http://informahealthcare.com/ada ISSN: 0095-2990 (print), 1097-9891 (electronic) Am J Drug Alcohol Abuse, 2014; 40(5): 374–379 ! 2014 Informa Healthcare USA, Inc. DOI: 10.3109/00952990.2014.933838
ORIGINAL ARTICLE
The relationship between drinking games and intentions to continue drinking, intentions to drive after drinking, and adverse consequences: Results of a field study John D. Clapp, PhD, FAAHB1, Mark B. Reed, PhD2, and Danielle E. Ruderman, MSW1 1
The Ohio State University College of Social Work, Columbus, Ohio and 2San Diego State University School of Social Work, San Diego, California, USA
Abstract
Keywords
Background: Drinking games have become a nearly universal aspect of excessive drinking on university campuses with 50–62% of college students reporting playing drinking games in the past month. Participation in drinking games has been correlated with numerous negative consequences and increased consumption of alcohol. Objectives: The present study addresses the influence of drinking games on three drinking-related outcomes: problems experienced the night of the drinking event, the intent to keep drinking, and the intent to drive after drinking. Methods: The data collected for the present study were part of a study testing environmental influences of drinking behaviors of young adults. A total of 226 randomly selected parties (representing 1725 partygoers) were selected for study inclusion. Three multilevel logistic regression models tested the relationship between drinking games and the three drinkingrelated outcomes. Results: Participants who reported playing drinking games were 1.58 times more likely to report continued drinking intentions than participants who did not play drinking games. If drinking games were observed at a party, participants were 2.38 times more likely to plan to drive while intoxicated. Additionally, participants who reported playing drinking games were 1.59 times more likely to report experiencing a drinking-related problem than participants who did not play drinking games. Conclusion: Drinking games have consequences beyond increasing the level of intoxication; they contribute to problematic behavior at individual and environmental levels. Preventing drinking games is warranted.
BrAC, college students, drinking games, drinking plans
Introduction Over the past two decades the etiology of drinking behavior and its attendant problems among college students has become fairly clear. Scholars from various disciplines have studied demographic, psychological, social, and environmental characteristics related to college students’ drinking to better understand how to prevent excessive drinking and related problems. Participation in drinking games is conceptually important as it includes psychological, social, and environmental aspects. Drinking games are widely popular on university campuses (1) and more than half of college students reporting playing drinking games in the past month (2–4). Drinking games are social drinking events whose main goal appears to be reaching intoxication as quickly as possible (3,5). The rules and structure of many drinking games often result in a ‘‘reversal of competence’’ where students repeatedly make more mistakes and drink more alcohol the longer the game is played (6). It has been suggested that drinking games promote Address correspondence to John D. Clapp, College of Social Work, The Ohio State University, 1947 College Road, Columbus, Ohio 43210, USA. E-mail: [email protected]
History Received 8 November 2013 Revised 4 June 2014 Accepted 8 June 2014 Published online 27 August 2014
the idea that negative aspects of alcohol consumption are actually positive (intoxication, social embarrassment, etc.) (7). Participation in drinking games has been correlated with an increased probability of experiencing alcohol-related problems such as hangovers, campus violations, risky sexual behavior, car accidents, interpersonal difficulties, and symptoms of alcohol dependence (1,2,7–10). Further, participation in drinking games might adversely impact lighter and moderate drinkers more than their heavy drinking counterparts (11). In 2011, LaBrie et al. (12) surveyed 2546 students from two west coast universities to determine alcohol risk factors that are associated with drinking games and prepartying (i.e., drinking prior to attending a drinking event). Drinking game participation during prepartying increased the likelihood of ‘‘blacking out’’ which was defined as temporary periods of memory loss that can occur during excessive drinking (12). Intention to continue drinking It is important to note that drinking games can be conceptualized as one aspect of the overall alcohol environment and it is important to consider other aspects of the
A field study on drinking games
DOI: 10.3109/00952990.2014.933838
environment as they relate to heavy drinking (6,13,14). Students are more likely to report heavy drinking at events that contain the following conditions: intoxicated people attending the event, the presence of illicit drug use, events featuring ‘‘bring your own beverage’’ (BYOB), as well as events in which drinking games are being played (13,15). These studies suggest that environmental conditions are associated with overall alcohol consumption. Thus, drinking games have the potential to impact students beyond game players, including partygoers who are simply attending events in which drinking games are taking place. Conceptually, drinking game participation might lead to intentions to continue drinking even after one is intoxicated. In a study conducted in three cities in the UK, individuals with higher blood alcohol contents were more likely to report a desire to continue drinking that evening than those who may have drank less (16). Given the relationship between drinking game participation and heavier alcohol consumption, it is probable that drinking games also contribute to intent to continue drinking net of other predictors. For instance, Reed et al. (17) found that individual and event-level characteristics influenced an individual’s intention to continue drinking after leaving a drinking event. Intention to drive Approximately 3.4 million college students drive under the influence of alcohol every year (18). Based on research by Engs and Hanson (11), drinking game participation was associated with driving after drinking. Although being a medium to heavy drinker is associated with drunk driving compared to those identifying as light drinkers (19,20), the disproportional negative influence of playing drinking games on lighter drinkers might place them at risk for DUI as well. The present study begins to address the association of drinking games with three drinking-related outcomes – problems experienced the night of the drinking event, the intent to keep drinking (after being surveyed in a field study), and the intent to drive after drinking – while considering both individual and environmental characteristics in a multi-level model framework. We anticipate that, net of other predictors, participation in drinking games and/or the presence of drinking games will positively correlate with self-reported problems the night of the drinking event, intentions to continue drinking, and intention to drive intoxicated.
