Explicating blame: examining the role of blame in the context of childhood obesity

As an important psychological construct, blame has not been clearly explicated and has been understudied in communication research. Situated in the context of childhood obesity, this dissertation aims to reveal the underlying structure of blame and examine its mediating role between media messages and social responses. A pilot study, using a convenience sample of college students, and a main study, using a general population sample from MTurk, which is marketplace for work that requires human intelligence, were carried out. The pilot study employed a 2 (outcome severity: high vs. low) ×2 (outcome preventability: high vs. low) between-participants experimental design (N =118), and the main study was of a 2 (outcome severity high vs. low) x 2 (outcome preventability high vs. low) ×2 (disease type: asthma attack vs. heart attack) design (N =373). Results showed that blame should be conceptualized as an amalgam of cognition and affect. Outcome preventability was found to influence the amount of blame, and that its effect was further enhanced when the outcome was described as more severe. Blame then led to greater policy support, information sharing behaviors, civic participation activities, and punishment. Blame was a significant mediator between outcome preventability and most of these social responses. There were also a few moderating effects, from need for cognition, need for affect, and moral identity, on the relationship between blame and social responses. Overall, this research extended the scope of the current research on blame and contributed to an understanding about the nature of blame and as well as its role in communication processes.

EXPLICATING BLAME: EXAMINING THE ROLE OF BLAME IN THE CONTEXT OF CHILDHOOD OBESITY by Miao Liu A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Communication The University of Utah December 2018 Copyright © Miao Liu 2018 All Rights Reserved The University of Utah Graduate School STATEMENT OF DISSERTATION APPROVAL The dissertation of Miao Liu has been approved by the following supervisory committee members: Ye Sun , Chair 4/30/2018 Date Approved Jakob D. Jensen , Member 4/30/2018 Date Approved Sara K. Yeo , Member 4/30/2018 Date Approved Leona Yi-fan Su , Member 4/30/2018 Date Approved Ming Wen , Member 4/30/2018 Date Approved and by Danielle Endres the Department/College/School of and by David B. Kieda, Dean of The Graduate School. , Chair/Dean of Communication ABSTRACT As an important psychological construct, blame has not been clearly explicated and has been understudied in communication research. Situated in the context of childhood obesity, this dissertation aims to reveal the underlying structure of blame and examine its mediating role between media messages and social responses. A pilot study, using a convenience sample of college students, and a main study, using a general population sample from MTurk, which is marketplace for work that requires human intelligence, were carried out. The pilot study employed a 2 (outcome severity: high vs. low) ×2 (outcome preventability: high vs. low) between-participants experimental design (N =118), and the main study was of a 2 (outcome severity high vs. low) x 2 (outcome preventability high vs. low) ×2 (disease type: asthma attack vs. heart attack) design (N =373). Results showed that blame should be conceptualized as an amalgam of cognition and affect. Outcome preventability was found to influence the amount of blame, and that its effect was further enhanced when the outcome was described as more severe. Blame then led to greater policy support, information sharing behaviors, civic participation activities, and punishment. Blame was a significant mediator between outcome preventability and most of these social responses. There were also a few moderating effects, from need for cognition, need for affect, and moral identity, on the relationship between blame and social responses. Overall, this research extended the scope of the current research on blame and contributed to an understanding about the nature of blame and as well as its role in communication processes. TABLE OF CONTENTS ABSTRACT.......................................................................................................................iii LIST OF TABLES ..........................................................................................................viii LIST OF FIGURES ............................................................................................................x ACKNOWLEDGMENTS..................................................................................................xii Chapters 1 INTRODUCTION ........................................................................................................... 1 Media Coverage, Public Perception, and Blame .....................................................3 Parent-focused Blame for Childhood Obesity..........................................................6 Blame in Communication Research......................................................................... 7 Chapter Outlines.......................................................................................................9 2 SITUATING BLAME AND EXPLICATING THE CONCEPT OF BLAME...............10 Situating Blame: Antecedents and Outcomes....................................................... 10 Antecedents of Blame................................................................................10 The severity of consequence…......................................................11 Causality........................................................................................12 Intentionality..................................................................................14 Preventability.................................................................................15 Outcomes of Blame....................................................................................18 Explicating the Construct of Blame.......................................................................22 Setting the Definitional Boundary.............................................................22 The Nature of Blame.................................................................................24 Blame as cognition.......................................................................24 Blame as affect.............................................................................26 Blame as both cognition and affect..............................................27 A Tentative Definition of Blame...............................................................28 Moral indignation as the affective aspect of blame......................31 Research Questions....................................................................................32 Chapter Summary.......................................................................................32 3 SOCIAL RESPONSES FOLLOWING BLAME: HYPOTHESES AND RESEARCH QUESTIONS.......................................................................................................................42 Social Responses Following Blame.......................................................................43 Policy Support ...........................................................................................43 Information Sharing Behavior....................................................................44 Civic Participation......................................................................................45 Punishment.................................................................................................46 Individual Traits as Potential Moderators.................................................49 Hypotheses and Research Questions......................................................................53 4 THE PILOT STUDY: USING A COLLEGE STUDENT SAMPLE............................56 Methods..................................................................................................................56 Power Analysis ..........................................................................................56 Participants.................................................................................................57 Message Design..................................................................................... 57 Measures...................................................................................... 58 Blame as cognition ........................................................................58 Moral indignation ..........................................................................59 Civic participation ........................................................................59 Information sharing behavior.........................................................60 Support for policies .......................................................................60 Need for affect ..............................................................................61 Moral identity ...............................................................................61 Analysis Strategies.....................................................................................62 Results....................................................................................................................64 Manipulation Check...................................................................................64 The Structure of Blame (RQ1a and RQ1b) ...............................................65 Antecedents of Blame (Hypothesis 1a to 1c) ..........................................67 Outcomes of Blame (H2-H4) ....................................................................67 Moderators.................................................................................................68 Summaries and Implications of the Pilot Study........................................68 5 THE MAIN STUDY: USING THE GENERAL POPULATION...............................85 Methods................................................................................................................85 Power Analysis....................................................................................... 85 Participants............................................................................................. 85 Message Design......................................................................................87 Measures................................................................................................88 Blame as cognition….....................................................................88 Moral indignation...........................................................................88 Policy support............................................................................. 88 v Information sharing behavior....................................................89 Civic participation...........................................................................90 Punishment......................................................................................90 Moderators.......................................................................................90 Need for cognition...........................................................................91 Moral identity..................................................................................91 Control variables.............................................................................91 Analysis Strategies ....................................................................................92 Results....................................................................................................................93 Manipulation Check...................................................................................93 The Structure of Blame (RQ1a and RQ1b) ...............................................94 Antecedents of Blame (Hypothesis 1a to 1c) ............................................95 Outcomes of Blame (H2-H4) ....................................................................96 Moderating Effects (R2a & R2b) ..........................................................................97 Need for cognition........................................................................97 Need for affect..............................................................................97 Moral identity……..….……………............................................98 Summaries of the Main Study…….…………............................................98 6 DISCUSSION ..............................................................................................................125 Summaries of Key Findings..................................................................................125 The Structure of Blame………………....................................................126 Antecedents of Blame..............................................................................127 Social Responses Following Blame.........................................................127 Policy support...............................................................................128 Information sharing behavior........................................................128 Civic participation.........................................................................129 Punishment....................................................................................129 Moderator Effects (RQ1&RQ2) ..............................................................130 Limitations and Future Directions.........................................................................133 Conclusion.............................................................................................................137 Appendices A: STIMULUS 1: HIGH PREVENTABILITY*SEVERE CONSEQUENCE….............141 B: STIMULUS 2: LOW PREVENTABILITY*MILD CONSEQUENCE…...…......…..142 C: STIMULUS 3: HIGH PREVENTABILITY*MILD CONSEQUENCE…...…….......143 D: STIMULUS 4: LOW PREVENTABILITY*SEVERE CONSEQUENCE…..............144 vi E: STIMULUS 1A: HIGH PREVENTABILITY*MILD CONSEQUENCE (ASTHMA ATTACK) ……………………………………………………………….…………..…145 F: STIMULUS 1B: HIGH PREVENTABILITY*SEVERE CONSEQUENCE (HEART ATTACK) ………………………..…………………………………………….……....146 G: STIMULUS 2A: LOW PREVENTABILITY*SEVERE CONSEQUENCE (ASTHMA ATTACK) ….…..………………………………………………………...………….....147 H: STIMULUS 2B: LOW PREVENTABILITY*MILD CONSEQUENCE（HEART ATTACK .......………………………………………………………………….……..148 I: STIMULUS 3A: HIGH PREVENTABILITY*SEVERE CONSEQUENCE (ASTHMA ATTACK）…………………………………………………………..………………...149 J: STIMULUS 3B: HIGH PREVENTABILITY*SEVERE CONSEQUENCE (HEART ATTACK）…………………………………………………………..……………..….150 K: STIMULUS 4A: LOW PREVENTABILITY*SEVERE CONSEQUENCE （ASTHMA ATTACK）……...…………………………………………..………..…151 L: STIMULUS 4A: LOW PREVENTABILITY*SEVERE CONSEQUENCE （ASTHMA ATTACK）………..……………………………………………….....…152 REFERENCES……………………………………………………………………..…..153 vii LIST OF TABLES 1: Definitions of Blame........................................................................................... 33 2: Correlations of the Key Variables for the Pilot Study............................ .............70 3: Fit Indices and Model Comparison for the Pilot Study Using Nontax Policy Support as the Outcome...................................................................................... ................71 4: Fit Indices and Model Comparison for the Pilot Study Using Civic Participation as the Outcome...................................................................................................... 72 5: Fit Indices and Model Comparison for the Pilot Study Using Information Sharing as the Outcome...................................................................................................... 73 6: Comparisons of Scales Used in the Pilot Study and the Main Study...................74 7: Results of Mean, Standard Deviations, and T-test of Heart Attack and Asthma Attack Group..................................................................................................... ...............100 8: Zero-Order Correlations Matrix for the Main Study (asthma attack group, N = 193) ...............................................................................................................................101 9: Zero-Order Correlations Matrix for the Main Study (heart attack group, N = 180) ................................................................................................................................102 10: Fit Indices and Model Comparison for the Main Study Using Punishment as the Outcome (Asthma Attack Group, N = 193) .....................................................................103 11: Fit Indices and Model Comparison for the Main Study Using Punishment as the Outcome (Heart Attack Group, N = 180) .........................................................................104 12: Fit Indices and Model Comparison for the Main Study Using Parent-Oriented Policy Support as the Outcome (Asthma Attack Group, N =193) .....................................105 13: Fit Indices and Model Comparison for the Main Study Using Parent-Oriented Policy Support as the Outcome (Heart Attack Group, N = 180) .............................106 14: Fit Indices and Model Comparison for the Main Study Using Public-Oriented Policy Support as the Outcome (Asthma Attack Group, N = 193) ....................................107 15: Fit Indices and Model Comparison for the Main Study Using Public-Oriented Policy Support as the Outcome (Heart Attack Group, N =180) ...............................108 16: Fit Indices and Model Comparison for the Main Study Using Civic Participation as the Outcome (Asthma Attack Group, N =193) ............................................................109 17: Fit Indices and Model Comparison for the Main Study Using Civic Participation as the Outcome (Heart Attack Group, N = 180) ...............................................................110 18: Fit Indices and Model Comparison for the Main Study Using Information Sharing Behavior as the Outcome (Asthma Attack Group, N =193) ........................................111 19: Fit Indices and Model Comparison for the Main Study Using Information Sharing Behavior as the Outcome (Heart Attack Group, N = 180) ……………............................112 20: Indirect Effects From Exogenous Variables to Outcome Variables Through Blame (asthma attack) ...............................................................................................................113 21: Indirect Effects From Exogenous Variables to Outcome Variables Through Blame (heart attack) ........................................................................................... ...............114 22: Research Questions and Hypotheses, and Summary of Findings......................139 ix LIST OF FIGURES 1: Single Process Affective Model.......................................................................38 2: Single Process Cognitive Model......................................................................39 3: The Dual Process Cognitive-Affective Model.................................................40 4: The Intertwined Process Cognitive-Affective Model......................................41 5: Interaction Effect Between Preventability and Outcome Severity on Perceived Level of Outcome Severity in the Pilot Study..................................................................79 6: Interaction Effect Between Preventability and Outcome Severity on Perceived Level of Preventability in the Pilot Study.........................................................................80 7: The Obtained Single Process Affective Model for the Pilot Study..................81 8: The Obtained Single Process Cognitive Model for the Pilot Study.................82 9: The Obtained Dual Process Model for the Pilot Study. ..................................83 10: The Obtained Intertwined Process Model for the Pilot Study........................84 11: Interaction Effect Between Preventability and Outcome Severity on Perceived Level of Outcome Preventability in the Asthma Attack Data..............................................115 12: Interaction Effect Between Preventability and Outcome Severity on Perceived Level of Outcome Severity in the Asthma Attack Data..........................................................116 13: Interaction Effect Between Preventability and Outcome Severity on Perceived Level of Outcome Preventability in the Heart Attack Data ...........................................117 14: Interaction Effect Between Preventability and Outcome Severity on Perceived Level of Outcome Severity in the Heart Attack Data..........................................................118 15: The Obtained Dual Process Model for the Asthma Attack Group in Main Study ...............................................................................................................................119 16: The Obtained Dual Process Model for the Heart Attack Group in the Main Study................................................................................................................................120 17: The Obtained Intertwined Model for the Asthma Attack Group in the Mai n Study................................................................................................................................121 18: The Obtained Intertwined Model for the Heart Attack Group in the Main Study................................................................................................................................122 19: Standardized Path Coefficients for the Intertwined Process Model with Five Outcomes for the Asthma Attack Group in the Main Study (N =193) ….………………123 20: Standardized Path Coefficients for the Intertwined Process Model with Five Outcomes for the Heart Attack Group in the Main Study (N =180) …….………………124 xi ACKNOWLEDGMENTS The dissertation has evolved through many years of research. I am most grateful to my dissertation chair, Ye Sun, for her unfailing support and generous guidance throughout the past 6 years. This dissertation would not have been possible without her patience and encouragement. She has devoted unimaginable time and resources to this dissertation project. I have been taught so much more than methods and theory in this dissertation project. I would also like send my sincere thanks to my doctoral committee member, Professor Jakob Jensen. He has helped me immensely during the past few years. Hearty thanks also goes to Professor Sara Yeo and Professor Leona Yi-fan Su. They read my work carefully and gave me very helpful suggestions and comments on my project, which has helped improve my dissertation greatly. I would also like to thank my committee member, Professor Ming Wen. She gave me much encouragement and support along the way. I would like to extend my special thanks to my parents, Juan Ge and Qiang Liu, for their unconditional love and continuous support from the beginning. Their constant encouragement helped me go through all the difficult times. I am proud to have such wonderful parents. Many thanks go to my friends at University of Utah, including Yuqing Qiu, Shuo Tan, Yiqing Yang, Penchan Phoborisut, Tingting Pang, and Yingxie Li. Their support meant so much to me when I was working on doing my dissertation. CHAPTER 1 INTRODUCTION Over the past few decades, childhood obesity has emerged as a serious national public health concern in the United States. The rate of childhood obesity has been rising despite increasing efforts to reverse the trend (Spark, 2011). In the United States, the percentage of obese children aged 6-11-years-old increased from 7% in 1980 to nearly 18% in 2012 (CDC, 2015). Globally, the number of overweight or obese children increased from 32 million in 1990 to 41 million in 2016 (WHO, 2017). The problem is worse for developing countries, where the increase of childhood obesity is 30% higher than that of developed countries (WHO, 2017). Childhood obesity has a wide range of consequences. Obese children are at higher risk of having chronic health conditions, including cardiovascular diseases, type 2 diabetes, high-blood pressure, asthma, sleep apnea, and bone and joint problems (CDC, 2018). Research has shown that childhood obesity is a strong predictor of adult obesity (SpruijtMetz, 2011). As declared by former surgeon Richard Carmona, “(b)ecause of the increasing rates of obesity, we may see the first generation that will be less healthy and have a shorter life expectancy than their parents” (Carmona, 2004). The negative impact of childhood obesity goes beyond health. Obese children are bullied more than their normal-weight peers, and they are more likely to suffer from social isolation, depression, 2 and lower self-esteem (Latner & Stunkard, 2003). At a broader level, childhood obesity places a substantial economic burden on the healthcare systems (Cawley, 2010). Over a lifetime, an obese child has a medical cost of $19000 compared to those children of normal weight (Finkelstein, Trogdon, Cohen, & Dietz, 2009). Causes of childhood obesity encompass a combination of genetic, behavioral, and socioenvironmental factors (Dehghan, FAkhtar-Danesh, & Merchant, 2005; Ebbeling, Pawlak, & Ludwig, 2002; Kumanyika, 2008). Genetic differences in the adult population explained a significant proportion of the variation in body mass index (BMI; Herbert et al., 2006; Locke et al., 2015; Silventoinen & Kaprio, 2009). One study showed that the Pima Indians were significantly more prone to obesity-risk factors than Caucasian and Asians (Knowler, Pettitt, Saad, & Bennett, 1990). It is also found that temporal increases in the consumptions of sugar-sweetened beverages interacted with the genetic predisposition to adiposity to influence obesity (Qi et al., 2012). Obesity is therefore a complex problem subject to intricately interacting genetic and nongenetic factors (Han, Lawlor, & Kimm, 2010). Given its severity and complexity, collaborative efforts with parents, schools, policy makers, health professionals, researchers, and the media have been suggested to address childhood obesity (Desjardins & Schwartz, 2007; Gruber & Haldeman, 2009). Aiming to contribute to this larger enterprise, this dissertation chooses childhood obesity as the context of inquiry, with specific interest in evoking social responses that potentially address the influences of family and social environment in treating the problem. This dissertation also situates the empirical investigation within the theoretical framework of blame. It seeks to explicate the construct of blame and understand its mediating role in a 3 process where media messages about childhood obesity elicit relevant social responses. Media Coverage, Public Perception, and Blame As with other social problems, media plays a vital role in shaping the public’s perceptions of obesity (Boero, 2013). Previous research has shown that news coverage of health issues influences perceived importance of these issues, policy preferences, and specific health behaviors (Holder & Treno, 1997; Leask, Hooker, & King, 2010; Li, Chapman, Agho, & Eastman, 2007; Major, 2009; Pierce & Gilpin, 2001; Weeks, Friedenberg, Southwell, & Slater, 2012). Over the past 30 years, there was an exponential increase in the news coverage of obesity and childhood obesity (Spark, 2011). A few content analyses of the media coverage of obesity converged to demonstrate that the media discourse on obesity highlight personal causes and solutions of obesity over societal attributions of responsibility (Kim & Wills, 2007; Lawrence, 2014; Yoo & Kim, 2012). News coverage of childhood obesity also exhibited such an imbalance, with newspaper articles attributing obesity to individual rather than social causes (Hawkins & Linvill, 2010). The dominant media frame of individual responsibility shapes the