“Liberal Bias” in Academia: Media Narrative vs. Social Science
“Available data do not support the claim that university professors are excessively and disproportionately liberal, much less that a majority of students are being educated by left-wing professors. So why do so many people have the impression that they are?”
That’s one of the questions taken up by Charlie Tyson and Naomi Oreskes in “The American University, the Politics of Professors and the Narrative of ‘Liberal Bias’,” published recently in Social Epistemology Review and Reply Collective.
They write:
The best survey data we have on the political attitudes of professors suggest a seed of truth in conservative complaints about the liberal faculty lounge—but just a seed. The American academy is more liberal than the general public, and it contains a larger proportion of liberals than other major professions. But critics of academic leftism routinely overstate their claims, in some cases grossly so. By and large, most professors self-identify as moderates of one kind or another. The mundane sociological truth is that academics are broadly representative of the demographic from which they are selected: highly educated people.
What explains why many people have what the authors call “an exaggerated impression of professorial leftism”? One factor has been the publicization of politically motivated poor quality studies:
We suggest that influential conservative analysts… have created a false impression of faculty political composition by committing one or more of the following methodological errors:
◦ Ignoring or eliding the preponderance of faculty who describe themselves as moderates by relying on the binary metric of party registration;
◦ Focusing only on elite, mostly private and northeastern institutions which may be more liberal overall, rather than institutions in the South, West, Midwest, or Great Plains;
◦ Focusing on humanities and social sciences departments within those institutions, and ignoring the natural sciences and engineering departments;
◦ Focusing disproportionately on particular subjects in the humanities and social sciences which conservatives find conspicuously liberal in orientation, such as English, or in which a (progressive) political critique is part of the discipline, such as women’s studies, and ignoring disciplines like economics or engineering which are often aligned with conservative principles such as laissez-faire economics;
◦ Omitting community colleges, theological institutions and seminaries, religiously affiliated or evangelical colleges (e.g. Wheaton, Brigham Young), and military institutions—institution types that together educate millions of students—from their analysis.[46]
These errors sometimes accompany basic sampling and methodological issues, including small sample size or biased sampling frame, poor response rate, response bias, and nonstandard question wordings.
When we examine studies (such as that performed by Gross and Simmons) that
a) Ask a series of standardly worded questions rather than relying on the crude metric of party registration;
b) Include faculty from a wide range of disciplines and institutional types; and
c) Have a large sample size, we find that they yield different results than the studies that claim to unmask professorial leftism.
Methodologically robust studies do not reveal the excessive leftism that less robust studies claim to do. The fact that these methodological errors are so widespread raises serious doubt as to whether these studies have been undertaken in good faith.
Apart from their critiques of attempts to show a liberal bias in academia, Tyson and Oreskes also comment on the idea that academia needs more ideological or political diversity:
Jonathan Haidt and his colleagues have argued that academia needs a measure of “political diversity” in order to minimize its blind spots. That might be so, but we would need a rubric for what that “diversity” might consist of. We cannot simply index the academy to trends in American politics and insist that the internal composition of the academy reflect the larger political culture.
In order for claims to survive in academia, they need to be attached to evidence. (When claims persist in scholarly discourse without sufficient evidence, we consider that an academic failure.) Americans of all political persuasions are committed to positions that evidence does not support. No reasonable scholar would claim that we should import empirically inadequate positions (much less empirically refuted ones) into academic discourse for the sake of some arbitrary standard of representativeness.
You can read the whole paper here.
(via Jason Stanley)
Related: Political Hostility and Willingness to Discriminate in Philosophy, Do Professors Penalize Conservative Students?, Demographic Diversity is Good for Philosophy, Professors Favor Free Speech, Political Uniformity and Religion in Philosophy, Lack of Political Diversity: A Problem?, Political Bias in Philosophy, Because They Are Universities, Response From A Conservative.
image: modified detail of geological chart by Levi Walter Yaggy
Imagine being mad at English because you think it’s overrun by leftists.
