Randomized Control Evaluation of Body-Worn Cameras


img 5956
A Pacifying Police Unit patrolling in Complexo do Alemão, Rio de Janeiro, Brazil
Photo credit: 
Elena Cryst


Alberto Diaz-Cayeros Headshot
Co-principle Investigator
Senior Fellow
  • Professor, by courtesy, Political Science
Principal Investigator
Senior Fellow
  • Professor, Political Science


Police use of deadly force in Brazil is extraordinary compared to the US. According to the Forum on Public Safety, the police in Brazil killed more than 11,000 people between 2009 and 2013, while law enforcement institutions in the US killed 11,000 over the past 30 years. Moreover, Brazil is one of the most violent countries in the world: more than 50,000 people were killed in Brazil in 2013, about one person every ten minutes.[1] How can technology work to reduce the incidence of police violence, reduce police corruption, and improve citizens’ trust to the police?


The Stanford International Crime and Violence Lab has initiated a partnership with the Secretary of Security and the Military Police of Rio de Janeiro focusing on the use of lethal force by the police and how the “pacification strategy” in the metropolitan area has impacted police-community relations, violent crime, and police use of lethal force, particularly in poor urban neighborhoods. Begun in 2008, the “pacification strategy” introduced Pacifying Police Unites (UPPs) in various favelas or slums, which have traditionally been dominated by armed groups, including drug trafficking gangs and militias. The 38 UPPs aim to regain control of the favelas from these armed actors and to improve security to its residents by establishing a more citizen-friendly police force.


To answer these questions on crime and technology, we will conduct a randomized control evaluation of the use of police body-worn cameras manufactured by Taser International Inc. These cameras capture video evidence from officers’ perspective. They weight approximately 108 grams and can be mounted on the officer’s collar, the hat, or specially designed Oakley sunglasses. The cameras capture data using a web-based computerized video management system developed by evidence.com. The software tracks and inventories all video cameras evidence. The system automatically uploads the officers’ videos at the end of their shifts. The knowledge generated by our study is not just of critical importance to the challenges of citizen protection in Brazil, but throughout Latin America as well as the United States. President Obama has just announced his request for federal funding to provide up to 50,000 body worn cameras to police forces throughout the country. The critics of this measure suggest that more information is critically needed on how these devices perform in different contexts.


To date the best-known randomized controlled evaluation on the use of body-worn cameras was conduced in the city of Rialto, California, and it showed great promise in restraining use of police force. But we do not know how this technology would work in a radically different context, characterized by extraordinary levels of violence, armed confrontations, and presence of drug trafficking. Furthermore, the use of police body-worn cameras has not been systematically evaluated in countries where there is a culture of disobedience to the laws. One of the difficult aspects of this technology is that police officers can turn the cameras on or off voluntarily and might hence purposely choose not to film certain interactions precisely because they do not want to be observed. It is not clear what type of protocols are necessary to oblige police officers to properly use the cameras and force them to turn them on every time they interact with citizens. Should officers switch the cameras on only when they interact with citizens, as in the Rialto study, or should they have the cameras on all the time, in which case there are issues related to the handling of massive amount of video footage? Our study will be able to answer these critical questions about protocol within the context of a randomized controlled evaluation.

The evaluation will be conducted in Rocinha, located in the South Zone of Rio de Janeiro. According to some estimates, Rocinha has more than 90 thousand inhabitants, one of the largest favelas in Brazil. The UPP in Rocinha was inaugurated in September of 2012 and currently has 510 police officers. Rocihna’s UPP experienced a mayor legitimacy crisis following the Amarildo scandal and the incarceration of its former commander, Major Edson Santos.[2] The Amarildo case was a major setback for government efforts to construct a good image for the UPP as “resident-friendly” police units. By charging the culprits and appointing a different commander, the authorities are signaling that they are committed to transform the police’s culture of violence and rebuild itsimage. The introduction of miniature cameras to tape interactions between UPP officers and Rocinha residents could be a step in the right direction to constrain police misbehavior and build trust among the community.

[1] Forum on Public Safety.

[2] Ten policemen of the UPP as well as Major Santos were convicted for torturing and killing Amarildo de Souza, who disappeared in July of 2013