Research and analysis

Stop and search data and the effect of geographical differences

Published 31 March 2021

Applies to England and Wales

Introduction

The Race Disparity Unit’s Ethnicity facts and figures website shows data about the experiences and outcomes for different ethnic groups in areas including crime and policing, education and employment.

One of the biggest disparities we have identified is the stop and search rate between Black and White people, which has widened in recent years.

This report looks at statistical factors that affect how we calculate and show:

• stop and search rates
• relative disparities between different ethnic groups in stop and search rates

An example of a relative disparity is that Black people were 5.8 times as likely as White people to be stopped and searched in Nottinghamshire in the year ending March 2020.

Conclusions

There are 3 main factors affecting how we interpret stop and search rates and the relative disparity between Black people and White people:

• geographic clustering
• population changes
• the number of people who do not report their ethnicity

The overall relative disparity between Black and White people in England and Wales could be misleading if geographical differences are not taken into account.

This is because the national relative disparity is influenced by how both stop and searches and ethnic minority group populations are clustered in particular areas, particularly London and areas with higher deprivation.

Stop and search is typically used in areas of higher deprivation or crime. Even at police force area level, there is significant variation in levels of stop and search, including areas where levels of stop and search are much more concentrated. The make-up of the populations in those areas are quite different too. (You can see a recent example of academic research into this topic).

For these reasons, ideally relative disparities should be analysed for small geographical areas. However, police force areas are currently the lowest level of geography available for this analysis.

Home Office has ambitions to provide more granular data in the future.

Geographic clustering

Local and national disparities between Black and White people appear contradictory because of geographic clustering of populations and stop and searches. This applies to:

• Black people in the Metropolitan Police Force area in London
• White people outside London

In future updates of our data, we will report on relative disparities only within police force areas and not for England and Wales as a whole.

We will use either or both of:

• the median stop and search relative disparity for the 43 police force areas
• a range of values for the 43 police force areas – for example, the maximum and minimum values for a particular year

We will continue to quote the national rate for every 1,000 people by ethnicity.

Population changes

The stop and search rates themselves are based on 2011 Census population data, which is now nearly 10 years out of date.

In addition, the rates do not distinguish whether someone who is stopped and searched in a particular police force area is a resident or visitor in that area – the rate is based only on the resident population of each area.

Stop and search: latest facts and figures

In the year ending March 2020, there were nearly 564,000 stop and searches in England and Wales, excluding:

• data for Greater Manchester, whose police force were not able to provide complete data
• stops for vehicles only – ethnicity data is not collected for this type of stop and search

20.8% of all stop and searches involved Black people.

There were 54 stop and searches for every 1,000 Black people compared with 6 stop and searches for every 1,000 White people.

The most common reason for a stop and search was suspicion of possessing drugs. Other reasons include suspicion of having stolen property or having an offensive weapon.

Figure 1: Stop and search rates for every 1,000 Black people and White people (England and Wales, including British Transport Police, April 2009 to March 2020)

Note: Chart excludes data for Greater Manchester for all years.

The stop and search rate for Black people was around 6 times higher than the rate for White people in the year ending March 2010.

This is the ‘relative disparity’ – the stop and search rate for Black people divided by the rate for White people.

Table 1: Stop and search rates for every 1,000 Black people and White people, and relative rates (England and Wales, including British Transport Police, April 2015 to March 2020)

Year ending	Rate for every 1,000 Black people	Rate for every 1,000 White people	Relative rate
March 2016	33	5	6.5
March 2017	31	4	8.4
March 2018	30	3	9.5
March 2019	39	4	9.7
March 2020	54	6	8.9

Note: Table excludes data for Greater Manchester for all years.

Table 1 shows that in the 3 years to March 2018, the stop and search rate went down slightly faster for White people than for Black people.

The stop and search rate for Black people was 6.5 times higher than the rate for White people in the year ending March 2016, and 8.9 times higher in the year ending March 2020. This figure was lower than that for March 2019 (9.7 times).

