|
Article Excerpt
RESEARCHERS HAVE SHOWN THAT COLLEGE students greatly overestimate the percentage of their peers who drink heavily (Baer et al., 1991; Borsari and Carey, 2003; Perkins and Berkowitz, 1986). A recent study demonstrated such misperceptions at more than 100 institutions of higher education (Perkins et al., 1999). Prevention experts have expressed concern that these misperceptions, by creating a distorted view of subjective drinking norms, might drive up student alcohol consumption (Perkins, 2003).
This perspective is consistent with the "Theory of Planned Behavior," which cites subjective norms as a key determinant of behavior (Ajzen, 1985). Subjective norms encapsulate the perceived expectations of other persons or groups who approve or disapprove of the behavior (Fishbein and Ajzen, 1975). These perceived expectations, in turn, are informed by observations of what other people say or do in particular circumstances. The process of collecting, analyzing, storing, and recalling this social information is highly prone to error (Nisbett and Ross, 1980), which can result in misperceptions of which behaviors are typical. Yet, subjective norms derived from an inaccurate assessment of people's behavior will still be an important influence on social behavior (Ajzen, 1988).
Perkins and Berkowitz (1986) called for social norms marketing (SNM) campaigns that would convey accurate information about student drinking patterns, arguing that correcting misperceptions of subjective drinking norms would drive down alcohol consumption. Several colleges and universities have reported success with campuswide media campaigns, with student surveys revealing both more accurate perceptions of actual drinking behavior on campus and concomitant decreases in reported heavy episodic drinking (e.g., Fabiano, 2003; Haines and Barker, 2003; Haines and Spear, 1996; Johannessen and Glider, 2003; Perkins and Craig, 2003). To date, however, this research has involved one-group pretest-posttest comparisons or small-scale quasi-experimental designs.
Several studies have reported failed SNM campaigns, but these are far from definitive. One study tested a residence hall-based campaign rather than a campuswide effort, as SNM proponents have advocated; moreover, the campaign lasted only 6 weeks (Clapp et al., 2003). A second study evaluated a limited campaign featuring three mailed greeting cards, each with a different SNM message (Werch et al., 2000). In a third study, a senior administrator at each of 98 institutions stated whether the campus had employed an SNM program. The researchers then compared changes in student drinking at institutions with, versus without, a campaign, but they failed to examine the scope, intensity, or quality of the purported SNM programs (Wechsler et al., 2003). Indeed, one of the included SNM campaigns is the subject of a case study report analyzing its many flaws in both design and implementation (Russell et al., 2005). The present study--a large-scale, randomized trial involving 18 institutions of higher education--was designed to meet the clear need for more rigorous research on the effectiveness of SNM campaigns to reduce college student drinking.
Method
We invited 218 institutions of higher education with a Golden Key International Honour Society chapter to participate in the study. Golden Key's national office had conducted a small pilot campaign, and its chapters could provide student volunteers to help implement the campaigns. Nearly one third of the invited institutions applied, which provided a diverse group from which to choose. The 18 participating institutions were selected on the basis of level of expressed interest, acceptance of the randomization procedure, and the absence of substantial SNM campaigns in the past. In some cases, suitable institutions were not included, because they could not be reasonably matched with another applicant institution before randomization. The Institutional Review Boards at Education Development Center, Inc. (EDC), and all 18 institutions approved the study procedure.
The 18 institutions represented all four U.S. census regions (Northeast = 11.1%; Midwest = 33.3%; West = 22.2%; and South = 33.3%); 72.2% were public institutions, whereas 27.8% were private. Undergraduate enrollments were as follows: 11.1% were from institutions with enrollment below 5,000 students; 22.2% were from institutions with enrollment at the 5,001-10,000 level; 50.0% were from institutions with enrollment at the 10,001-20,000 level; and 16.7% were from institutions with enrollment above 20,000. According to the institutions, men composed 40%-65% of the student body (mean = 47.9%). Nonwhite students composed 0%-61% (mean = 22.4%), while those ages 18-24 years old composed 45%-100% (mean = 77.6%). The percentage of residential students ranged from 1% to 100% (mean = 52.9%).
