Abstract:
Objective: Heavy drinking is prevalent on college campuses, and collegiate sporting events may precipitate heavy drinking. Despite this, relatively few studies have examined student drinking on the days of sporting events. In 2003, Syracuse University won the men's National College Athletic Association basketball championship; ongoing data collection allowed an investigation of alcohol consumption at Syracuse University during the two Final Four game days, when the semifinals and championship games are played. The goals of the study were to examine the level of alcohol use on these days and to examine factors related to game-day consumption. Method: As a part of an ongoing study, 206 undergraduate students completed several questionnaires, then returned daily drinking diaries at 1-week intervals for 4 consecutive weeks. Results: Alcohol consumption levels on the two game days exceeded what is typical on campus. Further analyses conducted using zero-inflated negative binomial regression modeling demonstrated that heavier drinkers were more likely to drink alcohol, and drink alcohol heavily, on both game days. Furthermore, lack of impulse control was independently associated with an increased likelihood of drinking on both days. Conclusions: Although results from this study should be considered preliminary, these data document heavier drinking associated with high-profile athletic events. Sporting events may be a particularly opportune time and venue for collegiate risk-reduction programs.
Full Text :COPYRIGHT 2005 Alcohol Research Documentation, Inc.
HEAVY DRINKING IS PREVALENT on college campuses nationwide (Douglas et al., 1997; Meilman et al., 1997; Wechsler et al., 1995), and college students report a wide range of negative consequences due to alcohol use (Presley et al., 1994). Collegiate sporting events appear to be a particularly heavy-drinking context; sports-related alcohol use can occur in many places (e.g., at home, at a bar, at the stadium) and celebratory drinking is a frequently endorsed reason for drinking (Rabow and Duncan-Schill, 1995). Furthermore, celebratory drinking is more often related to alcohol intoxication compared to other drinking motives (Hunter, 1990). Therefore, sporting events that have important meaning may be a particularly risky situation for heavy drinking, increasing the frequency of alcohol-related rioting on college campuses after important sporting events.
Although the relationship between elevated alcohol consumption and participation in collegiate sports is clearly documented (e.g., Leichliter et al., 1998; Nelson and Wechsler, 2001), relatively few studies have examined college student drinking during sporting events. College students who are sports fans are more likely to drink heavily and experience alcohol-related consequences compared with those who are not sports fans, and students who attended schools where 40% or more of the survey respondents identified themselves as sports fans were more likely to experience negative consequences as a result of others' alcohol use (Nelson and Wechsler, 2003). However, additional research has found no statistically significant relationship between degree of interest in sports and alcohol consumption (Wann, 1998).
The National College Athletic Association (NCAA) men's basketball championship tournament runs from March through the first week in April. In 2003, Syracuse University won the national championship by defeating the University of Texas on Saturday, April 5, in a semifinal game and the University of Kansas on Monday, April 7, in the championship game. Data collection that was in progress provided an opportunity to investigate alcohol consumption at Syracuse University on these two days relative to consumption in the weeks surrounding them. The goals of the present study were to examine the level of alcohol use on these two game days and to explore factors predictive of drinking.
Method
Participants
Participants were 206 Syracuse University undergraduate students enrolled in an Introduction to Psychology course. The students received course credit in exchange for participating in the study; all gave informed consent. The sample was predominantly female (64%), white (83%) and freshman (77%). The average (SD) age among participants was 18.8 (0.98) and participants ranged in age from 18 to 25 years; 25% were involved in the Greek system (i.e., members of a fraternity or sorority).
Procedure
Participants were enrolled in an ongoing study involving questionnaire assessments and 4 weeks of self-monitoring using Daily Drinking Diaries (DDD). Questionnaire data were collected in small group testing sessions between March 26 and April 4, 2003. Measures included demographics, alcohol use and alcohol-consequences variables. Additionally, several measures that have been shown to be related to alcohol consumption were administered, including the Marlowe-Crowne Social Desirability Scale (MCSDS; Crowne and Marlowe, 1964), the Short Self-Regulation Questionnaire (SSRQ; Carey et al., 2004), the Impaired Control Scale (ICS; Heather et al., 1998), the Eysenck Impulsivity Scale (EIS; Eysenck et al., 1985) and the Reasons for Drinking Questionnaire (RDQ; Farber et al., 1980). Participants received extensive instructions about completing the DDD. For each drinking episode, participants estimated the number of standard drinks consumed in a waking day. Participants returned the weekly diary sheets to a locked box. A total of 175 participants (85%) returned drinking records for all 4 weeks, and these data were subsequently included in the analyses.
