Abstract
A lack of political diversity in psychology is said to lead to a number of pernicious outcomes, including biased research and active discrimination against conservatives. We surveyed a large number (combined N = 800) of social and personality psychologists and discovered several interesting facts. First, although only 6% described themselves as conservative “overall,” there was more diversity of political opinion on economic issues and foreign policy. Second, respondents significantly underestimated the proportion of conservatives among their colleagues. Third, conservatives fear negative consequences of revealing their political beliefs to their colleagues. Finally, they are right to do so: In decisions ranging from paper reviews to hiring, many social and personality psychologists said that they would discriminate against openly conservative colleagues. The more liberal respondents were, the more they said they would discriminate.
Research in social and personality psychology often bears directly on important political debates. Social–personality psychologists have studied the nature of prejudice and discrimination (Allport, 1954), the origins of ideology (Jost, Glaser, Kruglanski, & Sulloway, 2003), and the intuitive underpinnings of people’s moral convictions (Haidt, 2001). Thus, the political beliefs of researchers can have large consequences for research. Critics have argued that social–personality psychologists are overwhelmingly politically liberal (left-wing) and that this lack of diversity leads to ideologically biased selection of research questions, selective interpretation of evidence, and even to discrimination against conservative (right-wing) students and faculty (Haidt, 2011; Redding, 2001; Tetlock, 1994).
This issue seems to recur roughly every 10 years—most recently, in a provocative talk given by Jonathan Haidt (2011) at the annual meeting of the Society for Personality and Social Psychology (SPSP; APA Division 8). During his talk, Haidt asked the political conservatives present to raise their hands. In an audience of more than 1,000, only three hands went up. Haidt also described two other attempts he had made to locate conservatives in social psychology: searching the web using the term “conservative social psychologist” and asking 30 social psychologists to name a conservative colleague. Combined, these latter two methods uncovered one conservative social psychologist. Following earlier critics of psychology’s liberal bias, Haidt argued that this “statistically impossible lack of diversity” has serious negative consequences, including the unwillingness to consider “taboo” hypotheses and discrimination against politically conservative students. Haidt suggested setting the explicit goal that 10% of SPSP members be political conservatives by 2020.
The Political Ideology of Social and Personality Psychologists
Whether one agrees or disagrees with this goal, it is clearly a problem that we know so few of the relevant facts. First, we have little reliable data on the political ideology of social– personality psychologists. Haidt’s (2011) demonstration shows that very few are willing to openly identify as conservative. However, the picture may be quite different when individuals are asked to report their attitudes privately and anonymously. Furthermore, the liberal–conservative dichotomy may mask a great deal of meaningful variation. For example, one might hold liberal views on social issues (e.g., skepticism regarding cultural traditions and sexual restrictiveness) but more conservative views on economic issues (e.g., opposition to regulation and wealth redistribution) (Ball & Bellamy, 2003; Duckitt, 2001; Jost et al., 2003; Lipset, 1960/1981). A third important political domain is foreign policy, where conservatives tend to be hawks and liberals doves. This domain has received less attention from psychologists, but international security is among the most important criteria on which voters evaluate politicians (Hurwitz & Peffley, 1987; Mueller, 1986; Rosenwasser & Seale, 1988). Our first aim, therefore, was to assess the ideology of social–personality psychologists in each of these three domains and overall.
Accuracy in Perceived Ideology
Our second aim was to determine whether our respondents accurately perceived the ideology of their colleagues. There are several reasons to expect systematic inaccuracy. First, people generally believe that other members of their in-group are more extreme in their opinions than they themselves are. That is, people often believe that they form the more moderate part of a group (Keltner & Robinson, 1996; Robinson, Keltner, Ward, & Ross, 1995). Second, when people perceive a strong social norm—such as the norm of political liberalism among psychologists—they often do not express their (private) disagreement with it, without realizing that other people may be doing the same thing. The result is a norm that is seen as much more widely supported than it actually is (Prentice & Miller, 1996).