public’s understanding of who is responsible. A study using a representative sample showed that the majority of the participants (over 70%) attributed obesity to personal causes, such as overeating and physical inactivity (Hilbert, Rief, & Braehler, 2008). An attribution analysis, which determines the causes of 10 different diseases (AIDS, Alzheimer’s disease, blindness, cancer, child abuser, drug addiction, heart disease, obesity, paraplegia, and Vietnam War syndrome), revealed that obesity was rated high by the public on personal responsibility and controllability, along with AIDS, child abuse, and drug addiction 4 (Weiner, Perry, & Magnusson, 1988). Obese individuals were also rated low on being liked, evoking little pity and more anger, and eliciting low help-giving intentions. Studies have also shown that when the media story emphasized other factors, such as genetics (Jeong, 2007) or societal factors contributing to obesity (Sun, Krakow, John, Liu, & Weaver, 2016), it helped significantly reduce personal attributions as well as increase helping behaviors, interpersonal discussion about obesity, as well as participatory behaviors for obesity prevention. Through various forms, blame has been present in various media channels. Considering media has an unneglectable role in shaping the public’s interpretation of a social issue, blame could directly or indirectly influence the public’s understanding of a social issue when it appears as a core theme in media messages. Blame in news coverage often takes the form of assigning responsibility to individuals or entities by providing relevant facts and evidence (Holton, Weberling, Clarke, & Smith, 2012). Blame also takes the form of stigmatizing and shaming people by using dehumanizing or biased words in some media messages (Russell, Cameron, Socha, & McNinch, 2014). For example, explicit blame directed at obese people that is embedded in stigmatizing messages has appeared in obesity prevention campaigns. Even though these public health campaigns might be effective in terms of raising awareness of the obesity problem, they nevertheless are susceptible to ethical consideration because such campaigns often involve stigmatization toward obese people and fat shaming. A controversial ad campaign launched in Georgia against childhood obesity in 2011 is an exemplar of an ethical trade-off (Grinberg, 2012). This campaign used the stigmatizing messages such as “It’s hard to be a little girl if you’re not” and “Being fat takes the fun out of being a kid.” Thus, blame as a message strategy 5 should be adopted with ethical and moral considerations. Linking media coverage of health-related issues and the public’s responses, blame has often appeared in such discussions either as a general idea or as a focal concept. One frequent reference is the notion of victim-blaming, which occurs when media coverage treats those who are suffering from the consequence as perpetrators themselves. Lung cancer patients who are smokers, for example, are regarded as blameworthy, so are AIDS patients, alcoholics, drug users, and obese individuals (Cataldo, Jahan, & Pongquan, 2012; Herek, Capitanio, & Widaman, 2003; Kirkwood & Brown, 1995; Saguy, Gruys, & Gong, 2010; Schellenberg & Bem, 1998; Schiller, Crystal, & Lewellen, 1994; Weiner, 2005). For health issues, blaming serves as a means of interpreting risks or diseases in an accessible and comprehensible way, and therefore making them possibly controllable (Nelkin, 1991). Blame has been frequently used in media messages to guide individuals to take actions that could benefit their health. One study found that the images blaming individuals for causing obesity (e.g., close-up images of obese adults eating junk food) served as fear appeal among people who are not overweight, prompting them to increase healthy behaviors (Young, Subramanian, & Hinnant, 2016). With a target for blame, individuals identify causes of the problematic conditions, locate the responsible agents, and figure out possible solutions. Who or what to blame for health issues is often a site of contestation. Lung cancer patients felt unjustly blamed for their illness, for example, and held tobacco companies with unscrupulous policies responsible for many lung cancer cases (Chapple, Ziebland, & McPherson, 2004). One the other hand global marketing campaigns promoting caloriedense, nutrient-poor, and high-sugar diets are blamed for childhood obesity (Harris, 6 Pomeranz, Lobstein, & Brownell, 2009). These industries keep evoking the values of free choice and individual responsibility. Such “blame wars” are prevalent and ongoing in various public health domains. Parent-focused Blame for Childhood Obesity As mentioned earlier, childhood obesity is attributable to a host of factors. This dissertation focuses on addressing family influence, or in other words, the role of parents, with news stories with varied emphasis on the irresponsible behavior of parents in causing the harmful consequences to the child featured in the story. I chose this focus for the following reasons. Among the various factors that lead to childhood obesity, the role of parents has been emphasized by both media and health professionals. Despite children’s autonomy, legally and ethically the responsibility to make health-related choice, such as providing healthy food and encouraging physical activity, falls to the parents (Barry, Brescoll, & Gollust, 2013; Perryman & Sidoti, 2015). As Featherstone (2003) suggested, “contemporary social policy frameworks aim to maintain children as the responsibility of individual families” (p.3). However, not all parents promote or model healthy behaviors to their children (Crossman, Anne Sullivan, & Benin, 2006). A report revealed by the leading global network for physicians, SERMO found that 69% of doctors thought parents were either completely or mostly to blame for childhood obesity. In addition, the health conditions of overweight or obese children who do not have obese-related illness are often neglected by their parents (Siddique, 2016). A recent study revealed that as high as 78.4% of parents 7 perceived their obese children as having “normal weight” (Duncan, Hansen, Wang, Yan, & Zhang, 2015). Research suggested that bad eating habits were the main cause of childhood obesity (Kuźbicka & Rachoń, 2013). The increases in parents’ working hours, the popularity of frozen food, and inadequate intake of vegetables and fruit, significantly contribute to children’s poor eating habits (Mariz et al., 2015). Parental modeling theory contended that parents’ own behaviors could have a powerful effect on children’s behaviors (Crossman et al., 2006). One study revealed that parental feeding practices and parents’ own behaviors significantly impacted children’s obesity risk (Anzman, Rollins, & Birch, 2010), and another study showed that parent inactivity was a strong and positive predictor of children’s physical inactivity (Fogelholm, Nuutinen, Pasanen, Myöhänen, & Säätelä, 1999). Crossman et al. (2006) studied the effect of family environment and adolescents’ behaviors on their weight status over 6 years using data from the United States National Longitudinal Study of Adolescent Health. Their study found that family environment exerted strong influence on adolescents’ weight that lasted into their young adulthood. Blame in Communication Research Most of the existing research on blame has been in the field of social psychology. In communication research, attention to the construct of blame has been rather limited, with some studies using content analysis or discourse analysis to map out how media coverage attributes blame to different agents (Holton, Weberling, Clarke, & Smith, 2012; Lau, 2009; O’Reilly, 2014; Romer, Jamieson, Riegner, Emori, & Rouson, 1997; Rowland, 2012). Some other studies examined how news content influenced audiences’ attribution 8 of blame (An, 2011; Gilmore, Meeks, & Domke, 2013; Hameleers, Bos, & Vreese, 2017; Wachs, Cooky, Messner, & Dworkin, 2012). How blame as a psychological response to mediated communication, which in turn triggers social responses or communicative behaviors, has rarely been studied. For communication researchers, understanding the role of blame in the communication process could have both theoretical and practical significance. Theoretically, blame as a psychological response may provide another explanatory mechanism that accounts for media influence on the public’s responses to certain social issues. These issues are often represented by communicative behaviors. In the context of communication research, blame as the innate motivation to communication is understudied. The communicative behaviors such as information sharing that are driven by blame not only reflect one’s attitude toward a particular social issue but also represent one’s willingness to engage in this issue. This study is designed in a way to capture the explanatory power of blame in predicting communicative behaviors. Practically, understanding the antecedents and consequences of blame could shed light on message strategies with implications for changing people’s attitudes and promoting desired behaviors. Because blame is a psychological response that is associated with the causes and consequence of a negative event, health communication researchers could use blame as a focal concept in public health campaigns and health education to shape the public’s understanding of health issues and promote healthy behaviors. This dissertation aims to explore the nature of blame as a psychological response to mediated communication about a negative event. Drawing on the existing literature of blame in social psychology (Alicke, 2000; Malle, Guglielmo, & Monroe, 2014; Shaver, 9 1985), this dissertation has three goals. First, it seeks to advance a concept explication of blame and examine its empirical properties. Specifically, I compare four conceptual models of blame that have been proposed in prior research to answer the question of what is the underlying structure of blame. Second, the project examines blame as a mediating construct between exposure to media message and social responses relevant to the issue of childhood obesity. Third, it explores whether this communication process, mediated by blame, varies with certain individual traits variables. Chapter Outlines Following the introduction chapter, Chapter 2 lays out the theoretical framework. Following a literature review on the antecedents and outcomes of blame, I will propose a tentative definition of blame and two research questions on the nature of blame. Chapter 3 introduces the four social responses that this dissertation focuses on as outcome variables (e.g., policy support, civic participation, information sharing, and punishment), and outlines the hypotheses and research questions. Chapter 4 reports the goals, the design, and the results of the pilot study (using a college student sample). Chapter 5 presents the methods and the results of the main study (using an MTurk sample, which are the participants that recruited from marketplace for work that requires human intelligence). Finally, Chapter 6 summarizes the key findings of this dissertation and discusses the limitations, implications, and directions for future research. CHAPTER 2 SITUATING BLAME AND EXPLICATING THE CONCEPT OF BLAME This chapter lays out the theoretical framework for this dissertation. It consists of two parts. First, to get a more complete picture of blame, it is necessary to examine what precedes and follows blame. A review of the existing research on the antecedents and outcomes of blame is provided first. Situated in this process, blame is a psychological response elicited by these antecedents and leading to outcomes of interest. In the second part, I focus on the question what is blame. Based on a review of previous conceptual and operational work on blame, this dissertation proposes and discusses the explication of blame that guides the empirical efforts of this project. Situating Blame: Antecedents and Outcomes Antecedents of Blame Previous research has proposed various conditions under which blame is generated. For example, Shaver (1985) suggested that the extent of blame depended on the association between actor and outcome, causality, intentionality, and foreseeability. Alicke’s (2000) work proposed that individuals assess a series of “blame criteria” such as causality, intentionality, and controllability, before assigning blame (Alicke, 2000; Lagnado & 11 Channon, 2008; Malle et al., 2014; Mikula, 2013 Shaver, 1985). Cutting across various theoretical frameworks are a few antecedents of blame, such as intentionality, controllability, foreseeability, and causality. Below I highlighted four antecedents of blame that have been discussed in prior research, though sometimes under slightly different labels, which include the severity of consequence, causality, intentionality, and preventability. The severity of consequence. Broadly speaking, blame is a psychological response to a negative event. A negative event is typically defined as one that has caused, or is anticipated to cause, undesirable or harmful consequences. The severity of consequence, according to Anderson (2014), is an important determinant of blame assignment. Studies have shown that more blame will be assigned to an agent if his or her behavior leads to a harmful consequence (Malle et al., 2014; Mazzocco, Alicke, & Davis, 2004). Cushman (2008), in distinguishing blame from similar notions such as wrongness, found that judgments about blame depended substantially on the information about consequences. When a harmful event is absent, malicious intentions are not sufficient to trigger blame and punishment. Tennen and Affleck (1990) proposed that outcome severity, by motivating search for causal explanations, heightened blame judgment. In one study, participants read a fictitious scenario in which drivers with Michelin tires were seriously injured in car accidents caused by tire blowouts (Laufer, Gillespie, McBride, & Gonzalez, 2005). It was found that the perceived outcome severity was positively associated with the attribution of blame to the tire firm. Furthermore, such link could be explained by the defensive attribution hypothesis (Shaver, 1970), which proposes that when a negative event results in a more severe outcome, more blame will be attributed to a potentially responsible party. In line with the defensive attribution hypothesis, Laufer, Silver, and Meyer (2005) found 12 that the association between outcome severity and blame was motivational in nature and related to harm protection. They suggested that individuals blamed a responsible party more for an accident as the consequence of an action became more severe in order to believe such an accident would not happen to them. Robbennolt’s (2000) meta-analysis of 75 studies found that the severity of consequence was a positive predictor of a similar construct, responsibility assignment. In another meta-analysis that examined the relationship between the outcome severity of car accidents and responsibility assignment, it was also found that individuals tended to attribute more responsibility to the perpetrator for a severe than for a mild accident (Burger, 1981). In one study conducted by Gino et al. (2009), participants read a story about a pharmaceutical researcher who purposely changed the data points in order to make the analysis results reach statistical significance before the drug went to market. It was shown that the same behaviors were judged more blameworthy, less ethical when they produced bad outcomes (e.g., kill six patients), than when they produced good outcomes (e.g., no side effects of the drug), even if the outcomes were determined by chance. How the negative value of a certain outcome is determined, it should be noted, is a subjective judgment. The severity of consequence is defined in relative terms – it is the difference between what should have happened, and what has actually happened. In other words, the severity of consequence is a function of the perceived discrepancy between the “ought event” and the actual event. Yet, the “ought” and the “actual” are both individual properties. Causality. For the blame to occur after a negative event is encountered, the perceiver must establish that an agent has caused the event via either action or inaction 13 (Malle et al., 2014; Shaver; 1985). Causality judgment connecting an agent and the negative outcome is a precondition for blame to occur (Alicke, 2000; Malle et al., 2014; Shaver,1985). The effect of causal judgment on blame attribution has been shown in abundant research. In the context of marital difficulties and negative spouse behaviors, causal attributions for relational conflicts were found to predict the assignment of blame (Fincham & Bradbury, 1987). Another study conducted by Pearl and Lebowitz (2014) revealed that the causal attributions of obesity led to various levels of blame assignment and predicted corresponding policy support. They found that personal responsibility attributions of obesity were associated with a higher level of self-blame, and that biological attributions were associated with decreased self-blame and policy support. Using the scenario based on the famous incident in which a woman was scalded by hot coffee from McDonald’s, Kalsher, Phoenix, Wogalter, and Braun (1998) varied the causal information pertaining to McDonald’s and examined the extent to which such variation affected individuals’ attribution of blame. In their study, participants who read the fictitious story in which McDonald’s was framed as responsible— serving coffee that was 40 degrees higher than home-brewed coffee and was involved in 700 individual coffee-related burn lawsuits, assigned more blame to McDonald’s than to the consumer. In comparison, less blame was assigned to McDonald’s when it was described having decreased the serving temperature for its coffee and designed more effective warning label. Causal judgments can be complex (Gerstenberg & Lagnado, 2010). Lagnado and Channon (2008) explored how individuals selected causes from chains of events involving human agents. Specifically, this study asked whether people’s attribution focused on an earlier or later event in the causal chain when other elements were equal. Assessing a 14 scenario where the death of a man was described as being caused by the combination of an early event (i.e., drug overdose purposefully given by his wife) and a later event (i.e., the failure of ambulance to arrive in time), participants assigned greater blame to the later event than to the earlier event (Lagnado & Channon, 2008). Their findings were consistent with Alicke’s (2000) claim that proximity increases perceived causal control and this in turn heightens blame assignment. In another study, Gerstenberg and Lagnado (2010) sought to unpack the complexity of causal attributions by testing three models, with increasing sophistication of judgment made against each other. They found evidence for the most sophisticated model, where individuals applied both directly available causal information (i.e.., that agents contributed to the outcome) and counterfactual thinking (i.e., how the agents would have contributed to the outcome, if the situation were different) to the judgment-making process. Intentionality. Establishing the causal role of an agent is a necessary but not a sufficient condition for blame assignment (Mandel, 2009). For example, if a criminal activity is caused by a perpetrator with biological defects such as intellectual disability, the perpetrator will not be blamed in these situations (Alicke, Buckingham, Zell, & Davis, 2008). Intentionality of the agent is thus integral to blame judgment. Early attribution theory recognized the importance of intentionality in social perception (Heider, 1958). Heider (1958) made the distinctions between personal intentional causality and impersonal unintentional causality. Previous studies have shown that individuals were more likely to blame in situations of intentional norm violations than unintentional violations (Boyle, Dahlstrom, & Kellaris, 1998; Darley & Shultz, 1990; Ohtsubo, 2007; Shultz & Wells, 1985). In the aforementioned study conducted by Lagnado and Channon (2008) that 15 involved a scenario of a man dying from a drug overdose, intentionality, along with location in a causal chain, predicted the assignment of blame. The perpetrator was blamed more when the overdose was described as intentional versus unintentional (i.e., his wife gave him an overdose on purpose versus by mistake) with other elements being equal. One study conducted by Cho and Gower (2006) investigated how news framing of a corporate crisis influenced blame judgment. More organizational blame was elicited when the crisis of a car company was reported as a result of a corporate transgression that involved unsafe working conditions and caused serious employee injuries, compared to a report of an unintentional event (i.e., the car accident story involving the death of a driver after the brakes on the vehicle failed). Preventability. Aside from intentionality, the extent to which the agent is perceived to have obligations or capabilities to prevent the negative event is another important consideration (Malle et al., 2014). Judgment of preventability follows intentionality judgement, which bifurcates information processing into two independent tracks: an independent track and an unintentional track (Malle et al., 2014). When people perceive a behavior as unintentional, they infer the information whether the agent could have prevented the norm-violating event or should have prevented it (Malle et al., 2014). Preventability differs from intentionality in that preventability describes one’s ability to prevent the happening of a negative event whereas intentionality focuses on the mental state of the agent (being purposive or not) when she or he caused the harm. Preventability assumes that the individual at the locus of responsibility is sufficiently rational and autonomous (Moore, 2010). Controllability is another similar concept to preventability. According to Malle et al. (2014), preventability is a more accurate description of the 16 prerequisite for blame because controllability may not be directly linked to responsibility. To illustrate his point, Malle et al. (2014) provided a good example: Gina was very sad when Jesse mentioned the immaculate health of her 80-year-old father, because Gina’s 80year-old father just died. In this case, even though Jesse could control her utterance, she was not responsible for Gina’ sadness. Here it should be noted that the term foreseeability is an akin concept (Malle, Guglielmo, & Monroe, 2014), as both refer to the notion of volitional outcome control (Alicke, 2000). This dissertation will just use the term preventability. Preventability assessment constitutes a type of “counterfactual thinking” (Roese, 1994), consisting of the perception that one could have taken a better course of action to avoid the negative event. One study found that making participants believe that AIDS was a preventable disease increased their tendency to blame the AIDS victims (Berrenberg, Rosnik, & Kravcisin, 1990). In another study, an obese person was less likely to be blamed when observers believed that his or her obesity was a result of uncontrollable medical conditions; but if observers believed that following a diet could improve his or her situation, even in the presence of the medical condition, the obese person would be blamed for not changing his diet (Malle et al., 2014). In another scenario, where a person had been shot dead during a surgical procedure, the negligent physician who failed to save the patient received greater blame than the nonnegligent physician (Alicke, Buckingham, Zell, & Davis, 2008). Among these four antecedents, outcome severity is a characteristic of the event, whereas causality, intentionality, and preventability are largely assessments of “personal control” (Alicke, 2000, p.557). These perceptions, it should be noted, are subjective and 17 individuated assessments. Even though scholars generally agreed in prescribing how to blame people, however, “everyday blamers are capable of violating virtually every rational prescription that moral philosophers, legal scholars, and rational decision theorists hold dear” (Alicke, 2008). Malle et al. (2014) contended that individuals do not need to go through a complicated process of acquiring the relevant information for making blame judgment. There are four types of shortcuts that enable individuals to make quick blame judgment, including hierarchy, event-implied information, multiple-concept information, and preset values (Malle et al., 2014). For instance, in the case that causal agency, intentionality, and justifications are made obvious, the perceiver only needs to do a little computational work between identifying the norm violation and making a moral judgment (Malle et al., 2014). How one interprets the cause and effect, or assigns responsibility or culpability, is contingent upon perceiver characteristics such as their preexisting knowledge about the issue or event, their mood, their relationship to the agent involved, attribution style, and so on (Burger, 1981; Choi, Nisbett, & Norenzayan, 1999; Hammond, Berry, & Rodriguez, 2011). These perceptions are susceptible to biases and can sometimes be intricately intertwined. For example, Knobe (2004) found that individuals tended to consider negative actions more intentional than positive or neutral actions. People often describe blameworthy behaviors as “intentional” to justify the blame they have already assigned (Malle & Nelson, 2003). Alicke (2000) used the concept “culpable control” to emphasize that judgments of “personal control” in a blame situation are in themselves unwittingly influenced by the unconscious, spontaneous evaluations that are based on one’s preexisting attitudes and biases (Alicke, 2000). These unconscious and spontaneous evaluations 18 encourage a blame-validation model of processing in which observers selectively interpret the available evidence and work backward to alter their perceptions of personal control in a way that supports the attribution they want to make. Consistent with Alicke’s (2000) work, Mantler, Schellenberg, and Page (2003) found that judgments for a negative event followed a specific sequence: controllability–responsibility–blame and this attribution sequence was influenced by personal biases and social attitudes. Outcomes of Blame Blame has been linked to a host of outcomes in a variety of contexts. Most of the extant research focuses on coping experiences or outcomes following the negative event. One area of research presents evidence on the effect of self-blame in the recovery from traumatic experiences such as sexual assaults (Branscombe, Wohl, Owen, Allison, & N’gbala, 2003; Malcarne, Compas, Epping-Jordan, & Howell, 1995). Cantón-Cortés, Cantón, and Cortés (2012) found blame on self and that of family for child sexual abuse influenced posttraumatic stress disorder (PTSD) scores, which varied with characteristics of the abuse, such as exhibitionism, touching, oral, and penetration. Similarly, Christy (2016) found a positive relationship between self-blame and PTSD symptoms among sexual assault victims. Along the same line, the study conducted by Branscombe et al. (2003) revealed that increased self-blame decreased the psychological well-being among women who had been raped. One study found that self-blame served as the mediator between children’s experiences of incestuous sexual abuse and their self-mutilating behavior later in life (Shapiro, 1987). Weinberg (1994) examined the relationship between attribution of blame and self-reported recovery from bereavement in the context of 19 mourning after someone’s death. The results of this study revealed that mourners’ selfblame was negatively associated with recovery from bereavement, irrespective of the cause of death. In an effort to compare the difference between self-blame and responsibility in psychological adjustment, Voth and Sirois (2009) examined the relationship between the role of attributions and avoidant coping strategies among people with inflammatory bowel disease. It was shown that attribution of self-blame was directly related to increased avoidant coping, which was in turn associated with poor adjustment, and beliefs about responsibility were associated with decreased use of avoidance coping strategies and subsequently improved psychological adjustment. As a more extreme effect, self-blame could also result in suicide or suicidal attempt (Unnithan, 1994; Yen & Siegler, 2003). When people have no other sources to blame for their misery, the assignment of self-blame leads to suicidal outcomes (Henry & Short, 1954). It was found that a combination of self-blame, which is characterized by feelings of failure, disappointment, guilt, and self-criticism, and negative life events (frustration, stress) led to suicide or suicidal ideation (Unnithan, 1994; Ballard, Patel, Ward, & Lamis, 2015). One study showed that self-blame during childhood was related to overall adulthood symptomatology and the presence of suicide attempts among the individuals who were abused by an immediate family member before 10 years of age (Barker-Collo, 2001). In this line of research, the type of blame plays an important role. For example, Janoff-Bulman (1979) proposed that it was necessary to distinguish two types of self-blame: behavioral self-blame (i.e., attributing blame to one’s behavior) and characterological selfblame (i.e., attributing blame to one’s character). His study found that behavioral selfblame was a predictor of effective coping and avoiding a negative outcome in the future 20 whereas characterological self-blame was associated with negative adjustment among rape victims. Another study that examined the psychological adjustment to the rape, however, challenged Janoff-Bulman’s (1979) findings by showing that behavioral self-blame and characterological self-blame are both associated with poor adjustment (Meyer & Taylor, 1986). Other-blame affects coping differently. Bulman and Wortman (1977) found that among spinal-cord-injured accident victims, self-blame was a predictor of positive coping, such as accepting the reality of their injuries and dealing positively with paralysis, whereas blaming others led to negative coping, such as denying the extent of their injuries despite medical evidence to the contrary or showing no interest in improving their condition. Other studies also demonstrated that the maladaptive effects of other-blaming (Affleck, Allen, McGrade, & McQueeney, 1982; Bulman & Wortman, 1977; Thompson, 1985). Attribution research in general showed that causal attributions to other people were negatively associated with coping (Bulman & Wortman, 1977; Madden & Janoff-Bulman, 1981). In their study that explored the association between causal attributions to others and a mothers’ adjustment to the birth of a child with Down syndrome, Hall, French, and Marteau (2003) found that other-blame was associated with anger, anxiety, depression, parenting stress, and more negative attitudes toward their children. Thompson (1985) conducted a longitudinal study on people whose apartment had been partially or totally destroyed by a major fire and showed that people who blamed others coped poorly and experienced more negative emotions. It should be noted, however, that other-blame can also be an effective means to rationalize one’s own misconduct and thus improve “coping” in that sense. For example, 21 some rapists subscribe to the beliefs of holding rape victims responsible for bringing suffering on themselves by exhibiting sexually provocative appearances (Bandura, 1999). By viewing oneself as faultless and the misconduct as compelled by forcible provocation, blaming one’s antagonists or environment circumstance can serve self-exonerative purposes (Bandura, 1999). In studying psychiatric inpatients who had committed serious criminal acts, it was found that external attribution (blaming the crime on society) was positively associated with psychoticism scores, mental element attribution (blaming the crime on mental illness) was found to be positively associated with socialization and social desirability, and guilt feeling (regret at the offence) was positively associated with neuroticism and depression (Gudjonsson, 1984). Other than the various coping outcomes, as reviewed above, there is scanty research on other behavioral outcomes following blame. Some research has examined aggression behaviors as a result of blame. A study conducted by Miller, Markman, and Handley (2007) showed that self-blame predicted the subsequent victimization among female college students who had been sexually assaulted when they were adolescents. In other circumstances, it was found that both self-blame and partner blame were not associated with the frequency nor severity of physical aggression among female victims of domestic violence (Cascardi & O’Leary, 1992). Studies have also found that blame was positively associated with revenge behavior and was negatively associated with reconciliation and forgiveness in the workplace (Aquino, Tripp, & Bies, 2001; Bradfield & Aquino, 1999). It was found that the relationship between blame and revenge was moderated by situation variables in the organization settings, such as power difference between victim and offender (Aquino et al., 2001). 22 Despite a growing literature on the effect of other-blame on different outcomes, Hall et al. (2002) suggested that the process by which blaming others might influence outcomes has yet to be determined. How blame may lead to behavioral outcomes in communication processes has not been studied. I will take up this question in Chapter 3 and discuss what social responses may follow blame in the context of childhood obesity. Explicating the Construct of Blame Setting the Definitional Boundary Existing conceptualizations and operationalizations of blame can be mapped along a few dimensions. First, blame can be approached as either a situational or a trait construct. As a situational variable, blame is a psychological reaction to a specific negative event. Most research on blame falls under this category. Other research has examined the tendency toward blaming as an enduring individual characteristic. A series of studies drew their ideas from Ellis's (1962) study that conceptualized blame proneness as a type of irrational beliefs. Blame proneness, as an individual characteristic, was linked with emotional distress (Ellis, 1962), anxieties (Deffenbacher et al., 2011), or competitiveness and impatience (Smith & Brehm, 1981). Another related construct is blame accessibility, measured as the reaction time for the task of classifying words as blameworthy in Meier and Robinson’s (2004) study. They found that among people who were low in agreeableness, blame accessibility was positively related to hostile feelings. Other studies developed the scales of tendency of blaming a specific population. For example, the AIDS victim-blaming scale, developed by Mulford, Lee, and Sapp (1996), included the victimblaming subscale and the society-blaming subscales, which captured the individual 23 tendencies to blame AIDS victims and the society in which victims live. Their study revealed that both subscales were positively related to the tendency to discriminate against AIDS patients and negatively associated with social responsibility-taking. As a situational construct, blame is a target-specific reaction. As mentioned earlier, extant research has differentiated self-blame and other-blame, also referred to as internal blame and external blame (Burgess, Wojslawowicz, Rubin, Rose-Krasnor, & BoothLaForce, 2006). Here “other” can encompass a host of entities, including a specific person involved in a scenario (e.g., a perpetrator; Berns, 2001), a group of people (e.g., standersby in a crime scene; Gilbert, 2006), general others (e.g., Muslims for terrorism; Strelan & Lawani, 2010), impersonal entities such as government, food industry or policies (Barry, Brescoll, Brownell, & Schlesinger, 2009), or other situational factors such as the relationship with the perpetrator and the work environment (Aquino, Tripp, & Bies, 2001). Who or what to blame is subject to event characteristics as well as individuals’ interpretative schemata, biases, or idiosyncrasies. Jones and Nisbett (1971) observed the differences between actor and observers’ perceptions of causes for blameworthy action. They pointed out that the actor tended to attribute the causes of their behaviors to situational cues while observers tended to consider the actor’s behavior as a manifestation of the actor’s stable dispositions. This dissertation focuses on blame as a situational construct, and a response directed at others. Whereas self-blame and its implications for personal well-being and interpersonal relationships are important to understand, self-blame is typically a private judgment. Other-blame, on the other hand, can be a powerful social judgment, one that could be malleable to media influences, and with implications for communication 24 processes. Media discussions such as “who is to blame for the war” were in themselves a “blame war.” As an example, for the Syria crisis, the Western and Eastern media assigned blames in disparate ways (Shahtahmasebi, 2016). Another illustrative example was the blame game between scientific research and parents for the decreased rate of childhood vaccinations. Wakefield’s fraudulent study published on The Lancet (which claimed a link between MMR (Measles, Mumps, and Rubella) vaccines and an increased risk of autism (Wakefield et al., 1998), the media that propagated the false findings, and parents have all been made targets of blame (Godlee, Smith, & Marcovitch, 2011; Smith et al., 2011). Given its potentially profound impact on judgments about social issues and the following communicative actions, situational other-blame is taken as the focus of this dissertation. The Nature of Blame What is blame, however, remains a question despite the existing body of research on blame. Surveying the literature, I summarize the existing conceptualizations of blame into three categories. One line of literature conceptualizes blame as a cognitive judgment (Malle et al., 2014; Mikula, 2003; Shaver, 1995), equating blaming with responsibility attribution. Other researchers have discussed blame as a certain type of negative emotion, particularly anger (Averill, 1983; Sheikh & McNamara, 2014). A third way of conceptualization treats blame as a blend of cognition and affect (Weiner, 1995). Table 1 lists existing definitions of blame from prior research. Blame as cognition. Much of the prior research on blame highlights it as a cognitive judgment that centers on ascribing responsibility for a negative outcome. Shaver (1985), tracing back to Heider’s early attribution theory, construed blame as a type of 25 attribution. As attributional judgment, blame is a cognitive task in which individuals make causal explanations for behaviors and events (Latridis & Fousiani, 2009). Malle, Guglielmo, and Monroe (2012) more broadly conceived blame as a cognitive system for social perceivers — “a system of concepts and processes… in inferring mental states from behavior” (p.313). As one can see in Table 1, many definitions highlight the cognitive nature of blame. For example, Lazarus (1991) defined blame as an appraisal that grows out of accountability and imputed control in the context of frustration and threat. Knaus (2006) claimed that “blame is a process of objectively finding fault before assigning consequences” (p.178). Malle at al. (2014) discussed that blame is “cognitive and social,” and “fundamentally relies on social cognition” (p.176). Scrutinizing these definitions also showed that although some definitions basically equate blame with responsibility, others suggest implicit differentiations. Compared to responsibility judgment, blame has an inherent negative valence. For example, blame is a “negative attitude” (Bayles, 1982, p.7), an expression of “disapproval of … bad behavior or character” (Sher, 2005, p.72), or “a class of responses to morally faulty actions” (Scanlon, 2013, p.1). Responsibility attribution, on the other hand, could encompass negative, neutral, or positive judgments. Responsibility assignment for a positive action can lead to praise (Pizarro, Uhlmann, & Bloom, 2003). A person could be regarded as responsible for certain outcomes but is assigned little blame (Weiner, 1995). A factor analysis made by Harvey and Rule (1978) revealed that causal responsibility was different from blame by showing that the clusters associated with responsibility were distinct from those associated with blame. Blame, therefore, should be differentiated from responsibility or attribution 26 judgment. When blame was treated as a cognitive response, researchers either considered affect as the antecedent or the outcome of blame (Camodeca, Goossens, Schuengel, & Terwogt, 2003; Frijda, 1993; Lussier, Sabourin, & Wright, 1993; Weber, 2004), or treated affect in a parallel fashion with the cognitive aspect of blame in response to a normviolating event (Olthof, Ferguson, & Luiten, 1989). Additionally, one study revealed that anger and blame attribution had a bidirectional effect on each other (Quigley & Tedeschi, 1996). Malle et al. (2014) suggested that the negative affect co-occurs with the detection of a norm violation is neither blame nor it could generate blame. Blame as affect. Other research on blame has focused on the negative emotions as a defining characteristic of blame. For example, Averill (1983) claimed that “Anger, more than anything else, is an attribution of blame” (p.1150). According to him, anger follows from a value judgment of perceived wrongdoings. Besharat, Eisler, and Dare (2001) defined self-blame as “an affect (rather than judgment or belief) which is negatively toned” (p.210). In Smith's (2013) definition, blame is “a way of responding emotionally to the perceived disregard or disrespect” (p.31). Fingarette (1963) stated that “Blame…involves an affective tone reminiscent of anger” (p.118). Existing research suggested that how blame manifested as an affective response depended on the context (Janoff-Bulman, 1979; Sheikh & McNamara, 2014). Inherent to self-blame are feelings of shame and guilt (Dickerson, Gruenewald, & Kemeny, 2004; Gudjonsson & Singh, 1989; Leskela, Dieperink, & Thuras, 2002) , whereas anger is often the defining emotion of other-blame (Berenbaum, Fujita, & Pfennig, 1995; Sheikh & McNamara, 2014). More specifically about self-blame, the State Shame and Guilt Scale 27 (SSGS; Marschall, Sanftner, & Tangney, 1994) was developed to assess characterological and behavioral self-blame, where characterological self-blame was measured by assessing state shame and behavioral self-blame was measured by state guilt. A set of emotions such as anger, disgust, and contempt have been associated with conceptualizations of blame (Sheikh & McNamara, 2014). According to Wydo (2003), depression is blame turned inward and anger is blame turned outward. Other prior research also considers blame as equivalent to anger (Averill, 1983). From a constructivist view of emotion, anger involves both subjective experience and social evaluation (Averill, 1983). Averill (1983) argued that anger differs from other similar negative emotions primarily because of the attribution of blame implicit in anger and in the commitment to action. Other scholars also indicated that anger involves an attribution of blame (Ortony et al., 1988; Rule & Nesdale, 1976). For instance, Ortony, Clore, and Collins (1990) suggested that the cognitive structure of anger is “disapproving of someone else’s blameworthy action and being displeased about the related undesirable event” (p.148). Blame as both cognition and affect. Implicitly or explicitly, some researchers pointed out that a combination of cognitive evaluation and affective reaction characterizes blame. Sheikh and McNamara (2014), for example, contended that if an agent regarded herself as responsible for some negative events but showed no emotional response, she might not sincerely blame herself for her action in the given scenario. In Friedlander, Heatherington, and Marrs (2000) definition, they referred to blame as both “a powerful attribution” and emotional responses including “shame, guilt, anger, and defensiveness” (p. 134). Some empirical measures of blame incorporated an affective component in addition 28 to cognitive component (Gudjonsson & Singh, 1989; Mantler et al., 2003). The Gudjonsson Blame Attribution Inventory-Revised (GBAI-R) scale (Gudjonsson & Singh, 1989), for example, consists of both cognitive and affective aspects. The cognitive subscale includes internal attribution (e.g., assigning responsibility for the crime on mental illness or poor self-control) and external attribution (e.g., blaming the crime on social circumstances, victims, or society), whereas the affective subscale assesses the feelings of regret and remorse concerning the offence. Another study conducted by Mantler et al. (2003) also included an affective component, guilt, in addition to cognitive items—in measuring blame for a person with AIDS or lung cancer (e.g., “William should not feel guilty for being ill”). Most explicitly, Weiner (1995) defined blame as “a blend construct” of ascribed responsibility and anger. Weiner (1995) noted that “blame appears to be a cognition similar to responsibility, as well as an affect akin to anger” (p.42). In other words, blame as a unique construct is not reducible to mere responsibility assignment, or anger, but rather is characterized by the copresence of both. The varied conceptualizations of blame, hence variations in operational measures, can be a hinderance to theory development. Shaver and Drown (1986) pointed out that researchers tended to treat causality, responsibility, and blameworthiness as interchangeable concepts. A clear conceptualization and empirical validation of the construct of blame is in order. A Tentative Definition of Blame A review of previous research suggested that blame as a unique construct should be differentiated from responsibility and anger on both theoretical and empirical grounds 29 (Mantler et al., 2003; Weiner, 1995). According to Sheikh and McNamara (2014), the unpleasant feelings that people have when they blame others, such as regret and shame, are not side effects of blame but a necessary component of blame itself (similarly, pride is central to praising oneself for a success). Most existing literature on blame shares the common ground that blame involves the attribution of responsibility (Kanus, 2006; Malle et al., 2014; Shaver, 1985). Based on the previous literature, it is postulated that to “blame” is to assign a responsibility judgment that is accompanied by negative affect such as anger (other-blame) or guilt (self-blame). Synthesizing the previous theoretical and empirical work on blame, this dissertation tentatively defines blame as a psychological response, consisting in attributional judgment and negative affect, directed at an agent perceived to be causally responsible for a normviolating event. “Norm” here broadly refers to a perceiver’s expectation for the involved social actor’s behaviors, which could be grounded in moral standards, religious beliefs, or other social or personal values. In the context of this dissertation, I focus on other-blame with potential implications for communicative behaviors. This proposed definition is in alignment with Weiner’s (1995) conceptualization of blame as a blend construct with both cognitive and affective aspects. In terms of its role in message processing, blame is functionally akin to other psychological constructs such as psychological reactance. Elicited by a norm-violating event, blame (whether self-blame or other-blame) is a negative experience that entails a motivation to restore the norm retrospectively or prospectively. Shame or guilt associated with self-blame, for example, could be related with wishful, hypothetical thinking regarding what one could have done, and/or lead to future actions to re-establish a standard 30 for behavior. In the framework of media effects, blame could be conceived as a mediating psychological state that can be elicited by a message and shape subsequent social attitudes and communicative behaviors. With this conceptualization of blame, the operational framework of blame in this project has two key parts. First, a measurement model of blame comprising both cognitive and affective components will be tested against alternative models. In the context of childhood obesity, the cognitive component is the attribution of responsibility for the described problem, and the affective component is moral indignation, which will be elaborated below. Second, blame will be examined as a mediating construct, a response to a media message with consequences on attitudinal and behavioral outcomes. Following the analytical process of examining the construct of psychological reactance outlined in Dillard and Shen's (2005) study, this study approaches the operational explication of blame by comparing four models in which blame is a mediating construct. The four models differ in terms of the empirical specification of the structure of blame. The first two models pose blame as consisting of either a cognitive or an emotional component. The third and fourth models involve both cognitive and affective components. The third model conceives of cognitive and affective components as separable, parallel paths, with distinctive effects on outcomes; whereas the fourth model regards blame as a latent construct that consists of cognition and affect in an intertwined fashion (i.e., as “a blend construct”). In other words, in the fourth model, blame is construed as an alloy of cognition and affect, the effects of which on attitudinal or behavioral outcomes cannot be separated. Conceptual depictions of the four models are included in Figure 1 to Figure 4, corresponding to Models 1 to 4. 31 Moral indignation as the affective aspect of blame. In this dissertation, the affective component of blame is represented by moral indignation. Such “moral emotions,” also referred as “social emotions” (Weiner, 2005), pertain to “the interests or welfare either of society as a whole or at least of persons other than the judge or agent” (p.276). Moral emotions are often triggered by third-party norm violations that may not directly affect the observer himself or herself (Rozin, Lowery, Imada, & Haidt, 1999), and often lead to prosocial actions for the benefit of others or maintaining social order (Haidt, 2003). More specifically, moral indignation, as a type of punitive emotion, constitutes “a discomfort or disquiet experienced by those who adhere to social norms concerning others’ deviance from those standard” (Krinsky, 2016, p.18). Moral indignation was proposed as a social affect that includes a class of feelings that human beings develop as they live out their lives collectively (Krinsky, 2016). Moral indignation is constitutively linked to the expectations that people hold others to (Wallace, 1994), embodies an evaluative judgment principle (Solomon, 2008), and is likely to occur when one’s normative principles are violated (Hwang, Pan, & Sun, 2008). Moral indignation can motivate certain action tendencies, often with the goal of moral sanction (Krinsky, 2016; Wallace, 1994). Moral indignation can also lead to public denunciation and has the social function to bind people collectivity (Krinsky, 2016). The public discourse on the negative consequences of childhood obesity, as the results of the “vice” of parents and “dangerous” food-laden environment, may evoke moral indignation due to our cultural belief that children are vulnerable and need to be protected (Furedi, 2013). The misfortune of children, when attributed to irresponsible behaviors of their parents, is likely to activate one’s moral alarm system and evoke moral indignation. 32 Research Questions Since the conceptualization of blame as an amalgam of cognition and affect has not been tested before, it is unclear whether such conceptualization will be supported by data. Lacking guidance from previous research, the first set of research questions were proposed to explore the nature of blame. RQ1a: Should blame be conceptualized and operationalized as a cognition, affect, or both? RQ1b: If blame has both cognitive and affective component, how are the two aspects combined? Chapter Summary This chapter provided a review of existing research on blame, with specific focus on (1) the antecedents and outcomes of blame that have been studied, and (2) various conceptualizations of blame. Based on the review, I proposed to study situational, otheroriented blame in this research project, and advanced my explication of blame. Research questions related to the nature of blame were advanced. As the outcomes of blame reviewed in this chapter were mostly drawn from social psychological research and thus not directly pertinent to communication research, the next chapter will introduce and describe the key outcome variables used in this dissertation. Then hypotheses and research questions will be outlined. 