University of Southern California Dissertations and Theses

Developing and testing a heuristic-systematic model of health decision making: the role of affect, trust, confidence and media influence

(USC Thesis Other)

Developing and testing a heuristic-systematic model of health decision making: the role of affect, trust, confidence and media influence

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content DEVELOPING AND TESTING A HEURISTIC-SYSTEMATIC MODEL OF HEALTH DECISION MAKING: THE ROLE OF AFFECT, TRUST, CONFIDENCE AND MEDIA INFLUENCE Copyright 2013 By Nien-Tsu Nancy Chen A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (COMMUNICATION) August 13, 2013 Nien-Tsu Nancy Chen 11 TABLE OF CONTENTS ACKNOWLEDGEMENTS ......................................................................................................... vi ABSTRACT ................................................................................................................................. viii INTRODUCTION .......................................................................................................................... 1 CHAPTER 1: HEALTH DECISION MAKING IN CONTEMPORARY SOCIETY ............ 4 Contemporary Perspectives on Decision Making: Bounded Rationality, Heuristics, and Dual- Process Models ............................................................................................................................ 7 Theories of Health Decision Making and Health Behavior Change .......................................... 15 CHAPTER 2: FLU VACCINATION AS AN EXEMPLIFICATION OF THE CONUNDRUM OF HEALTH DECISION MAKING IN MODERN TIMES ...................... 20 Seasonal Influenza Vaccine: A Vaccine for All with Uncertain Efficacy ................................. 21 H1N1 Swine Influenza Vaccine: Historical Roots for Vaccine Safety Concern ....................... 25 H5N1 Avian Influenza Vaccine: Will It Prevent the Next Pandemic? ...................................... 29 CHAPTER 3: SOCIO-PSYCHOLOGICAL THEORIES OF HEALTH BEHAVIOR CHANGE AND THEIR APPLICATION IN THE CONTEXT OF FLU VACCINATION 32 Health Belief Model.. ................................................................................................................. 33 Theory of Reasoned Action and Theory of Planned Behavior .................................................. 38 An Integrated Model of Health Behavioral Prediction .............................................................. 42 CHAPTER 4: INCORPORATING AFFECT, TRUST, CONFIDENCE AND MEDIA INFLUENCE INTO SOCIO-PSYCHOLOGICAL THEORIES OF HEALTH BEHAVIOR CHANGE. ..................................................................................................................................... 50 Integral Affect in Judgment Making .......................................................................................... 52 Social Trust and Social Confidence in Shaping Risk Perceptions and Health Behaviors ......... 60 Media Influence over Health Decision Making ......................................................................... 74 CHAPTER 5: RESEARCH DESIGN ........................................................................................ 92 Procedures .................................................................................................................................. 92 Participants ................................................................................................................................. 93 Measures .................................................................................................................................... 94 Analysis ................................................................................................................................... 104 CHAPTER 6: RESEARCH RESULTS ................................................................................... 110 Modeling H1N1 Flu Vaccination Intention in 2009 ................................................................ 110 Modeling Seasonal Flu Vaccination Behavior in 2009 ........................................................... 116 111 Modeling H5Nl Flu Vaccination Intention in 2012 ................................................................ 122 Correlates for Seasonal Flu Vaccination Behavior in 2012 ..................................................... 129 CHAPTER 7: DISCUSSION AND CONCLUSION ............................................................... 131 REFERENCES ........................................................................................................................... 140 APPENDIX A: ITEMS FROM ANHCS SURVEY FIELDED IN 2009 ............................... 158 Core Questions ......................................................................................................................... 158 Module Questions .................................................................................................................... 159 APPENDIX B: ITEMS FROM ANHCS SURVEY FIELDED IN 2012 ............................... 161 Core Questions ......................................................................................................................... 161 Module Questions .................................................................................................................... 162 IV LIST OF TABLES Table I. Socio-demographic Characteristics of Samples Compared to US Population .. 94 Table 2. Endogenous Variables from the 2009 and 2012 Surveys and Their Means and Standard Deviations .......................................................................................................... 95 Table 3. Polychoric Correlations among Variables from the 2009 Survey .................... 108 Table 4. Polychoric Correlations among Variables from the 2012 Survey .................... 109 Table 5. Standardized Gamma Estimates for Significant Paths Predicting H1N1 Flu Vaccination Intention in 2009 ......................................................................................... 115 Table 6. Standardized Gamma Estimates for Significant Paths Predicting Seasonal Flu Vaccination Behavior in 2009 ........................................................................................ 121 Table 7. Standardized Gamma Estimates for Significant Paths Predicting H5N1 Flu Vaccination Intention in 2012 ......................................................................................... 128 Table 8. Estimates (Odds and Log-odds and) from Logistic Regression Analysis of Correlates of Seasonal Flu Vaccination Behavior in 2012. N ~ 584 ............................. 130 v LIST OF FIGURES Figure 1. Health Belief Model components and linkages ................................................ 35 Figure 2. Theory of Planned Behavior components and linkages .................................... 40 Figure 3. An Integrative Model of Behavioral Prediction components and linkages ...... 44 Figure 4. Affect in predicting H5Nl flu vaccination intention in 2012 ........................... 60 Figure 5. Technology Acceptance Model components and linkages ............................... 70 Figure 6. Social trust in predicting H5Nl flu vaccination intention in 2012 ................... 73 Figure 7. Social confidence in predicting HlNl flu vaccination intention in 2009 ......... 73 Figure 8. Social confidence in predicting seasonal flu vaccination behavior in 2009 ..... 73 Figure 9. Influence of passive media exposure over HlNl flu vaccination intention in 2009 ................................................................................................................................... 89 Figure 10. Influence of passive media exposure over seasonal flu vaccination behavior in 2009 ................................................................................................................................... 89 Figure 11. Influence of active attention to HlNl news over HlNl flu vaccination intention in 2009 ............................................................................................................... 90 Figure 12. Influence of active attention to HlNl news over seasonal flu vaccination behavior in 2009 ............................................................................................................... 90 Figure 13. Influence of passive media exposure over H5Nl flu vaccination intention in 2012 ................................................................................................................................... 91 Figure 14. Standardized beta estimates for significant paths predicting HlNl flu vaccination intention in 2009 .......................................................................................... 114 Figure 15. Standardized beta estimates for significant paths predicting seasonal flu vaccine behavior in 2009 ................................................................................................ 120 Figure 16. Standardized beta estimates for significant paths predicting H5Nl flu vaccination intention in 2012 .......................................................................................... 127 VI ACKNOWLEDGEMENTS I owe a debt of gratitude to the great minds and wonderful friends I have encountered during the past six years at the USC Annenberg School for Communication & Journalism. My thanks go out to my dissertation committee members, including: • My advisor, Michael Cody, for his advice on how to navigate academia, for being always supportive to my research interests, and for showing me the qualities and dedication that make a great educator • Sheila Murphy, for introducing me to the wonderful world of research methods and for providing invaluable guidance while I worked on developing instruments for my dissertation and other projects • Doe Mayer, for opening my eyes to a bottom-up, participatory approach to research and for encouraging me to critically examine my pre-existing assumptions about communication practices I wish to convey special thanks to Peter Monge and Poong Oh for their advice on data analysis, and to Tom Goodnight for introducing me to sociological perspectives that inform my inquiry. Katya, Amanda, Nina and Yujung, you make up the best peer support group one could have while trying to survive and enjoy the dissertation writing process. Beyond dissertation research, I want to acknowledge other sources of mentorship that have guided my intellectual journey. Sandra Ball-Rokeach, your commitment to grounded research and to finding theory-based solutions to address social challenges is inspirational. Thank you for your vision. Michael Parks and other members of the Alhambra Project team, our weekly conversations have allowed me to stay current about Vll innovations, opportunities and challenges in journalism. Past and current members of the Metamorphosis Project team, thank you for being great partners to bounce ideas off and to collaborate with on various projects. Much appreciation also goes out to my former boss and colleagues at the Taiwan Centers for Disease Control -you helped ignite my passion for health communication and motivated me to pursue doctoral education in this area. Last but not least, I would like to acknowledge the bottomless love and support from my family and friends in Taiwan, Australia, the United States and other parts of the world. Special thanks to my parents who have always given me the freedom to choose my own path, make mistakes and be an independent spirit. V111 ABSTRACT This project aims to extend major socio-psychological theories of health behavior change to take into account the more intuitive, heuristic and affect-laden route of decision making. Despite extensive research showing that human beings tend to perceive the world and make judgments based on two different and interacting routes of information processing, established health behavior theories have focused on the systematic route at the expense of the heuristic route. Based on a review of literature from the fields of communication, psychology, sociology and risk management, this research identifies integral affect, social trust, social confidence, and media influence as important forces that shape the heuristic, and often also the systematic, process of health decision making. An extended, dual-process model of health decision making is subsequently proposed. In order to test this model, data were collected from a national probability sample of adults living in the United States regarding their decisions to vaccinate against three flus (i.e. seasonal influenza, the HlNl swine flu and the H5Nl bird flu). The results suggest that integral affect, social trust and social confidence are powerful predictors of vaccination decisions, influencing intention or behavior not only directly and but also indirectly by coloring individuals' beliefs regarding vaccine and in the case of integral affect, also their perceptions regarding the vaccine-preventable flu. The analysis also indicates considerable media influence over individuals' health perceptions and decisions. However, it is active attention to health content in the media rather than passive exposure that seems to function as the primary mechanism behind these media effects. 1 INTRODUCTION This work proposes and empirically tests a heuristic-systematic model of health decision making. Extensive research from the fields of communication, psychology and risk management has indicated that individuals tend to make decisions through two different and interacting routes of information processing, one systematic and rule-based, and the other heuristic and affect-laden. However, established health behavior change theories tend to focus on the systematic route at the expense of the heuristic route. The time is ripe to investigate the relevance of the dual-mode perspective to health decision making. As a first step in theory extension, this project suggests the inclusion of integral affect (i.e. affect targeted at a particular stimulus), social trust, social confidence, and media influence in established health behavior models. Psychologists, sociologists, risk analysts and communication scholars have identified these variables as important heuristic forces in shaping judgments and decisions (Damasio, 1994; Earle & Siegrist, 2008; Earle, Siegrist, & Gutscher, 2007; Forgas, 1995; Shrum, 2009; Siegrist, Cousin, Kastenholz, & Wiek, 2007; Walsh-Childers & Brown, 2009). Using vaccination against three different flus (i.e. the HlNl swine flu, seasonal influenza and a hypothetical H5Nl bird flu) as case studies, this extended model's predictive power in different contexts is empirically investigated, and the model's implications for practice and future research are subsequently discussed. Chapter 1 provides an overview of contemporary perspectives on decision making from the disciplines of sociology, psychology, economics, communication, risk 2 management and health. Key premises and concepts, including risk, active trust, reflexivity, bounded rationality, heuristic, and dual-process models of information processing are described. This chapter also provides a classification of health decision making theories based on their level of analysis, statistic or dynamic nature, and focus on intention versus behavior enactment. Acknowledging the value of these different theoretical orientations, justification is provided for this project's focus on the socio psychological determinants of vaccination behavior. This introductory chapter sets up the context for more detailed literature reviews in subsequent chapters. Chapter 2 articulates why vaccination constitutes a suitable candidate behavior for testing the extended model. It also traces the history and controversy surrounding the development of the three flu vaccines studied. Furthermore, the epidemiology of the flu against which each vaccine protects is described. Chapter 3 reviews major socio-psychological models of health behavior change. Their respective premises are presented, and findings from studies applying these theories in predicting flu vaccination behavior are summarized. In addition, these theories are compared and contrasted in terms of their strengths and weaknesses, and several hypotheses were advanced in an effort to investigate discrepancies across different theories and to clarify certain causal relationships within a theory. Chapter 4 offers the rationales for including integral affect, social trust, social confidence, and media influence in an extended heuristic-systematic model of health decision making. Theory and research from the fields of psychology, sociology, communication and risk management that explicate the role of these variables in influencing the heuristic and systematic routes of information processing are presented. Based on this review, hypotheses and research questions are put forward regarding how these variables should be incorporated into existing models of health behavior change. 3 Chapter 5 describes the empirical component of this project. It outlines the survey instruments and procedures used to gathered data, the operationalization of predictor and outcome variables, characteristics of the study participants, and the structural equation modeling techniques and hierarchical logistic regression analysis used to test the hypotheses and investigate the research questions. Chapter 6 reports findings from the statistical analyses. Integral affect, social trust, social confidence, and media coverage were found to have indirect effects and sometimes also a direct impact on flu vaccination intention and behavior. The theoretical and practical implications of these research findings are discussed in Chapter 7, which also addresses the study's limitations and directions for future research. 4 CHAPTER 1: HEALTH DECISION MAKING IN CONTEMPORARY SOCIETY In describing the current epoch, prominent sociologists have coined concepts such as "risk society," "second modernity" and "reflexive modernity" (Beck, 1992a, 1992b, 1999; Beck, Giddens, & Lash, 1994; Beck & Grande, 2010; Giddens, 1990, 1994). Cutting across those concepts is an acknowledgement that despite many regional and national differences, contemporary societies are characterized by risk and uncertainty, which are largely the unintended side effects of rapid industrial and technological development since the 17th century. Such development ushered in the age of modernity by transforming the principles and institutions of previous modes of social organization. However, "discontinuous transformations within modernity" (Beck & Grande, 2010, p. 414) have been increasingly noted by scholars since the 1990s, leading to calls to reconsider the present epoch as a reflexive second modernity, where the basic institutions of an earlier modernity come under increasing scrutiny and challenge. It is within this socio-historical context that the following theoretical and empirical exploration of health decision making takes place. Therefore, a brief overview of the unique conditions that characterize the reflexivity oftoday's world is presented first. As Giddens (1990, p. 21) argues, modernity is characterized by "the 'lifting out' of social relations from local contexts of interaction and their restructuring across indefinite spans of time-space." Different from previous modes of social life, new communication and transportation technologies have enabled individuals living in diverse time zones and locales to coordinate their social and economic activities. This coordinating capability, together with the demand of industrial capitalism, drove 5 individuals to contributing their specialized knowledge to the production or delivery of socioeconomic goods without having to acquire a more general, comprehensive skillset to meet their own everyday needs. As specialization in the division of labor and social differentiation grew, it was no longer feasible for people to manage their lives based on information obtained from face-to-face interactions with others and from their local experiences, as many social interactions now occurred remotely and virtually. Instead, individuals came to entrust abstract systems and modern institutions, such as government agencies, scientific communities and industrial groups, in making key decisions on their behalf(Giddens, 1990; Renn & Levine, 1991). Trust in those institutions, in Luhmann's (2000, p. 103) words, offers a "structural reduction of complexity" in modern life. Instead of trying to acquire and assess the multitude of information needed for making decisions in every domain of life, individuals simply expected that their basic needs would be met through the decisions made by those with access to expert knowledge. Since the middle of the 20th century, however, a "chain of publicly revealed catastrophes, near-catastrophes, whitewashed security faults and scandals" (Beck, 1992b, p. 104) have led to a growing awareness of the unintended consequences of scientific industrial development, which was once assumed to produce certain and exclusively positive outcomes (e.g. endless economic growth for the nation-state and a more secured, affluent life for individuals). For example, events such as global warming and other ecological disasters caused by industrial pollution and meltdowns of nuclear power plants built to sustain economic growth have exposed the weaknesses of modern institutions in managing such development and in meeting individuals' essential concerns for health and 6 safety. The public is increasingly cognizant of the political nature of institutional decision making, as expert disagreements over alternative modes of scientific inquiry, data interpretation and problem-solving approach are made increasingly visible to the public through the mass media and new communication platforms. It has also become apparent that even the best-trained scientists are confronted with many "unknowns" and "unknownables" in their inquiries (Beck, 1999). As a result, institutional decision making that favors one solution over another is always based on some extra-scientific criteria and is never politically 'neutral,' and the lay public is increasingly mindful of this complexity. As witnessed over the past several decades, many institutional decisions have led to new risks that were unexpected and unmanageable by the decision maker. Such shortcomings, brought home to lay individuals through personal and mediated experience, have made them increasingly reflexive (Giddens, 1990). Reflexive individuals neither take for granted the assumptions and promises of modernity, nor trust the ability of modern institutions in delivering social functions indiscriminately. Rather, they assume more personal responsibility in dealing with risk that permeates their environment. Aided with access to exponentially growing sources of news and information, they constantly scrutinize the activities of institutions and alternative social movements. Based on such information, individuals actively choose which perspective and social agent to trust in making decisions consequential to their lives, and they continue to revise those choices based on new, incoming information (Giddens, 1998). In other words, trust in institutions has become critical and conditional in the current epoch. 7 To summarize, we live in the age of risk society because risk, as the aggregation of unintended consequences of scientific-industrial development, has gained centrality in our collective consciousness and become the lens through which we see our environment (Jaeger, Renn, Rosa, & Wehler, 2001). By throwing the assumptions and institutions of modernity into flux, risk has necessitated the distinction between a first modernity and a second modernity. At the individual level, a wholesale faith in experts and scientific industrial development has been replaced by active trust in modern institutions, where people constantly reevaluate the effectiveness and trustworthiness of governmental, industrial and scientific organizations as well as the decisions they espouse. At the macro level, risk has "become a defining concept in public and political debate" that is largely fought out in the mass media, where "[p ]ressure groups seek to attract media attention in their campaigns for safety measures, experts complain of media 'scare-mongering,' industries and government bodies employ special 'risk communicators' in an attempt to maintain (or woo) public confidence, and journalists themselves describe the attractions of scientific controversy and risk disputes" (Eldridge & Reilly, 2003, p. 138). It is under this context of risk, uncertainty, active trust and reflexivity that our discussion of health decision making now takes place. Contemporary Perspectives on Decision Making: Bounded Rationality, Heuristics, and Dual-Process Models Judgment and decision making are an integral part of everyday life, and every decision made is consequential for an individual in meeting their immediate or long-term goals. As a result, researchers in diverse disciplines, ranging from economics to 8 psychology to health to risk management, have sought to understand how and why people make particular choices under certain circumstances. Since the middle of the 20th century, there has been increasing cross-fertilization among disciplines (Kahneman, 1991; Payne & Bettman, 2007), with scholars from all fields giving increasing acknowledgement to the ideas of bounded rationality and satisficing (Simon, 1955, 1956, 1982, 2000), heuristics and biases (Gilovich, Griffin, & Kahneman, 2002; Kahneman, Slovic, & Tversky, 1982; Kahneman & Tversky, 1973; Tversky & Kahneman, 1974) and the dual-process theories of thinking, knowing, information processing and judgment making (Chaiken, 1980; Epstein, 1994; Kahneman & Frederick, 2002; Petty & Cacioppo, 1986; Sloman, 1996, 2002; Stanovich & West, 2002). Economist Herbert A. Simon's works on judgment making under constraints have been influential since the 1950s, when he advanced a critique of the "unbounded" rationality underpinning traditional economic theory. According to this traditional approach: [An individual] is assumed to have knowledge of the relevant aspects of his environment which, if not absolutely complete, is at least impressively clear and voluminous. He is also assumed to have a well-organized and stable system of preferences, and a skill in computation that enables him to calculate, for the alternative courses of action that are available to him, which of these will permit him to reach the highest attainable point on his preference scale. (Simon, 1955, p. 99) 9 However, Simon (1955, 1956, 1982) has effectively demonstrated that in reality, judgments and decisions tend to be made under both external constraints (e. g. limited access to relevant information in the environment or time constraint) and internal constraints (e.g. limited cognitive capabilities in processing and evaluating incoming information). Consequently, rather than searching for an alternative that would lead to the maximum level of expected utility, individuals tend to engage in "satisficing," where they settle on a decision with a satisfactory level of expected utility without expending too much time and energy in reaching that decision. The process of satisficing requires a conceptualization of rationality in human judgment that is bounded, situational, procedural and instrumental (Over, 2007; Payne & Bettman, 2007). According to Over (2007), this type of rationality refers to having reliable means to achieve goals, with both the means and the goals varying across individuals and contexts. In other words, this definition of rationality considers not only the end result but also the process of decision making (Simon, 1978, 2000). A thought process is considered rational in so far as it functions as an effective instrument for realizing one's goal in a particular context. Simon's work has shaped much of the research on human information processing and judgment making since the 1950s, including the seminal work on heuristics and biases by Amos Tversky and Daniel Kahneman (1974). Those psychologists' research has demonstrated how boundedly rational individuals use heuristics as an effective mechanism in achieving their goals. Although some of the earliest definitions of 'heuristic' can be found in the computer science and problem-solving literature (Simon & Newell, 1958), the concept gained traction from Tversky and Kahneman's research beginning in the 1970s. 10 Heuristics are commonly defined as short-cuts that an individual, organization or machine uses to solve problems or attain goals (Keren & Teigen, 2007). Heuristics are typically contrasted with algorithms in the case of artificial intelligence or normative rules of utility maximization in the case of human decision making. These are "explicit and detailed rules that guarantee a correct result, but could be effortful and time consuming, and hence impractical in situations characterized by limited cognitive resources" (Keren & Teigen, 2007, p. 4 ). Given the limited computational capacities and incomplete access to relevant information of humans or machines, heuristics become a necessary tool for reaching a satisfactory decision when it comes to problem-solving and goal attainment. Tversky and Kahneman coined three canonical heuristics in their early works, which are mainly concerned with probability and frequency judgments by individuals in situations where uncertainty abounds and a 'correct' answer may not be available. In articulating the impetus behind their research and the significance of what has come to be known as the 'heuristics and biases program,' Tversky and Kahneman (1973) state: Most important decisions men make are governed by beliefs concerning the likelihood of unique events. The "true" probabilities of such events are elusive, since they cannot be assessed objectively. The subjective probabilities that are assigned to unique events by knowledgeable and consistent people have been accepted as all that can be said about the likelihood of such events. Although the 11 "true" probability of a unique event is unknowable, the reliance on heuristics ... biases subjective probabilities in knowable ways. A psychological analysis of the heuristics that a person uses in judging the probability of an event may tell us whether his judgment is likely to be too high or too low. We believe that such analyses could be used to reduce the prevalence of errors in human judgment under uncertainty. (p. 231) In other words, the employment of a particular heuristic typically leads to judgments that are skewed in a specific way. Through several dozen studies, Kahneman and Tversky have identified representativeness, availability, and anchoring and adjustment as three heuristics commonly employed by both lay individuals and experts in making judgments (Kahneman & Tversky, 1972, 1973; Tversky & Kahneman, 1971, 1973, 1974). The representativeness heuristic refers to individuals' tendency to judge the likelihood of an event based on the extent to which it bears similarity to the essential features of the process or population from which it originated. The availability heuristic is employed when a person estimates the frequency or probability of an event by the ease with which relevant instances or associations can be brought to mind. The anchoring and adjustment heuristic refers to people's tendency to be influenced by a reference value offered to them either implicitly or explicitly, such that they begin with this 'anchor' value and make subsequent adjustments to reach their own estimate of the frequency, probability, value or magnitude of an event. Tversky and Kahneman point out that while those heuristics are adaptive and ecologically valid under certain circumstances, their employment tends to slant our judgment in one way rather than another and result in systematic biases. 12 In recent years, Kahneman has moved from identifying commonplace heuristics to developing a theory of heuristics in judgment making (Kahneman & Frederick, 2002, 2005). Kahneman and Frederick (2002, p. 53) argue that the employment of most heuristics results from the process of attribute substitution, where "[ w ]hen confronted with a difficult question people often answer an easier one instead, usually being unaware of the substitution ... We will say that a judgment is mediated by a heuristic when an individual assesses a specified target attribute of a judgment object by substituting another property of that object- the heuristic attribute- which comes more readily to mind." In proposing the mechanism of attribute substitution, Kahneman and Frederick (2002) suggest that the "affect heuristic" (Slovic, Finucane, Peters, & MacGregor, 2007) should replace anchoring and adjustment as a general purpose heuristic because the latter does not work through the substitution of one attribute for another. They define affect as the specific quality of' goodness' or 'badness' experienced "as a feeling state (with or without consciousness) and ... demarcating a positive or negative quality of a stimulus" (Slovic, Finucane, Peters, & MacGregor, 2002, p. 397). They suggest that reliance on such feelings constitutes the employment of the affective heuristic in judgment making. Despite decades of research demonstrating the primacy of affect in conditioning people's preferences and its independent- albeit often conjoint- operation from cognition (Murphy & Zajonc, 1993; Zajonc, 1980, 1984), Slovic et al. (2002) point out that major 13 theories on judgment making across disciplines have continued to focus on the cognitive processes at the expense of the affective mechanism. The incorporation of affect into decision making theories is long overdue considered the consensus among most social and cognitive psychologists that people perceive and judge in two different modes (E. R. Smith & DeCoster, 2000). The first mode has been variously labeled as the intuitive (Tversky & Kahneman, 1983), heuristic (Chaiken, 1980), peripheral (Petty & Cacioppo, 1986), experiential (Epstein, 1994) or associative (Sloman, 2002) system, or simply System 1 (Stanovich & West, 2002). The second mode has been given names such as the extensional (Tversky & Kahneman, 1983), systematic (Chaiken, 1980), central (Petty & Cacioppo, 1986), analytical (Epstein, 1994) or rule-based (Sloman, 2002) system, or simply System 2 (Stanovich & West, 2002). In reviewing the hypotheses of dual-mode models, Smith and DeCoster (2000) find much agreement over the characterization of the two modes. One mode of thinking is considered to draw on associations that are structured by global similarity and congruity and that are learned through repeated experiences. This mode of thinking occurs pre-consciously, retrieving not only properties of the judgment target but also well-learned affective response toward the target automatically. As a result, individuals are aware of the outcome but not the process of their judgment activity, and this outcome is often strongly shaped by affect. The other model of thinking is perceived to draw on symbolically represented rules that are structured by language and logic and that can be learned in just one or few experiences. This mode of thinking occurs optionally when individuals have adequate cognitive capacity and motivation, and they are often consciously aware of the steps and logic leading to their judgment outcome. 14 Under this general framework, Smith and DeCoster (2000) acknowledge that many models include theoretical constructs and postulates that are specific to a particular topic area (e.g. persuasion and attitude change, information processing and judgment making, impression formation and stereotypical perception) in psychology. For example, the Elaboration Likelihood Model (Petty & Cacioppo, 1986) and Heuristic Systematic Model (Chaiken, 1980) were developed to encapsulate information processing and subsequent judgment making when individuals are exposed to persuasive messages, whereas the two-system models of reasoning proposed by Sloman (2002) and Stanovich and West (2002) are primarily concerned with judgment making in everyday contexts, where a persuasive attempt may not exist. While the latter type of theory provides a better fit with the more general scope of discussion in the present project, relevant hypotheses from dual-process models developed in the persuasion context are also drawn upon to advance certain arguments. An even greater difference exists when one considers various dual-mode models' assumptions on the temporal and logical relations between the two processing routes. Some theories consider the two modes as alternatives and argue that individuals process information either one way or the other but not both (M. B. Brewer, 1988; Fazio, 1986). Other models conceptualize a sequential relation, where associative processing occurs automatically and rule-based processing can optionally follow to correct or override the fast conclusions drawn initially (Devine, 1989; Gilbert, 1989; Stanovich & West, 2002; Wegener & Petty, 1995). Still other models propose that that both types of processing occur simultaneously (Chaiken, 1980; Petty & Cacioppo, 1986; Sloman, 1996), sometimes adding to and sometimes cancelling out each other's effects on judgment making. 15 While research findings on the temporal relations between the two modes of thinking remain inconclusive, there is far more theoretical and empirical support for some type of interaction between the associative and rule-based processing than for treating them as either-or alternatives. Furthermore, there is an agreement among those advocating an interactive approach that associative processing is automatic so it always occurs when a judgment is called for. However, rule-based processing can follow and either strengthen or the overturn the initial judgment. Although some argue that the reshaping of the initial judgment occurs through many iterative exchanges between automatic and systematic processing, these multiple iterations would be difficult to untangle and model using the analytical tools available for the present project. Therefore, this project adopts the perspective of a sequential relationship between these two thinking modes as advocated by Stanovich and West (2002), while acknowledging the need for future research to explore the potentially iterative interaction between the two modes. Theories of Health Decision Making and Health Behavior Change There is no shortage of theories and models on health decision-making and behavior change (Glanz, Rimer, & Viswanath, 2008; Noar & Zimmerman, 2005). Recognizing the inadequacy of socio-demographic determinants in accounting for individual differences in health behaviors, researchers have devoted considerable effort to identifying other contributing factors and processes in health decision making since the 1940s. Over the decades, a multitude of models have been proposed and empirically tested (Cummings, Becker, & Maile, 1980; Noar & Zimmerman, 2005). They can be crudely categorized by their level of analysis (Glanz et al., 2008), static or dynamic nature (Prochaska, Redding, & Evers, 2008), and focus on intention formation vis-a-vis behavior enactment (Armitage & Conner, 2000). 