Pando

YouTube’s science communication channels have a diversity problem

It turns out science communication faces the same challenges on YouTube as it does in academia.

By Chris Stokel-Walker , written on November 5, 2020

From The Culture Desk

YouTube is often heralded as the great leveller, lowering barriers to entry and destroying discrimination. But a new academic study suggests that the platform’s science community mirrors the world of science in real life – with all its problems of representation.

Kaiping Chen of the University of Wisconsin-Madison, and colleagues at Tufts University and Carnegie Mellon University, have set out to look at how so-called “knowledge producers” are represented on the world’s biggest video sharing platform.

As part of a study, they analysed the videos posted in four key scientific topics: climate change, GMOs, vaccination and human genome editing, monitoring who posted them and how they connected. In all, they analysed nearly 10,000 science videos on YouTube. 

“I wanted to look at whether YouTube solved the diversity issue in science, or just replicated what is offline,” says Chen. “What I found was if we look at the profile of people, if they’re presented in their profile picture, I see there’s more diversity. But if I look at who connects with who, it’s still segregated. It’s still dominated by a certain group of people who are connected with each other and promote each other intensively.” 

The researchers coined the phrase “segregated inclusion” to encapsulate the problem.

Each YouTuber found in the search of videos was connected to the others surfaced through the research dragnet. Of the group’s “superconnectors”, who were cited by more than five other YouTube channels, only 12.5% were women.

While the range of creators publishing videos on scientific topics seems much more diverse than in real life – where, for instance, the proportion of Black science researchers in faculty positions in US universities’ STEM laboratories is half the proportion of the general population – there are tribal divides as to who is promoted by the community.

And concerningly, there appears to be a lack of diversity in who YouTube’s algorithm promotes as key science communicators. 

“Those of us in the science YouTube space have long been aware that diverse representation is a major area for improvement,” says Rohin Francis, a UK cardiologist who posts under the Medlife Crisis channel to his 310,000 subscribers. “This is reflective of barriers to success both on YouTube and in science as a whole.”

Unfortunately, science communication is often considered a 'less important' obligation.

Francis says it isn’t surprising that YouTube’s science communicators appear more diverse than the core group of researchers and academics in educational institutions worldwide. “Women and non-white scientists are disproportionately allocated science communication roles, often regarded as ‘less important’ obligations,” he says, “so we should expect a medium like YouTube to demonstrate a more representative balance of voices.” 

Highly popular science communicators, including Jim Al-Khalili, who is also a respected academic in the field, have previously told reporters that early on in their career, they were warned against pursuing a path explaining science to the public alongside their academic work because it was seen as less prestigious. 

“When I started in science communication in the early- to mid-90s, I was very active as a researcher and my colleagues warned me against doing too much communicating,” Al-Khalili told Wired in 2018. “They said: ‘Why are you doing that newspaper piece, or talking on radio? You should be focusing on getting this paper finished and running this computer code to get these results.’ I was very research active. I wanted to know why I couldn’t do both.”

Francis does query some parts of the study. “There are many methodological problems with this piece of research, and I would question some of the authors’ conclusions, but it is important that some quantification of the problem has occurred,” he says. “YouTube is a major source of education worldwide, and inherent biases can have wide-reaching implications.”

Chen believes that YouTube – and YouTubers – need to do more to ensure diverse voices are heard from the wider scientific community. 

“There are two ways of fixing this,” she says. “One is looking at the producers. I would highly suggest when those big YouTubers want to promote their videos, when they think about who they’re going to recommend, try to break out from their echo chamber.”

She also sees an issue with the platform as a whole. YouTube’s recommendation algorithm reportedly dictates what people watch up to 70% of the time spent on the platform – and it’s pushing people to predominantly white male creators, says Chen.

“We need to think about whether we should not always push out the voices of those with huge numbers of subscribers,” she says. “Maybe we should think about whose voices we might want to overrepresent more. This is where I think the group of engineers in the industry need to have more social science-driven thinking to see how the algorithm can help solve the representation problem.”

The issue isn’t one that can be fixed tomorrow, she admits. “I know it’s a big challenge for the company, but fundamentally, if we really want to solve the problem, we need a conversation between academia and the companies.”

For more from Pando, you can subscribe to our weekly newsletter. We'll send it to you every Friday.