Methods The data for the present study were collected as part of a multi-year study testing environmental influences of the drinking behaviors of young adults. The field methodology used for the present study has been described in detail in two other papers (15,21). The San Diego State University Institutional Review Board approved the use of human subjects for the research described in this paper. Study procedures A total of 113 randomly selected parties (representing a total of 1725 partygoers) were selected for study inclusion. Research staff located parties along a 6.2-mile route
375
surrounding a local public university in Southern California which included large apartment complexes, fraternity/sorority housing, and single-family homes rented by students. Research staff conducted surveys on Thursday, Friday, and Saturday evenings between the hours of 9 and 11 pm each night. Typically, research staff identified seven parties on a survey night; however, not all identified parties were included in data collection. In order to gain access to a party, two members of the research staff would approach a residence and would speak with the party host. The researchers explained the purpose of the study to the host and obtained informed consent. An incentive of a $20 gift card was offered to party hosts. If a host refused to participate, the research team went on to the next party identified. Once inside the party, research staff surveyed a convenience sample of participants and attempted to survey as many partygoers as possible in 30 minutes. On average, nearly 35% of partygoers were surveyed in each party – this was based on approximate head counts of partygoers. A small minority (11.6%) of partygoers refused to participate in the study. While partygoers were being interviewed, two other members of the research staff collected observational data concerning party characteristics. Study measures Participants completed a short paper-and-pencil survey comprised of items measuring demographic characteristics, drinking-related behaviors, alcohol consumption history, and questions about cigarette use. For the present study, we focused on the following measures: participant gender, age, race/ethnicity (White, Latino/Hispanic, Asian, African American, Pacific Islander, Native American, other), student status (student/non-student), fraternity/sorority membership, played drinking games (yes/no), number of prepartying locations, number of heavy drinking episodes (defined as the number of times the participant consumed 5 or more drinks in one setting over a two-week period), intention to continue drinking (yes/no), whether the participant reported that he/she planned to drive from the party, and whether the participant had experienced any problems that evening. We did not define what a ‘‘problem’’ might entail; however, participants could describe the problem. Given the small number of some of the racial/ethnic categories, we collapsed race/ethnicity into a binary variable (white/non-white). All of these measures have been shown to be associated with drinking behaviors in previous empirical work (16,22,23). We also included an objective measure of intoxication by asking participants to give a breath sample using a handheld Breath Alcohol Concentration (BrAC) test unit (CMI Intoxilyzer SD-400; CMI, Inc., Owensboro, KY, USA). BrAC values were not visible to participants or research staff; however, participants with obvious alcohol-related impairment (e.g. slurred speech, stumbling gait, etc.) were excluded from study participation. In order to reduce the effects of mouth alcohol on the BrAC measure, we had participants wait at least 10 min after consuming an alcoholic beverage before they gave the breath sample. For each party, a team of two members of the research staff observed and collected data concerning the characteristics of
376
J. D. Clapp et al.
the parties studied. The inter-rater reliability for these observations was good and averaged 0.77. During the observation, the research staff assessed a myriad of partyrelated characteristics; however, the following observations were considered for the present study: party size (small/large), whether the party was themed (yes/no), whether drinking games were being played (yes/no), and whether there were many people who were intoxicated at the party (yes/no). We have also demonstrated significant associations between these ecological variables and drinking behavior in our previous empirical work (16,22,23). Data analysis Because the individual-level (i.e. level-1) data are nested within party-level (i.e. level-2) data, a multilevel modeling approach (MLM) is necessary for data analysis. We followed a MLM model building strategy as suggested by Tabachnick and Fidell (24). First, we tested an intercepts-only model (null model) and calculated intraclass correlation coefficients (ICCs) to determine whether variability in our dependent measures existed at both the individual and party levels of our data. Because the outcome variables examined in this study were dichotomous, we calculated intraclass correlation coefficients utilizing methods suggested by Hox (25). Interceptsonly models yielding significant parameter estimates indicates the use of MLM to analyze the data is warranted. The next step in this process was to identify significant bivariate predictors (at the individual and party levels) of our three outcome measures. For the last step, we then tested full MLM models including all significant level-1 and level-2 variables from the bivariate analyses. We used HLM for windows (ver. 7.0) to perform the multilevel modeling procedures. For the present study, we ran three multilevel logistic regression models testing the relationship between playing drinking games (level-1) and three outcome variables: intending to continuing to drinking, intending to drive while intoxicated (BrAC 0.08), and reporting experiencing a problem that evening. Binomial models utilizing a Bernoulli distribution analysis routine were used for all three models given that each of the outcome variables were dichotomous. Only participants who indicated consuming alcohol during the evening were selected for inclusion in the study (n ¼ 1568 individuals and 113 parties; 91% of all study participants). Furthermore, participant records with missing data were excluded listwise in all MLM models. For the models testing the relationship between playing drinking games and intending to continue drinking and playing drinking games and intending to drive while intoxicated, the following level-1 (individual) covariates were assessed: participant gender, age, race/ethnicity, student status, fraternity/sorority membership, number of heavy episodic drinking (HED) occasions, number of prepartying locations, BrAC, and time. We included time to control for the fact that drinking behaviors would likely change as a function of time of night. For the model testing the relationship between playing drinking games and experiencing a problem, we included demographic variables (gender, age, race/ethnicity, student status, fraternity/sorority membership) as well as other level-1 covariates that are assumed to
Am J Drug Alcohol Abuse, 2014; 40(5): 374–379
Table 1. Participant demographics, drinking characteristic and party characteristics. Participant characteristics Gender Female Race Non-white Age Mean (SD) Student status Student Fraternity/sorority Yes Played drinking games Yes Number of prepartying locations Mean (SD) HED episodes Mean (SD) BrAC Mean (SD) Intend to keep drinking Yes Intend to drive intoxicated (0.08) Yes Experienced problems Yes Party characteristics Size of party Small Large Drinking games Yes Many people intoxicated Yes Theme party Yes
Percent (%)
n
40.51
627
22.3
345
20.21 (2.02)
1568
89.2
1386
18.7
293
33.2
496
0.93 (1.18)
1568
4.87 (4.37)
1490
0.083 (0.062)
1498
88.2
1306
3.1
48
8.1
122
50.4 49.6
57 56
24.8
28
77.9
88
13.3
18
be temporally related to the outcome variable: number of predrinking locations, BrAC, and time. For all three models, we included the following level-2 (party characteristics) variables: size of party, people intoxicated, theme party, and drinking games being played.
Results Participant demographics, drinking behaviors, and party characteristics are shown in Table 1. More than half of the participants were male while more than three-quarters were white. Close to 90% of study participants were college students and close to 20% self-identified as a member of a fraternity or a sorority. The average age of the participants was slightly older than 20 (M ¼ 20.21, SD ¼ 2.02). A large majority of the participants intended to keep drinking after leaving the party while the average BrAC of participants was 0.083 (SD ¼ 0.062). The average number of heavy drinking episodes in the past two weeks was 4.87 (SD ¼ 4.37) and participants reported an average of less than one prepartying drinking location (M ¼ 0.93, SD ¼ 1.18). Slightly more than 3% of the sample planned to drive while intoxicated (BrAC 0.08) and nearly one in 10 (8.1%) reported experiencing a problem during the evening. In terms of party characteristics, about half of the parties were large, more than 10% had a theme, nearly 80% had many intoxicated partygoers, and almost 1 in 4 had people playing drinking
A field study on drinking games
DOI: 10.3109/00952990.2014.933838
377
Table 2. Results of multilevel logistic regression testing the relationship between playing drinking games and intending to continue drinking.
Table 3. Results of multilevel logistic regression testing the relationship between playing drinking games and driving while intoxicated.
