As we train the next generation of STEM professionals, we use a filter that selects against marginalized folks, on account of their ethnicity, income, gender, and other aspects of identity. This, I hope you realize, is an ethical and pragmatic problem, and constrains a national imperative to maintain competitiveness in STEM.
When we are working for equity, this usually involves working to remediate perceived deficiencies relative to the template of a well-prepared student — filling in gaps that naturally co-occur with the well-established inequalities that are not going away anytime soon. These efforts at mitigation are bound to come up short, as long as they’re based on our current Deficit Model of STEM Recruitment.
You might be familiar with the phrase “deficit model” in terms of science communication with the public. The “Information Deficit Model,” in a thimble, is about learning what the audience doesn’t know and then teaching stuff to fill in the gaps. When it comes to public outreach and informal science education, it doesn’t really work. Let me give an example.
Let’s say you meet someone who thinks global climate change isn’t a result of carbon pollution. You take a few minutes to explain the science, but out some graphs, and get them to read all of the evidence for anthropogenic climate change laid out crystal clear at the NASA Climate site. In the end, your pal won’t point out any technical errors, but also, won’t be convinced. It might seem bizarre, but you can give someone all of the facts in the world, and many folks won’t adjust their world view to accommodate these facts. The same story could be told for vaccines, evolution, the quality of education available in your public school district, and Russian manipulation of the White House. Even for more prosaic topics, like how osmosis works or what causes seasons, the deficit model doesn’t work that well in the long run.
For another example, here’s a story of mine about a guy who insisted that a snake was an ant. It’s funny, but also tragic. There’s no shortage of scholarship about how and why the deficit model is a recipe for failure in science communication.
We use another deficit model when we are working to diversify STEM, but this is something that doesn’t get much discussion.
When we recruit and select candidates for admission into graduate STEM programs, we look at what people have to offer, in terms of their talents, accomplishments, and potential. It’s standard to look at grades, scores, publications, and experiences. This approach, whether you look at it as a matter of addition or subtraction, involves identifying deficits and seeing whether holes need to be fixed and if we can fix them.
If a student is applying to your lab with a low-ish GPA and not as many publications as other students, you might think, “Well, there are legitimate reasons why their grades are lower than other students, and they haven’t had an opportunity to publish based on the opportunities available to them.” Which is a reasonable mindset. Some programs are committed to bringing in first-generation students and from underrepresented groups who have these “deficiencies,” though many are fighting over the small number of marginalized students who have managed to build a portfolio that rises above these structural disadvantages. When high-prestige institutions boast of their diversity, that’s typically a consequence of investment into finding and attracting people who fit their demographic desires but also lack the perceived quantitative and qualitative “deficits.”
This can fix the diversity problem for a single institution, but across the system, doesn’t move the needle whatsoever, it just takes these students away from other institutions that aren’t going to recruit them. The only way to change the system is to recruit students who, otherwise, would not be recruited. The underlying socioeconomic and cultural inequities that result in lower average metrics for members of underrepresented groups are not going to change overnight, so if we recruit the students from all groups who are most highest achieving — or least deficient — then we will continue to perpetuate our inequities.
The failure of this deficit model of STEM recruitment — rejecting individuals who fail to impress on the basis of their grades, scores, and such — is that there is no clear evidence that these metrics are cause success in STEM! Wherever you might see a correlation, there are more parsimonious causal explanations involving social capital.
The real drawback comes from our reliance on the Deficit Model of STEM Recruitment to prepare students from underrepresented groups to compete for space in the academy.
To avoid talking about this in abstract terms, let’s return briefly to the discussion about the lack of institutional diversity among the folks receiving NSF graduate fellowships (which I first discussed here). The upshot is that underrepresented minority (URM) students are disproportionately found in low-prestige regional public universities, such as my own, and that these students are getting an absurdly small piece of the pie. Instead, what we have is a larger scale version of the university-level phenomenon, of institutions supporting the fraction of URM students that have numbers on the par with their white competitors at higher-prestige institutions. Which won’t move the needle, because it’s won’t bring new people into the game who otherwise wouldn’t be in the game.
