Case-Based Article Buzz Agents in a Teen-Driven Social Marketing Campaign: Positive Campaign Attitude Leads to Positive Changes in Health Outcomes Social Marketing Quarterly 2016, Vol. 22(3) 218-235 ª The Author(s) 2016 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/1524500416637052 smq.sagepub.com Amy Struthers1 and Ming Wang1 Abstract A research team in Nebraska developed a public health campaign for teens focused on obesity prevention, based on social marketing and buzz marketing principles, to test a series of hypotheses postulating that positive attitude toward the campaign among the most engaged members of the target audience, the buzz agents, will lead to positive attitudinal as well as positive self-reported behavioral changes involving fruit and vegetable consumption and physical activity. The team implemented the Whatcha doin? social marketing campaign in a select number of public high schools in Nebraska from 2007 to 2013. Results from 6 years of annual survey data of buzz agents in these schools showed positive associations between attitude toward the campaign and changes in health attitudes and behavior. Keywords buzz marketing, buzz agent, social marketing, health communication, teen wellness Introduction From 1971 to 2005, the obesity rates among U.S. adolescents aged 12–19 increased from 6.1% to 17.8%, according to the Centers for Disease Control and Prevention (CDC) National Health and Nutrition Examination Survey (Ogden & Carroll, 2010). In response, government agencies and public health officials at the federal, state, and local levels began searching for more innovative ways to communicate persuasive messages of obesity prevention. In the fall of 2006, a research partnership between the advertising faculty at a large Midwestern university in the United States and the state’s Department of Health and Human Services (DHHS) was formed to explore more effective ways of communicating public health messages that would address, at the state level, this growing crisis of adolescent obesity. The collaboration evolved into a teen-driven 1 College of Journalism & Mass Communications, University of Nebraska–Lincoln, Lincoln, NE, USA Corresponding Author: Ming Wang, College of Journalism & Mass Communications, University of Nebraska–Lincoln, Lincoln, NE 68588, USA. Email: [email protected] Struthers and Wang 219 social marketing campaign, Whatcha doin?, implemented in 32 high schools across the state over a 6year period, funded by the CDC and jointly administered by state public health officials and the lead author of this article. The campaign utilized buzz agents, ordinary students from participating high schools, to engage their peers in the promotion of fruit and vegetable consumption and physical activity. The rest of the article first describes the application of buzz marketing in social marketing and its relevance to the theoretical framework of this project. After that, the authors discuss the development of the campaign using social marketing benchmarks. The article then details recruiting buzz agents and implementing the campaign with the social marketing mix. Using 6 years of survey data, this study finds that buzz agents’ positive attitude toward the campaign is largely associated with improved health attitudes and behaviors. This article concludes with a discussion of the lessons learned from engaging buzz agents and the issue of scaling up the campaign. Applying Buzz Marketing to Social Marketing Buzz Marketing and Its Application in Social Marketing Among young adult consumers, research indicates both a lack of interest in messages concerning health and nutrition and a reliance on trusted family members and friends as information sources (Brennan, Dahl, & Eagle, 2010). Buzz marketing, as a carefully planned, multidimensional strategic communication strategy that utilizes trained ‘‘buzz agents’’ to convey persuasive messaging to peers through created events (Ahuja, Michels, Walker, & Weissbuch, 2007), can facilitate purposeful development of positive word of mouth. Characterized by a lack of commercial bias as well as speed and interactivity, buzz marketing is considered to be more trustworthy and authentic than other sources of information about products, services, and ideas (East, Hammond, & Lomax, 2008; Silverman, 2001). Buzz marketing as a social marketing strategy has been applied in social marketing campaigns with encouraging results, such as the ‘‘Liverpool’s Challenge’’ campaign to fight obesity (Thomas, 2009) and the HEALTHY intervention to reduce the risk for type 2 diabetes (DeBar et al., 2009). The buzz marketing strategy thus offers promise for social marketing campaigns. Theoretical Framework The focus on buzz agents in this campaign was partly informed by the Diffusion of Innovation theory (Rogers, 1983). According to this theory, adoption of a product, behavior, or idea follows an S-shaped curve, where the adoption process is driven by five types of individuals: innovators, early adopters, early majority, late majority, and laggards. More importantly, this process typically moves through these five stages sequentially. In other words, for a behavior to be adopted by a majority of the public, it has to be adopted by a small group of innovators first, after which its followers will include early adopters. Buzz marketing strategy is consistent with the diffusion of innovation theory in focusing on identifying and growing innovators and early adopters first to a critical mass and then letting them, commonly referred to as buzz agents, spread the idea or behavior to their peers to reach a majority stage. Campaign Development Social Marketing Benchmarks What differentiates social marketing from other types of marketing activities? Andreasen (2002) argues that social marketing campaigns take into account six criteria: (1) targeting and segmentation, 220 Social Marketing Quarterly 22(3) (2) behavior change, (3) competition, (4) exchanges, (5) audience research, and (6) the marketing mix. The research team actively incorporated these strategies in developing the Whatcha doin? campaign. Target Audience: Teens Recent research in social marketing has underscored a consumer-centric perspective, ‘‘inviting consumers to be true partners in determining how to best meet their health needs’’ (Grier & Bryant, 2005, p. 336). Through target audience analysis, the research partners agreed that the project would focus on the state’s teen population. This was identified as a demographic that was ignored in 