Arnold Kling  

Flynn Effect

Working Poor... Health Care and Moral Values...

Following Marginal Revolution's recommendation, I bought Ian J. Deary's Intelligence, which is a summary of research on IQ testing.

For me, the most interesting chapter was on the Flynn effect (see also this post), which is that IQ scores have been increasing by about three points per decade. This fact has generally been masked by the fact that researchers periodically "re-norm" test scores to set the mean at 100.

As an economist, my first instinct is to treat the Flynn effect as a result of economic growth. That is, like height or longevity, IQ is something that increases incrementally as the standard of living rises.

Some folks leap on the Flynn effect as a demonstration that IQ is not purely genetic. I think that, as with height or longevity, most individual differences are genetic. However, environmental factors are significant when one looks at averages of groups under very different economic circumsances. Or, to put it another way, there may be very little that you can do at the margin to improve your child's IQ, just as there is very little at the margin that you can do to increase your child's height, even though environmental factors affect both.

Deary raises the possibility that the Flynn effect reflects a secular increase in test-taking ability, with "true" intelligence unchanged. He writes (p. 109)

Compared with a mean of 100 in 1992, the mean for the population in 1942 would be almost at a level that indicated mental handicap for the average person. (It is this consideration that makes me very sceptical about the veracity of these supposed 'IQ gains.')

Deary also points out that the Flynn effect is contra-indicated by the decline in SAT scores. However, since SAT tests are not taken by a random sample of students, perhaps it is the SAT decline that is illusory.

Assume that Deary (the expert) is wrong in his skepticism, and that in fact intelligence has, like longevity or height, been increasing, because of economic growth. In addition, one would presume that average IQ in turn feeds back into economic growth. In fact, this book, which finds a correlation of .7 between average IQ and GDP across countries, treats all of this correlation as causality from IQ to GDP.

Assuming that causality runs in both directions, this implies a sort of virtuous circle. A positive feedback process might be the answer to an important puzzle in economics: the failure of economies to converge. The economic growth models that emphasize capital accumulation all have the characteristic that poor countries should grow faster than rich countries. A model that takes into account IQ might explain how rich countries would grow faster than poor countries.

Discussions of IQ often are tinged with issues of racism. In this regard, it is useful to read Thomas Sowell's critique of The Bell Curve, in which Sowell makes the point that the Flynn effect contradicts a racially deterministic view of intelligence.

For Discussion. What factors might cause the Flynn effect to slow down?

TRACKBACKS (3 to date)
TrackBack URL:
The author at Stumbling and Mumbling in a related article titled Intelligence and incentives writes:
    Arnold Kling has some more thoughts on the Flynn effect here. The question is: how can we reconcile this with this (worth the free registration)? Here’s one possibility. There is an economic explanation for the Flynn effect, but not (just) [Tracked on November 27, 2004 8:10 AM]
COMMENTS (12 to date)
Jim Glass writes:

I just read through a history of World War I and it sure seemed like everyone in the west today is about 30 IQ points smarter than then (which is a lot, two standard deviations).

I mean, geeze, I sure hope so.

Robert writes:

If the average 100 years ago was "100" then 3%/year would make it 134 today: genius level. Going the way, in 1904 it was 77. Given the remarkable achievements of the 1st decade of the 20th century, I believe it is mostly a testing artifact. I would also think that it would imply a narrower bell curve with the extremes seeing a less significant change.

Just like the Japanese growth spurt after the introduction of more meat, this phenomena levels off when the limitations to genetic potential are removed.

Lawrance George Lux writes:

The Flynn effect stands as a feat of acculturation, rather than Intelligence increase. The effect exists because of the rapid technological advance. The older Generation was confronted then, and today, with what they needed to know to survive and participate. The Younger Generation reflects the technological advances. The Flynn effect will slow when the rate of technological advance slows. lgl

razib writes:

the flynn effect might also be partly genetic:

Ronnie Horesh writes:

Not completely off-topic are these questions asked of 11-years old candidates to an English Grammar school in 1898. These are the arithmetic questions; the complete set is at :

1. Multiply 642035 by 24506.

2. Add together £132 4s. 1d., £243 7s. 2d., £303 16s 2d., and £1.030 5s. 3d.; and divide the sum by 17. (Two answers to be given.)

