Professor Mary Armstrong speaks on the ‘persistence of bias’ in STEM

The STEM field is not the only discipline with a susceptibility for bias, but it may get away with it more often. (Photo from Creative Commons)

“What would it mean to say that many aspects of science, technology, and engineering are inherently biased?” Women’s, Gender and Sexuality Studies (WGSS) and English professor Mary Armstrong asked her audience over Zoom. 

Armstrong addressed this open-ended question in her talk this past Tuesday, entitled, “Gendered by Design: The Secret Life of Things”.

Armstrong began her talk by explaining the ways in which people reflect on the history of bias in STEM fields. She said people tend to think about the mistakes made in the days of blatant bias in science, such as eugenics and race purity, as something of the past.

“We are fairly good at recognizing the mistakes they’ve made [in the past], not as good at recognizing the mistakes they are making right now,” Armstrong said.

This shallow relationship with understanding bias in STEM creates a false sense of objectivity, which Armstrong said was especially dangerous and costly.

Armstrong explained that for many, even the notion of understanding science as biased is, by definition, hard to come to terms with. Armstrong noted that the notion “is terrifying to us because it implies that science…is out of our control.”

Armstrong said that STEM is not alone in its susceptibility to cultural bias. But, unlike other academic disciplines, it tends to get away with it due to the false understanding of its objectivity.

“It is very difficult to examine your perspective when you think you are dealing only with facts, and you are objective,” Armstrong said.

Throughout the lecture, Armstrong employed many examples of STEM failing in being objective. One example involved the case of western engineers designing stoves for communities in the global South. The stoves, which were built with the intention of combatting climate change and alleviating the responsibilities of women, were not being used by the communities they were designed for, despite the stove being mechanically designed “perfectly,” which confused the engineers.

“You built a perfect stove in the eyes of a Western engineer,” Armstrong said. “If you had learned the facts, which are that in a particular culture, women cook and never do repair work, that would be inappropriate. When the stove breaks, the men in the household say, ‘Okay, let’s just go back to the way it was because I am not really sure how to fix this,’ and so it all falls apart.”

Armstrong posited that the engineers failed to understand and learn about the communities they were serving and rather extended their perspectives as absolute. The comfort that STEM has with doing this is, once again, Armstrong said, rooted in the idea that scientists are completely objective.

Another example Armstrong used was the implicit bias in voice-activated software.

“There are bodies of voice data on which these machines are trained, it’s called a corpus. So the average corpus has more than 70% male voices. Now is that objective? Does 70% sound like 50-50 to you?” Armstrong asked.

This has caused frustration, as many voice-activated machines failed to recognize or respond to women. 

Armstrong said this was not a case of overt bias against women; rather, it was a bias that was unintended.

“Nobody was sitting in a room saying, ‘I know what we’ll do, we will get male voices to do this, and we will mess those women up’…I think it is a failure, and I think it is bias coming into science and technology and making it less good, and potentially harming people because of the bias that nobody intended,” Armstrong said.

The persistence of bias in STEM is sometimes puzzling because no one benefits from it. In fact, employees may actually lose money by having to go back and correct the problem, Armstrong noted.

So the question is, how can more inclusive designs become incorporated into STEM fields?

Armstrong said that the burden of this task is often misplaced on women and people of color.

“If I had a nickel for every time somebody said ‘We are going to have diverse teams in engineering. That is going to solve everything.’ The role of women and people of color and women of color is not to correct the dominant perspectives of white men. Their mere presence is not some weird antidote to bias,” Armstrong said.

Armstrong did, however, emphasize the great importance that diversity has in academic environments.

“Women and people of color, because of social conditions, have new perspectives. This means new knowledge and insights the marginal user might actually know best,” Armstrong said.

But, as Armstrong concluded, the benefits of diversity can sometimes lead to tokenism. 

So how can the college best facilitate a world which avoids this “bad science masquerading as objectivity?”

The answer, according to Armstrong, is to learn.

“It is important to know about what I just told you because everyone…needs to learn how to think outside of their own box and understand that the ‘universal’ user is likely white, male, Western and abled,” Armstrong said.

Leave a Reply

*