We selected and assigned 18 institutions to pairs based on the best overall matches, taking into account U.S. census region, public versus private sector, undergraduate enrollment size, and student demographics (gender, race/ ethnicity, residence, and membership in fraternities/sororities). We gave priority to the student demographic variables, which are known to be predictive of heavy episodic drinking (Wechsler et al., 2002). As a result, some matched pairs differed on certain school characteristics that are not predictive of student drinking (e.g., public vs private sector). We randomly assigned one institution from each pair to be a treatment site (nine treatment sites, nine control sites).
Administrators at the control group institutions signed an agreement to refrain from issuing SNM messages during the study period. Newspaper content analyses, activity reports, and annual key informant interviews confirmed that the control institutions did not undertake any substantial SNM campaign efforts.
SNM intervention
The treatment institutions ran campaigns for 3 academic years (Fall 2000-Spring 2003), following these rules: (1) The core message will report a normative behavior for all undergraduates and correct an identified misperception. (2) All SNM materials will include the core message, the campaign logo ("Just the Facts"), a brief description of the student survey, and the survey definition of a "drink." (3) The marketing plan will employ credible, far-reaching, and cost-effective campus media venues (e.g., posters, newspaper advertisements, emails, presentations). (4) All messages, materials, and delivery channels will be pilot-tested and then approved by the EDC-based research staff.
We used the following methods to ensure effective implementation of the campaigns: (1) an in-person training event for campus-based personnel; (2) a guidebook on how to develop, implement, and track the campaign; (3) frequent contact with site-based staff via telephone, mail, and email, including weekly campaign updates and reminders; (4) a rigorous checkpoint schedule to ensure the quality of all messages and materials; and (5) required progress reports each semester.
Core messages were based on one of two student survey questions: (1) "What is the number of drinks you consume in a week?" and (2) "When you party, how many drinks do you usually have?" Each SNM message followed this format: "[Percentage/proportion] of [Institution Name] students [have/drink] [drinking level] [when they party/per week]." Example: "67% of XYZ University students have 4 or fewer drinks when they party." Seven of the nine institutions developed a core message based on the "party" question; specified drinking levels were zero to two drinks (n = 1), zero to three drinks (n = 1), zero to four drinks (n = 4), and zero to five drinks (n = 1). Two institutions used a core message based on the "drinks per week" question; one institution specified zero to three drinks, and the other one used zero to four drinks.
We used the semester reports to calculate each institution's average weekly activity score (AWAS), a measure of campaign intensity: For each type of campaign activity, we counted the number of 7-day units per semester in which it occurred one or more times. Next, we added the total across all 18 activities and divided by the number of weeks in the institution's semester. We then calculated the mean AWAS across all six semesters of the campaign. We could not calculate overall audience reach, because most activities (e.g., posters, table tents) were not directed to students individually.
Student survey
Students at all 18 institutions completed the Survey of College Alcohol Norms and Behavior (SCANB) during the spring semesters of 2000 (baseline) and 2003 (posttest). For both administrations, registrars provided a list of all matriculated, degree-seeking undergraduates, including both full- and part-time students; those with an out-of-state current address were excluded. A random sample of 300 students per institution, stratified to produce proportional representation by class year, received the SCANB by mail (N = 5,400).
A cover letter with each survey mailing served as the informed consent document. To preserve anonymity, no identifying information was put on either the survey instrument or its stamped return envelope. Rather, with each survey mailing, students received a separate postcard with a unique code number and instructions to mail the postcard separately from the survey itself to indicate that they had completed the survey or did not wish to participate.
Contact with the students was as follows: (1) "teaser" postcard announcing the survey 2-3 days before the first survey mailing; (2) first survey mailing, sent approximately 3 weeks after the start of each school's semester; (3) reminder postcard sent to nonrespondents; (4) second survey mailing, sent to arrive 2 weeks before...
|