16 To elaborate, some theories focus exclusively on the intrapersonal, interpersonal, group, community or institutional level of analysis, whereas others consider more than one level of influence or take on an ecological perspective (Glanz et al., 2008). For example, theories such as the Health Belief Model (HBM) (Janz & Becker, 1984; Rosenstock, 1974) and Protection Motivation Theory (PMT) (R. W. Rogers, 1975, 1983) are concerned exclusively with intrapersonal psychological predictors of health behaviors. The Theory of Planned Behavior (TPB) (Ajzen, 1991), a general behavioral prediction model that has been widely applied to the health context, has a similar psychological orientation. To assess interpersonal or group influence over health decision making, the social networks and social support approaches are frequently utilized by researchers (Heaney & Israel, 2008). For community-level analysis of health behaviors, researchers have resorted to theories such as the Diffusion of Innovations (E. M. Rogers, 2003) mass media effects (Finnegan & Viswanath, 2008). While each individual theory offers valuable insights, there is growing awareness that the complex and interactive forces shaping health behaviors in real life are best captured by ecological models, which "incorporate constructs from models that focus on 17 psychological, social, and organizational levels ofinfluence ... along with consideration of environments and policy in the broader conununity" (Sallis, Owen, & Fisher, 2008, pp. 466-467). Despite their recognized values, ecological models have not been widely validated or utilized in empirical research, as multilevel studies require far more resources to conduct and present far more methodological challenges to the investigators than single level studies (Sallis et a!., 2008). Another direction for theoretical integration focuses on the process of behavior change and the utility of different theories in predicting decisions at various stages of change. These theories are dynamic rather than static because they conceptualize behavior change as taking place over time rather than as a discrete event. For example, the Transtheoretical Model (TTM) of Health Behavior Change (Prochaska & V elicer, 1997) posits that people progress through the six stages of"precontemplation," "contemplation," "preparation," "action," "maintenance" and ''termination" in order to successfully adopt a new behavior, even though the progression is not always linear, inevitable or irreversible. To understand the progression from the precontemplation stage (i.e. not thinking about behavior change at all) to the contemplation stage (i.e. thinking about changing behaviors in the near term), theories like HBM, PMT and TPB might be useful because they articulate the psychological conditions (e.g. the pros and cons of adopting a preventive behavior) that need to be met for individuals to feel motivated to change behaviors. On the other hand, social support and resources in one's social networks might be critical to help individuals move from the action stage (i.e. having made recent modifications in behavior) to the maintenance stage (i.e. preserving with 18 behavior modifications and working to prevent relapse). As Prochaska et al. (2008, p. 103) argue, "[n]o single theory can account for all complexities of behavior change. A more comprehensive model is most likely to emerge from integration across major theories." Arguably, stage theories like the TTM are more appropriate for understanding behavior modifications that are relatively complex and need to be sustained over time (e.g. abstaining from smoking or engaging in regular exercise) compared to behaviors that are simple and one-time in nature (e.g. getting a flu shot). Up till this point, health decision making theories and health behavior change theories have been used interchangeably. However, as made clear by the stage theories and ecological perspective, behavior modifications often result from multiple decisions made at different stages of change, and the final decision to adopt a health behavior is not always translated into action due to environmental constraints. However, as Conner and colleagues (Armitage & Conner, 2000; Conner & Norman, 1996) point out, the decision to perform a health behavior is often used as the dependent variable in empirical research on health behavior change, as it is much harder for investigators to ascertain the actual performance of a health behavior by research participants compared to relying on their self-reported intention to act. In order to narrow the gap between intention and behavior, the present project uses flu vaccination as the health behavior under investigation. Research has suggested that the predictive power of different health behavior models are context-dependent (Janz & Becker, 1984; Weinstein, 1993), and flu vaccination represents the kind of behavior that can be well accounted for by theories focusing on socio-psychological factors. This 19 is due to the behavior's discrete nature that requires minimal levels of preparation or maintenance, as well as its typical accessibility and relatively lack of structural constraints to perform (N. T. Brewer et al., 2007; Janz & Becker, 1984). In addition, its annual or seasonal occurrence calls for a deliberate decision to act, whereas health behaviors carried out on a daily or weekly basis may become habitual and no longer require a conscious decision (Ouellette & Wood, 1998). As an initial attempt to expand existent theories to include the heuristics of affect, trust and confidence as well as the role of media influence, flu vaccination offers a relatively straightforward example for empirical test. Consequently, the present project will use prominent socio-psychological models as the basis for theory extension, while acknowledging that future research should build upon this initial theoretical extension to account for multilevel and/or stage-based factors and processes that shape more complex health behaviors. The following two chapters will provide an overview of the three flu vaccinations studied in this project, followed by a detailed review ofliterature on socio-psychological models of health behavior change and their application in the context of flu vaccination. CHAPTER 2: FLU VACCINATION AS AN EXEMPLIFICATION OF THE CONUNDRUM OF HEALTH DECISION MAKING IN MODERN TIMES 20 Vaccine is one modern technological innovation that touches on the life of almost everyone and that exemplifies the conundrum of reflexive modernity. Following a golden era of rapid development in vaccine science and widespread support for vaccination that resulted in significant declines in illnesses and deaths from vaccine-preventable diseases, a minority but highly vocal opposition against vaccination emerged in the 1970s (Poland & Jacobson, 20 II). This opposition has rallied around various issues, such as the alleged -and at rare occasions proven -injuries and deaths linked to vaccination, the D D D D and other developmental disorders, and the revelation that many vaccines contain mercury as a preservative. Consequently, a conspiracy theory has developed in the US where the government is accused of colluding with the pharmaceutical industry in exponentially expanding the types of officially recommended vaccines as a way of profitingrfueir :dia1Jbofur:tliJ:milooversies, it is perhaps not surprising that a national survey found that more than one-third of parents in the US were concerned about the general safety of childhood vaccines and 35% believed that children were getting more shots than were good for them (Kennedy, Brown, & Gust, 2005). Despite their concerns, however, 88% of the parents surveyed remained supportive of mandatory vaccination as a pre requisite for elementary school entry. While general support for childhood vaccination remains high, some studies suggest that individuals are less supportive of "hot lots" of vaccine that have been associated with well-publicized negative events (Critcher, 2007), 21 such as the measles-mumps-rubella (MMR) vaccine (Mnookin, 2011), and feel more uncertain about newly developed shots and their potential side effects, such as the human papillomavirus (HPV) vaccine (Bigman, Cappella, & Hornik, 2010; McRee, Brewer, Reiter, Gottlieb, & Smith, 2009). Taking into account that vaccination decisions might be based upon a combination of factors, of which some can be generalized across the broad while others are context-dependent, the present project compares the acceptance of three flu vaccines -the seasonal influenza shot which has been available for public use in the US since the 1940s, the HlNl swine influenza shot that was developed, approved and mass-distributed speedily in 2009, and a hypothetical vaccine against H5Nl avian influenza. Before delving into existent studies that have utilized health behavior change theories to predict adaptive behavior or behavioral intention regarding those vaccines, an overview of the development of each vaccine and the disease it helps prevent is warranted to provide context for subsequent discussions. Seasonal Influenza Vaccine: A Vaccine for All with Uncertain Efficacy Each winter season, up to 20% of the US population catches influenza or the flu (CDC, 2011c). While many individuals consider influenza a nuisance and associate it with mild respiratory symptoms, it can lead to serious complications or even deaths. In the US, the flu was associated with an average of 36,000 deaths annually between 1990 and 1999, and an average of226,000 hospitalizations annually between 1970 and 2001 (Cox & Subbarao, 1999). Individuals aged 65 years or older are at the highest risk for influenza-related complications, and they account for about 90% of all influenza-related 22 deaths (Cox & Subbarao, 1999) and over 60% of influenza-related hospitalizations (CDC, 20 12d). Young children and individuals with certain chronic health conditions are also at higher risk for serious complications from influenza (CDC, 2011 b). Vaccination is the most effective strategy for preventing seasonal influenza and its complications (Cox & Subbarao, 1999; Nichol & Treanor, 2006). The influenza virus family includes many subtypes, and different subtypes tend to be active in different flu seasons. Consequently, the World Health Organizations (WHO) has established a network to monitor influenza virus activity worldwide, and each February, the WHO uses its global surveillance data to predict three virus strains that are most likely to genetically resemble major circulating strains for the coming winter season in the Northern Hemisphere (Demicheli, Di Pietrantonj, Jefferson, Rivett, & Rivett, 2007). This gives vaccine manufacturers enough time to mass produce and distribute a vaccine that simultaneously tackles these three virus strains for the upcoming winter season. WHO's predictions, however, do not always turn out to match the circulating strains well, and, consequently the efficacy of the influenza vaccine varies from year to year depending on the degree of resemblance between the vaccine strains and the actual circulating strains. Indeed, the US Centers for Disease Control and Prevention (CDC, 2013, para. 1) acknowledges that "it's possible that no benefit from vaccination may be observed" during years where the vaccine and circulating strains are not well matched. The effectiveness of the influenza vaccine also differs across population groups, as the elderly and individuals with certain chronic health conditions might develop less immunity after vaccination compared to younger, healthier persons (CDC, 2013). 23 Research has found that in years where the vaccine strains are similar to the circulating strains, the seasonal influenza shot can protect up to 90% of healthy individuals from illness. However, the same shot is only 30%-40% effective in protecting individuals aged 65 years or older (CDC, 2012c). Nonetheless, the CDC recommends influenza vaccination for the elderly and for people with high-risk health conditions because it has considerable efficacy in reducing cases of serious complications and deaths from the flu. The evaluation of the efficacy of the influenza vaccine is further complicated by the fact that there are currently two types of vaccine available in the US. The trivalent inactivated influenza vaccine (TIV) has been available since the 1940s, whereas the live attenuated influenza vaccine (LAIV) was approved for use recently in 2003. TIV contains killed virus so it cannot cause influenza in vaccinated individuals, but some research has found it less effective in preventing illness compared to LAIV (CDC, 2012c). LAIV, on the other hand, contains live flu virus so it can cause mild flu symptoms in vaccinated individuals. As a result, LAIV is approved for administration on healthy, non-pregnant individuals aged 2 to 49 years only and cannot be used on young children or the elderly. The CDC recommends repeated annual influenza immunization for populations considered to be at high risk for influenza-related complications as well as their close contacts. The definition of these populations has expanded over the years to include almost everyone now. In 2000, influenza vaccination was recommended for the elderly (ages 65 and above) and individuals with certain high-risk health conditions, who totaled around 73 million or one-fourth of the US population at that time (Harper, Fukuda, Uyeki, Cox, & Bridges, 2004). Since then, individuals between the ages of 50 and 64 24 years (Bridges, Fukuda, Cox, & Singleton, 2001 ), healthy children between the ages of 6 and 23 months (Harper eta!., 2004), healthy children 24 months through 4 years of age (N. M. Smith eta!., 2006), and healthy children and youth from 5 to 18 years old (Fiore eta!., 2008) had been added to the list of target groups for seasonal flu vaccination. As a result, influenza vaccination was recommended for 85% of the US population by the 2009-2010 flu season. In the following year, the CDC further expanded its vaccination recommendations to include everyone over the age of 6 months based on evidence suggesting higher levels of susceptibility of healthy young adults to the influenza A(H1N1) virus strain that emerged in 2009 to cause a pandemic (Fiore eta!., 2010). Notwithstanding the CDC's recommendations, seasonal flu vaccination rates have been relatively low among all population groups. During the 2010-2011 winter influenza season, for example, vaccination coverage was estimated at 51.5% among children aged 6 months through 17 years, 28.6% among adults aged 18 through 49 years (36.8% for those with high-risk conditions), 42.7% among individuals between 50 and 64 years old, and 64.9% among the elderly, which together averaged to a national coverage of 41.8% (CDC, 2012a). Many studies have been conducted to investigate the reasons behind the relatively low levels of acceptance of the seasonal flu shot despite its ready availability in clinics, hospitals, pharmacies, grocery stores and other places. Continuing skepticism about the vaccine's efficacy, worries about its side effects and influenza-causing potential, and not knowing that the vaccine is recommended for people like oneself have been identified as key barriers to more widespread acceptance of the flu shot (Adler & Winston, 2004; Drociuk, 1999; Fiebach & Viscoli, 1991; Zimmerman eta!., 2003). 25 HlNl Swine Influenza Vaccine: Historical Roots for Vaccine Safety Concern The HlNl influenza is a pandemic flu. Different from seasonal influenza that occurs annually in the winter season and is caused by virus strains already in widespread circulation, pandemic influenza occurs infrequently and is caused by a newly mutated virus strain to which most individuals have no previous exposure and therefore little pre existing immunity (DHHS, 2013a). Consequently, individuals from all population groups are vulnerable to becoming sick. In fact, patterns observed during the three pandemics of the 20th century suggest that healthy young adults, who are usually not at risk for serious complications from seasonal influenza, might be more susceptible to illness during a flu pandemic (DHHS, 2013b). In the spring of2009, a new strain of influenza virus A(HlNl) began spreading across the US and the world. This new strain was a result of genetic re-assortment between human and swine flu viruses so the disease caused by it was often referred to as the HlNl swine flu. By June 11, 2009, nearly 30,000 HlNl flu cases had been confirmed in 74 different countries, prompting the WHO to declare that the first influenza pandemic of the 21st century had arrived (Chan, 2009). Given that production of the 2009-2010 seasonal flu vaccine was already underway and that this vaccine did not protect against the newly emerging HlNl virus, development of a separate HlNl flu shot began soon after the identification of the new strain in late April 2009. By the end of June 2009, several pharmaceutical companies had started manufacturing the newly developed HlNl vaccine (Singleton eta!., 2010), and in mid-September, the US Food and Drug Administration (FDA) approved the use ofHlNl flu shots manufactured by four companies in the country (FDA, 2009). Nationwide distribution of the newly approved H1N1 flu shots began soon after, on October 5 (Singleton eta!., 2010). 26 In the initial weeks ofH1N1 vaccine distribution, the supply was limited so groups considered to be at the highest risk for H1N1 influenza-related complications were prioritized to receive the shot. Based on indications from past flu pandemics and observations of higher attack rate and hospitalization rate among the younger populations compared to the elderly during the first few months of the pandemic, the following groups were prioritized to receive first batches of the H1N1 vaccine: children and young adults aged 6 months through 24 years, individuals between 25 and 64 years old with underlying health conditions that put them at higher risk for influenza-related complications, pregnant women, persons in close contact with infants younger than 6 months, and healthcare and emergency medical personnel (CDC, 2009c ). On the other hand, the CDC recommended individuals not in the high risk groups to protect themselves against H1N1 influenza by adopting more general precautions, such as washing their hands more frequently with soap, avoiding close contact with sick people, and getting a seasonable flu shot that has been available since early September 2009 (CDC, 2009a). Even though the seasonal flu shot did not offer immunity against H1N1 influenza, it reduced individuals' chances of experiencing illness from seasonal influenza and their likelihood of becoming co-infected with seasonal flu and H1N1 flu viruses, which could lead to further re-assortments and mutations of those viruses and consequently a worsening of the pandemic. The CDC thus included seasonal flu vaccination as a preventive measure against H1N1 influenza. By the start of 2010, amply 27 supply of the H1N1 shots had become available so individuals not in the priority groups were also urged to take up the vaccine (Singleton eta!., 2010). The production of the H1N1 flu shots followed the same procedures as those behind seasonal flu shots, and like the seasonal influenza vaccine, the H 1 N 1 vaccine also came in the TIV and LAIV varieties, with the former licensed for use in anyone aged 6 months or older and the latter approved for use in healthy, non-pregnant individuals ages 2 through 49 years only (CDC, 2010). Estimates ofH1N1 vaccine efficacy at the population level in the US are not yet available, but research conducted across seven European countries suggests that the vaccine was between 65% and 100% effective in preventing infection with the pandemic flu depending on population groups, with efficacy being higher among individuals younger than 65 years of age and without underlying health conditions that put them at a higher risk for influenza-related complications (Kissling et a!., 20 11 ). According to CDC estimates, 34.2% of all individuals in the priority groups had received the H1N1 vaccine by the end ofthe 2009-2010 pandemic flu season, whereas the national coverage across all population groups averaged around 27.0% in the US (CDC, 2011a). This was much lower than the national coverage of seasonal influenza vaccination ( 41.2%) during the same period (CDC, 20 11a) and resulted in a waste of more than 70 million doses of the H1N1 influenza vaccine (Poland & Jacobson, 2011). With the relative low vaccine coverage rate, approximately 20% of the US population was estimated to have caught the H1N1 flu during the pandemic, resulting in over 274,000 hospitalizations and about 12,500 deaths (Shrestha eta!., 2011). Furthermore, as 28 projected, individuals younger than 65 years of age experienced much greater rates of disease attack, hospitalization and mortality in comparison to the impact of seasonal influenza. On the other hand, the elderly suffered a lower level of negative consequences from the H 1 N 1 flu compared to seasonal influenza (Shrestha et a!., 2011 ). Studies suggest that concerns about side effects and safety of the H1N1 vaccine were major barriers to vaccine uptake (Blendon, SteelFisher, Benson, Bekheit, & Hernnann, 2009, 2010; Maurer, Uscher-Pines, & Harris, 2010). This is perhaps not surprising given the rapid development and approval of the H1N1 flu shots and the historical reference to the swine flu vaccine 'debacle' of 1976 that is sometimes evoked in media reports (CDC, 2009b; Pollack, 2009; Roan, 2009). In 1976, there was a government effort to vaccinate nearly the entire US population in anticipation of a swine flu pandemic. However, this mass vaccination program was halted after 10 weeks because the vaccine appeared to increase the risk for Guillain-Barre syndrome, a rare neurological condition that caused temporary paralysis and could be fatal. More than 500 individuals were thought to have developed Guillain-Barre syndrome following swine flu vaccination, and 25 of these people eventually died. On the other hand, the pandemic threat never materialized, and ultimately there were only about 200 swine flu cases and one fatality during the 1976-1977 flu season (Roan, 2009). To date, scientists have been unable to figure out how exactly the 1976 swine flu vaccine triggered Guillain-Barre syndrome, and this uncertainty provided the ground for public concern that history could repeat with the 2009 H 1 N 1 influenza vaccine. Consequently, it is not surprising that research found a substantially lower uptake of the H1N1 influenza vaccine compared to the seasonal influenza vaccine during the 2009-2010 flu season despite the widespread belief that the H1N1 flu was more serious than the seasonal flu (Maurer eta!., 2010). H5Nl Avian Influenza Vaccine: Will It Prevent the Next Pandemic? 29 Prior to the H1N1 influenza outbreaks in 2009, public health authorities and medical scientists had considered the H5N1 avian influenza virus the most likely culprit to trigger the next flu pandemic (WHO, 2005a, 2005b). Avian influenza viruses are believed to have caused the three flu pandemics of the 20th century through either genetic reassortment with a human flu virus or direct adaptation to the human body (Peiris, de Jong, & Guan, 2007). Consequently, the WHO was alarmed when a new strain of H5N1 avian influenza virus caused a small human outbreak in Hong Kong in 1997, infecting 18 and killing six (Nerlich & Halliday, 2007). While this human outbreak was quickly contained through the culling of the entire poultry population in farms and markets throughout Hong Kong, the HSN 1 virus has continued to be detected in poultry and wild birds in Asia (Peiris et a!., 2007). As the world entered the new millennium, epidemiologists became increasingly concerned about an overdue flu pandemic based on historic data. In 2003, surveillance of influenza virus activity began to show widespread outbreaks ofH5N1 influenza in domestic flocks and wild birds in Asia, Middle East, Europe and Africa. Concurrently, cases of sporadic human infections with the same avian virus had been reported to the WHO beginning November 2003 (CDC, 2007). While most of the human infections were believed to have been caused by direct contact with diseased birds, there was evidence suggesting that limited human-to-human transmission of the virus had taken place. This 30 led to increasing concern among scientists that the H5N 1 avian virus could soon mutate and acquire the ability to transmit efficiently among humans, thus triggering the next pandemic. By December 2004, the WHO stepped in by urging all countries to develop or update their influenza pandemic preparedness strategies (Nerlich & Halliday, 2007). In the following year, the WHO released a series of documents to specifically assess the pandemic threat of avian influenza to humans and to guide global preparations against such a pandemic (WHO, 2005a, 2005b). One key recommendation from the WHO was that vaccines should be developed against various subtypes of H5N 1 virus circulating among birds in Asia, and human trials should be completed before the onset of a pandemic. This would enable rapid production of vaccines in the event that any of these subtypes triggered a pandemic (Peiris eta!., 2007). Responding to this call for action, pharmaceutical companies such as Sanofi Pasteur and research institutes like the National Institutes of Health (NIH) in the US have been working on developing and testing inactivated and live-attenuated vaccines against different subtypes of the H5Nl virus (Fauci, 2005; Peiris eta!., 2007). Furthermore, governments in developed countries like the US have contracted pharmaceutical companies to produce a substantial quantity of those candidate vaccines as soon as they completed clinical trials. In doing so, there would be a national vaccine stockpile to help with rapid containment of disease should a pandemic occur (Fauci, 2005). As it turned out, the first flu pandemic of the 21st century was caused by the HlNl swine flu virus in 2009, not the H5Nl avian flu virus. Health experts have acknowledged that up to date, "[e]ven though HPAI H5Nl viruses are spreading among 31 poultry and wild birds and this increases the possibility of human exposures to infected birds or poultry, it has not increased the ability ofHPAI H5Nl viruses to infect and transmit between people" (CDC, 2012b, para. 3). Nonetheless, countries like the US still keep a stockpile of HSN 1 vaccines in case that any subtype of this virus ever acquires the ability to transmit easily among humans and cause the next pandemic. 32 CHAPTER 3: SOCIO-PSYCHOLOGICAL THEORIES OF HEALTH BEHAVIOR CHANGE AND THEIR APPLICATION IN THE CONTEXT OF FLU VACCINATION Socio-psychological models are underpinned by the assumption that individuals' perceptions of their social environment are a key force in shaping their behaviors (Conner & Norman, 1996). Despite those models' limitations in accounting for multilevel or multi -stage influence over behavior, they have informed much empirical research in the health domain and have provided the foundation for more dynamic or comprehensive theoretical frameworks. This is largely due to the recognition that the individual constitutes an essential unit of analysis in understanding health decision making and behavior change, and that all larger social units are composed of individuals (Rimer, 2008). Socio-psychological models of health decision making are largely rooted in the rationalist, utility-maximizing tradition. Conner and Norman (1996) point out that in those models, The health behaviours to be predicted are considered to be the end result of a rational decision-making process based upon deliberative, systematic processing of the available information. Most assume that behaviour and decisions are based upon elaborate, but subjective, cost-benefit analysis of the likely outcomes of differing courses of action ... It is assumed that individuals generally aim to maximize utility and so prefer behaviours which are associated with the highest expected utility. (p. 7) 33 While the idea of bounded rationality is acknowledged in the subjectivity of cost and benefit analysis in health decision making, most socio-psychological models of health decision making have not sought to incorporated findings from research on heuristics and biases or on dual-mode processes of judgment making. It is thus the present project's goal to bridge this gap. Before articulating potential ways for theory integration, however, a review of major socio-psychological models of health behavior change is warranted. Health Belief Model The Health Belief Model (HBM), described as a "socio-psychologic [sic] theory of decision making to individual-related health behaviors" (Harrison, Mullen, & Green, 1992, p. 107), is one of the most frequently cited and researched health behavior theories. HBM was developed in the 1950s to investigate reasons behind failure to use preventive and detection healthcare services, such as influenza immunization, the Pap test to screen for cervical cancer and chest x-ray to detect tuberculosis (Rosenstock, 1966, 1974). Moving beyond the traditional focus on socio-demographic determinants of behavior in research, HBM identified a set of subjective, psychological factors that predicted health behavior. The premise of the HBM is that motivation is key to the adoption of a health behavior, and individuals are motivated to the extent that they perceive the health problem in question- for example contracting the flu- as serious (perceived severity), feel themselves to be at risk (perceived susceptibility), believe that the preventive or health-enhancing behavior being recommended - for example flu vaccination- is effective and beneficial (perceived benefits) and that the physical and psychological costs 34 of the advised behavior are reasonable (perceived barriers). Over the years, the HBM has been expanded to include two more constructs to enhance its predictive power- "self efficacy" and "cues to action" (Rimer & Glanz, 2005). Coined by psychologist Albert Bandura (1977), the term self-efficacy refers to individuals' belief in their own capability to overcome obstacles in performing a behavior. This concept has been added to several health behavior change theories, such as the extended HBM (Rosenstock, Strecher, & Becker, 1988) and PMT (Maddux & Rogers, 1983 ), to underscore the fact that oftentimes the performance of a behavior requires both motivation and ability. Bandura considers self-efficacy as resulting from having the knowledge and skills for performing a particular behavior as well as having the opportunity to practice this behavior, and he accords self-efficacy prominence in the change process: Perceived self-efficacy affects every phase of personal change. It determines whether people even consider changing their behavior, whether they can enlist the motivation and perseverance needed to succeed should they choose to do so, and how well they maintain the changes they have achieved. (2002, pp. 144-145) This claim has been supported by research demonstrating that strategies aimed at increasing self-efficacy constitute one of the most effective mechanisms for altering health behaviors (Bandura, 2002; Rimer & Glanz, 2005). The other addition to the HBM, cues to action, refers to anything that prompts individuals to practice a health behavior. They can be either internal, such as subjective perception of one's bodily state, or external, such as a vaccination reminder for from one's 35 doctor, family or fiiends (Rimer & Glanz, 2005; Rosenstock et al., 1988). Furthermore, recent research indicates that the mass media can be an important source of cues to action to activate health behaviors among the general public in today's media saturated environment (Chen & Murphy, 20 12). Figure 1 provides an illustration of the constructs and relationships proposed by the extended HBM. Figure 1. Health Belief Model components and linkages Perceived susceptibility Perceived severity Perceived benefits Health behavior Perceived barriers Self-efficacy Cues to action Although self-efficacy and cues to action are now commonly acknowledged components of the HBM, their average effect size have not been examined in meta analyses, which were largely conducted in the 1980s and included only the four original vcuiables of perceived susceptibility, perceived severity, perceived benefits and perceived barriers (Champion & Skinner, 2008). Fmthermore, no published meta-analyses have assessed the predictive power of the model as a whole- they focused, instead, on individual constructs. While the original HBM constructs were found to be significant 36 predictors of behavior in meta-analyses, their average effect size tended to be small, with Harrison eta!. (1992) reporting that none of the constructs explained more than 10% of the variance in behavior. Furthermore, the predictive power of each variable appears to vary according to the type of behavior studied (e.g. preventive vs. sick role behaviors) or the temporal design of a study (i.e. prospective vs. retrospective studies). For example, Janz and Becker (1984) found that across studies, perceived barriers and perceived severity constituted, respectively, the strongest and the weakest predictor of behavior. However, perceived susceptibility is a more powerful predictor of preventive behaviors (e.g. vaccination and smoking cessation) than perceived benefits, whereas the reverse is true for sick role behaviors (e.g. compliance with regimens prescribed by a physician). Harrison eta!. (1992), on the other hand, found that compared to prospective studies, the effect size for perceived benefits and perceived barriers in retrospective studies tended to be larger while the effect size for perceived severity was relatively small. As the current project utilizes a retrospective, self-report design and focuses on the preventive behavior of flu vaccinations, the expectations are that perceived susceptibility and perceived barriers would serve as the strongest behavioral predictors while perceived severity would be the weakest component. HBM has indeed been applied to a number of studies on flu vaccine acceptance, but research findings have been inconsistent. For example, Fiebach and Viscoli (1991) argued that the HBM did not explain seasonal influenza vaccination very well because in their study, they found that perceived susceptibility and perceived severity were not associated with vaccination behavior, even though cue to action operationalized as 37 doctor's recommendation, perceived effectiveness of the shot in preventing influenza and perceived barriers operationalized as concern for vaccine side effects were. On the other hand, Rundall and Wheeler (1979) concluded the HBM was useful in explaining vaccination uptake during the 1976-1977 swine influenza epidemic. They found that the HBM variables explained 34% of the variance in swine flu vaccine behavior, with perceived susceptibility and perceived barriers operationalized as danger associated with vaccination being the strongest predictors of behavior. Another study conducted during the same swine flu epidemic also concluded that all four original HBM variables were effective in explaining vaccination behavior, but most of their impact was mediated through vaccination intention (Cummings, Jette, Brock, & Haefner, 1979). It is worth noting that the socio-psychological constructs identified as most relevant in motivating behavior change by the HBM have informed the development of other theories. For example, the Protection Motivation Theory (PMT) (R. W. Rogers, 1983), which has informed many health interventions, can be considered a hybrid theory of the HBM and Bandura's work on self-efficacy (Boer & Seydel, 1996). However, PMT is a theory developed to guide the design of persuasive health communication rather than to understand health behavior change per se. Consequently, it articulates a different set of rules for linking intention and behavior with their predictors of perceived susceptibility, perceived severity, perceived benefits of a maladaptive behavior, perceived costs of adaptive behavior, response-efficacy, and self-efficacy compared to the HBM (Boer & Seydel, 1996; Witte, Meyer, & Martell, 2001). Given the current project's focus on health decision making in everyday context, with or without exposure to persuasive messages, PMT's utility is more limited and therefore a review of the theory is not provided here. Theory of Reasoned Action and Theory of Planned Behavior 38 While HBM was developed to predict health behavior, studies citing HBM as their theoretical basis have often employed intention, rather than actual action, as their dependent measure. On the other hand, behavioral intention is accorded an independent and explicit role in the Theory of Reasoned Action (TRA), which was developed as a general framework for explaining all types of voluntary behaviors (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975). According to this theory, intention is the best predictor of behavior and encompasses all motivational factors behind the performance of a behavior. These motivational factors can be grouped under two categories - attitude toward the behavior and subjective norm pertaining to this behavior. TRA's developers (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975) followed an expectancy-value tradition and considered an individual's attitude toward a behavior as a function of their most important beliefs about the consequences of performing this behavior (expectancy) as well as the subjective value they place on those consequences (value). For example, an expected consequence of getting vaccinated against seasonal influenza could be that one would not contract the flu during the winter months, and some individuals might value this positive consequence very much while others might care less about it. Similarly, subjective norm is co-determined by one's beliefs about whether significant others think that he or she should perform the behavior (expectancy) and one's motivation to comply with those others (value). For example, some individuals might believe that their parents want them to get vaccinated against the flu, but not everyone is highly motivated to comply with their parents. 39 Building upon TRA, Ajzen (1991), later developed the Theory of Planned Behavior (TPB) to include "perceived behavioral control" (PBC) as a key construct in predicting behaviors that are not completely under a person's volitional control. Again, PBC was conceptualized as a product of individuals' beliefs about the presence or absence of the resources or opportunities needed to perform a particular behavior (expectancy) and the perceived importance of each resource or opportunity in facilitating behavioral performance (value). For example, one might believe that flu vaccination is not offered in a location close to his or her home, and having convenient access to vaccine service might be more important to some people (e.g. those who lack the means of transportation) than others. PBC's influence over behavior is context dependent and can be either direct or indirect. Ajzen (1991) argues that PBC would have little effect on behaviors that are under high volitional control. However, as volitional control decreases, PBC becomes an increasingly powerful and independent predictor of behavior. As Armitage and Conner (2001, p. 473) point out, "[t]his is based on the rationale that increased feelings of control will increase the extent to which individuals are willing to exert additional effort in order successfully to perform a particular behavior." Furthermore, PBC can exert indirect influence over behavior via intention, as individuals are more likely to form the intention to act in the first place when feeling in control of the environmental factors that influence 40 the enactment of a behavior. Figure 2 provides an illustration of the relationships among the TPB constructs. Figure 2. Theory of Planned Behavior components and linkages Behavioral beliefs & Attitude outcome evaluation Normative Subjective Behavior beliefs & motivation norm to comply Control Perceived beliefs & behavioral perceived control facilitation Meta-analyses have demonstrated that both TRA and TPB explain a moderate amount of variance in intention and behavior, and that the addition ofPBC in TPB results in a significant improvement in the model's predictive power. For example, Armitage and Conner (2001) analyzed 185 published empirical tests ofTPB and found that the model on average accounted for 27% of the variance in behavior and 39% ofthe variance in intention. Furthermore, PBC on average added 2% to the prediction of behavior and 6% to the prediction of intention. On the other hand, Godin and Kok (1996) focused their analysis on 87 studies involving the application ofTPB in the health context, and they found that the model on average explained 34% of the variance in behavior and 41% of the variance in intention. Moreover, the addition ofPBC on averaged added 12% to the prediction ofbehavior and 13% to the prediction of intention. 