The Research, Development and Statistics Directorate exists to improve policy making, decision taking and practice in support of the Home Office purpose and aims, to provide the public and Parliament with inform a t i o n necessary for informed debate and to publish information for future use. Findings are produced by the Research, Development and Statistics Directorate. For further copies visit: http://www.homeoffice.gov.uk/ rds/pubintro1.html Young people, crime and antisocial behaviour: findings from the 2003 Crime and Justice Surv e y 2nd Edition Ruth Hayward and Clare Sharp The 2003 Crime and Justice Survey (C&JS) is a nationally representative self-report offending survey of around 12,000 people in England and Wales. The main purpose of the survey is to provide a measure of criminality in the general population (Budd et al., 2005). In addition the survey also asked young people aged 10 to 25 (n=4,574) about their involvement in antisocial behaviour. Research elsewhere shows a link between low level delinquency and more serious offending. Moreover, antisocial behaviour can itself be a concern, particularly if people view it as a sign of more serious crime risks, and it results in social and economic costs. Evidence suggests that such behaviour is concentrated among young people. Key points 29% of young people said they had committed at least one act of antisocial behaviour in the previous year. The most common antisocial behaviour was causing a public disturbance (15%), followed by causing neighbour complaints (13%). More serious incidents such as joyriding and carrying a weapon were much less common. Of those responsible for antisocial behaviour, the majority (68%) only committed one type of behaviour. Only 9% (2.4% of the sample) committed three or more different types. Males reported higher levels of antisocial behaviour across all types of behaviour. A third of males admitted at least one behaviour, compared with a fifth of females. 14- to 16-year-olds were more likely to commit antisocial behaviour than other age groups. Two-fifths of them reported at least one act of antisocial behaviour in the last 12 months. The following factors were strongly associated with antisocial behaviour: disruptive school environment; delinquent peers; drug use; risky alcohol use; negative relationship with parents; delinquent personality traits; living in a household in financial difficulties; living in a high disorder area; and being a victim of crime. About 17% of young people had committed antisocial behaviour but no more serious offence. 12% of young people had committed both antisocial behaviour and offences. 9% had committed offences, but no antisocial behaviour. Crown copyright 2005 ISSN 1473-8406 Printed by: TABS In 2003, the Government published an action plan outlining priority areas to reduce antisocial behaviour over the next three years. This is backed by new powers in the Anti-Social Behaviour Act 2003. Defining antisocial behaviour is challenging. People have different expectations and levels of tolerance. What one person may find offensive or distre s s i n g, another person might view as innocuous. The views expressed in these findings are those of the authors, not necessarily those of the Home Office (nor do they reflect Government policy)
Box 1 Types of behaviour asked about Noisy or rude in a public place Joyriding Behaviour leading to a neighbour complaining Carrying some form of weapon Graffiti Racial harassment (attacked/threatened/rude) The Crime and Disorder Act 1998 defined antisocial behaviour as acting in a: manner that caused or was likely to cause harassment, alarm or distress to one or more persons not of the same household. The C&JS asked about various types of behaviour that could be considered antisocial (Box 1). Some of these racial harassment and carrying a weapon fall outside the focus of the Home Office Action Plan on antisocial behaviour but are the subject of other govern m e n t initiatives on violent crime and racially aggravated crime. The survey also asked about fare evasion as a benchmark for measuring more serious antisocial behaviour. Prevalence of antisocial behaviour The C&JS found that 29% of 10- to 25-year-olds had committed at least one of the types of behaviour listed in Box 1 in the 12 months before interview. Around one in seven had been rude or noisy in a public place or caused a neighbour to complain. All other types of antisocial behaviour were committed by less than one in ten. Fare evasion was much more common, committed by 21%. Males have higher levels across all types of antisocial behaviour, though this is most apparent for carrying weapons, public disturbance and neighbour complaints. Over a third of males had committed at least one antisocial act, compared with around a fifth of females (Table 1). The gender diff e rence holds across age groups, and is consistent with other evidence on offending and delinquency (Farrington and Painter, 2004). For antisocial behaviour, the gender gap is most pronounced among 10- to 11-year-olds and 20- to 25-year-olds. The most problematic age groups are 14- to 16-year-olds (41% committing an antisocial act) and 17- to 19-yearolds (35%). There are some differences across behaviour types. 