Young adults, alcohol, crime anddisorder

ANNA RICHARDSON AND TRACEY BUDD, Drugs and AlcoholResearch Unit, Home Office, London, UK

ABSTRACTBackground Alcohol-related crime is increasingly being recognized as a problem incities and towns with popular entertainment districts. Crime and disorder linked toalcohol has been particularly connected with binge drinking or heavy sessionaldrinking. Research evidence indicates that it is the young adult age group who aremost likely to be involved in crime and disorder and the most likely to binge drink.This paper examines the relationship between binge drinking and criminal and disor-derly behaviour among 18- to 24-year-olds.Method Secondary analysis was undertaken of the 1998/1999 Youth LifestylesSurvey, a large-scale, representative, household survey of 12- to 30-year-olds livingin England and Wales.Results Binge drinking, and especially male binge drinking, among 18- to 24-year-olds is statistically related to offending behaviour. In the 12 months prior to interview39% of binge drinkers admitted to committing an offence and 60% admitted criminaland/or disorderly behaviour during or after drinking alcohol. Multivariate analysisfound that binge drinking remains strongly associated with criminal and disorderlybehaviour even after taking other relevant factors into account. Individuals who gotdrunk at least once a week had more than five times the odds of being involved infighting or violent crime. For offences or disorderly behaviour that took place duringor after drinking alcohol an individual had a seven times greater chance of breaking ordamaging something and a five times greater chance of being involved in a fight ifhe/she got drunk at least once a week.Discussion These findings suggest that frequency of drunkenness is a better predictorof offending behaviour than frequency of drinking per se. Using frequency of drunk-enness as the basis for defining binge drinking reveals that a large minority of youngadults who binge drink also become involved in offending or disorderly behaviour.Binge drinking is particularly associated with crimes of violence. The relationshipbetween binge drinking and crime and disorder is stronger among young males thanfemales.

Criminal Behaviour and Mental Health, 13, 5–17 2003 © Whurr Publishers Ltd 5

CBMH 13.1_crc 25/9/03 2:28 pm Page 5

Introduction

The relationship between alcohol, crime and disorder is complex. Whilstevidence of an association is well documented in the research literature, theextent and nature is a matter of continued debate.

The British Crime Survey indicates that a third of violent incidentsbetween strangers and a fifth of violent incidents between acquaintances takeplace in or around a pub or club. In half of all stranger incidents and a third ofacquaintance incidents the victim considered the offender to be ‘under theinfluence’ of alcohol (Mattinson, 2001). Canadian-based research byPernanen (1991) found that in violence occurring in bars and taverns drinkingby both victim and offender was more likely than in violence occurring inother contexts. Furthermore, much of the research evidence suggests that thebinge-drinking culture, broadly regarded as high levels of consumption over ashort period of time, mainly associated with young adults, is linked with crimeand disorder.

There has been previous research on binge drinking, but usually with smallor select samples. Deehan et al. (2002) found that the majority of thosearrested late at night in city-centre areas were intoxicated with alcohol andalmost half were aged between 18 and 24 years. Studies of student populationshave examined the links between binge drinking and a range of social harms,including involvement in crime. Studies of students in the US and the UKhave shown that high levels of binge drinking are associated with factors suchas illicit drug taking or involvement with the police (Marlatt et al., 1995;Webb et al., 1996; Wechsler et al., 2000). Wechsler et al. (1995) identified alink between binge drinking (5[males]/4[females] or more drinks in a row) andstudents who experienced problems such as injury, committing vandalism andbeing involved with the police. Studies carried out in Accident andEmergency settings have demonstrated that binge drinking is connected toincreased risk of injury arising from an assault (Cherpital, 1993; Shepherd andBrickley 1996; Borges et al., 1998). Shepherd and Brickley (1996) undertookresearch in a large city-centre A&E department which compared males aged18–35 attending with assault injuries with controls and found thatconsumption of more than 10 units of alcohol in the six hours prior to assaultwas associated with injury.

