Résumé : From Bucharest to Skopje, anti-government protests have been sweeping Southeast Europe(SEE) since 2011. Considering that the region had been regarded to have a weak civil societybefore the protests, the mobilisations came as a surprise to academics and practitioners alike.Yet, these protests were not only noteworthy for happening in a region, seemingly endowedwith a weak civil society, but also for the fact that they remained largely peaceful. Even untilthe mid-1990s, in contrast, the Romanian government had violently dispersed protests usingmilitant miners, and protests in Albania, Bulgaria, and Yugoslavia had a similar outcome.Tracking the frequency with which governments have used lethal or non-lethal violence duringprotests, it can be shown that governments in SEE appear to have become more tolerant of thistype of extra-parliamentary political participation. Data collected from protest events showsthat at least since the early 2000s there has been a period of government restraint. This came atthe same time as the EU opened accession pathways for most countries in the region. Thus, thethesis discusses the reasons for the decline in repression. It explores whether the temporalcorrelation between increased EU engagement and the decline of physical repression in theregion was coincidental or can be linked. In this context, the thesis focuses on the role of EUaccession conditionality, since it has been assumed that EU conditionality could have ademocratizing and, therefore, potentially constraining effect on governments. Even though thethesis concludes that EU-accession conditionality impacted repression, the nature of thatimpact is complex. It shows that to understand how EU-accession conditionality impactedrepression, we cannot treat repression as a monolithic concept. Although EU conditionalityreduced visible forms of repression, such as lethal and non-lethal violence, it had little effecton opaque, institutional forms of repression. Studying how media outlets throughout the regionfunctioned as proxies of the state and how governments relied on them to quell dissent, thethesis demonstrates that the fall of physical repression did not result in a reduction innonphysical forms of coercion. It develops the argument that repression followed the logic ofsymbolic compliance. Hence, SEE leaders, keen not to jeopardize their EU accession prospects,rarely used violence to repress dissent, maintaining an image of democratic governance.However, this was only a smokescreen since at the same time, below the surface and withoutBrussels' attention, they were attempting to demobilize protesters via institutional channels.Accordingly, the thesis contends that repression, rather than being reduced, was bifurcated intodifferent types. To support this argument, the thesis employs a comprehensive methodologythat includes a regression analysis, qualitative comparative analysis (QCA), and framesanalysis.