Procedural Justice: A Guide for Police Leaders — and How to Explain What It Actually Is

In 2016, 63 in every 100 people in England and Wales thought their local police did a good job. By 2025 that had fallen to 49, and it has barely moved in two years. Crime over the same period did not rise to explain it. Officer numbers, after the cuts of the 2010s, were largely rebuilt. Whatever has been eroding confidence, it is not simply “more crime, fewer police” — and procedural justice, now tested in randomised trials on three continents, points at a more useful explanation. People do not mostly decide whether to trust the police by totting up crime statistics. They decide it, encounter by encounter, from how they were treated while it was happening.

2016	63
2019	58
2023	51
2024	49
2025	49

That is the whole case for procedural justice in one sentence, and the rest of this guide is about making it usable: what the theory actually says, the trials that have tried to break it, and — because a leader’s real job is not believing the research but getting it into the head of a nineteen-year-old constable at 2am — a plain way to explain it on parade.

What procedural justice actually is

The theory has a single author of record, the American psychologist Tom Tyler, whose 1990 book Why People Obey the Law asked a question policing had mostly assumed the answer to: why do people comply with the law and with police officers, when the odds of being caught for any given offence are low and the penalties, discounted for that low probability, are rarely severe enough to explain it alone? Tyler’s answer, tested since across dozens of countries, courts and police forces, was that compliance runs less through fear of punishment than through legitimacy — the belief that the authority has the right to be obeyed — and that legitimacy is built overwhelmingly from how people are treated, not from what is decided or how effective the institution appears to be.

Tyler broke “how people are treated” into four components that have since become the standard vocabulary of the field, used by the College of Policing and by Police Now’s own officer training material alike:

• Voice — did they let you explain, or tell your side, before deciding?
• Neutrality — did the decision look consistent and unbiased, not made up on the spot for you specifically?
• Respect — were you treated with dignity, as a person rather than a problem?
• Trustworthy motives — did the officer’s reasons look like they were about your safety or the law, rather than a quota, a grudge or convenience?

None of the four requires a force to give anything up. An officer who searches someone, arrests them, or turns them away from a cordon can do every one of those four things and still search, arrest or turn them away. Procedural justice is not a theory about being lenient. It is a theory about how the power gets used, and the finding that keeps replicating is that the how matters to people about as much as the what — sometimes more.

The evidence: three trials, three continents

Theories about policing are cheap. What makes procedural justice unusual is how much of it has survived contact with a randomised controlled trial — the gold standard the rest of this site’s evidence-based-policing coverage, from Lawrence Sherman’s foundational work onward, keeps returning to.

Queensland, 2013. Criminologist Lorraine Mazerolle and colleagues ran the first randomised field trial of procedural justice in a live police encounter, in the Queensland Community Engagement Trial (QCET). Sixty roadside random breath-test operations were randomly assigned to business-as-usual policing or to a scripted procedurally just version — officers gave drivers a brief, standardised explanation of why the stop was happening and what would follow, then surveyed them afterwards. Drivers who got the scripted version rated the stop as significantly fairer, and 17.4% reported that the stop had changed their view of the police for the better, against 12.5% in the control group — a statistically significant gap produced by nothing more than what the officer said in the first thirty seconds of a stop that was going to happen regardless.

Manchester, 2013. Closer to home, the Greater Manchester Police procedural justice training experiment, run by Emma Wheller and Paul Quinton for the College of Policing, randomised 339 officers to a fourteen-hour communication-skills course and left 237 as a control group. The training focused on how officers dealt with victims of crime. Trained officers showed statistically significant shifts toward prioritising quality of service, building empathy and rapport, and making fair decisions — moderate effect sizes on a seven-point scale, the kind of shift a force can actually produce with a training budget rather than a change in the law.

Three US cities, 2017–2020. The largest and most rigorous test to date is a multicity randomised trial led by George Wood, David Weisburd and Lawrence Sherman, published in PNAS in 2022. Twenty-eight officers in Tucson, Cambridge (Massachusetts) and Houston received forty hours of procedural justice training and were sent to work the same 120 randomly assigned crime hot spots as a matched control group over nine months. The results were the sharpest yet produced:

• 60%+ Fewer arrests made by procedurally trained officers statistically significant, p<0.001
• 14% Relative fall in crime in their hot spots during the trial period
• d=−0.47 Drop in residents perceiving police harassment moderate-to-large effect
• d=0.39 Rise in observed voice, neutrality & respect independently coded encounters

Officers who had been through the training were independently observed giving people more voice, behaving more neutrally and showing more respect. They made well over half as many arrests as the control group for policing the same hot spots, residents reported markedly less perceived harassment and excessive force, and crime in their patch fell — all without the training touching a single power an officer holds. That combination, fewer arrests and less crime together, is the finding that gets a chief constable’s attention, because it says fairness is not a trade-off against effectiveness. In this trial it went the other way.

