Why does IQ cluster around 100 points?

[Speakpigeon]

It clusters around 100 per definition.

Why are people with an average IQ not great at solving complex problems, and why is this true for a majority of people?

Our social organisation is characterised at least for now by social inequalities, including living conditions, educations, health etc. As I see it, intelligent people have an objective interest in the renewal of the population of people who are of average intelligence, essentially to do the jobs they don't fancy doing but are nonetheless necessary to the economy and to the privileged few. Our social organisation is effectively re-enforcing the unbalance in the natural intelligence spread. The increase in IQ scores seems to me to be a direct effect of the industrial, democratic and meritocratic politics that has developed over the last few centuries and that prevails now. It was in the interests of intelligent people to develop our industrial, and therefore technological, and therefore scientific capabilities and therefore expertise, with a huge increase in the number of students as well as better paid workers. Isn't that good enough?

But even if intelligent people somehow made sure the less intelligent ones vanished, there would still be a natural spread, only less pronounced, and perhaps with a much smaller population.

I'm not sure that would solve the problems humanity is facing, though.

Again, it's emotions that decide what we want to do. So, if war, then war, intelligent or not.
EB

 

[Loren Pechtel]

Nothing says anybody has to get the mean. In a normal bell curve it's the most common result but it doesn't have to be. Consider that computer literacy for seniors class I mentioned--had there been a test I don't think there would have been anyone at the mean as it would have fallen between the clueless and the ones who got it.

Of course you would.

You would have the people who are really clueless. Then a little more who are a little less clueless.

You will get a peak where people are midway between clueless and not clueless.

Then you would get get less and less people who are less and less clueless until you have a few who were the least clueless.

What you fail to understand is that clueless was a binary state. Either they got the idea of a control key or they didn't. This was before rodents and menus invaded our systems, the control key was the only way to accomplish many tasks. If you couldn't comprehend that ctrl-i (insert) was a different key than i (the letter) you were going to be badly lost.

 

[ronburgundy]

Because the answers are intercorrelated.

The answers on an IQ test are also intercorrelated and yet you get a unimodel normal distribution.

I don't disagree with your main point that a normal distribution is what you'd expect anyway, but you can't infer that intelligence is normally distributed from the distribution of IQ scores. If we had an absolute measure of intelligence, which we don't, it could have a bimodal distribution, or a heavily skewed distribution, or one with a long plateau of essentially uniform frequency - IQ scores would still be normally distributed because the tests are specifically designed to make that happen.

Incidentally @rousseau: a normal distribution at any one point in time is also what you'd expect if intelligence was currently undergoing heavy selection in one direction, so the shape of the curve, even if IQ values were absolute, does nothing to infer that the population is at or near any kind of "sweet spot".

First, given that nearly every measurable trait that has a continuum of values is normally distributed in the human population (or any animal population), the odds are extremely high that intelligence is normally distributed.

Second, if IQ scores are produced by one's level of intelligence, then their normal distribution is evidence that the thing producing them is normally distributed. The fact that the scores are converted into a relative scale does not inherently create a normal distribution.
The actual tests produce absolute scores that range from zero to some maximum possible score of number of test items answered correctly. If the underlying intelligence were non-normal, those raw scores of correct items would be non-normal. And if those raw scores were non-normal, then the distribution of z-scores they are converted into (essentially what IQ scores are), would be non-normal. Standardizing scores does not normalize them, it merely converts the raw scores into a score that reflects distance from the mean rather than from zero.

 

[Jokodo]

I don't disagree with your main point that a normal distribution is what you'd expect anyway, but you can't infer that intelligence is normally distributed from the distribution of IQ scores. If we had an absolute measure of intelligence, which we don't, it could have a bimodal distribution, or a heavily skewed distribution, or one with a long plateau of essentially uniform frequency - IQ scores would still be normally distributed because the tests are specifically designed to make that happen.

Incidentally @rousseau: a normal distribution at any one point in time is also what you'd expect if intelligence was currently undergoing heavy selection in one direction, so the shape of the curve, even if IQ values were absolute, does nothing to infer that the population is at or near any kind of "sweet spot".

First, given that nearly every measurable trait that has a continuum of values is normally distributed in the human population (or any animal population), the odds are extremely high that intelligence is normally distributed.

I'm not denying that...