2006. Adolescents were clearly not the target audience for existing obesity prevention programs; rather, younger audiences, particularly those under age 12, were observed to have more exposure to age-appropriate marketing materials concerning healthy food choices and appropriate levels of physical activity. Efforts to create healthy habits at as young an age as possible had created a gap in programming during the high school years. And this was at a time when young adults were already known to be increasingly skeptical of mass media, making it particularly challenging to provoke behavior change through the use of more traditional communication channels (Beard, 2003). Behavior Change: Increasing Fruit and Vegetable Consumption and Physical Activity Data collected at the state level, in particular the 2005 Youth Risk Behavior Survey, found only 36.5% of high school students reported achieving the recommended 60 min of physical activity on 5 or more days per week. The study also found only 13.5% of high school students consumed the recommended daily fruit and vegetable allowance of four and a half to six cups per day. The research team agreed that a campaign tightly focused on impacting attitudes and behaviors about these three specific activities would contribute the most to obesity prevention for this project. Concept Development Research: Initial Primary Data Collection The research team conducted in-depth interviews about high school students’ attitudes toward fruit and vegetable consumption and physical activity. Insights from this qualitative research contributed to the development of the campaign strategy and to the choice of buzz marketing as the communication channel. The strategy developed for the teen high school audience was that the decisions to eat fruits and vegetables and to be physically active are up to each individual and can be realized in many unique ways, not just those defined by adults, and any of them will contribute to a fun, cool lifestyle as modeled by peers. The research team then developed three distinct creative approaches for communicating the message. These concepts were presented to the project partners at the state DHHS who provided feedback that was incorporated into the design concepts. The refined concepts were then tested on two focus groups from 24 high school students. Participants were asked questions about their own attitudes toward fruit and vegetable consumption and physical activity and toward the educational materials concerning obesity that were present in their high schools. They were also asked about the influence of peers on their own behavior. Finally, they were shown the campaign creative concepts and asked to react both to the creative work and to the concept of teens marketing the message to other teens. These data were used to arrive at a visual look and feel for additional testing of the pilot campaign. In addition, the data revealed a high level of comfort with and confidence in a peer-to-peer communication channel, such as buzz marketing. Struthers and Wang 221 Competition The insight distilled from the research that came to drive the campaign development was the consistently articulated need for teens to feel they were mature enough to make good individual choices concerning health and nutrition decisions. Their desire to be unique was balanced by a fear of being too out of the norm: ‘‘I want to be an individual, just like everybody else.’’ This presented an internal psychological competition of conflicted thoughts among the target audience, in addition to such external competition as abundant availability of unhealthy food in the school environment and the perceived social norm that ostracizes healthy eaters. The campaign needed to walk a fine line between encouraging individual choice and staying within the accepted norms of each teen group. Exchanges The success of a social marketing campaign hinges upon creating compelling exchanges with target audiences by maximizing the benefits and minimizing the costs (Andreasen, 2002). Through an exchange analysis, the research team became concerned about the time and effort expected from buzz agents. In order to overcome such barriers, some incentives or rewards needed to be present (Andreasen, 2002). Providing buzz agents with advance information of the campaign and allowing them to tailor the tactics to their own environment, thereby boosting their self-efficacy, would arguably achieve that purpose. This tested well in focus groups, reinforcing Rosen’s (2000) theory that anticipation is an important element in a buzz campaign. The activities targeted at the general student population were also intended to make sure the benefits outweighed the costs. For instance, to reduce the cost of spending time looking for a place to do some physical activity, the research team provided each school with green stability balls to be placed in classrooms and libraries to keep the students active during class. Buzz agents were encouraged to use these stability balls to show to their peers how being physically active could be fun. Campaign Strategy: Peer Buzz Agents The research partners identified resistance to top-down persuasion commonly found in high school students compared to those at other grades as a potential threat to health intervention campaigns. On the other hand, the research team has found peer influence to be a powerful force at this developmental stage. Use of peer educators or student peer communicators (DeBar et al., 2009) in health campaigns has been found to be effective in some studies. For instance, one project paired teen mentors with younger children to promote healthier eating and more physical activity, finding that mentored children had higher levels of health knowledge, attitudes, and body mass index (Smith, 2011). A peer-led health promotion program among elementary students, called Healthy Buddies, also reported more healthy living knowledge among students in intervention schools than in comparison sites (Campbell et al., 2012). Another tobacco cessation project found schools that used peer supporters achieved a 22% reduction in smoking among teens compared with control schools (Starkey, Audrey, Holliday, Moore, & Campbell, 2009). Testing the Buzz Agents Idea This formative research provided feedback on the role of the buzz agents in this campaign. Teen focus group participants viewed the buzz agents concept as a positive element of the campaign, consistent with other research that shows teens like being buzz agents and that they treat the role as a job (Ahuja et al., 2007). Students found it important that buzz agents should ‘‘live the lifestyle,’’ stating that peer-to-peer marketing was only honest if the person selling something truly believed in what they were selling. 