This summary of the Gross and Simmons study is misleading. Contrary to what the authors claim, this study clearly demonstrates that faculty are “disproportionately liberal.” On a 7 point scale measuring political orientation, 44% of faculty are in the two most liberal categories. This compares to 23% of the general population (from the same question on ANES at around the same time).
Further, the authors of the piece you cite characterize faculty ideology this way: “By and large, most professors self-identify as moderates of one kind or another.” But only 47% of professors are moderates according to Gross and Simmons. If the above quote is accurate, it would also be accurate to say that “by and large” most professors are liberal.
(To anticipate a reply, it is not appropriate to use PhD holders as the relevant baseline for measuring overrepresentation and bias. For, by this measure, we would find the Black faculty are not underrepresented in philosophy.)
You said “contrary to what these author say, the study clearly demonstrates that faculty are ‘disproportionately liberal’.”
The authors in fact said, that the study demonstrates (‘G&S note that…’) that faculty are ‘disproportionately liberal.'”
Following is the relevant quote from the article. Would you please respond to what I’ve said above?
——
Gross and Simmons found a much more centrist professoriate than is alleged in conservative discourse. Some 44 percent of professors described themselves as “extremely liberal” or “liberal” (9 percent and 35 percent, respectively); 46 percent described themselves in centrist terms (18 percent as “slightly liberal,” 17 percent as “middle of the road,” and 11 percent as “slightly conservative”); and 9 percent described themselves as “conservative” or “extremely conservative” (8 percent and 1 percent respectively).[20] In other words, liberals do outnumber conservatives, but the largest cohort of faculty—46 percent—are moderates, spanning the terrain between center-left and center-right.
Gross and Simmons note that American professors are disproportionately liberal—though probably less left-leaning on a number of issues than their counterparts in a number of other countries, including Denmark, Norway, Spain, and New Zealand.[21] From our vantage point, what is most striking about their data is how much more moderate the American professoriate appears, in contrast with what conservative critics—and probably most liberals—believe of this population.
By “this summary,” I was referring to the passages quoted by Justin. Such as: “Available data do not support the claim that university professors are excessively and disproportionately liberal…”
“Excessive” is a normative word and it would take much more than descriptive data to support the claim that professors are excessively liberal.
Including the second conjunct makes this quote misleading to my mind. Available data do support the second conjunct and the entire statement is merely true on a technicality: the first conjunct is something that data simply do not speak to.
I would have likewise objected if they had written that available data do not support the claim that university professors are insufficiently and disproportionately conservative.
Imagine if someone wrote that available data do not support the claim that philosophy faculty are excessively and disproportionately white. I would hope someone would take issue with that characterization.
“We point to a more mundane explanation for academia’s leftward tilt: the fact that highly educated people with advanced degrees tend to be liberal.”
I’m genuinely confused as to how this ‘explanation’–IF it turns out to be anything other than a Moliere’s Doctor explanation–is at all inconsistent with the relevant conservative criticism. Conservatives say that academia brainwashes young people and makes them more liberal, and that this is largely because of the leftward tilt. The authors respond by acknowledging the tilt and then saying that “Sociologists have long argued that higher education has a liberalizing effect on social and political views…”. Huh? Isn’t this exactly what the relevant critics charge?
Two more quotes:
1. “Available data do not support the claim that university professors are excessively and disproportionately liberal, much less that a majority of students are being educated by leftwing professors.”
2. “We do not deny that professors are more liberal than the general public, and that academia is among the most liberal of the major professions.”
You can only make these two claims consistent by engaging in a little equivocation and misrepresentation. The normative question of whether the overrepresentation is “excessive” is of course technically open, but it is 100% clear that when conservative critics say that there is overrepresentation, they mean with respect to the American population in general. That is, they mean “more liberal than the general public”. They do not mean “with respect to highly educated elites”. This is kind of ironic, because willful misrepresentation of your opponent’s position is a hallmark of…. bias.
