Stop and Search: Exploring Disproportionality

Transcription

1 Stop and Search: Exploring Disproportionality Jean Hine Community and Criminal Justice Division July 2015

2 Stop & Search: Exploring Disproportionality Page 2

3 Table of Contents Acknowledgements... 4 Foreword... 5 Executive Summary... 7 Introduction Who gets stop searched? Why are people stop searched? Outcomes of Stop Searches Ethnicity and difference Change over time Changes over time within ethnic groups Explaining ethnic difference Conclusions APPENDIX A: Methodology APPENDIX B: Leicestershire Stop & Search Record APPENDIX C: Stop & Search Grounds Categories APPENDIX D: Information Sheet and Interview Questions Stop & Search: Exploring Disproportionality Page 3

4 Acknowledgements This work would not have been possible without the full support of Leicestershire Police and its staff. Particular thanks are due to Nick Glynn, Steve Riley, and Chris Haward who steered the work through its various stages within the Force; to Mr Webb for his clever data work, and to Lynne Woodward for her support in publishing the report. I am also grateful for the time and effort of the many officers who took part in the research, either by interview or by contributing to the legal/appropriate sub study. I was ably supported in different phases of the work by the assistance of Joanna Welford, Jessica Urwin, Hannah Baghum and Dennis Anthony and thank them for their contribution. The responsibility for the analysis and content of the report is, however, my own, and I hope it sheds a little light on the complex question of disproportionality. Jean Hine July 2015 Stop & Search: Exploring Disproportionality Page 4

5 Foreword In 2010 the Equality and Human Rights Commission produced a scrutiny report on the use of Stop and Search powers. Since then the Force has been striving to improve its practices. Leicestershire Police carried out a range of work to identify alternative working practices and areas of improvement around the use and recording of stop search powers. One of these initiatives was to commission Jean Hine from to review the data collected by the force on the use of the power. At the same time officers were receiving further training and the organisation continued to work closely with the Equality & Human Rights Commission, the College of Policing and. The De Montfort research presented in this report used data about stops and searches conducted between October 2011 and June 2013 and identified a number of potential areas of improvement. The findings of the research and its recommendations led to a number of further changes in systems and procedures around the use of stop and search. In June 2014 we also adopted the Best Use of Stop and Search Scheme, a voluntary scheme which is part of a range of measures introduced by the Home Secretary to help improve the way police forces use stop and search. All of these changes have led to a dramatic reduction in the use of stop and search, from around 28,000 a year at the time of the initial report of the Equality & Human Rights Commission, to around 3,000 in the last year. We are not complacent however, and continue to strive to ensure that stop and search is used appropriately and fairly. Our use of the power is under constant scrutiny both internally by the Force and by external organisations and groups. Leicestershire Police recognises the impact that being stopped and searched can have on an individual and the wider community, and to manage this we continue to engage with community groups and partners to ensure that our use of stop and search is appropriate, proportionate and effective. Phil Kay Assistant Chief Constable Leicestershire Police August 2015 Stop & Search: Exploring Disproportionality Page 5

6 Stop & Search: Exploring Disproportionality Page 6

7 Executive Summary Research Aims and Method The Equality and Human Rights Commission review of stop and search (EHRC 2010) 1 identified Leicestershire Police as having some of the highest black/white disproportionality ratios (p5). The Force implemented a range of initiatives to address the issues and commissioned this research to explore why the data showed disproportionality. This study is based on a quantitative analysis of the records of stop searches made by officers. The study involved a comprehensive analysis of the records of 13,310 stop searches from the Leicestershire Police database (suitably anonymised), covering the period 1 October 2011 to 30 June A sub study was undertaken to assess the legality and appropriateness of stop searches whereby 52 police officers, drawn across ranks and functions, peer reviewed the textual description of the grounds for 5,237 stop searches. They had no information except the textual grounds as written by the officer. These records cannot incorporate the nuanced considerations that may justify a stop search in any given situation, but quantitative analysis does reveal interesting patterns worthy of more detailed consideration. Key findings from the quantitative analysis of records are given below. Findings Why were people stop searched? Most people were stop searched under the Misuse of Drugs Act 1971 (53%) or PACE powers (46%). The object of the search was recorded as drugs (54%), stolen goods (25%), going equipped (12%), and offensive weapon (7%). Firearms were the object of the search in less than 1% of cases. Examination of the textual grounds recorded for stop searches revealed that three broad categories of grounds accounted for the bulk of stop searches: o acting on information was the grounds in 32% of cases; o subject behaviour was the grounds in 26% of cases; and o smell of cannabis was the grounds in 26% of cases. 85% of the 9,503 individuals in the analysis had just one stop search recorded; 11% had two stop searches and 4% had three or more. The highest number of stop searches for an individual was 10, occurring for just two people over the 21 month period. Multiple searches were mainly carried out by different officers, sometimes from different LPUs. 1,349 officers carried out the stop searches, 68% conducted between one and ten searches. Less than 5% of officers recorded more than 30 searches over the 21 months of the study. Three officers recorded over 100 searches, few of which were on the same individuals. A higher proportion of their searches were on Asian people than overall, most of who were searched for drugs. Who was stop searched? The vast majority of those stop searched were male (94%). 70% were aged 25 or younger; 25% were juveniles aged 17 or under and 9% were children 14 or younger. Officer perceived ethnicity was 68% White European, 19% Asian, 10% African Caribbean and 2% Dark European, Oriental, Arab or Other. Perceived ethnicity was not present in 2% of the records. Self reported ethnicity had high agreement with officer perceived ethnicity of White European (97%) and perceived Asian (85%). African Caribbean agreement was 57%: 23% of these subjects chose a selfidentification of mixed heritage (4% of stop searches). Most of these (58%) were perceived by the officer as African Caribbean, with 91 (18%) perceived as Asian. This may impact disproportionality calculations. Most of those stop searched (89%) were residents of Leicestershire. Just over a third of the stop searches (35%) related to a vehicle. 1 Equality & Human Rights Commission Stop and think: A critical review of the use of stop and search powers in England and Wales. London, EHRC. Stop & Search: Exploring Disproportionality Page 7

