The so-called “demarcation question” – separating science from not-science – is an old and thorny one. Fields (often) want to be labeled as scientific because science is afforded a certain kind of cultural authority along with material resources (read: big grant money). Sometimes this question is about creating boundaries between science and pseudo-science (debating say ESP or homeopathy). Other times, it’s about sectioning off science from the humanities, into “two cultures” that are understood as distinct and even oppositional. In the classic “two cultures” division of academia though, the social sciences are a bit of an outlier. Are economics, psychology, sociology, anthropology, geography, anthropology, and history sciences? Are some of them sciences and not others? What criteria would we use to demarcate the scientific ones from the not-science, whatever we decide to call it?
Well, apparently the answer is that a field counts as science if men do it.
In a new paper, discussed by the second author in a post on the Brookings website, economists Diyi Li and Cory Koedel examine gender and racial representation gaps at top public universities in six fields: “biology, chemistry, economics, educational leadership and policy, English, and sociology” which they claim were “selected to be inclusive of science, technology, engineering, and mathematics (STEM) and non-STEM fields.” They classify the first three – biology, chemistry and economics – as STEM, and the last three as not-STEM. So, they categorize economics in the STEM side and sociology in the not-STEM side. What justifies this division? The authors explain in footnote 2 (emphasis added):
We classify economics as a science, technology, engineering, and mathematics (STEM) field, as in Nelson and Brammer (2010). Although the classification of economics may not seem as obvious as the other fields, economics is a mathematically oriented field, and empirically it has more in common with STEM than non-STEM fields in terms of student and faculty composition. For example, Arcidiacono and Koedel (2014) find that incoming economics majors are similar to other STEM entrants in terms of pre-entry qualifications. In addition, Butcher, McEwan, and Weerapana (2014) and Koedel (2011) document that the grade distributions in economics courses align with grade distributions in other STEM fields and differ from grade distributions in non-STEM fields. Data from the 2016 Science and Engineering Indicators report published by the National Science Foundation (National Science Board, 2016) also show that enrollment patterns in economics share key features with other STEM fields that are not shared by non-STEM fields, most notably, a sharp increase in foreign-student enrollment and a decline in the proportion of bachelor’s degrees awarded to women (see chap. 2 in particular). Ceci, Ginther, Kahn, and Williams (2014) also illustrate the similarity of economics with other STEM fields.
That is, economics is defined as a STEM field because it is dominated by men (and because it has low representation of Black and Latinx faculty relative to White and foreign faculty).
There are so many problems with this, and sociology twitter went to work naming them. “Sampling on the dependent variable” is a clear one as Peter Levin notes. By defining STEM fields in terms of their gender representation, you make it harder to find variation within STEM (or not STEM). Clayton Childress linked to this chart from an old Kieran Healy post:
Clayton notes that if you instead sampled Psych and Molecular Bio as your STEM fields and Philosophy and Religion as your non-STEM field, you’d suddenly be arguing that the humanities need to look to the sciences for guidance on recruiting, retaining, and promoting women! [Note biology in general is an included field.]
By claiming that the six fields that Li and Koedel selected are representative of STEM vs. non-STEM, the authors erase important variation and downplay field-specific histories and cultures as important causes of gender and racial inequality production. Koedel ends the Brookings piece by calling for policies focused on STEM in particular:
Earnings differences between individuals with STEM and non-STEM degrees, combined with the disparate diversity conditions between STEM and non-STEM fields among faculty, point toward policies that specifically promote faculty diversity in STEM fields as worthy of attention if the end goal is to improve outcomes for disadvantaged-minority and female students.
And similarly, in the paper, Li and Koedel write:
We conclude by briefly addressing the policy implication of our finding that diversity is particularly lacking in STEM fields. If a rationale for policies to improve faculty diversity is to provide role models for underrepresented students, and if it is presumed that students will gravitate toward such role models, the current diversity imbalance in higher education implies that students from underrepresented groups may be nudged toward lower-paying, non-STEM fields. This would serve to perpetuate an already-existing imbalance in the workforce, both in academia and in the broader labor market (e.g., also see Bayard, Hellerstein, Neumark, & Troske, 1999; Carnevale, Fasules, Porter, & Landis-Santos, 2016). If an aim of diversifying the faculty is to promote better long-term outcomes for underrepresented students, targeted efforts to increase diversity in STEM fields may need to be an explicit objective. However, STEM-specific considerations do not seem to be prominent in current policy discussions on faculty diversity.