To add to the other well-deserved criticisms, “focusing on humanities and social sciences departments within those institutions, and ignoring the natural sciences and engineering departments” does not seem like a methodological error. I’m not steeped in right-wing circles, but did anyone ever worry that chemistry professors were out there indoctrinating students?
I was wondering the same. And it strikes me there’s a missed connection to the fact that those most likely to argue for gutting liberal arts departments probably aren’t liberals.
Yes. Conservatives often believe that university science profs are attempting to indoctrinate students into a liberal worldview via evolution, big bang theory, etc. I was, as an undergrad, hailed by some religious students in my physics class as a hero for arguing with the professor when he said time doesn’t exist. They were sure I was defending Jesus from him, when in fact I was just having a non-religious philosophical debate with the guy.
How would chemistry fit in? A chemistry professor’s “teachings” will be under suspicion of functioning to support atheism and liberalism if the conservative student gets any whiff of an idea that the chemist in question thinks chemical facts help explain things like genetics, evolution, etc.
Fair enough. I will therefore weaken my claim: ignoring the natural sciences and engineering departments *need not* be a methodological error. It depends on which particular conservative worry is being tested. The worry that science professors might be indoctrinating students into atheism is very different from the worry that social sciences and humanities professors are indoctrinating students into left-wing politics, and if it’s only the latter you’re worried about–as I suspect is the case of many conservatives–then it makes total sense to ignore the natural sciences and engineering departments.
> Focusing only on elite, mostly private and northeastern institutions which may be more liberal overall, rather than institutions in the South, West, Midwest, or Great Plains;
It is a problem if elite institutions are excessively liberal.
> “The problem is that the conservative position is refuted by an overwhelming body of scientific evidence, and has been so now for more than two decades.”
1. To the extent that the conservative view represents values and not factual claims, it cannot be refuted by empirical evidence.
2. Humanities professors are not experts at empirical claims, yet they are disproportionately liberal. Why?
> whether these studies have been undertaken in good faith.
I question the authors’ (an English PhD at Harvard and Naomi Oreskes, a historian of science) good faith when they apparently don’t include any non-liberal or social/political scientist in writing this piece.
Yes, yes, I know that anecdotes are not the plural of data, but give me a break. As a moderate liberal in academia, I find it patently obvious that social science and humanities professors are overwhelmingly liberal, both in number and degree. To paraphrase a recent comment by a Harvard economist, when I see social science results such as this I have to roll my eyes and ask “what went into making that sausage?”
These lines are curious:
“The American academy is more liberal than the general public, and it contains a larger proportion of liberals than other major professions. But critics of academic leftism routinely overstate their claims, in some cases grossly so. By and large, most professors self-identify as moderates of one kind or another.”
If the first two claims are true, how should we interpret the claim about the majority of professors who self-identify as moderates? Do they view themselves as moderates relative to their academic colleagues? Or do they view themselves as moderates relative to the general public? I recall, for instance, Professor Stanley arguing that most of his Yale colleagues cannot possibly be considered on the left because they by-and-large supported Hillary Clinton for President. That may align with an academic view of what it means to be on the left. But why privilege that perspective when we are attempting to understand the ideological bent of the academy itself?
So often in these attempts to defend the ideological status quo in the academy, one comes away with the impression that such defenders do not see the bias that is so evident to their heterodox colleagues because they themselves swim in it. In such an environment, it is unsurprising that critiques of academic leftism will be perceived as grossly overstated. But why should we trust those who are part of the problem with an unbiased diagnosis?
What if, instead of considering university professors as a whole, we focus in on the smaller set of humanities professors who specifically have grant money that they can use to hire post docs; mentor aspiring academics; promote their work; and achieve wider influence in academia, the arts, or politics. Of the set that consists of humanities professors with grant money, what proportion are conservative, and what proportion are liberal?