41 Despite TPB's ability to explain a considerable amount of variance in intention and behavior, various propositions have been made for theory extension to further enhance the predictive power of this model. For example, moral norms or personal beliefs about what is right or wrong (Conner & Armitage, 1998; Godin, V ezina-hn, & Naccache, 2010), descriptive norms or one's perceptions of what most people do in a given situation (Cooke, Sniehotta, & Schuz, 2007), past behavior (Cooke et al., 2007; Gallagher & Povey, 2006), emotions such anticipated regret (Cooke et al., 2007; Gallagher & Povey, 2006; Godin et al., 2010; Sandberg & Conner, 2008), and self identity (Conner & Armitage, 1998) have all been nominated as potential predictors to be incorporated into the TPB. However, for any of these variables to become part of the established TPB, more empirical testing and theoretical articulation are needed. Even though TRA and TPB were developed in a non-health context, both models have been widely applied to study health behavior change, including vaccination acceptance. For example, Montano (1986) conducted a two-wave longitudinal survey to identify the TRA variables that predicted seasonal flu vaccination intention and actual vaccine uptake during the 1982-1983 flu season. Depending on how attitude was operationalized (i.e. a summary score computed from multiplying behavioral beliefs and their evaluations vs. a single global measure), attitude and social norm together accounted for between 52% and 62% of the variance in vaccination intention, and between 28% and 31% of the variance in vaccination behavior. Gallagher and Povey (2006), on the other hand, applied TPB in investigating flu vaccination intention among older adults. They found that attitude, subjective norm and perceived behavioral control together accounted 42 for 48% of the variance in vaccination intention. However, when looking at individual predictors, only subjective norm had a statistically signification relationship with intention. Gallagher and Povey also sought to test an extended TPB by adding anticipated regret and past flu vaccination behavior into their analysis. They found both additional variables to be significant predictors of intention, and the extended model accounted for 72% of the variance in intention. While this extended model was powerful in explaining vaccination intention, many researchers have cautioned against the inclusion of past behavior in behavioral prediction models because it adds no independent explanatory value and provides no useful information for designing effective health interventions (Ajzen, 1987). An Integrated Model of Health Behavioral Prediction Looking at the HBM, TRA and TPB, clear similarities and differences emerge in the constructs and relations they propose. All theories acknowledge the importance of perceived costs and perceived benefits as proximal determinants of health decisions. However, subjective norms are not part of the formulation in HBM; TRA does not consider the role of self-efficacy/perceived behavioral control; and perceived susceptibility, perceived severity and cues to action are missing in action in TRA and TPB. It is possible to advance an integrated model that includes constructs identified by all these theories, and there has been notable effort in this direction. For example, Oliver and Berger (1979) conducted a prospective study during the 1976-1977 swine flu epidemic to test the respective power of TRA and HBM in predicting the uptake of the swine flu vaccine. They found that vaccination intention explained about one-third of the 43 variance in vaccination behavior, and that all constructs in HBM and TRA were significant predictors of intention. However, HBM accounted for only 29% of the variance in intention, whereas TRA explained 55% of the variance. When both the HBM and TRA variables were entered into the same multiple regression analysis, attitudes, subjective norms as well as susceptibility and severity perceptions were found to be significant predictors of vaccination intention, and this integrated model accounted for more than 56% of the variance in intention. Those findings were not surprising given that perceived susceptibility and perceived severity constituted the two HBM components not accounted for by TRA, unlike perceived benefits and perceived barriers that could be considered subsumed under the attitude variable. Oliver and Berger (1979) therefore called for future research to combine non-redundant constructs from TRA and HBM into an integrated model with greater predictive power. In recent years, Fishbein and colleagues (Fishbein & Cappella, 2006; Fishbein & Yzer, 2003) have led a concerted effort in developing an integrative model of behavioral prediction. In this model, perceived susceptibility, perceived severity and cues to action are considered "distal" determinants, whose effects on decision-making are fully mediated by the proximal predictors of attitude, perceived norm and self-efficacy (see Figure 3). However, some researchers argue that perceived susceptibility and perceived severity constitute key considerations in health decision-making (N. T. Brewer et a!., 2007; Norman & Conner, 1996) so when shifting from a general theory of behavioral prediction to a more specific model of health behavior change, these variables should be given more weight. 44 Perhaps the truth is somewhere in between, as suggested by the specific hypotheses and empirical evidence provided by Ronis and colleagues (Ronis, 1992; Ronis & Harel, 1989). These researchers suggested that in the health context, the effects of perceived susceptibility and perceived severity on intention or behavior should be fully mediated by perceived benefits. Their rationale was that even when individuals felt vulnerable toward a health threat and considered its consequences to be setious, they would not perform a health-protective behavior unless they believed in its effectiveness in reducing that health threat. However, this rationale was only partially supported by empirical evidence. Figure 3. An Integrative Model of Behavioral Prediction components and linkages / Distal Behavioral variables: ~ beliefs & ~ Attitude Skills -Demographic , I outcome I variables I evaluations , -Culture I ~ I -Attitudes ' toward targets I I I (e.g. stigma) I Nonnative Perceived Behavior - Personality, -----) beliefs & ~ ... Intention _..,; moods & nonn ,. , \ motivation emotions \ \ to comply -Other \ ' ~ ~ individual \ \ differences \ \ variables (e.g. \ \ Self- Environ- perceived risk) ~ Efficacy --) efficacy mental - Exposure to beliefs constraint media & other interventions ~ ~ Using survey data collected fi·om a probability sample of 619 women living in metropolitan Detroit, Ronis and Harel (1989) found that the impact of perceived severity on breast examination behavior was completely mediated by perceived benefits, whereas 45 the influence of perceived susceptibility on behavior was direct and not mediated by perceived benefits at all. These researchers reasoned that breast examination was a detection behavior, whose key benefit was to identify breast cancer early so that it could be treated promptly before advancing in severity. Consequently, it was logical for the impact of perceived severity to be mediated by perceived benefits but not that of perceived susceptibility. However, when it came to preventive behaviors, such as immunization, Ronis and Hare! (1989) suggested that an analogous but reversed pattern should be found. Under such circumstances, perceived severity should have a direct impact on behavior, whereas the effect of perceived susceptibility would be mediated. In a subsequent study on the preventive behavior of dental flossing, however, Ronis (1992) did not find this reversed pattern of relationships. Rather, he found that perceived susceptibility continued to have both a direct effect and an indirect effect via perceived benefits on flossing behavior. Other studies utilizing an experimental design (Ronis, 1992) or targeting the intention and behavior for sun protection (K. M. Jackson & Aiken, 2000, 2006) also found that the impact of perceived severity tended to be fully mediated by perceived benefits, whereas perceived susceptibility was likely to have both a direct and an indirect influence over intention or behavior. Based on the inconclusive evidence on whether susceptibility and severity perceptions serve as distal or immediate determinants of health behavior, the present project seeks to clarify the relationships among perceived susceptibility, perceived severity, perceived benefits, intention and behavior. 46 Before presenting the hypotheses, however, it needs to be pointed out that data for the current project were collected during two different time periods: during the H1N1 influenza pandemic in October and November 2009, and outside of an influenza season in April and May 2012. In 2009, data were collected to assess individuals' perceptions regarding the H1N1 flu as well as their preventive behavior and intention. As mentioned in Chapter 2, up till mid November 2009, the H1N1 flu shot was short in supply so individuals not in the priority groups were unlikely to have access to this vaccine. Consequently, the CDC advised the majority of the population to protect themselves by adopting general precautions, such as washing their hands more frequently with soap, avoiding close contact with sick people, and getting the seasonal flu shot that had been available since the start of September (CDC, 2009a). Even though the seasonal flu shot did not offer immunity against H1N1 influenza, it helped ensure that individuals would be in a better health condition as they would be less prone to severe illness from seasonal influenza. Therefore, uptake of the seasonal flu vaccine during the 2009-2010 flu season and intention to receive the H1N1 flu vaccine once it became available were both considered as an indication of prevention against H1N1 influenza. In 2012, data were collected to assess individuals' perceptions regarding the seasonal flu and a hypothetical H5N1 bird flu. Given the hypothetical nature of the H5N1 flu, study respondents were asked about their intention to receive a vaccine against this flu. On the other hand, information was obtained on respondents' actual uptake of influenza vaccine in the previous influenza season (i.e. the 2011-2012 flu season). Given the retrospective nature of this information, empirical analysis is conducted to serve the limited purpose of identifying correlates for seasonal flu shot uptake rather than to test any causal relationship. 47 During both data collection periods, perceived benefit of vaccine was measured as perceived effectiveness of vaccine in general, whereas perceived cost of vaccine was measured in terms of safety concern over vaccine in general. Given the available data, the following hypotheses are tested using information collected from the 2009 survey: Hl(a): The effect of perceived severity of the HlNl flu on intention to vaccinate against this flu is fully mediated by perceived effectiveness of vaccine. Hl(b): The effect of perceived severity of the HlNl flu on uptake of the seasonal flu vaccine is fully mediated by perceived effectiveness of vaccine. H2(a): The effect of perceived susceptibility to the HlNl flu on intention to vaccinate against this flu is partially mediated by perceived effectiveness of vaccme. H2(b): The effect of perceived susceptibility to the HlNl flu on uptake of the seasonal flu vaccine is partially mediated by perceived effectiveness of vaccine. Furthermore, the following two postulates are tested using information gathered through the 2012 survey: Hl(c): The effect of perceived severity of the hypothetical H5Nl bird flu on intention to vaccinate against this flu is fully mediated by perceived effectiveness of vaccine. H2( c): The effect of perceived susceptibility to the hypothetical HSN 1 flu on intention to vaccinate against this flu is partially mediated by perceived effectiveness of vaccine. 48 In addition to the need to better map out the relationships among distal and proximal predictors of health decisions, consensus is lacking on how perceived susceptibility and perceived severity relate to each other. While they are commonly treated as independent predictors of behavior in the HBM, some scholars suggest that these two variables should be multiplied to create a single variable "perceived threat" in accordance with the expectancy-value structure (Rimer, 2008). Other researchers argue that the effect of perceived susceptibility is moderated by the influence of perceived severity (Kruglanski & Klar, 1985; Weinstein, 1988), as "severity must reach a certain magnitude to figure in health decisions, but once the magnitude has been reached, decisions are solely a function of perceived susceptibility" (Sheeran & Abraham, 1996, p. 34). Despite being theoretically plausible, empirical studies have typically found no interaction between perceived susceptibility and perceived severity in the context of health decision making (R. W. Rogers & Mewborn, 1976; Ronis, 1992; Ronis & Hare!, 1989). In explaining this lack of multiplicative effects, Ronis and Hare! (1989) suggested that when it comes to health decision making, individuals might not consider susceptibility and severity simultaneously. Instead, they might rely on rules of thumb such as "ifl'm not likely to get it, I'm not going to worry about it" or "if it isn't going to be serious, don't worry about it"- in making decisions about behavior change. On the 49 other hand, Weinstein (2000) argues that the relationship between susceptibility and severity is complicated and context dependent, and interaction effects are unlikely to be detected in correlational studies that employ a between-subject design. Given all these arguments and previous research findings, perceived susceptibility and perceived severity are not expected to have a multiplicative effect on vaccination decisions in the present study, which employs a between-subject cross-sectional design. Last but not least, some scholars have suggested that perceived susceptibility and perceived severity might be all but two attributes of perceived threat, which might involve not only cognitive but also affective assessments (Weinstein, 1993). This suggestion is in line the psychometric tradition in risk management, which argues that individuals consider the perceived threat of a risk based on the three factors of 'exposure' -a cognitive assessment- as well as 'dread' and 'unknown-ness,' which are more emotional in nature (Slovic, Fischhoff, & Lichtenstein, 1985). Nonetheless, research on major health behavior models has not tapped much into the affective quality of risk perceptions to date. Recognizing the importance to pursue this line of inquiry, it is beyond the scope of the present project and will need to be investigated in future studies. CHAPTER 4: INCORPORATING AFFECT, TRUST, CONFIDENCE AND MEDIA INFLUENCE INTO SOCIO-PSYCHOLOGICAL THEORIES OF HEALTH BEHAVIOR CHANGE 50 There is a general agreement among psychologists that individuals perceive the world, process information and make decisions based on two distinct modes of thinking (E. R. Smith & DeCoster, 2000), one automatic and associative and the other systematic and rule-based. Most established health behavior change theories are, however, cognitively oriented and focus exclusively on the process of systematic decision making (Peters, Lipkus, & Diefenbach, 2006). The time is ripe for health behavior models to embrace the dual-mode perspective so that they can offer a more comprehensive explanation of behavior and become more powerful tools of prediction. This chapter proposes the incorporation of integral affect, social trust, social confidence and media influence as a first step in extending extant theory, given their powerful influence over health decision making as suggested by empirical evidence. The dual-mode perspective was touched on in Chapter I, but its application to health decision making will be explicated in greater detail here. Bodenhausen, Sheppard, and Kramer (1994) and Chaiken (1980) use the term 'systematic' versus 'heuristic' thinking to articulate the distinction between the two routes of social information processing and judgment making. Systematic thinking involves "effortful cognitive activity whereby the person draws upon prior experience and knowledge in order to carefully scrutinize all of the information relevant to determining the central merits" (Petty, Brinol, & Priester, 2009, p. 132) of a position or decision. For example, when 51 systematically thinking about whether to get vaccinated against the flu, individuals may try to estimate their susceptibility to the flu and the severity of the likely consequences from getting ill. They may also contemplate the effectiveness of the vaccine and its potential side effects as well as think about whether their significant others would want them to get the flu shot and if they can easily obtain a shot. It is after a careful appraisal of all those considerations that people arrive at a decision about flu vaccination. Now, if we take a moment to reflect on how we decided on getting the flu shot last winter, did we go through such an effortful thought process? The answer is likely to be 'no' for most people. It is widely acknowledged among researchers that it is neither possible nor adaptive for human beings to engage in slow, systematic thinking in every judgment making situation (Forgas, 1995; Peters et al., 2006; Petty & Cacioppo, 1986). In fact, humans tend to resort to systematic thinking only when simpler information processing strategies prove inadequate to a judgment task. Forgas (1995) suggests that several sets of factors co-determine if systematic thinking will be involved in making a judgment: the features of a judgment target (e.g. familiarity, topicality and complexity), characteristics of the individual engaged in the judgment task (e.g. personal involvement, cognitive capacity and motivational goal) and situational demands (e.g. need for accuracy, time pressure, and social desirability). He states that systematic processing is more likely when "the target is complex or atypical and the judge has no specific motivation to pursue, has adequate cognitive capacity, and is motivated to be accurate, possibly because of explicit or implicit situational demands" (Forgas, 1995, p. 47). 52 On the other hand, in many decision making contexts, there is no demand for detailed and precise considerations, the judgment target may be familiar or highly typical, the personal relevance of the judgment may be low, and the individual making the judgment may have no specific motivational goals and only limited cognitive capacity. Under such circumstances, individuals tend to engage in heuristic thinking. That is, rather than conducting a thorough search and a careful evaluation of all knowledge and past experience relevant in making a particular decision, individuals often rely on a few heuristics or mental shortcuts to arrive at a decision (Kahneman eta!., 1982; Tversky & Kahneman, 1974 ). Kahneman and Frederick (2002) suggest that affect, availability and representativeness are the three most frequently evoked mental shortcuts in making various types of decisions. The following section seeks to discuss the relevance of the affect heuristic in health decision making, whereas the role of the availability heuristic will be explicated in the later section on media influence over health-related perceptions and judgments. Integral Affect in Judgment Making It is necessary to first define "affect" and related constructs in order to delimit the scope of the current project. In scholarly writing, terms such as "affect," "emotions," "feelings" and "moods" are often used interchangeably. However, Forgas (1995) points out that affect is a more inclusive construct that can be used to refer to both moods and emotions. Simply put, affect is "a feeling that something is good or bad" (Slovic, Finucane, Peters, & MacGregor, 2004, p. 311 ). This feeling can be free-floating and not linked to a particular stimulus, as seen in moods. Peters eta!. (2006, p. S 141) name this 53 type of affect "incidental affect." Affect can also be experienced in relation to a particular stimulus and is typically based on a person's prior experiences and thoughts related to that stimulus- this type of affect is called "integral affect." Based on this definition, emotion can be considered a kind of integral affect. However, for an affect to be classified as an emotion, it needs to fulfill additional psychological and functional criteria (Nabi, 1999). The present study focuses on integral affect, as this type of affect appears to be what analysts (Alhakami & Slovic, 1994; Kahneman & Frederick, 2002) are referring to when they call for the inclusion of an affect heuristic in decision making theory. While research has suggested that discrete emotions (Nabi, 1999, 2003, 2010; Witte & Allen, 2000; Witte eta!., 2001) and moods (Salovey & Birnbaum, 1989) can also play a role in shaping health decisions, they are beyond the scope of the current project. When employing an affective heuristic in making a judgment, individuals rely on how they feel about the judgment target rather than what they think (Alhakami & Slovic, 1994; Kahneman & Frederick, 2002). For example, individuals may decide to get a flu shot merely because they feel good about getting the vaccine, and they may decide against having a colonoscopy simply because they experience negative affect when thinking about having this procedure done. Under such circumstances, it is affect that functions as the informational input in the decision making process, rather than statistics, evidence or logic. Consequently, this has been termed the "affect heuristic" (Forgas, 1995) or "affect-as-information" (Peters eta!., 2006) approach to decision making. Damasio (1994), one of the most prominent contemporary neuropsychologists, has provided one, if not the most, widely cited hypothesis on the origin of this heuristic 54 route to decision making. He argues that a lifetime of personal experience, education and socialization have enabled individuals to connect various stimuli to different future outcomes and their associated affective states in an automatic fashion. Therefore, when making a decision: [B]efore you apply any kind of cost/benefit analysis to the premises ... something quite important happens: When the bad outcome connected with a given response option comes to mind, however fleetingly, you experience an unpleasant gut feeling. Because the feeling is about the body, I gave the phenomenon the technical term somatic state ... and because it "marks" an image, I called it a marker. .. What does [this] somatic marker achieve? It forces attention on the negative outcome to which a given action may lead, and functions as an automated alarm signal which says: Beware of danger ahead if you choose the option which leads to this outcome ... The automated signal protects you against future losses ... and then allows you to choose among fewer alternatives. There is still room for using a cost/benefit analysis and proper deductive competence, but only after the automated step drastically reduces the number of options. (p. 173) Damasio's work suggests that the heuristic and systematic routes to decision making are not mutually exclusive and that the influence of affect extends beyond the heuristic process. There is indeed empirical evidence supporting the impact of affect on systematic thinking. In the context of health-related judgment making, for example, Salovey and Birnbaum (1989) found that individuals experiencing positive affect, in the form of 55 mood, had more positive perceptions of their health efficacy, illness symptoms and possible future disease. Croyle and Uretsky (1987) also found that experiment participants induced to have a positive mood rated their health status more favorably compared to participants induced to have a negative mood. Situating such findings within the associative network theory of memory, Salovey, O'Leary, Stretton, Fishkin, and Drake (1991) argue that mood congruent information tends to be more accessible in the memory network and therefore more likely to be used in making subsequent judgments. In the context of persuasion, research based on the Elaboration Likelihood Model (ELM) demonstrates that under high elaboration conditions where individuals engage in systematic processing of persuasive messages, having a positive mood indirectly predicts a more positive attitude toward the advocated position by influencing the positivity of the thoughts induced by the persuasive messages (Petty, Schumann, Richman, & Strathman, 1993). On the other hand, under low elaboration conditions where individuals engage in heuristic thinking, positive mood directly leads to a more positive attitude toward the advocated position. When it comes to integral affect (i.e. affect directed toward a particular stimulus), Alhakami and Slovic (1994) and colleagues (Finucane, Alhakami, Slovic, & Johnson, 2000) have studied individuals' judgments of various hazardous technologies, such as pesticides, cell phones, food preservatives, natural gas and nuclear power plants. They found that individuals' judgments of a technology's risk and benefit tended to be negatively correlated despite the fact that risk and benefit were usually qualitatively different and therefore unrelated or positively correlated in reality. Furthermore, they 56 discovered that the magnitude of correlation between the perceived risk and perceived benefit of a hazard was associated with the strength of the positive or negative affect attached to that hazard by an individual. In other words, both the valence and intensity of affect attached to a stimulus served to influence subsequent judgment making. While Alhakami and Slovic (1994) and colleagues (Finucane eta!., 2000) regard those research findings as evidence of the affect heuristic at work, other scholars point out that this could also be explained by the indirect influence of affect over systematic thinking, a phenomenon they term the "affect-priming" (Forgas, 1995) or "affect-as-spotlight" approach to decision making (Peters et a!., 2006). In fact, there is evidence that integral affect influences intention to adopt a new technology both directly and indirectly through the perceptions of benefit and risk. For instance, Siegrist eta!. (2007) found that individuals with more positive affect toward food and food packaging modified through nanotechnology tended to perceive those items as more beneficial and less risky. Perceived benefit, in turn, positively predicted intention to purchase those items, whereas perceived risk negatively influenced buying intention. In addition, affect was found to have a positive impact on purchasing intention over and beyond the influence of perceived benefit and perceived risk. Similar findings have come from the field of economic behavior, and research in this field has shed light on when the influence of initial affective reactions could be eliminated by subsequent systematic thinking. For example, Ganzach (2000) conducted a series of studies where graduate students majored in finance were asked to consider investing in unfamiliar foreign stock markets. Those students' perceived return from a 57 market was found to negatively associate with their perceived risk, in spite of their learning of the positive relationship between risk and return in global stock markets in the classroom. The author argues that this inverse relationship is rooted in an overall impression of a market as either good or bad. However, when the causal schema suggesting that higher risk leads to higher return is implicitly evoked through experimental manipulation, the perceived risk and return from investing in an unfamiliar market become positively associated (Ganzach, 2000). Moreover, research participants' risk and benefit judgments of familiar stock market portfolios were found to be positively correlated. Taken together, these findings suggest that the influence of affect is likely to be eliminated in systematic decision making when effort is made to implicitly or explicitly highlight cognitive schemas that are at odds with judgments made based on affective reactions, or when the judges are well versed about the judgment target and its attributes. Despite all this research on the different ways in which affect can influence decision making, there has been little effort to incorporate affect into established health behavior change models. It is therefore the goal of the present study to begin the process of theory extension by hypothesizing about the direct and indirect influence of affect over vaccination decisions and by empirically testing those hypotheses. Before presenting the hypotheses, one needs to reiterate that data for the current project were collected during two different periods. Information on integral affect toward getting the flu shot was collected in 2012 only, and therefore all hypotheses below are tested using data from the 2012 survey: H3: Integral affect toward flu vaccination has a direct and positive effect on intention to vaccinate against the hypothetical HSN 1 flu. H4: Integral affect toward flu vaccination positively predicts perceived effectiveness of vaccine in general. 58 HS: Integral affect toward flu vaccination negatively predicts safety concern over vaccine in general. Perceived risk and perceived benefit have been the focus of investigation in studies looking at the role of affect in judgment making, and these studies are primarily from the field of risk management and risk communication (Alhakami & Slovic, 1994; Finucane et a!., 2000; Ganzach, 2000; Siegrist et a!., 2007; Siegrist, Keller, & Cousin, 2006). In the context of health decision making, however, it is plausible that affect can also shape other perceptions considered important in predicting health behavior, for example social norm and self-efficacy: H6: Integral affect toward flu vaccination positively predicts subjective norm regarding immunization against the hypothetical H5N1 flu. H7: Integral affect toward flu vaccination positively predicts self-efficacy regarding immunization against the hypothetical H5N1 flu. On the other hand, it is less than clear as to how integral affect would influence perceived susceptibility and perceived severity. lf we assumed that the experience of positive affect would make positive thoughts more accessible in the memory network compared to negative thoughts, as argued by proponents of the mood congruency hypothesis (Salovey eta!., 1991), we would expect positive affect toward flu vaccination 59 to be associated with more optimistic and thus lower estimates of perceived susceptibility and perceived severity. Simultaneously, we would expect positive affect to result in a more positive decision regarding influenza immunization. However, lower levels of perceived susceptibility and perceived severity serve to reduce rather than increase motivation to vaccinate against the flu, as suggested by established health decision making models such as HBM. This inconsistency suggests the need to empirically clarify how affect influences susceptibility and severity perceptions: RQ 1: Does affect toward flu vaccination influence perceived susceptibility to the hypothetical H5N1 flu and if so, in what direction? RQ2: Does integral affect influence perceived severity of the hypothetical H5N1 flu and if so, in what direction? Hypotheses 1 to 7 and research questions 1 and 2 pertaining to the predictors ofH5N1 flu vaccination intention are summarized in Figure 4. Figure 4. Affect in predicting H5Nl flu vaccination intention in 2012 Integral affect Perceived effectiveness Safety concern Self-efficacy Hypothetical H5Nl flu vaccination intention 60 Note. Dotted lines indicate the hypotheses regarding the mediational effects of perceived effectiveness of vaccine. Social Tt·ust and Social Confidence in Shaping Risk Perceptions and Health Behaviors Trust is a frequently evoked concept in explaining health behaviors. For example, trust in one's healthcare providers has been linked to greater utilization of preventive healthcare services (Musa, Schulz, Hanis, Silvetman, & Thomas, 2009) and more regular healthcare visits (Halbett, Annstrong, Gandy, & Shaker, 2006), whereas distrust in public health authorities has been found to prevent individuals from following the authorities' recommendations during a public health emergency (Public Health- Seattle & King County, 2008). However, research on the role oftmst in influencing health behavior has been largely atheoretical, and there has been no concerted effmt to incorporate trust into existing models of health behavior change. 61 On the other hand, many communication and risk management studies have investigated the influence of trust over beliefs, attitudes and behaviors based on well developed or developing theoretical frameworks. An overview ofliterature from these disciplines can thus offer insights on how to incorporate trust into established health behavior models. Before delving into the literature, it needs to be acknowledged that trust has often been used interchangeably with the terms 'confidence' and 'credibility,' even though many scholars acknowledge that these concepts are interrelated but not identical with one another. Sociologist Luhmann's (1979, 2000) definitions of trust and confidence have influenced much research in risk management and communication (Earle et al., 2007). In such, trust results from the active choice of the trusting party who weighs the risk against the potential advantages in allowing the trustee to act on one's behalf, without having any control over the trustee's action. Confidence, on the other hand, is based on the prediction that the other party will behave in a certain way in a particular situation based on its past performance and/or on its obligation to answer to control mechanisms, such as law and politics in modern society. According to Luhmann (2000, p. 97), Both concepts refer to expectations which may lapse into disappointment. The normal case is that of confidence. You are confident that your expectations will not be disappointed: that politicians will try to avoid war, that cars will not ... suddenly leave the streets and hit you on your Sunday afternoon walk. You cannot live without forming expectations with respect to contingent events and you have to neglect, more or less, the possibility of disappointment. You neglect 62 this because it is a very rare possibility, but also because you do not know what else to do. The alternative is to live in a state of permanent uncertainty ... Trust, on the other hand, requires a previous engagement on your part. It presupposes a situation of risk. You may or may not buy a used car which turns out to be a 'lemon' ... You can avoid taking the risk, but only if you are willing to waive the associated advantages ... The distinction between confidence and trust thus depends on perception and attribution. If you do not consider alternatives, you are in a situation of confidence. If you choose one action in preference to others in spite of the possibility of being disappointed by the action of others, you define the situation as one of trust. However, Luhmann acknowledges that the distinction between trust and confidence is often subjective and less than clear cut in everyday life. For example, many parents take for granted that they have to follow the schedule recommended by health authorities in vaccinating their children and perceive no alternative, whereas some parents consider themselves having the choice of not vaccinating their children at all or vaccinating them based on a different schedule. In other words, vaccinating one's children based on the recommended schedule can be a situation of either trust or confidence in the health authorities depending on whether the parent sees a possibility for active choice. Furthermore, trust can morph into confidence and vice versa. For example, Renn and Levine (1991, p. 179) argue that "[p]eople have confidence in a source if their prior investment of trust in that source has not been disappointing over a longer period of time." On the other hand, Luhmann (2000, p. 98) points out that "a relation of confidence 63 may turn into one of trust if it becomes possible to avoid that relation. Thus elections may to some extent convert political confidence into trust, at least if your party wins." It has become apparent that distinguishing trust from confidence requires careful attention to the issues of perceived alternative, agency and temporal fluidity, and this distinction always involves a layer of subjectivity. The concept of credibility, on the other hand, seems to be defined by its communal nature. Renn and Levine (1991, p. 180) argue that "both trust and confidence are necessary conditions for the assignment of credibility to a source. Credibility is a product of long-term evidence and commonly shared experience that a source is competent, fair, flexible to new demands, and consistent in its task performance and communication efforts." This definition of credibility as a perception shared by a community is echoed by J. X. Kasperson, Kasperson, Pidgeon, and Slovic (2003 ). Despite this distinct definition, trust is sometimes used interchangeably with credibility in communication literature. For example, Giffin (1967) argues that trust in the information source by the information receiver is the theoretical equivalent of 'source credibility,' a construct that has been extensively investigated by communication scholars since the 1940s. Even more conceptual overlaps emerge when one compares the dimensions of source credibility and trust as proposed by different scholars. For example, in reviewing fifty years of research on source credibility in the persuasion context, Pornpitakpan (2004) concludes that credibility is defined by the dimensions of perceived 'expertise' and 'trustworthiness.' Expertise refers to an individual's beliefs about the source's experience, qualifications and training that put them in a position to know the 64 truth. On the other hand, trustworthiness has an affective undertone and pertains to a person's assessments of whether the source is honest, open, disinterested and fair in its communication and how likely the source will be inclined to tell the truth as they see it (D. J. O'Keefe, 2002). These two dimensions find their counterparts in Frewer's (2003) definition of trust, which is codetermined by competence (i.e. the level of knowledge in a subject area) and honesty (i.e. the extent of truthfulness in sharing information in the subject area). It has become apparent that despite effort to conceptually distinguish among trust, confidence and credibility, they are treated as more or less similar constructs m many cases. Adding to the complexity of definition is the many types of trust and confidence that have been suggested by scholars. For example, Uslaner (2002, 2003) calls for a distinction between strategic trust (i.e. trust based on knowledge) and moralistic trust (i.e. trust based on perceived shared values with others) as well as for a differentiation between particularized trust (i.e. trust in within-group members) and generalized trust (i.e. trust in out-group members or people in general). Frewer and Salter (2007), on the other hand, propose a three-tier scheme for classifying trust. The first tier involves individualized trust, which is simply trust in another person or a product. For example, one might simply trust that vaccine is an effective product to prevent illness. At the second level, trust is system-oriented and directed toward institutions. For example, one might trust that the FDA would always ensure the effectiveness and safety of vaccines before approving them for public use. At the third level, trust is relational and results from personal experience of interacting with another person or group. For example, one 65 might trust the quality and safety of food purchased from a particular farm based on direct interaction with the farm's owner and workers. Alternatively, Earle eta!. (2007) derive two broad categories of trust from their literature review of key works on the topic: within group trust that can be further divided into interpersonal trust and social trust, and across-group trust that takes the form of general trust. These authors also suggest that the ideas of social trust and general trust can be used to develop their counterparts of 'social confidence' (i.e. the belief that institutions will continue to act in predictable and consistent ways based on their past performance) and 'general confidence' (i.e. the belief that things are generally under control and events will occur as expected). By looking across the aforementioned typologies of trust, it has become apparent that the concept of social trust (Earle eta!., 2007) is closely aligned with the idea of system-oriented trust (Frewer & Salter, 2007), where the trustee is an institution. On the other hand, interpersonal trust (Earle eta!., 2007) is similar to the construct of particularized trust (Uslaner, 2003) or relational trust (Frewer & Salter, 2007), where trust is largely based on direct experience of interacting with others. Furthermore, general trust (Earle eta!., 2007) can be considered identical with generalized trust (Uslaner, 2003), where one believes that people in general, including strangers, are trustworthy. While recognizing the multilayered nature of trust and the likelihood that various types of trust may play a role in health decision making under different circumstances, the present project focuses on social trust. This type of trust underpins individuals' faith in modern science, medicine and technology and in the ability of modern institutions to manage 66 these new developments. Therefore, social trust likely influences the making of all health decisions and should thus be given a place in theory. In addition, the present project focuses on health authorities as the target of social trust. While other institutions (e.g. pharmaceutical industry and scientific community) are also involved in the development, manufacturing and delivery of vaccines for public use, health authorities occupy a particularly interesting role. Like many agencies in the public sector, they are simultaneously the generator and regulator of vaccine-related risk (Frewer, 2003). They create risk by recommending the public to receive various vaccines that might have side effects on a small number of individuals, and at the same time they also regulate risk by seeking to ensure that new vaccines meet various efficacy and safety standards before giving them the seal of approval for mass marketing. Therefore, it is important to empirically investigate the extent to which individuals are willing to place their trust in health authorities despite their apparent conflict of interest and the potential impact of this trust on health decision making. So how can social trust potentially shape health decisions and behaviors? Siegrist, Cvetkovich, and Roth (2000) define social trust by emphasizing its impersonal nature, as something placed upon a "class of people ... with formal responsibilities within organizations who may not be personally known to the individual making the trust attribution" (p. 354). They argue that social trust is a key heuristic for lay individuals who need to make decisions about complex health risks on a daily basis in modern society: Social trust is the willingness to rely on those who have the responsibility for making decisions and taking actions related to the management of technology, the 67 environment, medicine, or other realms of public health and safety ... it is generally true that being able to determine who to trust is most important in those situations where the individual lacks the interest, time, abilities, knowledge, or other resources to personally make decisions and take actions. Science and technology are areas where many individuals seem to lack such resources. (Siegrist et a!., 2000, p. 354) In addition to functioning as a heuristic, Cvetkovich and Winter (2007) argue that social trust can be understood as a "social emotion," which operates in a way similar to affect. This perspective is echoed by Earle et a!. (2007), who posit that trust is experienced as a positive affect. It can therefore directly influence individuals' acceptance of an innovation just like affect. Trust can also indirectly impact acceptance by biasing systematic thinking and shaping the perceived benefit and perceived risk of the innovation in question (Earle eta!., 2007). Indeed, there is much empirical evidence demonstrating the key role of trust in influencing judgments and behaviors related to various risks. For example, Earle eta!. (2007) have advanced a Trust, Confidence and Cooperation (TCC) model, which posits that social trust and social confidence are two key heuristics influencing individuals' willingness to cooperate with the policies of companies, government agencies or other social institutions responsible for the creation and regulation of risks. This model has been tested in studies investigating people's response to environmental and health risks, such as Arctic oil exploration (Earle & Siegrist, 2006) and EMF produced by cell phone base stations (Siegrist, Earle, & Gutscher, 2003). Findings from those studies indicate 68 that both social trust and social confidence positively predict people's acceptance of health risks. Furthermore, there is a causal link from trust to confidence because trust consistently conditions individuals' confidence assessments. Researchers (Earle & Siegrist, 2008; Earle eta!., 2007) have also found that the influence of trust over cooperative behavior is greater when individuals are less familiar with the risk in question. This is perhaps not surprising given that individuals have less information and past experience to consult with in assessing the ability of institutions in managing novel risks. As a result, social trust is used as a heuristic in guiding confidence assessment and behavioral response to the risk. Beyond studies based on the TCC model, other research has demonstrated that trust can simultaneously shape risk-related decisions via both the direct and indirect routes. For example, Katsuya (2002) found that individuals with higher levels of knowledge about and interest in nuclear power judged the acceptability of this technology based on their perceptions of the risk and benefit associated with nuclear power, as well as their trust in the government agencies and electric power companies concerned. On the other hand, those with lower levels of knowledge about or interest in nuclear power did not take perceived risk into account when judging the acceptability of this technology and relied more heavily on social trust in making this assessment. Pursuing a similar line of research, Siegrist and colleagues have developed a Technology Acceptance Model based on their analyses of the influence of trust over individuals' response to nanotechnology, gene technology and food technology (Siegrist, 1999, 2000; Siegrist eta!., 2007; Siegrist eta!., 2000). These studies suggest that the effect of trust on acceptance behavior is either 69 fully (Siegrist, 1999, 2000; Siegrist eta!., 2000) or partially (Siegrist eta!., 2007) mediated through perceived risk and perceived benefit of the technology in question (see Figure 5). Furthermore, trust is found to be a common cause for risk and benefit perceptions in many circumstances, sometimes fully accounting for the correlation between these two variables. One explanation for the varying power of trust to account for the association between perceived benefit and perceived risk lies in the degree of knowledge individuals have with a particular technology, with the influence of trust being negatively associated with knowledge. This explanation is plausible given findings from research that compared various hazardous technologies and activities -trust turned out to be less predictive of perceived risk and perceived benefit and less able to explain their association in the case of familiar hazards compared to unfamiliar ones (Siegrist & Cvetkovich, 2000). In sum, while trust can play a decisive role in influencing many risk related decisions, its impact is dependent upon the nature of the target risk as well as the level of knowledge or familiarity an individual has about the target. 