While this type of research provides a strong indication that binge drinkingis associated with offending behaviour, it is difficult to generalize from theresults or to explore the extent of the link between binge drinking and crimeand disorder. The focus therefore of this paper is to present the results of theexploration of this link based on a nationally representative sample of 18- to24-year-olds.

6 Richardson, Budd

CBMH 13.1_crc 25/9/03 2:28 pm Page 6

Method

Design

The 1998/99 Youth Lifestyles Survey (YLS) is a nationally representativesurvey of 4848 12- to 30-year-olds living in private households in England andWales. The survey was designed to measure self-reported offending but alsocollected basic sociodemographic details as well as information associated withoffending, including alcohol and drug use. This provided a unique opportunityto explore the relationship between patterns of offending behaviour andalcohol consumption. A more detailed account of the methodology can befound in Flood-Page et al. (2000) The full technical report details the samplingstrategy and includes a copy of the questionnaire (Stratford and Roth, 1999).For the purposes of this study we were interested primarily in the 18- to 24-year-old age group (n = 1336).1

Measures

An inherent problem in researching alcohol, and in particular binge drinking,is definition. Within the research community there is no standard approach todefining or measuring binge drinking (Midanik, 1999; Murgraff et al., 1999).The YLS enabled us to explore the merits of two alternative definitions ofbinge drinking, one based on unit consumption and one based on frequency ofdrunkenness. The unit measure was based on the average number of unitsdrunk per drinking day in the week prior to interview. Binge drinkers weredefined as men who had drunk more than eight units on at least one day in thepast week and women who had drunk more than six units on at least one dayin the past week. This measure is consistent with measures used in other UK-based surveys, such as the General Household Survey and the Health Surveyfor England, and allowed further robust statistical analysis. However, becauseof the questions asked in the YLS an average measure had to be adopted whichwill necessarily underestimate the number of binge drinkers.2 For thefrequency of drunkenness based measure those who reported getting drunk atleast once a week, several times a month or once or twice a month wereclassified as binge drinkers.

Initial analysis was undertaken using both definitions. The results indicatedthat patterns of offending behaviour were similar using both definitions ofbinge drinking. However, the frequency of drunkenness measure only isreported on here. There are several reasons for this. The unit definitionmisclassifies some binge drinkers owing to the averaging approach; takes noaccount of important individual factors, such as weight, alcohol tolerance and

Young adults, alcohol, crime and disorder 7

CBMH 13.1_crc 25/9/03 2:28 pm Page 7

gender, which may affect the influence that alcohol has on an individual; andis based on the last week’s experience. In contrast the drunkenness definitioncaptures the impact that alcohol has on individuals in terms of feelings ofintoxication and also measures the frequency of behaviour over a period of 12months, better reflecting typical behaviour. However, it should be noted thatone of the limitations of the drunkenness definition is that ‘feeling drunk’ is asubjective term and therefore interpretation may differ from person to person.Moreover, this definition (the same applies to the unit definition) does notcapture the social context in which alcohol is consumed (e.g. location, atmos-phere and presence of other drinkers), a factor that has been identified in theliterature as influencing the relationship between alcohol and criminal ordisorderly behaviour (Midanik, 1999; Raistrick et al., 1999).

Analysis

The YLS provides two ways of investigating crime and disorder. First, respon-dents were asked if they had committed various offences in the 12 monthsprior to interview and, second, respondents who had drunk alcohol in theprevious 12 months were directly asked if they had been involved in specifieddisorderly or criminal acts during or after drinking alcohol. With both thesemeasures bivariate analysis was undertaken to examine which sociodemo-graphic and lifestyle factors, including drinking behaviour, were associatedwith offending and disorderly behaviour.3

Forward-stepwise logistic regression was undertaken to establish whetherbinge drinking remained predictive of involvement in crime and disorder,taking into account various other variables associated with criminal and disor-derly behaviour. Frequency of drinking and frequency of drunkenness wereentered into the models to enable us to identify which was most strongly relatedto offending behaviour. Other independent variables were selected based onprevious research (see for example, Flood-Page et al., 2000). However, in inter-preting the results it should be noted that some factors that may impact onoffending and disorderly behaviour were not collected in the YLS.

