Peer-reviewed article · 2025
How Does Community Policing Affect Police Attitudes? An Experimental Test and a Theory of Bureaucrat-Citizen Contact
Dotan Haim, Matthew Nanes, and Nico Ravanilla · American Political Science Review 119(4): 1983-2001
Winner, 2023 APSA Southeast Asian Politics Group Best Paper Award (Open Category)
In brief
Most research asks how community policing changes citizens' views of the police. This field experiment in the Philippines flips the question: how does sustained contact with citizens change police officers' own attitudes?
Abstract
In contrast to the expansive work on how community policing affects citizens' attitudes toward police, existing research says little about how community policing affects officers' attitudes toward citizens. We examine officer-facing outcomes using an experiment in the Philippines, in which a random subset of a province's 705 officers were assigned to intensive community policing activities for seven months. Treatment officers saw improved understanding of citizen concerns, but did not develop greater empathy or trust toward civilians, nor an increased sense of accountability for citizen-facing misconduct. We build from the experiment to develop an inductive theory of bureaucrat-citizen contact, relying on qualitative observations and exploratory analyses of heterogeneous effects. We propose that contact with citizens is only likely to improve attitudes among frontline bureaucrats who are not ex-ante embedded in their communities. Moreover, contact may have negative effects when it reveals threats to bureaucrats' personal safety.
Research annotationQuestion, variables, design, findings, and mechanism at a glance▾
- Research question
- How does sustained contact with citizens change police officers' own attitudes toward those citizens?
- Independent variable (X)
- Assignment to seven months of intensive community policing under the One Sorsogon program, a bundled treatment comprising a two-day training, six monthly problem-oriented policing meetings with barangay leaders, citizen-facing foot patrols and visits between meetings, and a culminating community town hall.
- Dependent variable (Y)
- Five pre-specified officer attitudes: sharing citizens' public safety concerns, trust in citizens, empathy for citizen concerns, perceived organizational accountability for citizen-facing misconduct, and viewing corruption as serious misconduct, plus a composite Officer Attitude Index combining the four attitudinal indices on a 100-point scale.
- Identification strategy
- Field experiment in Sorsogon Province, Philippines. Within the 99 barangays assigned to community policing, 198 officers were randomly drawn from the province's 705 eligible officers, two per barangay, blocking on police station and on rank, one senior and one junior officer per team. Randomized replacements for reassigned officers bring the intent-to-treat group to 245 officers. The estimator is OLS on the officer, regressing each endline attitude on the treatment indicator, the same attitude measured at baseline, block fixed effects, and a missing-baseline indicator, with cluster-robust standard errors clustered at the station level. Qualitative observation of the meetings and exploratory heterogeneity analysis build the inductive theory.
- Main findings
- Treated officers identified 24 percent more public safety concerns overlapping with those named by citizens, against a control mean of 1.07 matched issues. There was no effect on trust, empathy, accountability, or views of corruption. The design is powered to rule out large effects: on the composite Officer Attitude Index, the largest change within the 90 percent confidence interval is a 2.5-point increase, about 4 percent, or a 2.7-point decrease, on the 100-point scale.
- Mechanism
- Contact improves attitudes only among bureaucrats not already embedded in the communities they serve, and can work in reverse when contact reveals threats to officers' personal safety. The intervention raised the Officer Attitude Index substantially among officers from outside Sorsogon and did nothing for officers living in the province, with the gap driven by trust and empathy. Where New People's Army presence was higher, community policing reduced officers' trust in civilians and raised their belief that citizens were rebel sympathizers, while trust rose marginally where the insurgency was absent.
- Why it matters
- Turns the large literature on how policing reform affects citizens around to the officer-facing side, and specifies when the contact hypothesis should and should not be expected to hold for state agents.
- Speaks to
- contact hypothesis and intergroup contact theorystreet-level bureaucracypolice reform and bureaucratic behaviorcitizen-state relationsfield experiments on governance
Cite
@article{haim2025policing,
title = {How Does Community Policing Affect Police Attitudes? An Experimental Test and a Theory of Bureaucrat-Citizen Contact},
author = {Dotan Haim and Matthew Nanes and Nico Ravanilla},
journal = {American Political Science Review},
year = {2025},
doi = {10.1017/S000305542400145X},
}