8 Did the recorded grounds show the stop was legal and appropriate? 64% of grounds were assessed as legal, and 23% as not legal, 13% could not make a judgement on the information available 61% were assessed as appropriate, 18% as definitely not appropriate and 21% had can t say responses, suggesting this is more difficult to judge than legality. Sergeants were more likely to judge grounds as legal (79%), than Inspectors (60%)and PCs (67%). Stop searches based on subject behaviour were most likely to be seen as not legal (30%), compared with acting on information (18%) and smell of cannabis (14%). Outcomes and Effectiveness Something was found in 19% of the stop searches overall: in 23% of drug searches, 16% of stolen property searches, 10% of offensive weapon searches and 9% of going equipped. The overall rate of arrest was 10%, remaining very stable across the 21 months of the study despite a substantial reduction in the numbers of stop searches undertaken. A higher arrest rate (17%) followed stop searches where the grounds were based on information or intelligence compared with 10% for suspicious behaviour and 4% for smell of cannabis. Where something was found an arrest followed in 34% of cases. Actions other than arrest followed 15% of the stop searches, with street caution being the most common. Overall, some formal police sanction followed the search in 25% of cases. Stop searches where something was reported as found were more often assessed as legal (77%) than those where nothing was found (65%). The outcome with the highest rating of legality was street caution (84%). 71% of the stop searches which resulted in an arrest were assessed as legal Ethnic difference The perceived African Caribbean group stop searched had a higher proportion of under 18 year olds (30%) than the White Group (25%) or the Asian group (20%). 40% of those self identified as mixed heritage were under 18 years old. The stop search of Asian people was more often linked to a car (50%) than for White people (31%) or African Caribbean (28%). Drugs was the object of the search for 76% of Asian subjects compared with White subjects (48%) and African Caribbean (55%). The African Caribbean group was more likely to be stop searched on the grounds of acting on information (41%) than the White group (34%) or the Asian group (23%), and less likely to be stop searched on the grounds of suspicious behaviour (19%) than White people (27%) or Asians (26%). The stop searches on perceived African Caribbean (74%) and Asian(72%) subjects were more often rated as legal than the stop searches on White European (65%) subjects. There was little difference between the groups in something being found: 21% of Asian, 19% of White,16% of African Caribbean. There were higher arrest rates for African Caribbean and mixed heritage subjects and lower arrest rates for Asian subjects, whether or not something was found. Change over time The range of initiatives introduced by the Force led to a dramatic reduction in the number of stop searches, from 3,152 in the quarter October to December 2011 to 1,140 in the quarter April to June There was remarkable consistency within that reduction with the balance of ethnicities staying much the same resulting in a continuation of disproportionality. There was an increase in under 18s in the perceived African Caribbean group, an increase in the year olds in the Asian group, and an increase in the 26 and over ages in the White group. Stop & Search: Exploring Disproportionality Page 8

9 There has been a steady increase in the proportion of stop searches which were acting upon information from 29% in the first quarter of the study to 38% in the last quarter. This increase is greater for perceived White and Asian subjects than for African Caribbean. Stop searches based on suspicious behaviour reduced over time. The rate at which something was found and arrest remained much the same. Recommendations Recording practice should be addressed to ensure a) that all required information is collected, and b) that the recording of grounds provides sufficient information to confirm the legality and appropriateness of the search. Supervising officers (usually sergeants) should take a more active role in ensuring that officer practice is appropriate and correctly recorded. The higher arrest rate of those perceived to be African Caribbean whether something was found or not is a cause for concern and should be investigated further. The disproportionate use of stop search for young males should be reviewed. Not only does this have a negative impact on young people s views of the police but it may also affect ethnic disproportionality. As a high proportion of the stop searches of African Caribbean people are based on information, the sources of the information, call centre procedures and the process by which this is passed to officers should be reviewed. Ultimately it is the individual officer who takes the decision about which individuals to stop and whether to conduct a search. Officers talked about developing a nose or intuition about when this was appropriate. This instinct develops from experience, and particularly by learning from experienced officers when first working on the streets. Officers should be aware of the danger of unconscious bias impacting upon these decisions. There is no evidence to suggest that individual officers are acting disproportionately. However, interviews with police officers revealed a colour blind approach to their work, when a more nuanced ethnic awareness may be more appropriate. Stop & Search: Exploring Disproportionality Page 9

10 Stop & Search: Exploring Disproportionality Page 10

11 Introduction Stop and search is a topic of much current political interest with opponents arguing that this police tactic is not only ineffective, in that such a low proportion of stop searches result in an arrest, but that the procedure actually harms public perceptions of the police and community relations, particularly with ethnic minority communities who feel disproportionately targeted by the police in this respect. Disproportionality is a major discussion point. Statistics about race and stop and search have been published regularly by the Home Office since One element of these reports is to relate the number of stop searches of different ethnic groups to the population of those groups, with the assumption that equal treatment of different groups would be demonstrated by similar proportions of the populations being subject to stop and search. The figures have consistently shown however that on the basis of these calculations, Black and Asian people are more likely to be stopped and searched than White people, leading to claims of disproportionate treatment of these groups. Some have argued that this is evidence of institutional racism 3. This topic has frequently been addressed and researched by academics, interest groups, and the media, and in 2009 was investigated by the Equality and Human Rights Commission (EHRC) which undertook a major review of stop and search 4. That report focussed on geographical and force differences in rates of stop and search and particularly their impacts on Black and Asian communities. Leicestershire was one of the police forces identified as having some of the highest black/white disproportionality ratios (p5). The Force 5 was subsequently contacted by the EHRC and agreed to take a range of initiatives aimed at reducing both the overall number of stop searches and reducing the disproportionality ratios for Black and Asian people. The Force also decided to commission this piece of research to investigate these issues further. Research Aims and Method The aim of the research is to identify possible drivers of disproportionality, and from this suggest opportunities for changes in policy and practice that could impact upon these rates. The objectives of the research were to explore: factors about the individual and the circumstances of the stop and search event to identify any that may be drivers of race disproportionality. the outcomes of stop and search. the impact of officer and process factors on race disproportionality. police officer and manager experiences and views of stop and search and its role. The main source of data is the database of stop and search records introduced by the Force in October Initial analysis was undertaken on data for the six months October 2011 to March 2012, and following a report on this data the research was extended to cover the period October 2011 to June 2013, enabling a review of change over time. Most of the data was used exactly as recorded in the Force system, except for the most important item, the grounds for the stop, which was collected and recorded in text format just as the officer had written it on the form on which they record details of the stop 6. A rigorous coding exercise was undertaken to identify a set of codes which reflected the range of grounds entered on the records. Full details of this process, together with details of all the data included in the analysis, are provided in the 2 Ministry of Justice (2013) Statistics on Race and the Criminal Justice System Available at and cjs 2012.pdf 3 See Waddington, PAJ, Stenson, K & Don, D (2004). In Proportion: Race, and Police Stop and Search. British Journal of Criminology 44(6) for a discussion of this term. 4 Equality & Human Rights Commission (2010). Stop and think: A critical review of the use of stop and search powers in England and Wales. London, EHRC. 5 The term Force indicates Leicestershire Police throughout this report. 6 Leicestershire Police Stop and Search Record (P1 Form) Stop & Search: Exploring Disproportionality Page 11

