Binge Drinking Among College Students: A Comparison of California With Other States

College students' alcohol consumption, especially binge drinking, continues to be a serious problem for institutions of higher education (IHEs). Studies consistently report that from 86% to 95%of all college students drink alcoholic beverages.^1,2 Alcohol use by college students is associated with academic problems^3,4 emotional problems, violence, and physical injuries.^4,5 Nearly half of college freshmen rode in a car with someone legally drunk.³ Binge drinking is identified as the Number 1 health hazard for American college students.^3,4,6,7 Other research suggests that the wide availability and accessibility of alcohol makes alcohol a "gateway" drug that leads to the use of other dangerous drugs.⁸

The Harvard School of Public Health College Alcohol Study examined the nature, extent, and associated problems of heavy episodic or "binge" drinking.^4,6,9 Although such national-scale data are important to public health policy makers, persons involved in developing programs to prevent or reduce the impact of alcohol-related problems need more specific information about characteristics of their target population if they are to measure the impact of their programs. In addition, an analysis of national-level data, by region, may shed light on factors associated with excessive alcohol use.

In this article, we report the results of a California-specific analysis of the College Alcohol Study. We were interested in answering the following questions: Are patterns of binge drinking among California college students different from those found elsewhere? If so, in what way or ways? And what are the implications of these differences?

METHOD

The approach we used for the College Alcohol Study (CAS) has been detailed in several publications.^4,6,9 Briefly, the CAS examined the drinking behaviors in a sample of 195 4-year colleges selected as representing a cross-section of American higher education. Of the 140 campuses that ultimately participated in the study, 15 were located in California. We mailed a 20-page survey instrument to randomly selected students at each university; the questionnaire items addressed students' drinking behavior, including frequency of binge drinking, and problems the students experienced as a result of their own drinking and the drinking of other students. The overall response rate for the survey was 69%.

For the survey, we defined a drink as a 12-ounce can or bottle of beer, a 4-ounce glass of wine, a 12-ounce bottle or can of wine cooler, or a shot of liquor taken straight or in a mixed drink. The study adapted a measure of heavy episodic, or binge, drinking used in other national studies of alcohol use by adolescents and young adults. Previous national college studies have typically used five or more drinks in a row on at least one occasion in the past 2 weeks as the measure of heavy episodic drinking.^10,11 The cutoff for women was adjusted to four or more drinks so that the expected dose effects would more closely approximate those of men.⁷ Alcohol problems related to binge drinking include students getting into arguments, getting injured, forgetting where they were or what they did, missing classes, and engaging in unplanned or unprotected sex.

For this study, we compared data gathered from California students with that gathered from students attending colleges in 39 other states and the District of Columbia. The responses from the 15 California campuses represented 1,864 student responses and was the largest subset of data from any state in the survey.

We used chi-square comparisons of California colleges with all other colleges to determine the extent of differences between the California sample and all other campuses and used a stepwise multiple logistic regression technique (SAS PROC LOGISTIC) to determine whether the students in California had a lower risk of binge drinking after we adjusted for possible confounding factors. The model for this analysis included select demographic variables and the use of tobacco as a high-risk behavior, all of which could potentially explain the differences in rates.

In the model, all variables were dichotomized. The demographic variables were age (under 24 years of age = 1, over 24 = 0); sex (male = 1, female = 0), White (White = 1; other ethnic groups = 0). The high-risk behavior of smoker (smoker = 1, and nonsmoker = 0), noncommuter (lives on campus = 1, elsewhere = 0), and never-married (never married = 1, otherwise = 0). We also considered two-way interactions by crossing California and the main effects.

RESULTS

National Sample

One hundred forty of the 195 campuses solicited (72%) responded; two thirds of these were public, one third private, and 4% were historically Black institutions. Approximately two thirds were located in suburban/urban settings and one third in small town or rural settings. We received responses from 17,592 students, 84% of whom had consumed an alcoholic drink during the previous school year. Nearly half (44%) of all students were binge drinkers, and 19% were frequent binge drinkers (binged three or more times in the past 2 weeks). Participation in binge drinking ranged from 1% of respondents on some campuses to 70% on others.

California Sample

Demographics and Smoking Variable

Our first task was to compare characteristics of students sampled in California with those from the national survey. We found several statistically significant differences between California and the remainder of the nation in the proportions of students who were under age 24, were White, were smokers, were unmarried, and were living in campus housing (Table 1). In general, the sample of California college students tended to be older, more likely non-White, more likely married, and less likely to live on campus.

Alcohol Consumption and Binge Drinking

Overall, California college students reported lower levels of alcohol consumption than the levels from the combined group of other states (Table 2). California colleges had more students who abstained from alcohol than other states had (19.2% v 15.3%). Among students who had had a drink in the past year, fewer California students (13.4%) reported that they drank on 10 or more occasions in the previous 30 days, compared with those from all other colleges (18.1%). In addition, the California students who had been drunk three or more times in the past month were fewer in number (12.2%) than students in other states (23.9%). Among students who had drunk in the past year, fewer of those from California colleges indicated the reason for drinking was that they "drank to get drunk" (32.6% v 39.7%).

