The sociologist’s fallacy

The sociologist’s fallacy is the tendency to interpret a correlation between a social variable and a phenotype as causal, without considering that genetics could mediate the relationship. In Making sense of heritability, Neven Sesardic gives several good examples of this. A report from the Nuffielf Council on Bioethics (2002), for example, engages in the fallacy when it notes the following:

While it is often claimed that on average, black individuals score slightly lower on IQ tests than white individuals, who in turn score lower score lower than people from East Asia, there are also studies which show that, if black individuals and white individuals are closely matched on socioeconomic status, the differences in IQ are substantially reduced…The authors [of a study cited in the text] conclude that “socioeconomic differences are largely responsible for the usually reported differences in intellectual performance.

Apparently, it never crossed the authors’ mind that genetics could mediate the relation between SES and IQ. As it is, we might call this the sociologist’s first fallacy, since this fallacy has a corollary, which we can call the sociologist’s second fallacy. The sociologist’s second fallacy is the tendency to attribute any undesired “gap” found between a designated minority group and Whites after controlling for all know social variables to the mysterious effects of racism, without considering alternative possibilities. Here is a nice example of that:

Racism and Birth Weight

A large number of epidemiological studies have attempted to account for this disparity in terms of maternal age, education, lifestyle, and or socio-economical position. However, the results make it clear that these variables account for only a small proportion of the difference…Researchers involved with these studies argue that their findings “suggest that growing up as a woman of color in the US is somehow toxic to pregnancy, and imply a social etiology for racial/ethnic disparities in prematurity that is not solely explained by economics or education.

Apparently, it never crossed the author’s mind that genetics could contribute to the gap. After all, West SubSaharan Africans have a higher frequency than Europeans (80% to 30%) of the low-birth-weight-risk allele C825T on the GNB3 gene — not that those social constructs called ‘whites’ and ‘blacks’ are in any way related to Europeans and West Africans. When it comes to patterns of low-birth-weight one can see the following illuminating trend:

From: Rowe (2005)

This trend is just what hereditarians would predict. Magnusa, et al. (2001) estimates a .25 heritability for birth-weight with a .26 mother-child correlation and a .12 father-child correlation. Unsurprisingly, in the above sample the mixed children’s birth-weight falls between the population means and mixed children with White mothers have a higher mean birth-weight than mixed children with White fathers.

References

Magnusa, et al., 2001. Paternal contribution to birth weight

Rowe, 2005. Under the Skin : On the Impartial Treatment of Genetic and Environmental Hypotheses of Racial Differences