Duarte et al. document that the majority of psychological researchers in the United States are politically liberal. They present illustrative cases where the researchers' liberal orientations have led to biased and tendentious research. A corresponding bias could easily be found in economics or other fields where the majority of researchers are conservative (Zipp & Fenwick 2006); whether the issue is particularly severe in social psychology is hard to tell. Bad science is ubiquitous (Goldacre Reference Goldacre2008). To combat bad science, it is essential to use scientific rigor to identify scientific errors, methodological flaws, and unfounded claims. We strongly support attempts to improve current reviewing procedures and scientific self-correction mechanisms.
Duarte et al., however, propose a different strategy. They argue that increasing political diversity in a research environment will improve scientific validity. Specifically, they assume that politically diverse positions “cancel out each other” (sect. 4.1, para. 4) and that a mix of politically opposing positions in a research environment will generate better approximations to scientific truth. We strongly disagree with this “cancelling-out” hypothesis for several reasons.
First, we can find no empirical evidence to support this assumption in the academic domain. Even if diversity is beneficial in some domains outside of science, as the examples from organizational psychology cited by Duarte et al. suggest, it does not follow that this would be beneficial for science.
Second, the call for greater diversity is commonly motivated by a desire to increase social justice and equity rather than to search for scientific truth. Facilitating the access of, say, women to academia is a political issue; nobody should be discriminated against, for example, by being excluded from an academic position on the basis of features such as gender or ethnicity. The question of whether the gender composition of a research environment has an effect on scientific quality is not part of the discrimination argument. In fact, discrimination based on gender or ethnicity is considered unfair precisely because it is generally assumed that gender or ethnicity have no bearing on academic achievement.
Third, several bizarre conclusions follow from the cancelling-out hypothesis. For example, will collaboration between evolutionary theorists and Intelligent Design advocates cancel out their respective biases and generate a more truthful theory somewhere in the middle? Scientific truth is not a matter of political diversity and compromises unless one assumes a radical constructivist position (Lennon Reference Lennon1997). And why limit diversity to political diversity? Why not increase religious diversity and add a religious fundamentalist to a psychology department dominated by atheists, hoping that their orientations will cancel out? We cannot see how this combination would improve scientific outcomes. The history of science rather demonstrates that religious or political diversity is a hindrance to scientific progress; the role of the church in the great scientific revolutions from Galileo to Darwin may serve as a case in point. Instead of cancelling each other out, those with opposing political viewpoints will likely denigrate each other, and their particular biases will be stretched to greater extremes. It seems that Duarte et al. are neglecting a crucial distinction: Political diversity is a manifestation of conflicts of interest, not of biased knowledge, and a compromise of interests does not imply a convergence on truth.
Fourth, the cancelling-out hypothesis suggests that all political orientations are comparable and on par with one another, in particular with respect to their stance on the scientific method and their ability and willingness to contribute to scientific research. However, some political orientations, most notably those that are closely associated with religious beliefs, are in effect opposed to the scientific method as a privileged route to knowledge (Gauchat Reference Gauchat2008). Of course, there has always been academic debate about what constitutes a legitimate scientific method. But although science is in a continuous state of flux, there is an accepted core of legitimate methodology; the case of Intelligent Design is a good example of an attempt to gain trustworthiness and political influence by declaring oneself to be genuinely “scientific.” If we wanted to increase political diversity in our institutions, who would decide which parties to admit to the diversity mix and which to exclude? What are the criteria for determining whether a political orientation will contribute to the cancelling-out mechanism?
There is a more fundamental argument for why we see the call for political diversity in science as misguided and ultimately as politically dangerous. Simply speaking, one cannot choose one's gender, skin color, or ethnicity. Which party to vote for in an election is, by contrast, a matter of choice. The idea that people can freely change their vote as personal preferences or political circumstances change is a pillar of democratic societies. We are not born with our political partisanship, albeit some findings suggest a – weak and mediated – genetic influence on political orientations (Oskarsson et al. Reference Oskarsson, Cesarini, Dawes, Fowler, Johannesson, Magnusson and Teorell2014). Treating political categories such as liberal or conservative as if they were categories like gender and ethnicity – that is, genetically determined and immutable – is committing a kind of naturalistic fallacy. Recommending, as Duarte et al. do, that the political composition of academic teams be actively regulated is to falsely take political orientations as “given” – as facts of nature. What if researchers change their political orientation over time, and what about the possibility that this change may occur as a result of their scientific pursuit itself? We think that any attempt to externally control the degree of political diversity in a group of thinking and developing individuals is doomed to fail.
Why is it politically dangerous to try to actively increase political diversity? Selecting candidates according to political orientation, be it for an academic or any other type of position, has rightly been viewed as a distinguishing hallmark of totalitarian regimes. We do not insinuate any such intentions to Duarte et al.; however, selecting for political diversity necessarily implies assessing individuals' political orientations. And who assesses the assessor? The very idea of political selection bears the seed of political control, of abuse, and of fabricating academic careers that are uncorrelated with scientific achievement.
