You see surveyors lying about the observations in front of them, I see surveyors with an obviously incorrectly low evaluation of what constitutes “average” intelligence
Also the term “fairly low” isn’t the same thing as “below average” it clearly carriers all kinds of implications as to that person’s ability to function in the world.
That makes Caplan’s argument even weaker. There wasn’t even an “average” option on the coding? Why should the surveyors even assume that a score of “3” should be “average” then? It’s not like the average SAT score is 800 (right in the middle), so why should someone surveying assume that the average pollee should be “3”?
It seems it would be easy to perceive a ~90 IQ person as average. 105-120 as fairly high, etc. It also seems reasonable that few respondents would be much below 90. Is it equally likely for a respondent have a 70 IQ as it is for one to have a 130 IQ?
Just considering "who can or would answer any question" would bias (strongly?) to the right side of the curve.
LOL. I don't think, as cynical as I am, that I'd say a great majority of people are profoundly stupid. However, growing up on a farm, where practical wisdom is extolled as much as sheer "brains", and then spending a second career in Academia, my take is that people at about 1 to 1.5 standard deviations above the mean have a very real tendency to think they're much smarter than average, when they're obviously not.
I especially found this to be true of my grad students. Those who were at least 2 s.d. above average were noticeably better students. (Those who put in real effort but were 1.5-2.0 s.d above were very good students, but their raw intellectual ability was not that much greater than average.)
At least that is my anecdotal experience. I had, over the years, a few students who were in tha 3.0+ s.d. crowd, and they stood out like like a tall pine in a bunch of shrubbery.
My guess is that if you aren't in a profession that more or less depends on g, then unless you spend a lot of time around those below the mean you probably don't have good comparisons for actual average intelligence.
Well that is the problem either way democracy- collective stupidity but, on the other hand, Aristotle did say that sometimes there is more wisdom in the common people than the elites - which is obviously true as we look around.
“Half the population is below the 50th percentile of intelligence.” is only unequivocally true if intelligence could be measured with such high accuracy AND precision that you can give a unique score to every human alive, and so find the 1 human from 7 billion + who is the exact median-IQ person. In practice IQ measurement is both not so precise, nor so accurate, so a bell-curve distribution leaves a large chunk of people (as median is also the mode) who are exactly on the 50th percentile.
So more than half the population is average IQ or below. But equally, more than half the population is average IQ or above. Both these formulations are more accurate, and the latter one allows for stating the bald truth in a socially acceptable way.
My understanding is that most standard IQ tests are graded on a curve, so that the distribution of results is forced to have a mean of 100 and a standard deviation of 15.
I don't know if the raw scores typically follow a normal distribution.
My work with this is that if you have thousands (or millions) of random test takers the raw scores really do fall pretty much onto a normal bell curve. But we do set the mean arbitrarily at 100.
And not a lot of those below 100 take IQ tests. When I was in grade school, many decades ago, we began taking the Iowa Standard Basic Tests in 4th grade and it continued to 8th grade. Then in HS only those intending to go to college took anything remotely like an IQ test.
And the Iowa Tests weren't very good for measuring the top and bottom ends.
I also took a battery of tests before entering the Air Force that probably could be used as IQ indicators.
Is it just my impression or are women generally more prone to be swayed by social desirability bias?
https://jmpolemic.substack.com/p/serenity-self-indulgence
Women are more agreeable.
You see surveyors lying about the observations in front of them, I see surveyors with an obviously incorrectly low evaluation of what constitutes “average” intelligence
Also the term “fairly low” isn’t the same thing as “below average” it clearly carriers all kinds of implications as to that person’s ability to function in the world.
That makes Caplan’s argument even weaker. There wasn’t even an “average” option on the coding? Why should the surveyors even assume that a score of “3” should be “average” then? It’s not like the average SAT score is 800 (right in the middle), so why should someone surveying assume that the average pollee should be “3”?
It seems it would be easy to perceive a ~90 IQ person as average. 105-120 as fairly high, etc. It also seems reasonable that few respondents would be much below 90. Is it equally likely for a respondent have a 70 IQ as it is for one to have a 130 IQ?
Just considering "who can or would answer any question" would bias (strongly?) to the right side of the curve.
I must be a misanthrope. I find it obvious to the point of irrefutability that the great majority of people are profoundly stupid.
LOL. I don't think, as cynical as I am, that I'd say a great majority of people are profoundly stupid. However, growing up on a farm, where practical wisdom is extolled as much as sheer "brains", and then spending a second career in Academia, my take is that people at about 1 to 1.5 standard deviations above the mean have a very real tendency to think they're much smarter than average, when they're obviously not.
I especially found this to be true of my grad students. Those who were at least 2 s.d. above average were noticeably better students. (Those who put in real effort but were 1.5-2.0 s.d above were very good students, but their raw intellectual ability was not that much greater than average.)
At least that is my anecdotal experience. I had, over the years, a few students who were in tha 3.0+ s.d. crowd, and they stood out like like a tall pine in a bunch of shrubbery.
My guess is that if you aren't in a profession that more or less depends on g, then unless you spend a lot of time around those below the mean you probably don't have good comparisons for actual average intelligence.
Well that is the problem either way democracy- collective stupidity but, on the other hand, Aristotle did say that sometimes there is more wisdom in the common people than the elites - which is obviously true as we look around.
Classic post and timely republishing after Scott Alexander posted:
https://www.astralcodexten.com/p/how-to-stop-worrying-and-learn-to
“Half the population is below the 50th percentile of intelligence.” is only unequivocally true if intelligence could be measured with such high accuracy AND precision that you can give a unique score to every human alive, and so find the 1 human from 7 billion + who is the exact median-IQ person. In practice IQ measurement is both not so precise, nor so accurate, so a bell-curve distribution leaves a large chunk of people (as median is also the mode) who are exactly on the 50th percentile.
So more than half the population is average IQ or below. But equally, more than half the population is average IQ or above. Both these formulations are more accurate, and the latter one allows for stating the bald truth in a socially acceptable way.
Yes it's painful. So, uncomfortable scientists write whole studies based on emotion:
Aporia Magazine Is hereditarian research "abhorrent"?
A point-by-point response to a recent paper.
https://www.aporiamagazine.com/p/is-hereditarian-research-abhorrent?utm_source=post-email-title&publication_id=828904&post_id=154344851&utm_campaign=email-post-title&isFreemail=true&r=6mos7&triedRedirect=true&utm_medium=email
Does "intelligence" naturally fit a bell curve? Or are IQ tests just deliberately scored to produce a bell curve?
Naturally normal: https://x.com/cremieuxrecueil/status/1831755817508598142.
HOW would THAT be done?
My understanding is that most standard IQ tests are graded on a curve, so that the distribution of results is forced to have a mean of 100 and a standard deviation of 15.
I don't know if the raw scores typically follow a normal distribution.
My work with this is that if you have thousands (or millions) of random test takers the raw scores really do fall pretty much onto a normal bell curve. But we do set the mean arbitrarily at 100.
And not a lot of those below 100 take IQ tests. When I was in grade school, many decades ago, we began taking the Iowa Standard Basic Tests in 4th grade and it continued to 8th grade. Then in HS only those intending to go to college took anything remotely like an IQ test.
And the Iowa Tests weren't very good for measuring the top and bottom ends.
I also took a battery of tests before entering the Air Force that probably could be used as IQ indicators.