Message Design and Audience Engagement with Tobacco Prevention Posts on Social Media

Article

Abstract

Understanding the appropriate medium to communicate health promotion messages is vital for improving personal and societal health. As increasingly more people utilize social media for health information, public health practitioners use these platforms to engage an existing audience in health promotion messages. In this study, the relational framing theory was used as a lens for studying how message framing may influence social media audience engagement. Specifically, we assessed how posts from Tobacco Free Florida’s Facebook page were framed as either dominant-submissive or affiliate-disaffiliate to an implied audience of either smokers, nonsmokers, active quitters, or a mixed audience, and the extent to which a direct call for engagement, in terms of a request to comment, like, or share the post, was used for audience engagement. A three-way interaction for the level of engagement through comments was significant, F(3217) = 7.11, p < .001, ηp2 = .09, and showed that framing, a call for engagement, and varying implied audience choice played a role in audience engagement with smoking cessation posts on social media. Implied audiences of Tobacco Free Florida’s posts included smokers, those who are trying to quit, and nonsmokers as health promotion can be targeted at the individual’s health, social support infrastructure, or the well-being of the society, and implications for strategic message design and audience targeting are discussed.

Keywords

Social media engagement Message design Message framing Smoking cessation 

Introduction

The purpose of health education and promotion efforts is to encourage individuals to make decisions that contribute to their own and community well-being [1]. For example, an organization may urge a person to quit smoking for reasons that benefit the individual, such as to reduce that person’s likelihood of developing a smoking-related disease, or the community, such as to decrease the amount of secondhand smoke in the environment. Understanding the appropriate medium to communicate health promotion messages is vital for improving personal and societal health. Social networking sites are potential platforms for health promotion messages, which are becoming increasingly pervasive [2]. As more people utilize social media for health information, the CDC recommends that practitioners use these platforms to engage an existing audience in health promotion messages [2, 3]. Within the context of social media, audience engagement is conceptualized as an interactive process resulting in a two-way communication between a health promotion organization and the public [4]. In order for an organization to accomplish their health promotion goals, Neiger et al. recommend that organizations frame a chosen message to a specific intended audience and evaluate its efficacy at generating high user-engagement levels [4]. Their advice falls within the broader approach of message tailoring, which seeks to deliver specific content to individuals that aligns with their personal background and needs [5]. Evaluating how an organization frames messages on social media can provide insights and facilitate effective interaction with a chosen audience.

In this study, we used the relational framing theory (RFT) as a lens for studying how message framing may influence audience engagement on social media. RFT states that individuals process messages to make inferences about their relationships across two independent frames, dominance-submissiveness and affiliation-disaffiliation [6]. The dominance-submissiveness frame can be interpreted as “the degree to which one person controls, influences, or has status over the other,” while the affiliation-disaffiliation frame “captures the appreciation, esteem, or solidarity one person has for the other” [7]. The theory posits that people interpret messages as primarily dominant-submissive or affiliate-disaffiliate instead of a blend of the two frames, and that frames serve as mental shortcuts to facilitate quick interpretations of potentially ambiguous messages communicated in fast-paced environments [6, 7]. RFT also specifies that people will gauge the level of communication engagement, which can act as an intensifier to bolster the person’s interpretation of either frame [6]. Extended to the context of communication on social media, user comments have been shown to indicate audience engagement and amplify message effects [8].

The researchers in this study assessed how posts from Tobacco Free Florida’s (TFF) Facebook page were framed as either dominant-submissive or affiliate-disaffiliate to an implied audience of either smokers, nonsmokers, active quitters, or a broader mixed audience for messages that could be relevant and applicable to any audience including smokers and nonsmokers, and the extent to which a direct call for engagement, in terms of a request to comment, like, or share the post, was used for audience engagement.

RQ1: What are the differences in the use of a call for engagement, message framing, and implied audience as message strategies for smoking cessation communication on social media?

While the first research question focused on how TFF framed messages to their implied audience, we also aimed to assess the audience’s response. The number of comments per post was used as an indicator of social media audience engagement, to evaluate the influence of message framing on audience engagement with Facebook posts.

RQ2: What message strategies are most effective in engaging specific audiences in smoking cessation communication on social media?

Method

Data were collected from TFF Facebook page, which predominantly features posts by the TFF staff. Excel Power Query was used to collect official TFF posts that appeared between July 2015 and June 2016 (N = 233) and their associated meta-data (i.e., post date and number of comments per post). Data collection and analysis protocol have received approval from an institutional review board.

