Bucknell University Economics 312 Class Survey on

Drinking Habits, Attitudes, and Perceptions

Amy M. Wolaver *Comments and Suggestions Welcome


Additional analysis can be found at the following web pages:


In the wake of the Harvard Core Studies on College Alcohol Use (1993-1999), faculty and administrators, parents, and the media have paid serious attention to the incidence of alcohol use and abuse on campuses across the country. Bucknell is no exception, indeed, the data show that Bucknell has one of the highest rates of frequent, heavy drinking(1) in the United States (Napolitano, 1999). Alcohol use and dependence has both short and long run effects. In the short run, alcohol use is related to students inflicting harm upon themselves and others (Weschler, et al, 1999, 2000), in the long run, alcoholism may affect occupational choice, lower total educational attainment, reduce earnings, and affect marital status, among other possible adverse consequences (Mullahy & Sindelar, 1991).

For their major class project, the students in EC312, Health Economics, in the spring 1999/2000 semester designed and administered a survey of their fellow students in April, 2000 to try to further analyze this problem within the context of Bucknell. We sent 524 surveys to students randomly drawn from the campus directory. To encourage participation, one $100 and two $50 lottery prizes were offered to students who returned the survey by April 3. Twelve of the surveys were returned because the students had left campus since the fall. Of those possible 512 respondents, 183 students returned the questionnaire wholly or partially filled out. The surveys were anonymous; neither no one can directly tie any individual response to a particular student.

The following is a presentation of some of the main results of the survey, although the data contains more information that you may find interesting. Copies of the questionnaire and the codebook for the data in Word documents are posted on this web site to download. You can also download the data in excel format for your own analyses. The first section details some general and specific concerns about the reliability of survey data on this subject. Subsequent sections outline the reported drinking habits of Bucknell students, perceptions of the extent of drinking on campus, and attitudes about drinking and the 10 point plan.

Validity of the Data

As with any self-reported survey data, several problems can arise. First, the response rate was 36%, somewhat lower than we had hoped(2). Low response rate in and of itself only means that the information presented below is less precisely measured than we would like. A far greater worry is that the responses themselves are biased and/or untruthful. Normally, researchers in the area of substance abuse worry that young adults overestimate their habits (Weschler et al, 2000). There is good reason to believe that the opposite may be the case here. Bucknell students are well aware of Bucknell's participation in the Harvard College Alcohol Study, the results, and the policy response in the form of the 10 point plan. If students fear future reprisals as a result of the Econ312 class survey, respondents may underestimate their consumption in order to avoid further perceived onerous sanctions. Working in the opposite direction is the timing of the survey. While the questions on drinking behaviors and perceptions asked about students' usual actions, the survey went out shortly after house party weekend, traditionally a higher than normal consumption period, and a recall bias could result that inflates students' memories of their typical drinking.

Examining some of the characteristics of the students that responded to the survey and comparing them to information on all Bucknell students can help us determine if there are any obvious biases in the data. Table 1 shows averages for several demographic characteristics and compares them to published information about the Bucknell student body as a whole.

Table 1: Study Characteristics Compared to Bucknell Student Body:

Variable EC312 Data Bucknell Student Body
% Female 59.6 49%
% Freshmen 25.6 26.5%
% Sophomore 21.9 26.7%
% Junior 23.0 23%
% Senior 29.0 23%
% who reside in:



83 percent of undergraduates live on campus (including mods).
Fraternity House 12.6
Mods 7.7 8%
Off campus apartment 13.6 12%-17%
Ethnic/Racial composition(3):



Asian 4.4

12% of students belong to minority groups
Hispanic/Spanish Origin 0.5
African American 0
Other 11.6
% in a Fraternity or Sorority 45.9 41%
% of non-Greeks who plan to or who have thought about rushing (N=102) 42.1 - an additional 23% of total sample, making estimated rush statistics 69% of students A range of 62-78% sophomores in 1997-1999 registered for rush
Grade point average between:

2.0 and 2.5


Bucknell average gpa:

3.2 in 1997

2.51 and 3.0 17.5
3.01 and 3.5 47.5
3.51 and 4.0 29.5

Sources: Bucknell Facts, Committee on Greek Life, Media guide.

The sample slightly over-represents Greeks, whites, women, and seniors, but with the exception of females, the sample averages are within a few percentage points of the actual population means. Nonetheless, for the reasons outlined above, care should be taken in interpreting the results given below.

