The goal of the project described above was to learn more about college student alcohol use by conducting a "first-of-a-kind" BAC survey of a representative sample of students, then to use that information to develop a program to reduce excessive drinking as well as drinking-driving by students on a single campus.
The use of directly measured BAC data was particularly valuable in several ways. First, our plan to obtain BAC measurements was a strong selling point for the project with campus administrators. Although there were concerns about this unusual approach to studying student drinking, there was also great interest in having a substantial data base with objective information about student alcohol use. The perceived advantage of such information was based on a genuine desire to know more about the issue and, therefore, to be able to direct programs and resources appropriately.
Another benefit of having BAC information is that there was greater flexibility in developing the social norms message. On other campuses that have created social norms programs, program developers have been severely constrained in the kinds of information they can provide to students as evidence of normative behavior. As a short-term program, messages that state something like "When they party, xx% of UNC students have four or fewer," are perfectly appropriate. However, social norms are not based only on "poster factoids." Norms are evident in many ways in the social environment. For a program to "engineer" a correct perception of a norm, it needs to become richer, more diverse and more enduring than can be accomplished by repeating, year after year, the same basic message. Evidence available from the BAC survey can easily serve that purpose. In addition to generalized information about self-reported drinking, it is possible to use situation-specific information, as was the case in the first year of the "2 out of 3" program. Because situation-specific data were collected, it will also be possible to craft messages about drinking-driving, drinking in bars, drinking at parties – all related to BAC values or to self-reported drinking, or both. Designated driver messages can be created as well. For example, messages like "When they had been drinking, 99 out of 100 UNC students did not drive with an illegal BAC," are possible. Similar messages can, of course, be developed based on self-report data. However, because of the general nature of the questions typically asked, in combination with individual misunderstandings of impairment and DWI laws, the data may be unable to support normative messages that are as compelling as those afforded by the BAC data.
A particular strength of having BAC data was that the social norms program that resulted had the potential to be far more credible than typical norms programs. BAC data currently are noteworthy – both to the media and the target population – because they are so unique. They are unique both in their objective quality and in the way they are obtained (time-specific personal interview rather than retrospective account in response to mailed questionnaire). This is important, since a message to students (or others) that drinking is not so extensive as most everyone seems to believe is likely to be disputed. In the present case, numerous individuals who have learned about the BAC survey have been skeptical of the method, simply assuming that it was somehow defective.12 However, in contrast to the typical survey upon which social norm messages are usually based, the BAC survey is more difficult for students to question.13 Because the operation is highly visible to drinkers, they know that it is taking place. A substantial proportion of the student population (about 18%) participated in the survey and many more observed the interview teams moving around campus and saw interviews being conducted. Consequently, belief in the veracity of the "2 out of 3" fact may have increased after the Fall 1999 survey. Until the survey was conducted, no first or second year student had seen first-hand evidence of the data collection, so it was only members of the Junior and Senior classes who had known about the survey from which the "2 out of 3" information was drawn. This may help to account for the greater tendency among Seniors to believe "2 out of 3" accurately represents drinking among UNC-CH students.
Another benefit of having collected BAC data is that they provide a solid set of information against which to evaluate the effects of programs – both the social norms program described here as well as other initiatives that may occur in the future. The finding that several self-reported measures of alcohol use did not change while objective measures of drinking declined suggests that self-report data may be relatively insensitive to changes in drinking behavior. It is not clear why this would be the case. It may have to do with the fact that self-report information is probably not very precise when it comes to actual amount of alcohol consumed, even if number of "drinks" consumed can be accurately recalled and reported. For example, in 1999 students may have been drinking at a somewhat slower rate, so that the same number of drinks would not produce such high BACs. Or the alcohol content in their drinks may have been somewhat lower, or there may have been fewer instances of drinking from common containers (e.g., kegs) so the "drinks" they had were on average not so large. Whatever the explanation, it is clear that heavy drinking, as measured by BAC values, declined and that was not reflected in any of the self-report data.
