More transparency on Stop and Search and ethnicity in Scotland?
For the first time, detailed information on stop and search at a local level has been made available through Police Scotland’s Local Policing Management Information reports. Carol Young explores what this data tells us about race equality in stop and search.
Institutional racism in stop and search has long been seen as one area where England and Scotland diverge in terms of race equality. Inequalities are well evidenced south of the border, with people from Black communities six times more likely to be stopped and searched in England and Wales in 2010.
In Scotland, stop and search is generally more prevalent than it is in England and Wales. This may be at least partly due to differences in legal powers, with non-statutory stop and search being prohibited in England and Wales since 2003. Looking at Police Scotland’s preferred measure of stop and search, the rate per 10,000 people, stop and search rates are almost seven times higher in Scotland than in England and Wales (1,285 stop and searches per 10,000 people compared to 180). Comparing the two areas with the highest rate of stop and search, Cleveland in England recorded 570 stop and searches per 10,000 people, whereas Glasgow in Scotland recorded 3,712. These are the latest comparable figures, for 2012/13. More recent figures for Scotland from Police Scotland’s Management Information 2013/14 show a slight drop, to 1,206 per 10,000.
Despite the relative prevalence of stop and search, to date national statistics have suggested that stop and search has little impact on race equality in Scotland. According to Police Scotland, over 2013/14 “95.9% of stop and searches recorded were conducted on persons of white ethnicity which is very close to the proportion of ethnic white people in the Scottish population in the 2011 Scottish Census (96%).” This is as far as that report’s overview of stop and search by ethnicity goes. Although statistical tables give more detail, the written analysis of race equality is purely about white communities.
This decision to focus on white groups as a whole and only at national level obscures what’s actually happening to specific communities. The Scottish Police Authority’s (SPA) May 2014 Scrutiny Review on stop and search fared no better, consolidating all minority ethnic communities into one category when in fact there are dramatic differences between them. The new data paints a complex picture, with specific ethnic groups experiencing stop and search rates far higher than the Scottish average in some areas. This was particularly pronounced in the west of Scotland, with Black communities in South Ayrshire five times more likely to be stopped and searched compared to the average rate.
 Non-statutory stop and search is a supposedly ‘consensual’ form of stop and search where no suspicion is required. A useful overview is given on p.19 of the SCCJR’s 2014 evaluation of police practice on stop and search.
 Rate for England and Wales calculated using rate per 1,000 from UK Government Police Powers and Procedures publication for 2012/13 The National Situation
At national level, by removing ‘other white’ groups from the white category used in Police Scotland’s overview, we can see that white Scottish people are actually over-represented in stop and search. The rate of stop and search per 10,000 people in the population for this group was 1,300.4. In other words, the number of stop and searches for this group equates to about 13% of all white Scottish people (bearing in mind that some people will have been stopped and searched several times; these figures do not represent individual people). Black communities were over-represented nationally to a slightly lesser degree, with a rate equivalent to around 11% of the Black population.
However, some communities were very under-represented, with equivalent rates of less than 2% for the Chinese community and 4% for people of mixed ethnicity.
On the face of it, the most over-represented group appeared to be the ‘other’ category. In accordance with the Scottish Census, this would represent the Arab population as well as any other ethnic groups out with the basic categories (all white, Asian, Chinese, Black and mixed groups). In 2013/14, the ‘other’ category recorded stop and searches equivalent to over a third of that population (33.9%).
However, in discussion with Police Scotland, it appears that this information is unreliable. Concerns about how well officers understand the ‘other’ category and consequent problems with recording practice have been raised. Clarity on the extent of the supposed over-estimation for this category is badly needed. It is assumed that work will be undertaken to address this in future; this work should ideally encompass recording practice overall, as problems with the perceptions or assumptions of officers could potentially affect the whole process.
The data below expands on Police Scotland’s approach to assessing racial proportionality of stop and search. The ethnicity categories listed below are those used by Police Scotland, which can be analysed by compiling equivalent sets of data from the Census. Comparing these figures shows how inadequate the ’95.9% white’ overview is, with white Scottish and Black ethnic communities experiencing far higher rates of stop and search than any of the other groups.