Results

Sample characteristics

Analysis is based on the subset of 18- to 24-year-olds (n = 1334).4 There was aneven split between male and females. In total, 93% (n = 1223) of respondentsdescribed themselves as white, 1% (n = 35) as black, 4% (n = 41) as Asian and2% (n = 34) as none of these. The majority (60%; n = 814) reported being inwork or training, with a further 23% (n = 266) in higher or further education;8.3% (n = 111) were unemployed. Some 83% (n = 1119) were single and 15%(n = 238) reported having children.

8 Richardson, Budd

CBMH 13.1_crc 25/9/03 2:28 pm Page 8

Extent and nature of binge drinking

As discussed above, the frequency of drunkenness definition was chosen as themeasure of binge drinking. Those who admitted getting very drunk at least oncea month were classified as binge drinkers. A threefold typology was developedbased on frequency of drinking and frequency of drunkenness. In this typology39% (n = 508) of 18- to 24-year-olds were classified as binge drinkers (thosewho felt very drunk at least once a month), 42% (n = 551) as regular drinkers(those who drank at least once a month but felt very drunk less often) and 18%(n = 196) as occasional or never drinkers (those who drank alcohol less thanonce a month, including those who had not drunk at all). Binge drinkers weremore likely to be male (61%), white (98%), single (92%) and less likely to havechildren (6%) than their non-binge-drinking counterparts.

Binge drinkers reported drinking more frequently than non-binge drinkers:95% drank at least once a week, with a fifth drinking nearly every day of theweek. Binge drinkers were also more likely to drink several types of drink andwere more likely to drink beer and spirits than non-binge drinkers. Whilebinge drinkers were as likely as regular drinkers to spend time in pubs and bars(95% vs. 92%), binge drinkers were far more likely to attend night clubs,dances or parties (91% vs. 77%). In terms of usual drinking place and drinkingcompanion binge drinkers were more likely to drink within the confines of thenight-time economy and more likely to drink with friends and work colleaguesthan other regular drinkers.

Binge drinking and general offending

Overall, a quarter of all 18- to 24-year-olds admitted committing at least one ofthe offences asked about in the YLS in the previous 12 months. Young menwere significantly more likely to admit to offending than young women (33%compared with 13%). Examining the link between drinking patterns andoffending behaviour, 18- to 24-year-old binge drinkers were almost three timesmore likely to admit committing an offence than other regular drinkers.Focusing on the broad offence categories of violent crime (including the sub-category of fights), criminal damage and theft, Table 1 shows the percentage ofbinge drinkers compared with other regular drinkers and occasional or non-drinkers who admitted participation in each offence type. For all of the offencegroups binge drinkers were significantly more likely to admit to an offence.The difference was particularly marked for fights, with binge drinkers beingfive times as likely to admit to an incident than a regular drinker. For thosewho occasionally or never drank, prevalence of offending was far lower withonly 8% (n = 14) admitting offending in the 12 months prior to interview.

Male binge drinkers were far more likely to offend than female bingedrinkers (49% of male binge drinkers admitted committing an offencecompared with 22% of women binge drinkers). Again, the biggest discrepancy

Young adults, alcohol, crime and disorder 9

CBMH 13.1_crc 25/9/03 2:28 pm Page 9

was for taking part in a fight: male binge drinkers were more than 10 timesmore likely to have been involved in fighting than female binge drinkers. Malebinge drinkers were also significantly more likely to offend than other regularmale drinkers. For example, 25% of male binge drinkers reported beinginvolved in a violent crime compared with 7% of regular male drinkers. Foryoung women the relationship between binge drinking and offending was notsignificant for any offence.

