In 2018, a conflict erupted at Google when a male software engineer proclaimed that women were less suited to working in technology. Many people came to his defense by citing a study from earlier that year that purported to find that, in countries with higher gender equality, fewer women worked in Science/Technology/Engineering/Math (STEM) fields (it says nothing about non-binary people). On its face, the study supported those who argue against diversity initiatives: when given a choice, it appears that women will less often work in STEM.
Don’t fall for it.
Here are some plots included in the infamous paper. The one on the left shows the relationship between gender equality as measured by the Global Gender Gap Index (GGGI) and the gender gap in performance on science exams between individuals, illustrating that countries with more equal genders had larger gender gaps in school. On the right they plot GGGI against percent of women graduating with STEM degrees, which shows a lower percentage of such graduates in more equal countries.
Similarly, they show here that countries that are more equal according to GGGI have a stronger bias toward boys among student ratings of science self-efficacy, which measures the extent to which students believe they are good at science.
In short, the authors of this study argue that general measures of gender equality and specific metrics of science ability and interest are inversely related. That is, if women are more equal to men in a given country, they think they are worse at science and demonstrate less interest in STEM. This would be a very provocative finding, given widespread efforts to increase the number of women in STEM fields, and since its publication has been used by opponents of such efforts. It seems to support their arguments. But does it?
First and most simply, consider that the gender equity metric they use may not measure what they claim it does. GGGI measures “gender-based gaps in access to resources and opportunities” and it may not be the case that highly developed countries should have higher scores. While everyone in wealthier countries has better access to resources overall, wealth may not necessarily correspond to such access on a gendered basis. Put more simply: rising tides do not always lift all boats.
The GGGI includes:
- Labour force participation by gender
- Wage equality for similar work
- Estimated earned income by gender
- Representation of women among legislators, senior officials, managers, professional and technical workers
- Literacy rate by gender
- Enrollment in primary, secondary, and tertiary education
- Sex ratio at birth and healthy life expectancy by gender
- Women in parliament, ministerial positions, and as heads of state
These are all very powerful metrics that establish how good it is to be a woman in a given country, but they barely touch on STEM and say nothing at all about perceptions. What do people in each country think is the appropriate role of women? How do women and men disagree or agree about this? How free do women feel to pursue their interests? What kinds of support exist for women to have families? How do religious expectations interplay with cultural expectations? Which career fields are growing fast in a given region? What are the barriers to entry in different careers?
In addition to all of these unanswered questions, we can pick apart a variety of variables that potentially confound the apparent pattern in this graph.
What percent of all graduates work in STEM? Maybe it’s higher for both genders in the countries with low gender parity than high gender parity, making it appear as if women tend to work in STEM when really it’s just that everyone works in STEM. (We know this is the case in at least some of the counties included on the chart). (5)
What percent of the population gets degrees? Maybe in some of these countries it’s a very small number, which could distort results.
How about the relative economic viability of careers? Perhaps technical, high-paying jobs are the only ones for which college makes sense in a given country, distorting STEM degree attainment. Maybe there is an assumption that for other careers, you don’t need college.
And what about barriers to college for women? Maybe in countries with higher rates of female participation in STEM than expected, it’s very hard for women to attend college, so the only degrees pursued are those that are absolutely required for their desired career. If an aspiring writer doesn’t need a degree, she won’t get one, but to be an engineer, perhaps she has to.
What kind of support is there for STEM learning earlier in education? Maybe all secondary school graduates in a country with a newer school system graduate more qualified for STEM degrees, or maybe some countries have an earlier funnel that filters people out based not on interest or skills but something else (more on this later).
And then there are issues with the graphs themselves. They appear to be missing data, as each plot is missing some countries present in the others. Why were these countries left out? Of those that are included, there is a clear clustering pattern, hinting at regional variation that suggests Scandinavia and the Middle East may be outliers. Very little correlation remains with these removed.
An explanation of why the Nordic model may be an outlier has to do with a paradox not in the behavior of Scandinavian women but rather in their tax system. The tax model was constructed for couples in which one partner works full-time and the other part-time, and due to historic gender role expectations, the woman was often the one working part-time. When modeled by parents, such patterns continue into the next generation(1,2). Making this even worse, fields such as education, health, and elderly care that were historically filled with women became public monopolies, complete with a wage-setting system that limits pay in these careers. This trend then spreads throughout the economy, in which people see women earning less and assume that there is a reason for it(2). Swedish researchers have found that the glass ceiling is “more severe in the Scandinavian countries with their generous family policies than in…other comparable countries”; childcare and maternity leave policies make it a more reasonable choice to reduce work hours after having a child, but taking time out of her career eliminates a woman from the competition for top jobs later on. In short, the welfare state that makes these countries rank so highly for gender parity can actually hinder women in their careers.
All of this is a great start toward debunking this study, but it does not actually show true errors, just suggest possible problems. Let’s get more specific.
