Across all forces reporting, 24.5% of completed reviews concluded the force's own decision was not reasonable and proportionate. Not that the original complaint was proved. That the force's handling of it did not stand up.
The four things this section measures
Complaints are only one of the records a force leaves behind. We publish four, because each one answers a question the others cannot, and because a force that looks unremarkable on one can look very different on another.
Complaints, and what survives review
How many complaint cases a force logged, how many went to review, and how often the review found the force's own handling was not reasonable and proportionate. Nationally that is 24.5%, and the spread between forces is far wider than the average suggests.
Stop and search, and who gets stopped
Thirty-six months of searches with outcomes, plus the share ending in no further action, nationally 70.8%. We compute each force's ethnic disparity against the resident population of its own area using the 2021 Census, which is the comparison that makes a rate mean anything.
Use of force, and deaths after contact
Home Office use of force reports over a five-year series, and the IOPC's record of deaths during or following police contact. The second is the gravest thing any of this data records, and it is presented with its categories intact rather than collapsed into one number.
Where the search disparity is widest
How many times more likely a Black person is to be searched by this force than their share of the force area's resident population would produce. A ratio of 1.0 would mean searches exactly match the population. This is a measured difference in outcomes; it is not proof that any officer acted unlawfully, and it does not identify a cause.
| # | Police force | Black ratio | Searches |
|---|---|---|---|
| 1 | City of London Police | 10.6× | 5,984 |
| 2 | Norfolk Constabulary | 9.2× | 6,204 |
| 3 | Dorset Police | 9.1× | 4,892 |
| 4 | Sussex Police | 8.2× | 19,348 |
| 5 | West Mercia Police | 7.9× | 10,540 |
| 6 | Surrey Police | 7.7× | 16,875 |
| 7 | Warwickshire Police | 7.0× | 3,833 |
| 8 | Lincolnshire Police | 6.8× | 6,574 |
| 9 | Suffolk Constabulary | 6.8× | 6,373 |
| 10 | Wiltshire Police | 6.6× | 5,762 |
All 36 forces, with the denominators and the forces we cannot publish
Deaths during or following police contact
The IOPC's published categories for 2004/05 to 2024/25, England and Wales. These are the categories as the IOPC defines them. The category does not itself imply that the police caused a death or acted wrongly.
| Category | Deaths |
|---|---|
| Deaths following other police contact | 1,299 |
| Apparent suicides following custody | 1,149 |
| Road traffic incidents | 612 |
| Deaths in or following custody | 399 |
| Fatal police shootings | 44 |
How to complain, in three routes
They are not alternatives. Which one applies depends on what happened and how serious it was, and using one does not stop you using another.
- Complain to the force. Most complaints start and finish here. A force must normally record a complaint and give you a written outcome, and you do not need a reason, a solicitor, or a form.
- Ask for a review. If the outcome is wrong, you can ask for it to be reviewed, by the force's own local policing body or by the IOPC, depending on the case. Nationally 24.5% of reviewed outcomes do not survive that, which is the strongest argument on this page for actually using the route.
- Bring a civil claim. A separate process with its own strict time limits, which can run out while a complaint is still being handled. If a claim is a possibility, take advice on the deadline early.
The three routes in full Build a complaint letter
Start with your force
Complaint handling varies enormously between forces. Find yours to see how it handles complaints, how often its decisions are overturned, what it does when it stops and searches people, and how that compares with its neighbours.
See all 44 forces Find my force by postcode
Data last refreshed . Every figure on this page is recomputed from the source datasets, not edited by hand. Found something wrong? Tell us.