Why consider online social marketing as a strategy?

A plethora of strategies to date have: focused on all levels of schooling (the individual, class, and whole school); encompassed proactive, reactive and peer-support and mentoring strategies; targeted different cyber/bullying participant roles (bully, victim, bystander); employed curriculum, inclusion and whole school/multidisciplinary approaches; and explicitly taught social skills to individuals and peers (Polanin, Espelage, & Pigott, 2012; Smith, Schneider, Smith, & Ananiadou, 2004; Thompson & Smith, 2011; Ttofi & Farrington, 2011; Vreeman & Carroll, 2007; Yeager, Fong, Lee, & Espelage, 2015). Often, however, what might appear to work in one context/setting is not transferable or scalable in another (Evans, Fraser, & Cotter, 2014; Vreeman & Carroll, 2007), highlighting the difficulty in determining efficacy across sociocultural contexts and over time and place.

With the advent of cyberbullying in the early 2000s, a new sense of school-related time and place emerged: that which occurs beyond the school day, gate, and jurisdiction. While there is recognized overlap and interplay between what happens in school settings between peers and what happens online (Cross et al., 2009; Smith, 2014; Spears et al., 2008, 2009; Waasdorp & Bradshaw, 2015), to date, prevention and intervention strategies for bullying and cyberbullying have been predominantly school- or class-based, with many drawing on traditional bullying approaches to attempt to reduce cyberbullying. It is relevant therefore to reflect on the efficacy of school-based bullying interventions first.

What do meta-analyses of school-based cyber/bullying intervention research tell us?

Research on traditional antibullying programs is significant, but there is a paucity of evaluative research on cyberbullying interventions. Both, however, report mixed results in terms of efficacy. Merrell, Gueldner, Ross, and Isava (2008) found some evidence from 16 intervention studies reviewed from a 25-year period, which supported the interventions in enhancing students’ social competence, self-esteem, and peer acceptance; in enhancing teachers’ knowledge of effective practices, feelings of efficacy regarding intervention skills, and actual behavior in responding to incidences of bullying at school. They concluded however that interventions were likely to influence knowledge, attitudes, and self-perceptions, rather than actual bullying behaviors. Cook, Williams, Guerra, Kim, and Sadek (2010) in their meta-analysis of the predictors of bullying and victimization concluded that intervening simultaneously at multiple levels (individual, peer ecology, and broader aggression contexts) was the most promising. Ttofi and Farrington (2011) rigorously reviewed the effectiveness of 44 school-based programs to reduce bullying finding that overall, school-based antibullying programs were found to be effective in decreasing bullying and victimization by 20-23% and 17-20%, respectively Only certain elements however were found to be having the effect, and they further argued that “work with peers” should not be used, as it was associated with an increase in victimization, (2011, p. 44). Smith, Salmivalli, and Cowie (2012) challenged this arguing that abandoning work with peers, could undermine existing successful, integrated whole school approaches. This view also does not take account of the importance of youth voice: an enabling and empowering approach to youth-centered, codesigned approaches to cyberbullying, which have evolved from an understanding of children’s rights (United Nations General Assembly, 1989; Spears, Slee, Campbell, & Cross, 2011), and most notably of late, children’s digital rights (Third, Bellerose, Dawkins, Keltie, & Pihl, 2014; Spears & Kofoed, 2013).

