Narrowing the gender gap in computing

Psychology, R, invisibility cloaks, the Pied Piper and a Trojan horse

Andy Field

University of Sussex

The pioneers of modern tech

Image from Guest, O., & Forbes, S. H. (2023). Teaching coding inclusively: if this, then what? https://doi.org/10.31235/osf.io/3r2ez

Ada Lovelace (1840s): Creating (arguably) the first computer program. Worked with Charles Babbage on his proposed mechanical general-purpose computer, the Analytical Engine. Grace Hopper (1950s): Inventing the compiler Frances ‘Betty’ Snyder Holberton, Jean Jennings Bartik, Kathleen ‘Kay’ McNulty Mauchly Antonelli, Marlyn Wescoff Meltzer, Ruth Lichterman Teitelbaum and Frances Bilas Spence (1940s): programmers of the ENIAC computer (a secret US Army project during World War II)

Trends over time¹

• Computing occupations
• STEM degrees
• Retention

Y-axis should be labelled proportion, not percentage.

In non-stem, over nearly 20 years a proportion of 0.8-0.9 are retained (so 10-20% dropout) In STEM, retention is at about 0.7 (30% dropout) after 3 years, and close to 50% dropout over 10 years

1. Figures from Ashcraft et al. (2016). Women in tech: The facts. NCWIT. https://ncwit.org/resource/thefacts/

“Cracking the gender code”¹

“Only by tailoring courses to girls’ specific needs can we boost their commitment to computing.”

• Accenture report with Girls Who Code: Cracking the Gender Code (2016)
• Interviews/Surveys
  • 4,000 girls in junior high and high school,
  • 500 boys aged 12–18,
  • 650 male and female college students,
  • 2,000 working women < 30 years (500 working in computing)
  • 250 teachers and principals
  • 2,200 parents of girls aged 12–18

1. Accenture & Girls Who Code. (2016). Cracking the gender code.

Accenture & Girls Who Code. (2016). Cracking the gender code. www.accenture.com/us-en/about/inclusion-diversity/cracking-gender-code

Junior high = 10-14 years High school = 14-18 years College = 18+

Cracking the gender code: recommendations

Junior high (10-14 years)

• Action 1: Deepen girls’ hands-on computing experience
  • More exposure to games = more likely to go into coding (But …)
  • Introduce coding through computer games

• Action 2: Change girls’ perceptions of computing
  • Challenge stereotypes
  • Role models

• Action 3: Support parents in understanding the wider role of computing

“… there is little awareness among any group that applying computing and coding to real societal problems can help change the world, which can be a powerful hook for girls and a magnet for women.”

High school (14-18 years)

• Action 1: Redesign computing courses to appeal to girls

“81% of high school girls who studied computing over the summer were interested in studying computing at college, compared to 52% of those who only studied computing at school.”

• Action 2: Create grassroots campaigns to motivate peer group action

“Definitely [girls want] more clubs and after schools programs. We should be making a connection between ladies of distinction and computing or coding, because a lot of the girls want to feel empowered.” (Teacher)

• Action 3: Attract more women teachers

“62% of girls in high school who have had someone encourage them to study computing and coding say they are likely to major in it at college, compared to only 15% who have had no role model.”

College (18+ years)

• Action 1: Give computing courses a makeover

“This includes engaging young women by describing computing courses – and potential careers in the field – in ways that capture women’s strong interest in problem-solving, especially around real-world and social issues.”

• Action 2: Offer female students immersion programs (allow exposure to coding on non-coding courses and make it easy to switch).

• Action 3: Create female mentorship and role model programs

“… the influence of role models is strong among women who don’t study computing at college but then go on to pursue a computing career.”

3 Recurring themes

• Exposure to coding in non-coding environments
  • Games, clubs
  • Non-coding courses

• Capture women’s interests
  • Teach the wider context of computing
  • Teach within a more female-centric context

• Provide role models and mentors
  • Challenge stereotypes
  • More female teachers/mentors
  • Inspiration

Question

Where does Psychology fit in?