To start with, we can take the data and analysis of this study and see how it doesn’t fully support its own conclusions. They find:
- girls match or outperform boys in science in 2/3 of countries
- more girls test as capable of STEM degrees than enroll in almost all countries
- if rational preferences led to level of STEM participation, women should earn 34 percent of STEM degrees, versus the actual global average of 28 percent
Even if you accept the premise that interest varies how the study authors say it does, there are still fewer women in STEM than we expect.
Next, we can see that in all countries except Romania and Lebanon(3), boys’ best average subject was science and girls’ was reading, even though the average girl was just as good at science as the average boy (in other words, she was even better at reading). If an academic system funnels or even just encourages students based on their best subject, this will have the effect of steering girls away from science.
Anecdotally, this almost happened to me: during the transition to middle school, when selecting advanced “honors” courses, my guidance counselor advised students not to take both honors English and honors math, and since I had the highest reading scores in my sixth grade class, she was going to enroll me in honors English. Luckily my parents overruled her, because I excelled at math as well — just not to the same extent as I did in reading. What if I had attended a school where you had to choose an exclusive track at that point? I very well might not have a STEM career now. This is the case across much of Europe. (4)
Also important to note is that interest as expressed by life path does not necessarily correlate to actual interest. As it relates to gender in various careers:
- the number of women already in a field predicts the stereotypes people have about that field
- perceived sexism in a specific degree program predicts the expressed interest in the field by girls (1,2)
- parents are less likely to think their daughters will be interested in STEM fields, which predicts later attitudes and efficacy at science among those daughters (1)
In short, gender differences in fields and among adults lead girls who would otherwise be interested in STEM to abandon such interest. This is a negative reinforcing cycle that keeps girls out of STEM.
Why would this be less severe in some countries than others, particularly those that have greater gender difference overall? Perhaps these countries also have less history with traditional STEM careers, so the examples and stereotypes are less prevalent. Or, conversely, perhaps all careers have strong stereotypes that drive women away from them, so there is no relative pressure on STEM fields. Perhaps they have simply lucked into surpassing the necessary ratio to eliminate the aforementioned parent-stereotype and number-of-women factors, leading to a reinforcing cycle of improvement of the number of women in the field.
Finally, as Scientific American very clearly explains, “rather than reveal a paradox, this research helpfully illustrates that the barriers that keep women out of government or the boardroom may not be the same barriers that keep them out of science” (1).
Even in the US, shown as relatively equal:
- women get less credit than men for the same math performance
- female STEM majors are less likely to get a response from potential PhD advisors (1)
- professors and managers are less likely to interview candidates with identical resumes if the name attached is female(1)
- those women who are hired despite all of this are paid less
All of these biases are even more extreme for non-white and LGBTQ women(1).
Women in Western countries do not express their interest in STEM freely.
I’ve poked holes in the paradox and offered some alternative explanations, but this trend does exist in some form in some data. How, then, can we explain this measured paradox?
The Implicit Association Test (IAT) measures how tightly two concepts are tied together in people’s minds. An analysis of more than 500,000 responses to a version of the IAT of associations between men or women and STEM, with results from 34 countries, found that people associate science with men rather than women, but also found different trends than the GGGI would predict. Sweden is presented as more gender equal than the US by the GGGI, but the IAT shows the opposite, and the country with the least-biased scores was Jordan(5). The Netherlands scores high on composite indices of gender equality, but has the strongest explicit and second strongest implicit gender-science stereotypes, which corresponds to a gender imbalance with roughly 5x as many male scientists as female scientists in both employment and education. This study also found that metrics like GGGI could not explain gender-science stereotypes and including GGGI as a (potentially) confounding variable in their analysis “did not nullify relationships between women’s representation in science and these stereotypes.” In other words, the implicit and explicit stereotype metrics for a nation predict female participation in STEM more strongly than national gender equity does, even when national gender equity would otherwise suggest an opposite trend.
Chances are, at some point you’ll find yourself dating or befriending or working with a woman who works in STEM in some capacity, or even who has just a casual interest.
If you’re thinking in the back of your mind that her interest isn’t likely to be as sincere as a man’s, you’re creating an invisible barrier in your interaction with her.
So don’t buy into the propaganda! Women, like all people, can be interested in anything, and anyone who argues otherwise is trying to make the complex world simple in order to meet their own agenda.
- https://www.pnas.org/content/111/12/4403
- https://blogs.scientificamerican.com/voices/countries-with-less-gender-equity-have-more-women-in-stem-huh/
- https://www.theatlantic.com/science/archive/2018/02/the-more-gender-equality-the-fewer-women-in-stem/553592/
- https://www.ncbi.nlm.nih.gov/pubmed/25867167
- https://capx.co/what-jordan-peterson-gets-wrong-about-the-nordic-gender-paradox/