33 Table 1 Definitions of Blame Author Caplan & Nelson, (1973) (Bayles, 1982) Definition Cognition/affe ct/behavior Individual blame is defined as the tendency to cognition hold an individual responsible for his or her problems, rather than the system of which the individual is a part. Blame is a negative attitude towards a person cognition for a character trait; it is appropriate if the person is blameworthy. Shaver (1985) The assignment of blame to a person who has cognition helped to bring about a misfortune or morally reprehensible event is a social judgment that must incorporate decisions about causality and responsibility. It presupposes a particular set of actions (those that produce negative consequences), a specific level of personal causality (single causation at the intentional level), a special combination of the dimensions (responsibility, causation, knowledge of the consequences, intentionality, voluntary choice, and the capacity to distinguish right from wrong), and the failure to have an adequate justification or excuse. Zuk & Zuk(1989) Blame is a powerful attribution that, when expressed, provokes shame, guilt, anger, and defensiveness, obstructs intimacy, and hinders problem-solving. cognition Darley & Shultz (1990) In cases where it is judged that the harm was done intentionally, blame is a joint function of moral responsibility, the presence of net harm, and justification for the harm. cognition Shultz(1990) Blame refers to a decision that a person is at fault, given that he has caused and is responsible for the harm. cognition Lazarus (1991) Blame is an appraisal based on accountability and imputed control. cognition 34 Table 1 Continued Author Definition Skorupski (1993) Blame is a response whose justifying base requires that the object be something that has been done or brought about but morally ought not to have been done or brought about. Weiner (1995) Blame is a “blend-construct” of ascribed responsibility and anger. Feiring, Taska, & Self-blame is conceptualized as a cognitive Lewis (1996) attribution by a survivor, in which she places blame for the abuse/assault on herself; it is understood as internal, stable, and global. Alicke (2000) Blame is an aspect of everyday conduct evaluation that identifies behavior as morally wrong or socially opprobrious. Besharat, Eisler, Self-blame can be defined as an affect (rather & Dare (2001) than judgment or belief) which is negatively toned (like guilt) but which is more associated with a sense of responsibility and causality for event, illness, or interpersonal interaction than is guilt. Obermann Attribution of blame refers to how individuals (2011) come to view their transgressions as reactions to provocations or as defense against the victim's initial norm violation. Runciman, Blame is a deeply rooted human response to Roughead, harm, particularly in health care, where the Semple, & patient expects that he or she will be helped. Adams (2003) Vander Ven & Mother-blame is defined as the propensity to Vander Ven explain negative outcomes for children by (2003) focusing on the failures of mothers. Chockler & We present a definition of blame that considers Halpern (2004) whether agent a performing action b is to blame for an outcome ϕ. The definition is relative to an epistemic state for a, which is taken, roughly speaking, to be a set of situations before action b is performed, together with a probability on them. The degree of blame is then essentially the expected degree of responsibility of action b for ϕ (except that we ignore situations where ϕ was already true or b was already performed). Cognition/Affect /Behavior cognition cognition&affect cognition cognition affect cognition not specified cognition cognition 35 Table 1 Continued Author Definition Devooght (2004) Moral blame refers to the prevailing moral system and is aiming at the goodness of life of individuals. Blameworthiness refers to the extent to which a person deserves to be punished for committing a wrongful act. Blame is a judgment that a harm doer deserves punishment. Roberts & Stalans (2004) Ohbuchi, Hanaizumi, & Hock (2004) Knaus (2006) Sher (2005) Garnefski & Kraaij (2007) Curry & CorralCamacho (2008) Renaud (2008) Vogel et al. (2008) Campbell, Dworkin, & Cabral (2009) Blame is a process of objectively finding fault before assigning consequences. To blame someone is simply to express disapproval of his bad behavior or character in a way that is calculated to mitigate or improve it. Self-blame refers to thoughts of putting the blame for what you have experienced on yourself. Other-blame refers to thoughts of putting the blame for what you have experienced on the environment or another person. Blameworthiness refers to the perceived culpability of offenders and how deserving they are of particular punishments. Self-blame refers to the victims’ belief that they are personally responsible, either by act or omission, for the traumatizing event. Blame is defined as blaming, accusing, or criticizing the partner, using critical sarcasm; or making character assassinations. Self-blame is conceptualized as meta-construct that stems from all levels of this ecological model, including individual-level factors (e.g., sociodemographic, biological/genetic factors), assault characteristics (e.g., victim-offender relationship, injury, alcohol use), microsystem factors (e.g., informal support from family and friends), meso/exosystem factors (e.g., contact with the legal, medical, and mental health systems, and rape crisis centers), Cognition/Affect /Behavior cognition cognition cognition cognition cognition cognition cognition cognition behavior not specified 36 Table 1 Continued Author Mikolajczak (2009) Moses (2011) Dunning (2012) Hostler (2012) Malle, Guglielmo, & Monroe (2012) Warner, VanDeursen, & Pope (2012) Definition macrosystem factors. (e.g., societal rape myth acceptance), and chronosystem factors (e.g., sexual revictimization and history of other victimizations). Others-blame refers to rejecting on others the responsibility for the occurrence of the problem and/or our incapacity to solve it. Parental self-blame is defined as parents holding themselves responsible for causing, contributing to, or exacerbating their children’s mental disorders. Blame is a form of social judgment that is “a symbolic act, one that holds tacit implications for the self. As such, people treat evaluation of others as evaluations of themselves, even when the self is not being explicitly judged, and thus manage their judgments of others to affirm and retain flattering self-images.” Victim blaming is a devaluing act that occurs when the victims of a crime or an accident is held responsible, in whole or in part, for the crimes that have been committed against them. Blame is grounded in the capacity to have a “theory of mind”—a system of concepts and processes that aid a human social perceiver in inferring mental states from behavior. Behavior blaming refers to attributing responsibility for victimization to the victim’s behavior (e.g., believing that a victim of domestic abuse provoked her attacker by arguing with him), character blaming refers to denigrating the character of a victim (e.g., believing that an abuse victim deserved her fate because she is a bad person). Cognition/Affect /Behavior cognition cognition cognition cognition cognition cognition 37 Table 1 Continued Author Definition Tomai & Forbus (2007) Blame is a moral condemnation that follows from responsibility for a morally reprehensible outcome but may be mitigated by justification or excuse. Bell (2013) Blame refers to expressions of a judgment of blameworthiness and some hostile attitude” such as resentment, indignation, or contempt. cognition Pickard (2013) When we say another is ‘to blame’ we may mean one of three things: (i) is a judgement about another (They are blameworthy). (2) is about us and what we should do. (we should blame them). (3) blame is desirable relative to a given end (we actually do blame them). Blame refers to a class of responses to morally faulty actions. cognition Smith (2013) Blame is a way of responding emotionally to the perceived disregard or disrespect manifested in someone's behavior toward oneself or others. emotion Malle et al. (2014) Blame is a unique type of moral judgment and has four properties: It is both cognitive and social; it regulates social behavior; it fundamentally relies on social cognition; and, as a social act, it requires warrant. cognition Bauman & Mullen (2014) Private blame is a means to mitigate personal risk of exploitation from cheating, social and other forms of opportunism. cognition Scanlon (2013) Cognition/Affe ct/Behavior cognition not specified 38 Figure 1. The Single Process Affective Model. 39 Figure 2. The Single Process Cognitive Model. 40 Figure 3. The Dual Process Cognitive-Affective Model. 41 Figure 4. The Intertwined Process Cognitive-Affective Model. CHAPTER 3 SOCIAL RESPONSES FOLLOWING BLAME: HYPOTHESES AND RESEARCH QUESTIONS As contended by Nelkin and Gilman (1988), “locating blame is in effect a quest for order and certainty in an anxious and disruptive situation” (p. 363). Since individuals experience blame when there is a norm-violating event, they are inherently positioned to restore the social norm that has been violated by the wrongdoer. In this dissertation, four social responses that could be elicited by blame in the context of childhood obesity were studied, including policy support, information sharing behavior, civic participation, and punishment, all of which share the goal of keeping children protected and healthy. As noted earlier, there is scant research on the outcomes of other-blame, and even less so in communication research. This chapter will first discuss in detail these four social responses that might follow blame. Then, it will briefly highlight some individual trait variables that may serve as potential moderators. After that, I will formulate hypotheses and research questions to be examined in this dissertation project. 43 Social Responses Following Blame Policy Support Broadly speaking, policy support is a type of social responsibility taking behavior. It is a way for the public to engage in public affairs and achieve desired outcomes for social issues. The power of public policy on improving population health has been well documented in previous research (Eyler & Brownson, 2016). Developing and implementing health-related policies is imperative to address various complex health challenges (Eyler & Brownson, 2016). Regarding the issue of obesity, there was little support among Americans for public policies targeting obesity, as most still viewed obesity as primarily a result of individual behaviors (Oliver & Lee, 2005). Given the urgency of the obesity epidemic, however, public policy change is a most effective avenue for addressing the problem (Kersh et al., 2011). Health researchers and practitioners have called for policy actions such as controlling the conditions of sale (e.g., limiting what schools can offer), restricting advertising of high-fat, low nutrient foods that target young children or using other alternatives to increase awareness of what they are eating, subsidizing healthier alternatives that have much higher per-calorie costs, and restricting or banning certain ingredients (e.g., trans fats; Frieden, Dietz, & Collins, 2010; Kersh, Stroup, & Taylor, 2011). Although there is no direct evidence on the effect of blame on policy support, separate sets of studies showed that causal attribution and anger can be predictors of policy support. Causal narratives about why people are overweight and obese were found to generate support for policy change (Barry, Brescoll, Brownell, & Schlesinger, 2009; Niederdeppe et al., 2013; Oliver & Lee, 2005). Nabi (2003) showed that participants 44 primed with anger showed stronger support for both protective and retributive policy initiatives. Goodall and Reed's (2013) study also found that elicited anger was positively associated with support for individual-oriented and societal-oriented alcohol-control policies. Another study conducted by Skitka, Bauman, Aramovich, and Morgan (2006) showed that anger predicted policy responses to terrorism, such that greater anger led to stronger support for expanding the war beyond Afghanistan. Information Sharing Behavior Information sharing behavior is defined in this study as the purposive transmission of health information to others. It is a type of supportive communication that contributes to building and consolidating social relationships. Again, the relationship between blame and information sharing behavior has not been studied yet. Previous research has suggested that negative emotions such as anger could propel information sharing behavior. First, in general, people were more willing to share or communicate information that aroused emotion (Berger, 2011; Hasell & Weeks, 2016; Ibrahim, Ye, & Hoffner, 2008). Second, physiological arousal, an integral part of emotional experiences, was shown to stimulate social transmission of content, whether it is related or unrelated to a source of emotions (Berger, 2011). Berger and Milkman (2012) argued that negative emotions with heightened arousal should have a positive effect on social transmission and increased the likelihood of sharing information. Their experiment revealed that induced anger was positively related to the likelihood of sharing a story about customer service experiences with United Airlines. A study of partisan news on social media showed that the use of proattitudinal partisan news online during the 2012 U.S. presidential election was associated with increased anger, 45 which in turn facilitated information sharing about the election on the social media (Hasell & Weeks, 2016). Civic Participation Previous studies found that blame was related to participatory behaviors (Kingree & Thompson, 2000; Levin, Sinclair, & Alvarez, 2016). To elucidate the underlying association between blame and civic participation, Levin et al. (2016) explained that responsibility attributions politicized political evaluations and prepared voters to participate in rational retrospective voting. Other research has shown that civic participation was contingent upon blame (Gurin, Miller, & Gurin, 1980; Levin et al., 2016). Gurin et al. (1980) suggested that civic participation should be related to group identification only after it had been politicized by attribution of blame for the group’s position. There is scanty research that examined the direct link between blame and civic participation. One study revealed that blame attribution toward political party drove political participation in the 2010 general election among Californians (Levin et al., 2016). Even though not much research has focused on the direct link between blame and civic participation, many studies assessed the relationship between the cognitive component of blame, which is often manifested as attribution of responsibility, and civic participation. Crystal and DeBell (2002) indicated that attributions of social responsibility are an area of cognition importance to understanding individual differences in levels of civic participation. It was found that that contributions to public good were strongly related to feelings of responsibility (Fleishman & Payne, 1980). Crystal and Debell (2002) studied sources of civic orientation among American youth, and it was found that individualistic 46 attribution of social responsibility predicted running for student government office, and collective attribution of social responsibility predicted endorsement of public and private citizenship activities. Punishment Punishment has been studied as an outcome of blame in extant literature. Punishment refers to “a societal reaction to an act that is considered harmful to the entire society” (Schichor, 1995, p.9), with the objective to lower the likelihood of an unwanted response recurring in the eliciting stimulus setting (Jones-Smith, 2014). In Weiner's (2006) view, punishment is a “debt” that the transgressor owed to the victim. In this research, punishment is defined as a type of negative social sanction that regulates people’s social conduct in conformity with social norms and prevents the reoccurrence of a norm-violating event in the future. Conceptually, punishment is inherently linked to the notion of blame. For instance, legal punishment was conceptualized as a response to a judgment of the perpetrator’s blameworthiness (Shoemaker, 2012). “Punishment is, as it were, blame carried out or acted upon” (Warmke, 2011, p.651). Zaibert (2009) stated that to punish “is to do something about what we find blameworthy” (p.385). Fincham and Jaspars (1980) proposed the “entailment model,” which proposed that causal judgment leads to blame, and in turn, leads to punishment. Punishment has its practical functions. Punishment has two kinds of deterrent effects. The first type of deterrence is directed at a specific person who has committed a crime to keep him or her from future lawbreaking. The second type of deterrence is referred 47 as “general deterrence,” which is directed toward the general public and potential offenders who have not yet violated the law (Shichor, 2005). Weiner (2006) identified two goals of punishment: retribution and utilitarianism. Retribution considers a past misdeed, with an aim of retaliation (Weiner, 2006). Utilitarianism focuses on the future, with a goal of lowering the likelihood of the misdeed by the perpetrator in the society (Murphy & Coleman, 2013). There has been plenty of evidence suggesting that punishment follows blame. Shultz, Wright, and Schleifer (1986) found that children assigned more punishment for intentional harm. Critchlow's (1985) study found that stronger blame judgment led to harsher punishment. Taylor and Kleinke (1992) investigated the relationship between blame and punishment given to drunk drivers involved in car accidents. The results showed that attribution of blame was positively associated with punishment for the drunk driver, including a fine and prison sentence. Punishment is contingent upon judgment of the causal role, or the intentionality, of an actor in a negative event, as well as the severity of consequences. If an actor’s behavior was presented as a necessary condition for harmful consequence, he was more likely to be considered as deserving of punishment (Shultz, Schleifer, & Altman, 1981). A series of studies showed that people assigned more punishment for rapists when the alleged rapist was intentional (Hogue & Peebles, 1997; Kleinke, Wallis, & Stalder, 1997). One study conducted by Fragale, Rosen, Xu, and Merideth (2009) reported that observers’ judgment of intentionality of the wrongdoers was positively linked to the opinions about the wrongdoer’s deserved punishment. In their study that examined punishment goals to O. J. Simpson, perceived controllability was found to influence inferences about defendant’s 48 responsibility, which in turn, instigated a set of responses leading to punishment (Graham et al., 1997). Punishment is also affected by the severity of consequence (Taylor & Kleinke, 1992). The severity of consequence of a car accident involving drunk driving was found to directly influence punishment for the drunk driver (Taylor & Kleinke, 1992). Tetlock, Self, and Singh (2010) found that punishment was positively correlated with the severity of consequences resulting from norm transgression. Negative affect, particularly anger, was found to be associated with punishment (Feigenson & Park, 2006; Harth, Leach, & Kessler, 2013; Lerner, Goldberg, & Tetlock, 1998). Bennett (2009) stated that punishment is the expression of moral condemnation. Weiner (2005) suggested that anger along with responsibility inferences (either directly or indirectly) could lead to punishment behaviors. Feigenson and Park (2006) found that anger could have direct effects on punitive actions. Another study found that anger was positively associated with punishment tendencies, whereas guilt and pride were not (Harth et al., 2013). In a study about a scenario where a paralyzed couple died in an accident because a taxi drive refused to take them, anger was a significant predictor of financial penalties against the taxi driver (Williams, Lees-Haley, & Price, 1996). Lerner et al. (1998) found that anger-primed participants made more punitive attributions than did neutral-emotionprimed participants. In one study, participants were exposed to 30 hypothetical norm violation scenarios that varied in outcome severity and moral domain (e.g., loyalty to one’s in-group, purity (Konishi et al., 2017). It was found that moral emotions, including moral outrage and moral disgust, promoted some forms of punishment (e.g., imposing a hypothetical fine). 49 Individual Traits as Potential Moderators Three individual trait variables, including need for affect, need for cognition, and moral identity, were included in this study to examine their moderating roles in the relationships between the antecedents of blame and blame and the relationships between blame and subsequent social responses. Need for affect (NFA) refers to “the general motivation of people to approach or avoid situations and activities that are emotion inducing for themselves and others” (Maio & Esses, 2001, p.2). Need for affect is described as the “affective counterpart to the need for cognition” (Appel & Richter, 2010, p.107). Individuals differ in the extent to which they interpret affective experiences, which vary in their intensity, quality, stability, specificity, clarity, and valence (Maio & Esses, 2001). NFA includes two dimensions: emotion approach and emotion avoidance. Past research found that emotion approach and emotion avoidance had a distinct effect on outcomes that were relevant to the experiences and information people seek (Bartsch, Appel, & Storch, 2010; Haddock, Maio, Arnold, & Huskinson, 2008). NFA has been linked to various outcomes. For example, NFA was found to predict eudaimonia and hedonism among individuals who consume media entertainment (Oliver & Raney, 2011). Some research investigated the moderating effect of NFA on media effects. One study conducted by Rosenbaum and Johnson (2016) revealed that NFA and need for cognition moderated the effect of spoilers on enjoyment among media users, with people high in NFA enjoyed unspoiled stories more than spoiled stories and those low in need for cognition preferred spoiled stories. Previous research has shown that NFA moderated the persuasive effect of narratives. One study revealed that NFA (approach tendency) moderated the persuasive effects of a fictional narrative compared to a belief- 50 irrelevant story and the persuasive effects of a story with high emotional content compared to a story with low emotional content (Appel & Richter, 2010). Similar to narrative persuasion, a persuasive message that contains the antecedents of blame is likely to evoke negative affect, which is proposed as the affective component of blame. NFA is thus considered as a relevant moderator in the relationships between the antecedents of blame and blame. Maio and Esses (2011) proposed that avoidance tendency might be associated with negative affect and approach tendency might be associated with positive affect. It is possible that individuals with a higher level of emotion approach tendency are more likely to perform positive behaviors, and those with emotion avoidance tendency are more likely to take actions that could result in negative outcomes, such as punishment. Building on this line of research, NFA is proposed as a potential moderator between blame-behavior relations. A second moderator included in this study was need for cognition (NFC). Need for cognition refers to “a need to structure relevant situations in meaningful, integrated ways” (Cohen, 1955, p.291). It is a “need to understand and make reasonable the experiential world” (Cohen, 1955, p.291). Individuals with high NFC are more likely to engage in elaborative processing and thus more likely to use the framing scenarios as the basis for their judgment (Boyle, Dahlstrom, & Kellaris, 1998). Jones (1991) suggested that individuals with high NFC tend to rely on issue-relevant information when making their ethical decisions. Therefore, individuals who are high in NFC are more prone to contextual bias (Boyle, Dahlstrom, & Kellaris, 1998). On the contrary, some people are “cognitive misers” (Waller, 2005), who are less motivated to engage in and enjoy effortful cognitive activities (Cacioppo, Petty, Feinstein, & Jarvis, 1996). Research has shown that NFC 51 moderated the effect of argument quality on persuasion such that there were more favorable evaluations for implicit conclusions over explicit conclusions for individuals who were high in NFC (Martin, Lang, & Wong, 2003). Sargent (2004) found that individuals high in NFC were less supportive of punishment than those low in need for cognition. The moderating role of NFC were examined in a series of studies. Because it is generally found that NFC predicted heuristic-based persuasive messages (Salerno & McCauley, 2009), Key, Edlund, Sagarin, and Bizer (2009) tested the moderating role of NFC to the heuristic-based compliance technique and did not find a moderating effect. One study examined the moderating effects of both NFA and NFC to affect- and cognition-based persuasive messages (Haddock, Maio, Arnold, & Huskinson, 2008). Their study demonstrated that affective-based message was more persuasive among individuals who were high in NFA and low in NFC, whereas a cognitive-based message was more persuasive among individuals who were low in NFA and high in NFC. In this dissertation, blame is proposed to have a cognitive component, which involves cognitive analysis of attribution of responsibility. After exposing to the antecedents of blame, how individuals experience blame may be dependent upon their NFC level, as blame requires complicated cognitive processing. The moderating role of NFC in attitude-behavior relations and perceptionbehavior relations has been identified in several studies (Cho & Park, 2014; Haug et al., 2010; Salerno and McCauley, 2009; Verplanken,1989). For example, one study found that NFC moderated the linkages between instrumental beliefs, social influence factors, and behavioral intentions of adopting smartphone (Cho & Park, 2014). Specifically, perceived usefulness had a stronger effect on the behavioral intention of using a smartphone for highNFC people, whereas perceived ease of use and subjective norms had a stronger effect on 52 people low in NFC. Based on the previous research, NFC is included as a potential moderator in the relationships of blame and social responses. Moral identity is a self-regulatory tendency that motivates moral behavior and provides a basis for social identification that people use to construct their self-definitions (Aquino & Reed II, 2002). Moral identify was found to be associated with various moral actions, such as donation behavior and volunteer activities (Aquino & Reed II, 2002). People with a stronger moral identity were expected to be more sensitive to moral violations and thus are more likely to take actions to keep their moral identity. Aquino and Reed II’s study (2002) revealed two dimensions of moral identity: symbolization and internalization. The symbolization dimension taps into “public self,” referring to “a general sensitivity to the moral self as a social object whose actions in the world can convey one has these characteristics”; and the internalization dimension focuses on “private self,” referring to “the self-importance of the moral characteristics” (Aquino & Reed II, 2002, p.1427). Previous research has shown that symbolization and internalization were associated with different outcomes (Vitell et al., 2009; Winterich, Aquino, Mittal, & Swartz, 2013). Symbolization was found to be more influential on behaviors that were subject to public scrutiny, particularly those driven by social reward or social recognition, and internalization was found to be related to donation behavior (Winterich et al., 2013). The moderating role of moral identity was examined by previous research. Xu and Ma (2016) explored how moral identity interacts with moral decision to influence moral behavior. They suggested that moral judgment is guided by two competing principles: deontology and utilitarianism. Individuals who use deontology to guide their behaviors tend to follow universal moral rules, and individuals who use utilitarianism to guide their behaviors are 53 more likely to focus their attention on the ends of a moral act. It is found that deontology was more likely to lead to moral behavior among individuals who have strong moral identity. In a similar study, Reynolds and Ceranic (2007) found that moral identity interacted with moral judgments to influence moral behaviors. Specifically, internalization of moral identity interacted with both consequentialism and formalism to lead individuals to the most extreme or idyllic of solutions. Reed, Aquino, and Levy (2007) examined the link between moral identity and judgments of charitable behaviors and it was found that consumers with higher organizational status tended to give money versus time, and this tendency was weaker for those with a high self-important moral identity. Another study found that dishonest behavior on charity donation was moderated by moral identity such that individuals who were high in moral identity made more efforts to uphold a moral selfimage by donating more to charity after having lied (Mulder & Aquino, 2013). Because moral identity is an individual-difference variable that represents one’s moral character and blame often occurs when there are moral violations, it is likely that the effect of blame antecedents on blame depend on the level of moral identity. As previous studies have shown (Reynolds & Ceranic, 2007; Xu & Ma, 2016), the effect of moral judgment on moral behaviors was contingent upon moral identity. It is likely that blame as a psychological response, which is associated with violation of moral standard, interacts with moral identity, to influence prosocial behaviors. Hypotheses and Research Questions This dissertation proposes the following research questions and hypotheses. First, it focuses on two antecedents