Protesting that is fit to be published: issue attention cycle and nationalist bias in coverage of protests in Ukraine after Maidan

ABSTRACT

There is an established tradition in social movement research to study protests through event datasets constructed from newspaper data. However, studies evaluating the magnitude of bias in newspaper reporting on protests have been largely confined to Western Europe and North America and have predominantly focused on time-invariant bias. I extend the past scholarship by analyzing the bias in online newspapers’ reports on protest events in Ukraine prior to and following the Maidan protests, focusing on issue attention cycles. I find that the number of mentions per protest event in the national media dropped after Maidan, while the coverage of ideological protests increased, indicating an issue attention cycle. The regional sources displayed less bias; however, they over-reported protests against separatism and Russian intervention. This presents researchers with a trade-off between using selective national outlets which adhere more to journalism standards and less selective regional outlets featuring more unpredictable patterns of bias.

KEYWORDS:

• Protest
• protest event analysis
• selection bias
• Ukraine
• Maidan

Acknowledgments

I would like to thank the participants of the workshop at the Research Centre for East European Studies at the University of Bremen, Viktoriia Moshkina, and the anonymous reviewer for comments on earlier versions of this article.

Supplementary material

Supplemental data for this article can be accessed here.

Notes

1. Hereinafter, I use the word “campaign” to denote a group of events, unified by a single set of issues and protest actors.

2. Going beyond studies of protest and conflict, Maier (2010) systematically compares the content of the five most popular news websites to a broad selection of print newspapers. Comparing an equal number of sources (i.e. five websites and the five most popular print newspapers), websites collectively published 68% more front-page news.

3. I regard 21 February 2014 as the date when the Maidan protests ended – that is, when then-President Viktor Yanukovych fled the country, resulting in the main movement demand being fulfilled.

4. The exclusion of the Maidan period from the analysis is driven by two reasons. First, the Maidan protests were both extraordinarily large and violent. There is a consistent finding in the literature on selection bias that large and violent events are over-reported. Hence, over-reporting of the Maidan protests would be entirely consistent with the previous literature and would generally generate few theoretically meaningful insights. Furthermore (and second), the Maidan protests stood out from the protests in 2013 and the protests in 2014 that followed in terms of the dominant issues, raising primarily political (for EU association, against the government) and civil rights (against police violence) demands. Ideological protests, which are the focus of this article, peaked after Maidan (Table 1).

5. That is, the purge of politicians who worked in the government under Yanukovych.

6. See, for instance, Waisbord (2011) for examples from post 9/11 United States, or Bose (2011) for description of journalists’ practices in reporting on the India-Pakistan conflict.

7. Given the regional political divide in Ukraine, a reader may hypothesize that the heightened attention to anti-separatist protest may be unequal throughout the regions, to the point that in some, the media would over-report events in favor of separation. However, I find no significant differences in the rate of reporting of pro- and anti-separatist events in the regional media in the four macro-regions of Ukraine (Eastern, Western, Southern, and Central). See Table OS2 in the online supplementary material.

8. For three regions – Khmelnytskyi, Sumy, and Vinnytsia regions – fewer than seven regional newspapers that featured an RSS-feed were available during the period under study. For these regions, at least five regional sources were monitored at any point in time. Exclusion of these regions from the analysis produces no substantive differences in the model coefficients (see the supplementary material), as I model the rate of reporting rather than the number of reports. Any effect of the difference in the number of sources is controlled by the regional dummy variables. I thus report the model results for all regions.

9. Ten percent (10%) of the websites had unknown owners. According to Ukrainian legislation, online media are not obliged to register as mass media and hence are not obliged to disclose their ownership structure.

10. With the exception of three regions, which are controlled by regional dummy variables in the regression analyses.

11. For instance, if an event has been reported only by regional sources but not by national ones, the number of mentions in the national sources will be equal to zero. However, by definition each event is mentioned by at least one national or regional source to appear in the dataset.

12. While there is an overlap between ideological and political protests – 52.7% of events raising an ideological demand also included political demands – as I will show below, the reporting of these categories peaked at different periods and varied independently.

13. Small-scale subnational protest in the former Soviet Union has recently attracted interest. See, for instance Economist (2019), linking decreasing government popularity in Russia to dissatisfaction with bread-and-butter issues in the regions.

Additional information

Funding

This publication has been produced as part of the “Comparing Protest Actions in Soviet and Post-Soviet Spaces” research project, which was organized by the Research Centre for East European Studies at the University of Bremen, with financial support from the Volkswagen Foundation;Volkswagen Foundation [Az90200];