Why does IQ cluster around 100 points?

[Jokodo]

No, you don't get it. It is not the only tool we have. It's not even a tool for that purpose - it's a tool for an entirely different purpose. Cracking hazelnuts open with a chainsaw is a more efficient endeavor than what you're trying to do here.

IQ tests do measure people's problem solving skills and then map them onto a scale designed to tell where a person sits on the population's distribution!

No, I got it the first time you said it, and I got it when I wrote the original post.

The title of the thread is "Why does IQ cluster around 100 points?"

If you'd gotten it, you'd question is equally meaningful as the question "why do number of the type 10^n all get written with one '1' and exactly n '0's? What does this tell us about the distribution of prime numbers?"

 

[rousseau]

No, you don't get it. It is not the only tool we have. It's not even a tool for that purpose - it's a tool for an entirely different purpose. Cracking hazelnuts open with a chainsaw is a more efficient endeavor than what you're trying to do here.

IQ tests do measure people's problem solving skills and then map them onto a scale designed to tell where a person sits on the population's distribution!

No, I got it the first time you said it, and I got it when I wrote the original post.

The title of the thread is "Why does IQ cluster around 100 points?"

If you'd gotten it, you'd question is equally meaningful as the question "why do number of the type 10^n all get written with one '1' and exactly n '0's? What does this tell us about the distribution of prime numbers?"

Yes it was poor wording, which took us 5 days to work out rather than carefully reading the post and inferring it's meaning. My bad.

 

[Jokodo]

The title of the thread is "Why does IQ cluster around 100 points?"

If you'd gotten it, you'd question is equally meaningful as the question "why do number of the type 10^n all get written with one '1' and exactly n '0's? What does this tell us about the distribution of prime numbers?"

Yes it was poor wording, which took us 5 days to work out rather than carefully reading the post and inferring it's meaning. My bad.

I'm sorry to be blunt, but it doesn't get much better from there.

 

[rousseau]

The title of the thread is "Why does IQ cluster around 100 points?"

If you'd gotten it, you'd question is equally meaningful as the question "why do number of the type 10^n all get written with one '1' and exactly n '0's? What does this tell us about the distribution of prime numbers?"

Yes it was poor wording, which took us 5 days to work out rather than carefully reading the post and inferring it's meaning. My bad.

I'm sorry to be blunt, but it doesn't get much better from there.

Well in fairness I wrote it in about 45 seconds between coding tasks, but it's still a fairly straight-forward question that shouldn't have been this hard to gauge.

But at this point I probably should have known better.

 

[untermensche]

Sure. That's why you wouldn't get a sharp two-camps selection. As I said. But we were talking about a bimodal distribution, not a binary categorical difference.

I'm talking about the shape of the curve made with number of people on the y-axis and score on the x-axis.

In any written test with a sufficiently powered random group there will be a single peak.

All you could do is skew the peak and curve to the right or left a little.

You could not create two peaks.

If there is an end spike at a score of O that only shows the test was not a good measuring tool. It failed to measure many.

You will still have one peak.

And every test has questions that most people can't solve.

Giving "weight" to them will not create a peak.

The peak is numbers of people with a score not the difficulty of the question.

And the score is what? The sum of the points awarded for questions they were able to answer.

You are still dealing with a normal distribution of scores. Like you will always find a normal distribution of height from a random group.

If you know of any publicly available (anonymised and machine-readable) detailed result set of a random group of several thousand test takers, with the individual's answers to each question specified, I bet you dollars to dimes that I can come up with a weighting that produces a bimodal distribution with less than half a day's work (though the implementation may have to wait till my newborn sleeps through the night).

You'll get some kind of normal curve. That is all you could get.

And maybe also a single spike at zero which only shows the test failed to measure many.

 

[Jokodo]

I'm talking about the shape of the curve made with number of people on the y-axis and score on the x-axis.

In any written test with a sufficiently powered random group there will be a single peak.

All you could do is skew the peak and curve to the right or left a little.

You could not create two peaks.

If there is an end spike at a score of O that only shows the test was not a good measuring tool. It failed to measure many.

And the score is what? The sum of the points awarded for questions they were able to answer.

You are still dealing with a normal distribution of scores. Like you will always find a normal distribution of height from a random group.

