Quantitative analyses have the potential to contribute to transitional justice mechanisms, via empirical evidence supporting the memory of victims, allocating proportional responsibility among perpetrators, determining legal responsibility, and supporting historical memory and clarity. However, most data available in transitional justice settings are incomplete. Conducting quantitative analyses relying solely on what is observable and knowable leads to not only incomplete but often incorrect analytical results. This can harm rather than contribute to transitional justice mechanisms. This article outlines different types of data, the ways in which observable data, on their own, are insufficient for most quantitative analyses of interest, presents these limitations via a case study from Syria, and introduces statistical methods to overcome these limitations.