Jen DuBois By: Jen DuBois

Women in Data Science in 2020

Over the past year, QuantHub has seen two women from diverse backgrounds win our first data challenge.  We hosted a highly informative webinar with two successful female data science managers and we launched a Women in Data Science Scholarship.  As a data science and engineering skills assessment platform, focused on helping companies hire the right talent for the right role, we’re well aware of the movement to promote more women into the data science field and tech in general. In fact, many of our customers are female data science leads.

So in light of the upcoming International Women’s Day, we thought it timely to take a look at the issue of women in data science and tech.  What’s the status in 2020? Have we made progress recruiting women into the field? What is being done and what remains to do? Where are the challenges and successes?

In this article we’ve brought together a variety of information that speak to these questions and which help to summarize what the data science arena looks like for women in 2020.

Why Do We Need Women in Data Science?

First off, let’s establish why we should care about the representation of women in data science (besides the fact that it is simply fair to afford women equal opportunities).

1.     It’s good for company financial performance

In 2014, McKinsey started studying the impact of gender diversity on company performance. In that first study and since then, McKinsey has consistently found that companies with superior financial performance have more gender diversity than average or underperforming companies. Gender diverse companies are 21% more likely to outperform 4th quartile companies on their EBIT margin and 27% more likely to outperform them on long term value creation.

It’s no secret that in today’s world “being a data-driven company” is code for ensuring the success of the bottom line for any company. It makes sense then that any company claiming to create value out of data have a decent representation of women in data science roles and in the company in general.

2.     Women are really good at coding

When we spoke to Linda Burtch of BurchWorks recruitment firm last year, she told us that in her experience,

“Women are great coders. They are naturally organized thinkers and are into the details. These traits go well with coding.”

Linda’s hunch was confirmed in a recent New York Times piece, “The Secret History of Women in Coding”.   The fascinating and well researched article revealed that several decades ago most coders were women.  Did you know that in the early days of IBM, women were the coder gender of choice?

A woman coder at IBM in the 1960s

This was because of the very limited capacity of computers back then.  Code had to be very very precisely written and coders had to have the capacity to envision how the computer would translate the code they were writing mentally in their heads.  The computing industry identified that women possessed the kind of picky and iterative capacity that was needed for coding on the earliest computers.  In fact at M.I.T in the 1960’s the majority of government labelled “career programmers” were women.

Why women started to disappear from programming jobs in the mid-1980s is further explained in the New York Times article. We suggest you have a read. But suffice to say it wasn’t due to lack of coding abilities!

3.     Data science needs diversity

It’s clear that data science is a team sport that requires a plethora of diverse skills.  Women certainly bring traditionally female skill strengths such as the aforementioned organizational skills and high intuition that contribute to data science success.  In addition, having women on a data science team helps to ensure that data and AI products are representative of the population at large and therefore help to reduce bias in the development of data science and AI solutions.

Kavita Sangwan, Director of Technical Programs, AI and Machine Learning at Intuit explained her thoughts on  this in a Medium post writing,

The more our organization represents the full breadth of our customer base, the better we’re able to deliver on Intuit’s mission to power prosperity for people of all kinds.

Put simply data-driven solutions must reflect not only male perspectives and needs, but female ones as well if they are to avoid biased outcomes.

Current Stats and Trends for Women in Data Science

OK, so we know why it’s important to include women in the data science field, but do the stats reflect their importance now?  Here’s a few datapoints.

The Current State of Women in Data Science

The infographic on the left put together by Betterbuys.com last year is a pretty good snapshot of the current state of women in data science. Although some of the data is a little old, we find nevertheless that it represents more or less the status quo.

We suspect that the 26% female representation may have something to do with the high turnover of women in data science roles. Many other sources such as recruiting agency and consulting firm studies also indicate a range of around 20-30% participation of women in “data-related roles”.

Note that women are twice as likely to leave a tech role as men and 50% of those will leave the industry altogether. Not. Good.  We’ll get to why this may be a little later.