222 Social Marketing Quarterly 22(3) Recruiting Buzz Agents Fall 2007 became the launch window, a time when teens would be most energized by a new school year. That time frame then required the development of a process for quickly recruiting and training in the fall the essential element of the campaign, the buzz agents. The solution was to work through each high school principal to identify a teacher coordinator who would serve as an on-the-ground recruiter for and manager of buzz agent teams in each participating school. With the ability to fit the campaign into their school in a customized way, a variety of teachers representing different subject matter expertise proposed a diverse range of student groups as buzz agent teams, including a student council, a nutrition class, a physical education class, and student clubs. Over the 6 years of the campaign implementation and the involvement of 32 high schools, these processes worked smoothly and provided the team with a valuable cross section of high school-aged buzz agents. These students served as the chief communications channel for the campaign concepts. Campaign Implementation—Social Marketing Mix The hallmarks of a social marketing campaign are the 4Ps, or the marketing mix: product, price, place, and promotion (Glassman & Braun, 2010). Product A product is what is offered to a market to meet a want or need; it can include core benefits of behavior, goods or services, or product elements that can help lead to behavior change (Lee & Kotler, 2015). The product in the Whatcha doin? campaign was a desired behavior change: teens choosing to consume more fruits and vegetables and to be more physically active. The researchers developed this product by linking it with the benefit of showing peers they were part of a cool movement. The idea was packaged with tangible items or ‘‘augmented products’’ (Lee & Kotler, 2015): branded campaign collateral such as stability balls, T-shirts, stress balls, stickers, and many other items. Buzz agents demonstrated and exemplified the benefits of adopting this product and encouraged their peers to join them through peer-to-peer modeling. Price Price involves a cost–benefit analysis to consumers. The campaign tried to show that the benefits outweighed the costs. Students felt they did not have time for physical activity; buzz agents found time during breaks to engage their peers in physical activities. Students also found fruit and vegetables undesirable; buzz agents conveyed the benefits to the students and were active consuming fruit and vegetables themselves. Place Place refers to ‘‘where and when the target audience will perform the desired behavior, acquire any related goods, and receive any associated services’’ (Lee & Kotler, 2015, p. 315). In the case of this campaign, the high school setting was a bounded system where the target audience was consistently located. The campaign was built to show the targeted behavior in a location where it was easier for the buzz agents to perform and model the desired behavior and where the collateral materials could be more accessible to the target audience. For example, each school received 5–10 green stability balls that were placed around the building (e.g., classroom, cafeteria, library, and principal’s office) to get students talking and to keep them active during class. Struthers and Wang 223 Figure 1. Students exercised with stability balls in the Whatcha? doin campaign. Promotion Teacher coordinators received tools and materials for their buzz agent teams to help promote the campaign within their schools, from detailed packets filled with suggested ways to execute the campaign to branded collateral such as stability balls, T-shirts, static clings, stickers, and campaign signs (see Figures 1 and 2 for example). Teams were required to work with their teacher coordinators each week to first slowly roll out the campaign using stealth tactics and then to stage frequent, branded ‘‘random acts’’ involving fruits, vegetables, and physical activity. In order to keep the campaign new and alive, each phase of the campaign offered a new goal so that interest never died as the campaign progressed. Both buzz agents and teacher coordinators participated in periodic conference calls, e-mail exchanges, and site visits throughout the campaign each year. Rosen (2000) developed a number of principles to drive a successful buzz marketing campaign. These included introducing fresh, creative materials associated with the message on a frequent basis; creating 224 Social Marketing Quarterly 22(3) Figure 2. Students voted to adopt this T-shirt design for the 2012 Whatcha doin? campaign. plenty of tools to make it easy for the target audience to spread the message; building a sense of anticipation about the bigger communications story; developing activities that are fun for people to talk about and share; and even being a little outrageous in order to more effectively engage the audience. Following these principles, the campaign showed the fun, cool, and even unexpected lifestyle accessible yet unique to each teen that could come from increased fruit and vegetable consumption and increased physical activity. The campaign tagline, Whatcha doin’? How you do it is up to you, encouraged high schoolers to take action in their own way but to certainly take action. A strong visual brand identity is critical in marketing to adolescents, including things like an eyecatching logo associated with the campaign message (L’Engle, Pardun, & Brown, 2004). A bright lime green splash evolved as the campaign logo (see Figure 3), and all materials created thereafter used a well-defined guide to brand standards including color palette, font, and logo. According to Rosen (2000), the most important aspect of a buzz campaign is to keep the buzz alive. To do that, the Whatcha doin? campaign was crafted to be ever changing and full of wacky and random activities that kept the student body constantly guessing. Buzz agents were encouraged to use their own imaginations and to stage many small low-cost but unexpected activities. Dubbed by the research teams as ‘‘random acts,’’ these events were meant to catch people off guard and also empowered the buzz agents to work with their teacher