Analysis strategy
Analysis of alcohol consumption data using ordinary least squares (OLS) regression can be problematic due to extreme non-normality of data, heteroskedasticity of residuals and excessive numbers of participants who abstain. Therefore, hypothesis testing was conducted using zero-inflated negative binomial (ZINB) regression models (cf., Cameron and Trivedi, 1998). This model incorporates two distinct components: (1) a logistic component to predict the likelihood of an always-zero score; that is, in this sample, the likelihood that a student would never consume alcohol on the game day and (2) a negative binomial component to predict consumption conditional on the likelihood that the student may consume alcohol. As such, the model can predict simultaneously both who chose to drink and how much they consumed. Likelihood Ratio and Wald Tests are used for hypothesis testing with ZINB models. The regression model was built using four hierarchical steps. First, drinking was regressed on gender. Second, drinking was regressed on gender, impulse control and drinking motives. Third, drinking was regressed on gender, typical drinking and drinking frequency. Fourth, drinking was regressed on gender, impulse control, drinking motives, typical drinking and drinking frequency. Such analyses allowed for understanding of independent and joint prediction of consumption with other covariates. Each set of hierarchical models was run on drinking data for both the semifinal game day and championship game day.
Results
Alcohol-use summary statistics from the questionnaire assessment (N = 206) and the DDD (n = 175) are below. For the questionnaire assessments, male students compared to female students had significantly higher peak consumption. The mean (SD) peak consumption for men was 11.5 (5.8) versus 6.6 (3.9) for women (F = 53.5, 1/204 df, p < .001). For typical consumption, the mean for men was 5.6 (3.3) compared with 3.4 (2.7) for women (F = 26.3, 1/204 df, p < .001). The DDD data revealed a similar pattern; male students compared to female students had significantly higher peak consumption. The mean peak consumption for men was 12.3 (7.1) compared with 7.7 (4.9) for women (F = 28.7, 1/173 df, p < .001). For typical consumption, the mean for men was 6.5 (3.5) compared with 3.5 (2.8) for women (F = 17.5, 1/173 df, p < .001).
Descriptive drinking data
On April 5 (date of the semifinal game), 69.7% of the sample consumed alcohol, and the mean consumption was 5.7 (5.9) drinks, with men consuming more than women. For men, the mean was 7.7 (7.2) compared with 4.7 (4.8) for the women (F = 10.8, 1/173 df, p < .01). On April 7 (date of the championship game), 66.3% of the sample consumed alcohol and the mean consumption was 4.6 (5.51) drinks, with men consuming more than women. The mean for men was 5.6 (6.3) compared with 4.0 (4.2) for women (F = 4.3, 1/173 df, p < .05). Restricting the sample to just those who reported drinking revealed mean consumption to be 8.2 (5.5) and 6.9 (4.8) drinks on the respective game days. The quantity of consumption on the two game days exceeded consumption for a typical Saturday (the mean was 3.2; 46.9% consumed alcohol) and Monday (the mean was 0.3; 7.7% consumed alcohol).
Predicting game-day drinking
No significant relationships emerged between criterion variables and the following demographic variables: ethnicity, Greek affiliation, grade point average and age. To reduce collinearity of predictors, SSRQ, ICS, EIS, and positive-and negative-reinforcement RDQ scores were submitted to factor analysis, and factor scores based on regression weighting (cf., Hamilton, 1992) were generated to represent impulse control (SSRQ, ICS and EIS) and drinking motives (positive and negative reinforcement). No significant relationships between gender and impulse control or drinking motives emerged. Impulse control was significantly correlated with typical quantity (r = 0.24, p < .01) and frequency (r = 0.30, p < .001), and drinking motives correlated with typical quantity (r = 0.40, p < .001) and frequency (r = 0.45, p < .001). Drinking motives and impulse control were correlated (r = 0.53, p < .001) as were drinking frequency and typical quantity (r = 0.64, p < .001).
Results of the ZINB regression analyses are presented in Table 1. For semifinal game-day drinking, the first step was significant ([chi square] = 13.4, 2 df, p < .001). Gender did not significantly predict abstainers, but among those who drank, men consumed more alcohol than women. Adding impulse control and drinking motives in the second step (i.e., Step 2a) significantly increased the predictive power of the model ([chi square] = 25.1, 4 df, p < .001). In this model, men still consumed more alcohol than women, and stronger impulse control was associated with an increased likelihood of abstaining from drinking. Adding typical quantity and drinking frequency in the second step (i.e., Step 2b) significantly increased the predictive power of the model ([chi square] = 83.5, 4 df, p < .001), with heavier typical quantity and drinking frequency associated with lower likelihood of abstaining and higher consumption. In the third step, which included all five variables, it is evident that the alcohol-use variables significantly increase prediction over Step 2a ([chi square] = 67.3, 4 df, p < .001), whereas adding the personality variables to Step 2b did not ([chi square] = 9.0, 4 df, NS). In this model, gender did not have a predictive effect; higher impulse control was associated with increased likelihood of abstaining from drinking, and heavier drinkers in general were less likely to abstain from alcohol and more likely to drink larger quantities.