14- to16-year-olds are particularly likely to be responsible for public disturbance and graffiti. Both these age groups are also more likely to evade fares. Overall, young people who classified themselves as Asian were less likely to commit fare evasion, public disturbance and graffiti than white, black or mixed race groups (although not significantly less likely than the black group). No significant diff e rences were found for neighbour complaints, racial harassment, joyriding and carrying weapons. Further analysis by ethnicity was limited due to low numbers of respondents in the non-white groups. The frequency of antisocial behaviour For each type of antisocial behaviour admitted, respondents were asked the number of times they had committed it in the last 12 months. For each type, around a half or more of perpetrators had only committed it once or twice (Figure 1). Racial harassment was committed most frequently. 21% of those who racially harassed someone Table 1 Prevalence of antisocial behaviour in the previous 12 months by gender, age group and ethnicity 2 Public Neighbour Graffiti Racial Joyriding Carrying Any ASB Fare disturbances complaint harassment weapons excluding evasion Percentages fare evasion Base number 1 Male 19* 16* 4 2 2 10* 35* 23 2,209 Female 12* 10* 3 1* 1 2* 23* 19 2,183 10 11 9* 13 2* 1 0 3* 22* 6* 585 12 13 18 11 5 1 <0.5* 5 29 11* 707 14 16 26* 13 10* 3 3* 9* 41* 28* 971 17 19 18* 14 4 3 3* 9* 35* 37* 787 20 25 9* 13 1* 1 1* 4* 22* 20 1,342 10 16 19* 13 6* 2 1 6 32* 17* 2,263 17 25 12* 14 2* 2 1 5 26* 25* 2,129 White 16 14 4 2 1 6 30 22 3,983 Mixed 13 13 3 1 1 5 26 24 231 Asian 10* 10 1* 3 1 5 22* 15 452 Black 12 10 3 1 2 6 25 21 197 Other 8* 11 1* 2 0 5 23 18 109 All 15 13 4 2 1 6 29 21 4,392 2 Note: 1 Base numbers are for any ASB, base numbers differ slightly for specific ASBs depending on the numbers of respondents who answered don t know or refused. 2 The base for all ethnic groups comes to 4,972, as the analysis was carried out using an additional booster sample of people from ethnic minorities. * indicates they are significantly different from All at the 5% level.
F i g u re 1 Frequency of antisocial behaviours committed as a percentage of those admitting to the act Notes: 1. Frequency of use was not asked in regard to the carrying of weapons. 2. The %s for those who answered don t know or refused are not presented, and therefore figures don t all sum to 100%. had done so at least five times. Overall, around 3% of all 10- to 25-year-olds had committed an act of antisocial behaviour five or more times (based on any of the five ASBs in Figure 1). Number of antisocial behaviour acts committed Of those who had acted antisocially, 68% had only committed one type of antisocial act and 23% had committed two. Only 9% had committed three or more (equivalent to 2.4% of the sample). Males and 14- to 19-year-olds were most likely to commit three or more types (Table 2). Table 2 Number of antisocial behaviour types committed among those admitting to committing at least one antisocial behaviour O n e Tw o Three or more Percentages Base numbers Male 62 27 11 748 Female 77 17 6 473 10 11 71 24 6 120 12 13 70 21 9 194 14 16 60 28 12 363 17 19 67 20 13 254 20 25 74 22 4 290 All 68 23 9 1,221 Antisocial behaviour and risk factors A considerable body of re s e a rch has examined factors associated with offending and delinquency. This is also possible in the C&JS, which collected a wealth of inform a t i o n about respondents lifestyles and circumstances. Table 3 shows the groups with higher than average levels of antisocial behaviour (as measured by the percentage of those committing any antisocial behaviour in the last 12 months). Results are given separately for 10 to16 and 17- to 2 5 - y e a r-olds because of diff e rences in some of the questions asked. The characteristics associated with higher risk are generally consistent across the individual types of antisocial b e h a v i o u r, and are also similar for fare evasion. The risk factors identified accord with those found in other studies (Youth Justice Board, 2001). However, it is not possible to say from the evidence whether these factors cause antisocial behaviour. In some cases the relationship might exist because it reflects an underlying risky lifestyle. In others, a high risk factor may be the result of an i n d i v i d u a l s antisocial behaviour; for instance, being suspended/expelled from school. What is clear is that these groups are more likely to commit antisocial b e h a v i o u r, which has implications for targeting of resources to reduce it. The relative importance of the risk factors was examined using multivariate analysis. This makes it possible to examine the pattern of relationships between several variables simultaneously, taking into account the overlap between them. Logistic regression was used to identify the risk factors independently associated with antisocial behaviour. Two models were examined one for 10 16s and one for 17 25s. Both models explained a large amount of the variance in antisocial behaviour. Many of the factors discussed above remained significant in the model, though not all (Table 4). The most significant factors are discussed below. Gender and age group For 17 to 25s, males and 17 to 19s showed a higher likelihood of antisocial behaviour. For 10 to 16s, being male was an important factor. 3
