Might schooling raise IQ?

Children who have more schooling may see their IQ improve, Norwegian researchers have found.

Although time spent in school has been linked with IQ, earlier studies did not rule out the possibility that people with higher IQs might simply be likelier to get more education than others, the researchers noted.

Now, however, “there is good evidence to support the notion that schooling does make you ‘smarter’ in some general relevant way as measured by IQ tests,” said study author Taryn Galloway, a researcher at Statistics Norway in Oslo.

Findings from the large-scale study appear in this week’s online edition of the Proceedings of the National Academy of Sciences.

…In 1955, Norway began extending compulsory middle school education by two years. Galloway and her colleague Christian Brinch, from the department of economics at the University of Oslo, analyzed how this additional schooling might affect IQ.

Using data on men born between 1950 and 1958, the researchers looked at the level of schooling by age 30. They also looked at IQ scores of the men when they were 19.

“The size of the effect was quite large,” she said. Comparing IQ scores before and after the education reform, the average increased by 0.6 points, which correlated with an increase in IQ of 3.7 points for an addition year of schooling, Galloway said.

The summary is here, the paper is here, and for the pointer I thank Michelle Dawson.


Might schooling raise IQ test scores?


It's not like psychometricians haven't looked into this before... the consensus is that schooling has a significant but *temporary* effect on IQ. The gains fade more and more the further you get from the "shock" of increased schooling.

I don't think those kind of effects necessarily fade away. It seems that school reforms are one of the things that could explain the Flynn effect. On the other hand, a very similar school reform apparently had little effect on IQ in Finland: http://hse-econ.fi/pekkarinen/Tracking_revised_nov_10_2010%20(2).pdf

However, I have trouble parsing that report. Was the effect 0.6 points or 3.7 points or something else?

Do the IQ test designers take special care to make the skills tested impervious to the beneficial effects of schooling? The questions I remember on an IQ test seemed to pertain to skills like verbal, quantitative, spatial, pattern recognition etc. Why do we expect these talents to remain un-impacted by schooling?

Math SAT scores have been rising over the last decade and a half. Certainly, students are exposed to more advanced math earlier on average today than a generation ago.

If I have understood correctly, the total effect of the reform was an increased average IQ by 0.6 points for all (male) students. However, if you only look at those who were forced to stay longer in school instead of dropping out after seven years, the gain was 3.7 IQ points per extra year. This was not the case for the majority of kids, who would continue with some kind of non-mandatory schooling, such as high school, regardless of the reform.

Arthur Jensen found a similar thing more than a generation ago with the improvement of educations in the South for blacks: in black Southern families, younger brothers were outscoring on IQ tests older brothers who had had to go to inferior Jim Crow schools.

That could've been just the Flynn effect.

Could be -- the Flynn Effect had been discovered in the 1940s, but wasn't widely considered until, say, Lynn started using it to down-adjust Japanese scores on recent tests in, what, the 1980s? But my vague recollection is that Jensen's discovery of the impact of the end of Jim Crow segregated schools in the South in the 1960s was a pretty big effect for a short period of time.

Also, since we don't have a definite explanation of what drives the Flynn Effect, we shouldn't rule out factors as just being the Flynn Effect, when the Flynn Effect could well be driven by factors such as more schooling, better nutrition, more mental stimulation, and so forth. What we are looking for is factors that correlate with the rise in raw IQ scores being above or below the norm associated with the Flynn Effect in general.

There is also Green, Hoffman, Morse, Hayes, & Morgan, 1964 which studied blacks who had been artificially deprived of an education when some Virginia schools shut down and compared them to black students at other schools that remained open. They found a 0.4 standard deviation decline in IQ for every year of schooling missed.

Given this, I don't think the sentence above -- "earlier studies did not rule out the possibility that people with higher IQs might simply be likelier to get more education than others, the researchers noted" -- is correct.

I hope this works in adult life too, because I am still (or again) attending university at age 36.

Definitely seems true in my case. Once I graduated from undergrad, I was not well known for being a thinking person. After a couple of grad degrees, it seems that people consider me a source of wisdom. This could be simply accumulated knowledge and experience, but I have noticed in myself the willingness and ability to understand a whole lot more. Hmm, as I think about this some more, could this be again the value of accumulated knowledge?

Noticing many grey hairs?

Grey hair is sexy. I wish mine would turn grey quicker.

Try Arsenic. Or Mercury.

None of the works listed really comes at a loss to consumers-- they're readily available in affordable (and in the case of music and film, well-restored) editions nearly everywhere.

Compare to post-1920 works that HAVE slipped into the public domain. Stanley Donen's "Charade," for instance. Public domain editions have been put out by every fly-by-night DVD manufacturer, leading to a loss for consumers.