of blame: outcome severity and preventability. In the context 54 of childhood obesity, greater blame may arise when the negative consequences were more severe, and when what happened to the child in the story could have been prevented by the parents. These two dimensions may also jointly influence the amount of blame. The first set of hypotheses were therefore proposed to examine the effects of two antecedents of blame: H1a: Outcome severity positively predicts blame. H1b: Outcome preventability positively predicts blame. H1c: There is an interaction effect between outcome severity and preventability on blame. Second, I proposed a set of hypotheses to examine the potential social responses following blame, including support for childhood-obesity related policies, information sharing behavior, civic participation, and punishment. In addition, how some individual trait variables may have moderating effects are also examined. Based on the review of relevant studies in this chapter, the following hypotheses and research questions were proposed. H2: Blame positively predicts support for public policies addressing the problem of childhood obesity. H3: Blame positively predicts information sharing behavior about childhood obesity. H4: Blame positively predicts civic participation behaviors for preventing childhood obesity. H5: Blame positively predicts the tendency to punish parents of the obese child. RQ2a: Will the effects of outcome severity and outcome preventability specified in 55 H1 be contingent upon individual traits, including (a) need for affect, (b) need for cognition, and (c) moral identity? RQ2b: Will the effects of blame on social responses specified in H2-H5 be contingent upon individual traits, including (a) need for affect, (b) need for cognition, and (c) moral identity? CHAPTER 4 THE PILOT STUDY: USING A COLLEGE STUDENT SAMPLE A pilot study was conducted in Spring 2017 using a convenience sample of undergraduate students to evaluate the feasibility of the experimental design and assess the empirical properties of the key measures. The pilot study is an important component of a research project as it can help identify potential problems associated with the research design (Hassan, Schattner, & Mazza, 2006). Specifically, this pilot study serves three purposes: (1) to identify potential problems with the study design, especially in terms of manipulation check of stimulus materials, (2) to empirically assess the validity and reliability of some key constructs, (3) to get preliminary findings from a college student sample. Ultimately, the major goal of the pilot study is to inform what changes, if any, need to be made for the main study. Methods Power Analysis Using G*power software (Faul, Erdfelder, Buchner, & Lang, 2009), the results of power analysis suggested that a sample size of 119 and 158 would be adequate to detect a medium effect size of f = .25 given the statistical power 0.8 and 0.9. Because the main purpose of the pilot study was to test the feasibility of the main study rather than 57 conducting a full-scale study, the pilot study was slightly undersampled. Participants For the pilot study, 131 undergraduate students were recruited from three departments at the University of Utah (Department of Communication, Department of Sociology, and School of Business). Participants were given a small amount of course extra credit for participating in the study. The Web-based experiment was embedded in an online survey created on Qualtrics (Qualtrics, 2018). Participants were sent a URL link to the survey and completed it in a location of their choice with Internet connection. The final sample of the pilot study consisted 118 participants after removing duplicates and invalid responses. Responses with completion time less than five minutes were removed. Of the final sample, 56.8% were female and 43.20% were male, and the average age was 25.27 (SD = 4.29). The majority of the participants (67.80%) were White and Caucasian, 14.40% were Asian, 2.5% were Native Hawaiian or Pacific Islander, and 0.8% were Black or African American; Hispanic participants accounted for 16.80 % of the sample. The majority of the participants (72%) were never married. Most of the participants (89%) reported having no children. Message Design The experiment employed a 2 (outcome severity: fatal vs. near-fatal consequence) × 2 (outcome preventability: high vs. low) between-participants design. The experimental stimuli were created by modifying an online news article (Slater, 2014). Both exemplification theory (Zillmann, Knobloch, & Yu, 2001) and the identified victim effect 58 (Small, Loewenstein, & Slovic, 2007) suggest that in terms of involving people to solve social issues, a brief story depicting specific individuals is more persuasive than statistical descriptions about the issue. The basic storyline described what happened to a 7-year old boy who was obese. Four versions of the story (Appendix A-D) were created, which differed in terms of whether the boy died or almost died of a heart attack due to obesity (outcome severity manipulation), and whether the boy’s obesity was largely genetic or attributable to irresponsible parenting (outcome preventability manipulation). Manipulation strategies included varying headlines, story leads, and quotes as suggested by previous research (Nelson & Oxley, 1999; Slothuus, 2008). The four versions were roughly comparable in length (word count: from 284 to 303 words). The information about childhood obesity, the background information about the obese child and his parents, and the recommendations for preventing childhood obesity stayed constant across the four versions of the story. Measures For each key construct, I conducted a confirmatory factor analysis (CFA) using Stata 12 (Stata 12, 2011) to validate the factor structure of its measures. Four model fit indices were used to assess the model fit, including chi-square, RMSEA (Browne & Cudeck, 1992), CFI (Bentler, 1990), and SRMR (Hu & Bentler, 1999). A value of 0.06 or lower for RMSEA, less than 0.09 for SRMR, 0.95 or higher for CFI, was generally recognized as a good fit (Hu & Bentler, 1999). Scale reliability was assessed and Cronbach’s alpha was reported. Blame as cognition. The cognitive aspect of blame was measured by an eight-item 59 scale of blameworthiness adapted from previous research (e.g., Abrams, Viki, Masser, & Bohner, 2003; Gerber, Cronin, & Steigman, 2004; Transgurd, 2010). Participants were asked to indicate their agreement with each of the statements (1= “strongly disagree,” 7= “strongly agree”). Sample items included the following: “Ryan’s parents were to blame for what Ryan suffered”; “Ryan’s parents had no good reason for letting this happen.” The CFA showed the one-factor scale of blame as cognition had a good fit, χ2 (11) = 6.41, p >.05, RMSEA= 0.02, CFI = .99, and SRMR = .03. These items demonstrated high internal consistency ( = .88) and were averaged into the cognitive scale of blame (M = 3.43, SD = .61). Moral indignation. Moral indignation scale was adapted from the media indignation scale proposed by Hwang et al. (2008). Participants were asked to respond to the question “While you were reading the story, to what extent did you experience each of the following emotions?” (0 = “none of this feeling” and 4 = “a great deal of this feeling”), and five items were included: contemptuous, angry, disgusted, infuriated, and outraged. The CFA showed a one-factor structure, χ2(3) = .67, p > .05, RMSEA = .00, CFI = 1, and SRMR = .01. The five items were averaged into one scale ( = .91, M = 3.01, SD = 1.02). Civic participation. The civic activities that participants involved were measured by three items to the question “If given a chance, how likely are you to do each of the following?” (1 = “very likely,” 5 = “very likely”). The items included the following: “volunteer for a public education campaign to increase the awareness of childhood obesity,” “join an on-campus student organization that works on issues related to childhood obesity,” and “make a small donation to a nonprofit health organization that fights childhood obesity.” Because the civic participation scale only has three items, which is the minimum 60 requirement for doing confirmatory factor analysis, the CFA model was not identified. These items, with a high reliability (= .85), were averaged into a scale of civic participation (M = 3.01, SD = 1.16). Information sharing behavior. Participants’ information sharing behavior was measured by the likelihood of “doing each of the following information-related activities” (1 = “very likely” and 5 = “very likely”). Four items were included, such as “share this story about Ryan on my Facebook or Twitter, or another social media platform.” The results of the CFA suggested a good model fit after adding correlated errors, χ2 (1) = 0.25, RMSEA= 0, CFI = .99, and SRMR = .007. The items demonstrated adequate reliability ( = .78, M = 2.87, SD = 1.04). Support for policies. Two sets of policy support questions were included: policies requiring additional tax payment, and policies without requiring additional tax. For the first, four items were adapted from a previous study to gauge participants support for policies related to childhood obesity that require paying an $50 per year (Barry et al., 2013). Sample items included “Prohibit sales of fast food and soft drinks in public school cafeterias and stores and use federal tax dollars to compensate the schools for the revenues they now make on these products” and “Use federal funds to open and sustain full service grocery stores in communities with limited access to healthy foods.” The results of the CFA showed a poor model fit, χ2(2) = 9.27, p < .01, RMSEA = .18, CFI = .94, and SRMR = .06. Because the ideal sample size for the CFA is 200 (Loehlin, 1998), the poor fit could be due to the relative sample size used in the pilot study. The reliability of tax policy support scale was .74, M = 3.76 and SD = .86. For policies without requiring additional tax, six items were adapted from a 61 previous study (Barry et al., 2013). Sample items included “encouraging public school governing authorities to create partnerships for the purpose of increasing parental involvement in schools,” and “using federal funds to create a national network of summer camps for low-income children that emphasize good nutrition and exercise.” The results of the CFA showed that nontax policy support had a good fit after adding some correlated errors, χ2(4) = 5.66, p > .05, RMSEA = .06, CFI = .99, and SRMR = .03. The reliability of nontax policy support scale was .82, M = 3.84, SD = .84. Individual traits variables, including need for affect and moral identity, were included as moderators. Need for affect. A 10-item short form of the need for affect (NFA) Questionnaire developed by Appel, Gnambs, and Maio (2002; 1= “not at all like me,” 5= “a lot like me”) was used to measure participants’ need for affect. NFA includes two subscales: emotion approach (e.g., “Emotions help me get along in life.”) and emotion avoidance (e.g., “I find strong emotions overwhelming and therefore try to avoid them.”). CFA results demonstrated that the two-factor model had a fair fit, χ2(34) = 51.99, p < .05, RMSEA = .07, CFI = .96, and SRMR = .07. The items of emotion approach and emotion avoidance demonstrated adequate reliability ( = .84, M = 3.69, SD = .81, for emotion approach;  = .83, M = 2.44, SD = .91, for emotion avoidance). Moral identity. Participants’ moral identity was measured by a 10-item scale developed by Aquino and Reed II (2002), which had been validated in a few prior studies (Baker, 2015; Xing & Keung, 2014). Previous research showed two factors, symbolization and internalization. Participants were asked to visualize a person who has the following characteristics: caring, compassionate, fair, friendly, generous, hardworking, helpful, 62 honest, and kind. Sample items included “Having these characteristics is not really important to me” (internalization subscale), and “I am actively involved in activities that communicate to others that I have these characteristics” (symbolization subscale). The results of the CFA showed that the two-factor model had a reasonable fit, χ2(31) = 59.31, p < .01, RMSEA = .07, CFI = .95, and SRMR = .07. The reliability of the symbolization subscale was α = .80 (M = 3.51, SD= .77), and for internalization subscale, α = .82, M = 3.27, and SD = .43. Analysis Strategies Two-way ANOVAs in which preventability and the severity of consequence were entered as independent variables were used for manipulation check. Structural equation modeling technique was used to test the validity of the proposed four conceptual models (see Figure 1 to Figure 4) using maximum likelihood estimators in Mplus 7.4 Because the Dual Process Cognitive-affective model subsumed both Single Process models, comparing the Dual Process Cognitive-affective model and the Intertwined Process model should yield all the possible results. To test the goodness of fit for the hypothesized models, four fit indices were considered for evaluating the model fit: Root Mean Square Error of Approximation (RMSEA), standardized root mean square residual (SRMR), comparative fit index (CFI) and Bayesian Information Criterion (BIC; Raftery, 1995). RMSEA and SRMR are absolute fit indices that determine how well an a priori model fits the sample data (McDonald & Ho, 2002) and demonstrates which proposed model has the most superior fit. CFI is a relative fit index comparing a chi-square for the model tested to one from a so-called null model 63 (also called a “baseline” model or “independence” model; Tanaka, 1993). A value of 0.06 or lower for RMSEA, less than 0.09 for SRMR, 0.95 or higher for CFI, was generally recognized as a good fit (Hu & Bentler, 1999). CFI values range from 0 to 1, with better overall fit indicated by higher values. According to Hu and Bentler (1999), CFI values of 0.95 or higher is considered as a good-fitting model and cutoff values close to 0.09 for SRMR. BIC is commonly used for choosing an optimal model from the alternatives (Lin, Huang, & Weng, 2017). Better fitting models have more negative BIC values. The advantage of BIC is that is can be used for non-nested model comparisons. Another advantage of BIC is that even for small sample sizes, the BIC will favor more parsimonious models than the AIC or chi-square difference tests (O’Connell & McCoach, 2008). Lower BIC indicates better model fit. Raftery (1995) recommended that a BIC difference of 5 offered evidence of model improvement and any value above 10 was strong evidence that the model with the lower BIC fit best. To examine the effect of blame on social response variables, I performed SEM analyses with all the social response variables, including policy support, information sharing behavior, and civic participation, in the model. All the social response variables were entered into the model simultaneously. For this set of analyses, blame was treated as a mediating variable between the antecedents of blame, including the severity of consequence, preventability, and their interaction term on the social responses variables. The interaction effect between individual trait variables and the antecedents of blame on blame was tested using Mplus. XWITH command in Mplus was used to test the interaction effect between latent variable and observed variables. I chose to use XWITH 64 command instead of using a multigroup comparison to test the moderation effect because the moderators included in the analysis were continuous variables and dichotomizing them into groups of high versus low would cause loss of variance or information. MacCallum, Zhang, Preacher, and Rucker (2002) pointed out that dichotomization of quantitative measures is not a desirable approach as it could result in loss of information about individual differences, loss of effect size and power, and loss of measurement reliability. Descriptive statistics were examined for all variables to identify and correct outliers. The examination of frequency distribution showed that there was no outlier for all the variables included in the analysis. Table 2 presents the correlation matrix of each variables in the analysis. Results Manipulation Check Manipulation check was done on perceived severity and perceived preventability. The perceived outcome preventability was measured by three items: “Ryan’s tragedy was entirely avoidable,” “What happened to Ryan could have been prevented,” and “Ryan’s situation would not have occurred to other children with more responsible parents” (1= “strongly disagree,” 5= “strongly agree”). The perceived outcome severity was measured by four 5-piont semantic differential questions. Four word pairs included the following: trivial/ serious, severe/slight, negligible/dreadful, and grim/mild. The alpha reliability of the preventability scale and the severity of consequence scale were .70 and .73, respectively. The results of two-way ANOVAs showed that there was no significant main effect for preventability on the perceived level of preventability, F (1,114) = 2.42, p > .05, and 65 the severity of consequence on the perceived level of consequence, F (1,114) = 2.73, p > .05. There was a significant interaction on the perceived level of preventability, F (1,117) = 4.35, p < .01 (Figure 5). In the nonsevere consequence condition, participants in the high preventability condition perceived the negative outcome as more preventable than those in the low preventability condition, p <.05 (M = .3.95, SD = .78, compared to M = 3.46, SD = .78). In the severe consequence condition, there was no significant difference between two groups. Similarly, there was also a significant interaction on the perceived outcome severity (Figure 6), F (1,117) = 7.29, p < .01. Therefore, the manipulation check showed that the severity of consequence and preventability manipulation were only partially successfully under certain conditions. The Structure of Blame (RQ1a and RQ1b) Research Question 1a posed the following question: should blame be conceptualized and operationalized as a cognition, affect, or both? Because the Dual Process model subsumed both Single Process models, assessing the Dual Process model and the Intertwined Process model should provide the evidence for the nature of blame. The Dual Process model contained cognitive and affective aspects of blame as separate paths. In the Intertwined Process model, blame was treated as a latent construct composed of cognition and affect. Nontax policy support was used as the outcome variable because the CFA analysis showed that the scale for policy change requiring taxes was a poor fit. If the data supported the Single Process Cognitive Model, then there should be a significant path to and from cognition and nonsignificant paths to and from affect. If the data supported the Single 66 Process Affective Model, there should be a significant path to and from affect and nonsignificant paths to and from cognition. Figures 7-10 present the path coefficients for the four models, including the Single Process Affective model, the Single Process Cognitive model, the Dual Process model, and the Intertwined Process model. It was found that preventability predicted cognition in the dual process model and other paths were not significant (Figure 9). Then I tested the Single Process Affective Model (Figure 7) and the Single Process Cognitive Model (Figure 8), and it was found preventability predicted affect and cognition separately, but other paths were not significant. I also tested the SEM models using civic participation and information sharing as the outcome variables. When using information sharing as the outcome, it was found that the path from emotion to information sharing was significant (b=.06, p <.001) in the Dual Process model but other paths were not significant; affect predicted information sharing (b= .51, p <.001) in the Single Process Affective model and preventability predicted cognition (b= .09, p <.05) in the Single Process Cognitive model but other paths in both models were not significant; all the paths in the Intertwined Process model were not significant. When using civic participation as the outcome, both affect (b=.48, p <.01) and cognition (b=-.52, p <.001) predicted civic participation but other paths were not significant in the Dual Process model; preventability predicted cognition in the Single Process Cognitive model (b=.09, p <.05) and affect in the Single Process Affective model (b=.09, p <.05) but other paths in both models were not significant; all the paths in the Intertwined Process model were not significant. In response to RQ1a, therefore, there was no conclusive evidence about the nature of blame. RQ1b asked the following: If blame has both cognitive and affective component, how are the two aspects combined? In answering RQ1b, I compared the Dual Process 67 Model and the Intertwined Process Model. The results were in favor of the Intertwined Process Model over the Dual Process Model based on both absolute (i.e., RMSEA, SRMR) and relative (CFI, BIC difference) indices (see Table 3-5). The answer to RQ1b therefore is that blame should be conceptualized as an alloy of both cognition and affect. It should be noted that such results were only suggestive, but there was more support for the Intertwined Process Model. Antecedents of Blame (Hypotheses 1a to 1c) The first set of hypotheses aimed to examine the effects antecedents, preventability and severity, on blame. In the final intertwined model (Figure 10), the path from preventability to blame was significant (H1b), b = .17, p < .05; outcome severity did not predict blame (H1a), b = -.04, p > .05, nor was there an interaction effect (H1c), b=. -.01, p > .05. Therefore, H1b was supported, but H1a and H1c were not supported. Outcomes of Blame (H2 – H4) This set of hypotheses predicted the effect of blame on policy support (H2), information sharing (H3), and civic participation (H4). SEM analyses were conducted with all the social response variables, including policy support, information sharing behavior, and civic participation, in the model. As the measure of punishment was not included in the pilot study, H5 could not be tested with pilot study data. The results showed that blame did not predict nontax policy support, b = .05, p > .05. The path from blame to civic participation in childhood obesity prevention activities was not significant either, b = .08, p > .05. The SEM analysis with information sharing behavior as the outcome variable 68 showed a significant path from blame, b = .25, p < .01. Therefore, H3 was supported, and there was no support for H2 or H4. Moderators (RQ2a and RQ2b) The moderator analysis revealed that the interaction term between preventability and symbolization of moral identity had a positive effect on blame, b = .96, p < .05, despite the fact that preventability and symbolization did not have a significant main effect on blame. Internalization of moral identity and need for affect were not found to be significant moderators between the relationships of the antecedents of blame and blame. Neither subscales of need for affect was significant moderators of the effect of blame on subsequent social responses, nor were the subscales of moral identity significant moderators in any scenario. Need for cognition was not included in the pilot study. Summaries and Implications of the Pilot Study Findings from the pilot study can be summarized as follows. First, in terms of manipulation check, it was found that both manipulations of the severity of consequence and preventability were only partially successful such that the interaction of the severity of consequence and preventability had a significant effect on the perceived levels of outcome severity and preventability. Second, in assessing measurement of key constructs, it was found that most key constructs had good construct validity. Third, preliminary findings from this college students sample showed that (1) the Intertwined Process model received the most support, but whether blame should be conceptualized as an alloy of cognition and affect needs further investigation; (2) outcome preventability was a significant antecedent 69 of blame; (3) blame had a significant effect on information sharing but had no significant effect on policy support or civic participation; and (4) symbolization of moral identity moderated the effect of preventability on blame. The pilot study helped revise the study design in three ways. First, as manipulation check showed that preventability effect disappeared for severe consequences (i.e., death of the child), I speculated that one reason is that death as a severe outcome could have been too overwhelming, to the extent that the difference between high versus low preventability was no longer relevant or important. Therefore, in the main study, the manipulation of outcome severity was changed to nonfatal consequences (e.g., severe heart attack vs. mild heart attack). Second, modifications were made to the key measures based on empirical, conceptual, and practical considerations. For example, two items in blame as cognition were removed since both of them were similar to the other items in terms of underlying meaning and some of the items were reworded to better reflect the cognitive aspect of blame. A comparison between the items used pilot versus main study can be seen in Table 6. Third, the pilot study used a single-message design, a limitation that could pose a challenge to the generalizability of this study. O’Keefe (2002) suggested that a welldesigned study should have multiple instances of each message type. Therefore, in the main study, I added another set of stimuli constructed based on the same experimental design. The second set of experimental materials focus on a different health risk associated with childhood obesity: asthma attacks. Variable 1.Preventability 2.Severity 3.Preventability By Severity 4.Cognition 5.Affect 6.Non-tax Policy Support 7.Tax Policy Support 8.Civic participation 9.Information Sharing *p< 0.05, **p < 0.01 .46** 1 .53** -.05 -.13 .25** 1 .02 -.06 -.08 .01 -.05 .02 .56** -.02 -.06 -.12 .05 -.12 .05 .60** .18 .16 -.01 .04 .18 .69 4 1 3 1 1 .02 2 Table 2 Correlations of the Key Variables for the Pilot Study (N =118) 1 .05 .08 .10 .04 5 1 6 .69** .37** .12 1 7 .41** .23** 1 8 .61** 1 9 70 71 Table 3 Fit Indices and Model Comparison for the Pilot Study Using the Nontax Policy Support as the Outcome (N =118) Chi-square The Dual Process Model The Intertwined Model a RMSEAa CFIb BICc BIC difference 30.56(df=3) .24 .84 .09 1257.93 6.31(df=6) .02 .99 .03 1230.46 27.47 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b SRMRb 72 Table 4 Fit Indices and Model Comparison for the Pilot Study Using Civic Participation as Outcome (N =118) Chi-square The Dual Process Model The Intertwined Model a RMSEAa CFIb BICc BIC difference 29.07(df=4) .23 .86 .09 1294.06 15.56(df=6) .12 .95 .05 1270.63 23.43 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b SRMRb 73 Table 5 Fit Indices and Model Comparison for the Pilot Study Using Information Sharing as Outcome (N =118) Chi-square The Dual Process Model The Intertwined Model a RMSEAa CFIb SRMRb BIC difference 26.50(df=4) .22 .87 .09 1266.38 4.15(df=6) 0 1 .03 1238.98 27.4 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b BICc 74 Table 6 Comparisons of Scales Used in the Pilot Study and the Main Study Scales The Pilot Study Moral Indignation 1. 2. 3. 4. 5. 6. Blame as cognition 1. Overall, Ryan's parents were responsible for the event occurred in the described scenario. 2. Ryan's parents had control over the event that occurred in the described scenario. 3. Ryan's parents acted carelessly in the described scenario. 4. Ryan's parents were at fault for the consequence in the described scenario 5. Ryan's parents were to blame for the consequence in the described scenario. 