Table 3 High risk characteristics (% committed any antisocial behaviour in last 12 months) D rug and alcohol factors 1 0 1 6 1 7 2 5 Family/friend factors 1 0 1 6 1 7 2 5 % % % % Taken any drug in the last year 70 45 Friends in trouble with the police 58 40 Taken any Class A drug in the last year * 48 Negative parental relationship 1 54 na Felt drunk at least once a month 66 36 Parents are perceived to have a Go to pub/bar at least once a week 41 31 relaxed attitude to certain behaviours 2 48 na Personality 1 0 1 6 1 7 2 5 School factors 1 0 1 6 1 7 2 5 Low score on adult personality questions 4 na 40 Ever been suspended/expelled 60 40 Low score on child personality questions 5 40 na Have skipped school at least once 62 32 (if still at school asked about last 12 months; if left school asked about ever) Poor school environment 3 50 na Victimisation 10 16 17 25 Homelessness 10 16 17 25 Victim of force or violence 46 47 Ever homeless for a month or longer * 33 Income 10 16 17 25 Area 10 16 17 25 Households with financial difficulties 42 42 Nothing to do in the area 43 na Living in area with three or more disorder problems 6 45 35 Gender 10 16 17 25 Males 37 34 All 32 26 Note: * indicates that base number too low to report results; na indicates that question not asked of this age group. 1. Scale based on whether parents praise, listen, treat fairly, want to know where young person is, and whether parents often argue/fight with each other. 2. Scale based on how parents would feel about young person starting a fight, using graffiti, skipping school and smoking cannabis. 3. Scale based on whether school has clear rules, pupils hit teachers, teachers praise good work and how easy to truant. 4. Scale based on taking risks, speaking without thinking, giving into temptation, not thinking about the consequences, boredom, lack of confidence, not having people you can talk to about problems, easily losing patience, not upset if see others suffering. 5. Scale based on disobedience, worrying, not helping others, boredom, not upset if see others suffering. 6. Respondents asked whether noisy neighbours, teenagers hanging around causing problems, people sleeping rough on streets, people being racially harassed in the street, people using or selling drugs, people being drunk or rowdy were common in their area. Associating with delinquent peer groups This was an important factor, especially for 10 to 16s and is consistent with other studies, though the nature of the association remains open to debate. Individuals might be influenced by delinquent peers who lead them to act antisocially. Alternatively they may have existing antisocial tendencies and simply seek out likeminded individuals. A poor relationship with pare n t s 10- to 16-year-olds with a poor relationship with p a rents was linked with a higher likelihood of antisocial b e h a v i o u r. Whether or not a poor parental re l a t i o n s h i p contributes to the young person s antisocial behaviour, or is a result of it, is difficult to assess, although re v i e w s of longitudinal re s e a rch have indicated pare n t a l relationships influence the young person s behaviour ( Youth Justice Board, 2001). Poor school environment This was important for 10 to 16s (e.g. poor teachers, lack of clear rules). A stronger likelihood of being involved in antisocial behaviour was found for those with a poorer school environment. Drug use and risky alcohol consumption These were associated with a greater likelihood of engaging in antisocial behaviour for both age gro u p s, with alcohol being a stronger factor in 10- to 16-yearolds. For some people use of drugs or alcohol might d i rectly contribute to antisocial behaviour. On the other hand, drug and alcohol use could simply be another manifestation of risky behaviour. Personality aspects Some individuals may be more prone to antisocial behaviour than others, based on diff e rent aspects of their personality. Hyperactivity, impulsivity, sensationseeking, poor concentration and risk-taking have all been linked to antisocial behaviour. In the C&JS, a scale m e a s u re of delinquent personality was customised. Those who were classified as having various adverse characteristics (e.g., prone to risk-taking and disre g a rd for others) had a higher likelihood of committing antisocial behaviour than others. Victims of crime The odds of committing antisocial behaviour are higher for victims of crime, particularly violent crime. 4