Drinking Habits of Bucknell Students

Table 2 shows the basic drinking habits of students currently, at the beginning of their freshmen year, and during their senior year in high school. Drinking habits were assessed using categories for "usual number of drinks" at each point in time and the number of nights per week the student typically drinks. As in the Harvard study, large percentages of Bucknell students are relatively heavy drinkers. The data are not directly comparable with the Harvard study for several reasons. First, the students determined that broader categories would encourage better responses than open-ended questions. Second, the Harvard study bases the definition of binge drinking on respondents' behavior in the two weeks prior to the survey, not the students usual reported behavior. If our study had used the same format, we may have had more students falling into the binge drinking category. Since the Bucknell data overrepresents females and females tend to drink less than males, the numbers reported below may also be smaller than the rates we are familiar with from the Harvard study.

Despite these differences, the data are on the order of the Harvard College Alcohol Study findings for Bucknell. In the 1997 survey, 67% of Bucknell students were defined at "binge" drinkers. If one assumes that half of the respondents in the 3-5 drink category would meet the minimum Weschler, et al definition of binging (4 drinks for women and 5 for men in one sitting), roughly half of the students who responded are currently "binge drinkers."

Table 2: Drinking Habits of Bucknell Students, Spring 2000

Percent of students who report:
Usual Drinks in an evening currently Usual Drinks in an evening beginning of freshmen year Usual Drinks in an evening senior year in high school
Abstain/no drinks 17.8 25.6 40.6
1-2 Drinks 14.4 13.3 15.6
3-5 Drinks 35.6 28.9 27.8
6-9 Drinks 23.9 26.6 12.8
10+ Drinks 8.33 6.7 3.33
Number of nights per week usually drink currently Number of nights per week usually drank beginning of freshmen year Number of nights per week usually drank senior year in high school
0 Nights 21.3 27.9 49.7
1 Night 23.0 18.0 33.9
2 Nights 30.6 24.6 11.5
3 Nights 18.0 23.0 4.4
4 or more Nights 7.1 6.6 0.6
Average nights drink per week* 1.66 1.62 0.72

Source: Author's calculations from Econ312 Survey.

* Assigns a value of 4 nights for the 4 or more nights per week category.

These simple statistics show several patterns. First, drinking on average steadily increases from the end of high school throughout the college years. The percent of abstainers drops from 40.6% to 25.6% at the start of college, and then further drops to 17.8%. The percentage of students who report moderate consumption of 1-2 drinks per evening remains fairly steady, while the percent of students drinking 3 or more drinks per evening increases. Similarly, using the measure of nights the student typically drinks during a week, consumption increases, an average of almost one night per week from high school to Bucknell.

Bucknell students' high school drinking habits do not differ significantly from the reported national averages-again if one assumes that half of the students in the 3-5 drinks per evening category would count as a binge drinkers, roughly 30% of Bucknell students binged in high school, comparable to national statistics from the National Institute for Alcohol Abuse and Alcoholism (NIAAA) from 1997. Even though Bucknell students arrive with similar drinking habits, they soon surpass their national peers.

The average differences in drinking over time hide changes within the distribution; not all students increase their intake of alcohol; indeed, some students lower their intake over the course of their college years. Table 3 shows the distribution of the change in drinking behavior over two time periods: senior year in high school to beginning of freshmen year and beginning of freshmen year to the time of the survey.

Table 3: Changes in Drinking Habits of Bucknell students over time

Percent of students who
Usual Number of Drinks in a Typical Evening Usual Nights Drink Per Week
Between high school and beginning of freshmen year Between beginning of freshmen year & survey date Between high school and beginning of freshmen year Between beginning of freshmen year & survey date
Decrease drinking habit 5.6 18.9 3.9 25.1
Do not change drinking habit 53.3 54.4 41.0 45.9
Increase drinking habit 41.1 26.7 55.2 29.0

The majority of students did not report increasing the number of drinks they typically consume in an evening over either time period, but very few- 5.6%- reported decreasing their consumption upon arrival at Bucknell. The good news is that a significant minority of students-18.9%- have reported a decrease in their consumption over the course of their study at Bucknell. However, a larger minority-26.7%- continues to increase their consumption in between the beginning of their freshmen year and the survey date. The patterns are similar for the frequency of drinking, although there are somewhat higher increases in this category.

