Questions have been raised about the Implicit Association Test (IAT), a psychological test which purports to find implicit bias in individuals. Olivia Goldhill, writing for Quartz, an online business journal, reports that the IAT has a low reliability, or test-retest, score; where perfect reliability would score as 1, and strong reliability would score as 0.7-0.8, the race IAT has a reliability score of 0.44, or unacceptable. Goldhill also reports that several meta-analyses have found that the IAT is a poor predictor of behavior.
I have my own criticism of the well-known race IAT, which you can take online at the Project Implicit Web site. I took this test online, and scored as having a low to moderate bias in favor of African Americans. As much as I’d like to think I’m Mr. Egalitarian, I had a problem with the test: it required me to make fast judgments about low-resolution photos of facial characteristics, and I know myself well enough to know that I have poor facial recognition ability — I once passed my mother and younger sister on the street and only recognized them when I realized that these two women were laughing at me — so any test that requires me to recognize facial characteristics is not going to produce accurate results.Regardless of the strengths and weaknesses of the actual test, I’m still skeptical of using tests for implicit bias to implement organizational change. In my experience, that’s not the way organizational change actually happens: it’s not as easy as administering a test, identifying who has implicit bias, and then watching the complete eradication of bias. If it were that easy, we already would have eradicated racism, sexism, etc.
“What the ‘Bias of Crowds’ Phenomenon Means for Corporate Diversity Efforts,” an article by Liz Kofman (a change management consultant with a doctorate in sociology), suggests a different path towards changing organizational biases that I find more pragmatic. Writing for Behavioral Scientist, an online non-profit magazine, Kofman identifies three recommendations for organizations wishing to get rid of bias.
First of all, Kofman suggests that we “focus on changing processes, not people.” In other words, forget all those training sessions where you make individuals in the organization confront their inner biases; instead, change your organizational processes to reduce chances for bias. Why don’t more organizations do this? I suspect it’s because it’s much easier to hold a workshop on implicit bias than it is to do the hard and detailed work of changing organizational processes. It’s fine to hold those workshops, and Unitarian Universalist congregations wishing to address bias should continue to offer, for example, the excellent “Beloved Conversations” class developed by Mark Hicks at Meadville/Lombard Theological School. Just don’t expect one workshop to take the place of lots of rather boring but necessary detail work.
Kofman’s second recommendation is to “prioritize process change and stick to it.” She points out that this is not easy; it takes “organizational will and discipline to implement and sustain … new processes.” Furthermore, Kofman says, an organization needs to focus on a few key process changes, making those a priority; otherwise, it’s easy to get overwhelmed with too many changes and then nothing happens. Prioritizing process changes, and sticking to them, has proved to be an insurmountable problem for most Unitarian Universalist congregations I’ve known: lay leadership changes from year to year and so priorities change from year to year; new and attractive projects arise and draw attention away from ongoing projects. It’s easier to do that high-profile capital campaign, or to add solar panels on your building, than it is to stick to the hard work of implementing new organizational processes designed to reduce racism and sexism.
And this brings us to Kofman’s third recommendation: “provide resources and incentives for change management.” Because “everyday processes influence the bias of crowds” in any organization, you need to change those everyday processes; but too often there are not resources to help people change those processes, in addition to which there’s little incentive for change. Take for example a Unitarian Universalist congregations which wishes to become less white. Clearly, one thing you’ll need to do it completely overhaul the intake process — how potential members are greeted on their first visit, the processes used to integrate newcomers into the congregational culture, and so on. All that is hard work, so one critical resource required for change will be staff time, from both paid and volunteer staff; and because staff time is a limited resource, other projects will have to received fewer staff hours. And how will you provide incentives for those staffers, particularly for the volunteer staffers? None of this is easy.
To summarize: While Implicit Association Tests might be fascinating, they are probably not particularly useful tools for implementing organizational change. Instead, congregations seriously committed to, e.g., becoming less white, should pay attention to the change management technique of process change.