Recent efforts to add a geo-spatial dimension to studies of protest have given social movement scholars the chance to draw some really interesting conclusions. Dan Myers and Beth Caniglia found how close your protest needed to be to New York City to have a hope of appearing in the New York Times. Kenneth Andrews and Michael Biggs determined how sit-ins rapidly diffused from city to city in the south in 1960. And Robert Sampson, Doug McAdam, Heather MacIndoe, and Simon Weffer-Elizondo established what neighborhood characteristics really mattered in where collective action occurred over 30 years in Chicago. While these studies asked different questions and focused on different places, they had one major component in common: they hinged on the painstaking collection of data from a variety of sources to identify the location and characteristics of large numbers of protest events.
What if you want to study a contemporary movement? Good news–you may be able to crowdsource (some of) the painful part! The Guardian has posted the beginnings of a geographic array of Occupy Wall Street protest sites worldwide, including estimates of the number of protesters, indications of the duration of the protest, and (if available) photos and images related to the events. It seems they did some of the initial locating and coding the old-fashioned way; they dug through media accounts. But now they’ve opened it up to “the crowd.” People who know of OWS sites and events can enter that information directly into the database. As word spreads and more people upload information, the data coverage for this particular protest wave could become quite the analytic treasure trove–and the crowdsourced data collection approach could become a model for other protest data collection in the future.
3 responses to “Where is Wall Street? Let the Crowd Tell Us”
Interesting approach to gathering data on protest. It would be very interesting to track this data versus media sources (newspapers, readers, etc.) and to track when and from where entries are added to the database to see who is paying attention to among activists. Not to mention the potential diffusion analysis of actual accumulated data, of course!
This is really interesting, Matt. I noticed that the Guardian was tracking this months ago, and it looks like they have really refined it since then. I’m curious what you and others think about this type of data as a reliable source for statistical analysis. Is there potential to take data assembled by a media source and/or crowd source to create variables in our models (i.e. locations of all occupy sites, how long they lasted, etc.)? Would it stand up in the peer review process?
Interesting questions, Kevin. I think we’ll probably need a phase of methodological development that looks something like what DJM suggests above–a process of doing some old-school digging and comparing that to the crowd-sourced data to see what, if any, differences there are. If the match is good (or the crowd-sourced data seems to be better), I would expect to see more scholars utilizing a technique like this. We’d probably see an expectation develop of doing at least some data reliability checking in any project. So, maybe you pull a 5% random sample of the crowd-sourced data and cross-check that with other sources to make sure it seems like your new data is in line with the match rate other projects have gotten. Another way to think about it might be to use the crowd-sourced data as a set of “leads” to follow up on. The crowd would suggest where things had happened, but then we could follow up using multiple other sources (police records, local news media, local blogs… heck, maybe even Facebook pages and tweets) to confirm that something happened and what exactly that something was. Of course, doing that adds a lot of work for the researcher. I expect the data would be far more comprehensive, but perhaps the painstaking part of data collection will always be with us.