If you know of any publicly available (anonymised and machine-readable) detailed result set of a random group of several thousand test takers, with the individual's answers to each question specified, I bet you dollars to dimes that I can come up with a weighting that produces a bimodal distribution with less than half a day's work (though the implementation may have to wait till my newborn sleeps through the night).

You'll get some kind of normal curve. That is all you could get.

And maybe also a single spike at zero which only shows the test failed to measure many.

5000 euros of Mine against 500 of yours that I can? Based on a real result Set of actual test takers, not some made up data, pseudorandom or otherwise?

 

[untermensche]

5000 euros of Mine against 500 of yours that I can? Based on a real result Set of actual test takers, not some made up data, pseudorandom or otherwise?

Show me what shape you think you can create.

Score on the x-axis

Number of people with that score on the y-axis

How many questions?

 

[Loren Pechtel]

IQ is not problem solving skills. IQ is one particular scale to express problem solving skills, one that happens to be calibrated on the distribution in the population. The fact that the median is is (close to) 100 and the standard deviation (close to) 15 literally only tells us that the people who shipped the latest series of tests did their job right....

There is no reason sores on a test should be a normal curve.

Why just one peak, not two, or three or fifty?

That is the point of the OP.

When there are many basically random factors involved you get a bell curve as the result. Many genes control intelligence, that gives you a bunch of random factors and thus a bell curve is to be expected.

 

[Loren Pechtel]

Intelligence is HOW you solve the problem.

No test can look at that.

If you get the right answer it is just assumed you solved it like everybody else. Or worse how you solved it doesn't matter to the test makers.

This eliminates natural differences.

No two people have the same intelligence.

Even if they both have the same exact score on a test.

Take people and give them an IQ test and then ask them to fix a car.

You will get a similar distribution but a completely different order.

It doesn't matter how people solve a problem, only that they solve it.

 

[untermensche]

Intelligence is HOW you solve the problem.

No test can look at that.

If you get the right answer it is just assumed you solved it like everybody else. Or worse how you solved it doesn't matter to the test makers.

This eliminates natural differences.

No two people have the same intelligence.

Even if they both have the same exact score on a test.

Take people and give them an IQ test and then ask them to fix a car.

You will get a similar distribution but a completely different order.

It doesn't matter how people solve a problem, only that they solve it.

Intelligence is how you solve problems.

Not a score on a test.

 

[steve_bank]

A mix of genetics and environment, from my experience environment rules out. t. Nature vs nurture. Part of it is certainly culture and environment. What you are exposed to as a kid matters. I can trace my technical problem solving skills baack to toys I had as a kid.

Does stimulus as a kid influence how the brain wires itself? The ability to learn language by immersion fades as we get into the teens.

https://en.wikipedia.org/wiki/Intelligence_quotient
https://www.buzzfeednews.com/article/peggy/list-of-average-iqs-per-occupation
https://www.quora.com/How-do-the-av...ession-Which-professions-have-the-highest-IQs

There was a controversial book called The Bell Curve. It showed that occupations tended to be witching IQ bands. According to the curve whatever you think of police they trend to be above average.

https://en.wikipedia.org/wiki/The_Bell_Curve

Try the MENSA workout

https://www.mensa.org/workout
https://www.mensaiqtest.net/

IQ tests are a mix of spatial reasoning, logical reasoning, reading comprehension, math, pattern recognition. Mostly high school level stuff in more than one dimension.

 

[steve_bank]

Intelligence is HOW you solve the problem.

No test can look at that.

If you get the right answer it is just assumed you solved it like everybody else. Or worse how you solved it doesn't matter to the test makers.

This eliminates natural differences.

No two people have the same intelligence.

Even if they both have the same exact score on a test.

Take people and give them an IQ test and then ask them to fix a car.

You will get a similar distribution but a completely different order.

It doesn't matter how people solve a problem, only that they solve it.

Indeed. There are general approaches but in the end only results matter.

 

[steve_bank]

Why do human heights cluster about a mean?

https://www.bing.com/images/search?...orm=IEQNAI&selectedindex=0&exph=0&expw=0&vt=0

A bell or Gaussian empirical curve be it IQ or physical characteristic infers a random variable. Randomness somewhere in the process.

 

[untermensche]

Intelligence is HOW you solve the problem.

No test can look at that.

If you get the right answer it is just assumed you solved it like everybody else. Or worse how you solved it doesn't matter to the test makers.

