What does an IQ of 100 (average) actually imply?

I’ve always wondered about this- people run around saying the average IQ is 100, and that half of everyone is below that.

I know the second part’s not necessarily true unless the median is also 100, so don’t crucify me on that. What I do want to know is what does an IQ of 100 imply? Are there things that aren’t comprehensible to a person of 100 IQ that are to a person of say… 130? Does it take people longer? Is it more of a matter of drawing the wrong conclusions about things and making incorrect decisions? Is it a matter of lack of insight into things?
I’m just really unsure what that 100 IQ means in practical terms. Anyone able to shed some light on it?

The design of IQ tests involves a whole lot of complicated math (statistics and especially correlations in particular). The question sets are designed to differentiate people along a normal curve for a particular mental skill.

The shortest way to put it is that the questions or tasks are ranked in difficulty, not by the people who design them, but by the results of actual test takers from the population of interest. There are certain questions that most people tend to get right even if they fall below the mean overall, questions that people that score 100 overall tend to get correct, and questions that only people who score over 100 tend to get correct.

You would have to look at a specific test to get concrete examples but there are say, math problems that people who score 100 overall are unlikely to get right while those that score over 130 are quite likely to answer correctly. Those specific examples would give you an idea of the differences in ability but the main point is that there is a gradient in difficulty among the questions or tasks and people tend to get most of the questions right up to a certain difficulty point and most of them wrong above it.

You may think I am just restating the obvious about your question but most people don’t understand how IQ tests and the math that goes into them works. That is the extremely condensed version. To know what a particular IQ score actually represents, you have to specify the specific test and score in question and then identify the point where most of the answers switch from right to wrong.

If you just want wild generalities, a score of 100 is generally good enough to earn an undergraduate degree somewhere but not high enough to go to medical school, get a reputable PhD, or major in engineering as an undergrad.

Also, an individual IQ score, I believe, is not only a result of the questions you answer correctly, but also your age and the speed with which you completed the test.

I may be mistaken on this so don’t quote me on it, but that’s what I’ve heard.

Age generally isn’t used anymore although it was among early IQ tests. Don’t feel bad, the majority of psychology probably don’t even realize that. There are timed IQ tests or at least timed sections however and speed is sometimes a factor in the score but not always.

Getting an IQ score of 130 implies that you have better test-taking skills than someone getting an IQ score of 100.

The general populace believes that IQ tests have some sort of mysterious cachet that allows them to measure a pure abstract quality known as “intelligence” free from any cultural influence or educational opportunity. Despite such claims, no such test exists.

The test-writer first has to define what constitutes this “intelligence” they wish to measure. Is it vocabulary? Is it math ability? Is it memory? Is it pattern recognition? Then they construct a test to measure their notion of intelligence. Different IQ tests measure whatever the author has defined as intelligence.

Test taking is also a learned skill. Let me give you a personal example: In early grade school they gave us intelligence tests to determine whether you would go into the smart, dumb, or average class the next year. I did poorly on the first test because I hadn’t been taught the code the tests use. For example, a circle with dots on it is supposed to represent a cookie. I knew what a cookie was, I could identify a real life cookie. I just had no idea that in the test-taking world a circle with dots on it was supposed to represent a (chocolate chip) cookie. In the next year, in spelling class the standard school picture for each word was drawn for us and I learned that the circle with dots in it was a cookie and I did much better in the next year’s test. I did not have more “intelligence” I just learned a skill that was useful for the test. (And think about how a student from a culture that did not have cookies would do on the test.)

Any given version of an IQ test has some sort of cultural bias built into it. It measures what the culture of the test-writer values. It also measures to some extent life experience and educational opportunities. A person who has never had a math class may very well excel at math if given the opportunity, but he will get the math questions wrong on the test.

In other words, you need to look at the assumptions that the given test-writer has made and what the writer is seeking to measure.

See here, Intelligence quotient - Wikipedia, then if you have more questions, ask away.

I taught a class not two years ago where students were still trying to say that (mental age/chronological age) was part of the calculation. Ohh, Psych 101, what are you teaching?

The median and mode are also 100 if the population is normal. This is a core assumption of IQ and parametric statistics. Normal here doesn’t literally mean a nice symmetrical Gaussian, just that it is “close enough” to not be considered skewed.

IQ of 130 = you scored better than 97.72% of people, and worse than 2.28%. Unless you took the Stanford-Binet, where SD = 16, so you “only” did better than 96.99%, worse than 3.01%. And the measure is “performance on an IQ test,” not intelligence directly, although there is correlation. A score of 100 is good enough to be Secretary of the Interior.

But the entire purpose of inventing IQ was to compare at age levels. It was designed to be a guide to figure out who might need extra help at their class level. Thus someone with an IQ of 85 in the first grade would need extra attention. It was not originally intended to be an absolute nor a permanent value.