Good point!
Molly, the vast majority of departments are overwhelmingly on the left. We are both friends with people who have declared publicly (on Facebook and elsewhere) that when they see a job application from someone they think is a right winger or a libertarian they pitch their applications. For the most part, people with grant funds aren’t making unilateral decisions about permanent faculty hires, promotions are still following the usual protocols, etc.
Chris Surprenant, I have no idea who you are referring to. We have Facebook friends who do what you describe? Who are these people? If that’s what they’re doing, it is unethical, just as it would be unethical for right-wing professors to manipulate job searches in order to hire other right-wing professors.
“Focusing only on elite, mostly private and northeastern institutions which may be more liberal overall, rather than institutions in the South, West, Midwest, or Great Plains”
Of course the critics focus on elite, northeastern institutions; the critiques I’ve seen are more or less explicit that that’s where the problems are. The alleged liberal bias at elite institutions is an issue, given the cultural influence exerted by and within those institutions, given the disproportionate share of grant funding they receive, and given that tenured academics employed elsewhere disproportionately received their PhDs from those institutions. Are the authors suggesting that a pernicious liberal bias at prestigious universities is okay so long as it’s not also found in less prestigious ones.
I’d add that appealing to survey outcomes that indicate the distribution of political preferences within universities misses the point. The complaints from conservatives and others relate to a lack of tolerance of views on controversial topics that are unsanctioned by activists within the academy. People are incentivised to falsify their preferences in a leftward direction: i.e. either to pledge their allegiance to an orthodox leftwing position on one of these issues or at least to keep any reservations to themselves. The alleged problem is not a given political orientation is over- or under-represented in a numerical sense, but that there are costs associated with failing to toe the line. Perhaps this problem is merely imagined by conservatives (or exaggerated by them), but survey outcomes won’t show this. In fact to the extent that preference falsification is prevalent, you would expect the outcomes of anonymously conducted surveys to reveal a less severe political skew than appearances might otherwise suggest.
Find me a psych department with 15 percent republican faculty and I’ll believe you. Not gonna happen. Jerry kroth, psychology prof
I’d like to devote more time to this article in the future, as I’ve written on similar themes and I think the more we can have a reasonable conversation about some of the basic facts, the better off we’ll be. For now, I’d like to highlight four things.
1. A more thorough engagement with Duarte et al. (“Political Diversity Will Improve Social Psychological Science”), the BBS article that partly kicked-off the Heterodox Academy, would have helped Tyson and Oreskes make their case.
They write:
In a context in which that BBS essay was published, a claim like this really needs to take account of arguments like those set forth in that essay, the response essays from other social scientists in the same journal, and the authors’ rejoinder. That’s to say nothing of the documentation of similar problems by Heterodox Academy members since 2015. It’s great that philosophers want to get involved in these debates, but we have to be sure that we’re engaging with the work that’s already been done by researchers in other fields. It’s noteworthy, for instance, that there was almost near-universal agreement among the respondents, both with Duarte et al.’s descriptive claim about what was going on in social psychology, and with their argument that it sometimes has a negative impact on the work in that field. The lack of serious engagement with that essay did Tyson and Oreskes’ argument a substantial disservice, particularly as they decry ‘cherry-picked data’ and levy accusations of bad faith at their interlocutors. We can do better, it seems to me.
2. This claim exhibits the same sort of shortcoming:
Instead of dismissing talk of ‘political diversity’ with a claim like this (it’s the only time the term is even mentioned in the essay, let alone used), Tyson and Oreskes could have begun by looking over the page labelled “The Problem” at the Heterodox Academy, where they would have found headings like “The Solution”, “Open Inquiry”, “Viewpoint Diversity”, and “Constructive Disagreement”. Plenty to talk about there.
