymptote giving you a handful of extremely smart people (‘geniuses’) in a sufficiently large population. The fixing of IQ as a normal distribution likely skews the mean (which by the definition of IQ is 100) low, making people who are above the mean appear more intelligent than the statistics should properly justify. A better way to assess IQ (whatever it is) would be to state a result in terms of a variance or standard deviation from the reference population. This would make it independent of any assumption of how intellect is distributed and would just tell you how high or low your score is in comparison to the mean and breadth of the distribution.

As to whether IQ is a useful metric for any practical purpose I don’t have an opinion formed by actual data, thought he anecdota I’ve experienced suggests that high IQ does not correspond to academic or career success as much as hard work, determination, and focus does. There are plenty of people who believe that being smart will reap awards by itself without effort, and a lot of people of otherwise mean intellectual abilities who persevere over challenge and hardship to do impressive work which is celebrated. And I say this as someone who has a score from the WISC and S-B tests that is well above the supposed mean. I would rather hire someone who has worked hard to accomplish goals and feels comfortable questioning his or her own conclusions over an arrogant self-proclaimed genius who feels entitled to adulation for scoring well on a test and responds defensively to challenge or question.

Stranger

[Flyer]

@Marvin_the_Martian

When i was in high school the entire school took the Lorge-Thorndike IQ test. This was a standardized test taken in a group setting (fill in the circle with a #2 pencil), so likely rife with cultural bias and other flaws. I went to a private high school, so I don’t know if it was common across all schools.

ETA: neither of my kids took a standardized IQ test in school, but they both had ones administered by Psychologists - one for special needs purposes (needed a measured IQ below a threshold to qualify for state services) and one for qualifying for the gifted program. If you average their results it comes out right around 100…

I took that one, or one very similar, in 7th grade. The result I got was very close to the result on the Mensa entrance exam when I took that about 8 years later.

@Chronos

There are no IQ tests online. There are plenty of things online that claim to be IQ tests. On the most commonly-used one, last I checked, answering every question completely randomly still gives you a 104.

That result is not terribly surprising, considering that it’s close to the average score. In other words, it’s probably a well-designed test.

I don’t remember the name, but I once did one of those online tests. The result was just one point different from my result on the Mensa entrance exam.

[Thudlow_Boink]

@Chronos

On the most commonly-used one, last I checked, answering every question completely randomly still gives you a 104.

@Flyer

That result is not terribly surprising, considering that it’s close to the average score. In other words, it’s probably a well-designed test.

I don’t follow your logic. How is it a well-designed test if an average person would get the same score from guessing completely randomly as he would get, or should get, from trying his best?

[Chronos]

You think that the average person can’t do any better than random guesses, on every single question on a test? Most tests, random guesses will give you not only a failing grade, but a very badly failing grade.

[pulykamell]

I’m 41, and I don’t think I know anyone my age who took an IQ test. I mean, sure, I took one when I was a kid from a book called “Know Your IQ” or something like that, but that doesn’t count for shit. No standardized IQ tests in school, just something called the Iowa Tests and then later California Assessment Test or something like that.

[pulykamell]

@Thudlow_Boink

I don’t follow your logic. How is it a well-designed test if an average person would get the same score from guessing completely randomly as he would get, or should get, from trying his best?

Yeah, that makes no sense at all. So somebody who gets something like an 80 on the test is “actively” stupid in the sense that they have to try to do worse than chance.

[usedtobe]

@Colibri

Thanks for the clarification. However, I note that the actual name is “Black Intelligence Test of Cultural Homogeneity,” not “Black Intelligence Test Correlated for Honkies,” as usedtobe had it. And his question is not one that is on the actual test. So it would be subject to moderation even if he had made it clear what he was referring to.

The meaning of the last two letters is open to “to whom you are speaking”.
I remembered that there was a “real academic meaning”.
The other (correlated…) was used in some circles.

The test was defined about the time that “Standardized Testing” came under fire.

The two names was another dig at the culturally defined testing.

[Flyer]

@Thudlow_Boink

I don’t follow your logic. How is it a well-designed test if an average person would get the same score from guessing completely randomly as he would get, or should get, from trying his best?

You (and the other people who responded) are conflating multiple points.

For starters, the average person is not very likely to be an active poster on a website like this. The average Doper is virtually certain to be more curious, better read, and more intelligent than the average person. The subset of the overall population that is represented here is largely self-selecting for those things.

Furthermore, an IQ test is not the same thing as an academic test. You will not find a bell curve on an academic test–the results WILL be skewed if the teacher is in any way effective. An academic test measures your actual knowledge about a specific subject. An IQ test measures how effectively your brain works.

Since a disproportionate number of IQ scores are near the average, any degree of randomness – including random answering – in any given test-taking attempt is by definition likely to result in a score that is near average. That is precisely how a bell distribution works–any random result has a “higher than average” chance (so to speak) of being average.

[eburacum45]

That seems to make no sense. If the questions are multiple choice and there are 4 options, then random responses will only be right 25% of the time. This cannot possibly correlate with an average score unless everyone answers at random, which I am fairly sure is not the case.

[thelurkinghorror]

@eburacum45

That seems to make no sense. If the questions are multiple choice and there are 4 options, then random responses will only be right 25% of the time. This cannot possibly correlate with an average score unless everyone answers at random, which I am fairly sure is not the case.