12 methodological appendix to this report. The analysis has raised a series of questions about the quality and reliability of the data, which are discussed at appropriate points in the report. A further important issue with the data was created by the changes implemented to protect the identity of people included in the database. One of these was that the date of birth of the subject could not be provided, so the age at the date of file extraction was calculated and provided instead. The age of the subject at the time of the search was estimated by subtracting the time between the stop search and the date of file extraction from the calculated age provided. This statistical analysis was complemented by interviews and focus groups with a sample of police officers and managers in the Force. Again, details of the sample and method are provided in the methodological appendix. Stop & Search: Exploring Disproportionality Page 12

13 Who gets stop searched? Gender and Age The data set for the period 1 October 2011 to 30 June 2013 comprised records about 13,310 stop searches. The vast majority of those stop searched were male, with just 6% 7 being female. The recorded age of the subjects of stop searches ranged from 10 to 75, with over two thirds of them (70%) aged 25 or younger. A quarter of those stopped and searched (25%) were juveniles aged 17 or under at the time of the stop. Of these 71 were children aged 10, 11 or 12 and a further 504 were aged 13 or 14. Overall 9%, or almost 1 in 10 of the recorded searches were on children aged 14 or younger. A quarter (24%) were aged 18 20, 20% were aged 21 25, and 30% were over 26 years old at the time of the stop. Age information was not available for 211 stop searches. Ethnicity Officers were required to record their perception of the ethnicity of the person stop searched according to the categories of White European, Dark European, African Caribbean, Asian, Oriental, Arab or Other. The results for this sample are given in Table 1 below. Table 1: Officer perception of ethnic group Perceived ethnicity Frequency Percent White European % Dark European 143 1% African Caribbean % Asian % Oriental % Arab 8 0.1% Other % Not completed 249 2% Total % The officer perceived ethnicity of the majority of subjects, where this was recorded, was White European (68%), with 19% perceived as Asian, 10% perceived as African Caribbean and 2% recorded as Dark European, Oriental, Arab or Other. Perceived ethnicity was not recorded in 2% of the records. This level of nonrecording of officer perceived ethnicity is surprising, given this is totally within the officer control and thus would be expected to be completed in all cases. The subjects of the searches were asked to self report their ethnicity according to the census categories. Comparison of the officers perceived ethnicity with self reported ethnicity is presented in Table 2 below. It shows high agreement for officer perceived ethnicity of White European, and less agreement for all other categories. 7 All percentages have been rounded to whole numbers except for very small figures of less than 1% Stop & Search: Exploring Disproportionality Page 13

14 Table 2: Comparison of Officer Perception of Ethnic Group & Self Reported Ethnicity Self Reported Ethnicity Asian African Caribbean White European Officer Perceived Ethnicity Dark European Oriental Arab Other Not completed Asian 2162 (85%) 60 ( 5%) 41 (<1%) 15 (10%) 0 1 (13%) 2 ( 3%) 44 (18%) Black 51 ( 2%) 721 (57%) 18 (<1%) 3 ( 2%) (10%) 15 (6%) White 31 ( 1%) 20 ( 2%) 8690 (97%) 69 (48%) 1 (6%) 2 (25%) 4 ( 6%) 61 (24%) Mixed 91 (4%) 301 (23%) 46 (<1%) 23 (16%) 1 (6%) 2 (25%) 27 (38%) 30 (12%) Other 6 ( <1%) 2 (<1%) 17 (<1%) 9 ( 6%) 9 (50%) 3 (38%) 6 (9%) 3 ( 1%) Not stated 208(8%) 166(13%) 190 (2%) 24 (17%) 5 (28%) 0 25 (35%) 96 (39%) Total 2549 (100%) 1270 (100%) 9002 (100%) 143 (100%) 18 (100%) 8 (100%) 71 (100%) 249 (100%) The shaded cells in Table 2 above show the proportion of the officer perceptions that are matched by the self identified ethnicity. Where officers assess the subject to be White European this was the selfidentification in 97% of cases; where officer perception was Asian self identification agreed in 85% of cases, but where officer perception was African Caribbean the agreement on self identification was 57%. Much of this latter lack of agreement was because 23% of these subjects chose a self identification of a mixed heritage. This comparison highlights some limitations of the categories for officer perceived ethnicity. Dark European, Oriental and Arab do not have direct matches in the census categorisation. Very few people were perceived as Oriental or Arab but Dark European was recorded more often. Most of those perceived as Dark European identified themselves as White (48%) with others spread across the categories. There is always a question of the reliability of the data, and both officer perceived and self reported ethnicity can be recorded inaccurately either purposely, or by mistake. However, assuming that all assessment is genuine and not intentionally misleading, the results suggest that officers are good at accurately identifying White ethnicity, but less good at others. As this data is the basis for calculating disproportionality it is important that it is accurate. Table 2 above shows that in the cases with no officer perceived ethnicity, self reported ethnicity was provided in most instances, meaning just 96 cases had no information at all about ethnicity. Of the missing officer perceived ethnicity 24% self identified their ethnicity as White, 18% as Asian, 6% as Black and 12% as a mixed category. Self reported ethnicity was not recorded in 5% of the records. Looking at the three main officer perceived categories, the group least likely to give a self identified ethnicity was those perceived by the officer as African Caribbean (13%), followed by Asian (8%), with those perceived as White rarely (2%) not having a self identification. This raises the question of why this information is less likely to be available for these minority ethnic groups. The overall impact of these missing data, both officer perceived and selfidentified, is small, but there is a greater impact on Asian and Black/African Caribbean 8 groups than White European. The officer identified categories do not include options for mixed race, but as Table 2 above illustrates, this was the self reported choice for 521 people (4% of those stop searched). Most of these people (58%) were perceived by the officer as African Caribbean, with 91 (18%) perceived as Asian. Detailed examination of the self identification for these cases shows that the officer perceived category for these individuals was often part of the self identified mix: 76% of the self reported White Black Caribbean group were perceived by the officer as African Caribbean, 56% of the self reported White Black African Group were perceived as African Caribbean, and 53% of the self reported White Asian group were perceived as Asian. Although the actual 8 The terms Black and African Caribbean are used interchangeably. Stop & Search: Exploring Disproportionality Page 14

