This paper describes an experiment to determine the impacts of body-worn cameras on fatality rates stemming from police-citizen encounters, with details about the research methodology and a discussion of the outcomes.
The study reported here assessed the effects of body-worn cameras (BWCs) on rates of fatalities arising from police-citizen encounters. While existing experimental research has not examined this outcome because it is so rare, the staggered roll-out of BWCs across U.S. law enforcement agencies provides an opportunity for quasi-experimental analysis. Difference-in-difference (DID) analyses using Poisson models compare changes in U.S. law enforcement agencies’ fatality counts with changes in BWC acquisition. Using a federal law enforcement survey focused on body worn cameras (LEMAS-BWCS) and media-sourced data on fatal encounters from fatalencounters.org (FE), the research examines agencies acquiring BWCs between 2013-2014 and 2015-2016 and those that did not acquire them up to 2016 and had no plans to do so. The article includes a fixed effects annual panel data analysis with data from 2005-2006 to 2018-2019 and two two-group analyses focusing on a pre-treatment period (2010-2011 to 2012-2013) and a post-treatment period (2016-2017 to 2018-2019). The latter includes a propensity score matched comparison. Two out of three DID analyses showed statistically significant negative effects of BWCs on citizen fatalities, and the propensity score matched two-group analysis returned a non-significant negative effect. The research found some evidence for BWC effects on citizen fatalities. However, there are important validity threats to this conclusion. These include the possibility that BWC acquisition serves as a marker for other policy changes focused on BWC-acquiring agencies in the 2013-2014 to 2015-2016 period and beyond. Publisher Abstract Provided