Ttofi and Farrington (2011) also noted that effects were highest in age-cohort designs and weakest when more stringent designs (randomized controlled trials) were employed. This highlights the complexities of working and measuring impact in real-world contexts such as schools, where manipulation and control of variables, and establishing implementation fidelity may be problematic, and suggests that research design for these settings, may need consideration as to the most appropriate approach for the context. Polanin et al. (2012) suggested that overall, the reviewed programs were successful, at both a practical and statistically significant level, but found that there were differential effects for high school samples, suggesting that the programs were not as effective for this older age group. In an holistic approach to improving well-being, Durlak, Weissberg, Dymnicki, Taylor, and Schellinger (2011) in their meta-analysis of 213 school-based universal social and emotional learning (SEL) programs found that, compared to controls, SEL participants demonstrated significantly improved social and emotional skills, attitudes and behaviors, and academic performance. This supports a strategic role for promoting holistic protective and preventative attitudinal factors, such as respect for self and others, among all children and youth, as distinct from specifically targeting cyber/bullying and victimization behaviors. Evans et al. (2014) extended Ttofi and Farrington’s meta-analyses, and reported that up to 45% of the 22 studies examined (controlled trials published from June 2009 to April 2013) had no demonstrated program effects for bullying perpetration, and 30% showed no program effects on victimization. More recently, Yeager et al. (2015) found that antibullying programs may not be effective beyond 8^th grade, and that “aging-up” existing materials to be used with older adolescents may be a waste or even misuse of resources, supporting findings by Polanin et al. (2012). Yeager et al. also noted that there have been important advances in psychological theory and improvements to behavior-change techniques, which are underutilized in existing intervention programs.

Jiménez-Barbero, Ruiz-Hernández, Llor-Zaragoza, Pérez-García, and Llor-Esteban (2016) conducted a meta-analysis of 14 high-quality random clinical trials (RCTs) cautiously finding that there were beneficial, but discrete outcomes for schools with regard to victimization, warning, similar to Merrell et al. (2008) that “most of the mean effect sizes were too weak to be considered significant” (2016, p. 173).

Yet, Rigby and Smith (2011) noted that there had been decreases in the prevalence of reported bullying in many countries, consistent with the small reductions in peer victimization following the implementation of antibullying programs. Thompson and Smith (2011; 2012) reviewed the use and effectiveness of antibullying strategies employed in the UK, finding there was much variation across individuals and contexts with regard to the proactive, reactive, and peer-support strategies employed within a whole school approach: specifically recommending, [that] a ‘toolkit of strategies providing a range of interventions was fundamental (2011, p. 8). Rigby and Johnson (2016) similarly examined the prevalence and effectiveness of anti-bullying policies in a convenience sample of Australian schools. Intervention methods, reported by teachers, included: use of sanctions; strengthening the victim; mediation and restorative practices, and less frequently, specific methods such as the Support Group Method or Method of Shared Concern. Proactive approaches employed in the schools included having and using anti-bullying policies, classwork related to bullying, encouraging reporting of bullying, and peer-support initiatives. Reactive approaches were found to be inconsistent, with no single method preferred by schools. Parent responses suggested negative attitudes toward the school concerning how bullying had been dealt with. Alarmingly, students reported that after reporting being bullied to the school, in 70% of cases, it continued.

In sum, the findings from meta-analyses of bullying interventions across all levels of schooling point to many being trialed over time, with very limited effects overall. This has implications for cyberbullying interventions that are underpinned by these existing approaches.

Della Cioppa, O'Neil, and Craig (2015) reviewed 12 international cyberbullying intervention programs against criteria established by Craig, Pepler, and Shelley (2004), which emphasized scientific merit (e.g., control group, multiple informants; randomized assignment) and ease of implementation (e.g., availability of a manual, training, program maintenances, assessment, and who delivered the program) and found most studies lacked scientific merit and most met less than half the criteria, which were premised upon elements of traditional bullying interventions. Of note: few programs were able to engage in a whole school approach; most prevention programs reviewed did not transcend the school (e.g., to family, media), which is a problem when cyberbullying occurs via digital/social media; and most programs targeted 12-13 year olds, meaning that little understanding of program effectiveness is known beyond this age group, for particular time-points (school transitions). Four research-practice gaps were identified: (1) research and theory have been lacking; (2) there are few evidence-based programs tailored to diverse developmental stages; (3) few programs focus on relationships beyond peers and teachers; (4) there is lack of research on what works for whom and critical ingredients or processes for success, including deficits in knowledge about program implementation and processes of change.

If cyberbullying is a school-based problem as argued by Smith et al. (2008), then employing largely school-based prevention and intervention approaches is logical; however, on balance, existing school-based interventions have been shown to be limited in their program effects and cyberbullying is more than a school-based problem today; it is a worldwide societal and public health concern (Spears, Taddeo, Daly, et al., 2015).