Number of Stop and Searches - 578,145
% of total stop and search - 90.2
% of Scottish population - 83.9
Proportionality* - 1300.4
Number of Stop and Searches - 36,310
% of total stop and search - 5.7
% of Scottish population - 12.1
Proportionality* - 569.1
Number of Stop and Searches - 791
% of total stop and search - 0.1
% of Scottish population - 0.4
Proportionality* - 395.5
Number of Stop and Searches - 7,047
% of total stop and search - 1.1
% of Scottish population - 2.0
Proportionality* - 658.6
Number of Stop and Searches - 3,914
% of total stop and search - 0.6
% of Scottish population - 0.7
Proportionality* - 1087.2
Number of Stop and Searches - 539
% of total stop and search - 0.1
% of Scottish population - 0.6
Proportionality* - 158.5
Number of Stop and Searches - 4,751
% of total stop and search - 0.7
% of Scottish population - 0.3
Proportionality* - 3393.6**
Number of Stop and Searches - 9,202
% of total stop and search -1.4
% of Scottish population -n/a
Proportionality* - n/a
* Rate of stop and search per 10,000 population in the equivalent Census category – over-represented groups are highlighted in bold
** The ‘other’ category is statistically unsound due to suspected flaws in the recording process
Diverse local areas, diverse stop and search rates
Local figures released by Police Scotland in July 2014 allow us to look closer at how stop and search plays out across Scottish Local Authority areas.
Proportionality data isn’t published at local level for the most nationally over-represented group, ‘other’, due to the previously mentioned concerns about data quality. Of the remaining groups, the white Scottish group were statistically the most over-represented at national level; however, they were only over-represented in half of the 32 local authority areas. In four of the areas where they were relatively over-represented, people in the Black ethnicity category were even more over-represented.
The national figure for the white Scottish category (1,300 per 10,000) is boosted by unusually high levels of stop and search for this group in certain areas. The three Ayrshire council areas, Inverclyde and Renfrewshire all recorded between 2,000 and 3,000 stop and searches per 10,000 white Scottish people resident in the area. In Glasgow, where stop and search rates are far higher, this group were targeted at a rate of 4,241 per 10,000 people. This is likely to be due to what the recent SPA Scrutiny Review described as an east / west divide, with policing in the west of Scotland more focussed on stop and search. The same report identifies a trend towards searches for alcohol; research could potentially explore whether this is linked with the disproportionate targeting of the white Scottish community in some areas.
The approach in the west is arguably seen as part of a tougher approach to crime reduction. Stop and search is credited with halving knife crime in Glasgow between 2006 and 2013. However, whether the extensive targeted stop and searches are an effective part of evidence-led policing or a human rights violation is contentious. The SPA Scrutiny Review concluded that overall, Glasgow police use stop and search to an unreasonable degree. Glasgow accounts for 19.5% of all crime in Scotland but 35% of stop and search activity.
For the Black ethnicity category, national levels of stop and search appear to be slightly below average, but local figures again tell a different story. Glasgow also showed disproportionate levels of stop and search for this group, although at 1,526 per 10,000 people in the Black population, nowhere near the disparity for the white Scottish community in this area.
In some regions, particularly smaller or more rural areas, only a small number of stop and search incidents aimed at Black communities were recorded. It’s common practice for researchers to suppress data which involves low numbers, on the grounds that these are not statistically reliable. If we choose to only include areas where more than 30 stop and searches were carried out against people from Black communities, the Black ethnicity category showed disproportionality in stop and search in 14 areas. However, if we opt to include those areas with under 30 stop and searches, the majority (24 out of 32) show over-representation of these communities in stop and search.
