Of Construct Validity and Jason Richwine's Piss Poor IQ Race Science Hustle

Former Heritage Foundation policy wonk Jason Richwine is having an exceptionally bad day when both Ta-Nehisi Coates and Daniel Drezner are calling out his piss poor social science.

Richwine's claim that "Hispanics" are less intelligent than "whites" and as such, any immigration reform the encourages their coming to the country will be a net drain on the U.S. economy is very old school. It reeks of discredited claims about "racial stock" and "national character."

Drezner works through Richwine's problematic methodology, using the academic version of the cobra clutch to ether his race science sophistry:

I've perused parts of Richwine's dissertation, and … well … hoo boy. Key terms are poorly defined, auxiliary assumptions abound, and the literature I'm familiar with that is cited as authoritative is, well, not good. It's therefore unsurprising that, until last week, Richwine's dissertation disappeared into the ether the moment after it was approved. According to Google Scholar, no one cited it in the four years since it appeared. Furthermore, Richwine apparently didn't convert any part of it into any kind of refereed or non-refereed publication. Based on the comments that Weigel and others have received from Richwine's dissertation committee, one wonders just how much supervising was going on.

Efforts by the polite white supremacists in the human biodiversity crowd, and the less delicate advocates of race science such as Pat Buchanan, concoct a link between intelligence and race are intellectually hamstrung in a number of ways.

Most importantly, race is social construction and a myth. Human beings have not existed long enough to be divided into separate and distinct racial "species." Moreover, how do we categorize a cohort such as Hispanics which are an ethnic/cultural/census category that encompasses many different "racial" groups?

IQ purports to measure "intelligence." However--extreme outliers being noted for how they complicate this narrative--what is performance on an IQ test actually revealing about a given test taker?

There is a long tradition of the in-group devising seemingly objective tests for "intelligence" which are really just a way of advancing a particular racial project. As black psychologists and others suggested in the 1970s, IQ tests are actually just a way of measuring cultural assimilation to Whiteness.

Ultimately, Richwine's work could be taught in an Introduction to Data Analysis course as an example of a failure in construct validity. Efforts to link race and intelligence also introduce questions of confounding and proxy variables. In all, race science sophistry is attempting to empirically model something--what I am not sure--but is really just furthering "the mismeasurement" of human beings in the service of legitimating white supremacy and racial domination.

I love Charles Mills' explanation of the fallacy of the race concept. Consequently, I return to it often.

As Mills implies, researchers cannot accurately measure the relative intelligence of different racial groups when the very categories are not fixed, change over time and context, and in many ways are arbitrary distinctions made by Power in an effort to legitimate the exploitation and subordination of some groups, and the natural authority of others.

Why is race science still with us? And is it compelling because the label of "science" gives a veneer of neutrality to a bigoted and regressive racial and political project?