The “Good Bartenders” program was designed and implemented by a public-private partnership between Fundación Bavaria (FB), the social responsibility unit of Colombia’s largest brewery, and the Secretariat of Security, Coexistence, and Justice of Bogotá (SSCJ). This program provides materials and training to bartenders and offers food and non-alcoholic drinks to patrons.

Randomized assignment was carried out in two stages, both stratified by locality. In the first stage, we randomly select treatment and control police quadrants from eligible locations. The assignment procedure classified these 221 locations into 109 treatment and 112 control quadrants. In the second stage we randomly select street segments within the 109 treated quadrants that contain bars that will receive the intervention. This two-stage design allows estimating both direct and spillover effects. Our procedure therefore creates three groups from an eligible sample of 5,987 streets: 228 treated street segments located within treated police quadrants (directly treated units), 2,730 control street segments located within treated police quadrants (indirectly treated units), and 3,029 control street segments located within untreated police quadrants (control units). For more details on the randomization procedure, see our Pre-Analysis Plan.

We highlight that this experiment would not have been possible without collaboration from the Secretariat of Security, Coexistence, and Justice and Fundación Bavaria. Their shared information and enthusiasm in designing and testing the effectiveness of a program using experimental methods allows us to contribute evidence on the relationship between alcohol consumption and alcohol-attributable violence. Throughout this project, our partners provided information and insightful feedback.