Barlett (2017) noted the lack of theory-driven research involved in cyberbullying, and argued interventions, which are underpinned by theory, can be more successful. Citing various studies, e.g., Schultze‐Krumbholz, Schultze, Zagorscak, Wölfer, and Scheithauer (2016) who had incorporated the Theory of Planned Behavior; Doane, Pearson, and Kelley (2014) (the Theory of Reasoned Action); Kowalski, Giumetti, Schroeder, and Lattanner (2014) (General Aggression Model); and Patchin and Hinduja (2011) (General Strain Theory), he suggested that these studies enabled broader psychological, aggression theories, and attitude-based theories to predict cyberbullying behavior. This does not necessarily mean however that these theories readily translate to the online setting.

Online social marketing (OSM) approaches are embedded in public health strategies (VicHealth, 2016) and are underpinned by contemporary psychological theories such as the Nudge Theory (Thaler & Sunstein, 2008) and other transtheoretical models of behavior change, such as those proposed by Prochaska and DiClemente (1983), Ajzen (1985, 1991), and Perugini and Bagozzi (2001).

Meta-analyses such as these, together with the changing online context that is never “off,” highlight opportunities for moving beyond the school setting, and for extending evaluation techniques to: (1) include the online environment as the context in which young people socialize and operate, before, during, and after school; (2) cocreate, coresearch and codesign strategies with youth, so the programs closely resonate with the target audience and age contexts; (3) consider the methodological and ethical challenges of conducting randomized controlled studies with youth online, incorporating informed parental consent; and (4) advance thinking in relation to psychological theory as applied to online studies, such as Nudge Theory (Thaler & Sunstein, 2008) and Goal-Directed Behavior models (Perugini & Bagozzi, 2001).

These collective efficacy findings demonstrate that school-based approaches are necessary, but not universally successful and, while they may demonstrate some success at the individual level, it may not be sufficient especially in a climate of shifting digital and social media communications. Online social marketing (OSM) approaches specifically target attitudes and beliefs as a means of eventually leading to actual behavior change. They also work closely with end-users (in this case, youth) to ensure the relevance, age-appropriateness, credibility, and longevity of the campaigns and messages for maximum impact and reach. In other words, they are not adult driven or designed around adult conceptions of the problem, but are firmly oriented in a participatory codesign approach: with and by young people toward youth-identified issues. If bullying and cyberbullying are systemic group processes (Vreeman & Carroll, 2007), then intervention at only one level, even if that level is the whole school, is unlikely to create change, when their social environment includes the digital setting. Only focusing on schools does not take account of the breadth and depth of the blended on- and offline social interactions young people experience. If most reductions in bullying tend to be relatively small, and related to the children being victimized rather than those engaging in bullying, then it stands to reason that approaches that target broader social relationship issues, such as respect for self and others across all settings, may have merit.

The safe and well online study

The Safe and Well Online Study (2011–2016), a collaboration led by the University of South Australia with Western Sydney University; Queensland University of Technology; and Zuni, a Digital Marketing Agency, trialed and tested an online methodology to examine: (1) whether authentic online social marketing campaigns relevant to young people (end-users) could be successfully cocreated using a participatory design approach (Hagen et al., 2012); and (2) the extent to which safety and well-being messages could positively influence young people’s attitudes and beliefs, and potentially behaviors, in online settings. To do this, innovative methods and measures, including an underpinning theoretical model (Model of Goal-Directed Behavior, Perugini & Bagozzi, 2001) and bespoke online data gathering techniques, which respected the ethical and privacy concerns of conducting online research involving young people, were required. The Model of Goal-Directed Behavior (Perugini and Bagozzi (2001) builds upon the Theory of Planned Behavior (Ajzen, 1991) and proposes that attitudes, anticipated emotions, social norms, and perceived behavioral control are mediated by the desire to perform a particular behavior (see Fig. 6.7). This model is explained briefly:

What did the program look like: four online social media campaigns

Using Bruner’s notion of the Spiral Curriculum (1960) four sequentially developed, youth-centered, codesigned online campaigns were produced (Keep It Tame!; Appreciate A Mate; Something Haunting You? and Goalzie) as part of an inter-related, holistic preventative strategy/program: to nudge young people toward being considerate and respectful of self and others (see Fig. 6.1). The Spiral Curriculum is predicated on cognitive theory, and features revisiting a topic several times, with increasing complexity, so that new learning forms relationships with old learning and previous contexts. The benefits associated with this approach are that information is reinforced and solidified each time the content/message is revisited; there is a logical progression from simple to complex ideas; and early knowledge can be applied to later concepts.