Why fairness beats force as a strategy

The mechanism behind all three trials is the same one Tyler set out originally, and it is worth stating plainly because it cuts against an old policing instinct that visible enforcement is what buys compliance. Research led by Jonathan Jackson and Ben Bradford, running through decades of British survey data, finds that the public’s readiness to cooperate with police — to report crime, to give evidence, to accept an officer’s decision on the night rather than fight it — tracks their sense that the police are legitimate far more closely than it tracks their sense that the police are effective. And legitimacy, in turn, is driven overwhelmingly by procedural fairness rather than by outcomes. A force that gets better at catching criminals without getting fairer in the process does not reliably become more trusted. A force that gets fairer, even without a single extra clear-up, tends to.

This is not a new insight dressed in new language. It is Robert Peel’s founding premise — that a police force holds its authority on loan from the public and depends on manner rather than force to keep it — confirmed with survey data and, now, with randomised trials Peel never had access to. The link runs directly through the site’s own coverage of neighbourhood policing and public confidence, which draws on the same LSE research programme. What procedural justice adds is precision: not just that fairness matters, but which four things fairness is actually made of, and evidence that training officers in them changes measured behaviour, not just attitudes on a survey.

Where it is tested hardest: stop and search

If procedural justice is right, its clearest test should be the power most likely to feel unfair when it is not: stop and search. It is. A 2020 study by Kath Murray and Susan McVie, comparing adolescents in Scottish and English cities, found that young people who experienced high-volume stop and search in Scotland held more negative views of the police and rated their stops as less procedurally fair than their English counterparts — and, using structural equation modelling, that the damage to trust and perceived legitimacy came specifically from stops experienced as unfair in manner and decision-making, not simply from the number of stops. The same study found early evidence that this loss of legitimacy was associated with an increase in offending behaviour: procedural injustice does not just cost a force goodwill, it may actively undermine the crime-control case for using the power at all.

That finding sits behind current national policy. The NPCC and College of Policing’s Police Race Action Plan commits to training for officers of every rank that goes beyond the lawfulness of a stop to its fairness and respect — decision-making, communication, de-escalation and cultural awareness — precisely because the power itself is not the whole story of whether a community experiences it as fair. Search rates and outcome rates matter and should be published and scrutinised. But the Murray/McVie finding is a reminder that a force can hit every legal test for a search and still lose the encounter, if the ten seconds before the search were rushed, unexplained and delivered without eye contact.

The caveat: it is not a script

A guide that only reported the wins would be worth less than this one aims to be, and the multicity trial above supplies its own caveat. Despite the drop in arrests and perceived harassment, the researchers found no statistically significant difference in perceived police legitimacy between the procedurally trained officers’ hot spots and the control group’s. Behaviour changed. Crime fell. People felt less harassed. But the deeper, harder-to-move sense of whether the police are a rightful authority did not shift in nine months, which is a useful corrective to any leader hoping a training day is a legitimacy fix rather than the start of a much longer discipline.

There is a sharper critique worth taking seriously too. In a widely cited 2012 paper, Anthony Bottoms and Justice Tankebe argued that legitimacy is not something an officer generates unilaterally inside a single encounter at all — it is dialogic: an ongoing back-and-forth in which the police make a claim to authority and communities respond to it, shaped by history, group identity and the accumulated weight of every previous encounter that community has had with the police, not reset to zero each time an officer says the right four things. Taken together, the fair reading is this: procedural justice is real, well-evidenced and free to implement, but it is not a substitute for addressing the disparities — in who gets stopped, who gets force used against them, who gets believed — that shape what an individual officer’s fairness is being asked to overcome before the encounter has even started. Treat it as the discipline that makes every encounter a little less likely to do damage, not as the fix for damage already done.

How to put it in a uniformed officer’s head

None of this is any use to a leader if it stays in a technical report. The genuine skill is compressing four decades of research into something a sergeant can say in a two-minute slot at parade, and something an officer can carry into a doorway at 2am without reaching for a script.

That is close to the script QCET’s officers were given for a roadside breath test, and it is deliberately unglamorous: it names the officer, states the real reason, offers voice, and promises an explanation of the outcome — voice, trustworthy motives, respect and neutrality in four sentences, before a single question has been asked. It does not soften the power. It does not slow the stop down by more than the time it takes to say it. What it changes is whether the person on the other end of it experiences the next two minutes as something done to them or something explained to them, and the evidence above says that difference is not a courtesy. It is most of what determines whether they trust the police afterwards.

The one-line version, for the officer who will not remember a table on a briefing slide: would I accept this, done to me, this way, by this person? If the answer is no, something in the voice, the neutrality, the respect or the reason given needs to change before the power is used, not after.