Intensity of Drinking Behaviors

Further analysis shows that the students who consume the most drinks per night consume more often than their more temperate peers. Table 4 shows the frequency of drinking in number of nights per week, given the number of drinks typically consumed in one sitting.

Table 4: Frequency of Consumption by Typical Drinks per Evening

0 Drinks in an evening 1-2 drinks 3-5 Drinks 6-9 Drinks 10 or more Drinks
0 Nights 96.9 15.4 4.7 0 0
1 Night 3.1 46.2 35.9 11.6 6.7
2 Nights 0 34.6 39.1 41.9 20.0
3 Nights 0 0 18.8 32.6 46.7

4+ Nights

0 3.9 1.6 14.0 27.7
Average nights per week* .03 1.31 1.77 2.49 2.93
Sample size 32 26 64 43 15

*Figured assuming that highest category is 4 nights per week.

Most students (61.6%) who report 1-2 drinks or fewer also report drinking only once a week or less. Conversely, the heaviest drinkers are also the more frequent drinkers. Students who typically consume 6 or more drinks all report drinking at least one night a week, and the vast majority of these students drink 2 or more nights per week. Most alarmingly, three quarters of the most intense drinkers, those who drink 10 or more drinks in an evening, also report drinking three or more nights a week. While this category represents only 8.3 % of the entire sample and should not be viewed as representative of a large group of the student body, the overall figures indicate that the frequency and intensity of drinking are positively related on this campus.

Is "Binge Drinking" Defined Correctly?

The Harvard Study authors have shaped the public discourse about college drinking with their definition of binging as 4/5 drinks or more in a sitting for women/men, but some question the validity of this assessment (Weschler et al, 2000). Their definition is based on the point at which problems increase for drinkers. Bucknell students are no exception, shown in Figure 1.

The data show patterns that are consistent with the Harvard study definition- the biggest increases in the fraction of students experiencing most of the problems come at 3-5 drinks. Hangovers, falling behind in classwork, and doing something you regretted are possible exceptions, with large increases coming at 1-2 drinks per sitting, but the overall pattern supports the Harvard definition. The drinking style of heavier drinkers also follows this pattern, with increasing percentages of students answering that "getting drunk" is an important/very important reason to drink as the response to usual drinks per night increases. Figure 2 shows that relationship.

More disturbing is the relationship between the number of drinks that a student reports is necessary to "get drunk" and their definition of binge drinking. Only 27% of students define binge drinking as equal to or less than the number of drinks it takes them to "get drunk," and as students' self-perceived tolerance for alcohol increases, so does their definition of what binge drinking is. Figure 3 displays the relationship between these two variables:

Views, Beliefs and Attitudes

The above figure leads to the final topic of analysis: students' views and attitudes about drinking on campus; information about students' knowledge of and opinions about the 10 Point Plan follows this section. One of the primary objectives of the data collection was to learn more about the unique situation at Bucknell. Table 5 outlines some of the responses to various questions about knowledge and attitudes.

Table 5: Views, Beliefs, & Attitudes of Bucknell Students, April 2000

Percent of Students Who Indicate:
There are enough non-alcoholic options on campus 42.5
Uptown has contributed to reducing consumption & providing a satisfactory alternative activity 37.4
Binge Drinking is:
Under 4 Drinks in on Night 1.7
Between 5 and 7 Drinks 33.7
Between 8 & 10 Drinks 35.4
More than 10 Drinks 29.3
Typical Number of Nights Respondent Binges, under respondent's own definition of binging
0 Nights per Week 66.7
1 Night per Week 20.2
2 Nights per Week 8.7
3 or more per Week 4.4
Have experience pressure to drink from the following sources:
Greek Hazing 7.1
Hazing within another organization 13.1
Friends, roommates, hallmates 60.1
School work pressures/stress 32.2
Relationship problems 18.6
Atmosphere of party 66.7
Celebration 64.5
Agree that Bucknell students overall have a drinking problem 36.3
Fraction of respondents that believe X% of all Bucknell students drink heavily
0-20% 11.5
21-40% 38.5
41-60% 33.5
61-80% 14.8
81-100% 1.65
Fraction of respondents that believe X% of my friends drink heavily
0-20% 43.2
21-40% 21.4
41-60% 14.8
61-80% 17.3
81-100% 4.4

One oft-repeated student explanation for the high rates of binging on campus is the relative isolation of campus and lack of other activities. However, 42.5% of Bucknell students feel that there are enough non-alcoholic activities on campus, and 37% believe that Uptown has contributed to campus, somewhat easing concerns on this score. The student paper on alternative activities has more analysis on this topic.