Impact of geographical differences

To be able to interpret relative disparities for stop and search, it is important to be aware of geographical differences in:

• the use of stop and search
• the ethnic diversity of the population – some ethnic minority populations are concentrated in certain police force areas

Stop and searches are concentrated in certain areas

The overall use of stop and search is concentrated in London (the Metropolitan Police force area). Stop and searches of Black people are even more concentrated there.

In the year ending March 2020:

• 48.9% of all stop and searches in England and Wales were made by the Metropolitan Police in London
• 80.0% of all stop and searches involving Black people were made by the Metropolitan Police in London
• 90.0% of all stop and searches involving Black people took place in 8 police force areas (out of 43 in England and Wales, excluding Greater Manchester)

These figures include stop and searches by the British Transport Police, and exclude those in Greater Manchester who could not provide a complete return in the year ending March 2020.

Figure 2: Police force areas with the highest numbers of stop and search of Black people (England and Wales, including British Transport Police and excluding Greater Manchester, year ending March 2020)

Note: Chart excludes data for Greater Manchester.

Stop and search rates vary by police force area

In the Metropolitan Police force area in the year ending March 2020:

• there were 71.2 stop and searches for every 1,000 Black people
• there were 17.6 stop and searches for every 1,000 White people
• the relative disparity between Black and White people was 4.0

The relative disparity between Black and White people was highest in the following areas:

• Dorset, where the stop and search rate for Black people was 23 times the rate for White people
• West Mercia (14 times higher)
• Warwickshire (13 times higher)

However, despite having high relative disparities between Black and White people, none of these areas were among the police forces making the largest numbers of searches of Black people. The number of searches of Black people was small, as was the number of Black people living in these areas (based on Census 2011 estimates). Users of the data need to exercise caution when interpreting disparities based on areas with small numbers.

Figure 3: Stop and search rates for every 1,000 Black people and White people, by police force area (England and Wales, not including Greater Manchester, City of London and British Transport Police, year ending March 2020)

Note: Chart excludes data for Greater Manchester.

How regional differences affect the overall disparity

When we calculate the relative disparity for England and Wales, we are dividing the overall stop and search rate for Black people by the rate for White people.

There is an assumption that stop and searches, and the population, are spread fairly evenly across the country. However, as we have seen, there are big differences between police force areas in:

• numbers of stop and searches
• stop and search rates for each ethnic group
• ethnic minority and White populations

Table 2: Stop and search rate for every 1,000 Black people and White people, and percentage of stop and searches, by police force area (England and Wales, year ending March 2020)

	Stop and search rate for every 1,000 Black people	Percentage of all stop and searches of Black people (%)	Stop and search rate for every 1,000 White people	Percentage of all stop and searches of White people (%)	Relative disparity between Black and White people
Metropolitan Police	71.2	80.0	17.6	30.7	4.0
All other police forces combined (excluding BTP and Greater Manchester)	25.3	18.3	4.6	67.9	5.4
All forces (including BTP and excluding Greater Manchester)	54.1	100.0	6.1	100.0	8.9

Note: The ‘All forces’ figures in this table include stop and searches by the British Transport Police (BTP) and exclude those in Greater Manchester. These figures will differ from the aggregate of the Metropolitan Police Force and all other police forces in this table as those totals exclude stop and searches by the BTP.

If we simplify our comparison of areas to the Metropolitan Police force area against all other areas combined, we find the stop and search rate for Black people was:

• 4.0 times higher than for White people in the Metropolitan Police force area
• 5.4 times higher than for White people in all other police force areas combined

Both figures are lower than the relative disparity between Black and White people in England and Wales as a whole (8.9). This is a statistical paradox, and is a feature of many datasets where the outcomes for different population sub-groups vary substantially.

The paradox in stop and search data is because the police force areas that influence the rates for Black people are different to the areas that influence the rates for White people.

For Black people, the police force area which most influences the national disparity is the Metropolitan Police force area. This is where 4 out of 5 stop and searches of Black people happen, and where nearly 60% of Black people live.

For White people, it is the very low stop and search rate in police forces outside the Metropolitan Police force area that most strongly influences the national disparity. In these areas, the stop and search rate for White people is 4.6 for every 1,000 people, and around 90% of White people live in them.