Reasons for Conservative Underrepresentation
To foreshadow our findings, there is more ideological diversity among social–personality psychologists than our respondents think—that is, they overestimate the field’s liberalism. In fact, conservatives seem to be a well-hidden minority. But why would conservatives hide their views? One salient explanation is a hostile political climate against conservatives (and even moderates). Nonliberals may feel unable to publicly express their views, and the liberal majority may actively discriminate against openly conservative individuals.
Obvious as this explanation might seem, it is not uncontroversial. For example, Gilbert (2011) argued that in the absence of data showing discrimination against conservatives, an equally plausible explanation for liberal overrepresentation is self-selection, as “liberals may be more interested in new ideas, more willing to work for peanuts, or just more intelligent, all of which may push them to pursue the academic life while deterring their conservative peers.” Similarly, Jacquet (2011) speculated that “there are very few conservatives in social psychology possibly not because it’s a hostile environment but because the field of social psychology self-selects for liberals and might even create them.” We do not doubt that self-selection may play some role (see Gross & Fosse, 2012), and we agree that the field’s liberal tilt is, on logical grounds, not enough to conclusively show a hostile climate or discrimination against conservatives (although such large statistical disparities are legally considered prima facie evidence of discrimination; Teamsters v. United States, 1977). We therefore investigated whether conservative social–personality psychologists do indeed experience a hostile climate and whether the more liberal might be willing to actively discriminate against their conservative colleagues.
Summary and Overview
In sum, we set out to answer three questions. First, how left-wing is the political ideology of social–personality psychologists? Second, do social–personality psychologists accurately perceive the ideology of their colleagues? Third, is there a perceived or actual hostile climate for—or even willingness to discriminate against—conservative social–personality psychologists? To answer these three questions, we contacted all 1,939 members of the SPSP electronic mailing list and invited them to complete two brief online surveys.
Survey 1
Survey 1’s primary purpose was to assess the political ideology of social–personality psychologists on social issues, economic issues, and foreign policy. We asked respondents to report their ideology (rather than asking about specific policy questions) because self-rated ideology is highly predictive of attitudes on specific issues (see Jost, 2006), because it allows straightforward classification of respondents (as liberal, moderate, or conservative), because it facilitates international comparisons, and because it maximizes response (by minimizing survey length).
Respondents
Of those contacted, 508 individuals participated (mean age = 36.8 years; 53.6% of the sample was female). This response rate (26.2% of those contacted) is close to that observed in previous studies using similar methods (e.g., Klein & Stern, 2005). Compared with the demographics of the entire SPSP membership in 2011 (gender, age, and country of residence), our sample represented the population quite well (see supplementary analyses in the Appendix).
Political ideology
Respondents were asked whether they considered themselves to be (1) Very liberal, (2) Liberal, (3) Somewhat liberal, (4) Moderate, (5) Somewhat conservative, (6) Conservative, or (7) Very conservative separately for “social issues,” “economic issues,” and “foreign policy.” As expected, respondents were liberal on average in all three domains. Yet as Figure 1 shows, we found an overwhelmingly liberal majority only on social issues: Here only a handful of respondents described themselves as moderate (5.5%) or conservative (3.9%). But in the two other domains, we found a considerable amount of diversity. On economic issues, 18.9% were moderates and 17.9% were conservative (i.e., right of Moderate). Similarly, on foreign policy, 21.1% were moderates and 10.3% were conservative.
Political diversity by domain (Survey 1). Histogram bars show response counts; the solid lines show the true sample mean, and the dotted lines show what participants imagined the mean to be. Pie charts show responses recoded as “liberal” (1–3), “moderate” (4), and “conservative” (5–7).
Correlations between self-ratings in each domain were medium to large, positive, and significant (all ps < .001). Economic conservatism correlated with social conservatism at r = .53 and with foreign-policy conservatism at r = .67; social and foreign-policy conservatism correlated at r = .53.