I’m not saying that Tyson and Oreskes would have been convinced by what is said–still less that they should be. But it’s not accurate to suggest that the solution on offer is to “index the academy to trends in American politics and insist that the internal composition of the academy reflect the larger political culture”. By not engaging with positions of the sort laid out in the Heterodox Academy statement (or the dozens of discussions on political diversity in the academy that have been going on for the last few years), Tyson and Oreskes come off as disingenuous when making this suggestion (particularly when it comes on the heels of a remark about putative tolerance for scientific racism).
3. I think the discussion of oppositional epistemology is illuminating, and perhaps a point for future consideration. I’d be curious to see what advocates of viewpoint diversity have to say about standpoint epistemology as it manifests in, say, women’s studies and the conservative who rails at a perceived lack of representation in the academy.
4. Tyson and Oreskes’ observation on changes in the political landscape in the academy, and especially the shift from the McCarthyism of the 1940s and 50s to the situation today, is also productively drawn into the conversation. We might also compare events like the Berkeley free-speech movement of the 1960s and the Berkeley protest/riots of 2017. I don’t know of any social science research that’s addressed the forces lying behind these shifts, but Tyson and Oreskes are right to remind us that things were quite different during the period of the red scare.
When are we going to grapple with the really important problem with university professors? They are excessively and disproportionately highly educated. I know this is just anecdotal, but in every place I’ve taught (4 research universities and 1 liberal arts college), every member of the Philosophy Department had a Ph.D. Every single one! What is more, the vast majority of members of the other departments I was familiar with at those institutions also had doctoral degrees. Now, I haven’t done any rigorous studies, but I get the impression that this high a proportion is actually quite common at other colleges and universities as well. I haven’t actually checked on this further point either, so I could be wrong about it, but I also get the impression that not even half of the general population of the US possess doctoral degrees. Maybe not even a third. This is simply shameful. Something must be done to redress this imbalance. Down with this sort of thing!
There’s worse. Utilitarians are over represented in philosophy departments!
Nicolas,
You’re right–that IS worse!
Sadly, we’re vastly underrepresented in philosophy departments. I have visited many departments to give talks where the students were basically invited to poke me through the bars of my cage. In many places, I am the first utilitarian these poor students have seen. Some of them have to make do with the likes of Richard Galvin and John Harris. Can you even imagine the mangled version of utilitarianism that they get from these people? The horror!
Haha. You’re still overrepresented relative to the general population don’t you thin? My comment was of course tongue in cheek, in the spirit of yours.
I took it in the spirit in which it was intended. 🙂 But actually, I do believe that utilitarians are underrepresented relative to the general population. It’s a tricky issue, because many people obviously haven’t given much thought to whether their views are utilitarian, as opposed to deontological (or even, gasp, virtue ethical). My view, of course, is that utilitarianism is the real ‘commonsense morality’.
I see. You may be right (sociologically)! If so then you’re clearly being unfairly discriminated against!
It is worth pointing out that, given that the Fact of Reason is universal, Kantians are necessarily underrepresented (or at least not overrepresented).
The running argument against the study seems to hold that the views of both conservatives and liberals can be equally well-supported by reasons and that the subsequent (real or perceived) disparity in representation in higher education deserves some kind of redress.
I’m sorry–has this been established? Outside of appeals to population and similar fallacies, why think that conservative or liberal views will hold up equally well to the kind of scrutiny that we encourage in higher education?
I think it is important to distinguish between the issue of whether there is a disparity on the one hand and the question of whether (if there is one) the disparity is justified. The running argument against this study is simply that there is such a disparity, its claims to the contrary not withstanding. This running argument can be accepted (and, I think, based upon the comments of this thread, is accepted) by even those who think that conservative views will not hold up to intense scrutiny.
1. Even some conservative positions typically espoused by religious people does not assume theism. For example, the philosophical arguments for and against abortion does not assume the existence of God. In fact, most political debates tend to cite scientific evidence showing the extent to which a fetus is a person.
