A Health Communication Intervention To Reduce ...

Glassman, T., Haughton, N., Wohlwend, J., Roberts, S., Jordan, T., Yingling, F., & Blavos, A. (2013). A health communication intervention to reduce high-risk drinking among college students. Journal of Student Affairs Research and Practice, 50(4), 355–372. doi:10.1515/jsarp-2013-0026 Available at http://dx.doi.org/10.1515/jsarp-2013-0026 Innovations in Research and Scholarship Feature

A Health Communication Intervention To Reduce High-Risk Drinking Among College Students Tavis Glassman, University of Toledo Noela Haughton, University of Toledo Jennifer Wohlwend, Lourdes University Stephen Roberts, University of Toledo Timothy Jordan, University of Toledo Faith Yingling, Bowling Green State University Alexis Blavos, University of Toledo

This investigation examined the effect of a health communication intervention on the alcohol consumption patterns of first-year college students. Researchers employed a quasi-experimental design consisting of students in three residence halls at two Midwestern universities. Between-group comparisons revealed students receiving the intervention drank less and experienced fewer alcohol-related consequences. A systematic high-dose health communication campaign represents a promising prevention strategy to implement on college campuses.

Tavis Glassman, Associate Professor, Health Education, University of Toledo (UT). Noela Haughton, Assistant Professor, Research & Measurement, UT. Jennifer Wohlwend, Lecturer, Health Education, Lourdes University. Stephen Roberts, Professor Emeritus, Public Health, UT. Timothy Jordan, Professor, Health Education, UT. Faith Yingling, Director of Wellness Connection, Department of Recreation and Wellness, Bowling Green State University. Alexis Blavos, Graduate Assistant, Health Education, UT. Correspondence concerning this article should be addressed to Tavis Glassman at [email protected]

JSARP 2013, 50(4)

© NASPA 2013 http://journals.naspa.org/jsarp doi:10.1515/jsarp-2013-0026 Brought to you by | University of Toledo Authenticated | [email protected] author's copy Download Date | 11/18/13 3:39 PM

355

High-Risk Drinking Among College Students

High-risk drinking or binge drinking, defined as consuming five or more drinks on a single occasion during the past two weeks (Johnston, O’Malley, & Bachman, 2001) remains a serious problem at many colleges and universities (Nelson, Ziming, Hang, Weitzman, & Wechsler, 2009). According to the Core Alcohol and Drug Survey, in 2010, 82% of students consumed alcohol in the past year with 44% engaging in high-risk drinking (Core Institute, 2010). Consequences of excessive alcohol consumption by college students include automobile accidents, fights, falls, vandalism, noise, injuries, sexual assaults, unwanted pregnancies, passing out, and aspirating vomit (Green, 2009). One estimate puts deaths for college students at 1,700 per year due to alcohol (Green, 2009; Wechsler & Nelson, 2008). Another estimate indicates 2,600 people below the age of 21 die each year as a result of alcohol-related incidents (Hingson & Zha, 2009). High-risk drinking also results in other harmful consequences. Self-reported data reveal that the proportion of students who were found responsible for alcohol-related issues with residence life staff, other school officials, or campus police during the last year changed little from 1994 (14%) (Presley & Meilman, 1994) to 2008 (13%) (Core Institute, 2010). Similarly, the percentage of students who missed class during the last year due to alcohol consumption changed little from 1994 (30%) (Presley & Meilman, 1994) to 2008 (28%) (Core Institute, 2010). Getting hurt or injured from an alcohol related-event decreased slightly from 20% in 1999 to 16% in 2006 (Core Institute, 2010). One area showing marked improvement involves drinking and driving. Indeed, self-reported drinking and driving during the previous year decreased from 35.6% in 1994 (Presley & Meilman, 1994) to 22.3% in 2008 (Core Institute, 2010). Perhaps this decrease is due in part to the local and national campaigns designed to reduce drunk driving (National Drunk and Drugged Driving Prevention Month December, 2004; Tay, 2005; Weber, Dillow, & Rocca, 2011), including the Friends Don’t Let Friends Drive Drunk campaign where the researchers reported a 10% decrease in alcoholrelated fatalities between 1990 and 1991(Kotler & Lee, 2008). Many university administrators struggle with the issue of alcohol abuse by students and have implemented campus policies and intervention programs in an attempt to decrease this problem (Wechsler, Seibring, I-Chao Liu, & Ahl, 2004). Social norming campaigns are commonly implemented as intervention strategies, which expose students to messages designed to correct misconceptions about the amount of alcohol consumed by their peers (Glassman & Braun, 2010; Wechsler & Nelson, 2008). A typical social norms message states “most [insert school mascot] have 0 to 4 drinks when they party,” but many students report neither believing nor trusting these messages (Glassman, Dodd, Miller, & Braun, 2010). Many social norms campaigns fail to properly develop campaign materials using health communication or social marketing theory, such as properly segmenting the audience or providing a desirable exchange (Glassman & Braun, 2010). A onesize-fits-all approach tends to be less effective than a message tailored to a specific group. If the message does not provide an appealing benefit, it is unlikely the audience will adopt the recommended practice or behavior (Kotler & Lee, 2008). While implementing social norms campaigns at universities is a common practice (Wechsler & Nelson, 2008), the results of these campaigns have

