Most people, but especially political liberals, view diversity of almost any kind as an intrinsic good. But Duarte et al. recognize that greater diversity of political views in social psychology should not be seen as an end in itself. In no way diminishing contemptible cases of politically conservative students made to feel unwelcome in our field, the preeminent value of diversity in this case is its potential to produce better science. Duarte et al.'s core argument is that a more politically diverse social psychology will serve as an antidote to liberal bias and help the field more quickly and efficiently “converge upon the truth” (sect. 1, para. 5).

Their argument rests on two key assumptions. The first is that social psychological research is widely vulnerable to political bias. While only a small percentage of social psychological research has an explicitly political focus, it is important to remember that only a few decades ago climate science would have seemed irrelevant to partisan politics. As partisan hostility increasingly insinuates itself into everyday American life (Iyengar & Westwood Reference Iyengar and Westwood2014), its potential to ensnare previously apolitical scientific questions in the web of the ongoing culture war will grow as well. Moreover, social psychological research is uniquely susceptible to political bias because its fundamental motivating assumption – that human behavior and outcomes are largely determined by social forces – lies precisely on the intellectual fault line of left–right ideological conflict. Any research that bears on the role of individual versus situational determinants of human outcomes is vulnerable. It is hard to dispute, for example, that liberal sympathies in social psychology contributed to the field's initial reluctance to accept research demonstrating substantial genetic contributions to intelligence and personality (e.g., Kamin Reference Kamin1974).

More formally, the persistent intuition that political ideology biases the interpretation of scientific data has now been confirmed by dozens of experiments over the past five decades (Lord & Taylor Reference Lord and Taylor2009; MacCoun Reference MacCoun1998), and there is little reason to believe that social psychological researchers are immune to these effects. While some evidence suggests that liberals are dispositionally less prone than conservatives to motivated reasoning (e.g., Jost et al. Reference Jost, Glaser, Kruglanski and Sulloway2003), a recent meta-analysis by our research group examining more than 30 studies of politically biased evidence evaluation found clear evidence of partisan bias in both liberals and conservatives, and at virtually identical levels (Liu et al. Reference Liu, Clark, Wojcik, Chen, Grady and Ditto2014). Moreover, several studies have provided intriguing evidence that partisan bias becomes more rather than less pronounced with greater topical knowledge and cognitive sophistication (e.g., Kahan et al. Reference Kahan, Peters and Dawson2013; Taber & Lodge Reference Taber and Lodge2006), as does a general insensitivity to one's own biases (West et al. Reference West, Meserve and Stanovich2012). Early studies documenting biased evidence evaluation used psychology researchers as participants (e.g., Abramowitz et al. Reference Abramowitz, Gomes and Abramowitz1975; Mahoney Reference Mahoney1977), and this more recent research further confirms that high levels of knowledge, intelligence, and perceived objectivity do not necessarily provide protection from bias, as most people likely assume. Instead, they may simply allow scientists with strong ideological commitments to unknowingly deploy their considerable cognitive skills in biased fashion to become particularly resistant to attacks on those commitments.

Thus, we agree with Duarte et al. regarding the potential for political bias to impede the progress of scientific discovery in social psychology. We have considerably less confidence, however, in their subsequent assumption that increasing the representation of non-liberals in the field will effectively address the problem.

There is certainly wisdom in Duarte et al.'s assertion that increasing the number of conservative social psychologists would increase the likelihood of identifying flaws in research with embedded liberal biases. The anecdotal examples of liberal bias they cite are consistent with research on motivated skepticism (Ditto & Lopez Reference Ditto and Lopez1992; Ditto et al. Reference Ditto, Scepansky, Munro, Apanovitch and Lockhart1998) showing that a primary source of biased judgment is our tendency to uncritically evaluate information that confirms our prior beliefs and preferences.

But political bias is both implicit and tenacious, and there is little reason to believe that either liberal social psychological researchers, or any newly minted conservative ones, will be easily disabused of the tendency to expect and prefer empirical results that confirm their political views, and find flaws in results that do not. Social psychology has seen many theoretical controversies and data have resolved few of them (Greenwald Reference Greenwald2012). Increasing the minority influence of conservatives in the field may lead to more diverse viewpoints being represented in the literature and a more challenging peer-review process, but rather than leading the field to converge on some universally accepted “truth,” it seems more likely to engender theoretical conflict and a divided literature, with each side defending their operationalizations, methods, and data while disparaging those of the other side. Calls for greater civility and scientific humility are valuable, but another fear is that a prevailing liberal bias will be replaced by an “equivalency bias” favoring the view that liberals and conservatives are equally bestowed with psychological strengths and weaknesses. This may ultimately prove to be the case, but it may not, and defaulting to such an equivalency bias in place of a liberal one will leave our science no better off.

In the mid-1900s, psychologists were optimistic that integration by itself would improve interracial relations, until research and real-world experience revealed that contact produces beneficial results only under specified conditions (Dovidio et al. Reference Dovidio, Gaertner and Kawakami2003). Analogously, additional efforts will be required to approximate a social psychology free of political bias, and there is important convergence here with ongoing efforts to acknowledge and combat researcher bias more generally (Simmons et al. Reference Simmons, Nelson and Simonsohn2011).

Duarte et al. offer an important critique and some initial plans of attack, but the challenge remains to develop strategies that allow the signal of data to rise above the noise of ideological conflict. Making our field more welcoming to scholars of all political persuasions is intrinsically right, and it will surely lead to new questions and novel data. But in times so partisan, and for a field as entangled in ideology as social psychology, convergence upon the truth is likely more than even liberals can expect from diversity.