PROMOTING HEALTHY EATING BEHAVIORS USING INFORMATION AND COMMUNICATION TECHNOLOGY (ICT) SUCCESSION THEORY AND MEDIA RICHNESS THEORY DURING COVID-19 PANDEMIC By Mengyan Ma A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Media and Information—Doctor of Philosophy 2021 ABSTRACT PROMOTING HEALTHY EATING BEHAVIORS USING INFORMATION AND COMMUNICATION TECHNOLOGY (ICT) SUCCESSION THEORY AND MEDIA RICHNESS THEORY DURING COVID-19 PANDEMIC By Mengyan Ma Adapting from Stephens and Rain’s (2011) revised model of complementary ICT use, the current study used a 2 (succession: single modality vs. complementary modalities) x 2 (order: text first vs. video first) x 3 (block repetition) mixed factorial design to examine how different modalities (video and text) used on social media influence individuals’ perceived information overload, information effectiveness, attitudes, viral behavioral intention (VBI), and intention to follow suggestions. A total of 399 participants were recruited from Amazon Mechanical Turk (mTurk) and 359 of them (Mage = 42.15) were retained for data analysis. Findings suggested that modality succession did not make any difference on information overload, information effectiveness, attitudes, VBI, nor intentions to follow persuasive message suggestions. Modality order did not influence any outcome variables either. Contrary to Media Richness theory, results suggested that video was not more effective than text in lessening information overload, nor improving information effectiveness, attitude toward message, VBI, or intention to follow suggestions. Information overload negatively predicted attitude toward message and intention to follow suggestions while information effectiveness positively related to attitude. More results, implications, and limitations are discussed in the study. Copyright by MENGYAN MA 2021 ACKNOWLEDGEMENTS It has been a long journey during the past six years. I passed my comprehensive exam at the end of my second year in the doctoral program (2017). Then I gave birth to my first baby. I took a year off and came back to school and started working on my dissertation proposal. My proposal was approved at the end of 2019. I planned to conduct an in-person experiments in 2020 after my second baby was born. However, the COVID-19 pandemic broke out at the beginning of 2020. Both of my kids stayed home. I could not find any time to work on my dissertation. After half a year, my daughter went back to the daycare. I was back to work on my dissertation while my mother-in-law took care of my son at home. However, restricted by the health and safety guidance for research during the COVID-19 pandemic, I could not conduct in-person experiments. I met my advisor once or twice per week to figure out how to transfer the in-person experiments to online experiments. We talked with the Qualtrics specialist, yet the solution were not optimal from a resource and logistics standpoint. It was impossible to run my approved study online. In the meantime, I got a job offer from the University of Wisconsin-Eau Claire with a stard date in August 2021. My advisor and I both knew that I must finish my dissertation soon. We met very frequently, proposed a new study, discussed the new design with my committee, and got approved from the committee. Then I started my pretest study data collection, analyzed pretest data, and followed by main study data collection. My advisor and I worked closely via Zoom the past year. I learned a lot from him, not only about new knowledge or skills but passion and curiosity about research. I could not think of any other words but “thank you” to my wonderful advisor Dr. Saleem Alhabash for his guidance and support. I couldn’t have done my dissertation or my program without him. I would like to thank to iv my committee members: Dr. Esther Thorson, Dr. Jingbo Meng, and Dr. Elizabeth Dorrance Hall, for their suggestions and comments, which strengthened my dissertation. Although I have changed my topic again and again, they never maintained their support of me. They kept offering great suggestions as well as encouragement. Last but not the least, I would like to thank my big family: my parents, my parents-in-law, my husband, and my two little kids. Without their support and understanding, I wouldn’t be able to chase my dream. They are always there behind me. v TABLE OF CONTENTS LIST OF TABLES......................................................................................................................viii LIST OF FIGURES.......................................................................................................................x CHAPTER ONE: INTRODUCTION..........................................................................................1 CHAPTER TWO: LITERATURE REVIEW.............................................................................6 2.1 Definition of Healthy Eating.................................................................................................6 2.2 Promoting Healthy Eating Behaviors through Different Communication Channels............7 2.3 Social Media for Healthy Eating Promotion.......................................................................11 2.3.1 Definition of Social Media..........................................................................................11 2.3.2 Social Media Use ........................................................................................................12 2.3.3 Social Media and Food Marketing..............................................................................14 2.3.4 Using Social Media to Promote Healthy Behaviors ...................................................16 CHAPTER THREE: THEORETICAL APPROACHES ........................................................20 3.1 ICT Succession Theory.......................................................................................................20 3.2 Media Richness Theory ......................................................................................................25 3.2.1 Media Richness Theory and Health Communication .................................................27 3.2.2 Video Message and Text Message..............................................................................28 CHAPTER FOUR: CONCEPTUAL MODEL, RESEARCH QUESTIONS AND HYPOTHESES ............................................................................................................................30 CHAPTER FIVE: METHODS ..................................................................................................36 5.1 Pretest Study .......................................................................................................................36 5.1.1 Measurement ...............................................................................................................36 5.1.2 Pretest Results .............................................................................................................38 5.2 Main Study..........................................................................................................................39 5.2.1 Stimuli Design.............................................................................................................39 5.2.2 Sample and Participants ..............................................................................................40 5.2.3 Procedure.....................................................................................................................40 5.2.4 Measurement ...............................................................................................................42 CHAPTER SIX: RESULTS........................................................................................................44 6.1 Descriptive Statistics...........................................................................................................44 6.2 Scale Reliability ..................................................................................................................45 6.3 Effect of Modality Succession (Condition) ........................................................................46 6.4 Modality Order Effect.........................................................................................................48 vi 6.5 Serial Mediation Analysis: Effect of Modality Succession on Behavioral Intentions........49 6.6 Simple Mediation Analysis: Relationship among Information Overload, Information Effectiveness, and Attitudes......................................................................................................52 6.7 Supplemental Moderated Mediation Analysis: Effect of Information Load and Effectiveness on Behavioral Intentions ....................................................................................53 CHAPTER SEVEN: DISCUSSION...........................................................................................59 7.1 Modality Succession and Succession Order .......................................................................59 7.2 Moderated Mediation Model ..............................................................................................60 7.3 Implications.........................................................................................................................63 7.4 Limitation and Future Research..........................................................................................64 APPENDICES..............................................................................................................................69 Appendix A: List of True/False Recognition Questions for Message Comprehension, Pretest Study .........................................................................................................................................70 Appendix B: Experimental Stimuli in Main Study...................................................................72 BIBLIOGRAPHY......................................................................................................................103 vii LIST OF TABLES Table 1 Eigenvalues, Cronbach's αs and Means (SDs) for Pretest Study......................................75 Table 2 Pretest Results for the Effect of Modality and Message Repetition using Repeated Measures ANCOVA .......................................................................................................77 Table 3 Main Study Design ...........................................................................................................78 Table 4 Stimuli ..............................................................................................................................78 Table 5 Demographics for Main Study Sample ............................................................................79 Table 6 Descriptive Results for Health Status and Snacking Behavior for Main Study ...............81 Table 7 Eigenvalues, Cronbach's αs and Means (SDs) for Main Study ........................................82 Table 8 Repeated Measures ANCOVA Results for the Effect of Modality Succession on Information Overload ......................................................................................................85 Table 9 Repeated Measures ANCOVA Results for the Effect of Modality Succession on Information Effectiveness ...............................................................................................86 Table 10 Repeated Measures ANCOVA Results for the Effect of Modality Succession on Attitude toward Message.................................................................................................87 Table 11 Repeated Measures ANCOVA Results for the Effect of Modality Succession on VBI 88 Table 12 Repeated Measures ANCOVA Results for the Effect of Modality Succession on Intention to Follow Suggestions......................................................................................89 Table 13 Bivariate Correlation Coefficients among Variables......................................................90 Table 14 Repeated Measures ANCOVA Results for the Effect of Modality Order on Information Overload ..........................................................................................................................92 Table 15 Repeated Measures ANCOVA Results for the Effect of Modality Order on Information Effectiveness ...................................................................................................................93 Table 16 Repeated Measures ANCOVA Results for the Effect of Modality Order on Attitude toward Message...............................................................................................................94 Table 17 Repeated Measures ANCOVA Results for the Effect of Modality Order on VBI.........95 viii Table 18 Repeated Measures ANCOVA Results for the Effect of Modality Order on Intention to Follow Suggestions .........................................................................................................96 Table 19 Serial Mediation Analysis for the Effect of Modality Succession on Intentions to Follow Suggestions (Behavioral Intentions) Mediated by Information Load, Information Effectiveness, Attitudes, and VBI ...................................................................................97 Table 20 Indirect Effects for Serial Mediation Analysis for the Effect of Modality Succession on Intentions to Follow Suggestions (Behavioral Intentions) Mediated by Information Load, Information Effectiveness, Attitudes, and VBI.....................................................99 Table 21 Simple Mediation Analysis for the Effect of Information Load on Attitude toward the Message, Mediated by Information Effectiveness ........................................................100 Table 22 Moderated Mediation Analysis for the Effect Perceived Information Load on Intentions to Follow Suggestions, Moderated by Information Effectiveness and Mediated by Attitudes and VBI..........................................................................................................101 Table 23 Indirect Effects of Moderated Mediation Analysis for the Effect Perceived Information Load on Intentions to Follow Suggestions, Moderated by Information Effectiveness and Mediated by Attitudes and VBI ....................................................................................102 ix LIST OF FIGURES Figure 1 Stephens and Rains’ Revised Model of Complementary ICT Use (2011) .....................22 Figure 2 Proposed Theoretical Model ...........................................................................................31 Figure 3 Interaction of Attitude and Information Effectiveness on VBI.......................................55 Figure 4 Interaction of Information Overload and Information Effectiveness on Intention to Follow Suggestions .........................................................................................................56 Figure 5 Interaction of Attitude and Information Effectiveness on Intention to Follow Suggestions......................................................................................................................57 Figure 6 Moderated Mediation Model...........................................................................................58 x CHAPTER ONE: INTRODUCTION Obesity is prevalent among adults, adolescents, and children in the United States (U.S.). Forty percent of U.S. young adults (aged 20-39), 44.8% of middle-aged adults (aged 40-59), 42.8% of older adults (aged 60 and over) are obese (Hales et al., 2020). Though overall prevalence of obesity was similar for men and women, severe obesity was more prevalent among women compared to men. Another report shows that obesity affected about 14.4 million children and adolescents from 2017 to 2018 (Fryar et al., 2020). More notably, 21.2% of adolescents (aged 12–19), 20.3% of youth (aged 6–11), and 13.4% of children (aged 2–5) are classified as obese by the Centers for Disease Control and Prevention (CDC) (Fryar et al., 2020). Over the past two decades, the percentage of obesity increased from 30.5% to 42.4%, while percentage of severe obesity has risen from 4.7% to 9.2% (Hales et al., 2020). The CDC (2010) estimates that by 2050, one in three adults in the United States will be obese. Obesity is associated with heightened risks for heart disease, Type 2 diabetes, stroke, osteoarthritis, several types of cancer, including cancer of the breast, colon, kidney, and pancreas, as well as multiple myeloma and Hodgkin’s lymphoma (U.S. Department of Health and Human Services, 2013). Given the prevalence of obesity, coupled with increasing health risks, it is critical to identify ways to prevent and control weight gain. Past research showed that good nutrition, comprised of a balanced diet of nutritionally rich foods, is one of the key ways to reduce obesity and prevent its harmful risks (Lent et al., 2012; Millimet et al., 2010; Watt et al., 2013). Poor eating habits, such as lower consumption of vegetables and fruit as well as overconsumption energy-dense, nutrient-poor (EDNP) foods such as junk food and sweetened-beverages are considered leading causes of obesity (Chang & Nayga, 2009, 2010; Henderson et al., 2009). Scholars emphasize that consuming nutritional 1 foods and limiting intake of fat and sugar to keep a balanced diet are critical to reduce risks of chronic diseases (e.g., cardiovascular, obesity). Healthy eating is one of the potential solutions to the obesity problem (Lichtenstein et al., 2006). Therefore, encouraging people to live healthily by motivating them to purchase and consume more nutrient-dense foods and fewer EDNP foods is a critical public health endeavor. The risks of obesity have further intensified during the past year of the COVID-19 global pandemic. The World Health Organization (2021) indicated that healthy eating behavior is extremely important during the COVID-19 pandemic, considering the importance of food and drink choice and intake in enhancing the body’s mechanism to prevent, fight, and recover from infections. Although no foods or dietary supplements can directly prevent or cure the COVID- 19, healthy eating is crucial to improving one’s immune systems. Several studies have suggested that good nutrition can reduce the likelihood of developing chronic diseases, including obesity, heart disease, diabetes, and some types of cancer. Data from the COVID-19 Symptom Study (2020) indicated that the COVID-19 lockdown in 2020 increased snacking behaviors in the United States. More specifically, a large sample of respondents (N = 97,000) completed a nutrition survey through the COVID Symptom Study app. About one-third of the sample reported “an increase in snacking with an average of seven pounds of weight gained from March to June 2020” (“Lockdown eating insights: Has the pandemic turned us into a nation of snackers?”, 2020). Ammar and colleagues (2020) conducted an international online survey across Europe, North Africa, Western Asia, and North America in April 2020. Results showed that home confinement predicted unhealthy eating behaviors including increased unhealthy food consumption, binge eating, snacking between meals, and having additional meals per day (Ammar et al., 2020; Sidor & Rzymski, 2020). As the pandemic has shaped, changed, and 2 reshaped people’s behaviors, including eating and snacking habits and behaviors, it is important to find ways to enhance people’s awareness of and promote the enactment of healthy eating behaviors. Therefore, the current study focuses on investigating effects of health message modalities on attitude and behavioral intention toward healthy eating during the COVID-19 pandemic. As health-related messages are increasingly disseminated through social media platforms and other digital technologies (e.g., Boulos et al., 2011; Divecha et al., 2012; Donelle & Booth, 2012; Guse et al., 2012; Evers et al., 2013; Pagoto et al., 2014; Weymann et al., 2014), the current study does not focus on examining message-related factors (e.g., appeals), sources, and receivers of the messages. Instead, this study investigates channel-related factors through the application and extension of information communication technology (ICT) succession and media richness theories by manipulating the modality of repetitious health-related communication attempts geared toward changing attitudes and behavioral intentions. There are numerous ways to motivate healthy eating and ultimately address the public health concerns associated with obesity in the United States including health communication (see e.g., Public Health Association’s the Extras advertising campaign) and public policy (e.g., Let’s Move). Schiavo (2013) defined health communication as “a multifaceted and multidisciplinary approach to reach different audiences and share health-related information with the goal of influencing, engaging, and supporting individuals, communities, health professionals, special groups, policymakers and the public to champion, introduce, adopt, or sustain a behavior, practice, or policy that will ultimately improve health outcomes” (p.7). For example, health professionals and scholars disseminate health information to influence people’s individual health choices through public health campaigns, health education, and health provider-patient communication, among others. As they have been shifting the nature and speed of 3 communication between individuals and health care organizations, social media play a potentially important role in health promotion and education. Korda and Itani (2013) highlighted the necessities of incorporating theoretical approaches and outcome analysis into social media applications for health promotion. With that in mind, the current study aims to test the ICT Succession theory (Stephens, 2007) and Media Richness theory (Daft & Lengel, 1986) in the context of using social media to promote healthy eating behaviors. More specifically, this study aims to investigate if complementary modality use, specifically the combination of video and text message, enhances the effectiveness of healthy eating promotion campaigns, compared to repetitious presentation of the same message in the same modality (video-video and text-text). In their revised ICT succession theoretical model, Stephens and Rains (2011) found that complementary ICT use (i.e., sequential use of different channels to deliver a persuasive message) decreases information overload, in turn increases information effectiveness, which then positively predicts message-congruent attitude and behavioral intentions. To summarize, given the prevalence of obesity in the United States, health professionals have used a myriad of tactics to promote healthy eating behaviors, especially through social media. However, limited research has incorporated theoretical approaches for social media applications in the health communication. Therefore, I adapt Stephens and Rain’s (2011) ICT Succession and Media Richness theories to examine how different modalities (video and text) used on social media influence individuals’ attitudes and, in turn, intentions to eat healthy. With increasing emphasis on healthy eating behaviors to prevent obesity and diabetes, the current study underscores the importance of repetitive messages integrated with different modalities in motivating and maintaining behavior change. 4 This study is organized in seven chapters. Following this introductory chapter, Chapter Two reviews relevant literature regarding different communication approaches to promotion of healthy eating attitude and behaviors. I also review recent studies focused on social media use and social media advertising spending by the food industries. Given that people are highly affected by social media food marketing, the current study emphasizes the importance to fight against unhealthy food advertising by deciphering message strategies to promote healthy eating behaviors on social media. In Chapter Three, I describe the theoretical models informing this study. Chapter Four summarizes the study’s hypotheses and research questions. In Chapter Five, I describe the study design, participant recruitment criteria, study procedure, and measurement. Chapter Six presents the study’s results and findings. Lastly, in Chapter Seven, I discuss the study’s major findings, limitations, and future research directions. 5 CHAPTER TWO: LITERATURE REVIEW In this study, I draw from multiple bodies of literature to inform the current study on healthy eating promotion. Specifically, I first define healthy eating. Next, I identify a variety of communication channels, including mass communication and interpersonal communication, used to promote healthy behaviors. Furthermore, I review previous studies on social media and discussed the necessity to investigate social media’s role in healthy eating promotion. 2.1 Definition of Healthy Eating Researchers defined healthy eating behavior as choosing a diet low in fat and high in fiber, fruit, and vegetables (Conner & Norman, 2005). Following a balanced diet and refraining from the consumption of processed foods are considered important elements of healthy eating (Lappalainen et al., 1998). In Povey and colleagues’ (1998) study, participants defined healthy eating behavior as eating fresh foods, food containing vitamins, food containing minerals, natural foods, a balanced diet, varied diet, and vegetables and fruit. Falk and colleagues (2001) found that participants perceived healthy eating as eating low-fat, natural, balanced, nutritional foods to prevent disease, manage an existing disease, and control weight. According to college students, additional definitions of healthy eating behaviors include the consumption of organic food and eating with others (House et al., 2006). In the study of House et al. (2006), one participant said that examples of healthy foods were broccoli because it prevents cancer and fish because of its Omega-3 fatty acids. Studies show that red meat consumption associates with cancers such as colorectal, esophagus, and lung (Marmot et al., 2007), obesity (Wang & Beydoun, 2009), and Type 2 diabetes (Aune et al., 2009), while no or less red meat consumption was found to be associated with lower risks of cancer, diabetes, obesity, cardiovascular disease, cholesterol, and high blood pressure (Craig, 2009; Key et al., 2006; Tilman & Clark, 2014). Combining the 6 above findings with the “healthy eating plate” guidance provided by Harvard University (12 for 2012: Twelve tips for healthier eating, 2012), I define healthy eating as (1) increasing intake of vitamin C, whole grains, and protein; (2) limiting consumption of sugary drinks, fatty foods, salt, and red meat; and (3) avoiding overeating. 