Attitudes toward and Engagement with Misinformation

• In Event: Poster Session: Public Opinion and Comparative Politics

Abstract

There is a large volume of extant research on misinformation in elections in western democracies. However, less is known about misinformation in countries in the Global South. This study addresses this gap in the literature with evidence from India. This research explores two aspects of misinformation in the Indian electoral context: 1) attitudes toward misinformation and 2) engagement with misinformation. For the former, I use face-to-face survey data from the Pew Research Center study (n = 3,490) conducted several months before the 2019 election in fall 2018. I first consider the reported importance of various media for political information as dependent variables. Models predict the importance of newspapers, TV, radio, social media, and in-person discussion, controlling for demographics and reported closeness to a party. Results show that partisans, compared to non-partisans are likely to think access to “mobile phones, internet and social media have made people easy to manipulate with false information and rumors” which was the actual question wording. People who rely on newspapers, television and in-person discussions for political news are also more likely to be concerned about misinformation.

Using the Pew data, I then model attitudes toward misinformation to estimate the effect of closeness to a party, and the importance of traditional and social media for political news, controlling for demographics. Attitudes were measured with the question:
“Let’s think for a moment about some experiences that might happen to people when they use their mobile phones. Do you think people should be concerned about exposure to false or incorrect information or not? And is that very concerned or only somewhat concerned?” The response scale ranged from 1=not concerned 2=somewhat and 3=very concerned. Results show that BJP partisans, compared to non-partisans, are more likely to be concerned about exposure to false or incorrect information whereas INC partisans and other partisans were not significant.

In the second section of the paper, I use Facebook posts labelled by fact-checker firms as misinformation to show the extent to which Facebook users engage with misinformation. I obtained misinformation-labelled Facebook posts (N = 4,478) during the 2019 Indian national election campaign that I identified using fact-checked stories from the websites of 10 different fact-checker firms. I content analyzed each of these posts to determine the tone of the post and the target, e.g., if it was positive or negative toward the target party. A second trained coder tested 15% of posts, taking a random sample, for reliability which was high (Cohen’s Kappa = 0.93). Each post is labelled pro-party or anti-party with respect to the party targeted in the post. Results show that many political parties – at least 5 parties – and their affiliates or supporters deployed both positive or pro-party and negative or anti-party misinformation posts on Facebook. Models of engagement with misinformation predicting Shares, Likes and Comments, show that compared to pro- or anti- misinformation posts about other parties, pro-BJP misinformation was most often shared and liked, and interaction effects were significant: Pro-BJP misinformation posts that mentioned incumbent Narendra Modi were highly shared, and pro-INC misinformation posts that mentioned lead challenger Rahul Gandhi were highly commented on and liked. However, anti-BJP posts that mentioned Modi were more often shared while anti-INC posts that mentioned Gandhi were more often liked.

I discuss the implications of these findings for understanding social media strategies employed by political parties and their affiliates who were using misinformation as a campaigning tool. Overall, this study suggests that respondents to the Pew survey in India were right to be concerned about misinformation in the upcoming 2019 campaign, because on Facebook, which has some 410 million users, political parties and their affiliates often used posts containing pro- or anti-party misinformation claims.

Author

• Kiran Arabaghatta Basavaraj, University College London