Abstract
In this paper, homicide related statistics from official sources in Turkey have been compared, in an attempt to identify trends in homicides. Up until now no study has looked at national homicide trends by triangulating different data sources. Data has been extracted from a variety of sources, covering data from agencies from every step of the criminal justice process: Cause of death data, police, prosecution and adjudication data, and prison data. Time period covered included the last 20 years, however due to the lack of availability the time range was different for different types of data. It was found that all those types of data suffer from some limitations, and that it is very difficult to compare data from different sources. In particular, it was found that changes in what is counted and how, made it hard to identify trends over longer periods of time. Further, it was found that administrative decisions that determine how data are collected complicate interpretation and reduce the validity of data. Counting units are inconsistent across different sources making cross-checking problematic. Finally human error in data entry was found to reduce data reliability. Despite these drawbacks, an attempt is made to identify potential trends.