Table 4 Risk factors for anti-social behaviour Model 1: 10- to 16-year- o l d s 1 Model 2: 17- to 25-year- o l d s 2 Risk factors Odds ratio 3 Risk factors Odds ratio Taken drugs in last year 3.25 Friends in trouble with police 2.68 Felt drunk once a month or more 2.66 Felt drunk but less than once a month 1.76 (comparison group: never felt drunk/don t drink) Negative relationship with parents 2.12 Poor school environment 1.86 Victim of force or violence 1.66 Victim of non-violent personal crime 1.59 (comparison group: not victim of personal crime) Disorder problems in area 1.57 Low score on personality factors 1.55 Males 1.45 Taken Class A drugs in last year 2.58 Taken non-class A drugs in last year 2.47 (comparison group: not taken drugs in last year) Household getting into financial difficulty 2.56 Household just getting by 1.42 (comparison group: household managing well) Victim of force or violence 2.03 (comparison group: not victim of personal crime) Males 1.92 Low score on personality factors 1.83 Aged 17 to 19 1.74 Felt drunk once a month or more 1.62 Friends in trouble with police 1.51 Notes: 1. Forward stepwise LR. 1,204 cases in model. Nagelkerke R squared=0.282; overall percentage=77%. Variables which were not important in the model were: age group, truancy, expelled or suspended from school, parental attitudes to certain behaviours, whether there is much to do in the area and financial difficulties. 2. Forward stepwise LR used. 1,871 cases in model. Nagelkerke R squared=0.241; overall percentage=77%. Variables which were not important in the model were: truancy, expelled or suspended from school, disorder problems in the area and experiences of homelessness. 3. The odds ratio coefficients refer to the probability of committing an ASB over the probability of not. For categorical independent variables, odds ratios represent the odds of being in the outcome category (committed ASB) compared with the reference category. Factors are listed in order of importance according to the odds ratio. 5 Those who behave antisocially may be at greater risk of becoming victims because their generally risky lifestyles make them vulnerable. More directly, their antisocial behaviour might make them a target for other offenders (e.g., revenge). Individuals may also behave antisocially in response to their victimisation. For further discussion on the links between victimisation and offending, see Wood (2005). The type of are a The area in which respondents live was also import a n t for 10 to 16s. Young people living in areas they themselves considered to have relatively high levels of d i s o rder had a higher likelihood of committing antisocial behaviour than those in other are a s. Financial difficulties This was only a significant factor for the 17 to 25 age gro u p. Area and antisocial behaviour The C&JS has data on the type of area where offenders live (although it cannot be assumed that antisocial behaviour is always carried out in the perpetrators home are a ). Overall, the prevalence of antisocial offenders was similar across rural and urban areas, and only slightly higher in council areas compared to non-council areas (derived from ACORN see Methodological note). However, young people living in urban areas and areas characterised by a high proportion of council housing were far more likely to admit to racial harassment, graffiti and causing neighbour complaints than young people in other areas. Looking at ACORN areas in more detail, those living in council estates and low income areas had the highest levels of neighbour complaints, graffiti, racial harassment and carrying weapons. These areas were also characterised by a high level of perceived disorder (37% said there were three or more disorder problems in these areas, compared with 18% or less in other areas). Perceptions of high d i s o rd e r, as measured by the British Crime Surv e y (Nicholas and Walker, 2004), were also more prevalent in council estates and low income areas. In contrast to the above, non-council and rural areas had a higher pro p o rtion of young people admitting fare evasion. Antisocial behaviour and more serious offending Previous research suggests that those prone to antisocial behaviour are at risk of progressing to more serious offending behaviour (e.g., Farrington, 1995). The 2003 C&JS can be used to examine the association between antisocial behaviour and offending. In some cases, antisocial behaviour may overlap with one of the core offences covered by the survey (robbery, assault, burglary,
