The Women in Data Science Diversity Gap. All Biology?
Are biological differences between men and women responsible for the gender equality crisis in AI and data science?
A question that popped up during an Omdena webinar addressing the topic “Why we need more female leaders in Artificial Intelligence”.
Discussing women in data science
A webinar attendee noted that the male and female gender have biological and personality differences. As a result, more men are inclined to study STEM and thus are better at math and science in comparison to their female counterparts.
We believe that the biological argument of why there are more men in AI and data science is flawed and harms the discussion around gender diversity and equality.
The initial debate around the topic started while I was joining an Omdena challenge with the World Resources Institute (WRI) on connecting land conflicts with policies for land restoration.
One of Omdena’s mission points is to increase the ratio of women in their AI challenges (currently 35 %).
In this article, co-authored by Michael Burkhardt, we will present various perspectives and data points to evaluate the topic.
Important numbers on gender diversity
There have been press releases from some of the big tech companies (FAAMNG — Facebook, Amazon, Apple, Microsoft, Google, and Netflix) reporting a higher percentage of female employees in their workforce.
These numbers are global and not inclusive to for example the US workforce.
Other countries have a higher workforce of women than in the USA. In India, 35% of developers are women while in the USA only 16% are. The other factor for these high percentages is that technical roles have a broad definition.
When it comes to the people researching and creating artificial intelligence — a technology where algorithmic bias becomes a key issue — women account for only an estimated 12 percent of AI researchers.
The value of diversity of thought
Barbara Oakley, instructor of the popular course, Learning How to Learn, states:
People who pivot from one field of learning into another can often bring with them helpful perspectives. If something doesn’t come naturally to you and it feels a bit hard, then you are using different neural patterns to learn and you may be more creative as a result.
Who is better at math?
She also has an interesting take on how early average differences in the development of skills may link to the issue of gender diversity in STEM:
When little boys and girls are growing up, their math and science are equivalent, on average, but the boys’ verbal abilities lag some whereas the little girls’ verbal abilities are ahead. The boy thinks that ‘I am better at math and science’; and the little girl thinks, ‘I am better at verbal than math and science’ — which is true, even though [the boys and girls] have the same abilities in math and science. And we tend to develop passions about things that we feel we are good at and that come easy. This is where the message to broaden out from your passions could play a key role, so that girls and young women continue to engage with science and maths.
This statement is supported by data from the PISA math test, showing that girls only scored about 2 percent, on average, lower than boys.
More data, charts, and commentary here.
Before writing this article, I first contacted Barbara Oakley and she referred me to the research of David Geary.
David is a cognitive-developmental and evolutionary psychologist, professor, author, speaker and consultant on mathematical cognition and learning as well as the biological bases of sex differences.
Referencing one of his papers: women in countries with higher gender inequality have more motivation for financial freedom that STEM fields can provide. They measure gender inequality using the World Economic Forum’s Global Gender Gap Index.
It says in countries that are more gender-equal, women are more empowered to choose what they enjoy the most. This is similar to what Oakley states — that we tend to go with what we are good at, working at it, and continuing on that path as we get older. They found that girls and boys have similar abilities in science literacy in most nations.
This leads me to think that a girl is more likely to push herself into STEM, in favor of economic independence, despite negative self-perception or an environment telling her she has a better ability at reading or non STEM fields.
How to increase the Women in Data Science ratio?
Further research by Geary suggests balancing diversity in STEM fields will require, “more than improving girls’ science education and raising overall gender equality. The generally overlooked issue of intraindividual differences in academic competencies and the accompanying influence on one’s expectancies of the value of pursuing one type of career versus another need to be incorporated into approaches for encouraging more women to enter the STEM pipeline. In particular, high-achieving girls whose personal academic strength is science or mathematics might be especially responsive to STEM-related interventions.”
The other side of the argument
There are some, such as Jordan Peterson, who was mentioned in our webinar, that preach ideas such as order is masculine and chaos is feminine, and white privilege is a farce. With regards to math and science, he writes agreeing with Geary’s research on gender-equal countries having less diversity in STEM fields, though he states and quotes research on male/female personality differences being the cause for occupational choices and success in those fields.
He asks why should we bother with creating an ideal utopia, “where every occupation and every stratum of authority within every occupation is manned (so to speak) by 50% men and 50% women.”
James Damore, the Google engineer who was fired for writing a memo echoes some similar sentiments that the reason there are fewer women in tech fields is not because of bias and discrimination but because of psychological differences between men and women.
Discrimination to reach equal representation is unfair, divisive, and bad for business.
Our take on more diversity in Artificial Intelligence
We are not arguing there are no differences between genders that could be biological and psychological, or that only the environment is an affecting factor.
I believe embracing differences people may have, not stereotyping them, seeing the strength in diversity in the workforce is actually good for business, not just an ethical thing to do. In fields such as AI, Machine Learning and Computer Science, diversity is crucial.
Gender bias has already entered into algorithms.
Having joined several Omdena challenges with 40 to 50 engineers building AI for Good solutions for tough social challenges, I’ve experienced first hand the value of diversity of thought.
Some other great initiatives are the Women in Data Science Conference (WiDS) and Women in AI (WAI) community.
In addition to gender diversity, this also applies to diversity in race, ethnicity, and special needs.
About Omdena
Omdena is a global platform where AI experts, engaged citizens and aspiring data scientists from diverse backgrounds collaborate to build AI-based solutions to humanity’s toughest problems. Join us here.