Variable
Variable
Level 1 Gender Female Race/ethnicity White Played drinking games Yes No. of HED occasions No. of prepartying locations BrAC Time of night
Coefficient Odds ratio p Value
95% CI
0.24
0.79
0.22 0.54–1.16
0.31
1.36
0.146 0.90–2.08
0.46 0.14 0.21 0.07 0.57
1.58 1.15 1.24 1.07 0.57
0.032 50.001 0.038 50.001 0.002
1.04–2.41 1.08–1.22 1.01–1.52 1.03–1.11 0.40–0.081
Coefficient Odds ratio p Value
Level 1 Gender Female Student status Student No. of HED occasions Level 2 Observed drinking games Yes
95% CI
0.42
0.66
0.93 0.12
0.39 1.13
0.008 0.20–0.79 50.001 1.05–1.20
0.87
2.38
0.003 1.36–4.18
0.25
0.32–1.34
Table 4. Results of multilevel logistic regression testing the relationship between playing drinking games and experiencing problems.
games. Themes were primarily sexual in nature and included ‘‘Pimp and Ho’’ and ‘‘Victoria Secret’’ parties. Intentions to continue drinking As noted above, the first step in this model building process was to test an intercepts-only model to determine if significant variability existed in the intentions to continue drinking outcome measure at both the party and individual levels. The intercepts-only model was significant and yielded an ICC of 0.121, indicating that 12.1% of the variance in the outcome measure could be explained at the party level while 87.9% could be explained at the individual level. The bivariate analyses for the intending to continue to drink outcome variable yielded a set of seven statistically significant level-1 predictors: gender, race/ethnicity, number of HED occasions, number of prepartying locations, BrAC, time, and playing drinking games (results not shown). None of the level-2 (party-level) variables were significantly associated with intending to continue to drink. The results of the multivariate model are shown in Table 2. Participant gender and race/ ethnicity were not significantly associated with intending to continue drinking in the multivariate model. In contrast, the number of HED occasions in the past two weeks as well as the number of prepartying locations were both positively associated with the increased odds of intending to continue to drink. Participants’ level of intoxication (BrAC) was also predictive of higher odds of continued drinking intentions while time was negatively associated. The relationship between playing drinking games and reporting intentions to continue drinking was also significant net of the contributions of model covariates. Specifically, participants who reported playing drinking games were 1.58 times more likely to report continued drinking intentions than participants who were not playing drinking games (Table 2). Planning to drive intoxicated The intercepts-only model was not significant but yielded an ICC of 0.065, indicating that 6.5% of the variance in the outcome measure could be explained at the party level while 93.5% could be explained at the individual level. The results of the bivariate analyses (not shown) for the planning to drive while intoxicated outcome variable yielded a set of four significant level-1 predictors: participant gender, student status, number of HED occasions, and
Variable Level 1 Gender Female Played drinking games Yes No. of prepartying locations Time of night
Coefficient Odds ratio p Value 95% CI
0.5 0.46 0.13 0.41
0.6
0.012 0.41–0.89
1.59 1.14 1.51
0.029 1.05–2.41 0.047 1.00–1.30 0.002 1.17–1.95
BrAC. Drinking games were not significantly associated with the outcome variable in the bivariate analysis and thus was not retained for the multivariate model. At the party level, the observation of drinking games being played was significantly associated with planning to drive intoxicated. Results of the multivariate model are shown in Table 3. Gender was not significant in the multivariate model; however, being a student reduced the odds of planning to drive while intoxicated. As expected, increasing occasions of HED were predictive of a higher odds of planning to drive while intoxicated. Lastly, after controlling for all level-1 covariates, participants who observed games being played at a party were 2.38 times more likely to drive while intoxicated than participants who did not observe drinking games. Problems experienced The intercepts-only model was significant and yielded an ICC of 0.093, indicating that 9.3% of the variance in the experienced problems tonight outcome measure could be explained at the party level while 90.7% could be explained at the individual level. Results of the bivariate analyses yielded a set of four significant predictors: gender, number of prepartying locations, playing drinking games, and time. There were no level-2 variables significantly associated with planning to drive intoxicated in the bivariate results. As shown in Table 4, women were less likely to report experiencing a problem while the number of prepartying locations and time were both significantly predictive of increased odds of experiencing a problem. Net of the influence of these predictor variables, participants who reported playing drinking games were 2.25 times more likely to report experiencing a problem than participants who did not play drinking games.
378
J. D. Clapp et al.
Discussion This study examined the relationship between playing drinking games and partygoers’ event-level self-reported problems and their plans to continue drinking and to drive intoxicated. This study begins to examine the association between drinking games and problem outcomes. The study is, to our knowledge, the first to show an environmental association between drinking games and a problem outcome. That is, the presence of drinking games itself is associated with problem outcomes for individuals regardless of whether they actually participated in the games. The study is somewhat unique in the current literature in that the data were collected in actual party settings and included breath measures of blood alcohol. This event-level design allows for closer temporal measurement of relevant variables, provides a ‘‘hard’’ measure of level of intoxication, and allows for the assessment of both event and individual level factors as they relate to dinking games. One unique aspect of the study was our ability to include an environmental measure of party games being played in our models. These aspects of the study can be viewed as strengths of the overall design, which contribute to the ecological validity of our findings. Overall, our sample of partygoers were mostly underage college students who drank fairly heavily both in the two weeks prior to being surveyed and at the event where they were interviewed. At the multivariate level, participating in drinking games predicted both intentions to continue drinking and reporting experiencing problems, net of other predictors, most notably including intoxication (BrAC level). In each of these models, the multivariate coefficient for drinking games was the strongest predictor of the outcome of interest. For intentions to drive with a BrAC of 0.08 or greater, the presence of drinking games at the party level (as opposed to playing drinking games) was a significant and large contributor to the model. Thus, for two of the three criterion variables, drinking games at the individual level had large and independent relationships with negative outcomes. For the third, drinking games at the environmental level was strongly and independently associated with the intent to drive intoxicated variable. One common challenge associated with studying drinking games within the overall context of drinking events is determining when gaming occurs chronologically in the course of what are often dynamic, multi-location events. The retrospective and cross-sectional nature of most studies examining drinking behavior makes identifying such sequencing difficult. Although the extant literature clearly shows a relationship between drinking games, heavy drinking, and problems, it is unclear what exact role such games have in the overall sequence of a drinking event. This study begins to look at drinking games in context. However, more work needs to be done to examine games within an overall drinking event. Our results suggest that drinking games are not simply a mechanism to increase intoxication – they are associated with problematic behavior at the individual and environmental levels. Although many prevention approaches common to college campuses focus on reducing drinking to intoxication and overall alcohol consumption, controlling drinking games presents a unique challenge. At the individual level,
Am J Drug Alcohol Abuse, 2014; 40(5): 374–379
Zamboanga et al. (9) suggested expectancy challenges might be a viable option to prevent drinking game participation. Based on our study, an environmental approach might also be warranted to complement individual-level approaches. Potential environmental approaches might include social host policies, which hold party hosts accountable for over-serving guests at their home. Such approaches have been widely adopted to prevent underage drinking at the community level. Such policies might be coupled with host training that either discourages drinking games or alters them (e.g. beer pong played with water, etc.). Framing drinking game participation as a risk-factor for DUI might facilitate the political will for developing social host policies targeting drinking game participation. Future research is needed to further test such interventions. The present study is not without limitations. First, although the ecological validity of the study is arguably high, the external validity of the study is limited to a single large urban campus. Drinking games might function differently – especially in relation to drinking and driving outcomes – on smaller purely residential campuses. Second, our measure of problems experienced is general and not directly related to drinking games. In event-level designs like ours, a next day follow-up, or ecological momentary assessments coupled with a series of potential problems, would be beneficial (26). Finally, further studying how drinking game participation might change intentions to keep drinking would require a design in which intentions to drink for the evening were identified prior to any drinking behavior and then assessed throughout the event. Future research might address these issues to better understand the unique role that game participation plays in alcohol consumption and alcoholrelated problems. In conclusion, this study adds to the growing body of work concerning the risks associated with drinking games and how drinking game participation may increase the likelihood of continued harmful drinking, driving under the influence, and drinking-related problems.