coordinator and school administration on surprise ‘‘secret’’ activities. By utilizing the Whatcha doin? signs, stickers, and T-shirts to properly brand the random acts, students in the schools linked the randomness with the campaign, which effectively kept both the buzz and the campaign alive (see Figure 4 for a random act that engaged students). Hence, the placement in the marketing mix was not static in this campaign; rather, it was quite dynamic in that buzz agents could decide when and where to stage the acts. Struthers and Wang 225 Figure 3. Whatcha doin? Campaign Logo. Campaign Time Line The research team conducted a pilot study in four schools at one school district involving 32 buzz agents in 2007 before executing the full scope of the campaign. Upon receiving positive feedback from school partners and participants through quantitative and qualitative survey instruments, the research team kept running the campaign for 5 more years on a bigger scale involving more buzz agents in more schools throughout our Midwestern state. With all their responsibilities and materials, the buzz agents had the most intense engagement in the campaign of any students. Research Hypotheses Knowing that the success of a buzz campaign rested in these ‘‘buzzers’’’ hands, the research team began to ask whether the buzz agents would enjoy their participation in campaign activities so much that it would actually impact their own attitudes and behavior around the campaign’s goals: increased fruit and vegetable consumption and physical activity. Such evaluations are in line with the need for a ‘‘return on investment’’ (ROI) analysis for social marketing campaigns. Buzz agents played a critical role in this social marketing campaign; accordingly, the research team invested most of its resources on this core group of students. It will be hard to believe the campaign had an impact on the majority of the high school students if attitude toward the campaign did not even lead to improved health attitudes and behaviors among these buzz agents. Hence, we expected that engagement with the campaign would foster liking for the campaign among these buzz agents, which in turn would lead to an improvement in health attitudes and an increase in health behaviors. Utilizing data from their responses to year-end surveys, the research team tested the following centered on buzz agents: 226 Social Marketing Quarterly 22(3) Figure 4. Students got involved in the Whatcha doin? campaign with buzz agents. Hypothesis 1: Attitude toward the Whatcha doin? buzz campaign is positively associated with an improvement in health attitudes toward (a) being physically active, (b) eating fruit, and (c) eating vegetables. Hypothesis 2: Attitude toward the Whatcha doin? buzz campaign is positively associated with an increase in health behaviors, such as (a) the amount of physical activity, (b) fruit intake, and (c) vegetable intake. Method Sample Survey respondents, who were the buzz agents from participating schools, were unique due to the new configuration of participating schools and students each year of the campaign. The volunteer teacher coordinators were tasked with assembling a team within their own school, resulting in different students assuming buzz agent duties each year. All of the teacher coordinators from the 32 participating schools over the 6-year campaign were asked to administer the survey to their school buzz agents near the end of the campaign each academic year. As the liaison between the researchers and the buzz agents, the teacher coordinators could choose to administer the survey or not. The campaign coordinator was charged with inviting the teacher coordinators to participate, distributing the appropriate number of the paper survey instruments to them all and following up with requests to administer and return the surveys in postage-paid envelopes provided with the surveys. A financial incentive was used to encourage participation by the teachers: A US$100 gift card was Struthers and Wang 227 Table 1. The Number of Schools that Participated in the Whatcha doin? Campaign Each Year and the Number of Buzz Agents Who Completed the Surveys. Campaign Year Schoolsa 2007–2008 2008–2009 2009–2010 2010–2011 2011–2012 2012–2013 4 12 15 13 12 20 (4) (11) (6)b (9)b (10) (10) Buzz Agents (Response Rate) 32 97 74 78 73 71 (100%) (58.79%) (67.89%) (53.06%) (42.44%) (34.98%) a Number in parenthesis indicates the number of schools where buzz agents returned survey questionnaires. Pretest and posttest surveys were administered in these 2 years. Number in parenthesis indicates the number of schools where buzz agents returned posttest survey questionnaires that were analyzed in this study. b offered to those who returned surveys to the campaign coordinator. The number of teacher coordinators who chose to participate impacted response rates significantly from year to year, as did the effectiveness of the campaign coordinator, a unique individual in each of the 6 years of the campaign. The mix of participating schools each year was different, including a wide variety of rural and urban schools across the state. These schools ranged in student population from 136 to more than 1,700. The Nebraska Department of Education (2012) recorded an increase across the state in the number of students on free or reduced lunch during the 6-year campaign, from 37.33% the year the campaign launched (2007–2008) to 43.79% the last year reported (2011–2012). Data The Whatcha doin? buzz campaign was implemented each academic year in participating high schools in Nebraska from 2007 to 2013. A survey instrument was distributed to buzz agents each year to assess the effectiveness of campaign tactics used in each school, buzz agents’ evaluations of the campaign, and their own health attitudes and behaviors. Formative evaluation provoked evolution of the questionnaire each year, but a core set of questions was included in each survey in order to track longitudinal changes in key indicators. To evaluate the effectiveness of the whole Whatcha doin? campaign, the research team combined cross-sectional data collected from the buzz agents over the 6-year period into one data set (n ¼ 425) to test the hypotheses. Longer survey instruments were distributed to buzz agents in W2009–2010 and W2010–2011 that included additional measures of health behavior. The responses to the new questions from these 2 years were combined together into another data set to provide an alternative test of Hypothesis 2. The number of participating