Duarte et al. document that the majority of psychological researchers in the United States are politically liberal. They present illustrative cases where the researchers' liberal orientations have led to biased and tendentious research. A corresponding bias could easily be found in economics or other fields where the majority of researchers are conservative (Zipp & Fenwick 2006); whether the issue is particularly severe in social psychology is hard to tell. Bad science is ubiquitous (Goldacre Reference Goldacre2008). To combat bad science, it is essential to use scientific rigor to identify scientific errors, methodological flaws, and unfounded claims. We strongly support attempts to improve current reviewing procedures and scientific self-correction mechanisms.
Duarte et al., however, propose a different strategy. They argue that increasing political diversity in a research environment will improve scientific validity. Specifically, they assume that politically diverse positions “cancel out each other” (sect. 4.1, para. 4) and that a mix of politically opposing positions in a research environment will generate better approximations to scientific truth. We strongly disagree with this “cancelling-out” hypothesis for several reasons.
First, we can find no empirical evidence to support this assumption in the academic domain. Even if diversity is beneficial in some domains outside of science, as the examples from organizational psychology cited by Duarte et al. suggest, it does not follow that this would be beneficial for science.
Second, the call for greater diversity is commonly motivated by a desire to increase social justice and equity rather than to search for scientific truth. Facilitating the access of, say, women to academia is a political issue; nobody should be discriminated against, for example, by being excluded from an academic position on the basis of features such as gender or ethnicity. The question of whether the gender composition of a research environment has an effect on scientific quality is not part of the discrimination argument. In fact, discrimination based on gender or ethnicity is considered unfair precisely because it is generally assumed that gender or ethnicity have no bearing on academic achievement.
Third, several bizarre conclusions follow from the cancelling-out hypothesis. For example, will collaboration between evolutionary theorists and Intelligent Design advocates cancel out their respective biases and generate a more truthful theory somewhere in the middle? Scientific truth is not a matter of political diversity and compromises unless one assumes a radical constructivist position (Lennon Reference Lennon1997). And why limit diversity to political diversity? Why not increase religious diversity and add a religious fundamentalist to a psychology department dominated by atheists, hoping that their orientations will cancel out? We cannot see how this combination would improve scientific outcomes. The history of science rather demonstrates that religious or political diversity is a hindrance to scientific progress; the role of the church in the great scientific revolutions from Galileo to Darwin may serve as a case in point. Instead of cancelling each other out, those with opposing political viewpoints will likely denigrate each other, and their particular biases will be stretched to greater extremes. It seems that Duarte et al. are neglecting a crucial distinction: Political diversity is a manifestation of conflicts of interest, not of biased knowledge, and a compromise of interests does not imply a convergence on truth.
Fourth, the cancelling-out hypothesis suggests that all political orientations are comparable and on par with one another, in particular with respect to their stance on the scientific method and their ability and willingness to contribute to scientific research. However, some political orientations, most notably those that are closely associated with religious beliefs, are in effect opposed to the scientific method as a privileged route to knowledge (Gauchat Reference Gauchat2008). Of course, there has always been academic debate about what constitutes a legitimate scientific method. But although science is in a continuous state of flux, there is an accepted core of legitimate methodology; the case of Intelligent Design is a good example of an attempt to gain trustworthiness and political influence by declaring oneself to be genuinely “scientific.” If we wanted to increase political diversity in our institutions, who would decide which parties to admit to the diversity mix and which to exclude? What are the criteria for determining whether a political orientation will contribute to the cancelling-out mechanism?
There is a more fundamental argument for why we see the call for political diversity in science as misguided and ultimately as politically dangerous. Simply speaking, one cannot choose one's gender, skin color, or ethnicity. Which party to vote for in an election is, by contrast, a matter of choice. The idea that people can freely change their vote as personal preferences or political circumstances change is a pillar of democratic societies. We are not born with our political partisanship, albeit some findings suggest a – weak and mediated – genetic influence on political orientations (Oskarsson et al. Reference Oskarsson, Cesarini, Dawes, Fowler, Johannesson, Magnusson and Teorell2014). Treating political categories such as liberal or conservative as if they were categories like gender and ethnicity – that is, genetically determined and immutable – is committing a kind of naturalistic fallacy. Recommending, as Duarte et al. do, that the political composition of academic teams be actively regulated is to falsely take political orientations as “given” – as facts of nature. What if researchers change their political orientation over time, and what about the possibility that this change may occur as a result of their scientific pursuit itself? We think that any attempt to externally control the degree of political diversity in a group of thinking and developing individuals is doomed to fail.
Why is it politically dangerous to try to actively increase political diversity? Selecting candidates according to political orientation, be it for an academic or any other type of position, has rightly been viewed as a distinguishing hallmark of totalitarian regimes. We do not insinuate any such intentions to Duarte et al.; however, selecting for political diversity necessarily implies assessing individuals' political orientations. And who assesses the assessor? The very idea of political selection bears the seed of political control, of abuse, and of fabricating academic careers that are uncorrelated with scientific achievement.