70 Figure 5. Technology Acceptance Model components and linkages Perceived risk Social trust Teclmology acceptance Perceived benefit Vaccination provides an interesting context for examining the effect of social trust on decision-making regarding a technological innovation that is charactetized by both familiarity and uncertainty. Most individuals have personal experience with vaccination and would consider it a familiar pharmacological product. It is therefore not surprising that a study conducted by Slovic in 1987 found that among 30 potentially hazardous technologies and activities, vaccination was considered one of the least risky hazards. However, the anti-vaccinationmovement, which gained momentum in the 1990s with events like the MMR vaccine controversy (Gross, 2009) and FDA's acknowledgement of mercury content in certain vaccines (Kirby, 2005), has increased public concem about vaccination. These events have also coincided with a rapid expansion of recommended vaccines for children and to a lesser extent for adults. Under these intersecting trends of growing uncertainty about vaccine and its increasing ubiquity, it is of interest to assess empirically the level of public trust in the institutional regulator of vaccine risk (i.e. the FDA in the US) and the extent to which this trust shapes individuals' perceptions, intention and behavior regarding vaccination. 71 Before presenting the hypotheses, it needs to be noted that items measuring trust in the FDA were included in the 2012 survey only. However, respondents were asked about their social confidence in public health authorities at the federal, state and local levels (i.e. the CDC, their state health department, and their local health department, respectively) in the 2009 survey. Consequently, social confidence is used in lieu of social trust in hypotheses and analyses pertaining to HlNl influenza with an acknowledgement that these constructs are closely related but not identical. The first three of the following hypotheses pertain to social trust and are thus tested with data collected in 2012 only, whereas the next four hypotheses pertain to social confidence and are tested using information collected in 2009 only: H8: Social trust in the FDA positively predicts perceived effectiveness of vaccine in general. H9: Social trust in the FDA negatively predicts safety concern over vaccine in general. HlO: Social trust in the FDA has a direct and positive effect on intention to get vaccinated against the hypothetical HSN 1 bird flu. Hll: Social confidence in public health authorities positively predicts perceived effectiveness of vaccine in general. Hl2: Social confidence in public health authorities negatively predicts safety concern over vaccine in general. Hl3(a): Social confidence in public health authorities has a direct and positive effect on intention to get vaccinated against the HlNl flu. Hl3(b ): Social confidence in public health authorities has a direct and positive effect on uptake of the seasonal flu vaccine during the HlNl flu pandemic. 72 In addition, a corollary to treating social trust as a social 'affect' is the need to investigate the potential relation between social trust and integral affect (Frewer, 2003). Siegrist et al. (2007) argue that the feeling of trust contributes to the overall affect toward a judgment target. They demonstrated this causal link in their study on people's acceptance of nanotechnology food and food packing. If this finding can be generalized to the context of vaccine acceptance, the following hypothesis should receive support: Hl4: Social trust in the FDA positively predicts individuals' affect toward flu vaccination. The aforementioned hypotheses pertaining to the influence of social trust over intention to vaccinate against the hypothetical H5Nl flu are summarized in Figure 6. The postulates regarding the impact of social confidence on intention to vaccinate against the HlNl flu and on seasonal flu shot uptake in 2009 are summarized in Figures 7 and 8, respectively. Figure 6. Social trust in predicting H5Nl flu vaccination intention in 2012 Perceived effectiveness HlO Safety concern Hypothetical H5Nl flu vaccination intention Figure 7. Social confidence in predicting HlNl flu vaccination intention in 2009 Perceived effectiveness Social confidence H13 a Safety concern HlNl flu vaccination intention Figure 8. Social confidence in predicting seasonal flu vaccination behavior in 2009 Perceived effectiveness Social confidence H13 Safety concern Seasonal flu vaccine uptake 73 74 Media Influence over Health Decision Making In reviewing the diverse array of studies focusing on the health effects of the mass media, Walsh-Childers and Brown (2009) propose that these effects can be understood along three dimensions: level of influence (personal/public), intention of the message producer regarding effects (intended/unintended), and outcome (positive/negative). To elaborate, the media can shape individuals' health-related perceptions and behaviors as well as public policies and public opinions about health issues. The producers of media content might intend to change the audience's beliefs and actions, such as in the case of advertising, public service announcements and entertainment-education programming. On the other hand, they might have no intention to exert any influence but nonetheless do, such as in the case of promoting unhealthy body images to girls by casting mostly skinny actresses in advertisements, films and television shows. Lastly, the individual- or public-level outcomes attributed to media influence are considered either positive or negative by public health professionals. In the present project, the focus is on unintended individual-level effects that are the outcomes of everyday exposure to health content in the news media, specifically in television news. This type of media effects deserves more attention, as it is less investigated and under-theorized compared to the concerted effort that has been devoted to studying intended media effects on health at the micro and macro levels (Rice & Atkins, 2001; Singhal, Cody, Rogers, & Sabido, 2004; Wallack, Dorfman, Jernigan, & Themba, 1993). There is a consensus among scholars that media effects are multifaceted, ranging from the effect of passive exposure or 'volume effect' (R. E. Kasperson eta!., 1988) to agenda-setting, priming and framing (Scheufele & Tewksbury, 2007) to cultivation (Gerbner, Gross, Morgan, Signorielli, & Shanahan, 2002), just to name a few. In addition, since the time ofMcLuhan (1964), there has been a myriad of studies demonstrating a 'channel effect,' where the unique features of different mass media channels are found to have differential influence over the audience beyond the content that they deliver. 75 With an acknowledgement of the diverse ways through which the mass media can influence audience's perceptions and behaviors, the present investigation focuses on the volume effect of television news on individuals' beliefs, intention and action regarding different types of flu vaccination. One reason for this focus is that compared to print, television news more frequently conveys information in the form of case studies or extreme examples that are more emotion-provoking, visual and memorable (McCabe & Fitzgerald, 1991; Shrum, 2002). Consequently, health content on television may have greater influence over the heuristic, affect-laden route of information processing while simultaneously having an impact on systematic processing. The current project's focus is also a result of the quantitative nature of the available data, as well as promising hypotheses that can be derived from existing literature for empirical test. As Stryker (2003) points out, the volume effect of news coverage on health behaviors can be considered on a temporal continuum. Intensive coverage of a 'media event,' such as a public figure's diagnosis with and treatment for cancer, HIV/AIDS or another illnesses, has been found to generate short-term effects on public awareness, interest, beliefs and behaviors pertaining to those diseases and their prevention (Brown & 76 Potosky, 1990; Kalichman, 1994; Stoddard, Zapka, & Schoenfeld, 1990). On the other hand, news coverage can work in a more gradual and cumulative manner, affecting trends in health perceptions and behaviors in the long term. For example, longitudinal research has shown that the rise and fall in the volume of news reports on health harming behaviors (e.g. illicit drug use, smoking and binge drinking) can influence public perceptions about those behaviors and the percentage of the population engaging in abstinence or cessation (Fan & Holway, 1994; Pierce & Gilpin, 2001; Stryker, 2003; Y anovitzky & Stryker, 200 1 ). So what are the potential mechanisms underlying the volume effect of health coverage in the media as demonstrated by these existing studies? Yanovitzky and Stryker (200 1) explain the direct effect of news coverage on health behavior in terms of Bandura's (2002) concept of social modeling, where people can learn how to perform a desirable health behavior or how to refrain from an undesirable behavior by just being repeatedly exposed to relevant news coverage. Other studies suggest that direct media influence over preventive health behavior can be explained in terms of cue to action (Chen & Murphy, 2012; Tang & Wong, 2004), as articulated in the extended HBM (Rosenstock et al., 1988). Unlike social learning that can lead to enduring behavior change, cue to action is likely to prompt immediate behavior change but is not expected to have a lasting behavioral impact. Leading to the 2009 data collection period for this research, there had been high volumes of media reports on HlNl influenza and its prevention for about six months. Consequently, the following hypotheses should hold regardless of whether it is social modeling or cue to action that lies behind the volume effects of exposure to mass media coverage: 77 HIS( a): Individuals with higher degrees of exposure to televised health news will have greater levels of intention to receive the HINI flu vaccine. HIS(b ): Individuals with higher degrees of exposure to televised health news will have greater levels of uptake of the seasonal flu shot. On the other hand, there was no media coverage of HSN I influenza vaccination during the 20 12 data collection, as this is a hypothetical vaccine that had not been manufactured for public use. Therefore, it is unlikely that exposure to health news on television would have a direct impact on individuals' intention to receive this vaccine. Nonetheless, the following question is empirically assessed: RQ3: Do individuals with higher degrees of exposure to televised health news on average have greater levels of intention to receive the HSNI flu shot? In addition to having a direct effect, the mass media may also influence vaccination intention or behavior indirectly by shaping individuals' perceptions regarding their susceptibility to the disease targeted by a vaccine and the severity of the possible consequences of this disease. Media influence over individuals' vulnerability perceptions has indeed been demonstrated by research. Studies have revealed that the news media give certain health risks, causes of death and types of crime coverage that is disproportional to their actual rates of occurrence (Combs & Slovic, 1979; Frost, Frank, & Maibach, 1997; Lowry, Nio, & Leitner, 2003; Miller & Reilly, 1995), and that this media over-representation is associated with audience members' overestimation of the 78 prevalence of these risks and their perceived personal susceptibility to these risks (Lowry et a!., 2003 ). One psychological mechanism that has been frequently cited to explain the association between media over-representation and audience's overestimation of personal vulnerability is the availability heuristic, which refers to "a judgmental heuristic in which a person evaluates the frequency of classes or the probability of events by availability, i.e., by the ease with which relevant instances come to mind" (Tversky & Kahneman, 1973, p. 207). Research has consistently demonstrated that human beings are cognitive misers who tend to rely on a small subset of all information that is in their memory when forming judgments (Eveland, 2002; Shrum, 2002, 2009), and that this subset contains pieces of information that are most frequently evoked, recently activated, or vivid and imaginable to the individual (Shrum, 2002, 2009; Tversky & Kahneman, 1973, 1974). Given the pervasiveness of the mass media in contemporary society and their tendency to provide vivid, dramatic representations of events, it has been suggested that events frequently portrayed in the media are easily retrievable in people's memory and are therefore often used in judgment making (Shrum, 2002, 2009). While this argument has received substantial support from studies that reveal a positive association between frequent exposure to news coverage or dramatic programming involving crime and individuals' crime-related vulnerability perceptions (Einsiedel, Salomone, & Schneider, 1984; Guo, Zhu, & Chen, 2001; Lowry eta!., 2003; Weaver & Wakshlag, 1986), empirical evidence also suggests that this association might be moderated by several factors. 79 For example, Tyler and Cook (1984) have advanced the "impersonal impact hypothesis" of media effects based on their research on perceived vulnerability toward crime. This hypothesis posits that mass-mediated information is likely to heighten individuals' societal-level vulnerability judgments (Tyler, 1980, 1984), whereas their personal-level vulnerability evaluations are primarily determined by their direct experiences as well as those of friends (Tyler, 1984). However, empirical evidence has questioned the generalizability of the impersonal impact hypothesis. For example, Stapel and Velthuijsen (1996) found that exposure to newspaper stories about a risk would heighten both societal-level and personal-level vulnerability assessments in so far as these stories included messages that were vivid or self-relevant to the readers. On the other hand, Weaver and Wakshlag (1986) demonstrated that when an individual had experienced crime exclusively through television representations, exposure to crime-related programs was positively associated with heightened concerns for personal safety and crime victimization. However, the effects of television exposure on crime perceptions lessened when individuals had experienced crime directly or through interpersonal conversations. In a related argument, Sjoberg and Engelberg (2010) suggested that in the context of technological risks, media influence over vulnerability perceptions was contingent upon the nature of mass-mediated information, such that media coverage had an impact on perceptions only when it focused on a previously unknown risk or transmitted a new piece of information about a known risk. What has emerged from a review of these studies on media influence over perceptions related to crime and hazardous technologies is that repeated exposure to media coverage of a risk 80 tends to have a stronger amplifying effect on vulnerability perceptions when individuals are unfamiliar with the risk in question and have no personal or interpersonal experience with it. The extent to which findings on media influence over perceptions of crime and hazardous technologies can apply to the health context is investigated in the current project. As a newly emerging disease with a pandemic potential that was later actualized, the H1N1 flu received media coverage unparalleled by any other disease or health condition between April and November 2009. In fact, it even managed to beat all other issues of national importance to become the most covered topic in the US news media for one week in late April 2009 (Kaiser Family Foundation & Pew Research Center, 2009). The novelty ofH1N1 influenza and the unprecedented media attention to this disease in 2009 provide the ideal conditions for testing the hypothesis that frequent exposure to health news on television during this time would lead to greater levels of perceived vulnerability toward H1N1 influenza: H16: Individuals with greater degrees of exposure to televised health news on average perceive higher levels of susceptibility to the H1N1 flu. Given the media's tendency to focus on mortalities, hospitalizations and serious cases in reporting diseases and taking into consideration criticisms directed at media hypes during public health emergencies (Vasterman, Yzermans, & Dirkzwager, 2005), exposure to news coverage of H 1 N 1 influenza is also likely to influence judgments regarding the severity of the H1N1 flu, such that: Hl7: Individuals with greater degrees of exposure to televised health news on average perceive the HlNl flu to have more serious consequences. 81 On the other hand, it is unlikely that exposure to health-related television content would have much influence on individuals' vulnerability perceptions regarding the H5Nl flu because there was no coverage of this hypothetical disease in the media during the data collection period in 2012. Nonetheless, the following questions are empirically investigated: RQ4: Do individuals' degrees of exposure to televised health news influence their levels of perceived susceptibility to the hypothetical H5Nl flu? RQS: Do individuals' degrees of exposure to televised health news influence their levels of perceived severity of the hypothetical H5Nl flu? Besides susceptibility and severity perceptions regarding various flus, it is of interest to investigate if the media also influence cost and benefit perceptions regarding vaccination. In the American context, the mass media have long been considered a culprit in raising public skepticism about vaccination by exaggerating the risk associated with certain vaccines or by giving anti-vaccinationists a voice (Bostrom & Atkinson, 2007; Heller, 2008; Poland & Jacobson, 2011 ). However, little research has been conducted to demonstrate that the media indeed hold sway over audience perceptions of vaccine. In the British context, Critcher (2007) considered the media's desire to provide 'balanced' coverage by giving voice to both sides of the MMR vaccine controversy a key factor behind a surge in the number of people believing that there was equal evidence for and against an alleged link between the MMR vaccine and autism. However, this claim has 82 not been supported by rigorous research. It is therefore the goal of the present project to investigate if frequent exposure to health-related media content has an impact on individuals' perceptions regarding the safety and effectiveness of vaccination in general. Recognizing that during the H 1 N 1 influenza pandemic, there was much greater media coverage on flu vaccination -mostly pertaining to the newly developed and relatively untested H1N1 flu vaccine (Mullen, 2009; Stein & Agiesta, 2009)- it is possible that exposure to televised health news would hold greater sway over individuals' vaccine related perceptions during the pandemic compared to outside of an influenza season. Therefore, the following research questions are investigated: RQ6(a): Does exposure to televised health news have an influence over safety concern regarding vaccine in general during the H1N1 influenza pandemic? RQ6(b ): Does exposure to televised health news have an influence over safety concern regarding vaccine in general outside of an influenza season in 2012? RQ7(a): Does exposure to televised health news have an influence over the perceived effectiveness of vaccine in general during the H1N1 influenza pandemic? RQ7(b ): Does exposure to televised health news have an influence over the perceived effectiveness of vaccine in general outside of an influenza season in 2012? Another route through which news coverage can indirectly influence individuals' health behavior is through changing norm perceptions surrounding a behavior. The validity of this pathway has been demonstrated in research showing positive and causal relationships among the volume of news coverage, the level of perceived social disapproval and the rate of abstinence from health harming behaviors (Stryker, 2003; 83 Y anovitzky & Stryker, 200 1 ). It is therefore of interest to investigate if exposure to health news has an impact on perceived social approval of vaccination behavior. In the present project, subjective norm was assessed in the 2012 survey only. However, given the lack of television coverage of the hypothetical H5N1 flu, exposure to health coverage is not expected to have an impact on subjective norm regarding H5N1 flu vaccination. Nonetheless, the following question is empirically assessed: RQ8: Does exposure to televised health news have an influence over subjective norm regarding H5N1 influenza vaccination? It is also of interest to investigate if exposure to televised health news has an impact on individuals' self-efficacy in getting different flu vaccines. Ideally health coverage in the media, especially in the regional and local media, should provide information on how and where to access health services and resources, including flu vaccination service. Such know-how information can have a positive impact on the audience's self-efficacy. Again, data on self-efficacy were collected in the 2012 survey only. However, in light of the lack of media coverage on the hypothetical H5N1 flu vaccine, exposure to televised health news is unlikely to have any impact on individuals' self-efficacy to get this vaccine. Nonetheless, the following question is empirically investigated: RQ9: Does exposure to televised health news have an influence over self-efficacy in getting the H5N1 influenza vaccine? 84 Last but not least, it is of import to investigate if news media influence major heuristics -social confidence, social trust and integral affect- that are previously hypothesized to influence health decision making. Luhmann (2000) suggests that the news media play a critical role in building or destroying trust in this day and age because in lieu of personal interaction, individuals have to judge the competence and goodwill of institutions largely based on information obtained through the mass media. Renn and Levine (1991) further suggest that exposure to news coverage tends to have a negative rather than positive impact on institutional trust and confidence, given the media's tendency to focus coverage on incidents pertaining to institutional abuses and breaches of trust. Therefore, the current project investigates the potential role of exposure to health news on television in shaping individuals' levels of trust in the FDA, the federal health agency responsible for approving and monitoring vaccines and drugs for public use: RQ10: Does exposure to televised health news have an influence over social trust in the FDA? Although data collected during the H1N1 flu pandemic did not include measures of social trust, items were available to assess social confidence in health authorities. It is plausible that exposure to health news on television would influence individuals' confidence assessments of these institutions, and therefore the present project asks: RQ 11: Does exposure to televised health news have an influence over individuals' social confidence in health authorities? 85 Furthermore, if trust is considered a 'social affect' that is partially shaped by media coverage, it is possible that intrapersonal affect toward vaccination can also be influenced by exposure to television news, especially given its emotion-laden imagery and sound. RQ12: Does exposure to televised health news have an influence over individuals' affect toward flu vaccination? Until this point, media effects have been considered from the perspective of passive exposure to television news. It has long been acknowledged that mass communication research is often based on two opposing conceptions of the audience: passive versus active. Theories like cultivation (Gerbner eta!., 2002) and agenda-setting (Dearing & Rogers, 1996; McCombs & Shaw, 1972) tend to consider media audiences as passive recipients who are unwittingly influenced by heavy exposure to media messages, whereas theories such as uses and gratifications (Katz, Blumler, & Gurevitch, 1973; Rubin, 1993, 2002) and media systems dependency (Ball-Rokeach, 1985) accord an active role to the audience in actively selecting, interpreting and retaining media content based on their preexisting goals and expectations. While acknowledging that there are many types of audience activities that might function as intervening mechanisms through which the media come to influence the audience (Biocca, 1988; Rubin, 2009), it is beyond the scope of the present project to provide a comprehensive examination of audience activities regarding health news on television. Rather, the current research focuses on the audience activity of selective attention to news content. Research has indicated that active attention has an effect over and beyond passive exposure in amplifying media influence over knowledge gain (Chaffee & Schlender, 1986) and in 86 strengthening the framing effect (Scheufele & Tewksbury, 2007). With regard to risk perceptions, studies have found that a greater degree of active attention to crime-related programming on television predicts a higher level of perceived crime rates and personal fear of crime, controlling for the frequency of passive exposure to crime programming (G. J. O'Keefe, 1984; G. J. O'Keefe & Reid-Nash, 1987). Therefore, data were collected on how much attention individuals pay to news coverage on H1N1 influenza during the 2009 data collection period to test the following hypotheses: H18: Controlling for passive exposure to televised health news, individuals who pay more attention to H1N1-related news on average perceive higher levels of susceptibility to the H1N1 flu. H19: Controlling for passive exposure to televised health news, individuals who pay more attention to H1N1-related news on average perceive the H1N1 flu to have more severe consequences. H20: Controlling for passive exposure to televised health news, individuals who pay more attention to H1N1-related news on average have greater levels of intention to receive the H 1 N 1 flu shot. H21: Controlling for passive exposure to televised health news, individuals who pay more attention to H1N1-related news on average have greater levels of uptake of the seasonal flu shot during the H1N1 flu pandemic. It is also of interest to investigate if active attention to H1N1-related news shapes individuals' beliefs about vaccine in general and their social confidence in health authorities. The news media had regularly reported on government actions and 87 recommendations in curbing the novel flu during the H1N1 influenza pandemic, and they also devoted much attention to the H1N1 vaccine, due to its relatively new and untested nature (Geracimos, 2009; Mullen, 2009; Stein & Agiesta, 2009; Stein & Laris, 2009). Consequently, the following questions are answered empirically: RQ13: Controlling for passive exposure to televised health news, does attention to H1N1-related news have an influence over the perceived effectiveness of vaccine in general? RQ 14: Controlling for passive exposure to televised health news, does attention to H1N1-related news have an influence over safety concern regarding vaccine in general? RQ 15: Controlling for passive exposure to televised health news, does attention to H1N1-related news have an influence over social confidence in health authorities? The aforementioned hypotheses and research questions regarding the influence of passive exposure to televised health news over intention to vaccinate against the H1N1 flu and over actual uptake of the seasonal flu shot in 2009 are present in Figures 9 and 10, respectively. Figure 11 and 12 summarize the hypotheses and research questions pertaining to the effects of active attention to H1N1-related news on H1N1 flu vaccination intention and on the uptake of seasonal influenza vaccine in 2009, respectively. For the sake of clarity, only causal paths leading from passive exposure are included in Figures 9 and 10, including a causal link from passive exposure to active attention to H1N1-related news because this is likely to be the relationship between these two variables. In Figure 13, the hypotheses and research questions pertaining to the 88 influence of passive exposure to televised health news over intention to vaccinate against the hypothetical H5Nl flu are presented. To sum up, the present project seeks to advance and test a multi-stage model of vaccination decision making, with exposure to health news on television being the first stage, integral affect and social confidence/social trust occupying the second stage, susceptibility and severity perceptions constituting the third stage, perceived effectiveness and perceived safety of vaccine as well as social norm and self-efficacy comprising the fourth stage. The final stage is the decision to vaccinate against a particularly flu. Variables at an earlier stage are hypothesized to directly or indirectly predict variables at subsequent stages depending on a number of contextual factors. Furthermore, in the context of the HlNl flu, data are available to test the health effects of active attention to HlNl-related news over and beyond the impact of passive exposure to televised health news. Figure 9. Influence of passive media exposure over HlNl flu vaccination intention in 2009 Exposure to health Attention toHlNl news H15 a confidence Perceived e:ff ecti v e ness Safety concern HlNl flu vaccination intention 89 Figure 10. Influence of passive media exposure over seasonal flu vaccination behavior in 2009 Exposure to health news H15 confidence Perceived e:ff ecti v e ness Safety concern Seasonal flu vaccine uptake Figure 11. Influence of active attention to HlNl news over HlNl flu vaccination intention in 2009 Attention toHlNl news confidence Perceived e:ff ecti v e ness Safety concern HlNl flu vaccination intention Figure 12. Influence of active attention to Hl Nl news over seasonal flu vaccination behavior in 2009 Attention toHlNl news confidence Perceived e:ff ecti v e ness Safety concern Seasonal flu vaccine uptake 90 Figure 13. Influence ofpassive media exposme over H5N1 flu vaccination intention in 2012 Exposure to health news Perceived susceptibi· lity Perceived effective- Subjective norm Self· efficacy H5Nl:Ou vaccination intention 91 92 CHAPTER 5: RESEARCH DESIGN Procedures This study utilizes data collected through the Annenberg National Health Communication Survey (ANHCS). ANHCS is an ongoing cross-sectional survey of adults in the US on topics pertaining to health-related knowledge, perceptions, behaviors, policy preferences and media exposure. Knowledge Networks, a web-based research company, is responsible for sample recruitment and survey administration on behalf of the creators of the ANHCS instrument- the Aunenberg Schools at the University of Pennsylvania and the University of Southern California ("ANHCS 2007 User's Guide & Codebook.," 2009). Although ANHCS participants are asked to fill out the survey online, they were initially selected through random-digit dialing. Recruits without Internet access are provided the necessary hardware and connection service to enable their participation in the survey. Since April 2005, between 250 and 350 individuals selected from a previously established research panel designed to be representative of the US adult population have responded to ANHCS each month. In October and November 2009, a module was added to the core section of ANHCS to collect information on respondents' perceptions and behaviors related to the HlNl flu. A total of 626 individuals responded to ANHCS during these two months. In April and May 2012, another module was added to the core section of ANHCS to collect data on respondents' perceptions and behaviors related to seasonal influenza and a hypothetical H5Nl bird flu. A total of 646 individuals responded to ANHCS during this period. 93 Participants Participants with missing values on any variables of interest were removed before statistical analyses, resulting in a final sample size of 571 individuals in the 2009 dataset and 584 in the 2012 dataset. Table 1 summarizes the socio-demographic characteristics of the final samples. The two samples are similar in all characteristics, but compared to the overall composition of the US population (US Census Bureau, 2013 ), the samples include individuals who are older and more educated. In addition, non-Hispanic Whites make up a larger proportion of the samples while minorities comprise a smaller percentage compared to the population as a whole. Those discrepancies are due to non-response, non-coverage and other problems common to RDD sampling ("ANHCS 2007 User's Guide & Code book.," 2009). To adjust for those deviations, post-stratification sample weights were created by Knowledge Networks using the age, sex, race/ethnicity and education benchmarks from the US Census Bureau, and those weights were subsequently applied to the sample data before preforming further statistical analyses. 94 Table I. Socio-demographic Characteristics of Samples Compared to US Population 2009 2012 us Sample Sample population Characteristics (N~57J) (N~584) Male 46.7% 46.3% 49.2%* Race Non-Hispanic White 77.8% 77.1% 63.7%* Non-Hispanic Black 7.5% 8.2% 12.2%* Hispanic 8.2% 7.6% 16.3%* Other 6.5% 7.1% 7.8%* Median age 50 51 37* Median income 55,000 55,000 52,7621 Education Less than high school 9.1% 8.2% 14.6% 1 High school graduate 28.4% 28.6% 28.6% 1 Some college 31.3% 27.8% 28.6%1 Bachelor's degree or higher 31.2% 35.4% 28.2% Note. *Data from the 2010 Census. Tnata from the 2007-2011 American Community Survey 5-year estimates. Measures Table 2 presents the endogenous variables that are available from the 2009 and 2012 surveys, respectively. The exact items used to measure all these variables are provided in Appendices A and B. Strictly endogenous variables from the 2009 survey. Intention to receive the HJNJ flu shot. Respondents were asked to indicate how likely they were to get a free HINI shot when it becomes available to them on a 10-point scale, ranging from I (not likely at all) to 10 (extremely likely). Sixteen respondents reported having already been vaccinated against the HI N I flu, and their responses were subsequently recoded as I 0. 