Beyond bubble bursting

Simon Knight writes on bias and diversity.

16 August 2017

It was good to see both Christian Jarrett’s recent Research Digest coverage of confirmation bias in political discourse and Poppy Noor’s excellent piece with Michal Kosinski on information bubbles (June 2017). In the Research Digest, Jarrett draws attention to the personal choices we make (avoiding a focus on algorithmically instigated filter bubbles), flagging that (1) many of us live in information bubbles, and (2) actively avoid exposure to views that we do not agree with.

The two pieces are somewhat at odds though, as Michal flags, and I argue in a recent Australasian Science op-ed, the evidence on both of these claims is not particularly strong in naturalistic contexts (social media use and media consumption). Evidence supports the claim that we have a preference for information that affirms our prior beliefs. However, most people are pretty centrist in their media consumption, with social media potentially exposing us to – rather than insulating us from – a diversity of perspectives.

Moreover, there are concerning implications of emphasising deliberate exposure to diverse views. My recent online piece for this magazine discussed some cognitive issues that people face in evaluating and reconciling multiple sources. In addition, a recent paper by Pennycook et al. indicates that ‘prior exposure increases perceived accuracy of fake news’ – i.e. they find novel support for the concern that exposure to falsehoods increases the likelihood that those falsehoods are believed. This ‘mere exposure’ strategy has been effectively used by climate change deniers and tobacco lobbyists amongst others to manufacture false equivalence leading to lasting damage. I wondered, then, if instead of exposure to ideological diversity, we should be more concerned with demographic diversity. In the former case, we risk encouraging exposure to the small group of deniers who repeat information from single sources, manufacturing a false equivalence and maintaining uncertainty where little exists. In the latter, we would instead focus on the ways in which lack of diversity in media sources silences and marginalises groups and the implications of that.

We should be prepared to engage with other’s ideas, but this must be with an eye on the evidence. A recent US report found that, regardless of political affiliation and demographic factors, those who placed value on evidence were less partisan on contentious issues like climate change. Other work from researchers, including Dan Kahan, has recently indicated the importance of curiosity in not only reading (mere exposure), but being critically receptive to others’ perspectives. Rather than a focus on ‘bursting our bubble’ generally, we need strategies that encourage engagement with evidence.

Simon Knight
University of Technology Sydney