Acknowledgements This study was funded in part by a grant to Dr. Clapp from the National Institutes of Health (NIAAA, RO1 AA013968).
Declaration of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this paper.
References 1. Cameron JM, Heidelberg N, Simmons L, Lyle SB, Mitra-Varma K, Correia C. Drinking game participation among undergraduate students attending National Alcohol Screening Day. J Am College Health 2010;58:499–506. 2. Borsari B. Drinking games in the college environment: a review. J Alcohol Drug Educat 2004;48:29–51. 3. Ahern NR, Sole ML. Drinking games and college students. J Psychosoc Nursing 2010;48:17–20. 4. LaBrie JW, Ehret PJ, Hummer JF. Are they all the same? An exploratory, categorical analysis of drinking game types. Addict Behav 2013;38:2133–2139. 5. Correia CJ, Cameron JM. Development of a stimulated drinking game procedure to study risky alcohol use. Experim Clin Psychopharmacol 2010;18:322–328.
A field study on drinking games
DOI: 10.3109/00952990.2014.933838
6. Green TA, Grider SA. Reversal of competence in college drinking games. Play Culture 1990;3:117–132. 7. Nagoshi CT, Wood MD, Cote CC, Abbit SM. College drinking game participation within the context of other predictors of alcohol use and problems. Psychol Addict Behav 1994;8:203–213. 8. Cameron JM, Leon MR, Correia CJ. Extension of the stimulated drinking game procedure to multiple drinking games. Experim Clin Psychopharmacol 2011;19:295–302. 9. Zamboanga BL, Schwartz SJ, Ham LS, Borsari B, Van Tyne K. Alcohol expectancies, pregaming, drinking games, and hazardous alcohol use in a multiethnic sample of college students. Cognit Ther Res 2010;34:124–133. 10. Johnson TJ, Sheets VL. Measuring college students’ motives for playing drinking games. Psychol Addict Behav 2004;18:91–99. 11. Engs RC, Hanson DJ. Drinking games and problems related to drinking among moderate and heavy drinkers. Psycholog Rep 1993; 73:115–120. 12. LaBrie J, Hummer J, Kenney S, Lac A, Pederson E. Identifying factors that increase the likelihood for alcohol-induced blackouts in the prepartying context. Substance Use Misuse 2011;46:992–1002. 13. Clapp JD, Reed MB, Holmes MR, Lange JE, Voss RB. Drunk in public, drunk in private: the relationship between college students, drinking environments and alcohol consumption. Am J Drug Alcohol Abuse 2006;32:275–285. 14. Clapp JD, Min JW, Shillington AM, Reed MB, Croff JK. Person and environment predictors of blood alcohol concentrations: a multi-level study of college parties. Alcoholism: Clin Experim Res 2008;32:100–107. 15. Clapp JD, Shillington AM. Environmental predictors of heavy episodic drinking events. Am J Drug Alcohol Abuse 2001;27: 310–313. 16. Bellis MA, Hughes K, Quigg Z, Morleo M, Jarman I, Lisboa P. Cross-sectional measures and modelled estimates of blood alcohol
17. 18.
19. 20. 21.
22. 23.
24. 25. 26.
379
levels in UK nightlife and their relationships with drinking behaviours and observed signs of inebriation. Substance Abuse Treat, Prevent, Policy 2010;5:1–8. Reed MB, Clapp JD, Martell B, Hidalgo-Sotelo A. The relationship between group size, intoxication and continuing to drink after bar attendance. Drug Alcohol Depend 2013;133:198–203. Hingson RW, Zha W, Weitzman ER. Magnitude of and trends in alcohol-related mortality and morbidity among U.S. college students aged 18–24 1998–2005. J Stud Alcohol Drugs (Suppl.) 2009;16:12–20. LaBrie JW, Kenney SR, Mirza R, Lac A. Identifying factors that increase the likelihood of driving after drinking among college students. Accident Analysis Prevent 2011;43:1371–1377. Clapp JD, Johnson M, Voas RB, Lange JE, Shillington A, Russel C. Reducing DUI among US college students: results of an environmental prevention trial. Addiction 2005;100:327–334. Clapp JD, Holmes MR, Reed MB, Shillington AM, Freisthler B. Measuring college students’ alcohol consumption in natural drinking environments: field methodologies for bars and parties. Evaluation Rev 2007;31:469–489. Clapp JD, Lange J, Min JW, Shillington A, Johnson M, Voas R. Two studies examining environmental predictors of heavy drinking by college students. Prevent Sci 2003;4:99–108. Clapp JD, Min JW, Trim R, Reed MB, Lange JE, Shillington AM, Croff J. Predictors of error in estimates of blood alcohol concentrations: a replication. J Alcohol Drug Stud 2009;70: 683–688. Tabachnick BG, Fidell LS. Using multivariate statistics, 5th ed. Boston, MA: Allyn & Bacon; 2007. Hox J. Multilevel analysis: techniques and applications. Mahwah, NJ: Lawrence Erlbaum Associates, Inc.; 2002. Stone AA, Shiffman S. Ecological momentary assessment in behavioral medicine. Annals Behav Med 1994;16:199–202.
