Independent Research: The Eugenics of Statistics as a Social Ideology
This entry, by Akio*, frames our contemporary relationship to statistics, given the historical relationship of statistical analyses to eugenic figures, methods, and conclusions.
*Akio Tamura-Ho ‘23 (they/them) is a senior at Yale University studying the History of Science, Medicine, and Public Health.
The tools and methods of statistical analysis underpin a significant proportion of academic research, and the vast majority of published STEM research. “Key stats” are deployed in news articles as facts or indisputable pieces of evidence. Statistics drive the professional and research field of data science, which has been alternately hailed as the “sexiest” and most promising career path of the 21st Century. Although many have argued against the valorization of statistics as objective truth, these critiques seldom extend beyond a self-satisfied xkcd comic that smugly points out that correlation does not equal causation. Millions of students have likely been presented with this image or a similar kind of lesson, through which teachers of statistics can assure themselves that any issues surrounding their field can be attributed to ignorant viewers or irresponsible practitioners. A responsible practice of statistics can avoid these problems, and return to its comfortable status as the method through which we strive toward the empirical truth.
However, historians of science and scholars working in critical science and technology studies have identified the ways in which foundational statistical methodologies, epistemologies, and ways of thinking have also naturalized eugenic attitudes into the public consciousness. For instance, Francis Galton, an influential statistician, developed important statistical concepts such as correlation and regression. He was also an avowed eugenicist, and deployed statistics as a social and scientific ideology in support of eugenic conclusions. In 1907, Galton delivered a lecture entitled: “Probability, the Foundation of Eugenics”, and argued that race-based differences in intelligence could be evidenced through the statistical concept of the normal distribution or bell curve. In 1993, The Bell Curve, written by Charles Murray and a Harvard psychologist, was published. The book continued Galton’s work by applying the concept of the bell curve to race science in modern America. And in 2017, Sam Harris (a popular podcaster listened to by millions) invited Charles Murray to appear as a guest on his show. Entanglements between statistics and eugenics haunt us today.
How might it transform our relationship to statistics if we are to think of the field’s epistemology and methods not only as historical tools of eugenics, but as potentially eugenic themselves? That is, how might embedded structures in the field of statistics guide us to eugenic conclusions? Such an idea suggests that not only do we need to reimagine how we use and deploy statistics, we also need to be prepared to rework and reimagine statistics as a field and heuristic. During the summer, I received funding from Yale Education Studies to conduct archival research into the relationship between statistics and eugenics (guided by Professor HoSang), examining material from the American Eugenics Society when it operated out of Yale. My key concern was: how did eugenicists at high levels institutionalize their ideologies in academia through statistical methods? Throughout the summer, I read sources from the 1920s-1950s, including papers from the Second International Congress of Eugenics (1921), papers relating to Ellsworth Huntington (the former President of the American Eugenics Society and Yale geography professor), and past issues of historical journals like Eugenics Quarterly and Social Forces. In addition to this research, this grant also enabled me to attend meetings with the Anti-Eugenics Collective, and learn from their collective experience and labors.
Reading archival materials for statistical thinking as a social ideology, I found that statistics proliferated in materials published by the AES—both as a kind of infallible positioning of self-same objectivity and also as a language that was employed across a variety of disciplines in order to render eugenic assertions as neutral. The value of statistics as a factual and predictive tool emerges in an article by Ellsworth Huntington from 1933, in which he writes, “The secret of future events is often hidden in trends which are now scarcely visible.” Analyzing data on the number of children per family, Huntington presents a variety of line graphs and tables before using them to draw speculative and predictive conclusions about birth rates among “distinguished men” compared to “the rank and file”. Data visualizations were a central tool of the American Eugenics Society that appeared again and again as I looked through archival material across several decades—stark imagery that communicated ideological messages while presenting their content as “objective”. At one point, Huntington even urges the importance of using “the very best graphic illustrations”.
“There can be little question that the increase in families shown by these statistics is real,” Huntington writes—the statistics are self-evident, they are conflated with Huntington’s eugenic analysis, wrapping themselves in a veneer of neutral representation. At a 1936 conference, Huntington argued that “certain statistical facts indicate that this prolonged perpetuation of unusual achievement from colonial times down to the present is biological rather than environmental,” while at the same conference the Statistician-in-Charge of the Milbank Memorial Fund, Dorothy G. Wiehl, drew out observations of “high correlation” and called for “the painstaking collection of data which are at present lacking for any precise and objective answer to the foregoing questions.” The language of statistics performed categorical and boundary work as well, delineating groups of people who are rendered “normal” or “abnormal”. A Statistical Directory of State Institutions for the Defective, Dependent, and Delinquent Classes shows the collaborative mechanisms of eugenics, statistics and carcerality as people labeled as “feeble-minded” or “criminalistic” are sorted into the categories such as “white or colored” in extensive tables. Inherent to the logic of this discrete sorting—the categorical variable-ization of human beings—is the promise of correlative and eventually predictive analysis. The language of statistics is indelibly tied to eugenic projects that seek to stratify and cull the general population according to white supremacist necropolitics. And although there is not space in this post to elaborate on all of the examples of statistical thinking that I found in the archive, exposing and untangling the ways in which statistics leads us toward a certain kind of categorical, correlative, and insulated way of thinking invites us to continue questioning the built-in logics of data and quantitative analysis. Otherwise, a statistical practice that doesn’t hold these tensions and troubles at the forefront risks perpetuating eugenic thinking.