Adolescent male hazardous drinking and participation in organised activities: Involvement in team sports is associated with less hazardous drinking in young offenders

Background There is a lack of research investigating organised activity participation and associated alcohol use in vulnerable groups. Aims The purpose of this research was to test and compare associations between participation in organised activities and indicators of hazardous drinking between young offenders and young non-offenders. Methods Two groups of 13 – 18 year-old males were recruited in Cardiff, UK: 93 young offenders and 53 non-offenders from secondary schools matched on estimated IQ, sex and socioeconomic status. Indicators of hazardous drinking were measured using the Fast Alcohol Screening Test (FAST). Organised activity participation and externalising behaviour was measured by the Youth Self Report. The Wechsler Abbreviated Scale of Intelligence was also administered. Results Young offenders participated in fewer organised activities and had higher FAST scores than non-offenders. Young offenders and non-offenders signi ﬁ cantly differed on mean FAST scores if they participated in no organised activities but not if they participated in at least one team sport. Externalising behaviour problems were unrelated to participation in organised activities. Conclusions Although young offenders were less likely to have participated in organised activities, for them, participation in a team sport was associated with less hazardous drinking. Vulnerable youths who might bene ﬁ t most from sporting activities actually access them the


Introduction
Participation in constructive leisure activities that take place outside the school curriculum, also known as organised activities (Bohnert et al., 2010;Eccles et al., 2003), is generally associated with many benefits (Farb and Matjasko, 2011), including school involvement (Mahoney and Cairns, 1997), academic achievement and psychological adjustment (Fredricks and Eccles, 2006b).Participation in organised activities during adolescence, however, may also be associated with poor health-related outcomes, such as increased alcohol use (Eccles and Barber, 1999;Wichstrøm and Wichstrøm, 2009;Farb and Matjasko, 2011).Early adolescent alcohol use is associated with violence, injuries, drink driving (Gruber et al., 1996) and later alcohol misuse (DeWit et al., 2000).Alcohol consumption is one of the top 10 leading causes of death in the world (Mathers et al., 2009) and England and Wales have some of the highest adolescent drinking rates in Europe (World Health Organisation, 2004, 2009).In England and Wales, offending populations report higher levels of alcohol use (Lader et al., 2000), and young offenders in particular report more hazardous drinking compared with older offenders (Plant and Taylor, 2012).
Adolescence participation in organised activities, such as sport, has been found to be associated with higher levels of alcohol use (Peretti-Watel et al., 2002;McCaul et al., 2004;Terry-McElrath et al., 2011) and an increased growth in alcohol use over time (Eccles and Barber, 1999;Barber et al., 2001;Denault et al., 2009;Mays and Thompson, 2009;Mays et al., 2010), but findings have not been consistent (Darling, 2005;Fredricks and Eccles, 2006b;Barnes et al., 2007).The relationship between sports and alcohol use also varies according to the type of sport.It has been shown that team sports are associated with higher levels of alcohol use than other types of sport (Peretti-Watel et al., 2002, 2003;McCaul et al., 2004;Wichstrøm and Wichstrøm, 2009).Despite the weight of evidence, however, sports-based interventions have been widely adopted in the UK to reduce crime and substance use among youth (Kelly, 2011(Kelly, , 2013)).The relationship between sports participation and delinquency is inconclusive (Burton and Marshall, 2005;Gardner et al., 2009;Sønderlund et al., 2013) and there remains a lack of evidence on the effectiveness of such interventions on reducing alcohol use (Smith and Waddington, 2004).
the exclusion of many at-risk youths, characterised by 'externalising' or 'challenging' behaviours.Truanting youths are harder to recruit into studies generally and school-based research specifically (Jimerson et al., 2000).Externalising behaviour predicts less participation in organised activities such as sport (Fredricks and Eccles, 2006b) and higher attrition rates from studies (Peck et al., 2008).It is questionable whether young people less represented in research, such as young offenders, have similar relationships between hazardous drinking and participation in sport.It has been suggested that sport participation increases supervision, establishes appropriate social norms and curbs drinking through orientation towards success (Eccles et al., 2003;Mahoney et al., 2003;Wichstrøm and Wichstrøm, 2009).On this basis, young offenders might also benefitf r o m participation in such activities and engage less in hazardous drinking than young offenders who do not participate.
Our aim, therefore, was to understand the relationship between participation in organised activities and levels of hazardous drinking in a group of vulnerable adolescents previously under-represented in research.It was predicted that young offenders would show higher levels of hazardous drinking and lower levels of participation in organised activities than non-offenders matched for age, sex and socioeconomic status.It was hypothesised that, within groups, participation in at least one organised activity would be associated with lower scores indicative of hazardous drinking for both young offenders and nonoffenders, but participation in a team sport compared with no participation in any activity would be associated with a higher prevalence of hazardous drinking in non-offenders and lower levels in young offenders.The role of externalising behaviour in predicting participation was investigated to help understand these relationships.

