history and social science

One of the commenters on that Ta-Nehisi Coates piece from the other day lectured Coates about how he should "learn some statistics." It was, as you can imagine, not a helpful comment, and whether the commenter would have said the same to a white writer or not, it echoed traditional condescension towards black writers and black intellectuals in a way that was really unfortunate. This is a sticking point in arguments about race all the time. People complain that they are judged on the symmetry between their ideas and historical racism rather than their intent. But we live in history, not in a land of pure ideas, and the insistence on sensitivity and care in our racial dialogue is a necessary product of a legacy of oppression.

I've been thinking about statistical analysis of social phenomenon and history lately. Coates, as much as any other blogger, bases his work on history. History and the social sciences, particularly those pursued quantitatively like experimental psychology, have always been tense neighbors. Why? Because the purpose of history is to establish context. History's enduring message is that no condition can be understand without understanding the conditions that created it. History is an effort to contextualize. Statistical examinations of human life are just the opposite: they rely on the decontextualization of sampling and stratification, where we divide humanity into various groups, based on demographics and features, on the theory that this is the key to getting to the causal relationships that we want to find.

The people who conduct such research (and I read and interact with many) are often themselves quite candid about the limitations of stratification and the existence of uncontrolled variables. The important question, always, is the details, and reasonable people can reasonably disagree. When we point out that poverty has a large impact on a variety of life outcomes, particularly metrics of education or intelligence, you often hear the reply that poverty has been accounted for in the research. What that means in practice, typically, is stratifying the sample for income level, and then "comparing like with like." The question is whether these stratifying mechanisms are actually accounting for the influence of confounding variables adequately. You might point out, for example, that poverty is a holistic phenomenon that extends far beyond the simple question of income strata. When people talk about the role of parentage, people will say, "we've controlled for parent's educational level." But surely, parenting contains a vastly larger amount of variation than can be explained with that control.

Adjudicating those disputes has to be conducted by people with a deeper grasp of and greater expertise in the philosophy of social science than me. I do want to say: that there are people who dispute the degree to which key variables can be adequately controlled for in social scientific research, and people who point out that there are some potentially key variables that researchers have almost no ability to investigate, such as childhood lead exposure in adult subjects. These people are perfectly mainstream scholars. People asserting the case for racial inferiority through these mechanisms often express them with considerable certitude over the experimental mechanism, even when the most anodyne parts of the data analysis are subject to legitimate debate.

But even beyond the specific and limited questions involved with stratification of variables in the social sciences, there is a broader question of locating observed results in the context of history. Even if we are perfectly confident that individual variables have been isolated, when it comes to end results, it's our responsibility to place their observed values in a broader social context that helps to explain discrepancies. That is more true than ever when it comes to race. In a very real sense, the effort to combat racism has been the effort to insist on history: the history of slavery, the history of Spanish conquest of indigenous American people, the history of the decades in which Southern black people lived under slavery in all but name, the history of systems of racial discrimination, the history of immigration, the history of our war with Mexico.... Isolating variables can be a key part of socially just research. It is from the attempt to isolate variables that we can say with great confidence that poverty has an impact on education, for example. But we must return always to the reality of a history that has witnessed systematic and relentless oppression of nonwhite people. What aggravates me so much about many who raise the race and IQ question (often while refusing to speak plainly about their feelings on it at all) is the shortsightedness of their considerations, the denial of history. It happens so frequently that, yes, I think it is fair to ask about their motives.

Will Wilkinson put it well:

I don't think the subject or conclusion of Mr Richwine's dissertation is out of the bounds of reasonable discourse. Yet I think a suspicion of racism is perfectly reasonable. Grad students can choose from an infinite array of subjects. Why choose this one? Who are especially keen to discover a rational basis for public policy that discriminates along racial lines? Racists, of course. Anyone who chooses this subject, and comes down on the side vindicating racist assumptions, volunteers to bring suspicion upon himself, to expose his work to an extraordinary level of scrutiny.

Precisely so. It is not racist to ask these questions. James Flynn, one of the most important researchers of the question of human intelligence in history, has used this sort of research precisely to agitate for social justice and left-wing politics. But it is perfectly natural, in a country with such a long legacy of racism, to expect those arguing that race leads to inferior outcomes in as existential a quality as intelligence to be held to very stringent consideration. That is particularly true when, as in the case of Jason Richwine, that argument is levied in the service of further discrimination, a reactionary call against immigration and deepening racial diversity in the United States.

I think that the idea that is being combated here is not merely the question of whether non-white people are prone to low intelligence, but an idea that is rarely voiced but frequently floats around in the ether the way taboo ideas do. The idea is the notion that, whatever historical oppression for nonwhite people we can accurately identify, they've "had long enough," that all these decades after the Civil Rights Acts— after people decided that we had "solved" racism— nonwhite people should have figured it out. But this attitude depends on an ahistorical account: black and Hispanic Americans have never reached outcome parity with white Americans. We can't claim that they have failed to maintain the conditions of American middle class lifestyles when they have never enjoyed them. I could list a dozen metrics for quality of life and economic security on which black and Hispanic Americans have never enjoyed equality. To act as though it is their fault for not meeting those standards when they are working from a legacy of entrenched and deliberate exclusion is precisely to deny history.

What we ask for, when we dispute the conclusions pursued by the race science crowd with such zeal, is that they be forced to live in history, that they not try to argue as though human life is a series of disconnected variables but rather that it is an interconnected fabric of phenomena that cannot be separated without rending the garment. If we force those discussing race and IQ to live in the history of racial oppression, we are merely asking them to occupy the same position as those they are describing. People of color have never had any choice but to live in history.