15 numbers involved relatively small, they do, again, raise questions about their potential impact on disproportionality calculations. Residence To protect the identity of people stop searched, full address details were not included in the data set, but the first four characters of the postcode were provided. About a quarter of the people who were stop searched during the period of study did not have postcode information recorded, but for the others it has been possible to assess the region where they live. Where information has been provided, 89% of the people who were stop searched were residents of Leicestershire. Where people from other areas had been stop searched by Leicestershire officers, the most often cited postcodes are Derby, Nottingham, Coventry and Birmingham, but stop searches have been carried out on people who had given postcodes from as far away as Edinburgh, Brighton and Llandudno. Vehicle presence A vehicle registration mark (VRM) was present in 35% of the stop searches meaning the search was linked to a motor vehicle. Where the VRM was present 58% of those searched were recorded as the driver of the vehicle and 40% were passengers. The remaining 2% were recorded as unattended. There were also 30 passengers and 5 drivers recorded where there was no VRM present, another small inconsistency in the data. The overall proportions in each category are presented in Table 3 below. Table 3: Vehicle involvement Vehicle Driver or Passenger Frequency Percent Driver % Passenger % Unattended 112 1% Neither % Total % Stop & Search: Exploring Disproportionality Page 15

16 Why are people stop searched? Search Power and Object of the Search The power to stop and search is governed by three main Acts: the Firearms Act 1968, the Misuse of Drugs Act 1971, and the Police & Criminal Evidence Act Additional powers are granted by the Road Traffic Act 1988, the Criminal Justice and Public Order Act 1994, and the Terrorism Act Officers are required to record the powers under which they have stop searched each individual. Table 4 shows that the power used most often was Drugs, accounting for 53% of all the stop searches. This was closely followed by PACE powers which accounted for 46% of the stop searches. Other powers were rarely used. Table 4: Search powers used Search Power Frequency Percent Drugs % PACE % Prevention violence 130 1% Firearms 83.6% Terrorism 12.1% Other 17.1% Total % Unsurprisingly then, the object of most searches was recorded as drugs (54% of cases). The next most often cited object of the search was stolen goods, in 25% of cases, followed by going equipped in 12% of cases, and offensive weapon in 7% of cases. Firearms were the object of the search in less than 1% of cases. Powers used and the object of the search match for most powers, but the stop searches under PACE had a wider range of objects recorded, as demonstrated in Table 5. This variation does raise the question as to whether some of the powers are incorrectly recorded, eg searching for drugs under PACE. Table 5: Object of searches under PACE powers Grounds Stolen Offensive Going Drugs Firearms Other Total property Weapon equipped 55% 4% 0.1% 13% 26% 2% 100% The grounds for searches are recorded in the database as textual information exactly as written by the officer. These were categorised in detail as described in the methodological appendix. There were a small number of cases (235) where the grounds mentioned did not fit any of the coded categories. The number of times each category was mentioned is presented in Table 6 below. The total number of grounds mentioned is much more than the number of cases because 57% of the stop searches mentioned more than one of the categorised grounds, 36% mentioned two, 16% mentioned three and 5% had more than three. Stop & Search: Exploring Disproportionality Page 16

17 Table 6: Frequency of Items mentioned in Grounds Grounds category Number % of cases Subject Behaviour % Smell of Cannabis % Acting on Information % High or Recent Crime % Location of Incident % Ran or Drove Off 853 6% Known Offender 848 6% With Known Offender 258 2% Perceived Threat 231 2% Secluded or Remote Place 230 2% Time of Incident 213 2% The most often mentioned grounds, with very similar numbers of reports, are Subject Behaviour, Smell of Cannabis and Acting on Information, each being mentioned in just over a third of cases. Subject Behaviour contains a wide range of conduct by the person stop searched and the initial categorisation (see Appendix) proved to account for only around a half of the activities which were classified as suspicious. The subcategories for Acting on Information were a better match to officer descriptions, and within this category the most frequently cited ground was responding to a call, which occurred in just over a half of the cases in that category: from a police briefing or operation (17%), from the public (17%), from a shop, hospital or security (7%), or from somewhere unspecified (10%). The second most frequently cited sub group of Acting on Information was that the subject fitted a description, which accounted for 32% of those in this category and might also be understood as responding to a call. To simplify further analysis, each case was classified according to one category of grounds, with priority being given in the order of Table 7 below, which also indicates the number of cases which fell into each of those categories. Table 7: Main grounds category Grounds Number % Acting on Information % Subject Behaviour % Smell of Cannabis % Crime in area 831 6% Secluded area % Known Offender 155 1% Location 273 2% Time of Incident % Ran/drove off 185 1% With Known Offender 318 2% None of these grounds mentioned 235 2% Total % This table shows that when considering the main grounds for the stop search, three categories account for 85% of all stop searches: Acting on Information, Subject Behaviour, and Smell of Cannabis. Stop & Search: Exploring Disproportionality Page 17

18 Grounds Legal and Appropriate? An important issue in relation to stop searches is that they are conducted lawfully and appropriately. A Policing Foundation Briefing 9 explains that there must be a legal basis for any search and that an officer must have reasonable suspicion based on facts, information and/or current intelligence, which are relevant to the likelihood that a person possesses a prohibited article, is about to commit, or has committed a crime. (p3) The officer s recording of the grounds for the stop search is the only way to assess this. This assessment is conducted on a day to day basis by the officer s supervisor who checks all stop search records and approves them or has a discussion with the officer about their recording and stop search practice. The decision is recorded in the system and shows that 98% of the stop search records were accepted, and 2% rejected. As the legitimacy of the search is a critical consideration an additional exercise was introduced to address this. All the grounds from the period October 2011 to March 2012 were distributed to a range of grades of police officers in batches of 100 and they were asked to assess whether the grounds were a) legal and b) appropriate. These were distributed to 52 officers and returned by 47, resulting in a final sample of assessed cases of 4,700. Forty five per cent of these of the records were assessed by Inspectors, 26% by PCs, 15% by Senior Managers, 8% by Sergeants, and 6% by non police who were members of a Stop Search Reference Group established by the Force. To assess the level of agreement between individuals each set of records included a small sample of six cases that were the same across several sets. In total 36 records were assessed by between five and nine officers of a range of grades. For the purposes of the exercise the only information which was given to the assessors was the textual description of the grounds provided on the record. They had no information about the individual stopped, the officer, local policing unit (LPU), location of the stop, or the outcome. The results of this exercise demonstrated that people often found it difficult to make a judgement based solely on what was written on the record. Of the 36 records in the inter rater sub sample, 12 had 100% agreement that the search was legal and two had 100% agreement that the search was not legal. There were differences of opinion in the other 22 cases (61%). Appropriateness seems to be harder to assess, with two cases which had 100% agreement about legality having less agreement about appropriateness. For the full 4,700 cases that were assessed, 64% were assessed as legal, and 23% as not legal: 8% were assessed as Can t say and 5% were not assessed. Giving the not assessed cases the benefit of the doubt, the minimum proportion of cases assessed as not legal on the basis of what is recorded on the form was 23%, and is likely to have been higher. In terms of appropriateness the results show that there are some cases where the stop search was assessed as legal, but its appropriateness was questionable, as overall 61% were assessed as appropriate. This issue has a greater proportion of can t say responses or not assessed, at 21%, suggesting again that this is more difficult to judge from the record than legality. 18% were assessed as definitely not appropriate. The criteria for the decision are different, but these figures are not radically different from those reported from the HMIC Review of Stop & Search Powers which assessed that 27% of the records which they reviewed did not have reasonable grounds for suspicion recorded 10. There was some difference between officer grades in their assessments, as can be seen in Table 8 below. 9 Police Foundation (2012). Stop and Search. The Briefing Series 2, Edition HMIC (2013). Stop and Search Powers: Are the police using them effectively and fairly? London: HMIC. p30 Stop & Search: Exploring Disproportionality Page 18