So then how does an IQ test differ from an achievement test?

If a given child got an 100% on their math test but an 85 on their IQ test, would you assign them to a special tutoring program?

If I understand the question in the OP, it is what is this “intelligence” that the “performance on an IQ test” allegedly correlates with? And my question is, if you can’t rigorously define and quantify this intelligence, how can you claim that the test score correlates with it?

No, I would send him to the principal’s office because he is obviously cheating his ass off on math tests :). That scenario shouldn’t come up much in the real world. If it did, you would have to look closely at why the two results were so different. Either the math test was too easy for everyone (or he really was cheating) or he didn’t do as well as he should have on the IQ test maybe because he didn’t take it seriously, didn’t understand the directions, or was sick that day.

Total IQ is a composite score anyway. It is perfectly possible to be above average in math while having deficiencies in other areas that bring the overall score way down, especially verbal. You would have to look at the case from that angle as well.

There is the standard saying that IQ tests only measure performance on IQ tests but really isn’t true. IQ scores or at least subsections of them correlate very highly with many other kinds of tests that aren’t considered IQ tests. You can translate SAT and GRE scores into close approximations of an IQ score for example just like you can for any test that measures cognitive ability along a normal curve because most of them are highly correlated. That includes most well-designed tests whether they call themselves IQ tests or not. They won’t tell you about social skills or speaking skills for example but they measure their target skill set quite reliably.

They correlate with a whole lot of things to different degrees. The highest correlations are with other well-defined tests like you might take in school or to get into school. Not surprisingly, there is a moderately strong positive correlation between IQ and academic performance but it is a complicated subject. Having a high IQ certainly doesn’t guarantee high grades or even completion of a program because there is a work ethic component that it doesn’t measure and is highly important as well.

In laymen’s terms, the basic idea is to measure raw cognitive capability in isolation and that is all they were ever meant to do. They don’t promise or deny success but they do suggest certain statistical outcomes especially at the highest and lowest levels. There is something very similar to a minimum IQ requirement to be successful in some of the more academically rigorous fields particularly in science, math, engineering, medicine, and many other professional disciplines. An IQ test or equivalent can tell you whether a given individual has a realistic shot of completing a rigorous program given other favorable factors like work ethic.

IQ tests can also help identify people who have cognitive deficiencies and may need special programs to help them. That is better than just someone arbitrarily deciding that someone is ‘retarded’ IMHO.

The middle part of the curve which includes people doesn’t tell you a tremendous amount about individuals as opposed to populations. You can say that a higher IQ is better as general rule but differences of less than one standard deviation (generally 15 points) can be overshadowed by other factors that aren’t measured by IQ tests.

Let me see if I can give a clearer primer on the way IQ tests really work because there is so much misunderstanding about them in general. IQ are basically one giant statistics problem with many sub-parts. IQ tests don’t have any goals or values of their own. They are just a tool that we know measures something highly repeatable even though it isn’t always the most important factor in any given task.

Their main task is separating individuals along a normal curve as cleanly as possible. How that is done is a series of statistical problems of its own. Test designers create questions but they don’t assign weights or importance to them. They give the questions to samples of people and measure how well each one differentiates people along the normal curve. You never want a question that everyone gets right just like you never want one that everyone gets wrong because that tells you nothing. You want questions that people below a certain level get wrong and those above it get right. You generate thousands of these questions (or sometimes tasks like putting puzzle blocks together) and just do the math on which tasks differentiate people the most efficiently along a normal curve.

You build the test with best questions you have according to the math and set 100 as your mean (ideally). It then becomes a potential predictive tool but the test itself has no assumptions about what those scores mean in the outside world. They are just a statistical phenomenon.

Once thousands of people have taken the test, you start studies to see what correlations there are with whatever outcome data you are interested in (grades, graduation rates, lifetime income, overall happiness etc.) Once those are compiled, you have a set of correlations you can use to describe what was being measured in the first place.

IQ theorists don’t assign any mystical traits to an IQ score. They just call the overall score ‘G’ for general trait and have other abstract names for the sub-measures. ‘G’ is a very reliable measure that you can take in many different ways for individuals and it tends to stay incredibly stable over a lifetime barring exceptional events. Nobody really knows what these traits are except as statistical phenomenon. The are just factors like you see in advanced math or physics classes. Some people want them to be more than that or sometimes less but they aren’t. They are highly repeatable across many different measuring tools no matter what they are so that makes them a valid scientific measure.

