thinkevolution.net

NOTE: This post was originally published on February 28, 2007, on an old version of this site.

In discussing Stephen Jay Gould and Richard Lewontin’s article “The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme” with a group of undergraduates the other day, one of the students asked if this critique, published in 1979, is still relevant. Gould and Lewontin’s basic thesis is that too many biologists fall prey to the seduction of describing everything they see as an adaptation. They accuse their colleagues of inappropriately dissecting organisms into discrete traits, then conjuring adaptive explanations for the existence of these traits—a process that not only neglects to provide evidence to justify such an adaptive explanation, but that also fails to consider how evolution is a whole-organism process. Well, today I have a definitive answer: it is as relevant as ever.

Today Marty Moss-Coane interviewed guest Sharon Moalem on her show Radio Times. Dr. Moalem, author of Survival of the Sickest, a book about how many human diseases are in fact adaptive “complicated blessings,” paraded a series of sophomoric explanations for how susceptabilities to many human diseases have in fact been selected for over our evolutionary history.

For example, Moalem argues that the relatively high frequency of genes that underlie the disease hemochromatosis in populations of European descendants may be the result of selection for these genes during the plagues of Europe. Hemochromatosis is a disorder in which iron accumulates in body tissue, and macrophages—which are a part of the immune system—suffer an iron deficiency. Because some pathogens may depend on iron to escape attack by the immune system, and because descendants from countries with high plague casualties show high frequencies of hemochromatosis, Moalem makes the argument that Europeans with iron-deficient macrophages were more likely to survive the plague… and thus we see a disproportionate representation of hemochromatosis in European descendants. But wait a minute! Correlation must not be confused with causation. If the genes that cause hemochromatosis are so detrimental, why were they not purged from the population before the plagues began? Maybe new mutations are popping up, or maybe selection is too weak to purge these alleles from the population. Or perhaps there has been positive selection for some of these genes, but we have no direct evidence that it was to avoid dying from the plague. It may be the case that the genes that underlie hemochromatosis also perform a beneficial function elsewhere in the body.

Furthermore, Moalem uses spurious scientific data to spin his adaptation stories. For example, he reports that fewer boy babies than girl babies were born in New York city after 9/11. Apparently there are some data suggesting that gestation of a male fetus incurs a higher cost to the mother, and that under stress, a mother will be more likely to miscarry if she’s carrying a high-cost son. Nevermind that this phenomenon deserves skeptical examination—the statistics that claim a significant gender skew in New York births can be easily dismissed, thus debunking Moalem’s claim that the 9/11 attacks stressed out pregnant mothers and caused a slight decrease in the boy/girl newborn ratio. The failure of this study is quite straightforward. The authors compute birth ratios for 81 different time intervals in New York city: 71 intervals leading up to 9/11, and 10 inervals after 9/11. The authors claim that the lowest boy/girl ratio occurred in one of the 10 post-9/11 intervals, and that it deviated “significantly” from the average ratio over all 81 intervals. Their data do show that the lowest ratio occurred post-9/11, but it is not significant. Fluctuations in birth ratios, just as in any natural phenomena, will occur due to randomness. (And in fact, the sex ratio went up more often than it when down in the 10 intervals after 9/11!) Statistics allow us to infer whether a fluctuation differs enough from the average to suggest something unusal, and when looking at multiple events, a statistician must correct for multiple tests. Quite simply, these authors performed the correct test but utterly failed to adjust the type I error value to accommodate the type of data they were examining. The truth is that there is absolutely no evidence that birth ratios changed at all in response to the 9/11 attacks. To see what p-values should have been reported, see this criticism.

Radio Times is broadcast every weekday from the Philadelphia NPR station WHYY, 90.9 FM. You can listen to this particular show from the Radio Times archive.

This entry was posted on Thursday, April 19th, 2007 at 1:00 am and is filed under News and reports, The adaptationist programme. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.