Very few students report hazing in Greek or other organizations, although it is entirely plausible that respondents would be more reluctant to report these activities than others. Considerable peer pressure to drink is noted in the response to the friends, roommates, and hallmates and the "atmosphere of party" response.

Bucknell students have different beliefs about what constitutes "binge" drinking, and under their own definitions, a third of students binge at least once weekly. One proposed policy to decrease drinking is social norming. Most students nationally over-estimate the fraction of their peers who are heavy drinkers and may feel more peer pressure to consume because of these beliefs. How do Bucknell students perceive the drinking habits of their peers? Comparing this figure to their perceptions about the rates of heavy or problem drinking on campus provides a mixed picture. Thirty-six percent of respondents feel that Bucknell students overall have a problem with drinking; 33.5% of students correctly estimate (based on the current survey) that between 41-60% of all BU students are heavy drinkers while 16.4% over-estimate the fraction of heavy drinkers.

Asking students about their reasons for drinking can give further insight into the drinking "style" on campus. Table 6 shows the percentage of students who feel that each response is a important or very important reason to drink alcohol.

Table 6: Percent of Bucknell Students Who Rate Various Reasons to Drink as Important or Very Important

To Escape 13.0
To Relax 40.8
To Get Drunk 24.2
To Have a Good Time 57.9
To Relieve Boredom 16.9
To Celebrate 54.9
For the Taste 22.6
As a Reward 18.0
To Fit in 6.8
For Ease with Men/Women 17.5

A quarter of students indicate that "getting drunk" is an important reason to drink. The most often cited important reasons are "to have a good time" and "to celebrate." The former is interesting to note, given the high percentage of problems experienced on campus as a result of the respondent's own drinking and the drinking of others around the respondent. Students do not respond that relieving boredom is an important/very important reason to drink, which suggests that alternative activities are not the only initiative needed to reduce binging at Bucknell.

Students were asked to rate the importance of various reasons to choose to limit or not to drink. Table 7 shows the percent of students that rated each reason as important or very important:

Table 7: Percent of Students Rating Each Reason Not to Drink or to Limit Drinking as Important or Very Important

While on a date 50.3
While Driving 96.7
Cost 43.0
Fear of Being Caught/Legal Consequences 33.5
Cutting Down 5.6
Athletics 35.8
Alcohol Problems 7.3
Fattening 30.2

Only 7.3% of students indicate that having problems is an important reason for them to cut down, despite the fact that many indicated elsewhere that they had experienced several problems. Thirty-three and a half percent of students indicate that the fear of being caught is a determinant. One suggestion from a classmember was that the legal consequences were often fines, that were subsequently paid by the parents of offenders. The drinker often do not experience any serious discomfort as a result of their behavior. My class suggested that community service punishments (especially ones assigned for early Saturday mornings) might be a more effective deterrent. Athletics was only important to 35.8% of students, but not all students are athletes, so this estimate probably underrates the effectiveness of the athletic department in limiting drinking.

How do students perceive the acceptability of abstinence and how does drinking fit in the social life on campus? Table 8 outlines student responses that pertain to these questions.

Table 8: Percent of Bucknell Students that Agree/Strongly Agree with Following Advice for New Freshmen

Non Drinkers are Admired 18.8
You must Drink to fit in 30.3
Drinking is Important 21.6

Clearly, drinking is an important part of the social scene on campus, and few students perceive that abstention is considered an admirable trait among their peers.

Awareness and Opinions About the 10 Point Plan

Information was collected about students awareness and perception of the 10 point plan. First, students were simply asked whether they had heard of and whether they had read the 10 point plan. They were then asked to mark which goal they felt best described Bucknell's policy toward alcohol and to rate the effectiveness of accomplishing this goal. A (slightly) abbreviated version of the 10 point plan was reproduced and a further set of questions was asked about the effectiveness of several of the points in the policy and whether they had received any of several forms of alcohol education. Table 9 presents the answers to these questions.

