54 Police Body Camera Statistics

These police body camera figures cover the measurement periods and populations identified in the available source material. Estimates, surveys and forecasts retain their reported scope.

Key Police Body Camera Statistics

The main findings are:

  • 47% of 15,328 general-purpose state and local law-enforcement agencies had acquired body-worn cameras in 2016 (United States general-purpose law-enforcement agencies).
  • All local departments serving 1 million or more residents reported body-camera use in 2020 (U.S. local departments serving 1 million or more residents).
  • About 86% of agencies that acquired body cameras had a formal policy or one under development in 2016.
  • 77% of agencies without cameras cited video storage or disposal costs as a reason for not obtaining them (2016).
  • A six-month British randomized trial assigned 430 police shifts to body-camera treatment or control conditions (Six-month trial; Large British police force).
  • A Campbell systematic review estimated a 16.6% reduction in citizen complaints, with a 95% confidence interval from 0.7% to 30.0% reduction (Systematic review evidence through 2019; Studies of police body cameras).
  • 47.4% of general-purpose agencies had body-worn cameras, compared with 69.0% with dashboard cameras and 38.1% with personal audio recorders (2016).
  • Body-camera use was 90.3% among departments serving 500,000-999,999 residents in 2020 (United States local police departments).
  • 93% of agencies with a camera policy specifying events required recording traffic stops (2016).
  • 74% of agencies without cameras cited hardware costs (2016).
  • The British trial estimated a 50% reduction in the odds of force when cameras were present (Six-month trial; Large British police force).
  • Perth’s trial policing sub-district covered 21 square kilometers, while Bunbury’s covered more than 1,000 square kilometers (2016 trial; Perth and Bunbury, Western Australia).
  • About 45% of U.S. law-enforcement agencies had at least some body cameras in service in 2016.
  • Body-camera use was 94.3% among departments serving 250,000-499,999 residents in 2020 (United States local police departments).
  • About 85% required recording officer-initiated citizen contacts (2016).

Contents

Agency acquisition, not necessarily full deployment.

Survey estimates; recording-device categories are not mutually exclusive.

DOJ OIG summary of BJS; in-service differs from acquired.

Full deployment means all intended personnel.

Categories describe deployment status at survey time.

Different BJS survey years and instruments; not a continuous annual series.

BJS reports the increase; do not interpret it as percentage-point growth.

Officer-weighted estimate, not the share of departments.

BJS 2020 sheriffs’ statistical-table highlight; deployment is not identical to authorization.

Police body camera statistics by agency size and population served

BJS flags the estimate as based on the comparison group.

Population-served grouping; some estimates have BJS caution flags.

BJS marks the difference from the comparison year as statistically significant.

Department-level estimate.

Small-department estimate; interpret with BJS significance/caution notes.

Police body camera policies, recording rules and video access

Formal policy includes policies under development.

Conditional on having the relevant policy provision.

Conditional policy estimate.

Conditional on event-recording policy and local department type.

Conditional on event-recording policy.

Policy presence, not measured compliance.

Policy presence, not actual officer behavior.

Direct access is defined as access without filing a formal request.

Conditional on agency having acquired cameras.

Multiple access categories could be selected.

Police body camera costs, barriers and alternative recording technology

Reasons are non-mutually-exclusive self-reports.

Non-mutually-exclusive self-reported reason.

Self-reported, non-exclusive reasons.

Self-reported reason; not an objective need assessment.

Alternatives are non-mutually-exclusive.

Alternative technologies, not body-camera adoption.

Police body camera outcomes in randomized trials

  • A six-month British randomized trial assigned 430 police shifts to body-camera treatment or control conditions (Six-month trial; Large British police force). (Testing the effects of police…)
  • The British trial estimated a 50% reduction in the odds of force when cameras were present (Six-month trial; Large British police force). (Testing the effects of police…)
  • The British trial estimated a 35% reduction in overall weighted force under treatment (Six-month trial; Large British police force). (Testing the effects of police…)
  • The British trial detected 40% more force in treatment conditions for handcuffing non-combatant suspects (Six-month trial; Large British police force). (Testing the effects of police…)
  • The Western Australia randomized trial lasted six months and involved 498 officers (Six-month trial in 2016; Perth and Bunbury, Western Australia). (The results of a randomized…)
  • The Rialto randomized trial assigned 988 officer shifts to camera and control conditions over 12 months (12-month trial; Rialto, California). (The Effect of Police Body-Worn…)
  • Rialto body-camera program cost about $90,000, including 70 camera units, docking, management software and training (Program implementation; Rialto, California). (Program Profile: Police Body-Worn Cameras…)
  • In the Rialto protocol-adherent subgroup, use-of-force rates were 37% lower than in control conditions (12-month randomized trial subgroup; Rialto, California). (Program Profile: Police Body-Worn Cameras…)
  • When Rialto officers used discretion over activation, use-of-force rates were 71% higher than in control conditions (12-month randomized trial subgroup; Rialto, California). (Program Profile: Police Body-Worn Cameras…)

Single force and shift-randomized design.

Estimate is for use of force per arrest in that force.

Trial estimate; weighted-force definition matters.

Subcategory result; authors discuss possible reporting/transparency effects.

Officers acted as their own controls and treatment was randomized across shifts.

Single mid-sized department; shift-level randomization.

Reported program cost; no inflation adjustment or lifecycle-cost estimate.

Subgroup result conditional on following the activation protocol.

Subgroup result; should not be generalized beyond the study setting.

Police body camera evidence on complaints, force and justice outcomes

  • A Campbell systematic review estimated a 16.6% reduction in citizen complaints, with a 95% confidence interval from 0.7% to 30.0% reduction (Systematic review evidence through 2019; Studies of police body cameras). (Body-worn cameras’ effects on police…)
  • Perth’s trial policing sub-district covered 21 square kilometers, while Bunbury’s covered more than 1,000 square kilometers (2016 trial; Perth and Bunbury, Western Australia). (The results of a randomized…)
  • The Australian trial setting included a Perth population of over 2 million and a Bunbury population of about 45,000 (2016 trial; Perth and Bunbury, Western Australia). (The results of a randomized…)
  • Rialto complaints fell from 0.70 to 0.07 per 1,000 police-public contacts in the pre/post analysis (Before versus during 12-month trial; Rialto, California). (The Effect of Police Body-Worn…)
  • The Washington, D.C., trial estimated 74 additional uses of force per 1,000 officers per year, with a standard error of 87 (Trial follow-up, annualized). (A randomized control trial evaluating…)
  • The Washington, D.C., trial estimated 57 additional complaints per 1,000 officers per year, with a standard error of 41 (Trial follow-up, annualized). (A randomized control trial evaluating…)
  • In the Miami Beach trial, 30.1% of 2,605 charges ended in conviction or adjudication withheld, while 13.5% ended in a formal conviction (Six-month trial with 12-month follow-up; Miami Beach, Florida). (Local Police Departments, Procedures, Policies,…)

Review notes it is unclear whether fewer complaints reflect better interactions or changed reporting.

Study-setting context, not a camera outcome.

Local context; these are study populations, not national estimates.

Pre/post estimate is not the same as a randomized between-group effect.

Estimate was not statistically significant and is compatible with no effect.

Estimate was not statistically significant; complaint reporting and administrative capture matter.

Overall sample outcomes; not an estimate of the camera effect by itself.

Table of contents