The Justice, Equity, Diversity, and Inclusion (JEDI) Outreach Group Corner is a regular component of Amstat News in which statisticians write about and educate our community about JEDI-related matters. If you have an idea or article for the column, email the JEDI Corner manager.

David Banks is a professor in the department of statistical science at Duke University who is deeply grateful for his longstanding engagement with the ASA.

I am very proud of the ASA for its recent JEDI initiative and the Anti-Racism Task Force. Both make the association better and fairer, but they also introduce complexity. It is easy to say we oppose racism and inequity, but operationalizing such principles is not straightforward. As a specific example, I point to the fact that the ASA has never elected a statistician of Chinese descent as its president.

When I raise this point with friends, some point out that two have been nominated but lost the election (Xiao-Li Meng and Christy Chuang-Stein). As a democratic organization, we have to respect the process, but that position leaves the door open to the ASA Committee on Nominations inadvertently nominating pairs of candidates in which the outcome is possibly determined by implicit biases.

The existence of such biases has been widely documented. When a white person runs against a Black person, or a male against a female, or one ethnicity against another, or a cis-gendered person against a trans-gendered person, or a young person against an old person, such biases are almost inevitable. The ASA membership probably wants to figure out a fair way to minimize the impact of such implicit bias on its election processes. This will be difficult to do.

In the past, the ASA seems to have taken deliberate steps to eliminate one potential source of bias. We made the decision that the office of the president should rotate among the academic, industry, and government sectors. That plan ensured different kinds of stakeholders were regularly represented at the highest executive levels and that a populous sector did not dominate statisticians in smaller sectors. Nothing is perfect, but I think that system has served the ASA well on the whole.

The naïve approach is to nominate opponents who have the same race, gender, ethnicity, age, and so forth. But as we know from the design of experiments, that much balance is hard to achieve. Since a potential candidate may decline the nomination, it will be completely impracticable to seek such a solution.

Alternatively, one might hypothetically try to estimate the magnitude of the implicit biases the ASA membership collectively feels for or against different races, genders, ethnicities, etc., and then seek a pair of nominees such that the net implicit bias between the two is nearly zero. That poses an interesting statistical challenge and seems like an interesting research topic that is amusingly self-referential to statistical inference. Of course, this idea is still absurdly impractical.

According to the ASA website, the Anti-Racism Task Force is charged with the following:

• Developing recommendations that ensure the communications and activities of the association’s groups align with its position on justice, equity, diversity and inclusion
• Proposing mechanisms for the association to inform the public about how statistics and data science can contribute to—or, if used responsibly, help fight against—racial and ethnic bias in society
• Developing recommendations for ASA infrastructure and policy that will help drive positive cultural change within the ASA and remove structural barriers to justice, equity, diversity, and inclusion

Nearly all of us want to achieve these goals. Members of the JEDI group and Anti-Racism Task Force have shown a strong beginning, and their existence is a necessary and laudable first step. The follow-through, however, is technically difficult and poses complex ethical challenges I don’t think anyone really knows how to address.

The good news is the world is changing, and implicit biases may now sometimes advantage the previously disadvantaged, which brings me back to the Chinese statisticians. By almost any standard one can imagine, Chinese statisticians have been regularly passed by for leadership in the ASA, despite their prominence in our profession and their long history of outstanding contribution. Of course, we absolutely need to support underrepresented minorities, ethnicities, and genders, but when an overrepresented group is consistently overlooked, it should be no surprise that some Chinese statisticians might begin to perceive it as prejudice.