For championship game-day drinking, the first step was significant ([chi square] = 9.9, 2 df, p < .01). Gender did not significantly predict abstainers, but among those who drink, men consumed more alcohol than women. Adding impulse control and drinking motives in the second step (i.e., Step 2a) significantly increased the predictive power of the model ([chi square] = 38.1, 4 df, p < .001). In this model, men still consumed more alcohol than women, stronger impulse control was associated with both decreased consumption and likelihood of abstaining from drinking, and stronger drinking motives was associated with a decreased likelihood of abstaining. Adding typical quantity and drinking frequency in the second step (i.e., Step 2b) significantly increased the predictive power of the model ([chi square] = 49.9, 4 df, p < .001), with typical drinking predicting heavier consumption, and drinking frequency predicting lower likelihood of abstaining and heavier consumption. In the third step, which included all five variables, it is evident that the alcohol-use variables ([chi square] = 50.9, 4 df, p < .001) and the personality variables ([chi square] = 17.0, 4 df, p < .01) both significantly add predictive power. In this model, men were less likely to abstain, stronger impulse control was associated with a higher likelihood of abstention, and heavier drinkers were less likely to abstain from alcohol and more likely to drink larger quantities.
Discussion
Important sporting events were associated with much heavier drinking than is typically seen on this college campus. The average student consumed 5.7 drinks on the day of the semifinal game and 4.6 drinks on the day of the championship game. On the day of the championship game (a Monday), more than 8 times as many students consumed alcohol than normally would. Not surprisingly, heavier drinkers were both more likely to drink alcohol and drink alcohol heavily on the two game days. On the day of the semifinal game (a Saturday) both typical consumption and drinking frequency predicted decision to drink and how much alcohol was consumed. On the day of the championship (a Monday), frequency of drinking predicted the decision to drink, and typical amount consumed was related to both the decision to drink and how much alcohol was consumed. Personality characteristics also had a significant influence on game-day consumption, but only when considered independent of typical consumption and drinking frequency. Before controlling for typical alcohol use, lack of impulse control was a significant predictor of choosing to drink (on both game days) and heavier drinking (on championship game day). Furthermore, number of drinking motives was associated with decreased likelihood of abstention on the day of the championship game. Importantly, even after controlling for alcohol use, lower impulse control was associated with increased likelihood of drinking on both game days. Thus, both previous drinking history and personality characteristics influence which students may drink heavily in this context. Although the results could be interpreted as typical drinking patterns mediating the relationship between these two factors and game-day consumption, we are reluctant to interpret the results as such, given the strictly correlational nature of these data.
Results from this study should be considered with its limitations in mind. Although this study had the advantage of being a prospective data collection study that coincided with Syracuse University winning the national championship, this study was not designed to examine drinking during sporting events. As such, one should not draw any causal conclusions about the effects of winning a national championship on college student drinking, only that the event was correlated with heavy drinking. Also, predictive analyses were limited to the variables available; although the original study included variables that are theoretically and empirically related to college student alcohol use, we did not include variables that were specifically related to celebratory and/or sport-related drinking. However, the correlational data presented here clearly document a large increase in drinking on the days of the semifinal and championship games. Therefore, it appears that sporting events, and particularly sporting events with significant meaning on campus, may be a good window of opportunity for prevention efforts.
Acknowledgment
The authors thank Michael Carey for the inspiration for this study. TABLE 1. Results of hierarchical zero-inflated negative binomial
regression modeling
Semifinals drinking
Hierachial
steps Abstained Quantity LR [chi square]
Step 1 [chi square] = 13.4
([double dagger]), 2 df
Gender -0.17 0.45
([double
dagger])
Step 2a [chi square] = 25.1
([double dagger]), 4 df
Gender -0.18 0.46
([double
dagger])
Impulse control 0.80 -0.07
([dagger])
Drinking motives -0.45 0.26
Step 2b [chi square] = 83.5
([double dagger]), 4 df
Gender 0.69 0.15
Typical drinking -0.26 * 0.07
([double
dagger])
Drinking frequency -0.15 * 0.05
([double
dagger])
Step 3 [chi square] = 67.3
([double dagger]),
4 df (a)
[chi square] = 90,
4 df (b)
Gender 0.67 0.16
Impulse control 0.95 -0.001
([dagger])
Drinking motives 0.80 0.08
Typical drinking -0.30 0.07
([dagger]) ([double
dagger])
Drinking frequency -0.14 * 0.04
([double
dagger])
Championship drinking
Hierachial
steps Abstained Quantity LR [chi square]
Step 1 [chi square] = 9.9
([dagger]), 2 df
Gender 0.34 0.46
([double
dagger])
Step 2a [chi square] = 38.1
([double dagger]), 4 df
Gender 0.50 0.53
([double
dagger])
Impulse control 0.77 * -0.20 *
Drinking motives -1.18 * 0.10
Step 2b [chi square] = 49.9
([double dagger]), 4 df
Gender 1.03 * 0.17
Typical drinking -0.05 0.08
([double
dagger])
Drinking frequency -0.17 0.06
([dagger]) ([double
dagger])
Step 3 [chi square] = 50.9
([double dagger]),
4 df (a)
[chi square] = 17.0
([dagger]), 4 df (b)
Gender 0.96 * 0.22
Impulse control 0.77 * -0.14
Drinking motives -0.53 -0.10
Typical drinking -0.01 0.08
([double
dagger])
Drinking frequency -0.13 * 0.05
([double
dagger])
Notes: Impulse control = composite of self-regulation, impulsivity
and impaired control; Drinking motives = composite of positive and
negative reinforcement motives. (a) Compare with Step 2a; (b) compare
with Step 2b.