Copyright of American Journal of Drug & Alcohol Abuse is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
ORIGINAL ARTICLE
The relationship between drinking games and intentions to continue drinking, intentions to drive after drinking, and adverse consequences: Results of a field study John D. Clapp, PhD, FAAHB1, Mark B. Reed, PhD2, and Danielle E. Ruderman, MSW1 1
The Ohio State University College of Social Work, Columbus, Ohio and 2San Diego State University School of Social Work, San Diego, California, USA
Abstract
Keywords
Background: Drinking games have become a nearly universal aspect of excessive drinking on university campuses with 50–62% of college students reporting playing drinking games in the past month. Participation in drinking games has been correlated with numerous negative consequences and increased consumption of alcohol. Objectives: The present study addresses the influence of drinking games on three drinking-related outcomes: problems experienced the night of the drinking event, the intent to keep drinking, and the intent to drive after drinking. Methods: The data collected for the present study were part of a study testing environmental influences of drinking behaviors of young adults. A total of 226 randomly selected parties (representing 1725 partygoers) were selected for study inclusion. Three multilevel logistic regression models tested the relationship between drinking games and the three drinkingrelated outcomes. Results: Participants who reported playing drinking games were 1.58 times more likely to report continued drinking intentions than participants who did not play drinking games. If drinking games were observed at a party, participants were 2.38 times more likely to plan to drive while intoxicated. Additionally, participants who reported playing drinking games were 1.59 times more likely to report experiencing a drinking-related problem than participants who did not play drinking games. Conclusion: Drinking games have consequences beyond increasing the level of intoxication; they contribute to problematic behavior at individual and environmental levels. Preventing drinking games is warranted.
BrAC, college students, drinking games, drinking plans
Introduction Over the past two decades the etiology of drinking behavior and its attendant problems among college students has become fairly clear. Scholars from various disciplines have studied demographic, psychological, social, and environmental characteristics related to college students’ drinking to better understand how to prevent excessive drinking and related problems. Participation in drinking games is conceptually important as it includes psychological, social, and environmental aspects. Drinking games are widely popular on university campuses (1) and more than half of college students reporting playing drinking games in the past month (2–4). Drinking games are social drinking events whose main goal appears to be reaching intoxication as quickly as possible (3,5). The rules and structure of many drinking games often result in a ‘‘reversal of competence’’ where students repeatedly make more mistakes and drink more alcohol the longer the game is played (6). It has been suggested that drinking games promote Address correspondence to John D. Clapp, College of Social Work, The Ohio State University, 1947 College Road, Columbus, Ohio 43210, USA. E-mail: [email protected]
History Received 8 November 2013 Revised 4 June 2014 Accepted 8 June 2014 Published online 27 August 2014
the idea that negative aspects of alcohol consumption are actually positive (intoxication, social embarrassment, etc.) (7). Participation in drinking games has been correlated with an increased probability of experiencing alcohol-related problems such as hangovers, campus violations, risky sexual behavior, car accidents, interpersonal difficulties, and symptoms of alcohol dependence (1,2,7–10). Further, participation in drinking games might adversely impact lighter and moderate drinkers more than their heavy drinking counterparts (11). In 2011, LaBrie et al. (12) surveyed 2546 students from two west coast universities to determine alcohol risk factors that are associated with drinking games and prepartying (i.e., drinking prior to attending a drinking event). Drinking game participation during prepartying increased the likelihood of ‘‘blacking out’’ which was defined as temporary periods of memory loss that can occur during excessive drinking (12). Intention to continue drinking It is important to note that drinking games can be conceptualized as one aspect of the overall alcohol environment and it is important to consider other aspects of the
A field study on drinking games
DOI: 10.3109/00952990.2014.933838
environment as they relate to heavy drinking (6,13,14). Students are more likely to report heavy drinking at events that contain the following conditions: intoxicated people attending the event, the presence of illicit drug use, events featuring ‘‘bring your own beverage’’ (BYOB), as well as events in which drinking games are being played (13,15). These studies suggest that environmental conditions are associated with overall alcohol consumption. Thus, drinking games have the potential to impact students beyond game players, including partygoers who are simply attending events in which drinking games are taking place. Conceptually, drinking game participation might lead to intentions to continue drinking even after one is intoxicated. In a study conducted in three cities in the UK, individuals with higher blood alcohol contents were more likely to report a desire to continue drinking that evening than those who may have drank less (16). Given the relationship between drinking game participation and heavier alcohol consumption, it is probable that drinking games also contribute to intent to continue drinking net of other predictors. For instance, Reed et al. (17) found that individual and event-level characteristics influenced an individual’s intention to continue drinking after leaving a drinking event. Intention to drive Approximately 3.4 million college students drive under the influence of alcohol every year (18). Based on research by Engs and Hanson (11), drinking game participation was associated with driving after drinking. Although being a medium to heavy drinker is associated with drunk driving compared to those identifying as light drinkers (19,20), the disproportional negative influence of playing drinking games on lighter drinkers might place them at risk for DUI as well. The present study begins to address the association of drinking games with three drinking-related outcomes – problems experienced the night of the drinking event, the intent to keep drinking (after being surveyed in a field study), and the intent to drive after drinking – while considering both individual and environmental characteristics in a multi-level model framework. We anticipate that, net of other predictors, participation in drinking games and/or the presence of drinking games will positively correlate with self-reported problems the night of the drinking event, intentions to continue drinking, and intention to drive intoxicated.
