De-escalation technology: the impact of body-worn cameras on citizen-police interactions

How to cite

Fetzer, Thiemo; Souza, Pedro; Soto, Caterina; Barbosa, Daniel. 2021. "De-escalation technology: the impact of body-worn cameras on citizen-police interactions." AEA RCT Registry. AEARCTR-0007785. https://doi.org/10.1257/rct.7785-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Thiemo Fetzer* , Pedro Souza , Caterina Soto , Daniel Barbosa

* Primary investigator

Registry ID
AEARCTR-0007785
DOI
https://doi.org/10.1257/rct.7785-1.0
First registered
Last updated
Status
Completed
Trial dates
2018-09-03 – 2018-12-10
Countries
Brazil (State of Santa Catarina (Municipalities of Florianópolis, Tubarão, Biguaçu, São José and Jaraguá do Sul))
Registry record
http://www.socialscienceregistry.org/trials/7785

Abstract

We implemented a RCT of body-worn cameras in the state of Santa Catarina, Brazil. In partnership with the State Military Police (PMSC), we randomized the roster of 450 police officers with public facing duties across 5 police precincts to either a treatment or a control group in a 1:3 ratio. To estimating learning effects, we also randomly provided blackout days, in which no officers would be allowed to wear cameras during their shifts. Our analysis is done at the police dispatch/event level, in which police-citizen interactions unfold.

Intervention

We implemented a RCT of body-worn cameras in the state of Santa Catarina, Brazil. In partnership with the State Military Police (PMSC), we randomized the roster of police officers (approx. 450) with public facing duties across 5 police precincts to either a treatment or a control group in a 1:3 ratio. We also randomly provided blackout days, in which no officers would be allowed to wear cameras during their shifts. Our analysis is done at the police dispatch/event level, in which police-citizen interactions unfold.

Experimental design

- BWC Assignment: From the roster of 450 sworn police officer with public facing duties we stratified them into blocks based on (i) precinct, (ii) officer main activity type (tactical, preventive or ostensive operations), (iii) above/below median rank, (iv) above/below number of previous internal investigations and (v) gender. We further collapse blocks with less than 3 officers, resulting in a total number of 40 stratifications blocks. We then randomly allocated 1/3 of officers inside each block to the treatment group.

- Blackout Days: We stratified the experimental period days based on days of the week and randomly selected days that BWC would not be available for all officers in the police precincts. In this way, we ensure that we have shifts (12h each) in which treatment is not administered to any officer during those periods, either by directly using BWCs or by patrolling with officers wearing BWCs.

Design details

Randomization method
Computer Randomization
Randomization unit
(1) BWC assignment at the officer level; (2) Blackout day assignment at two 12 hour shift (1 day)
Clusters
40 stratification blocks
Observations
180 events per day on average before the officer level randomization x 98 experimental days = 17640 events
Arms
301 treated officers, 149 control officers. 24 blackout days, 74 no-blackout days.
Primary outcomes
(i) Indicator if the event generated a police report; (ii) Indicator if there was any victim reported; (iii) "Citizen Behavior"; (iv) Use of Force; (v) Indicator for Arrests and Handcuffs; (vi) Inverse covariance-weighted Index combining (iii), (iv) and (v)

Cited by

No citation data available for this registration yet.

Keywords