* p < .05; ([dagger]) p < .01; ([double dagger]) p < .001.
* This research was supported, in part, by National Institute on Alcohol Abuse and Alcoholism grant R01 AA 12518 to Kate B. Carey.
References
CAMERON, A.C. AND TRIVEDI, P.K. Regression Analysis of Count Data, New York: Cambridge Univ. Press, 1998.
CAREY, K.B., NEAL, D.J. AND COLLINS, S.E. A psychometric analysis of the Self-Regulation Questionnaire. Addict. Behav. 29: 253-260, 2004.
CROWNE, D.P. AND MARLOWE, D. The Approval Motive: Studies in Evaluative Dependence, New York: John Wiley & Sons, 1964.
DOUGLAS, K.A., COLLINS, J.L., WARREN, C., KANN, L., GOLD, R., CLAYTON, S., Ross, J.G. AND KOLBE, L.J. Results from the 1995 National College Risk Behavior Survey. J. Amer. Coll. Hlth 46: 55-66, 1997.
EYSENCK, S.B., PEARSON, P.R., LASTING, G. AND ALLSOPP, J.F. Age norms for impulsiveness, venturesomeness and empathy in adults. Pers. Indiv. Diff. 6: 613-619, 1985.
FARBER, P.D., KHAVARI, K.A. AND DOUGLASS, F.M. A factor analytic study of reasons for drinking: Empirical validation of positive and negative reinforcement dimensions. J. Cons. Clin. Psychol. 48:780-781, 1980.
HAMILTON, L.C. Regression with Graphics: A Second Course in Applied Statistics, Belmont, CA: Wadsworth, 1992.
HEATHER, N., BOOTH, P. AND LUCE, A. Impaired Control Scale: Cross-validation and relationship with treatment outcome. Addiction 93: 761-771, 1998.
HUNTER, G.T. A survey of the social context of drinking among college women. J. Alcohol Drug Educ. 35 (3): 73-80, 1990.
LEICHLITER, J.S., MEILMAN, P.W., PRESLEY, C.A. AND CASHIN, J.R. Alcohol use and related consequences among students with varying levels of involvement in college athletics. J. Amer. Coll. Hlth 46: 257-262, 1998.
MEILMAN, P.W., PRESLEY, C.A. AND CASHIN, J.R. Average weekly alcohol consumption: Drinking percentiles for American college students. J. Amer. Coll. Hlth 45: 201-204, 1997.
NELSON, T.F. AND WECHSLER, H. Alcohol and college athletes. Med. Sci. Sports Exer. 33: 43-47, 2001.
NELSON, T.F. AND WECHSLER, H. School spirits: Alcohol and the collegiate sports fan. Addict. Behav. 28: 1-11, 2003.
PRESLEY, C.A., MEILMAN, P.W. AND LYERLA, R. Development of the Core Alcohol and Drug Survey: Initial findings and future directions. J. Amer. Coll. Hlth 42: 248-255, 1994.
RABOW, J. AND DUNCAN-SCHILL, M. Drinking among college students. J. Alcohol Drug Educ. 40 (3): 52-64, 1995.
WANN, D.L. A preliminary investigation of the relationship between alcohol use and sport random. Social Behav. Pers. 26: 287-290, 1998.
WECHSLER, H., DOWDALL, G.W., DAVENPORT, A. AND CASTILLO, S. Correlates of college student binge drinking. Amer. J. Publ. Hlth 85: 921-926, 1995.
Received: July 8, 2004; Revision: September 23, 2004.
DAN J. NEAL, PH.D., ([dagger]) DAWN E. SUGARMAN, B.A., JOHN T. P. HUSTAD, M.S., CATHERINE M. CASKA, AND KATE B. CAREY, PH.D.
Center for Health and Behavior, Syracuse University, Syracuse, New York
([dagger]) Dan J. Neal is now with the Department of Psychology, the University of Texas at Austin. Correspondence may be addressed to him at the University of Texas at Austin, 1 University Station A8000, Austin, TX 78712-0187, or via email at: neal@psy.utexas.edu.
Source Citation:Neal, Dan J., Dawn E. Sugarman, John T.P. Hustad, Catherine M. Caska, and Kate B. Carey. "It's all fun and games ... or is it? Collegiate sporting events and celebratory drinking *." Journal of Studies on Alcohol 66.2 (March 2005): 291(4). Academic OneFile. Gale. Boise State Univ/ Albertsons Lib. 12 Mar. 2008.
