Exposure to extremely partisan news from the other political side shows scarce boomerang effects. Andreu Casas, Ericka Menchen-Trevino, Magdalena Wojcieszak. Oct 2020. http://andreucasas.com/CasasEtAl-ExtremeSitesExperiment.pdf
Rolf Degen's take: https://twitter.com/DegenRolf/status/1322422135873298433
Abstract: Research shows that a narrow information diet may be partly to blame for the growing political divides in the United States, suggesting exposure to dissimilar views as a remedy. These efforts, however, could be counterproductive, exacerbating attitude and affective polarization. Yet findings on whether such boomerang effect exists are mixed and the consequences of dissimilar exposure on other important outcomes unexplored. To resolve this debate, we designed an experiment, in which one should certainly observe boomerang effects. We incentivized liberals to read political articles on extreme conservative outlets (Breitbart, The American Spectator, and The Blaze) and conservatives to read extreme left-leaning sites (Mother Jones, Democracy Now, and The Nation). We explored the effects on attitude and affective polarization, as well as on perceptions of the political system, support for democratic principles, and well-being. Overall we find little evidence of boomerang effects, suggesting they are an exception rather than the norm.
7 Discussion
In the US, greater harmony between different political factions is needed more than ever. To achieve this ever-eluding goal, scholars and practitioners encourage exposure to dissimilar political views, with the hope that encountering views that challenge one’s beliefs will minimize extremity and interparty hostility. Although some scholars caution against this approach, suggesting that dissimilar exposure can increase polarization, the findings about the existence of such boomerang effects are mixed and limited in scope. Given the crucial societal and political implications of this largely inconclusive debate, we set out to solve it with an innovative experimental design combining incentivized over time exposure to extreme news domains from across the political aisle (Breitbart, TheAmerican Spectator, and The Blaze for liberals; and Mother Jones, Democracy Now, and The Nation for conservatives), pre-, post-, and intermediate surveys, and trace data on actual online exposure of the participants. Although this design is highly counterfactual (after all, most liberals are unlikely to regularly visit Breitbart), it was well suited to detecting boomerang effects if these are in fact a likely outcome of exposure to dissimilar views. The design also allowed us to test whether the studied dissimilar exposure has effects on broader societal outcomes and on individual well-being, and also among those for whom these effects emerge (attending to a systematic set of political predispositions to ascertain potential heterogeneous treatment effects). In short, despite the over time nature of the treatment (i.e., fourteen days), accounting for intended treatment effects as well as the levels of compliance (see Bail et al. (2018)), and testing attitude polarization on a range of salient political issues and affective polarization with several indicators and toward various outgroups, we show that dissimilar exposure is unlikely to intensify political conflict or have any discernible effects on the societal and individual outcomes tested. After going to extreme news sites of the opposing ideology every second day for a twoweek period, people did not radicalize their issue attitudes nor their feelings towards the out-party and the supporters o the opposing ideology. Although we did find that people slighlty polarized their perceptions of those holding opposing views on a few political issues (such as climate change and immigration), these effects were not systematic across different measurements and did not generate a pattern that would suggest the existence of relevant boomerang effects ( < 0.2 standard deviation changes). Furthermore, although many observers fear that strong partisans are most likely to radicalize and drive political conflict (Garrett et al., 2014) and the work on motivated reasoning suggest that individuals with strong priors are most likely to counter-argue dissimilar information and become more extreme as a result (Taber and Lodge, 2006), we do not find pronounced heterogeneous effects. Because we tested party-, ideology-, and political identity strength as relevant covariates, we are confident that these largely null effects are not due to any specific measurement. In a similar vein, our treatment did little to influence participants’ perceptions of the political system, in terms of their support for interparty compromise, attributing malevolent intentions to the outparty, or seeing the polity as polarized. It also did not shift their support for key democratic principles, such as freedom of speech or freedom of press (even though our pre-registered expectation was that those exposed to our treatment would be more inclined to ban their political opponents from the media and to have search engines and social media platforms avoid displaying or promoting articles from some media outlets). Relatedly, extreme dissimilar exposure also did not significantly worsen participants’ wellbeing, even though – again – we predicted that it would make them feel more anxious or dissatisfied or increase negative or unhealthy behaviors. It is crucial to emphasize that apart from not being statistically significant, all the observed average effects are of a very small magnitude (¡ .2 and ¡ .1 standard deviation changes). The findings are a great contribution to the existing literature and theorizing on the potential negative effects of exposure to counter-attitudinal information. Contrary to some evidence, which finds exposure to opposing views to exacerbate attitude and affective polarization (Levendusky, 2013; Bail et al., 2018; Garrett et al., 2014), and in line with other existing work (Guess and Coppock, 2018; Wood and Porter, 2019), we conclude that these types of boomerang effect are the exception rather than the norm (and of a very small magnitude if they do emerge). Extending past work by incorporating people’s evaluations of the outlets and accessed articles (based on both short surveys and also their open-ended thoughts and emotions), we conclude that this consistent lack of boomerang effects may be due to people’s largely neutral or even positive reactions to the outlets and their content. We wanted to test the effect of an extreme counter factual and selected these 6 news sites because they are located at the extreme of each ideological side (with the exception of very minor fringe/niche sites). Nevertheless, despite representing the extreme of each ideological side, and despite often being vilified by one’s partisan group, the partisans we studied often valued the information they consumed in them. In addition, this study also makes a relevant contribution to the growing body of work that uses trace data to study people’s attitudes and behavior. Rather than relying on a forced exposure experiment that shows people mock sites with counter-attitudinal news articles, we incentivized exposure, accounted for compliance, and exposed them to real news stories that actually appeared in news outlets of the opposing ideology. The trace data donated by the participants some weeks later additionally allowed us to assess compliance using precise behavioral data, complementing the self-reported compliance measures we collected. Furthermore, because we analyzed the rich qualitative data contributed by the participants in response to the news they read, we are able to shed important insight into participants’ thoughts and feelings generated by extreme exposure to the other side, insight that was previously unavailable as virtually all existing work on boomerang effects relies on close-ended survey responses. It is designs like ours (combining online traces, systematic experimental treatments, self-reports at several time points, and unrestricted reactions to content) that are most apt to accurately portraying the existence (or rather lack thereof) of boomerang effects. At a time where key stakeholders such as social media companies (Farr, 2018-10-16; Wood and Ethan, 2020), news organizations (Goodman and Chen, 2010), and governments (Rendall, 2015; Commission, 2013) are designing policies decreasing or increasing exposure to dissimilar views in order to reduce mass polarization, we strongly believe that the findings reported here can help inform the decision-making process moving forward.