Working closely with young people in a participatory codesign phase for each campaign (see Hagen et al., 2012), each process: (1) identified the “problem” from a youth perspective, which entailed reframing it so that it met their concerns (see Fig. 6.2); (2) determined why it occurred, who it affected, and the magnitude of the problem through a series of workshops; and subsequently (3) codesigned the messaging and the look and feel of the campaign, considering what might work, and for whom in conjunction with design and marketing experts, for delivery in the online space.

Campaign 1 Keep It Tame! (pilot study) (Spears, Taddeo, Barnes, Scrimgeour, et al., 2015) preparatory process identified that the problem with cyberbullying, from youth perspectives, was a lack of respect for self and others. This reframed cyberbullying for the purposes of campaign development from a problem behavior (what not to do), to a positive, strengths-based behavior as a preventative attitudinal underpinning (desired behavior: respect for self and others), with the campaign challenge to Keep It (their use of technology) Tame! (see Fig. 6.3)

Campaign 2 Appreciate A Mate (Spears, Taddeo, Barnes, Collin, et al., 2015) built upon this theme by providing opportunity for explicitly showing respect for others, through distributing positive affirmations via a purpose-created and codesigned app (see Fig. 6.4). Youth rely on their peers for mutual acceptance, validation, and inspiration, which is underpinned by a desire to belong and be valued in their social contexts (Slee, Campbell, & Spears, 2012). Young people identified that image sharing was a problem related to cyberbullying, and determined that body image generally, and how others “saw” them off and online, was one of their most significant concerns (Ivancic, Perrens, Fildes, Perry, & Christensen, 2014). By mobilizing existing popular youth digital practices, such as “liking,” to facilitate desired behaviors and attitudes, in this case, positive body image, self-esteem, and respect for self and others (see https://vimeo.com/94621287), young people could spread positivity through various social media sites, such as Instagram, Tumblr, Facebook, or Twitter.

Rather than tell young people not to cyberbully, Appreciate A Mate sought to reframe negativity in relation to image sharing and subsequent body image, through promoting positive, youth-inspired, and designed affirmations as messages of support

In Campaign 3, Something Haunting You? (Spears et al., 2016a), young people identified that reporting and reaching out for help when facing difficult peer situations, such as bullying and cyberbullying, was challenging (Connolly, Hussey & Connolly, 2014; Cortes & Kochendorfer-Ladd, 2014; Smith, 2014; Spears et al., 2016a). Young people reframed help seeking from a negatively perceived, reactive response, to a positive, “ordinary” and “everyday” action for dealing with stressors, which can follow you around and “haunt you like a Zombie” (see Fig. 6.5). Through using dark, edgy humor and a cultural trope (the Zombie), together with graphics (enhanced comics), and existing digital practices (YouTube videos) relevant to the target audience of young men, young people were encouraged to practice help seeking (Deal With It) in various scenarios and to access a Survival Guide, an ecosystem of support, and online resources with links to reputable formal and informal evidence-informed supports (http://www.somethinghauntingyou.com/).

In Campaign 4, Goalzie (Spears et al., 2016b), young people identified that they needed to practice helping themselves and others, particularly in relation to supporting others when facing such peer difficulties as bullying and cyberbullying. Goalzie was codesigned to facilitate and enable rehearsing of self-help: through setting goals and engaging with friends who can help you achieve those goals (see Fig. 6.6). It gamified help seeking and encouraged fun, peer-to-peer interactions and goal setting, by giving players the opportunity to challenge friends and be challenged by them in fun ways: practicing and normalizing, setting and achieving goals, and seeking help in the process.

