Francis Galton, Ronald Fisher, and Karl Pearson are well-known for their contributions to science and the origins of statistics, including linear model methods of regression, correlation, and analysis of variance (ANOVA). Meanwhile, a look at history reveals Galton as a pioneer of the eugenics movement, who believed that “humanity could be improved through selective breeding” (Gillham, 2001). Pearson and Fisher followed by spending their careers, in the early 20th century, as vocal eugenicists. Pearson once claimed, “if you bring the white man into contact with the black … you get superior and inferior races living on the same soil” (Cleather, 2020). For those of us who have calculated Pearson’s r countless times, this was a stunning revelation. The second author of this project had taught university statistics for 15+ years – and even once took a class on the history of statistics – and still had no idea that the biographies of our method founders were so tainted.
This might lead one to question whether Pearson’s apparent objective use of data to make inferences about social groups was biased by the interests and groups that he was a part of (Delzell & Poliak, 2012). Today, despite their histories, these founders’ names are also still attached to statistical terms and tests with little question regarding their appropriateness. Some may believe that such historical figures cannot be judged by modern-day standards, and that their contributions to the field should remain honored (Adamson & Lipoff, 2020). Alternatively, others may believe in a humanitarian obligation to reconsider, when monikers denote those involved in egregious activity.
We often intend the field of statistics to be an objective playing field free of bias in measurement and analysis. Therefore, theoretically the personal views of statistical founders may not be considered a relevant conversation for a methods course. However, implementation of anti-racist pedagogy in classrooms is ongoing, inspiring institutional and social change (see Kishimoto, 2018). Even if race is not our independent variable, it certainly once was for Pearson and Fisher, and continues to be such in analyses conducted every day. There are implications for the conclusions we draw. Can racial differences – or other demographic differences for that matter – ever cause changes in behavior? This is a slippery issue, compounded by the implication that Pearson and Fisher would have surely answered “yes” to this question.
So how should we proceed, and what should we recommend? We decided to ask students.
Methods
204 undergraduate students were recruited from a psychological statistics course, across multiple terms. Toward the end of each term and as part of the curriculum, they were shown a factual video detailing the history of common statistical methods and their founders, as well as the founders' associations with eugenics. Afterward, students were given an open-ended prompt to share their perspectives on whether statistical terms, such as Fisher’s F and Pearson’s r, should be renamed or retained. Qualitative themes were pulled from their responses.
Results
Overall, students favored renaming these statistics, although views were mixed. Of the 204, 96 students generally supported renaming, 54 opposed, and 46 were ambivalent toward the issue. Themes reflected a tension between honoring historical scientific contributions and addressing the harm that may now be associated with the methods or terminology.
In favor of re-evaluating our terms, one student said, “I believe history should not be rewritten or deleted as it shows the wrongful and horrible treatment of people of color. Still, the people who were involved should not get the notoriety that comes from their creation.” Some students believed that renaming terms was only a first step, calling into question the validity of these statistical methods. For example, one argued: “It's wise to take the step of renaming methods in statistics from racist statisticians further and examine how their creations might have supported the world around them and if these equations need to be changed as well.” Another was a bit more guarded: “We're currently seeing a movement of renaming and taking down sports teams, statues, etc. I think (renaming Pearson’s r) is a great start, but just a start.”
Others presented more nuanced views. One opined, “Given the creators' attitudes, I believe that renaming these methods would be ineffective, as any hidden biases inside them would remain unchanged. As a result, I believe the techniques should be revisited in order to identify any such biases and make a meaningful change.” Another stated, “I don’t think the name should be changed because if keeping these names around reminds statistics teachers to discuss the history of eugenics, then I think it’s potentially a good segue into important discussions.” Indeed, the lesson left some students at a crossroads: “Math, itself, is supposed to be factual and unbiased, so learning about how some methods of statistics are rooted in racism and bigotry sounds hypocritical.”
Conclusion
These results highlight the complexity of the issue presented to students. We intend to forward their attitudes as a starting point for consideration of changes in how we view the methods revisited here. Hypothetically, what would happen if we renamed Pearson’s r as just r, or the ANOVA F as say, the MS ratio? If the changes made it all the way into our textbooks and computer packages, folks would surely notice and wonder why they occurred. This may serve as a starting point of further action, such as re-evaluations of the appropriateness or potential biases inherent in the methods. Or it may just call attention to, and bring awareness of, the questionable world views of the methods’ founders. Either way, inaction is not recommended, at least not by the whole of our modest sample of undergraduate student guides. At minimum, we hope to inspire ongoing conversations on these issues.
References
Adamson, A. S., & Lipoff, J. B. (2021). Reconsidering named honorifics in medicine—the troubling legacy of dermatologist Albert Kligman. JAMA Dermatology, 157(2), 153–155. https://doi.org/10.1001/jamadermatol.2020.4570opens in new window
Cleather, D. (2020, March 11). Is statistics racist?. A number of the most important… | by Daniel Cleather | the startup | medium. Is Statistics Racist?https://medium.com/swlh/is-statistics-racist-59cd4ddb5fa9opens in new window
Delzell, D. A., & Poliak, C. D. (2012). Karl Pearson and eugenics: Personal opinions and scientific rigor. Science and Engineering Ethics, 19(3), 1057–1070. https://doi.org/10.1007/s11948-012-9415-2opens in new window
Gillham, N. W. (2001). Sir Francis Galton and the birth of eugenics. Annual Review of Genetics, 35, 83–81. https://doi.org/https://doi.org/10.1146/annurev.genet.35.102401.090055opens in new window
Kishimoto, K. (2018). Anti-racist pedagogy: From faculty’s self-reflection to organizing within and beyond the classroom. Race Ethnicity and Education, 21(4), 540–554. https://doi.org/10.1080/13613324.2016.1248824opens in new window
Contact
Fatima Alfaro-Varnes, MA
Pepperdine University
Mark P. Otten, PhD
California State University, Northridge
Email
Amelia S.L. Vitale, MSW
California State University, Northridge
Eleni R. Benchek, MA
California State University, Northridge