356

doi:10.1515/jsarp-2013-0026 http://journals.naspa.org/jsarp © NASPA 2013 Brought to you by | University of Toledo Authenticated | [email protected] author's copy Download Date | 11/18/13 3:39 PM

JSARP 2013, 50(4)


been mixed. Confounding issues include theoretical fidelity and mitigating factors such as outlet density (DeJong et al., 2009; Scribner et al., 2011). While the efficacy of social norms campaigns remains in question, other health communication campaigns show promise. A campaign implemented in the southeastern United States from 2004 to 2008 used extensive formative research to design prevention materials. After conducting numerous focus groups, researchers found that college students were more concerned with the social consequences associated with overconsumption of alcohol than physical health effects (Dodd, Glassman, Arthur, Webb, & Miller, 2010). Messages were subsequently developed to accentuate the social costs associated with excessive drinking, such as embarrassment, unattractiveness, and annoyance. The slogan for this campaign, which appeared on all materials, included a moderation message stating: “Less is More.” Researchers chose to use this slogan because students adamantly rejected abstinence messages; consequently, the goal of the campaign focused on moderation rather than asceticism. Although the campaign resulted in substantial decreases in high-risk drinking and alcohol-related consequences, the lack of a control group and use of only one location for this study limited the internal and external validity of the findings (Glassman, Dodd, Miller, & Braun, 2010). To ameliorate these shortcomings, the investigators of the current study designed an extensive health communication campaign using control groups to address the limitations associated with nonexperimental investigations. Researchers sought to determine whether a series of integrated health messages would decrease self-reported alcohol consumption rates and reduce alcohol-related consequences among first-year college students living in the residence halls. Evidence suggests that challenging the perceived benefits of alcohol use decreases alcohol consumption, which results in fewer alcohol related consequences (Brown, Christiansen, & Goldman, 1987; Fromme, Stroot, & Kaplan, 1993; Gilles, Turk, & Fresco, 2006; Ham, 2009; Ham, Zamboanga, Olthuis, Casner, & Bui, 2010; Reingle et al., 2009). This health communication intervention was predicated on the expectancies construct within social cognitive theory (Bandura, 2004).

Methods The study population included first-year students living at two public universities in the Midwest. The cohorts included an on-campus intervention and on-campus control group from one university and an external control group from the second, to control for contamination effects (internal control group inadvertently exposed to the message on campus). Students self-selected their on-campus housing on a first-come, first-served basis. A sample size analysis was conducted separately for each cohort (residence hall) using a margin of error of 5% and confidence level of 95%, necessitating a minimum sample size of 235, 240, and 267 in the intervention, internal control, and external control groups, respectively. A total of 741, 700, and 923 students resided in the on-campus intervention, on-campus control, and off-campus control residence halls respectively. Fall and spring pretest and posttest survey response rates for each residence hall, were 333 (45%) and 301

JSARP 2013, 50(4)


357


(41%), 413 (59%) and 436 (62%), and 695 (75%) and 395 (43%). The number of completed surveys exceeded the minimum values yielded from the a priori sample size calculations.