The WiDS Initiative is Gaining Momentum

Stanford’s Women in Data Science Initiative has been around for about 5 years but has gained a huge amount of support.  The summary of its accomplishments and the growing level of interest shown in the graphic below is indicative of where the movement to include women in the field is going.

 

 

 

 

 

Large companies are (finally) actively recruiting female data scientists

WiDS has certainly caught the attention of global companies who historically have had low rates of females working in tech as demonstrated in the image below.

Companies such as Microsoft, Walmart, Facebook, Intuit and Google have gotten behind the women in data science movement. In one recent interview, John Hoegger, Principal Data Science Manager at Microsoft explained how Microsoft is promoting the idea of women in data science,

We make sure that we have women on every set of interviews…What’s it like to be a woman on this team? If it’s all men, you can’t answer that question. I’ve now got a team of 30 data scientists and half of them are women.

In addition, tech companies like Slack and Amazon that have low female representation have started adopting the “Rooney Rule” which mandates including females minorities in any job interview process. It remains to be seen whether it will be enforced with rigor.

Female representation is sloooowly increasing

In their 2019 Salary Survey, Burtch Works recruitment agency found that just 17% of Data Scientists and 26% of Predictive Analytics Professionals are women. (The difference between the two roles is that Data Scientists do more coding and work with larger, messier data sets).

Despite these low figures, there’s hope.  According to Burtch Works roughly a third of entry-level jobs in these two roles are filled by women, meaning that the new generation of female graduates is finding more roles in data science.

Another bright spot and sign of increasing female representation is female enrollment in top computer science degree programs. Colleges such as Harvey Mudd and Carnegie Mellon have been actively and successfully recruiting women to their computer science degree programs for several years. 50% of Carnegie Mellon’s computer science enrollment is now female and 40% of Harvey Mudd’s is.

Females in senior data science roles is still lacking

The story at the other end of the job experience spectrum, however, is different.  Women are greatly lacking at higher levels of data science management, with mid to senior-level management positions showing just about 15% female representation. The high rate of turnover described previously is no doubt the root cause of this statistic. Women are difficult to retain in data science roles.

Speculation as to why this is includes the need for work/life balance that comes about when women have a family and the lack of paid leave for fathers, both of which are at odds with the demanding and intensely competitive nature of the data science field.  We discuss this more later in the next section.

What’s Keeping Women out of Data Science?

With all the hype around hiring and keeping women in data science and tech roles, you’d think the numbers would be growing more rapidly.  The reality is, as many studies show, that the usual factors which inhibit women in the field linger.

Research by the Alan Turing Institute in the UK validates that the reasons for these lower percentages of females in data science range from sexism, bullying, and sexual harassment to gender pay gap, slow career progression for women, male-dominated office culture, lack of access to mentors, lack of paid maternal leave and gender bias in hiring. All familiar themes.

A few novel findings about why data science is not raking in more female applicants came out recently in a 2020 Boston Consulting Group study.  It found that almost 50% of female STEM students perceive data science to be “overly theoretical and low impact”.  The reality is that many many companies still do not place enough emphasis on being data-driven nor do they place high value on data science initiatives.

BCG found that this lack of investment in data science is a turnoff to 75% of women, while men are more ambivalent.  In addition, the field of data science itself is a turnoff because women get the sense that it is a more competition-focused field than other jobs.

The study found two additional reasons why women shy away from careers in data science.  One is that they don’t have a good understanding of what a “data science career” is and what the day-to-day life of a Data Scientist in the workplace entails. The other is that many companies still lack a culture of collaboration in their analytics teams, a situation that are more apt tend to avoid.

How Can Companies Increase the Number of Females Working in Data Science?

A number of suggestions and ideas have been put into practice to try and recruit more women into data science. The following are a few ideas that have been successful it would seem.

The “right” kind of diversity initiatives

A 2016 Harvard Study on diversity showed that the initiatives in the diagram below had the greatest impact on increased diversity in the workplace. They represent familiar themes in the data science world: mentoring, putting more women in the recruitment pipeline, managerial decisions to create diverse teams, and creating diversity manager roles.

