Extreme Binge Drinking among 12^th-Grade Students in the U.S.: Prevalence and Predictors

Abstract

Importance

The prevalence of underage alcohol use has been studied extensively but binge drinking among youth in the U.S. is not yet well understood. In particular, adolescents may drink much larger amounts than the threshold (5 drinks) often used in definitions of binge drinking. Delineating various levels of binge drinking, including extreme levels, and understanding predictors of such extreme binge drinking among adolescents will benefit public health efforts.

Objective

To examine the prevalence and predictors of 5+ binge drinking and of 10+ and 15+ extreme binge drinking among 12th graders in the U.S.

Design

A non-clinical nationally representative sample.

Setting

High school seniors in the annual Monitoring the Future study between 2005 and 2011.

Participants

The sample included 16,332 12th graders (modal age 18) in the U.S. Response rates were 79–85%.

Main Outcome Measures

Prevalence of consuming 5+, 10+, and 15+ drinks in a row in the past two weeks.

Results

Between 2005 and 2011, 20.2% of high school seniors reported 5+ binge drinking, 10.5% reported 10+ extreme binge drinking, and 5.6% reported 15+ extreme binge drinking in the past 2 weeks. Rates of 5+ binge drinking and 10+ extreme binge drinking have declined since 2005, but rates of 15+ extreme binge drinking have not. Students with college-educated parents were more likely to consume 5+ drinks but less likely to consume 15+ drinks than students whose parents were not college educated. Students from more rural areas were more likely than students from large metropolitan areas to drink 15+ drinks. Socializing with substance-using peers, number of evenings out with friends, substance-related attitudes, and other substance use (cigarettes, marijuana) predicted all three levels of binge and extreme binge drinking.

Conclusions

Binge drinking at the traditionally defined 5+ drinking level was common among high school seniors representative of all 12^th graders in the contiguous U.S. A significant segment of students also reported extreme binge drinking at levels two and three times higher. These data suggest the importance of assessing multiple levels of binge drinking behavior and their predictors among adolescents in order to target effective screening and intervention efforts.

Adolescent alcohol consumption is a major public health problem in the U.S. and a high priority for organizations such as the Office of Surgeon General,¹ Centers for Disease Control and Prevention,² World Health Organization,³ American Academy of Pediatrics,⁴ and National Institute on Alcohol Abuse and Alcoholism.⁵ Approximately 5,000 persons under the age of 21 die each year from alcohol related-fatalities,⁶ while problems linked to underage drinking were estimated to cost about $62 billion in 2001.⁷ Underage drinking is also a predictor of alcohol problems and early mortality in adulthood.^8–11

Consuming a large amount of alcohol in a single sitting (binge or heavy episodic drinking) confers acute risks (eg, impaired driving, alcohol poisoning, injury) and long-term risks (eg, alterations to the developing brain, liver damage, alcohol dependence^{2, 12–15}). In alcohol studies, binge drinking is commonly defined as 5 or more drinks¹⁶ (or 4+ for women and 5+ for men^{16, 17}) based on the approximation that consuming 5 drinks in a 2-hour period would lead to a blood alcohol concentration (BAC) of up to 80 mg/dL (.08%) for the typical adult.¹⁷ The 5+ measure has been a valuable tool for research predicting consequences of alcohol use.^{16, 18–20} However, sole reliance on a 5+ binge drinking threshold obscures meaningful variance in the quantity of alcohol consumed per occasion. Serious acute consequences of alcohol use are considerably more likely at very high levels of alcohol use.^{21, 22} Despite the known risks, the extent of adolescent alcohol use at the high end of binge drinking remains unclear.

