Assessing Stability in the Patterns of Selection Bias in Newspaper Coverage of Protest During the Transition from Communism in Belarus*

Abstract

Analyses of selection bias in the coverage of protest events in Minsk, Belarus between 1990 and 1995 are presented. The rapid changes characterizing Minsk during its transition from Communism made it an ideal location for investigating the stability of the patterns of selection bias. Police records of 817 protest events were used to create a protest event dataset, and Minsk's four daily newspapers were read for the entire period in order to establish estimates of event coverage. Results show that large events, events with strong sponsors and, in two of the four newspapers, events accompanied by arrests are each more likely to receive coverage. These effects remain stable through phases of the transition for the combined coverage in any of the papers. The selection factors of event size and event sponsorship also display stability across media source, although the impact of arrests is not always consequential.

Automated identification of media bias in news articles: an interdisciplinary literature review

“…Researchers then gather the relevant news data, and coders systematically read the news texts, annotating parts of the texts that indicate instances of media bias relevant to the analysis being performed. Afterward, the researchers use the annotated findings to accept or reject their hypotheses [90,91]. In a deductive content analysis, researchers devise a codebook before coders read and annotate the texts [92,93].…”

“…After linking events across the datasets, a comparison enables researchers to deduce factors that influence whether a specific news outlet reports on a given event. For instance, several studies compare demonstrations mentioned in police reports with news coverage on those demonstrations [32,90,91]. During the manual content analyses, the researchers extracted the type of event, i.e., whether it was a rally, march, or protest, the issue the demonstration was about, and the number of participants.…”

The Interdependency of Mass Media and Social Movements

“…Remarkably, many students of social movements see the selection and description bias not as phenomena that have to be studied in their own right but rather as data problems. Instead of taking the variation in coverage as an starting point to theorize about the relationships between protest and mass media, the main question that drives most investigations is how useful media data, especially newspaper data, are to serve as an 'objective' measure for the occurrence and characteristics of protest events (see among others the discussions in Earl et al 2004;McCarthy et al 2008;Ortiz et al 2005;Strawn 2008) (Koopmans & Rucht 2002). The resulting data on protest have been used in many groundbreaking studies, especially those that focus on political opportunity structures, one of the classical theories in the study of social movements (some of the most well-known examples include Kriesi et al 1995;McAdam 1982).…”

Multi-channelled forceful intervention, frames and protest success—a fuzzy-set qualitative comparative study of 40 anti-demolition protests in China

“…Considering that existing studies are primarily single case studies, cross-case comparison helps more systematically assess the explanatory conditions for protest success. In addition, the selection bias associated with this method has been empirically investigated in a previous study (McCarthy et al 2008), which helps clarify potential bias and the generalizability of present findings. In fact, we do not attempt to reveal the conditions for all successful protests but merely aim to explain the success of socially influential protests.…”