Bumps and Bruises

In 2006, Dejong et al. released a study that examined social norms marketing campaigns on university campuses. The study compared universities using a social norms campaign to a control group of universities. The results of the study showed that the universities who were using a social norms program had either no increase or a very small increase in perceptions of drinking norms, actual consumption, or harm relating to alcohol use; the control group of schools had substantial increases in all three. This study helps to show that the changes we have seen at MSU are not just the result of some secular trend in lowering alcohol consumption.

Despite all of the successes reported above, there have also been bumps and bruises along the way. One particular problem manifests itself in the results between our 2004 and 2005 surveys. In 2004 we found that the average number of drinks consumed was 5.3; in 2005 that number had dropped to 4.5. Those numbers do not necessarily sound bad. However, when we examined what might have caused such a drastic difference we found a few problematic methodological elements. The survey we administered in 2004 was the NCHA survey, which provided a definition of what was considered one drink. The 2005 survey, however, did not define what was considered a drink. When we later conducted some cognitive interviews with focus groups, we found that in the absence of a definition, students were considering a drink to be one container (e.g., a solo cup) rather than a particular amount (i.e., a twelve ounce can).

Another problem that manifests itself through our focus group interviews was that our language was not matching up with our students’ language. Students had different definitions regarding partying vs. socializing. For students partying and socializing are not one and the same. Partying means that alcohol was consumed; socializing does not. On account of this definitional inconsistency, there is a large measurement error regarding the number of drinks consumed at the last social event. In our 2006 survey, the percentage of students who stated that they had one or more drinks the last time they partied/socialized was around 84%. We also asked and additional question: “Since the last time that you ‘partied’/socialized was there another more recent occasion when you got together with friends and did not drink alcohol?” 85% of the students answered the previous question with a “yes.” When taking those numbers into account we get a drastically different estimate for the percentage of students who did not drink the last time they socialized/partied; that number increases from 16% to 88%.

The preceding examples are given to demonstrate how important the crafting of messages and surveys can be. It is incredibly useful and informative to involve focus groups from the very beginning and to continue their use throughout. Also, the evaluative research you conduct in addition to providing you with measurable change can also stir new research questions. Recently, we have begun looking at different personality types. One personality type that is particularly interesting to us is the thrill-seeking personality. In preliminary investigations we have found that risk-taking and thrill-seeking is positively correlated with several risky drinking behaviors and negatively correlated with many protective behaviors. Further research needs to be done, but we are very interested and concerned about how this might affect our normative messages.