Methods The data for the present study were collected as part of a multi-year study testing environmental influences of the drinking behaviors of young adults. The field methodology used for the present study has been described in detail in two other papers (15,21). The San Diego State University Institutional Review Board approved the use of human subjects for the research described in this paper. Study procedures A total of 113 randomly selected parties (representing a total of 1725 partygoers) were selected for study inclusion. Research staff located parties along a 6.2-mile route
375
surrounding a local public university in Southern California which included large apartment complexes, fraternity/sorority housing, and single-family homes rented by students. Research staff conducted surveys on Thursday, Friday, and Saturday evenings between the hours of 9 and 11 pm each night. Typically, research staff identified seven parties on a survey night; however, not all identified parties were included in data collection. In order to gain access to a party, two members of the research staff would approach a residence and would speak with the party host. The researchers explained the purpose of the study to the host and obtained informed consent. An incentive of a $20 gift card was offered to party hosts. If a host refused to participate, the research team went on to the next party identified. Once inside the party, research staff surveyed a convenience sample of participants and attempted to survey as many partygoers as possible in 30 minutes. On average, nearly 35% of partygoers were surveyed in each party – this was based on approximate head counts of partygoers. A small minority (11.6%) of partygoers refused to participate in the study. While partygoers were being interviewed, two other members of the research staff collected observational data concerning party characteristics. Study measures Participants completed a short paper-and-pencil survey comprised of items measuring demographic characteristics, drinking-related behaviors, alcohol consumption history, and questions about cigarette use. For the present study, we focused on the following measures: participant gender, age, race/ethnicity (White, Latino/Hispanic, Asian, African American, Pacific Islander, Native American, other), student status (student/non-student), fraternity/sorority membership, played drinking games (yes/no), number of prepartying locations, number of heavy drinking episodes (defined as the number of times the participant consumed 5 or more drinks in one setting over a two-week period), intention to continue drinking (yes/no), whether the participant reported that he/she planned to drive from the party, and whether the participant had experienced any problems that evening. We did not define what a ‘‘problem’’ might entail; however, participants could describe the problem. Given the small number of some of the racial/ethnic categories, we collapsed race/ethnicity into a binary variable (white/non-white). All of these measures have been shown to be associated with drinking behaviors in previous empirical work (16,22,23). We also included an objective measure of intoxication by asking participants to give a breath sample using a handheld Breath Alcohol Concentration (BrAC) test unit (CMI Intoxilyzer SD-400; CMI, Inc., Owensboro, KY, USA). BrAC values were not visible to participants or research staff; however, participants with obvious alcohol-related impairment (e.g. slurred speech, stumbling gait, etc.) were excluded from study participation. In order to reduce the effects of mouth alcohol on the BrAC measure, we had participants wait at least 10 min after consuming an alcoholic beverage before they gave the breath sample. For each party, a team of two members of the research staff observed and collected data concerning the characteristics of
376
J. D. Clapp et al.
the parties studied. The inter-rater reliability for these observations was good and averaged 0.77. During the observation, the research staff assessed a myriad of partyrelated characteristics; however, the following observations were considered for the present study: party size (small/large), whether the party was themed (yes/no), whether drinking games were being played (yes/no), and whether there were many people who were intoxicated at the party (yes/no). We have also demonstrated significant associations between these ecological variables and drinking behavior in our previous empirical work (16,22,23). Data analysis Because the individual-level (i.e. level-1) data are nested within party-level (i.e. level-2) data, a multilevel modeling approach (MLM) is necessary for data analysis. We followed a MLM model building strategy as suggested by Tabachnick and Fidell (24). First, we tested an intercepts-only model (null model) and calculated intraclass correlation coefficients (ICCs) to determine whether variability in our dependent measures existed at both the individual and party levels of our data. Because the outcome variables examined in this study were dichotomous, we calculated intraclass correlation coefficients utilizing methods suggested by Hox (25). Interceptsonly models yielding significant parameter estimates indicates the use of MLM to analyze the data is warranted. The next step in this process was to identify significant bivariate predictors (at the individual and party levels) of our three outcome measures. For the last step, we then tested full MLM models including all significant level-1 and level-2 variables from the bivariate analyses. We used HLM for windows (ver. 7.0) to perform the multilevel modeling procedures. For the present study, we ran three multilevel logistic regression models testing the relationship between playing drinking games (level-1) and three outcome variables: intending to continuing to drinking, intending to drive while intoxicated (BrAC 0.08), and reporting experiencing a problem that evening. Binomial models utilizing a Bernoulli distribution analysis routine were used for all three models given that each of the outcome variables were dichotomous. Only participants who indicated consuming alcohol during the evening were selected for inclusion in the study (n ¼ 1568 individuals and 113 parties; 91% of all study participants). Furthermore, participant records with missing data were excluded listwise in all MLM models. For the models testing the relationship between playing drinking games and intending to continue drinking and playing drinking games and intending to drive while intoxicated, the following level-1 (individual) covariates were assessed: participant gender, age, race/ethnicity, student status, fraternity/sorority membership, number of heavy episodic drinking (HED) occasions, number of prepartying locations, BrAC, and time. We included time to control for the fact that drinking behaviors would likely change as a function of time of night. For the model testing the relationship between playing drinking games and experiencing a problem, we included demographic variables (gender, age, race/ethnicity, student status, fraternity/sorority membership) as well as other level-1 covariates that are assumed to
Am J Drug Alcohol Abuse, 2014; 40(5): 374–379
Table 1. Participant demographics, drinking characteristic and party characteristics. Participant characteristics Gender Female Race Non-white Age Mean (SD) Student status Student Fraternity/sorority Yes Played drinking games Yes Number of prepartying locations Mean (SD) HED episodes Mean (SD) BrAC Mean (SD) Intend to keep drinking Yes Intend to drive intoxicated (0.08) Yes Experienced problems Yes Party characteristics Size of party Small Large Drinking games Yes Many people intoxicated Yes Theme party Yes
Percent (%)
n
40.51
627
22.3
345
20.21 (2.02)
1568
89.2
1386
18.7
293
33.2
496
0.93 (1.18)
1568
4.87 (4.37)
1490
0.083 (0.062)
1498
88.2
1306
3.1
48
8.1
122
50.4 49.6
57 56
24.8
28
77.9
88
13.3
18
be temporally related to the outcome variable: number of predrinking locations, BrAC, and time. For all three models, we included the following level-2 (party characteristics) variables: size of party, people intoxicated, theme party, and drinking games being played.
Results Participant demographics, drinking behaviors, and party characteristics are shown in Table 1. More than half of the participants were male while more than three-quarters were white. Close to 90% of study participants were college students and close to 20% self-identified as a member of a fraternity or a sorority. The average age of the participants was slightly older than 20 (M ¼ 20.21, SD ¼ 2.02). A large majority of the participants intended to keep drinking after leaving the party while the average BrAC of participants was 0.083 (SD ¼ 0.062). The average number of heavy drinking episodes in the past two weeks was 4.87 (SD ¼ 4.37) and participants reported an average of less than one prepartying drinking location (M ¼ 0.93, SD ¼ 1.18). Slightly more than 3% of the sample planned to drive while intoxicated (BrAC 0.08) and nearly one in 10 (8.1%) reported experiencing a problem during the evening. In terms of party characteristics, about half of the parties were large, more than 10% had a theme, nearly 80% had many intoxicated partygoers, and almost 1 in 4 had people playing drinking
A field study on drinking games
DOI: 10.3109/00952990.2014.933838
377
Table 2. Results of multilevel logistic regression testing the relationship between playing drinking games and intending to continue drinking.
Table 3. Results of multilevel logistic regression testing the relationship between playing drinking games and driving while intoxicated.
