Stop and search by ethnicity

The national picture

Searches in the 12 months to June 2026, against the resident population of each group from the 2021 census. Ethnicity is as the officer recorded it at the time, which is not the same as how the person would describe themselves.

Two things to know about these numbers. 14 of the 41 forces filed only part of the year, the largest being Metropolitan Police Service at 10 of 12 months and about 26% of all searches in the country. Rather than drop them and lose forces that size, each force's population is scaled to the months it actually filed, so a force reporting ten months is measured against ten months of residents. And 8.5% of searches have no ethnicity recorded at all: they are counted in the totals and can never appear in a rate. The Mixed row is not a national figure, because only 6 forces use that category and the rest place those searches elsewhere.

What the find rate shows

The rate at which a search turns something up is the most useful column in the table, because it is the one thing here that describes the search rather than the searched. If a group were being searched on weaker grounds than another, you would expect officers to find something less often when they did.

That is not what the figures show. Searches of Black people found something 30.8% of the time and searches of White people 27.4% of the time, with Asian people at 31.0%. The gap in how often people are searched is 5.7 to one. The gap in how often those searches find anything is a couple of points, and it does not run in the direction the disparity might lead you to expect.

That comparison cannot settle why the disparity exists. Find rates say nothing about the people who were never stopped, and a search can be lawful, productive and still part of a pattern worth questioning. What the figures do rule out is the simplest defence of the gap, that the extra searches are simply better targeted.

Every force compared

How many times more often Black people were searched than White people, per resident, in each force area. Ordered by that ratio.

Greater Manchester Police and Police Service of Northern Ireland are absent because they publish no stop and search data at all. Searches are counted where they were carried out, residents where they live, so the two columns do not describe the same set of people.

Questions

Does this show which ethnic groups commit more crime?

No, and no dataset on this site can answer that. Police recorded crime contains no information about the people involved in an offence, and a stop and search record describes a police decision rather than a proven one. A group being searched more often tells you where officers looked. It cannot tell you what was there to be found among everyone they did not stop.

Why is ethnicity officer-defined?

Because it is recorded at the moment of the stop, by the officer, from appearance. The police also collect self-defined ethnicity separately, and the two disagree often enough to matter. The resident populations these searches are divided by come from the census, where people describe themselves. So the numerator and the denominator are not measured the same way, which is a real limitation of every published version of this comparison, including the Home Office's own.

What counts as finding something?

An arrest or another formal sanction resulting from the search. A search that ends with no further action counts as finding nothing, as does one where the outcome was never recorded, which means the find rates here are a floor rather than an exact figure.

Where does this come from?

Stop and search records published by each force at data.police.uk under the Open Government Licence, and Census 2021 population by ethnic group from the Office for National Statistics. Both are listed on data sources. How the figures work explains the rest of the method.