Second, if IQ scores are produced by one's level of intelligence, then their normal distribution is evidence that the thing producing them is normally distributed. The fact that the scores are converted into a relative scale does not inherently create a normal distribution.

They are not solely the product of one's level of intelligence. They are the product of one's level of intelligence when applied to a test that has been carefully crafted in order to be useful for the purpose of easily assessing where your level of intelligence is relative to your peers, by people working under the (reasonable but unproven and unprovable) assumption that intelligence should be normally distributed, who will evaluate test designs based on how closely they achieve that outcome, and who are building on a century's worth of experience in designing test from that perspective. And even so, the raw scores may not have a normal distribution: This review paper lists some of the methods that have been used to achieve normal distributions from whatever the raw scores may be: http://www.ets.org/Media/Research/pdf/Angoff.Scales.Norms.Equiv.Scores.pdf

The actual tests produce absolute scores that range from zero to some maximum possible score of number of test items answered correctly. If the underlying intelligence were non-normal, those raw scores of correct items would be non-normal.

And sometimes they are (the raw scores). And sometimes the range includes non-negative numbers (if a penalty for false answers to discourage guessing is included). And the weighting of different sub-components of intelligence varies (e. g. some tests deliberately exclude or minimise the role of "verbal intelligence" since their designers suspects that its results are too much a function of vocabulary size and thus upbringing rather than measuring anything intrinsic to the person, others attach a great deal of weight to it). IQ tests don't measure just one thing, they measure a number of faculties with varying degrees of intercorrelation. Each of them individually is very likely to be normally distributed, but it's already much less certain that all are.

And if those raw scores were non-normal, then the distribution of z-scores they are converted into (essentially what IQ scores are), would be non-normal.

Not all IQ tests operate with so simple a conversion algorithm. Some use essentially hard-coded mappings from raw scores to IQ scores. Here's an example: https://www.us.mensa.org/AML/?LinkServID=8E2A4401-C884-F2CB-AC116EDD5F77A0F8 - other examples can be found in links I provided throughout this thread. In this example, five extra points in the raw scores from 75 to 80 buys you 14 points on the IQ scale, while 5 extra points on the raw score between 50 and 55 only buys you 4 IQ points. A similarly arbitrary mapping from raw scores to IQ, but one where the spread is largest in the mid-range of raw scores, would already get you a moderately bimodal distribution.

Standardizing scores does not normalize them, it merely converts the raw scores into a score that reflects distance from the mean rather than from zero.

If conversion to z-scores and subsequent scaling is all that is done to derive the final score. Not all tests operate like that, and the ones that do are the product of a century worth of experience in how to make sure you get a normal distribution without too much post-processing.

 

[Trausti]

Polygenic scores mediate the Jewish phenotypic advantage in educational attainment and cognitive ability compared with Catholics and Lutherans.

 

[steve_bank]

If you believe a single test of 100 or so questions can tell how smart you are avoiding a debate over what smart is, the I'd like to sell you AZ bridge in Brooklyn. Cheap. An intelligent person IMO would never get hung up in the question of IQ and IQ resting.

The only true comparisons are what you achieve. IQ and IQ tests are pure vasnity.

As I understand it if want your IQ tested you can go to a credentialed service which administers a battery of tests.

 

[fromderinside]

Your thoughts?

Demand characteristics. Kurtosis ignored.

 

[Trausti]

If you believe a single test of 100 or so questions can tell how smart you are avoiding a debate over what smart is, the I'd like to sell you AZ bridge in Brooklyn. Cheap. An intelligent person IMO would never get hung up in the question of IQ and IQ resting.

The only true comparisons are what you achieve. IQ and IQ tests are pure vasnity.

Not untrue; but there is a strong correlation between IQ and achievement. If IQ was scattered all over the range for high achieving individuals, that would suggest IQ is meaningless. Yet, that's not the case.

 

[untermensche]

If you believe a single test of 100 or so questions can tell how smart you are avoiding a debate over what smart is, the I'd like to sell you AZ bridge in Brooklyn. Cheap. An intelligent person IMO would never get hung up in the question of IQ and IQ resting.

The only true comparisons are what you achieve. IQ and IQ tests are pure vasnity.

Not untrue; but there is a strong correlation between IQ and achievement. If IQ was scattered all over the range for high achieving individuals, that would suggest IQ is meaningless. Yet, that's not the case.