Whilst bivariate analysis indicates that there is a significant relationshipbetween offending behaviour and binge drinking, other research has shownthat factors such as educational achievement, illicit drug use and associationwith delinquent peers are also linked to offending behaviour. Multivariateanalysis is required to isolate the independent influence of these variousfactors. Logistic regression was chosen to determine whether drinkingbehaviour, particularly binge drinking, was predictive of offending among 18-to 24-year-olds once other relevant factors had been taken into account. Afternon-significant variables had been removed from the analyses (by forwardelimination) five variables were retained in the final regression model. Table 2presents the factors that proved to be significantly predictive of involvementin any offence, violent crime, fights and theft.5

Frequency of drunkenness was strongly associated with overall offendingbehaviour even when other factors were controlled for. Frequency of drinkingwas not associated with any offending behaviour. Getting drunk at least once aweek increased an individual’s odds of being involved in a violent crime or afight by more than five times, relative to those who admitted getting drunk lessthan once a month. These odds reduced to around three times for those whogot drunk once or twice a month. The only offence type that frequency ofdrunkenness was not predictive of was theft. However, unlike fighting orviolent crime drug use was predictive of theft.

Several other variables also proved predictive of offending behaviour. Beingmale was the only factor that was more predictive of all offence groups than

10 Richardson, Budd

Table 1: Offending among 18- to 24-year-olds, by drinking pattern

Percentages Binge drinker Other regular Occasionally or drinker never drinks

Any offence in last 12 months 39 14 8Violent crime 17 4 2Taken part in a group fight in public place 15 3 2Theft 11 6 3Criminal damage 4 < 1 1

Base n 355 381 181

Source: 1998/1999 Youth Lifestyle Survey.

CBMH 13.1_crc 25/9/03 2:28 pm Page 10

frequency of drunkenness. Knowing a person who had been involved inoffending was significantly predictive in all groupings. Being expelled orexcluded from school significantly increased the odds of becoming involved inviolent crime only.6

Binge drinking and offending during or after drinking alcohol

The YLS also allows the examination of a more direct link between alcoholintake and offending behaviour by analysis of the questions that ask aboutdisorderly behaviour during or after drinking alcohol. Four acts were askedabout: being involved in a heated argument, fighting, breaking something andtaking something.7 Overall, 38% (n = 448) admitted committing at least oneof the four acts in the12 months prior to interview, with males being morelikely to commit an act than women (50% vs. 26%).

Patterns of drinking were strongly associated with the likelihood ofinvolvement in crime and disorder during or after drinking. Binge drinkerswere significantly more likely to admit to each of the four behaviours thanother regular drinkers. Table 3 shows the percentage of those reporting each ofthe acts by the threefold drinking typology.8 Binge drinkers were almost fourtimes as likely to take part in a fight than regular drinkers and over five timesas likely to admit to criminal damage and eight times as likely to admit totheft. Unsurprisingly, occasional drinkers were far less likely to report disor-derly behaviour during or after drinking (13%).

Young adults, alcohol, crime and disorder 11

Table 2: Offending in the last 12 months – predictive factors for 18- to 24-year–olds

Exp(B) – odds ratios Any offence Violent crime Fighting Theft

Frequency of drunkenness (base: less than once a month)

At least once a week 3.02* 5.63* 5.74* nsOnce or twice a month 1.95* 3.12* 2.85* nsOther factorsMale 2.83* 6.83 * 6.62* 3.87*Know delinquent 2.39* 3.31* 2.78* 3.03*Expelled/excluded ns 2.04* ns nsUsed a drug in last 12 months 4.02* ns ns 2.36*Hosmer and Lemeshow # # # #

Base n 867 867 867 867

Notes: *Indicates statistic is significant, p = <0.05; ns indicates the variable was not signif-icant, p = > 0.05; #indicates Hosmer and Lemeshow test is not significant, therefore model isa good fit. Employment status, ethnicity, children, educational qualifications, amount ofspending money, frequency of drinking and going to pubs or clubs were also tested but didnot come out in the model.Source: 1998/1999 Youth Lifestyle Survey.

CBMH 13.1_crc 25/9/03 2:28 pm Page 11

Male binge drinkers were significantly more likely to have committed any ofthe behaviours than female binge drinkers (69% vs. 45%). Furthermore, in allcategories, male binge drinkers were significantly more likely to have reportedinvolvement in criminal or disorderly behaviour during or after drinking thanregular male drinkers. Male binge drinkers were around three times more likelyto be involved in fighting than other regular male drinkers. Among women therelationship between binge-drinking status and offending and disorderlybehaviour during or after drinking was significant (unlike for generaloffending). In all categories, apart from taking something, female binge drinkerswere significantly more likely to have committed the act than regular femaledrinkers.