...That's the crux of the public domain-- it isn't so bad, just as long as people ARE making money on it. If Lolita/Night of the Hunter/Lady and the Tramp are still making money, they're still available and well-preserved for consumers.

Orphaned works are the only real tragedy. Check out the sad state of Tin Pan Alley sheet music of the '20s, '30s, and beyond. Many university libraries have spent a great deal of time digitizing said works. They're catalogued online, but absolutely unavailable to the public, due to the fact that they're not in the public domain.

But no one is profiting. The sheet music is not available in free digitized form, and it is not available in for-profit form.

...The most sensible solution for the public domain, I think, is for more hurdles for renewing, with higher fees-- if every author of a book (or estate of an author) from the 1920s until now was required to pay a $100 fee to retain the copyright for the book, the vast majority of books would not be renewed.

This would unfree the orphaned works, and relieve the public from this deadweight loss.

Oops. Wrong thread...

Good comment, though!

Off the top of my head, this could have interesting implications for the choice between human capital theory or screening theory; ie does education make you smarter or just serve as a signal of innate ability.

How about checking the ability of the student to do well on tests generally? Years of filling in the bubble & pacing training will dramatically affect scores without reflecting any actual change in the abilities that IQ tests are supposed to measure. I think that this issue neatly dovetails with the temporary nature of the IQ boost associated with extra schooling.

As for the achievements of Jim Crow schools verses integrated, compare the scores after 15 & 30 years before you trumpet that we are doing better with our compassion than we did with our hatred. Jim Crow was an outrage & an embarrassment. And the results we produce now are worse--much worse.

I have read of IQ for a long time. I have a strong idea of what IQ correlates to and my PhD studies were in cognitive neuroscience. And yet, for all this, I still have little idea what IQ is actually measuring. Or even what we want it to measure.

In particular, are we trying to measure some innate, fixed ability? Are we trying to measure abstract problem solving, stripped of context?

An example of the problem: San-bushmen have highly sub-normal IQs. Do we expect that this means deficient problem-solving skills? What about the very concrete problems of finding buried tubers based on complex environmental heuristics?

A second example: After an IQ of 100, there is almost no correlation between IQ and measured creativity. This is a highly unintuitive result. Is the problem with creativity testing? With IQ? With both?

I recall once reading that an IQ of 120 is sufficient brain power to win a Nobel Prize. But that was about 20 years ago so maybe it's been superseded by more recent research.

Not that surprising. Discovery can be pretty arbitrary and historically contingent.

We want IQ tests to predict a wide range of outcomes in the modern world, which they tend to do. In general, the real world seems to have been moving in the direction of what IQ tests tend to measure, which is why, remarkably, these tests are still effective in their second century.

The problem is with creativity. It has costs.

> An example of the problem: San-bushmen have highly sub-normal IQs. Do we expect that this means deficient problem-solving skills? What about the very concrete problems of finding buried tubers based on complex environmental heuristics?

The San prove nothing. Even a pretty dumb person can learn a sophisticated activity if he spends decades at it, is taught in hands-on practical fashion by teachers who have spent their whole life grubbing, has little else to do (like TV), and his dinner depends immediately on his success.

And further, if the San were so smart, why do, when they go to the local cities and villages and camps, they fail to become prominent businessmen or leaders or artisans or scholars? (I understand the outcome is more like they become the local wino.)

"the very concrete problems of finding buried tubers" and similar are routinely solved by wild hogs. Does that indicate that hog intelligence is on par with human?

Obviously not. This is a skill which does not require any abstract thinking whatsoever.

Isn't the SAT technically an IQ test? How does this square with the so called "super tutors" who can boost your score by hundreds of points?

No tutor will promise you anywhere close to a perfect score; why is that? Because while you can stop sabotaging yourself and learn some tricks (just like you can on even matrix IQ tests) and deal with test anxiety etc, you still face hard upper limits on what you can think.

Even if you know exactly how a matrix test is constructed (it's not hard to find out: they're basically a grab-bag of 10 or so transformations and properties, which problems can even be automatically generated by a program), at some point you simply can't apply your knowledge because there are too many little dots and shapes to keep in your head or you simply don't think of the right combination of transformations. And you get questions wrong, and your IQ gets measured anyway.

More generally, test-retest effects are everywhere. Tests still work.

Yes, but a perfect score doesn't mean a perfect IQ. It means your IQ is higher than that test measures. Presumably there are problems that even Mark Zuckerberg would miss. Conversely, someone who scored in the 98th percentile might be able to move up to a perfect score if they apply some of the tricks.

Is Mark Zuckerberg now the benchmark for "high IQ"?

He had a perfect score on the SAT. Others did too, but he's well-known.