Others’ ideology
On a separate page, respondents used the same scales to rate their beliefs about the attitudes of the average social–personality psychologist in their country of residence. 1 As Figure 1 shows, participants believed the average social–personality psychologist to be more liberal than they themselves were on economic issues, M self = 3.1, M other = 2.9, F(1, 501) = 11.02, p = .001, d = .21. This was also true for foreign policy, M self = 2.9, M other = 2.6, F(1, 498) = 18.88, p < .001, d = .27. Only for social issues were perceptions of the average opinion accurate, M self = 2.1, M other = 2.1, F(1, 502) = 1.57, p = .21. (Variations in degrees of freedom across these questions are due to missing data.)
Summary
Social–personality psychologists are more ideologically diverse than they think. Although conservatives are a minority, they are a substantial minority on economic issues and foreign policy. It is important to note that ideology in the latter two domains has important consequences for psychological research. When Tetlock (1994) argued that the lack of political diversity leads to biased research, his two examples of such (putatively) ideologically biased research mapped onto these domains: White’s (1984) work on deterrence in international relations; and research on modern racism, which Tetlock argued improperly defined opposition to government regulation and subsidies as racist.
On both economics and foreign policy, respondents overestimated the liberalism of their colleagues. In our second survey, we examined why this might be the case—in other words, why the conservative minority might be hard to see. Specifically, we investigated whether more conservative social– personality psychologists conceal their views for fear of negative consequences. We also investigated whether these fears might be well founded.
Survey 2
Six months after our initial data collection, we again contacted all members of the SPSP electronic mailing list. We briefly describe the measures below along with the results. For the full version of each measure, please see the online supplemental materials at http://pps.sagepub.com/supplemental.
Respondents
Of those contacted, 292 individuals participated (mean age = 38 years; 58% of the sample was female). Complete data including demographics were obtained from 266. (Lower participation compared with Survey 1 is probably due to the fact that Survey 2 was described as taking 10 minutes rather than 3.) In the analyses, we used data from all participants who answered the relevant question(s), so degrees of freedom vary. As in Survey 1, our sample’s demographics reflected those of the SPSP membership (see Table 1).
Demographics (%) for Each Survey and the Entire SPSP Membership
Note. The Society for Personality and Social Psychology (SPSP) classifies members as “Undergraduate,” “Graduate,” or “Member”; thus, the only statistics available for comparison purposes in position demographic in the overall SPSP sample are for those classified as graduate students.
Political ideology
We measured political ideology with the same seven-point measures as in Survey 1 but now also added a fourth, “overall,” item. As in Survey 1, there was a conservative minority that was largest for economic issues and foreign policy (see Fig. 2), indicating that the two samples are comparable. In the following analyses, we use participants’ ratings of themselves “overall” as our primary measure of political ideology. For all analyses, combining all political ideology ratings into one composite measure (α = .91) yields nearly identical results.
Political diversity by domain (Survey 2). Responses are recoded as “liberal” (1–3), “moderate” (4), and “conservative” (5–7).
Hostile climate
We asked respondents how much they felt a hostile climate toward their political beliefs in their field, whether they were reluctant to express their political beliefs to their colleagues for fear of negative consequences, and whether they thought colleagues would actively discriminate against them on the basis of their political beliefs (all responses were on a seven-point scale: 1 = Not at all, 4 = Somewhat, 7 = Very much). Scores on a composite of these three questions (α = .93) correlated significantly with political orientation, r(289) = .50, p < .0001: The more conservative respondents were, the more they had personally experienced a hostile climate. Treating ideology as a categorical variable, conservatives experienced a significantly more hostile climate (M = 4.7) than did liberals (M = 1.9), t(17.61) = 5.97, p < .0001; or moderates (M = 3.7), t(30.43) = 2.06, p = .05. Moderates also experienced a more hostile climate compared with liberals, t(26.10) = 5.40, p < .0001.