Young offenders
Ninety-three young offenders, mean age 16 years (standard deviation [SD] 1.02, range 13-18), were recruited from a local Youth Offending Team (YOT).Each local authority in England and Wales has a YOT that is governed by the Youth Justice Board and reports to the Ministry of Justice.The main priority of a YOT is to prevent recidivism by identifying and meeting the needs of youngsters.Participants recruited from the YOT had been convicted of at least one offence.
There were very small numbers of young female offenders, which is usual (Zheng and Cleveland, 2013), so only male offenders were included in the analyses.Participants who had an estimated intelligence quotient (IQ) score of less than 70, and therefore possible learning disabilities, were also excluded.

Organised activities and hazardous drinking
Non-offenders Fifty-three non-offenders, mean age 15.1 years (SD 1.3, range 13-18), were recruited from local schools in the catchment area of the YOT, ensuring that socioeconomic factors were similar across study groups.They had had no past or current contact with the criminal justice system.

Organised activities
Participation in organised activities was measured using the Youth Self Report (YSR) (Achenbach and Rescorla, 2001), a two-part questionnaire measuring behavioural problems and competencies in youth aged 11-18 years.Participants were asked to list up to three 'organisations, clubs, teams or groups' that they belonged to.Those who listed football, rugby, hockey, cricket or basketball were categorised as participating in at least one team sport.Those who reported nonteam activities (kickboxing, gym, swimming, fishing, snooker, skittles, pool, rowing, motorbikes, sign language, computer club, music group, cadets and youth group) were categorised as doing other activities.Those who did not report any participation were categorised as doing 'no activities'.These three groups were mutually exclusive.Those who reported at least one team activity were categorised as participating in a team sport even if they reported participation in other non-team activities.Participants were then described according to three binary variables, one for each activity category, which were the dependent variables in subsequent analyses.Three-way comparisons between those who participated in sports, other activities and no activities have been conducted previously (Gardner et al., 2009).In our sample, team sports were more popular and the small number of other types of activities reported did not allow for comparison.

Externalising behaviour
The YSR questionnaire (Achenbach and Rescorla, 2001) is a reliable and valid measure of emotion and behaviour in the past 6 months, with 112 items, each rated on a three-point scale.The externalising scale is a hierarchical score drawn from delinquent and aggressive behaviour sub-scales.Raw scores on the externalising scale were converted into t-scores and used in the current analyses.The current eight-syndrome taxonomic model of the YSR meets the criteria for a good fit to data from 30,243 youths in 23 societies (Ivanova et al., 2007).

Indicators of hazardous drinking
Hazardous drinking was measured by the Fast Alcohol Screening Test (FAST; Hodgson et al., 2002).This assesses drinking using four questions about frequency of bingeing and negative effects resulting from alcohol use in the past year; scores range from 0 to 16, with a score of 3 or more indicating hazardous drinking.The Hallingberg et al.
FAST has been shown to have good test-retest reliability (>0.80), internal consistency (Cronbach's alpha, 0.77), specificity and sensitivity for identifying hazardous drinkers (Hodgson et al., 2002) and to correlate with breath alcohol concentration (Moore and Cusens, 2010) and other established alcohol measures in samples of young people (Bowring et al., 2013).The FAST is also used to measure changes in drinking patterns (Brendryen et al., 2013;Moore et al., 2013).

Estimated intelligence quotient
Intelligence quotient was estimated by creating a scaled score for vocabulary and matrix reasoning sub-tests of the Wechsler Abbreviated Scale of Intelligence (Wechsler, 1999), adjusted for the age of the participant.These scores were summed to create an IQ estimate.

Offending history
The offence frequency rate for young offenders was measured by dividing the number of offences committed by the participant's age.Offence severity was measured by the Youth Justice Board's seriousness scale.

Procedure
This study was approved by Cardiff University's School of Psychology's Research Ethics Committee.Information about the study, including the tasks, their right to withdraw at any time and the confidentiality of the data collected was provided.For offenders and non-offenders alike, both the young people and their parent/guardian were required to give consent before participation.

Young offenders
Young offenders were recruited by caseworkers who worked at the YOT.The eligibility criterion was that the participant's behaviour had resulted in any contact with the criminal justice system, whether a court conviction, a reprimand or a final warning.

Non-offenders
Contacts within local schools provided summaries of the research to students.Those who expressed an interest in participation were given more information.The inclusion criterion was absence of any criminal justice system contact as a perpetrator of delinquent or offending behaviour.
For both groups, interviews were conducted individually with each participant in a private room.
Participants were given £5 store vouchers for every hour of their participation.
There was a tendency for the offenders to have lower estimated IQ scores than the non-offenders (mean 87.The young offenders had committed an average of 12 offences (SD 11, range 1-52), yielding a mean offending rate of 0.76 (SD 0.67, range 0.06-3.06).For the following analyses in this study, the offence frequency rate was transformed with a log transformation in order to reduce skewness.The mean offence severity score was 5.35 (SD 1.28, range 2-8).
Table 1 shows the results of the logistic regressions used to investigate measures associated with participation patterns.Age, estimated IQ and externalising behaviour were entered as the independent variables.None of them were associated with participation in any activity or more specifically with a team sport.A likelihood ratio chi-square test found that both models fit the data better than an empty model (χ 2 4 = 27.48,p < 0.001; χ 2 4 = 25.11,p < 0.001, respectively).