A codebook was developed to code the posts for message framing, the presence of a call for engagement, and implied audience. Coding for message framing was guided by prior research [6, 9]. First, 20 posts from a different time period were coded, and categories were discussed to ensure the clarity of the coding rules. Then, 50 posts from the data reported in this study were coded to assess intercoder reliability [10], which reached high Cohen’s kappa values between .84 and .92 for the coding categories. Finally, we evenly split and coded the remaining posts. Dominance-submissiveness (D-S) frame was operationalized as communication from the position of power and evidence that uses linguistic markers associated either with dominance, including self-confidence, determination, persuasiveness, headstrong presentation of opinions, and confidence, or with submissiveness manifested in the use of hedges and indirect, cautious language [9, 11]. Affiliation-disaffiliation frame (A-D) was operationalized as including the communication of similarity or dissimilarity and focusing on commonality or difference in values and norms beyond smoking cessation [12]. The presence of a call for engagement was operationalized as featuring a direct request that audience members share a post or submit a response to a posted question. The dependent variable, audience engagement, was operationalized as the number of audience comments in response to a particular post. Finally, posts were coded for the implied audience of nonsmokers, active quitters, smokers, or a mixed audience. Examples of post representative of each coding category are presented in Table 1.
Table 1.

Examples of representative messages for post frame, audience, and engagement call coding categories

Coding category

Example

Frame

 Dominance/Submissiveness

Five years after you quit smoking, your risk of bladder cancer is cut in half!

 

While on your quit journey, you may face tough moments. If you quit, how did you overcome these challenges?

 Affiliation/Disaffiliation

Happy Parents’ Day! Share if you are a tobacco-free parent.

 

Protect your loved ones from secondhand smoke. It does not just hurt the smoker. #SHSExposed.

Audience

 Nonsmokers

Tobacco-free can mean a longer life with the ones you love. Happy Father’s Day to all of the tobacco-free dads.

 Active quitters

Your body can reverse the damage caused by smoking. Have you noticed any changes?

 Smokers

Pack-a-day smokers in Florida can save $20,000 in 10 years by quitting smoking.

 General audience

Today is the 15th anniversary of the 2001 Surgeon General’s Report on Women and Smoking.

 Engagement call

We can help you quit any form of tobacco, including chew and dip. Ask us how!

Results

Three separate chi-squire tests and subsequent review of the standardized residuals were performed to assess whether message strategies differed among the TFF posts (RQ1). The first test examined the relationship between the implied audience and message frame and was significant, χ 2 (3, N = 233) = 33.52, p < .001. A-D frame was used more often to address nonsmokers and a mixed audience but less often to address active quitters and smokers. The second test examined the relationship between the implied audience and presence of a call for engagement and was also significant, χ2 (3, N = 233) = 22.47, p < .001. Calls for engagement were used more often to address nonsmokers and a mixed audience but less often to address smokers. The third test examined the relationship between message frame and presence of a call for engagement and was also significant, χ 2 (1, N = 233) = 14.25, p < .001. Calls for engagement were more frequent in combination with A-D frame and less frequent with D-S frame.

An ANOVA tested whether call for engagement and message framing would lead to different levels of engagement among implied message audiences (RQ2, Table 2). The test showed significant main effects of a call for engagement, F(1232) = 11.02, p < .001, ηp2 = .05, and implied audience, F(3230) = 5.03, p < .01, ηp2 = .07, but not message framing, F(1232) = .17, p = .68, ηp2 = .01. Post hoc analyses revealed that messages addressed to nonsmokers and messages with a call for engagement received a significantly greater number of comments. Two-way interactions of implied audience with call, F(3230) = 4.05, p < .01, ηp2 = .05, and frame, F(3230) = 7.33, p < .001, ηp2 = .09, were significant. The presence of a call resulted in a significantly greater number of comments when messages were addressed to nonsmokers and active quitters. However, a presence of a call had no effect on smokers or a mixed audience. Based on the frame used, messages that used D-S frame generated significantly more comments when addressed to nonsmokers than to any other audience, and messages that used A-D frame generated significantly more comments when addressed to active quitters.
Table 2.

