 Hello, everyone. It's great to be here. My name is Rene Krasen. I'm a PhD student at the University of Washington in Seattle. I'm going to be presenting qualitative interview study on governance capture in a self-governing community, looking at a qualitative comparison of the Serbo-Croatian Wikipedia. For those of you who are unfamiliar with the Croatian Wikipedia case, for about a decade, at least 2011 to 2020, a small group of far-right editors took over the governance mechanisms of the Croatian Wikipedia. They installed themselves and their allies in administrative positions, and they systematically start introducing far-right bias, including historical revisionism about World War II, into the project's article base. Finally, they harassed and stamped out dissenting editorial voices, so it's hard for editors that were not ideologically aligned with this core group to contribute to the project. After years of calls from the foundation to intervene, the situation started getting some media attention. Eventually, the offending editors, many of them, most of them were removed from positions of power, many of them were blocked outright, issued global blocks. The foundation did a commissioned a retrospective assessment of the situation to figure out what happened. They found out that essentially while there was systematic issues with the content on the project, but really this was a problem of governance, they called it project capture, that the mechanisms by which the project was governed were taken over. And as Serbo Croatian is a plural-centric language, and as four separate language editions, three of which correspond to these national, the national variants of the language, so there's, in addition to Serbo Croatian, there's also Serbian, Croatian and Bosnian. The report found that the situation on Croatian Wikipedia was unique while there are issues with historical revisionism and vandalism in these other projects, none of them experienced the project takeover that Croatian Wikipedia did. So the question that arose was why? Why did Croatian Wikipedia succumb to capture while other editions in the same language family that shared similar characteristics did not fail in the same way? To answer this question, we did an interview study, we interviewed people essentially in two groups. First group were people who were involved in global governance processes on Wikipedia. So this could be by being a steward, or by being a foundation staff member that was dealing with disinformation issues, or people that were in other informal cross-wiki groups. The second group of people we interviewed were editors and admins specifically on these Croatian and Serbian projects. Through our interviews, we came up, we did a grounded theory analysis of the interview and came up with three themes or propositions that together we propose explain why some projects like Croatian Wikipedia may be more vulnerable than others to this form of governance capture. So the first characteristic of propositions perceived value. Is the project even worth expending the effort to capture? So through our interviews, we found that the four projects were not equally attractive targets for this form of disinformation campaign that promoted nationalist ideology. Croatian and Serbian both had a critical mass of editors and a community in which national narratives resonated. So they were in this way of the four projects, the two most attractive targets of potential targets for capture. The first one that differed in these two other respects. First was bureaucratic openness. How easy was it for contributors outside the core founding team to ascend to local governance positions. On Serbian Wikipedia, there was an early effort by the founders of the project to open up the project's bureaucracy to people outside of the core founding team. So it was a transparent process to ascend to administrator positions, bureaucrat, check user positions, and there was an effort to recruit this initial diverse group of administrators. But that didn't happen. The initial group was quite insular and grew more insular over time such that it was, there was no process for people to ascend to these positions with elevated user rights unless without relying on sort of their existing relationships. The third factor, or the second one on which these two projects differed was institutional formalization. So to what degree did the project preferred these personalistic more informal forms of organization over formal ones. Serbian Wikipedia formalized a lot of its sort of governance processes so they had more rules, and they specifically had rules that constrained administrator behavior there was a process to remove administrators that were not acting in good faith. They also outlined a blocking policy that outlined when you could, what was a good faith use of the blocking policy which was a big problem on Croatian Wikipedia. So instead, rules were invoked kind of ad hoc at the whims of the admins there was no sort of the organization governance was very informal project had no Wikipedia chapter, which Serbian Wikipedia had and which landed some legitimacy or scrutiny from outside of the actual project into the processes in the project. So, we wanted to rep we represented these two propositions on which the two projects differed on this two by two matrix. And one is that each of these quadrants represents the relative a window of a likelihood of capture. So projects that have more personalized institutions and insular bureaucrats like the Croatian Wikipedia are highest risk. Maybe and projects that are more likely to have to have this are these medium size projects that have a critical mass of editors that they're attractive enough to be targeted, but they have not yet developed the sorts of resources that larger projects have lowest risk of capture things are projects like English Wikipedia and German Wikipedia who have open bureaucracies and these more formal processes in place, as well as Serbian Serbian Wikipedia. So, the next step for this study is a quantitative study that empirically tests these relationships. Because obviously this is our kind of theory for what we're have for what was happening based on our interview results. But the next step is testing it across many many Wikipedia additions to see if we can to see if it generalizes beyond our immediate study. And I'm looking forward to feedback and questions.