Many academic disciplines have few if any conservatives in their ranks. The disparity between conservative underrepresentation in the academy and their significant presence in the general population has been well-documented, but the question of why the disparity exists is more difficult to answer.1 Using a nationwide sample of university faculty collected from 1959–1964 by Henry A. Turner and Charles B. Spalding, we find evidence that self-selection—conservatives choosing careers outside of academia—explains only part of the disparity. A strong self-selection position would suggest that party affiliation is stable over time and usually established prior to entering graduate school.2 These data, however, show that of the faculty who changed their party affiliation, the majority did so after becoming faculty members. This finding strongly suggests that pre-existing political characteristics of each discipline, not just of individuals who enter those disciplines, influence the likelihood of faculty changing their party affiliation. In these data, disciplines with relatively high percentages of Democrats (e.g., political science, sociology, law) had the highest conversion of faculty from Republican to Democrat. In the fields more evenly split in party affiliation (e.g., biosciences, geosciences), conversion from Democrat to Republican was the lowest.
These data tell us that about 40 percent of faculty changed their party identification, with about 60 percent of those who changed parties converting to the Democratic party. Translated into comparative ratios, political science had the highest conversion ratio of 5.58:1, followed by sociology at 4.81:1. The hard sciences had the lowest ratios of conversion. Analyzed a different way, using divergence from parental party affiliation, a similar pattern emerged: conversion ratios of 3:1 favoring Democrats were found.3 Thus, self-selection, or differences in party affiliation before becoming a faculty member, can account for only part of the political disparity across disciplines between the number of conservatives in the academy as opposed to the number of conservatives in the general population.
Although these data are old, we believe they reveal some crucial information that has been overlooked by recent studies and that bears further investigation. A large swath of studies has documented glaring disparities in the political leanings and party affiliations of university faculty. Studies from 1960 to 1972, for instance, found Democrat-to-Republican ratios ranged between 1.3 to 2.04, while later studies revealed substantially larger ratios, especially in the humanities and social sciences. Christopher Cardiff and Daniel Klein’s analysis of tenure-track faculty at eleven California universities found an average Democrat-to-Republican ratio of 5:1.4 Ratios ranged from a low of 0.9:1 at Pepperdine to a high of 8.7:1 at UC Berkeley. Cardiff and Klein estimated a ratio of 10:1 in humanities departments to 1.3:1 in business. Similarly, they found dramatic disparities when examining specific disciplines. The ratio of Democrats to Republicans in sociology was 44:1, in ethnic studies 16:1, while ratios of 1:1 to 3:1 were found across business, engineering, medicine, and economics. More recently, Mitchell Langbert and colleagues examined the voter registration of faculty in forty U.S. colleges across the fields of economics, history, journalism/communication, law, and psychology.5 They found an overall disparity of 11.5:1 in favor of registered Democrats. Ratios ranged from a low of 4.5:1 for economics to a high of 33.5:1 for history.
On many campuses and across many academic departments, especially those at more selective institutions, political conservatives and members of the Republican party are entirely absent.6 The causes of these disparities are often reduced to two possibly connected perspectives: self-selection and discrimination. Self-selection refers to mechanisms that propel individuals with similar traits, viewpoints, and dispositions to make similar choices. Writing in 1958, Paul Lazarsfeld and Wagner Thielens argued that political disparities in the social sciences reflected “a natural selection” of liberals into the academy.7 The social sciences, they reasoned, were attractive settings to individuals interested in challenging social institutions and crafting imaginative social theories. Conservatives, they reasoned, had little interest in such matters and thus did not share the liberal “academic mind.” Echoing contemporary sentiments, Lazarsfeld and Thielens concluded that “occupational self-selection” was the driving force behind the political disparities that were already prevalent in their analysis of 2,451 social scientists.8
Contemporary research also draws heavily from the self-selection paradigm. Matthew Woessner and April Kelly-Woessner (2009), for example, analyzed data from a national sample of college freshmen and seniors. With a focus on the 13 percent of the overall sample of seniors who answered affirmatively that they would likely pursue a Ph.D., the authors found that 24 percent self-identified as being on the far left, 18 percent were liberal (42 percent overall liberal), while 9 percent were conservative, and 11 percent were on the far right (20 percent, overall conservative). Based on these differences, Woessner and Kelly-Woessner argued that conservatives self-select out of academic careers early in their education, largely by selecting into more “practical” fields, such as engineering and medicine.9 They also argued that because conservatives and liberals differ in their “philosophy of life,” where conservatives attach less priority to academic pursuits such as writing original works, they were less likely to find scholarly pursuits attractive. Similar findings emerged from an analysis of data from the National Longitudinal Study of Adolescent Health.10 Almost one-half of those seeking a doctoral degree were self-described as liberal compared to 18 percent who described themselves as conservative.11