schools, the number of buzz agents who completed the surveys in each cross-sectional wave, and the response rate each year are reported in Table 1. Measures: Independent Variables The independent variables included measures of demographics, campaign year, and campaign attitude. Descriptive statistics are reported in Table 2. Demographic variables. Three demographic variables were controlled for grade, sex, and race. For the whole sample from 2007 to 2012, buzz agents ranged from Grade 9 to Grade 12, with the majority from Grade 10 (Grade 9: 19.3%; Grade 10: 35.5%; Grade 11: 26.4%; Grade 12: 18.8%). This variable can also be viewed as a proxy for age. Most of the buzz agents were females (71.43%) and 228 Social Marketing Quarterly 22(3) Table 2. Descriptive Statistics. Variables n Whole sample (W2007–2008–W2012–2013) Grade 420 Male 420 White 420 Year 425 Campaign attitude 420 Physical attitude 414 Fruit attitude 411 Vegetable attitude 413 Physical behavior 411 Fruit behavior 410 Vegetable behavior 404 Subsample (W2009–2010 and W2010–2011) Grade 148 Male 149 White 149 Year 152 Campaign attitude 152 Daily physical activity 152 Daily cups of fruit 149 Daily cups of veggies 148 Min. Max. Mean SD 9 0 0 1 1 0 0 0 0 0 0 12 1 1 6 5 1 1 1 1 1 1 10.45 0.29 0.72 3.65 4.05 0.66 0.63 0.51 0.50 0.64 0.44 1.006 0.45 0.45 1.58 0.67 0.47 0.48 0.50 0.50 0.48 0.50 9 0 0 1 1 0 0 0 12 1 1 2 5 7 12 6 10.49 0.26 0.62 1.51 3.86 2.64 1.73 1.50 1.00 0.44 0.49 0.50 0.72 1.74 1.28 1.07 Whites (71.9%). Similar demographic profiles hold for the subsample of buzz agents from 2009 and 2010. The majority came from 10th grade (33.8%) and were females (73.8%) and Whites (62.4%). Campaign year. Since the 6 years of cross-sectional data were combined together, a campaign year variable was included ranging from 1 (denoting W2007–2008) to 6 (denoting W2012–2013) to control for the possibility that the campaign had stronger effects in later years as the campaign matured. A dichotomous campaign year variable was used for the subsample from W2009–2010 and W2010–2011. Campaign attitude. This is the key explanatory variable in this study. Each year, buzz agents were asked ‘‘how would you describe your experience as a buzz agent?’’ Responses were measured on a 5-point scale from loved it, liked it, no opinion, didn’t like it, to hated it. This variable was reverse coded so that a higher value indicated more positive experiences. Buzz agents were also asked ‘‘overall, what is your attitude toward the Whatcha doin? campaign?’’ The same response categories were offered, and the same recoding procedure was followed. These two questions were strongly correlated (r ¼ .62, p < .001), and therefore, they were averaged to create a campaign attitude index (W2007–2008 to W2012–2013: M ¼ 4.05, SD ¼ 0.67; W2009–2010 to W2010–2011: M ¼ 3.86, SD ¼ 0.72). Measures: Dependent Variables The focus of this project is on the health attitudes and health behaviors as outcomes of this buzz campaign. Each year, buzz agents were asked whether they personally changed the following during the course of the Whatcha doin? campaign: (1) attitude toward being active, (2) attitude toward eating fruit, (3) attitude toward eating veggies, (4) amount of physical activity, (5) fruit intake, and (6) vegetable intake. Response categories included ‘‘increased,’’ ‘‘decreased,’’ and ‘‘stayed the same.’’ Respondents who indicated an increase to each of these questions were coded as 1, while the other two categories were coded as 0 because the researchers were interested in whether the intervention increased physical activity or fruit/vegetable intake or not. Struthers and Wang 229 Admittedly, some information was lost by collapsing two categories together. Nevertheless, the researchers believe it is justified for at least two reasons. First, theoretically, the team is more interested in whether the campaign resulted in an improvement in any of the health indicators; as a result, binary coding sufficed. Furthermore, very few buzz agents, and at times none, indicated a decrease in these indicators. Hence, dichotomous coding is warranted. Overall, 66.2% of buzz agents improved their attitude toward being active, 62.8% improved their attitude toward eating fruit, and 51.3% improved their attitude toward eating veggies. In terms of health behaviors, 49.9% increased the amount of physical activity, 63.9% increased their fruit intake, and 43.8% increased their veggie intake. Evidently, positive attitudinal and behavioral changes regarding vegetables lagged behind physical activity and fruit intake. In addition, several new items to measure health behaviors were included in W2009–2010 and W2010–2011. For physical activity, buzz agents were asked whether they did any of the following on a daily basis: (1) sports practice, (2) marching band, (3) personal running, (4) biking, (5) Rollerblading, (6) skateboarding, (7) workout videos, (8) housework/chores, (9) fitness center, (10) farm work, and (11) Physical Education (PE)/gym class. For each of these activities, responses were coded 1 if the buzz agent did it and 0 if the agent did not. These 11 items were summed to create a daily physical activity index (M ¼ 2.64, SD ¼ 1.74). Buzz agents from W2009–2010 and W2010–2011 were also asked how many cups of fruits (M ¼ 1.73, SD ¼ 1.28) and vegetables they ate on an average day (M ¼ 1.50, SD ¼ 1.07). To make sure that respondents had the same anchor point, prompts were provided suggesting that one cup of fruit is equal to a baseball, a medium apple, or a large orange, and a half cup of fruit is similar to 4 large strawberries or 16 grapes. For vegetables, the prompts were that one cup is equal to a baseball, 12 baby carrots, or a medium potato, and a half cup is similar to one cup of lettuce or 5 broccoli florets. These questions from W2009–2010 and W2010–2011 provided the research team with additional measures of the dependent variables to test Hypothesis 2, enabling the researchers to validate the findings that relied on a different set of outcome measures. Results To test the hypotheses, a series of logistic regression models were run on the outcome variables controlling for grade, sex, race, and campaign year for the 6-year whole sample. Furthermore, several ordinary least squares regression models were run on the pooled data from W2009–2010 and W2010–2011 to provide additional tests of campaign effects on health behavior. Health Attitudes Table 3 shows results from three logistic regressions on the health attitude variables