ANOVA results for call, frame, and implied audience effects on the number of comments per post

Tested effect

n

M (SD)

F

p

η2

Call for engagement (C)

  

11.02

.001

.05

Absent

201

50.07 (16.30)

   

Present

32

157.77 (28.05)

   

Frame (F)

  

.17

.68

.01

Dominance-

167

97.27 (19.23)

   

submissiveness

 

110.57 (26.13)

   

Affiliation-disaffiliation

66

    

Implied audience (A)

  

5.03

.01

.06

Nonsmokers

19

169.91 (30.96)

   

Active quitters

41

148.23 (37.56)

   

Smokers

118

61.63 (36.07)

   

Mixed

55

35.90 (23.25)

   

C × F

  

.44

.51

.01

C × A

  

4.05

.01

.05

F × A

  

7.33

.001

.09

C × F × A

  

7.11

.001

.09

Finally, the ANOVA indicated that a three-way interaction for the level of engagement through comments was significant, F(3217) = 7.11, p < .001, ηp2 = .09. For messages without a call for engagement, no between-subject differences were observed for A-D frame, but for D-S frame messages addressed to smokers attracted more comments than messages addressed to active quitters, SE = 25.91, p < .001, 95% CI [5.83, 107.95]. For messages that contained a call for engagement, comments to both frames differed based on implied audience. For D-S frame, messages addressed to nonsmokers attracted more comments than messages to active quitters, SE = 105.08, p < .001, 95% CI [281.56, 695.78], smokers, SE = 93.99, p < .001, 95% CI [311.96, 682.45], and a mixed audience, SE = 470.67, p < .001, 95% CI [263.56, 677.78]. For A-D frame, messages addressed to active quitters attracted more comments than messages to nonsmokers, SE = 111.46, p < .001, 95% CI [149.83, 589.17], smokers, SE = 157.62, p < .05, 95% CI [1.84, 623.17], and a mixed audience, SE = 98.93, p < .001, 95% CI [218.33, 608.31].

Discussion

The goal of this study was to examine the differences in audience engagement in relation to varying message strategies in the context of the posts on the Tobacco Free Florida Facebook page. This study showed that differences in message design strategies can explain variability in the level of audience engagement. The examination of the interaction of a call for engagement, message framing, and implied audience led to the following insightful findings. First, the data showed that A-D frame was used more often to address nonsmokers and a mixed audience but gained more comments when addressed to active quitters. Second, a call for engagement was used more often to address nonsmokers and a mixed audience. This strategy resulted in more comments from nonsmokers in combination with D-S frame. However, in the presence of A-D frame, it had an effect only when addressed to active quitters. Finally, messages to current smokers accounted for more than a half of all posts, but only D-S framing without a call for engagement led to increased commenting for messages to this implied audience. Engaging with nonsmokers could support the development of social capital for those who are attempting to quit, and findings presented here suggest that strategic message framing decisions can lead to more effective engagement with target audiences.

This study has a number of theoretical and practical implications, which will be discussed next, but two limitations should be acknowledged. First, this study presents a focused evaluation of the posts as part of a tobacco cessation campaign, but audience responses can further attenuate or amplify the effects, which could be assessed by the examination of the content of audience comments [8, 13]. Therefore, building on the results of this study, future research could look at the sentiment and themes of the audience comments. Second, the posts were coded to assess implied audience, but the extent to which these audience groups associate with the messages that target them is not evident. Subsequent experimental studies would be necessary to assess the effects of audience segmentation.

Previous research on the relational frame theory has focused on the content and linguistic features of messages [7, 9], but the role of the intended audience has not received much attention. This study showed that the differences in audience engagement could not be explained by the main effect of frame but rather by its interaction with the intended audience. This finding leads to a proposition that implied message audience moderates the relationship between message framing and audience engagement. In addition, prior research indicates that a person’s characteristics or perceptions can influence their interpretation of a message within the dominant-submissive or affiliate-disaffiliate frames [6, 7]. This finding is an important consideration for smoking cessation, as smoking is a highly stigmatized behavior [14]. A smoker may already view a message sender as judging their behavior, which could influence their interpretation of how the post and willingness to engage on social media with the organization.

Health promotion organizations aim to convey information to people with the goal of influencing behavior and improving societal health. For example, TFF uses social media to facilitate smoking cessation. One step to this positive health change can be interactions between organizations and individuals on pervasive social media sites [4]. TFF engaged with smokers, nonsmokers, and active quitters through their Facebook page. The results of this study showed that message framing does influence audience engagement on social media. The practical implications of these results can inform the social media engagement practices of health promotion organizations, such as TFF, by providing additional evidence to support efforts that focus on understanding their audience, tailoring messages to a specific group, and evaluating the efficacy of their approach. These suggested guidelines reinforce Neiger et al.’s recommendations for health promotion practitioners [4].