95 Table 2. Endogenous Variables from the 2009 and 2012 Surveys and Their Means and Standard Deviations Variables 2009 Survey (N~571) 2012 Survey (N~584) M SD M SD Outcome HlNl vaccination intention y 5.16 3.43 N variables Seasonal flu vaccine uptake y 30.2% N in current flu season Seasonal flu vaccine uptake N y 40.5% in previous flu season H5Nl vaccination intention N y 4.84 3.11 Endogenous Perceived effectiveness of y 3.72 092 y 3.65 0.97 variables vaccine in general Safety concern over vaccine y 3.11 0.90 y 3.23 0.93 in general Social norm regarding N y 9.23 13.79 seasonal flu vaccination Social norm regarding N y 6.44 14.35 H5Nl flu vaccination Self-efficacy for getting N y 7.46 2.89 seasonal flu vaccine Self-efficacy for getting N y 5.80 2.89 H5Nl vaccine Perceived susceptibility to y 4.15 2.16 N HlNl flu Perceived susceptibility to N y 4.74 2.47 seasonal flu Perceived susceptibility to N y 4.25 2.33 H5Nl flu Perceived severity ofHlNl y 3 09 2.14 N flu Perceived severity of N y 603 2.89 seasonal flu Perceived severity of H5Nl N y 6.5 2.99 flu Social trust in FDA N y 9.24 2.64 Social confidence in health y 909 2.48 N authorities Integral affect toward flu N y 5.92 3.30 vaccination Passive exposure to y 2.38 107 y 2.18 106 televised health news Active attention to HlNl- y 4.17 1.41 N related news Note. Y and N indicate yes and no, respectively. 96 Uptake of seasonal flu vaccine in the current flu season. Participants were asked to give a yes/no answer to the question: "Have you been vaccinated against the flu for the most recent winter flu season (i.e. starting in October 2009)?" Respondents with an affirmative answer were coded as 1, and those with a negative answer were coded as 0. Endogenous variables from the 2009 survey. Perceived effectiveness of vaccine in generaL Respondents were asked to indicate their level of agreement to the statement "in general, getting vaccinated is an effective way to prevent illness." Answers were given on a 5-point scale, ranging from 1 (strongly disagree) to 5 (strongly agree). Safety concern over vaccine in generaL Respondents were asked to indicate their level of agreement to the statement "in general, getting vaccinated can make you sick." Answers were given on a 5-point scale, ranging from 1 (strongly disagree) to 5 (strongly agree). Perceived susceptibility to HJNJ flu. Respondents were asked "how likely are you to catch the HlNl flu between now and the end of March 2010," and the answer was given on a 10-point scale, ranging from 1 (not likely at all) to 10 (extremely likely). Perceived severity of HJNJ flu. Respondents were asked "if you did catch the HlNl flu, how likely are you to die," and the answer was given on a 10-pont scale, ranging from 1 (not likely at all) to 10 (extremely likely). Social confidence in health authorities. Three items were used to assess confidence levels in health authorities' capabilities to deal with the HlNl flu at the local, state and federal levels. Respondents were asked how much confidence they have in each 97 of the following agencies' ability to deal effectively with the HlNl flu, including (a) the Centers for Disease Control and Prevention or CDC, (b) the health department of their state government, and (c) the health department of their local government. Answers were given on a 4-point scale, ranging from 1 (not confident at all) to 4 (very confident). To assess if these three items can be explained by a single underlying factor, a confirmatory factor analysis was conducted using LISREL 9.1. All paths in this saturated model were significant at p < .05, and all factor loadings were satisfactory (A.x~ .67 for confidence in the CDC, A.x~ .82 for confidence in state health department, and A.x~ .87 for confidence in local health department). Consequently, each respondent's summative score on the three items was computed and used in further analyses. Passive exposure to televised health news. Passive exposure was measured with the question: "Some local or national television news programs include special segments of their newscasts which focus on health issues. About how often have you watched such health segments in the past 30 days?" Answers were given on a 4-point scale, ranging from 1 (not at all) to 4 (a few times a week). Attention to HJNJ flu-related news. Respondents were asked to indicate how often they have seen or heard any news about the HlNl flu during the past 30 days on a 6-point scale, ranging from 1 (not at all) to 6 (almost every day). Strictly endogenous variables from the 2012 survey. Intention to receive a hypothetical HSNJ bird flu shot. The following statement introduced the question on H5Nl flu vaccination intention: "Since 2003, H5Nl avian influenza, also known as the bird flu, has sickened individuals in 15 countries and led to hundreds of deaths. Public health authorities are concerned that avian influenza might spread to the US this coming winter flu season (between October 2012 and March 2013)." Respondents were then asked "if a vaccine against avian influenza was developed and approved for use by September 2012, how likely are you to get this new vaccine for the coming flu season, assuming that it is readily available and free?" Answers were indicated on a 10-point scale, ranging from 1 (not likely at all) to 10 (extremely likely). 98 Uptake of seasonal flu vaccine in the previous flu season. Participants were asked to give a yes/no answer to the question: "Have you been vaccinated against the flu for the most recent winter flu season (i.e. starting in October 2011 )?" Respondents with an affirmative answer were coded as 1, and those with a negative answer were coded as 0. Endogenous variables from the 2012 survey. Perceived effectiveness of vaccine in generaL Respondents were asked to indicate their level of agreement to the statement "in general, getting vaccinated is an effective way to prevent illness." Answers were given on a 5-point scale, ranging from 1 (strongly disagree) to 5 (strongly agree). Safety concern over vaccine in generaL Respondents were asked to indicate their level of agreement to the statement "in general, getting vaccinated can make you sick." Answers were given on a 5-point scale, ranging from 1 (strongly disagree) to 5 (strongly agree). 99 Subjective norm regarding seasonal flu vaccination. Three referent groups considered important in influencing individuals' vaccination behavior (de Wit, Vet, Schutten, & van Steenbergen, 2005; Gallagher & Povey, 2006), including one's family members, friends, and health care provider, were used to construct two sets of items for measuring subjective norm: normative belief (NB) items and motivation to comply (MC) items. The NB items assessed the extent to which the respondents thought that each referent group would approve of them getting the seasonal flu vaccine for the coming flu season. Answers were given on a 10-point scale, ranging from I (not at all) to 10 (a great deal). They were subsequently transformed into five points on a bipolar scale, ranging from -2 to +2, according to convention (Montano & Kasprzyk, 2008). MC items assessed the extent to which each referent group's opinion on health mattered to the respondents. Answers were given on a 10-point scale, ranging from I (not at all) to 10 (a great deal). They were subsequently collapsed into five points on a unipolar scale, ranging from I to 5, according to the convention that dictated the NB items and MC items to have the same number of points on their respective scales (Montano & Kasprzyk, 2008). For each referent group, a respondent's score on the NB item was multiplied by his or her score on the corresponding MC item to obtain a composite score. Finally, each respondent's composite scores on the three referent groups (i.e. family members, friends and health care provider) were added together to create a summative score that reflected the respondent's overall subjective norm regarding seasonal flu vaccination. Subjective norm regarding HSNJ flu vaccination. The same referent groups and the same computation procedures were used to assess individuals' overall subjective 100 norm pertaining to getting the hypothetical HSN 1 flu shot. The NB items asked respondents to imagine that they are getting the newly-developed avian flu vaccine for this coming flu season and to indicate on a 10-point scale how much they think each of the following people would approve of them getting this vaccine: (a) your family members, (b) your friends, and (c) your health care provider. The M C items asked respondents to indicate on a 10-point scale how much each referent group's opinion on health mattered to them. Self-efficacy regarding seasonal flu vaccination. Respondents were asked how confident they were that they could get the seasonal flu vaccine in the coming winter flu season if they wanted to. Answers were indicated on a 10-point scale, ranging from 1 (not at all confident) to 10 (extremely confident). Self-efficacy regarding HSNJ flu vaccination. Respondents were asked how confident they were that they could get the hypothetical H5Nl flu vaccine in the coming winter flu season if they wanted to. Answers were indicated on a 10-point scale, ranging from 1 (not at all confident) to 10 (extremely confident). Perceived susceptibility to seasonal flu. Respondents were asked "how likely are you to catch seasonal influenza this coming flu season if you didn't get the seasonal flu vaccine," and the answer was given on a 10-point scale, ranging from 1 (not likely at all) to 10 (extremely likely). This is a conditional question (i.e. susceptibility conditioned upon not having received the seasonal flu vaccine) because scholars (N. T. Brewer et al., 2007) have argued that unconditional susceptibility items tend to underestimate the relationship between susceptibility and preventive behavior. 101 Perceived susceptibility to HSNJ flu. Respondents were first told to "[t]hink about the coming flu season (between October 2012 and March 2013) and imagine that avian influenza has indeed spread to the US." They were then asked "how likely are you to catch seasonal influenza this coming flu season if you didn't get the seasonal flu vaccine." Answer was given on a 10-point scale, ranging from 1 (not likely at all) to 10 (extremely likely). Again, this is a conditional question that follows the methodological advice ofN. T. Brewer et al. (2007). Perceived severity of seasonal flu. Respondents were asked to think about their ability to carry out your usual activities and assess how much their usual activities would be affected if they got seasonal influenza this coming winter flu season. Answers were indicated on a 10-point scale, ranging from 1 (not at all disrupted) to 10 (extremely disrupted). Perceived severity of HSNJ flu. Respondents were asked to think about their ability to carry out your usual activities and assess how much their usual activities would be affected if they got the hypothetical bird flu this coming winter influenza season. Answers were indicated on a 10-point scale, ranging from 1 (not at all disrupted) to 10 (extremely disrupted). Social trust in FDA. To ensure that respondents were aware of the responsibility of the FDA as the federal regulator of vaccines, the following explanation preceded the four items measuring trust in the FDA: "The Food and Drug Administration (FDA) is the federal government agency responsible for approving and regulating vaccines and medicines in the US. Please indicate your level of agreement with the following 102 statements." The proceeding statements included two items adopted from Siegrist et al. (2003), one item adopted from Siegrist et al. (2000), and one item developed by the investigator of the current project. Respondents were asked to indicate their level of agreement with the following statements: (a) the FDA can be trusted, (b) the FDA communicates honestly about possible side effects of vaccines; (c) should it turn out that a vaccine is harmful, the FDA would openly and honestly inform the public; and (d) should it turn out that a vaccine is harmful, the FDA would protect the interests of pharmaceutical companies. Answers were indicated on a 5-point scale, ranging from 1 (strongly disagree) to 5 (strongly agree). To assess if these four items can be explained by a single underlying factor, a confirmatory factor analysis was conducted using LISREL 9.1. Although all paths in this model were significant at p < .05, only items a, band chad a satisfactory factor loading (A.x~ .80 for item a, A.x~ .94 for item b, and A.x~ .77 for item c). Factor loading for item d was relatively low (A.x~ -.32), and this item was subsequently dropped when computing each respondent's additive score on the social trust items to be used in further analyses. Integral affect toward flu vaccination. Respondents were asked to indicate their feeling toward getting vaccinated for the flu on a 1 0-point scale, ranging from 1 (extremely negative) to 10 (extremely positive). Passive exposure to televised health news. Passive exposure was measured with the question: "Some local or national television news programs include special segments of their newscasts which focus on health issues. About how often have you watched such health segments in the past 30 days?" Answers were given on a 4-point scale, ranging from 1 (not at all) to 4 (a few times a week). Exogenous variables. 103 The exogenous variables are hypothesized to predict all endogenous variables in the theoretical models of HlNl flu vaccination intention, seasonal flu vaccination behavior and H5Nl flu vaccination intention. Socio-demographic variables. In both the 2009 and 2012 surveys, respondents were asked about their age, gender, race/ethnicity, education level, and annual household income. These socio-demographic characteristics have been found to influence individuals' uptake of various flu vaccines (CDC, 20lla, 2012a), the extent to which they are susceptible to different flus and to severe complications from these flus (CDC, 2009c, 20llb), their level of social confidence (Lipset & Schneider, 1983) and their frequency of watching televised news (The Pew Research Center, 2013). Consequently, they were included in all statistical analyses. In the core section of ANHCS, respondents were asked to provide their age as a numerical number and select from a number of categories their education attainment and annual household income. Mid-point transformation was originally performed for both education and income to turn them into a continuous variable. However, the transformation of income into a continuous variable created a multicollinearity problem in further statistical analyses, even with the computation of an asymptotic covariance matrix and polychoric correlations. Thus, income was maintained as a categorical variable, with eight categories (i.e. less than 25,000, between 25,000 and 49,999, between 104 50,000 and 74,999, between 75,000 and 99,999, between 100,000 and 124,999, between 125,000 and 149,999, between 150,000 and 174,999, and 175,000 or more). Furthermore, to facilitate subsequent analyses, gender and race/ethnicity were both dummy-coded, such that male respondents were coded as 1 and females as 0, and that non-Hispanic White respondents were coded as 1 while all other races/ethnicities were coded as 0. Risk factors for influenza-induced complications. In both the 2009 and 2012 surveys, respondents were asked if they have suffered from a range of health conditions, including (a) diabetes, (b) HIV/AIDS, (c) heart disease, (d) stroke, (e) cancer, (f) kidney disease, (g) lung or breathing problems, and (h) obesity (BMI of 40 or higher). These conditions put individuals at a higher risk for developing severe complications from infection with seasonal or novel flus (CDC, 2009c ). An affirmative response to each of these conditions was coded as 1, whereas a negative response was coded as 0. A summative score was computed for each respondent by adding up his or her scores on all these health conditions. Analysis The hypothesized relationships among the endogenous and exogenous variables are summarized in Figure 4 and Figures 6 to 13. The models for seasonal flu vaccine uptake in 2009, H1N1 flu vaccination intention in 2009 and H5N1 bird flu vaccination intention in 2012 were tested using structural equation modeling (SEM) techniques. The correlates for seasonal flu vaccine uptake in the 2011-2012 influenza season were identified using hierarchical logistic regression analysis because the temporal connection between the outcome variable (i.e. seasonal flu vaccine uptake in the previous flu season) 105 and the predictor variables did not allow modeling of causal relationships. Rather, regression analysis was used to assess the associations among these variables to enable suggestive observations to be made. For data analyzed using SEM, they were first screened with PRE LIS 2. The vaccination intention variable for HlNl and H5Nl flus were found to have a bi-modal distribution, and the 'risk factors for influenza-induced complications' variable was found to be positively skewed in both the 2009 and 2012 datasets. Given the non-normal distribution of these variables and the dichotomous nature of variables such as gender, race/ethnicity and seasonal flu vaccine uptake, PRE LIS 2 was used to prepare data for analysis with LISREL 9.1. The polychoric correlations and asymptotic covariance matrix computed by PRELIS were subsequently input into LISREL 9.1, and the weighted least squares method was used for model estimation. The polychoric correlations used in those analyses are presented in Tables 3 and 4. The goodness-of-fit of the hypothesized models to the observed data were assessed in three ways. First, the significance of the chi-square statistic was examined, with a non-significant i suggesting a good fit. Given that the i statistic is sensitive to sample size, Wheaton, Muthen, Alwin, and Summers (1977) offer the rule of thumb of considering a ito degrees-of-freedom ratio of 5 or smaller indicative of a satisfactory fit. Second, the goodness-of-fit index (GFI) was examined, with values 2.90 suggesting a good fit (Hu & Bentler, 1995). Lastly, the standardized root mean squared residual (SRMR), which reflected the average discrepancy between the observed and hypothesized correlations among the variables, was inspected. A SRMR S: .08 suggests a 106 satisfactory fit (Hu & Bentler, 1999). In cases where any of these measures indicated a less than acceptable fit, the theoretical model was revised based on the modification indices provided by LISREL. A modification was made when it was theoretically plausible, would result in a large reduction in chi-square, and was linked to a change in path coefficient of considerable magnitude. The significance of individual parameters were assessed using the ttests with the significance level set at p < .05, and these parameters were presented as standardized coefficients to enable their comparison with one another. The coefficient of determination reflecting the amount of variance in the strictly endogenous variable explained by all hypothesized relationships in each model was also inspected. For data analyzed using hierarchical logistic regression, they were first screened with SPSS 19.0. The 'risk factors for influenza-induced complications' variable was found to be positively skewed (skewness~ 2.58) so square-root transformation was performed on this variable. The transformed, normally-distributed variable was used in subsequent statistical analyses. For these analyses, the independent variables were entered in three blocks (i.e. a block of socio-demographic and physiological variables, a block of variables from established health behavior theories including perceived susceptibility, perceived severity, perceived effectiveness of vaccine and safety concern, subjective norm and self-efficacy, and the final 'heuristic' block of health news exposure, social trust and integral affect). The significance level for all statistical tests was set at p < .05, and all unstandardized coefficients and their odds ratios were presented to enable comparison. The R -squared and chi -square statistics indicating the amount of variance in 107 the outcome variable explained by the regression and the statistical significance of each additional block in improving the amount of variance explained were also inspected. 108 Table 3. Polychoric Correlations among Variables from the 2009 Survey 2 3 4 5 6 7 8 9 10 II 12 13 14 15 I. Exposure to televised health news 1.00 2. Seasonal flu shot uptake 0.12 1.00 3. Perceived effectiveness of vaccine 0.07 0.28 1.00 4. Safety concern over vaccine 0.00 -0.28 -0.19 1.00 5. Perceived susceptibility to HINI flu 0.01 0.02 0.03 0.07 1.00 6. Perceived severity ofHINI flu 0.07 0.13 -0.02 -0.02 0.29 1.00 7. HlNl flu vaccination intention 0.06 0.37 0.34 -0.27 0.27 0.28 1.00 8. Attention to BINI-related news 0.26 0.22 0.08 -0.06 0.13 -0.02 0.04 1.00 9. Age 0.20 0.37 0.19 -0.07 -0.12 0.03 0.02 0.22 1.00 10. White 0.00 0.25 0.22 -0.09 -0.07 -0.23 -0.02 0.17 0.34 1.00 II. Male -0.08 -0.07 -0.14 -0.04 -0.01 -0.05 -0.02 -0.09 -0.14 -0.08 1.00 12. Education -0.05 0.20 0.19 0.06 -0.01 -0.29 -0.06 0.13 0.12 0.41 0.00 1.00 13. Income -0.13 0.04 0.11 0.05 0.01 -0.16 -0.13 0.05 0.00 0.25 0.06 0.43 1.00 14. Risk factors for flu complications 0.20 0.06 0.19 0.02 -0.03 0.16 0.12 -0.07 0.26 0.04 -0.07 -0.12 -0.14 1.00 15. Social confidence 0.04 0.16 0.30 -0.15 -0.09 0.00 0.18 -0.12 0.04 0.06 -0.04 0.04 0.01 0.07 1.00 109 Table 4. Polychoric Correlations among Variables from the 2012 Survey 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1. Exposure to televised health news 1.00 2. Integral affect toward flu vaccination 0.13 1.00 3. Perceived effectiveness of vaccme 0.13 0.49 1.00 4. Safety concern over vaccine -0.12 -0.33 -0.19 1.00 5. H5Nl flu vaccination intention 0.06 0.66 0.37 -0.27 1.00 6. Self-efficacy for H5Nl flu vaccination 0.04 0.20 0.21 -0.07 0.29 1.00 7. Perceived susceptibility to H5Nl flu 0.07 0.48 0.34 -0.13 0.56 0.32 1.00 8. Perceived severity ofH5Nl flu 0.09 0.25 0.26 0.08 0.31 0.22 0.37 1.00 9.Age 0.25 0.24 0.22 -0.06 0.20 0.12 0.21 0.19 1.00 10. White -0.05 0.14 0.33 -0.10 0.09 0.18 0.10 0.11 0.26 1.00 11. Male -0.11 -0.01 -0.06 -0.02 0.05 0.05 0.04 0.03 0.04 0.09 1.00 12. Education 0.00 0.10 0.24 -0.15 0.18 0.12 0.06 0.08 0.02 0.15 0.14 1.00 13. Risk factors for flu complications 0.25 0.24 0.25 0.03 0.14 0.09 0.23 0.15 0.39 0.16 -0.03 -0.09 1.00 14. Subjective norm for H5Nl flu vaccination 0.12 0.45 0.40 -0.20 0.53 0.35 0.41 0.25 0.18 0.27 0.01 0.15 0.08 1.00 15. Income -0.04 0.06 0.18 -0.03 0.05 0.18 0.07 0.01 0.02 0.23 0.07 0.38 -0.04 0.14 1.00 16. Social trust 0.01 0.27 0.37 -0.21 0.25 0.09 0.16 -0.02 0.01 0.09 0.01 0.05 -0.08 0.31 0.07 1.00 110 CHAPTER 6: RESEARCH RESULTS Modeling HlNl Flu Vaccination Intention in 2009 To test Hl(a) and H2(a), which postulated that the effects of perceived severity and perceived susceptibility on HlNl flu vaccination intention were mediated through perceived effectiveness of vaccine, the theoretical model was first fitted to the data without the two mediational paths. Perceived severity (b ~ .19, p < .05), perceived susceptibility (b ~ .23,p < .05) and perceived effectiveness of vaccine (b ~ .29,p < .05) were all found to be significant predictors of vaccination intention. Subsequently, the mediational paths were added to the model. However, both additional paths were found to be non-significant, and the magnitude of the coefficients indicating the respective direct effects of perceived susceptibility, perceived severity and perceived effectiveness on vaccination intention did not change. Therefore, Hl(a) and H2(a) were rejected. This initial model had an acceptable GFI at .93 and SRMR at .03. However, the chi-square statistic indicated an unsatisfactory model fit (x 2 /df ~ 80.97/4). Thus, the modification indices were examined to identify plausible adjustments to the model. The modification indices suggested that the largest reduction in model x 2 would result from adding a causal path between perceived susceptibility and perceived severity or by allowing the error term of these two variables to correlate. As discussed at the end of Chapter 3, perceived susceptibility and perceived severity could be both tied to an underlying factor that can be conceptualized as 'perceived threat' (Weinstein, 1993). Consequently, the error term of these two variables were allowed to correlate to account for this potential common cause. With this additional path, the model's global fit 111 statistics improved substantially but were still not completely satisfactory (x 2 /df~ 21.00/3, SRMR ~ 0.02, GFI ~ 1.00). A further examination of the modification indices indicated that allowing the error term of perceived effectiveness and safety concern regarding vaccine to correlate would result in the largest reduction in model x 2 This addition was justifiable based on research showing a correlation between individuals' benefit and risk perceptions when asked to assess a hazardous technology or activity (Ganzach, 2000; Siegrist et al., 2007). Previous studies have suggested that this correlation can be largely explained by integral affect (Alhakami & Slovic, 1994; Finucane et al., 2000) and social trust (Siegrist, 1999, 2000; Siegrist et al., 2007; Siegrist & Cvetkovich, 2000; Siegrist et al., 2000) as the common underlying factors. Even though the present analysis included social confidence as a proxy measure for social trust, it did not include integral affect. Therefore, the correlation between perceived effectiveness and perceived risk was unlikely to be fully explained by variables in the model. After adding this correlational path, all global fit indices became satisfactory (x 2 /df~ 2.72/2, SRMR ~ 0.01, GFI ~ 1.00). An examination of the modification indices suggested that no other modification would significantly improve the model fit. Thus, non-significant paths were removed from this revised model one by one. After all non-significant paths were eliminated, the model fit remained good (x 2 /df ~ 53.12/47, SRMR ~ 0.03, GFI ~ .99), and there was no substantial change in the magnitude of the significant parameter estimates. This final model explained 26.3% of the variance in HlNl vaccination intention. All significant beta coefficients are presented in Figure 14, whereas all significant gamma coefficients are summarized in Table 5. 112 Regarding individual parameters, Hll, Hl2 and Hl3(a) were all supported, as social confidence was found to positively predict perceived effectiveness of vaccine in general (b ~ .27, p < .05), negatively predict safety concern over vaccine in general (b ~ .15, p < .05), and positively predict intention to vaccinate against the HlNl flu (b ~ .09, p < .05). On the other hand, Hl5(a), Hl6 and Hl7(a) were all rejected, as passive exposure to televised health news did not predict intention to vaccinate against the HlNl flu, perceived susceptibility to this flu, or perceived severity of this flu. RQ6( a) and RQ7( a) asked whether passive media exposure influenced individuals' perceptions of the effectiveness and safety of vaccine, and there was no empirical evidence for this influence. RQ 11 inquired if passive exposure to televised health news affected individuals' social confidence in health authorities. Again, there was no support for this effect. On the other hand, active attention to HlNl-related news was found to be a predictor of several variables. Support was found for Hl8, which postulated a positive relationship between individuals' degrees of active attention to HlNl flu-related news and their levels of perceived susceptibility toward this flu (b ~ .15, p < .05). However, Hl9 and H20 were rejected, as increased attention to HlNl-flu related news resulted in neither greater perceived severity regarding this flu nor higher intention to vaccinate against it. RQ13 and RQ14 asked if active attention to HlNl-flu related news had any impact over individuals' perceptions of the effectiveness and safety of vaccine in general, and there was no evidence of such an impact. RQ15 inquired if active attention to HlNl flu related news would shape individuals' confidence in health authorities, and 113 individuals with higher degrees of active attention were found to be less confident in the health authorities concerned (b ~ -.12, p < .05). All HBM variables were found to significantly predict HlNl vaccination intention, with perceived effectiveness of vaccine being the most powerful predictor (b ~ .28, p < .05), followed by perceived susceptibility to HlNl influenza (b ~ .23, p < .05), safety concern over vaccine (b ~ -.21, p < .05), and perceived severity of HlNl influenza (b ~ .19, p < .05). Among the socio-demographic and physiological variables, only income had a direct effect on HlNl influenza vaccination intention (standardized y ~ .12, p < .05). The influence of all other exogenous variables on vaccination intention was mediated by HBM constructs and active attention to HlNl-flu related news, which was in turn positively predicted by passive exposure to televised health news (b ~ .26, p < .05). Figure 14. Standardized beta estimates for significant paths predicting HI N 1 flu vaccination intention in 2009 Exposure to health news 0.26 Active attention to HlNl news Social confidence Perceived susceptibility Perceived severity Perceived effectiveness Safety concern HlNl flu vaccination intention 114 Table 5. Standardized Gamma Estimates for Significant Paths Predicting HJNJ Flu Vaccination Intention in 2009 Age Education Social confidence Attention to H1N1 news 017 Exposure to health news 0.16 Perceived effectiveness 0.16 Safety concern 0.12 Perceived susceptibility -0.17 Perceived severity -0.22 H1N1 flu vaccination Intention Income Male -0.11 -0.11 -0.12 Risk factors for flu White complications 0.12 0.12 -0.13 -0.13 -0.17 0.14 0.19 013 115 116 Modeling Seasonal Flu Vaccination Behavior in 2009 To test Hl(b) and H2(b), which postulated that the effects of perceived severity and perceived susceptibility on seasonal flu vaccine uptake were mediated through perceived effectiveness of vaccine, the theoretical model was first fitted to the data without the two mediational paths. Perceived severity (b ~ .15, p < .05), perceived susceptibility (b ~ .12,p < .05) and perceived effectiveness of vaccine (b ~ .12,p < .05) were all found to be significant predictors of vaccination intention. Subsequently, the mediational paths were added to the model. However, both additional paths were found to be non-significant, and the magnitude of the coefficients indicating the respective direct effects of perceived susceptibility, perceived severity and perceived effectiveness on seasonal flu vaccine uptake did not change. Therefore, Hl(b) and H2(b) were rejected. This initial model had an acceptable GFI at .98 and SRMR at .03. However, the other goodness-of-fit statistic was unsatisfactory (;(/df~ 84.0114). Thus, the modification indices were examined to identify plausible adjustments to the model. The modification indices suggested that the largest reduction in modell would result from adding a causal path between perceived susceptibility and perceived severity or by allowing the error term of these two variables to correlate. For the same reason as discussed in the previous section, the error term of these two variables were allowed to correlate to account for their potential common cause of 'perceived threat' (Weinstein, 1993). With this additional path, the model's global fit statistics improved considerably but were still not completely satisfactory (x 2 /df~ 20.45/3, SRMR ~ 0.02, GFI ~ 1.00). 117 A further examination of the modification indices indicated that allowing the error term of perceived effectiveness and safety concern regarding vaccine to correlate would now result in the largest reduction in model x 2 As explicated in the previous section, this addition was justifiable based on research from the field of risk management (Alhakami & Slovic, 1994; Finucane et al., 2000; Ganzach, 2000; Siegrist, 1999, 2000; Siegrist et al., 2007; Siegrist & Cvetkovich, 2000; Siegrist et al., 2000). After adding this correlational path, all global fit indices became satisfactory (x 2 /df~ 2.59/2, SRMR ~ 0.01, GFI ~ 1.00). An examination of the modification indices suggested that no other modification would significantly improve the model fit. Thus, non-significant paths were removed from this revised model one by one. After all non-significant paths were eliminated, the model fit remained good (x 2 /df ~ 59.49/45, SRMR ~ 0.03, GFI ~ .99), and there was no substantial change in the magnitude of the significant parameter estimates. This final model explained 29.7% of the variance in HlNl vaccination intention. All significant beta coefficients are presented in Figure 15, whereas all significant gamma coefficients are summarized in Table 6. Regarding individual parameters, Hll, Hl2 and Hl3(b) were all supported, as social confidence was found to positively predict perceived effectiveness of vaccine in general (b ~ .27, p < .05), negatively predict safety concern over vaccine in general (b ~ .15, p < .05), and positively predict seasonal flu vaccine uptake (b ~ .10, p < .05). On the other hand, Hl5(b ), Hl6 and Hl7(b) were all rejected, as passive exposure to televised health news did not predict seasonal flu vaccine uptake, perceived susceptibility to this flu, or perceived severity of this flu. RQ6(a) and RQ7(a) asked whether passive media 118 exposure influenced individuals' perceptions of the effectiveness and safety of vaccine in general, and there was no empirical evidence for this influence. RQ 11 inquired if exposure to televised health news affected individuals' social confidence in health authorities. Again, there was no support for this effect. Support was found for Hl8, which postulated a positive relationship between individuals' degrees of active attention to HlNl flu-related news and their levels of perceived susceptibility toward this flu (b ~ .15, p < .05). On the other hand, Hl9 was rejected, as increased attention to HlNl-flu related news did not result in greater perceived severity regarding this flu. However, increased attention to HlNl-flu related news did have a positive impact on the likelihood of seasonal flu vaccine uptake, as postulated by H21 (b ~ .12,p < .05). RQ13 and RQ14 asked if active attention to HlNl flu related news had any impact on individuals' perceptions of the effectiveness and safety of vaccine in general, and there was no evidence of such an impact. RQ15 inquired if active attention to HlNl-flu related news would shape individuals' social confidence in health authorities, and individuals with higher degrees of active attention were found to be less confident in the health authorities concerned (b ~ -.12, p < .05). All HBM variables were found to significantly predict seasonal flu vaccination behavior, with safety concern over vaccine being the most powerful predictor (b ~ -.24, p < .05), followed by perceived severity ofHlNl influenza (b ~ .14, p < .05), perceived susceptibility to HlNl influenza (b ~ .12, p < .05), and perceived effectiveness of vaccine in general (b ~ .11, p < .05). Among the socio-demographic and physiological variables, only age (standardized y ~ .30, p < .05) and education (standardized y ~ .18, p 119 < .05) had a direct effect on seasonal flu vaccination behavior. The influence of all other exogenous variables on vaccination behavior were mediated by HBM constructs and active attention to HlNl-flu related news, which was in turn positively predicted by passive exposure to televised health coverage (b ~ .26,p < .05). Figure 15. Standardized beta estimates for significant paths predicting seasonal flu vaccine behavior in 2009 Exposure to health news 0.26 Active attention to HlNl news Social confidence Perceived susceptibility Perceived severity Perceived effectiveness Safety concern Seasonal flu vaccine uptake 120 Table 6. Standardized Gamma Estimates for Significant Paths Predicting Seasonal Flu Vaccination Behavior in 2009 Age Education Social confidence Attention to H1N1 news 017 Exposure to health news 0.16 Perceived effectiveness 0.16 Safety concern 0.12 Perceived susceptibility -0.17 Perceived severity -0.22 Seasonal flu vaccine uptake 0.30 0.18 Income Male -0.11 -0.11 Risk factors for flu White complications 0.12 0.12 -0.13 -0.13 -0.17 0.14 0.19 013 121 122 Modeling H5Nl Flu Vaccination Intention in 2012 To test Hl(c) and H2(c), which postulated that the effects of perceived severity and perceived susceptibility on H5Nl flu vaccination intention were mediated through perceived effectiveness of vaccine, the theoretical model was first fitted to the data without the two mediational paths. Perceived severity (b ~ .07, p < .05) and perceived susceptibility (b ~ .23, p < .05) were found to positively predict H5Nl flu vaccination intention, but perceived effectiveness of vaccine was not a significant predictor of vaccination intention. Therefore, there was no ground for testing the mediational effect of perceived effectiveness, and Hl(c) and H2(c) were rejected. This initial model had an acceptable GFI at .94 and SRMR at .05. However, the chi-square statistic indicated an unsatisfactory model fit (x 2 /df~ 274.93/31). Thus, the modification indices were examined to identify plausible adjustments to the model. The modification indices suggested that allowing the error term of perceived susceptibility and perceived severity to correlate would result in the largest reduction in model x 2 that was theoretically justifiable, for the same reasons articulated in the previous section. After adding this correlational path, the model's global fit statistics improved but were still not completely satisfactory (x 2 /df~ 226.79/30, SRMR ~ 0.05, GFI ~ .95). A further examination of the modification indices indicated that adding a causal path from perceived susceptibility to subjective norm regarding H5Nl flu vaccination would result in the largest reduction in model x 2 that was theoretically justifiable. Given the hypothetical nature of the H5Nl flu, there was unlikely that study participants could resort to concrete experience to assess how much their family, friends and health care provider would approve of them getting a vaccine against this flu. Therefore, it is plausible that they might base their subjective norm judgments partially on their perceived susceptibility toward this disease, as one's family, friends and health care provider would likely encourage vaccination against a disease to which one was vulnerable. After adding this causal path, the model's global fit statistics improved but were still not completely satisfactory (x 2 /df~ 186.85/29, SRMR ~ 0.04, GFI ~ .96). 