2.2 Promoting Healthy Eating Behaviors through Different Communication Channels Schiavo (2013) defined communication as (1) an exchange of information between individuals by speaking, writing, or using a common system of signs and behaviors; (2) spoken or written messages; (3) an act of communicating; (4) a sense of mutual understanding and sympathy; and (5) a mean of access or communication. Communication channels include mass media communication and interpersonal communication, among others (Schiavo, 2013). Reardon and Rogers (1988) contrasted mass and interpersonal communication channels, where they defined mass media communication as mediated messages affecting a large audience, and interpersonal communication as any communication activity encompassing face-to-face (FtF) communication affecting one or a few. Pearce (2009) argued that, in contrast to interpersonal communication, mass communication allows a single source (e.g., a person, group of people, organization) to transmit information through some type of medium (e.g., newspaper, magazine, book, radio, television, film, internet) to many anonymous receivers (Pearce, 2009). In contrast, interpersonal communication is defined as “the production and processing of verbal and nonverbal messages between two or a few persons” (Braithwaite & Schrodt, 2014, p.7). Interpersonal communication comes in two channels: (1) face-to-face (FtF) communication, and (2) technology-mediated communication including mobile phone, short text messages, telephone, and the internet’s interactive services (Berger, 2005). 7 The proliferation of digital technologies, social media, and mobile devices have blurred the lines among different communication channels. Computer-mediated communication (CMC), defined as any human communication that happens through two or more electronic devices (McQuail, 2010), affords the transmission of communication messages that transcend the distinction between mass communication and interpersonal communication, as both types of communication could happen through CMC channels. Recently, health professionals and scholars have been investigating the effects of different communication channels on health promotion. Next, I review health promotion across different channels. There are numerous examples of successful mass communication campaigns aiming to promote healthy eating attitudes and behaviors. A mass media campaign in New Orleans targeting African American women aged 18-49 years old successfully promoted positive attitudes toward fruit and vegetable consumption (Beaudoin et al., 2007). The “1% or Less” campaign in Wheeling, West Virginia increased the sale of low-fat milk and encouraged high-fat milk drinkers to switch to low-fat milk (Reger et al., 1999). In the early to mid-1800s, Alcott and Graham launched a healthy eating campaign using mass media options of their time such as magazines and books (Noar, 2006). Despite the considerable success of mass media campaigns in promoting healthy eating behaviors, these effects are often investigated in a cross-sectional manner; therefore, it is unclear whether these campaigns have enduring effects on promoting healthy eating behaviors. Within the CMC context, health communication experts have embarked on leveraging the varied affordances to ICTs to disseminate information about healthy eating to different populations. For example, a number of investigations assess the effectiveness of using smartphone applications imbedded with health messages to promote a large population’s healthy 8 behaviors (Boschen & Casey, 2008; Gerber et al., 2009; Ly et al., 2012), thus harnessing the greater accessibility and relatively inexpensive implementation costs of mobile technologies (Smith, 2013). Specifically, health promotion efforts mediated by digital technologies such as computers and smartphones is referred to as e-health communication (Neuhauser & Kreps, 2003). For example, in one study, Gilliland et al. (2015) developed a smartphone application, called “SmartAPPetite”, to send participants daily messages about healthy eating, healthy recipes, and information about local food vendors. Based on these efforts, participants reported an increase in healthy food consumption and a decrease in unhealthy food consumption. Nevertheless, Gilliland and colleagues (2015) discussed limitations of their studies that they were unable to ensure continued engagement in the program. The examples presented here point to important aspects of unidimensional applications of digital technologies to disseminate information regarding health eating. However, such interventions may be limited in their ability to foster continuous engagement with health-related messages (Gilliland et al., 2015; Mendelsohn, 1973). Such limitations point to the lack of comprehensive applications of digital technologies with all their affordances to encompass mass-mediated and interpersonal communication capabilities to motivate greater awareness, attitude change, and adherence to health-related messages. In other words, such interventions are limited in that they do not amplify the integrative role of interpersonal communication that could potentially enhance the effectiveness of digitally mediated health communication related to food-related attitudes and behaviors. Indeed, interpersonal communication in the context of family relationships influences an individual’s health choices as well. Past research found that positive interpersonal communication in families can improve individual’s weight management efforts (e.g., exercise, 9 eating healthy) (Dailey et al., 2011; Dailey et al., 2010) and diabetic patients’ glucose control (Parchman et al., 2009). Therefore, another way to promote healthy eating behaviors and reduce overconsumption of EDNP foods is through interpersonal communication within the family context. Research has found other benefits of interpersonal communication on health promotion and the key role it plays alongside mass communication. Scholars observed that many mass communication campaigns facilitated interpersonal communication around health topics, which in turn successfully influenced behavioral outcomes (Backer et al., 1992; Noar, 2006; Rogers & Storey, 1987). In the realm of interpersonal communication, scholars investigated parent-child communication about healthy eating behaviors. Parent-child communication has been found as primary mean through which children’s eating habits are affected (Flora & Schooler, 1995). Parents influence children’s eating habits through direct communication such as sharing nutritional knowledge and health concern (Davison & Birch, 2001). Additionally, Riesch and colleagues (2006) indicated that improving parent-child communication processes could decrease individual risky behaviors, and facilitate conversation about factors causing health-risk behaviors, for example unhealthy eating behaviors (Riesch et al., 2006). Miller-Day (2006) suggested that family communication patterns could strengthens children’s susceptibility to commercial messages. According to priming theory, children who are highly exposed to unhealthy food commercials are more likely to have preference of unhealthy foods (Harris et al., 2009). Parents’ active mediation (i.e., communication) could influence children’s interpretation of advertising messages (Austin, 1993). Specifically, parents decipher the purpose of advertising messages that helps children critically understand the message. In other words, parent-child communication lessens unhealthy food advertising effects on children. 10 The penetration of new ICTs such as email and social media has been enriching patterns of interpersonal communication (Ackerson & Viswanath, 2009). Social media has been shifting the nature and speed of communication between individuals and health care organizations. For example, people search or scan through health-related Facebook posts published by individuals or organizations, and they engage with the posts by liking, commenting, sharing on their own page, or sharing through private methods such as Messenger and emails. In that light, health- related information has been easily, widely, rapidly, and continuously spreading around a large population via social media. The following section details the necessity to understand the role of social media for health promotion. 2.3 Social Media for Healthy Eating Promotion 2.3.1 Definition of Social Media Mangold and Faulds (2009) defined social media as “a wide range of online, word-of- mouth forums including blogs, company sponsored discussion boards and chat rooms, consumer- to-consumer e-mail, consumer product or service ratings websites and forums, Internet discussion boards and forums, [and] microblogs” (p.358). Lariscy and colleagues (2009) defined social media as “online practices that utilize technology and enable people to share content, opinions, experiences, insights, and media themselves.” (p. 314) Carr and Hayes (2015) defined social media as “Internet-based channels that allow users to opportunistically interact and selectively self-present, either in real-time or asynchronously, with both broad and narrow audiences who derive value from user-generated content and the perception of interaction with others” (p.8). Freberg (2016) suggested that “social media can provide a personalized, online networked hub of information, dialogue, and relationship management. These new 11 communication technology tools allow individual users and organizations to engage with, reach, persuade, and target key audiences more effectively across multiple platforms” (p. 773). Kleinberg (2008) discussed that social media share some common features with traditional communication platforms while maintaining unique aspects. “Social media allows users to participate to an extent not seen previously in traditional media” (Freberg, 2018, p. 7). Aside from establishing and maintaining relationship with others, people use social media to generate their own content and interact with others’ generated content in the online community (Waters et al., 2009). Empirical evidence has provided initial support for the notion that people use social media for specific reasons such as social interaction, entertainment, passing time, information sharing, self-expression, self-documentation, medium appeal and convenience (Alhabash & Ma, 2017). 2.3.2 Social Media Use In 2018, the average number of social media accounts per Internet user was 8.5 (Statista, 2019). In 2019, U.S. adults spent an average of 47 daily minutes on social networking (Statista, 2021a). Mobile social network usage is estimated to increase to 53 minutes per day in 2022. In the United States, 230 million Americans are active social media users while 225 million are active mobile social media users (Statista, 2020). According to Statista (2020), 90% of U.S. young adults (age 18-29) used social networks in 2019. More specifically, 82% of U.S. young adults have Facebook accounts; 89% of U.S. adults age 30-33 have Facebook accounts; 89% of U.S. adults age 45-54 have Facebook accounts; 84% of U.S. elders (age 55-64) have Facebook accounts. Social media are also widely used by U.S. adolescents (Anderson & Jiang, 2018). More specifically, over 8 in 10 U.S. adolescents use YouTube, 72% use Instagram, 69% use Snapchat, 51% use Facebook, and 32% use Twitter. 12 Users share meal and dietary experiences on social media (Mejova et al., 2015). For example, social media users post photos of foods and restaurants on Instagram. A survey of U.S. consumers (N = 3,008) about which social media sites they use for researching new food products, recipes, and nutritional information (Wunsch, 2020), showed that 38% of respondents indicated that they mainly use Facebook for researching food, followed by Pinterest (38%), YouTube (18%), Instagram (12%), Twitter (8%), and other (6%). Kunst (2020) found that 32% of 4,184 American users expressed their opinions about food and drinks on the Internet in the past four weeks such as liking a food/drink-related article on social media, whereas 17% expressed opinions about the health and medicine (Kunst, 2020). The notion that individuals who disclose their own health-related information and engage in healthy eating related posts on social media reported a higher intention to eat healthy has been reflected in different strands of literature (Krishnan & Zhou, 2019; Watanabe-Ito et al., 2020), suggesting that social media work as a self-regulatory tool. More notably, posting healthy eating behaviors on social media discloses an individual’s offline eating behaviors to his/her followers, which reinforces their adherence to healthy behaviors offline (Krishna & Zhou, 2019). Empirical evidence has provided support for the notion that people intend to build a positive image of themselves in front of others, which indirectly influences their own healthy behaviors (Krishnan & Zhou, 2019; Thøgersen-Ntoumani & Ntoumanis, 2007). Due to the fact of peer education and peer support, children developed a greater intention to eat healthy after interactions with others on social media, suggesting that social media hold promise for the future of health promotion (Krishnan & Zhou, 2019; Watanabe-Ito et al., 2020). In addition to individuals’ personal use of social media, it is important to examine how advertisers are capitalizing on social media to promote purchase and consumption of unhealthy 13 food and drink products (see a social media campaign “Fanta: It’s an orange thing”). Health communicators should not only focus on their own message design but understand who their competitors are, as well as how to compete with advertising messages that cause unhealthy eating (for example repetitive advertising exposure), and how can health communicators garner target audience’s attention among a great number of message producers including unhealthy food advertisers. With that in mind, it is important to examine the ways in which food and beverage advertisers are using social media for marketing and advertising, which is expanded in the following section. 2.3.3 Social Media and Food Marketing Advertising on social media has grown exponentially over the past few years, and globally. Social media ad spending is expected to rise from $97.66 billion in 2020 to $138.41 billion in 2025 (Statista, 2021b). In the United States, social media ad spending is predicted to increase from $39.68 billion in 2020 to $55.32 billion in 2025 (Statista, 2021b). Statista (2021b) also shows that most social media ad spending is generated in the United States with a market volume of $44.61 billion in 2021, followed by China ($28.33 billion). These statistics show that advertisers and marketers heavily invest in social media advertising. There are several potential explanations for this trend. First, social media advertisements, along with organic presence (e.g., managing a company’s Facebook page), are cheaper than advertising on traditional channels (e.g., TV, newspapers, magazines). Second, social media advertising reaches larger audiences and could lead to more lasting impressions of ads, brands, or products because of its frequent exposure. Finally, sociotechnical systems like Facebook and Twitter provide more precise measures of advertising effectiveness because of their use of census data rather than sampled data. 14 People are constantly bombarded with mediated messages promoting both healthy and unhealthy food items. From television programs, cooking shows, billboard advertisements, junk food logos on the highway, food brand ads on social media, to user–generated posts of food items; people are constantly exposed to messages that attempt to persuade them to purchase products. Scholars have investigated media effects on eating habits for decades. Dixon et al. (2007) found that healthy food advertisements can promote positive attitudes and beliefs of healthy foods, while heavier usage and higher frequency of commercial TV viewing were associated with more positive attitude toward junk food. Brown and Witherspoon (2002) concluded that the effects of media on eating habits can sometimes be subtle and other times quite powerful, particularly they found that media effects heavily contribute to unhealthy eating behaviors. Harris (2009) found that food advertising has priming effects triggering people’s automatic snacking behavior. More recent literature also encompasses empirical evidence on how unhealthy food ads leads to unhealthy eating behaviors (Cairns et al., 2009; Norman et al., 2016; Sadeghirad et al., 2016). Another study illustrated that social media influencer promotion of unhealthy foods immediately affects children’s unhealthy food intake, whereas promotion of healthy food has no effect (Coates et al., 2019). Two theories can be used to explain the media effects on eating behaviors: cultivation theory and social learning theory. According to cultivation theory, repetitive exposure to consistent media portrayals and themes influences people’s perceptions in the direction of the media portrayals (Gerbner et al., 1994). Additionally, scholars have used social learning theory to explain people’s reactions to a broad category of advertised brands. Social learning theory proposes that modeled behaviors (e.g., eating behavior portrayed in advertisement) will motivate similar behavior among audiences (Bandura, 1986, 2001; Buijzen et al., 2008). 15 Despite the fact that much of the research on food advertising centers around the promotion of unhealthy food items (i.e., junk food) (Harris et al., 2009; Kelly et al., 2010; Lewis & Hill, 1998), limited research on food advertising deals with populations other than children. Given the fact that adults age 30-54 constitute the group with the highest adoption of social media and social networking sites (SNSs), messages placed by social media food advertisers, even though target younger age groups, might still reach adults and thus influence their eating and buying behaviors (Statista, 2020). Therefore, it is worthwhile to uncover the ways in which health communicators can disseminate messages to persuade consumers of the importance and implications of healthy eating behaviors among adults. 2.3.4 Using Social Media to Promote Healthy Behaviors Health communication scholars have investigated how websites and social media sites operate in producing and circulating information to promote healthy behaviors (e.g., Boulos et al., 2011; Divecha et al., 2012; Donelle & Booth, 2012; Guse et al., 2012; Evers et al., 2013; Pagoto et al., 2014; Weymann et al., 2014). After apprehending the great reach and influence of these digital technologies, health professionals used social media to (a) provide health information, (b) provide answers to medical questions, (c) provide online consultations, (d) collect patient experiences and opinions, and (e) reduce stigma (e.g., Crutzen & De Nooijer, 2011; Kratzke & Cox, 2012; Buhi et al., 2013; Chou et al., 2013; Korda & Itani, 2013; Moorhead et al., 2013; Epton et al., 2014; Smith et al., 2014). Moorhead and colleagues (2013) have referred social media use for health communication as “the general public, patients, and health professionals communicating about health issues using social media platforms such as Facebook and Twitter” (p. 2). In their study, Moorhead et al. (2013) indicated that there was a dearth of information summarizing the uses, 16 benefits, and limitations of social media for health communication from original research in that they conducted a systematic review of 98 primary studies that published between January 2002 and February 2012. Seven uses of social media for health communication were identified in their study: “(1) provide health information on a range of conditions, (2) provide answers to medical questions, (3) facilitate dialogue between patients to patients, and patients and health professionals, (4) collect data on patient experiences and opinions, (5) health intervention, health promotion and health education, (6) reduce stigma, and (7) provide online consultations” (Moorhead et al., 2013, p. 9). They have also discussed six benefits of using social media for health communication, for instance social media increases interactions with others, provides more available, shared, and tailored information, and increases accessibility (Chou et al., 2009; Kontos et al., 2010; Kukreja et al., 2011; Scanfeld et al., 2010). Abundant studies on social media documented that users, patients, and health professionals are also using social media to communicate about health-related information (e.g., Dawson, 2010; Fox, 2011; Giustini, 2006; Green & Hope, 2010; Heidelberger, 2011; Levac & O'Sullivan, 2010; McNab, 2009; Scanfeld et al., 2010). For instance, the World Health Organization used Twitter during the influenza A (H1N1) pandemic (McNab, 2009), and health professionals obtained information to inform their clinical practice (Giustini, 2006; Green & Hope, 2010). Freeland-Graves and Nitzke (2013) suggested that food and nutrition practitioners use a variety of accessible communication technologies, including social media, to communicate with professional colleagues and the public. Even though it is a herculean task to fight against unhealthy food and drink marketing on social media when public health budgets pale compared to global corporations, Dunlop and colleagues (2016) indicated that public health organizations and social marketing campaigns 17 have put great efforts in promoting healthy eating behaviors through social media. Studies on small-scale healthy eating promotion campaigns which integrated social media platforms such as Facebook and Twitter suggested a comparatively high engagement rate on social media of people who were exposed to these campaigns (George et al., 2016; Tobey & Manore, 2014). Targeting U.S. college students, Napolitano et al. (2013) designed weight-loss messages delivered via Facebook and text messaging, suggesting preliminary efficacy of weight loss and message acceptance among participants. A study of 33 Caucasian women aged between 22-73 indicated that providing useful recipe ideas, improved lifestyle, were a credible source of information and allowed intention with a dietitian are considered as main advantages of using healthy eating blogs written by a dietitian (Bissonnette-Maheux et al., 2015). As suggested by Chan and colleagues (2020), during the COVID‐19 pandemic, “social media has the potential, if responsibly and appropriately used, to provide rapid and effective dissemination routes for key information” (p. 1582). To conclude, I am certain that social media is a terrific fit with the consideration that the final goal of my study is to uncover effective communication strategies to promote healthy eating behaviors during the pandemic. Despite the evidence that social media plays a potentially important role in health promotion and education, Korda and Itani (2013) argued that social media, like traditional media, requires careful applications. Additionally, they highlighted the necessities of evaluating effectiveness across various social media and incorporating theoretical approaches and outcome analysis into the social media application for health promotion. With that in mind, my study aims to test ICT Succession theory and Media Richness theory in the context of using social media to promote healthy eating behaviors. 18 In this chapter, I first reviewed other scholars’ definitions of healthy eating and then provided my definition of healthy eating. Next, I discussed studies on mass communication campaigns and smartphone applications that promote healthy eating behavior with their limitations as well as the penetration of social media, which informed me the necessity to understand and promote healthy eating by social media. In the next chapter, I will review the ICT Succession theory and Media Richness theory applied in the current study. 19 CHAPTER THREE: THEORETICAL APPROACHES In this chapter, I, first, review two core theoretical approaches applicable to in the current study: ICT Succession theory and Media Richness theory. Next, I propose a conceptual model adapted from the “revised model of combinatorial ICT use”. 3.1 ICT Succession Theory Message repetition is important for effective persuasion (Cacioppo & Petty, 1979; Flora et al., 1997; Harrison, 1977; Hornik, 2002; Rogers & Storey, 1987; Zajonc & Rajecki, 1969). Sequential ICT use refers to repeatedly communicating a message through the same ICT or complemental ICTs (Stephens, 2007; Stephens et al., 2008; Westerman et al., 2008). Repeating the same ICT can increase message fatigue, whereby recipients become tired of the message (Stephens & Rains, 2011). Message fatigue in mass media campaigns has been found to have adverse effects on health promotion (Ray et al., 1971). Over the past two decades, scholars have developed theories to explain how people choose a specific ICT in workplace (e.g., Daft & Lengel, 1986; Fulk et al., 1990). However, Zhu (2019) argued that the concentration on the selection of a single ICT limits our understanding of the combinatorial ICT use in the real workplace. Instead, it is important to capture the complexities of multiple ICTs use from a combinatorial perspective. Zhu (2019) defined combinatorial use of ICT as “the strategic integration of multiple information and communication technologies in processes to achieve communication goals across a variety of organizational settings” (p.624). ICT Succession theory extends combinatorial ICT use by incorporating the nature of complementary modalities of ICTs (Zhu, 2019). More notably, the ICT Succession theory suggests that maximizing complementary modalities of ICTs over time influence persuasion attempt. Stephens (2007) defined complementary ICT use as a specific class of sequential 20 multiple ICT use where the successive ICT is reflected in presentation of the same persuasive message through different communication channels with modality-expanding capabilities (Stephens, 2007). Modality expansion relates to the complementary forms that offering “expanded cues or higher richness, which in turn provide error-reducing redundancy for equivocal and uncertain tasks” (Stephens et al., 2008, p. 197). ICT Succession theory (Stephens, 2007) suggests that the multiple modalities (i.e., visual and audio) present when using complementary ICTs may serve to better reinforce message content. Furthermore, complementary ICTs guarantee the second message appear similar but unique and, thus, help retain audiences’ attention and ensure that they engage both messages. For example, people who receive a text message (text-based information) about a certain topic followed by a phone call (audio information) are more likely to attend to the message than people who receive the message twice but only from a single ICT (either text message or phone call). Stephens and Rains (2011) suggested a revised model of complementary ICT use (see Figure 1), where the complementary ICT use reduces message-induced information overload, thus enhances information effectiveness, which in turn positively affects attitudes and behavioral intentions. Information overload is defined as a situation where people have more information than they can process (Eppler & Mengis, 2004). Information effectiveness refers to the formal and informal sharing of meaningful and timely information between a message sender and a receiver (Sharma & Patterson, 1999). An attitude is “a process of individual consciousness which determines real or possible activity of the individual in the social world” (Strauss, 1945, p. 329). Behavioral intention is defined as “the degree to which a person has formulated conscious plans to perform or not perform some specified future behavior” (Warshaw & Davis, 1985, p. 214). 21 Figure 1 Stephens and Rains’ Revised Model of Complementary ICT Use (2011) Stephens and Rains (2011) recruited 148 undergraduate students from two large southwestern universities. They designed persuasive messages to encourage participants to use career service center at universities. They conducted a 4 (ICT sequence) x 2 (message repetition) between-subjects study1. The four ICT sequence conditions were e-mail then FtF (complementary ICT use), FtF then e-mail (complementary ICT use), e-mail then e-mail (repeated ICT use), and FtF then FtF (repeated ICT use). Results suggested that complementary media effects overrode the simple effects of multiple message exposure. More specifically, participants in the complementary ICT condition (FtF then email and email then FtF) reported lower information overload than participants in the repeated ICT condition (FtF then FtF and email then email). Furthermore, participants in the complementary ICTs condition were more likely to visit Career Services than participants in the repeated ICT condition. However, no significant differences were found between the complementary ICT condition and the repeated ICT condition in terms of attitudes, or information effectiveness. Additionally, there were no 1 More specifically, in the first step: in group 1 to 4, researchers e-mailed participants a message; in group 5 to 8, researchers read a message FtF to participants. Second step: in group 1 and 5, researchers delivered a different message with the same topic through e-mail; in group 2 and 6, researchers delivered a different message with a different topic through e-mail; in group 3 and 7, researchers read a different message with the same topic FtF to participants; and in group 4 and 8, researchers read a different message with a different topic FtF to participants. 22 order effects of the four combinations for attitudes. However, order effects were found in relation to information overload and behavioral intentions. Participants in the email then email condition reported more information overload than participants in the FtF then email condition and participants in the email then FtF condition, while there was no significant difference in information overload between the email then email and FtF then FtF conditions. In terms of the intention to visit Career Services, participants in the email then email condition reported significantly lower intentions than participants in the three other conditions. Results suggested that the revised model of complementary ICT use marginally fitted the data. Although the SRMR met Hu and Bentler’s (1998, 1999) fit criteria, the p value for the model’s chi-square was at .05, and the CFI was slightly below .96. Stephens and Rains (2011) suggested that the results of this revised model should be interpreted tentatively. An in-depth interview study was conducted to understand experienced ICT users’ reasons for discrete, sequential, and follow-up ICT use in the workplace (Stephens et al., 2008). A total of 66 participants (male=71%) from 64 different organizations (e.g., banking, software development, management consulting, fish farming, sales) were recruited from the United States and Norway. Findings suggested three major ICT sequence pairs: Web-email, computer-paper, and FtF-phone. Although information (e.g., prepare for meeting) was the most frequent reason related to most sequence pairs, participants used sequential ICTs primarily for persuasion. For instance, participants mentioned that they felt they could be persuasive when they followed-up a FtF meeting with an e-mail. To the best of my knowledge, even though scholars emphasize the benefits of using multiple media to promote health behaviors (Flora et al., 1997; Hornik, 2002; Rogers & Storey, 1987), ICT Succession theory has not been explicitly tested in the realm of health 23 communication. The theory suggested that the multiple modalities (i.e., visual and audio) present when combining multiple media is more effective in reducing fatigue and, thus, positively influencing persuasive outcomes. It is worthwhile to extend the ICT Succession theoretical framework to the health communication field. The current study extends the ICT succession theory by highlighting the affordances of social media, through the emphasis on an intermediary factor – VBI– that is thought to influence information processing and persuasiveness of a message shared via social media. Alhabash and colleagues (2015, p. 523) defined VBI as the “desire to enact online behaviors that contribute to a message’s virality, such as pressing the like button, sharing the video, and commenting on it.” In a study of civic behavioral intention, VBI expressed toward YouTube videos was the strongest predictor of offline behavioral message adherence intentions (Alhabash et al., 2015). As discussed earlier, people who are willing to engage in healthy eating related posts on social media are more likely to eat healthy (Krishnan & Zhou, 2019), given social media self-regulator role and users intentions to harmonize their online and offline personas. In the original ICT Succession study (Stphens & Rains, 2011), the authors manipulated complementary ICT use by repeating the same message either on the same channel (email vs. face to face) or across two channels (email then FtF, FtF then email). However, such a manipulation is confounded as it does not exhaustively identify whether the effect of ICT succession is because of channel succession or differences in modalities, given that FtF communication provides more nonverbal cues than email communication. Both factors were identified as the umbrella for the integration of channel succession and the unique affordances for each channel. To further explicate the influence of modality, the next section provided a brief review of Media Richness theory. 24 3.2 Media Richness Theory Media Richness theory (Daft & Lengel, 1986), or Information Richness theory (Daft & Lengel, 1983), has been applied in the realm of interpersonal communication and organizational communication. The theory describes the relative efficiency of different communication media for the purpose of reducing equivocality and enhancing information effectiveness (or communication effectiveness). In this case, equivocality refers to “the degree to which a decision-making situation and information related to it are subject to multiple interpretations” (Walther, 2011, p. 448). Meanwhile, information effectiveness is defined as the formal and informal sharing of meaningful and timely information between a message sender and a receiver (Sharma & Patterson, 1999). Daft and Lengel (1984, 1986) investigated the vital role of uncertainty and equivocality reduction in successful information processing in organizations. Uncertainty refers to the lack of information that could be reduced by amount of information, whereas equivocality is the confusion or lack of understanding that could not be reduced by amount of information but could be reduced by the richness of information (Daft et al., 1987). Using email to communicate would not cause uncertainty because it could provide sufficient information, however it would not be helpful for equivocality reduction due to its lack of nonverbal cues (Ishii et al., 2019). Daft and Lengel (1986) explained that Media Richness theory deals with four dimensions/criteria: (1) the number of cues provided by a medium or channel, (2) the immediacy of feedback afforded by the channel, (3) how naturally people communicate through the channel, and (4) how personalized the messages delivered through the channel are, specifically the degree to which a message addresses a specific individual (Walther, 2011). Walther (2011) defined rich media as multimodal or greater bandwidth media which support multiple verbal and nonverbal 25 cue systems. In other words, Daft and Lengel (1986) contended that the using multiple media streams or channels increases individuals’ reception, attention, and comprehension of messages. Based on these criteria, face-to-face communication is deemed the richest channel, followed by video, audio, text, and image. Daft et al. (1987) suggested that for a higher equivocality task, the richest medium (i.e., FtF) would be most effective, conversely, for a lower equivocality task a leaner medium would be most effective. The most important premise from the Media Richness theory is that effective managers should chose the most appropriate medium by matching its richness to the level of equivocality of task (Daft et al., 1987). A total of 132 undergraduate students (male=62%) were recruited for a lab experiment (Dennis & Kinney, 1998). Participants were randomly assigned as dyads, and dyads were randomly assigned into one of four conditions: video-immediate feedback, video-delayed feedback, CMC-immediate feedback, and CMC-delayed feedback. No matter which conditions they were in, they were required to complete one lower equivocality task and one higher equivocality task. Higher equivocality task was a version of the undergraduate admissions task, where participants needed to rank undergraduate program applicants from best to worst and reach a decision for admission. Lower equivocality task was similar to those used on Scholastic Aptitude Tests. Participants first completed both tasks individually and made an individual decision. Then they were introduced to their assigned partner through the medium of their condition and familiarized with each other. Next, they completed the first task as a dyad, completed an individual questionnaire and made another individual decision. Last they completed the second task as a dyad, completed individual questionnaire and made an individual decision. Although participants reported differences in media’s richness as predicted by the Media Richness theory, no significant difference was found between richer and leaner media in 26 their performance on the higher equivocality task. In contradiction to Media Richness theory, results showed that matching media richness to task equivocality did not improve performance, which was consistent with empirical studies of Kinney and Watson (1992) and Valacich et al. (1994). 3.2.1 Media Richness Theory and Health Communication Ishii and colleagues (2019) argued that Media Richness theory was born in organizational setting but continues to inspire research in other fields such as interpersonal relationship development (e.g., Liu & Yang, 2016; Pettergrew & Day, 2015; Sheer, 2010, 2011; Zhou et al., 2016) and online course design (e.g., Balaji & Chakrabarti, 2010; Brinker et al., 2015; Lan et al., 2011), especially during the COVID-19 pandemic (e.g., Fernandez & Shaw, 2020). Media Richness theory has also been applied in health communication from a different angle other than the equivocality (Huo et al., 2018). Huo et al. (2018) found that information richness positively predicted individuals’ understanding of health knowledge on social media. More specifically, people perceived health knowledge with clearer expression and higher content richness (e.g., vivid knowledge representation with pictures, cartoons, and videos; Griffith & Neale, 2001; Lan & Sie, 2010) as higher-quality knowledge, and they put less cognitive efforts with the help of rich knowledge. People who were exposed to a richer health knowledge presentation tended to gain a deeper cognition and understanding of the knowledge than others (Huo et al., 2018). Media Richness theory has been used in a 2 (media richness: high vs. low) x 2 (interactivity: high vs. low) between-subject experiment study of promoting physical activity (Lu et al., 2014). A total of 205 college students from a large east coast university of the United States participated in this study. About half of participants were female and the rest were male. Participants were randomly assigned into four groups. Participants in low richness groups saw 27 stationary images, while participants in high richness groups saw 360-degree video. Participants in low interactivity groups saw a fixed image, whereas participants in high interactivity groups were able to control the visual image. After viewing the website, they completed a survey about their opinions of the website and their behavioral intentions. Overall, researchers found that media richness predicted participants’ intention to visit the fitness center while interactivity predicted their intention to recommend it. 3.2.2 Video Message and Text Message Scholars suggested that disseminating health information through online videos embedded with verbal, vocal, and visual elements (Waters & Jones, 2011) could potentially reach a large segment of the target audience because online videos allow users tagging, commenting, sharing, etc. (Chin et al., 2010; Keelan et al., 2007; Lange, 2007). O’Mara (2013) suggested that social media have created new opportunities for using digital video to promote health behaviors in Australia. Digital videos which were limited to one-way communication channels such as broadcast television are accessible through interactive platforms such as YouTube and Facebook nowadays. Users can easily upload their own videos, as well as view and share others’ videos on these platforms (Burgess & Green, 2009; Department of Health, 2010; O’Mara, 2013). Although it is unclear if text-based social media message is effective in promoting healthy behaviors, a meta-analysis study of 19 interventions suggested that text messaging-based health promotion intervention were generally efficacious (Head et al., 2013). Based on the Media Richness theory and study of Huo et al. (2018), I predicted that video-based social media post, which provided a clearer expression and higher content richness, would be more effective than text-based post. 28 In this chapter, I reviewed two core theories (i.e., ICT Succession theory and Media Richness theory) applied in the current study. In the next chapter, I describe my conceptual model and present the study’s hypotheses and research questions. 29 CHAPTER FOUR: CONCEPTUAL MODEL, RESEARCH QUESTIONS AND HYPOTHESES Regarding the manipulation of ICT succession, I was faced with conceptual and operational challenges. First, per the original premise of ICT succession theory, complementary ICT use is defined and operationalized as the repetition of the same message across channels that vary in affordances and the number of cues (e.g., email vs. FtF). The challenge with such an operationalization entails threats to internal validity should the current study manipulate ICT succession by exposing participants to text-based messages on email vs. a video-based message on a social media platform like Facebook. Restricting text-message exposure to email and video exposure to social media would constitute a confound to the manipulation of the independent variable by manipulating two elements within the same condition – channel and modality. The second challenge lies in the alternative experimental strategy of manipulating message modality (text vs. video) within the same channel, where participants would receive the same message twice either with the same modality (text vs. video) in the same channel or with varied modality (text then video vs. video then text). While this might be a deviation from the original premise of ICT succession, this might provide an opportunity to extend ICT succession to studying within- channel repeated exposure by investigating the effect of media richness on processing of sequential and complementary modality use. In both cases, I argue that the repetition of the same message with varied modalities (richness) would yield the most effective persuasive outcomes compared to the repetition of the same message with the same modality. This theoretical argument warranted the proposal of the following hypotheses and conceptual model (see Figure 2 for the proposed theoretical model). Hence, the independent variable in this study is modality 30 succession rather than complementary ICT use considering that I did not use different ICTs but same or different modalities sequentially. Figure 2 Proposed Theoretical Model It is worthwhile to promote healthy eating behaviors through social media as explained in Chapter Two, however in recent years, active social media users complained that their social media feeds were too overloaded, which discouraged their continuous use of social media (Bontcheva et al, 2013; Fu et al., 2020; Grineva & Grinev, 2012; Rodriguez et al., 2014). As reviewed in Chapter Three, Stephens and Rains (2011) suggested a significant effect of complementary ICT use on information overload and intention. In other words, using complementary modalities rather than a single modality would decrease information overload. Additionally, Media Richness theory suggested that media richness positively predicted intention to visit a fitness center (Lu et al., 2014) and intention to use e-book readers (Lai & Chang, 2011). The richer the media/modality, the greater the behavioral intention. Fishbein and Ajzen developed the Theory of Reasoned Action (TRA), which explains behavior by identifying the primary determinants of behavior and the sources of these determinants, and by organizing the relations between these variables (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1977; Yzer, 2013). The TRA proposes that an individual’s attitude toward a certain behavior and subjective norms predict his/her intention to perform or not perform that 31 behavior (O'keefe, 2002). Ajzen extended TRA by introducing a third element, perceived behavioral control, which also determines behavioral intention, in addition to attitudinal and normative influences (Ajzen, 1985, 1991). Ajzen’s proposed model is called Theory of Planned Behavior (TPB) (1985,1991). The Integrated Behavioral Model (IM) is an updated version of the behavioral intention theory. As recommended by Glanz and colleagues (2015), it includes constructs from TRA/TPB as well as from other influential theories. All three theories discuss behavioral intention predictions. Interventions designed to promote healthy eating have utilized the TRA, TPB and IM (Anderson et al., 2005; Angelopoulos et al., 2009; Beaulieu & Godin, 2012; Gratton et al., 2007; Jemmott III et al., 2011; Karimi-Shahanjarini et al., 2013; Kothe et al., 2012; Kothe et al., 2011; Prelip et al., 2011; Spiegel & Foulk, 2006; Tsorbatzoudis, 2005). TPB has also been applied in a number of studies to understand food-related behaviors, for example fat intake (Armitage & Conner, 1999a, 1999b; Louis et al,, 2007; Paisley et al., 1995; Paisley & Sparks, 1998), milk consumption (Raats et al., 1995), organic vegetable consumption (Sparks & Shepherd, 1992), chip consumption (Towler & Shepherd, 1992), biscuit and wholegrain bread consumption (Sparks et al., 1992), sugar intake (Beale & Manstead, 1991). To promote behavioral intentions related to healthy eating, health professionals should improve target audiences’ attitude toward the behavior. Based on all discussed above, I predicted: H1: Participants exposed to video and text (complementary modalities) will report (a) lower information overload, (b) higher information effectiveness, (c) more favorable attitude toward healthy eating promotion message, (d) higher VBI, and (e) higher behavioral intention to follow suggestions than those exposed to the video-only and text-only conditions, respectively. 32 Stephens and Rains’ (2011) found a significant order effect of the ICT use on information overload regarding to the messages that encourage visiting Career Service and participants’ intention to visit Career Service but no significant order effect on attitude toward the message nor information effectiveness of the message. In my study of promoting healthy eating behaviors, I wondered: RQ1: What, if any, is the effect of modality order in the complementary modality condition on (a) information overload, (b) information effectiveness, (c) attitudes, (d) VBI, and (e) behavioral intentions? Additionally, Stephens and Rains’ (2011) revised model of complementary ICT use suggested a serial mediation from complementary ICT use through information overload, information effectiveness, and attitudes to behavioral intentions. More specifically, complementary ICT use reduced information overload, thus enhanced information effectiveness, which in turn positively affected attitudes and behavioral intentions. In my study, I predicted: H2: Modality succession affects behavioral intentions through the serial mediation of information overload, information effectiveness, attitudes toward message, VBI, and intention to follow suggestions. H3: Information overload will be negatively associated with information effectiveness. H4: Information effectiveness will be positively associated with attitudes toward the message. H5: Information overload will be negatively related to attitudes toward the message. The notion that attitudes positively predicted VBI was supported by a study of 835 young Chinese consumers (Mage= 21.34), where their attitudes toward forwarding the entertaining electronic messages positively predict their intention to share the electronic messages. VBI on 33 social media was found as the strongest predictor of behavioral intentions such as civic behavioral intention (Alhabash et al., 2015) and intention to eat a healthy diet (Krishnan & Zhou, 2019). Therefore, I predicted that: H6: Attitude toward healthy eating promotion message will be positively associated with VBI (i.e., like, share and comment). H7: VBI will be positively associated with intention to follow suggestion from the message. Referring to the Stephens and Rains’ (2011) study, where information effectiveness regarding to the messages about Career Services mediated the relationship between how overloaded they felt with the messages and their attitude toward the Career Services, I predicted that: H8: The relationship between information overload and attitudes toward the message will be mediated by information effectiveness, such that lower information overload would be associated with higher information effectiveness, which would, in turn, lead to more favorable attitudes toward the message. To summarize, I proposed that modality succession would reduce information overload. More specifically, people exposed to messages in two different modalities sequentially (complementary modalities) would report lower information overload than people who were exposed to messages in the same modality sequentially (single modality). I proposed that information overload negatively predicted information effectiveness. The lower the information overload, the higher the information effectiveness. I proposed that information effectiveness positively related to attitude toward the message. People who reported greater information effectiveness regarding to the message had more favorable attitude toward the message. I 34 predicted that attitude toward the message positively influenced VBI, which positively related to behavioral intention to follow suggestions from the message. Furthermore, I predicted that information effectiveness mediated the relationship between information overload and attitude toward message. In the next chapter, I described methods including pretest and main study design, stimuli design, recruitment, procedure, measurement, and results. 35 CHAPTER FIVE: METHODS In this chapter, I describe the experimental method used in this study including pretest and main study design, stimuli design, recruitment, procedure, and measurement. 5.1 Pretest Study The pretest study was designed to select three out of five video/text messages that were most effective and did not significantly different from each other to use in the main study. First, I chose five videos that promote healthy eating behavior during the COVID-19 pandemic from YouTube and had a transcription company (Rev.com) transcribe the videos into texts. Next, I designed five newsletters by tailoring texts to ensure close numbers of words across the five newsletters and adding two screenshots from the videos for each newsletter. Upon the approval from the MSU Institutional Review Board (IRB), I recruited 55 college students from a large midwestern university for an online survey. In the survey, participants were randomly assigned into two groups: video only or text only. More specifically, participants either watched five videos or read five text-based newsletters, and evaluated the videos/newsletters in terms of message comprehension, information overload, information effectiveness, attitude toward the message, intention to follow the message, and intention to eat a healthy diet. 5.1.1 Measurement The following section provide operational definitions of the main variables used in the pretest study. Where noted, reliability metrics from previous studies are included, yet Table 1 provides the list of items for each variable along with validity and reliability indices. Participants answered two target and two foil questions for each video/text message. If they got the correct answer, they got 1 point, otherwise 0 point for that question. For each message participants got 0 to 4 points in total. An example question was “Omega-3 has been 36 found to lower depression.” Questions for all five video/text messages are listed in the Appendix A. Participants’ perception of information overload was assessed using Stephens and Rains’ scale (Cronbach’s α= .83) (2011). More specifically, participants rated their agreement with statements on a seven-point Likert-scale ranging from strongly disagree (1) to strongly agree (7): “The information I received about healthy eating behavior needs too much explanation to be useful;” “The information I received requires me to make too many decisions;” “The information I received has too much information;” “The information I received is more discussion than I wished;” “The information I received is more information than I need;” “The information I received is about the right amount of information I need.” The final item was reverse coded. Participants were asked to evaluate the message they just saw ranging from strongly disagree (1) to strongly agree (7): “This message is effective/detailed/useful/high quality/well supported/weak/uninformative.” (Stephens & Rains, 2011) The last two items were reverse coded. A higher number indicated a greater information effectiveness of the message (Cronbach’s α = .96). Adapted from the attitude scale (Stephens & Rains, 2011), participants rated the following statements on a seven-point Likert-type scale ranging from strongly disagree (1) to strongly agree (7): “This message is helpful to my health;” “This message is a valuable resource;” “This message is important for being healthy;” “This message is a good reference for meal preparation;” “This message has something positive to me.” A higher number represented a more positive attitude toward the message (Cronbach’s α= .78). Adapted from the scale of Conner et al. (2002), participants rated the following statements ranging from strongly disagree (1) to strongly agree (7), definitely do not (1) to 37 definitely do (7), or unlikely (1) to likely (7): “I intend to follow what the message suggests in next seven days;” “I will try to follow what the message suggests in next seven days;” “I want to follow what the message suggests in the next seven days;” “I expect to follow what the message suggests in next seven days;” “How likely is it that you will follow what the message suggests in next seven days?” A higher number indicated a greater intention to follow the message (Cronbach’s α ranged from .89 to .95). Adapted from a previous scale (Conner et al., 2002), participants rated the following statements ranging from strongly disagree (1) to strongly agree (7), definitely do not (1) to definitely do (7), or unlikely (1) to likely (7): “I intend to eat a healthy diet in next seven days;” “I will try to eat a healthy diet in next seven days;” “I want to eat a healthy diet in next seven days;” “I expect to eat a healthy diet in next seven days;” “How likely is it that you will eat a healthy diet in next seven days?” A larger number indicated higher intention to eat a healthy diet (Cronbach’s α ranged from .89 to .95). 5.1.2 Pretest Results Fifty-five participants completed the survey with 28 evaluated the videos and 27 evaluated the newsletters. The average age of participants was 21.02 (SD = 1.80). About two- third of participants were female. Most participants (72.7%) identified as White, followed by Asian (21.8%), Native Hawaiian or Pacific Islander (3.6%), and other (0.1%). Most participants were undergraduate students: freshmen 21.8%, sophomore 27.3%, junior 29.1%, and senior 20.0%. Only 1.8% of participants reported themselves as Master’s student. In the pretest study, I ran repeated ANOVA analysis between the selected three groups to make sure that they do not differ from one another (non-significant interaction effect of repetition and modality) on information overload F(2, 52) = 2.08, ns, information effectiveness 38 F(2, 52) = 2.16, ns, attitude toward the message F(2, 52) = 2.83, ns, intention to follow the suggestions F(2, 52) = 1.38, ns, intention to eat a healthy diet F(2, 52) = 2.075, ns, and message comprehension F(2, 52) = 1.289, ns (see Table 2). 5.2 Main Study This design operationalized complementary ICT use as the variability in modality of the promotional health message within the boundaries of the same channel (sequential exposure). With that in mind, this current study used a 2 (succession: single modality vs. complementary modalities) x 2 (order: text first vs. video first) x 3 (block repetition) mixed factorial design, with modality manipulated between subjects, order nested in the complementary condition, and block repetition was manipulated within-subject. Here, succession was manipulated as the exposure to the same message twice, either as both exposure in text format or video format, or exposure to both text and video versions of the message in a complementary way (with order counterbalanced). Participants were randomly assigned to one of four groups (see Table 3). For all three repetition blocks, participants were exposed to the health messages on Facebook only. As illustrated before, I saw this study as a potential extension of ICT succession that encapsulated the new affordances of digital and social media in relation to disseminating content of varying modalities within the same channel. For analysis purposes, and to test the study’s hypotheses and answer the research questions, where ICT succession modality was tested, I used a three-level factor: text-only, video-only, and text and video (combining the two order factors). 5.2.1 Stimuli Design Three videos/texts were selected from the pretest study. I first looked into Academy of Nutrition and Dietetics’ Facebook account “Eat Right Nutrition” (https://www.facebook.com/EatRightNutrition/) and found its average number of likes and 39 shares. Then I used the website (https://zeoob.com/generate-facebook-status-post/) to design text-based Facebook posts with mimic likes and shares. The text-based Facebook post include a screenshot of the video and transcribed texts. Examples are attached in the Appendix A. To design the Facebook video posts, I used a screenshot of the top banner of the text posts including the account name, publication date and time, and bottom banner of the text posts including the numbers of likes and shares. Next, I built the video posts on Qualtrics by adding the top and bottom banners into a blank box and inserting the videos in between in that videos look like Facebook video posts. Details of stimulus were included in Table 4. 5.2.2 Sample and Participants The sample size of 252 was determined using an a priori power analysis with the following parameters: Effect size (f) = .23, alpha error probability = .04, Power = .95. Taking the response rate into consideration, I oversampled by 30% to 35%. To this end, I recruited 399 participants in total. The study was designed as three conditions (video only, text only, video and text) but four groups. More specifically group 3 and group 4 together were combined into condition 3. Participants who failed the attention check questions or did not complete the survey were removed from the data analysis. The data analysis included a total number of 359 participants. More specifically, there were 124 participants in group 1, 116 participants in Group 2, 59 participants in group 3, and 60 participants in group 4. Participants were nearly equally assigned into condition 1 (n=124), condition 2 (n=116), and condition 3 (group 3 and group 4 combined, n=119). 5.2.3 Procedure Participants were adult users of Amazon Mechanical Turk (mTurk). Forty-five minutes were allotted for per participant. I required that participants be masters to do my task and their 40 HIT approval rate percentage for all requesters’ HITs was greater than 85%. Upon recruitment, participants were provided with the IRB-approved consent form electronically, where they indicated their willingness to participate in the study by clicking on an agreement statement. Upon completing the consent procedure, participants were randomly assigned into four groups based on predetermined quota: text only, video only, text followed by video, and video followed by text. A timer question was used to ensure that participants spend an adequate amount of time with each stimulus. For the text-based messages, the timer was set to 15 seconds, after which, the ability to move forward to the next element was enabled. As for the video messages, the ability to move forward to the next element was restricted with the same amount of time for the video. The video and text were embedded into Facebook posts. Each participant was presented with three different sets of messages (two same messages in each set). The messages were either texts or videos. Upon exposure to each experimental block of stimuli, participants were asked the information overload, information effectiveness, attitude toward the message, VBI, and intention to follow the suggestions after each set of messages. Considering that the posts were about promoting healthy eating during the COVID-19 pandemic and participants’ opinions regarding the COVID-19 might influence how they process the information in the posts. At the end of the survey, I asked perceived susceptibility of the Coronavirus and anxiety about the Coronavirus, which were used as control variables in the data analysis. Same as opinions about COVID-19, participants in different age (Phillips & Sternthal, 1977), gender (Sun et al., 2010), race and socio-economic status might process the health message posts differently, I asked participants’ demographic information and control it in the data analysis. The survey took 25 minutes on average. 41 5.2.4 Measurement The main study used the same measures from the pretest for information overload, information effectiveness, attitude toward the message, intention to follow suggestions that had been assessed in the pretest study. Reliability metrics from previous studies are included, yet Table 7 provides the list of items for each variable along with validity and reliability indices. The main study included the following new variables: VBI, perceived susceptibility of COVID-19, and anxiety about COVID-19 were also tested in the main study. Considering the new features of Facebook, I adjusted Alhabash and colleagues’ VBI scale (2015) by adding several new items, participants answered the questions on a seven-point Likert scale ranging from strongly disagree (1) to strongly agree (7): “This message is worth sharing with others;” “I will recommend this message to others;” “I will press on the ‘Like’ button (emoji embedded) on Facebook;” “I will press on the ‘Love’ button (emoji embedded) on Facebook;” “I will press on the ‘Haha’ button (emoji embedded) on Facebook;” “I will press on the ‘Wow’ button (emoji embedded) on Facebook;” “I will press on the ‘Sad’ button (emoji embedded) on Facebook;” “I will press on the ‘Angry’ button (emoji embedded) on Facebook;” “I will ‘comment’ on this message on Facebook;” “I will ‘share’ this message on Facebook.” Adapted from Champion’s perceived susceptibility scale (1999), participants rated the following statements using a Likert scale, ranging from strongly disagree (1) to strongly agree (5): “I feel I will get the Coronavirus sometime during my life;” “My chances of getting the Coronavirus in the next few months are great;” and “It is likely that I will get the Coronavirus.” (Cronbach’s α = .85) Using Lee’s (2020) Coronavirus Anxiety Scale (CAS), I asked participants to rate five statements on a 5-point frequency scale (0=“not at all” to 4 “nearly every day over the last 2 42 weeks”): “I felt dizzy, lightheaded, of faint, when I read or listened to news about the coronavirus;” “I had trouble falling or staying asleep because I was thinking about the coronavirus;” “I felt paralyzed or frozen when I thought about or was exposed to information about the coronavirus;” “I lost interest in eating when I thought about or was exposed to information about the coronavirus;” and “I felt nauseous or had stomach problems when I thought about or was exposed to information about the coronavirus.” A higher number on this scale represented higher anxiety about the Coronavirus (Cronbach’s α = .85). Additionally, participants answered following questions: birth year, gender, race/ethnicity, education, household income, marital status, number of dependents, current employment status, height and weight, perception of their own health (i.e., excellent, very good, good, fair or poor), snacking behavior, and who cooks at home. 43 CHAPTER SIX: RESULTS 6.1 Descriptive Statistics Three participants who did not answer two attention check questions correctly and 37 participants who did not complete the survey were excluded from the data analysis. Looking into the Boxplots, no extreme outliers (extend more than three box-lengths from the edge of the box) were observed on information load, information effectiveness, attitude toward message, VBI, or intention to follow suggestions. Per Pallant (2020) suggestion, and given that that the 5% trimmed mean and the mean value were not very different, all remaining cases were retained for statistical analyses. The final data analysis included a total number of 359 participants. More specifically, there were 124 participants in Group 1, 116 participants in Group 2, 59 participants in Group 3, and 60 participants in Group 4. This achieves the expected proportion of random assignment mentioned in the methods section. Groups 1 and 2 each constituted one third of the whole sample. Group 3 and Group 4 each constitute 1/6 of the whole sample, for a combined one-third of the sample. Overall, participants were nearly equally assigned into Condition 1 (n=124), Condition 2 (n=116), and Condition 3 (Group 3 and Group 4 combined, n=119). The average age of the sample (N=359) was 42.15 with the standard deviation of 10.98. About 48% of participants were female, while 52% were male. Most participants identified as White (61.9%), followed by Asian (26.9%), Black or African American (6.7%), American Indian or Alaska Native (1.4%), other (0.3%), and mixed races (2.8%). About half of the sample had a bachelor’s degree. Most participants (55.8%) reported themselves as married or in a domestic partnership. About 70% of the sample had no dependents. Two-thirds of the sample were employed for wages. I have calculated participants’ BMI using their self-reported height and 44 weight according to CDC’s formula (2020). According to the standard weight status categories (CDC, 2020), most participants (42.2%) reported as normal weight or healthy weight (BMI falls between 18.5-24.9), followed by 30% reported as overweight (25.0-29.9), 19% reported as obesity (30.0 and above), and 8.9% reported as underweight (below 18.5). All demographic results were included in Table 5 and other general questions such as health status were included in Table 6. All means (SDs) of independent variable, dependent variables and covariates were displayed in Table 7. 6.2 Scale Reliability After running the factor analysis (direct oblimin), one item of information overload (“The information I received is about the right amount of information I need.”) was removed because it did not load well. Two items of information effectiveness (“This message is weak.” “This message is uninformative.”) were removed. Four items of VBI (“I will press on the "Haha" button on Facebook.” “I will press on the "Wow" button on Facebook.” “I will press on the "Sad" button on Facebook.” “I will press on the "Angry" button on Facebook.”) were removed because they loaded on a different factor, which sent me back to the original VBI scale (Alhabash et al., 2015). Eigenvalues and percentage of variance explained are reported in Table 7. All eigenvalues were above 1. Cronbach’s α is regarded as an important concept in the assessment and questionnaire evaluation (Tavakol & Dennick, 2011). The acceptable values of Cronbach’s α have been reported by different studies, ranging from .70 to .95 (Bland & Altman, 1997; Cohen et al., 1996; DeVellis, 2016). All scales of this study showed acceptable internal consistency, with the Cronbach’s α reported above .90 (see Table 7). 45 6.3 Effect of Modality Succession (Condition) To test H1a-e regarding to the effects of modality succession, data for information overload (H1a), information effectiveness (H1b), attitude toward message (H1c), VBI (H1d), and intention to follow suggestions (H1e) were submitted to identical, yet separate, 3 (modality succession) x 3 (block repetition) repeated measures analysis of covariances (ANCOVAs) with modality succession as a between-subject factor and block repetition as a within-subject factor, and while controlling for participants’ age, gender, BMI, race, education, marital status, number of dependent, employment status, household income, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19 (see Table 8-12). These analyses were run among Condition 1, Condition 2, and Condition 3. As explained in the method section, Group 3 and Group 4 combined into Condition 3. Preliminary checks were conducted to ensure that there was no violation of the assumptions of normality, linearity, homogeneity of variances, homogeneity of regression slopes, and reliable measurement of the covariates. Covariates were not too strongly correlated with one another (r ranged from -.35 to .34, see Table 13). In all analyses where the assumption of sphericity was violated, Hynh-Feldt-adjusted degrees of freedom were reported. H1a predicted that participants exposed to video and text (complementary modalities) will report lower information overload than those exposed to the video-only and text-only conditions. The main effect of modality succession on information overload was not significant, F(2, 343) = .92, ns. H1a was not supported. The following covariates had significant main effects on information overload: race: F(1, 343) = 43.06, p < .001, η2p = .11; anxiety about COVID-19: F(1, 343) = 52.82, p < .001, η2p = .13; and marital status: F(1, 343) = 4.39, p < .05, η2p = .01 (see Table 8). 46 H1b predicted that participants exposed to video and text (complementary modalities) will report higher information effectiveness than those exposed to the video-only and text-only conditions. The main effect of modality succession on information effectiveness was not significant, F(2, 343) = .49, ns. H1b was not supported. The following covariates had significant main effects on information effectiveness: F(1, 343) = 26.87, p < .001, η2p = .07; employment: F(1, 343) = 6.64, p < .05, η2p = .02; and age: F(1,343) = 6.43, p < .05, η2p = .02 (see Table 9). H1c predicted that participants exposed to video and text (complementary modalities) will report more favorable attitude toward healthy eating promotion message than those exposed to the video-only and text-only conditions. The main effect of modality succession on attitude toward message was not significant, F(2, 343) = .35, ns. H1c was not supported. The following covariates had significant main effects on attitude toward message: race: F(1, 343) = 17.25, p < .001, η2p = .05; and health status: F(1, 343) = 4.57, p < .05, η2p = .01 (see Table 10). H1d predicted that participants exposed to video and text (complementary modalities) will report higher VBI than those exposed to the video-only and text-only conditions. The main effect of modality succession on VBI was not significant, F(2, 343) = .05, ns. H1d was not supported. The following covariates had significant main effects on VBI: age: F(1, 343) = 11.94, p < .001, η2p = .03; race: F(1, 343) = 47.05, p < .001, η2p =.12; anxiety about COVID-19: F(1, 343) = 14.13, p < .001, η2p = .04; and employment: F(1, 343) = 4.32, p < .05, η2p = .01 (see Table 11). H1e predicted that participants exposed to video and text (complementary modalities) will report higher behavioral intention to follow suggestions than those exposed to the video- only and text-only conditions. The main effect of modality succession on intention to follow suggestions was not significant, F(2, 343) = .18, ns. H1e was not supported. The following 47 covariates had significant main effects on intention to follow suggestions: race: F(1, 343) = 13.33, p < .001, η2p = .04; employment: F(1, 343) = 7.29, p < .01, η2p = .02; age: F(1, 343) = 5.78, p < .05, η2p = .02; and health status: F(1, 343) = 5.15, p < .05, η2p = .02. No interaction effect of message repetition and modality succession (condition) was found on information overload, information effectiveness, attitude toward message, VBI, nor intention to follow suggestions (see Table 12). 6.4 Modality Order Effect To answer R1a-e related to the effects of modality order, I ran repeated measures analysis of covariances (ANCOVAs) between Group 3 (video followed by text) and Group 4 (text followed by video) with controlling participants’ age, gender, BMI, race, education, marital status, number of dependent, employment status, household income, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19 (see Table 14-18). The between- subject factor was the modality order (video followed by text, text followed by video), and the within-subject variables consisted of three scores on the information overload, information effectiveness, attitude toward message, VBI, and intention to follow suggestions, assessed after the stimulus was displayed. RQ1a asked what the effect of modality order in the complementary modality condition on information overload is. The main effect of modality order on information overload was not significant, F(1, 104)=.01, ns. The following covariates had significant main effects on information overload: anxiety about COVID-19, F(1, 104) = 28.00, p < .001, η2p = .21; BMI, F(1, 104) = 6.64, p < .05, η2p = .06; and race, F(1, 104) = 6.43, p < .05, η2p = .06. There was an interaction effect of repetition and order on information overload, F(2, 103) = 4.32, p < .05, η2p = .08 (see Table 14). 48 RQ1b asked what the effect of modality order in the complementary modality condition on information effectiveness is. The main effect of modality order on information effectiveness was not significant, F(1, 104) = .42, p = .52, η2p = .004. The following covariates had significant main effects on information effectiveness: race, F(1, 104) = 12.72, p < .01, η2p = .11; repetition, F(2, 103) = 1.20, p < .05, η2p = .02; age, F(1, 104) = 4.77, p < .01, η2p = .07; and employment, F(1, 104) = 4.68, p < .05, η2p = .04. There was an interaction effect of repetition and order on information effectiveness, F(2, 103) = .03, p < .01, η2p = .07 (see Table 15). RQ1c asked what the effect of modality order in the complementary modality condition on attitude toward message is. The main effect of modality order on attitude toward message was not significant, F(1, 104) = .65, p = .53, η2p = .01. Race had a significant main effect on attitude toward message, F(1, 104) = 5.28, p < .05, η2p = .05 (see Table 16). RQ1d asked what the effect of modality order in the complementary modality condition on VBI is. The main effect of modality order on VBI was not significant, F(1, 104) = .07, ns. The following covariates had significant main effects on VBI: race, F(1, 104) = 12.89, p < .01, η2p = .11; and anxiety about COVID-19, F(1, 104) = 8.37, p < .01, η2p = .07 (see Table 17). RQ1e asked what the effect of modality order in the complementary modality condition on intention to follow suggestions is. The main effect of modality order on intention to follow suggestions was not significant, F(1, 104) = .18, p = .67, η2p = .002. Health status had a significant main effect on intention to follow suggestions, F(1, 104) = 5.16, p < .05, η2p = .05 (see Table 18). 6.5 Serial Mediation Analysis: Effect of Modality Succession on Behavioral Intentions H2 predicted that modality succession affects behavioral intentions through the serial mediation of information overload, information effectiveness, attitudes toward message, VBI, 49 and intention to follow suggestions. H3-7 predicted relationships among modality succession, information overload, information effectiveness, attitude toward message, VBI, and intention to follow suggestion. To test these hypotheses, I ran a serial mediation model using Hayes’ (2021) PROCESS macros for SPSS (version 3), using the Model 6 specification with 10,000 bootstrap samples. In the model, modality succession (MS) was treated as an independent variable, information overload (IL), information effectiveness (IE), attitude toward message (Am), and viral behavioral intention (VBI) as mediators, and behavioral intention to follow suggestions (BI) as a dependent variable, participants’ age, gender, BMI, race, education, marital status, number of dependent, employment status, household income, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19 as control variables (see Table 19). H2 predicted that modality succession affects behavioral intentions through the serial mediation of information overload, information effectiveness, attitudes toward message, VBI, and intention to follow suggestions. The final regression model predicting intention to follow suggestions (BI) was statistically significant and fully explained 63% of the variance (R = .79, R2 = .63, MSE = .74, F(19,339) = 30.43, p < .001). Results, summarized in Table 20, showed that attitude toward message was strongest predictor of intention to follow suggestions (Coefficient = .69, SE = .11, t = 6.35, p < .001, CILL–UL = .48 to .91), suggesting that the more favorable the attitude toward message, the higher the intention to follow suggestions. In this study, I have only reported unstandardized coefficients according to Hayes’ suggestion (2017, p. 43). VBI was positively related to intention to follow suggestions, Coefficient = .26, SE = .04, t = 6.22, p < .001, CILL–UL = .18 to .34. The more the likelihood to engage with the Facebook post, the greater the intention to follow suggestions. Information overload was negatively associated with intention to follow suggestions, Coefficient = -.12, SE = .04, t = -2.98, p < .01, CILL–UL = -.20 to 50 -.04. The more the information overload, the less the intention to follow suggestions. Education was positively related to intention to follow suggestions, Coefficient = .08, SE = .04, t = 2.03, p < .05, CILL–UL = .003 to .16. The higher the education, the higher the intention to follow suggestions. As for the serial mediation, results of bias–corrected confidence interval with a bootstrap sample of 10,000 for the total indirect effect of modality succession on intention to follow suggestions for text only covers zero, Coefficient = .07, SE = .17 , t = .43 , CILL–UL = -.26 to .41; the total indirect effect of modality succession on intention to follow suggestions for complementary modalities (video and text combined) covers zero, Coefficient = .10, SE = .17 , t = .57, CILL–UL = -.24 to .43. All serial paths were not significant as the bias–corrected confidence intervals cover zero (see Table 20). H3 predicted that information overload will be negatively associated with information effectiveness, which was not supported (Coefficient = -.05, SE = .04, t = -1.21, p =.23, CILL–UL = -.14 to .03). H4 predicted that information effectiveness will be positively associated with attitude toward the message, which was supported (Coefficient = .96, SE = .03, t = 35.60, p < .001, CILL–UL = .90 to 1.01). This indicated that the higher the information effectiveness, the more favorable the attitude toward the message. H5 predicted that information overload will be negatively related to attitudes toward the message, which was supported (Coefficient = -.07, SE = .02, t = -3.22, p < .05, CILL–UL = -.11 to -.03). The higher the information overload, the less favorable the attitude toward the message. H6 predicted that attitude toward message will be positively associated with VBI, which was supported (Coefficient = .89, SE = .13, t = 6.75, p < .001, CILL–UL = .63 to 1.15). The more favorable the attitude toward message, the more the likelihood to engage on Facebook post. H7 predicted that VBI will be positively associated with 51 intention to follow suggestion from the message, which was supported (Coefficient = .26 SE = .04, t = 6.22, p < .001, CILL–UL = .18 to .34). The more the likelihood to engage on Facebook post, the greater the intention to follow the suggestions from the message. 6.6 Simple Mediation Analysis: Relationship among Information Overload, Information Effectiveness, and Attitudes H8 predicted that the relationship between information overload and attitudes toward the message will be mediated by information effectiveness, such that lower information load will be associated with higher information effectiveness, which will, in turn, lead to more favorable attitudes toward the message. To test this hypothesis, I ran a serial mediation model using Hayes’ (2021) PROCESS macros for SPSS (version 3), using the Model 4 specification with 10,000 bootstrap samples. In the model, information overload (IL) was treated as an independent variable, information effectiveness (IE) as mediator, attitude toward message (Am) as a dependent variable, participants’ age, gender, BMI, race, education, marital status, number of dependent, employment status, household income, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19 as control variables (see Table 21). The final regression model predicting attitude toward message (Am) was statistically significant and explained 81% of the variance (R = .90, R2 = .81, MSE = .20, F(15, 343) = 98.50, p < .001). Results of bias–corrected confidence interval with a bootstrap sample of 10,000 for the total effect of information overload on attitude toward message were below zero, Coefficient = -.12, SE = .05, t = -2.62, p < .01, CILL–UL = -.21 to -.03. Results, summarized in Table 22, showed that information overload significantly predicted attitude toward message, Coefficient = -.07, SE = .02, t = -3.21, p < .01, CILL–UL = -.11 to -.03, suggesting that the higher the information overload, the less favorable the attitude toward message. Information effectiveness was 52 positively associated with attitude toward message, Coefficient = .96, SE = .03, t = 35.72, p < .001, CILL–UL = .90 to 1.01. The higher the information effectiveness, the more favorable the attitude toward message. Whereas, information overload did not predict information effectiveness, Coefficient = -.05, SE = .04, t = -1.28, p=.20, CILL–UL = -.14 to .03. The indirect effect result of information overload on attitude toward message covers zero, Coefficient = -.05, SE = .05, CILL–UL = -.14 to .03. Therefore, the indirect effect from information overload to attitude toward message through information effectiveness was not significant. 6.7 Supplemental Moderated Mediation Analysis: Effect of Information Load and Effectiveness on Behavioral Intentions Although no significant variances was created by the manipulation, the average scores on information overload, information effectiveness, attitude, VBI, and intention to follow suggetions differentiate among conditions. As suggested by O’Keefe (2003), if there is variance in the psychological states (e.g., information overload) we can still assess the relationship between the psychological states and outcomes. I was interested in understanding relationships among information overload, information effectiveness, attitude, VBI, and behavioral intentions. Therefore, a supplemental analysis of measured variables was conducted. In this supplemental analysis, I ran a moderated mediation analysis using Hayes’ (2021) PROCESS macros for SPSS (version 3), using the Model 92 specification with 10,000 bootstrap samples. In the model, information overload (IL) was treated as an independent variable, information effectiveness (IE) as a moderator, attitude toward message (Am) and viral behavioral intention (VBI) as mediators, and behavioral intention to follow suggestions (BI) as a dependent variable, participants’ age, gender, BMI, race, education, marital status, number of dependent, employment status, 53 household income, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19 as control variables (see Tables 22 and 23). The final model predicting intention to follow suggestions (BI) was statistically significant and explained 64% of the variance in behavioral intentions to follow the message’s suggestions of consuming healthy snacks during COVID-19 (R = .80, R2 = .64, MSE = .71, F(20, 338) = 30.45, p < .001). Results, summarized in Table 23, showed that information overload significantly predicted attitude toward the message, Coefficient = -.23, SE = .11, t = -2.13, p < .05, CILL–UL = -.44 to -.03, suggesting that the higher the perceived information overload, the less favorable the attitude toward the message. Information effectiveness was positively associated with attitude toward message, Coefficient = .88, SE = .05, t = 16.16, p < .001, CILL–UL = .78 to .99. The higher the information effectiveness, the more favorable the attitude toward the message. However, the interaction between information overload and information effectiveness did not predict attitude toward message, Coefficient = .03, SE = .02, t = 1.52, p=.13, CILL–UL = -.01 to .06. Results also showed that information effectiveness negatively predicted VBI, Coefficient = -.92, SE = .30, t = -3.01, p < .01, CILL–UL = -1.51 to -.32, suggesting that the higher the information effectiveness, the greater the likelihood to engage with the message online. The interaction of attitude toward message and information effectiveness positively affected VBI, Coefficient = .17, SE = .05, t = 3.76, p < .001, CILL–UL = .08 to .26. As shown in Figure 3, in instances when participants indicated less favorable attitudes toward the message, no differences were detected in their VBI. However, when participants indicated more favorable attitudes toward the message, the perception of high information effectiveness resulted in greater VBI. However, information overload did not predict VBI, Coefficient = -.11, SE = .26, t = -.41, p= .66, 54 CILL–UL = -.62 to .41. Attitude toward the message did not predict VBI, Coefficient = .13, SE = .24, t = .53, p= .60, CILL–UL = -.34 to .60. The interaction of information overload and information effectiveness did not predict VBI, Coefficient = .03, SE = .04, t = .71, p= .48, CILL– UL = -.06 to .12. Viral Behavioral Intentions 5.5 5 4.5 4 3.5 3 2.5 Low Att (4.53) Moderate Att (5.67) High Att (6.53) Low IE (4.47) Moderate IE (5.47) High IE (6.40) Figure 3 Interaction of Attitude and Information Effectiveness on VBI Lastly, results showed that information overload negatively predicted intention to follow suggestion, Coefficient = -.63, SE = .21, t = -3.03, p < .01, CILL–UL = -1.03 to -.22, suggesting that the higher the information overload, the less the likelihood to follow suggestion. Attitude toward message positively predicted intention to follow suggestion, Coefficient = 1.31, SE = .28, t = 4.75, p < .001, CILL–UL = .77 to 1.85, suggesting that the more favorable the attitude, the more the likelihood to follow suggestion. The interaction of information overload and information effectiveness positively predicted intention to follow suggestions, Coefficient = .08, SE = .04, t = 2.38, p < .05, CILL–UL = .01 to .15. In this regard, when information load was low, participants expressed higher behavioral intentions when they perceived the message to be less effective, than moderate, and highly effective, respectively. When information load was perceived moderately, 55 no differences were found across different perceptions of information effectiveness. Finally, when perceived information load was high, participants’ behavioral intentions were higher in instances when participants perceived the message to be highly effective, followed by moderate and low effectiveness, respectively (see Figure 4). Behavioral Intentions 5.6 5.4 5.2 5 4.8 4.6 4.4 4.2 4 Low IL (1.20) Moderate IL (2.47) High IL (4.67) Low IE (4.67) Moderate IE (5.47) High IE (6.40) Figure 4 Interaction of Information Overload and Information Effectiveness on Intention to Follow Suggestions The interaction between attitude toward message and information effectiveness negatively predicted intention to follow suggestions, Coefficient = -.13, SE = .05, t = -2.53, p < .05, CILL–UL = -.24 to -.03. In contrast to the interaction between load and effectiveness, when attitude was brought into the mix, participants who expressed less favorable attitudes toward the message expressed higher behavioral intentions when they perceived the message to be highly effective, followed by moderate and low level of information effectiveness, respectively. No differences were detected in behavioral intentions as a function of information effectiveness at moderate levels of attitudes toward the message. Finally, when participants expressed more favorable attitudes toward the message, the highest behavioral intentions were expressed for 56 participants who indicated the lowest level of perceived information effectiveness, followed by moderate and high levels, respectively (see Figure 5). Behavioral Intentions 6 5.8 5.6 5.4 5.2 5 4.8 4.6 4.4 4.2 4 Low Att (4.53) Moderate Att (5.67) High Att (6.53) Low IE (4.67) Moderate IE (5.47) High IE (6.40) Figure 5 Interaction of Attitude and Information Effectiveness on Intention to Follow Suggestions VBI did not relate to intention to follow suggestions, Coefficient = -.14, SE = .24, t = -.59, p = .56, CILL–UL = -.61 to .33. Information effectiveness did not relate to intention to follow suggestions, Coefficient = .21, SE = .25, t = .84, p = .40, CILL–UL = -.28 to .70. The interaction of VBI and information effectiveness did not predict intention to follow suggestions, Coefficient = .07, SE = .04, t = 1.76, p = .08, CILL–UL = -.01 to .16. The final model is depicted in Figure 6. 57 Figure 6 Moderated Mediation Model 58 CHAPTER SEVEN: DISCUSSION Given the prevalence of obesity in the last decades and weight gain during the COVID-19 pandemic in the United States, I was interested in incorporating theoretical approaches for social media application to promote healthy eating behaviors. More specifically, I adapted Stephens and Rain’s (2011) ICT Succession Model with integrating the Media Richness theory to examine how different modalities (video and text) used on social media relate to individual’s attitudes toward the message and, in turn, intention to follow the suggestions from the messages. With increasing emphasis on healthy eating to prevent obesity and diabetes, the current study investigated effects of repetitive messages integrated with different modalities on motivating attitude and behavioral change. 7.1 Modality Succession and Succession Order Different from previous findings (Stephens & Rain, 2011), results of this study showed that the modality succession (video only, text only, video and text combined) did not make any difference on information overload (H1a), VBI (H1d), nor intention to follow suggestions (H1e). In alignment with Stephens and Rains’ study (2011), modality succession had no significant main effect on information effectiveness (H1b) or attitude (H1c). Per the original premise of ICT succession theory, complementary ICT use is defined and operationalized as the repetition of the same message across channels that vary in affordances and the number of cues (e.g., email vs. FtF). However, such a manipulation is confounded as it does not exhaustively identify whether the effect of ICT succession is the result of channel succession or differences in modalities (media richness). To avoid the confound between modality and channel, my study only tested content in different modalities on a single channel, which deviated from the original approach of ICT succession theory. It is possible that ICT succession theory is more applicable to the combinatorial variability in both channels and modality rather than a test of succession within the 59 same channel. Future research should decipher this effect and isolate its nature in an experimentally valid way. Stephens and Rains (2011) found no order effects on attitudes nor information effectiveness. Whereas there were order effects on the overload and behavioral intentions. In my study, modality order (video followed by text vs. text followed by video) did not influence any of the outcome variables (RQ1a-e). It is plausible that the exposure to different modalities of the same message within the same channel was not strong enough to affect outcome variables across different order conditions. In other words, given that participants were exposed to the same message in different modalities and different orders within the same channel, participants still perceived the messages to be similar, and the repetitious exposure did not result in variability in the outcome variables. Furthermore, my results suggested that video was not more effective than text in lessening information overload, nor improving information effectiveness, attitude toward message, VBI, or intention to follow suggestions, which did not support findings from Lu et al. (2014). Again, taking the within-channel effects of various modalities, it is plausible that participant’s experience of viewing text and/or video messages within a Facebook context did not yield significant variability in any of the outcome variables. 7.2 Moderated Mediation Model With the controlling of modality succession, the serial mediation model suggested significant relationship between attitude and VBI, and relationship between VBI and behavioral intention. In contrast, in the moderated mediation model, without controlling modality succession, the significant relationship between attitude and VBI, and relationship between VBI and behavioral intention disappeared. Given the result that the modality succession did not 60 predict any outcome variables, I will focus on discussing the moderated mediation model in this chapter. VBI was a significant mediator in the serial mediation model but not in the moderated mediation model, suggesting that VBI does not operate on its own. The interesting thing here is that perceived information overload and information effectiveness do matter for influencing behavioral intentions, yet this effect is strengthened if both information overload and information effectiveness influence attitudes. What is even more interesting is that attitude, on its own, is not a significant predictor of VBI, yet does so by interacting with information effectiveness. So, in essence, what this means is that the combination of high perceived effectiveness that leads to higher favorability (attitudes) can evoke greater VBI. The other interesting pattern from the data is the significant two-way interaction between information overload and information effectiveness. What the figure tells us is that in cases when participants experienced lower levels of information overload, or perceived the information to be less cognitive loaded, positive effects on behavioral intentions were observed even in cases when information effectiveness was low. In fact, the lower the information effectiveness, the more likely they were to follow the suggestions, compared to when the effectiveness was high. On the other hand, when information overload was high (overload), information effectiveness was crucial in determining behavioral intentions, such that to evoke greater adherence to the message, the information should be perceived as highly effective by participants. This effect is a classic example of elaboration likelihood, where the availability of cognitive resources determines the elaborative route taken to process the information (Cacioppo et al., 1996; O’Keefe, 2013; Petty et al., 1992; Petty et al., 1983; Petty & Wegener, 1999). When participants were not overloaded, they processed information superficially through peripheral cues, whereas, when the availability 61 of cognitive resources was low (overload), higher effectiveness was needed to acquire the desired effects on behavioral intentions. This is further illustrated by the two-way interaction between attitude and information effectiveness. For participants who developed less favorable attitude toward message, it was important to perceive the message as highly effective for that to positively influence behavioral intentions, whereas higher effectiveness worked in opposite direction when attitude toward message was high. This suggests that the development of positive message attitude, despite perceptions of information effectiveness, is a sufficient condition for influencing behavioral intentions. The last argument I would make is that the moderated mediation model (see Figure 6) shows that information overload, information effectiveness, and their interaction affect attitude, which in turn affects behavioral intentions. VBI, as a novel variable, is critical to examine here, because it is shown to not be part of the sequential mediation of the effects of information overload and information effectiveness on behavioral intentions, and more importantly, that this effect is pertinent on how attitude and information effectiveness interact to influence it, yet VBI did not directly influence behavioral intentions. Contrary to prior studies (e.g., Alhabash et al, 2015; Krishnan & Zhou, 2019) that found a direct relationship between VBI and offline behavioral intentions and as a mediator to the relationship between attitudes and behavioral intentions, the current study provides alternative evidence. There are a few plausible explanations here. First, it is possible that regarding processing message related arguments and seeing how they influence planned behavior, the role of VBI is distinct from the attitude-behavior relationship. VBI seems to operate as a different type of outcome variable, time-bound by the affordances of the social media platform. Second, it is also plausible that participants perceived positive health messages as normative, thus were ready to 62 engage with them online, yet planning for healthy eating might not be as convenient or attainable. This points to a potential disconnect between online and offline behaviors as they relate to healthy eating. Finally, and by way of understanding effective health communication on social media, it is possible that the ways in which participants were influenced online and offline by the messages could yield valuable insights regarding segmenting the target audience. In other words, it is possible that a portion of the sample in this study were more interested in engaging online than enacting changes in their offline behaviors regarding healthy eating, while others allocated greater cognitive resources to evaluating the persuasive arguments and decided to allocate their cognitive resources to planning their behavior change offline. Such a plausible explanation is worthy of further exploration, especially amid prior research within the activism literature arguing that engaging online could substitute offline behavior change in what is called slacktivism (Knibbs, 2013). With that in mind and considering the finite amount of cognitive resources available for processing persuasive messages, it is plausible that individuals are dividing their resources based on the context of their planned behavior, where some focus on engaging online, while others plan for offline behavior change. 7.3 Implications In the original ICT Succession study, Stephens and Rains (2011) manipulated complementary ICT use by repeating the same message either on the same channel (email vs. face to face) or across two channels (email then FtF, FtF then email). However, such a manipulation is confounded as it does not exhaustively identify whether the effect of ICT succession is because of channel succession or differences in modalities, given that FtF communication provides more nonverbal cues than email communication. My study aimed to investigate the effect of different modalities and their succession (modality followed by a 63 different modality) within a single channel on information overload, information effectiveness, attitude, and behavioral change. Results show that different modalities or succession did not make any difference. To understand the mechanism of ICT succession theory, it would be interesting for future research to compare effects of messages in same modality across different channels. Although my study did not fully support the ICT succession theory, it is worthwhile to confirm that information overload negatively predicted attitude toward message and intention to follow suggestions while information effectiveness positively predicted attitude. To promote people’s attitude and behavioral change toward healthy behaviors, health professionals should dedicate their efforts in lessening information overload and improving information effectiveness. Health communicators should define and focus on their target outcomes. Planned behavior should be weighted more heavily when aiming at changing behavior, whereas VBI is more suitable when the goal is to raise awareness. 7.4 Limitation and Future Research There are some limitations of this study and could be improved in future research. At first, the manipulation might be unsuccessful in several reasons. Under the health and safety guidance for research during the COVID-19, this study was transferred from an in-person to an online experiment study. Participants completed the study as a self-administered questionnaire. Although I enforced a timer that disabled the “Next” button for a particular period congruent with the video length and estimated reading time of the text-based posts, there were no guarantees that participants watched the entire video or read the message. The videos were not set as auto-play so that it was possible that participants did not watch the videos. Violated the external validity, participants were forced to watch videos or read texts until the timer was over. 64 Participants in group 1 were shown the same videos twice sequentially while participants in group 2 were shown the same texts twice sequentially. A repeated measures ANCOVA showed that participants spent less time on the second exposure of the same message, F(2, 343) = 9.02, p < .001, η2p = .05. In condition 1 (video only), participants spent an average time of 107.25 seconds (SD = 3.922) on the first exposure while 99.81 seconds (SD = 2.926) on the second exposure of the same message. In condition 2 (text only), participants spent an average time of 75.28 seconds (SD = 4.08) on the first exposure while 38.317 seconds (SD = 3.04) on the second exposure. In condition 3 (video and text combined), participants spent an average time of 86.60 seconds (SD = 4.06) while 74.80 seconds (SD = 3.03) on the second exposure. Although in condition 3 (group 3 and group 4) participants were shown to the same message twice but in different modalities (video followed by text or text followed by video), participants could easily identify the two messages were the same. Participants might have not paid attention to the second exposure after they realized that was the same video or text. The first message had 9 suggestions, while there were 5 in the second message and 4 in the third message. The measurement of intention to follow suggestions might not be accurate considering that participants might not be sure about what exact “suggetions” I was referring to. One of the study’s major shortcomings lies in the very reason why I altered the operationalization of succession from the original ICT Succession model. In their original study, participants received the same message, yet the combinatorial nature of the stimuli made it seem like two factors (modality and channel) were manipulated, thus posing threats to internal validity. However, it is plausible that this very design alteration yielded non-significant effects of the manipulation here. The nature of the current study in that participants immediately saw the same message either in the same modality or in varying modalities, could have limited the 65 potential effects of media richness. In other words, had this study either incorporated the message repetition factor within a newsfeed environment, where participants would be exposed to the same message in similar or varying modalities yet separated by other content, the effects could have been different. Additionally, it is also plausible that ICT Succession works more effectively when both modality and channel are combined. For example, future research could incorporate FtF and digital channels for presentation of the persuasive messages. Future research could potentially combine the testing of ICT succession as originally operationalized, alongside the modality succession described in this study. Another disadvantage of this study was that the study purpose was too obvious. To avoid bias effect of a single message, I showed participants three sets of messages, whereas Stephens and Rains’ study (2011) only delivered one set of messages. In other words, participants in my study saw a total of 6 videos, 6 texts, or 3 videos and 3 texts. Participants were bombarded with healthy eating promotion posts and immediately after they were asked to indicate their perception of the information effectiveness, attitude toward message, VBI, and intention to follow suggestions (to eat healthy). It is plausible that participants were able to identify the study purpose, which was promoting healthy eating behavior, thus hindering the testing of the effect of modality. To fulfill the social desirability, no matter which group participants were assigned to, they indicated they had favorable attitude toward message, intention to engage on the post, and intention to follow suggestions. The manipulation did not make any significant difference. My pretest study recruited from college students while main study recruited from a more diverse population. The stimuli that decided based on results of college students might not work the same way to the main study participants. Future study should make sure the pretest study’s sample to be as similar as possible to the main study’s sample in terms of age range, employment 66 status, etc. Last, all measures were based on self-report. In other words, they all came from one person’s mind. According to Fuller and colleagues (2016), “researchers place most concern in the possibility that common method variance may falsely inflate observed relationships among measures. If so, biased results could cause a researcher to falsely conclude that a relationship exists (enhancing type I error).” (p. 2) Future research could have explicit measure, for example providing participants opportunities to hit the “like” and “share” buttons on Facebook or asking them to create a shopping list of foods. Future researchers can conduct in-person experiments to exert greater control in comparison with online experiments. More specifically, researchers can ensure participants watch the entire video and read the texts, and encourage participants from navigating to distracted or irrelevant elements in their physical environment, other digital technologies, and the same screen through which the stimuli are presented. My stimulus comprised of screenshots of fake Facebook posts embed into Qualtrics survey. During the in-person experiments, participants can have more realistic experience with social media where they can really press “like”, “share” and comment on the posts. Additionally, researchers can have participants work on multimedia such as social media, emails, and websites. In this case, researchers can test effects of different modalities across different communication channels such as text on social media, video on social media, text on health organization’s website, and video text on health organization’s website. To lessen the effect of social desirability bias, it will be interesting to assess other outcome variables such as healthy eating behavior by providing healthy foods/drinks and unhealthy foods/drinks before they leave the lab and asking participants to take three things away as a part of their compensation. If a healthy food/drink is picked, then the participant gets 1 67 point, otherwise 0 point. Each participant will get a total point of healthy eating behavior varying from 0 to 3. 68 APPENDICES 69 Appendix A: List of True/False Recognition Questions for Message Comprehension, Pretest Study Video 1/ Text 1 Q1: Omega-3 has been found to lower depression. True False Q2: Folate can reduce the risk of depression. True False Q3: Vitamin D does not help reduce with anxiety and depression. True False Q4: Eating junk food does not affect mental health. True False Video 2/ Text 2 Q1: Dried peas and beads, canned chickpeas, cannellini beans, kidney beans are good sources of fiber and protein and macro and micronutrients. True False Q2: Canned chicken, tuna, sardines, salmon are good sources of fiber. True False Q3: Pasta, quinoa, barley, oats, rice, couscous are good sources of protein. True False Q4: It is not recommended to consume a lot of protein during the pandemic. True False Video 3/ Text 3 Q1: Vitamin A, Vitamin C, Vitamin D, and Vitamin E play a pivotal role in a healthy functioning immune system. True False Q2: Minerals iron, selenium, magnesium, and copper are important for immune system. True False 70 Q3: Vitamin B2, B6, B9, and B12 are not important for the immune system. True False Q4: Vitamin B6 is folate. True False Video 4/ Text 4 Q1: We should make sure that we're eating three balanced meals a day to keep ourselves from getting too hungry, to keep ourselves from getting too snacky. True False Q2: We could build a healthy snack with some protein, some complex carbohydrates and sit down and really enjoy that. True False Q3: Research shows that we 80% more likely to eat what we see first. True False Q4: When we want a snack, it’s OK eat in front of the computer screen. True False Video 5/Text 5 Q1: Eating highly processed foods, sugary foods, and fast food lines make you feel tired and sluggish. True False Q4: Nutrient-dense foods are clean energy sources for you. True False Q3: Healthy eating is expensive and difficult. True False Q4: Healthy eating is not going to help you fight against COVID-19. True False 71 Appendix B: Experimental Stimuli in Main Study 72 73 74 Table 1 Eigenvalues, Cronbach's αs and Means (SDs) for Pretest Study Variable Name Message 1 Message 2 Message 3 Message 4 Message 5 Information Overload: The information I Eigenvalue 3.78 4.03 3.12 3.53 3.68 received… % of Var. Exp. 62.78 67.19 52.02 58.79 61.39 about healthy eating behavior needs too much Cronbach’s α .83 .78 .79 .81 .84 explanation to be M (SD) 3.26(1.27) 3.58(1.28) 3.37(1.26) 3.12(1.10) 2.77(1.08) useful. requires me to make too many decisions. has too much information. is more discussion than I wished. is more information than I need. Information Effectiveness: This message is… Eigenvalue 4.08 4.49 4.03 4.43 4.04 effective. % of Var. Exp. 58.26 64.15 57.61 63.31 57.74 detailed. Cronbach’s α .51 .58 .53 .51 .63 useful. M (SD) 4.33(.72) 4.26(.83) 3.96(.81) 4.36(.65) 4.00(.94) in high quality. well supported. Attitude toward the message: This message Eigenvalue 3.96 3.84 3.89 3.46 4.00 is… % of Var. Exp. 79.23 76.71 77.85 69.12 80.08 helpful to my health. Cronbach’s α .95 .94 .94 .90 .95 a valuable resource. M (SD) 5.19(1.32) 5.12(1.40) 4.55(1.40) 5.43(1.14) 4.73(1.53) important for being healthy. a good reference for meal preparation. has something positive to me. Intention to Follow Suggestions from the Eigenvalue 4.61 4.58 4.48 4.52 4.46 Message % of Var. Exp. 92.13 91.54 89.64 90.48 89.13 I intend to follow what the message suggests in Cronbach’s α .98 .98 .98 .98 .98 next seven days. M (SD) 4.40(1.77) 4.12(1.69) 3.76(1.66) 4.71(1.65) 4.38(1.56) I will try to follow what the message suggests in next seven days. I want to follow what the message suggests in next seven days. 75 I expect to follow what the message suggests in next seven days. How likely is it that you will follow what the message suggests in next seven days? Intention to Eat A Healthy Diet Eigenvalue 4.30 4.40 4.28 4.23 4.39 I intend to eat a healthy diet in next seven days. % of Var. Exp. 86.04 88.01 85.63 84.61 87.77 I will try to eat a healthy diet in next seven Cronbach’s α .97 .97 .97 .96 .97 days. M (SD) 5.02(1.62) 4.87(1.65) 5.52(1.39) 5.01(1.51) 4.97(1.61) I want to eat a healthy diet in next seven days. I expect to eat a healthy diet in next seven days. How likely is it that you will eat a healthy diet in next seven days? 76 Table 2 Pretest Results for the Effect of Modality and Message Repetition using Repeated Measures ANCOVA Message 1 Message 4 Message 5 Repetition Modality Rep X Modality Overload Video-only 2.44 (1.26) 2.63 (1.22) 2.46 (1.04) F(2, 52) = F(1, 53) = F(2, 52) = Text-only 3.27 (1.35) 2.85 (1.23) 2.62 (1.03) 1.59, .21 2.41, .126 2.08, .136 Overall 2.85 (1.36) 2.74 (1.22) 2.54 (1.03) Effectiveness Video-only 5.35 (1.14) 5.61 (1.21) 5.15 (1.18) F(2, 52) = F(1, 53) = F(2, 52) = Text-only 4.89 (1.12) 5.09 (.79) 4.07 (1.24) 9.90, .00 8.64, .01 2.16, .13 Overall 5.12 (1.14) 5.35 (1.05) 4.62 (1.32) Attitude Video-only 5.57 (1.30) 5.65 (1.06) 5.34 (1.17) F(2, 52) = F(1, 53) = F(2, 52) = Text-only 4.79 (1.23) 5.20 (1.20) 4.10 (1.64) 9.23, .00 9.79, .003 2.83, .07 Overall 5.19 (1.32) 5.43 (1.14) 4.73 (1.53) Intention to Follow Suggestions Video-only 4.91 (1.63) 4.85 (1.68) 4.81 (1.27) F(2, 52) = F(1, 53) = F(2, 52) = Text-only 3.87 (1.80) 4.57 (1.65) 3.93 (1.73) 1.35, .27 4.33, .04 1.38, .26 Overall 4.40 (1.77) 4.71 (1.65) 4.38 (1.56) Intention to Eat A Healthy Diet Video-only 5.39 (1.42) 5.30 (1.32) 5.49 (1.24) F(2, 52) F(1, 53) = F(2, 52) = Text-only 4.64 (1.75) 4.71 (1.65) 4.44 (1.80) = .091, .91 4.18, .05 2.08, .14 Overall 5.02 (1.62) 5.01 (1.51) 4.97 (1.61) Message Comprehension Video-only 2.07 (.47) 2.61 (.50) 2.00 (.47) F(2, 52) = F(1, 53) F(2, 52) = Text-only 1.93 (.55) 2.70 (.72) 2.15 (.36) 23.525, .00 = .153, .70 1.29, .28 Overall 2.00 (.51) 2.65 (.62) 2.07 (.42) 77 Table 3 Main Study Design Stimulus 1st 2nd 3rd 4th 5th 6th Group Exposure Exposure Exposure Exposure Exposure Exposure Group 1 Video 1 Video 1 Video 2 Video 2 Video 3 Video 3 Group 2 Text 1 Text 1 Text 2 Text 2 Text 3 Text 3 Group 3 Video 1 Text 1 Video 2 Text 2 Video 3 Text 3 Group 4 Text 1 Video 1 Text 2 Video 2 Text 3 Video 3 Table 4 Stimuli Message 1 Message 2 Message 3 Lengths in Seconds of Video 84 99 79 Number of Words in Text 206 239 263 Number of Suggestions 9 5 4 78 Table 5 Demographics for Main Study Sample N 359 Age-M(SD), range 42.15(10.98), 22-75 Sex-(n, %) Female 47.8 Male 52.2 Hispanic/Latino Yes 4.4 No 95.6 Race/Ethnicity American Indian or Alaska Native 1.4 Asian 26.9 Black of African American 6.7 White 61.9 Other 0.3 Mixed Races 2.8 Education-(%) Less than high school 0.3 High school graduate 12.2 Some college no degree 13.3 Associate's degree/two-year degree 9.7 Bachelor's degree 48.6 Graduate degree 15 Other 0.8 Marital Status-(%) Single, never married 35.3 Married or domestic partnership 55.8 Widowed 0.8 Divorced 6.7 Separated 1.4 Dependent-(%) Yes 31.1 No 68.9 Employment Status-(%) Employed for wages 66.4 Self-employed 21.4 Out of work and looking for work 2.2 Out of work but not currently looking for work 0.6 A homemaker 4.4 A student 0.6 Military 0.3 79 Retired 3.3 Unable to work 0.8 Income-(%) Less than $10,000 10.3 $10,000 - $24,999 18.3 $25,000 - $49,999 28.1 $50,000 - $74,999 21.4 $75,000 - $99,999 11.1 $100,000 - $ 124,999 5 $125,000 - $149,999 3.3 $150,000 or more 2.5 80 Table 6 Descriptive Results for Health Status and Snacking Behavior for Main Study N 359 Health Status-(%) Excellent 11.9 Very good 36.7 Good 35.6 Fair 14.7 Poor 1.1 BMI-(%) Underweight (below 18.5) 8.9 Normal or healthy weight (18.5-24.9) 42.2 Overweight (25.0-29.9) 30 Obesity (30.0 and above) 19 Cook-(%) Myself 71.4 Parents 13.6 Significant other 14.2 Child 0.3 Roommate 1.7 Friend 0.3 Snack between breakfast and lunch-(%) Everyday 13.9 5-6 days a week 7.8 3-4 days a week 17.5 1-2 days a week 33.9 Never 26.9 Snack between lunch and dinner-(%) Everyday 17.8 5-6 days a week 19.2 3-4 days a week 26.7 1-2 days a week 28.6 Never 7.8 81 Table 7 Eigenvalues, Cronbach's αs and Means (SDs) for Main Study Variable Name Message 1 Message 2 Message 3 Information Overload: The information I received… Eigenvalue 4.09 4.08 3.90 about healthy eating behavior needs too much explanation to be % of Var. Exp. 81.85 81.54 77.93 useful. Cronbach’s α .94 .94 .93 requires me to make too many decisions. M (SD) – All 2.92(1.82) 2.86(1.81) 2.74(1.68) has too much information. M (SD) – Group 1 2.97(1.87) 2.91(1.81) 2.67(1.66) is more discussion than I wished. M (SD) – Group 2 3.02(1.72) 2.85(1.73) 2.92(1.63) is more information than I need. M (SD) – Group 3 2.62(1.79) 2.83(1.85) 2.33(1.64) M (SD) – Group 4 2.94(1.93) 2.80(1.94) 2.92(1.83) Information Effectiveness: This message is… Eigenvalue 3.71 3.75 4.00 effective. % of Var. Exp. 74.27 75.06 80.05 detailed. Cronbach’s α .91 .92 .94 useful. M (SD) – All 5.53(1.17) 5.64(1.14) 5.09(1.37) in high quality. M (SD) – Group 1 5.45(1.20) 5.60(1.17) 5.15(1.38) well supported. M (SD) – Group 2 5.64(1.17) 5.60(1.19) 4.91(1.40) M (SD) – Group 3 5.43(1.21) 5.71(1.13) 5.07(1.52) M (SD) – Group 4 5.56(1.10) 5.78(.99) 5.33(1.12) Attitude toward the message: This message is… Eigenvalue 3.96 3.79 4.04 helpful to my health. % of Var. Exp. 79.15 75.86 80.78 a valuable resource. Cronbach’s α .93 .91 .94 important for being healthy. M (SD) – All 5.59(1.22) 5.54(1.23) 5.32(1.37) a good reference for meal preparation. M (SD) – Group 1 5.44(1.39) 5.52(1.19) 5.39(1.35) has something positive to me. M (SD) – Group 2 5.66(1.17) 5.50(1.31) 5.18(1.44) M (SD) – Group 3 5.60(1.11) 5.45(1.35) 5.22(1.49) M (SD) – Group 4 5.77(1.05) 5.74(1.00) 5.53(1.16) 82 Table 7 (cont’d) Variable Name Message 1 Message 2 Message 3 This message is worth sharing with others. % of Var. Exp. 76.35 76.14 77.37 I will recommend this message to others. Cronbach’s α .94 .94 .94 I will press on the "Like" button on Facebook. M (SD) – All 4.03(1.89) 4.10(1.89) 3.83(1.92) I will press on the "Love" button on Facebook. M (SD) – Group 1 3.95(1.88) 3.98(1.92) 3.86(1.92) I will "share" this message on Facebook. M (SD) – Group 2 4.06(1.95) 4.18(1.88) 3.75(1.97) I will "comment" on this message on Facebook. M (SD) – Group 3 4.01(1.88) 3.99(1.87) 3.86(2.01) M (SD) – Group 4 4.19(1.86) 4.31(1.85) 3.92(1.79) Intention to Follow Suggestions from the Message Eigenvalue 4.49 4.55 4.56 I intend to follow what the message suggests in next seven days. % of Var. Exp. 89.85 90.90 91.17 I will try to follow what the message suggests in next seven days. Cronbach’s α .97 .98 .98 I want to follow what the message suggests in next seven days. M (SD) – All 4.85(1.81) 5.07(1.66) 4.88(1.71) I expect to follow what the message suggests in next seven days. M (SD) – Group 1 4.85(1.81) 4.98(1.76) 4.94(1.73) How likely is it that you will follow what the message suggests in next M (SD) – Group 2 5.13(1.54) 5.06(1.62) 4.78(1.74) seven days? M (SD) – Group 3 5.08(1.57) 4.98(1.73) 5.03(1.71) M (SD) – Group 4 5.06(1.43) 5.35(1.47) 4.79(1.67) Perceived Susceptibility of COVID-19 Eigenvalue 2.67 I feel I will get the coronavirus sometime during my life. % of Var. Exp. 89.14 My chances of getting the coronavirus in the next few months are great. Cronbach’s α .94 It is likely that I will get the coronavirus. M (SD) – All 3.12(1.63) M (SD) – Group 1 3.13(1.70) M (SD) – Group 2 3.21(1.62) M (SD) – Group 3 2.90(1.46) M (SD) – Group 4 3.15(1.69) 83 Table 7 (cont’d) Variable Name I felt dizzy, lightheaded, of faint, when I read or listened to news about % of Var. Exp. 81.95 the coronavirus. I had trouble falling or staying asleep because I was thinking about the Cronbach’s α .94 coronavirus. I felt paralyzed or frozen when I thought about or was exposed to M (SD) – All 1.80(1.37) information about the coronavirus. I lost interest in eating when I thought about or was exposed to M (SD) – Group 1 1.76(1.33) information about the coronavirus. I felt nauseous or had stomach problems when I thought about or was M (SD) – Group 2 1.69(1.16) exposed to information about the coronavirus. M (SD) – Group 3 1.76(1.47) M (SD) – Group 4 2.11(1.70) 84 Table 8 Repeated Measures ANCOVA Results for the Effect of Modality Succession on Information Overload Hypothesis Error df F Sig. Partial df ƞ2 Repetition 2 684 .32 .72 .002 Repetition * Condition 4 686 .94 .44 .01 Age 1 343 2.79 .10 .01 Gender 1 343 .49 .48 .001 Race 1 343 43.06 .00 *** .11 Hispanic 1 343 .08 .78 .000 BMI 1 343 .52 .47 .002 Education 1 343 1.76 .19 .01 Marital Status 1 343 4.39 .04 * .01 Dependent 1 343 .12 .73 .000 Employment 1 343 1.33 .25 .004 Income 1 343 2.98 .09 .01 Health Status 1 343 .21 .65 .001 Susceptibility 1 343 .74 .39 .002 Anxiety 1 343 52.82 .00 *** .13 Condition 2 343 .92 .40 .01 Notes: * p < .05; ** p < .01; *** p < .001. Control variables include age, gender, race, BMI, employment, income, marital status, number of dependent, education, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19. 85 Table 9 Repeated Measures ANCOVA Results for the Effect of Modality Succession on Information Effectiveness Hypothesis Error df F Sig. Partial df ƞ2 Repetition 2 343 .23 .80 .001 Repetition * Condition 4 686 2.22 .11 .01 Age 1 343 6.43 .01 * .02 Gender 1 343 .94 .33 .003 Race 1 343 26.87 .00 *** .07 Hispanic 1 343 .00 .97 .000 BMI 1 343 .00 .97 .000 Education 1 343 3.07 .08 .01 Marital Status 1 343 1.01 .32 .003 Dependent 1 343 .54 .46 .002 Employment 1 343 6.64 .01 ** .02 Income 1 343 .09 .77 .000 Health Status 1 343 3.23 .07 .01 Susceptibility 1 343 .24 .62 .001 Anxiety 1 343 .17 .68 .000 Condition 2 343 .49 .61 .003 Notes: * p < .05; ** p < .01; *** p < .001. Control variables include age, gender, race, BMI, employment, income, marital status, number of dependent, education, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19. Because the assumption of sphericity was violated (p < .001), I report the Huynh-Feldt adjusted degree of freedom. 86 Table 10 Repeated Measures ANCOVA Results for the Effect of Modality Succession on Attitude toward Message Hypothesis Error df F Sig. Partial df ƞ2 Repetition 2 686 .69 .50 .002 Repetition * Condition 4 686 1.40 .23 .01 Age 1 343 4.48 .04 .01 Gender 1 343 .17 .68 .000 Race 1 343 17.25 .00 *** .05 Hispanic 1 343 .37 .54 .001 BMI 1 343 .35 .55 .001 Education 1 343 2.06 .15 .01 Marital Status 1 343 .59 .44 .002 Dependent 1 343 .25 .62 .001 Employment 1 343 3.18 .08 .01 Income 1 343 .00 .97 .000 Health Status 1 343 4.57 .03 * .01 Susceptibility 1 343 .49 .48 .001 Anxiety 1 343 .02 .88 .000 Condition 2 343 .35 .71 .002 Notes: * p < .05; ** p < .01; *** p < .001. Control variables include age, gender, race, BMI, employment, income, marital status, number of dependent, education, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19. Because the assumption of sphericity was violated (p < .01), I report the Huynh-Feldt adjusted degree of freedom. 87 Table 11 Repeated Measures ANCOVA Results for the Effect of Modality Succession on VBI Hypothesis Error df F Sig. Partial df ƞ2 Repetition 2 686 .05 .95 .000 Repetition * Condition 4 686 .70 .59 .004 Age 1 343 11.94 .00 *** .03 Gender 1 343 .14 .71 .000 Race 1 343 47.05 .001 *** .12 Hispanic 1 343 .01 .91 .000 BMI 1 343 .95 .33 .003 Education 1 343 .66 .42 .002 Marital Status 1 343 .38 .54 .001 Dependent 1 343 3.54 .06 .01 Employment 1 343 4.32 .04 * .01 Income 1 343 .00 .96 .000 Health Status 1 343 1.25 .26 .004 Susceptibility 1 343 .54 .46 .002 Anxiety 1 343 14.13 .00 *** .04 Condition 2 343 .05 .95 .000 Notes: * p < .05; ** p < .01; *** p < .001. Control variables include age, gender, race, BMI, employment, income, marital status, number of dependent, education, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19. Because the assumption of sphericity was violated (p < .01), I report the Huynh-Feldt adjusted degree of freedom. 88 Table 12 Repeated Measures ANCOVA Results for the Effect of Modality Succession on Intention to Follow Suggestions Hypothesis Error df F Sig. Partial df ƞ2 Repetition 2 686 .47 .62 .001 Repetition * Condition 4 686 1.26 .29 .01 Age 1 343 5.78 .02 * .02 Gender 1 343 .69 .41 .002 Race 1 343 13.33 .00 *** .04 Hispanic 1 343 .21 .65 .001 BMI 1 343 .62 .43 .002 Education 1 343 .08 .78 .000 Marital Status 1 343 1.08 .30 .003 Dependent 1 343 1.66 .20 .01 Employment 1 343 7.29 .01 ** .02 Income 1 343 2.27 .13 .01 Health Status 1 343 5.15 .02 * .02 Susceptibility 1 343 .60 .44 .002 Anxiety 1 343 3.06 .08 .01 Condition 2 343 .18 .83 .001 Notes: * p < .05; ** p < .01; *** p < .001. Control variables include age, gender, race, BMI, employment, income, marital status, number of dependent, education, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19. Because the assumption of sphericity was violated (p < .01), I report the Huynh-Feldt adjusted degree of freedom. 89 Table 13 Bivariate Correlation Coefficients among Variables Variables 1 2 3 4 5 6 7 8 9 10 1 Succession - 2 Information Overload -.03 - 3 Effectiveness .04 .11* - 4 Attitude .04 .01 .90** - 5 VBI .03 .29** .66** .67** - 6 Behavioral Intention .04 .04 .66** .72** .67** - 7 Age .04 -.30** .02 .03 -.04 -.01 - 8 Gender .02 .11 * -.03 -.01 .05 -.02 -.24** - 9 Race .00 -.56** -.28** -.22** -.46** -.25** .30** -.16** - 10 Hispanic .08 -.05 .02 .05 .00 .03 .11* -.04 -.03 - 11 BMI .05 -.30** -.08 -.04 -.22** -.17** .23** -.09 .30** -.02 12 Income -.05 -.28 ** -.09 -.06 -.14** .05 .03 -.09 .34** -.02 13 Employment .08 -.08 -.08 -.05 -.07 -.13* .33** -.18** .03 .01 14 Dependent .00 -.04 -.10 -.08 -.14** -.15** .03 .08 .12* -.03 15 Marital Status -.02 .09 .09 .07 .10 .10 .30** -.22** -.01 .05 16 Education .02 .26 ** .03 .02 .16** .16** -.12* .11* -.35** -.04 17 Health Status -.10* -.05 -.13* -.14** -.13* -.20** .24** -.05 .13* .03 18 Susceptibility -.03 .29** .02 -.03 .12* .03 -.10 -.03 -.18** -.10 19 Anxiety .05 .58** .11* .06 .33** .16** -.29** .03 -.39** -.13* **Correlation is significant at the 0.01 level (2-tailed) *Correlation is significant at the 0.05 level (2-tailed) 90 Table 13 (cont’d) Variables 11 12 13 14 15 16 17 18 19 1 Succession 2 Information Overload 3 Effectiveness 4 Attitude 5 VBI 6 Behavioral Intention 7 Age 8 Gender 9 Race 10 Hispanic 11 BMI - 12 Income .07 - 13 Employment .12 * -.17** - 14 Dependent -.03 -.15** .06 - 15 Marital Status .03 -.07 .25** -.24** - 16 Education -.29** .08 -.11* -.12* .04 - 17 Health Status .34** -.19** .13* .11* .05 -.12* - 18 Susceptibility -.17** -.08 -.12* -.02 -.08 .08 .17** - 19 Anxiety -.33** -.21** -.04 .02 .08 .18** -.06 .44** - **Correlation is significant at the 0.01 level (2-tailed) *Correlation is significant at the 0.05 level (2-tailed) 91 Table 14 Repeated Measures ANCOVA Results for the Effect of Modality Order on Information Overload Hypothesis Error df F Sig. Partial df ƞ2 Repetition 2 103 .08 .92 .002 Repetition * Order 2 103 4.32 .02 * .08 Age 1 104 .61 .44 .01 Gender 1 104 .95 .33 .01 Race 1 104 6.43 .01 * .06 Hispanic 1 104 .10 .75 .001 BMI 1 104 6.64 .01 * .06 Education 1 104 .06 .81 .001 Marital Status 1 104 .39 .53 .004 Dependent 1 104 1.78 .19 .02 Employment 1 104 2.68 .11 .03 Income 1 104 3.34 .07 .03 Health Status 1 104 1.59 .21 .02 Susceptibility 1 104 .06 .80 .001 Anxiety 1 104 28.03 .00 *** .21 Order 1 104 .01 .93 .000 Notes: * p < .05; ** p < .01; *** p < .001. Control variables include age, gender, race, BMI, employment, income, marital status, number of dependent, education, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19. 92 Table 15 Repeated Measures ANCOVA Results for the Effect of Modality Order on Information Effectiveness Hypothesis Error df F Sig. Partial df ƞ2 Repetition 2 103 1.20 .31 * .02 Repetition * Order 2 103 .03 .001 *** .07 Age 1 104 4.77 .03 * .04 Gender 1 104 .01 .91 .000 Race 1 104 12.72 .001 *** .11 Hispanic 1 104 .16 .69 .001 BMI 1 104 .02 .90 .000 Education 1 104 .13 .72 .001 Marital Status 1 104 1.09 .30 .01 Dependent 1 104 .77 .38 .01 Employment 1 104 4.68 .03 * .04 Income 1 104 .31 .58 .003 Health Status 1 104 3.38 .07 .03 Susceptibility 1 104 .08 .78 .001 Anxiety 1 104 .83 .37 .01 Order 1 104 .42 .52 .004 Notes: * p < .05; ** p < .01; *** p < .001. Control variables include age, gender, race, BMI, employment, income, marital status, number of dependent, education, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19. 93 Table 16 Repeated Measures ANCOVA Results for the Effect of Modality Order on Attitude toward Message Hypothesis Error df F Sig. Partial df ƞ2 Repetition 2 208 .65 .53 .01 Repetition * Order 2 208 .04 .97 .000 Age 1 104 2.98 .09 .03 Gender 1 104 .02 .89 .000 Race 1 104 5.28 .024 * .05 Hispanic 1 104 .90 .35 .01 BMI 1 104 .66 .42 .01 Education 1 104 .08 .78 .001 Marital Status 1 104 .03 .86 .000 Dependent 1 104 .60 .44 .01 Employment 1 104 1.01 .32 .01 Income 1 104 .07 .80 .001 Health Status 1 104 2.37 .13 .02 Susceptibility 1 104 .01 .92 .000 Anxiety 1 104 .32 .58 .003 Order 1 104 1.50 .22 .01 Notes: * p < .05; ** p < .01; *** p < .001. Control variables include age, gender, race, BMI, employment, income, marital status, number of dependent, education, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19. Because the assumption of sphericity was violated (p < .01), I report the Huynh-Feldt adjusted degree of freedom. 94 Table 17 Repeated Measures ANCOVA Results for the Effect of Modality Order on VBI Hypothesis Error df F Sig. Partial df ƞ2 Repetition 2 208 2.39 .09 .02 Repetition * Order 2 208 1.35 .26 .01 Age 1 104 3.82 .05 .04 Gender 1 104 .77 .38 .01 Race 1 104 12.89 .001 *** .11 Hispanic 1 104 .62 .43 .01 BMI 1 104 1.24 .27 .01 Education 1 104 2.70 .10 .03 Marital Status 1 104 .00 .97 .000 Dependent 1 104 .70 .41 .01 Employment 1 104 .02 .89 .000 Income 1 104 .04 .85 .000 Health Status 1 104 .02 .88 .000 Susceptibility 1 104 .28 .60 .003 Anxiety 1 104 8.37 .01 ** .07 Order 1 104 .07 .80 .001 Notes: * p < .05; ** p < .01; *** p < .001. Control variables include age, gender, race, BMI, employment, income, marital status, number of dependent, education, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19. Because the assumption of sphericity was violated (p < .05), I report the Huynh-Feldt adjusted degree of freedom. 95 Table 18 Repeated Measures ANCOVA Results for the Effect of Modality Order on Intention to Follow Suggestions Hypothesis Error df F Sig. Partial df ƞ2 Repetition 2 208 .40 .67 .004 Repetition * Order 2 208 2.30 .10 .02 Age 1 104 .45 .51 .004 Gender 1 104 .00 .97 .000 Race 1 104 3.40 .07 .03 Hispanic 1 104 1.78 .19 .02 BMI 1 104 .28 .60 .003 Education 1 104 2.19 .14 .02 Marital Status 1 104 .11 .74 .001 Dependent 1 104 .41 .53 .004 Employment 1 104 .36 .55 .003 Income 1 104 1.63 .20 .02 Health Status 1 104 5.16 .03 * .05 Susceptibility 1 104 .13 .72 .001 Anxiety 1 104 .21 .65 .002 Order 1 104 .18 .67 .002 Notes: * p < .05; ** p < .01; *** p < .001. Control variables include age, gender, race, BMI, employment, income, marital status, number of dependent, education, health status, perceived susceptibility of COVID-19, and anxiety about COVID-19. Because the assumption of sphericity was violated (p < .01), I report the Huynh-Feldt adjusted degree of freedom. 96 Table 19 Serial Mediation Analysis for the Effect of Modality Succession on Intentions to Follow Suggestions (Behavioral Intentions) Mediated by Information Load, Information Effectiveness, Attitudes, and VBI Information Attitudes toward Predictor Information Load (IL) Effectiveness (IE) Message (Am) Coeff (SE) CILL-UL Coeff (SE) CILL-UL Coeff (SE) CILL-UL R (R2) .70 (.49) .37(.14) .90(.81) MSE 1.33 .83 .21 df 15, 343 16,342 17,341 F 21.51 3.45 86.44 p < .001 <.001 <.001 (constant) 3.31(.93) 1.49,5.14 6.89(.75) 5.42,8.36 -.00(.42) -.82,.82 X1 (text-only) .05(.15) -.25,.34 -.04(.12) -.28,.20 -.00(.06) -.12,.12 X2 (succession) -.16(.15) -.46,.14 .07(.12) -.17,.31 -.02(.06) -.14,.10 Inf. Load (IL) -- -- -.05(.04) -.14,.03 -.07(.02)** -.11,-.03 Inf. Effective. (IE) -- -- -- -- .96(.03)*** .90,1.01 Attitudes (Am) -- -- -- -- -- -- VBI -- -- -- -- -- -- BMI -.01(.01) -.03,.01 .00(.01) -.02,.02 .004(.004) -.003,.01 Age -.01(.01) -.03,.01 .01(.01)* .003,.02 -.002(.003) -.01,.004 Race -.35(.05)*** -.46,-.25 -.24(.05)*** -.33,-.15 -.01(.02) -.05,.04 Susceptibility .04(.04) -.05,.13 -.02(.04) -.08,.05 -.01(.02) -.04,.07 Anxiety .42(.06)*** .31,.54 .04(.05) -.06,.14 .02(.03) -.03,.07 Education .07(.05) -.04,.18 -.07(.04) -.16,.01 .01(.02) -.03,.05 Marital Status .17(.08)* .01,.34 .08(.07) -.06,.20 .004(.03) -.06,.07 Dependents .05(.14) -.23,.33 -.08(.11) -.30,.14 .02(.06) -.09,.13 Employment -.05(.04) -.13,.03 -.09(.03)** -.15,-.02 .01(.02) -.02,.05 Income -.08(.04) -.16,.01 -.01(.04) -.08,.05 .003(.02) -.03,.04 Hispanic .09(.31) -.53,.70 -.01(.25) -.49,.48 .18(.12) -.06,.42 Gender .09(.13) -.16,.35 -.10(.10) -.30,.11 .06(.05) -.05,.16 Health .04(.08) -.12,.19 -.11(.06) -.23,.01 -.04(.03) -.10,.03 Notes: * p < .05; ** p < .01; *** p < .001. 97 Table 19 (cont’d) Viral Behavioral Intention Intention to Follow Predictor (VBI) Suggestions (BI) Coeff (SE) CILL-UL Coeff (SE) CILL-UL R (R2) .78(.61) .79(.63) MSE 1.23 .74 df 18,340 19,339 F 28.90 30.43 p <.001 <.001 (constant) -1.02(1.01) -3.02,.97 .61(.79) -.94,2.16 X1 (text-only) .06(.15) -.23,.35 .10(.11) -.12,.33 X2 (succession) -.003(.15) -.29,.35 .02(.11) -.20,.25 Inf. Load (IL) .06(.05) -.05,.16 -.12(.04)** -.20,-.04 Inf. Effective. .15(.14) -.13,.43 -.06(.11) -.28,.16 (IE) Attitudes (Am) .89(.13)*** .63,1.15 .69(.11)*** .48,.91 VBI -- -- .26(.04)*** .18,.34 BMI -.02(.01) -.03,.002 -.01(.01) -.02,.01 Age .02(.01)** .01,.03 .001(.01) -.01,.01 Race -.25(.06)*** -.36,-.14 -.02(.05) -.11,.07 Susceptibility -.02(.04) -.10,.07 -.01(.03) -.07,.06 Anxiety .25(.06) .13,.37 .09(.05) -.01,.18 Education .01(.05) -.09,.11 .08(.04)* .003,.16 Marital Status -.003(.08) -.16,.16 .07(.06) -.06,.19 Dependents -.28(.14)* -.56,-.01 -.07(.11) -.28,.14 Employment -.04(.04) -.12,.04 -.06(.03)* -.12,-.002 Income .01(.04) -.07,.09 .07(.03)* .00,.13 Hispanic -.20(.30) -.79,.40 .07(.23) -.39,.53 Gender .11(.13) -.13,.36 -.10(.10) -.29,.09 Health .03(.08) -.12,.18 -.07(.06) -.19,.04 Notes: * p < .05; ** p < .01; *** p < .001. 98 Table 20 Indirect Effects for Serial Mediation Analysis for the Effect of Modality Succession on Intentions to Follow Suggestions (Behavioral Intentions) Mediated by Information Load, Information Effectiveness, Attitudes, and VBI X1 (text-only) X2 (succession) Indirect Effects Effect BootSE BootCIll-ul Effect BootSE BootCIll-ul MS  IL  BI -.01 .02 -.05,.03 .02 .02 -.02,.06 MS  IE  BI .002 .02 -.04,.05 -.004 .02 -.05,.04 MS  Am  BI -.001 .04 -.09,.08 -.01 .04 -.01,.07 MS  VBI  BI .02 .04 -.06,.10 -.001 .04 -.08,.08 MS  IL  IE  BI .000 .002 -.003,.004 -.001 .002 -.01,.004 MS  IL  Am  BI -.002 .01 -.02,.01 .01 .01 -.01,.03 MS  IL  VBI  BI .001 .003 -.01,.01 -.002 .004 -.01,.004 MS  IE  Am  BI -.03 .09 -.20,.14 .05 .08 -.11,.21 MS  IE  VBI  BI -.002 .01 -.02,.01 .003 .01 -.01,.02 MS  Am  VBI  BI -.000 .01 -.03,.03 -.003 .01 -.03,.02 MS  IL  IE  Am  BI -.002 .01 -.02,.01 .01 .01 -.01,.03 MS  IL  IE  VBI  BI -.000 .001 -.002,.001 .000 .001 -.001,.003 MS  IL  IE  VBI  BI -.000 .001 -.002,.001 .000 .001 -.001,.003 MS  IL  Am  VBI  BI -.01 .03 -.07,.04 .02 .03 -.04,.07 MS  IL  IE  Am  VBI  BI -.001 .003 -.01,.004 .002 .003 -.003,.01 Notes: MS = Modality Succession; IL = Information Load; IE = Information Effectiveness; Am = Attitudes; VBI = Viral Behavioral Intention; BI = Intentions to follow suggestions; * p < .05; ** p < .01; *** p < .001. 99 Table 21 Simple Mediation Analysis for the Effect of Information Load on Attitude toward the Message, Mediated by Information Effectiveness Information Effectiveness Attitude toward Message Predictor (IE) (Am) Coeff (SE) CILL-UL Coeff (SE) CILL-UL R (R2) .38 (.14) .90(.81) MSE .83 .20 df 13, 346 15,343 F 4.35 98.50 p < .001 <.001 (constant) 6.83(.742) 5.37,8.29 .01(.41) -.80,.82 Inf. Load (IL) -.05(.043) -.14,.03 -.07(.02)** -.11, -.03 Inf. Effective. (IE) -- -- .96(.03)*** .94,.10 BMI .001(.01) -.01,.02 .004(.004) -.003,.01 Age .01(.01)* .003,.02 -.002(.003) -.01,.004 Race -.24(.05)*** -.32, -.15 -.01(.02) -.05,.04 Susceptibility -.02(.04) -.09,.05 -.01(.02) -.04,.03 Anxiety .05(.05) -.05,.14 .02(.02) -.03,.07 Education -.07(.04) -.15,.02 .01(.02) -.03,.05 Marital Status .07(.07) -.06,.20 .01(.03) -.06,.07 Dependents -.08(.11) -.30,.14 .02(.06) -.09,.13 Employment -.09(.03)** -.15, -.02 .01(.02) -.02,.05 Income -.02(.04) -.09,.05 .01(.02) -.03,.04 Hispanic .02(.25) -.46,.50 .17(.12) -.07,.41 Gender -.09(.10) -.30,.11 .06(.05) -.05,.16 Health -.12(.06) -.24,.003 -.03(.03) -.09,.03 Notes: * p < .05; ** p < .01; *** p < .001. 100 Table 22 Moderated Mediation Analysis for the Effect Perceived Information Load on Intentions to Follow Suggestions, Moderated by Information Effectiveness and Mediated by Attitudes and VBI Attitudes toward Viral Behavioral Intention to Follow Predictor Message (Am) Intention (VBI) Suggestions (BI) Coeff (SE) CILL-UL Coeff (SE) CILL-UL Coeff (SE) CILL-UL R (R2) .90 (.81) .79(.62) .80(.64) MSE .20 1.17 .71 df 16, 342 18,340 20,338 F 92.87 31.07 30.45 p < .001 <.001 <.001 (constant) .42(.49) -.55,1.38 3.61(1.55) .56,6.66 -.10(1.27) -2.59,2.39 Inf. Load (IL) -.23(.11)* -.44,-.02 -.11(.26) -.62,.41 -.63(.21)** -1.03,-.22 Inf. Effective. (IE) .88(.05)*** .78,.99 -.92(.30)** -1.51,-.32 .21(.25) -.28,.70 Attitudes (Am) -- -- .13(.24) -.34,.60 1.31(.28)*** .79,1.85 VBI -- -- -- -- -.14(.24) -.61,.33 IL x IE .03(.02) -.01,.06 .03(.04) -.06,.12 .08(.04)* .01,.15 Am x IE -- -- .17(.05)*** .08,.26 -.13(.05)* -.24,-.03 VBI x IE -- -- -- -- .07(.04) -.01,.16 BMI .004(.004) -.004,.01 -.02(.01) -.03,.001 -.01(.01) -.02,.01 Age -.001(.003) -.01,.004 .02(.01)* .01,.03 .001(.01) -.01,.01 Race -.005(.02) -.05,.04 -.23(.06)*** -.34,-.12 .03(.04) -.11,.01 Susceptibility -.01(.02) -.04,.03 -.08(.04) -.09,.07 -.004(.03) -.07,.06 Anxiety .02(.02) -.03,.06 .26(.06)*** .15,.38 .08(.05) -.01,.17 Education .005(.02) -.04,.05 .004(.05) -.10,.11 .06(.04) -.02,.14 Marital Status .003(.03) -.06,.07 -.03(.08) -.18,.12 .09(.06) -.03,.21 Dependents .02(.06) -.09,.13 -30(.13)* -.56,-.03 -.07(.11) -.28,.14 Employment .01(.02) -.02,.04 -.05(.04) -.13,.03 -.06(.03)* -.12,-.005 Income .01(.02) -.03,.04 .03(.04) -.13,.03 .07(.03)* .01,.14 Hispanic .18(.12) -.06,.42 -.16(.29) -.74,.42 .08(.23) -.37,.53 Gender .05(.05) -.05,.15 -.09(.12) -.15,.33 -.08(.10) -.27,.11 Health -.03(.03) -.09,.03 .03(.07) -.11,.18 -.06(.06) -.17,.05 Note: Interaction 1 = IL x IE, Interaction 2 = Am x IE, Interaction 3 = VBI x IE; * p < .05; ** p < .01; *** p < .001. 101 Table 23 Indirect Effects of Moderated Mediation Analysis for the Effect Perceived Information Load on Intentions to Follow Suggestions, Moderated by Information Effectiveness and Mediated by Attitudes and VBI Indirect Effects IE Effect BootSE BootCIll-ul IL  Am  BI Low (4.47) -.08* .03 -.15,-.02 Moderate (5.47) -.05* .02 -.08,-.02 High (6.40) -.02* .01 -.05,-.004 IL  VBI  BI Low (4.47) .01 .02 -.03,.04 Moderate (5.47) .02 .02 -.01,.05 High (6.40) .03 .02 -.01,.08 IL  Am  VBI  BI Low (4.47) -.02* .01 -.04,-.002 Moderate (5.47) -.02* .01 -.04,-.01 High (6.40) -.02* .01 -.04,-.01 Notes: IL = Information Load; IE = Information Effectiveness; Am = Attitudes; VBI = Viral Behavioral Intention; BI = Intentions to follow suggestions; * p < .05; ** p < .01; *** p < .001. 102 BIBLIOGRAPHY 103 BIBLIOGRAPHY Ackerson, L. K., & Viswanath, K. (2009). The social context of interpersonal communication and health. Journal of Health Communication, 14(sup1), 5-17. doi:10.1080/10810730902806836 Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In Action control (pp. 11-39): Springer. Ajzen, I. (1991). The theory of planned behavior. Organizational behavior and human decision processes, 50(2), 179-211. Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behaviour. Alhabash, S., Baek, J. H., Cunningham, C., & Hagerstrom, A. (2015). To comment or not to comment?: How virality, arousal level, and commenting behavior on YouTube videos affect civic behavioral intentions. Computers in human behavior, 51, 520-531. Alhabash, S., & Ma, M. (2017). A tale of four platforms: Motivations and uses of Facebook, Twitter, Instagram, and Snapchat among college students?. Social media+ society, 3(1), 2056305117691544. Ammar, A., Brach, M., Trabelsi, K., Chtourou, H., Boukhris, O., Masmoudi, L., ... & Hoekelmann, A. (2020). Effects of COVID-19 home confinement on eating behaviour and physical activity: results of the ECLB-COVID19 international online survey. Nutrients, 12(6), 1583. Anderson, M., & Jiang, J. (2018). Teens, social media & technology overview 2018. 2018. Pew Research Center: Washington DC. URL: http://assets. pewresearch. org/wp‐content/uploads/sites/14/2018/05/31102617/PI_2018, 5. Anderson, A., Porteous, L., Foster, E., Higgins, C., Stead, M., Hetherington, M., . . . Adamson, A. (2005). The impact of a school-based nutrition education intervention on dietary intake and cognitive and attitudinal variables relating to fruits and vegetables. Public health nutrition, 8(6), 650-656. Angelopoulos, P., Milionis, H., Grammatikaki, E., Moschonis, G., & Manios, Y. (2009). Changes in BMI and blood pressure after a school based intervention: the CHILDREN study. European Journal of Public Health, 19(3), 319-325. Armitage, C. J., & Conner, M. (1999a). Distinguishing perceptions of control from self‐efficacy: Predicting consumption of a low‐fat diet using the theory of planned behavior. Journal of applied social psychology, 29(1), 72-90. 104 Armitage, C. J., & Conner, M. (1999b). The theory of planned behavior: Assessment of predictive validity and perceived control. British journal of social psychology, 38(1), 35-54. Aune, D., Ursin, G., & Veierød, M. B. (2009). Meat consumption and the risk of type 2 diabetes: a systematic review and meta-analysis of cohort studies. Diabetologia, 52(11), 2277- 2287. Austin, E. W. (1993). Exploring the effects of active parental mediation of television content. Journal of Broadcasting & Electronic Media, 37(2), 147–158. Backer, T. E., Rogers, E., & Sopory, P. (1992). Designing health communication campaigns: What works? : Sage Publications. Balaji, M. S., & Chakrabarti, D. (2010). Student interactions in online discussion forum: Empirical research from'media richness theory'perspective. Journal of interactive online learning, 9(1). Bandura, A. (1986). Social foundations of thought and action. Englewood Cliffs, NJ, 1986(23- 28). Bandura, A. (2001). Social cognitive theory of mass communication. Media psychology, 3(3), 265-299. Beale, D., & Manstead, A. S. (1991). Predicting mothers' intentions to limit frequency of infants' sugar intake: Testing the theory of planned behavior. Journal of applied social psychology, 21(5), 409-431. Beaudoin, C. E., Fernandez, C., Wall, J. L., & Farley, T. A. (2007). Promoting healthy eating and physical activity: short-term effects of a mass media campaign. American Journal of Preventive Medicine, 32(3), 217-223. Beaulieu, D., & Godin, G. (2012). Staying in school for lunch instead of eating in fast-food restaurants: results of a quasi-experimental study among high-school students. Public health nutrition, 15(12), 2310-2319. Berger, C. R. (2005). Interpersonal communication: Theoretical perspectives, future prospects. Journal of Communication. Bissonnette-Maheux, V., Provencher, V., Lapointe, A., Dugrenier, M., Dumas, A. A., Pluye, P., ... & Desroches, S. (2015). Exploring women’s beliefs and perceptions about healthy eating blogs: A qualitative study. Journal of medical Internet research, 17(4), e87. Bland, J. M., & Altman, D. G. (1997). Statistics notes: Cronbach's alpha. Bmj, 314(7080), 572. Boschen, M. J., & Casey, L. M. (2008). The use of mobile telephones as adjuncts to cognitive behavioral psychotherapy. Professional Psychology: Research and Practice, 39(5), 546. 105 Bontcheva, K., Gorrell, G., & Wessels, B. (2013). Social media and information overload: Survey results. arXiv preprint arXiv:1306.0813. Braithwaite, D. O., & Schrodt, P. (2014). Engaging theories in interpersonal communication: Multiple perspectives: Sage Publications. Brinker, D. L., Gastil, J., & Richards, R. C. (2015). Inspiring and informing citizens online: A media richness analysis of varied civic education modalities. Journal of Computer- Mediated Communication, 20(5), 504-519. Brown, J. D., & Witherspoon, E. M. (2002). The mass media and American adolescents’ health. Journal of adolescent health, 31(6), 153-170. Boulos, M. N. K., Resch, B., Crowley, D. N., Breslin, J. G., Sohn, G., Burtner, R., ... & Chuang, K. Y. S. (2011). Crowdsourcing, citizen sensing and sensor web technologies for public and environmental health surveillance and crisis management: trends, OGC standards and application examples. International journal of health geographics, 10(1), 1-29. Buhi, E. R., Trudnak, T. E., Martinasek, M. P., Oberne, A. B., Fuhrmann, H. J., & McDermott, R. J. (2013). Mobile phone-based behavioural interventions for health: A systematic review. Health Education Journal, 72(5), 564-583. Buijzen, M., Schuurman, J., & Bomhof, E. (2008). Associations between children's television advertising exposure and their food consumption patterns: A household diary–survey study. Appetite, 50(2-3), 231-239. Burgess, J., & Green, J. (2018). YouTube: Online video and participatory culture. John Wiley & Sons. Cacioppo, J. T., & Petty, R. E. (1979). Effects of message repetition and position on cognitive response, recall, and persuasion. Journal of personality and social psychology, 37(1), 97- 109. doi:10.1037/0022-3514.37.1.97 Cacioppo, J. T., Petty, R. E., Feinstein, J. A., & Jarvis, W. B. G. (1996). Dispositional differences in cognitive motivation: The life and times of individuals varying in need for cognition. Psychological Bulletin, 119(2), 197. Cairns, G., Angus, K., Hastings, G., & World Health Organization. (2009). The extent, nature and effects of food promotion to children: a review of the evidence to December 2008. World Health Organization. Carr, C. T., & Hayes, R. A. (2015). Social media: Defining, developing, and divining. Atlantic journal of communication, 23(1), 46-65. CDC. (2010). Number of Americans with diabetes projected to double or triple by 2050. Retrieved from https://www.cdc.gov/media/pressrel/2010/r101022.html 106 CDC. (2020). About Adult BMI. Retreived from https://www.cdc.gov/healthyweight/assessing/bmi/adult_bmi/index.html Champion, V. L. (1999). Revised susceptibility, benefits, and barriers scale for mammography screening. Research in nursing & health, 22(4), 341-348. Chan, A. K., Nickson, C. P., Rudolph, J. W., Lee, A., & Joynt, G. M. (2020). Social media for rapid knowledge dissemination: early experience from the COVID‐19 pandemic. Chang, H. H., & Nayga, R. M. (2009). Television viewing, fast-food consumption, and children's obesity. Contemporary Economic Policy, 27(3), 293-307. doi:http://dx.doi.org/10.1111/j.1465-7287.2009.00157.x Chang, H. H., & Nayga, R. M. (2010). Childhood obesity and unhappiness: The influence of soft drinks and fast food consumption. Journal of Happiness Studies, 11(3), 261-275. doi:http://dx.doi.org/10.1007/s10902-009-9139-4 Chin, A., Keelan, J., Tomlinson, G., Pavri-Garcia, V., Wilson, K., & Chignell, M. (2010). Automated delineation of subgroups in web video: A medical activism case study. Journal of Computer-Mediated Communication, 15(3), 447-464. Chou, W. Y. S., Hunt, Y. M., Beckjord, E. B., Moser, R. P., & Hesse, B. W. (2009). Social media use in the United States: implications for health communication. Journal of medical Internet research, 11(4), e48. Coates, A. E., Hardman, C. A., Halford, J. C., Christiansen, P., & Boyland, E. J. (2019). Social media influencer marketing and children’s food intake: a randomized trial. Pediatrics, 143(4). Cohen, R. J., Swerdlik, M. E., & Phillips, S. M. (1996). Psychological testing and assessment: An introduction to tests and measurement: Mayfield Publishing Co. Conner, M., & Norman, P. (2005). Predicting health behaviour: McGraw-Hill Education (UK). Conner, M., Norman, P., & Bell, R. (2002). The theory of planned behavior and healthy eating. Health psychology, 21(2), 194. Craig, W. J. (2009). Health effects of vegan diets. The American Journal of Clinical Nutrition, 89, 1627S-1633S. doi:10.3945/ajcn.2009.26736N Crutzen, R., & De Nooijer, J. (2011). Intervening via chat: an opportunity for adolescents’ mental health promotion?. Health promotion international, 26(2), 238-243. Daft, R. L., & Lengel, R. H. (1983). Information richness. A new approach to managerial behavior and organization design. Daft, R. L., & Lengel, R. H. (1986). Organizational information requirements, media richness and structural design. Management science, 32(5), 554-571. 107 Daft, R. L., Lengel, R. H., & Trevino, L. K. (1987). Message equivocality, media selection, and manager performance: Implications for information systems. MIS quarterly, 355-366. Dailey, R. M., McCracken, A. A., & Romo, L. K. (2011). Confirmation and weight management: Predicting effective levels of acceptance and challenge in weight management messages. Communication Monographs, 78(2), 185-211. doi:10.1080/03637751.2011.564638 Dailey, R. M., Richards, A. A., & Romo, L. K. (2010). Communication with significant others about weight management: The role of confirmation in weight management attitudes and behaviors. Communication Research, 37(5), 644-673. Davison, K. K., & Birch, L. L. (2001). Childhood overweight: a contextual model and recommendations for future research. Obesity Reviews, 2(3), 159-171. doi:10.1046/j.1467-789x.2001.00036.x Dawson, J. (2010). Doctors join patients in going online for health information. New Media Age, 7. Dennis, A. R., & Kinney, S. T. (1998). Testing media richness theory in the new media: The effects of cues, feedback, and task equivocality. Information systems research, 9(3), 256- 274. Department of Health. (2010) Social Media Action Plan. Department of Health, State Government of Victoria, Melbourne. DeVellis, R. F. (2016). Scale development: Theory and applications (Vol. 26): Sage publications. Divecha, Z., Divney, A., Ickovics, J., & Kershaw, T. (2012). Tweeting about testing: Do low‐income, parenting adolescents and young adults use new media technologies to communicate about sexual health?. Perspectives on sexual and reproductive health, 44(3), 176-183. Dixon, H. G., Scully, M. L., Wakefield, M. A., White, V. M., & Crawford, D. A. (2007). The effects of television advertisements for junk food versus nutritious food on children's food attitudes and preferences. Social science & medicine, 65(7), 1311-1323. Donelle, L., & Booth, R. G. (2012). Health tweets: an exploration of health promotion on twitter. Online J Issues Nurs, 17(4). Dunlop, S., Freeman, B., & Jones, S. C. (2016). Marketing to youth in the digital age: The promotion of unhealthy products and health promoting behaviours on social media. Media and Communication, 4(3), 35-49. Eppler, M. J., & Mengis, J. (2004). The concept of information overload: A review of literature from organization science, accounting, marketing, MIS, and related disciplines. Information Society, 20, 325-344. 108 Epton, T., Norman, P., Harris, P., Webb, T., Snowsill, F. A., & Sheeran, P. (2015). Development of theory-based health messages: three-phase programme of formative research. Health promotion international, 30(3), 756-768. Evers, C. W., Albury, K., Byron, P., & Crawford, K. (2013). Young people, social media, social network sites and sexual health communication in Australia:" This is funny, you should watch it". International Journal of Communication, 7, 18. Falk, L. W., Sobal, J., Bisogni, C. A., Connors, M., & Devine, C. M. (2001). Managing Healthy Eating: Definitions, Classifications, and Strategies. Health education & behavior, 28(4), 425-439. Retrieved from https://journals.sagepub.com/doi/abs/10.1177/109019810102800405. doi:10.1177/109019810102800405 Fernandez, A. A., & Shaw, G. P. (2020). Academic leadership in a time of crisis: The coronavirus and COVID‐19. Journal of Leadership Studies, 14(1), 39-45. Fishbein, M., & Ajzen, I. (1977). Belief, attitude, intention, and behavior: An introduction to theory and research. Flora, J., & Schooler, C. (1995). Influence of health communication environments on children's diet and exercise knowledge, attitudes, and behavior. Communication and health outcomes, 187-213. Flora, J. A., Saphir, M. N., Schooler, C., & Rimal, R. N. (1997). Toward a framework for intervention channels: Reach, involvement, and impact. Annals of epidemiology, 7(7), S104-S112. Fox, S. (2011). The social life of health information, 2011 (pp. 1-33). California Healthcare Foundation. Freberg, K. (2016). Social media. In C. Carroll (Ed.), Encyclopedia for corporate reputation (pp. 773–776). Thousand Oaks, CA: SAGE. Freberg, K. (2018). Social media for strategic communication: Creative strategies and research- based applications. Sage Publications. Freeland-Graves, J. H., & Nitzke, S. (2013). Position of the academy of nutrition and dietetics: total diet approach to healthy eating. Journal of the Academy of Nutrition and Dietetics, 113(2), 307-317. Fryar, C.D., Carroll, M.D., Afful, J. (2020) Prevalence of overweight, obesity, and severe obesity among children and adolescents aged 2–19 years: United States, 1963–1965 through 2017–2018. NCHS Health E-Stats. Fu, S., Li, H., Liu, Y., Pirkkalainen, H., & Salo, M. (2020). Social media overload, exhaustion, and use discontinuance: Examining the effects of information overload, system feature overload, and social overload. Information Processing & Management, 57(6), 102307. 109 Fulk, J., Schmitz, J., & Steinfield, C. W. (1990). A social influence model of technology use. Organizations and communication technology, 117, 140. Fuller, C. M., Simmering, M. J., Atinc, G., Atinc, Y., & Babin, B. J. (2016). Common methods variance detection in business research. Journal of Business Research, 69(8), 3192-3198. George, K. S., Roberts, C. B., Beasley, S., Fox, M., Rashied-Henry, K., & Brooklyn Partnership to Drive Down Diabetes (BP3D). (2016). Our health is in our hands: a social marketing campaign to combat obesity and diabetes. American Journal of Health Promotion, 30(4), 283-286. Gerber, B. S., Stolley, M. R., Thompson, A. L., Sharp, L. K., & Fitzgibbon, M. L. (2009). Mobile phone text messaging to promote healthy behaviors and weight loss maintenance: a feasibility study. Health informatics journal, 15(1), 17-25. Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1994). Growing up with television: The cultivation perspective. Gilliland, J., Sadler, R., Clark, A., O'Connor, C., Milczarek, M., & Doherty, S. (2015). Using a smartphone application to promote healthy dietary behaviours and local food consumption. BioMed Research International, 2015, 11. doi:10.1155/2015/841368 Giustini, D. (2006). How Web 2.0 is changing medicine. Retrieved from https://www.bmj.com/content/333/7582/1283?ijke__= Glanz, K. (2015). Health behavior: Theory, research, and practice: John Wiley & Sons. Gratton, L., Povey, R., & Clark‐Carter, D. (2007). Promoting children's fruit and vegetable consumption: interventions using the Theory of Planned Behaviour as a framework. British journal of health psychology, 12(4), 639-650. Green, B., & Hope, A. (2010). Promoting clinical competence using social media. Nurse educator, 35(3), 127-129. Griffith, T. L., & Neale, M. A. (2001). 8. Information processing in traditional, hybrid, and virtual teams: From nascent knowledge to transactive memory. Research in organizational behavior, 23, 379-421. Grineva, M., & Grinev, M. (2012, April). Information overload in social media streams and the approaches to solve it. In 21st International World Wide Web Conference, Lyon, France. Guse, K., Levine, D., Martins, S., Lira, A., Gaarde, J., Westmorland, W., & Gilliam, M. (2012). Interventions using new digital media to improve adolescent sexual health: a systematic review. Journal of adolescent health, 51(6), 535-543. Hales, C.M., Carroll, M.D., Fryar, C.D., Ogden, C.L. (2020). Prevalence of obesity and severe obesity among adults: United States, 2017–2018. NCHS Data Brief, no 360. Hyattsville, MD: National Center for Health Statistics. 110 Harris, J. L., Bargh, J. A., & Brownell, K. D. (2009). Priming effects of television food advertising on eating behavior. Health psychology, 28(4), 404. Harrison, A. A. (1977). Mere exposure. In Advances in experimental social psychology (Vol. 10, pp. 39-83): Elsevier. Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford publications. Hayes, A. (2021). The PROCESS macro for SPSS, SAS, and R. Retrieved from https://www.processmacro.org/download.html. Head, K. J., Noar, S. M., Iannarino, N. T., & Harrington, N. G. (2013). Efficacy of text messaging-based interventions for health promotion: a meta-analysis. Social science & medicine, 97, 41-48. Heidelberger, C. A. (2011). Health care professionals’ use of online social networks. Dakota State University. Henderson, J., Coveney, J., Ward, P., & Taylor, A. (2009). Governing childhood obesity: Framing regulation of fast food advertising in the Australian print media. Social science & medicine, 69(9), 1402-1408. doi:http://dx.doi.org/10.1016/j.socscimed.2009.08.025 Hornik, R. (2002). Public health communication: Evidence for behavior change: Routledge. House, J., Su, J., & Levy-Milne, R. (2006). Definitions of healthy eating among university students. Canadian Journal of Dietetic Practice and Research, 67(1), 14-18. Hu, L., & Bentler, P. M. (1998). Fit indices in covariance structure modeling: Sensitivity to underparameterized model misspecification. Psychological Methods, 3, 424–453. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55 Huo, C., Zhang, M., & Ma, F. (2018). Factors influencing people’s health knowledge adoption in social media: the mediating effect of trust and the moderating effect of health threat. Library Hi Tech. Ishii, K., Lyons, M. M., & Carr, S. A. (2019). Revisiting media richness theory for today and future. Human Behavior and Emerging Technologies, 1(2), 124-131. Jemmott III, J. B., Jemmott, L. S., O’Leary, A., Ngwane, Z., Icard, L., Bellamy, S., . . . Tyler, J. C. (2011). Cognitive-behavioural health-promotion intervention increases fruit and vegetable consumption and physical activity among South African adolescents: a cluster- randomised controlled trial. Psychology and Health, 26(2), 167-185. Karimi-Shahanjarini, A., Rashidian, A., Omidvar, N., & Majdzadeh, R. (2013). Assessing and comparing the short-term effects of TPB only and TPB plus implementation intentions 111 interventions on snacking behavior in Iranian adolescent girls: a cluster randomized trial. American Journal of Health Promotion, 27(3), 152-161. Keelan, J., Pavri-Garcia, V., Tomlinson, G., & Wilson, K. (2007). YouTube as a source of information on immunization: a content analysis. jama, 298(21), 2482-2484. Kelly, B., Halford, J. C., Boyland, E. J., Chapman, K., Bautista-Castaño, I., Berg, C., ... & Summerbell, C. (2010). Television food advertising to children: a global perspective. American journal of public health, 100(9), 1730-1736. Key, T.J., Appleby, P.N., & Rosell, M.S. (2006). Health effects of vegetarian and vegan diets. Proceedings of the Nutrition Society, 65, 35-41. doi:10.1079/PNS2005481 Kinney, S. T., & Watson, R. T. (1992). The effect of medium and task on dyadic communication (Doctoral dissertation, University of Georgia). Kleinberg, J. (2008). The convergence of social and technological networks. Communications of ACM, 51(11), 66–72. Knibbs, K. (2013). Slactivists, unite! Social media campaigns aren’t just feel-good back patting. Digital Trends. Kontos, E. Z., Emmons, K. M., Puleo, E., & Viswanath, K. (2010). Communication inequalities and public health implications of adult social networking site use in the United States. Journal of health communication, 15(sup3), 216-235. Korda, H., & Itani, Z. (2013). Harnessing social media for health promotion and behavior change. Health promotion practice, 14(1), 15-23. Retrieved from https://journals.sagepub.com/doi/abs/10.1177/1524839911405850 Kothe, E. J., Mullan, B., & Butow, P. (2012). Promoting fruit and vegetable consumption. Testing an intervention based on the theory of planned behaviour. Appetite, 58(3), 997- 1004. Kothe, E. J., Mullan, B. A., & Amaratunga, R. (2011). Randomised controlled trial of a brief theory-based intervention promoting breakfast consumption. Appetite, 56(1), 148-155. Kukreja, P., Sheehan, A. H., & Riggins, J. (2011). Use of social media by pharmacy preceptors. American journal of pharmaceutical education, 75(9). Kunst, A. (2020). Products/topics talked about online in the U.S. 2020. Statista. Retrieved from https://www.statista.com/forecasts/997053/products-topics-talked-about-online-in-the-us Kratzke, C., & Cox, C. (2012). Smartphone technology and apps: rapidly changing health promotion. Global Journal of Health Education and Promotion, 15(1). 112 Krishnan, A., & Zhou, X. (2019). Modeling the effect of health antecedents and social media engagement on healthy eating and quality of life. Journal of Applied Communication Research, 47(4), 365-380. Lai, J. Y., & Chang, C. Y. (2011). User attitudes toward dedicated e‐book readers for reading: The effects of convenience, compatibility and media richness. Online information review. Lan, Y. F., & Sie, Y. S. (2010). Using RSS to support mobile learning based on media richness theory. computers & education, 55(2), 723-732. Lan, Y. F., Hung, C. L., & Hsu, H. J. (2011). Effects of Guided Writing Strategies on Students' Writing Attitudes Based on Media Richness Theory. Turkish Online Journal of Educational Technology-TOJET, 10(4), 148-164. Lange, P. G. (2007). Publicly private and privately public: Social networking on YouTube. Journal of computer-mediated communication, 13(1), 361-380. Lappalainen, R., Kearney, J., & Gibney, M. (1998). A pan EU survey of consumer attitudes to food, nutrition and health: an overview. Food Quality and Preference, 9(6), 467-478. Lariscy, R. W., Avery, E. J., Sweetser, K. D., & Howes, P. (2009). An examination of the role of online social media in journalists’ source mix. Public Relations Review, 35(3), 314–316. DOI: https://doi.org/10.1016/j.pubrev.2009.05.008. Lee, S. A. (2020). Coronavirus Anxiety Scale: A brief mental health screener for COVID-19 related anxiety. Death studies, 44(7), 393-401. Lent, M., Hill, T. F., Dollahite, J. S., Wolfe, W. S., & Dickin, K. L. (2012). Healthy children, healthy families: Parents making a difference! A curriculum integrating key nutrition, physical activity, and parenting practices to help prevent childhood obesity. Journal of Nutrition Education and Behavior, 44(1), 90-92. doi:http://dx.doi.org/10.1016/j.jneb.2011.02.011 Levac, J. J., & O'Sullivan, T. (2010). Social media and its use in health promotion. Revue interdisciplinaire des sciences de la santé-Interdisciplinary Journal of Health Sciences, 1(1), 47-53. Lewis, M. K., & Hill, A. J. (1998). Food advertising on British children’s television: a content analysis and experimental study with nine-year olds. International journal of obesity, 22(3), 206-214. Lichtenstein, A. H., Appel, L. J., Brands, M., Carnethon, M., Daniels, S., Franch, H. A., . . . Wylie-Rosett, J. (2006). Diet and lifestyle recommendations revision 2006: A scientific statement from the american heart association nutrition committee. Circulation, 114(1), 82-96. Retrieved from https://search.proquest.com/docview/17263548?accountid=12598 Liu, D., & Yang, C. C. (2016). Media niche of electronic communication channels in friendship: A meta-analysis. Journal of Computer-Mediated Communication, 21(6), 451-466. 113 "Lockdown eating insights: Has the pandemic turned us into a nation of snackers?" (2020). Retrieved from https://covid.joinzoe.com/us-post/lockdown-snacking Louis, W., Davies, S., Smith, J., & Terry, D. (2007). Pizza and Pop and the Student Identity: The Role of Referent Group Norms in Healthy and Unhealthy Eating. The Journal of Social Psychology, 147(1), 57-74. Lu, Y., Kim, Y., Dou, X. Y., & Kumar, S. (2014). Promote physical activity among college students: Using media richness and interactivity in web design. Computers in Human Behavior, 41, 40-50. Ly, K. H., Carlbring, P., & Andersson, G. (2012). Behavioral activation-based guided self-help treatment administered through a smartphone application: study protocol for a randomized controlled trial. Trials, 13(1), 62. Mangold, W. G., & Faulds, D. J. (2009). Social media: The new hybrid element of the promotion mix. Business Horizons, 52, 357–365. Marmot, M., Atinmo, T., Byers, T., Chen, J., Hirohata, T., Jackson, A., . . . & Zeisel, S. (2007). Food, nutrition, physical activity, and the prevention of cancer: a global perspective. Retrieved from http://discovery.ucl.ac.uk/4841/ McNab, C. (2009). What social media offers to health professionals and citizens. World Health Organization. Retrieved from https://www.scielosp.org/article/bwho/2009.v87n8/566- 566/en/. McQuail, D. (2010). McQuail's mass communication theory: Sage publications. Mendelsohn, H. (1973). Some reasons why information campaigns can succeed. Public Opinion Quarterly, 37(1), 50-61. Mejova, Y., Haddadi, H., Noulas, A., & Weber, I. (2015). # foodporn: Obesity patterns in culinary interactions. In Proceedings of the 5th International Conference on Digital Health 2015 (pp. 51-58). Miller-Day, M., & Marks, J. D. (2006). Perceptions of Parental Communication Orientation, Perfectionism, and Disordered Eating Behaviors of Sons and Daughters. Health Communication, 19(2), 153–163. doi:http://dx.doi.org.proxy2.cl.msu.edu/10.1207/s15327027hc1902_7 Millimet, D. L., Tchernis, R., & Husain, M. (2010). School nutrition programs and the incidence of childhood obesity. Journal of human resources, 45(3), 640-654. Retrieved from https://search.proquest.com/docview/816376337?accountid=12598 Moorhead, S. A., Hazlett, D. E., Harrison, L., Carroll, J. K., Irwin, A., & Hoving, C. (2013). A new dimension of health care: systematic review of the uses, benefits, and limitations of social media for health communication. Journal of medical Internet research, 15(4), e85. 114 Napolitano, M. A., Hayes, S., Bennett, G. G., Ives, A. K., & Foster, G. D. (2013). Using Facebook and text messaging to deliver a weight loss program to college students. Obesity, 21(1), 25-31. Neuhauser, L., & Kreps, G. L. (2003). Rethinking communication in the e-health era. Journal of health psychology, 8(1), 7-23. doi:10.1177/1359105303008001426 Noar, S. M. (2006). A 10-year retrospective of research in health mass media campaigns: Where do we go from here? Journal of Health Communication, 11(1), 21-42. doi:10.1080/10810730500461059 Norman, J., Kelly, B., Boyland, E., & McMahon, A. T. (2016). The impact of marketing and advertising on food behaviours: evaluating the evidence for a causal relationship. Current Nutrition Reports, 5(3), 139-149. O'Keefe, D. J. (2002). Persuasion: Theory and research (Vol. 2): Sage. O'Keefe, D. J. (2003). Message properties, mediating states, and manipulation checks: Claims, evidence, and data analysis in experimental persuasive message effects research. Communication Theory, 13(3), 251-274. O'Keefe, D. J. (2013). The elaboration likelihood model. The SAGE handbook of persuasion: Developments in theory and practice, 137-149. O'Mara, B. (2013). Social media, digital video and health promotion in a culturally and linguistically diverse Australia. Health Promotion International, 28(3), 466-476. Pagoto, S., Schneider, K. L., Evans, M., Waring, M. E., Appelhans, B., Busch, A. M., ... & Ziedonis, M. (2014). Tweeting it off: characteristics of adults who tweet about a weight loss attempt. Journal of the American Medical Informatics Association, 21(6), 1032- 1037. Pallant, J. (2020). SPSS survival manual: A step by step guide to data analysis using IBM SPSS. Routledge. Parchman, M. L., Flannagan, D., Ferrer, R. L., & Matamoras, M. (2009). Communication competence, self-care behaviors and glucose control in patients with type 2 diabetes. Patient education and counseling, 77(1), 55-59. Paisley, C., Lloyd, H., Sparks, P., & Mela, D. (1995). Consumer perceptions of dietary changes for reducing fat intake. Nutrition Research, 15(12), 1755-1766. Paisley, C. M., & Sparks, P. (1998). Expectations of reducing fat intake: The role of perceived need within the theory of planned behavior. Psychology and Health, 13(2), 341-353. Pearce, K. J. (2009). Media and mass communication theories. Encyclopedia of Communication Theory. Thousand Oaks, 623-627. 115 Pettegrew, L. S., & Day, C. (2015). Smart phones and mediated relationships: The changing face of relational communication. Review of Communication, 15(2), 122-139. Petty, R. E., Cacioppo, J. T., & Haugtvedt, C. (1992). Ego-involvement and persuasion: An appreciative look at the Sherifs’ contribution to the study of self-relevance and attitude change. Social judgment and intergroup relations: Essays in honor of Muzafer Sherif, 147-174. Petty, R. E., Cacioppo, J. T., & Schumann, D. (1983). Central and peripheral routes to advertising effectiveness: The moderating role of involvement. Journal of Consumer Research, 10(2), 135-146. Petty, R. E., & Wegener, D. T. (1999). The elaboration likelihood model; Current status and controversies. Dual-process theories in social psychology, 41-72. Phillips, L. W., & Sternthal, B. (1977). Age differences in information processing: a perspective on the aged consumer. Journal of Marketing Research, 14(4), 444-457. Povey, R., Conner, M., Sparks, P., James, R., & Shepherd, R. (1998). Interpretations of healthy and unhealthy eating, and implications for dietary change. Health education research, 13(2), 171-183. Prelip, M., Slusser, W., Thai, C. L., Kinsler, J., & Erausquin, J. T. (2011). Effects of a school‐based nutrition program diffused throughout a large urban community on attitudes, beliefs, and behaviors related to fruit and vegetable consumption. Journal of school health, 81(9), 520-529. Raats, M. M., Shepherd, R., & Sparks, P. (1995). Including Moral Dimensions of Choice Within the Structure of the Theory of Planned Behavior1. Journal of applied social psychology, 25(6), 484-494. Ray, M. L., Sawyer, A. G., & Strong, E. C. (1971). Frequency effects revisited. Journal of Advertising Research, 11(1), 14-20. Reardon, K. K., & Rogers, E. M. (1988). Interpersonal versus mass media communication a false dichotomy. Human communication research, 15(2), 284-303. Reger, B., Wootan, M. G., & Booth-Butterfield, S. (1999). Using mass media to promote healthy eating: a community-based demonstration project. Preventive Medicine, 29(5), 414-421. Riesch, S. K., Anderson, L. S., & Krueger, H. A. (2006). Parent-Child Communication Processes: Preventing Children’s Health-Risk Behavior. Journal for Specialists in Pediatric Nursing, 11(1), 41–56. Rodriguez, M. G., Gummadi, K., & Schoelkopf, B. (2014). Quantifying information overload in social media and its impact on social contagions. In Proceedings of the International AAAI Conference on Web and Social Media (Vol. 8, No. 1). 116 Rogers, E. M., & Storey, J. D. (1987). Communication campaigns. Sadeghirad, B., Duhaney, T., Motaghipisheh, S., Campbell, N. R. C., & Johnston, B. C. (2016). Influence of unhealthy food and beverage marketing on children's dietary intake and preference: a systematic review and meta‐analysis of randomized trials. Obesity Reviews, 17(10), 945-959. Scanfeld, D., Scanfeld, V., & Larson, E. L. (2010). Dissemination of health information through social networks: Twitter and antibiotics. American journal of infection control, 38(3), 182-188. Schiavo, R. (2013). Health communication: From theory to practice (Vol. 217): John Wiley & Sons. Sharma, N., & Patterson, P. G. (1999). The impact of communication effectiveness and service quality on relationship commitment in consumer, professional services. Journal of services marketing, 13(2), 151-170. Sheer, V. C. (2010). Hong Kong adolescents' use of MSN vs. ICQ for developing friendships online: Considering media richness and presentational control. Chinese Journal of Communication, 3(2), 223-240. Sheer, V. C. (2011). Teenagers' use of MSN features, discussion topics, and online friendship development: The impact of media richness and communication control. Communication Quarterly, 59(1), 82-103. Sidor, A., & Rzymski, P. (2020). Dietary choices and habits during COVID-19 lockdown: experience from Poland. Nutrients, 12(6), 1657. Smith, M., Morita, H., Mateo, K. F., Nye, A., Hutchinson, C., & Cohall, A. T. (2014). Development of a culturally relevant consumer health information website for Harlem, New York. Health Promotion Practice, 15(5), 664-674. Smith, A. (2013). Smartphone ownership–2013 update. Pew Research Center: Washington DC, 12, 2013. Sparks, P., Hedderley, D., & Shepherd, R. (1992). An investigation into the relationship between perceived control, attitude variability and the consumption of two common foods. European Journal of Social Psychology, 22(1), 55-71. Sparks, P., & Shepherd, R. (1992). Self-identity and the theory of planned behavior: Assesing the role of identification with" green consumerism". Social psychology quarterly, 388-399. Spiegel, S. A., & Foulk, D. (2006). Reducing overweight through a multidisciplinary school‐based intervention. Obesity, 14(1), 88-96. Statista. (2019). Global social media account ownership from 2013 to 2018. Retrieved from https://www.statista.com/statistics/788084/number-of-social-media-accounts/ 117 Statista. (2020). Social media usage in the United States. Retrieved from https://www.statista.com/topics/3196/social-media-usage-in-the-united-states/ Statista. (2021a). U.S. daily media usage time via mobile 2018-2020. Retrieved from https://www.statista.com/statistics/469983/time-spent-mobile-media-type-usa/ Statista. (2021b). Social Media Advertising. Statista. Retrieved from https://www.statista.com/outlook/amo/advertising/social-media-advertising/spain Stephens, K. K. (2007). The successive use of information and communication technologies at work. Communication Theory, 17(4), 486-507. Stephens, K. K., & Rains, S. A. (2011). Information and communication technology sequences and message repetition in interpersonal interaction. Communication Research, 38(1), 101-122. Stephens, K. K., Sørnes, J. O., Rice, R. E., Browning, L. D., & Sætre, A. S. (2008). Discrete, sequential, and follow-up use of information and communication technology by experienced ICT users. Management Communication Quarterly, 22(2), 197-231. Strauss, A. (1945). The concept of attitude in social psychology. The Journal of Psychology, 19(2), 329-339. Sun, Y., Lim, K. H., Jiang, C., Peng, J. Z., & Chen, X. (2010). Do males and females think in the same way? An empirical investigation on the gender differences in Web advertising evaluation. Computers in Human Behavior, 26(6), 1614-1624. Tavakol, M., & Dennick, R. (2011). Making sense of Cronbach's alpha. International journal of medical education, 2, 53. Thøgersen-Ntoumani, C., & Ntoumanis, N. (2007). A self-determination theory approach to the study of body image concerns, self-presentation and self-perceptions in a sample of aerobic instructors. Journal of Health Psychology, 12(2), 301-315. Tilman, D., & Clark, M. (2014). Global diets link environmental sustainability and human health. Nature, 515, 518-522. doi:10.1038/nature13959 Tobey, L. N., & Manore, M. M. (2014). Social media and nutrition education: the food hero experience. Journal of nutrition education and behavior, 46(2), 128-133. Towler, G., & Shepherd, R. (1992). Modification of Fishbein and Ajzen's theory of reasoned action to predict chip consumption. Food Quality and Preference, 3(1), 37-45. Tsorbatzoudis, H. (2005). Evaluation of a planned behavior theory-based intervention programme to promote healthy eating. Perceptual and motor skills, 101(2), 587-604. 118 U.S. Department of Health and Human Services. (2013). Managing Overweight and Obesity in Adults: Systematic Evidence Review from the Obesity Expert Panel. Retrieved from https://www.nhlbi.nih.gov/health-topics/managing-overweight-obesity-in-adults Valacich, J. S., Mennecke, B. E., Wachter, R. M., & Wheeler, B. C. (1994). Extensions to media richness theory: A test of the task-media fit hypothesis. In Proceedings of the 27th Hawaii International Conference on System Sciences (HICSS-27). Part 4 (of 5) (pp. 11- 20). Publ by IEEE. Wang, Y., & Beydoun, M. A. (2009). Meat consumption is associated with obesity and central obesity among US adults. International Journal of Obesity, 33, 621-8. doi:10.1038/ijo.2009.45. Warshaw, P. R., & Davis, F. D. (1985). Disentangling behavioral intention and behavioral expectation. Journal of experimental social psychology, 21(3), 213-228. Waters, R. D., Burnett, E., Lamm, A., & Lucas, J. (2009). Engaging stakeholders through social networking: How nonprofit organizations are using Facebook. Public Relations Review, 35, 102–106. Walther, J. B. (2011). Theories of computer-mediated communication and interpersonal relations. The handbook of interpersonal communication, 4, 443-479. Watanabe-Ito, M., Kishi, E., & Shimizu, Y. (2020). Promoting Healthy Eating Habits for College Students Through Creating Dietary Diaries via a Smartphone App and Social Media Interaction: Online Survey Study. JMIR mHealth and uHealth, 8(3), e17613. Waters, R. D., & Jones, P. M. (2011). Using video to build an organization's identity and brand: A content analysis of nonprofit organizations' YouTube videos. Journal of Nonprofit & Public Sector Marketing, 23(3), 248-268. Watt, T. T., Appel, L., Roberts, K., Flores, B., & Morris, S. (2013). Sugar, stress, and the supplemental nutrition assistance program: Early childhood obesity risks among a clinic- based sample of low-income hispanics. Journal of Community Health, 38(3), 513-520. doi:http://dx.doi.org/10.1007/s10900-012-9641-1 Westerman, D., Van Der Heide, B., Klein, K. A., & Walther, J. B. (2008). How do people really seek information about others?: Information seeking across Internet and traditional communication channels. Journal of Computer-Mediated Communication, 13(3), 751- 767. Weymann, N., Härter, M., & Dirmaier, J. (2015). Quality of online information on type 2 diabetes: a cross-sectional study. Health promotion international, 30(4), 821-831. World Health Organation. (2021). #HealthyAtHome: Healthy Diet. Retrieved from https://www.who.int/campaigns/connecting-the-world-to-combat- coronavirus/healthyathome/healthyathome---healthy- diet#:~:text=While%20no%20foods%20or%20dietary,and%20more%20active%20lives 119 Wunsch, N. (2020). Top social media sites for food information among U.S consumers 2017. https://www.statista.com/statistics/922632/social-media-food-information-source-us- consumer/ Yzer, M. (2013). Reasoned action theory. The SAGE handbook of persuasion: Developments in theory and practice, 120-136. Zajonc, R. B., & Rajecki, D. W. (1969). Exposure and affect: A field experiment. Psychonomic Science, 17(4), 216-217. Zhou, Z., Su, C., Zhou, N., & Zhang, N. (2016). Becoming friends in online brand communities: evidence from China. Journal of computer-mediated communication, 21(1), 69-86. Zhu, Y. (2019). Combinatorial use of communication technologies in organizations. Corporate Communications: An International Journal. 120