Recommendation 1

Exposure to coding in non-coding environments

Psychology is a trojan horse for computing. You think you’re getting a beautiful wooden horse full of psychology. You welcome it in, only for a bunch of R to jump out and stab you ….

The reality of studying psychology

• 75% Psychology

• 25% Statistics, research methods, statistical computing
  • General R (|>, here, knitr)
  • Computing (Functions, loops, , Markdown)
  • Distributions, sampling theory, probability
  • Data viz (ggplot2, GGally)
  • Descriptive statistics (datawizard, dplyr, forcats, readr, tibble, tidyr)
  • Associations (correlation)
  • Gaussian linear models (afex, broom, car, effectsize, emmeans, ggfortify, lm(), modelbased, parameters, t.test())
  • Logistic regression (glm(), broom)
  • Mediation (lavaan)
  • Robust methods (robust, WRS2)
  • Factor analysis (psych)

Teach using interactive tutorials¹

1. The discovr package www.discovr.rocks/discovr/

R integrates aspects of coding and computing.

Student quotes

I was taken aback at first but it has grown to be my favourite part of psychology.

Learnt many coding skills I didn’t think I would be able to.

Learning R was difficult at times but overall I found it fun as it’s a completely new and quite unexpected activity and the way it is taught to us is really engaging and quite satisfying.

INJECT R INTO MY VEINS!!

Recommendation 2

Capture women’s interests

I was trying to think of a metafor for Psychology, and I keep coming back to the story of the pied piper. Probably because of the megadeth song.

I feel as though psychology lures in innocent sweet children with its sweet and harmonious symphony only to lure them to the mountain of data science from which some will never return

HESA student data 2019-20¹

• Proportion of women
• Number of women

It seems like computing has an issue with both attracting female school-leavers, and retaining women that choose computing as a career

Needless to say there’s a lot of research into why these problems exist.

I’m going to focus on one particular piece of work

1. Data from www.hesa.ac.uk/data-and-analysis/students/what-study

Why might more women than men choose psychology?¹

“… capture women’s strong interest in problem-solving, especially around real-world and social issues.”

Do women have strong interest in problem-solving, especially around real-world and social issues?

Su et al. (2009) massive meta-analysis looking at differences in interests in men and women. Let’s focus on the RIASEC model of vocational interests.

Technical manuals for 47 interest inventories were used, yielding 503,188 respondents.

Men prefer working with things and women prefer working with people, producing a large effect size (d = 0.93) on the Things–People dimension. On the Data-Ideas dimension there was a small effect towards women favouring ideas over data (d = -0.1)

Men showed stronger Realistic (d = 0.84) and Investigative (d = 0.26) interests Women showed stronger Artistic (d = -0.35), Social (d = -0.68), and Conventional (d = -0.33) interests.

Age effects: as age increases the sex differences on Things-People decrease as do differences on social interests.

Diagram shows the main areas of model. Larger circles show larger pooled effect sizes representing male-female differences. The key finding is women are more interested in people, and men in things.

RIASEC model of vocational interests Realistic - interest in working with things and gadgets or working outdoors; Investigative interest in science, including mathematics, physical and social sciences, and biological and medical sciences; Artistic interest in creative expression, including writing and the visual and performing arts; Social interest in helping people; Enterprising interest in working in leadership or persuasive roles directed toward achieving economic objectives; and Conventional interest in working in well-structured environments, especially business settings.