from the whole sample. Hypothesis 1 predicts that attitude toward the Whatcha doin? buzz campaign is positively associated with an improvement in health attitudes. As can be seen in Model I results, campaign attitude is indeed positively associated with improved attitude toward being physically active (B ¼ .92, SE ¼ .19, p < .001). An odds ratio of 2.50 suggests that buzz agents were 2.5 times as likely to improve their attitude toward being physically active when their campaign attitude increased one unit. Hence, Hypothesis 1a is supported. Results from Model II show that campaign attitude is also positively associated with improved attitude toward eating fruit (B ¼ .96, SE ¼ .19, p < .001). An odds ratio of 2.60 suggests that buzz agents were 2.6 times as likely to improve their attitude toward eating fruit when their campaign attitude increased one unit, confirming Hypothesis 1b. In a similar vein, campaign attitude is positively associated with improved attitude toward eating vegetables (B ¼ .60, SE ¼ .17, p < .001), as Model III in Table 3 demonstrates. An odds ratio of 1.83 230 Social Marketing Quarterly 22(3) Table 3. Effects of Attitude Toward Whatcha doin? Campaign on Health Attitudes From 2007 to 2012. Model I: Improved Attitude Toward Being Physically Active Independent Variables Grade Male White Year Campaign attitude Constant Cox and Snell R2 Nagelkerke R2 n B (SE) eB 0.072 (0.113) 0.074 (0.250) 0.208 (0.251) 0.104 (0.071) 0.917*** (0.187) 2.806* (1.405) .088 .122 404 0.931 1.077 1.231 1.110 2.502 0.060 Model III: Improved Attitude Toward Eating Vegetables Model II: Improved Attitude Toward Eating Fruit eB B (SE) 0.036 0.074 0.578* 0.103 0.958*** 2.887* (0.109) (0.241) (0.262) (0.069) (0.188) (1.370) .086 .118 401 0.964 0.928 0.561 1.108 2.606 0.056 B (SE) 0.134 0.042 0.052 0.084 0.601*** 4.067** (0.104) (0.229) (0.240) (0.065) (0.170) (1.305) .045 .060 403 eB 1.144 1.043 0.949 1.087 1.825 0.017 Note. Entries are coefficients from logistic regressions on the pooled sample of all buzz agents in the study. For Model I, respondents who indicated that they ‘‘improved’’ their attitudes toward being active were coded as 1; those who chose ‘‘got worse’’ or ‘‘stayed the same’’ were coded as 0. For Models II and III, the same coding protocol was followed for attitudes toward eating fruit and vegetables. *p < .05. **p < .01. ***p < .001. Table 4. Effects of Attitude Toward Whatcha doin? Campaign on Health Behavior From 2007 to 2012. Model I: Increased Amount of Physical Activity Independent Variables Grade Male White Year Campaign attitude Constant Cox and Snell R2 Nagelkerke R2 n B (SE) eB 0.055 (0.109) 0.299 (0.242) 0.083 (0.251) 0.233*** (0.068) 1.049*** (0.194) 4.704*** (1.382) .128 .170 401 0.946 1.348 1.087 1.262 2.855 0.009 Model II: Increased Fruit Intake B (SE) 0.140 0.012 0.387 0.168* 0.766*** 4.292** (0.110) (0.243) (0.257) (0.070) (0.180) (1.285) .074 .101 400 Model III: Increased Vegetable Intake eB B (SE) eB 1.150 0.988 0.679 1.183 2.152 0.014 0.056 (0.105) 0.433 (0.231) 0.176 (0.243) 0.101 (0.066) 0.496** (0.171) 3.222* (1.307) .037 .050 396 1.058 1.541 0.839 1.106 1.641 0.040 Note. Entries are coefficients from logistic regressions on the pooled sample of all buzz agents in the study. For Model I, respondents who indicated that they ‘‘increases’’ the amount of physical activity were coded as 1, and those who chose ‘‘decreased’’ or ‘‘stayed the same’’ were coded as 0. For Models II and III, the same coding protocol was followed for fruit and vegetable intake. *p < .05. **p < .01. ***p < .001. suggests that buzz agents were about twice as likely to improve their attitude toward eating vegetables when their campaign attitude increased one unit. Again, Hypothesis 1c is supported. Health Behaviors Table 4 reports results from three logistic regressions on the health behavior variables. Hypothesis 2 predicts that attitude toward the Whatcha doin? buzz campaign is positively associated with an increase in health behaviors. Model I shows that campaign attitude is indeed positively associated with increased amount of physical activity (B ¼ 1.05, SE ¼ .19, p < .001). An odds ratio of Struthers and Wang 231 Table 5. Effects of Attitude Toward Whatcha doin? Campaign on Health Behavior in 2009 and 2010. Model I: Amount of Daily Physical Activities Independent Variables Grade Male White Year Campaign attitude Constant Adjusted R2 n B (SE) 0.080 0.210 0.660* 0.530 0.619** 1.588 .108 145 (0.148) (0.320) (0.299) (0.295) (0.202) (1.669) b .045 .053 .181 .151 .252 Model II: Cups of Fruit Eaten Per Day B (SE) 0.005 0.114 0.391 0.275 0.479** 0.097 .071 143 (0.111) (0.241) (0.226) (0.222) (0.152) (1.249) b .004 .039 .146 .107 .266 Note. Entries are coefficients from ordinary least squares regressions on the pooled sample of buzz agents from W2009–2010 and W2010–2011. Model I tests campaign effects on the amount of daily physical activity and Model II on the amount of fruit, measured in cups, eaten on an average day. *p < .05. **p < .01. ***p < .001. 2.86 means that buzz agents were almost 3 times as likely to increase their physical activity when their campaign attitude increased one unit, lending support to Hypothesis 2a. Consistent with Hypothesis 2b, campaign attitude is shown in Model II to be positively associated with increased fruit intake (B ¼ .77, SE ¼ .18, p < .001). An odds ratio of 2.15 denotes that buzz agents were about twice as likely to eat more fruits when their campaign attitude increased one unit. As Hypothesis 2c predicts, campaign attitude is also positively associated with increased vegetable intake (B ¼ .50, SE ¼ .17, p < .01), as shown in Model III. An odds ratio of 1.64 suggests that buzz agents were 1.64 times as likely to eat more vegetables when their campaign attitude increased one unit. Next, we turn to additional health behavior indicators included in W2009–2010 and W2010–2011 to see if the same pattern holds with alternative measures of health behavior change. The results are reported in Table 5. After controlling for grade, sex, race, and year, we still find statistically significant positive association between campaign attitude and the amount of daily physical activity (B ¼ .62, SE ¼ .20, p < .01), lending further support to Hypothesis 2a. The model explains 10.8% of the variance in daily physical activity, and campaign attitude is the strongest predictor (b ¼ .25). Just as Hypothesis 2b predicts, campaign attitude is also positively associated with the amount of fruits eaten every day (B ¼ .48, SE ¼ .15, p < .01). This model explains 7.1% of the variance in the amount of fruits eaten and campaign attitude is again the strongest predictor (b ¼ .27). Lastly, we tested the effect of campaign attitude on the amount of vegetables eaten every day. Unfortunately, we did not find a statistically significant relationship between these two variables (B ¼ .14, SE ¼ .13, ns). Hence, Hypothesis 2c is not supported by the data from this subsample. Discussion Lesser Impact on Vegetable Consumption The purpose of this study was to describe the evolution of a social marketing campaign using buzz marketing tactics, including buzz agents, and to discern whether attitude toward such a social marketing campaign by the participating buzz agents would be positively associated with an improvement in health attitudes and behaviors regarding being physically active, eating fruit, and eating vegetables. 