Variable
Variable
Level 1 Gender Female Race/ethnicity White Played drinking games Yes No. of HED occasions No. of prepartying locations BrAC Time of night
Coefficient Odds ratio p Value
95% CI
0.24
0.79
0.22 0.54–1.16
0.31
1.36
0.146 0.90–2.08
0.46 0.14 0.21 0.07 0.57
1.58 1.15 1.24 1.07 0.57
0.032 50.001 0.038 50.001 0.002
1.04–2.41 1.08–1.22 1.01–1.52 1.03–1.11 0.40–0.081
Coefficient Odds ratio p Value
Level 1 Gender Female Student status Student No. of HED occasions Level 2 Observed drinking games Yes
95% CI
0.42
0.66
0.93 0.12
0.39 1.13
0.008 0.20–0.79 50.001 1.05–1.20
0.87
2.38
0.003 1.36–4.18
0.25
0.32–1.34
Table 4. Results of multilevel logistic regression testing the relationship between playing drinking games and experiencing problems.
games. Themes were primarily sexual in nature and included ‘‘Pimp and Ho’’ and ‘‘Victoria Secret’’ parties. Intentions to continue drinking As noted above, the first step in this model building process was to test an intercepts-only model to determine if significant variability existed in the intentions to continue drinking outcome measure at both the party and individual levels. The intercepts-only model was significant and yielded an ICC of 0.121, indicating that 12.1% of the variance in the outcome measure could be explained at the party level while 87.9% could be explained at the individual level. The bivariate analyses for the intending to continue to drink outcome variable yielded a set of seven statistically significant level-1 predictors: gender, race/ethnicity, number of HED occasions, number of prepartying locations, BrAC, time, and playing drinking games (results not shown). None of the level-2 (party-level) variables were significantly associated with intending to continue to drink. The results of the multivariate model are shown in Table 2. Participant gender and race/ ethnicity were not significantly associated with intending to continue drinking in the multivariate model. In contrast, the number of HED occasions in the past two weeks as well as the number of prepartying locations were both positively associated with the increased odds of intending to continue to drink. Participants’ level of intoxication (BrAC) was also predictive of higher odds of continued drinking intentions while time was negatively associated. The relationship between playing drinking games and reporting intentions to continue drinking was also significant net of the contributions of model covariates. Specifically, participants who reported playing drinking games were 1.58 times more likely to report continued drinking intentions than participants who were not playing drinking games (Table 2). Planning to drive intoxicated The intercepts-only model was not significant but yielded an ICC of 0.065, indicating that 6.5% of the variance in the outcome measure could be explained at the party level while 93.5% could be explained at the individual level. The results of the bivariate analyses (not shown) for the planning to drive while intoxicated outcome variable yielded a set of four significant level-1 predictors: participant gender, student status, number of HED occasions, and
Variable Level 1 Gender Female Played drinking games Yes No. of prepartying locations Time of night
Coefficient Odds ratio p Value 95% CI
0.5 0.46 0.13 0.41
0.6
0.012 0.41–0.89
1.59 1.14 1.51
0.029 1.05–2.41 0.047 1.00–1.30 0.002 1.17–1.95
BrAC. Drinking games were not significantly associated with the outcome variable in the bivariate analysis and thus was not retained for the multivariate model. At the party level, the observation of drinking games being played was significantly associated with planning to drive intoxicated. Results of the multivariate model are shown in Table 3. Gender was not significant in the multivariate model; however, being a student reduced the odds of planning to drive while intoxicated. As expected, increasing occasions of HED were predictive of a higher odds of planning to drive while intoxicated. Lastly, after controlling for all level-1 covariates, participants who observed games being played at a party were 2.38 times more likely to drive while intoxicated than participants who did not observe drinking games. Problems experienced The intercepts-only model was significant and yielded an ICC of 0.093, indicating that 9.3% of the variance in the experienced problems tonight outcome measure could be explained at the party level while 90.7% could be explained at the individual level. Results of the bivariate analyses yielded a set of four significant predictors: gender, number of prepartying locations, playing drinking games, and time. There were no level-2 variables significantly associated with planning to drive intoxicated in the bivariate results. As shown in Table 4, women were less likely to report experiencing a problem while the number of prepartying locations and time were both significantly predictive of increased odds of experiencing a problem. Net of the influence of these predictor variables, participants who reported playing drinking games were 2.25 times more likely to report experiencing a problem than participants who did not play drinking games.
378
J. D. Clapp et al.
Discussion This study examined the relationship between playing drinking games and partygoers’ event-level self-reported problems and their plans to continue drinking and to drive intoxicated. This study begins to examine the association between drinking games and problem outcomes. The study is, to our knowledge, the first to show an environmental association between drinking games and a problem outcome. That is, the presence of drinking games itself is associated with problem outcomes for individuals regardless of whether they actually participated in the games. The study is somewhat unique in the current literature in that the data were collected in actual party settings and included breath measures of blood alcohol. This event-level design allows for closer temporal measurement of relevant variables, provides a ‘‘hard’’ measure of level of intoxication, and allows for the assessment of both event and individual level factors as they relate to dinking games. One unique aspect of the study was our ability to include an environmental measure of party games being played in our models. These aspects of the study can be viewed as strengths of the overall design, which contribute to the ecological validity of our findings. Overall, our sample of partygoers were mostly underage college students who drank fairly heavily both in the two weeks prior to being surveyed and at the event where they were interviewed. At the multivariate level, participating in drinking games predicted both intentions to continue drinking and reporting experiencing problems, net of other predictors, most notably including intoxication (BrAC level). In each of these models, the multivariate coefficient for drinking games was the strongest predictor of the outcome of interest. For intentions to drive with a BrAC of 0.08 or greater, the presence of drinking games at the party level (as opposed to playing drinking games) was a significant and large contributor to the model. Thus, for two of the three criterion variables, drinking games at the individual level had large and independent relationships with negative outcomes. For the third, drinking games at the environmental level was strongly and independently associated with the intent to drive intoxicated variable. One common challenge associated with studying drinking games within the overall context of drinking events is determining when gaming occurs chronologically in the course of what are often dynamic, multi-location events. The retrospective and cross-sectional nature of most studies examining drinking behavior makes identifying such sequencing difficult. Although the extant literature clearly shows a relationship between drinking games, heavy drinking, and problems, it is unclear what exact role such games have in the overall sequence of a drinking event. This study begins to look at drinking games in context. However, more work needs to be done to examine games within an overall drinking event. Our results suggest that drinking games are not simply a mechanism to increase intoxication – they are associated with problematic behavior at the individual and environmental levels. Although many prevention approaches common to college campuses focus on reducing drinking to intoxication and overall alcohol consumption, controlling drinking games presents a unique challenge. At the individual level,
Am J Drug Alcohol Abuse, 2014; 40(5): 374–379
Zamboanga et al. (9) suggested expectancy challenges might be a viable option to prevent drinking game participation. Based on our study, an environmental approach might also be warranted to complement individual-level approaches. Potential environmental approaches might include social host policies, which hold party hosts accountable for over-serving guests at their home. Such approaches have been widely adopted to prevent underage drinking at the community level. Such policies might be coupled with host training that either discourages drinking games or alters them (e.g. beer pong played with water, etc.). Framing drinking game participation as a risk-factor for DUI might facilitate the political will for developing social host policies targeting drinking game participation. Future research is needed to further test such interventions. The present study is not without limitations. First, although the ecological validity of the study is arguably high, the external validity of the study is limited to a single large urban campus. Drinking games might function differently – especially in relation to drinking and driving outcomes – on smaller purely residential campuses. Second, our measure of problems experienced is general and not directly related to drinking games. In event-level designs like ours, a next day follow-up, or ecological momentary assessments coupled with a series of potential problems, would be beneficial (26). Finally, further studying how drinking game participation might change intentions to keep drinking would require a design in which intentions to drink for the evening were identified prior to any drinking behavior and then assessed throughout the event. Future research might address these issues to better understand the unique role that game participation plays in alcohol consumption and alcoholrelated problems. In conclusion, this study adds to the growing body of work concerning the risks associated with drinking games and how drinking game participation may increase the likelihood of continued harmful drinking, driving under the influence, and drinking-related problems.
