CoChat

Procedural justice training reduces police use of force and complaints against officers

Open access
Wood, George, Tyler, Tom R., Papachristos, Andrew V. · 2020 · Proceedings of the National Academy of Sciences

The DOI matches its Crossref record; 31 of the 31 references it cites were found in a registry or an index; 3 of the 24 sources its claims cite were checked for support.

Abstract

Existing research shows that distrust of the police is widespread and consequential for public safety. However, there is a shortage of interventions that demonstrably reduce negative police interactions with the communities they serve. A training program in Chicago attempted to encourage 8,480 officers to adopt procedural justice policing strategies. These strategies emphasize respect, neutrality, and transparency in the exercise of authority, while providing opportunities for civilians to explain their side of events. We find that training reduced complaints against the police by 10.0% and reduced the use of force against civilians by 6.4% over 2 y. These findings affirm the feasibility of changing the command and control style of policing which has been associated with popular distrust and the use of force, through a broad training program built around the concept of procedurally just policing.

Overview

CoChat Summary

TL;DR

A large-scale staggered rollout of 1-day procedural justice training in the Chicago Police Department reduced complaints, sustained or settled complaints, and use of force reports among trained officers over 24 months. The effects were durable and larger for early adopters, supporting officer retraining as a viable strategy for changing field behavior.

Key findings

  • Training reduced complaints by −11.6 per 100 officers over 24 months (95% CI: −15.60, −7.45; P < 0.001), a 10.0% reduction equivalent to approximately 732 fewer complaints.
  • Training reduced sustained or settled complaints by −1.67 per 100 officers over 24 months (95% CI: −2.81, −0.40; P = 0.008), a 15.5% reduction from 0.39 to 0.32 per 100 officers per month, or 105 fewer such complaints.
  • Training reduced mandatory use of force reports by −7.45 per 100 officers over 24 months (95% CI: −12.40, −3.37; P = 0.002), a 6.4% reduction equivalent to 486 fewer uses of force.
  • Effects were durable over 24 months and more pronounced for early adopters: after 18 months, early adopters had 9.54 fewer complaints, 1.28 fewer sustained/settled complaints, and 8.72 fewer uses of force per 100 officers, compared with 6.04, 0.74, and 4.51 for late adopters.

Methods

  • The study used a staggered adoption design: 8,480 CPD officers were assigned to a 1-day procedural justice training from January 2012 through March 2016, with 138 more trained after the evaluation period; analysis covered 8,618 officers in 328 clusters over 63 months (January 2011–March 2016).
  • Outcomes were complaints received, complaints sustained or resulting in a settlement payout, and mandatory officer-filed use of force reports (Tactical Response Reports), aggregated by training-date cluster and month.
  • Effects were estimated with an interactive fixed effects model, reporting average treatment effect on the treated per 100 officers per month and cumulative ATT over 24 months, with 2,000 block bootstraps at the cluster level and placebo tests introducing training 3 and 5 months early.

Limitations

  • The paper states that outcome data do not necessarily capture the full range of misconduct or use of force: many people who believe they were mistreated do not file complaints, complaint filing can be costly and complicated, and TRR filing is mandatory but compliance is not guaranteed.
  • The paper states that complaints reflect civilian assessments that may not align with legally or procedurally inappropriate behavior, though sustained or settled complaints reflect demonstrated violations; it also notes obstacles to settlements and racial disparities in complaint dispositions.
  • The paper states that officers appointed during the study period were excluded, so estimates concern retraining serving officers rather than training new officers, and that a revised later training module could not be evaluated due to insufficient follow-up data. The text also makes evident that late adopters were observed for shorter posttraining periods and that spillover from early adopters may have attenuated effects among late adopters.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1073/pnas.1920671117