Perceived hostile climate for conservatives
We asked respondents whether they felt that there was a hostile climate toward the political beliefs of social–personality psychologists “who would rate themselves as ‘somewhat conservative,’ ‘conservative,’ or ‘very conservative’ overall.” We used the same three hostile climate questions as above but with “conservative social–personality psychologists” as the target. Composite scores of perceived hostile climate for conservatives (α = .85) were significantly correlated with political orientation, r(263) = .28, p < .0001: The more liberal respondents were, the less they believed that conservatives faced a hostile climate. This correlation was driven entirely by more conservative respondents’ greater personal experience of a hostile climate: Controlling for personal experience, the relationship disappeared (r = −.01), suggesting that the hostile climate reported by conservatives is invisible to those who do not experience it themselves.
Stated willingness to discriminate
Participants answered four questions assessing their stated willingness to discriminate against conservatives. The first two questions asked whether, when a participant was reviewing a grant application or paper, a feeling that it took a “politically conservative perspective” would negatively influence the decision to award the grant or accept the paper for publication. The third asked whether the participant would be reluctant to invite “a colleague who is generally known to be politically quite conservative” to participate in a symposium (on an unspecified topic). The fourth question asked whether, in choosing between two equally qualified job candidates for one job opening, the participant would be inclined to vote for the more liberal candidate (over the conservative). For each of these four questions, respondents also indicated whether they thought that other members of their department would be willing to discriminate against conservatives. All responses were on seven-point scales (1 = Not at all, 4 = Somewhat, 7 = Very much). For each question, means differed significantly from 1, and a substantial percentage of respondents chose the scale midpoint (Somewhat) or above (see Table 2). Furthermore, the more liberal respondents were, the more they said they were willing to discriminate against conservatives on each question: paper reviews, r(279) = −.32, p < .0001; grant reviews, r(280) = −.34, p < .0001; symposium invitations, r(277) = −.20, p = .001; hiring decisions, r(279) = −.44, p < .0001.
Stated Willingness to Discriminate Against Conservatives
Note. All ratings were on scales from 1 (Not at all) to 7 (Very much), with the midpoint labeled 4 (Somewhat). All means differ significantly from 1 (ps < .001).
General Discussion
Like most scientists (Cardiff & Klein, 2005), social and personality psychologists are on average more liberal than the general population (Gallup, 2010). Yet we also find in two studies that their political ideology is more diverse than often assumed. On economics and foreign policy, a sizeable minority described themselves as moderate or conservative. Indeed, although only 6% described themselves as conservative “overall,” on economic and international affairs our sample already meets or exceeds the 10% “quota” that Haidt (2011) suggested as a 10-year target for ideological diversity. Why, then, did Haidt have such difficulty finding more than a handful of conservative colleagues? The current results suggest one answer: Members of the conservative minority are reluctant to express their political beliefs publicly. Survey 2 shows why: Hostility toward and willingness to discriminate against conservatives is widespread. One in six respondents said that she or he would be somewhat (or more) inclined to discriminate against conservatives in inviting them for symposia or reviewing their work. One in four would discriminate in reviewing their grant applications. More than one in three would discriminate against them when making hiring decisions. Thus, willingness to discriminate is not limited to small decisions. In fact, it is strongest when it comes to the most important decisions, such as grant applications and hiring. This hostile climate offers a simple explanation of why conservatives hide their political opinions from colleagues. Given that all academics depend on the opinions of their colleagues—who judge their papers, grants, and job applications—and given that such judgments are typically made by multiple reviewers (most of whom are liberal), this means that outspoken conservatives face a very serious problem. Hence, the more conservative respondents are, the more they hide their political opinions.
Limitations
One might argue that our discrimination questions overstate respondents’ willingness to discriminate, because we asked only whether they would evaluate papers and grant applications that seemed to take a conservative perspective negatively, not how they would evaluate work with a liberal perspective. That is, perhaps any political perspective is seen as improper. Three things speak against this: First, there is a priori no reason to believe that objections to taking a political perspective per se should be correlated with ideology, yet we find a strong correlation. Liberals are much more negative toward work with a conservative perspective. Second, the paper and grant review questions correlate positively with the other two discrimination questions (rs > .45), which are not susceptible to this alternative explanation: Research should be neutral, but researchers need not be. Finally, it is easier to detect bias in material that opposes one’s beliefs than in material that supports it (Lord, Ross, & Lepper, 1979). Work that supports liberal politics may thus seem unremarkable, whereas work that supports conservatism is seen as improperly ideological. There is also the more subtle possibility that respondents interpreted “conservative social psychologist” as describing a conservative activist bent on promoting a political agenda. 2 We cannot rule this out entirely, but we should note that we also asked participants for comments (see below), and none advanced this interpretation.