1. Su et al. (2009). Psychological Bulletin, 135(6), 859–884. doi.org/10.1037/a0017364

Psychology is about people, problem solving & social issues

• Does social media negatively affect mental health?
  • Longitudinal Study of Young People in England (LSYPE2), Viner, et al. (2019)

• Does playing “hard to get” make you more desirable?
  • Birnbaum et al. (2020)

• Do good friendships buffer against poor family relationships?
  • Kaufman et al. (2022)

• Can music be used to regulate emotions?
  • Cook et al. (2019), Sharman & Dingle (2015)

• Does scary TV lead to anxiety?
  • Field & Pierce (2016)

Psychology and coding

“… applying computing and coding to real societal problems can help change the world, which can be a powerful hook for girls and a magnet for women.”

• The skills acquired through learning will be taught within the context of Psychology examples.
• Psychology is a science that looks to addresses societal and inter- and intra-personal problems.

Student quotes

I really enjoyed the practicals and surprisingly found coding not that scary but yet interesting.

I personally just really enjoyed learning R. I know this isn’t something most students enjoy learning, but it has been one of my favourite parts of university so far and I am quite upset to part with it (for now!).

Tidy Tuesday¹

Julia Silge (SILL-GHEE)

1. https://juliasilge.com/blog/

Text analysis of descriptions of chocolate alongside ratings of it. So, data set people rate chocolate then describe it.

Text cloud shows high ratings associated by lots of people with words balanced, complex, creamy. Low ratings associated with the words pastey, chemical, bitter.

Recommendation 3

Provide role models and mentors

I need a more positive metaphor, so I’m going for a cloak of invisibiloity. Who doesn’t like one of those?

So, for some students, their interests and career desires lay hidden under a cloak of invisibility. they think they want a career in psychology but only because data science is hidden to them as an option. Mentors remove the invisibility cloak revealing options and confidence they never thought they had!

Role models

“62% of girls in high school who have had someone encourage them to study computing and coding say they are likely to major in it at college, compared to only 15% who have had no role model.”

Our undergraduates do not meet a make faculty member teaching stats/computing until autumn term of year 2 (where they meet me)

Across their core methods modules, they are taught by 2 male faculty compared to 7 females (this is roughly in line with the ratio of 80% female intake).

This may also help to challenge some of the stereotypes in male students

Student quotes

I used to have a panic attack every workshop last term and now I got 82!!!!!!!!! WTF When you said I’m capable of a 1st in this module I thought it was BS like there was no way that was going to happen. I’m shakinggggg, I think I’m gunna cry. THANK YOU (you are literally the best!!! the stats lecturer !!!! the best R explainer !!!! the best listener !!!! the absolute BEST!!!!) I wish I could give you hug!!!!! (IOU 1 HUG).

Yesterday, I had a technical interview where I was required to code a massive data set with 110,000 participants. I then had to report it and present it back in their board room to the CEO. Today… out of 5 finalist graduates… they chose me to be their new data analyst!!! I spoke to the CEO over the phone, and he commented that he couldn’t believe I had been coding for only a couple of years, and that my work was far more advanced than any other graduate he has had - and that’s because of you!! Taught by the best!

I didn’t realise that R was a part of my course. As a lazy student, I didn’t even read my prospectus before joining, so it was a complete (unpleasant) surprise that I was studying statistics and coding. During secondary school, I barely scraped a C in Higher Maths GCSE. As such, I was absolutely terrified to learn that I now had to learn coding and statistics … I now work as a Data Analyst in the R Programming Language. I love my job.

What do they say about my colleagues?

Role models: R-ladies¹

“… the influence of role models is strong among women who don’t study computing at college but then go on to pursue a computing career.”

What can you do within your company?

Being mindful of not adding disproportionately to the workload of women

• Recognise and celebrate women’s achievements/roles as much as mens
• Make sure people in junior roles are aware of women in senior roles
• Look outside of your company for activities and opportunities that support women

1. https://rladies.org/

Too long, didn’t listen

Ask yourself these questions

• Are there women in non-coding roles who I could support to try coding?
• Do our training materials motivate and engage women?
• Who are the role models inside and outside of my group?
• Am I doing enough to challenge stereotypes around coding?
• Can we do more outreach (STEM ambassadors)?