Diagnosis versus Ideological Diversity

Ideational homogeneity recapitulates itself. Regardless of issues of fairness, the lack of ideological diversity within the discipline of political science has a negative influence on research, whether in the choosing of “interesting” or “relevant” research programs, the manner in which peer review proceeds, or the manner in which certain types of political phenomena and political actors are perceived. We have practical examples of how ideological homogeneity likely aided in major incidents in political science research in recent years, and we can note a tendency toward “diagnosis” of political ideologies that are not part of the homogeneity. The potential danger to the political science discipline’s aspiration to scholarly rigor and impartial examination of political phenomena is the creation of “blind spots” due to lack of ideological diversity.

IDEOLOGICAL BLIND SPOTS: TWO CASES

Two recent incidents within political science indicate the dangers inherent in a lack of ideological diversity: the LaCour fabrication (McNutt 2015), and the Verhulst, Eaves, and Hatemi erratum to their original article (Verhulst, Eaves, and Hatemi 2015). Generally, these two incidents can be viewed as illustrating certain problems: the former, of research dishonesty; the latter, as structural problems in peer review or simply as an outlier one-off instance with little relevance for the discipline in general. However, in both cases, we can surmise that a lack of ideological diversity created an environment in which these instances could occur, insofar as both cases fall into ideological blind spots within the discipline.

We can view “ideological blind spots” as referring to those topics, areas, and populations for which an intuitive “general consensus” exists among political scientists about the “correct” views. These blind spots need not be the result of malicious intent and may not even be conscious. Rather, they arise from structural pressures and peer dynamics. In a population in which the “typical” member is left-of-center to progressive in ideology, certain areas—inequality, race/class/gender, and so forth—are viewed as “serious” areas for scholarship, whereas areas that may be of more interest to right-of-center and conservative members—the benefits of tradition/custom, limitation in administrative powers, and the like—tend to be perceived as minor areas of focus, perhaps not terribly “serious” or interesting to the scholarly community. What is “serious” or “marginal,” then, is defined by a skewed peer population along the lines of ideology. This leads to structural pressures as well: these peers will be the “gatekeepers” of journals and publishers, judging “relevance” by the measures of the peer group. Over time, less scholarship of attraction to conservatives will appear whereas more that is attractive to progressives will be published. The end result is a homogenized peer group that assumes certain ideological presumptions as “givens,” not necessarily based on certain axioms or arguments but rather merely by the general consensus of a population sharing the same fundamental political assumptions. The consequence can be reviewers and others perceiving some results in research as “too good to check” because—in effect—they want to believe the results are true, regardless of whether the presented work is sound.

In the case of the Verhurslt, Eaves, and Hatemi (2015) article, an issue of codebook error in compiling various datasets resulted in the finding of a correlation between conservative political ideology and traits of psychoticism. These results received wide attention and have been cited in numerous journals, including Behavioral and Brain Sciences, American Political Science Review, Political Behavior, and Policy Studies Journal, among others. Their erratum noted that the finding actually is reversed—that a correlation was found between progressive ideology and traits of psychoticism. In the initial publication and in later erratum, the authors emphasized that a correlation of such traits does not equate with causation. However, there is concern about whether other researchers citing the original piece were as careful to make such a distinction. What is notable is how long it took for this mistake to be noticed. Whereas this may reflect problems in the peer-review process regarding the analysis of statistical methodology, it also reflects a potential problem of ideological blind spots. That is, in a disciplinary population overwhelming progressive in political perspective, it intuitively “makes sense” that conservatives—the Other—would share psychotic traits.

The LaCour fabrication provides another example (for an overview of the case, see Retraction Watch 2015). “When Contact Changes Minds: An Experiment on Transmission of Support for Gay Equality” was published in Science in 2014 (LaCour and Green 2014) and received significant attention, including notices in the New York Times, Washington Post, and This American Life, among others. The study supposedly illustrated that a 20-minute conversation with a gay canvasser was enough to create a large and ongoing change in attitude on same-sex marriage among residents in the County of Los Angeles. Although various peer reviewers or others noted some “concerns” regarding the work—given its significant variation from previous research and theory on political attitudes—it was accepted for publication,¹ and LaCour himself was offered a position at Princeton University. The fabrication came to light only when David Broockman and Joshua Kalla requested LaCour’s data—not to check its veracity but rather to expand on his findings (Broockman, Kalla, and Aronow 2015). In other words, the fabrication was revealed not by skepticism but instead by agreement. As the case gained notice through investigations of the Retraction Watch webpage, Green eventually requested that the article be retracted and LaCour’s position at Princeton did not materialize.