Gale Document Number:A132963341
Objective: Heavy drinking is prevalent on college campuses, and collegiate sporting events may precipitate heavy drinking. Despite this, relatively few studies have examined student drinking on the days of sporting events. In 2003, Syracuse University won the men's National College Athletic Association basketball championship; ongoing data collection allowed an investigation of alcohol consumption at Syracuse University during the two Final Four game days, when the semifinals and championship games are played. The goals of the study were to examine the level of alcohol use on these days and to examine factors related to game-day consumption. Method: As a part of an ongoing study, 206 undergraduate students completed several questionnaires, then returned daily drinking diaries at 1-week intervals for 4 consecutive weeks. Results: Alcohol consumption levels on the two game days exceeded what is typical on campus. Further analyses conducted using zero-inflated negative binomial regression modeling demonstrated that heavier drinkers were more likely to drink alcohol, and drink alcohol heavily, on both game days. Furthermore, lack of impulse control was independently associated with an increased likelihood of drinking on both days. Conclusions: Although results from this study should be considered preliminary, these data document heavier drinking associated with high-profile athletic events. Sporting events may be a particularly opportune time and venue for collegiate risk-reduction programs.
Full Text :COPYRIGHT 2005 Alcohol Research Documentation, Inc.
HEAVY DRINKING IS PREVALENT on college campuses nationwide (Douglas et al., 1997; Meilman et al., 1997; Wechsler et al., 1995), and college students report a wide range of negative consequences due to alcohol use (Presley et al., 1994). Collegiate sporting events appear to be a particularly heavy-drinking context; sports-related alcohol use can occur in many places (e.g., at home, at a bar, at the stadium) and celebratory drinking is a frequently endorsed reason for drinking (Rabow and Duncan-Schill, 1995). Furthermore, celebratory drinking is more often related to alcohol intoxication compared to other drinking motives (Hunter, 1990). Therefore, sporting events that have important meaning may be a particularly risky situation for heavy drinking, increasing the frequency of alcohol-related rioting on college campuses after important sporting events.
Although the relationship between elevated alcohol consumption and participation in collegiate sports is clearly documented (e.g., Leichliter et al., 1998; Nelson and Wechsler, 2001), relatively few studies have examined college student drinking during sporting events. College students who are sports fans are more likely to drink heavily and experience alcohol-related consequences compared with those who are not sports fans, and students who attended schools where 40% or more of the survey respondents identified themselves as sports fans were more likely to experience negative consequences as a result of others' alcohol use (Nelson and Wechsler, 2003). However, additional research has found no statistically significant relationship between degree of interest in sports and alcohol consumption (Wann, 1998).
The National College Athletic Association (NCAA) men's basketball championship tournament runs from March through the first week in April. In 2003, Syracuse University won the national championship by defeating the University of Texas on Saturday, April 5, in a semifinal game and the University of Kansas on Monday, April 7, in the championship game. Data collection that was in progress provided an opportunity to investigate alcohol consumption at Syracuse University on these two days relative to consumption in the weeks surrounding them. The goals of the present study were to examine the level of alcohol use on these two game days and to explore factors predictive of drinking.
Method
Participants
Participants were 206 Syracuse University undergraduate students enrolled in an Introduction to Psychology course. The students received course credit in exchange for participating in the study; all gave informed consent. The sample was predominantly female (64%), white (83%) and freshman (77%). The average (SD) age among participants was 18.8 (0.98) and participants ranged in age from 18 to 25 years; 25% were involved in the Greek system (i.e., members of a fraternity or sorority).
Procedure
Participants were enrolled in an ongoing study involving questionnaire assessments and 4 weeks of self-monitoring using Daily Drinking Diaries (DDD). Questionnaire data were collected in small group testing sessions between March 26 and April 4, 2003. Measures included demographics, alcohol use and alcohol-consequences variables. Additionally, several measures that have been shown to be related to alcohol consumption were administered, including the Marlowe-Crowne Social Desirability Scale (MCSDS; Crowne and Marlowe, 1964), the Short Self-Regulation Questionnaire (SSRQ; Carey et al., 2004), the Impaired Control Scale (ICS; Heather et al., 1998), the Eysenck Impulsivity Scale (EIS; Eysenck et al., 1985) and the Reasons for Drinking Questionnaire (RDQ; Farber et al., 1980). Participants received extensive instructions about completing the DDD. For each drinking episode, participants estimated the number of standard drinks consumed in a waking day. Participants returned the weekly diary sheets to a locked box. A total of 175 participants (85%) returned drinking records for all 4 weeks, and these data were subsequently included in the analyses.
Analysis strategy
Analysis of alcohol consumption data using ordinary least squares (OLS) regression can be problematic due to extreme non-normality of data, heteroskedasticity of residuals and excessive numbers of participants who abstain. Therefore, hypothesis testing was conducted using zero-inflated negative binomial (ZINB) regression models (cf., Cameron and Trivedi, 1998). This model incorporates two distinct components: (1) a logistic component to predict the likelihood of an always-zero score; that is, in this sample, the likelihood that a student would never consume alcohol on the game day and (2) a negative binomial component to predict consumption conditional on the likelihood that the student may consume alcohol. As such, the model can predict simultaneously both who chose to drink and how much they consumed. Likelihood Ratio and Wald Tests are used for hypothesis testing with ZINB models. The regression model was built using four hierarchical steps. First, drinking was regressed on gender. Second, drinking was regressed on gender, impulse control and drinking motives. Third, drinking was regressed on gender, typical drinking and drinking frequency. Fourth, drinking was regressed on gender, impulse control, drinking motives, typical drinking and drinking frequency. Such analyses allowed for understanding of independent and joint prediction of consumption with other covariates. Each set of hierarchical models was run on drinking data for both the semifinal game day and championship game day.