Recent studies using multiple cut-offs for binge drinking have found variable consequences (eg, high risk for injuries at 5+/8+ drinks for women/men²³) and have shown that considerable numbers of people engage in high levels of binge drinking (eg, in a 19- to 30-year-old sample, 14.7% reporting 10+ and 5.6% reporting 15+ drinks in a row in the past 2 weeks.²⁴ Significant gender differences at various levels of binge drinking in a sample of first-term college freshmen were found (33.7% of women vs. 40.6% of men at 4+/5+ drinks, 8.2% vs. 19.9% at 8+/10+, and 1.8% vs. 7.6% at 12+/15+).²⁵ An increasing number of studies document risky single-occasion drinking (10+ /11+)²⁶ and event-specific drinking²⁷ (eg, drinking during Spring Break,^28–32 21^st birthdays,^33–37 sporting events,^{38, 39} local and national holidays^{19, 40}) with particularly high levels of alcohol use. Thus far, a large proportion of the research on binge drinking^{16, 25, 41–44} and the small number of studies on more extreme binge drinking^{20, 25} focus on college students, often from single universities.

To date, research has not examined extreme binge drinking, or the prevalence of consuming dangerously high levels of alcohol in one sitting, specifically among adolescents. Despite declines in overall alcohol use among teens in the last decade and a half, alcohol exposure remains high. In 2011, 70% of U.S. 12^th-graders reported using alcohol in their lifetime, 51% reported ever being drunk, and 22% reported binge drinking (5+ in a row) in the past 2 weeks.⁴⁵ Rates are generally similar or higher in other Western countries.^{46, 47}

Drawing on theoretical perspectives and comprehensive reviews,^48–52 as well as empirical literature cited below, this study examines common U.S. sociodemographic predictors and high school risk and protective factors associated with binge drinking and extreme binge drinking. To identify the adolescents at highest risk, it compares empirically supported risk factors for consuming 5+, 10+, and 15+ drinks in a row. For traditionally defined binge drinking (5+), demographic findings have shown that male gender, White race/ethnicity, and higher parental education/SES are consistent predictors of greater alcohol use among adolescents.^{45, 49, 53–55} Binge drinking also differs in relation to geographic region and urbanicity: The Midwest and Northeast have the highest rates of 5+ binge drinking,⁴⁵ and rural high school students have higher rates of alcohol abuse, particularly for males.⁵⁶ More religious students exhibit lower levels of binge drinking than their peers.^57–59 Risk and protective factors within the school and peer context also predict greater adolescent alcohol use,^{49, 60–62} including lower average grades, plans to attend college, and higher frequency of skipping school. During high school, students who believe their friends get drunk are more likely to drink themselves.^63–65 Additional risk factors include a lower level of disapproval and a lower perceived risk of binge drinking.^45,62,66 Use of other substances, including cigarettes and marijuana, is also consistently correlated with heavier drinking.^50,62

Method

This is the first national study to examine extreme binge drinking among adolescents. It utilizes data from Monitoring the Future (MTF⁴⁵) to examine the prevalence and predictors of 5+ binge and 10+ and 15+ extreme binge drinking among nationally representative samples of American high school seniors from 2005 to 2011. Annually since 1975, MTF has used questionnaires administered in classrooms to survey nationally representative samples of about 16,000 American high school seniors (at modal age 18) each year.⁶⁷ Measures assessing 5+ binge have been included since the inception of the study and are consistent with other studies¹⁶; measures of extreme binge drinking were added in 2005.

The analyses used cohorts of 12^th-graders from the high school classes of 2005 to 2011 who answered questions regarding 5+ binge and 10+ and 15+ extreme binge drinking. Measures of 10+ and 15+ extreme binge drinking were included on one of six questionnaire forms. Multiple questionnaire forms were used to decrease respondent burden and were randomly assigned within classrooms to individuals. Analyses accounted for the complex multistage sample design, and the data were weighted to adjust for differential selection probabilities. Response rates for surveys from 2005–2011 were 79–85%, with nearly all non-response due to absenteeism. The weighted sample (N=16,332) is 52.3% girls, and 64.5% White, 11.0% Black, 13.1% Hispanic, and 11.5% Other race/ethnicity. The study was approved by a University of Michigan IRB.