This raises questions about the potential for data suppression in research to minimise the appearance of inequalities. Although stop and search numbers may be low in some regions, in the vast majority of these there are sufficient numbers of people in each of the ethnic communities covered by the Census categories to enable analysis. There was no appearance of over-representation in any of the areas with small Black populations; the practice of using rates per 10,000 population doesn’t carry the same risk of skewing due to population size that would be found in percentage based measures. Given that stop and search impacts the people targeted on a personal level, the priority should be to explore fully what these figures can tell us about the experience of people in Scotland’s communities. Everyone resident in Scotland should be entitled to expect that their risk of stop and search is fair in comparison to people in other communities. Even where the number of incidents in an area is relatively low, if the rate of stop and search for particular groups is disproportionate then this should be examined and the reasons made public. Any indication of racial profiling should be dealt with by Police Scotland swiftly and decisively. Looking at the ten areas with the highest stop and search rates, there is a clear argument for action to investigate disproportionality. All of the areas listed are in the west of Scotland, specifically across Ayrshire and in Glasgow, Inverclyde and Renfrewshire. Most notably, in South Ayrshire Black communities were stopped and searched at an astonishing rate of 6,043 stop and searches per 10,000 people. This was far and away the highest rate in any area, for any ethnic group. Black communities were affected by the highest rates of disproportionality in four of the top ten areas, but tended to face worse disproportionality and so rank higher on the list, with white Scottish communities making up the bottom four.
Ten highest stop and search rates per 10,000 people, Scotland, 2013/14:*
South Ayrshire, Black communities: 6,043
Glasgow, white Scottish community: 4,241
East Ayrshire, white Scottish: 2,948
Inverclyde, Black communities: 2,932
East Ayrshire, Black communities: 2,925
North Ayrshire, Black communities: 2,837
South Ayrshire, white Scottish community: 2,684
Inverclyde, white Scottish community: 2,603
North Ayrshire, white Scottish community: 2,541
Renfrewshire, white Scottish community: 2,402
These high levels of stop and search stand in sharp contrast to those of the Chinese community. Only Glasgow and Edinburgh recorded more than 30 stop and searches overall for this community, with the highest rate per 10,000 being 330 (in Glasgow). At national level, Chinese communities were targeted in very few stop and searches; just 158 per 10,000 in the equivalent Census category in comparison to 1,087 in Black communities. This shows clearly how consolidating all minority ethnic groups into one category at national level can create the misleading impression that there is no ethnic disparity.
Measuring representation by ethnicity using Census data is far from straightforward, even at local level. It has previously been confirmed that there are problems with over-recording of stop and search, including accusations that figures are being deliberately massaged, but there are also practical difficulties.
A common concern is that the two sources of data measure different things. As mentioned previously, stop and search data measures the number of stop and searches overall and will therefore include individuals who have been stopped and searched multiple times (current recording practice means the extent to which this happens can’t be analysed). Census data, on the other hand, measures each individual.
Of course, the idea that some disproportionality in stop and search may be due to repeat incidents involving the same individuals raises another set of concerns and questions. Ideally, if national level data on ethnicity could be cross-referenced with positive results (i.e. where something potentially incriminating has been found) and whether searches are statutory or non-statutory (which can be used as an indicator of whether there was reasonable suspicion) it might be possible to assess a degree of reasonableness in addition to simple proportionality.
Another caveat relates to residence. Unlike Census data, stop and search figures will include individuals who aren’t resident in Scotland. So, for example, in the Scottish Borders a massive 43% of all stop and searches involve people of ‘white other’ ethnicity, but this group makes up only 20% of the population. It’s possible that most of this disparity is due to stop and searches against people who live across the border in England.
Despite these issues, comparison with Census data is an essential starting point for analysis of stop and search by ethnicity in Local Authority areas. Thorough analysis at both statistical and practical level within Police Scotland should be able to identify potential causes of disproportionality. This might relate to differences in how communities tend to spend their time, for example attendance at football matches or socialising in places where stop and search is more prevalent (such as town centres at pub closing time). Recording practice by police officers might also be an issue. Without further investigation, it’s impossible to say – and impossible to rule out racial profiling.
Despite the picture of equality that’s being painted nationally, some communities in the west of Scotland are experiencing unreasonably high levels of stop and search. The new local statistics clearly demonstrate this. A rate of stop and search equivalent to six out of ten people in Black communities, as seen in South Ayrshire, cannot be ignored. If Police Scotland is serious about tackling racial inequality, it’s time to start using data on stop and search effectively.