Results
Alcohol-use summary statistics from the questionnaire assessment (N = 206) and the DDD (n = 175) are below. For the questionnaire assessments, male students compared to female students had significantly higher peak consumption. The mean (SD) peak consumption for men was 11.5 (5.8) versus 6.6 (3.9) for women (F = 53.5, 1/204 df, p < .001). For typical consumption, the mean for men was 5.6 (3.3) compared with 3.4 (2.7) for women (F = 26.3, 1/204 df, p < .001). The DDD data revealed a similar pattern; male students compared to female students had significantly higher peak consumption. The mean peak consumption for men was 12.3 (7.1) compared with 7.7 (4.9) for women (F = 28.7, 1/173 df, p < .001). For typical consumption, the mean for men was 6.5 (3.5) compared with 3.5 (2.8) for women (F = 17.5, 1/173 df, p < .001).
Descriptive drinking data
On April 5 (date of the semifinal game), 69.7% of the sample consumed alcohol, and the mean consumption was 5.7 (5.9) drinks, with men consuming more than women. For men, the mean was 7.7 (7.2) compared with 4.7 (4.8) for the women (F = 10.8, 1/173 df, p < .01). On April 7 (date of the championship game), 66.3% of the sample consumed alcohol and the mean consumption was 4.6 (5.51) drinks, with men consuming more than women. The mean for men was 5.6 (6.3) compared with 4.0 (4.2) for women (F = 4.3, 1/173 df, p < .05). Restricting the sample to just those who reported drinking revealed mean consumption to be 8.2 (5.5) and 6.9 (4.8) drinks on the respective game days. The quantity of consumption on the two game days exceeded consumption for a typical Saturday (the mean was 3.2; 46.9% consumed alcohol) and Monday (the mean was 0.3; 7.7% consumed alcohol).
Predicting game-day drinking
No significant relationships emerged between criterion variables and the following demographic variables: ethnicity, Greek affiliation, grade point average and age. To reduce collinearity of predictors, SSRQ, ICS, EIS, and positive-and negative-reinforcement RDQ scores were submitted to factor analysis, and factor scores based on regression weighting (cf., Hamilton, 1992) were generated to represent impulse control (SSRQ, ICS and EIS) and drinking motives (positive and negative reinforcement). No significant relationships between gender and impulse control or drinking motives emerged. Impulse control was significantly correlated with typical quantity (r = 0.24, p < .01) and frequency (r = 0.30, p < .001), and drinking motives correlated with typical quantity (r = 0.40, p < .001) and frequency (r = 0.45, p < .001). Drinking motives and impulse control were correlated (r = 0.53, p < .001) as were drinking frequency and typical quantity (r = 0.64, p < .001).
Results of the ZINB regression analyses are presented in Table 1. For semifinal game-day drinking, the first step was significant ([chi square] = 13.4, 2 df, p < .001). Gender did not significantly predict abstainers, but among those who drank, men consumed more alcohol than women. Adding impulse control and drinking motives in the second step (i.e., Step 2a) significantly increased the predictive power of the model ([chi square] = 25.1, 4 df, p < .001). In this model, men still consumed more alcohol than women, and stronger impulse control was associated with an increased likelihood of abstaining from drinking. Adding typical quantity and drinking frequency in the second step (i.e., Step 2b) significantly increased the predictive power of the model ([chi square] = 83.5, 4 df, p < .001), with heavier typical quantity and drinking frequency associated with lower likelihood of abstaining and higher consumption. In the third step, which included all five variables, it is evident that the alcohol-use variables significantly increase prediction over Step 2a ([chi square] = 67.3, 4 df, p < .001), whereas adding the personality variables to Step 2b did not ([chi square] = 9.0, 4 df, NS). In this model, gender did not have a predictive effect; higher impulse control was associated with increased likelihood of abstaining from drinking, and heavier drinkers in general were less likely to abstain from alcohol and more likely to drink larger quantities.
For championship game-day drinking, the first step was significant ([chi square] = 9.9, 2 df, p < .01). Gender did not significantly predict abstainers, but among those who drink, men consumed more alcohol than women. Adding impulse control and drinking motives in the second step (i.e., Step 2a) significantly increased the predictive power of the model ([chi square] = 38.1, 4 df, p < .001). In this model, men still consumed more alcohol than women, stronger impulse control was associated with both decreased consumption and likelihood of abstaining from drinking, and stronger drinking motives was associated with a decreased likelihood of abstaining. Adding typical quantity and drinking frequency in the second step (i.e., Step 2b) significantly increased the predictive power of the model ([chi square] = 49.9, 4 df, p < .001), with typical drinking predicting heavier consumption, and drinking frequency predicting lower likelihood of abstaining and heavier consumption. In the third step, which included all five variables, it is evident that the alcohol-use variables ([chi square] = 50.9, 4 df, p < .001) and the personality variables ([chi square] = 17.0, 4 df, p < .01) both significantly add predictive power. In this model, men were less likely to abstain, stronger impulse control was associated with a higher likelihood of abstention, and heavier drinkers were less likely to abstain from alcohol and more likely to drink larger quantities.