The relationship between offending and disorderly behaviour during orafter drinking was further tested by multivariate analysis. In a similar way togeneral offending, logistic regression was used to isolate the independentinfluence of different factors.

As with the results from the general offending model, frequency of drunk-enness remained very strongly associated with criminal and disorderlybehaviour during or after drinking even after other factors had been controlledfor. Young adults who got drunk at least once a week were almost seven timesas likely to admit to criminal damage, five times as likely to admit fighting andfour times as likely to become involved in an argument than those who gotdrunk less than once a month. Those who got drunk once or twice a monthwere around three times as likely to be involved in any of these behaviours.Unlike the general offending model, frequency of drinking proved predictive.However, this was only apparent in the ‘any consequence’ category, with those

12 Richardson, Budd

Table 3: Crime and disorder while drinking among 18- to 24-year-olds, by drinking patterns

Percentages Binge drinker Other regular Occasionally drinker drinks*

In the past 12 months has during or after drinking …

…got into a heated argument 50 23 12…got into a fight 26 7 3…broken, destroyed or damaged something

belonging to someone else 17 3 1…taken something belonging to

someone else 8 1 1…done any of the above 60 25 13

Base n (those who had drunk last 12 months) 508 551 196

Notes: *Drunk alcohol less than once a month in the previous 12 months.Source: 1998/1999 Youth Lifestyle Survey.

CBMH 13.1_crc 25/9/03 2:28 pm Page 12

who drank at least once a week being twice as likely to admit to at least one ofthe specified acts relative to those who drank less than once a month.

The only other factors that were predictive of all behaviours were beingmale and using illicit drugs in the previous 12 months. However, age of firstdrink was associated with being involved in heated arguments, age waspredictive of fighting, with the 18- to 20-year-old age group having more thandouble the odds of getting into a fight than their younger counterparts, andthose with no or unspecified qualifications had higher odds of involvement infights. Table 4 shows the full results.9

Young adults, alcohol, crime and disorder 13

Table 4: Crime and disorder while drinking – predictive factors for 18- to 24-year-olds

Exp(B) – odds ratios Any Heated Got into Broken, consequence argument a fight damaged

Frequency of drinking (base: less than once a month)

At least once a week 1.98* ns ns nsOnce or twice a month 1.64 ns ns ns

Frequency of drunkenness (base: less than once a month)

At least once a week 4.33* 3.79* 5.02* 6.80*Once or twice a month 3.09* 3.03* 3.27* 3.18*

Age of first drink (base: 15–24)

1–12 1.46* 1.60* ns ns13–14 1.76* 1.87* ns ns

Qualifications (base: higher)

Other ns ns 8.18* nsNone ns ns 4.42* nsGCSE ns ns 1.11 nsA levels ns ns 1.38 ns

Other factors

Male 2.08* 1.60* 4.33* 3.93*18–20 ns ns 2.32* nsUsed a drug 1.75* 1.81* 2.58* 3.01*

Hosmer and Lemeshow # # # #

Base n 1237 1237 1237 1237

Notes: *indicates statistic is significant, p = <0.05; ns indicates the variable was not signif-icant, p = > 0.05; #indicates Hosmer and Lemeshow test is not significant, therefore model isa good fit. Ethnicity, children, employment status, going to a pub or club and going out in theevening were also tested but did not come out in the model.Source: 1998/1999 Youth Lifestyle Survey.

CBMH 13.1_crc 25/9/03 2:28 pm Page 13

Discussion

The adoption of frequency of drunkenness as a measure for binge drinking isnot common in alcohol research but was chosen for various reasons. First, wefound no strong reason to adopt the unit definition as bivariate analysisindicated that patterns of offending were similar using both measures. Second,it was felt that the drunkenness definition overcame many of the limitations ofa unit-based measure and third it allowed for a more in-depth investigation ofwhether it is drinking itself or intoxication that is linked to offendingbehaviour.