Is the plumber who makes more than the doctor under achieving? Is the bodybuilder who works at a warehouse underachieving? Is the stay at home mom underachieving?

There is no objective measurement of achievement.

And no test that measures the full complexity of human intelligence.

People that do good on IQ tests tend to do better on most written tests. Period! They are not better leaders or deserving of leadership. They are not better people.

So while high test scores will get you very far in school which will allow you to start at a higher place in the real world after school there is no evidence it means you will achieve more.

IQ tests do not look at social intelligence. They do not look at creativity. They do not look at emotional stability. They do not look at tenacity and the endurance to stay on a task.

Some of what IQ tests measure is merely the desire to do well on the test. They measure to a degree how much the person cares about the test.

But they do not look at much that really matters beyond memory and exposure to certain kinds of mental problems at an early age.

Extremely high IQ scores really don't translate to much beyond notoriety. The human with the super high IQ score still suffers all the same human problems and weaknesses of the ordinary scorer. They are not necessarily high achievers. They are not special humans with special powers.

They can score high on tests.

My IQ scores in school were in the low 130's. Scores that high or higher include a lot of people. It is no great feat.

And there is no evidence a higher score is necessary to understand anything or perform any task.

 

[fromderinside]

Two thoughts. Exceptions prove rules and your status is not evidence.

 

[untermensche]

My status is only given because I am tired of people saying I have some problem with IQ tests because I was not very good at them.

I have a problem with IQ tests because they do not measure intelligence.

They measure memory and prior exposure to similar problems. And they measure motivation and caring about the test.

A person can work hard and raise their IQ score.

It is not a measurement of something static that actually exists like height.

 

[Loren Pechtel]

My status is only given because I am tired of people saying I have some problem with IQ tests because I was not very good at them.

I have a problem with IQ tests because they do not measure intelligence.

They measure memory and prior exposure to similar problems. And they measure motivation and caring about the test.

A person can work hard and raise their IQ score.

It is not a measurement of something static that actually exists like height.

IQ tests are a pretty good measure of your ability to apply your mind to many situations. You're not going to be able to raise yourself above your potential but if you're not up to it then hard work can raise your IQ.

 

[untermensche]

My status is only given because I am tired of people saying I have some problem with IQ tests because I was not very good at them.

I have a problem with IQ tests because they do not measure intelligence.

They measure memory and prior exposure to similar problems. And they measure motivation and caring about the test.

A person can work hard and raise their IQ score.

It is not a measurement of something static that actually exists like height.

IQ tests are a pretty good measure of your ability to apply your mind to many situations. You're not going to be able to raise yourself above your potential but if you're not up to it then hard work can raise your IQ.

No.

Not "many" situations.

A very few situations. A very narrow range of situations. Situations that are all reduced to a written test.

Test taking intelligence has no correlation to real world intelligence. They do not look at creativity, leadership skills, ability to stay on a real world task longer than the time to take the test.

But test scores are given a lot of value because people have to be separated some way.

There must be losers. And people with large memories must be found. Memory is something some professions require.

So a meaningless method of separating people and creating losers who have limited opportunity must be invented.

Social intelligence is just as important as anything on an IQ test.

That is the intelligence people use in the real world.

The only reason test scores seem to relate to real world success is because high test scores allow you to start higher.

But we could do it differently and use some kind of social intelligence indicator to separate people and then end up with totally different people succeeding and not succeeding.

 

[Loren Pechtel]

My status is only given because I am tired of people saying I have some problem with IQ tests because I was not very good at them.

I have a problem with IQ tests because they do not measure intelligence.

They measure memory and prior exposure to similar problems. And they measure motivation and caring about the test.

A person can work hard and raise their IQ score.

It is not a measurement of something static that actually exists like height.

IQ tests are a pretty good measure of your ability to apply your mind to many situations. You're not going to be able to raise yourself above your potential but if you're not up to it then hard work can raise your IQ.

No.

Not "many" situations.

A very few situations. A very narrow range of situations. Situations that are all reduced to a written test.

I've taken an IQ test that was mostly not written.

Test taking intelligence has no correlation to real world intelligence. They do not look at creativity, leadership skills, ability to stay on a real world task longer than the time to take the test.

You're not going to be too good at the first two without intelligence.

The only reason test scores seem to relate to real world success is because high test scores allow you to start higher.

But we could do it differently and use some kind of social intelligence indicator to separate people and then end up with totally different people succeeding and not succeeding.