Participants Table 1 includes demographic data for all participants. For the combined sample population, 59% (n=1,426) of the respondents were female and 41% male (n = 957). The majority of respondents were White (78%) and 22% were Students of Color (African American/Black, Asian/Pacific Islander, American Indian/Alaskan Native, Other). At pretest, 99% were under 21 years old with 98% under 21 at posttest, which was expected since the surveys were administered to first-year students. The racial category was divided into White and Students of Color due to the limited number of minority participants. A significantly higher percentage of non-Caucasians resided in the treatment cohort than the control residence halls, x2(2, N = 2,366) = 185.083, p < .001. There were also statistically significant differences in gender between all three residence halls, x2= (2, N = 2,382) =, p < .001 with the percentage of females at disproportionately high rates in the intervention and external control group. Finally, participants were coded under 21 years of age or greater than or equal to 21 years of age based on the legal drinking age, x2 (2, N = 2,380) =13.818, p = .001). Table 1 Participant Demographics Intervention Fall 2010

Internal Control

Spring 2011

Fall 2010

External Control

Spring 2011

Fall 2010

Spring 2011

n

%

n

%

n

%

n

%

n

%

n

%

Male

123

37

127

42

204

49

233

53

135

27

135

34

Female

210

63

174

58

209

51

203

47

370

73

260

66

White

203

61

166

56

351

85

366

85

429

86

329

84

Minority

129

39

132

44

61

15

67

15

72

14

62

16

331

99

285

96

411

99

428

98

505

100

394

99

1

1

13

4

2

1

28

2

0

0

2

1

Gender

Race

Age Under 21 ≥ 21

358


JSARP 2013, 50(4)


Research Design Researchers employed a quasi-experimental design collecting pretest and posttest data (Creswell, 2012) to examine self-reported attitudes and behaviors towards alcohol consumption. The dependent variables for this investigation were alcohol consumption and alcohol-related consequences. The alcohol consumption measures included: the average number of days alcohol was consumed during a week, the average number of drinks consumed in a session, binge drinking (five or more drinks at one sitting), and maximum number of drinks consumed. The five alcohol-related consequence dependent variables were: embarrassed by behavior, spent more money than intended, went to school drunk, blacked out, and had a hangover (Presley & Meilman, 1994). The independent variable for this study was the residence hall students lived in during the academic year.

Measures A comprehensive literature review of college student drinking and health communication campaigns was conducted to establish face validity of the survey. In addition, to enhance content validity the survey was analyzed by a variety of experts including a college health promotion specialist, a student affairs administrator, an alcohol and drug researcher, and a psychometric expert. The four-page survey, in booklet form, consisted of 54 items divided into various sections. Demographics comprised Part A and included items pertinent to college students, such as Greek status, academic major, grade point average, and employment. Part B included items from the Core Alcohol and Drug Survey, a standardized instrument developed specifically for college students (Presley & Meilman, 1994). More specifically, six items solicited responses on drinking habits, such as average number of drinks per week. The alcohol-related consequence items included missing work and getting into fights, and used a Likert-type response scale with options: frequently, sometimes, rarely, never, and do not drink. Internal reliability estimates ranged from good to excellent as follows: drinking habits .69 (pretest) and .74 (posttest); consequences .97 (pretest) and .97 (posttest).