How did we do it: methodology

Each campaign development and evaluation strategy comprised 5 components:

(1) A thematic, scoping literature review in consultation with key stakeholders, to identify current, key issues relevant to young people

(2) A Participatory Design (PD) study (Hagen et al., 2012) involving over 500 young people, researchers, digital strategists, key stakeholders, and creative agency professionals in focus groups, generative workshops, online discussions, and peer-research activities;

(3) An Age-Cohort study employing:

a. pre-post experimental survey design using items related to each campaign theme; internet use and practices; social respect (user-designed scale); as well as measures of: mental health and well-being (Lovibond & Lovibond, 1995); social connectedness (adapted from Lee, Dean, & Jung, 2008; Lee, Draper, & Lee, 2001); help seeking (adapted from Rickwood, Deane, Wilson, & Ciarrochi, 2005); and cyberbullying experiences (Cross et al., 2009),

b. random allocation to control and campaign exposure groups by employing coding during a short “contained” (prepublic release) research period

c. the Model of Goal-Directed Behavior (Perugini & Bagozzi, 2001) to underpin prediction of attitude and behavioral change (see Fig. 6.7)

d. a Unique ID assigned to each participant and mapped across the campaign platform and survey instrument to enable matching across pre-post surveys and campaign engagement analytics, with the Model of Goal-Directed Behavior (Perugini & Bagozzi, 2001).

(4) a Digital Data Collection component, which progressively refined active and passive digital data collection methods over the four campaigns to explore campaign engagement; including the following innovations from Campaigns 3 & 4:

a. capturing and matching young people’s “real time” engagement, with the campaigns in the contained (prepublic release) research period, with survey responses at the individual (Unique ID) participant level;

b. extending the passive (nonidentifiable) analytic data collection beyond the contained period, to “in the wild” (postpublic release) to explore organic campaign engagement in naturalistic settings

c. linking of the Unique ID with self-report measures and Google Analytics (Campaigns 2 and 3) and a bespoke back-end database (Campaign 4), with time-stamped data touchpoints, to ascertain extent and nature of individual engagement with the campaigns.

(5) a postcampaign qualitative study exploring key messaging and outcomes through interviews, and or focus groups.

The pilot cohort study (Campaign 1) recruited participants (N = 165) aged 12-18 years (M = 14.53, SD = 1.44, n = 163; 2 missing) from South Australian schools, and trialed and tested the validity of instruments and scales when employed in online surveys; and the efficacy of employing the Model of Goal-Directed Behavior (MGB) (Perugini & Bagozzi, 2001) in an online setting to predict attitude and behavioral change (Spears, Taddeo, Barnes, Scrimgeour, et al., 2015). Data gathered provided comparative evidence of online practices and the mental health and well-being context of this pilot group. Subsequent campaigns recruited young people via their parents, from dedicated online research panels, ensuring representative samples required to test the campaigns. Over the course of the Safe and Well Online Study (2011–2016), over 5000 young people participated with informed parental consent (see Table 6.1). In addition, 15 sector partners acted as key stakeholders, and 9 schools contributed to the user testing and participatory design workshops.

What did we find and what does it mean practically²?

Participatory Design: Campaign Development

Through the Participatory Design process (Hagen et al., 2012; Spears, Taddeo, Collin, et al., 2016) negative peer issues, such as cyberbullying, were reframed as positive messages for their peers, to be respectful of self and others; be affirming; see help seeking as a positive, strengths-based action; and recognize that with practice, goal setting for help seeking is easier. Collectively, this study found that shifting adult-defined and expert-driven practices of designing interventions for youth to those designed with and by youth, enabled young people to act as coresearchers alongside the adults, refining the project themes, aims, campaign concepts, development, production, and review together, so that meaning was shared, and knowledge transferred, further maximizing the likelihood of relevancy and currency in the messaging of the campaigns.

The Cohort Studies: The well-being context³

The contexts into which the four campaigns were delivered (2012–2016) demonstrated that young people were largely well and had a positive outlook on life; the majority were socially connected and within normal ranges for mental health. Those more likely to be socially connected, demonstrated respectful behavior and had a greater awareness of the norms around social respect. Approximately one in five however reported an absence of hope and lack of confidence in society. This was particularly the case for cyber bully-victims, and females, who, demonstrated significantly poorer mental health than males. Nearly half the young people in the pilot cohort study accessed the internet after 11pm, with one-third of these reporting they did so more than four nights a week. Higher levels of antisocial behavior online, and significantly higher levels of cybervictimization, were evident for those regularly online after 11pm than those who were not. This finding suggests that it may not be how often young people go online that is a factor in cyberbullying, but when they do.