Because it has been found that those who don't do well on an intelligence test aren't going to do well at the fields that need intelligence.

 

[untermensche]

It's not an intelligence test.

It's an IQ test and it only measures a small fraction of intelligence.

It does not measure social intelligence or creativity or intellectual honesty or intellectual endurance or so many things that are important.

 

[Loren Pechtel]

It's not an intelligence test.

It's an IQ test and it only measures a small fraction of intelligence.

It does not measure social intelligence or creativity or intellectual honesty or intellectual endurance or so many things that are important.

Saying it doesn't measure certain important things doesn't mean it's not an intelligence test. Only one of those things you list is an aspect of intelligence.

 

[untermensche]

It's not an intelligence test.

It's an IQ test and it only measures a small fraction of intelligence.

It does not measure social intelligence or creativity or intellectual honesty or intellectual endurance or so many things that are important.

Saying it doesn't measure certain important things doesn't mean it's not an intelligence test. Only one of those things you list is an aspect of intelligence.

It looks at a tiny sliver of intelligence.

But doing well on tests is given undeserved reverence.

There is no correlation between doing good on IQ tests and doing good in life.

But people that do good on IQ tests tend to do good on many kinds of written tests so these people tend to do well in school which gives many of them pathways that others can't take.

Some opportunities are only given to people with high test scores.

And since only people with high test scores get the opportunity and because some people that have high IQ scores will randomly also have other skills, like good social intelligence, there becomes a false impression that high IQ scores also translates to "success".

But really it is just the exclusive opportunity translating into success.

 

[rousseau]

Allow me to rephrase:

You: "Couldn't it be that high IQ people are too socially inept to succeed?"
Me: "If anything, the evidence suggests that high IQ is positively correlated to social intelligence."*
You: "But that's just what I'm saying!"

*To be fair, the correlation, it seems, can only be inferred up to an IQ range of around 120-130. Beyond that, the data is too sparse to make any meaningful conclusions, though some studies suggests that some kind of bifurcation happens, with a bimodal distribution of low and high social competence among extremely high IQ individuals. Also, I seem to have brought that up in a different post.

I'd differentiate social intelligence from crystallized and fluid intelligence, which are two different skills. My point was that the more critical aspect of our neurophysiology is being able to exist in a group, and so our evolution has historically been more about social intelligence, than the kind of mathematical/logical intelligence that IQ measures.

I'd guess that usually 'just enough' mathematical intelligence will do, but having no social intelligence is a non-starter. This re-iterates the point because the vast majority of people are more socially intelligent than mathematically intelligent.

 

[rousseau]

So if you want a meaningful discussion, you should be asking "why are people poor at task XYZ", not "why are people stupid"?

That's fair, but 'why are people poor at task XYZ' is what I've been trying to imply from the very beginning, without being explicit about what they're poor at. We're still talking about how I'm wording the question, rather than just groking the meaning and talking about it.

Most people have weak mathematical/logical intelligence beyond very basic reasoning skills. Why?

 

[untermensche]

Allow me to rephrase:

You: "Couldn't it be that high IQ people are too socially inept to succeed?"
Me: "If anything, the evidence suggests that high IQ is positively correlated to social intelligence."*
You: "But that's just what I'm saying!"

*To be fair, the correlation, it seems, can only be inferred up to an IQ range of around 120-130. Beyond that, the data is too sparse to make any meaningful conclusions, though some studies suggests that some kind of bifurcation happens, with a bimodal distribution of low and high social competence among extremely high IQ individuals. Also, I seem to have brought that up in a different post.

I'd differentiate social intelligence from crystallized and fluid intelligence, which are two different skills. My point was that the more critical aspect of our neurophysiology is being able to exist in a group, and so our evolution has historically been more about social intelligence, than the kind of mathematical/logical intelligence that IQ measures.

I'd guess that usually 'just enough' mathematical intelligence will do, but having no social intelligence is a non-starter. This re-iterates the point because the vast majority of people are more socially intelligent than mathematically intelligent.

Social intelligence is ultimately how things are sorted out in the real world.

If you can score high on an IQ test but have poor social intelligence, unless you can write or paint or do some other thing that does not require social interaction you will not go far.

The people given opportunity because they have learned and retain superior test taking skills and also have social intelligence can more easily rise to positions of authority.

In a top down society.

Where work places are rigid top down dictatorships.