Fewer California students reported binge drinking at all (29.9% v 45.2%), and they were less likely to report binging three or more times (10% v 20.6%). Among the California students who had had a drink in the past month, fewer "usually binged" when drinking (21.5%), compared with students from all other colleges (41.8%). Fewer California students had been binge drinkers in high school, before coming to college (24.5% v 32.5%). And among Californians, a lower percentage of nonbingers in high school had begun to binge drink in college (21% v 33%).

Adjusted Ratios for Binge Drinking
and Selected Variables

Although binge drinking in California appears to be less common on the surface, we found that after controlling for demographic variables and for smoking as a risk factor, the decrease in reported binge drinking in California was not statistically significant when compared with the remainder of the nation (p = .3534). Among the factors assessed in the multivariate analysis (MANOVA), age and marital status showed greater changes in magnitude of the relative risk, whereas the remaining variables showed very little change. This finding indicates that the most important contribution to the lowered relative risk of binge drinking among California students is attributable to the fact that the California students generally tended to be older than other students in the national survey (Table 3).

Problems Related to Drinking

California students who had had a drink in the past year reported fewer problems caused by their own drinking than did students at colleges outside of California (Table 4). California students were less likely to report problems affecting educational, psychological, or social functioning and safety. About half as many students in California colleges reported having had five or more problems, excluding hangover but including drinking and driving (9.8% v 16.8%).

Drinking and Driving

Drinking and driving data for California students also differed from those of the rest of the country (Table 5). Students from California were slightly less likely to report having driven after having five or more drinks (6.0% v 9.6%) and having been passengers with a driver who was "high or drunk" (12.8% v 19.1%). We found no statistically significant difference between California college students and those from the rest of the nation who reported having driven after drinking during the previous 30 days (25.1% v 26.7%).

Secondary Effects of Drinking

The problems students had experienced as a result of other students' drinking were less often reported by California students than among the national sample (Table 6). Fewer California students than those from other areas surveyed had had a serious argument or quarrel; been pushed, hit, or assaulted; or had had their property damaged. In addition, fewer students in California than those from the combined group of other states had experienced an unwanted sexual advance.

DISCUSSION

A number of factors that may affect the validity of the findings should be noted. A possible limitation of the study is the use of self-reports in measuring binge drinking; however, a number of studies have established the validity of self-reports of alcohol and substance use.^12-14 Findings indicate that if a self-report bias exists, it is largely limited to the heaviest use group¹⁵; thus, possible bias should not affect such a conservative estimate of binge volume as five drinks. Furthermore, the heavy drinking rates found in the national study (44%) correspond very closely to those found in other national studies (41%).^10,11

Nonresponse of binge drinkers was another area of potential methodological weakness. We assessed this possibility by comparing early and late responders to the survey.⁴ No significant differences were found in the percentages of nondrinkers, non-binge-drinkers, and heavy drinkers when we compared early and late responders. In addition, when a short form of the questionnaire was sent to nonresponders, their binge drinking rate was comparable to that of students who had responded to the full-length survey.

One hundred forty of the 195 colleges that were asked to do so participated in the Harvard study. Colleges that did not participate did not differ from those that did in terms of location, commuter status, public or private enrollment, or size of the community. The only difference that we found was enrollment size-colleges with fewer than 1,000 students were less likely to have participated.

The data we present in this study suggest that demographics, such as age, marital status, and commuter status, account for why the California students drank and binged less frequently, exhibited fewer problems associated with heavy episodic drinking, and experienced fewer secondhand effects from binge drinking than did their counterparts nationwide. The findings from this study provide a striking example of the need to consider local demographic factors in planning for college health services. Older, married, and commuting students have health needs different from those of younger students who live on campus; programs should appropriately segment the population to be served.

In addition, these findings suggest that a significant proportion of college students in California do indulge in alcohol in ways that place their own health (and the health of others) at risk. Even though a larger proportion of California college students were older, non-White, married, and living off campus, the comparability between California and the remainder of the nation when these factors are controlled suggests a serious problem within the younger, on-campus, nonmarried, and more-predominately White students.

Finally, one of the most important aspects of this study is the compelling demonstration that national findings may not reflect local reality. This study seems to confirm what we intuitively know: Regional differences between California and the rest of the nation exist, even in this area of alcohol-related health risk behaviors. Although this effect is attributable to demographic variables, it does highlight the fact that college populations represent diverse settings that do not reflect the common media stereotype of college students as a homogeneous group of post-high-school young adults. Thus, developing an ongoing national system of surveillance of college students' risky behaviors is imperative. Such a surveillance system, however, must be coupled with corollary surveillance systems at the state and local levels if college health professionals are to make programmatic decisions with intelligence and precision.

REFERENCES