19 Table 8: Officer Grade by Assessment of Legality of the Stop on the Basis of the Recorded Grounds Assessed as Legal Inspector Nonpolice PC Senior Sergeant Total N Can t Say/Not assessed 12% 10% 20% 5% 9% 13% No, not legal 28% 35% 13% 29% 12% 23% Yes, legal 60% 55% 67% 66% 79% 64% This table shows that PCs are the least likely to judge the grounds as not legal (13%), and the non police assessors being the most likely to make this judgement (35%). What cannot be judged from this work is whether the groups have different criteria for making such judgements, or whether affiliation with those generating the records makes them more likely to give the benefit of doubt. Sergeants are much more likely to assess the recorded grounds as legal than any other group, which again raises the question of whether this is different judgement criteria. It is important to know because Sergeants are the grade which generally authorises the recorded stop search. There is a similar and more marked pattern in relation to assessment of appropriateness, as shown in Table 9 below. Table 9: Officer Grade by Assessment of Appropriateness of the Stop on the Basis of the Recorded Grounds Assessed as Appropriate Inspector Nonpolice PC Senior Sergeant Total N Can t Say/Not assessed 18% 17% 34% 18% 10% 21% No, not appropriate 23% 32% 7% 20% 11% 18% Yes, appropriate 60% 52% 60% 62% 79% 61% All groups of assessors, except Sergeants, have a lower proportion of cases that are deemed to be appropriate than were thought to be legal, with PCs being very unlikely to assess the grounds as not appropriate and more likely not to make a judgement. Sergeants assessed exactly the same number and proportion as legal and appropriate and shifted very slightly on the balance between not appropriate and can t say. The data about judgements were added to the full record about stop searches and explored to see whether there were any links between these data and the legality and appropriateness of the stop search. This process inevitably reduced the number of cases where all relevant information was available, but on this basis the stop searches on African Caribbean and Asian subjects were more often rated as legal and as appropriate than the stop searches on White European subjects, as Tables 10 and 11 show. Table 10: Officer Perceived Ethnicity of Subject by Assessment of Appropriateness of the Stop on the Basis of the Recorded Grounds Assessed as Legal African Caribbean Asian White European N Can't Say 7% 6% 9% No, not legal 19% 22% 26% Yes, legal 74% 72% 65% Stop & Search: Exploring Disproportionality Page 19

20 Table 11: Officer Perceived Ethnicity of Subject by Assessment of Appropriateness of the Stop on the Basis of the Recorded Grounds Assessed as Appropriate African Asian White Caribbean European N Can't Say 7% 7% 10% No, not appropriate 18% 19% 21% Yes, appropriate 74% 74% 69% What cannot be said here is whether this is a reflection of actual practice on the ground or of recording practice. It may indicate that officers are aware of the scrutiny of stop searches conducted on African Caribbean and Asian subjects and are more thorough in documenting the grounds for their decision. Examination of the category of grounds recorded reveals a difference in both legality and appropriateness between the three main categories of grounds. Where the main grounds was categorised as the officer acting on the basis of Subject Behaviour 30% of the cases were assessed as not legal. For the category of Acting on Information 18% were assessed as not legal, and where the category was Smell of Cannabis 14% were assessed as not legal. The figures for appropriateness are much the same, demonstrating that the recording of grounds for stop searches based on Subject Behaviour are the least likely to be seen as legal or appropriate. Stop searches where something was reported as found have a higher assessment of legality at 77% compared with 65% where this is not reported, though what is perhaps most surprising here is the number of cases (127) which were assessed as not legal and yet something was reported as being found, which is 12% of the stop searches assessed as not legal. The figures for appropriateness are very similar. In terms of overall outcome, there are also differences. The outcome with the highest rating of legality was street caution, where the stop search was assessed as legal in 84% of the 300 cases. Of the 435 stop searches which resulted in an arrest, 71% were assessed as legal, so here again there was a substantial number of arrests where the grounds recorded do not satisfy the assessor as to their legality. Subjects Hanging Out In Groups The grounds mention that the person stop searched is with one or more other people in 30% of cases. Mention in the grounds may not be an accurate reflection of the numbers and proportion of individuals who are stop searched with other people as some officers may not have mentioned this even though the person was not alone, but it does enable some analysis of the question. Being with others is rarely the only grounds given for the stop, and the extent to which it is mentioned does vary according to the main category of grounds. Where the main category is Acting on Information 26% of the stop searches mentioned the subject was with others; where the main category was suspicious behaviour being with others was mentioned in 33% of the stop searches, and where the main category was Smell of Cannabis being with others was mentioned in 31% of the stop searches. Day and Time of stop As would be expected, the number of stop searches undertaken varies by day of the week and time of day, as shown in the charts below. Stop & Search: Exploring Disproportionality Page 20

21 Chart 1: Number of Stop Searches by Day of the Week Monday is the day of the week with least stop searches, gradually increasing to Friday and Saturday being the days when stop searches are most likely to happen. Chart 2: Number of Stop Searches by Time of Day Relatively few stop searches are recorded between 4am and 10am in the morning, with most taking place at night between 8pm and 1am in the morning. Multiple Stops Identifying information about the individuals stop searched was provided by means of an ID code created by Leicestershire police. This was missing for 210 cases, perhaps because they had not provided a date of birth. The remaining records referred to 9,593 individuals, 85% of whom had just one stop search recorded during the period of study, and 15% of whom had more than one stop search. The largest number of stop searches for an individual was ten, received by two people. The number of individuals receiving each number of multiple stop searches is given in table 12 below. Stop & Search: Exploring Disproportionality Page 21