As I raised these concerns, all kinds of folks were like “whoah hold on wait a moment this is fair, it just reflects the composition of the applicant pool!” and I was like “Yeah, precisely, that’s the problem.”
At this moment, the standard approach to fixing this diversity problem is the Deficit Model — to fix the holes in the records of the applicants.
And so, I am told every day that my job is to fix holes in the applications of the students training in my lab, much like one would tell a CalTrans worker to fix potholes. I need to get my students to author first authored pubs so they can compete against undergrads from big labs where their names can readily get added to papers. I need to get my students tutoring to pull up their test scores and their grades so that they might have the same GPA and scores as well-heeled students who go to a school with far more resources, don’t work 30 hours per week, and come from a family that invested great resources into this sort of achievement for the past 20 years. I’m told I need to train my students how to act so that they fit in, so that other scientists can choose to view them as potential peers.
In other words, I need to fix my students so that they conform to the template that the system uses to identify future scientists. This template, though, is rigged against them.
This pretty much makes me sick, but yet, if I want my students to succeed, then this is what the system requires of me. If I’m training students to do great science, then that might mean that their grades would suffer. If I’m doing what it takes to get a students’ name on a paper, that might mean that I’m cutting corners on actual mentorship and developing the capacity of my students to work independently.
As a professor of URM students in an underfunded regional public university, the impossible is asked of me every day: I’m being asked to groom my students so that on paper, they look as if they are competitive to white students from wealthy universities on the basis of the template developed to evaluate these wealthy white students. What does success look like for me? When one of my amazing students squeaks through the system, given the benefit of the doubt, even though their numbers don’t match up. I’ve done pretty well at this, because I have no choice if I’m going to help my students, but the system as a whole is still failing my students.
Herein lies the problem with the Deficit Model of STEM Recruitment: No matter what kind of effort and resources are applied, there is no way that the average URM applicant from a regional public university is going to develop a record as competitive as the average non-URM candidate from a prestigious and well-funded small liberal arts college or major research institution.
As long as we compare URM and first-generation students with metrics that have been used to evaluate students who have been handed more advantages in life, then we are always going to perpetuate the inequities we are trying to eliminate. At best, we can shrink the gap. However, after any efforts are applied to shrink that gap, then the wealthy institutions will invest even more into increasing that gap.
Let’s take another quick look at the NSF graduate fellowships. It turns out that, this year, they did a pretty good job of awarding fellowships. Grad students in STEM are 11% URM, but 25% of the fellowships went to URM students. That’s pretty darn good, right? Well, yes, and no. Why do I not think this is 100% peachy? Because take a look at the undergraduate institutions of the students that are landing these fellowships. They are, overwhelmingly, coming from higher-prestige institutions that have relatively low rates of minority enrollment, compared to less-prestigious institutions with less funding. In other words, fellowships are going to minority students, but they’re going to students who had already cracked the code of making their way into prestigious academic institutions. Would these students be going to a strong graduate program without this fellowship? That’s an important question to ask, and I don’t know the answer. Regardless, the underclass of talented students who didn’t make their way into prestigious universities — which is an environment that has much higher URM enrollment — remains nearly shut out of this kind of federal support. And please don’t fool yourself into thinking that students enrolled at at UCLA or Harvard or Pomona College are there because they are smarter or better or work harder than the students at my university.
Even if we keep giving support to URM students from prestigious institutions, we’ll always be playing a game of catchup. But let’s not be silly, we know that the entitled will keep their privileged access to resources for the common good. When we rely on metrics that are designed to evaluate people who succeed in the very system that they’ve developed, we’re enabling majoritarian rule.
When we evaluate students who have experienced systemic disadvantages using the metrics of those who don’t experience these systemic disadvantages, there will always be a deficit that we cannot fill, no matter what handicap we try to apply.