6. Ryan's parents should have foreseen the consequence in the described scenario. 7. Ryan's parents acted properly. 8. Ryan's parents' behaviors were responsible for the consequence in the described scenario. 1. Ryan’s parents acted irresponsibly with regard to Ryan's health. 2. Ryan’s parents were at fault for the consequence described in the story. 3. Ryan's parents were to blame for what Ryan suffered. 4. Ryan's parents were largely responsible for what happened to Ryan in the story. 5. Ryan’s parents can be excused. 6. Ryan’s parents had no good reasons for letting this happen. Need for affect Emotion approach 1. Emotions help me get along in life. 2. It is important for me to know how others are feeling. Same infuriated angry disgusted outraged contemptuous annoyed The Main Study Same 75 Table 6 Continued Scales Perceived level of preventability Perceived level of severity Moral identity The Pilot Study The Main Study 4. I enjoy a task that involves coming up with new solutions to problems. 5. I prefer my life to be filled with puzzles that I must solve. Same 1. Ryan's tragedy was entirely avoidable. 2. What happened to Ryan could have been prevented. 3. Ryan's situation would not have occurred to other children with more responsible parents. 4. Ryan's parents could have done something to prevent this from happening. Think of what happened to Same Ryan depicted in this story. In your opinion, the consequence depicted in the story was.... -serious, -severe. -dreadful. -grim. Internalization 1. It would make me feel good to be a person who has these characteristics. 2. Being someone who has these characteristics is an important part of who I am. 3. I would be ashamed to be a person who has these characteristics. 4. Having these characteristics is not really important to me. 5. I strongly desire to have these characteristics. 76 Table 6 Continued Scales The Pilot Study The Main Study Symbolization Same 1. I often wear clothes that identify me as having these characteristics. 2. The types of things I do in my spare time (e.g., hobbies) clearly identify me as having these characteristics. 3. The kinds of books and magazines that I read identify me as having these characteristics. 4. The fact that I have these characteristics is communicated to others by my membership in certain organizations. 5. I am actively involved in activities that communicate to others that I have these characteristics Punishment Policy Support 1. Ryan’s parents should be punished 2. Ryan’s parents should be fined. 3. Ryan’s parents should receive public reprimand. 4. Ryan’s parents should be penalized. Nontax policy support Parent-oriented policy support 1. Require all public schools 1. Require parents who have to offer physical education obese children to bring them classes for students at least to regular medical checkups. three times per week. 2. Require parents with obese 2. Prohibit use of cartoon children to participate in characters to sell unhealthy community programs on food high in fat and sugar appropriate nutrition and to children in physical activities. advertisements or package labeling. 77 Table 6 Continued Scales Policy Support The Pilot Study 3. Require that television stations provide free time for public-service announcements on healthy eating and exercise in proportion to the food advertising they carry during children’s programming. 4. Impose a tax on junk food similar to existing government taxes on cigarettes and alcohol. 5. Prohibit advertising of unhealthy foods high in fat and sugar in schools. 6. Require grocers to add a surcharge to high-sugar, high-fat foods and use the revenues to reduce their prices for fresh fruits. The Main Study 3. Develop more public health campaigns to raise parents’ awareness about childhood obesity. 4. Develop more parentcentered education programs to prevent childhood obesity. Public-oriented Policy Support 1. Increase taxes on fast food and soft drinks in public school cafeterias and stores. 2. Prohibit advertising of unhealthy foods high in fat and sugar in and around schools. 3. Use federal funds to increase access to healthy foods in neighborhoods around schools. 4. Use federal funds to increase the number of eligible children enrolled in school meal programs. Communicative behavior 1. Share this story about 1. Share this story about Ryan Ryan on my Facebook or on your Facebook or Twitter, or another social Twitter, or another social media platform. media platform 2. Forward this story about 2. Forward this story about Ryan via email to friends, Ryan via email to friends, family, or other people you family, or other people you know. know. 3. Talk to others (friends, 3. Talk to others (friends, family, colleagues, etc.) family, colleagues, etc.) about the problem of about the problem of childhood obesity. childhood obesity Talk to others (friends, family, colleagues, etc.) 78 Table 6 Continued Scales 4. 5. 6. 7. Civic participation 1. 2. 3. The Pilot Study Share this story about Ryan on my Facebook or Twitter, or another social media platform. Forward this story about Ryan via email to friends, family, or other people you know. Talk to others (friends, family, colleagues, etc.) about the problem of childhood obesity. Talk to others (friends, family, colleagues, etc.) about the importance of preventing childhood obesity. Volunteer for a public education campaign to increase the awareness of childhood obesity. Join an on-campus student organization that works on issues related to childhood obesity. Make a small donation to a non-profit health organization that fights childhood obesity. The Main Study 4. Talk to others (friends, family, colleagues, etc.) about the importance of preventing childhood obesity. 5. Pass on information about childhood obesity to others (friends, family, colleagues, etc.). 1. Volunteer for a public education campaign to increase the awareness of childhood obesity. 2. Wear or display a "Childhood Obesity Prevention" badge or sticker. 3. Sign on a petition for using more federal fund to tackle childhood obesity. 4. Register an email listserve to receive messages about childhood obesity prevention and treatment. 79 Figure 5. Interaction Effect Between Preventability and Outcome Severity on Perceived Level of Outcome Severity in the Pilot Study (N = 118). 80 The Perceived Level of Preventabilty 4 3.9 3.8 3.7 3.6 Non-Severe 3.5 Severe 3.4 3.3 3.2 Not Preventable Preventable Outcome Preventability Figure 6. Interaction Effect Between Preventability and Outcome Severity on Perceived Level of Preventability in the Pilot Study (N = 118). 81 Figure 7. The Obtained Single Process Affective Model for the Pilot Study. 82 Figure 8. The Obtained Single Process Cognitive Model for the Pilot Study. 83 Figure 9. The Obtained Dual Process Model for the Pilot Study (N = 118).1 1 For both the Dual Process Model and the Intertwined Process Model, preventability (with the preventable condition was coded as 1 and nonpreventable condition was as 0), severity (with the severe condition was coded as 1 and less severe condition was coded as 0), and the interaction term between preventability and the severity of consequence were specified as latent exogenous variables. The measurement paths for preventability, severity, and interaction term were set to 1.00. In the DualProcess model, cognition and affect were corrected for measurement error by fixing the error term of the corresponding manifest variables to 1- 𝑎 times its variance. In both Dual Process and Intertwined Process Model, nontax policy support was corrected for measurement error fixing the error term of the corresponding manifest variable at 1- 𝑎 times its variance. 84 Figure 10. The Obtained Intertwined Model for the Pilot Study (N = 118). CHAPTER 5 THE MAIN STUDY: USING THE GENERAL POPULATION With modifications of the study design based on the pilot study, the main study was launched in December 2017 using a sample of general population recruited from Amazon’s Mechanical Turk (MTurk). Details of the main study and results from the analyses are reported in this chapter. Methods Power Analysis Using G*power software (Faul, Erdfelder, Buchner, & Lang, 2009), the results of the power analysis suggested that a sample size of 128 and 171 would be adequate to detect a medium effect size of f = .25 (Cohen, 1988) given the statistical power 0.8 and 0.9. To detect a small effect size (f = .01) would require a sample of 787 and 1053 participants given the statistical power 0.8 and 0.9. Participants A total of 433 participants were recruited via Amazon’s Mechanical Turk (MTurk) and asked to complete a web-based survey from January 3, 2018 to January 13, 2018. MTurk is a marketplace for work that requires human intelligence (Alonso & Mizzaro, 86 2009). Researchers have been increasingly using MTurk to collect data for social science research. Previous research showed that MTurk enables researchers to obtain high-quality data inexpensively and rapidly (Buhrmester, Kwang, & Gosling, 2011). Each participant was given one dollar for completing the survey. The survey, designed in Qualtrics, an online survey software (Qualtrics, 2018), was embedded in MTurk. After signing an online consent form, participants were randomly assigned to eight experimental conditions and instructed to read a news story about childhood obesity. To prevent bots filling out the surveys, CAPTCHA, which is a computer program that determines whether or not the user is human, was embedded in the survey (Captcha, 2018). Bots were automatically screened by CAPTCHA and unable to fill out the survey. Moreover, the time that each participant spent on reading the news story was recorded. Data were screened based on the time spent by the participants on reading the experimental materials. Given the length of the news story, the participants who spent less than 10 seconds on the reading the news story were removed from the data analysis. The data were first cleaned and screened following the recommendation established by Tabachnick and Fidell (2001). After data cleaning procedures, a total of 373 participants were included in the analysis. Based on the power analysis, this sample size will allow me to detect a medium to small effect size. On average, participants spent 14.3 minutes completing the survey. The majority of the participants were White (75.9%), 13.10% were Asian, 11.50% were African Americans, 2.10% were American Indians or Alaska Natives, and 1.1% were mixed race. There were 48.80% participants were male and 50.9% were female. Of the total participants, only 7% had Hispanic, Latino, or Spanish origin. The age of participants ranged from 19 to 73 (M = 37.13, SD = 11.17). Of all the participants, the majority of 87 participants had no children, accounting for 51.50%, with 17.40 % that had two children, 16.90% that had one child, and 14.20% that had three or more children. The majority of the participants had bachelor’s degree or higher (54.2%), 21.20 % with some college but no degree, 21.90% with an associate degree in college, and 11.80 % of the participants were high school graduates. Among all the participants, 41.50 % of them had annual household incomes in the range of $40,000–80,000, and 40.50% were in the range of less than $40,000 per year. Message Design The experiment employed a 2 (obesity-related health risk: heart attack vs. asthma attack) × 2 (outcome preventability: high vs. low) × 2 (outcome severity: severe vs. less severe) between-participants design. Messages were created by modifying an online news article. The basic storyline described what happened to a 7-year old boy who was obese. Eight versions of the story were created (Appendix E-L), which differed in terms of whether the boy suffered a severe or mild obesity-related health risk (outcome severity manipulation), whether the boy’s obesity was largely genetic or attributable to irresponsible parenting (preventability manipulation), whether the boy suffered a heart attack or an asthma attack (risk type manipulation). Manipulation strategies included varying headlines, story leads, and quotes as suggested by previous research (Nelson & Oxley, 1999; Slothuus, 2008). The eight versions were roughly comparable in length (word count: from 316 to 356 words). The information about childhood obesity, the background information about the obese child and his parents, and the recommendations for preventing childhood obesity stayed constant across the eight versions of the story. 88 Measures For each key construct, confirmatory factor analysis (CFA) was conducted using Mplus 7.4 to validate the factor structure of its measures (Muthén & Muthén, 2015). Four model fit indices were used to assess the model fit, including chi-square, RMSEA (Browne & Cudeck, 1992), CFI (Bentler, 1990), and SRMR (Hu & Bentler, 1999). Scale reliability was then assessed and Cronbach’s alpha was reported. A value of 0.06 or lower for RMSEA, less than 0.09 for SRMR, 0.95 or higher for CFI, was generally recognized as a good fit (Hu & Bentler, 1999). Blame as cognition. Blame as cognition was measured by a six-item scale that was developed based on the previous study (Abrams, Viki, Masser, & Bohner, 2003; Gerber, Cronin, & Steigman, 2004; Trangsrud, 2010). Sample items include “Ryan’s parents had no good reasons for letting this happen” and “Ryan's parents were largely responsible for what happened to Ryan in the story.” The factor model had a satisfactory fit, χ2(9) = 18.62, p < .05, RMSEA = .05, CFI = .99, and SRMR = .02. The reliability of the scale was .90 (M = 4.15, SD = .79). Moral indignation. The media indignation scale proposed by Hwang et al. (2008) was adapted to measure moral indignation. Participants were asked to respond to the question “While you were reading the story, to what extent did you experience each of the following emotions?” (0 = “none of this feeling” and 4 = “a great deal of this feeling”), and five items were included: contemptuous, angry, disgusted, infuriated, and outraged. The model had an acceptable fit, χ2(5) = 13.96, p < .05, RMSEA = .07, CFI = .99, and SRMR = .02. The five items were averaged into one scale (𝛼 = .86, M = 3.02, SD = 1.12). Policy support. Two sets of policy support were included, one targeting parents of 89 obese children, and the other aiming at improving the social environment. Parent-oriented policy support was measured on a four-item scale (1 = “strongly oppose” to 7 = “strongly support”) that was developed based on the previous studies (Davison, Jurkowski, Li, Kranz, & Lawson, 2013; Jurkowski et al., 2013; Koller & Mielck, 2009). Participants were asked “Please tell us to what extent you may oppose or support these proposals.” Sample items include “Require parents who have obese children to bring them to regular medical checkups” and “Develop more parent-centered education programs to prevent childhood obesity.” The results of the CFA revealed a satisfactory fit after including some correlated errors, χ2(1) = 1.12, p >.05, RMSEA = 0.02, CFI =1, and SRMR = .005. The reliability of the scale was .77 (M = 4.07, SD = .78). Support for policies to improve social environment was measured on a four-item scale (1 = “strongly oppose” to 7 = “strongly support”) that was adapted from previous studies (Barry, Brescoll, & Gollust, 2013). Participants were asked “Please tell us to what extent you may oppose or support these proposals.” Sample items included “Prohibit advertising of unhealthy foods high in fat and sugar in and around schools” and “Use federal funds to increase the number of eligible children enrolled in school meal programs.” The results of the CFA showed a satisfactory fit, χ2(1) = 0.29, p >.05, RMSEA = 0, CFI =1, and SRMR = .003. The reliability of the scale was .76 (M = 3.73, SD = .93). Information sharing behavior. Information sharing activities were measured by the likelihood of “doing each of the following information-related activities” (1 = “very likely” and 5 = “very likely”). Five items were included, including “Forward this story about Ryan via email to friends, family, or other people you know” and “Share this story on my Facebook or Twitter, or another social media platform.” The factor model showed 90 a good fit,χ2(2) = 0.69, p > .05, RMSEA = 0, CFI =1, and SRMR = .002. The reliability of information sharing behavior was .91(M = 3.06, SD = 1.18). Civic participation. The civic activities of each participants were measured by four items to the question “If given a chance, how likely are you to do each of the following?” (1 = “very likely,” 5 = “very likely”). The items included “Volunteer for a public education campaign to increase the awareness of childhood obesity,” “Wear or display a ‘Childhood Obesity Prevention’ badge or sticker,” “Sign on a petition for using more federal fund to tackle childhood obesity,” and “Register an email listserve to receive messages about childhood obesity prevention and treatment.” The results of the CFA showed that civic participation had a good fit, χ2(2) = 3.95, p > .05, RMSEA = .05, CFI = .99, and SRMR = .01. The five items were averaged into one scale (𝛼 = .88, M = 2.69, SD = 1.19). Punishment. Punishment was measured by four items developed specifically for this study. The punishment scale included the following items: “Ryan's parents should be punished,” “Ryan’s parents should be fined,” “Ryan's parents should receive public reprimand,” and “Ryan’s parents should be penalized.” The 5-point response scale was anchored at 1= “strongly disagree” and 5 = “strongly agree.” The results of the CFA suggested a satisfactory fit, χ2(1) = 2.67, p > .05, RMSEA = .07, CFI = .99, and SRMR = .006. These items demonstrated high internal consistency ( = .93, M = 3.03, SD = 1.29). Moderators. Three individual trait variables were included as moderators, including need for affect, need for cognition, and moral identity. Need for affect (NFA) was measured by a 10-item short form of the NFA Questionnaire that is developed by Appel, Gnambs, and Maio (2002; 1= “not at all like me” and 5= “a lot like me”). The twofactor model indicated an acceptable fit, χ2(31) =105.08, p < .05, RMSEA = .08, CFI = .96, 91 and SRMR = .06. The reliability of emotion approach subscale was .83 (M = 3.47, SD = .88), and the reliability of emotion avoidance subscale was .87 (M = 2.44, SD = 1.04). Need for cognition. Need for cognition of each participant was measured by a five item short need for cognition scale (Nir, 2011). The 5-point scale was anchored at 1= “not at all like me” and 5= “a lot like me.” Sample items in need for cognition included “I prefer complex to simple problems” and “I like having the responsibility of handling a situation that requires a lot of thinking.” The model had a good fit, χ2(2) = 3.26, p > .05, RMSEA = .04, CFI = .99, and SRMR = .001. The reliability of need for cognition was .91 (M = 3.48, SD = 1.07). Moral identity. Moral identity was measured by a 10-item scale developed by Aquino and Reed II (2002) with two subscales: symbolization and internalization. A CFA was performed to cross-validate the two-factor solution suggested by Aquino and Reed II (2002). The model indicated a better fit after adding the covariance between the error terms, χ2(26) = 35.29, p > .05, RMSEA = .03, CFI = .99, and SRMR = .04. The reliability for internalization and symbolization subscale were .79 (M = 4.21, SD = .73) and .86 (M = 3.16, SD = .87), respectively. Control variables. Participants were asked to report their demographic variables, including age, gender, race and ethnicity, income, and education level. In addition, participants’ ideology, marital status, the number of children, and religious belief were also measured in this study. Religious belief was measured on a 5-point scale ranging from 1 (not at all) to 5 (a lot). Participants were asked “How much guidance does religion provide in your everyday life?” Participants were also asked to report their political and economic ideology, which is anchored at a 7-point semantic differential ranging from 1 (very liberal) 92 to 7 (very Conservative). The index of ideology was created (M = 3.43, SD = 1.51). The zero-order correlation analysis of the composite measure of blame and all the demographic and individual characteristic variables revealed that blame was negatively correlated with marital status in the heart attack data and was positively correlated with income in the asthma attack data. Therefore, marital status was entered as a control variable in the heart attack data and income was controlled for in the asthma attack data. Analysis Strategies The analytic process was almost identical with that in the pilot study. Some of the analytical details included in Chapter 4 are not repeated here. Two-way ANOVAs in which preventability and the severity of consequence were entered as independent variables were conducted for manipulation check. Four models of blame were tested using Mplus 7.4 (Muthén & Muthén, 2015). And model fit was assessed based on conventional criteria, including RMSEA, SRMR, CFI, and BIC. To test the moderating effects between individual characteristics and blame, XWITH command in Mplus was used. To test the effect of blame on social responses, parent-orientated policy support, public-oriented policy support, punishment, information sharing behavior, and civic participation were all entered into SEM analysis as the outcome variables simultaneously. Descriptive statistics were examined for all variables to identify and correct outliers. The examination of frequency distribution showed that there was no outlier for all the variables included in the analysis. 93 Results Manipulation Check A manipulation check was conducted to examine whether the experimental manipulations were successful via two-way ANOVAs. The perceived level of preventability was measured by three items: “Ryan’s tragedy was entirely avoidable,” “What happened to Ryan could have been prevented,” and “Ryan’s situation would not have occurred to other children with more responsible parents” (1= “strongly disagree” and 5= “strongly agree”). The perceived severity of consequence was measured by four 5-piont semantic differential questions. Four word pairs included the following: trivial/ serious, severe/slight, negligible/dreadful, and grim/mild. The alpha reliability of the preventability scale and the severity of consequence scale were .85 and .88, respectively. The results of two-way ANOVAs showed that there was a significant main effect for preventability on the perceived level of preventability in the asthma attack data (Figure 11), F (1, 193) = 44.27, p < .001, and the severity of consequence on the perceived severity of consequence (Figure 12), F (1,193) = 38.06, p < .001. Participants in the high preventability condition (M = 4.26, SD = .59) perceived the outcome as more preventable than those in the nonpreventable condition (M =3.62, SD = .76). Participants in the severe condition (M = 4.29, SD = .67) perceived the consequence depicted in the news story as more severe compared to those in the nonsevere condition (M = 3.57, SD = .92). There was also a significant main effect for preventability on the perceived level of preventability in the heart attack data (Figure 13), F (1, 180) = 13.43, p < .001, and the severity of consequence on the perceived severity of consequence (Figure 14), F (1,180) = 19.12, p < .05. Overall, the results showed that the two manipulations were successful. 94 Participants in the severe condition (M = 4.53, SD = .61) perceived the consequence depicted in the news story as more severe compared to those in the nonsevere condition (M = 4.09, SD = .77). Participants in the high preventability condition (M = 4.44, SD = .56) perceived the outcome as more preventable than those in the nonpreventable condition (M = 3.88, SD = .76). The Structure of Blame (RQ1a and RQ1b) To compare the four conceptual models of blame, punishment was used as the outcome variable in the model. To see whether the analyses should be done on the combined data across the two health risks (asthma attack and heart attack), I compared the key variables across the two subsets. As can be seen in Table 7, there were statistically significant differences on most of the variables. Therefore, the analyses reported below were conducted on each health risk separately. Table 8 (asthma attack data) and Table 9 (heart attack data) present the correlation matrix of the variables in the SEM analyses. To address RQ1a, whether blame consists of only cognition, or affect, or both, I examined the Dual Process model, which subsumed both of the single process models. As explained in Chapter 4, as the Dual Process Model showed significant paths from both cognition and affect (see Figure 15 and Figure 16), that constituted evidence that blame should be conceptualized as both cognition and affect. To examine RQ1b, which asks how the cognitive and affective aspects of blame are combined, the Dual Process model and Intertwined model were compared against each other. For both asthma attack data (Table 10) and heart attack data (Table 11), the Intertwined Process model demonstrated a much better fit. The BIC difference (31.18 and 95 36.46, respectively) was strong evidence in favor of the Intertwined model. Figures 15-18 displays the Dual Process model and the Intertwined Process Model with the standardized parameter estimates for the asthma attack group and the heart attack group. To ensure the data could produce consistent results of model comparisons across each outcome variable, I conducted several sets of SEM analyses using parent-oriented policy support, public-oriented policy support, civic participation, and information sharing behavior as the outcome variables. Each outcome variable was entered into SEM analysis separately. The results of SEM analysis showed that the Intertwined Process models with parent-oriented policy support, public-oriented policy support, civic participation, and information sharing behavior as the outcome variables had a better fit compared to the Dual Process models (Table 12 to Table 19). Consistent with findings from the pilot study, therefore, the data from the main study showed that blame should be conceptualized as a combination of cognition and affect, and its effect on subsequent outcome variables cannot be separated. Antecedents of Blame (Hypotheses 1a to 1c) The results from the final Intertwined Process models (Figure 17 and Figure 18) showed that outcome severity did not predict blame for the heart attack group, b = -.06, p > .05, but significantly predicted blame for the asthma attack group, b = .07, p < .05. Therefore, H1a was partially supported. Outcome preventability was a significant predictor of blame for both the heart attack data (b = .21, p < .001) and the asthma attack data (b = .18, p < .001). H1b was supported. The interaction effect between outcome severity and preventability on blame was significant for both heart attack (b =. 08, p < .001) and asthma 96 attack (b =. 