Five Biases Pushing Women out of STEM

March 28, 2015 10:43 AM

By now, we’ve all heard about the low numbers of American women in science, technology, engineering, and math (STEM). My own new research, co-authored with Kathrine W. Phillips and Erika V. Hall, indicates that bias, not pipeline issues or personal choices, pushes women out of science – and that bias plays out differently depending on a woman’s race or ethnicity.
posted by sciatrix (24 comments total)

What a novel methodology: simply asking people about their experiences.

One criticism I have is that since she is selecting on the dependent variable (women who are in science) she cannot make any inferences about why women leave science, since the people studied have not left science. Having said that, we have every reason to believe that the pressures to leave science that women who stay in science face will be stronger for those who leave science due to those pressures. Making that claim, however, will require more research.
posted by MisantropicPainforest at 11:00 AM on March 28, 2015 [3 favorites]

What a fantastic article! I absolutely love the way they incorporated intersectionality here - it's long overdue for this type of study. That figure about nearly 50% of Black and Latina female scientists having been mistaken for admin or janitorial staff at some point is absolutely heartbreaking. Agreed that this study doesn't speak to why women leave science in terms of causality, but they hit the nail on the head all the same - I've considered leaving science for most of these reasons at one point or another.

The first bias (having to prove it again and again) is so pervasive and so difficult to deal with. At least in my field, men are just assumed to have some baseline level of competence that is absolutely not granted to women. It's especially a problem with field work, but also with research overall. I have at least weekly conversations with female colleagues about how much more we feel we have to prove ourselves to get the same amount of respect our male colleagues get just for existing. It comes with a lot of second-guessing, like "maybe I'm just being too emotional about this, maybe I just can't compete, maybe it's just that I'm really not as good as that guy", etc etc. Even when we support each other and know that this is a thing that happens, it still undermines our confidence and makes it very tough to know when we're being treated poorly and when we really aren't measuring up.

I see it in the careers of female professors in my department too, so it's not just an early-career thing - they're tenured and they still constantly have to prove themselves anew to their male colleagues, especially if they have children, to keep from being shunted into maternal teaching/mentoring roles by default, getting stuck with all the administrative crap nobody wants to do, etc. Of course they're badass and overcome this stuff, but I don't want to have to be a superhero just to have a regular career.

Thanks so much for posting this!
posted by dialetheia at 11:10 AM on March 28, 2015 [21 favorites]

This is a great article!

If organizations are truly interested in retaining and advancing women, they will approach the issue of gender bias the same way they do other business issue: develop objective metrics and hold themselves to meeting them.

The reluctance to do so, even in soft or partial ways, speaks volumes.

At least in my field, men are just assumed to have some baseline level of competence that is absolutely not granted to women. It's especially a problem with field work, but also with research overall.

I see this all the time as well, and it must underlie why men dominate the prestigious, better paid parts of this work -- which of course requires field work and is full of subtle and non-subtle challenges to prove yourself.
posted by Dip Flash at 11:28 AM on March 28, 2015 [4 favorites]

This reminds me of a Sociological Images post about how students evaluate female instructors differently than male instructors. What strikes me particularly is how the authors found that men were twice as likely to be offered a position that required math than women--and how this overlaps with the belief that math requires "genuis," which women are not perceived to have nearly as often.

This hits home for me, because as an instructor most of the positive and negative feedback I've received has been about my helpfulness and presentation. Although, once I was criticized for "showing off" how much I know about the field I specialize in--and I honestly don't think I would have gotten that comment if I wasn't female.

tough to know when we're being treated poorly and when we really aren't measuring up

This is so insidious.

I'm lucky that I'm in a department that has a good gender balance and is more progressive/welcoming than many, and I still sometimes wonder, when someone decides to explain something to me that they really should expect me to know, whether it's just because they're explainers, or because I'm a woman.
posted by Kutsuwamushi at 11:32 AM on March 28, 2015 [8 favorites]

The tightrope is what I have the most problem with. I rarely have the imposter syndrome because my experience has shown me where I excel over others. So I come across as too direct. My last performance evaluation I was told I need to work on my communication. Not that I didn't communicate enough, but that some people felt my tone was too harsh.

My latest strategy is to act excited about everything. That improvement? I love it! Suggested new procedure? That's a great idea! I have never used an exclamation point so much in my life*. I also complain as little as possible and try to vent just to my partner.

It seems to be working. I will find out this week if I get a tech-heavy promotion.

*I hear Zinsser in my head: "The Exclamation Point... has a gushy aura, the breathless excitement of a debutante commenting on an event that was exciting to only her."
posted by Monday at 11:49 AM on March 28, 2015 [10 favorites]

I have to add this Bitch in Business thread from ask.mefi just in case anyone missed it. It's amazing.
posted by amanda at 11:58 AM on March 28, 2015 [8 favorites]

Of course they're badass and overcome this stuff, but I don't want to have to be a superhero just to have a regular career.

Yeah, I think this is an important point. For someone with the biases working against them, it takes a huge amount of effort just to maintain a perceived level of competence in colleagues' eyes. I posit that the more effort you have to put in just to be allowed to continue doing your current job (nevermind getting a promotion), the more likely you are to drop out (or be let go).

Whereas for people who have the biases working in their favour (e.g. white males), even the semi-incompetent ones can coast along up their career ladders without too much trouble. From what I've seen they have to have a pattern of making pretty egregious mistakes for people to even start wondering about their level of competence.
posted by mantecol at 12:10 PM on March 28, 2015 [16 favorites]

For me the worst thing is that my male colleagues don't talk to me. We're not a big place: 14 scientists. Yet years can go by without the other scientists talking to me. It's great that most of the technicians are women and will talk to me every day, sometimes more than once!
posted by acrasis at 12:18 PM on March 28, 2015 [9 favorites]

After all, the percentage of women in computer science has actually decreased since 1991.

I've read this stat before, and it depresses me every time. Thanks for posting this article. Every time someone posts something like this to MeFi, it gives me some more insight into challenges faced by women working in the technology field. As a manager in said field, it's another prompt for a bit of self-reflection. Am I causing any of these challenges? Am I failing to stamp them out when other people are causing them? Do I know any other people managers who need to read this too?
posted by FishBike at 12:34 PM on March 28, 2015 [1 favorite]

I think the issues with women having children are a symptom of a greater problem in the sciences; scientists are not supposed to have anything else in their lives but science. You are considered not-serious if you want to work a normal schedule, have regular hobbies on the side, go on vacations that aren't scientific conferences or field work, and so on. It's pretty twisted that this has apparently now extended to being not-serious if you attempt the transgression of having a family, but it doesn't surprise me at all. It's a natural extension of the concept.

I think in this case, it might be better to fight the root cause of the issue (the idea that scientists are supposed to sacrifice their entire personal lives to the job) rather than merely the gender-related component.
posted by Mitrovarr at 12:49 PM on March 28, 2015 [9 favorites]

I understand the author's reason for framing that way, but I'm curious what "pipeline issues" that aren't "bias" look like.
posted by PMdixon at 12:50 PM on March 28, 2015 [1 favorite]

I think in this case, it might be better to fight the root cause of the issue (the idea that scientists are supposed to sacrifice their entire personal lives to the job) rather than merely the gender-related component

Good thing we don't have to choose between the two!
posted by MisantropicPainforest at 1:15 PM on March 28, 2015 [13 favorites]

We had our departmental retreat this weekend, which among other things involves a day of research talks and a series of awards (monetary and not) given to the graduate students. I find myself somewhat frustrated that of the major awards, those for departmental service went to women (one being myself), while best research talk and best TA of the year both went to men.

What is maybe more frustrating is how grateful I am that at least the amount of adminstrative work I do for the department has finally been recognized in some way. I gave up a long time ago on having my research considered on equal terms.
posted by pemberkins at 10:45 PM on March 28, 2015 [2 favorites]

Mitrovarr- you're right. I didn't realize how much I had sacrificed to science until I sat down recently to figure out when I could retire and realized that, thanks to all the low-paying no-benefit post-docs I took to stay in the field, I can never retire. The problem is, my male colleagues only had to take one or two postdocs.
posted by acrasis at 7:55 AM on March 29, 2015 [3 favorites]

PMdixon: "I understand the author's reason for framing that way, but I'm curious what "pipeline issues" that aren't "bias" look like."

It's all bias, just a matter of when and where by whom and how much. The authors of the paper appear to prefer the 'industry is biased' model, and seem to dismiss pipeline arguments. It can be both, and this matters, because remedies at one stage of a pipeline can only be expected to stop the loss, not reverse it. For example, If STEM degree programs are driving women out, the only way you could reverse that damage in industry is by hiring people without degrees into the job. I've yet to see an academic paper suggest that we should stop requiring STEM degrees for STEM jobs.

Abraham Wald studied airplane bomber survivability with an eye to prioritizing plating. Being an intellectual not an airman, all Wald had to examine were the planes that return from bombing runs. He divided planes into sections, and for every returning plane counted bullet holes. A naive analysis might look at these planes and place additional armor where the guns appear to be focused. But Wald was smarter than that. He realized his data had a bias not mentioned in the article: survivor bias. For the planes returning, every bullet hole is additional evidence that the section was sufficiently strong. Instead, we needed to examine more closely the surfaces that appeared unscathed.

Survivor bias is a huge problem with these surveys. Whatever trauma and bullets they've received, by definition when you recruit from the for Women in the Sciences, your respondents women are in the sciences. They survived their metaphorical European bombing run. In a perverse way, these surveys offer support for the opposite of what the authors suggest: none of the complaints contained within were strong enough to dissuade these women.

But it could just be that they were personally resilient, and that people are more complicated than airplanes (likely!). So what we'd prefer is to go out and find women who didn't take the career path and find out why. Or perhaps even find people who dropped out of a STEM program, and ask them why. To get a large enough sample size, you'd probably need to hit up multiple universities, and dealing with multiple laywers, FERPA and IRBs seems unsurmountable.

The best trick I can think of is that women tend to have a fair number of women friends. Instead of asking Women in Science how terrible their experience was, we could ask them how many of their friends dropped out of STEM, and why.
posted by pwnguin at 1:01 PM on March 29, 2015 [7 favorites]

Okay, so the point of the 'pipeline' metaphor is really so those downstream segments can raise their hands and say "not my fault, it was like that when I got here," or am I being overly cynical?
posted by PMdixon at 7:01 PM on March 29, 2015

Yeah. I mean, "pipeline problems" are real. There's a large amount of research showing that girls start getting steered away from math and science beginning in elementary school, and at every stage thereafter. But you're correct that many people in industry and higher ed. use "the pipeline" as an excuse to ignore their own responsibility.
posted by mbrubeck at 10:10 PM on March 29, 2015

For example, If STEM degree programs are driving women out, the only way you could reverse that damage in industry is by hiring people without degrees into the job.

Or you could establish a fellowship or grant program for women in your relevant fields. Or you could establish a grant program for departments to put together programs that don't drive out women so much. Or you could work with departments two near your HQ or research facilities to put together pilot programs that don't drive out women so much.
posted by ROU_Xenophobe at 7:09 AM on March 30, 2015 [1 favorite]

ROU_Xeno, my position is that these programs encourage qualified candidates to stay in the pool, but don't really produce a larger pool of candidates. I don't think your suggestions help, mathematically. I'll use CS an example, because I know that department best. If 4 percent of women graduate in CS, and 3 percent of grad school applications are women, then to reverse the damage you need to raise applications to at least 5 percent.

Fellowships and grants seem like the sort of thing that could move the needle up, certainly. But in that example, you can at best entice more of the 4 percent to stay in. The timing of grants and fellowships are far too uncertain to expect any changes in behavior of undergraduates; no intelligent person will stay in a degree program they'd otherwise drop by assuming they'll win a fellowship. And there's a strong chance that you will not even accomplish your objective -- if you hand out fellowships to the top women applicants, you're going to have substantial overlap with the 3 percent who would have otherwise gotten paid GTA or GRA positions.

If you want 5 percent though, you really need the 4 percent in CS, plus an extra 1 percent form outside. However, those without a STEM background are going to be assigned an inordinate amount of deficiencies. To give you an idea, we have an online post-bacc program in CS. We get a lot of people who matriculated in something and decided they needed a job. It takes two years of nothing but CS classes full time to graduate. That's basically the size of the deficiency a non-STEM major would need.

All this says nothing about the fact that 5 percent is a shitty goal, and 50 percent is the number to reach. Whatever we do, we need to do it like ten times harder. It seems reasonable that if you want to raise grad school admissions, undergraduate ratios, freshman applications, and high school preparedness all factor in.
posted by pwnguin at 9:01 AM on March 30, 2015

So you're saying that when you have a problem, you should only consider solutions that would each individually solve the entire problem more-or-less instantaneously?
posted by ROU_Xenophobe at 9:28 AM on March 30, 2015

Or, more to the point, industry can do things to raise grad school admission, and do things to improve undergraduate ratios, and do things to improve freshman applications, and do things to improve high school preparedness, and do things to improve earlier preparedness, and do things to improve earlier retention of girls.
posted by ROU_Xenophobe at 9:29 AM on March 30, 2015

Hey look, a pipeline!
posted by pwnguin at 11:03 AM on March 30, 2015

The article is about what happens once women have made it through the every narrowing pipeline. Basically, once women have made it through the "hard part" which pushes many women away, what is that factor that pushes them right out? You basically have a triangle-shaped pipeline. It's frankly ridiculous.
posted by amanda at 1:11 PM on March 30, 2015 [1 favorite]

Survivor bias is a huge problem with these surveys.

No it is not. Its a problem if you are making inferences about a population from a sample that is not representative of the population. But only selecting women who stay in science is not a problem if you are interested in making inferences about women who stay in science.
posted by MisantropicPainforest at 1:42 PM on March 30, 2015 [3 favorites]