ObSERV takes advantage of advances in the collection and analysis of subnational data from other parts of the world to address the theoretical and methodological challenges of previous research on election related violence and security arrangements in Africa. Relying on a methodological innovation that no other study or organization has used, our project employs probability sampling to deploy Election Observer Missions who, as a part of their election observation activities, will collect constituency level, geocoded data on types and perpetrators of violence, the presence or absence of security arrangements, mechanisms of violence prevention and other information. Their deployment will be preceded by district level assessments and categorization of security levels following a rigorous criteria for determining levels of pre-election violence. Pre-election security assessments at district level will generate better information about the electoral environment and afford greater protection to observers. More systematic data collection will also facilitate analysis (using machine learning techniques and standard statistical approaches) of the relationship between causes of violence and the type and quantity of local level violence; or between election related violence and its effects on electoral outcomes at the provincial and district level. We expect that our sampling strategy and our findings will not only improve scholarly work on election-related violence in Africa, but most importantly will drive policy changes that allow voters to exercise their democratic rights without fear.