Procedures The university’s Institutional Review Board approved this study. Three residence halls were chosen to participate in the study: one internal treatment hall; one internal control hall; and one external control hall at another university approximately 20 miles away. The two internal residence halls were selected and matched based on past demographic data. The internal residence hall that served as the treatment hall at the primary university was randomly selected between the two internal halls. The participants in the treatment group received health communication messages discouraging excessive alcohol consumption, whereas participants at the internal and external control groups were not exposed to health communication materials associated with the intervention. It was possible for students in the onsite control group (residence hall) to interact with students

JSARP 2013, 50(4)


359


from the intervention hall and discuss the messages. Students from the control group may have visited the intervention group, thus potentially seeing the messages, exposing themselves to the intervention and resulting in contamination effects. Despite the investigators’ efforts to match the demographics of the two internal residence halls, they differed demographically based on selfselection (Table 1). Data were collected via a paper and pencil survey. Participants provided informed consent by listening to the proctor’s instructions, reading the directions on the survey, and then answering questions listed on the survey. Baseline survey data were collected at each of the residence halls during orientation at the beginning of fall semester. Once baseline data were collected, poster holders were installed at 53 locations in the treatment residence hall, including bathrooms and hallways on each of 10 floors, study units, laundry rooms, recreation room, and a general gathering area. Follow-up survey data were collected at the end of the year (spring semester). In an attempt to increase the response rate for the surveys, a random drawing was held at each residence hall (Dommeyer, Hirao, Ikeda, Linkletter, & Watanabe, 2010; McCree-Hale, DeLaCruz, & Montgomery, 2010; Ritter et al., 2005; Ryu, Couper, & Marans, 2005), which included a total of 50 students who received $50 gift cards in the fall and spring, respectively.

Message. The health communication campaign messages were based on a related social marketing intervention conducted at another university (Glassman, Dodd, Miller, & Braun, 2010), revised and customized (school colors and images) for the student population in the present study, with graphic design assistance coming from the university communication department. Formative research, web-based confidential message testing (n = 72) with college students, was conducted to elicit their feedback, which led to further campaign revisions. The messages were designed to challenge positive alcohol expectancies, sometimes referred to as counter advertising. Each of the messages highlights the problems associated with excessive alcohol consumption, emphasizing the social consequences of drinking too much. One of the messages states: “When I drink too much… I feel guilty because my friends have to take care of me. I ruin everyone’s night and I look like a hot mess.” followed with the slogan, Less Is More. Central intercept. To help validate the results of this intervention, central intercept surveys were administered to assess students’ thoughts, feelings, opinions, and reactions to the health communication campaign. To conduct central intercept interviews, researchers indiscriminately recruited participants at the intervention residence hall to assess the prevention messages, using a convenience sampling technique. Questions in the central intercept survey included: “Where have you seen the message?” “How often have you noticed the message?” “What is the very first thing you noticed when you saw the message?” Other questions elicited student opinions of their own and others’ drinking habits after having seen the message and whether they had discussed the poster with others. Seven items were developed to measure the believability, understanding, and appeal of the message using Likert scales with responses ranging from strongly agree to strongly disagree as well as no opinion and not applicable. Fourteen waves of central intercept interviews were conducted 360


JSARP 2013, 50(4)


Table 2 Central Intercept Responses to Closed-Ended Questions (n = 340) To what extent do you agree with the following:

Strongly Agree

Agree

Not sure

Disagree

Strongly Disagree

No opinion

Not Applicable

I understand the message.

n = 223

n = 113

n=3

0

0

0

n=1

66%

33%

1%

n = 166

n = 143

n = 19

n=8

n=2

n=2

49%

42%

6%

2%

1%

1%

n = 120

n = 166

n = 28

n=9

n=3

n = 13

35%

49%

8%

3%

1%

4%

I think the message would appeal to my friends.

n = 64

n = 103

n = 101

n = 46

n = 13

n = 10

n=3

19%

30%

30%

14%

4%

3%

1%

The message appeals to me.

n = 79

n = 138

n = 32

n = 45

n = 20

n = 11

n = 15

23%

41%

9%

13%

6%

3%

4%

I decreased my drinking as a result of seeing the message.

n = 27

n = 32

n = 27

n = 51

n = 21

n = 18

n = 164

8%

9%

8%

15%

6%

5%

48%

My commitment to not drink was strengthened or reinforced by the message.

n = 61

n = 111

n = 33

n = 39

n=8

n = 22

n = 66

18%

33%

10%

12%

2%

7%

19%

I believe the message. I like the message.