22 Table 12: Number of Stop Searches for each Person No. of stop searches No. People % of people % % % % % % % % % Total % Those with the most stop searches had their first stop search early in the study, giving them more time to accrue further stop searches over time. A review of the cases with the most stop searches showed that they were generally carried out by different officers, sometimes in different LPUs. Local Policing Unit There is wide variation between Leicestershire Local Policing Units (LPUs) in terms of the numbers of searches which are carried out by their officers, as demonstrated in Table 13 below. Table 13: Number of Stop Searches for each LPU LPU Number % City New Parks (CA) 437 3% Beaumont Leys (CB) 830 6% Hinckley Road (CH) 580 4% Keyham Lane (CK) 857 6% Mansfield House (CM) 975 7% Spinney Hill (CN) % Euston Street (CU) % Welford Road (CW) 409 3% County Market Harborough (LA) 418 3% Blaby (LB) 718 5% Charnwood (LC) 726 6% Hinckley (LH) 762 6% Melton (LM) 636 5% NW Leics (LN) 837 6% Loughborough (LO) 685 5% Rutland (LR) 269 2% Wigston (LW) 453 3% Central Central Units (inc Belgrave) 650 5% Total % The Units which conduct the largest numbers of stop searches are all in Leicester City, with Euston Street and Spinney Hill having the largest numbers. Officers The number of stop searches undertaken by each officer varies considerably, from the lowest number of 1 to a maximum of 209 during the period October 2011 to June Figures for all the officers who undertook Stop & Search: Exploring Disproportionality Page 22

23 stop searches during the period are given in Table 14 below. It is not known from this data how many officers did not record any stop searches during this period. Table 14: Number of Stop Searches for each Officer No. of stop searches No of officers % % % % % % % % Over % Total % More than two thirds of the officers in the county (68%) had undertaken between 1 and 10 stop searches over the course of the 21 months of the study. A further 21% undertook between 11 and 20 stop searches, with relatively few (14%) having more than 20 stop searches recorded. There were 18 officers who recorded between 51 and 100 stop searches during this time and three officers who recorded more than 100 stop searches. In total these latter three officers accounted for 480 of all the stop searches recorded during this period, which is almost 4% of the total. The individual stop search pattern of the officers undertaking low numbers of stop searches, which is the overwhelming majority of officers, will not impact on the overall figures for the county, but individual officers who undertake high numbers of searches can have an impact and so their patterns have been explored in more detail. All three officers were based at the same LPU, where their combined stop searches accounted for 39% of all the stop searches recorded by that LPU during the 21 months of the study period. Their numbers of stop searches have reduced over the period of the study but not quite at the same rate as their colleagues: There have been fluctuations over the period, but in the first 6 months of the study these three officers recorded 33% of the stop searches for the LPU, and in the last six months that figure was 38%. There are differences between these individual officers in their pattern of stop searches, but there are also important similarities, perhaps the most significant for this study being the perceived ethnicity of the people they have stop searched. As Table 15, below, shows, all three of these officers have a much higher proportion of Asian subjects of stop searches than either the rest of their LPU or the county overall, with one officer having a particularly high proportion of Asian subjects at 79%. There may be a valid explanation for this, but it is a potential cause for concern and worthy of further investigation. Table 15: Perceived Ethnicity of Subject by High Stop Search Officers, compared with LPU and County Perceived Officer 1 Officer 2 Officer 3 Rest of Rest of Total Ethnicity LPU county N White 31% 30% 16% 36% 73% 69% Black 16% 7% 4% 16% 9% 10% Asian 52% 60% 79% 45% 16% 20% Other 1% 4% 0% 3% 2% 2% All stop search records are checked by a supervising officer, but it is unlikely that patterns such as this for individual officers would be discernible on a day to day basis. Perhaps a regular check for officers with high Stop & Search: Exploring Disproportionality Page 23

24 numbers of stop searches would identify individuals who may have a disproportionate impact on county patterns. These officers also display different patterns on other aspects of their stop searches. All three have a higher proportion that are vehicle related, again particularly Officer 3 who recorded this for 72% of the stop searches compared with 34% and 34% of the LPU and county respectively. The object of the search was drugs for a higher proportion of these officers stop searches than overall, especially for Officer 3 who recorded this as the object in 98% of the stop searches compared with 62% and 53% of the rest of the LPU and rest of the county respectively. There are also differences in the main grounds for the stop search, as shown in Table 16 below. Table 16: Main Grounds for Stop Search by High Stop Search Officers, compared with LPU and County Main grounds Officer 1 Officer 2 Officer 3 Rest of LPU Rest of county Total N Acting on info 14% 16% 5% 29% 34% 32% Behaviour 49% 43% 31% 29% 26% 26% Smell cannabis 22% 19% 61% 29% 25% 26% Other 15% 21% 1% 12% 14% 14% None of these mentioned 1% 1% 2% 1% 2% 2% The Smell of Cannabis is the main grounds in a very high proportion of Officer 3 s stop searches (61%) and Acting on Information is the main grounds in a low proportion of this officer s cases (5%). Acting on Information is also relatively low for the other two high stop searches officers, but not so radical at 14% and 16% compared with 29% for the rest of the LPU and 34% for the rest of the county. For these two officers it is the behaviour of the subject which is an increased grounds for the search, at 49% and 43% compared with around a quarter for the rest of the LPU and the rest of the county. Although these officers may be focussing more on certain groups or behaviours for their stop searches, they do not appear to be targeting particular individuals. The vast majority of the subjects of their stop searches (80% or more) are recorded just once, and the most that any individual is recorded is five times during the period of the study. These figures are not very different to the LPU and county figures. There are also differences in the outcomes of the searches of these officers, which will be discussed below. Stop & Search: Exploring Disproportionality Page 24