Three campaigns and three years later, using representative age cohort samples: (see Table 6.1) (Spears, Taddeo, Barnes, Collin, et al., 2015; Spears, Taddeo, Barnes, Scrimgeour, et al., 2015; Spears et al., 2016a, 2016b; Spears, Taddeo, Collin, et al., 2016) the contexts into which they were released had changed very little from the Pilot Study (N = 165): most young people were still largely healthy and well; the majority felt socially connected on- and offline; and were respectful of others. Consistent across all campaigns, parents/carers continued to provide a primary source of support for young people on a daily basis and most young people were in the normal or mild range for depression, anxiety, and stress.

However, approximately one in five continued to experience severe to extremely severe depression and anxiety: One in 10 experienced severe or extremely severe stress; and cyber bully-victims continued to be a highly vulnerable group, being significantly less socially connected and more anxious, stressed, and depressed than those with no experience of cyberbullying, and demonstrated significantly lower levels of respect for others (Spears, Taddeo, Barnes, Collin, et al., 2015; Spears, Taddeo, Barnes, Scrimgeour, et al., 2015; Spears et al., 2016a, 2016b; Spears, Taddeo, Collin, et al., 2016).

Building new knowledge to inform cyberbullying interventions: Nudging and Entry points

“Reach” is about reaching the intended audience, and “impact” relates to the effect that interventions may have on people, organizations, or systems. In terms of reach, the videos (Campaigns 1 & 3) were viewed over 1.76 million times; over 85,500 messages of positivity and appreciation were created and shared (Campaign 2), and over 1100 challenges were created via Goalzie (Campaign 4) during the contained trial periods (prepublic release). Impact, however, is noted at several levels. At the individual level, and similar to other previously noted interventions in this field, change was not able to be established to any significant effect and further highlights the complexity of measuring actual change online where the setting is fluid, difficult to contain, viral, and freely networked.

Most importantly though this study’s impact is at the broader system level and responds directly to the research-practice challenges highlighted by Della Cioppa et al. (2015) and noted previously: it has developed and trialed an innovative online research methodology, grounded in theory, which has demonstrated potential entry points for interventions to maximize their impact and effect. Careful consideration throughout was given to the methodology: sampling; tracking methods by Unique ID; approaches to operationalizing controlled experimental design in a contained online setting; and functionality and limitations of platforms.

To reiterate, the Model of Goal-Directed Behavior (MGB) (Perugini and Bagozzi (2001) proposed that attitudes, anticipated emotions, social norms, and perceived behavioral control are mediated by the desire to perform a particular behavior (see Fig. 6.7).

Investigations into relationships between key constructs of the model (MGB) (Perugini & Bagozzi, 2001) and campaign outcomes: respect; helping others feel good about themselves; help seeking and goal setting, demonstrated where best to target future initiatives and interventions, viz. social norms, attitudes, and perceived control. These were found to be key entry points necessary for nudging attitudinal and behavioral change for young people’s well-being, including cyberbullying.

Desire, in turn, is a determinant of intentions, with frequency and recency of past behavior contributing either directly or indirectly to determining the behavior. The desire then in this case to be respectful of self and others, to be affirming, to seek help and set goals, constitutes important motivators in the decision-making process to actually enact change. Predicting behavioral change rests on being able to measure these constructs in relation to the desired behavioral and attitudinal outcomes over an appropriate timeframe in a controlled environment: a complex task in an online setting, where users employ different devices, platforms, and systems, and the viral and interconnected nature of the web is fundamental.

Social marketing, especially using digital media, holds enormous potential for engaging young people, and it may be that nudging behavior in less obvious, but positive ways is more realistic than determining actual behavioral change. Thaler and Sunstein (2008) defined Nudge Theory as “any aspect of choice architecture that alters people’s behavior in a predictable way, without forbidding any options or significantly changing their economic incentives… Putting fruit at eye level counts as a nudge. Banning junk food does not” (p. 6).