Yet another benefit of the approach used here to learn about college student alcohol use is that it has helped to broaden our understanding of student drinking. It may also have contributed to the national debate about the most desirable approach to addressing student drinking. Promoting awareness of normative behavior is a relatively new approach. A more traditional approach is to use urgent messages about the dangers of student drinking, citing data that can exaggerate the extent of alcohol-related problems among students. During the fall of 1999, findings from the present study were cited by advocates of the normative focus in several forums on this issue. As noted above, reporting frightening (and sometimes misleading) statistics about student drinking may exacerbate the problem rather than help to reduce it. In the past, making the case that student drinking was a problem that needed to be addressed may have required quoting attention-getting figures. In the current climate it seems no longer to be the case that university administrators (or anyone else) are unwilling to admit that there is a problem with alcohol use on campuses in the U.S. Hence, there is less need to cite eye-catching statistics, which may inadvertently suggest the problem is larger than is actually the case.
Since we only have BAC data for the UNC-CH campus, it is not possible to address college student drinking in general. It does seem reasonable, however, to infer that the kinds of differences we found between the impression created by self-report data and that based on measured BAC data might be found on other campuses as well. It is important to recognize that we don’t consider the BAC data to be "correct," while other sources of information are wrong. Rather, these data help to broaden our understanding of student drinking, providing a richer picture than can be had only from self-report information and data concerning relatively rare incidents (e.g., alcohol-related deaths or injuries).
It should be noted that at the time this project was proposed four years ago, few thought that a campus would be willing to expose themselves to this form of data collection. Apprehension about what student BAC data would show was high among college administrators. This has changed and universities and colleges are relieved to hear the results of the UNC-CH project. It is our hope that the positive outcome this data collection has had for the UNC-CH campus will assist other campuses in deciding to collect similar data. The data and the project received accolades from the Board of Trustees and several favorable articles about the study appeared in local and national print media. Once there is a base of data from multiple campuses, researchers and practitioners will be better able to address college student drinking in general.
Another somewhat unique feature of the UNC-CH social norms program is the extent to which multiple program elements were woven together in a set of mutually reinforcing components. Several different channels were used to reach the ultimate population – undergraduate students at UNC-CH. The choice to focus initially on first year students was based on indications that they are most vulnerable, but also on evidence that they might be easier to convince with information that goes counter to commonly held beliefs. This group was then reached directly through two interconnected channels: A live, interactive presentation to convey an unusual message endorsed by students in their own words and a poster to be put up in students’ rooms, along with a financial incentive for doing so. By also making "miniature" posters (i.e., the "2 out of 3" stickers) available to all students, along with potential rewards for displaying them, the principal fact was delivered to students via another channel. Moreover, by offering financial rewards for knowing this fact, we sought to increase the proportion of students who would make an effort to retain the information about normative behavior. Several aspects of all these efforts were consciously designed to increase the likelihood that the information would reach students and be retained by them.
In an unusual approach for a university program, we also reached out to the general community with the same information about normative behavior provided to students. Because we sought to undermine misperceptions of student drinking norms that are held by the general public, not merely students, we believe it was important to do this. It is well-recognized among college and university alcohol program professionals that alumni can present a problem. For example, caught up in nostalgia they may inadvertently try to sustain the alcohol climate of their era on campus, through word and deed. By spreading the drinking norm message to alumni and the general public, many of whom are parents of present or future UNC-CH students, we believe this program is planting the seeds of future success on campus by reaching off campus into the broader community.
The responses of students surveyed during fall 1999 indicated that the program had an extremely high recognition rate (92% of first year students and 71% of the rest had heard of the program). Also, the vast majority of the students (78% of freshmen and 65% of the rest of the respondents) understood the meaning of the message, characterizing it as meaning either that drinking is less common than people think or that there is less pressure to drink. Both the degree of recognition and the extent of understanding are quite high, especially since the program had been in place for a relatively short period of time. We believe the main reasons for the success in getting this information into the population has to do with the way the program was structured, with integrated mutually supportive elements. In addition, reaching this audience was easier than others might be because the UNC-CH student population is geographically concentrated, and has one major information source to which most students attend: the student newspaper.