25 Outcomes of Stop Searches Property was recorded as found in 19% of the stop searches. The highest rates of property being found were for firearms and drugs, where something was found in 24% and 23% of cases. The lowest rate for something being found was where the object of the search was an offensive weapon, where property was found in 10% of cases. Where property was found the object of the search was drugs in two thirds (67%) of the stop searches. Stolen property was the object in 22% of the cases where property was found, and going equipped was the object in 6%. Firearms and offensive weapons were the object in the remaining 5% of cases where something was found. Where something was found, 34% of the people were arrested. Table 17 below presents the recorded outcome of the stop searches. An arrest was the outcome in 10% of cases overall, followed by Street Cautions in 7% of cases. Some other outcome was recorded in 8% of cases, and there was no formal outcome recorded in 75% of the stop searches. Table 17: Recorded Outcome of Stop Searches Outcome of stop Number % Nothing recorded % Arrested % Fixed Penalty Notice 191 1% Restorative Justice % Street Caution 890 7% Sec % Other disposal 834 6% Total % Unsurprisingly arrest was most likely to occur in cases where something was found. Almost two thirds (63%) of those arrested had something found recorded. The grounds for the stop search is related to the outcome of the search, as can be seen in Table 18 below. Table 18: Recorded Outcome of Stop Searches by Grounds Outcome of Stop Search Acting on Information Behaviour Smell cannabis Stop & Search: Exploring Disproportionality Page 25 Other None of these mentioned N Arrested 17% 10% 4.% 8% 12% 10% Fixed Penalty Notice 1% 1% 3% 1% 2% 1% Street Caution 2% 7% 14% 2% 10% 7% Other disposal 8% 7% 7% 6% 12% 7% No outcome recorded 72% 76% 72% 84% 65% 75% The highest arrest rate of 17% relates to stop searches that have been conducted in relation to some information or intelligence, what might be called intelligence led stop searches. The lowest arrest rate of 4% is for stop searches based on the Smell of Cannabis. The arrests based on the behaviour of the subject which arouses the suspicion of the officer falls in between these at 10%. The outcomes for the three officers with very high rates of stop searcher officers, particularly Officer 3, show a different picture to those overall. They record something being found in a higher proportion of their stop Total

26 searches than the rest of their LPU or the county. Officer 3 had something found in 38% of the stop searches, with Officer 1 recording something found in 22% of stop searches, and Officer 2 in 23% of cases. These figures compare with 17% for the rest of the LPU and 19% for the rest of the county. This did not however relate to a higher arrest rate, especially Officer 3, as can be seen from Table 19 below. Table 19: Outcome of Stop Search by High Stop Search Officers, compared with LPU and County Outcome of Stop Search Officer 1 Officer 2 Officer 3 Rest of Rest of Total LPU county N No outcome recorded 80% 77% 62% 78% 75% 75% Arrested 9% 10% 1% 10% 10% 10% Fixed Penalty Notice 2% 0.0% 11% 1% 1% 1% Street Caution 5% 11% 23% 7% 7% 7% Other disposal 4% 3% 4% 4% 7% 7% Officers 1 and 2 had slightly lower arrest rates, but for Officer 3 the arrest rate is extremely low at less than 1% of the stop searches. The proportions of cases with fixed penalty notices and street cautions are particularly high for Officer 3. These figures do suggest that more stop searches do not result in more arrest, and actually suggest that the opposite is the case, that fewer stop searches will result in a higher arrest rate, particularly if these are linked to Acting on Information. Stop & Search: Exploring Disproportionality Page 26

27 Ethnicity and difference A key focus of this analysis has been to explore differences between the main ethnic groups to see whether there are any features which may help explain the statistical disproportionality in stop search rates. The analysis has been undertaken using officer perceived ethnicity for the three groups of White, African Caribbean, and Asian. Other perceived ethnicities have not been included in the analysis because the numbers are small and they are not included in national disproportionality calculations. The total sample for this analysis is 12,821 consisting of 9,002 stop searches on subjects perceived to be White, 1,270 on subjects perceived to be African Caribbean, and 2,549 subjects perceived to be Asian. Subject characteristics There are some slight differences in subject characteristics between the ethnic groups. In terms of gender, there is a higher proportion of females amongst perceived White subjects (7%) than amongst perceived African Caribbean (5%) and Asian subjects (3%). There is a bigger difference in terms of age, with perceived African Caribbean subjects more likely to be younger and perceived White subjects more often older, as shown in the Chart below. Chart 3: Age group of subject of stop search compared by officer perceived ethnicity Almost 30% of the African Caribbean subjects were aged under 18 at the time of the stop search, compared with 25% of White subjects and less than 20% of Asian subjects. The differences in age become starker when looking at those who self identify their ethnicity as mixed. In this group 40% are officially children aged 17 or under. This proportion has fluctuated over time, between 28% and 52% in any one quarter but is generally consistent. For the last quarter in this study, March June 2013, the figure was 43% aged 17 and under. Stop searches of subjects who are perceived to be Asian are much more likely to be linked to a vehicle than those perceived to be either White or African Caribbean. A half (50%) of the Asian group had a VRM present, compared with 31% of the White Group and 28% of the African Caribbean group. The perceived African Caribbean and Asian groups were also more likely to have a Leicester postcode (69% and 72% respectively) and thus be resident in the area than the White group (64%), but the difference is linked more to not having a postcode recorded than to living outside the county. Search features The most significant differences between the perceived ethnic groups relate to the grounds for the stop search, as can be seen in the Chart below. The perceived African Caribbean group had the highest proportion of stop searches where the main ground was acting on some sort of information and the Stop & Search: Exploring Disproportionality Page 27

28 perceived Asian group had the highest proportion where the main grounds was Smell of Cannabis. The White group had a more even spread across the main grounds. Chart 4: Main Grounds for Stop and Search by Officer Perceived Ethnic Group The grounds frequently contains more than one reason for the stop search, and the prioritising of these for the main grounds masks the extent of some differences in the use of specific grounds. For instance, the Smell of Cannabis is often linked to suspicious behaviour or Acting on Information. When looking specifically at the Smell of Cannabis this is mentioned in 53% of the stop searches on perceived Asian subjects, just over half of their stop searches. This compares with 29% of perceived White subjects and 36% of perceived African Caribbean subjects, increasing the differences in grounds between the ethnic groups. These figures suggest that the Smell of Cannabis was combined with Acting on Information or suspicious behaviour in 15% of the stop searches on Asian subjects. The object of the search also varies between ethnic groups, with drugs being the object of the search for 76% of the perceived Asian group. Drugs were the object for smaller proportions of the perceived White group (48%) and the perceived African Caribbean group (55%). For the White group stolen property was the object of the search for 30% of subjects, compared with 22% of the African Caribbean group and 11% of the Asian Group. Although numbers are relatively small, a weapon was the object of the search for 14% of the African Caribbean group compared with 6% of the White group and 7% of the Asian group. The proportion of stop searches where there is mention in the grounds of being with others is much the same across the ethnic groups, being 30% of the perceived White group, 31% of the perceived African Caribbean group, and 31% of the perceived Asian group. It does however vary substantially by age, with this being mentioned more often for people in the younger age groups. Being with one or more others was mentioned in 43% of the stop searches recorded on those aged 17 and under, in 34% of the stop searches of year olds, in 26% of the stop searches of year olds, and in 21% of the cases of those aged over 25. There is wide variation between local policing units in the proportion of stop searches which are undertaken on the different ethnic groups, reflecting the population of the areas which they cover. The LPUs with the highest proportion of stop searches on perceived African Caribbean subjects are Mansfield House (19%), followed by Euston Street (16%), Beaumont Leys (15%), Keyham Lane (14%) and Spinney Hill (14%). The LPUs with the highest proportion of stop searches on perceived Asian subjects are Spinney Hill (52%), and Keyham Lane (42%), then Euston Street (29%) and Mansfield House (23%). These figures show that stop searches on ethnic minority groups are very localised in areas that have high ethnic minority populations. It has been suggested that disproportionality ratios would be different if calculated on a local area basis, and given the detailed data available from the recent census this would be an interesting exercise. Stop & Search: Exploring Disproportionality Page 28