Analytics engagement data from Campaign 3 demonstrated ongoing engagement with the messaging some years later, and passive data, obtained through real-time collection of data from users as they engaged with devices, websites, or Apps, generally demonstrated extensive engagement and reach, as noted at the beginning of this section. The nudge effect here thus relates to putting the youth-centered and driven campaign messaging online in engaging ways: metaphorically “putting the fruit at [their online] eye level.”

It is possible that these campaigns have “nudged” participants, as there is evidence to support return use and ongoing use of the campaign artifacts beyond the testing period: so-called in the wild, e.g., the videos and Appreciate a Mate messages are still being utilized, but since the project has finished, we are unable to continue to monitor uptake and spread.

Measures sensitive enough to capture shifts in attitudes in short time frames (the “nudge”) are needed, particularly given that attitudinal and behavioral change is a long-term endeavor and the online campaign testing periods were relatively short, due to needing to “contain” the integrity of the controlled experimental design.

Given the majority of young people across all campaigns reported positive attitudes and emotions toward behaving respectfully online, an appreciation of the social norms associated with respect, and feeling in control of their desire and intent to behave respectfully online, it is reasonable to suggest that these are the entry points for online social marketing campaigns to successfully nudge young people toward actual change, without telling them directly what they need to do. Further to this, the qualitative data (Study 5) found that friends played a key role in motivating others to behave respectfully, including giving disapproval for disrespectful behavior, effecting “nudging” (Thaler & Sunstein, 2008) young people’s behaviors in predictable ways.

Strengths, limitations, and conclusions

What have we learnt⁴?

Online social marketing approaches to intervention have merit, in that the affordances of social media are well suited to campaigns seeking to foster social connections as protective factors. They have reach far beyond the school settings, and can be revisited often, well beyond the initial release, reinforcing key messages over time, potentially nudging young people toward the desired outcomes. Campaigns that mirror youth language and behaviors (humor, “liking,” posting, sharing) resonate well and increase the likelihood of uptake and reach. Constructing cyberbullying initiatives and interventions with young people is critical –the way forward cannot be mapped without their input and voice. Young people provided a new way of approaching cyberbullying initiatives by applying a positive, strengths-based lens to this serious youth problem, and reframed it into desirable behaviors to be addressed positively through the campaigns such as respect for others, help seeking, and making others feel good about themselves. Determining behavioral change online proved elusive, however, and therefore not dissimilar to many offline interventions. However, the key findings from this study suggest key entry points for interventions to nudge young people toward change for well-being are addressing social norms, attitudes, and perceived control.

The ethical, design, and technical challenges in evaluating campaigns in real time and in the public domain are substantial. The research methodology, which brought together self-report and engagement data and applied both sets of data to the Model of Goal-Directed Behavior with a nationally representative sample of young people aged 12–18, is promising, and future research should continue to develop and test these innovative techniques. Young people’s engagement with campaigns, 2, 3, and 4 was mapped in real time; in naturalistic settings; at the individual level by unique ID; across relevant data touch points; then collected in databases and merged for analyses. This enabled campaign engagement data to be mapped against self-report data. In this way, data collection did not only rely on what young people told us, but also used passive data that provided a “real” indication of how young people were actually engaging with the campaigns.

There were however a number of project limitations, including the potential for self-selection bias, and ethical requirements restricting young people’s engagement with the full functionality of some of the campaigns: Campaign 4 especially where young people were required to have a Facebook account in order to establish peer connections for the challenges. Importantly, as a series of age-cohort designs, claims about cause-and-effect relationships cannot be made.

However, the project does provide detailed insights into the well-being contexts into which each of the campaigns were delivered. Further, the project has contributed to understandings about the use, potential benefits and limitations of using social marketing to address cyberbullying with young people; proposed areas for ongoing investigations, and highlighted the value of developing initiatives where youth participation and voice underpins intervention processes to maximize the potential reach and impact of cyberbullying initiatives.