Measures

Binge Drinking and Extreme Binge Drinking

5+ Binge was assessed with the question, “During the last two weeks, how many times (if any) have you had five or more drinks in a row?” 10+ Extreme Binge was assessed with the question, “During the last two weeks, how many times (if any) have you had 10 or more drinks in a row?” 15+ Extreme Binge was assessed with the question, “During the last two weeks, how many times (if any) have you had 15 or more drinks in a row?” Response options were None, Once, Twice, 3 to 5 times, 6 to 9 times, and 10 or more times. For these analyses, responses were dichotomized to None=0, Any=1. A drink was defined for respondents as any of the following, “a 12-ounce can (or bottle) of beer; a 4-ounce glass of wine; a 12-ounce bottle (or can) of wine cooler; a mixed drink, shot glass of liquor, or the equivalent.”

Demographics

Cohort year was a continuous variable for years from 2005 to 2011. Gender was coded as male=1 and female=0. Race/ethnicity was dummy coded as White (reference group), Black, Hispanic, and Other. Parent college education served as a proxy for socioeconomic status, with the maximum of mother or father education coded as some college or more=1, no college education=0. Geographic region was defined as the region where the respondent’s school was located: South (reference group), Northeast, Midwest, and West. Population density referred to the area surrounding the respondent’s school, classified based on U.S. Census categories as large Metropolitan Statistical Area (MSA, such as urban areas; reference group), Other MSA (such as suburbs), or non-MSA (such as rural areas). Religiosity was based on the self-reported importance of religion from not important=1 to very important=4.

High school risk and protective factors

Educational success and plans. Grades in high school were coded from D or lower=1 to A=9. College plans were coded as plan to graduate from a 4-year college or more=1, and plan on less than 4-year college graduation=0. The number of days students reported cutting school (ie, missing without an excuse) in the past four weeks was coded from none=1 to 11 or more=7. Social life and substance-related attitudes. Evenings out without parents in a typical week were coded from less than one=1 to 6 or 7=6. Students’ beliefs about how many of their friends get drunk at least once a week were coded as none=1 to all=5. Alcohol attitudes were measured by disapproval of binge drinking (ie, 5+ drinks) on the weekend (don’t disapprove=1 to strongly disapprove=3) and perceived risk of binge drinking on the weekend (no risk=1 to great risk=4). Finally, other substance use predictors were cigarette use in the past 30 days (none=1 to 2+ packs per day=7) and marijuana use in the past 30 days (none=1 to 40+ times=7).

Results

The prevalence rates of 5+ binge, 10+ extreme binge, and 15+ extreme binge drinking in the full sample and by gender, race/ethnicity, parental education, geographic region, and population density are shown in Table 1. Of 12^th graders, 20.2% reported consuming 5+ alcoholic drinks, 10.5% reported consuming 10+ drinks, and 5.6% reported consuming 15+ drinks in a row at least once in the past 2 weeks. Table 1 shows considerable variation in these rates as a function of the socio-demographic characteristics. Cohort year was negatively correlated with 5+ binge drinking and 10+ extreme binge drinking, but not 15+ extreme binge drinking. Chi-squared tests revealed significant differences (p < .001) by gender, race/ethnicity, parent education, and population density, with one exception; 5+ binge drinking rates did not vary by population density. Pairwise comparisons are shown in Table 1. Below, we consider in more detail socio-demographic differences in a multivariate context with the other covariates. We also examined interactions of all predictors by cohort year; only one out of 60 reached significance of p<.05. This indicates that the associations between predictors and outcomes have remained stable over time.

Table 1.