Discussion
Important sporting events were associated with much heavier drinking than is typically seen on this college campus. The average student consumed 5.7 drinks on the day of the semifinal game and 4.6 drinks on the day of the championship game. On the day of the championship game (a Monday), more than 8 times as many students consumed alcohol than normally would. Not surprisingly, heavier drinkers were both more likely to drink alcohol and drink alcohol heavily on the two game days. On the day of the semifinal game (a Saturday) both typical consumption and drinking frequency predicted decision to drink and how much alcohol was consumed. On the day of the championship (a Monday), frequency of drinking predicted the decision to drink, and typical amount consumed was related to both the decision to drink and how much alcohol was consumed. Personality characteristics also had a significant influence on game-day consumption, but only when considered independent of typical consumption and drinking frequency. Before controlling for typical alcohol use, lack of impulse control was a significant predictor of choosing to drink (on both game days) and heavier drinking (on championship game day). Furthermore, number of drinking motives was associated with decreased likelihood of abstention on the day of the championship game. Importantly, even after controlling for alcohol use, lower impulse control was associated with increased likelihood of drinking on both game days. Thus, both previous drinking history and personality characteristics influence which students may drink heavily in this context. Although the results could be interpreted as typical drinking patterns mediating the relationship between these two factors and game-day consumption, we are reluctant to interpret the results as such, given the strictly correlational nature of these data.
Results from this study should be considered with its limitations in mind. Although this study had the advantage of being a prospective data collection study that coincided with Syracuse University winning the national championship, this study was not designed to examine drinking during sporting events. As such, one should not draw any causal conclusions about the effects of winning a national championship on college student drinking, only that the event was correlated with heavy drinking. Also, predictive analyses were limited to the variables available; although the original study included variables that are theoretically and empirically related to college student alcohol use, we did not include variables that were specifically related to celebratory and/or sport-related drinking. However, the correlational data presented here clearly document a large increase in drinking on the days of the semifinal and championship games. Therefore, it appears that sporting events, and particularly sporting events with significant meaning on campus, may be a good window of opportunity for prevention efforts.
Acknowledgment
The authors thank Michael Carey for the inspiration for this study. TABLE 1. Results of hierarchical zero-inflated negative binomial
regression modeling
Semifinals drinking
Hierachial
steps Abstained Quantity LR [chi square]
Step 1 [chi square] = 13.4
([double dagger]), 2 df
Gender -0.17 0.45
([double
dagger])
Step 2a [chi square] = 25.1
([double dagger]), 4 df
Gender -0.18 0.46
([double
dagger])
Impulse control 0.80 -0.07
([dagger])
Drinking motives -0.45 0.26
Step 2b [chi square] = 83.5
([double dagger]), 4 df
Gender 0.69 0.15
Typical drinking -0.26 * 0.07
([double
dagger])
Drinking frequency -0.15 * 0.05
([double
dagger])
Step 3 [chi square] = 67.3
([double dagger]),
4 df (a)
[chi square] = 90,
4 df (b)
Gender 0.67 0.16
Impulse control 0.95 -0.001
([dagger])
Drinking motives 0.80 0.08
Typical drinking -0.30 0.07
([dagger]) ([double
dagger])
Drinking frequency -0.14 * 0.04
([double
dagger])
Championship drinking
Hierachial
steps Abstained Quantity LR [chi square]
Step 1 [chi square] = 9.9
([dagger]), 2 df
Gender 0.34 0.46
([double
dagger])
Step 2a [chi square] = 38.1
([double dagger]), 4 df
Gender 0.50 0.53
([double
dagger])
Impulse control 0.77 * -0.20 *
Drinking motives -1.18 * 0.10
Step 2b [chi square] = 49.9
([double dagger]), 4 df
Gender 1.03 * 0.17
Typical drinking -0.05 0.08
([double
dagger])
Drinking frequency -0.17 0.06
([dagger]) ([double
dagger])
Step 3 [chi square] = 50.9
([double dagger]),
4 df (a)
[chi square] = 17.0
([dagger]), 4 df (b)
Gender 0.96 * 0.22
Impulse control 0.77 * -0.14
Drinking motives -0.53 -0.10
Typical drinking -0.01 0.08
([double
dagger])
Drinking frequency -0.13 * 0.05
([double
dagger])
Notes: Impulse control = composite of self-regulation, impulsivity
and impaired control; Drinking motives = composite of positive and
negative reinforcement motives. (a) Compare with Step 2a; (b) compare
with Step 2b.
