What has been the most challenging aspect of researching democratic participation, and what new opportunities or insights has that challenge opened up in your WP/research?
1 Like
For our WP (WP5), one of the main challenges has been methodological: how can we study the relationship between protest, social movements and news media comparatively and at scale without losing the complexity of the phenomena we are analysing?
Protest Event Analysis traditionally relies heavily on human coding. Applying it to a multilingual corpus covering several countries, newspapers and years would have required an enormous investment of time and resources. We therefore developed an AI-assisted workflow combining Large Language Models with extensive validation, cleaning, deduplication and event-fusion procedures. This created new problems of its own. Automated classification can reproduce biases already present in journalistic discourse and in training data; categories that appear straightforward conceptually can become ambiguous when applied across languages and political contexts; and apparently minor methodological decisions can substantially affect the resulting representation of protest.
At the same time, precisely these difficulties have generated some of the most interesting opportunities in our research. Working at this scale has allowed us to move between different analytical levels: from national protest agendas to differences between newspapers; and from protest characteristics to the ways in which particular repertoires, actors and issues are framed by the media.
A second challenge has been avoiding the temptation to treat media representation as a one-directional process in which journalists simply “frame” movements. Our work increasingly suggests a relational perspective. Movements anticipate media reactions, design actions partly around visibility, criticize or endorse journalists, redistribute news coverage through social media, and sometimes attempt to bypass legacy media altogether. Journalists, in turn, react to movement tactics and communication strategies.
This has opened up the next phases of WP5. By combining the Protest Event and Frame Analysis with social media analysis and interviews with journalists and activists, we can study the news media not simply as sources of information about participation, but as an interface in which visibility, legitimacy, grievances and political participation itself are continuously negotiated.
One of the biggest challenges seems to be capturing the different perspectives and motivations behind participation. That challenge can also reveal useful patterns and show where more inclusive approaches to research could lead to better insights.
One challenge in WP6 is the dependence on social media platforms. This is particularly true when it comes to demobilisation in interfacing processes. We can only study what is visible and does not get censored or what happened through accounts that got deleted later due to security risks. I would expect that this particularly affects marginalized groups who self-censor for avoiding these risks.