1News as Data, News as ActionCurrent Issues in the Study of News Coverage and ProtestsPamela OliverPlan for the talk• Introduction – Co-evolution of news/politics/protest• News as data• Transitions– Data quality concerns– News holes – Looking for protest effects on news• News as action by strategic responsive actors among other strategic responsive actors– Agenda setting– Filtering communication– Reinforcing others’ actions – News holes, issue competition, and media/movement cycles• News as action and dataCoevolution of Protest, Politics, News• Movements change in coevolutionary relationships with political regimes, news media, and other actors• All may be thought of as sets of actions by diverse actors• Statistical distribution of actions evolve over time• Actions affect other actions– Diffusion of actions & ideas– Strategic interaction between different actors– Flows of resources or information– Others’ actions can have reinforcing effectsThe goal is to theorize the ways in which grassroots protest movements, institutional politics, and the news media co-evolve over time, and the ways the logic of news-writing effects the system.mediacoverageinstitutionalpoliticsmovementsToday’s talk poses the question, rather than answers it!News as Data: Protest Events Research• Newspapers are the most readily available source of information about protests and social movements• Newspapers are published routinely and regularly• A great deal of research takes newspapers as given sources of information• Event-oriented studies: code events from the news and then study them over time• Detailed studies of event series reveal important evidence about how processes work• I will flash through examplesDoug McAdam “Tactical Innovations”• Uses plots of the sequencing of events in the Civil Rights Movement to argue – New tactics explain steep rises in events– Dynamics of the CRM: protest first, then segregationist response, then federal response• Data are from the New York Times Index2Civil Rights Events Fig 1 (McAdam)Doug McAdam, “Tactial Innovations,” ASR 1093The “take off pointCivil Rights Events Fig 2 (McAdam)Doug McAdam, “Tactial Innovations,” ASR 1093Sit-ins drive the 1960 spikeCivil Rights Events Fig 3 (McAdam)Doug McAdam, “Tactial Innovations,” ASR 1093Civil Rights Events Fig 4 (McAdam)Doug McAdam, “Tactial Innovations,” ASR 1093The plot shows that movement actions, the solid line, generally leads an action peak, followed by segregationist and government actions1960s Black Urban Riots• Data compiled by newspaper clipping service from local newspapers• Data set originally constructed by Gregg Carter• Daniel J. Myers did extensive analysis using event-history modeling techniques. After the seasonal cycles are controlled, there are clear diffusion effects– Big riots diffused nationally– Smaller riots diffused regionally– Television broadcast areas account for the regional diffusion patternsMonthly Riot Counts 1964-19710204060801001201401964 1965 1966 1967 1968 1969 1970 1971Source: Data originally collected by Gregg Carter, compiled by Dan Myers31960s Riots, Weekly Counts (truncated)05101520Riots per Week, 1964-1971 (Newark and MLK riot weeks truncated) 1964 1965 1966 1967 1968 1969 1970 19711960s Riots Myers ASR 1997German Protests• Ruud Koopmans (ASR 1993) examines the mix of confrontational, demonstrative and violent actions over time• Argues the dynamics are driven by repressive actions by regime and facilitative actions by political and organizational elites• Data are coded from the Frankfurter RundschauProtests in Germany 1975-198902040608010075 77 79 81 83 85 87 89year/quarterfreqGerman new social movements protests, quarterly counts of events (Source:Ruud Koopmans)Koopmans ASR 1993 German Protest TypesConfrontationalDemonstrativeLight ViolenceHeavy ViolenceKoopmans ASR 1993 Responses to Protest Types% Repressed% FacilitatedDemonstrativeDemonstrativeConfrontationalConfrontationalViolentViolent4German Protest Events, Monthly CountsGerman Protests, Monthly Counts010203040506075 78 81 84 87evts_totGerman Protest Participants, Monthly TotalsParticipation in Protests, Monthly Totals02000004000006000008000001000000120000014000001600000180000075 78 81 84 87part_totIran Revolution• Karen Rasler (ASR 1996) examines the effects of repression on protest, arguing that it initially suppresses protest, but then there is a later rebound effect through spatial diffusion of protest to other areas• Data are several news sourcesRasler, Iran Revolution (ASR 1996)Rasler, Iran Revolution (ASR 1996)And MANY More• Event approaches and newspaper (news archive) data are the only viable source of longitudinal data on protest events (and many other types of events)• Underpinnings of quantitative event-oriented approaches to understanding dynamics of collective action and politics• Lots of interesting & important research problems to explore• BUT . . .5We’re Ignoring SomethingThe Elephant in the Living Room• These studies of over-time dynamics assume that newspapers capture “true” picture of events: “events in newspapers” are treated as equivalent to “events”• Assumption that “all” events are captured is obviously wrong, never actually defended • Authors generally argue (hope) that newspaper coverage of events is “unbiased” statistically– A constant % of events – OR Essentially random– OR AT LEAST a CONSISTENT selection structure (so changes over time are validly assessed)News as Data: The “Selection Bias” Problem• “Bias” used in the statistical sense to refer to the selection structure: from the pool of events of interest that actually occurred, what got into the news?• There is no data available to support the assumption of “unbiased” event sampling in newspapers, in fact theory & research suggest the opposite• Even the hope of a consistent selection structure over time is unsupported by dataSelection Bias Studies• Need an external reference source for comparison• Police permit records & Police logs– McCarthy, McPhail & Smith– Oliver et al Madison study• Or inter-media comparisons– Long tradition of comparing local & national news sources or of comparing different national sourcesMcCarthy, McPhail, Smith (ASR 1996)• Washington DC “First Amendment” Event Permits – Compare 1982 & 1991• New