Media Measurement Models of Protest Event Data

Abstract

Growing interest in quantitative studies of social movements and protest cycles attests to the vigor of protest event analysis as a strategy for investigating the protopolitical processes of collective claimsmaking in democratic states and emerging democracies. Increasing investments in protest events research has also led to growing concern about sources of measurement error that stem from reliance on media data sources. Using Blalock's conception of auxiliary measurement models, this article traces two alternative treatments of measurement bias in the literature on events analysis. The two approaches, characterized here as "media theory" and "representational," differ in the degree to which they accept media data as an adequate representation of protest event reality. Recent work that attempts to establish empirically the relationship between media data and alternative data sources on protest events promises an empirical base for examining the assumption of stable media bias that underlies representational approaches. Finally, the conclusions suggest some weaknesses in the empirical base for theorizing from media-based event analyses.

The Three Waves of Anti-Austerity Protest in Greece, 2010–2015

“…Admittedly, such protest event data are not issue-specific and lack details on their size, networks and claims, which a systematic study of LPEs can more reliably provide. It does, however, indicate continuing levels of decentralised yet widespread ‘everyday forms of resistance’ (Scott, 1985), which are typically not captured by mainstream newspaper reports (see Mueller, 1997). In the absence of any other major social movement organising protests at that time, it does not seem unreasonable to hypothesise that most of these protests targeted austerity.…”

Competition and Resource Partitioning in Three Social Movement Industries

“…In particular, studies have asserted that there are two main sources of bias in newspaper data: selection bias and description bias. Selection bias refers to the fact that not all protest events will be covered by a given newspaper and to the possibility that what is covered is not a random sample of all events that took place; such bias may vary over time (Mueller 1997). Description bias refers to the veracity of the reporting of events that are selected for coverage.…”

Europeanisation and social movements: The case of the Stop TTIP campaign

“…To measure the degree and paths (i.e., forms) of Europeanisation of the organisations against TTIP, we conducted a protest event analysis (PEA) between 2014 (when the issue became fully visible on the EU policy agenda) and May 2016 in six European countries and at the EU level making a total of 784 events coded. Following a longstanding tradition in social movement research (Tarrow 1989;Tilly 1978), and despite its limitations (McCarthy & Zald 1996;Mueller 1997), we use PEA because it allows the quantification of many protest properties, such as frequency, timing and duration, location and claims (Koopmans & Rucht 2002). In order to retrieve relevant information on the the anti-TTIP protest in Europe, we rely on a combination of sources: national quality newspapers scanned for relevant articles using keywords such as 'TTIP & protest' and 'TTIP & mobilisation' on the Lexis Nexis database 8 ; online information portals such as Euronews and Eurobserver; and the 'news sections' of the Stop TTIP websites in the six countries -namely, specific sections where 'news' from other newspapers or press agencies are reported.…”