Monday, June 02, 2008

Figuring out the Flynn Effect

When I was in Atlanta a few weeks ago, blogging for Scientific American about the Intel International Science and Engineering Fair, I interviewed a student named Julian Lyles Bass-Krueger who'd done a fascinating project on how Tetris (the computer game where you turn shapes to fit into variably-sized holes) makes you "smarter."

How so? He had half of a group of subjects play Tetris for 15 minutes (the other half did not play the game). Then he gave both groups a spatial reasoning assessment, of the sort you might see on an IQ test. The result? People who had played Tetris for 15 minutes scored about 55% higher on the spatial reasoning test.

While intelligence tests are usually based on multiple areas (not just spatial reasoning) this is obviously a huge difference. Bass-Krueger posited that this might be one factor in the Flynn Effect, the phenomenon that IQ scores seem to have been rising a few points per decade for much of the 20th century. After all, games like Tetris didn't exist 70 years ago. If many modern children are exposed to such games, their scores would naturally be higher.

It's an interesting theory, and there's probably some truth to it. The Flynn Effect is one of those concepts that's very difficult to tease out, and is often misunderstood. I have heard people reference it as a critique of the concept of intelligence. After all, if IQ scores are rising, then maybe IQ tests are easily gamed. People who've had the opportunity to do brain teasers as kids will naturally score better. Based on this idea, the use of IQ as a way to identify kids for gifted education also seems flawed. After all, who's to say if a child is gifted and needs educational intervention, or has just played a lot of Tetris?

But I think there's a larger factor in play that's probably more important. While it's a popular email forward to give an example of an 8th grade test from pioneer days (thus showing how dumb we modern folks are by contrast), one has to remember that when intelligence testing first took off in the earlier part of the 20th century, childhood mortality rates were still fairly bleak in industrial countries, including ours.

In 1936, according to historical records from the US government, infant mortality nationwide was 57.1 deaths (among children under age 1) for every 1000 live births. For non-white children, the rate in many states was twice as high (for instance, among non-white babies in Delaware in 1936, there were 110 deaths before age 1 for every 1000 live births). These days the number is 6.78 deaths per 1000 live births. Among black babies, the rate is twice as high: 13.6/1000. But even this unfortunately high number is many times lower than for children of all races 70 years ago.

Two things that we know affect brain development are pre-natal and infant health and nutrition -- factors that are highly correlated with infant mortality. Just as most of us are taller than our grandparents, it makes sense that our brains were better nourished when we were babies. And as it turns out, most of the Flynn Effect gains are concentrated on the lower part of the intelligence spectrum. That suggests that intelligence scores have gone up not because striving parents are teaching their kids brain games to get them into the top track in school, but because pregnant women and infants are not generally malnourished in developed countries anymore. They also receive better medical care (such as antibiotics that fight the infections that might otherwise lead to long term learning problems).

There is some evidence that the Flynn Effect has mostly stopped of late in richer countries, which would also be consistent with this hypothesis (there's less room for improvement these days).

But that brings us back to the Tetris question. If 15 minutes of Tetris can improve spatial reasoning scores by 55%, doesn't this put IQ scores under at least some suspicion? Yes...and no. All tests are blunt instruments. The fact that children's IQ scores tend to rise by 5-6 points the second time they take a test indicates that there is a learned familiarity boost. But this is pretty much a ceiling on the effect. Just as coaching can boost an SAT score 100 points, but not 600 points, playing brain games might boost an IQ score from 110 to 115, but not from 110 to 150.

In a previous post, I suggested that IQ tests be used as an initial screen for gifted education, but not the sole screen. Gifted education is not a reward for kids who work hard in school. Nor should it necessarily be based on a stark 130-and-over line in the sand. It should be an educational intervention for kids whose needs cannot be met in a normal classroom. My suspicion is that this is a subset of the 130+ crew and -- in good schools that allow grade skipping -- probably a smaller subset.

The Flynn Effect is a fascinating phenomenon. It may be a result of a more stimulating environment, but it's probably also the result of some more basic public health and welfare improvements. I don't believe, though, that it fundamentally changes the need to recognize the unique learning styles of extraordinarily bright children.


LaRita said...

Your last sentence says it all.

I would much rather have schools focus on how to best teach the gifted students than try to up their scores a few points by playing games.

I found your blog a while back while googling for alumni of Indiana Academy. My oldest son will be attending there this fall and I wanted to read things written by people who had been there.

When I first began looking into their school and goals, I think I cried a little... their approach toward gifted education includes many of the things that we have been BEGGING our son's teachers to do ever since he was in first grade. It was so reassuring to see that we were right all along in what we believed would work best for him.

I am looking toward this next school year with both excitement and fear... excitement at seeing what he will become in an environment so well suited to him, and fear about how much different it will be at home without him there.

I would love if you would write more about your experiences there... how being in a program that focused on gifted students changed where you were and where you ended up. Hearing more about your journey will help me be more secure that his leaving and our missing a good portion of his last two years of high school will be worth it in the end. And it might help other parents who are looking into similar programs in Indiana or other states make their decision as well.

Anonymous said...

What if it only ups scores locally?

Could it be possible for an activity to up IQ scores locally but not over a long period of time?

And then clearly if it has effects over a long period of time then we should be having all children play tetris, especially children with lower IQ levels so that their IQ scores can improve.

benjamindavidsteele said...

We actually don't know that the Flynn Effect will stop. We're still dealing with environmental factors.

During the mid decades of last century, there was an increase in lead pollution and hence childhood toxicity. This required around a two decade lag time for the effect to be seen in changes of adult average IQ and violent crime rates.

It was only during the last decades of last century (1970s-1990s), that serious lead regulations were implemented. With the same two decade time lag, we won't see the effect of lead toxicity fully ending until the coming years.

People forget about how environmental impact on children has such a long lag time. With possible epigenetic effects as well, there might even be a multi-generational lag time to see the full results of improvements. Plus, lead toxins remain in buildings and in the soil for generations, until cleaned up.

There will be a shifting up and down of average IQ, as there are shifts in environmental conditions. The temporary blips aren't that important. It's the larger trends across generations and centuries that will matter. The larger trend of the Flynn Effect has yet to be undermined, even if it briefly slowed down.

Time will tell.

By the way, I'd point out that some schools still have lead paint and the school playgrounds are more likely to have lead toxins in the soil. It is costly to clean up and many schools are underfunded, especially those in poor areas. Of course, those poor kids are more likely to also live in old houses with lead paint and to live in neighborhoods with accumulated environmental toxins, lead and otherwise.

If we really wanted to improve intelligence and cognitive ability for all kids, we'd require lead toxicity testing be done on all children. And we'd require all lead pollution to be cleaned up, at least in schools and residential areas.

Through simple environmental improvements, vast positive results could follow.