It is also the reason for the changes in the national disparity from the year ending March 2016, from over 6 times to around 9 times in the year ending March 2020.

In the year ending March 2016, the stop and search rate for Black people in the Metropolitan Police Force area was 42.4 for every 1,000 people. For White people outside this area, it was 4.0 for every 1,000 people.

The changes to the relative disparity in the year ending March 2020 have been most impacted by both:

• an increase in the stop and search rate for Black people in the Metropolitan Police Force area
• little or no change in the very low stop and search rate for White people outside the Metropolitan Police force area

If we present the statistics in the way outlined earlier:

• the median relative disparity between Black people and White people is 5.8 times (Nottinghamshire)
• the largest disparity was in Dorset (22.5 times)
• the smallest disparity was in North Wales (2.0 times)

You can read more about Simpson’s Paradox, a phenomenon in probability and statistics in which a trend appears in several different groups of data but disappears or reverses when these groups are combined.

You can also read about the modifiable areal unit problem, where summary statistics for different places (such as police force areas) are affected by how the size and shape of the places are defined.

Impacts of population changes

Using 2011 Census data

A further issue with stop and search rates is that they are calculated using population data from the 2011 Census.

The Census provides very accurate population data at a specific point in time. Data is available for small or less commonly-used geographical areas. This is very useful for calculating rates in police force areas which do not necessarily align with standard regions.

But there are 2 problems with using Census population data to work out stop and search rates:

• as time passes, the data no longer gives an accurate picture of the national and regional populations due to births, deaths, net migration and people moving house or changing their job
• population figures do not include visitors, students or a large influx of workforce – in authorised section 60 legislation areas, the population is often changeable or unknown

The Office for National Statistics is working on gaining population estimates split by ethnicity, using Annual Population Survey data, mid-year population estimates and data from the Census.

If these population estimates are accurate, we will assess their impact on stop and search rates. Using these estimates to calculate rates may allow population changes to be better reflected.

Section 60 legislation

The impact of population estimates is particularly relevant for the numbers of stop and searches under section 60 legislation.

Stop and searches take place under 3 legislative powers:

• section 1 of the Police and Criminal Evidence Act 1984 – the police can stop and search someone if they think they are carrying items like stolen property or drugs
• sections 44 and 47a of the Terrorism Act 2000 – the police can stop and search someone if they suspect an act of terrorism is about to take place (after authorisation by a senior officer)
• section 60 of the Criminal Justice and Public Order Act 1994 – the police can stop and search someone within an authorised area to prevent violence involving weapons

Section 60 is only authorised for use during a specific period and a defined area (for example, in an area where violent incidents are being anticipated), so the population of the area is often changeable or unknown. Because of this, we only show numbers of section 60 searches rather than rates.

Unreported ethnicity

In the year ending March 2020, ethnicity was not recorded in 17.3% of stop and searches in England and Wales (about 98,000, excluding data for Greater Manchester). This figure is similar to the number of stop and searches of Black people in that year.

This level of unreported data impacts the quality of the data. The extent of the impact depends on the actual distribution of the unreported cases across each of the ethnic groups (if we did know their ethnicity).

We are planning to produce a methods and quality report in the future looking at the impact of missing or unreported ethnicity on different datasets.

Next steps

This report shows that there are lessons to be learned around other outcomes which show considerable variation by geographic area.

We will continue to explore the data which shows this geographic variation, and will look at whether the same phenomena occur for other characteristics, such as gender and disability.

Policy actions could target different areas or people with different characteristics to take account of these disparities. Decision making can be improved by making sure that data is disaggregated to create effective interventions.

Further information

If you would like further information, or discuss this report in more detail, please contact Alice Whitfield (alice.whitfield@cabinetoffice.gov.uk) or Darren Stillwell (darren.stillwell@cabinetoffice.gov.uk).

Acknowledgements

The RDU is grateful for advice provided by Charles Lound and Martin Ralphs (Office for National Statistics), and Rosanna Currenti (Home Office).