Political science 2%
Performing Arts 16%
Computer science 26%
Using the labels liberal and conservative in their (somewhat odd) American usage, conservatives are most concentrated in the caring professions (Nursing, Education) and other social-practicality disciplines (Economics, Business). Liberals are most concentrated in the social-abstraction disciplines (Political science, English, Philosophy, Theology/Religion, Music, Sociology, History) with various hard sciences (Physics, Biology, Mathematics, Engineering, Computer Science, Chemisty) and some social-abstraction disciplines (Linguistics, Communications, Performing Arts) being intermediate cases.
This pattern seems to me to be explicable. Particularly if one compares to the wider society.
General US public self-identification
(Greens and Libertarians probably also have disproportionate faculty support).
Faculty are about as likely to be partisan identifiers, but far more likely to be ideology identifiers, than the general public. As ideological opinions are more characteristic of faculty, it can reasonably be expected that they matter more among faculty than in the general population. But the patterns are also wildly different between public and faculty, with striking levels of opinion hegemony among faculty.
Political beliefs per se often have little direct feedback from their wider effects if implemented to the believer. But if political beliefs act as status markers, then having particular political beliefs may have social consequences in particular milieus leading to selection effects (including self-selection effects). This effect is likely to be strongest where (1) the milieu is concerned with social matters and (2) lacks personal-consequence feedback from the wider social consequences of beliefs (an effect that tenure would obviously increase). Conversely, it is likely to be weakest where feedback on the wider social consequences of beliefs is strongest, reducing the power of opinions as status-markers.
So, if (US) liberal beliefs are status markers, selection processes are likely to be strongest in social abstraction disciplines and weakest in social practicality disciplines, yet still stronger in the faculty generally than the wider population. Not least because being insulated from “vulgar commerce” can be taken to be inherently of higher moral status (the "ivory tower" as bastion of moral cleanliness), provided that commerce is taken to be a marker of moral inferiority. Which writing about the evils of capitalism (and capitalists) would help to establish and reinforce.
The utility of such speaking and writing for the aforementioned status games leads to the “higher childishness” one sees within much academic discourse—ritual sneering at the notion of a “free market”, using labels to morally and intellectually differentiate the user from the phenomena being studied (e.g. ‘neoliberalism’), taking any negative phenomena (e.g. racism) to be a product of an undifferentiated thing labeled ‘capitalism’ and so on. Such status games would also encourage a penchant for interventionist policies, since such policies emphasizes the inadequacy of commerce, how knowing the interveners (actual and those in support) are and the ignorance and inadequacy of the choices of others.
Such pressures and temptations would shift the entire faculty leftwards but to varying degrees, depending on the discipline for the reasons discussed. Hence the above results.
ADDENDA: Let's put it another way. In how many academic seminars and common rooms are their people who have run a business where their own capital was at risk? In how many academic seminars and common rooms are their people who have seriously talked to people who have run a business where their own capital was at risk about that experience? How many academics have read books or articles or other written material by people who have run a business or have seriously talked to people who have run a business? How many academics have done none of those things yet pontificated about "capitalism" and the behaviour of "capitalists"? Would that be remotely acceptable practice in say, anthropology, sociology or pyschology about a subject other than "capitalism" and "capitalists"? No? Then why is it even remotely intellectually acceptable for that? Because it works for the status game.
SCARY QUESTION: How many academic economists would past this test? How many would past this test if they had not read Ronald Coase's article The Nature of the Firm?