26 May 2009

How to react when research says it supports stereotypes

The Intersection points out that Phillippe Rushton is at it again. Back in my grad schools days, there was tremendous debate about Rushton, an academic living in Canada, who published work that purported to confirm racial stereotypes.

Now, he has a new article that purports to confirm sexist stereotypes. (Additional: Whoops, didn’t notice the date and realize this is a couple of years ago.)

I support academic freedom. I think Rushton should be free to do his research. But that doesn’t mean he should be free from criticism of his research. In that spirit, I would like to include a couple of quotes from the most memorable book review I’ve ever read in a peer-reviewed journal or elsewhere, on Rushton’s earlier book, Race, Evolution, and Behavior. David Barash wrote:

(Rushton’s) argument is much weaker than it first appears, based on a great array of material resembling delicate needlework: superficially impressive, it boasts a discernible pattern, but is revealed on closer inspection to be composed entirely of holes. Thus, we get tabular material that cannot be evaluated and seems nothing less than absurd, such as purportedly serious heritability estimates for ‘pajama parties’, ‘striptease shows’, ‘learning Latin’, and ‘Bible truth’, these and 46 others somehow mysteriously ‘corrected for unreliability’ (Table 4.3). Defending such drivel, Rushton argues at length for what he calls the ‘principle of aggregation’, which, in his hands, means the pious hope that by combining numerous little turds of variously tainted data, one can obtain a valuable result; but in fact, the outcome is merely a larger than average pile of shit.

Among the sad substitutes for ‘data’ (Table 7.4) are the purported opinions of ‘orientals’ as to the ranking of blacks, whites, and orientals with regard to size of genitals (blacks first, whites second, orientals third) and the views of ‘whites’ as to brain size (whites first, orientals second, blacks third). These results may speak volumes as to the existence of racial stereotypes, but not a syllable as to their accuracy. Statistical tests are presented only rarely, when supportive; otherwise, differences of a few cc out of total cranial volumes on the order of 1400 cc are unleashed, as though noteworthy. This also ignores whether cranial volume or indeed, brain size, within the normal human range, correlates in any way with intelligence.

I don’t know which is worse, Rushton’s scientific failings or his blatant racism.

The review is well worth reading in full. It is the very definition of evisceration. And it’s all the more remarkable for being published in the normally staid pages of a peer-reviewed technical journal.

I suppose it’s possible that Rushton’s science has improved in the last decade. But that he didn’t gather any of his own data (again) does not make me hopeful. But I think the lesson offered by Barash is worth repeating. The first question that should always be asked with this kind of study is:

“Is the science sound?”

The Daily Mail article doesn’t ask this question at all. And as science journalism is cut back, there will probably a lot more of this sort of reporting.

Reference

Barash, D.P. 1995. [Review of Race, Evolution, and Behavior]. Animal Behaviour 49: 1131-1133. http://dx.doi.org/10.1006/anbe.1995.0143

No comments: