Abstract
We investigated the effects of message framing and online media channel on young adults’ perceived severity of human papillomavirus (HPV), perceived barriers and benefits of getting HPV vaccination, and behavioral intention to get vaccinated. An experiment was conducted with 142 college students. We found an interaction effect: The loss-framed message posted on Facebook was more effective in increasing the number of people who expressed their willingness to get HPV vaccination than the gain-framed message presented on Facebook. However, this framing effect was not found when the identical message was presented on an online newspaper. People’s perceptions of severity of HPV and barriers of getting HPV vaccination were also influenced, depending on which media channel the information was circulated.
Human papillomavirus (HPV) is the main cause of cervical cancer and genital warts (Centers for Disease Control and Prevention [CDC], 2016). It is estimated that 79 million Americans are currently infected with the virus and 14 million new infections occur yearly (CDC, 2016). Each year, in the United States alone, HPV causes 81% of about 33,000 newly diagnosed cases of cancer in the areas that HPV is frequently found (CDC, 2016). The Food and Drug Administration recommends the HPV vaccine for individuals who are 9 to 26 years old in order to reduce the probability of developing HPV-related cancers/sexually transmissible diseases. More specifically, the Advisory Committee in Immunization Practices recommended that HPV vaccination be initiated at age 11 or 12 years or as early as 9 years old. For those who have not been vaccinated previously, it is recommended for females aged 13 through 26 years and for males aged 13 through 21 years (Markowitz et al., 2014).
The HPV vaccine can prevent most cases of cervical, vaginal, and vulvar cancer in women and genital warts and anal cancer in both men and women (Mayo Clinic, 2016). However, the rate of vaccination has been relatively low. For example, only 40% of young women and 8.2% of young men aged 19 to 26 years old were vaccinated in 2014 (W. W. Williams et al., 2016). Thus, promoting HPV vaccination is critical in preventing HPV-related cancers. To ensure the success of public health promotion, using appropriate media channels and designing effective messages are two critical elements that go hand in hand. The main question is how to promote HPV vaccination among young adults to prevent devastating consequences later in their lives. How do we design and deliver messages that will motivate them to adopt this preventive measure (i.e., HPV vaccination)?
With the emergence of social media, particularly social networking sites (SNSs), health communication/promotion enters into a world of new media, presenting new ways to deliver health information and persuasion. Social media, Web-based applications that allow users to share content with other users, are considered to be an effective outlet that has greatly influenced how people are looking for health information and engaging in health-related behaviors (CDC, 2011). In the traditional media such as TV and newspapers, health information is merely conveyed to the public and expected to influence the public’s knowledge, attitudes, or behaviors. However, through social media like SNSs, people themselves not only become a messenger of health information but also express, share, and disseminate their own opinions, possibly serving as health advocates. Although SNSs (e.g., Facebook) are not specifically designed for health promotion, they become increasingly popular platforms to share and discuss health-related topics (The Nielsen Company, 2011). However, the impacts of how messages are delivered/shared via SNSs are understudied, particularly in the context of health communication and promotion.
Although the deliberation of health information is important, content/message itself is critical in promoting health. Message framing has been shown to influence how people respond to given messages (Detweiler, Bedell, Salovey, Pronin, & Rothman, 1999). Previous studies have found that loss-framed messages, which emphasize the negative consequences of not adopting a recommended behavior, are more effective than gain-framed messages, which emphasize the positive consequences of adopting a recommended behavior, for vaccination such as HPV vaccine, mumps and rubella, and H1N1 (Gerend, Shepherd, & Monday, 2008; Lechuga, Swain, & Weinhardt, 2011; Nan, 2012). Despite the promising effects of using loss-framing in encouraging vaccination, the underlying mechanism as to why loss-framing works better calls for further investigations.
According to health belief model (HBM), one’s decision to adopt or reject a recommended health behavior depends on several factors; perceived susceptibility, severity, benefits, and barriers (Janz & Becker, 1984). To encourage individuals to get vaccinated, messages should be designed to increase perceived severity of a disease and benefits of a recommended behavior while reducing perceived barriers of the recommended behavior. In this sense, message framing might highlight these key factors differently and in turn influence the public’s willingness to get vaccinated. Thus, in this study, we investigated whether using different message framing and media influences the public’s perceived severity/benefits/barriers and their willingness to get vaccinated.
Message Framing in Health Communication/Promotion
According to prospect theory (Kahneman & Tversky, 1979), “People respond differentially to messages depending on how these messages are framed” (Detweiler et al., 1999, p. 189). Two distinctive framing types were proposed (Tversky & Kahneman, 1981): gain-framing versus loss-framing. Gain-framing contains messages emphasizing benefits (or advantages) of adopting a recommended behavior (Rothman & Salovey, 1997). Loss-framing refers to messages emphasizing losses from the nonadaptation of a recommended behavior (Rothman, Bartels, Wlaschin, & Salovey, 2006; Rothman & Salovey, 1997).
Some found that gain-framed messages were shown to be more effective for prevention behavior (e.g., not smoking or wearing sunscreen) while loss-framed messages were more effective for illness detection behavior (e.g., mammogram or HIV testing; Rothman et al., 2006; Rothman, Martino, Bedell, Detweiler, & Salovey, 1999; Rothman & Salovey, 1997). However, although vaccination is considered as a prevention behavior, the findings of message framing effect on vaccination are shown to be the opposite from the original prediction: Loss-framing was found to be more effective than gain-framing on increasing individuals’ willingness to get vaccinated (Gerend et al., 2008; Lechuga et al., 2011; Nan, 2012). Although more recent studies found message framing effect interacts with other variables like mood and personal relevance (Wirtz, Sar, & Ghuge, 2015), temporal distance (Wen & Shen, 2016), and even cultural worldviews (Nan & Madden, 2014), loss-framing overall has been found to be more effective than gain-framing in the context of HPV vaccination. Therefore, the following hypothesis is proposed.
Although a consistent framing effect has been identified on individuals’ intention to get HPV vaccine, the underlying mechanism as to why loss-framing works better than gain-framing is still unknown. The main question is: What factors mediate the effect of framing on individuals’ intentions to get vaccinated?
Health Belief Model in the Context of HPV Vaccination Campaign
According to HBM, people’s preventive health behavior is influenced by several factors (Janz & Becker, 1984). The theory posits that for people to engage in preventive health behavior, they would need to believe that they (1) are susceptible to a disease (perceived susceptibility), (2) should be concerned about the severity of a disease (perceived severity), (3) should consider preventive health behavior to be a beneficial action (perceived benefits), and (4) need to overcome barriers such as cost, pain, and inconvenience (perceived barriers; Becker, 1974).
HBM has been empirically tested as the bases for health promotional campaigns to determine which factors are the strongest/weakest predictors. Previous studies presented that perceived barrier, benefit, and severity would be predictors for prevention behaviors (Harrison, Mullen, & Green, 1992; Janz & Becker, 1984; Reiter, Brewer, Gottlieb, McRee, & Smith, 2009). Janz and Becker (1984) found perceived barriers to be the strongest predictors while perceived severity was the weakest predictor. Harrison et al. (1992) and Carpenter (2010) found that perceived benefits and barriers were the strongest factors. In terms of vaccination behaviors, Reiter et al. (2009) found that perceived barriers to vaccination strongly correlated with vaccine initiation.
In this study, we suspected that different types of message framing may influence people’s perceived severity of HPV and perceived barriers and benefit of getting HPV vaccine. Since loss-framing is known to highlight losses of not getting HPV vaccine, we hypothesized that loss-framed messages would increase individuals’ perceived severity of HPV (Hypothesis 2) and decrease their perceived barriers of getting HPV vaccine (Hypothesis 3) while gain-framing would increase perceived benefits of getting HPV vaccine (Hypothesis 4).
Medium Effects: SNSs Versus Online Newspapers
It is important to know what would be the best media channels to deliver health messages to the public to garner the success of health promotion. In 1964, Marshall McLuhan coined the famous phrase, “The medium is the message,” advocating the importance of media characteristics. Traditionally, print media has been a popular selection for health promotion on various topics, such as antismoking (McAlister et al., 2004; Owen, 2000; Pierce, Macaskill, & Hill, 2002), melanoma prevention (Hill, White, Marks, & Borland, 2002), emergency contraception (Trussell, Koenig, Vaughan, & Stewart, 2001), alcohol and drug use reduction (Miller, Toscova, Miller, & Sanchez, 2000), cardiovascular prevention (Nishtar et al., 2004), seatbelt use (A. F. Williams, Wells, & Reinfurt, 2002), and vaccinations (Zimicki et al., 2002).
Newspapers have been considered high in accessibility, depth, reach, agenda-setting power, economy, efficiency, and credibility (Salmon & Atkin, 2003). Schooler, Chaffee, Flora, and Roser (1998) found that newspapers were the most powerful media for promoting health information, compared to booklets, television, public service announcements (PSAs), tip sheets, and television programs. However, the traditional print media, particularly print newspapers, have experienced a dramatic decline for the past decade. For example, since 2004, approximately 126 daily newspapers closed their operations, and according to Editor and Publishers’ DataBook listing, as of 2014, most newspapers provide online sites that have experienced growth in online and mobile traffic (Barthel, 2016), showing that people are more likely to read online newspapers than print newspapers.
However, the emergence of social media, such as SNSs, changed the media environment with unprecedented interactive features (Thompson et al., 2003). Although SNSs are not specifically designed for health topics, they become increasingly popular platforms for individuals to share health information with others in their online social networks (CDC, 2011). Boyd and Ellison (2008) define SNSs as Web-based services that allow individuals to (1) construct a public or semipublic profile within a bounded system, (2) articulate a list of other users with whom they share a connection, and (3) view and traverse their list of connections and those made by others within the system. (p. 211)
SNSs not only become popular platforms to find or disseminate health information but also encourage individuals to be advocates for health by liking, sharing, and commenting on a specific health topic in their social networks. In the National Research Corporation’s (2011) survey, one in five Americans were found to use SNSs as a source of health information. In particular, Facebook topped the list of available websites, with 94% of respondents indicating they have used it to gather health-related information. Due to its popularity and potential to reach many people, health communication practitioners are starting to investigate how SNSs like Facebook could be employed for health promotion purposes.
Not like traditional newspapers, SNSs like Facebook have very different interactive features that facilitate different types of interactions among users (Walther, Van Der Heide, Hamel, & Shulman, 2009). Although online newspaper users, with a certain degree of anonymity, can post comments about a topic and others can read overall comments, SNSs like Facebook facilitate more interactions among their “friends” by allowing users to respond to their network of people and assess the sentiment of people they presumably know. This may simulate interpersonal communication/interactions more closely than other types of media, including online newspapers.
Although online newspapers are increasingly adopting interactive features on their sites, SNSs allow users to build stronger social interactions with people in their social networks than other social media platforms like content-sharing sites (CSSs; Kaplan & Haenlein, 2010). For instance, Lee and Chen (2016) found that individuals’ intention to avoid an unhealthy behavior (i.e., smoking) was higher when a health message (i.e., a humorous antitobacco video) was presented on Facebook than when the identical video was featured on YouTube. They took this to suggest that the way health messages are circulated/shared online may influence how people perceive the severity or susceptibility of a given health topic. Therefore, we suspected that different types of media (i.e., online newspapers vs. SNSs) would yield different outcomes and we therefore proposed the following research questions.
Method
Participants
The sample consisted of 142 college students whose age ranged between 20 and 28 years, M = 22.44, SD = 1.22. Most participants were female (79%, n = 112) and in undergraduate courses (100%, n = 142). The biggest ethnic group was Caucasian (59.9%, n = 85), followed by Hispanic (25.4%, n = 36), African American (9.2%, n = 13), and Asian (5.6%, n = 8).
Participants were recruited from undergraduate classes at a large southwestern university and given extra credit for their participation. To sign up for the online experiment, the participants provided their e-mail addresses. The online experiment was designed with Qualtrics, an online survey program. The link to the experimental site was sent to the participants by e-mail. Once clicked, the site opens an informed consent form first, and after reviewing the consent form, the participants were asked to proceed by clicking a button at the bottom of the screen that randomly assigns them to one of four experimental conditions. Among 142 participants, 38 received a URL for a gain-framed message on a fake look-alike Facebook website, 39 received a URL for a loss-framed message on the fake Facebook website, 30 received a URL for a gain-framed message on a look-alike website of the New York Times website, and 35 received a URL for a loss-framed message on the fake site, modeled after the New York Times website.
Stimuli Development
Two different messages of HPV vaccine campaign were developed and posted in different media: 2 (message framing: gain message/loss message) × 2 (media type: SNSs/traditional media). At first, we developed two different HPV vaccine campaign messages using the information from the CDC website. The gain-framed message focused on the health benefits from getting the HPV vaccine, such as prevention of cervical cancer or genital warts. On the other hand, the loss-framed message emphasized the negative consequences of not getting the HPV vaccine, such as getting cervical cancer or genital warts.
Also, these messages were posted on fake Facebook and online newspaper sites to compare the effect of channel difference. In the case of the Facebook condition, the professional webmaster made a fake look-alike CDC Facebook page and posted the information and pictures about HPV vaccine from the actual CDC Web page. In the case of the online newspaper condition, an artificial New York Times Web page was created by using the front page from the actual New York Times website. The messages were presented in a typical PSA format in the online newspaper condition (i.e., Gain and New York Times: http://javacoffee.dothome.co.kr/nytimes/nytimes.php?type=1; Loss and New York Times: http://javacoffee.dothome.co.kr/nytimes/nytimes.php?type=2; Gain and Facebook: http://javacoffee.dothome.co.kr/fb/fb.php?type=1; Loss and Facebook: http://javacoffee.dothome.co.kr/fb/fb.php?type=2). After the participants were done reviewing the site, they were asked to answer a set of questions about the message they saw and their reactions to the site. In the end, the participants were briefed about the study and the nature of the site being created for research purposes.
Independent Variables
Message Framing
“Gain-framed message” (i.e., advantages or benefits) is defined as a message that emphasizes the benefits gained by adopting HPV vaccine. “Loss-framed message” (i.e., risks of not performing the behavior such as cervical cancer and genital warts) is defined by the emphasis on the losses that accompany the nonadoption of HPV vaccine. Schneider et al.’s (2001) measurement was adopted in addition to four new questions that we developed to increase the reliability of the measurement. Gain-framed message was measured with four-items (i.e., In general, the message emphasizes the positive consequences of getting HPV vaccine; In general, the message emphasizes the gain of getting HPV vaccine; In general, the message describes the advantages of getting HPV vaccine; In general, the message shows what I would gain if I get the HPV vaccine; Cronbach’s α = .94; r = .85, of Schneider et al.’s two original items). Loss-framed message was measured with four-items (i.e., In general, the message emphasizes the negative consequences of not getting HPV vaccine; In general, the message emphasizes the loss of not getting HPV vaccine; In general, the message describes the disadvantages of not getting HPV vaccine; In general, the message shows what I would lose if I don’t get the HPV vaccine; Cronbach’s α = .95, of four items; r = .85, of Schneider et al.’s two original items). The average scores of the four-items measured on a 7-point scale were used for each framing.
Media Channels
This study chose two different media channels to post the HPV vaccine campaign message to—Facebook, which is the most popular SNS (National Research Corporation, 2011), and New York Times (total circulation 1.6 million), which is the U.S. daily newspaper with the highest digital circulation (Moses, 2012).
Dependent Variables
HBM Factors
HBM factors were measured using the 14 items developed by Champion (1984). These variables included (1) barriers, including cost, safety, and time consumption (Cronbach’s α = .89); (2) severity, or whether people perceive cervical cancer or genital warts as being serious (Cronbach’s α = .81); and (3) benefits, specifically the effectiveness of the vaccine (Cronbach’s α = .85).
Behavioral Intention
Behavioral intention was operationally defined as the extent to which the participants intend to receive the HPV vaccine. To measure behavioral intention, this study employed and modified items from Fishbein and Ajzen (2010) scale of behavioral intention and a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree). Behavior intention was measured using the four items (i.e., I intend to get the HPV vaccine; I intend to search for the HPV vaccine information; I intend to counsel with a doctor about HPV vaccine; I intend to recommend the HPV vaccine to my friends and family; Cronbach’s α = .88).
Results
Manipulation Check
To check whether participants successfully perceived the gain- and loss-framed messages, this study conducted an independent samples t test. The results indicated that participants who saw the gain-framed message (M = 5.64, SD = 0.86) had a significantly higher level of perception in terms of gain than participants who saw the loss-framed message (M = 4.80, SD = 1.53), t(117) = 4.06, p < .01. Also, the perceptions of participants who saw the loss-framed message (M = 5.44, SD = 1.27) were at a significantly higher level of perception in terms of loss than those of the participants who saw the gain-framed message (M = 3.67, SD = 1.48), t(140) = −7.65, p < .01. As a result, gain- and loss-framed message manipulations were successful (Table 1).
Manipulation Check of Message Framing
Hypothesis 1 predicted that individuals would exhibit a higher level of behavioral intention to use the HPV vaccine when they are exposed to a loss-framed message than a gain-framed message. An independent samples t test showed that participants who viewed the loss-framed message (M = 4.68, SD = 1.45) had higher levels of behavioral intention than those who saw the gain-framed message (M = 4.26, SD = 1.32), t(140) = −1.77, p < .05. Therefore, Hypothesis 1was supported (Table 2).
Effect of Message Framing on Behavioral Intention
We predicted individuals would demonstrate a higher level of perceived severity of HPV (Hypothesis 2) and perceived barriers (Hypothesis 3) when they are exposed to a loss-framed message than when exposed to a gain-framed message. The independent samples t test result showed a significant difference in the level of perceived severity of HPV. Those who read the loss-framed message (M = 5.08, SD = 1.56) exhibited higher levels of perceived severity of HPV than those in the gain-framing condition (M = 4.57, SD = 1.29), t(140) = −2.13, p < .05. Therefore, Hypothesis 2 was supported.
However, no significant difference in perceived barrier was found (gain-framing condition, M = 2.87, SD = 1.03; loss-framing condition, M = 2.78, SD = 1.24), t(140) = 0.47, p = .64. Therefore, Hypothesis 3 was not supported.
Hypothesis 4 anticipated that individuals would exhibit a lower level of perceived benefits of getting the HPV vaccine in the loss-framing condition than those in the gain-framing condition, but no significant difference was found (gain-framing condition, M = 5.23, SD = 0.97; loss-framing condition, M = 5.33, SD = 1.06), t(140) = −.602, p = .55. Hypothesis 4 was not supported.
RQ1 examined the relationship between media channels and behavioral intention and found no statistically significant medium effect between newspaper (M = 4.49, SD = 1.39) and Facebook (M = 4.47, SD = 1.41), t(140) = .08, p = .94.
RQ2 investigated whether media channels affect all three HBM factors. We found a significant effect on perceived severity and perceived barriers. Participants who viewed the online newspaper (M = 5.08, SD = 1.36) perceived higher severity than those who saw the Facebook website (M = 4.63, SD = 1.51), t(140) = 1.84, p < .05. In the case of perceived barriers, participants who viewed the Facebook website (M = 2.65, SD = 1.04) perceived lower barriers than those who saw the newspaper (M = 3.02, SD = 1.24), t(125) = 1.91, p < .05 (Table 3).
Effect of the Media Channels on Health Belief Model Factors
RQ3 was intended to investigate the interaction effect between message framing and media channel using two-way analysis of variance. We found a significant interaction effect between message framing and media channel, F(1, 141) = 4.60, p < .05 (see Figure 1). Participants who viewed the loss-framed message on Facebook mostly exhibited a higher level of behavioral intention to get vaccinated (M = 4.90, SD = 1.33) than those in the gain-framing condition (M = 4.03, SD = 1.37), p < .05. However, the media framing effect was not shown in the online newspaper condition
Interaction Effect Between Message Framing and Media Channel
Discussion
We investigated how message framing (i.e., gain-framed messages vs. loss-framed messages) interacted with media channel (i.e., Facebook vs. online newspaper). We found that the loss-framing increased young adults’ perceived severity of HPV and their intention to get HPV vaccination than gain-framing. However, this framing effect was found only in the Facebook condition, not in the online newspaper condition.
We further found that those who read the message on an online newspaper (i.e., New York Times) exhibited higher levels of perceived severity of HPV than those who read the identical message on Facebook, while those who read the message on Facebook exhibited lower levels of barriers of getting the HPV vaccine than those who read the message on an online newspaper. Considering that individuals’ perceived barriers were found to be the strongest predictor for preventive behavior (Harrison et al., 1992; Janz & Becker, 1984), it appears that loss-framed messages circulated on Facebook somehow reduce individuals’ perceived barriers of getting the HPV vaccine and, in turn, increase their willingness to get vaccinated. This finding begs the question, Why? Perhaps this is due to the perceived peer endorsement/social support one feels on an SNS (i.e., Facebook) over an online newspaper. Although online newspapers provide some interactive features that allow subscribers to share their opinions with other anonymous users, SNSs like Facebook enable users to assess the comments/opinions of their “friends” in social networks. Even if loss-framed messages do not specifically contain any information about potential barriers of getting HPV vaccination, simply because the information was shared among their “friends” (i.e., presented in a form of Facebook posting) might have reduced their social anxiety about getting HPV vaccination or perceived social support for getting HPV vaccination to a certain degree. Further investigations are called for in the near future.
On the other hand, regardless of message framing, circulating promotional HPV vaccination messages in a form of PSA on a national online newspaper seems to increase individuals’ perceived severity of HPV although it does not seem to influence people’s willingness to get vaccination. Health promotion requires setting different goals depending on the stages of diseases, levels of public awareness, availabilities of preventive measures (i.e., vaccine), and so on. Increasing the public’s awareness as well as perceived severity of HPV, although it did not directly lead to the public’s willingness to take an action in the online newspaper condition, is a necessary step for disease prevention in public awareness campaigns.
Several limitations should be addressed. First, although the random assignment supposedly controls the potential influence of unknown third variables in experimental studies, other variables like individuals’ prior involvement/knowledge with HPV could have been measured to understand the impacts that these variables might have had. Second, although the stimuli were designed by a professional webmaster to closely simulate the look and feel of real sites, the fact that they are not the real sites with functional interactive features might have influenced the outcomes of this study. Third, the findings from this study are based on individuals’ self-reports after a brief exposure to the message in the different settings. Fourth, although college age is within the target age range where HPV vaccination is recommended, this age-group is less than optimal in terms of Advisory Committee in Immunization Practices’ recommended age range to initiate the vaccine. However, considering the fact that this is the age-group when individuals can make their own decisions about their health (parental consent not required), continuing efforts should be put into encouraging this group to get vaccinated. This is particularly so if they have not initiated the process. Last, there was a small sample size of male college students used in this study that makes it difficult to generalize the findings to a broader population of young men. Considering the fact that the rate of vaccination among young men is as low as 8%, it is recommended to deliberately include more male samples to test the replicability of the findings in the near future.
Despite these limitations, this research sheds light on how to promote HPV vaccination to the public, particularly young adults, online in terms of message design and media choice. It appears that to increase overall public awareness and perceived severity of HPV, using online newspapers might be a good way to propagate messages themselves. On the other hand, to encourage specific target publics to take a recommended action, it seems logical to find ways to promote and encourage the public to share content (messages that are written in loss-framing) on their SNSs. However, on social media, the ways that content is circulated are not strictly under the control of public health officials/practitioners, and therefore specific strategies to motivate individuals to share health information on HPV and HPV vaccine should be carefully developed to engage the public in the process.
Social media, particularly SNSs, present many opportunities to engage the public in health information dissemination/promotion due to their popularity and capacity (Lee & Chen, 2016). However, the selection of the medium as well as what types of messages are circulated on SNSs calls for consistent monitoring of context in which the messages are being shared. Partnerships with the public in many sectors of society will augment the success of public health promotion. In the era of social media, we no longer need to choose one medium over the others since most of traditional mass media (e.g., TV and newspapers) as well as emerging media (e.g., social media) are readily available online, blurring lines between the notions of traditional and new media at once. Collaborative efforts among the publics, health advocates, and health professionals through any means/media can create new ways to improve the health of our next generation. Scientific inquiries should be continued to inform us of better ways to improve our health through health promotion in the 21st century.