Alcohol and Other Drug Use Trends at the University of Pennsylvania

Survey Data from the Office of Health Education, 1999-00 to 2001-02 AY

For the past three years, the Director of Alcohol Policy Initiatives in conjunction with the Office of Health Education has conducted web-based surveys of alcohol and other drug (AOD) behaviors and attitudes among Penn undergraduates.  The surveys are designed to provide university stakeholders with AOD information in comparison to national datasets and in relation to health issues and academic effects on students.  This report presents selected data from the past three years in order to examine trends in key variables.

Surveys are administered each spring semester in early March to avoid the spring break holiday.  All full-time undergraduates are invited to participate in the survey administration and in 2000 and 2001 approximately one third of the group responded.  In 2002, the response rate for the dropped to approximately 25% for as yet unknown reasons.  Each year, despite the recent drop in response rate, the respondent pool has been highly representative of the undergraduate population as a whole when compared with data from Penn’s Office of Institutional Research.

Fig 1. Binge Drinking & Abstainer Rates Over Time
University of Pennsylvania

Trends in Key Alcohol Use Variables

In addition to Penn’s own surveys, data for Penn students is available from the Harvard College Alcohol Study, in which Penn has participated since 1993.  As a result, on certain variables, specifically binge drinking, data are available for nearly the past decade (Fig.1). 

T-tests for independent samples indicate that since 2000 year-to-year decreases in binge drinking rate are statistically significant at p<0.01, however, the proportion of students who report that they do not drink alcohol (abstainer rate) remains statistically unchanged. 

For the past three years, Penn’s surveys also have tracked how much alcohol students drink in general when they go out, as opposed to the very specific binge drinking measure.  Another focus of Penn’s efforts has been to closely monitor how students perceive the amount that their peers drink.

As is clear from Fig. 2, there have been no statistically significant changes in these variables since the 1999-00 academic year.  While many schools report that their students overestimate the amount their peers drink, this does not appear to be the case at Penn based on these measures.  In fact, when non drinkers are excluded from the measure of self-reported drinks per occasion we see that Penn students have been fairly accurate (within ½ a drink or less) at estimating how much other students drink.

Normative Advertising

One of the key ways that Penn has tried to impact students’ perception of their peers’ drinking is through its normative advertising campaign.  Logistic regression analysis of 2002 data does not indicate any statistically significant relationships between exposure to ads and binge drinking behavior or self-reported drinks per occasion.  Data from the 2002 survey also examined how students reacted to the ads as shown in Table 1 (p.6).

There is an ongoing argument in the communication literature as to how much of a role ad characteristics, individual viewer characteristics, and saturation of the environment play in creating exposure to advertising.  Certainly it is a combination of all three; however evidence from the government’s evaluation of the National Anti Drug Campaign (analyzed by a working group at the Annenberg School for Communication) suggests that simple environmental saturation is critical, especially where simple recognition is concerned.  The data above, which reflects simple recognition of each of Penn’s ads, appears to be consistent with this pattern.  Ads placed in multiple locations using multiple media (“flower power” and “tattoo”) have statistically significantly higher (p<0.05) recognition and attention-getting numbers than do those ads that were placed in only a single medium. 

Another important consideration is the fact that the “flower power” ad recognition across placements and attention-getting numbers are statistically significantly higher (p<0.05) than those for the “tattoo” ad, even though both ads were placed in three outlets and the “flower power” ad was never placed in the Daily Pennsylvanian (DP).  This suggests that individual ad characteristics, likely the bright colors, led to source confusion and the overall higher recognition for this ad.

Trends in Negative Consequences of Alcohol Use

Aside from actual measurement of drinking behaviors, several measures of how drinking affects students are included on Penn’s web-based surveys.  Since the 1999-00 school year, some of these indicators have shown statistically significant declines, as shown in Table 2 (p.7).  Related to individual wellbeing, the proportion of students who report having been hung over, having had a memory loss/blackout, and having been in a verbal argument as a result of their alcohol use all have significantly decreased.  As far as academic impact, the proportion of students who report missing a class to due their alcohol use during the academic year has significantly decreased since 1999-00. 

Other Drug Use Trends

Self-reported use of other drugs among survey respondents has shown no statistically significant change since 1999-00, as is evident in Fig. 3.  Marijuana continues to be the drug of choice for respondents, behind alcohol.  As Penn’s drug data has been consistent over time, it also is consistent with national estimates of drug use on campus (Table 3, p.8).

Conclusion

Data from the past three years of surveys at Penn indicate that on key usage indicators for alcohol and other drugs the university could be considered average.  Penn’s current binge drinking rate matches the national average of 44% as reported by the Harvard College Alcohol Study, and Penn’s drug use rates also are consistent with national averages as reported by the Monitoring the Future project from the National Institute on Drug Abuse.  So, while Penn’s challenges in the area of alcohol and other drug use among students are no worse than at many other universities, there remains work to be done.

Table 3. Drug Use Trends at Penn Compared to National Data

A: Excludes crack cocaine (prevalence = 1%)           B: Excludes LSD (prevalence = 4%)
C: Ecstasy only                                                             D: Excludes heroin (prevalence < 1%)

Note: Harvard and NIDA data ask about use in past 12 months, Penn surveys ask about use in academic year.