123 A subsequent examination of the modification indices indicated that adding a causal path from social trust to subjective norm regarding H5Nl flu vaccination would result in the largest reduction in model x 2 that was theoretically justifiable. Again, the hypothetical nature of the H5Nl flu made it difficult for study participants to ground their subjective norm judgments on concrete experience. Therefore, it is plausible that they might use how much they trust the FDA as a heuristic in making such judgments, as one's family, friends and health care provider would likely to want one to receive a vaccine that has been approved by a trustworthy authority. After adding this causal path, the model's global fit statistics were better but still not completely satisfactory (x 2 /df ~ 161.30/28, SRMR ~ 0.04, GFI ~ .97). A further examination of the modification indices indicated that adding a causal path from perceived susceptibility to self-efficacy for getting the H5Nl flu shot would result in the largest reduction in model x 2 that was theoretically justifiable. Again, there was little concrete experience for research participants to assess their efficacy in getting a vaccine that had never been on the market, and thus they might base their efficacy judgments on something else. It could well be the individuals who felt more susceptible 124 to H5Nl influenza could picture themselves trying harder to get a vaccine for this disease and thus reporting higher levels of efficacy. After adding this causal path, all global fit indices became satisfactory (x 2 /df~ 125.24/27, SRMR ~ 0.03, GFI ~ .97). An examination of the modification indices suggested that no other theoretically justifiable modification would significantly improve the model fit. Thus, non-significant paths were removed from this revised model one by one, and after all non-significant paths were eliminated, the model fit remained good (x 2 /df ~ 181.18/75, SRMR ~ 0.04, GFI ~ .96), and no substantial change in the magnitude of the significant parameter estimates were observed. This final model explained 56.6% of the variance in H5Nl vaccination intention. All significant beta coefficients are presented in Figure 16, and all significant gamma coefficients are summarized in Table 7. Regarding individual parameters, H3, H4, H5, H6 were supported, as integral affect toward flu vaccination was found to positively predict H5Nl flu vaccination intention (b ~ .42, p < .05) and perceived effectiveness of vaccine (b ~ .33, p < .05), negatively predict safety concern over vaccine (b ~- .30, p < .05), and positively predict subjective norm regarding H5Nl flu vaccination (b ~ .26, p < .05). On the other hand, H7 was rejected because no relationship was found between integral affect and self-efficacy for getting the H5Nl vaccine. RQl and RQ2 asked if and how integral affect shaped perceived susceptibility and perceived severity regarding H5Nl influenza, and positive relationships were established through statistical analyses, such that individuals experiencing more positive affect toward flu vaccination tended to perceive higher levels of personal susceptibility and severity regarding the H5Nl flu. In other words, empirical 125 findings suggested that rather than eliciting mood congruent thoughts (Salovey et al., 1991), affect worked in a more holistic and sophisticated way, where positive affect would elicit higher and less optimistic estimates of perceived susceptibility and perceived severity, which in turn led to greater adaptive intention to get vaccinated. H8 and H9 were supported, as social trust was found to positively predict perceived effectiveness of vaccine (b ~ .27, p < .05) and negatively predict safety concern over vaccine (b ~ -.12, p < .05). However, social trust did not have a direct impact on H5N 1 vaccination intention so H 10 was rejected. On the other hand, social trust was found to positively predict integral affect toward flu vaccination (b ~ .28, p < .05) so Hl4 was supported. RQ3 to RQlO and RQ12 inquired if exposure to televised health news had an impact on a range of variables, including H5N 1 flu vaccination intention, perceived susceptibility to H5Nl influenza, perceived severity of this flu, perceived effectiveness of vaccine, safety concern over vaccine, subjective norm toward H5Nl flu vaccination, self efficacy for getting the H5Nl flu shot, social trust in the FDA, and affect toward flu vaccination. Among all these constructs, exposure to health coverage was found to influence only safety concern over vaccine in general (b ~ -.10, p < .05). Among variables from established health behavior theories (i.e. HBM and TPB), perceived susceptibility to H5Nl influenza was found to be the most powerful predictor ofH5Nl vaccination intention (b ~ .23,p < .05), followed by subjective norm regarding this vaccine (b ~ .22, p < .05), perceived severity of the H5Nl flu (b ~ .07, p < .05), and safety concern over vaccine in general (b ~ -.06, p < .05). On the other hand, perceived 126 effectiveness of vaccine and self-efficacy for getting a H5Nl flu shot did not have a significant impact on vaccination intention. Among the socio-demographic and physiological variables, only education (standardized y ~ .09, p < .05) and being non Hispanic White (standardized y ~ -.08, p < .05) had a direct effect on intention to vaccinate against the H5Nl flu. The influence of all other exogenous variables on vaccination intention was mediated by HBM and TPB constructs as well as by integral affect and social trust. Figure 16 Standardized beta estimates for significant paths predicting H5N1 flu vaccination intention in 2012 Exposure to health news Social trust Integral affect Perceived susceptibility Perceived Perceived effectiveness Safety concern Self-efficacy Subjective norm H5Nlflu vaccination intention 127 128 Table 7. Standardized Gamma Estimates for Significant Paths Predicting H5Nl Flu Vaccination Intention in 2012 Risk factors for flu Age Education Income Male White complications Social trust 010 -0.10 Integral affect 0.15 010 0.21 Exposure to health news 022 -0.10 -0.13 0.18 Perceived effectiveness 0.19 -0.10 022 017 Safety concern -0.11 0.11 Perceived susceptibility 0.11 Perceived severity 0.12 Self-efficacy 013 013 Subjective norm 0.19 H5Nl vaccination intention 0.09 -0 08 129 Correlates for Seasonal Flu Vaccination Behavior in 2012 Hierarchical logistic regression analysis was used to identify correlates for seasonal flu vaccine uptake in the 2011-2012 influenza season. The unstandardized coefficients and odds ratios for the three consecutive models are available for review in Table 8. The Nagelkerke's R 2 indicated that the final model including all blocks explained about 66% of the variance in seasonal flu shot uptake (j ~ 394.26, p < .001, df ~ 15). All blocks significantly improved the amount of variance explained in the regression. In Model 2, the block of variables from established health behavior theories substantially improved the amount of variance explained from 19% to 46%. However, the Wald criterion indicated that only safety concern over vaccine in general (~ ~ -.64, p < .05), perceived susceptibility to seasonal influenza(~~ .23, p < .05), and subjective norm regarding seasonal flu vaccination(~~ .07, p < .05) made a significant contribution to prediction. In Model 3, the addition passive media exposure, social trust and integral affect increased the amount of variance in vaccination intention explained by the regression from 46% to 66%. However, the Wald criterion indicated that only integral affect toward flu vaccination(~~ .64, p < .05) made a significant contribution to prediction among the 'heuristic' variables. 130 Table 8. Estimates (Odds and Log-odds and) from Logistic Regression Analysis of Correlates of Seasonal Flu Vaccination Behavior in 2012. N ~ 584 Model 2 3 ~ exp(~) ~ exp(~) ~ exp(~) Gender Female (reference) Male -0.07 1.63 -0.38 0.68 -0.32 073 Age 0.04 * 1.02 003 * 1.04 0.04 * 1.04 Income 1.00 0.07 1.07 0.18 * 1.20 Education 0.04 1.04 -0 06 0.95 Race/ethnicity Non-White (reference) White -0.26 077 0.16 1.17 -0.05 0.96 Flu risk (square root) 0.43 * 1.54 0.33 1.39 0.23 1.26 Perceived susceptibility 0.23 * 1.26 0.07 1.07 Perceived severity 0.06 1.06 003 1.03 Perceived effectiveness 0.11 1.12 -0.29 0.75 Safety concern -0.64 * 0.53 -0.36 * 0.70 Self-efficacy -0 08 0.93 -0.03 0.97 Subjective norm 0.07 * 1.07 003 * 1.03 Health news exposure -0.17 0.85 Social trust -0 08 092 Integral affect 0.64 * 1.90 Intercept -3 06 * 0.05 -1.82 0.16 -3.24 * 0.04 Model/ 88.67 245.90 394.26 Df 6 12 15 Nagelkerke"sR 2 0.19 0.46 0.66 131 CHAPTER 7: DISCUSSION AND CONCLUSION This research challenges the tendency of established health behavior change theories to focus on the systematic route of decision making. It argues that the process of health decision making is more heuristic and affect-laden than suggested by major theories such as HBM, PMT, TRA, TPB and the Integrative Model of Behavioral Prediction. While they explicate important factors that shape health behaviors, they only tell part of the story. In order to begin giving a more comprehensive account of how individuals make health decisions, the present project proposes the inclusion of integral affect, social confidence, social trust and media influence in existent health behavior change theories. The rationale for the inclusion of these variables is based on extensive research from the fields of psychology, sociology, communication, risk management and economic behavior. To assess whether these variables function well as predictors of behavior and intention in the health context, data on decisions to vaccinate against three different flus were analyzed to provide empirical support. The choice of flu vaccine as the target health behavior is strategic, as this is a behavior that is relevant to and recommended to almost everyone by health authorities. It is also a behavior that is one-time and straightforward in nature, thus making the current study's focus on socio-psychological determinants of behavior (as opposed to structural determinants or a stage-based consideration of factors) justifiable. Empirical evidence suggests that integral affect is a powerful predictor of influenza vaccination intention and behavior. It influences intention to vaccinate against a 132 hypothetical H5Nl flu both directly and indirectly by shaping beliefs considered key determinants of health behavior by traditional theories. This is in line with psychologists' argument that human judgments are made through two routes of information processing, with affect-based processing always occurring automatically and frequently influencing subsequent rule-based processing that takes place optionally (Devine, 1989; Gilbert, 1989; Stanovich & West, 2002; Wegener & Petty, 1995). Of particular interest is the positive association observed between integral affect toward flu vaccination on the one hand and perceived susceptibility and perceived severity regarding the H5Nl flu on the other. According to one camp of psychologists (Salovey et al., 1991 ), affect would introduce thoughts that are congruent in valence. lf this was the case, positive integral affect toward flu vaccination should introduce more optimistic and therefore lower estimates of perceived susceptibility and perceived severity. However, positive integral affect was found to predict higher estimates of susceptibility and severity, which in turn resulted in greater intention to get the H5Nl flu shot as hypothesized by traditional health behavior theories. In other words, the functioning of affect appears to be more sophisticated and holistic, where it tends to influence thoughts in a way that would lead to a consistent final decision. Social trust and its close relative social confidence function in a way similar to affect, influencing behavior and intention both directly and indirectly by coloring systematic thinking. It is therefore not surprising that some scholars consider social trust a 'social affect' (Cvetkovich & Winter, 2007). During the 2009 HlNl influenza pandemic, individuals with higher levels of social confidence in public health agencies 133 had greater uptake of the seasonal flu vaccine and greater intention to receive the HlNl flu shot. Furthermore, these same people perceived vaccine in general to be more effective and reported lower levels of safety concern over vaccine compared to individuals low on social confidence. Similarly, in the case of intention to vaccine against a hypothetical H5Nl flu, individuals with higher levels of social trust in the FDA perceived vaccine in general as more effective and reported lower degrees of safety concern over vaccine. Furthermore, they had more positive subjective norm regarding vaccination against the hypothetical flu. Taken together, these findings are in line with research from the field of risk management that demonstrates perceived benefit and perceived risk of a hazardous technology or activity are often negatively correlated, despite that they are usually qualitatively different and therefore unrelated in real life (Alhakami & Slovic, 1994; Finucane et al., 2000). While some studies suggest that this negative correlation tends to disappear or becomes weaker when it comes to a familiar hazard or technology, this does not appear to be the case in the present research on the familiar pharmacological product of vaccine. Thus, more studies are needed to investigate how pervasive and context dependent is the correlation between perceived benefit and perceived risk regarding health technologies. On the other hand, it is worth noting that when only social confidence was taken into consideration in empirical analysis, perceived effectiveness of vaccine and safety concern over vaccine were found to correlate. However, when both social trust and integral affect were accounted for in the analysis, the correlation between perceived 134 effectiveness and safety concern disappeared. This finding is in line with research from risk management, which suggests that affect and trust might well be the two underlying causes that explain the correlation between perceived benefit and perceived risk regarding a hazardous technology or activity (Alhakami & Slovic, 1994; Finucane et al., 2000; Siegrist, 1999, 2000; Siegrist et al., 2007; Siegrist & Cvetkovich, 2000; Siegrist et al., 2000). Such findings have important implications for health communication practice. If integral affect and social trust are antecedents to risk and benefit perceptions of a health technology, then it becomes imperative for health communicators to include information that addresses those issues in their messages. As Peters et al. (2006) suggest, health professionals often prefer to provide 'objective facts' about a medical intervention and feel uncomfortable in offering contextual information and subjective interpretations that might "help their clientele take into account important [factual] information that might otherwise not enter into their decision processes" (p. S 145). The present research suggests that health communicators would achieve considerable success in influencing a person's benefit and risk perceptions regarding a health intervention if they could find ways to accentuate positive affect or alleviate negative feelings toward that intervention. To do this, health practitioners may look to the wealth of research from the fields of persuasion (Dillard & Pfau, 2002) and marketing and advertising (Mizerski & White, 1986) on how to incorporate affect-inducing sound, imagery, words, senses and events in their communication effort. On the other hand, scholars have suggested that social trust can be built through messages emphasizing that a social institution shares the values and goals of the population it serves (Earle & Siegrist, 2006; Earle et al., 2007; Siegrist et al., 2000). Health authorities like the FDA or CDC tend to focus their communication effort on 135 disseminating information about various diseases, official recommendations for their prevention or treatment, policy initiatives for reducing their burden, and the latest pharmacological products to curb these diseases. However, it might be worth devoting more time and energy to communicate an organization's values and vision to the public, if these are found to be shared by the public through research. Of course, there is the possibility that health authorities' goals are not always in line with the desires of the public. For example, the public might want health authorities to adopt the "precautionary principle" (W. Jackson, Raffensperger, & Tickner, 1999) in policy making. However, the uncertain and evolving nature of science makes it difficult for government regulators to obtain absolute proof of safety before allowing new medical technologies or products to be available for public use. This is especially the case when there is time pressure and lives at stake -for example, during a flu pandemic, newly developed prophylaxis products might not have time to gone through the standard clinical trial process but could potentially prevent many cases of illness and even deaths. Under such circumstances, the best strategy might be open and honest communication about the dilemma in institutional decision making and the provision of assurance of good intention and transparency on the part of the decision maker. Another interesting correlation observed in this research is between perceived susceptibility and perceived severity. This correlation is evident across the contexts of seasonal influenza, the HlNl flu, and the hypothetical H5Nl flu. It is possible that 136 susceptibility and severity perceptions could also be explained by an underlying common cause, maybe something resemble the construct of 'perceived threat' as proposed by Slovic et al. (1985), which encompasses an individual's cognitive as well as affective reaction toward a health condition. It has become apparent from the present research that traditional health behavior theories like the HBM and PMT tend to model lay individuals' decision making process based on the process of scientific decision making. That is, those theories consider perceived susceptibility, perceived severity, perceived benefit and perceived cost as more or less independent predictors of behavior. While this might be the case in real life where the benefit and risk of a health intervention are often qualitatively different and therefore uncorrelated, and where individuals' susceptibility to a disease is not necessarily linked to the possibility of him or her suffering severe consequences from the disease, in people's mind these various assessments seem to be related and be partially determined by a holistic affective reaction. The present research is one of the initial attempts seeking to understand this affective mechanism underlying lay individuals' health decision making. While there has been previous effort to include discrete emotions such as anticipated regret (Cooke et al., 2007; Gallagher & Povey, 2006; Godin et al., 201 0; Sandberg & Conner, 2008) into established health behavior theories as an independent and parallel predictor with cognitive beliefs, the present research demonstrates that the influence of affect might be more pervasive, where it influences health decisions both directly and indirectly by biasing cognitive thoughts. Of course, this project focuses on health decisions regarding vaccination, and more studies are needed in the future to 137 assess the influence of affect over different health behaviors (e.g. preventive, screening and treatment behaviors). Recognizing that the survey instrument utilized in this project has limitations in measuring affective reactions, future studies should seek to use other methods such as implicit association test and biometric sensor to complement self reported information. In addition to significant findings on integral affect, social confidence and social trust, this research also reveals considerable media effects on health decisions. During the HlNl influenza pandemic, active attention to HlNl-related news was found to be a double-edged sword. Individuals with greater levels of active attention reported greater uptake of the seasonal flu vaccine. They also perceived themselves as more susceptible to the H 1 N 1 flu, and this higher susceptibility in turn led to greater likelihood of getting the seasonal flu shot and higher intent to receive the HlNl flu vaccine. However, these same individuals also reported lower levels of social confidence in public health authorities, which in turn resulted in lower likelihood of getting the seasonal flu shot and lesser intent to get vaccinated against the HlNl flu. In the case of individuals surveyed outside of an influenza season, in April and May of 2012, those with greater exposure to televised health news had lower degrees of safety concern regarding vaccine. Given that active attention was not measured during this data collection period, it was unclear as to how much of this media effect resulted from passive media exposure versus active attention. Nonetheless, this finding serves to demonstrate that the media do not always work to elevate public concern about vaccine, in contrary to the claim of many critics (Bostrom & Atkinson, 2007; Heller, 2008; Poland 138 & Jacobson, 20 11; Critcher, 2007). At least in some occasions or time points, exposure to health coverage can actually encourage positive assessments of vaccine. A logical and ambitious follow-up study to the current project would be the implementation of longitudinal research that assesses individuals' perceptions regarding vaccine over time as well as their levels of active attention and passive exposure to health news in the media, and a content analysis of the news frames used in media reports regarding vaccine during the same period may shed additional light on how media influence individuals' vaccination-related judgments. Taken together, observations from the current project suggest that active attention to health content in the media can, at least during a pandemic situation, have both a direct and indirect impact on health behavior. In other words, the media can have more than a distal and indirect influence over individuals' health decision making, as suggested by the Integrative Model of Behavioral Prediction (Fishbein & Cappella, 2006; Fishbein & Yzer, 2003). This research also indicates that variables not explicitly articulated in major health behavior change theories, including integral affect, social confidence and social trust, can have powerful direct and indirect impacts on health behavior and intention. If these findings were replicated in future research, those factors should be incorporated into established health behavior theories to provide a more a more comprehensive account of the dual-mode processes of health decision making. Last but not least, this research set out to clarify the logical relationship among certain variables in established health behavior theories. While theories developed specifically to model health behaviors (e.g. HBM and PMT) consider perceived 139 susceptibility to a health condition and perceived severity of this condition key determinants of behavior, proponents of general behavioral prediction theories, such as TRA, TPB and the integrative model, argue that these perceptions are distal determinants of behavior. While some previous studies suggested that the influence of susceptibility and severity on behavior was largely mediated through perceived benefit, the present research found no evidence of such a mediational relationship across the contexts of vaccination against seasonal influenza, the HlNl flu and the hypothetical H5Nl flu. It appears that susceptibility and severity do function as proximal predictors of decisions in the health context, and they should thus be given due recognition in theories seeking to model health behaviors. 140 REFERENCES Adler, G. S., & Winston, C. A. (2004). Influenza vaccination and self-reported reasons for not receiving influenza vaccination among Medicare beneficiaries aged~ 65 years- United States, 1991-2002. MMWR, 53, 1012-1014. http:/ /www.cdc. gov /mmwr/preview /mmwrhtml/mm5343a3 .htm Ajzen, I. (1987). Attitudes, traits, and actions: Dispositional prediction of behavior in personality and social psychology. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 20, pp. 1-64). New York: Academic Press. Ajzen, I. (1991 ). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 5 0(2), 179-211. doi: 10.10 16/0749-5978(91 )90020-T Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice-Hall. Alhakami, A. S., & Slovic, P. (1994). A psychological study of the inverse relationship between perceived risk and perceived benefit. Risk Analysis, 14(6), 1085-1096. doi: 10.1111/j.l539-6924.1994.tb00080.x ANHCS 2007 User's Guide & Codebook. (2009). Retrieved March 28, 2013, from http:/ /anhcs.asc.upenn.edu Armitage, C. J., & Conner, M. (2000). Social cognition models and health behaviour: A structured review. Psychology & Health, 15(2), 173-189. doi: 10.1080/08870440008400299 Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behaviour: A meta-analytic review. British Journal of Social Psychology, 40( 4), 471-499. doi: 10.1348/014466601164939 Ball-Rokeach, S. J. (1985). The origins of individual media-system dependency: A sociological framework. Communication Research, 12( 4), 485-510. doi: 10.1177/009365085012004003 Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice Hall. Bandura, A. (2002). Social cognitive theory of mass communication. In J. Bryant & D. Zillmann (Eds.), Media effects: Advances in theory and research (2nd ed., pp. 121-153). Mahwah, NJ: Lawrence Erlbaum. Beck, U. (1992a). From industrial society to risk society: Questions of survival, social structure and ecological enlightment. Theory, Culture & Society, 9(1), 97-123. doi: 10.1177/026327692009001006 Beck, U. (1992b). Risk society: Towards a new modernity. London, UK: Sage. Beck, U. (1999). World risk society. Cambridge, UK: Polity Press. Beck, U., Giddens, A. , & Lash, S. (1994). Reflexive modernization: Politics, tradition and aesthetics in the modern social order. Cambridge, UK: Polity Press. Beck, U., & Grande, E .. (20 1 0). Varieties of second modernity: The cosmopolitan turn in social and political theory and research. The British Journal of Sociology, 61(3), 409-443. doi: 10.1111/j.l468-4446.2010.01320.x Bigman, C. A., Cappella, J. N., & Hornik, R. C. (2010). Effective or ineffective: Attribute framing and the human papillomavirus (HPV) vaccine. Patient Education and Counseling, 81(Suppll), S70-S76. doi: 10.1016/j.bbr.2011.03.031. Biocca, F. A. (1988). Opposing conceptions of the audience: The active and passive hemispheres of mass communication theory. In J. Anderson (Ed.), Communication yearbook (Vol. 11, pp. 51-80). Beverly Hills, CA: Sage. Blendon, R. J., SteelFisher, G. K., Benson, J. M., Bekheit, M. M. , & Herrmann, M. J. (2009). Harvard Opinion Research Program, Harvard School of Public Health: Public views of the HlNl vaccine topline results. Retrieved 2013, March 10, from http:/ /www.hsph.harvard.edu/horp/files/20 12/09/topline _- _public_ views_ of_ hlnl_ vaccine.pdf Blendon, R. J., SteelFisher, G. K., Benson, J. M., Bekheit, M. M. , & Herrmann, M. J. (2010). Harvard Opinion Research Program, Harvard School of Public Health: HlNl flu vaccination January poll top line results. Retrieved March 10, 2013, from http:/ /www.hsph.harvard.edu/horp/files/20 12/09/topline _- _public_ views_ of_ h ln 1_ vaccine_-_ update. pdf 141 Bodenhausen, G. V., Sheppard, L. A., & Kramer, G. P. (1994). Negative affect and social judgment: The differential impact of anger and sadness. European Journal of Social Psychology, 24(1), 45-62. doi: 10.1002/ejsp.2420240104 Boer, H., & Seydel, E. R. (1996). Protection motivation theory. In M. Conner & P. Norman (Eds.), Predicting health behaviour (pp. 95-120). Bristol, PA: Open University Press. Bostrom, A., & Atkinson, E. (2007). Trust and risk in smallpox vaccination. In M. Siegrist, T. C. Earle & H. Gutscher (Eds. ), Trust in cooperative risk management: Uncertainty and scepticism in the public mind (pp. 173-186). Sterling, VA: Earths can. Brewer, M. B. (1988). A dual process model of impression formation. In T. K. Srull & R. S. Wye (Eds.),Advances in social cognition (Vol. 1, pp. 1-36). Hillsdale, NJ: Erlbaum. Brewer, N. T., Chapman, G. B., Gibbons, F. X., Gerrard, M., McCaul, K. D., & Weinstein, N.D. (2007). Meta-analysis of the relationship between risk perception and health behavior: The example of vaccination. Health Psychology, 26(2), 136-145. doi: 10.1037/0278-6133.26.2.136 Bridges, C. B., Fukuda, K., Cox, N.J., & Singleton, J. A. (2001). Prevention and control of influenza: Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recommendations and Reports, 50(RR-4), 1-46. http:/ /www.cdc. gov /mmwr/preview /mmwrhtml/rr5004al.htm Brown, M. L., & Potosky, A. L. (1990). The presidential effect: The public health response to media coverage about Ronald Reagan's colon cancer episode. The Public Opinion Quarterly, 54(3), 317-329. doi: 10.2307/2749370 CDC. (2007). Avian influenza A virus infections of humans. Retrieved March 11, 2013, from http:/ /www.cdc.gov/flu/avian/gen-info/pdf/avian _flu_ humans.pdf CDC. (2009a). CDC says "Take3" steps to fight the flu. Retrieved March 11, 2013, from http:/ /warrencountyny.gov /healthservices/H 1 N 1-Take3. pdf CDC. (2009b). Safety of Influenza A (HlNl) monovalent vaccines- United States, October 1- November 24, 2009. 58( early release), 1-6. http:/ /www.cdc. gov /mmwr/preview /mmwrhtml/mm5 8e 1204al.htm 142 CDC. (2009c). Use of influenza A (HlNl) 2009 monovalent vaccine: Recommendations of the Advisory Committee on Immunization Practices (ACIP), 2009. MMWR Recommendations and Reports, 58( early release), 1-8. http:/ /www.cdc. gov /mmwr/preview /mmwrhtml!rr5 81 Oal.htm CDC. (2010). Key facts about 2009 HlNl flu vaccine. Retrieved March 11, 2013, from http:/ /www.cdc.gov/hlnlflu/vaccination/vaccine _ keyfacts.htm CDC. (2011a). Final estimates for 2009-10 seasonal influenza and influenza A (HlNl) 2009 monovalent vaccination coverage - United States, August 2009 through May 2010. Retrieved March 20, 2013, from http:/ /www.cdc. gov !flu/professionals/vaccination/ coverage_ 091 Oestimates.htm CDC. (20 11 b). People at high risk of developing flu-related complications. Retrieved March 18, 2013, from http://www.cdc.gov/flu/about/disease/high_risk.htm CDC. (2011c). Seasonal influenza: Questions & answer. Retrieved March 8, 2013, from http:/ /www.cdc.gov/flu/about/qa/disease.htm CDC. (2012a). Flu vaccination coverage, United States, 2011-12 influenza season. Retrieved March 18, 2013, from http:/ /www.cdc.gov/flu/professionals/vaccination/coverage _1112estimates.htm CDC. (2012b). Public health threat of highly pathogenic avian influenza A (H5Nl) virus. Retrieved 2013, March 20, from http://www.cdc.gov/flu/avianflu/h5nl-threat.htm CDC. (2012c). Vaccines and immunizations: Influenza. Retrieved March 18, 2013, from http:/ /www.cdc.gov/vaccines/pubs/pinkbook/flu.html#vaccine CDC. (20 12d). What you should know and do this flu season if you are 65 years and older. Retrieved March 18, 2013, from http:/ /www.cdc. gov !flu/ about/disease/65over.htm CDC. (2013). Vaccine effectiveness: How well does the flu vaccine work? Retrieved March 18, 2013, from http://www.cdc.gov/flu/about/qa/vaccineeffect.htm Chaffee, S. H., & Schlender, J. (1986). Measurement and effects of attention to media news. Human Communication Research, 13(1), 76-107. doi: 10.1111/j.l468- 2958.1986.tb00096.x Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source versus message cues in persuasion. Journal of Personality and Social Psychology, 39(5), 752-766. doi: 10.1037/0022-3514.39.5.752 Champion, V. L., & Skinner, C. S. (2008). The health belief model. In K. Glanz, B. K. Rimer & K. Viswanath (Eds.), Health behavior and health education: Theory, research, and practice (4th ed., pp. 45-65). San Francisco, CA: John Wiley & Sons. Chan, M. (2009). World now at the start of2009 influenza pandemic: Statement to the press by WHO Director-General Dr Margaret Chan. Retrieved March 10, 2013, from http:/ /www.who.int/mediacentre/news/statements/2009/hlnl _pandemic _phase6 _ 20090611/en/index.html Chen, N.-T. N., & Murphy, S. T. (2012). Examining the role of media coverage and trust in public health agencies in HlNl influenza prevention. In J. Merrick & M. Lemal (Eds.), Health risk communication (pp. 43-52). New York: Nova Science. Combs, B., & Slovic, P. (1979). Newspaper coverage Of causes of death. Journalism Quarterly, 56( 4), 837-849. doi: 10.1177/107769907905600420 Conner, M., & Armitage, C. J. (1998). Extending the theory of planned behavior: A review and avenues for further research. Journal of Applied Social Psychology, 28(15), 1429-1464. doi: 10.1111/j.l559-1816.1998.tb01685.x Conner, M., & Norman, P. (Eds.). (1996). Predicting health behaviour. Bristol, PA: Open University Press. 143 Cooke, R., Sniehotta, F., & Schuz, B. (2007). Predicting binge-drinking behaviour using an extended TPB: Examining the impact of anticipated regret and descriptive norms. Alcohol and Alcoholism, 42(2), 84-91. doi: 1 0.1093/alcalc/agl115 Cox, N.J., & Subbarao, K. (1999). Influenza. Lancet, 354(9186), 1277-1282. doi: 10.10 16/sO 140-6736(99)0 1241-6 Critcher, C. (2007a). 'Trust me, I'm a doctor': MMR and the politics of suspicion. In V. Bakir & D. M. Barlow (Eds.), Communicating in the age of suspicion: Trust and the media (pp. 88-102). New York: Palgrave MacMillan. Critcher, C. (2007b ). "Trust me, I'm a doctor:" MMR and the politics of suspicion. In V. Bakir & D. M. Barlow (Eds.), Communicating in the age of suspicion: Trust and the media (pp. 88-102). New York: Palgrave MacMillan. Croyle, R. T., & Uretsky, M. B. (1987). Effects of mood on self-appraisal of health status. Health Psychology, 6(3), 239-253. doi: 10.1037/0278-6133.6.3.239 Cummings, K. M., Becker, M. H., & Maile, M. C. (1980). Bringing the models together: An empirical approach to combining variables used to explain health actions. Journal of Behavioral Medicine, 3(2), 123-145. doi: 10.1007/BF00844986 Cummings, K. M., Jette, A. M., Brock, B. M., & Haefner, D. P. (1979). Psychosocial determinants of immunization behavior in a swine influenza campaign. Medical Care, 17(6), 639-649. Cvetkovich, G., & Winter, P. L. (2007). The what, how and when of social reliance and cooperative risk management. In M. Siegrist, T. C. Earle & H. Gutscher (Eds. ), Trust in cooperative risk management (pp. 189-209). Sterling, VA: Earthscan. Damasio, A. . (1994). Descarte's error: Emotion, reason and the human brain. New York: Penguin Group. de Wit, J. B. F., Vet, R., Schutten, M., & van Steenbergen, J. (2005). Social-cognitive determinants of vaccination behavior against hepatitis B: An assessment among men who have sex with men. Preventive Medicine, 40(6), 795-802. doi: 10.1016/j.ypmed.2004.09.026 Dearing, J. W., & Rogers, E. M. (1996). Agenda-setting. Thousand Oaks, CA: Sage. Demicheli, V., Di Pietrantonj, C., Jefferson, T., Rivett, A., & Rivett, D. (2007). Vaccines for preventing influenza in healthy adults. Cochrane database of systematic reviews. http://www. thecochranelibrary.com/userfiles/ ccoch/file/CDOO 1269. pdf Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of personality and social psychology, 56(1), 5. doi: 10.1037/0022-3514.56.1.5 DHHS. (2013a). About pandemics. Retrieved March 10, 2013, from http:/ /www.flu.gov/pandemic/about/index.html DHHS. (2013b). Pandemic flu history. Retrieved March 10, 2013, from http:/ /www.flu.gov/pandemic/history/index.html Dillard, J.P., & Pfau, M. (Eds.). (2002). The persuasion handbook: Developments in theory and practice. Thousand Oaks, CA: Sage. 144 Drociuk, D. (1999). Reasons reported by medicare beneficiaries for not receiving influenza and pneumococcal vaccinations-United States, 1996. MMWR, 48(39), 886-890. http://www.cdc.gov/mmwr/preview/mmwrhtml/mm4839a4.htm Earle, T. C., & Siegrist, M. (2006). Morality information, performance information, and the distinction between trust and confidence. Journal of Applied Social Psychology, 36(2), 383-416. doi: 10.lllllj.0021-9029.2006.00012.x Earle, T. C., & Siegrist, M. (2008). Trust, confidence and cooperation model: A framework for understanding the relation between trust and risk perception. International Journal of Global Environmental Issues, 8(1), 17-29. doi: 10.1504/IJGENVI.2008.017257 Earle, T. C., Siegrist, M., & Gutscher, H. (2007). Trust, risk perception and the TCC model of cooperation. In M. Siegrist, T. C. Earle & H. Gutscher (Eds. ), Trust in cooperative risk management: Uncertainty and scepticism in the public mind (pp. 1-49). Sterling, VA: Earthscan. Einsiedel, E. F., Salomone, K. L., & Schneider, F. P. (1984). Crime: Effects of media exposure and personal experience on issue salience. Journalism Quarterly, 61 (1 ), 131-136. doi: 10.1177/107769908406100ll8 Eldridge, J., & Reilly, J. (2003). Risk and relativity: BSE and the British media. InN. Pidgeon, R. E. Kasperson & P. Slovic (Eds. ), The social amplification of risk (pp. 138-155). New York: Cambridge University Press. Epstein, S. (1994). Integration of the cognitive and the psychodynamic unconscious. American Psychologist, 49(8), 709-724. doi: 10.1037/0022- 3514.63.4.534144768410.1037/0022-3514.63.4.5341993-04316-001 Eveland, W. P. (2002). The impact of news and entertainment media on perceptions of social reality. In J.P. Dillard & M. W. Pfau (Eds.), The persuasion handbook: Developments in theory and practice (pp. 691-727). Thousand Oaks, CA: Sage. Fan, D. P., & Holway, W. B. (1994). Media coverage of cocaine and its impact on usage patterns. International Journal of Public Opinion Research, 6(2), 139-162. doi: 10.1093/ijpor/6.2.139 Fauci, A. S. (2005). Race against time. Nature, 435(7041), 423-424. doi: 10.1038/435423a Fazio, R. H. (1986). How do attitudes guide behavior. In R. M. Sorrentino & E. T. Higgins (Eds. ), Handbook of motivation and cognition: Foundations of social behavior (Vol. 1, pp. 204-243). New York: Guilford Press. FDA. (2009). FDA approves vaccines for 2009 HlNl influenza virus: Approval provides important tool to fight pandemic. Retrieved March 10, 2013, from http:/ /www.fda.gov/newsevents/newsroom/pressannouncements/ucm 182399.htm Fiebach, N.H., & Viscoli, C. M. (1991). Patient acceptance of influenza vaccination. The American Journal ofMedicine, 91(4), 393-400. doi: 10.1016/0002- 9343(91)90157-S 145 Finnegan, J. R., & Viswanath, K. (2008). Communication theory and health behavior change: The media studies framework. InK. Glanz, B. K. Rimer & K. Viswanath (Eds.), Health behavior and health education: Theory, research, and practice (4th ed., pp. 364-387). San Francisco, CA Jossey-Bass. Finucane, M. L., Alhakami, A. S., Slovic, P., & Johnson, S. M. (2000). The affect heuristic in judgments of risks and benefits. Journal of Behavioral Decision Making, 13(1), 1-17. doi: 10.1002/(SICI)1099-0771(200001/03)13: 1 <l::AID BDM333<3.0.C0;2-S Fiore, A. E., Shay, D. K. , Broder, K. , Iskander, J. K. , Uyeki, T. M. , Mootrey, G. , ... Cox, N. S. (2008). Prevention and control of influenza: Recommendations of the Advisory Committee on Immunization Practices (ACIP), 2008. MMWR Recommendations and Reports, 57(RR-7), 1-60. http:/ /www.cdc. gov /mmwr/preview /mmwrhtml!rr57 e717 a1.htm Fiore, A. E., Uyeki, T. M., Broder, K., Finelli, L., Euler, G. L., Singleton, J. A., ... Bresee, J. S. (2010). Prevention and control of influenza with vaccines: Recommendations of the Advisory Committee on Immunization Practices (ACIP), 2010. MMWR Recommendations and Reports, 59(RR-8), 1-62. http:/ /www.cdc.gov/mmwr/preview/mmwrhtml!rr5908a1.htm Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley. Fishbein, M., & Cappella, J. N. (2006). The role of theory in developing effective health communications. Journal ofCommunication, 56, S1-Sl7. doi: 10.1111/j.1460- 2466.2006.00280.x Fishbein, M., & Yzer, M. C. (2003). Using theory to design effective health behavior interventions. Communication Theory, 13(2), 164-183. doi: 10.1111/j.1468- 2885.2003.tb00287.x Forgas, J.P. (1995). Mood and judgment: The affect infusion model (AIM). Psychological Bulletin, 117(1), 39-66. doi: 10.1037/0033-2909.117.1.39 Frewer, L. J. (2003). Trust, transparency, and social context: Implications for social amplification of risk. InN. Pidgeon, R. E. Kasperson & P. Slovic (Eds.), The social amplification of risk (pp. 123-137). New York: Cambridge University Press. Frewer, L. J., & Salter, B. (2007). Societal trust in risk analysis: Implications for the interface of risk assessment and risk management. In M. Siegrist, T. C. Earle & H. Gutscher (Eds.), Trust in cooperative risk management: Uncertainty and scepticism in the public mind (pp. 143-158). Sterling, VA: Earthscan. Frost, K., Frank, E., & Maibach, E. (1997). Relative risk in the news media: A quantification of misrepresentation. American Journal of Public Health, 87(5), 842-845. doi: 10.2105/AJPH.87.5.842 Gallagher, S., & Povey, R. (2006). Determinants of older adults' intentions to vaccinate against influenza: A theoretical application. Journal of Public Health, 28(2), 139- 144. doi: 10.1093/pubmed/fdl008 Ganzach, Y. (2000). Judging risk and return of financial assets. Organizational Behavior and Human Decision Processes, 83(2), 353-370. doi: 10.1006/obhd.2000.2914 Geracimos, A. (2009, November 5). HHS admits overstating flu-vaccine availability; health officials refuse to offer new estimates, The Washington Times, p. 1. 146 Gerbner, G., Gross, L., Morgan, M., Signorielli, N., & Shanahan, J. (2002). Growing up with television: Cultivation Processes. In J. Bryant & D. Zillman (Eds. ), Media effects: Advances in theory and research (2nd ed., pp. 43-67). Mahwah, NJ: Lawrence Erlbaum Associates. Giddens, A. (1990). The consequences of modernity. Stanford, CA: Stanford University Press. Giddens, A. (1994). Beyond left and right: The future of radical politics. Stanford, CA: Stanford University Press. Giddens, A. (1998). The third way: The renewal of social democracy. Cambridge, UK: Polity Press. Giffin, K. I. M. (1967). The contribution of studies of source credibility to a theory of interpersonal trust in the communication process. Psychological Bulletin, 68(2), 104-120. doi: 10.1037/h0024833 Gilbert, D. T. (1989). Thinking lightly about others: Automatic components of the social inference process. In J. S. Uleman & J. A. Bargh (Eds.), Unintended Thought (pp. 189-211). New York: Guilford Press. Gilovich, T., Griffin, D., & Kahneman, D. (Eds.). (2002). Heuristics and biases: The psychology of intuitive judgment Cambridge, UK: Cambridge University Press. Glanz, K., Rimer, B. K., & Viswanath, K (Eds.). (2008). Health behavior and health education: Theory, research, and practice (4th ed.). San Francisco, CA: Jossey Bass. Godin, G., & Kok, G. (1996). The theory of planned behavior: A review of its applications to health-related behaviors. American Journal of Health Promotion, 11(2), 87-98. doi: 10.4278/0890-1171-11.2.87 Godin, G., Vezina-Im, L.