This study also showed that RFT and social media audience analysis are applicable in the context of public health policy. Stakeholder engagement and feedback are essential components of policy development, and the evaluation of audience engagement can provide evidence of the topics or stakeholder groups that are active participants in the policy development dialog [15]. At the same time, social media analysis could serve as an indicator that engagement of particular groups might be insufficient. As this study showed, the number of comments to messages addressed to smokers is lower than that addressed to nonsmokers and active quitters, but the strategic use of framing could help address this inequity. In conclusion, message framing, a call for engagement, and varying implied audience choice played a role in audience engagement with smoking cessation posts on social media. Strategic communication and tailored message design can facilitate educational efforts and support stakeholder engagement in public policy debates.

References

  1. 1.
    World Health Organization. 2016. Health promotion.Google Scholar
  2. 2.
    Centers for Disease Control and Prevention. 2011. The Health Communicator’s Social Media Toolkit.Google Scholar
  3. 3.
    Perrin, Andrew. 2015. Social media usage: 2005–2015.Google Scholar
  4. 4.
    Neiger BL, Thackeray R, Burton SH, Giraud-Carrier CG, Fagen MC (2013) Evaluating social media’s capacity to develop engaged audiences in health promotion settings: use of Twitter metrics as a case study. Health Promot Pract 14:157–162. doi:10.1177/1524839912469378 CrossRefPubMedGoogle Scholar
  5. 5.
    Kreuter MW, Wray RJ (2003) Tailored and targeted health communication: strategies for enhancing information relevance. Am J Health Behav 27:S227–S232CrossRefPubMedGoogle Scholar
  6. 6.
    Dillard JP, Solomon DH, Samp JA (1996) Framing social reality: the relevance of relational judgments. Commun Res 23:703–723. doi:10.1177/009365096023006004 CrossRefGoogle Scholar
  7. 7.
    McLaren, Rachel M., and Denise Haunani Solomon. 2014. Relational framing theory. In Engaging Theories in Interpersonal Communication: Multiple Perspectives, ed. D. O. Braitwaite and P. Schrodt, 115–127.Google Scholar
  8. 8.
    Strekalova, Yulia A. 2016. Health risk information engagement and amplification on social media: News about an emerging pandemic on Facebook. Health Education & Behavior: 1090198116660310. doi:10.1177/1090198116660310.Google Scholar
  9. 9.
    Solomon DH, Dillard JP, Anderson JW (2002) Episode type, attachment orientation, and frame salience: evidence for a theory of relational framing. Hum Commun Res 28:136–152. doi:10.1111/j.1468-2958.2002.tb00801.x CrossRefGoogle Scholar
  10. 10.
    Krippendorff, Klaus. 2012. Content Analysis: An Introduction to Its Methodology. SAGE.Google Scholar
  11. 11.
    Burgoon JK, Johnson ML, Koch PT (1998) The nature and measurement of interpersonal dominance. Commun Monogr 65:308–335. doi:10.1080/03637759809376456 CrossRefGoogle Scholar
  12. 12.
    Mccroskey JC, Richmond VP, Daly JA (1975) The development of a measure of perceived momophily in interpersonal communication. Hum Commun Res 1:323–332. doi:10.1111/j.1468-2958.1975.tb00281.x CrossRefGoogle Scholar
  13. 13.
    Picanço L, Biancovilli P, Jurberg C (2016) Beyond the drama: the beautiful life in news feeds on cancer. J Cancer Educ:1–5. doi:10.1007/s13187-016-1094-2
  14. 14.
    Graham H (2012) Smoking, stigma and social class. Journal of Social Policy 41:83–99. doi:10.1017/S004727941100033X CrossRefGoogle Scholar
  15. 15.
    Riege A, Lindsay N (2006) Knowledge management in the public sector: stakeholder partnerships in the public policy development. J Knowl Manag 10:24–39. doi:10.1108/13673270610670830 CrossRefGoogle Scholar

Copyright information

© American Association for Cancer Education 2016

Authors and Affiliations

  1. 1.College of Journalism and CommunicationsUniversity of FloridaGainesvilleUSA

Personalised recommendations