Whatever the system invents or develops something to help out the students who experience systemic biases and disadvantages, students with more social capital and money will be able to take advantage of those developments as well. There is no way that prestigious institutions would going to sit idly by if folks from non-prestigious institutions would equal them in access to publicly funded resources. We will not ever see the day that high-potential STEM students from universities like mine will will be getting funded admission to graduate programs at the same rate as high-potential students from Caltech or Harvey Mudd. But if such a magical day would ever come to pass, it would be to the outrage of many people at these prestigious places because it would mean that their status would have lost its currency. (Even allowing in a slightly higher proportion of minorities at a place like Harvey Mudd is causing some faculty to go apoplectic and say that the “quality of the students” is declining. Imagine how these people would react if brown people at “low quality” institutions ever got access to the same resources as their own “high quality” students? It’s okay for the disadvantaged to get support, but the elite want to make sure their students always a couple tiers higher. Just take a look at how people allocate their charitable giving to universities to see how this dynamic plays out.)
How can we possibly identify people who are prepared to succeed without using metrics that are part of the deficit model? Instead of finding the people who are least lacking, we can identify the people who have the goods. We can do this by getting to know people as actual human beings. If a student spends a substantial amount of time working with you, then you’ll know whether they have what it takes to succeed. A recruitment weekend won’t be enough, but maybe a few weeks in the lab is. Maybe a whole summer. We have spent many years trying to fix the system so that we can give a fair chance to people who deserve one, and we still keep coming up short. If we just continue with these minor tweaks, we won’t be getting anywhere.
If you want to truly diversify, then we need to stop trying to fill in the holes based on perceived deficiencies. Instead, we need to focus on training complete scientists. We need to fundamentally change our mindset about what a successful student looks like in a way that doesn’t reflect systemic inequities — and then enact a training and recruitment agenda based on that mindset. We could continue investing our time and resources trying to get URM students to look more competitive against white students from private universities. But it’s doing a disservice to my students who want to be trained to become scientists, not to become what white people in charge think scientists should look like.
It’s dispiriting to see students consistently undervalued because of metrics that do not represent their worth or their capability. And the efforts to diversify STEM are kneecapped by disproportionate investment into people who came to the game with the benefit of higher resources. This isn’t because these more privileged students are prepared to become better scientists, but because the system is run by folks with the same background and are assessing potential in a way that students with fewer opportunities will always come up short.
I was chatting with a friend who was looking over the files of two undergraduate applicants. One went to a prestigious school, had multiple publications, did an REU in a prestigious faculty, and high test scores and grades. The other student was a first-gen URM student from a school that I hadn’t heard of, with had a strong letter of recommendation from a faculty member they worked with, but didn’t have the pubs and the fancy pedigree. How do you compare these two? That’s a challenge we have to meet, and it’s not an easy one. And ending that conversation with, “We have to be sure to be fair to the student who is working hard and doing very well, we can’t punish someone for being presented with opportunities” is not a mechanism for change.
You can’t measure what you can’t see, and when you are using these classic methods of assessing research potential, then you are literally overlooking a person because they hadn’t had the opportunity to excel. We already have so many Educational Opportunity Programs, funding programs that target URM students, and initiatives designed to prepare less advantaged students to become competitive. But no matter what we do, as long as we’re measuring students with this Deficit Model, there will always be this gap, because those with the advantages are always going to be leveraging those advantages to their opportunity. We’re working to shrink that gap, but others are working to expand that gap. And those expanding the gap will always have more resources, because it’s the resources themselves that generate the gap.
The deficit model of science communication shares a problem with the deficit model of STEM recruitment. No matter how hard to you try to fill in that gap, folks are going to make up their own minds using their own criteria. No matter how damn much you invest into filling in the “deficits” presented by URM students from a regional public university, they’re always going to get undervalued because the system is designed to value people who go through life with a greater set of opportunities.
Instead of doing everything we can to fill in the gaps, we need to find constructive forward-looking approaches, ones that will open the minds to welcome information that might otherwise go against the biases that we’ve grown up with. Using the deficit model buys into biases and reinforces them. It could be downright counterproductive.
For equity in STEM recruitment, we need to stop using the metrics that facilitate the systemic bias against marginalized students. What, precisely, could such a system of developing, recruiting, and evaluating scientists look like? I have a some ideas, and have written about them in some posts I’ve linked to here, but really, it’s not being done as far as I know, so I don’t know what it looks like. I realize it has to be radically different from what we’re doing now.
That’s a conversation that I’d like our scientific community to have in depth, because that’s the only way we can really move forward.