2. Many other conservative positions are not related to religion, for example, fiscal responsibility, small government.
3. Many conservative positions reflect preferences and values, for example, patriotism, and cannot be refuted by “evidence”.
4. Your ignorance of the diverse positions of conservatives might illustrate the bad consequences of liberal over-representation.
Wow, Xirui Zhao! If this is a reply to Matt, as its position in this thread (and your phrase ‘your ignorance’) suggests, it supports his implied point. He said nothing at all about religion. And how, exactly, does scientific evidence bear on the moral category of personhood? Perhaps you are confusing the biological category of human with the moral category of person? A common mistake. And what are you referring to by enclosing the word ‘evidence’ in quotation marks? And what does that have to do with the reasons (note the difference) that Matt is talking about? You might want to quit while you’re behind.
Alastair,
I was choosing to believe that comment was not directed as a response to me at all given its lack of germaneness as a reply.
I do want to offer a short correction: “… views will hold up equally well …” should read ” … views should hold up equally well … ” as I was intending to emphasize the is/ought problem in this argument.
Three thoughts:
1) There is quite a lot of evidence that political leanings (like religious leanings) depend a lot on your background and upbringing and aren’t just reached through the bright light of reason.
2) Humanities departments skew significantly more liberal than hard-science departments. Tempting though it would be to suppose that’s just because we’re smarter, I think it’s more likely to illustrate that there’s more at work than just people reaching the right conclusions.
3) Most academics (even most philosophers) don’t work directly on political issues. (And if the thought is that academics are bound to apply the same searching rigorous standards that characterise their research to all aspects of their life and views… well, all I can say is that we must have been to different faculty meetings.)
I think that people, on the right and left, tend to make the unjustified assumption that subsets of people will be representative of their supersets, in the absence of biased selection. For instance, when you first hear about the relatively high proportion of American piano manufacturers who are ethnic Germans, it’s hard not to suspect that this is the product of some bias, but as far as I can tell, it is not.
However, the best unbiased explanation for the lack of conservatives in the American academy is the James Demore’s “Google Manifesto” approach, in terms of differences in average interests and temperament. I haven’t seen this rigorously tested yet, though, and I’m agnostic about it.
“The mundane sociological truth is that academics are broadly representative of the demographic from which they are selected: highly educated people.”
This is the central empirical claim of the article. It is contradicted by the one source they provide for relevant data, the Pew survey. There, the percentage in the two most conservative categories does not change much between educational strata (vs total collapse in elite or progressive branches of academia). What is clearly seen in the survey data is a shift toward stronger liberalism within the non-conservative categories, so that the relative number of liberals per conservative increases from parity (1:1) to as much as a 3-to-1 imbalance when going from the least to the most educated strata. That’s fine and seems to match anecdotal observation. But the stuff the authors are trying to discredit, the Heterodox Academy data and surveys within highly liberal fields such as anthropology and psychology, show huge ratios like 20 or 80 liberals for every 1 conservative in multiple fields and at some universities. It’s blatantly false, a whitewash, to call those massive numbers a “mundane” matter of academics being “broadly representative of …. highly educated people” when education-based differences are off by more than an order of magnitude from the phenomena they are supposed to explain. There are no such giant ratios among doctors, engineers, statisticians or chess players, but there is a real and justified fear among (for example) anthropologists and professors of education of revealing, e.g., support for Trump or opposition to Black Lives Matter protests.
It seems to be true that humanities academics are more to the left than those in the sciences, and both to the left of the general population. I’m sure background plays a role in one’s views. But I would not admit to the charge that my socialist views are any sort of bias any more than my pro-vaccine stance is some sort of bias.
