Friday, August 15, 2008

Brooks vs. Science

David Brooks wrote this idiotic column earlier this week about American individualism and Chinese collectivism. The take-away paragraph was:
If you show an American an image of a fish tank, the American will usually describe the biggest fish in the tank and what it is doing. If you ask a Chinese person to describe a fish tank, the Chinese will usually describe the context in which the fish swim.

Mark Liberman dismantles this column at LanguageLog.
Question to Language Log: Is it correct that if you show an American an image of a fish tank, the American will usually describe the biggest fish in the tank and what it is doing, while if you ask a Chinese person to describe a fish tank, the Chinese will usually describe the context in which the fish swim?

Answer: In principle, yes. But first of all, it wasn't a representative sample of Americans, it was undergraduates in a psychology course at the University of Michigan; and second, it wasn't Chinese, it was undergraduates in a psychology course at Kyoto University in Japan; and third, it wasn't a fish tank, it was 10 20-second animated vignettes of underwater scenes; and fourth, the Americans didn't mention the "focal fish" more often than the Japanese, they mentioned them less often.

Read the whole thing. The blog has some other takedowns of Brooks's atrocious pop neuroscience. Relatedly, Liberman describes a study showing that the use of irrelevant neuroscientific jargon makes people perceive idiotic explanations as sensible.


Grobstein said...

Dude, I don't respect David Brooks as a scientist, but I don't think much of the Liberman post, either. Based on his and your billing, I was ready to read about how Brooks's assertions about neuropsychology or whatever were flatly ridiculous, and instead I learn that they're roughly supported by experimental literature. Perhaps he's overgeneralizing (this is after all what Brooks does) but there's nothing especially atrocious.

Actually, Brooks's whole scientific point is extremely modest: some societies ("like the United States or Britain") are more individualistic, while other societies ("like China or Japan") are more collectivistic. That's it. That's all the argument of his essay draws from the discussion of Nisbett's experiments.

So Liberman shows that Brooks flubbed some experimental designs, and that the differences he highlights are rough trends rather than absolute divides. So what?

Liberman is trying to generate some laughs at Brooks's expense (again, fine), but he's not making any good points. The "southern Italians" study is a good example of this. Its results tend to reinforce the trend identified in the overall literature, that the individuals in relatively collectivist societies have different cognitive styles. Liberman chooses it because southern Italy is not in Asia, and this creates an opportunity for ridicule.

I think Liberman's post is much more intellectually dishonest than Brooks's column. Brooks misstates the details of research, but harmlessly because he is only speaking in generalities, which he gets correct. Liberman, on the other hand, identifies many minor errors, and then exploits the negative halo effect and our derision to cast doubt on the overall thesis, without any specific justification. It's really nothing more than a potent ad hominem delivered with markers of scientific authority.

Grobstein said...

Liberman gives a fairer summary of his objections here:

From my point of view, the first problem (once we get past the carelessness with details) is that (as often) we're talking about modest differences in group averages, with lots of within-group variance, which are presented as if they were properties of all the individuals involved.

The second problem is that the experimental results seem to be all over the place, in fact, as a function of fairly small differences in experimental design, choice of subjects, etc. This leaves me with many questions about what's really going on.

A point that people often miss in discussions like this is that "modest differences in group averages, with lots of within-group variance" are not identical to "no differences." This error came up a lot, for example, in the Larry Summers kerfuffle. (I know you were against Summers on the grounds that presidents should be seen and not heard.) I think Liberman's post elides the difference.

Emblem Parade said...

Brooks' column is bad in dozens of different ways.

Besides the online "comments" to the column, the Times also published the best letters of objection it received. Between it all, I think the issue can be brought to rest.

I think my favorite "objection" letter was from a guy who was offended by the West-East equivalence Brooks set up for his argument. This guy thought it was wrong to see anything good about China. (Doesn't seem to matter, of course, that this supposed equivalence comes out as hypocritical.)

Zed said...

I don't think the Summers case applies. Summers's (valid) point was that slight differences in mean (and more so in variance) show up as vast asymmetries in the tails. Brooks is not talking about tails, he's saying something like -- Chinese people are wired for collectivism, therefore "development ... through Eastern, or collective, means" will be fundamentally different. This is about the average population, so you're looking at the distribution around 1 sigma and it does make sense to say that small differences in mean will be swamped by variation.

The Brooks column is not especially coherent (what do relationships have to do with fishtanks? and how is an army marching in step a proof of "collectivism"?) but I don't think the Summers defence applies to any of it.

I was also amused by the awfulness of the experiments.

Zed said...

Consider this: "And third, the American kids didn't usually make their choices on the basis of a shared named category like "animal", but rather they did this about 18% of the time, on average, while the Chinese kids did it about 12% of the time." I don't think it's an "overgeneralization" to call 18% "usually"; it's just a lie.

As Liberman notes elsewhere, the difference could well be a case of the Flynn effect.

It's conceivable that smallish differences in the number of individualists feed back into vast cultural differences. (In which case, yes, a Summers defence would apply.) But that's certainly not obvious to me.