criminal damage, thefts of and from vehicles, other miscellaneous thefts and drug dealing). For example, joyriding may, for some individuals, actually involve the theft of a vehicle. However, only a minority (19%) of those who said they had taken part in joyriding said they had actually stolen a vehicle when asked the antisocial behaviour question. This suggests that most joyriders in the sample admitted only to having been passengers. 62% of young people had not been involved in either antisocial behaviour or any of the core offences; 17% had been involved in antisocial behaviour but had not committed a core offence; 9% had committed an offence but no antisocial act. Only 12% had been involved in both. Males and 14-16 year olds were most likely to have been involved in antisocial behaviour and offending. Generally, factors associated with antisocial behaviour were also associated with offending behaviour. For 10- to 1 6 - y e a r-olds, the factors associated with both were : having delinquent friends, being a victim of force or violence, being male, using drugs in the last year, having a negative relationship with parents, and delinquent personality factors. For 17 to 25s, these were: being drunk on a frequent basis, being aged 17 to 19, being male, having delinquent friends, being a victim of force or violence, using drugs in the last year, and delinquent personality factors. Conclusions What counts as antisocial behaviour is not firmly agreed and there is a fine dividing line between antisocial acts and behaviour that might more appropriately attract a criminal label. The C&JS covers specific types of antisocial behaviour and shows that these are committed by a substantial minority of young people. It also provides an indication as to the types of young people most at risk. Given the link found between antisocial behaviour and other offences, tackling the former should help prevent young people progressing to more serious offending. Future annual sweeps of the C&JS are planned (the 2004 survey is underway) and in time it will become a useful source of data on trends in antisocial behaviour. The only trend information currently available is data from the British Crime Survey on people s perceptions of antisocial behaviour in their area (Nicholas and Walker, 2004). Longitudinal data from the C&JS will allow a more in-depth exploration of the links between lower level antisocial behaviour and offending. Methodological note The 2003 C&JS had a random probability sample design. The main sample comprised 10,079 people aged 10 to 65 living in private houses in England and Wales. The number of young people was boosted to around a half (N=4,574) as this is a group of key interest (weighting was applied to correct for this in analysis). In addition there was a booster sample of 1,882 non-white respondents. The response rate for the main sample was 74%, for the nonwhite sample 50%. Fieldwork (by BMRB Social Research and the National Centre for Social Research) took place between January and July 2003. The first part of the interview was interviewer administered; the second part including the more sensitive questions was self-administered. Computer assisted techniques were used. The anti-social behaviour module used AUDIO CASI whereby the questions and responses are pre-recorded and listened to by the respondents through headphones, as well as being presented on the computer screen. This was to make it easier for those with literacy problems to take part (Budd et al., 2005). Differences discussed in the text are significantly different at the 5% level unless otherwise stated. ACORN A Classification of Residential Neighbourhoods, produced by CACI Ltd, classifies households according to the demographic, employment and housing characteristics of the surrounding neighbourhood. 6 References Budd, T., Sharp, C. and Mayhew, P. (2005). Offending in England and Wales: First results from the 2003 Crime and Justice Survey. Home Office Research Study No. 275. London: Home Office. F a rrington, D.P. (1995). The Twelfth Jack Ti z a rd Memorial Lecture. The Development of Offending and Anti-social Behaviour from Childhood: Key Findings from the Cambridge Study in Delinquent Development. Journal of Child Psychology. Vol 36. No. 6. Farrington, D.P. and Painter, K.A., (2004). Gender Differences in Offending : Implications for risk-focussed prevention. Home Office Online-report. Nicholas, S. and Walker, A. (2004). Crime in England and Wales 2002/2003: Supplementary Volume 2: Crime, Disorder and the Criminal Justice System public attitudes and perceptions. Home Office Statistical Bulletin January 2004. London: Home Office. Youth Justice Board. (2001). Risk and protective factors associated with youth crime and effective interventions to prevent it. YJB Research note No. 5. London: Youth Justice Board for England and Wales. Wood, M. (2005). The victimisation of young people: findings from the Crime and Justice Survey 2003. Home Office Findings No. 246. London: Home Office. Ruth Hayward is a Research Trainee in the Offending Surveys and Research Section on placement from Surrey University. Clare Sharp is a Senior Research Officer in the Offending Surveys and Research Section of the Home Office Research, Development and Statistics Directorate.
Errata Page 6, paragraph 2, figure changed to 62% from 71%.