The founding of Heterodox Academy had roots in a collaboration between five social psychologists and one sociologist that produced a featured paper, and 33 responses to it, in Behavioral and Brain Sciences. A number of specific recommendations to improve social psychology were made, but the main thesis was that increasing political diversity would improve the field by improving the quality of the research. Now, there is a growing literature on how political beliefs can impact the research process (see here, here, here, and here).
Yet, one persistent criticism (see here) has been that increased viewpoint diversity in the academy will pit ideological biases against each other and lead to further polarization. New research, summarized below, suggests that politically polarized editorial teams produced higher quality Wikipedia articles on politics, social issues, and science. Although the debates between politically polarized teams were more competitive, they were also longer, more constructive, and more substantially focused compared to politically homogenous teams and teams composed of political moderates.
Shi, F., Teplitskiy, M., Duede, E., & Evans, J.A. (working paper). “The wisdom of polarized crowds.” arXiv, 29 Nov. 2017
Shi, F., Teplitskiy, M., Duede, E., & Evans, J.A. (2019). “The Wisdom of Polarized Crowds.” Nature Human Behavior. DOI: 10.1038/s41562-019-0541-6
Abstract (emphasis added)
As political polarization in the United States continues to rise, the question of whether polarized individuals can fruitfully cooperate becomes pressing. Although diversity of individual perspectives typically leads to superior team performance on complex tasks, strong political perspectives have been associated with conflict, misinformation and a reluctance to engage with people and perspectives beyond one’s echo chamber. It is unclear whether self-selected teams of politically diverse individuals will create higher or lower quality outcomes. In this paper, we explore the effect of team political composition on performance through analysis of millions of edits to Wikipedia’s Political, Social Issues, and Science articles. We measure editors’ political alignments by their contributions to conservative versus liberal articles. A survey of editors validates that those who primarily edit liberal articles identify more strongly with the Democratic party and those who edit conservative ones with the Republican party. Our analysis then reveals that polarized teams—those consisting of a balanced set of politically diverse editors—create articles of higher quality than politically homogeneous teams. The effect appears most strongly in Wikipedia’s Political articles, but is also observed in Social Issues and even Science articles. Analysis of article “talk pages” reveals that politically polarized teams engage in longer, more constructive, competitive, and substantively focused but linguistically diverse debates than political moderates. More intense use of Wikipedia policies by politically diverse teams suggests institutional design principles to help unleash the power of politically polarized teams.
Shi, Teplitskiy, Duede, and Evans briefly review the literature on how political affiliations shape consumption of political and non-political (e.g., culture and science) news. They also connect the development of “echo chambers” to psychological processes (e.g., motivated reasoning, heuristic based processing), and how opposing social identities can make communication counter-productive.
The review then turns to literature that “demonstrates that individuals from socially distinct groups embody diverse cognitive resources and perspectives that, when cooperatively combined in complex or creative tasks produce ideas, solutions, and designs that outperform those from homogenous groups [22, 23, 24, 25]” (Shi et al., p. 2).
Thus, there are reasons to suspect that political diversity could help and hamper collaborative efforts in complex ways:
In short, political diversity should increase access to fresh perspectives and information but may also undermine quality of discourse and engagement required to enjoy the performance benefits typically obtained by diverse groups (p. 2)
Data and Methods
To investigate the impact of political diversity on collaborative performance, Shi et al. studied the performance of 400,000 Wikipedia editorial teams. These teams collaborated on English-language articles in three domains: Politics, Social Issues, and Science. Political affiliation was identified through edit histories, under the assumption that political partisans would contribute more to articles consistent with their partisanship1. A machine learning algorithm, developed by Wikipedia’s internal researchers, was then employed to rate the quality of each article assessed. Article quality, the assessment of which was based on Wikipedia’s internal guidelines, was then related to the political diversity of the editorial team.
The collaboration process was investigated by computationally exploring the article “talk pages” where editorial conversations occur and editorial decisions are made. Using the text from “talk pages” Shi et al. were also able to examine the relationship between political polarization and 3 additional factors: debate intensity, information diversity, and use of Wikipedia institutions (e.g., policies and guidelines).
I present some of the key findings of Shi et al. below in list format:
- For all three kinds of articles (Political, Social Issues, and Science) higher polarization of the editorial team was associated with higher quality.
- Political articles saw the greatest improvement in quality from polarization, followed by Social Issues articles, and then Science articles.
- Polarized editorial teams reduced talk page semantic diversity (distinct meanings or issues discussed), but increased lexical diversity (the number of ways meanings or issues were discussed).
- Polarized editorial teams generated a larger volume of debate.
- Polarized editorial teams when balanced, with roughly equal numbers representing rival perspectives, were better able to reduce flare-ups in debate temperature.
And, perhaps most importantly:
Compared with politically homogenous or skewed teams, polarized teams debate fewer topics with more competing terminology and framings. They engage in more debate, which is less acrimonious. And they more frequently appeal to Wikipedia policies and guidelines to govern these interactions. (Shi et al., p. 9).
This study appears to provide “the first empirical, real-world evidence that political polarization can lead to productive collaboration” (Shi et al., p. 9), and that “frequent, intense disagreement within politically polarized teams foments focused debate  and, as consequence, higher quality edits that are more robust and comprehensive” (Shi et al., p. 10). It should be noted that it remains possible that the people who edited articles on Politics, Social Issues, and Science were different kinds of people in some way from those that did not. In other words, it remains possible that randomly assigning editors into politically polarized teams will not produce the benefits reported by Shi et al.
This does not mean that the collaboration itself was devoid of acrimony. Respondents from politically polarized editorial teams reported “pervasive displeasure in having to convince obstinate, competing partisans of points that they took to be self-evident” (Shi et al., p. 10). Although, it does seem that when such disagreements occurred balanced competition reduced this displeasure, as editors made more appeals to Wikipedia policies and norms. A final analysis by Shi et al., which included a quadratic term for political affiliation, suggested that the benefits of political diversity appear to have an upper bound, however the optimal level of polarization was above 95% of the editorial teams investigated.
These findings provide empirical evidence for one of Heterodox Academy’s main ideas: Political diversity can improve collaborative efforts and produce better quality research that is more robust and comprehensive.
1 = Shi et al. validated this measure by also surveying a random subset of the Wikipedia editors included in the main sample and asking for their political party affiliation (strong Democrat to strong Republican).