232 Social Marketing Quarterly 22(3) The results of the analysis of survey data from 6 years’ worth of buzz agents support the thesis that a positive attitude toward the campaign itself resulted in improved attitudes toward all three components of the campaign messaging. Attitudes toward increased fruit consumption and increased physical activity showed the most impact, vegetable consumption to a lesser extent. Perhaps more importantly, attitude toward the campaign itself had positive impacts on self-reported healthy behaviors. Again, fruit intake and physical activity levels experienced the most significant increases. However, results on vegetable consumption are mixed: Attitude toward the campaign was indeed associated with increased vegetable consumption in the overall sample, but this relationship disappeared when the researchers used a different measure of vegetable consumption among a subsample. This can speak to the challenge of the changing vegetable consumption behavior among teenagers. While findings regarding fruit intake and physical activity are quite robust, the unreliability of the impact on vegetable consumption suggests that scholars and public health professionals should devote more research to this area. Lessons From Engaging Buzz Agents Buzz agent engagement may provide a model for reaching the adolescent population with behaviorchanging obesity prevention messages. How did the research team cultivate buzz agents’ positive relationships with the Whatcha doin? campaign itself? First, the team made the campaign fun and engaging for these buzz agents. They were the first to see and experience all the elements of the campaign. In the campaign, buzz agents were made keepers of the secret, acting as those in their schools who could stealthily roll out the activities that launched the campaign. They were very handson, illustrating the messages to their peers on a frequent, regular basis through their staged random acts. They received fun, low-cost, branded collateral materials, such as T-shirts, stress balls, stability balls, stickers, and window clings, which they shared with their peers but also got to keep. Furthermore, buzz agents were encouraged to own the campaign themselves. Empowered to customize the campaign to their own school environment, working with their teacher coordinator and their school administration to choose from a wide array of potential activities and events created by the research team, buzz agents were responsible for the success or failure of the campaign in the long term and for weekly meetings and activities in the short term. The campaign depended on them. They were the owners of the campaign, not some higher authority dictating how they or their peers should lead their lives; as a result, they had an intense, immersive relationship with the campaign and its core messages. Buzz agents reported the kind of improvements in key obesity-prevention attitudes and behaviors that public health officials want to achieve. The Whatcha doin? campaign, by following the best practices for both social marketing and buzz marketing as recommended by scholars in past literature, produced an academic-year experience for the buzz agents that had significant impact on their health and nutrition attitudes and behaviors. Scaling Up the Campaign The issue of how to scale a buzz marketing public health campaign presents a number of concerns. The number of buzz agents relative to the total population of each school was relatively small, comprising at best an estimated 5% maximum of the whole student body. While this study reports only the data collected from this small subset of students in each school who were buzz agents, surveys were distributed to the entire student body in the first 2 of the 6 years of the campaign. Results from that study showed that student exposure to buzz marketing affected their attitudes toward the campaign, which in turn improved health attitudes (Wang & Struthers, 2014). Struthers and Wang 233 There were a number of challenges to surveying the entire student body of up to 15 high schools each year. High school students are, for the most part, minors, and so parental consent to participate in the research is required. Also, the researchers relied on teachers to administer the surveys and return them; these educators were required to use class time to administer and collect the survey. This resulted in lower response rates than desired. For these reasons, the yearly survey of the entire student body of each school was abandoned after the first 2 years, making it more difficult to report on any broader impact. To give more adolescents the same deep experience with a campaign they own and enjoy as the buzz agents have, the Whatcha doin? campaign could be incorporated into more classes as part of the curriculum. However, curriculum changes are difficult to achieve, and schools continue to work under the pressures of preparing students to succeed on state and national standardized tests, making the addition of new material difficult. The willingness of teachers to step up as coordinators may be another limitation to broader use of the buzz agent concept. This may be just one more additional project added to an already busy work life. Finally, the resource issue at the state level is also a concern. The campaign itself was relatively low cost: For example, the 2012–2013 year’s campaign was just under US$20,000 for the undergraduate intern and all the campaign collateral material. Yet there is a cost, and sometimes even US$20,000 is difficult for a state to put to such a campaign. It should be noted that