-A., & Naccache, H. (2010). Determinants of influenza vaccination among healthcare workers. Infection Control and Hospital Epidemiology, 31(7), 689-693. doi: 10.1086/653614 Gross, L. (2009). A broken trust: Lessons from the vaccine-autism wars. P LoS biology, 7(5), el000114. http:/ /www.plosbiology.org/article/info%3Adoi%2Fl 0.1371 %2Fjournal. pbio.lOO 0114 Guo, Z., Zhu, J. J. H., & Chen, H. (2001). Mediated reality bites: Comparing direct and indirect experience as sources of perceptions across two communities in China. InternationalJournal of Public Opinion Research, 13( 4), 398-418. doi: 10.1093/ijpor/13.4.398 Halbert, C., Armstrong, K. , Gandy, 0. H. Jr. , & Shaker, L. (2006). Racial differences in trust in health care providers. Archives of Internal Medicine, 166(8), 896-901. doi: 10.1001/archinte.l66.8.896 Harper, S. A, Fukuda, K., Uyeki, T. M, Cox, N. J, & Bridges, C. B. (2004). Prevention and control of influenza. MMWR, 5 3(RR -6), 1-40. http:/ /www.cdc. gov /mmwr/preview /mmwrhtml!rr5306al.htm 147 Harrison, J. A., Mullen, P. D., & Green, L. W. (1992). A meta-analysis of studies of the health belief model with adults. Health Education Research, 7(1), 107-116. doi: 1 0.1093/her/7.1.1 07 Heaney, C. A., & Israel, B. A. (2008). Social networks and social support. In K. Glanz, B. K. Rimer & K. Viswanath (Eds.), Health behavior and health education: Theory, research, and practice (4th ed., pp. 189-210). San Francisco: CA: Jossey Bass. Heller, J. (2008). The vaccine narrative. Nashville, VA: Vanderbilt University Press. Hu, L.-T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1-55. doi: 10.1080/10705519909540118 Hu, L.-T., & Bentler, P.M .. (1995). Evaluating model fit. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues, and applications (pp. 76-99). Thousand Oaks, CA: Sage. Jackson, K. M., & Aiken, L. S. (2000). A psychosocial model of sun protection and sunbathing in young women: The impact of health beliefs, attitudes, norms, and self-efficacy for sun protection. Health Psychology, 19(5), 469-478. doi: 10.1037/0278-6133.19.5.469 Jackson, K. M., & Aiken, L. S. (2006). Evaluation of a multicomponent appearance based sun-protective intervention for young women: Uncovering the mechanisms of program efficacy. Health Psychology, 25(1), 34-46. doi: 10.1037/0278- 6133.25.1.34 Jackson, W., Raffensperger, C., & Tickner, J. (Eds.). (1999). Protecting public health and the environment: Implementing the precautionary principle. Washington, DC: Island Press. Jaeger, C. C., Renn, 0., Rosa, E. A., & Wehler, T. (2001). Risk, uncertainty, and rational action. London, UK: Earthscan. Janz, N. K., & Becker, M. H. (1984). The health belief model: A decade later. Health Education & Behavior, 11(1), 1-47. doi: 10.1177/109019818401100101 Kahneman, D. (1991). Commentary: Judgment and decision making: A personal view. Psychological Science, 2(3), 142-145. Kahneman, D., & Frederick, S. (2002). Representativeness revisited: Attribute substitution in intuitive judgment. In T. Gilovich, D. Griffin & D. Kahneman (Eds.), Heuristics and biases: The psychology of intuitive judgment (pp. 49-81). Cambridge, UK: Cambridge University Press. Kahneman, D., & Frederick, S. (Eds.). (2005). A model of heuristic judgment. Cambridge, UK: Cambridge University Press. Kahneman, D., Slovic, P., & Tversky, A. (Eds.). (1982). Judgment under uncertainty: Heuristics and biases. Cambridge, UK: Cambridge University Press. Kahneman, D., & Tversky, A. (1972). Subjective probability: A judgment of representativeness. Cognitive Psychology, 3(3), 430-454. doi: 10.1016/0010- 0285(72)90016-3 Kahneman, D., & Tversky, A. (1973). On the psychology of prediction. Psychological Review, 80(4), 237-251. doi: 10.1037/h0034747 148 Kaiser Family Foundation & Pew Research Center. (2009, July). Health news coverage in the U.S. media. from http://www.kff.org/entmedia/upload/7950.pdf Kalichman, S.C. (1994). Magic Johnson and public attitudes toward AIDS: A review of empirical findings. AIDS Education and Prevention, 6(6), 542-557. Kasperson, J. X., Kasperson, R. E., Pidgeon, N., & Slovic, P. (2003). The social amplification of risk: Assessing fifteen years of research and theory. InN. Pidgeon, R. E. Kasperson & P. Slovic (Eds. ), The social amplification of risk (pp. 13-46). New York: Cambridge University Press. Kasperson, R. E., Renn, 0., Slovic, P., Brown, H. S., Erne!, J., Goble, R., ... Ratick, S. (1988). The social amplification of risk: A conceptual framework. Risk Analysis, 8(2), 177-187. doi: 10.1111/j.1539-6924.1988.tb01168.x Katsuya, T. (2002). Difference in the formation of attitude toward nuclear power. PoliticalPsychology, 23(1), 191-203. doi: 10.111110162-895X.00277 Katz, E., Blumler, J. G., & Gurevitch, M. (1973). Uses and Gratifications Research. The Public Opinion Quarterly, 37(4), 509-523. doi: 10.2307/2747854 Kennedy, A.M., Brown, C. J., & Gust, D. A. (2005). Vaccine beliefs of parents who oppose compulsory vaccination. Public Health Reports, 120(3), 252-258. doi: 10.2307/20056785 Keren, G., & Teigen, K. H. (2007). Yet another look at the heuristics and biases approach. In D. J. Koehler & N. Harvey (Eds.), Blackwell handbook ofjudgment and decision making. Blackwell Reference Online: Blackwell Publishing. Retrieved from http://www. blackwellreference.com/subscriber/tocnode.html ?id~g978140 5157 59 9_chunk_g97814051575997. Kirby, D. (2005). Evidence of harm: Mercury in vaccines and the autism epidemic: A medical controversy. New York: StMartin's Press. Kissling, E., Valenciano, M., Cohen, J. M., Oroszi, B., Barret, A.-S., Rizzo, C., ... Larrauri, A. (2011). I-MOVE multi-centre case control study 2010-11: Overall and stratified estimates of influenza vaccine effectiveness in Europe. P loS one, 6(11), e27622. http://www. plosone.org/article/info%3Adoi%2F 10.13 71 %2Fjournal. pone. 002762 2 Kruglanski, A. W., & Klar, Y. (1985). Knowing what to do: On the epistemology of actions. In J. Kuhl & J. Beckmann (Eds. ), Action control: From cognition to behavior (pp. 41-60). New York: Springer-Verlag. Lipset, S. M., & Schneider, W. (1983). The confidence gap: Business, labor, and government, in the public mind. New York: The Free Press. Lowry, D. T., Nio, T. C. J., & Leitner, D. W. (2003). Setting the public fear agenda: A longitudinal analysis of network TV crime reporting, public perceptions of crime, and FBI crime statistics. Journal ofCommunication, 53(1), 61-73. doi: 10.1111/j.1460-2466.2003.tb03005.x Luhmann, N. (1979). Trust and Power. New York: John Wiley & Sons. 149 Luhmann, N. (2000). Familiarity, confidence, trust: Problems and alternatives. In D. Gambetta (Ed.), Trust: Making and breaking cooperative relations (pp. 94-1 07). Oxford, UK: Department of Sociology, University of Oxford. Retrieved from http:/ /www.sociology. ox. ac. uk/papers/luhmann94-1 07. pdf. Maddux, J. E., & Rogers, R. W. (1983). Protection motivation and self-efficacy: A revised theory of fear appeals and attitude change. Journal of Experimental Social Psychology, 19(5), 469-479. doi: 10.1016/0022-1031(83)90023-9 Maurer, J., Uscher-Pines, L., & Harris, K. M. (2010). Perceived seriousness of seasonal and A(H1N1) influenzas, attitudes toward vaccination, and vaccine uptake among U.S. adults: Does the source of information matter? Preventive Medicine, 51(2), 185-187. doi: 10.1016/j.ypmed.2010.05.008 McCabe, A. S., & Fitzgerald, M. R. (1991). Media images of environmental biotechnology: what does the public see. In G. S. Sayler, R. Fox & J. W. Blackburn (Eds.), Environmental biotechnology for waste treatment (pp. 15-24). New York: Plenum Press McCombs, M. E., & Shaw, D. L. (1972). The agenda-setting function of mass media. Public Opinion Quarterly, 36(2), 176-187. doi: 10.1086/267990 McLuhan, M. (1964). Understanding media: The extensions of man. New York: New American Library. McRee, A.-L., Brewer, N. T., Reiter, P. L. , Gottlieb, S. L. , & Smith, J. S. (2009). The Carolina HPV immunization attitudes and beliefs scale (CHIAS): Scale development and associations with intentions to vaccinate. Sexually transmitted diseases, 37(4), 1. doi: 10.1097/0LQ.Ob013e3181c37e15 Miller, D., & Reilly, J. (1995). Making an issue of food safety: The media, pressure groups, and the public sphere. In D. Maurer & J. Sohal (Eds. ), Eating agendas: Food and nutrition as social problems (pp. 305-336). New York: Aldine de Gruyter. Mizerski, R. W., & White, J. D. (1986). Understanding and using emotions in advertising. Journal of Consumer Marketing, 3( 4), 57-69. doi: 10.1108/eb008180 Mnookin, S .. (2011). The panic virus. New York: Simon & Schuster. Montano, D. E. (1986). Predicting and understanding influenza vaccination behavior: Alternatives to the health belief model. Medical Care, 24(5), 438-453. doi: 10.2307/3764731 Montano, D. E., & Kasprzyk, D. (2008). Theory of reasoned ation, theory of planned behvior, and the integrated behavioral model. In K. Glanz, B. Rimer & K. Viswanath (Eds.), Health behavior and health education: Theory, research, and practice (pp. 67-96). San Francisco, CA: Jossey-Bass. Mullen, M. (2009, October 25). Aiming to dispel H1N1 vaccine fear; it's "made just like all the flu vaccines we have been making for 60 years," one expert argues, The Los Angeles Times, p. 1. Murphy, S. T., & Zajonc, R. B. (1993). Affect, cognition, and awareness: Affective priming with optimal and suboptimal stimulus exposures. Journal of Personality and Social Psychology, 64(5), 723-739. doi: 10.1126/science.7352271 150 Musa, D., Schulz, R., Harris, R., Silverman, M., & Thomas, S. B. (2009). Trust in the health care system and the use of preventive health services by older Black and White adults. American Journal of Public Health, 99(7), 1293-1299. doi: 10.2105/ AJPH.2007.123927 Nabi, R. L. (1999). A cognitive-functional model for the effects of discrete negative emotions on information processing, attitude change, and recall. Communication Theory, 9(3), 292-320. doi: 10.1111/j.l468-2885.1999.tb00172.x Nabi, R. L. (2003). Exploring the framing effects of emotion: Do discrete emotions differentially influence information accessibility, information seeking, and policy preference? Communication Research, 30(2), 224-247. doi: 10.1177/0093650202250881 Nabi, R. L. (2010). The case for emphasizing discrete emotions in communication research. Communication Monographs, 77(2), 153-159. doi: 10.1080/03637751003790444 Nerlich, B., & Halliday, C. (2007). Avian flu: The creation of expectations in the interplay between science and the media. Sociology of Health & Illness, 29(1), 46-65. doi: 10.1111/j.l467-9566.2007.00517.x Nichol, K. L., & Treanor, J. J. (2006). Vaccines for seasonal and pandemic influenza. The Journal of Infectious Diseases, 194, Slll-Sll8. doi: 10.2307/30086131 Noar, S. M., & Zimmerman, R. S. (2005). Health behavior theory and cumulative knowledge regarding health behaviors: Are we moving in the right direction? Health Education Research, 20(3 ), 27 5-290. doi: 10.1 093/her/cyg 113 Norman, P., & Conner, M. (1996). The role of social cognition models in prediting health behaviours: Future directions. In M. Conner & P. Norman (Eds.), Predicting health behaviour (pp. 197-225). Bristol, PA: Open University Press. O'Keefe, D. J. (2002). Persuasion: Theory and research (2nd ed.). Thousand Oaks, CA: Sage. O'Keefe, G. J. (1984). Television exposure, credibility and public views on crime. In R. Bostrom (Ed.), Communication yearbook (Vol. 8, pp. 513-536). Newbury Park, CA: Sage. O'Keefe, G. J., & Reid-Nash, K. . (1987). Crime news and real-world blues: The effects of the media on social reality. Communication Research, 14(2), 147-163. doi: 10.1177/009365087014002001 Oliver, R. L., & Berger, P. K. (1979). A path analysis of preventive health care decision models. Journal of Consumer Research, 6(2), 113-122. doi: 10.2307/2488869 Ouellette, J. A., & Wood, W. (1998). Habit and intention in everyday life: The multiple processes by which past behavior predicts future behavior. Psychological Bulletin, 124(1), 54-74. doi: 10.1037/0278-7393.19.4.922.1993-44102- 00110.1037/0278-7393.19.4.922 Over, D. (2007). Rationality and the normative/descriptive distinction. In D. J. Koehler & N. Harvey (Eds.), Blackwell handbook ofjudgment and decision making. Blackwell Reference Online: Blackwell Publishing. Retrieved from http://www. blackwellreference.com/subscriber/tocnode.html ?id~g978140 5157 59 9_chunk_g97814051575993. Payne, J. W., & Bettman, J. R. (2007). Walking with the scarecrow: The information processing approach to decision research. In D. J. Koehler & N. Harvey (Eds. ), Blackwell handbook ofjudgment and decision making. Blackwell Reference Online: Blackwell Publishing. Retrieved from 151 http://www. blackwellreference. com/subscriber/tocnode.html ?id~g978140 5157 59 9_chunk_g97814051575998. Peiris, J. S.M., de Jong, M.D., & Guan, Y. (2007). Avian influenza virus (H5N1): A threat to human health. Clinical Microbiology Reviews, 20(2), 243-267. doi: 10.1128/cmr.00037-06 Peters, E., Lipkus, I., & Diefenbach, M. A. (2006). The functions of affect in health communications and in the construction of health preferences. Journal of Communication, 56(Suppl1), S140-S162. doi: 10.1111/j.1460- 2466.2006.00287.x Petty, R. E., Brinol, P., & Priester, J. R. (2009). Mass media attitude change: Implications of the elaboration likelihood model of persuasion. In J. Bryant & M. B. Oliver (Eds.), Media effects: Advances in theory and research (3rd ed., pp. 125-164). New York: Routledge. Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: Central and peripheral routes to attitude change. New York: Springer-Verlag. Petty, R. E., Schumann, D. W., Richman, S. A., & Strathman, A. J. (1993). Positive mood and persuasion: Different roles for affect under high- and low-elaboration conditions. Journal of Personality and Social Psychology, 64(1), 5-20. doi: 10.1037/0022-3514.64.1.5 Pierce, J.P., & Gilpin, E. A. (2001). News media coverage of smoking and health is associated with changes in population rates of smoking cessation but not initiation. Tobacco Control, 10(2), 145-153. doi: 10.1136/tc.10.2.145 Poland, G. A., & Jacobson, R. M. (2011). The age-old struggle against the antivaccinationists. New England Journal of Medicine, 364(2), 97-99. doi: 10.1056/NEJMp1010594 Pollack, A. (2009, May 8). Fear of a swine flu epidemic in 1976 offers some lessons, and concerns, today, The New York Times. Retrieved from http:/ /www.nytimes.com/2009/05/09/health/09vaccine.html Pornpitakpan, C. (2004). The persuasiveness of source credibility: A critical review of five decades' evidence. Journal of Applied Social Psychology, 34(2), 243-281. doi: 10.1111/j.1559-1816.2004.tb02547.x Prochaska, J. 0., Redding, C. A., & Evers, K. E. (2008). The transtheoretical model and stages of change. InK. Glanz, B. K. Rimer & K. Viswanath (Eds.), Health behavior and health education: Theory, research, and practice (4th ed., pp. 97- 122). San Francisco, CA: Jossey-Bass. Prochaska, J. 0., & Velicer, W. F. (1997). The transtheoretical model of health behavior change. American Journal of Health Promotion, 12(1), 38-48. doi: 10.4278/0890- 1171-12.1.38 Public Health - Seattle & King County. (2008). Pandemic flu audience research. King County, W A: Public Health - Seattle & King County. 152 Renn, 0., & Levine, D. (1991). Credibility and trust in risk communication. In R. E. Kasperson & P. J. M. Stallen (Eds.), Communicating risks to the public: International perspectives (Vol. 4, pp. 175-218). Dordrecht, Netherlands: Kluwer Academic Publishers. Rice, R. E., & Atkins, C. K. (Eds. ). (200 1 ). Public communication campaigns (3rd ed. ). Thousand Oaks, CA: Sage. Rimer, B. K. (2008). Models of individual health behavior. In K. Glanz, B. K. Rimer & K. Viswanath (Eds.), Health behavior and health education: Theory, research, and practice (4th ed., pp. 41-44). San Francisco, CA: John Wiley & Sons. Rimer, B. K., & Glanz, K. (2005). Theory at a glance: A guide for health promotion practice (2nd ed.). Retrieved from http:/ /www.cancer.gov/cancertopics/cancerlibrary/theory.pdf Roan, S .. (2009, April 27). Swine flu 'debacle' of 1976 is recalled, The Los Angeles Times. Retrieved from http:/ /articles.latimes.com/2009/apr/27 /science/sci -swine history27 Rogers, E. M. (2003). Diffusion of innovations (5th ed.). New York: Free Press. Rogers, R. W. (1975). A protection motivation theory of fear appeals and attitude change. Journal ofPsychology, 91(1), 93-114. doi: 10.1080/00223980.1975.9915803 Rogers, R. W. (1983). Cognitive and physiological processes in fear appeals and attitude change: A revised theory of protection motivation. In J. T. Cacioppo & R. E. Petty (Eds.), Social Psychophysiology (pp. 153-176). New York: Guilford Press. Rogers, R. W., & Mewborn, C. R. (1976). Fear appeals and attitude change: Effects of a threat's noxiousness, probability of occurrence, and the efficacy of coping responses. Journal of Personality and Social Psychology, 34(1), 54-61. doi: 10.1037/0022-3514.34.1.54 Ronis, D. L. (1992). Conditional health threats: Health beliefs, decisions, and behaviors among adults. Health Psychology, 11(2), 127-134. doi: 10.1037/0278- 6133.11.2.127 Ronis, D. L., & Hare!, Y. (1989). Health beliefs and breast examination behaviors: Analyses of linear structural relations. Psychology & Health, 3(4), 259-285. doi: 10.1080/08870448908400385 Rosenstock, I. M. (1966). Why people use health services. The Milbank Memorial Fund Quarterly, 44(3), 94-127. doi: 10.2307/3348967 Rosenstock, I. M. (1974). Historical origins of the health belief model. Health Education & Behavior, 2(4), 328-335. doi: 10.1177/109019817400200403 Rosenstock, I. M., Strecher, V. J., & Becker, M. H. (1988). Social learning theory and the health belief model. Health Education & Behavior, 15(2), 175-183. doi: 10.1177/109019818801500203 Rubin, A. M. (1993). Audience activity and media use. Communication Monographs, 60(1), 98-105. doi: 10.1080/03637759309376300 Rubin, A.M. (2002). The uses-and-gratifications perspective of media effects. In J. Bryant & D. Zillman (Eds. ), Media effects: Advances in theory and research (pp. 525-548). Mahwah, NJ: Lawrence Erlbaum Associates. 153 Rubin, A. M. (2009). Uses-and-gratifications perspective on media effects. In J. Bryant & M. B. Oliver (Eds.), Media effects: Advances in theory and research (3rd ed., pp. 165-184). New York: Routledge. Rundall, T. G., & Wheeler, J. R. C. (1979). Factors associated with utilization of the swine flu vaccination program among senior citizens in Tompkins county. Medical Care, 17(2), 191-200. doi: 10.2307/3763890 Sallis, J. F., Owen, N., & Fisher, E. B. (2008). Ecological models of health behavior. In K. Glanz, B. K. Rimer & K. Viswanath (Eds.), Health behavior and health education: Theory, research, and practice (4th ed., pp. 465-486). San Francisco, CA: Jossey-Bass. Salovey, P., & Birnbaum, D. (1989). Influence of mood on health-relevant cognitions. Journal of Personality and Social Psychology, 57(3), 539-551. doi: 10.1037/0022-3514.57.3.539 Salovey, P., O'Leary, A., Stretton, M.S., Fishkin, S. A., & Drake, C. A. (1991). Influence of mood on judgments about health and illness. In J.P. Forgas (Ed.), Emotion and social judgments (pp. 241-262 ). New York: Pergamon Press. Sandberg, T., & Conner, M. (2008). Anticipated regret as an additional predictor in the theory of planned behaviour: A meta-analysis. British Journal of Social Psychology, 47(4), 589-606. doi: 10.1348/014466607X258704 Scheufele, D. A., & Tewksbury, D. (2007). Framing, agenda setting, and priming: The evolution of three media effects models. Journal of Communication, 57(1), 9-20. doi: 10.1111/j.0021-9916.2007.00326.x Sheeran, P., & Abraham, C. (1996). The health belief model. In M. Conner & P. Norman (Eds.), Predicting health behaviour (pp. 23-61). Bristol, PA: Open University Press. Shrestha, S. S., Swerdlow, D. L., Borse, R. H., Prabhu, V. S., Finelli, L., Atkins, C. Y., .. . Meltzer, M. I. (20 11 ). Estimating the burden of 2009 pandemic influenza A (H1N1) in the United States (April2009-April2010). Clinical Infectious Diseases, 52(suppl1), S75-S82. doi: 10.1093/cid/ciq012 Shrum, L. J. (2002). Media consumption and perceptions of social reality: Effects and underlying processes. In J. Bryant & D. Zillman (Eds. ), Media effects: Advances in theory and research (2nd ed., pp. 69-75). Mahwah, NJ.: Lawrence Erlbaum Associates. Shrum, L. J. (2009). Media consumption and perceptions of social reality: Effects and underlying processes. In J. Bryant & M. B. Oliver (Eds. ), Media effects: Advances in theory and research (3rd ed., pp. 50-73). New York: Routledge. Siegrist, M. (1999). A causal model explaining the perception and acceptance of gene technology. Journal of Applied Social Psychology, 29(10), 2093-2106. doi: 1 0.1111/j .1559-1816.1999.tb02297.x Siegrist, M. (2000). The influence of trust and perceptions of risks and benefits on the acceptance of gene technology. Risk Analysis, 20(2), 195-204. doi: 10.111110272- 4332.202020 Siegrist, M., Cousin, M.-E., Kastenholz, H., & Wiek, A. (2007). Public acceptance of nanotechnology foods and food packaging: The influence of affect and trust. Appetite, 49(2), 459-466. doi: 10.1016/j.appet.2007.03.002 154 Siegrist, M., & Cvetkovich, G. (2000). Perception of hazards: The role of social trust and knowledge. Risk Analysis, 20(5), 713-720. doi: 10.111110272-4332.205064 Siegrist, M., Cvetkovich, G., & Roth, C. (2000). Salient value similarity, social trust, and risk/benefit perception. Risk Analysis, 20(3), 353-362. doi: 10.111110272- 4332.203034 Siegrist, M., Earle, T. C., & Gutscher, H. (2003). Test of a trust and confidence model in the applied context of electromagnetic field (EMF) risks. Risk Analysis, 23( 4), 705-716. doi: 10.111111539-6924.00349 Siegrist, M., Keller, C., & Cousin, M.-E. (2006). Implicit attitudes toward nuclear power and mobile phone base stations: Support for the affect heuristic. Risk Analysis, 26( 4), 1021-1029. doi: 10.11111j.l539-6924.2006.00797.x Simon, H. A. (1955). A behavioral model of rational choice. The Quarterly Journal of Economics, 69(1), 99-118. doi: 10.2307/1884852 Simon, H. A. (1956). Rational choice and the structure of the environment. Psychological Review, 63(2), 129-138. doi: 10.1037/h0042769 Simon, H. A. (1978). Rationality as process and as product of thought. The American Economic Review 68(2), 1-16. Simon, H. A. (1982). Models of bounded rationality: Economic analysis and public policy (Vol. 1 ). Cambridge, MA: MIT Press. Simon, H. A. (2000). Bounded rationality in social science: Today and tomorrow. Mind & Society, 1(1), 25-39. doi: 10.1007/BF02512227 Simon, H. A., & Newell, A. (1958). Heuristic problem solving: The next advance in operations research. Operations Research, 6(1), 1-10. doi: 10.1287/opre.6.1.1 Singhal, A., Cody, M. J., Rogers, E. M., & Sabido, M. (Eds.). (2004). Entertainment education and social change: History, research, and practice. Mahwah, NJ: Lawrence Erlbaum Associates. Singleton, J. A., Santibanez, T. A., Lu, P. J., Ding, H., Euler, G. L., Armstrong, G. L., ... Balluz, L. (2010). Interim results: influenza A (HlNl) 2009 monovalent vaccination coverage-United States, October-December 2009. MMWR, 59, 44-48. http:/ /www.cdc. gov /mmwr/preview /mmwrhtml/mm5902a4.htm Sjoberg, L., & Engelberg, E. (2010). Risk perception and movies: A study of availability as a factor in risk perception. Risk Analysis, 30(1), 95-106. doi: 10.11111j.l539- 6924.2009.01335.x Sloman, S. A. (1996). The empirical case for two systems of reasoning. Psychological Bulletin, 119(1), 3-22. doi: 10.1037/0033-2909.119.1.3 Sloman, S. A. (2002). Two systems of reasoning. In T. Gilovich, D. Griffin & D. Kahneman (Eds.), Heuristics and biases: The psychology of intuitive judgment (pp. 379-396). Cambridge, UK: Cambridge University Press. Slovic, P. (1987). Perception of risk. Science, 236(4799), 280-285. doi: 1 0.1126/science.3563507 155 Slovic, P., Finucane, M. L. , Peters, E., & MacGregor, D. G. (2002). The affect heuristic. InT. Gilovich, D. Griffin & D. Kahneman (Eds.), Heuristics and biases: The psychology of intuitive judgment (pp. 397-429). Cambridge, UK: Cambridge University Press. Slovic, P., Finucane, M. L. , Peters, E., & MacGregor, D. G. (2007). The affect heuristic. European Journal of Operational Research, 177(3), 1333-1352. doi: 10.10 16/j .ejor.2005.04.006 Slovic, P., Finucane, M. L., Peters, E., & MacGregor, D. G. (2004). Risk as analysis and risk as feelings: Some thoughts about affect, reason, risk, and rationality. Risk Analysis, 24(2), 311-322. doi: 10.lllllj.0272-4332.2004.00433.x Slovic, P., Fischhoff, B., & Lichtenstein, S. (1985). Characterizing perceived risk. In R. W. Kates, C. Hohenemser & J. X. Kasperson (Eds.), Perilous progress: Managing the hazards of technology (pp. 91-125). Boulder, CO: Westview. Smith, E. R., & DeCoster, J. (2000). Dual-process models in social and cognitive psychology: Conceptual integration and links to underlying memory systems. Personality and social psychology review, 4(2), 108-131. doi: 10.1207/S 15327957PSPR0402 01 Smith, N. M., Bresee, J. S. , Shay, D. K. , Uyeki, T. M. , Cox, N. J. , & Strikas, R. A. (2006). Prevention and control of influenza: Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recommendations and Reports, 55(RR-10), 1-42. http:/ /www.cdc. gov /mmwr/preview /mmwrhtml!rr5 51 Oal.htm Stanovich, K. E., & West, R. F. (2002). Individual differences in reasoning: Implications for the rationality debate. In T. Gilovich, D. Griffin & D. Kahneman (Eds. ), Heuristics and biases: The psychology ofintuitivejudgment (pp. 421-440). Cambridge, UK: Cambridge University Press. Stapel, D. A., & Velthuijsen, A. S. (1996). "Just as if it happened to me": The impact of vivid and self-relevant information on risk judgments. Journal of Social and ClinicalPsychology, 15(1), 102-119. Stein, R., & Agiesta, J. . (2009, October 22). Despite HlNl fears, many worry about vaccination, The Washington Post, p. 1. Stein, R., & Laris, M .. (2009). Scramble to parcel out HlNl vaccine; doctors and patients frustrated as supply falls short of promises, The Washington Post, p. 1. Stoddard, A. M., Zapka, J. G., & Schoenfeld, M. E. (1990). Effects of a news event on breast cancer screening survey responses. In P. F. Engstrom, B. Rimer & L. E. Mortenson (Eds. ), Advances in cancer control: Screening and prevention research (pp. 259-268). New York: John Wiley & Sons. Stryker, J. E. (2003). Media and marijuana: A longitudinal analysis of news media effects on adolescents' marijuana use and related outcomes, 1977-1999. Journal of Health Communication, 8(4), 305-328. doi: 10.1080/10810730305724 Tang, C. S.-K., & Wong, C.-Y. (2004). Factors influencing the wearing offacemasks to prevent the severe acute respiratory syndrome among adult Chinese in Hong Kong. Preventive Medicine, 39(6), ll87-1193. doi: 10.1016/j.ypmed.2004.04.032 156 The Pew Research Center. (2013). The State of the News Media 2013: An annual report on American journalism. Retrieved March 28, 2013, from http:/ /stateofthemedia. org/ Tversky, A., & Kahneman, D. (1971). Belief in the law of small numbers. Psychological Bulletin, 76(2), 105-110. doi: 10.1037/h0031322 Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5(2), 207-232. doi: 10.1016/0010- 0285(73)90033-9 Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124-1131. doi: 10.1126/science.l85.4157.1124 Tversky, A., & Kahneman, D. (1983). Extensional versus intuitive reasoning: The conjunction fallacy in probability judgment. Psychological Review, 90( 4), 293- 315. doi: 10.1037/0033-295X.90.4.293 Tyler, T. R. (1980). Impact of directly and indirectly experienced events: The origin of crime-related judgments and behaviors. Journal of Personality and Social Psychology, 39(1), 13-28. doi: 10.1037/0022-3514.39.1.13 Tyler, T. R. (1984). Assessing the risk of crime victimization: The integration of personal victimization experience and socially transmitted information. Journal of Social Issues, 40(1), 27-38. doi: 10.1111/j.l540-4560.1984.tb01080.x Tyler, T. R., & Cook, F. L. (1984). The mass media and judgments of risk: Distinguishing impact on personal and societal level judgments. Journal of Personality and Social Psychology, 47( 4), 693-708. doi: 10.1037/0022- 3514.47.4.693 US Census Bureau. (2013). Population tables for the United States. Available from US Census Bureau American FactFinder. Retrieved March 28, 2013 http:/ /factfinder2. census. gov /faces/nav/jsf/pages/ community _facts.xhtml Uslaner, E. M. (2002). The mora/foundations oftrust. New York: Cambridge University Press. Uslaner, E. M. (2003). Trust and risk: Implications for management. In M. Siegrist, T. C. Earle & H. Gutscher (Eds.), Trust in cooperative risk management: Uncertainty and scepticism in the public mind (pp. 73-93). Sterling, VA: Earthscan. Vasterman, P., Yzermans, C. J., & Dirkzwager, A. J. E. (2005). The role of the media and media hypes in the aftermath of disasters. Epidemiologic Reviews, 27(1), 107-114. doi: 10.1093/epirev/mxi002 Wallack, L., Dorfman, L., Jernigan, D., & Themba, M. (1993). Media advocacy and public health: Power for prevention. Newbury Park, CA: Sage. Walsh-Childers, K., & Brown, J. (2009). Effects of media on personal and public health. In J. Bryant & M. B. Oliver (Eds. ), Media effects: Advances in theory and research (3rd ed., pp. 469-489). New York: Routledge. Weaver, J., & Wakshlag, J. (1986). Perceived vulnerability to crime, criminal victimization experience, and television viewing. Journal of Broadcasting & Electronic Media, 30(2), 141-158. doi: 10.1080/08838158609386616 157 Wegener, D. T., & Petty, R. E. (1995). Flexible correction processes in social judgment: The role of naive theories in corrections for perceived bias. Journal of Personality and Social Psychology, 68(1), 36-51. doi: 10.1016/S0065-2601(08)60017-9 Weinstein, N. D. (1988). The precaution adoption process. Health Psychology, 7( 4), 355- 386. doi: 10.1037/0033-2909.92.2.382 Weinstein, N.D. (1993). Testing four competing theories of health-protective behavior. Health Psychology, 12( 4), 324-333. doi: 10.1037/0278-6133.12.4.324 Weinstein, N.D. (2000). Perceived probability, perceived severity, and health-protective behavior. Health Psychology, 19(1), 65-74. doi: 10.1037/0278-6133.19.1.65 Wheaton, B., Muthen, B., Alwin, D. , & Summers, G. (1977). Assessing reliability and stabillity in panel models. In D. R. Heise (Ed.), Sociological methodology (pp. 84- 136). San Francisco, CA: Jossey-Bass. WHO. (2005a). Avian influenza: Assessing the pandemic threat. Retrieved from http:/ /whqlibdoc. who.int/hq/2005/WHO _CDS_ 2005.29.pdf WHO. (2005b ). Responding to the avian influenza pandemic threat: Recommended strategic actions. Retrieved from http:/ /www.who.int/csr/resources/publications/influenza!WHO _CDS_ CSR _ GIP _ 05 _ 8-EN.pdf Witte, K., & Allen, M. (2000). A meta-analysis of fear appeals: Implications for effective public health campaigns. Health Education & Behavior, 27(5), 591-615. doi: 10.1177/109019810002700506 Witte, K., Meyer, G., & Martell, D .. (2001). Effective health risk messages: A step-by step guide. Thousand Oaks, CA: Sage. Y anovitzky, I. , & Stryker, J. 0. (200 1 ). Mass media, social norms, and health promotion efforts: A longitudinal study of media effects on youth binge drinking. Communication Research, 28(2), 208-239. doi: 10.1177/009365001028002004 Zajonc, R. B. (1980). Feeling and thinking: Preferences need no inferences. American Psychologist, 35(2), 151-175. doi: 10.1111/j.1469-8986.1979.tb0151l.x Zajonc, R. B. (1984). On the primacy of affect. American Psychologist, 39(2), 117-123. doi: 10.1037/0003-066X.39.2.117 Zimmerman, R. K., Santibanez, T. A., Janosky, J. E., Fine, M. J., Raymund, M., Wilson, S. A., ... Nowalk, M. P. (2003). What affects influenza vaccination rates among older patients? An analysis from inner-city, suburban, rural, and veterans affairs practices. The American Journal ofMedicine, 114(1), 31-38. doi: 10.1016/S0002- 9343(02)01421-3 158 APPENDIX A: ITEMS FROM ANHCS SURVEY FIELDED IN 2009 Core Questions Exposure to televised health news AB2. Some local or national television news programs include special segments of their newscasts which focus on health issues. About how often have you watched such health segments in the past 30 days? Not at all ............................................ 1 Less than once per week ................... 2 Once per week ................................... 3 A few times a week .......................... .4 Seasonal flu vaccination behavior BD 1. Have you been vaccinated against the flu for the most recent winter flu season (starting in October 2009)? Yes ................................................. 1 ~ .................................................. 2 Socio-demographic and physiological characteristics PPAGE Age PPEDUCAT Education (Categorical) PPETHM Race/Ethnicity PPGENDER Gender PPHHSIZE Household Size PPlNCIMP Household Income PPHEDIAB Derived: Has R ever had [Diabetes] (doctor diagnosed only) PPHE HIV Derived: Has R ever had [HIV] (doctor diagnosed only) PPHEHPRB Derived: Has R ever had [Heart problem or disease] (doctor diagnosed only) PPHESTRO Derived: Has R ever had [Stroke] (doctor-diagnosed only) PPHECANC Derived: Has Rever had [Cancer] (doctor diagnosed only) PPHEKIDN Derived: Has R ever had [Kidney disease] (doctor diagnosed only) PPHELUNG Derived: Has R ever had [Lung or breathing problems] (doctor diagnosed only) PPHE BMI Derived: Body mass index (BMI; weight x 704.5 in numerator/height x height) Module Questions Perceived benefit and risk of vaccine 1. Thinking about the various types of vaccination that you have received or that have been recommended to you by a health care provider or the govermnent, please indicate your level of agreement with each of the following statements. 159 In general, getting Strongly Disagree Neither Agree Strongly vaccinated ... disagree agree nor disagree Is an effective way to prevent illness Can make you sick Perceived susceptibility and severity regarding HlNl influenza, and HlNl flu vaccination intention Agree 2. The HlNl or "swine" flu is caused by a new flu virus that has been spreading around the world since March 2009. Please indicate the likelihood of the following events on a 10-point scale, 1 being not likely at all and 10 being extremely likely: Not at Extremely all Likely Likely 2 3 4 5 6 7 8 9 10 1 How likely are you to catch the HlNl flu between now and the end of March 2010? If you did catch the HlNl flu, how likely are you to die? How likely are you to get the HlNl vaccine when it becomes available if it is free? 160 Active attention to HlNl-related news 3. In the past 30 days, how often have you seen or heard any news about the H 1 N 1 or "swine" flu? Not at all--------------------------------------------------------------------1 ()nce-------------------------------------------------------------------------2 llwice------------------------------------------------------------------------3 About once a week--------------------------------------------------------4 Several times a week------------------------------------------------------5 Almost every day----------------------------------------------------------6 Don't know----------------------------------------------------------------9 5 Social confidence in health authorities 4. How much confidence do you have in each of the following institutions' ability to effectively deal with the HlNl or "swine" flu? Not Not very Somewhat Very Don't confident confident confident confident know at all The Centers for Disease Control and Prevention or CDC The health department of your state government The health department of your local government 5. Have you been vaccinated against the HlNl or "swine" flu? Yes ....................................................... 1 No ......................................................... 2 161 APPENDIX B: ITEMS FROM ANHCS SURVEY FIELDED IN 2012 Core Questions Exposure to televised health news AB2. Some local or national television news programs include special segments of their newscasts which focus on health issues. About how often have you watched such health segments in the past 30 days? Not at all ............................................ 1 Less than once per week ................... 2 Once per week ................................... 3 A few times a week .......................... .4 Seasonal flu vaccination behavior BD 1. Have you been vaccinated against the flu for the most recent winter flu season (starting in September 2011)? Yes ................................................. 1 ~ .................................................. 2 Socio-demographic and physiological characteristics PPAGE Age PPEDUCAT Education (Categorical) PPETHM Race/Ethnicity PPGENDER Gender PPHHSIZE Household Size PPlNCIMP Household Income PPHEDIAB Derived: Has R ever had [Diabetes] (doctor diagnosed only) PPHE HIV Derived: Has R ever had [HIV] (doctor diagnosed only) PPHEHPRB Derived: Has R ever had [Heart problem or disease] (doctor diagnosed only) PPHESTRO Derived: Has R ever had [Stroke] (doctor-diagnosed only) PPHECANC Derived: Has Rever had [Cancer] (doctor diagnosed only) PPHEKIDN Derived: Has R ever had [Kidney disease] (doctor diagnosed only) PPHELUNG Derived: Has R ever had [Lung or breathing problems] (doctor diagnosed only) PPHE BMI Derived: Body mass index (BMI; weight x 704.5 in numerator/height x height) 162 Module Questions Perceived benefit and risk of vaccine 1. Thinking about the various types of vaccination that you have received or that have been recommended to you by a health care provider or the government, please indicate your level of agreement with each of the following statements. In general, getting Strongly Disagree Neither Agree Strongly vaccinated ... disagree agree nor Agree disagree Is an effective way to prevent illness Can make you sick Integral affect 2. We are interested in knowing your feeling toward getting vaccinated for the flu using the scale below. Extremely 2 3 4 5 6 7 8 9 Extremely Negative Positive 1 10 My feeling toward 'getting vaccinated for the flu' is .. Social trust in FDA 3. The Food and Drug Administration (FDA) is the federal government agency responsible for approving and regulating vaccines and medicines in the US. Please indicate your level of agreement with the following statements. Strongly Disagree Neither Agree Strongly disagree 2 agree nor 4 Agree 1 disagree 5 3 a. The FDA can be trusted b. The FDA communicates honestly about possible side effects of vaccines c. Should it turn out 163 that a vaccine is harmful, the FDA would openly and honestly inform the public d. Should it turn out that a vaccine is harmful, the FDA would protect the interests of pharmaceutical compames H5N 1 flu vaccination intention 4. The next question asks about a hypothetical flu vaccine, for H5Nl avian influenza. Since 2003, H5Nl avian influenza, also known as the bird flu, has sickened individuals in 15 countries and led to hundreds of deaths. Public health authorities are concerned that avian influenza might spread to the US this coming winter flu season (between October 2012 and March 2013). If a vaccine against avian influenza was developed and approved for use by September 2012, how likely are you to get this new vaccine for the coming flu season, assuming that it is readily available and free? Not at all 2 3 4 5 6 7 8 9 Extremely Likely Likely 1 10 Self-efficacy for vaccination 5. If you wanted to, how confident are you that ... Not at all 2 3 4 5 6 7 8 9 Extremely Confident Confident 1 10 a. you could get the SEASONAL flu vaccine this coming winter flu season (between October 2012 and March 2013)? b. you could get the new AVIAN flu vaccine this coming winter season (between October 2012 and March 2013)? Perceived susceptibility to seasonal and avian influenza 6. Think about the coming flu season (between October 2012 and March 2013) and imagine that avian influenza has indeed spread to the US. Please provide your best estimates to the following questions. 164 Notal Extremely all Likely Likely 2 3 4 5 6 7 8 9 10 1 a. How likely are you to catch SEASONAL influenza this coming flu season if you didn't get the seasonal flu vaccine? b. How likely are you to catch AVIAN influenza this coming flu season if you didn't get the avian flu vaccine? Perceived severity of seasonal and avian influenza 7. Thinking about your ability to carry out your usual activities, how much do you think your usual activities would be affected if you ... Not at all 2 3 4 5 6 7 8 9 Extremely Disrupted Disrupted 1 10 a. Got SEASONAL influenza this coming winter flu season (between October 2012 and March 2013)? b. Got AVIAN influenza this coming winter flu season (between October 2012 and March 2013)? Social norms regarding seasonal flu and H5Nl flu vaccination 8. How much do you think each of the following people would approve of you getting the seasonal flu vaccine for this coming flu season? 165 Notal 2 3 4 5 6 7 8 9 A Great all Deal I 10 a. Your family members d. Your friends c. Your health care provider 9. hnagine that you are getting the newly-developed avian flu vaccine for this coming flu season. How much do you think each of the following people would approve of you getting this avian flu vaccine? Notal 2 3 4 5 6 7 8 9 A Great all Deal I 10 a. Your family members d. Your friends c. Your health care provider 10. How much does each of the following people's opinion on health matter to you? Notal 2 3 4 5 6 7 8 9 A Great all Deal I 10 a. Your family members d. Your friends c. Your health care provider

