The Flynn effect has always been tinged with mystery. First popularized by the political scientist James Flynn, the effect refers to the widespread increase in IQ scores over time. Some measures of intelligence — such as performance on Raven’s Progressive Matrices in Des Moines and Scotland — have been increasing for at least 100 years. What’s most peculiar is how scores have increased:

1) Scores have increased the most on the problem-solving portion of intelligence tests.
2) Verbal intelligence has remained relatively flat, while non-verbal scores continue to rise.
3) Performance gains have occurred across all age groups.
4) The rise in scores exists primarily on those tests with content that does not appear to be easily learned.

What’s puzzling about this increase in general intelligence is that it appears where we’d least expect it. While one might assume that IQ scores could increase over time in terms of crystallized intelligence — the part of the test that measures particular kinds of knowledge, such as being able to count or vocabulary words — it’s actually increased on measures of fluid intelligence, which is the ability to solve abstract problems. This has led some psychologists, such as Ian Dreary, to conclude that “large differences in scores [between generations] are demonstrated in just those situations where similarity would be expected.” Flynn, meanwhile, marveled at the magical constancy of the effect: “It’s as if some unseen hand is propelling scores upward,” he wrote.

There is, of course, no unseen hand. In recent years, many psychologists have embraced the “multiplicity hypothesis” which argues that the Flynn effect is explained by a long list of factors, such as improvements in early education (especially for girls), removal of lead paint, increased sophistication of tests, better test taking attitudes and adequate nutrition.

However, despite the flurry of interest in the Flynn effect, one lingering question has remained unanswered: Does the effect apply to everyone? More specifically, does it apply to the right tail of the ability distribution, or those 5 percent of individuals who score highest on the IQ test? What makes this mystery particularly noteworthy is that many of the explanations for the Flynn effect seem particularly relevant to the left side of the bell curve, or those with below average scores. This suggests that most of the intelligence gains have come from solving low hanging fruit, fixing those glaring societal inequalities that meant millions of children lacked access to adequate food, education and medical care. Since we’ve made progress on these problems, one might suppose that the Flynn effect would start to fade, at least in developed nations. (All the low hanging fruit is gone, as Tyler Cowen might say.) Sure enough, some studies have concluded that the Flynn effect has begun to disappear in Denmark, Norway and Britain.

A brand new study, “The Flynn Effect Puzzle,” currently in press at Intelligence, and led by Jonathan Wai at Duke University, has found an interesting way to assess the right tail of the distribution. By looking at approximately 1.7 million scores of 7th grade students between 1981 and 2000 on the SAT and ACT, as well as scores of 5th and 6th grade students on the EXPLORE test, the psychologists were able to investigate the extent to which the Flynn effect exists in the right tail of the bell curve. The results were clear:

The effect was found in the top 5% at a rate similar to the general distribution, providing evidence for the first time that the entire curve is likely increasing at a constant rate. The effect was also found for females as well as males, appears to still be continuing, is primarily concentrated on the mathematics subtests of the SAT, ACT, and EXPLORE, and operates similarly for both 5th and 6th as well as 7th graders in the right tail.

In other words, the Flynn effect doesn’t appear to be solely caused by rising scores among the lowest quartile. Rather, it seems to be just as prevalent among the top 5 percent. The smartest are getting smarter.

What might be causing this aspect of the Flynn effect? Obviously, it’s harder to explain using the low hanging fruit hypothesis, since these students (even in 1981) were probably well fed and had access to math education. This leads the scientists to focus on new forms of “environmental stimulation” as a likely explanation:

Rowe and Rodgers (2002) noted that “If the rising mean were driven by the smart getting smarter, then the change might reflect the introduction of some qualitatively novel form of environmental stimulation. If the overall distribution increased in pace, the cause would lie in processes that affected everyone equally.” We find the rising mean of the entire distribution is partly driven by the smart getting smarter. This suggests some form of environmental stimulation may be at work in the right tail.

The question, of course, is what this stimulation might consist of? It obviously has to be extremely widespread, since the IQ gains exist at the population level. One frequently cited factor is the increasing complexity of entertainment, which might enhance abstract problem solving skills. (As Flynn himself noted, “The very fact that children are better and better at IQ test problems logically entails that they have learned at least that kind of problem-solving skill better, and it must have been learned somewhere.”) This suggests that, because people are now forced to make sense of Lost or the Harry Potter series or World of Warcraft, they’re also better able to handle hard logic puzzles. (The effect is probably indirect, with the difficult forms of culture enhancing working memory and the allocation of attention.) As Steven Johnson argued, everything bad is good for us, especially when the bad stuff has lots of minor characters and subplots. HBO is a cognitive workout.

That said, environmental stimulation remains an incomplete explanation. Even for those on the right side of the curve, intelligence gains probably have many distinct causes, from the complexity of The Wire to the social multiplier effect, which is the tendency of smart people to hang out with other smart people. (In this sense, gifted programs in schools might help drive IQ gains among the top five percent. The Internet probably helps, too.) The question, of course, is whether such factors have really changed over time. Has it gotten easier for smart people to interact with each other? Are those on the right side of the IQ distribution now more likely to have children together? Would the Flynn effect be even larger if we did more of [fill in the blank]? These questions have no easy answers, but at least we now know that they need to be answered.

In the meantime, it’s wise to be modest about what we know about population changes in IQ over time; the Flynn effect remains a paradox. As Flynn himself observed, “No one knows which role IQ gains over time will eventually play.”

Image: Dave Mosher/Flickr