Direct Comparison: Academic (Under) Represention Across Gender, Sexuality, Race and Ideology

Heterodox Academy strives to promote mutual understanding, constructive disagreement, and viewpoint diversity in institutions of higher learning – and especially within social research fields.

This mission includes promoting diversity in terms of the oft-discussed categories of race, gender and sexuality – as well as the less-discussed (but also pressing) lack of diversity along the lines of class and geography — and of course, the intersections between these categories (and we would be happy to partner with individuals or groups to help pursue these ends further).

However, up to now many of our efforts have concentrated on the issue of ideological diversity in institutions of higher learning. There are a few reasons for this.

First, there is general social agreement that discrimination on the basis of gender, sexuality or race is wrong. However, many in the academy believe it is acceptable or even desirable to exclude non-progressive perspectives. In fact, while noteworthy progress has been made since the 90’s in terms of representation for women and ethnic or racial minority groups, the ideological underrepresentation problem is actually growing worse.

We were concerned about these trends primarily because the lack of ideological diversity seemed to be undermining the quality and impact of social research. Therefore, much of our initial labor was to help our colleagues recognize that the lack of ideological diversity is a real problem, and to commit themselves to being part of the solution.

However, a question has come up time and again regarding the relative scale of these challenges: how does the lack of, say, political diversity measure up when compared to underrepresentation by race, gender or sexuality? To get at this question, we can compare rates of faculty identification across different identity measures.

Overrepresented Groups

With regards to race, research has shown that whites and Asians are overrepresented among tenured faculty.
In terms of gender, men are overrepresented among full-time tenured faculty, while women tend to be concentrated in adjunct or otherwise non-tenured positions.
Lesbians and gays are also overrepresented among faculty, especially in social research fields.
Non-Christian faiths (particularly Jews, Muslims, Buddhists and Hindus) tend to be overrepresented among faculty, but amount to relatively small numbers for both faculty and the general population. The religiously unaffiliated (atheists, agnostics, religious “nones”) represent a much larger share and are significantly overrepresented as well.
Progressives are vastly overrepresented in the academy. According to the most comprehensive study on faculty political leanings to date (Gross & Simmons 2014), progressives outnumber conservatives by 10:1 in the humanities and social sciences, despite the fact that conservatives outnumber liberals in the general U.S. population (and pretty much always have).
While we are not able to directly compare the socioeconomic backgrounds of faculty, we can reasonably infer from information about graduate school completion and school placement rates (see here, here) that faculty are disproportionately likely to come from families in the upper quartile of income (update: new study on socioeconomic background of faculty here. They do tend to come from well-off backgrounds. The more privileged the school, the more privileged the likely background).
We can similarly infer that faculty are likely to hail from cities and the suburbs over towns or rural areas… particularly from metropolitan areas along the West Coast or the Northeast Corridor running from Boston through Washington D.C. (see here, here, here for more).

Comparing Underrepresentation Rates

In order to more easily compare levels of underrepresentation among groups, we can create a common representation quotient by dividing the percentage of representation each group has among full-time faculty divided by the share these groups comprise of the total U.S. population.

On this scale, approaching “1” would signal parity between representation in the academy v. the broader society. Exceeding “1” would indicate overrepresentation, while approaching “0” would indicate severe underrepresentation. In descending order, the breakdown is as follows:

Group	Share of the U.S. Professoriate	Percentage of the U.S. Population	Representation Quotient	Source(s)
Non-Christian Faiths	13	6	2.17	Gross & Simmons 2009, Pew 2014
Asians & Pacific Islanders	11	6	1.83	U.S. Census Bureau, Myers 2016
Liberals	44	26[*]	1.69	Gross & Simmons 2014, Gallup 2018
Religiously Unaffiliated	31	23	1.35	Gross & Simmons 2009, Pew 2014
Whites	75	61	1.23	U.S. Census Bureau, Myers 2016
Men	59	49	1.2	U.S. Census Bureau, Steiger 2013
Women	41	51	0.8	U.S. Census Bureau, Steiger 2013
Christians	53	71	0.75	Gross & Simmons 2009, Pew 2014
Blacks	5	13	0.38	U.S. Census Bureau, Myers 2016
Conservatives	9	35	0.26	Gross & Simmons 2014, Gallup 2018
Hispanics	4	18	0.22	U.S. Census Bureau, Myers 2016

And so, if we were to visualize the over/under representation (with “0” indicating parity), it would look something like this:

This is the picture for U.S. colleges and universities overall. However, if we restrict our analysis to social science fields, the dynamic changes significantly:

Group	Share of the U.S. Social Science Professoriate	Percentage of the U.S. Population	Representation Quotient	Source(s)
Liberals	58	26[*]	2.23	Gross & Simmons 2014, Gallup 2018
Atheist/ Agnostic	39	23	1.7	Gross & Simmons 2009, Pew 2014
Whites	81	61	1.33	U.S. Census Bureau, National Science Foundation
Asians & Pacific Islanders	7	6	1.17	U.S. Census Bureau, National Science Foundation
Men	50	49	1.02	U.S. Census Bureau, Ginther & Khan 2015
Women	50	51	0.98	U.S. Census Bureau, Ginther & Khan 2015
Blacks	5	13	0.38	U.S. Census Bureau, National Science Foundation
Hispanics	4	18	0.22	U.S. Census Bureau, National Science Foundation
Conservatives	5	35	0.14	Gross & Simmons 2014, Gallup 2018

Visualized:

Looking at underrepresented groups: Women are even closer to parity in social research fields than in most other academic sectors. And while blacks and Hispanics are severely underrepresented among full-time faculty in the academy writ large, those faculty that are black and/or Hispanic tend to be heavily concentrated in social research fields (at the expense of STEM fields). However, there also happen to be more faculty employed in social research than most other occupational fields. Therefore, the representation quotients for blacks and Hispanics remain roughly unchanged, regardless of whether we are talking about social research fields in particular, or U.S. institutions of higher learning overall. While Asians remain overrepresented in social research fields, they are less-so than whites.

The non-religious are concentrated even more heavily in social research fields than the academy overall.

Finally, there is a significant downward shift in levels of political diversity when we restrict our analysis to social research fields. Liberals are hugely overrepresented: more than twice as common among social scientists as they are in the general population. Meanwhile, the representation quotient for conservatives in social science fields is 0.14, which is significantly lower than any other major population group measured. In other words, the lack of ideological diversity seems to be vastly more pronounced in social research fields than underrepresentation in terms of gender, sexuality, and race.

Of course, comparing relative rates of underrepresentation is both an empirical and a normative question. Here, we have only addressed the former dimension. For my reflections on the moral and practical considerations of this data, see my essay in the latest issue of Times Higher Education, “If conservative views are not represented in social research, leftists will suffer most.”