We believe it even more plausible that our measures underestimate willingness to discriminate, because respondents were asked directly and because presumably there is a strong norm against discrimination among psychologists. Consistent with this, perceived willingness to discriminate was even stronger when people were asked about the behavior of their peers—an indirect measure that bypasses some social desirability concerns.
There is good reason to believe that the results of these two surveys can be generalized to social–personality psychology as a whole: On important demographics—age, gender, and nationality—our sample is quite similar to the entire SPSP membership. This close correspondence suggests that it is unlikely that nonresponders systematically differed from responders, although we cannot rule out this possibility entirely (we should note that this is the case for any survey research; see, e.g., Keeter, Kennedy, Dimock, Best, & Craighill, 2006). We also looked for—and were unable to find—differences between those who described their area as primarily or exclusively “social psychology” and those who described it as primarily or exclusively “personality psychology.” The findings described here thus seem to be equally descriptive of psychologists in both areas (although limited, of course, to those who are SPSP members).
What about other areas within psychology? Here we have little firm empirical evidence. Cardiff and Klein (2005) obtained the political party registrations of 4,563 California faculty and found that among all psychology faculty (N = 295), Democrats outnumbered Republicans at a ratio of 8:1. This might seem high, but in Survey 2, we found a liberal to conservative ratio of almost 14:1 (for ratings of ideology “overall”). Keeping in mind the inherent problems of comparing these two very different samples (and of assuming that “liberal” and “conservative” self-descriptions are reliable proxies for Democratic and Republican voter registration), this comparison suggests that social–personality psychologists might be even more liberal on average than psychologists overall.
Implications
Is it a problem that conservative political opinion is not tolerated? If one believes that conservatives are simply wrong, perhaps not. After all, geologists are not obliged to accept colleagues who believe the earth is flat. But political or moral beliefs often do not have a truth value. A belief that the earth is flat is factually false; a belief that abortion should be prohibited is not. Neither is a belief that cultural traditions should be respected or that economic inequality is acceptable. It may also be that many aspects of conservative thinking can serve as inspiration for interesting research questions that would otherwise be missed. Finally, as offensive as it may seem to many (liberal) social psychologists, believing that abortion is murder does not mean that one cannot do excellent research.
But what is perhaps the most important issue—an issue that we believe most psychologists, regardless of political background, will find troubling—is that conservatives are barred from discussion and are forced to keep their political opinion to themselves, coupled with a denial of the severity of this issue. We found this in our quantitative data, as described in this article. But perhaps even more telling is what we found in our qualitative data. At the end of our surveys, we gave room for comments. Many respondents wrote that they could not believe that anyone in the field would ever deliberately discriminate against conservatives. Yet at the same time we found clear examples of discrimination. One participant described how a colleague was denied tenure because of his political beliefs. Another wrote that if the department “could figure out who was a conservative they would be sure not to hire them.” Various participants described how colleagues silenced them during political discussions because they had voted Republican. One participant wrote that “it causes me great stress to not be able to have an environment where open dialogue is acceptable. Although most colleagues talk about tolerance, and some are, there are a few vociferous voices that make for a closed environment.” Some respondents wrote that at times they felt reluctant to ask certain questions because these hinted at a conservative identity. But an environment that stimulates open discussion and where hypotheses may be raised regardless of their political implications can only benefit our field. Even those who fundamentally disagree with conservatism will agree that silencing political opponents will not convert them. By excluding those who disagree with (most of) us politically, we treat them unfairly, do ourselves a disservice, and ultimately damage the scientific credibility of our field.
Footnotes
Appendix
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.