Table 9: Bucknell Students' Awareness and Opinions about the 10 Point Plan

Percent of students who had heard of the 10 point plan 52.2
Of those who were aware of the 10 point plan, percent who had read it 56.5
Percent of students that responded that each of the following goals best described Bucknell's alcohol policy
Discourages alcohol use 51.9
Tolerates alcohol use 14.4
Encourages responsible drinking 11.1
Does little to discourage use 9.9
Prohibits all alcohol use 3.3
Prevents minors from drinking only 9.4
Bucknell is effective/very effective at accomplishing above goal 29.8
After reading the 10 point plan
Percent of students that agree/strongly agree with mission & intent of policy 44.2
Percent of students that agree/strongly agree that options have improved since inception of policy 31.3
Percent of students that agree/strongly agree that the university has successfully reduced problem drinking in the Greek system 15.0
Percent of students that agree/strongly agree that they are more aware of the problems associated with binge drinking 36.8
Percent of students who have attended/seen the following alcohol education materials
Mailing 50.0
Posters 55.0
Student Newspaper 33.9
Formal course on alcohol education 22.8

A slim majority were aware of the plan, but only 20.5% of students believe that the main intent of the plan is to encourage responsible drinking and to prohibit minors from drinking. Overall, students accept the 10 point plan's ideas; after reading the plan, 44% of students agreed or strongly agreed with the mission of the policy (an additional 33.7% were neutral, leaving only 22.1% who disagreed with the plan). The 10 point plan scored the lowest on the effectiveness of reducing problem drinking within the Greek system. Most students have been exposed to at least one form of educational material about drinking-only 25% have not attended a formal course, or seen a mailing, poster, or something in the student newspaper. Half of students were exposed to more than one of the above educational materials.


The above discussion touches on only some of the data available from this survey. You can find more Bucknell-specific analysis by following these links:

The information presented clearly indicates that Bucknell continues to have high rates of heavy drinking, under any definition. The rates are higher than national averages, despite students arriving on campus with previous drinking experiences similar to all college students. Students here have a much higher threshold for what they feel is "binge" drinking than do academic researchers. On this note, perhaps the most striking result is that a vast majority of students do not believe that problems are an important reason for cutting down on their drinking, even though they report many and multiple problems from their own and others' drinking. Educating students on the nature and frequency of problems associated with heavy drinking is one possible avenue for future efforts at reducing problem drinking on campus.

Additional Sources

Most of the findings of the Harvard College Alcohol Study are posted at: http://www.hsph.harvard.edu/cas/ Several are referenced below. The National Institute for Alcoholism and Alcohol Abuse also has several publications on line.

"Ten Point Plan to Reduce Alcohol Abuse at Bucknell" available at: http://www.orgs.bucknell.edu/bucknellian/sp99/3-18-99/news/3949.html

Mullahy, John, and Sindelar, Jody. 1991. "Gender Differences in Labor Market Effects of Alcoholism," Papers and Proceedings, American Economic Review 81(2): 161-165.

Napolitano, Heather. 1999. "Adams Tackles Student Alcohol Problems," Bucknellian Online News. Available at: http://www.orgs.bucknell.edu/bucknellian/sp99/3-18-99/news/5950.html.

National Institute of Alcohol Abuse and Alcoholism. 1997. "*" Alcohol Alert, Bulletin 37*. Available at: http://silk.nih.gov/silk/niaaa1/publication/aa37.htm.

Weschler, Henry, Dowdall, George W., Maenner, Gretchen, Gledhill-Hoyt, Jeana, and Lee, Hang. 1998. "Changes in Binge Drinking and Related Problems Among American College Students Between 1993 and 1997," Journal of American College Health 47: 57-68.

Weschler, Henry, Lee, Jae Eun, Kuo, Meichun, & and Lee, Hang. 2000. "College Binge Drinking in the 1990s: A Continuing Problem," College Alcohol Study Report 2000 #2, available at: http://www.hsph.harvard.edu/cas/test/rpt2000/CAS2000prt2.html.

1. Defined as 5 drinks in one sitting for men, 4 for women, more than once per week on average (Weschler, 1999*).

2. The Harvard Alcohol Study achieved an average response rate of 60 to 70% in 1993 and 1997. Their higher response rate may be due to several factors. The study sent follow-up post cards and offered larger cash prizes than the EC312 study. Bucknell students may be more reluctant to answer questions about their drinking behaviors for the same reasons outlined above.

3. Since the sample size is relatively small and since none of the student project proposals focused on race, we did not employ a commonly used technique known as over-sampling to assure us of a representative sample of any one racial group. Unfortunately, this resulted in no representation of African American students in our sample. It is also possible that some students did not feel comfortable with the standard classifications given and chose another category.