This eliminates natural differences.

No two people have the same intelligence.

Even if they both have the same exact score on a test.

Take people and give them an IQ test and then ask them to fix a car.

You will get a similar distribution but a completely different order.

It doesn't matter how people solve a problem, only that they solve it.

Indeed. There are general approaches but in the end only results matter.

That is an arbitrary opinion.

If you are looking at intelligence it is how you get the answer that matters.

A score says nothing about the intelligence.

 

[untermensche]

Why do human heights cluster about a mean?

https://www.bing.com/images/search?...orm=IEQNAI&selectedindex=0&exph=0&expw=0&vt=0

A bell or Gaussian empirical curve be it IQ or physical characteristic infers a random variable. Randomness somewhere in the process.

Look at any biological feature across a random subsection of a population of a large enough size and you will find normal distributions.

Look at the speed of geese or the strength of pitbulls or the intelligence of mice.

Or the score on a complex test.

Normal distributions.

That is life.

 

[Speakpigeon]

There is when the scale is designed to make it so.

No.

The scale does not force one peak.

The number of people with a certain score forces it.

But there is no reason why there should only be one peak. Why not a long plateau? Or multiple peaks?

Apparently, you're wrong. Here is what Wiki says.

By this definition, approximately two-thirds of the population scores are between IQ 85 and IQ 115. About 2.5 percent of the population scores above 130, and 2.5 percent below 70.
https://en.wikipedia.org/wiki/Intelligence_quotient

If two-third of the population falls by definition between IQ 85 and IQ 115, then by definition that's likely to produce one peak.

If you want to claim this is wrong, you should be able to articulate why you think it's wrong.
EB

 

[Speakpigeon]

Intelligence is HOW you solve the problem.

No test can look at that.

If you get the right answer it is just assumed you solved it like everybody else. Or worse how you solved it doesn't matter to the test makers.

This eliminates natural differences.

No two people have the same intelligence.

Even if they both have the same exact score on a test.

Take people and give them an IQ test and then ask them to fix a car.

You will get a similar distribution but a completely different order.

It doesn't matter how people solve a problem, only that they solve it.

Sure, but you're replying beside UM's point.

intelligence is a particular kind of capability to solve problems, absolutely not the fact that you solved a problem.
EB

 

[Jokodo]

Score on the x-axis

Number of people with that score on the y-axis

Yes, that's the format we're talking about.

Show me what shape you think you can create.

You know Saint Exupery's "Little Prince" and the drawing of the elephant inside a boa constrictor? Let's say a shape something like that but with a somewhat more expressed trough betwen the peaks. In numerical terms: I shall have succeeded if there's a trough at least 0.2 standard deviations wide (to avoid exploiting random noise), to both sides of which there are peaks of the same width with numbers of people at least 5% higher than inside the trough.

How many questions?

At least several dozen, but whatever real life dataset you can get your hands on. What's more important is the number of test takers be at least several thousands (the lower the number, the easier it is for me to exploit random noise, so this is working in your favour), preferably an order or two of magnitude above that, and that their results be available on a per-question basis.

So are you going to put your money where your words are?

 

[untermensche]

You know Saint Exupery's "Little Prince" and the drawing of the elephant inside a boa constrictor?

No I don't.

Show me the shape you think you can produce and give me a link to some researcher that has produced something besides a normal curve with a written test.

You're basically saying you can produce a lot of questions that people with a lower IQ score can answer that people with a higher score can't.

 

[untermensche]

Apparently, you're wrong. Here is what Wiki says.

By this definition, approximately two-thirds of the population scores are between IQ 85 and IQ 115. About 2.5 percent of the population scores above 130, and 2.5 percent below 70.
https://en.wikipedia.org/wiki/Intelligence_quotient

If two-third of the population falls by definition between IQ 85 and IQ 115, then by definition that's likely to produce one peak.

If you want to claim this is wrong, you should be able to articulate why you think it's wrong.
EB

Give a random population a test and you will always get a normal distribution. One peak.

For so-called IQ tests they arbitrary label the peak at 100 and arbitrarily label standard deviations from the mean.

Labeling the peak "100" does not create the peak or create a normal curve.

The scores produce a normal curve.

Every time.

It is a biological feature of populations which are minor variations around a central genetic "plan".