Can Trust Be Built Through Citizen Monitoring of Police Activity?

En Español

Background

New technologies can improve oversight, management of police and, through this process, enhance public confidence and gain citizens’ trust. Working in Brazil, in partnership with the Military Police of the state of Santa Catarina (PMSC), we propose to study the effects of two programs that make use of new technologies to build trust between citizens and the police and to improve public safety.

Common treatment arm – The Rede de Vizinhos (RdV – “Neighbor Network”) is a community policing program in which police officers engage with local communities through dedicated instant-messaging groups, a citizen training program and regular on-site group visits. It is based on the principle that efficient and more continuous communication with dedicated police contacts improves the quality of information shared, improves police response to crime and facilitates the coproduction of security through the implementation of prevention measures aiming to reduce crime at a fine, spatial neighborhood level within the cities.

Alternative treatment arm – The Igarapé Institute has developed CopCast, an open-source software application that turns Android devices into body-worn cameras (BWCs). We expect that the use of BWC technology to monitor policing activity on the ground will lead to increased oversight and improved police-community relations, preventing the escalation of use-of-force, facilitating criminal proceedings when video recordings of incidents are available and reducing the number of citizen complaints against police officers.

Both Rede de Vizinhos and CopCast aim to improve the relationship between the police and the communities they serve a) through better allocation of policing activity made possible through citizens sharing information with dedicated contacts in the police force and b) through improved oversight and monitoring of the police force.

Intervention Date: September 2017 – December 2018

Research Design

Common treatment arm – We will use a spatially targeted Facebook advertising campaign to induce the creation of Rede de Vizinhos neighborhood watch groups, using social media to solve the coordination problem around group formation. The advertisement campaign will be administered at a neighborhood level across the 295 municipalities of Santa Catarina. The unit of treatment will be a neighborhood targeted by the information campaign, inducing creation of RdV groups.

Alternative treatment arm – We will randomly assign CopCast BWCs across police officers and in randomly-selected shifts, BWCs will be unavailable for all officers within a police station. The trial is focused on five municipalities, with the possibility of future expansion, seeking to treat at least 300 officers. The (main) unit of treatment will be an individual event capturing an interaction between a citizen and police officers.

Hypotheses

Common Treatment arm:

• Do areas with RdV groups see improvements in (perception of) safety?
• Do RdV groups change the quantity and quality of citizen information sharing?
• Does RdV change beliefs about police capacity, intentions and effectiveness? Does it increase communal trust or trust in the police?
• Does RdV increase (perceived) physical police presence?

Alternative Treatment arm:

• Do areas with RdV groups see improvements in (perception of) safety?
• Do RdV groups change the quantity and quality of citizen information sharing?
• Does RdV change beliefs about police capacity, intentions and effectiveness? Does it increase communal trust or trust in the police?
• Does RdV increase (perceived) physical police presence?
• What is the effect of body-worn cameras on police use-of-force, behavior, efficiency and citizen complaints in particular shifts?
• Do treatment officers behave differently if they used BWC in the past, even if they are not using the app in a specific black out shift? How do treatment effects vary by officer level characteristics?
• How do BWCs impact dispatch group and team dynamics? In particular, can we observe any effect on control police officers who were allocated to a dispatch group in which at least one officer was treated? Do treatment effects persist even when officers are not wearing cameras?

Findings