Binge Drinking and Extreme Binge Drinking Prevalence by Demographic Subgroups

		5+ Binge	10+ Extreme Binge	15+ Extreme Binge
		%	%	%
Total (N=16,332)		20.2	10.5	5.6
By Cohort Year
	2005 (N=2,438)	22	10.5	5.5
	2006 (N=2,330)	22	12.6	7
	2007 (N=2,374)	20.5	10.8	5.2
	2008 (N=2,288)	19.2	10.1	5.2
	2009 (N=2,249)	20.7	10.4	5.8
	2010 (N=2,356)	18.5	9.6	6
	2011 (N=2,297)	18.1	9.6	4.4
By Gender
	Females (N=7,963)	15.0a	5.3a	2.3a
	Males (N=7,266)	24.7b	15.1b	8.3b
By Race/Ethnicity
	White (N=9,656)	23.8a	12.5a	6.3a
	Black (N=1,740)	7.6b	3.2b	1.8b
	Hispanic (N=2,112)	15.9c	7.7c	4.6c
	Other (N=1,794)	14.5d	7.0d	4.4d
By Parental Education
	No College Education
	(N=4,076)	18.1a	10.1a	6.0a
	Some College or More
	(N=10,724)	20.3b	9.9b	4.8b
By Geographic Region
	South (N=5,879)	18.4a	10.3a	5.9a,b
	Northeast (N=3,040)	24.2b	11.9a,b	5.2a
	Midwest (N=3,891)	22.7b	12.8b	7.3b
	West (N=3,522)	16.9a	7.3c	3.5c
By Population Density
	Large MSA (N=4,900)	20.4a	10.1a	4.9a
	Other MSA (N=7,877)	19.8a	9.9a	5.2a
	Non-MSA (N=3,556)	20.7a	12.5b	7.4b

Note.

^***p<.001,

^**p<.01. Measures assessed 5+, 10+, and 15+ drinks in a row in the past two weeks. There were missing data on gender (N=1103), race/ethnicity (N=1031), and parental education (N=1532); those with missing data on these variables tended to report higher rates of binge drinking. Correlation coefficients between cohort year and each of the binge drinking behaviors are r=.03*** for 5+ binge drinking, r=.02** for 10+ extreme binge drinking, and r=.01 for 15+ binge drinking. χ² tests for difference in proportions across each of the three thresholds of binge drinking by gender, race/ethnicity, parental education, and population density were all significant at p<.001, with one exception: there were no significant differences based on population density for 5+ binge drinking. Pairwise comparisons are shown by superscript values, where different superscript values indicate significant differences at p<.05.

Table 2 documents estimates of the partialled predictive power of the demographic and high school risk factors for binge drinking and extreme binge drinking, based on multivariate logistic regression models. Missing data dummy variables were included for all predictors with missing data, with means also assigned for continuous predictors. Students with missing data on gender were more likely to binge drink at all levels. Students with missing data on high school grades were less likely to engage in 5+ binge. Those with missing data on college plans were more likely to engage in 5+ binge. Students with missing data on days cut school were more likely to report 15+ binge. Students with missing data on disapproval of binge drinking were more likely to engage in all levels of binge drinking. Those with missing data on perceived risk were less likely to report 5+ binge drinking. Students with missing data on cigarette or marijuana use had greater odds of 15+ binge drinking. There were no significant differences for students with missing data on race/ethnicity, parental education, religiosity, evenings out, or friends get drunk compared to students who gave valid data for these measures.

Table 2.

Multivariate Logistic Regressions Predicting Binge Drinking and Extreme Binge Drinking among American 12^th-Grade Students