* p < .05; ([dagger]) p < .01; ([double dagger]) p < .001.
* This research was supported, in part, by National Institute on Alcohol Abuse and Alcoholism grant R01 AA 12518 to Kate B. Carey.
References
CAMERON, A.C. AND TRIVEDI, P.K. Regression Analysis of Count Data, New York: Cambridge Univ. Press, 1998.
CAREY, K.B., NEAL, D.J. AND COLLINS, S.E. A psychometric analysis of the Self-Regulation Questionnaire. Addict. Behav. 29: 253-260, 2004.
CROWNE, D.P. AND MARLOWE, D. The Approval Motive: Studies in Evaluative Dependence, New York: John Wiley & Sons, 1964.
DOUGLAS, K.A., COLLINS, J.L., WARREN, C., KANN, L., GOLD, R., CLAYTON, S., Ross, J.G. AND KOLBE, L.J. Results from the 1995 National College Risk Behavior Survey. J. Amer. Coll. Hlth 46: 55-66, 1997.
EYSENCK, S.B., PEARSON, P.R., LASTING, G. AND ALLSOPP, J.F. Age norms for impulsiveness, venturesomeness and empathy in adults. Pers. Indiv. Diff. 6: 613-619, 1985.
FARBER, P.D., KHAVARI, K.A. AND DOUGLASS, F.M. A factor analytic study of reasons for drinking: Empirical validation of positive and negative reinforcement dimensions. J. Cons. Clin. Psychol. 48:780-781, 1980.
HAMILTON, L.C. Regression with Graphics: A Second Course in Applied Statistics, Belmont, CA: Wadsworth, 1992.
HEATHER, N., BOOTH, P. AND LUCE, A. Impaired Control Scale: Cross-validation and relationship with treatment outcome. Addiction 93: 761-771, 1998.
HUNTER, G.T. A survey of the social context of drinking among college women. J. Alcohol Drug Educ. 35 (3): 73-80, 1990.
LEICHLITER, J.S., MEILMAN, P.W., PRESLEY, C.A. AND CASHIN, J.R. Alcohol use and related consequences among students with varying levels of involvement in college athletics. J. Amer. Coll. Hlth 46: 257-262, 1998.
MEILMAN, P.W., PRESLEY, C.A. AND CASHIN, J.R. Average weekly alcohol consumption: Drinking percentiles for American college students. J. Amer. Coll. Hlth 45: 201-204, 1997.
NELSON, T.F. AND WECHSLER, H. Alcohol and college athletes. Med. Sci. Sports Exer. 33: 43-47, 2001.
NELSON, T.F. AND WECHSLER, H. School spirits: Alcohol and the collegiate sports fan. Addict. Behav. 28: 1-11, 2003.
PRESLEY, C.A., MEILMAN, P.W. AND LYERLA, R. Development of the Core Alcohol and Drug Survey: Initial findings and future directions. J. Amer. Coll. Hlth 42: 248-255, 1994.
RABOW, J. AND DUNCAN-SCHILL, M. Drinking among college students. J. Alcohol Drug Educ. 40 (3): 52-64, 1995.
WANN, D.L. A preliminary investigation of the relationship between alcohol use and sport random. Social Behav. Pers. 26: 287-290, 1998.
WECHSLER, H., DOWDALL, G.W., DAVENPORT, A. AND CASTILLO, S. Correlates of college student binge drinking. Amer. J. Publ. Hlth 85: 921-926, 1995.
Received: July 8, 2004; Revision: September 23, 2004.
DAN J. NEAL, PH.D., ([dagger]) DAWN E. SUGARMAN, B.A., JOHN T. P. HUSTAD, M.S., CATHERINE M. CASKA, AND KATE B. CAREY, PH.D.
Center for Health and Behavior, Syracuse University, Syracuse, New York
([dagger]) Dan J. Neal is now with the Department of Psychology, the University of Texas at Austin. Correspondence may be addressed to him at the University of Texas at Austin, 1 University Station A8000, Austin, TX 78712-0187, or via email at: neal@psy.utexas.edu.
Source Citation:Neal, Dan J., Dawn E. Sugarman, John T.P. Hustad, Catherine M. Caska, and Kate B. Carey. "It's all fun and games ... or is it? Collegiate sporting events and celebratory drinking *." Journal of Studies on Alcohol 66.2 (March 2005): 291(4). Academic OneFile. Gale. Boise State Univ/ Albertsons Lib. 12 Mar. 2008
Gale Document Number:A132963341
This study was conducted at Syracuse University during the 2003 NCAA Men's Basketball Tournament. They wanted to determine if the amount of alcohol consumed before, during the game and afterwards, increased on game days that were deemed important for the school. Their basis being celebratory drinking is more socially accepted than binge drinking.
Again, it may be useable for some statistics. I would say it is reliable based on the methods and math involved in breaking everything down. Didn't learn anything that I hadn't already thought, other than some percentages I wasn't aware of. Again the reading didn't change how I had thought it would conclude.
No comments:
Post a Comment