Cites

31 of 31 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1. The staggered adoption of procedural justice training in the CPD; 8,480 officers were trained in 305 clusters across 49 mo. Once trained, clusters shift from the pretraining to posttraining condition. The rollout consisted of an initial pilot training program in month 14, followed by a first training phase from months 18 to 34, a 6-mo period in which no training occurred, and a second training phase from months 41 to 63. The study period ends at month 63, the last month for which outcome data are available, such that the 22 clusters trained after month 63 remain in the control condition throughout. The frequency of officers trained per month is shown in the top margin. In this visualization, we grouped clusters by the month in which they were trained. The frequency of clusters per training month is shown in the right margin.
Fig. 1. The staggered adoption of procedural justice training in the CPD; 8,480 officers were trained in 305 clusters across 49 mo. Once trained, clusters shift from the pretraining to posttraining condition. The rollout consisted of an initial pilot training program in month 14, followed by a first training phase from months 18 to 34, a 6-mo period in which no training occurred, and a second training phase from months 41 to 63. The study period ends at month 63, the last month for which outcome data are available, such that the 22 clusters trained after month 63 remain in the control condition throughout. The frequency of officers trained per month is shown in the top margin. In this visualization, we grouped clusters by the month in which they were trained. The frequency of clusters per training month is shown in the right margin.CC BY-NC-ND · PubMed Central
Fig. 2. (Top) Observed and counterfactual estimates of complaints, sustained or settled complaints, and use of force per 100 officers per month. Months are recalibrated to be relative to the onset of training. (Bottom) The ATT for each month is the estimated counterfactual frequency subtracted from the observed frequency in that month. Monthly ATT estimates are colored according to their value relative to zero. The 95% CIs are computed using 2,000 block bootstrap runs at the cluster level.
Fig. 2. (Top) Observed and counterfactual estimates of complaints, sustained or settled complaints, and use of force per 100 officers per month. Months are recalibrated to be relative to the onset of training. (Bottom) The ATT for each month is the estimated counterfactual frequency subtracted from the observed frequency in that month. Monthly ATT estimates are colored according to their value relative to zero. The 95% CIs are computed using 2,000 block bootstrap runs at the cluster level.CC BY-NC-ND · PubMed Central
Fig. 3. Placebo tests for the effect of training on complaints, sustained or settled complaints, and use of force. In the placebo tests, the training is artificially introduced before the observed onset of training. The placebo ATT is estimated for the period between the artificial onset and the true onset, denoted by the blue region. The P value for the placebo ATT is shown. (Top) Training is artificially introduced 3 mo before the true onset, and the placebo ATT is calculated for the period −3 mo to 0 mo. (Bottom) Training is artificially introduced 5 mo before the true onset, and the ATT is calculated for the period −5 mo to 0 mo. For all outcomes, we find no evidence for an effect of training in the placebo period, and the models pass the placebo test. The 95% CIs and P values are computed using 2,000 block bootstraps at the cluster level. The uncertainty for the placebo ATT is larger than the ATT in the full data because the placebo tests are informed by fewer observations in the pretraining condition.
Fig. 3. Placebo tests for the effect of training on complaints, sustained or settled complaints, and use of force. In the placebo tests, the training is artificially introduced before the observed onset of training. The placebo ATT is estimated for the period between the artificial onset and the true onset, denoted by the blue region. The P value for the placebo ATT is shown. (Top) Training is artificially introduced 3 mo before the true onset, and the placebo ATT is calculated for the period −3 mo to 0 mo. (Bottom) Training is artificially introduced 5 mo before the true onset, and the ATT is calculated for the period −5 mo to 0 mo. For all outcomes, we find no evidence for an effect of training in the placebo period, and the models pass the placebo test. The 95% CIs and P values are computed using 2,000 block bootstraps at the cluster level. The uncertainty for the placebo ATT is larger than the ATT in the full data because the placebo tests are informed by fewer observations in the pretraining condition.CC BY-NC-ND · PubMed Central
Fig. 4. Cumulative ATT per 100 officers in the 24 mo after training by adoption time. The cumulative ATT represents the total reduction in each outcome in the months since the onset of training. Early adopters are those officers trained in the first training phase, encompassing the first 17 mo of the rollout. Late adopters were trained in the second phase from months 41 to 63. The study period ends in month 63 for complaints and use of force and month 58 for sustained and settled complaints. Consequently, the cumulative ATT for late adopters extends for a maximum of 23 mo after training for complaints and use of force and 18 mo for sustained and settled complaints.
Fig. 4. Cumulative ATT per 100 officers in the 24 mo after training by adoption time. The cumulative ATT represents the total reduction in each outcome in the months since the onset of training. Early adopters are those officers trained in the first training phase, encompassing the first 17 mo of the rollout. Late adopters were trained in the second phase from months 41 to 63. The study period ends in month 63 for complaints and use of force and month 58 for sustained and settled complaints. Consequently, the cumulative ATT for late adopters extends for a maximum of 23 mo after training for complaints and use of force and 18 mo for sustained and settled complaints.CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

3 Valid · 0 Not valid · 15 Not checked · 60 Own results and interpretations

46 more claims

Create a free account to see all 78 claims and to verify the 3 references still open.

See all claims — free account
  • It may state two separate claims as one. 16
  • Probably not a substantial claim. 12

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

3 of 3

Citations

Cites

Show all 31 references

Research works printed but not linked: 4

  1. C. Lum , An Evidence-Assessment of the Recommendations of the President’s Task Force on 21st Century Policing (Center for Evidence-Based Crime Policy, Fairfax, VA, 2016).
  2. T. R. Tyler, J. Fagan, Why do people cooperate with the police? Ohio State J. Crim. Law 6, 231–275 (2008).
  3. J. D. McCluskey, Police Requests for Compliance: Coercive and Procedurally Just Tactics (LFB Scholarly Publishing LLC, New York, NY, 2003).
  4. S. Walker, V. W. Bumphus, The effectiveness of civilian review: Observations on recent trends and new issues regarding the civilian review of the police. Am. J. Police 11, 1–16 (1992).

Cited by

See all 219

Metadata

Authors
Wood, G., Tyler, T. R., & Papachristos, A. V.
Affiliations
  • Northwestern University (US)
  • Yale University (US)
Corresponding author
Wood, George
Venue
Proceedings of the National Academy of Sciences
Published
2020-04-20
Volume
117
Issue
18
Pages
9815-9821
ISSN
  • 0027-8424 (ISSN-L)
  • 1091-6490
OpenAlex
W3016948921
Cited by
215
References
27
Citations by year
2026: 14 · 2025: 31 · 2024: 49 · 2023: 24 · 2022: 45 · 2021: 34 · 2020: 18
FWCI
104.46
Type
journal article
Language
en
Open access
Locations
  • Proceedings of the National Academy of Sciences (publishedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • RePEc: Research Papers in Economics (submittedVersion)
Topics
  • Social Sciences › Political Science and International Relations › Policing Practices and Perceptions
  • Social Sciences › Law › Criminal Justice and Penology
  • Social Sciences › Sociology and Political Science › Crime Patterns and Interventions
Keywords
procedural justice, police use of force, citizen complaints against police, police training, police-community relations, distrust of police, transparency, neutrality, respect

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Procedural justice training reduces police use of force and complaints against officers. https://app.cochat.ai/papers/10.1073/pnas.1920671117
The paper
Wood, G., Tyler, T. R., & Papachristos, A. V. (2020). Procedural justice training reduces police use of force and complaints against officers. Proceedings of the National Academy of Sciences. https://doi.org/10.1073/pnas.1920671117

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.