Economics has a huge problem with sexism. Beatrice Cherrier has a great historical discussion here, framed around a discussion of Alice Wu’s recent undergrad thesis on the sexist cesspool that is the Econ Job Market Rumors site. Many other fields have problems with sexism, including STEM and non-STEM fields (and including fields with a greater representation of women). But Li and Koedel’s work folds the discipline-specific problems of economics into a broad “diversity in STEM” framing that tends to downplay any culpability for the field itself, and miss how some fields (Psychology, Biology) seem to have fewer problems. It naturalizes that inequality and deflects from the concrete and interrelated work that men have done in specific fields to make those fields seem more scientific on one hand, and keep women out on the other. Here the equation is simpler: Women don’t science, because science is defined as what women don’t do.
My point here is not to argue that we shouldn’t study gender and racial compositions of different fields, nor that we shouldn’t have more and better initiatives aimed at promoting inclusivity in STEM .* I’m also not just expressing bitterness that economics is treated as a science and sociology isn’t.** I’m happy to call us both social sciences, and endlessly debate whatever it is that means (especially if the NSF keeps funding us both!).
I’m writing this post because the lack of gender and racial diversity in academic fields is a big problem with real consequences in terms of both overt inequality, and in terms of what research gets done in the first place. Gender, class, and racial diversity in fields lead those fields to ask different questions, to notice different problems as they emerge. Is it surprising that it was women economists who pioneered and continue to push research on occupational segregation, comparable worth, and the gender wage gap? Would economics have been better able to see and understand the incoming mortgage crisis and financial crisis if it had been more racially diverse?It’s a hard question to answer, but it’s at least plausible.
To solve those problems of representation and diversity, we need better research than this study – research that avoids obvious pitfalls like sampling on the dependent variable, but also research that’s willing to learn from what’s already been done on these topics by scholars in the neighboring discipline (classified as STEM or not!). Li and Koedel do not cite a single paper in sociology, despite the massive empirical literature there on gender and racial diversity in STEM and the academy (e.g. Weeden et al, Cech et al, Morgan et al, and on and on). Economists, in their collective quest for superiority, are insular even when they are writing about and promoting diversity.
UPDATE: Cory Koedel responds on twitter here, which I encourage you to read. I acknowledge that the paper names other criteria for identifying STEM fields, that the paper is not primarily about the demarcation question, and that it makes other important points. I do not think that absolves it of the charge of sampling on the d.v., especially we are not exactly given a clear sense of how the fields were chosen or demarcated. We are in fact given more criteria than fields. His response also does not address the question of other STEM fields (e.g. biology) that have a different gender balance. When Koedel writes: “Finally, you can ignore the two in-between fields – econ and ed policy – and the implications are qualitatively unaffected.” that may be correct, but what happens if you include non-STEM fields like philosophy or STEM fields like biology? As Clayton noted, these inclusions would disrupt the STEM/non-STEM story in fatal ways. Koedel also does not discuss the casual and seemingly evidence-free assertions about there not being a conversation about diversity and STEM specifically (see Kim Weeden’s tweet here for one example involving $270 million from NSF). Last, when I wrote “That is, economics is defined as a STEM field because it is dominated by men”, I was not intending to paraphrase Li and Koedel, I was explaining my interpretation of what they practically did.
*Though there are a ton already. I’m not sure what Li and Koedel are arguing when they say that “STEM-specific considerations do not seem to be prominent” in discussions of faculty diversity. See, e.g., discussions of the “leaky pipeline.”
** As an aside, the limited available survey data suggest that differences in the perception of scientificity of economics and sociology are not widely held. That is, the average person thinks econ and soc are similarly scientific (or not). Here’s a post I wrote back in 2008, discussing GSS data from 2006, to discuss some of the reasons why economics gets treated more like a science. I don’t believe the questions have been asked again since, or in other surveys, though please point out more recent data in the comments if you know of any! The finding there was that similar proportions ranked econ and soc as “very” or “pretty” scientific (51% vs. 49%), but sociology had a higher volunteered “don’t know” answer. So to the extent economics is seen as more scientific, that perceptions lies with elites (policymakers, lawyers, academics, etc.). On the cultural authority of economics, see e.g. Fourcade et al and my paper with Elizabeth Popp Berman.