In over 99.5 percent of protest events in April and May, we recorded no injuries, arrests, or property damage — an unprecedentedly tiny fraction for a movement of this size and geographic dispersion.

Awesome, Jonas!!! Really cool paper 👏

Neither could it have been done w.o. a common template for identifying regimes and providing experts relevant information about time periods and regimes to code. The underlying HRD data are described here by @vldjuve.bsky.social , @torewig.bsky.social and me: journals.sagepub.com/doi/pdf/10.1...

This one of my favorite projects, leveraging output from two huge data collection efforts (NAVCO and HRD) and answering a question that has been on my mind for years. So happy it's out, and so happy for yet another smooth and rewarding collab with the great @chknutsen.bsky.social

Hence we conclude that "nonviolence is a key component in many successful campaigns for democracy, also absent full-fledged revolutions."

Crucially, when we distinguish between a) violent vs nonviolent mass movements and b) democratizing vs non-democratizing incumbent-led transitions, only nonviolent mass movements are associated with incumbents opening up their regimes through directed democratizaton.

We run a range of models with data covering the years 1900-2019 and find a really robust association between citizen mass movements and incumbent-led regime change. This sheds light on why mass action is so strongly associated with regime change and democratization, even if revolutions are so rare.