Higher Vocabulary ~ Higher Income
RSS
Search Text Case Sensitive Exact Words Include Comments
List of Bookmarks
Prompted by a comment below I was curious as to the correlation between intelligence and income. To indicate intelligence I used the GSS’s WORDSUM variable, which has a ~0.70 correlation with IQ. For income, I used REALINC, which is indexed to 1986 values (so it is inflation adjusted) and aggregates the household income. Finally, I limited my sample to non-Hispanic whites over the age of 30 (for what it’s worth, this choice also limited the data set to respondents from the year 2000 and later).
The results don’t get at the commenter’s assertions, because 10 out of 10 on WORDSUM does not imply that you’re that smart really. But the trendline is suggestive. Note that aggregated 0-4 because the sample size at the lower values is small indeed.
(Republished from Discover/GNXP by permission of author or representative)
Follow @razibkhan
English
Arabic
Chinese (Simplified)
Dutch
French
German
Hebrew
Hindi
Italian
Japanese
Korean
Persian
Polish
Portuguese
Russian
Spanish
Swedish
Turkish
Vietnamese
http://www.flickr.com/photos/51954263@N03/6927133195/in/photostream , from the GSS.
The relationship is obviously there. But the average implied IQ of someone in the top 1% is only 108. That’s sort of important, since it argues against the implicit argument a lot of people make that we’re in a perfect or near perfect meritocracy. (NOT claiming you made this argument, but it’s an implication that a lot of people make)
wow, 1%ers are morons!
and just to be clear, the correlation between income & IQ is going to be imperfect. other variables imperfectly correlated with IQ are important (e.g., time preference).
Dear Mr. Khan !
I believe that it is appropriate to mention Charles Murray’s thin, but important brochure “Income Inequality and IQ”, which, I am pretty sure, you know:
http://www.mega.nu/ampp/murray_income_iq.pdf
In that brochure Murray took a sub-sample of National Longitudinal Study of Youth (NLSY). Namely, he took only those particiaants of NLSY, who had siblings in NLSY. Thus he made “environment” as equal as it can possibly be.
Respectfully, F. r.
Conscientiousness together with IQ explain a lot(The Terman study, which had it’s own problems but is all I have to work from, had conscientiousness being roughly twice as important as IQ past 125). But there’s still stuff liberals can complain about. Poor kids with high SAT scores are less likely to go to college than rich kids with low SAT scores. Part of that is due to conscientiousness differences but not all of it.
I think liberals over-reach when pointing that out, because most low-income kids do not have high scores, it isn’t terribly important to income inequality. But it is an important indicator of the efficiency of our educational system.
“wow, 1%ers are morons!”
I suspect that most people would be surprised that the average K-8 teacher is about as smart as the average person in the top 1% of the income distribution.
It’s true that education and IQ strongly predict income. But it’s actually sort of shocking how little income predicts education and IQ. I don’t have numbers off hand, but I remember looking at the GSS and being really surprised how many low-income people had college and post-graduate degrees and how many high-income people didn’t.
[Part of this is that I used the total cumulative sample, things probably have gotten a bit better sorted over time, but I don’t think the effect was *that* large]
[ignore my comment – I did not noticed that REALINC was adusted to 1986 values]
Razib,
I know this point has been made before, but a household income of $70,000 is hardly the 1%. My wife and I live in a 1,260 square foot rowhouse, have two cars from 2003, and had a combined income of around $91,000 last year. This puts us in the top quartile of U.S. households for income, which is arguably upper-middle class (although my wife has a hard time believing this), but certainly not economically elite.
1 –
From what I can tell looking at that chart online, the high point is the 99th percentile, not the 100th percentile, which drops back down to 106. Of course, the sample sizes are probably too small for any percentile to be indicative. Still, it doesn’t look like there is any trend regarding IQ and income (or at best a weak one) above the 90th percentile.
Anyway, more generally…
To a great extent, I think our meritocracy works well at the level of the upper-middle class. No one can doubt that doctors, lawyers, engineers, and scientists have intelligence higher than the general public. I think it’s also generally the case that the smarter and more conscientious in all these professions all achieve more, and if they are so interested, earn more (although they may purposefully choose a less lucrative area within their field).
