A Matter of Convenience: Challenges of Non-Random Data in Analyzing Human Rights Violations during Conflicts in Peru and Sierra Leone

This chapter argues that reports of the experience of human rights violations and violence as an empirical record are likely to be biased (with selection bias often present), as well as incomplete and prone to significant measurement error. Two of the main fundamental errors are the underreporting of rights violations that have occurred and the overreporting of violations that have not. This chapter uses the cases of Peru and Sierra Leone to demonstrate how casualty reporting across different sources varies, inhibiting the ability to make valid statistical inferences. The implications drawn from this are twofold. When full enumerations or representative samples are not available, multiple sources of casualty records are vital to securing more accurate civilian casualty estimation. Nevertheless, limitations in the data available need to be translated into the conclusions that can be drawn about the contours of a conflict.