Sure, there are usually ceilings for these things. You have to have ceilings simply because you haven't tested the test on a large enough chunk of the population - how do you norm a test intended for the smartest man out of billions? etc. There are tests which claim to do it, but at that point, you're so rarefied you are disconnected from all the usual cross-checks and it's not clear you are still measuring the same mental correlations your IQ test was measuring with more normal people.

It's not technically an IQ test, it's a test that generally functions as something like an IQ test. And even actual IQ tests wouldn't be as accurate as they are if people spent hundreds of hours and thousands of dollars trying to learn to beat the test beforehand.

The correlation between the SAT and IQ tests is so high that ETS has never, ever calculated what that correlation might be. Just ask them and they'll tell you that they've never calculated it.

According to one study, the correlation between SAT scores and g factor scores calculated from the ASVAB battery is .82 (or ".86 corrected for nonlinearity"). This is similar to correlations between any two IQ tests. Link: http://pss.sagepub.com/content/15/6/373.short

That's high!

"San-bushmen have highly sub-normal IQs. Do we expect that this means deficient problem-solving skills? What about the very concrete problems of finding buried tubers based on complex environmental heuristics?"

Do you really think a population with an average IQ measured at 120 wouldn't have been able to either duplicate this achievement or (far more likely) surpassed it and moved far beyond this kind of thing as a case example of their problem solving acumen?

And isn't your assertion about creativity incorrect? I seem to remember a necessary condition for an individual to score high on tests of creativity an IQ floor of 110 -120.

"surpassed it and moved far beyond this kind of thing"

Such as: creating an agricultural civilization which can afford to put most human resources to the most effective work, rather than collecting tubers?

Nah, that never happened, we just all collect tubers.

Hmm... Looks like someone adopted the same nickname as me, so henceforth I am The Original D

This world is too small for two Ds.

I wonder what would happen to the study if the letters “IQ” and all the concepts related to it were removed and GA (for general ability) were substituted?

For a start, I think it would get rid of all the imputed and unjustified causation, the questionable correlations and reduce significantly the number of words used. We might be left with a reasonable large sample study (for which the Scandinavians are well known) a fairly clean independent variable and a timed measure of cognitive performance as the dependent variable. Then the authors could spend some time considering how reliable their study is and what kinds of independent criterion measures they might employ to validate it.

For me “IQ” is a little like Mark Twain’s complaint about the use of “very.” Get rid of it completely or substitute “damn” and you’ll have a much better paper.

The euphemism treadmill ...

Should not the IQ in general have gone up in WW2.; After all there was mass education by the military. From how to fire a rifle to how to fly a plane. None of this was available before ww2.

The military's experience was that draftees in WWII tended to have more on the ball than draftees in WWI. The 1917-18 draft was a pretty big shock to America's educated elite when they found out how backward a lot of Americans were. By WWII, literacy was more dominant among the young, the radio and talkie movies spread fast-thinking urban culture to the hinterlands, hookworm was much more under control, iodine supplementation in salt had brought down rates of cretinism, etc. Of course, IQ testing was also much more sophisticated by the latter part of WWII than in WWI when IQ tests were crude, so it's hard to get an apples to apples comparison, but the general impression of psychometricians was that the WWII generation was cognitively sharper than the WWI generation.

And average years of schooling had gone up substantially from WWI to WWII.

FWIW a friend of mine who has been teaching high school for 20 years things the current generation is more self aware than in the past (in a good way, not narcissistically).

Recent students appear to average higher on the SAT than their parents did, but the College Board keeps scraping the bottom of the barrel harder, so average scores are pretty flat.

In general, kids these days seem better than you might have expected in 1992: crime is down, abortions are down, no giant urban riots like LA in 1992, crack is out of fashion.

The smartest guys in WW2 were in "Catch 22".

Taking out too many student loans definitely isn't a sign of a high IQ.

Taking out too many student loans is a lack of emotional control. Emotional control doesn't correlate very well with IQ.

It is hard for me to imagine how school could increase one's total brain power over to other activities. After all studying girls is studying. That is because we are always thinking about something. After all studying girls is studying. School can surely change what we know but i do not know how it could change how much we know.

In fact IMO school reform should focus, not on trying to make people smarter and better at the same stuff, but it should focus on teaching them more valuable things.

I see somebody already mentioned this, but I was going to ask if they controlled for the Flynn effect.

Alternatively- might working (or having more booze as a result of the product of work) atrophy your IQ?

The Brinch et al paper doesn't cite Hansen, Heckman, and Mullen on the effect of schooling on intelligence, a paper which goes to great lengths to control for selection effects and other forms of endogeneity:


(published version is in J Econometrics 2004). The overall results are quite similar.

It IS cited in the paper, reference 11.

Comments for this post are closed