this project illustrates a model for a low-cost alternative to the creation and execution of a public health campaign, which is a series of choices prompted by the very lack of funding. The first choice was for the DHHS to collaborate with students and researchers at an institution of higher education rather than working with an ad agency. The costs could increase dramatically if a professional advertising agency were to create the campaign identity and design and produce the collateral materials. The choice of collateral pieces could also impact the total campaign cost; this project made low-cost choices such as branded stability balls, T-shirts, and simple digital cameras to provide to participating schools. It is also important to consider that a campaign using buzz marketing eschews traditional mass media, so the total cost will by definition be significantly less than producing television or radio spots, and then purchasing television or radio time, or producing glossy print ads and then purchasing print advertising placement in teen magazines. The ROI for the Nebraska DHHS was the number of high school students in Nebraska reached through the campaign, who improved their health attitudes and behavior. This campaign made sure buzz agents were committed first, who then influenced their peers. The evaluations reported in this article have shown campaign attitudes made a difference. Conclusion Despite these concerns, this project has shown that a positive attitude toward a social marketing campaign employing a peer-driven buzz marketing strategy can indeed improve health attitudes and behaviors among teens. While the small number of buzz agents may be perceived as producing too little impact, the data suggest that this concept is worth further testing. Given the scale of the nation’s obesity epidemic, future research on the use of buzz marketing in a social marketing public health context is warranted. Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) received no financial support for the research, authorship, and/or publication of this article. 234 Social Marketing Quarterly 22(3) References Ahuja, R., Michels, T., Walker, M., & Weissbuch, M. (2007). Teen perceptions of disclosure in buzz marketing. Journal of Consumer Marketing, 24, 151–159. Andreasen, A. R. (2002). Marketing social marketing in the social change market place. Journal of Public Policy and Marketing, 21, 3–13. Beard, F. K. (2003). College student attitudes toward advertising’s ethical, economic, and social consequences. Journal of Business Ethics, 48, 217–228. Brennan, R., Dahl, S., & Eagle, L. (2010). Persuading young consumers to make healthy nutritional decisions. Journal of Marketing Management, 26, 635–655. Campbell, A. C., Barnum, D., Ryden, V., Ishkanian, S., Stock, S., & Chanoine, J. (2012). The effectiveness of the implementation of healthy buddies, a school-based, peer-led health promotion program in elementary schools. Canadian Journal of Diabetes, 36, 181–186. DeBar, L. L., Schneider, M., Ford, E. G., Hernandez, A. E., Showell, B., Drews, K. L., . . . White, M. (2009). Social marketing-based communications to integrate and support the healthy study intervention. International Journal of Obesity, 33, S52–S59. East, R., Hammond, K., & Lomax, W. (2008). Measuring the impact of positive and negative word of mouth on brand purchase probability. International Journal of Research in Marketing, 25, 215–224. Glassman, T. J., & Braun, R. E. (2010). Confusion surrounding social marketing strategies and social norm theory: To prevent high-risk drinking among college students. Social Marketing Quarterly, 16, 94–103. Grier, S., & Bryant, C. (2005). Social marketing in public health. Annual Review of Public Health, 26, 319–339. Lee, N. R., & Kotler, P. (2015). Social marketing: Changing behavior for good (5th ed.). Thousand Oaks, CA: Sage. L’Engle, K., Pardun, C., & Brown, J. (2004). Accessing adolescents: A school recruited, home-based approach to conducting media and health research. Journal of Early Adolescence, 24, 144–158. Nebraska Department of Education. (2012). State of Nebraska: Free/reduced priced meals. Retrieved from http:// reportcard.education.ne.gov/20112012/pg_Demographics.aspx?Page¼StudentMeals&AgencyID¼00-0000000 Ogden, C., & Carroll, M. (2010). Prevalence of obesity among children and adolescents: United States, trends 1963-1965 through 2007-2008. Washington, DC: Center for Disease Control and Prevention. Rogers, E. M. (1983). Diffusion of Innovations (3rd ed.). New York: Free Press. Rosen, E. (2000). The anatomy of buzz: How to create word of mouth marketing. New York, NY: Random House. Silverman, G. (2001). The power of word of mouth. Direct Marketing, 64, 47–52. Smith, L. H. (2011). Piloting the use of teen mentors to promote a healthy diet and physical activity among children in Appalachia. Journal for Specialists in Pediatric Nursing, 16, 16–26. Starkey, F., Audrey, S., Holliday, J., Moore, L., & Campbell, R. (2009). Identifying influential young people to undertake effective peer-led health promotion: The example of a stop smoking in schools trial (ASSIT). Health Education Research, 24, 977–988. Thomas, J. (2009). Using social marketing to address obesity: The ongoing ‘Liverpool’s Challenge’ social marketing programme. Journal of Communication in Healthcare, 2, 216–227. Wang, M., & Struthers, A. (2014). Health buzz at school: Evaluations of a statewide teen health campaign. Paper presented at the 2014 Association for Education in Journalism & Mass Communication Convention, Montreal, Canada. Author Biographies Amy Struthers (MA, University of Nebraska-Lincoln) is associate professor in the College of Journalism and Mass Communications at the University of Nebraska-Lincoln. Her research agenda is focused on health and science communication, with a particular focus on public health messaging to Struthers and Wang 235 adolescents. Her research has been funded by the CDC, USDA and USAID and presented at the CDC National Conference on Health Communications, Marketing and Media. Ming Wang (Ph.D., University of Wisconsin-Madison) is assistant professor in the College of Journalism and Mass Communications at the University of Nebraska-Lincoln. His research interests focus on the impact of new communication technologies, particularly social and mobile media, on advertising, public relations and political communication. His research has been published in leading communication journals, such as Journal of Communication and Communication Research.