Let me change to a slightly different topic that is still relevant. In the 1930s the political orientation of the humanities and the sciences was reversed. Scientists were overwhelmingly socialists and often members of the Communist party. C.P. Snow’s Two Cultures book is remembered for its lament that the two groups didn’t talk to one another, but the real message was that scientists should be in charge politically, since the humanists couldn’t solve the big problems, such as poverty. Snow was a leftist himself and greatly admired his like-minded fellow scientists: “scientists are instinctively on the left” and “humanists are instinctively on the right.” He wrote that in 1960 when it was probably a stretch. It was plausible in the 30s and 40s, but it is certainly false now. It’s an interesting change in attitudes and I wonder why? I doubt that it is a mere change in bias; surely there’s something deeper going on.
I think that the transition away from a scientistic (but not necessarily scientific!) rhetoric on the left is part of the story. The idea of planning a society on the basis of articulated centralised reason and quantified metrics apparently tends to appeal to physicists, engineers, and behaviourist psychologists (less so to biologists and chemists?). So a planned economy based on centralised and nationalised industries tends to appeal to many scientists. This is different from the “better pay, better work, and more of it” appeal that traditional socialism (not just e.g. Scandinavian social democracy) once had for the working classes – that was more about the idea that THEY’D have more power, not the planners.
Also, one of the virtues of science since about 1500 AD has been its lack of respect for tradition, both in its practice (not treating the judgements of even Great and Virtuous people like Aristotle and Newton as authoritative over observation and logic) and in its rhetoric (from Francis Bacon onwards). If one also applies this attitude to social matters, it tends to encourage a disrespect for tradition, either in terms of economic organisation or non-economic social affairs.
Modern socialism has dropped almost all the scientistic rhetoric and at most stays pretty quiet about central planning. It’s much more associated now with post-modernism and other views that, rightly or wrongly, are seen by many scientists as anti-scientific. Also, for conservatives and libertarians interested in intellectual pursuits in the academy but who don’t like arguing all the time or feeling too “different”, the natural sciences are one place where they can generally work quietly and collegially. They need only the occasional and indirect mockery of their political views, which tend to be unknown to their colleagues anyway.
(For conservatives/libertarians who LIKE arguing all the time but who also like to calculate and measure things, there’s economics. I think that economics is maybe the most fun part of the modern academy because it has such a spread of political opinions and such a wealth of debate-hungry people who nonetheless share a methodological framework of how to argue with each other.)
In contrast, humanists have always tended an uneasy relationship with centralised power. Until relatively recently, at least in universities, they also tended to be strong admirers of tradition, as you’d expect from people whose chief role was the preservation, inculcation, and exegetical examination of a Canon. As the humanities have become more about innovation and iconoclasm, they have increasingly attracted the sort of radical thinkers who would tend to be socialists and extremely liberal on non-economic questions. The lack of scientism in modern socialism helps too. The most successful socialist rhetoric has a certain humane softness: the ideal is comfort and security for all in accepting and relaxed communities, in harmony with nature. That was always a part of socialism (see Bertrand Russell, George Orwell etc.) but I think that many scientistic socialists of 1920 would have found it unbearably soft in contrast to their modernistic and intense technocracy.
So, to summarise, I think that these are some factors: (1) socialism changed, (2) the humanities changed, (3) the sciences didn’t change in any relevant way, but became a relatively inviting place for non-socialists and non-liberals, and (4) economics is exceptionally fun if you enjoy debating political questions in a civilised but intellectually rigorous environment. I’m sure there are other factors that I’m missing out, e.g. the relative decline of academic incomes and the elongating path to stable academic employment, which turns off e.g. libertarians interested in increasing their financial independence and conservatives interested in raising above average-sized families.
Oh, I forgot: some of conservative/libertarian rhetoric became more scientistic-friendly. It’s very hard to find positivist right-wing thinkers prior to 1945; since then, not so much, especially among economists. For example, Milton Friedman was much better at making the case for capitalism towards scientists than, say, someone professor from Edmund Burke, just as Karl Marx was better at making the case for socialism towards scientists than Jacques Derrida.