The results of the research indicate that the vast majority of 18- to 24-year-olds drink alcohol but the majority do not frequently drink excessively.However, a sizeable minority (39%) do, and it is this group that the researchhas focused on. Findings clearly demonstrate a statistical association betweenbinge drinking and involvement in disorderly or criminal behaviour. While itis not possible to conclude a simple causal relationship between the two, i.e.not every 18- to 24-year-old who binge drinks consequently commits acriminal or disorderly act, the analyses provide convincing evidence of anenduring relationship even after other factors are taken into account. To put itsuccinctly, regular binge drinking is far more strongly associated with offendingbehaviour than many other factors associated with offending and regulardrinking per se. It therefore seems vital to consider patterns of drinkingbehaviour when examining links with crime and disorder.

Violent crime was most closely associated with binge drinking. Inexamining general offending behaviour binge drinkers were most likely to beinvolved in violent crime whereas other regular drinkers were most likely to beinvolved in theft. Interestingly, this was also reflected in the multivariateanalysis where frequency of drunkenness was predictive of all offence groupstested apart from theft, while using an illegal drug in the last 12 months waspredictive only of theft. This adds substance to the general perception thatalcohol use is linked to violent crime and drug use to acquisitive crime.

There is significant gender differentiation in the prevalence of offendingbetween men and women. Males generally are more likely to admit offendingthan women and in examining offending by alcohol intoxication this linkremained just as strong. Male binge drinkers were significantly more likely tohave been involved in each of the general offending categories and all of thebehaviours during or after drinking alcohol than other female binge drinkers.Among males, binge drinkers were significantly more likely to admit to alloffence categories and disorderly behaviour than other regular drinkers. Thelink between drinking and offending was not so evident among females. Forfemales there was no significant difference in general offending between bingedrinkers and other regular drinkers. However, female binge drinkers were signif-icantly more likely to admit to all forms of criminal and disorderly behaviourduring or after drinking, with the exception of theft, than regular drinkers.

14 Richardson, Budd

CBMH 13.1_crc 25/9/03 2:28 pm Page 14

Our results suggest that reducing the amount drunk in one session wouldhave an impact on levels of crime and disorder. Evidence from other research(for example Deehan et al., 2002) indicates that alcohol-related crime is aparticular problem for high-density areas of entertainment districts and it is inthis context that the greatest improvements would probably be seen. However,it should be remembered that drinking does not usually take place in isolationand it is unlikely that the pharmacological effects of alcohol alone are whollyresponsible for the link between alcohol intake and increased offending.Indeed, as the results demonstrate, almost 60% of binge drinkers have notoffended in the previous 12 months and 40% did not admit to disorderly orcriminal conduct after drinking. Individual, environmental and social andcultural factors have all been shown to have an impact on offending behaviour(Midanik, 1999; Homel et al., 2001), and thus need to be considered whenthinking of ways of addressing alcohol-related crime and disorder.10

Notes

1. This figure refers to only those who accepted self-completion (n = 1376 for all 18- to 24-year-olds in the sample).

2. The measure is based on total units consumed during the week divided by the number ofdrinking days. Therefore, for example, a male drinking 12 units across two drinking days wouldhave an average consumption of six units per drinking day and therefore not be classified as abinge drinker. However, it could be that 10 units were consumed on one day and two units onthe other. Other large surveys, such as the GHS and HSE, do not face this problem as they askabout the heaviest drinking day in the previous week.

3. All statistics presented are significant at least at the two-tailed 10% level unless otherwisestated.

4. Note that sample sizes for the analysis presented here differ slightly because of varyinglevels of missing information (refusal and don’t know responses). Moreover, in looking atgeneral offending there are fewer cases, due to a split-sample experiment (see Stratford andRoth, 1999 for further details).

5. The category ‘criminal damage’ was not used in logistic regression as base numbers weretoo low.

6. Multivariate analysis using the unit-based measure of binge drinking showed a similarpattern of results to the frequency-of-drunkenness-based measure presented.