Consider an alternate scenario regarding this research. All of the relevant facts remain the same but with one major difference: rather than research “showing” that attitudes on same-sex marriage could be changed in a 20-minute conversation, the research illustrated that populations could have their views on abortion markedly changed to a pro-life direction during a 20-minute conversation. We can safely postulate that the fabrication would never had gotten as far as it did in these circumstances. All of the relevant facts remain—the radically different results from previous scholarship, the questions about data, and so forth—but the “ideological blind spot” is gone, or at least different. We can summarize the difference simply. In the LaCour and Green case, the ideological presumption was that “people like them” (i.e., conservatives, people not in support of same-sex marriage) do not actually have political thought per se and therefore a simple and quick conversation would be enough to dissuade them from their “irritable mental gestures which seek to resemble ideas” (in Lionel Trilling’s phrase [1950, xv]). In the hypothetical case, the fabrication would be caught much sooner because the ideological presumption is that “people like us” (i.e., left-of-center or progressive, pro-choice; that is, the dominant ideological tendency within political science) reason through our views and therefore “people like us” would not be so easily changed in their perspectives on such a fundamental issue.

These ideological blind spots, brought about by the sheer lack of ideological diversity within the discipline, thus can have detrimental effects on scholarship. The discussion thus far has illustrated the problem of “wanting to believe,” but there is a more serious issue as well: the Otherizing of political difference and the valorization of “our” ideological perspectives as the “norm” or “true” types of political rationality.

These ideological blind spots, brought about by the sheer lack of ideological diversity within the discipline, thus can have detrimental effects on scholarship.

DIAGNOSING CONSERVATIVES: “THREAT OR MENACE?”

It may be the case that the baseline assumption for many political scientists is based on left-of-center or progressive political views, which can include a major focus on equality (in various forms), an expansive notion of legitimate governmental (particularly administrative) powers (Gray and Mattingly-Jordan 2018), and others. In itself, such a structural bias may not be an overwhelming problem—but only if other dissenting voices also can engage. Whether such meaningful engagement currently exists within political science as a discipline is an open question. There is at least the perception that systemic biases and gatekeeping exist to curb ideological diversity within the humanities and social sciences generally (Hess and Bell 2019) and in political science specifically (Gilley 2017).

Epistemic closure is particularly harmful within scholarly pursuits, and a lack of ideological diversity likely will create such closure. One of the best means to prevent such closure is ideological diversity—not merely that members of the discipline can “mimic” the views of those with different ideological perspectives but also that there are members who actually believe these different perspectives. The notion that participation by underrepresented populations is certainly not foreign to political science; we need only consider recent work focused on sexual, racial, and ethnic diversity whether from major political science organizations (APSA 2018) or in applied research (Mershon and Walsh 2015). It is notable, perhaps, that the discipline thus far has shown no similar interest in ascertaining the problems from ideological underrepresentation, particularly of those who are right-of-center or conservative.

An outcome of this ideological homogeneity is the “otherizing” of the outside ideology—in this case, conservatives and those to the political right. Rather than political ideologies, movements, and actors to be considered, they are instead dysfunctions that require diagnosis rather than understanding. This “diagnostic” perspective reflects an ongoing framing in much political science (and other) research presumes that conservatism/rightism as a dysfunctional “sickness” stymies or distorts a “healthy” political society, whereas progressivism/leftism is a “natural” or “healthy” manner of politics. To the credit of political science as a discipline, blatant otherizing generally is avoided—at least in empirical studies—but the cases discussed so far indicate that the diagnostic presumption may be influencing what is published and how. It is not untoward to note that the issue of directionality in psychotic traits and the questions regarding ease of changing deeply held beliefs likely passed peer review so easily because of the diagnostic presumption. Of course, those against same-sex marriage could be easily swayed and, of course, conservatives correlate with traits of psychoticism. After all, “those people” are dysfunctional populations “we” correct-thinking people need to treat rather than as populations that have different worldviews. If such a diagnostic perspective applied broadly, the issue of ideological diversity might not arise. However, the significant dearth in similar published research on dysfunctional or undesirable traits correlated with progressive or liberal beliefs is notable. Ideological homogeneity in political science sets an otherizing presumption for conservatives and rightists: rather than studying them indifferently as simply other actors in the political scene, the presumption is that “those people” are either a threat (i.e., too ignorant or prejudiced to know what is good) or a menace (i.e., inherently inclined toward destructive policies and acts).

