Digital media and perceptions of the United States among the Russian elite, 2004–2016

ABSTRACT

This paper seeks to explain why Russian elites’ exposure to online media for their news contributed to stronger pro-American attitudes than reliance on traditional media. Two causal mechanisms are tested using a repeated cross-section of elite surveys. One operates at the level of attitudes and is suggested by the field of political communication; the other emerges from the literature on cognitive psychology and operates at the level of beliefs by providing a cognitive map through which individuals process information and reach conclusions. I find that both mechanisms are relevant, with framing effects being particularly important to hardliners’ perceptions of security threats.

KEYWORDS:

• Threat perceptions
• anti-Americanism
• social media
• Russian elite
• foreign policy

Acknowledgments

The author is deeply thankful to Sharon Werning Rivera and William Zimmerman for the opportunity to contribute to this fascinating research project. The paper benefited greatly from the feedback of Timothy Frye, Sharon Werning Rivera, William Zimmerman, and the participants of the NSF Conference on Elite Attitudes toward Conflict and the West.

Notes

1. User-generated content and content produced by foreign and Kremlin-independent sources, as well as content produced by elites themselves, are distinctive features of new media.

2. For 2016, the response categories ranged from 1 (not once) to 5 (every day). To ensure consistency across the waves, the answer categories were aggregated to three categories before taking the average. The author is thankful to Sharon Rivera for this suggestion.

3. A linear model rather than a limited dependent variable model was used because we are interested in testing for differences in the means across the two groups rather than in the probability of observing the specific outcome. When a linear model specification is used for a limited dependent variable, standard errors become larger because the model is misspecified, which increases the probability of Type II error, i.e., failing to reject the null hypothesis of no effect when it is false.

Additional information

Funding

The National Science Foundation (Grant #1742798) provided generous conference travel funding.