Back when I was in Psych 101, we learned that there were other “tests” that bore no resemblance to what most people think of as tests, but which have very strong correlations with the results of an IQ test. One of them I remember, for instance, was to ask a child to draw a picture of a person (it might have specified a self-portrait, I can’t remember), and the test was then evaluated based on how much detail the portrait included (did the drawing have hair? Ears? Eyelashes? And so on). I doubt that anyone would say that drawing detailed portraits is just a reflection of “ability to do well on tests”, and yet it seems to be measuring the same thing that IQ tests do.

I have configured my browser to delete cookies. Anybody with an IQ of 100 should be able to do this.

I would LOVE to configure my browser to delete most cookies EXCEPT the chocolate chip ones. This is beyond the skillz of most people with IQ of 100, and may require in IQ of 120 to 130.

This is a case where you have to decide what the two scores mean with a little common sense. Although IQ results tend to be repeatable for a given individual, one test result could be off for a number of reasons. Sometimes an odd result comes in and the teacher and school staff have to reassure the parents that things are OK, just wait for the next test before over-reacting.

Sometimes it’s the other way around. In first grade, my youngest took a series of achievement tests (what we’re calling IQ tests, but with results given as a percentile). He got some amazingly low scores. A few were around 50. Two of the three reading scores were low, with reading comprehension being 21. I got called in because they were Very Concerned.

I looked at the scores. He’d gotten a 98 in science and an 89 in social studies. They were shocked that I wasn’t concerned about the reading scores. Reading is so basic to learning all the other subjects. I asked how he could have gotten such a high score on a written science test if his reading comprehension was really that low.

One of them actually tried guessing that maybe the science questions had been read aloud. I gave her The Look for that. She kept arguing, but the other staff members knew that the meeting was over. They were going to have to wait for the next set of academic tests, or for his grades to dip, before I’d agree to anything. He never tested that low again. Of course, the boy and I had words about the importance of putting a proper effort into those tests, even if they weren’t going to be part of his grade.

It wasn’t until he was in high school that he told me the full reason for the lowest score. He had decided that any story in a test was bound to be a boring one, and had answered the questions without reading the stories. From his score, I can deduce that eash question had five possible answers.

Question : I understand that the IQ tests are tightly timed. How does the distribution changes if one allows a lot more time? Would this only move the curve up? Basically I’m curious whether there can be very “intelligent” but slow people.

Test taking is definitely a learned skill. I’m a very good test taker. I can focus quite effectively for a few hours, an entire semester? Not so much. My predicted GPA based on my IQ and test scores was considerably higher than my actual GPA. It sounds like you took a poorly constructed or administered test, or perhaps at your level of mental development at the time you weren’t quite ready abstracting from cookies to circles and dots.

It’s not 1962 anymore. The test designers are well aware of these problems, and will try to eliminate them. There are cognitive tests which are appropriate for people who are illiterate or not yet reading. Just as there tests designed to measure ability in many specific domains (reading, math, spatial reasoning, processing speed, etc.) An additional thing to keep in mind is that IQ itself was original produced as an attempt to predict how “successful” an individual will be in society. In that case, a test biased towards that society would be appropriate.

I think many people’s negative reaction to IQ is because it says that some people are smarter than others. Nobody has much of a problem saying that some people can run faster than others, or even do complex things like play basketball better than others. But, suggesting that some people are intrinsically smarter seems to cause quite a stir. Lots of excuses are made, such as whatever is being measured isn’t real (whatever that means), or doesn’t matter.

I think in most of these tests, speed is part of what they’re measuring. Some people are smart enough to get all the answers if they have enough time, but not fast enough to get through the test in the time allotted.

This is why a lot of people with learning disabilities can be very intelligent, but not really test that way.

I know that one of the accommodations given to people on certain kinds of tests (specifically, learning disabilities related to reading) is more time.

Or not, as has been said it depends a lot on what the exact test asks for.

I’ve taken several which were all pictures. The advantage of that is, it doesn’t need to be translated in order to give it to people from different places; it can also be given to people who cannot read. The disadvantage is, it favors graphic people over verbal people.

One of the things I have to use and generate in my job is documents explaining processes. When everybody who’s going to work on a process has the same intelligence / communication / learning style, you can use documents which are all paragraphs, all process diagrams or all step-by-step tables. When you don’t know who is going to read it, you need to include all three… which leads to the pictures people* complaining that “this document says the same thing three times!” Yes, yes it does - if you want, you can stick to the pictures. The tables people and the words people will usually be slower at pictures (including all-pictures IQ tests) than the diagrams people, but that doesn’t mean they’re dumber; it means they have different kinds of intelligence (diversity, yay!)

  • IME, eh, I’ve simply never had a tables person or a words person complain that “the picture says the same thing, why do I have to look at it?” or “the table is just a point-by-point of the paragraphs, why do I have to read it?”, it’s always a pictures person complaining that “all these blahblahs say the same thing as the pictures, why do I have to read this shit?” You don’t…