3. Write out Length Measure, and reduce 217204 inches to miles, &c.

4. Find the G.C.M. of 13621 and 159848.

5. Find, by Practice, the cost of 537 things at £5 3s. 71/2d. each.

6. Subtract 37/16 from 51/4; multiply 63/4 by 5/36; divide 43/8 by 11/6; and find the value of 21/4 of 12/3 of 13/5.

7. Five horses and 28 sheep cost £126 14s., and 16 sheep cost £22 8s.; find the total cost of 2 horses and 10 sheep.

8. Subtract 3.25741 from 3.3; multiply 28.436 by 8.245; and divide .86655 by 26.5.

9. Simplify 183/4 – 22/3 ÷ 11/5 – 31/2 x 4/7.

10. Find the square root of 5.185,440,100.

11. Find the cost of papering the walls of a room 16ft long, 13ft 6in. wide, and 9ft high, with paper 11/2ft wide at 2s. 3d. a piece of 12yds in length.

12. A and B rent a number of fields between them for a year, the rent and other expenses amounting to £108 17s. 6d. A puts in 2 horses, 5 oxen and 10 sheep; and B puts in 4 horses, 1 ox, and 27 sheep. If a horse eats as much as 3 sheep and an ox as much as 2 sheep, how much should A and B each pay?

These papers were kindly sent in by Humphrey Stanbury, whose father took the exam, and passed.

Hi, Ronnie!

You said:

These papers were kindly sent in by Humphrey Stanbury, whose father took the exam, and passed.

I'm a little confused about the dates. You also said:

Not completely off-topic are these questions asked of 11-years old candidates to an English Grammar school in 1898.

Are you saying that Humphrey Stanbury's father took this exam in 1898, at 11 years old? If Humphrey Stanbury's father was age 11 when he took the exam, his father was born in 1887?

Something isn't adding up. When was Humphrey Stanbury born?

Let's say the father was age 25 when Humphrey Stanbury was born. That would make Humphrey Stanbury born in 1912. Thus he'd be 92 now.

It's not impossible. Econlib does have lots of readers ages 70-80--and maybe a few in their 90s, all welcome here! Or maybe my arithmetic or guesses are wrong. But something's not likely. Year 2004-Year 1898=106 years, which sounds like more than two generations.

He sent these papers to you when and how?


skh writes:

Robert, 3% a decade...not per year. Still, quite an increase.

Brad Hutchings writes:

Let's say lgl is right and Flynn's effect is consistent with rate of technological progress (or change). I'd refine that a little bit and hypothesize that Flynn's effect depends on increasing the number and variety of people you are likely to come in contact with. So decreases in costs of travel and communication and increases in interesting places to go and opportunities to participate in conversations would drive Flynn's effect. If my hypothesis were true, then Flynn's effect looks a lot like Moore's Law. Almost 30 years into Moore's Law, there are still giant steps that semiconductor designers and materials people have planned to keep up the pace for more than the foreseeable future. Similarly, with Flynn's effect, I could see the opening up of Africa, China, and Southeast Asia along with a sputtering but consistent trend of economic liberalization and growth in South America giving us all more interesting places to go. And perhaps in a couple generations, most of the world's people will speak English because they wanna participate in a world economy driven by that language. Or maybe technology makes the language differences immaterial. Either way, more people for us to communicate with.

spencer writes:

My question is this somewhat analogous to using a price index over time. If you use the CPI to compare prices today to prices two years ago or 5 years ago it will give you a pretty reasonable estimate of the impact of inflation. But if you try using it to make comparions over a period of time like 50 or a 100 years the validity breaks down because there are so many changes in the basket and quality changes in the basket. You can not compare living standards now with living standards 100 years ago without making some adjustment for health care for example, but that is really an impossible thing to do.

I do not know, but if you are comparing knowledge now to another time how do you make the adjusments. For example, how many of the current readers know how to preserve meat. 50 or a 100 years ago that was common knowledge that almost everyone knew. Is there a difference in intelligence in knowing how to cure meat and use a computer. I doubt it, but how do you compare intelligence that is based on a very different knowledge base?

Tim Worstall writes:

I’m sure that part of the Flynn efect comes from nutrition. We know very well that inadequate nutrition in the womb and childhood causes mental stunting. We also know that such nutrition has greatly improved over the past century. We are surprised by rises in IQ why?
The trackback above from Stumbling and Mumbling also contains a great argument.

Atanu Dey writes:

Somewhat related to what Spenser wrote about about different knowledge base, I would recommend a fascinating book "The Age of Missing Information" by Bill McKibben. In there he remarks that there is a lot of useful information that we are missing due to the overload of often useless information we have today.

Ronnie Horesh writes:
He sent these papers to you when and how?

Hi Lauren, sorry for the delay in communication. I had tried to post the URL of the article from which the exam questions were lifted as a link, but it didn't work. It comes from the current issue of the UK journal, 'the Spectator'. Look at go to the final colum and look at the article 'Dumbing down: the proof'. Can't tell you about Humphrey's age - the reference to his father etc was in the article itself, but your calculations seem right and it would seem he is in his eighties at least.

Comments for this entry have been closed
Return to top