07, p < .001). H1c was supported. When the outcome was regarded as preventable, participants in severe outcome condition experienced an even higher level of blame than those in nonsevere outcome condition. Outcomes of Blame (H2 to H5) The second set of hypotheses were proposed to explore the outcomes of blame. Figure 19 and Figure 20 display the standardized path coefficients for the Intertwined model with five outcomes of the asthma attack group and the heart attack group. On policy support (H2), the results of SEM analysis showed that blame positively predicted parentoriented policy support, b = .48, p < .001, and socioenvironmental policy support, b = .25, p < .01, for the asthma attack data. For heart attack data, blame positively predicted parentoriented policy support, b = .39, p < .001, but was not associated with socioenvironmental policy support, b = .18, p > .05. These results largely supported H2. For the relationship between blame and information sharing (H3), it was significant for both the asthma attack data, b = .29, p < .01, and the heart attack data, b = .49, p < .001. H3 was supported. Blame also had a significant effect on civic participation for both asthma attack (b = .22, p < .05) and heart attack (b = .29, p < .01) data, supporting H4. H5 predicted the effect of blame on punishment for parents. The path was significant for the asthma attack data, b = .82, p < .001, as well as for the heart attack data, b = .83, p < .001. Hence, H5 was supported. Results on indirect effects also supported blame as a mediating construct between message characteristics and social responses. Table 20 and Table 21 list the indirect effect from outcome preventability and outcome severity to each of the social response variables 97 for both data. As can be seen, blame was a significant mediator between preventability and most of the social response variables but not a significant mediator between the severity of the consequence and all the social response variables. Moderating Effects (RQ2a and RQ2b) The interaction effect between each of the moderators and the antecedents of blame on blame was tested using Mplus. Using XWITH command in Mplus, I further tested the interaction effects between each of the moderators and the latent construct of blame on each of the social response variables. It was only found that internalization of moral identity moderated the effect of the severity of consequence on blame. Among all the social response variables, there were moderating effect on parent-oriented policy support and information sharing behavior. I will report the significant findings below. Need for cognition. The findings showed that interaction between blame and need for cognition had a negative effect on parent-related policy support, B2 = -.17, p < .05. In other words, the effect of blame on parent-related policy support was weaker for those with higher need for cognition. It was also found that emotion approach moderated the effect of blame on information sharing such that the effect of blame on information sharing was stronger for participants who were high in emotion approach, B = .27, p < .05. This interaction effect did not emerge for the heart attack data. Need for affect. The two subscales of NFA, emotion avoidance and emotion approach, were analyzed separately. For asthma attack data, emotion avoidance moderated the effect of blame on parent-oriented public policy support, B = -.17, p < .05. The effect 2 Mplus produced unstandardized coefficient B for latent variable interaction. 98 of blame on parent-oriented policy support decreased for those who were inclined to avoid emotion. Again, this interaction effect was not present in the heart attack data. It was found that emotion approach moderated the effect of blame on information sharing behavior, B = 27, p < .05. The effect of blame on information sharing increased for those who were inclined to approach emotions. Moral identity. Moral identity has two subscales: symbolization and internalization. For the asthma attack data, it was found that the severity of consequence interacted with internalization of moral identity to influence blame such that the effect of the severity of consequence on blame was stronger for those who were high in internalization, b = 1.48, p < .05. For the asthma attack data, there was a significant interaction effect between symbolization and blame on parent-oriented policy support, B = -.24, p < .01. Blame, therefore, had less effect on parent-oriented policy support among those who were more likely to convey visible moral traits. Summaries of the Main Study Building upon the pilot study, the main study was set up with five goals: (1) to uncover the nature of blame by comparing four proposed models, (2) to examine the effect of antecedents of blame, (3) to explore the social responses that are associated with blame, (4) to examine the moderating effect of individual trait variables on the relationship between the antecedents of blame and blame, and (5) to examine the moderating effect of individual trait variables on the relationship between blame and four social responses. For the main study, both manipulations of the severity of consequence and preventability were successful. Among the four proposed models, the Intertwined Process model, which 99 conceptualizes blame as an alloy of cognition and affect, was significantly better than alternative models. Regarding the antecedents, outcome preventability, severity, and their interaction term were all significant predictors of blame in asthma attack data; and preventability and the interaction effect were significant for the heart attack data. Lastly, blame was found to predict both prosocial (policy support, civic participation, and information sharing) and punitive (punishment) behavioral intentions. The results of moderation analyses revealed a few moderating effects: internalization moderated the effect of the severity of consequence on blame; symbolization subscale of moral identity, emotion avoidance, and need for cognition all interacted with blame to influence parentoriented policy support; and emotion approach interacted with blame to influence information sharing behavior. 100 Table 7 Results of Mean, Standard Deviations, and T-test of Heart Attack and Asthma Attack Group M/SD (Heart Attack) 4.16(.73) M/SD (Asthma Attack) 3.95(.74) 4.31`(.73) 3.93(.73) 4.19(.81) 4.11(.77) Affect 3.2(1.12) 2.85(1.09) Punishment 3.23(1.27) 2.84(1.28) Parent-oriented Policy Support Public-oriented Policy Support Civic Participation 5.26(1.43) 5.17(1.50) 4.14(.82) 4.02(.75) 2.85(1.10) 2.61(1.19) Information Sharing 3.30(1.10) 2.92(1.20) Perceived Level of Preventability Perceived Level of Outcome Severity Cognition T-test t(371)=2.76, p <.05 t(371)=4.52, p <.05 t(371)=1.12, p >.05 t(371)=2.87, p <.05 t(371)=3.05, p <.05 t(371)=2.27, p >.05 t(371)=.27, p >.05 t(371)=3.87, p =.05 t(371)=9.23, p <.05 Variable 1.Preventability 2.Severity 3.Preventability By Severity 4.Cognition 5.Affect 6.Punishment 7.Parent-oriented Policy support 8.Public-oriented Policy Support 9.Civic Participation 10.Information Sharing *p< 0.05, **p < 0.01 .57** .18* .11 .11 -.03 .09 -.12 -.11 .03 -.04 .15* .11 1 2 .56** .32** .24** .28** 1 1 -.02 .05 .14* .003 .35** 1 .23** .21** .18** 3 .02 .10 .24 .26** 1 .39** .50** 4 .30 .14* .25** .35** 1 .54** 5 .23 .16* .17* .44** 1 6 Table 8 Zero-Order Correlations Matrix for the Main Study (asthma attack group, N =193). .44** .48** .39** 1 7 .30** 1 .39** 8 .80** 1 9 1 10 101 Variable 1.Preventability 2.Severity 3.Preventability By Severity 4.Cognition 5.Affect 6.Punishment 7. Parent-oriented Policy support 8.Public-oriented Policy Support 9.Civic Participation 10.Information Sharing *p< 0.05, **p < 0.01 .54** -.23 -.18 -.13 .04 1 .08 -.04 -.12 .56** .32** .33** .34** .03 .04 .10 .17* 2 1 1 -.07 1 3 .14 .12 .06 .07 .18* .18* .14 1 4 .06 .04 .06 .19* .42** .53** .47** .16 .35** .30** 1 .59** 5 .40** .09 .28** .33** 1 6 Table 9 Zero-Order Correlation Matrix for the Main Study (heart attack group, N =180) 8 9 10 .41** .57** .76** 1 .53** 1 .47** .53** 1 1 7 102 103 Table 10 Fit Indices and Model Comparison for the Main Study using Punishment as the Outcome (Asthma Attack Group, N =193) a RMSEAa CFIb Chi-square The Dual Process Model The Intertwined Model 32.49(df=4) .18 .92 .09 5.49(df=6) 0 1 .02 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b SRMRc Models BICd 2154.0 2 2122.8 4 BICdifferenc e 31.18 104 Table 11 Fit Indices and Model Comparison for the Main Study Using Punishment as the Outcome (Heart Attack Group, N =180) a Models Chi-square The Dual Process Model The Intertwined Model 38.24(df=4 ) 5.82(df=6) RMSEA CFIb a SRMRc .22 .90 .10 2034.58 0 1 .02 1998.12 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b BICd BICdifferenc e 36.46 105 Table 12 Fit Indices and Model Comparison for the Main Study Using Parent-Oriented Policy Support as the Outcome (Asthma Attack Group, N =193) a RMSEAa CFIb Chi-square The Dual Process Model The Intertwined Model 24.34(df=4) .17 .93 .08 4.93(df=6) 0 1 .03 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b SRMRc Models BICd 1856.4 5 1829.9 8 BICdifferenc e 26.47 106 Table 13 Fit Indices and Model Comparison for the Main Study Using Parent-Oriented Policy Support as the Outcome (Heart Attack Group, N =180) a RMSEAa CFIb BICd Chi-square The Dual Process Model The Intertwined Model 35.90(df=4) .20 .88 .09 2005.44 15.62(df=6) .09 .97 .05 1974.32 31.12 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b SRMRc Models BICdifference 107 Table 14 Fit Indices and Model Comparison for the Main Study Using Public-Oriented Policy Support as the Outcome (Asthma Attack Group, N =193) Models Chi-square The Dual 28.69(df=4) Process Model The 2.76(df=6) Intertwined Model a RMSEAa CFIb SRMRc .18 .91 .08 2297.92 0 1 .02 2263.29 36.63 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b BICd BICdifference 108 Table 15 Fit Indices and Model Comparison for the Main Study Using Public-Oriented Policy Support as the Outcome (Heart Attack Group, N =180) a Models Chi-square The Dual Process Model The Intertwined Model 22.69(df=4 ) 2.72(df=6) RMSE Aa .16 0 CFIb .93 .08 1 .03 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b SRMRc BICd 2098.1 0 2069.5 1 BICdifferenc e 28.59 109 Table 16 Fit Indices and Model Comparison for the Main Study Using Civic Participation as the Outcome (Asthma Attack Group, N =193) a RMSEAa CFIb Chi-square The Dual Process Model The Intertwined Model 29.56(df=4) .18 .92 .08 4.46(df=6) 0 1 .02 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b SRMRc Models BICd 2216.7 9 2182.7 1 BICdifferenc e 34.08 110 Table 17 Fit Indices and Model Comparison for the Main Study Using Civic Participation as the Outcome (Heart Attack Group, N =180) Models The Dual Process Model The Intertwined Model a Chi-square RMSEAa CFIb 26.76(df=4) .18 .92 11.15(df=6) .07 SRMRc BICd BICdifference .08 1998.78 .98 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b .04 1976.73 22.05 111 Table 18 Fit Indices and Model Comparison for the Main Study Using Information Sharing Behavior as the Outcome (Asthma Attack Group, N =193) Models The Dual Process Model The Intertwined Model a Chi-square 8.92(df=4) RMSEAa .09 2.81 (df=6) 0 CFIb .98 1 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b SRMRc .04 .01 BICd BICdifference 2195.38 2185.23 10.15 112 Table 19 Fit Indices and Model Comparison for the Main Study Using Information Sharing Behavior as the Outcome (Heart Attack Group, N =180) Models The Dual Process Model The Intertwined Model a Chi-square RMSEAa 26.60(df=4) .17 14.76(df=6) .09 CFIb .92 .97 RMSEA= Root Mean Square Error Approximation. CFI=Comparative Fit Index c SRMR= Standardized Root Mean Square Residual. d BIC= Bayesian Information Criterion. b SRMRc .09 .05 BICd BICdifference 1964.99 1946.33 18.66 113 Table 20 Indirect Effects From Exogenous Variables to Outcome Variables Through Blame Parentoriented Policy Support Outcome b=0.18 Preventability p < .01 Outcome b = 0.08 Severity p > .05 *Asthma attack data (N = 193) Publicoriented Policy Support b=0.109 p < .05 b = 0.04 p > .05 Information Civic Punishment Sharing Participation b=0.11 p < .05 b = 0.05 p > .05 b=0.08 p > .05 b = 0.04 p > .05 b=3.04 p < .001 b =0.14 p > .05 114 Table 21 Indirect Effects From Exogenous Variables to Outcome Variables Through Blame Parentoriented policy support Outcome b=0.21 Preventability p < .01 Outcome b = -.02 Severity p > .05 *Heart attack data (N = 180) Publicoriented Policy Support b=0.05 p > .05 b = -.01 p > .05 Information Civic Punishment Sharing Participation b=0.21 p < .01 b = -.03 p > .05 b=0.14 p < .01 b = -.48 p > .05 b= 0.35 p < .001 b = -.01 p > .05 115 The Perceived Level of Preventability 4.6 4.4 4.2 4 NonSevere 3.8 Severe 3.6 3.4 Not Preventable Preventable Outcome Preventability Figure 11. Interaction Effect Between Preventability and Outcome Severity on Perceived Level of Outcome Preventability in the Asthma Attack Data (N= 193). The Perceived Level of Outcome Severity 116 4.6 4.4 4.2 4 Not Preventable 3.8 Preventable 3.6 3.4 3.2 Non-Severe Severe The Severity of Consequence Figure 12. Interaction Effect Between Preventability and Outcome Severity on Perceived Level of Outcome Severity in the Asthma Attack Data (N= 193). 117 The Perceived Level of Preventability 4.6 4.4 4.2 4 Non-Severe Severe 3.8 3.6 3.4 Not Preventable Preventable Outcome Preventability Figure 13. Interaction Effect Between Preventability and Outcome Severity on Perceived Level of Outcome Preventability in the Heart Attack Data (N= 180). The Perceived Level of Outcome Severity 118 4.6 4.5 4.4 4.3 4.2 4.1 Not Preventable 4 Preventable 3.9 3.8 3.7 3.6 Non-Severe Severe The Severity of Consequence Figure 14. Interaction Effect Between Preventability and Outcome Severity on Perceived Level of Outcome Severity in the Heart Attack Data (N= 180). 119 Figure 15. The Obtained Dual Process Model for the Asthma Attack Group in the Main Study (N =193).3 3 For both the Dual Process Model and the Intertwined Process Model, preventability (with the preventable condition was coded as 1 and nonpreventable condition was as 0), severity (with the severe condition was coded as 1 and less-severe condition was coded as 0), and the interaction term between preventability and the severity of consequence were specified as latent exogenous variables. The measurement paths for preventability, severity, and interaction term were set to 1.00. In the DualProcess model, cognition and affect were corrected for measurement error by fixing the error term of the corresponding manifest variables to 1- 𝑎 times its variance. In both Dual Process and Intertwined Process Model, punishment was corrected for measurement error fixing the error term of the corresponding manifest variable at 1- 𝑎 times its variance. 120 Figure 16. The Obtained Dual Process Model for the Heart Attack Group in the Main Study (N =180). 121 Figure 17. The Obtained Intertwined Model for the Asthma Attack Group in the Main Study (N =193). 122 Figure 18. The Obtained Intertwined Model for the Heart Attack Group in the Main Study (N =180). 123 Figure 19. Standardized Path Coefficients for the Intertwined Process Model with Five Outcomes for the Asthma Attack Group in the Main Study (N =193). 124 Figure 20. Standardized Path Coefficients for the Intertwined Process Model with Five Outcomes for the Heart Attack Group in the Main Study (N =180). CHAPTER 6 DISCUSSION This dissertation set out to explicate the nature of blame and examine blame as a mediating construct between message features and social responses in the context of childhood obesity. Data were collected from two experimental studies to examine the hypotheses and research questions. News stories about a boy suffering from obesity were manipulated to capture two antecedents of blame: severity and preventability of the consequences. The pilot study, with a convenience sample of college students, was carried out to assess the feasibility of the experimental design. The main study, with modifications informed by the pilot study, was then conducted using participants recruited from MTurk (Mturk.com, 2018). The pilot study focused on one health risk associated with childhood obesity: heart attack. The main study included another obesity-related risk, asthma attack, in addition to heart attack (as a between-participants factor), to test the replicability of the findings. Summary of Key Findings This dissertation sets out to investigate four major sets of questions: (1) the underlying structure of blame as a psychological response, (2) whether and how preventability and the severity of consequence featured a negative event, as two theorized 126 antecedents induce blame to different degrees, (3) social responses that follow elicited blame, and (4) the potential moderating effects of individual difference variables. Table 22 summarizes the results of examining the hypotheses and research questions in the pilot study and the main study. The key findings are discussed below. The Structure of Blame (RQ1a & RQ1b) Based on the existing conceptualizations of blame, this dissertation proposes to test four models of blame: two single-process models with blame either as cognitive or affective, a dual process model in which blame consists of cognitive and affective components as separable, parallel paths, and an intertwined model where blame is an amalgam of both cognition and affect. In this project, the cognitive aspect of blame was measured as attribution of responsibility, and the affective aspect was represented by moral indignation. SEM analyses using Mplus for model comparison supported the Intertwined Process model in both the pilot and the main studies. However, it should be noted that the findings of the pilot study were only suggestive, as the Dual Process model and the Intertwined Process model could not provide direct evidence that blame constitutes both cognitive and affective components. Such results might be due to the ineffectiveness of experimental manipulation or lack of statistical power. In the main study, the model with blame as a latent variable consisting of both cognitive and affective components showed a substantially better fit than other alternative models. The factor loadings for cognition and affect were similar in the main study, suggesting that cognition and affect contributed similarly to the underlying factor of blame. This finding supported the theoretical account of blame as a blend of cognition and affect (e.g., Weiner, 1995). 127 Antecedents of Blame (H1a to H1c) Among the other antecedents of blame suggested by the previous literature (Alicke, 2000; Malle et al., 2014; Shaver, 1985), I chose to focus on two: the severity of consequence and the preventability of the negative event. Each of them was manipulated to have two levels: severe versus less severe consequences, which could have been more or less likely to be prevented. In the main study, for both heart attack and asthma attack data, there was a significant interaction effect between severity and preventability on blame: Preventability had a stronger positive effect on blame under severe condition as compared to nonsevere condition. In addition, there was also a main effect for preventability: When the consequence was depicted as preventable, participants were more likely to experience higher level of blame. There was a main effect for the severity of consequence in the asthma attack but not heart attack data. Why this main effect was not replicated across the two risk domains was not clear and should be a question for investigation in future research. Social Responses Following Blame (H2 to H5) This study investigated four types of social responses that could have implications for communication research: policy support, punishment, information sharing behavior, and civic participation. After confirming the factor structure of blame, I used findings from the intertwined model, where the outcome variables were theorized to result from the latent construct of blame. To summarize briefly, the pilot study generated rather limited results: blame was a significant predictor of information sharing but was not a predictor of policy support and civic participation. In the main study, it was found that blame was associated with policy support, punishment, information sharing behavior, and civic participation. 128 These findings all attested to the social implications of blame, which could motivate individuals to engage in communicative actions with the goal of collectively resolving a social problem. I will discuss each outcome variable in greater detail below. Policy support. The pilot study looked at policy support that requires paying additional tax or not requires paying additional tax. The main study focused on parentoriented and socioenvironmental policy support. Blame was not found to predict both support for policies that require paying additional tax and support for policies that do not require paying additional tax. The effect of blame on both parent-oriented policy support and socioenvironmental policy support were significant in the main study. For both heart attack and asthma attack, blame led participants to support policies that would require greater parental responsibility in addressing childhood obesity. Participants who blamed parents for causing the child’s obesity were more likely to support the public policies that focusing on the role of parents in preventing childhood obesity. Participants who experienced more blame were also more likely to support policies that focused on the legislative and regulatory actions to control childhood obesity in the asthma attack data, which was interesting and could be have practical importance as the message did not explicitly mention the responsibility of government and society. These findings have implications for message design for the purpose of increasing support for public policies for tackling the issue of childhood obesity. Information sharing behavior. Blame was shown to have a positive impact on sharing information related to childhood obesity. To my knowledge, no exiting research has directly assessed the relationship between blame and information sharing behavior. Previous research has focused on the influence of message characteristics, including source 129 credibility, argument quality, and behavioral predictor such as information seeking, on information sharing behavior (Ha & Ahn, 2011; Yang, Kahlor, & Griffin, 2014). Information sharing behavior is an increasingly important element in the collective effort of improving public health (Sun, 2014; Sun et al., 2016). Civic participation. Blame was found to have a positive effect on civic participation in childhood obesity prevention activities. To my knowledge, blame as the determinant of civic participation has not been studied before. This study provided new insights for understanding and promoting civic participation. The findings of this study were similar to the results from García-Cabrero, Sandoval-Hernández, and Martínez's (2017) study, which found that both affective (empathy) and cognitive responses (understand, critique, discuss and synthesize multiple sources of data) were determinants of civic participation. Civic participation is a complex multidimensional concept, taking various forms, and encompassing a variety of behaviors and actions (Arvanitidis, 2017). This study focused on the more explicitly communicative forms of civic participation (e.g., wear or display a "Childhood Obesity Prevention" badge or sticker, and register an email listserve to receive messages about childhood obesity prevention and treatment). Future research could continue this line of research by analyzing how blame could promote civic participation in other contexts and other forms of civic participation such as protesting and participating in fund-raising activities. Punishment. In addition to the above prosocial behaviors (e.g., policy support, information sharing behavior, and civic participation), blame also influenced retributive behavior (e.g., punishment). Blame was positively associated with punishment to the parents featured in the story. This finding was consistent with previous theoretical 130 propositions that blame increases intention to punish (Fincham & Jaspars, 1980; Weiner, 1985), and that moral indignation is intrinsically lined to punishment (Johnson & Newmeyer, 1975). It is important to note that punishment is a communicative behavior (Bourne, 2014; Wringe, 2017), a way of denouncing the wrongdoer, correcting the wrong, and protecting the violated norm. This dissertation explored blame as a driving force for communicative behaviors, and the above findings showed that blame indeed shape social responses that are integral to communication processes. Information sharing behavior and civic participation, the way they were measured in the study, captured directly communicative means such as sharing information or wearing a badge about preventing childhood obesity. Policy support and intention to punish, though not as explicitly communicative, are ways of expressing one’s attitude and preference about a social issue with implications for shaping public policies. The results of this dissertation contributed to the current communication research by understanding the role of blame in influencing communicative behaviors. Moderator Effects (RQ2a & RQ2b) The effect of moderators in the relationships between the antecedents of blame on blame and the relationships between blame and subsequent social responses were examined. Several important findings have emerged from the asthma attack data. First, internalization of moral identity was identified as a moderator between the relationship of the severity of consequence and blame. Aquino and Reed II (2002) suggested that internalization is often associated with moral outcomes. Because blame is intrinsically linked to the violation of one’s moral standard (Malle, Guglielmo, & Monroe, 2013), it seems justifiable that the 131 effect of the severity of consequence on blame was dependent on internalization of moral identity. Aquino, McFerran, and Laven (2011) indicated that internalization has been positively associated with moral reasoning. It is likely that individuals who are high in internalization devote more cognitive efforts for moral reasoning and thus the effect of the severity of consequence on blame was stronger for those individuals, given that blame requires complicated cognitive processing. In the pilot study, it was found that the effect of preventability on blame depended positively on symbolization. The effect of preventability on blame was stronger for the participants who were high in symbolization is most likely due to the fact that such participants adopted moral traits such as compassionate and caring as their social self-schema and thus were more sensitive to moral violations in a social context. Second, need for cognition was found to be a moderator between the relationship of blame and parent-related policy support. It was found that the effect of blame on parent-related policy support was attenuated by NFC. Lacking theoretical guidance, it is unclear why the relationship between blame and parent-related policy support was weaker for participants who were more likely to engage in more indepth cognitive activities. Third, emotion avoidance, a subscale of need for affect, moderated the effect of blame on support for policies focusing on the role of parents in childhood obesity prevention. Emotion avoidance is an individual’s tendency to avoid emotions. Previous research has found that negative affect was associated with policy support of social issues (Huddy, Feldman, Taber, & Lahav, 2005; Leiserowitz, 2006). People who tend to avoid emotion thus are expected to be less likely to support policies of a particular social issue. It is unsurprising that the relationship between blame and parentoriented public policy support was weaker for participants who tended to avoid emotion. 