		5+ Binge	10+ Extreme Binge	15+ Extreme Binge
		OR [95% CI]	OR [95% CI]	OR [95% CI]
Demographic Subgroups
	Cohort Year	0.99 [0.96, 1.02]	1.01 [0.97, 1.05]	1.01 [0.97, 1.06]
	Gendera
	Male	1.45 [1.28, 1.64]***	2.74 [2.32, 3.24]***	3.20 [2.61, 3.94]***
	Race/Ethnicityb
	Black	0.38 [0.27, 0.51]***	0.31 [0.21, 0.46]***	0.38 [0.22, 0.66]***
	Hispanic	0.86 [0.72, 1.04]	0.88 [0.68, 1.14]	1.18 [0.85, 1.65]
	Other	0.65 [0.54, 0.79]***	0.69 [0.53, 0.90]**	1.04 [0.75, 1.42]
	Geographic Regiond
	Northeast	1.14 [0.94, 1.40]	0.99 [0.79, 1.23]	0.72 [0.56, 0.94]*
	Midwest	1.05 [0.89, 1.23]	1.05 [0.86, 1.30]	1.07 [0.84, 1.38]
	West	1.06 [0.87, 1.30]	0.81 [0.59, 1.12]	0.67 [0.47, 0.96]*
	Population Densitye
	Other MSA	0.94 [0.80, 1.09]	0.98 [0.81, 1.19]	1.12 [0.90, 1.40]
	Non-MSA	0.93 [0.77, 1.11]	1.20 [0.95, 1.51]	1.51 [1.15, 1.99]**
	Parent College Educationc
	Some College or More	1.22 [1.07, 1.40]*	0.98 [0.82, 1.17]	0.79 [0.64, 0.99]*
	Religiosity	0.94 [0.88, 1.00]	1.00 [0.93, 1.09]	0.93 [0.84, 1.03]
High School Predictors
Educational Success and Plans
	Grades	1.01 [0.98, 1.04]	1.01 [0.97, 1.05]	1.02 [0.96, 1.08]
	College Plansf	0.97 [0.81, 1.15]	0.84 [0.69, 1.04]	0.84 [0.66, 1.08]
	Days Cut School	1.09 [1.05, 1.14]***	1.07 [1.02, 1.12]**	1.11 [1.05, 1.17]***
Social Life and Substance-related Attitudes
	Evenings Out	1.23 [1.18, 1.29]***	1.17 [1.11, 1.25]***	1.16 [1.07, 1.25]***
	Friends Get Drunk	1.72 [1.64, 1.81]***	1.83 [1.70, 1.96]***	1.83 [1.67, 1.99]***
	Disapprove of Binge	0.49 [0.45, 0.53]***	0.50 [0.44, 0.57]***	0.49 [0.41, 0.59]***
	Perceived Risk of Binge	0.70 [0.67, 0.74]***	0.68 [0.64, 0.73]***	0.72 [0.65, 0.78]***
Other Substance Use
	Cigarette Use	1.38 [1.31, 1.46]***	1.32 [1.23, 1.41]***	1.35 [1.26, 1.45]***
	Marijuana Use	1.23 [1.18, 1.27]***	1.20 [1.16, 1.24]***	1.17 [1.12, 1.23]***

Note.

^***p<.001,

^**p<.01,

^*p<.05.

^aReference category is female.

^bReference category is White.

^cReference category is no college education.

^dReference category is South.

^eReference category is large Metropolitan Statistical Area (MSA).

^fReference category is planning less than four-year-college graduation. Unweighted N = 16,330 (weighted N = 16,332). Dummy variables for missing data on gender, race/ethnicity, and all high school predictors were included but are not tabled.

Males were more likely than females to engage in all levels of binge drinking. Differences by race/ethnicity were: White students were more likely to engage in all levels of binge drinking than Black students and more likely to report 5+ binge and 10+ extreme binge drinking than students of most other race/ethnicities. White students and Hispanic students did not differ. The association between parental education and binge drinking differed across the thresholds of binge drinking. Students whose parents were college-educated had greater odds of engaging in 5+ binge drinking and lower odds of engaging in 15+ extreme binge drinking, with no difference in 10+ extreme binge drinking rates, compared to students whose parents were not college-educated. There were few differences by region or population density, although students in the Northeast and West were less likely to engage in 15+ extreme binge drinking than students in the South. Compared to students in large MSA areas, students in non-MSA (ie, more rural) areas had greater odds of engaging in 15+ extreme binge drinking. High school grades and college plans did not predict binge drinking at any threshold in multivariate analyses. For all binge levels, cutting school for more days, spending more evenings out with friends, and perceiving that more friends get drunk predicted greater odds of binge drinking. Disapproving and perceiving greater risk of 5+ binge drinking predicted lower odds in all cases. Finally, past month cigarette and marijuana users were more likely to report binge drinking at all levels.

While most predictors were in the same direction and of similar magnitude across the three thresholds of binge drinking, a notable exception was parental education: Students with more educated parents had higher odds of 5+ binge drinking but lower odds of 15+ extreme binge drinking. In addition, Other race/ethnicity students were less likely than White students to engage in 5+ binge and 10+ extreme binge, although there was no difference for 15+ extreme binge drinking. There were few regional differences in predictors of binge drinking, although students in the Northeast and West were less likely to engage in 15+ extreme binge drinking only than students in the South. Despite these few and important exceptions, the majority of risk factors had consistent patterns of association with the three thresholds of binge drinking and extreme binge drinking.