And many of them are not even multiple choice but free answer, so 25% is the maximum. Like if it asks you to complete a Fibonacci sequence, and you answer “w”.

[pulykamell]

@eburacum45

That seems to make no sense. If the questions are multiple choice and there are 4 options, then random responses will only be right 25% of the time. This cannot possibly correlate with an average score unless everyone answers at random, which I am fairly sure is not the case.

Agree. The explanation makes no sense on any level. To score above the average, all you need to do is answer a few questions you are certain of, and then randomly fill in the rest. To score below the mean, you have to answer a few questions purposefully wrong, and then randomly fill in the test. (Or, of course, simply have good or bad luck.) Nothing about “random answers” being the mean makes any sense to me. It’s not really much of a test if I can tell somebody with an IQ of 50 just to mark random squares on his Scantron sheet or answer all C and he ends up with a score of 104.

[Chronos]

Random Dopers might (or might not, but for the sake of argument) have a higher average IQ than the population at large, but that’s irrelevant when you’re talking about random guessing. I might be smart, but my dice aren’t. So why did that test think that my dice were about as smart as a typical human?

[pulykamell]

@eburacum45

That seems to make no sense. If the questions are multiple choice and there are 4 options, then random responses will only be right 25% of the time. This cannot possibly correlate with an average score unless everyone answers at random, which I am fairly sure is not the case.

The only way I could think of designing a test which results in a normal distribution (like the IQ test–it’s actually skewed slightly to the right, but good enough for government work) and a mean score that equals random chance is to design a test where the answers simply are random. Like, say, a “heads or tails” test for a coin flipped a hundred times.

[Chronos]

You could maybe also do it with misleading or trick questions, where most people would confidently think that the answer is one thing, while the actual answer is something else. But those are tough to put into multiple-choice format, because the mere presence of the right answer among the choices can give away the trick.

[pulykamell]

@Chronos

You could maybe also do it with misleading or trick questions, where most people would confidently think that the answer is one thing, while the actual answer is something else. But those are tough to put into multiple-choice format, because the mere presence of the right answer among the choices can give away the trick.

Yeah, if I think hard and imaginatively enough, there probably are other ways you can create a normal distribution centered around a mean that corresponds to random chance.

[md2000]

The only reason a person could do worse than random is if the answers had “looks right” incorrect answers:

2x4=?
(a) cheese
(b)9
©8
(d)a size of wood

People who were “almost” smart would consistently get wrong answers, not random ones. After all, you are aiming to distinguish more the upper scale than the lower. Presenting math or vocabulary questions to people who can’t read (what’s that, below 80?) is probably a waste of time. The questions would be targeted to separate the 100’s, 120’s, 140’s, etc. Therefore 25% would be average.

One cultural bias question I recall cited goes something like this - “Joe has a house on a lot 55x130; he wants to put up a hedge around the lot with two 6-foot gates and a 12 foot wide driveway opening. What is the total length of the hedge?”

This presumes a middle-school inner city type knows what a hedge is.

[Fiveyearlurker]

@pulykamell

I’m 41, and I don’t think I know anyone my age who took an IQ test. I mean, sure, I took one when I was a kid from a book called “Know Your IQ” or something like that, but that doesn’t count for shit. No standardized IQ tests in school, just something called the Iowa Tests and then later California Assessment Test or something like that.

Ok, I’m exactly the same age, so I’m guessing that I just never took one.

[susan]

The major IQ tests aren’t multiple choice. As thelurkinghorror described, you can’t choose a random answer because no answers are provided. The average score for guessing or generating random answers on a WAIS isn’t 104, it’s “invalid protocol.” You can’t answer a question like “Repeat this string of numbers” with “Gargoyle.” Well, you can, but it won’t generate an IQ range (though it might get you some descriptive labels).

[Pilgrim_Shadow]

@electronbee

Out of curiosity, why were you given these tests, was it due to your learning disabilities? I only ask as I have acquaintances who have children with learning disabilities (severe as in mute with grunts, can’t tie shoes, seizures, etc.) and it’s the same thing: kid is given an assessment test with the “happy” verdict that their child is actually in the cumulative “superior/high/advanced” range, with a few exceptions that they need to work on. Then commences some entirely non-empirical/non-scientific treatment. Cha-ching!

LOL! I always had a sneaking suspicion that the whole IQ thing was bogus, and I was the tardo the kids at the bus stop said I was.

Indeed, by the time I was in kindergarten, the experts were already telling my parents I was autistic or retarded or both. I went to Waldorf school K-5, which was nurturing, but they couldn’t teach for shit. So, I started going to public school in grade 6. That’s when they tested me for “special needs” this, that, the other.

I also developed an absolutely soul-crushing case of anxiety/depression disorders, which is why I had the battery of intelligence/psych tests done as an adult. It was a vain search for treatment of a tragic state of being only time would mitigate.

My current therapist suggested doing these tests again for measure, but I decided it wasn’t worth the money, and I’m through with all that testing BS.

[Tigers2B1]

If I had to guess - the online test in question yields a 104 score with random answers because an ‘average’ score on that test is much closer to 150 than 100.