Abstract

Several police practices, including stop-and-search actions, involve selecting citizens for controls. These selection procedures have faced extensive criticism for being biased with respect to socio-demographic factors. However, the different methodologies adopted in prior research to assess target selection have made it difficult to evaluate the actual existence of biases, especially due to difficulties in establishing a comparable benchmark population. To circumvent these limitations, we propose using covert systematic social observations as a method to investigate and evaluate the potential bias of police’s selection procedures. Using the weapon control performed by the Dutch police in Amsterdam in 2022 as a case, we present a method to detect selection biases based on ethnic profiles of citizens that can be reliably applied and has the benefit of allowing for varying police’s selection practices. In sum, given how covert on-site observations offer a realistic picture of how the situation unfolds, we recommend this as a method to examine stop-and-search actions and other bias-prone citizen selection procedures.