Acknowledgements This study was funded in part by a grant to Dr. Clapp from the National Institutes of Health (NIAAA, RO1 AA013968).
Declaration of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this paper.
References 1. Cameron JM, Heidelberg N, Simmons L, Lyle SB, Mitra-Varma K, Correia C. Drinking game participation among undergraduate students attending National Alcohol Screening Day. J Am College Health 2010;58:499–506. 2. Borsari B. Drinking games in the college environment: a review. J Alcohol Drug Educat 2004;48:29–51. 3. Ahern NR, Sole ML. Drinking games and college students. J Psychosoc Nursing 2010;48:17–20. 4. LaBrie JW, Ehret PJ, Hummer JF. Are they all the same? An exploratory, categorical analysis of drinking game types. Addict Behav 2013;38:2133–2139. 5. Correia CJ, Cameron JM. Development of a stimulated drinking game procedure to study risky alcohol use. Experim Clin Psychopharmacol 2010;18:322–328.
A field study on drinking games
DOI: 10.3109/00952990.2014.933838
6. Green TA, Grider SA. Reversal of competence in college drinking games. Play Culture 1990;3:117–132. 7. Nagoshi CT, Wood MD, Cote CC, Abbit SM. College drinking game participation within the context of other predictors of alcohol use and problems. Psychol Addict Behav 1994;8:203–213. 8. Cameron JM, Leon MR, Correia CJ. Extension of the stimulated drinking game procedure to multiple drinking games. Experim Clin Psychopharmacol 2011;19:295–302. 9. Zamboanga BL, Schwartz SJ, Ham LS, Borsari B, Van Tyne K. Alcohol expectancies, pregaming, drinking games, and hazardous alcohol use in a multiethnic sample of college students. Cognit Ther Res 2010;34:124–133. 10. Johnson TJ, Sheets VL. Measuring college students’ motives for playing drinking games. Psychol Addict Behav 2004;18:91–99. 11. Engs RC, Hanson DJ. Drinking games and problems related to drinking among moderate and heavy drinkers. Psycholog Rep 1993; 73:115–120. 12. LaBrie J, Hummer J, Kenney S, Lac A, Pederson E. Identifying factors that increase the likelihood for alcohol-induced blackouts in the prepartying context. Substance Use Misuse 2011;46:992–1002. 13. Clapp JD, Reed MB, Holmes MR, Lange JE, Voss RB. Drunk in public, drunk in private: the relationship between college students, drinking environments and alcohol consumption. Am J Drug Alcohol Abuse 2006;32:275–285. 14. Clapp JD, Min JW, Shillington AM, Reed MB, Croff JK. Person and environment predictors of blood alcohol concentrations: a multi-level study of college parties. Alcoholism: Clin Experim Res 2008;32:100–107. 15. Clapp JD, Shillington AM. Environmental predictors of heavy episodic drinking events. Am J Drug Alcohol Abuse 2001;27: 310–313. 16. Bellis MA, Hughes K, Quigg Z, Morleo M, Jarman I, Lisboa P. Cross-sectional measures and modelled estimates of blood alcohol
17. 18.
19. 20. 21.
22. 23.
24. 25. 26.
379
levels in UK nightlife and their relationships with drinking behaviours and observed signs of inebriation. Substance Abuse Treat, Prevent, Policy 2010;5:1–8. Reed MB, Clapp JD, Martell B, Hidalgo-Sotelo A. The relationship between group size, intoxication and continuing to drink after bar attendance. Drug Alcohol Depend 2013;133:198–203. Hingson RW, Zha W, Weitzman ER. Magnitude of and trends in alcohol-related mortality and morbidity among U.S. college students aged 18–24 1998–2005. J Stud Alcohol Drugs (Suppl.) 2009;16:12–20. LaBrie JW, Kenney SR, Mirza R, Lac A. Identifying factors that increase the likelihood of driving after drinking among college students. Accident Analysis Prevent 2011;43:1371–1377. Clapp JD, Johnson M, Voas RB, Lange JE, Shillington A, Russel C. Reducing DUI among US college students: results of an environmental prevention trial. Addiction 2005;100:327–334. Clapp JD, Holmes MR, Reed MB, Shillington AM, Freisthler B. Measuring college students’ alcohol consumption in natural drinking environments: field methodologies for bars and parties. Evaluation Rev 2007;31:469–489. Clapp JD, Lange J, Min JW, Shillington A, Johnson M, Voas R. Two studies examining environmental predictors of heavy drinking by college students. Prevent Sci 2003;4:99–108. Clapp JD, Min JW, Trim R, Reed MB, Lange JE, Shillington AM, Croff J. Predictors of error in estimates of blood alcohol concentrations: a replication. J Alcohol Drug Stud 2009;70: 683–688. Tabachnick BG, Fidell LS. Using multivariate statistics, 5th ed. Boston, MA: Allyn & Bacon; 2007. Hox J. Multilevel analysis: techniques and applications. Mahwah, NJ: Lawrence Erlbaum Associates, Inc.; 2002. Stone AA, Shiffman S. Ecological momentary assessment in behavioral medicine. Annals Behav Med 1994;16:199–202.
Copyright of American Journal of Drug & Alcohol Abuse is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