29 Outcome of the Search There are some differences in outcome between the perceived ethnic groups, but not as much as might be anticipated given the differences in grounds. Something was found in 19% of the stop searches on perceived White subjects, 16% of the perceived African Caribbean subjects and 21% of the perceived Asian group. These figures probably relate to the object of the search, as we have seen that where the object of the search is drugs something is more likely to be found. In terms of action following a search there was no difference in whether any following action was recorded, with some action following in 25% of White stop searches, 26% of African Caribbean stop searches and 26% of Asians stop searches. There were differences though in the type of action, as Table 13 below shows. Table 20: Recorded Outcome of Stop Searches Outcome of stop White N 6713 African Caribbean Asian Nothing recorded 75% 74% 74% Arrested 11% 13% 7% Fixed Penalty Notice 1% 1% 2% Restorative Justice 0.3% 1% 0.2% Street Caution 6% 4% 10% Sec27 0.4% 0.2% 0.4% Other disposal 6% 7% 8% A higher proportion of perceived African Caribbean subjects were arrested (13%) than either the White subjects (11%) or the Asian subjects (7%), which is perhaps surprising given that something was found in less of the stop searches for this group than the others. This may be related to the object of the search, so analysis was undertaken to look at the arrest rates where something was found and nothing found for different objects of the search for each of the ethnic groups. The results are presented in Table 21 below. Table 21: Proportion of arrests where something was found or not by object of search and ethnic group Object of the search Perceived Ethnic Group % of cases where something was found % of cases arrested where something was NOT found % of cases arrested where something WAS found Number of Arrests where something was NOT found/was found Stolen Property White 16% 7% 58% 159/251 Black 17% 11% 64% 26/30 Asian 13% 6% 56% 15/20 Drugs White 24% 1.7% 24% 58/245 Black 18% 3.5% 38% 20/48 Asian 25% 1.6% 17% 23/83 Going Equipped White 11% 5% 58% 55/76 Black 9% 4% 44% 4/4 Asian 11% 7% 50% 7/5 Offensive Weapon White 11% 13% 64% 59/37 Black 11% 10% 53% 16/10 Asian 6% 10% 82% 9/9 Although the numbers are small, particularly for the African Caribbean and Asian groups, and should be treated with caution, Table 21 shows that perceived African Caribbean subjects have higher arrest rates than Stop & Search: Exploring Disproportionality Page 29

30 Asian and White subjects when controlling for the object of the search AND whether anything was found. The largest numbers of arrests relate to Drugs and Stolen Property. The figures for Stolen Property show that something was found in 17% of the stop searches on perceived African Caribbean subjects and 16% of perceived White subjects much the same proportion. However the arrest rate for the African Caribbean Group is higher both where something was found and where nothing was found. The same is true where the object of the search was Drugs, and here the proportion of cases where something was found is lower for the African Caribbean group than for the Asian and White Groups. The same pattern is not apparent where the Object of the search was Going Equipped or Offensive Weapons, but here the numbers are very small and less reliable for determining patterns. This is an issue worthy of further investigation to understand why the arrest rates for the African Caribbean group are higher. It may be that the circumstances of the theft of the stolen property are different, or that different classes of drug were found, but there is no information on the stop search records to address this question. Mixed Ethnicity The standard classification for the assessment of race disproportionality uses the three broad categories of Asian, Black and White, based on the perceived ethnicity of the office making the stop search. As highlighted previously, a significant number (518) of the stop searches carried out in Leicestershire involved subjects whose self ethnicity is a mixed category. Further analysis of this group reveals some potentially important issues, the first of which is that this group is generally younger than other ethnic groups, including the perceived African Caribbean group. Overall, 14% of those who were stop searched during this study were aged 16 and under at the time of the search, but for the mixed ethnicity group this figure was 25%. Subjects aged 17 and 18 years made up 17% of the sample overall, but 21% of the mixed ethnicity group. Combined, these figures mean that 31% of the total sample was aged 18 or under when searched whereas 46% of those of mixed ethnicity were in these young age groups. In terms of the object of the search this is more likely to be an offensive weapon for this group (12% compared with 7% overall), but in terms of grounds categories there is little difference between this group and the overall figure. They are also more likely to have multiple stop searches than the overall figure, with 17% having more than one stop search recorded compared with 13% overall. When it comes to outcome, those of mixed ethnicity are less likely to have something found than the overall figure (14% compared with 19%), but are more likely to be arrested (14% compared with 10% overall). So we have the situation where this group was generally younger, was more subject to multiple searches that were less likely to reveal something illegal, and yet were more likely to be arrested. The data in this study cannot explain such a conundrum, but it is an issue worthy of further investigation. Stop & Search: Exploring Disproportionality Page 30

31 Change over time Since the intercession of the Equality and Human Rights Commission and the introduction of further training and guidance for officers in Leicestershire, the numbers of stop searches recorded have fallen substantially, from 1,201 stop searches in the month of October 2011 to 325 searches in the month of June 2013, as shown in the Chart below. Chart 5: Number of Stop Searches by Month This section explores the reduction in numbers. As the numbers in each month have become relatively small trends over time have been analysed on a quarterly basis, the total numbers for which can be seen in the chart below. Chart 6: Number of Stop Searches by Quarter Overall, the main finding from a review of features over time is consistency. Although there has been a substantial reduction in the number of stop searches the underlying features of those stops have not changed. Subject characteristics The proportion of female subjects stop searched fluctuated slightly between 5% and 7% but with no discernible pattern. There was greater fluctuation in the age groups of subjects, as shown in the Chart below, but no definitive trend. The biggest movement is in the proportion of under 18 year olds, which rose from 23% to a peak of 32% in July September 2012, but then fell to its lowest point in the following quarter, since when it settled back to its earlier levels. Given the timing of this peak it could have been related to school holidays. Stop & Search: Exploring Disproportionality Page 31