Unfortunately, the self-selection position is often invoked without a broader awareness of the more complex factors associated with individual choice. Often overlooked are the various and sometimes subtle incentives, attractors, and deterrents that affect individual decision making. Much the same way political parties attract individuals with similar beliefs and repel dissimilar others, incentives and attractors in certain disciplines may entice only those at the left end of the political spectrum. Conversely, these same attractors, which may include politically radicalized faculty and classes, would likely repel individuals who do not hold these orientations. Self-selection, then, occurs in a context where disciplinary attractors are heavily aligned with one political orientation or in the absence of any incentives that include other orientations. “Self-selection,” writes Peter Wood, “by no means rules out the possibility of bias: The most effective way to keep out a whole class of people who are unwelcome isn’t to bar entry, but to make sure that very few in that class will want to enter."12
Large numerical disparities may reflect not only self-selection but also political conversion—that is, change from one political viewpoint or political party to another. Given the heavy incentives favoring left-leaning viewpoints and party membership in certain fields, conservatives and Republicans entering those fields may be more likely to convert to left-leaning political parties. Indeed, Lazarsfeld and Thielans argued this very point: “Once [liberal] colleagues are in the majority, even a slight numerical differential may build up to a considerable effect on the uncommitted man.”13 Once employed in an academic department, friendships emerge and social networks expand that sponsor the “development of norms by mutual interaction.” These norms not only cause individuals to change their political orientations to mirror those of the dominate group, they result in a “slow atrophy of conservative potentialities.”14 Thus, professors may change their political loyalties for reasons that are linked to the preexisting political climate in an academic discipline.
Political disparities likely reflect a dynamic process involving a range of pre-existing traits and beliefs, disciplinary attractors and incentives, changes in viewpoints over time, and decisions to broadcast or to conceal one’s political views.15 The Turner-Spalding study constitutes a unique dataset that includes measures of faculty political affiliation, stability of political affiliation, changes in affiliation, and the timing of changes in political affiliations.
The data come from a mailed survey of a nationwide sample of university faculty selected randomly from the membership rolls of professional groups. From 1959 to 1964, Turner and Spaulding mailed 4,355 survey questionnaires to professors in botany, classics, economics, philosophy, history, math, engineering, geology, law, political science, and sociology.16 They received 2,647 surveys—a 61 percent response rate—of which 2,389 were usable. Sample sizes ranged from n=304 political scientists, n=223 economists, n=220 humanities professors, n=289 professors in the biosciences, n=241 in the geosciences, n=297 sociologists, and n=242 law professors.
Overall, 73 percent of faculty were Democrats and 27 percent were Republican. Table 1 (below) shows the distributions. In the outcome column, Democrats were clear majorities in political science (86 percent), sociology (94 percent), law (86 percent), and economics (83 percent). They also composed majorities in the classics (74 percent), the biosciences (62 percent), and the geosciences (56 percent). The ratios of Democrats to Republicans for each discipline followed a similar pattern. Sociology (12.4:1) had the highest ratio, followed by political science (6.05:1), law (6.00:1), economics (4.93:1), classics (2.86:1), biosciences (1.63:1) and geosciences (1.29:1).
Stability and Change in Political Party Membership Across Academic Disciplines
Note: Outcome column may not equal sum of selection plus conversion due to how the subjects’ initial political affiliation was measured. Source: Authors’ analysis of data from nationwide sample of university faculty collected from 1959–1964 by Henry A. Turner and Charles B. Spalding.
Stability in Party Identification
Self-selection arguments posit that political party identification forms prior to entrance into graduate school or prior to becoming a faculty member. These data demonstrate that self-selection is clearly part of the reason for ideological disparities across disciplines. Most notably, 60 percent of respondents reported never changing their political party, including 53 percent of Republican faculty and 63 percent of Democratic faculty. Across disciplines, 72 percent of political scientists, 65 percent of law professors and sociologists, 61 percent of economists and geoscientists, 63 percent of faculty in the classics, and 50 percent of those in the biosciences reported never changing their political party affiliation.
We computed the ratios of stable Democrats to stable Republicans across each discipline and overall. This ratio provides another estimate of the relative stability of party membership. For example, in sociology, 94 percent of Democrats, compared to only 6 percent of Republicans, reported never changing their party affiliation for a ratio of 14.70:1. In law the ratio was 6.56:1, while in political science the ratio was 6.26:1, followed by economics at 3.91:1, classics at 2.34:1, biosciences at 1.40:1, and geosciences at 1.00:1.
