Priming and Fake News: The Effects of Elite Discourse on Evaluations of News Media

Abstract

Fake news has become a prominent topic of public discussion, particularly among elites. Recent research has explored the prevalence of fake news during the 2016 election cycle and possible effects on electoral outcomes. This scholarship has not yet considered how elite discourse surrounding fake news may influence individual perceptions of real news. Through an experiment, this study explores the effects of elite discourse about fake news on the public’s evaluation of news media. Results show that exposure to elite discourse about fake news leads to lower levels of trust in media and less accurate identification of real news. Therefore, frequent discussion of fake news may affect whether individuals trust news media and the standards with which they evaluate it. This discourse may also prompt the dissemination of false information, particularly when fake news is discussed by elites without context and caution.

Notes

¹ Because the computed variable of total accurate identification of fake news is more ordinal than continuous, we also ran the analysis using an ordered probit regression and found no differences in the significance level across our results for H1a or H1b.

² We explored the influence of age in case there was an association between age and digital media literacy. We found no significant main effect of age on accurate identification of real news. In addition, age did not influence the significance level of these results.

³ Like H1a and H1b, this measure of media trust is more ordinal than interval. Therefore, we ran the analysis using an ordered probit regression, in which the priming treatment had a marginally significant main effect on media trust (p = .06) when controlling for type of article and political knowledge, and a marginally significant main effect of receiving fake articles (p = .07), when controlling for the priming treatment and political knowledge.

⁴ There were no changes in the results when controlling for when people took the survey and no interactions between the independent variables and when people took the survey. Controlling for age also produced no differences. Therefore, we report coefficients without these controls.

Additional information

Funding

This work was supported by the the Center for Media Engagement ($500).

Notes on contributors

Emily Van Duyn

Emily Van Duyn (M.Ed., Southern Methodist University, 2015) is a PhD student in the Department of Communication Studies at The University of Texas at Austin. Her research interests include how polarization and digital media influence political communication and public opinion expression.

Jessica Collier

Jessica Collier (M.A., University of North Carolina at Chapel Hill, 2016) is a PhD student in the Department of Communication Studies at The University of Texas at Austin. Her research interests include the influence of media effects, digital technologies, and partisan biases on political attitudes and behavior.