The question is more on the subject of the managerial/executive class, who as a whole have IQs only a few points above the national median, and are almost entirely responsible for the growth of income inequality in the U.S. All of my exposure to business school curriculum suggests it’s largely the bizzaro world clone of liberal arts, teaching graduates how to appear smart by using buzzwords, and how to apply solutions someone else came up, rather than think independently.
That is not to say that there aren’t smart managers or entrepreneurs. Indeed, I think in some industries, like tech and to a certain extent finance, there are a great many, drawn not just due to money, but the desire to do something new no one else has accomplished. However, being a manager in general involves progressively more time spent in meetings with both superiors and subordinates, and less time actually doing something. The real management innovations are often undertaken by others – industrial engineers or those with heavy math backgrounds, who may be well compensated, but not as much as high-level managers or executives.
I suppose the final question then is twofold.
1. Could management be something which smart people felt generally more engaged doing? If so, would we see significant productivity gains as a society?
2. If the above isn’t possible, is there any real reason why a job which does not actually require as high level of skill as many upper-middle class professional jobs should be compensated at a much higher level? Or is it really just so well rewarded because it’s a great soft landing for the wealthy with only moderate intellectual talents?
I know this point has been made before, but a household income of $70,000 is hardly the 1%
fwiw, #1 is referring to a different analysis that mine above.
According to this I should have the same vocabulary as both of my nephews. Just for the record, the fact that I barely make any money has nothing to do with my IQ or the vastness of my vocabulary. I have an IQ of at least 142 and an associate’s degree in Psychology and working on my bachelor’s. Plan on getting my Ph.D. in the future. Right now I work for the Department of Social Services, but don’t be fooled, it doesn’t pay as well as one might think.
#10, with that IQ you should take a statistics course! 🙂
@10
A general trend over a large population sample doesn’t mean that every member of the population will mirror that trend perfectly. Exceptions happen. If you’re going to flaunt an IQ number, might want to think about doing so in a less silly way.
I’m not trying to sound mean. Just got a degree in psych myself, and I always felt bad for the students who inadvertently provided comic relief for the class by saying stuff like that.
Half Sigma showed that vocabulary loses its correlation with income if education is controlled. I suppose you are just capturing the effect of aptitude entrance exams that test verbal ability.
#10 was meant as a joke, right? Please tell me someone is simply trying to be cute there.
Whether serious or not, it is a reminder that possessing intellect is disconnected with being educated.
Is the correlation of WORDSUM uniform across the scale? I’m pretty sure somebody with an extremely low score isn’t very smart, but at the other end of the scale, I’d expect an English major to have a higher WORDSUM score than an engineer, even though (at least in my experience) engineers tend to be considerably more intelligent.
One example of a distortion:
“most people would be surprised that the average K-8 teacher is about as smart as the average person in the top 1% of the income distribution”
Most K-8 teachers were education majors. Based on SAT derived IQ, people who enter college as education majors are the least intelligent of all college students. Still, I would expect them to have a strong vocabulary.
A problem with the theory “high IQ could lead to high income but much high IQ does not lead to much high income” is that it is very difficult to test, because the two factors that we are comparing “much high IQ” and “much high income” are both relatively rare and, to make things worse, many statistics, polls, etc have a ceiling problem, where it is impossible to differentiate between the “high” and the “very high” (like the wordsum=10, where you could find both “genius” and people with an IQ a bit above average).
The fact that these theory is very popular could be a point in its favour (meaning that matches with the conclusions that many people observe in the real world); however, it is possible that the reason being simply that the “genius” working as librarians or high school teachers are much more “visible” to the average persons than the genius earning millions in Wall Street or Sillicon Valley.
Or perhaps there is some confusion between IQ and Openness in that.
Since when is IQ synonymous with “merit”? To begin with, the term “meritocracy” is vague and non-scientific.
But for the sake of argument, let’s suppose that the term has some meaning consonant with the usual definition of “merit”. To then assert that the fact that the top 1% of earners have a mean IQ of 108 implies in itself that we are not a meritocracy is preposterous.
I’m not saying we are a meritocracy. (In politics, I’d say it’s virtually self-evident that we are not, by any rational definition of the term.) I’m saying that if one wishes to determine whether we are, it makes no sense to focus on a single variable such as IQ as one’s sole or main measure of “merit”.