These points are not presented to address the “fairness” (or lack thereof) in the ideological homogeneity of the discipline. Rather, these issues touch on the rigor and standards used within political science as a scholarly discipline. Just as the discipline has recognized that there may be many ways—overt and covert—in which minority or marginalized voices may be minimized within the field, leading to systemic absences in our research, it is reasonable to have a similar concern about voices of the ideological minority. Whether such systemic absences exist and/or have a negative impact on our scholarship requires investigation, of course. However, because our discipline has at its foundation the study of the political, examining the consequences (potential and actual) of ideological homogeneity surely is a sensible concern.

Just as the discipline has recognized that there may be many ways—overt and covert—in which minority or marginalized voices may be minimized within the field, leading to systemic absences in our research, it is reasonable to have a similar concern about voices of the ideological minority.

DIAGNOSIS VERSUS IDEOLOGICAL DIVERSITY

We may be unmoved by the considerations thus far presented. Indeed, we might even proclaim that the form of “diagnosis” discussed is positive for political science as a discipline: rather, that the discipline thrives by not having ideological diversity. Perhaps this is true, but such an assessment should be argued with reasons and evidence rather than presumed based on the brute fact of an already-existent ideological homogeneity (because such a brute fact, of course, would not be accepted as a defense for maintaining homogeneity in other areas).

However, we must keep in mind the dangers to political science as a scholarly discipline if it continues down the path of diagnosis and ideological homogeneity. Ideological homogeneity recapitulates itself: as a profession becomes dominated by one general political perspective, it will be “pushed” into further and “harder” forms of that ideology. It is looking to other disciplines and seeing the results of increased ideological hegemony that give us pause. The 1990s saw the Sokol Hoax published in Social Research, in which the author used postmodern gibberish to “argue” that gravity is a social construct. Recently, another major set of hoaxes revealed a similar action in what the authors described as “Grievance Studies,” managing to publish ridiculous articles in Gender, Place, and Culture; Sexuality & Culture; and Hypatia, among others. A particularly notable instance was the acceptance by Affilia of “Our Struggle Is My Struggle: Solidarity Feminism as an Intersectional Reply to Neoliberal and Choice Feminism,” which simply reprinted parts of Chapter 12 of Adolf Hitler’s Mein Kampf, with “fashionable buzzwords switched in” (Lindsay, Boghossian, and Pluckrose 2018). The hope is to prevent political science from falling into similar intellectual disarray; however, such an end result cannot be avoided merely on hope alone.

These other disciplines provide a cautionary tale on the dangers of ideological homogeneity. Within our own discipline, the substantial danger would be a notable decrease in the quality and rigor of political science research over time. With dissenting views nonexistent or wrestling to overcome the gatekeepers in journals and editorial boards, an unquestioning assumption of certain presumptions or norms as “facts” will leave the discipline open to serious mistakes (as in the Verhulst et al. case) or fraud (in the LaCour case). As long as research outputs have results that an ideologically homogeneous “we” finds satisfactory and salutary, questionable research will be published.

Political science as a whole is not in this bad of shape—yet. However, the ideological homogeneity of the field, which shows little sign of stopping, presents reasons for concern. There is time to mitigate or curtail these problems, but the first step is to acknowledge the problem. Whether the discipline has the capacity or even the will to make such an acknowledgment is an open question.

A prudent step forward would be to investigate empirically the issues regarding ideological homogeneity. Thankfully, our discipline already has the means by which to undertake such a study. It is certainly feasible to use the methods and techniques of previous studies and reports, such as APSA’s (2018) “Diversity and Inclusion Report,” to examine issues of ideological homogeneity in the way it has examined racial, gender, and other forms of homogeneity. Such an undertaking would provide a significant benefit: that is, to determine whether comparative ideological homogeneity in the discipline is indeed a problem. If the investigation finds that it is not, then the discipline can turn to examining why a perception of such bias exists and can attempt to mitigate it. More to the point, if the investigation does determine that there is a problem, our profession can move to alleviate these biases, perhaps in ways similar to how the field has addressed other biases.

Alternatively, the discipline could simply not investigate this matter. However, for a field that is focused on politics and has put particular emphasis on flushing out other potential implicit biases within it, the lack of desire to look into potential ideological bias would be most revealing indeed.