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Contagious: social norms about health in work group networks

PDF

Rediscovering the communication engine of neighborhood effects: how the interaction of residents and community institutions impacts health literacy and how it can be leveraged to improve health c...

PDF

The influence of a humorous intervention on parent-child sexual communication

PDF

Relationship formation and information sharing to promote risky health behavior on social media

PDF

The effects of a COVID-19 narrative message on beliefs and intentions among vaccine hesitant parents

PDF

Social media and health: social support and social capital on pregnancy-related social networking sites

PDF

Understanding normative influence of neighborhoods: a multilevel approach to promoting Latinas’ cervical cancer prevention behaviors in urban ethnic communities

PDF

The power of social media narratives in raising mental health awareness for anti-stigma campaigns

PDF

Mobile self-tracking for health: validating predictors, effects, mediator, moderator, and social influence

PDF

The role of social media in influencing consumer opinions and preferences in the fintech industry

PDF

Bounded technological rationality: the intersection between artificial intelligence, cognition, and environment and its effects on decision-making

PDF

The formation and influence of online health social networks on social support, self-tracking behavior and weight loss outcomes

PDF

Traditional media, social media or word-of-mouth? Examining the information sources that influence Chinese students' decision-making processes when applying to U.S. graduate schools

PDF

The effect of Twitter social support on health outcomes and its mediators: a randomized controlled trial of a social support intervention on Twitter for patients affected by cancer

PDF

The influencer break-down: a history, categorization, and analysis of success for health & wellness influencers

PDF

Examining influences on the decision-making process of Filipino students and parents in selecting a post-secondary institution

PDF

Towards health-conscious spaces: building for human well-being and performance

Asset Metadata

Creator Chen, Nien-Tsu Nancy (author)

Core Title Developing and testing a heuristic-systematic model of health decision making: the role of affect, trust, confidence and media influence

School Annenberg School for Communication

Degree Doctor of Philosophy

Degree Program Communication

Publication Date 07/23/2015

Defense Date 05/14/2013

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag affect,confidence,health behavior change,health decision making,heuristics,influenza vaccine,media effects,OAI-PMH Harvest,trust

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Cody, Michael J. (committee chair), Mayer, Doe (committee member), Murphy, Sheila T. (committee member)

Creator Email nientsuc@usc.edu,nntc56@hotmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-295240

Unique identifier UC11293314

Identifier etd-ChenNienTs-1813.pdf (filename),usctheses-c3-295240 (legacy record id)

Legacy Identifier etd-ChenNienTs-1813.pdf

Dmrecord 295240

Document Type Dissertation

Format application/pdf (imt)

Rights Chen, Nien-Tsu Nancy

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA