What is not but could be if

And if I can remain there I will say – Baby Dee

Obviously this is a blog that love the tabloids. But as we all know, the best stories are the ones that confirm your own prior beliefs (because those must be true).  So I’m focussing on  this article in Science that talks about how STEM undergraduate programmes in the US lose gay and bisexual students.  This leaky pipeline narrative (that diversity is smaller the further you go in a field because minorities drop out earlier) is pretty common when you talk about diversity in STEM. But this article says that there are now numbers! So let’s have a look…

And when you’re up there in the cold, hopin’ that your knot will hold and swingin’ in the snow…

From the article:

The new study looked at a 2015 survey of 4162 college seniors at 78 U.S. institutions, roughly 8% of whom identified as LGBQ (the study focused on sexual identity and did not consider transgender status). All of the students had declared an intention to major in STEM 4 years earlier. Overall, 71% of heterosexual students and 64% of LGBQ students stayed in STEM. But looking at men and women separately uncovered more complexity. After controlling for things like high school grades and participation in undergraduate research, the study revealed that heterosexual men were 17% more likely to stay in STEM than their LGBQ male counterparts. The reverse was true for women: LGBQ women were 18% more likely than heterosexual women to stay in STEM.

Ok. There’s a lot going on here. First things first, let’s say a big hello to Simpson’s paradox! Although LGBQ people have a lower attainment rate in STEM, it’s driven by men going down and women going up. I think the thing that we can read straight off this is that there are “base rate” problems happening all over the place. (Note that the effect is similar across the two groups and in opposite directions, yet the combined total is fairly strongly aligned with the male effect.) We are also talking about a drop out of around 120 of the 333 LGBQ students in the survey. So the estimate will be noisy.

I’m less worried about forking paths–I don’t think it’s unreasonable to expect the experience to differ across gender. Why? Well there is a well known problem with gender diversity in STEM.  Given that gay women are potentially affected by two different leaky pipelines, it sort of makes sense that the interaction between gender and LGBQ status would be important.

The actual article does better–it’s all done with multilevel logistic regression, which seems like an appropriate tool. There are p-values everywhere, but that’s just life. I struggled from the paper to work out exactly what the model was (sometimes my eyes just glaze over…), but it seems to have been done fairly well.

As with anything however (see also Gayface), the study is only as generalizable as the data set. The survey seems fairly large, but I’d worry about non-response. And, if I’m honest with you, me at 18 would’ve filled out that survey as straight, so there are also some problems there.

My father’s affection for his crowbar collection was Freudian to say the least

So a very shallow read of the paper makes it seems like the stats is good enough. But what if it’s not? Does that really matter?

This is one of those effects that’s anecdotally expected to be true. But more importantly, a lot of the proposed fixes are the types of low-cost interventions that don’t really need to work very well to be “value for money”.

For instance, it’s suggested that STEM departments work to make LGBT+ visibility more prominent (have visible, active inclusion policies). They suggest that people teaching pay attention to diversity in their teaching material.

The common suggestion for the last point is to pay special attention to work by women and under-represented groups in your teaching. This is never a bad thing, but if you’re teaching something very old (like the central limit theorem or differentiation), there’s only so much you can do. The thing that we all have a lot more control over is our examples and exercises. It is a no-cost activity to replace, for example, “Bob and Alice” with “Barbra and Alice” or “Bob and Alex”.

This type of low-impact diversity work signals to students that they are in a welcoming environment. Sometimes this is enough.

A similar example (but further up the pipeline) is that when you’re interviewing PhD students, postdocs, researchers, or faculty, don’t ask the men if they have a wife. Swapping to a gender neutral catch-all (partner) is super-easy. Moreover, it doesn’t force a person who is not in an opposite gender relationship to throw themselves a little pride parade (or, worse, to let the assumption fly because they’re uncertain if the mini-pride parade is a good idea in this context). Partner is a gender-neutral term. They is a gender-neutral pronoun. They’re not hard to use.

These environmental changes are important. In the end, if you value science you need to value diversity. Losing women, racial and ethnic minorities, LGBT+ people, disabled people, and other minorities really means that you are making your talent pool more shallow. A deeper pool leads to better science and creating a welcoming, positive environment is a serious step towards deepening the pool.

In defence of half-arsed activism

Making a welcoming environment doesn’t fix STEM’s diversity problem. There is a lot more work to be done. Moreover, the ideas in the paragraph above may do very little to improve the problem. They are also fairly quiet solutions–no one knows you’re doing these things on purpose. That is, they are half-arsed activism.

The thing is, as much as it’s lovely to have someone loudly on my side when I need it, I mostly just want to feel welcome where I am. So this type of work is actually really important. No one will ever give you a medal, but that doesn’t make it less appreciated.

The other thing to remember is that sometimes half-arsed activism is all that’s left to you. If you’re a student, or a TA, or a colleague, you can’t singlehandedly change your work environment. More than that, if a well-intentioned-but-loud intervention isn’t carefully thought through it may well make things worse. (For example, a proposal at a previous workplace to ensure that all female students (about 400 of them) have a female faculty mentor (about 7 of them) would’ve put a completely infeasible burden on the female faculty members.)

So don’t discount low-key, low-cost, potentially high-value interventions. They may not make things perfect, but they can make things better and maybe even “good enough”.

22 thoughts on “What is not but could be if

  1. > First things first, let’s say a big hello to Simpson’s paradox! Although LGBQ people have a lower attainment rate in STEM, it’s driven by men going down and women going up.

    Where do you see Simpson’s paradox? That would apply if the aggregate value was going down while the reverse was true for *both* subgroups, right?

    • The difference between straight and LGBQ groups are reversed between men and women. Everyone goes down (college students change majors), so it’s really just a matter of how much.

      • How is that related to Simpson’s paradox? Again, Simpson’s paradox would be having a common trend for both men and women, and having this trend reversed in the aggregate group.

        • To be more specific: Simpson’s paradox would be present in a situation like the following:
          – let the drop-out rate be lower for GM (3/5) than for SM (3/4) and lower for GW (1/5) than for SW (1/4).
          – let the population consist of 15 GM, 80 SM, 5 GW, 100 SW
          – the proportion of drop-outs will (9+1)/(15+5)=50% for the G aggregate, higher than (60+25)/(80+100)=47.2% for the S aggregate
          – G > S *but* GM < SM *and* GW < SW

        • Ok cool. I was working off the slightly different definition that the trend is different when the data is disaggregated. But yes, if it’s supposed to be opposite, then that’s not true in this case. The disaggregated trend is patly conflicting with the aggregated trend.

  2. “In the end, if you value science you need to value diversity”. As long as academic merit is not thrown under a bus, to ensure diversity I agree. But to be honest I am not sure if I should care about relatively small differences, that are likely noisy anyway.

  3. In the letiguous environment that we inhabit, our legal counsel warned us not to ask questions that coiuld be used by an applicant against us. Seemingly, harmless questions like “do you have a partner?” are particularly taboo.

    • Possibly because there is a long and well-understood practice of using the “do you have a husband” question as an excuse to not hire women, on the argument that a woman with a husband will subordinate her work life to her family life in a way that a man would not?

      And I mean really, what value is there is knowing if a potential future colleague has a spouse? If you are concerned about location issues, you can ask that directly… “do you think you would be happy living here?” is a fine question.

      I’m not saying that the only reason you might inquire about a potential colleagues’ spouse is for sexist reasons… I like to ask about my colleagues’ families from time to time. But even I recognize the history of the question in an interview setting, and so you just wait a little while until you’ve hired someone to get to know about their personal lives, as much as they choose to reveal that once in the security of their job.

      So anyway, I don’t see that as an example of a “harmless question” unless you are pretending that the history of academic employment discrimination didn’t happen or doesn’t matter any more. If that is the case, I strongly suggest reading Vivian Gornick’s “Women in Science”. Or even just asking any female international students you know about their experiences interviewing at non-US institutions where there are no rules in place to protect against sexist hiring practices.

      https://www.amazon.com/Women-Science-Then-Vivian-Gornick/dp/1558615873

      *Thread-Jack Note: that book is a really good read, in the sense of getting a chance to listen to smart women talk about their experiences in academia (not in the sense of making you feel great).

      • Did you really intend your comment to appear as a response to RS?
        After all, he is talking about a question being ‘seemingly harmless’ (‘seemingly’ normally being
        something you sometimes add specifically to imply “but no, it actually isn’t”). And his comment is about the _partner_
        question (the question the lead article more-or-less – although wrongly – endorses as an alternative to the husband question).

        • I guess I might have read it wrong. Although I think there is a difference between “seemingly harmless” and “seemingly, harmless”. But you might be right that I misinterpreted Rodney’s meaning when I read that as belittling the problem of asking candidates about partners, I’m really not sure now.

          Separately, I agree with you that as an example of good interview etiquette, suggesting we not gender the partner is probably missing the forest for the trees.

        • Sorry, the comma looked like noise to me but I now see that it’s significant.

          An (overly) optimistic view of the world, based on my (one person’s) corporate lifetime… The lesson that we should not ask such things (husbands, child-rearing plans, etc) seems to be consistently given and quickly accepted; there are hard discrimination problems but in the scheme of things this doesn’t (my experience, my industries) seem to be one. I genuinely wonder how much of a problem this (and I mean specifically: such questions, not anything broader) is today – at least in in larger, more professional, environments. I guess I’m naive but this struck me as entirely valid but somewhat antiquated advice. I’m shocked to hear that in (presumed: U.S.) academia it might not be.

          But to your forest/trees point, Dan’s discussion here seems weird. He sees the first problem but then jumps to offer an almost-as-bad alternative. (It really is 95% as bad – you get LGBT+ adverse implications partially mitigated, but leave everything else.) That’s a funny middle ground.

  4. Fine as far as it goes. But I’d really like to see the evidence about the broader issue.

    What’s the evidence that STEM has a diversity *problem*, as opposed to just unequal proportion? Is anybody familiar with whatever data there is?

    What’s the evidence that diversity of race and gender are beneficial to science *as* science? For all I know, it might make no difference at all to the knowledge generation process. For all I know, men might do better science than women, and Asians better than Whites, or any other combination.

    “…if you value science you need to value diversity. Losing women, racial and ethnic minorities, LGBT+ people, disabled people, and other minorities really means that you are making your talent pool more shallow”.

    Well, do you know? Yes, scientists as human beings should be welcoming and not discriminate by gender or race, but do we really know that this kind of diversity is important to science *as* science?

    And then the biased metaphors:
    – “Losing women…”: why is it *losing*? Maybe they’re *won over* to a cause that suits them better? Maybe they *escaped* STEM. Maybe they were *saved* from STEM?
    – “Leaky pipeline”: a pipeline is a structure whose function it is to transport something from A to B. And a leak causes a dysfunction. So are we to assume that a career in STEM has the function of transporting e.g. women from student to professorship, and that if a woman turns away from it, the system (or the woman?) has malfunctioned?

    • I think these are good points. Recently a friend posted a link to this article, which mainly was a link to the graph: https://www.nytimes.com/2018/02/05/upshot/even-in-family-friendly-scandinavia-mothers-are-paid-less.html

      what you see is that after birth of first child, women’s earnings on average drop 20% and never recover compared to women who didn’t have children.

      I pointed out that this was most likely actually caused by 20% of women leaving their job and not returning, and that 20% of women make less than the 20th percentile of income, and that you get to keep 100% of what you “produce” as a home worker, whereas after taxes especially in Scandinavia you only keep maybe around 50%. So my conclusion was that this story was at least potentially consistent with 20% of women finding that after children, they have a *better real income career* as a home care person and choose that route. For these people, their *real income* went UP compared to what it would be if they stayed in the labor market.

      This was of course not popular with all the highly educated career women who were discussing the graph. I think the take home message is that *people really really don’t get the concept of real income* and “if it isn’t measured in dollars it must be worthless” is a powerful but unfortunate heuristic.

      All that being said, there is as jrc says above, a long history of discrimination against women in hiring and in performance and compensation in jobs. The discrimination has definitely changed a lot over the last 40 years, and the types of jobs that women are doing has changed a lot over the last 40 years, and the experience that say 20 year olds have today straight out of college is very different than the experience their 70 year old grandmothers had in 1970. As Andrew likes to say, complete pooling of all of history over a period where there are strong time trends gives you a much bigger source of bias than is usually acknowledged. He usually mentions this in the context of the fetish for “unbiased estimates” in Econ, but it’s just as important to point out in the context of “maybe we shouldn’t be so naive about the mechanisms of things changing a lot in time”

      • The point about men’s results being different just points to really a general comparative advantage, cultural factors, and in some cases the fact that men are often slightly older than their spouses, and therefore slightly farther along in their careers and hence make slightly more even if you have total parity of treatment between women. Therefore, if the option is for one spouse to stay home, the choice will usually be the one who makes less money, and even a tiny but consistent edge for men (say 2%) would make it overwhelmingly more likely for women to be the ones who choose to take that route.

        In other words, *even in the absence of any discrimination* you might get the same result. Of course with discrimination you’ll get even stronger asymmetry.

        • Also, nice to see that the economists doing this research actually came to a similar conclusion:

          quoting from this alternative article about the same study: https://www.vox.com/2018/2/19/17018380/gender-wage-gap-childcare-penalty

          “What our evidence shows is that a lot of gender inequality is associated with choices that suggest different preferences,” Kleven says. “The holy grail is understanding whether those preferences are social norms, or something more intrinsic.”

          To bring things back to the original topic: certainly making the STEM workplace less gender biased and less full of assholes will change the preferences. But I do agree with Alex Gamma in saying that there’s a strong assumption in the conversation that STEM, particularly academic STEM is somehow obviously a good thing, and so of course we don’t use things like his terminology

          Maybe they *escaped* STEM. Maybe they were *saved* from STEM?

          But a lot of evidence out there is that today academic STEM is a cesspool of rent-seeking and NHST and that is true even in cancer biology and pharma and LHC and whatever not just psych and econ and social sciences that this blog focuses more on. Women may be much more perceptive of that fact on average, and may put up with it less, thereby *escaping* more frequently.

          I don’t honestly have a strong opinion on what the mechanisms are throughout all of this, I just know that *MECHANISM BLINDNESS* is getting us nowhere fast.

    • There are a couple issues I have with the argument that “maybe diversity isn’t *really* a problem”:

      1) Most social categories (e.g., ethnicities, genders, relative ability) are, well, socially constructed. They are contextual, have histories, and are under constant redefinition within all/parts of society. Thus almost prima facie asserting that the distribution of people with “innate science skills” just so happens to match specific contextual and historical patterns of access to power in specific falls apart. The latter category is a moving target, how could it always be right?

      2) Even for those categories that do have some physical basis (e.g., chromosomal sex, genital morphology, sensory sensitivity), I don’t see a reason why something like (scientific?) curiosity would be over- or under-represented.

      I’d put a lot of prior weight on scientific reasoning skills (and literacy, and numeracy, etc.) being more related to training/education and, I guess something like “freedom to pursue” scientific endeavors. I don’t have any literature on that (not my field – I’m an archaeologist), though, so if there has been research into the variable capacity of people to reason then I can update that. Scientific curiosity, though, the ability to come up with interesting questions about the world around us, do you see that as varying across populations?

      It seems to me that it’s more likely that we’re looking at differential access to power, decision-making, and intellectual freedom (something akin to the visibility argument Dan makes) rather than differences in capacity.

      Maybe at the level of faculty-job or graduate school applicants you might say that you’re reflecting the demographics of the applicants, but that’s why Dan’s arguments about making interventions at earlier levels is important. Ideally, we want people leaving STEM because they’re interested in something else, not because they don’t see themselves being able to succeed in science.

      Thus the argument that Dan makes: the deeper your pool of scientists, the more frequently rare events (like reproducible scientific breakthroughs or generally new ideas) occur, and that’s leaving aside added benefits from network effects / emergence. The best candidates in a large pool are more likely to be better than the best candidates in a smaller pool, unless as you say there are underlying distribution issues.

      But I’d argue that even if there are underlying distributional issues that relate to scientific curiosity and capacity for learning scientific reasoning, the odds that that distribution matches whatever definition of privileged social categories we make today is vanishingly small and temporally/contextually unstable.

  5. “These environmental changes are important. In the end, if you value science you need to value diversity. Losing women, racial and ethnic minorities, LGBT+ people, disabled people, and other minorities really means that you are making your talent pool more shallow.”

    This seems like a self-evidently bad argument, since it doesn’t take into account that such ‘inclusive’ policies (and lets be honest, in practice this almost always means reverse-discrimination) will drive away talented people who don’t want to work in environments where meritocracy is not paramount, and where everything is political. And its quite likely that these are the people who would be doing the best science in the first place.

    You make it sound like forced ‘diversity’ is simply a way to expand the talent pool without taking into account the second order effects of driving away people who might otherwise have been interested in those jobs.

  6. This seems like a self-evidently bad argument:

    1. Deepening the pool does not, by itself, imply any preference in hiring for any type of person.

    2. Thus there is no obvious second-order knock-on effect on academic choices for those with preferences for meritocracy, since the original suggestion made no claim on the nature of the actual hiring decision.

    3. You get from (1) to (2) by inventing an assumption/argument that “deepening the pool” implies “hiring less good candidates” based on, I assume, some sociological perspective about how “diversity” measures are implemented in scientific research organizations. But you do not actually state this part of the argument or even acknowledge its need for your broader argument’s logical coherence.

    4. You then argue that driving meritocratic-preference-havers from academia means driving away the most productive researchers. But this part of the argument relies on another unstated and not obviously true sub-argument, which probably proceeds along the lines of assuming there exists a correlation between preferences for meritocracy, on the one hand, and researcher (potential) quality on the other, along with an assumption that the meritocratic-preference-havers who are actually driven from choosing academia are the specific ones who would have been more productive than those who remain.

    5. Somehow you do still manage to state explicitly that your concerns are about second-order effects, without ever discussing their potential magnitude relative to any first-order effects. This makes your argument appear to assign zero value to any potential first order effects (for instance, any benefit of having a larger applicant pool). That is a weakness in your argument, given that having more applicants should, prima facie, to lead to weakly-improved applicant quality (weakly meaning greater than or equal to zero, because there would be more candidates so you can’t do worse). Your “second order effect” would have to dominate this primary effect for your argument to matter, but you make zero attempt to justify that and your own rhetoric (second order) works against it.

    None of my arguments provide evidence that your argument is wrong. But your argument, as presented, is incredibly lazy. It skips over precisely the important parts of the broader argument so as to, it appears to me at least, score cheap points without doing the necessary intellectual labor or without incurring the psychic costs of actually being clear about the chain of reasoning.

  7. I don’t think anyone who really studies the issue thinks that “leaky pipeline” is a good description of what happens. People move in and out of majors, from math to CS to econ or from biological sciences to public health etc. Or from nursing into biological sciences. And then they have all kinds of complicated career paths after college. Nowadays you will hear more about the tributary or watershed model that has many different ways of coming into a STEM career. People focus a lot on the outliers (people who knew they would be math majors when they were in 9th grade and continued through to a doctoral program) rather than what happens most of the time.

Leave a Reply

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