This degree of message recognition and comprehension is exceptional for any public information campaign. This success should be examined to identify strategies that may be of value for any public information and education program. First, we believe that the amount of effort (6 rounds of testing) that put into crafting the wording of the message was an important factor. We sought a clear and compelling single statement that also reflected the media savvy the younger audience is accustomed to seeing in modern advertising. A second factor may be the steps employed to present that message in a credible manner. For example, the presentation to incoming first year students was enhanced by the inclusion of a video of candid comments from real students about the validity of the message. A third factor may be the use of an incentive campaign in which students were rewarded for knowing or displaying the message. Incentive programs are not new to highway safety programs. Many such programs were valuable as part of successful occupant protection programs in the 1980's.
The success of any information-based program is contingent on the target population receiving, understanding, and accepting the message. In the follow-up survey, we found a substantial amount of skepticism about the initial message. There are two issues here. The non-believers are that group we would most expect to disagree, since they live in a world where drinking is common and has been for some time. Although there may be some self-justification in their unwillingness to believe the fact presented, that is probably not the main issue. Since humans tend to think the world is like what they see around them, the "2 out of 3" fact must seem way out of line to them. It doesn't represent their lives, or that of many of their friends. Or so it would seem. But the BAC distribution of non-believers in the follow-up survey was almost exactly the same as reported in the "2 out of 3" fact. So why don't they believe it? This may be because non-drinking is not nearly so salient as drinking when one thinks about the day, the night, the week. There is an abundance of research that indicates individuals are not very good statisticians, that is, that they do badly in judging the frequency with which things happen. This may simply be another manifestation of that phenomenon.
In the present case, and for other norms programs, the task continues to be to encourage students to see (and think about) things the way they really are. The lighter, less frequent, or non-drinkers are probably more inclined to accept the "2 out of 3" fact since it more closely parallels their lives. Still, many of them don't believe it either and many of the heavier drinkers do. In efforts to persuade non-believers, it is important not to oversimplify or misunderstand the lack of confidence in the "2 out of 3" fact. The variation within groups (e.g., heavier and lighter drinkers) is greater than the differences between them. It may be that the most effective way to counter non-belief is merely to sustain the approach over an extended period of time.
By stating that two out of three students return home with a .00 BAC, we are also saying that one out of three students does not. This raises the question of whether that statement may help to legitimize underage drinking, especially since most undergraduate students are not old enough to drink legally. Similarly, one might also ask what impression is given by the statement that most of the drinkers have four or fewer drinks. In fact, we chose to use the .00 BAC fact as our primary message based on feedback from students who felt the "4 or fewer drinks" message could be taken several ways.
It is important to understand that programs to emphasize campus drinking norms are not intended to replace the variety of other important elements of a comprehensive campus alcohol program. Because of its non-prescriptive nature, a social norms program can fit comfortably into a campus policy that prohibits alcohol possession on campus property, applies sanctions for alcohol-related transgressions, and provides treatment for persons who exhibit evidence of a problem with drinking (for example, by a DWI conviction). While a social norms message does not preach behavior change to the students, neither is it intended to encourage drinking. Rather, its purpose is to shake up the status quo. A desired outcome of the message is a reduction in the peer pressure felt by students who don’t drink or who drink moderately. Because it is currently believed that drinking is more common than is the case, rather than encouraging drinking, a norms message helps reduce subtle, but strongly felt pressures to drink. Another desired outcome is to reduce the comfort level of those who abuse alcohol and to help undermine their expectation that others will accommodate their behaviors. Many students are reluctant to complain about the noise, property damage and other problems that individuals who abuse alcohol inflict on those around them. This program in intended to serve the purpose of correctly placing the power of the majority in the hands of the true majority – those students who typically drink little or nothing and return home with a .00 BAC.