Address concerns specific to women explicitly during the hiring process

Companies should be much more specific in their communication with female candidates, directly addressing the concerns that women highlight in their feedback about job interviews.  For example, hiring managers should discuss what role data science has within the business because women want to have an impact (while men are more complacent in general).  Women also want to know how data scientists work together on use cases and how a career path in data science involves more than coding.

Companies should therefore provide female candidates with real-life working examples and proof of value and the tangible contribution of data science to the company. They should avoid talking theoretically about AI initiatives and being overly focused on theoretic questioning in interviews.

Create a women-friendly management and team culture

In the 2020 BCG study, they also noted that hiring managers

must create a visible culture within their data science teams that celebrates impact and shuns competitiveness, and then make this career opportunity very tangible and attractive to students of both genders.

In his article in American Banker “What Tech Managers Can Do About Gender Imbalance” former QuantHub webinar guest Jacob Kosoff, who is the Head of Model Risk Management and Validation at Regions Bank, talked largely about management and culture as the tool he used to grow his team from 13% female in 2014 to 46% female in 2019.

Here’s Jacob’s list of tips for management and culture:

  • Listen to what associates want and do not assume you know what they want.
  • Have honest career development conversations. Be proactive in asking associates about their talents, needs, interests and goals.
  • Regularly ask your team members to assess the workplace culture. Ask for their ideas on how to improve skills development and provide more flexibility.
  • Empower individuals to incorporate their outside-of-work passions.
  • Partner with individuals to help them find a neutral and honest mentor outside of your department.

Finally, many female Data Scientists recommend that management encourage divergent and diverse thinking to encourage a female friendly team culture.   They note that in traditional biased male-oriented teams,  people are not used to someone thinking differently and challenging the way they’re thinking. This is an issue that needs to be addressed if data leaders want their team to be a more inclusive one.

Anonymize skill testing and other job application factors

A newer concept being put forth to solve the problem of gender representation in STEM fields is that of anonymizing applicant information.  One study described in the chart on the right shows how the percent of women hired for research increased the more anonymized their applications were.

What can women do to increase their chances of working in data science roles?

Keep applying!

The reality is that there are going to be so many more jobs in machine learning and AI and so many more jobs that deal with data that the world is going to need more Data Scientists, Data Engineers and other data-related roles. Men simply cannot fill them all!

Focus on being confident and competent, not your gender

Claudia Perlich, PhD and advisor at Dstillery expressed this to female Data Scientists in an interview,

I want to come to work and do what I love and be recognized for what I bring to the table and not waste even one thought on the fact that I am a woman. Most successful women I know in the field seem to have this attitude and are very comfortable with themselves and their roles.

Other female Data Scientists often mention a lack of confidence they perceive in female candidates and a hesitation to apply for roles and promotions because they don’t possess every skill listed on the job description. Women need to find ways to avoid losing confidence such as learning new skills or participating in datathons.

Put your name in the hat for senior roles

Following on the previous point, it’s important to have a conversation with your manager or HR that says “I want to be in the pipeline for senior management.”  With the clear lack of female representation at the highest levels of data science it’s easy for unconscious bias to creep into the minds of your mostly male superiors and for assumptions to be made about what female Data Scientists want or don’t want career-wise.

And don’t just find a mentor that can make suggestions for what skills you should develop or what kinds of roles you should take on. Do as men do. Find a mentor whose relationship with you will effectively give you a public endorsement of authority and competence for a senior role.

Conclusion

In the end, there’s no silver bullet to increasing the representation of women in data science and tech. These are just a few observations about women in data science and suggestions for increasing their numbers.   Thankfully, after taking a look at the role of women in data science one thing is clear, it’s moving in the right direction. Who knows, maybe 2021 will bring statistics closer to 50%!

+ Add Comment

Leave a comment

Your email address will not be published. Required fields are marked *