Changes in Party Identification
To capture changes in party identification, respondents were asked when they had changed their political party, with response options ranging from 1= before college, to 4 = after starting teaching in college. As shown in Figure 1, relatively few professors changed party preferences before entering college. Instead, increasing proportions changed political parties as they transitioned from undergraduates to graduate students, and then to faculty. Clearly, however, of those who changed party affiliation, the majority reported changing their party affiliations after they started teaching in college. Approximately 55 percent of professors in the biosciences, geosciences, classics, economics, and law reported changing affiliations after taking on the job of professor. There are two exceptions to this pattern: faculty in political science and sociology reported changing their party affiliations across their educational trajectory, suggesting that party affiliation may be more strongly sorted across time for these fields. Yet it should be noted that these disciplines had the largest ratios of Democrats to Republicans and had large ratios of stable Democrats. Nonetheless, it is clear that most professors who changed party affiliation, did so in graduate school and, especially, after entering the professoriate.
Timing of Political Party Conversion Across Disciplines
One explanation for this finding is that individual professors are sensitive to the signals, incentives, and risks associated with belonging to a minority political party and that these incentives may be particularly stark in disciplines already dominated by one political party. We were able to tentatively address this possibility. Respondents were asked if information from their discipline affected their decision to change political parties. There was considerable variation in responses: only 21 percent of professors in the classics and the biosciences agreed, as well as only 34 percent of professors in the geosciences. However, a majority of faculty in law (54 percent), sociology (63 percent), political science (66 percent), and economics (71 percent) agreed that information from their discipline affected their choice to change political party. Academic disciplines with relatively high ratios of Democrats to Republicans likely provide the strongest motivation for professors to change their political parties. When considered in conjunction with the previous findings that many respondents changed political affiliation only after entering graduate school or, especially, after becoming professors, it appears reasonable to conclude that certain academic disciplines likely incentivize faculty to change their affiliation.
Changes in the timing and context of political party change, however, overlook the direction of party change. When comparing the percentages of faculty who reported changing parties across disciplines, we found that 63 percent in political science changed to the Democratic party, followed by 60 percent in economics, 58 percent in geosciences, 56 percent in classics, 55 percent in sociology, and 52 percent in law. Only in the biosciences was this pattern reversed: 56 percent of faculty who changed parties joined the Republican party. Overall, faculty converted to the Democratic party by a ratio of 2.98:1. For those who changed while in graduate school, the ratio equaled 3.97:1. The conversion ratio for those who changed from Republicans to Democrats after becoming a professor was 1.47:1.
Estimating Conversion Ratios
The data additionally afford us an alternative method of calculating conversion ratios. Specifically, respondents were asked about the political affiliations of their mothers and fathers. Using these measures, we found that only 41 percent of respondents with both parents who were Republicans remained Republican as professors, while 59 percent joined the Democratic party. Conversely, 81 percent of respondents with both parents who belonged to the Democratic party remained in the Democratic party as professors. Only 19 percent joined the Republican party. Using full parent data, we calculated a conversion ratio of 3.3:1 in favor of Democrats.
Political party conversion plays an underappreciated role in understanding overall political disparities across disciplines. Exactly why professors change their party affiliation is not well understood but prior analyses of these data showed that faculty reported changing parties because they agreed with the party platform and, interestingly, because of “knowledge gained in the profession.”17 When we examined responses to this question, we found that those who belonged to fields with more Democrats tended to join the Democratic party while those in disciplines characterized by political parity joined the Democratic or Republican party at roughly equal rates. Thus, we suspect that the political climate of a discipline imparts strong incentive pressures on faculty to join that party. If true, overall political disparities favoring Democrats could be sustained over time independent of levels of self-selection.
Lastly, we note that efforts to increase political and viewpoint diversity within the academy, but especially in the social sciences and the humanities, are likely to fail without explicit efforts to address the barriers to entry of individuals with heterodox views. These barriers may range from faculty political disparities, where, for example, centrist, libertarian, and conservative graduate students have no access to mentorship with faculty of similar beliefs or values, or they may be the product of partisan efforts to control entry into the discipline.18 Either way, barriers are likely to be a diverse mix of unintentional and intentional factors that coalesce into incentives and costs that differentially favor some students over others.
Addressing disincentives to entry, however, likely pales in comparison to altering broader disciplinary cultures that entice or reward political conversion in one direction but not another, or, worse still, that punish heterodox views. Having long ago exceeded the “slight numerical differential,” Lazarsfeld and Thielans hypothesized that this “may build up to a considerable effect on the uncommitted man.”19 Certain disciplines may be immune to or will simply reject internal efforts towards increased political diversity. This may be especially true for disciplines that have embraced political activism as a legitimate scholarly function, or for disciplines that are not merely overrepresented by faculty on the political left but by highly partisan faculty on the left. Nonetheless, our findings suggest that internal efforts to politically diversify the academy may have marginal effects at best, or, more likely, no measurable effects at all. Perhaps the best that can be hoped for, excluding legislative or policy intervention that makes political class a protected group, is that the academy more fully embraces tolerance and intellectual diversity.