There a dependency here with age, because peoples’ vocabulary and fluency continues to improve all through the adult life. Also older people are paid more because of maturity and experience, of course. Therefore your graph is incorporating age as well, as a hidden dependency. You need to subset by age and test again.
Obviously, richer people can afford to buy more words.
@19: Not just more words, but better words. You pull out your five-dollar words at the wrong party and some rich guy’ll trump your Abe with his Benjamins.
Vocabulary tends to grow with age due to increase exposure to written and spoken words. Income generally also tends to increase with age, up to the point of retirement. How have you corrected for age?
Behind every very wealthy 100 IQ executive there’s a much less wealthy 130 IQ techie whose ambition mainly revolves around items in World of Warcraft.
Obviously not really true but an element of truth maybe – especially with the importance of IT in generating competive advantage over the last 30 years.
Cool… correlation does not imply causation, though. It could be that people who make less money are immersed in cultures that tend to be less fluent/ eloquent. Ya never know! It’s interesting to look at, though.
Generally the whole “Income peaks when your IQ is at 130” thing is probably bullshit. The Terman study showed persistent and near linear gains past 130. Eminent scientists have higher IQ’s than their scientist peers, etc.
#24, the heuristic is simple. x = your IQ. all IQ above x is zero marginal returns on any domain you wish to measure 🙂
Spearman’s law of diminishing returns is a real phenomenon you know. g becomes less important as it increases and specific skills become more important. of course, that doesn’t mean that everyone at IQ x has those specific skills.
Does anyone have any data refutes half sigma findings on the GSS: http://www.halfsigma.com/2011/07/higher-verbal-ability-leads-to-lower-income.html
#25, I thin the reasoning is more:
“x = your IQ – 10 points. all IQ above x is zero marginal returns on any domain you wish to measure”
implicit in this theorem is that “you” are more intelligent than the social elite (and probably could be even a victim of your own intelligence).
“Eminent scientists have higher IQ’s than their scientist peers, etc.”
I think that this point is not much relevant – the essence of the theory “Income peaks when your IQ is at 130″ is that very high IQ people are more prone to choose arcane fields of work with (comparatively) little financial reward; then, to test that theory, you can’t compare people in the same field of work (scientists with scientists, for example).
Meanwhile, I made another calculation with the GSS.
Comparison of Means
Dependent: WORDSUM
Row:ISCO88(r:110-1120;1226-1240;1311-1314;2112-2460;3100-3480;4100-4223;5100-5220;6113-6150;7110-7520;8111-8400;9110-9333)
Selection Filter: RACE(1);HISPANIC(1);age(30-100);year(2000-2010)
Results:
110-1120 (politicians, I think, perhaps with some military mixed) – 6.94
1226-1240 (corporate managers) – 6.88
1311-1314 (managers of small enterprises) – 6.45
2112-2460 (professionals) – 7.46
3100-3480 (technicians and associate professionals ) – 6.84
4100-4223 (clerks) – 6.57
5100-5220 (service workers and shop and market sales workers ) – 6.02
6113-6150 (skilled agricultural and fishery workers) – 5.30
7110-7520 (craft and related trades workers ) – 5.84
8111-8400 (plant and machine operators and assemblers) – 5.29
9110-9333 (elementary occupations ) – 5.56
The only relevant point that I can conclude from that is “professionals” are indeed more intelligent (or, at least, have a higher vocabulary) than “corporate managers” (of course, it is very possible that these “professionals” more intelligent than “corporate managers” could have also a higher income)
The relative high wordsum of “elementary occupations” is an illusion created by a very high result in the category 9110 (street vendors, I think) – 7.11, higher than managers and politicians (probably some post-hippie types with college degrees).
Codes of ISCO-88:
http://www.ilo.org/public/english/bureau/stat/isco/isco88/publ4.htm
http://www2.warwick.ac.uk/fac/soc/ier/research/links/isco88/isco88.pdf
Like I said before, check out the Terman study(It followed a couple hundred 135+ IQ students for 80 years). http://infoproc.blogspot.com/2011/04/earnings-effects-of-personality.html links to a paper that unpacks income, personality, and education and their effect on income on the terman sample.
The impact of IQ on income is increasing past 135 and linear. It *probably* is lower than the impact below 135, but it’s still there.