The leader’s other half: internal procedural justice

The research most useful to a chief constable or a PCC is not about how officers treat the public at all. It is about how officers are treated by their own force. Ben Bradford and Paul Quinton’s 2014 study of an English constabulary found that officers’ sense of their own self-legitimacy — their confidence that their authority is rightfully held — was predicted most strongly by whether they perceived their senior management as procedurally fair. Officers with a stronger sense of self-legitimacy were, in turn, more supportive of constitutional, rights-respecting policing, and less supportive of using force against the community. The effect held even after accounting for how officers rated the organisation’s fairness generally — it was specifically procedural fairness from supervisors that carried the weight.

This is not a side finding. It means a force cannot train its way to procedurally just policing of the public while running an unfair, arbitrary or disrespectful internal culture — the disrespect a sergeant shows a probationer at a debrief, or the arbitrary way a duties rota gets changed, is plausibly part of the same causal chain that ends with a member of the public being spoken to rudely on a doorstep. For a leader, that reframes the whole subject: procedural justice is not a module you commission from a training provider and then consider delivered. It is a description of how you run the organisation, and the officers watching how you run it are the same officers who will, or will not, extend the same fairness downward.

A leader’s checklist

Whatever a force decides to do with this evidence, the same handful of questions travels well from the research to the parade room:

• Is it trained as a discipline, not a session? The trials that worked used sustained, scenario-based training (fourteen to forty hours), not a single slide deck — and the Manchester and multicity trials both suggest a refresher matters.
• Is it modelled from the top, not just taught to the front line? Bradford and Quinton’s finding means a leader’s own procedural fairness toward supervisors is part of the intervention.
• Is it measured, not assumed? Body-worn video, complaint patterns and stop-outcome disparities are the practical proxies for the “voice, neutrality, respect, trustworthy motives” the trials coded directly.
• Is it applied hardest where the stakes are highest? Stop and search, use of force and custody are where procedural injustice does the most measurable damage to trust — and, per Murray and McVie, potentially to reoffending.
• Is the caveat in the training too? Officers who are told procedural justice is a magic fix for legitimacy will notice, correctly, when legitimacy does not move — and stop believing the rest of it. Teach it as what the evidence actually says: a discipline that reliably reduces harm and improves the immediate encounter, working alongside the harder, slower work of closing gaps in who gets stopped, searched and believed.

The four pillars are not complicated, and that is rather the point: nothing here requires new legal powers, a bigger budget or an Act of Parliament. It requires an officer who, in the few seconds before saying anything, has asked the question a good sergeant should already be asking on their behalf — would I accept this, done to me, this way? Officers who can answer that straight, most days, are already doing the thing four decades of research keeps rediscovering: that people give the police their consent one interaction at a time, and can withdraw it the same way.

See more in public confidence and legitimacy and stop and search.

---

Sources and further reading

• Tom R. Tyler et al., “Procedural Justice, Legitimacy, and Effective Law Enforcement”, Psychological Science in the Public Interest, 2015 — the four-component model (voice, neutrality, respect, trustworthy motives) and the wider evidence review.
• College of Policing, “What works: policing with procedural and distributive justice” and “Procedural justice in stop and search”.
• Police Now, “Procedural Justice” insight briefing — how the four-part model is put to serving graduate-entry officers.
• Lorraine Mazerolle et al., “Shaping Citizen Perceptions of Police Legitimacy: A Randomized Field Trial of Procedural Justice”, Criminology, 2013 (the Queensland Community Engagement Trial).
• Emma Wheller, Paul Quinton, Alistair Fildes and Andy Mills, “The Greater Manchester Police procedural justice training experiment”, College of Policing technical report, 2013.
• George Wood, David Weisburd and Lawrence Sherman et al., “Reforming the police through procedural justice training: A multicity randomized trial at crime hot spots”, PNAS, 2022.
• Jonathan Jackson and Ben Bradford, “Legitimacy and the Social Field of Policing”, and Ben Bradford and Paul Quinton, “Self-Legitimacy, Police Culture and Support for Democratic Policing in an English Constabulary”, British Journal of Criminology, 2014.
• Kath Murray and Susan McVie, “Procedural justice, compliance with the law and police stop-and-search: a study of young people in England and Scotland”, Policing and Society, 2020.
• Anthony Bottoms and Justice Tankebe, “Beyond Procedural Justice: A Dialogic Approach to Legitimacy in Criminal Justice”, Journal of Criminal Law and Criminology, 2012 — the critique that legitimacy is dialogic, not generated unilaterally.
• NPCC and College of Policing, Police Race Action Plan.
• ONS, Crime Survey for England and Wales — public perceptions of the police, confidence and “good job” time series.
• Related on this site: the profile of Sir Robert Peel, the profile of Lawrence Sherman, and boosting public confidence through neighbourhood policing.