7. The category ‘taken something’ was not used in logistic regression as base numbers weretoo low.

8. Interestingly, the percentage of binge drinkers and regular drinkers admitting fighting andcriminal damage during or after drinking is higher than when responding to generaloffending (fights and criminal damage) in the last year. While it is not possible to identifywith certainty why this is so, there are several possible reasons. The two sets of questionsare worded differently and are in different sections of the questionnaire. Respondents mayfeel more comfortable admitting to such behaviour if they can excuse themselves throughtheir drunken state or may feel that their behaviour while drunk did not really constitute a‘crime’, thus failing to report to the ‘general offending’ set of questions. Moreover,reference to drinking in the question may prompt them to remember specific incidents thatthey failed to recall during the general questions.

9. Mutivariate analysis using the unit-based measure of binge drinking showed a similarpattern of results to the frequency-of-drunkenness-based measure presented.

Young adults, alcohol, crime and disorder 15

CBMH 13.1_crc 25/9/03 2:28 pm Page 15

10. This paper is based on Home Office Research Study No. 263, Alcohol, Crime and Disorder: AStudy of Young Adults (2003) by Anna Richardson and Tracey Budd.

References

Borges G, Cherpital, CJ, Rosovsky, H (1998) Male drinking and violence-related injury in theemergency room. Addiction 93: 103–112.

Cherpital CJ (1993) Alcohol and violence-related injuries: an emergency room study. Addiction88: 79–88.

Deehan A, Marshall E, Saville D (2002) Drunks and Disorder: A Description of Alcohol-relatedCrime and Disorder in Two Late Night City Centres, Police Research Series Paper 150.London: Home Office

Flood-Page C, Campbell S, Harrington V, Miller J (2000) Youth Crime: Findings from the1998/99 Youth Lifestyles Survey Home Office Research Study 209. London: Home Office.

Homel R, McIlwain G, Carvolth R (2001) Creating safer drinking environments. In Heather N,Peters TJ, Stockwell T, eds. International Handbook of Alcohol Dependence and Problems.Chichester: Wiley.

Marlatt GA, Baer JS, Larimer M (1995) Preventing alcohol abuse in college students; a harmreduction approach. In Boyd G, Howard J, Zucker R, eds. Alcohol Problems AmongstAdolescents. Hillsdale: Hove.

Mattinson J (2001) Stranger and Acquaintance Violence: Practice Messages from the British CrimeSurvey, Briefing Note 7/01. London: Home Office.

Midanik LT (1999) Drunkenness, feeling the effects and 5+ measures. Addiction 94: 887–897.Murgraff V, Parrott A, Bennett, P (1999) Risky single-occasion drinking amongst young people

– definition, correlates, policy, and intervention: a broad overview of research findings.Alcohol & Alcoholism 34: 3–14.

National Centre for Social Research (1998) Health Survey for England: The Health of YoungPeople ‘95–97, A Survey Carried Out on Behalf of the Department of Health. London: HMSO.

Office for National Statistics (ONS) (1998) General Household Survey. London: HMSO.Pernanen K (1991) Alcohol in Human Violence New York: Guilford Press.Raistrick D, Hodgson R, Ritson B, eds. Tackling Alcohol Together. London: Free Association

Press, 1999.Shepherd J, Brickley M (1996) The relationship between alcohol intoxication, stressors and

injury in urban violence. British Journal of Criminology 36: 546–566.Stratford N, Roth W (1999) The 1998 Youth Lifestyles Survey, Technical Report. London:

National Centre for Social Research.Webb E, Ashton CH, Kelly P, Karnali F (1996) Alcohol and drug use in UK university students.

Lancet 348: 922–925.Wechsler H, Lee JE, Kuo M, Lee H (2000) College binge drinking in the 1990s: a continuing

problem. Journal of the American College of Health 48: 199–210.Wechsler H, Dowdall G, Davenport A, Rimm EB (1995) A gender-specific measure of binge

drinking among college students. American Journal of Public Health 85: 982–985.

Address correspondence to: Anna Richardson, Drugs and Alcohol Research,Home Office, 50 Queen Anne’s Gate, London SW1H 9AT, UK.

16 Richardson, Budd

CBMH 13.1_crc 25/9/03 2:28 pm Page 16