This paper presents an approach to assess automatic fault location methods in power distribution systems. The proposed approach allows to consider the aleatory errors inherent to calculated, forecasted and measured values, typically required by the fault location methods for a more realistic and consistent evaluation. These errors can follow different mathematical distributions, according their source and the assumptions made. In order to demonstrate the proposed approach, a real power distribution system is considered, and additionally, one method that requires measurements at the power substation and one method that requires sparse voltage measurements are assessed.