Discussion

Our purpose was to quantify and draw attention to the prevalence of extreme binge drinking rates among the nation’s high school seniors, and to examine predictors of these behaviors. We estimate that over 1 in 10 high school seniors had 10+ drinks in a row and over 1 in 20 had 15+ drinks in a row at least once in the past two weeks. This means that among high school seniors reporting 5+ drinks in a row, over half report 10+ drinks in a row; and among those who report 10+ drinks in a row, over half report 15+ drinks in a row, all within a two-week period. In addition, some sub-groups (eg, Whites, males, students from more rural areas) show particularly high rates of extreme binge drinking. Such high levels of alcohol intake clearly put adolescents at risk for injuries and fatalities from alcohol-related motor vehicle accidents, homicide, suicide, alcohol poisoning, and drowning.⁶

Although 5+ binge drinking specifically, and frequency of drinking generally, have decreased among adolescents since record high levels in the late 1970’s and early 1980’s and have continued to decrease further since 2005,⁴⁵ 15+ extreme binge drinking has not shown such declines since 2005. This suggests that extreme binge drinking behavior may less affected by changing norms and more entrenched in specific adolescent subcultures,⁶⁸ as has been argued for the college level.⁶⁹

Many of the sociodemographic and risk factors predictive of 5+ binge drinking were similarly predictive of 10+ and 15+ extreme binge drinking. In particular, males and White adolescents engaged in more binge drinking at all levels than did females and Black, Hispanic, or Other race/ethnicity adolescents, similar to findings for lower-level drinking in prior empirical studies.^{45, 70, 71} Related behaviors and attitudes about drinking predicted binge drinking at all levels. The fact that the same risk factors predicted 5+ binge drinking and 10+ and 15+ extreme binge drinking suggests that additional, more prognostic, predictors for the different thresholds of binge drinking are needed.

We did find some predictors that varied in magnitude depending on the threshold of binge drinking. In particular, higher parent education was a risk factor for 5+ binge drinking but a protective factor against 15+ extreme binge drinking. This finding adds nuance to reports that higher SES youth are at greater risk for binge drinking,^{45, 53–55, 72} suggesting their risk is for the lower threshold of binge drinking; at the same time, lower SES youth and those from more rural areas may be at higher risk for very extreme binge drinking and concomitant consequences.

Clearly, a combination of classic binge drinking measures and assessments of extreme binge drinking is warranted to refine our understanding of such high levels of alcohol use among youth. Differentiating between levels of binge drinking, in terms of both behavioral predictors and resulting consequences, may help determine specific risks, and contribute to more effective screening and tailored intervention methods. The results of this study help to reconcile what seemed like conflicting findings—namely that reported levels of 5+ binge drinking were declining in recent years among adolescents⁴⁵ at the same time that medical emergencies involving alcohol use by teens were rising.⁷³

Important strengths of the study include the nationally representative data with the ability to examine demographic and regional differences across multiple recent years. Limitations of the study are, first, that high school dropouts were not included in the sampling frame. Because 5+ binge drinking tends to be higher among those who have early school difficulties,⁶⁰ our prevalence estimates may be conservative compared to a full population of 18-year-olds. Second, the same measures of binge drinking were used for males and females, although gender differences in average size and metabolism of alcohol mean that the same number of drinks is likely riskier for girls than boys. Finally, the available data were based on adolescent self-report of their consumption of 5, 10, or 15 or more drinks “in a row”, although an exact definition was not given. It also may be difficult for respondents to remember the number of drinks consumed at these high levels. Of course, this limitation comes with the benefits of large national samples. Future research should document the contexts of drinking, duration of the drinking occasion, and consequences experienced. Additional work is needed to assess extreme binge drinking in other populations, including young adults, and may consider a broader range of risk factors such as genetic, community, school, family, and mental health indicators to describe more clearly the etiology of the different levels of binge drinking. The documented rates of extreme binge drinking, and the fact that they have not changed across recent historical time, support the need for additional research to develop effective prevention and intervention strategies to reduce high-risk alcohol behaviors of adolescents.