Discussion
This cross-sectional study provided insights into participation rates for organised activities and indicators of hazardous drinking in a British sample of at-risk youth with low socioeconomic status.Young offenders in this study are particularly under-represented in other research; typically, such a group includes adolescents who are frequently excluded from school, involved in the criminal justice system from an early age and need additional support from numerous statutory agencies.To our knowledge, this is the first study to assess systematically the relationship between participation in organised activities and hazardous alcohol use in such a vulnerable cohort.
As predicted, young offenders participated in significantly fewer organised activities than non-offenders and had significantly higher levels of hazardous Organised activities and hazardous drinking drinking.For those who participated in no organised activities, young offenders had significantly higher scores on hazardous drinking than non-offenders.Young offenders did not, however, differ from non-offenders in FAST scores if they participated in at least one team sport.Participation in an organised activity in itself was not associated with levels of hazardous drinking.
It is interesting that the relationship between team activities and drinking was in opposite directions in offender and non-offender groups.It has previously been shown that students who participate in sports (Eccles and Barber, 1999;Barber et al., 2001;Fredricks and Eccles, 2006a;Denault et al., 2009;Mays et al., 2010) and team sports specifically (Peretti-Watel et al., 2002;Peretti-Watel et al., 2003;McCaul et al., 2004) report higher levels of alcohol use.In our study, there was an indication of this pattern amongst non-offenders but not offenders.The apparent advantage for the offender groups is consistent with previous observations that show those who are worse-off benefit most from taking part in organised activities (Mahoney and Cairns, 1997;Mahoney, 2000;Diamond and Lee, 2011).
Challenging behaviours have been shown to predict both alcohol use (Hawkins et al., 1992;Patrick and Schulenberg, 2010;Maslowsky and Shulenberg, 2013) and less participation in organised activities (Fredricks and Eccles, 2006b).In our study, externalising behaviour was not associated with participation in team sports or, indeed, any activity and cannot explain the observed relationships between participation in activities and alcohol use.As adolescents get older, they often reduce their participation in organised activities and increase their alcohol use (Peretti-Watel et al., 2002).Although our young offenders were moderately older than non-offenders, age was not associated with participation.
One limitation of our study was that we had to limit our analyses to males, and our results are unlikely to be generalisable to females.Female adolescents are less likely to participate in sports compared with male adolescents (Pate et al., 2000;Darling, 2005;Moore and Werch, 2005;Denault and Poulin, 2009) and the role of organised activities for young female offenders is an area that still needs to be explored.A second limitation was that other characteristics of activity participation, such as duration, level of competition and the amount of different activities participated in, were not measured.These have been shown to influence relationships between participation in organised activities and hazardous drinking (Peretti-Watel et al., 2002;McCaul et al., 2004;Bohnert et al., 2010).Third, we could only report cross-sectional relationships.Although participation may influence alcohol use, alcohol use may also influence participation.It remains that latent variables and self-selection factors other than externalising behaviour may explain observed relationships (Peretti-Watel, 2009).Longitudinal studies or controlled interventions would have the ability to disentangle these relationships and the results reported here suggest that such studies might be feasible.
Despite these limitations, we have shown that the relationship between participation in organised activities and hazardous drinking in smaller, atypical groups differ from that in the general population.Young offenders who participate in team sports may be less hazardous alcohol users because of the physical demands of the sport, the wider social support gained from an increase in social capital, the influence of social group norms and positive social role modelling (Eccles et al., 2003;Mahoney et al., 2003;Wichstrøm and Wichstrøm, 2009).
In conclusion, although young offenders were less likely to have participated in organised activities, for them, participation in a team sport was associated with less hazardous drinking.This suggests that vulnerable youths who might benefitm o s t from sporting activities actually access them the least.In order to change this, next steps would be to identify the barriers to participation, whether environmental, social and/or psychological that vulnerable young people face and determine the characteristics of activities that they find attractive.Future research should also aim to develop and evaluate sport interventions for vulnerable young people.

Table 1 :
Associations between participation patterns using logistic regressions.Dichotomous outcomes compared participation in no activity to participation in (a) an organised activity (n = 99) and (b) team sport (n = 85) Participation in an activity Participation in team sports

Figure 1 :
Figure 1: Participation rates in organised activities for young offenders and non-offenders

Figure 2 :
Figure 2: Participation in no organised activities or at least one organised activity and mean FAST scores for young offenders and non-offenders