One of the motivations for this project was to reduce driving after drinking among college students via a reduction in drinking. One of the encouraging findings from the BAC surveys was that, despite their reputation as heavy drinkers, driving after drinking is uncommon among UNC-CH students. Years of messages encouraging persons not to drive after drinking appear to have made an impression on UNC-CH students, as they have on the driving public in general. In both the 1997 and 1999 surveys, drivers exhibited low levels of drinking at all. It seems clear then, that students on the campus studied are doing a good, if not perfect, job of separating drinking from driving. There remains room for progress, but the magnitude of this problem is less than was anticipated.
Social norms programs to reduce student drinking have shown great promise. Their conceptual underpinning is strong and, although limited, evaluation data presently available suggest that they do produce decreases in problem behaviors. It is important to keep in mind, however, that this approach is likely to produce changes incrementally rather than suddenly. Consequently, to be effective a norms program needs to endure. Haines (1996) and Johannessen et al., (1999) both report continuing declines in student drinking over a several year period. A conceptual analysis of the manner in which social norms operate suggests that is all that might be expected. However, unlike "one-shot" programs that might, with enough resources, produce a greater short-term change, the effects of a social norms program are unlikely to dissipate like those of most programs do. This is because once a more realistic conception of normative behavior concerning alcohol use has developed on a campus, it will tend to be self-reinforcing. As more students, administrators and faculty members recognize that excessive drinking is not particularly common, the tenor of discussions on a campus will change. More alternative activities not involving alcohol will become available, both through student initiative and through administrative efforts to foster an environment that is more supportive of students who do not wish to drink or to rely mainly on drinking for entertainment (e.g., greater availability of alcohol-free housing, extended hours at campus recreation locations). Subsequently, all these kinds of efforts and activities themselves become part of a social norms program. For this to occur, however, it is important to make a concerted, focused effort to promote the fact that most students behave responsibly most of the time with respect to alcohol. The various messages that the campus community and its surroundings receive about student drinking, via multiple channels, should be carefully coordinated so as to be mutually reinforcing. We believe that one such program is now in place and poised to grow at UNC-CH.
This project was an attempt to demonstrate that a technique used to study drinking and driving – the roadside BAC survey – could be modified and used productively to study drinking among college students. It is now clear that this approach is both feasible and useful. Information obtained from the survey was then used to develop a multi-faceted social norms program to reduce student drinking, as well as driving after drinking. The program clearly benefitted from having objective data on drinking to draw on. A follow-up BAC survey documented a high degree of awareness among the target population. It detected a decline in excessive drinking among students on the campus, but also found a substantial degree of skepticism among students about the norm identified by the research.
As on most college campuses, a substantial proportion of students at UNC-CH report periodically drinking several drinks on an occasion. The results of the two BAC surveys indicate that despite this, even on weekend nights, a large majority of students have nothing to drink. Moreover, among those who do drink, reaching a high BAC is uncommon. These findings are both encouraging and a cause for concern. The fact that, on a typical night, nearly a quarter of students under the age of 21 have consumed alcohol is disturbing. On the other hand, the fact that less than 3% of students had BACs in the range that reliably produces observable signs of "drunkenness,"indicates that the problem is not as pervasive as many think. The finding that impaired driving, in this reputed heavy-drinking population, is rare and was similar to that found among the general driving population in North Carolina, even before the "2 out of 3" program, is certainly encouraging.
Finally, the benefits of social norm programs to reduce reported drinking among students have been demonstrated previously. The present study provided the first opportunity to document whether, and how, actual drinking changes in response to a carefully designed and implemented program to highlight the fact that for the large majority of students, moderate drinking or abstinence – rather than excess – is the norm. Findings from the present study must be considered preliminary, but results are encouraging. Future research, using direct BAC measurement, needs to follow a social norms program over a period of several years. That will allow a more definitive answer to the question of whether the long term effects that social norm program have on self-reported drinking represent true changes in drinking, or merely changes in verbal behavior, as true drinking norms are recognized and understood by students.