132 Fourth, emotion approach was found to moderate the effect of blame on information sharing behavior. Emotion approach captures one’s tendency to approach emotional engagement (Maio & Esses, 2001). Considering that individuals who are high in emotion approach tendency have stronger ability to share another person’s feelings, the positive moderating effect of emotion approach between the relationship of blame and information sharing behavior is probably due to the fact that information sharing behavior is a way to transmit one’s empathy. Thus, the effect of blame on information sharing behavior was stronger for participants who tended to explore emotions and share their emotions with others. Fifth, the relationship between blame and punishment was moderated by symbolization, a subscale of moral identity. Symbolization captured one’s social self that can be represented by the symbolic activities and was often positively associated with prosocial behaviors such as volunteering (Winterich et al., 2013). Contradictory to the findings of previous research, it is found that the effect of blame on parent-oriented policy support, which is a prosocial behavior, was weaker for participants who rated high in symbolization. One possible explanation is that compared to the internalization aspect of moral identity, symbolization is a less consistent predictor (Winterich et al., 2013). Interestingly, I did not find any significant moderating effect in the relationship between blame and social responses in the heart attack data. It suggested that the relationships between blame and social responses did not depend on the value of any moderators in the heart attack data. These findings about moderating effect, it should be noted, only excited for the asthma attack data. None of the moderators were significant for the heart attack data. How individual trait variables may influence the blame process, therefore, may be contingent 133 upon the featured health risks, or other differences between the heart attack versus asthma attack message. Future research could provide more explicit theorizing about the moderating roles of relevant individual trait variables, and more systematic empirical investigations. Limitations and Future Directions One limitation of this project was the nonrepresentative samples. The project included two types of samples, one consisting of college students and the other of U.S. adult population recruited from MTurk. Though MTurk provided a more heterogeneous sample, generalizability was still an issue. MTurk samples are different from a telephone survey sample that has been widely used to collect data from the general population (Simons & Chabris, 2012). It has been found that MTurk undersampled the elderly and relied more on self-selection for study participation (Simons & Chabris, 2012). The median age (Mdn=35) of the sample in this study was slightly younger than the median age (Mdn=37.90) of the U.S. population (U.S. Census Bureau, 2017). The MTurk sample in this study was more educated compared to the education level of general the U.S. population. All the respondents had a high school degree or above, and 67% of the respondents had associate degree or above compared to 41.21% of the U.S. population that has an associate degree or above and 11.04 % that do not have high school degree (U.S. Census Bureau, 2017). Another problem with Mturk samples is the prevalence of habitual survey takers. It was found that MTurk data collected between February 2011 and February 2013 showed on average, MTurk respondents reported having participated in 300 research studies (Rand et al., 2014). The fundamental motivation of MTurk participation might be 134 different from other types of samples, based on how often they participate in surveys as many MTurk respondents participate in survey experiments strictly to earn money while respondents such as university-based samples participate in survey experiments might for intrinsic rewards or other nonmonetary reasons (Mullinix, Leeper, Druckman, & Freese, 2015). Future research should use representative samples to increase the generalizability of the findings. Another limitation is the lack of replicability of findings across messages varying in other aspects. To address the limitations of single-message designs, I added to the main study another story featuring a different health risk. It was quite a limited effort, though. Results generated from asthma attack data and heart attack data were different to a certain degree. For example, the severity of consequences was positively associated with blame in the asthma attack data but not in the heart attack data. One possible explanation is that participants perceived a heart attack and asthma attack differently: A heart attack might be considered as a more severe consequence of childhood obesity than an asthma attack. According to the Centers for Disease Control and Prevention, there were 610,000 people that died of heart disease (CDC, 2017a) and 3,388 people that died of asthma in 2009 (CDC, 2017b). Thus, the participants might feel there was no significant difference between severe heart attack and mild heart attack, given that even mild heart attack is a severe consequence of childhood obesity. Both stories were in the format of a text without pictures. Future research may also use other types of experimental stimuli, such as visual or audio messages, or use textual information that is accompanied by graphic materials. How the blame process as investigated in this study may vary across different types of media, or other message 135 characteristics, remains a question for future research. The study also did not include other individual characteristics that could play moderating roles, such as anger accessibility (Schimel, Greenberg, & Martens, 2003), blame accessibility (Meier & Robinson, 2004), preexisting attitude toward childhood obesity, knowledge of the issue of childhood obesity, and so on. These could also potentially exacerbate or dissipate blame. Future research should explicitly incorporate some of these factors in the theoretical framework. For example, blame accessibility, which describes how quick one can access blame-related thoughts (Meier & Robinson, 2004), could influence the effect of preventability on the blame response such that those with a higher level of blame accessibility may be more likely to blame the parents after reading the news story. Controlling for blame accessibility could allow future research to examine the influence of blame-eliciting events on blame without the impact of how quick one could access blame. Even though the results of SEM analysis showed that the Intertwined Process model had better fit indices compared to all the other models and blame thus should be conceptualized as an amalgam of cognition and affect, it still raises the question that whether attribution of responsibility and moral indignation, both of which were produced by the antecedents of blame, occur simultaneously. The self-report measurements of cognition and affect could not provide an affirmative answer to this question. It should be noted that the self-report measurement of moral indignation could not accurately capture the participants’ emotional responses to the experimental stimuli. One weakness of selfreport measurement of affect is that people may not give honest answers when such answers are not socially desirable (Ciuk, Troy, & Jones, 2015). One way to address the 136 weakness of self-report measurements of affect and cognition is to use physiological measures. Recent research suggested that physiological measures are not subject to social desirability bias, and they can capture aspects of emotional response that are beyond respondents' conscious control (Ciuk et al., 2015). Another potential weakness of the measurement of moral indignation used in this study is that the measurement did not emphasize the target to which moral indignation directed. It is thus unclear as to where the moral indignation experienced by participants came from. The findings from this dissertation suggested several directions for future research on blame. First, future research should incorporate different message contexts to further validate the structure of blame and its mediating role between communication messages and behavioral outcomes. Replication of findings using different samples and message stimuli is important to further advancing of this line of research. Second, future studies could explore other potential outcomes of blame within different health contexts, such as drunk driving, smoking, and AIDS. Furthermore, this study only used behavioral intention measures, whereas future research could employ more proximal measures of actual behavior. As one example, researchers could embed a Twitter widget in the online survey so that the participants could actually choose to retweet or not. Third, future studies could consider incorporating visual stimuli. Visuals are effective tools for articulating ideological messages (Messaris & Abraham, 2001) and constitute additional storytelling (Gibson & Zillmann, 2000). Previous research suggested that visual elements influenced participants’ attribution of responsibility (Ben-Porath & Shaker, 2010) and emotional responses (Bucy & Newhagen, 1999). Future research could explore how certain visual features may be more or less effective in eliciting blame and subsequent communicative behaviors. 137 This dissertation has several important practical implications on addressing the issue of obesity. First, it offers insights on the message strategies that can be used for obesity prevention interventions. As shown in the findings of this study, blame positively predicted several prosocial behaviors, including policy support, civic participation, and information sharing behavior. Health practitioners should consider eliciting other-blame (e.g., blaming fast food restaurants for causing childhood obesity) through media messages as a way to engage the public in combating obesity epidemic and eliciting self-blame (e.g., blaming self for eating excessive junk food) as a strategy of enhancing self-regulation for obesity prevention. Since the public have limited knowledge about obesity, the presence of blame in media messages not only could help the public to understand the causes and consequences of obesity but also guide them to figure out ways to address the issue of obesity. Another practical implication emerged from this research is that as a predictor of information sharing behavior, blame could be utilized to raise the public awareness of obesity by launching social media campaigns that are embedded with the elements of blame, such as emphasizing the severe consequence of obesity in the campaign. Such social media campaigns could get messages out to a large number of people and motivate them to share obesity prevention-related information on social media. It worth noting that when health practitioners incorporate the elements of blame in health campaigns, they should craft the media messages in an unbiased way without shaming and stigmatizing people. Conclusion Conceiving of blame as an important construct for studying social responses in communication research, this dissertation takes up the task of explicating the nature of 138 blame in the context of childhood obesity. The current dissertation delved into the underlying structure of blame and has extended the scope of the current research on blame. It also developed a conceptual argument about the role of blame in shaping social responses. Two experimental studies were carried out to explore the nature of blame, the process of how blame occurred, and the social responses that were associated with blame. The key finding drawn from the current study is that blame should be conceptualized and measured as an amalgam of cognition and affect. This dissertation was an important step toward developing a comprehensive measure of blame within the communication literature. The results of SEM analysis showed that the severity of consequence, preventability, and their interaction term were positively associated with blame in the asthma attack data. In the heart attack data, it was found that preventability and the interaction term of preventability and the severity of consequence were significant predictors of blame. Blame was found to be a predictor of several social responses, including policy support, punishment, information sharing behavior, and civic participation. From a communication perspective, this study contributed to the understanding of blame as an important facilitator of communicative behaviors. The findings have practical implications for researchers and practitioners in promoting citizen involvement in various social issues. 139 Table 22 Research Questions and Hypotheses, and Summary of Findings RQ1a: Should blame be conceptualized and operationalized as a cognition, affect, or both? RQ1b: If blame has both cognitive and affective components, how are the two aspects combined? Pilot study: For RQ1a, there is no conclusive evidence about the nature of blame. For RQ1b, the Intertwined model where blame is a latent construct consisting of both causal attribution and moral indignation received the most support. Main study: RQ1: Both the Dual Process Model and the Intertwined Process Model suggested that blame has both affective and cognitive components. The Intertwined Process model was superior (both asthma attack and heart attack data). H1a: Outcome severity positively predicts blame. Pilot study: Not supported Main study: Supported for the asthma attack data but not for heart attack data. H1b: Outcome preventability positively predicts blame. Pilot study: Supported. Main study: Supported for both asthma attack and heart attack data. H1c: There is an interaction effect between outcome severity and preventability on blame. Pilot study: Not supported. Main study: Supported for both asthma attack and heart attack data. H2: Blame positively predicts support for public policies addressing the problem of childhood obesity. Pilot study: On policies not requiring tax: Not supported. Main study: On policies focusing on the role of parents in childhood obesity prevention: Supported. On policies improving social environment: Supported for asthma attack data. H3: Blame positively predicts information sharing behavior about childhood obesity. Pilot study: Supported. Main study: Supported for both asthma attack and heart attack data. H4: Blame positively predicts civic participation behaviors for preventing childhood obesity. Pilot Not supported. study: Main study: Supported for both asthma attack and heart attack data. H5: Blame positively predicts the tendency to punish parents of the obese child. Pilot study: Not included as the hypothesis in the pilot study. Main study: Supported for both asthma attack and heart attack data. 140 Table 22 Continued RQ2a: Will the effects of outcome severity and outcome preventability specified in H1 be contingent upon individual traits, including (a) need for affect, (b) need for cognition, and (c) moral identity? RQ2b: Will the effects of blame on social responses specified in H2-H5 be contingent upon individual traits, including (a) need for affect and (b) moral identity? Pilot study: Symbolization of moral identity moderated the effect of preventability on blame. None of the moderators had a significant effect on the relationship between blame and social responses. Main study: (a) There was a positive interaction effect between internalization and the severity of consequence on blame. (b) There was a negative interaction effect between emotion avoidance and blame on the policies focusing on the role of parents in preventing childhood obesity and there was a positive interaction between emotion approach and blame on information sharing behavior in the asthma attack data. (c) There was a negative interaction effect between symbolization and blame on parent-oriented policies in the asthma attack data. (d) There was a negative interaction effect between need for cognition and blame on parent-oriented policies in the asthma attack data. APPENDIX A STIMULUS 1A: HIGH PREVENTABILITY*SEVERE CONSEQUENCE APPENDIX B STIMULUS 2A: LOW PREVENTABILITY*MILD CONSEQUENCE APPENDIX C STIMULUS 3: HIGH PREVENTABILITY*MILD CONSEQUENCE APPENDIX D STIMULUS 4: HIGH PREVENTABILITY*SEVERE CONSEQUENCE APPENDIX E STIMULUS 1A: HIGH PREVENTABILITY*SEVERE CONSEQUENCE （ASTHMA ATTACK） APPENDIX F STIMULUS 1B: HIGH PREVENTABILITY*SEVERE CONSEQUENCE （HEART ATTACK） APPENDIX G STIMULUS 2A: LOW PREVENTABILITY*MILD CONSEQUENCE （ASTHMA ATTACK） APPENDIX H STIMULUS 2B: LOW PREVENTABILITY*MILD CONSEQUENCE （HEART ATTACK） APPENDIX I STIMULUS 3A: HIGH PREVENTABILITY*SEVERE CONSEQUENCE （ASTHMA ATTACK） APPENDIX J STIMULUS 3B: HIGH PREVENTABILITY*SEVERE CONSEQUENCE （HEART ATTACK） APPENDIX K STIMULUS 4A: LOW PREVENTABILITY*SEVERE CONSEQUENCE （ASTHMA ATTACK） APPENDIX L STIMULUS 4B: LOW PREVENTABILITY*SEVERE CONSEQUENCE （HEART ATTACK） REFERENCES Abrams, D., Viki, G. T., Masser, B., & Bohner, G. (2003). Perceptions of stranger and acquaintance rape: The role of benevolent and hostile sexism in victim blame and rape proclivity. Journal of Personality and Social Psychology, 84, 111-125. Affleck, G., Allen, D. A., McGrade, B. J., & McQueeney, M. (1982). Maternal causal attributions at hospital discharge of high-risk infants. American Journal of Mental Deficiency, 86, 757-580. Alicke, M. D. (2008). Blaming badly. Journal of Cognition and Culture, 8, 179–186. Alicke, M. D., Buckingham, J., Zell, E., & Davis, T. (2008). Culpable control and counterfactual reasoning in the psychology of blame. Personality and Social Psychology Bulletin, 34, 1371-1381. An, S.-K. (2011). Reducing anger and blame: The role of the morality news frame and crisis response strategy. Public Relations Review, 37, 169–171. Anderson, N. (2014). Unified social cognition. New York, NY: Psychology Press. Anzman, S. L., Rollins, B. Y., & Birch, L. L. (2010). Parental influence on children’s early eating environments and obesity risk: Implications for prevention. International Journal of Obesity, 34, 1116–1124. https://doi.org/10.1038/ijo.2010.43 Appel, M., & Richter, T. (2010). Transportation and need for affect in narrative persuasion: A mediated moderation model. Media Psychology, 13, 101–135. https://doi.org/10.1080/15213261003799847 Aquino, K., McFerran, B., & Laven, M. (2011). Moral identity and the experience of moral elevation in response to acts of uncommon goodness. Journal of Personality and Social Psychology, 100, 703-718. Aquino, K., & Reed II, A. (2002). The self-importance of moral identity. Journal of Personality and Social Psychology, 83, 1423-1440. 154 Aquino, K., Tripp, T. M., & Bies, R. J. (2001). How employees respond to personal offense: The effects of blame attribution, victim status, and offender status or revenge and reconciliation in the workplace. Journal of Applied Psychology, 86, 52-59. Arvanitidis, P. A. (2017). The concept and determinants of civic engagement. Human Affairs, 27, 252–272. https://doi.org/10.1515/humaff-2017-0022 Averill, J. R. (1983). Studies on anger and aggression: Implications for theories of emotion. American Psychologist, 38, 1145-1160. Ballard, E. D., Patel, A. B., Ward, M., & Lamis, D. A. (2015). Future disposition and suicidal ideation: Mediation by depressive symptom clusters. Journal of Affective Disorders, 170, 1–6. https://doi.org/10.1016/j.jad.2014.08.029 Bandura, A. (1999). Moral disengagement in the perpetration of inhumanities. Personality and Social Psychology Review, 3, 193–209. Barker-Collo, S. L. (2001). Adult reports of child and adult attributions of blame for childhood sexual abuse: Predicting adult adjustment and suicidal behaviors in females. Child Abuse & Neglect, 25, 1329–1341. Barry, C. L., Brescoll, V. L., Brownell, K. D., & Schlesinger, M. (2009). Obesity metaphors: How beliefs about the causes of obesity affect support for public policy. Milbank Quarterly, 87, 7–47. https://doi.org/10.1111/j.14680009.2009.00546.x Barry, C. L., Brescoll, V. L., & Gollust, S. E. (2013). Framing childhood obesity: How individualizing the problem affects public support for prevention. Political Psychology, 34, 327–349. https://doi.org/10.1111/pops.12018 Bartsch, A., Appel, M., & Storch, D. (2010). Predicting emotions and meta-emotions at the movies: The role of the need for affect in audiences’ experience of horror and drama. Communication Research, 37, 167–190. Bauman, C. W., & Mullen, E. (2014). Reconsidering motivation to blame and the distinction between private and public blame. Psychological Inquiry, 25, 197– 200. Bayles, M. D. (1982). Character, purpose, and criminal responsibility. Law and Philosophy, 1, 5–20. Bell, M. (2013). The standing to blame: A critique. In D. J. Coates & N. A. Tognazzini (Eds), Blame: Its nature and norms (pp.263-281). New York, NY: Oxford University Press, 155 Bennett, C. (2008). The apology ritual: A philosophical theory of punishment. Cambridge, UK: University Press. Ben-Porath, E. N., & Shaker, L. K. (2010). News images, race, and attribution in the wake of Hurricane Katrina. Journal of Communication, 60, 466–490. Bentler, P. . (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238–246. https://doi.org/10.1037/0033-2909.107.2.238 Berger, J., & Milkman, K. L. (2012). What makes online content viral? Journal of Marketing Research, 49, 192–205. Berns, N. (2001). Degendering the problem and gendering the blame: Political discourse on women and violence. Gender & Society, 15, 262–281. Berrenberg, J. L., Rosnik, D., & Kravcisin, N. J. (1990). Blaming the victim: When disease-prevention programs misfire. Current Psychology, 9, 415–420. https://doi.org/10.1007/BF02687197 Besharat, M. A., Eisler, I., & Dare, C. (2001). The self- and other-blame scale (SOBS). The background and presentation of a new instrument for measuring blame in families. Journal of Family Therapy, 23, 208–223. https://doi.org/10.1111/14676427.00179 Boero, N. (2013). Obesity in the media: Social science weighs in. Critical Public Health, 23, 371–380. https://doi.org/10.1080/09581596.2013.783686 Bourne, R. (2014). Communication, punishment, and virtue. Journal of Religious Ethics, 42, 78–107. Boyle, B. A., Dahlstrom, R. F., & Kellaris, J. J. (1998). Points of reference and individual differences as sources of bias in ethical judgments. Journal of Business Ethics, 17, 517–525. Bradfield, M., & Aquino, K. (1999). The effects of blame attributions and offender likableness on forgiveness and revenge in the workplace. Journal of Management, 25, 607–631. Branscombe, N. R., Wohl, M. J., Owen, S., Allison, J. A., & N’gbala, A. (2003). Counterfactual thinking, blame assignment, and well-being in rape victims. Basic and Applied Social Psychology, 25, 265–273. Browne, M. W., & Cudeck, R. (1992). Alternative ways of assessing model fit. Sociological Methods & Research, 21, 230–258. Bucy, E. P., & Newhagen, J. E. (1999). The emotional appropriateness heuristic: 156 Processing televised presidential reactions to the news. Journal of Communication, 49(4), 59–79. Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s Mechanical Turk: A new source of inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6, 3–5. https://doi.org/10.1177/1745691610393980 Bulman, R. J., & Wortman, C. B. (1977). Attributions of blame and coping in the" real world": Severe accident victims react to their lot. Journal of Personality and Social Psychology, 35, 351-363. Burger, J. M. (1981). Motivational biases in the attribution of responsibility for an accident: A meta-analysis of the defensive-attribution hypothesis. Psychological Bulletin, 90, 496–512. https://doi.org/10.1037/0033-2909.90.3.496 Burgess, K. B., Wojslawowicz, J. C., Rubin, K. H., Rose-Krasnor, L., & Booth-LaForce, C. (2006). Social information processing and coping strategies of shy/withdrawn and aggressive children: Does friendship matter? Child Development, 77, 371– 383. https://doi.org/10.1111/j.1467-8624.2006.00876.x Cacioppo, J. T., Petty, R. E., Feinstein, J. A., & Jarvis, W. B. G. (1996). Dispositional differences in cognitive motivation: The life and times of individuals varying in need for cognition. Psychological Bulletin, 119, 197-253. Campbell, R., Dworkin, E., & Cabral, G. (2009). An ecological model of the impact of sexual assault on women’s mental health. Trauma, Violence, & Abuse, 10, 225– 246. Caplan, N., & Nelson, S. D. (1973). On being useful: The nature and consequences of psychological research on social problems. American Psychologist, 28, 199-211. Captcha. (2018). reCAPTCHA: Easy on humans, hard on bots. Retrieved from https://www.google.com/recaptcha/intro/ Carmona, R. (2004, March 2). HHS - Office of the Surgeon General - speeches and presentations. Retrieved from https://www.surgeongeneral.gov/news/testimony/childobesity03022004.html Cascardi, M., & O’Leary, K. D. (1992). Depressive symptomatology, self-esteem, and self-blame in battered women. Journal of Family Violence, 7, 249–259. Cawley, J. (2010). The economics of childhood obesity. Health Affairs, 29, 364–371. CDC. (2015). Childhood obesity facts. Retrieved from https://www.cdc.gov/healthyschools/obesity/facts.htm 157 CDC. (2017a). Heart disease facts & statistics. Retrieved from https://www.cdc.gov/heartdisease/facts.htm CDC. (2017b). Asthma - about asthma. Retrieved from https://www.cdc.gov/asthma/faqs.htm Chapple, A., Ziebland, S., & McPherson, A. (2004). Stigma, shame, and blame experienced by patients with lung cancer: Qualitative study. BMJ, 328, 1470. https://doi.org/10.1136/bmj.38111.639734.7C Cho, H., & Park, B. (2014). Testing the moderating role of need for cognition in smartphone adoption. Behaviour & Information Technology, 33, 704–715. https://doi.org/10.1080/0144929X.2013.825643 Cho, S. H., & Gower, K. K. (2006). Framing effect on the public’s response to crisis: Human interest frame and crisis type influencing responsibility and blame. Public Relations Review, 32, 420–422. https://doi.org/10.1016/j.pubrev.2006.09.011 Chockler, H., & Halpern, J. Y. (2004). Responsibility and blame: A structural-model approach. Journal of Artificial Intelligence Research, 22, 93–115. Choi, I., Nisbett, R. E., & Norenzayan, A. (1999). Causal attribution across cultures: Variation and universality. Psychological Bulletin, 125, 47-63. Christy, A. (2016). The indirect effect of social support on PTSD through self-blame in sexual assault survivors and the moderating role of gender (Master thesis). Northern Illinois University, Illinois, the United States. Retrieved from https://search.proquest.com/docview/1824369529/abstract/9403D1E0F26B46DC PQ/1 Ciuk, D., Troy, A., & Jones, M. (2015). Measuring emotion: Self-reports vs. physiological indicators. Rochester, NY: Social Science Research Network. Retrieved from https://papers.ssrn.com/abstract=2595359 Critchlow, B. (1985). The blame in the bottle: Attributions about drunken behavior. Personality and Social Psychology Bulletin, 11, 258–274. Crossman, A., Anne Sullivan, D., & Benin, M. (2006). The family environment and American adolescents’ risk of obesity as young adults. Social Science & Medicine, 63, 2255–2267. https://doi.org/10.1016/j.socscimed.2006.05.027 Crystal, D. S., & DeBell, M. (2002). Sources of civic orientation among American youth: Trust, religious valuation, and attributions of responsibility. Political Psychology, 23, 113–132. Curry, T. R., & Corral-Camacho, G. (2008). Sentencing young minority males for drug 158 offenses: Testing for conditional effects between race/ethnicity, gender and age during the US war on drugs. Punishment & Society, 10, 253–276. Cushman, F. (2008). Crime and punishment: Distinguishing the roles of causal and intentional analyses in moral judgment. Cognition, 108, 353–380. Darley, J. M., & Shultz, T. R. (1990). Moral rules: Their content and acquisition. Annual Review of Psychology, 41, 525–556. Davison, K. K., Jurkowski, J. M., Li, K., Kranz, S., & Lawson, H. A. (2013). A childhood obesity intervention developed by families for families: Results from a pilot study. The International Journal of Behavioral Nutrition and Physical Activity, 10, 3. https://doi.org/10.1186/1479-5868-10-3 Deffenbacher, J. L. (2011). Cognitive-behavioral conceptualization and treatment of anger. Cognitive and Behavioral Practice, 18, 212–221. https://doi.org/10.1016/j.cbpra.2009.12.004 Dehghan, M., Akhtar-Danesh, N., & Merchant, A. T. (2005). Childhood obesity, prevalence, and prevention. Nutrition Journal, 4, 24. Desjardins, E., & Schwartz, A. L. (2007). Collaborating to combat childhood obesity. Health Affairs, 26, 567–571. Devooght, K. (2004). On responsibility-sensitive egalitarian ethics. Ethics and Economics, 2, 1-21. Dickerson, S. S., Gruenewald, T. L., & Kemeny, M. E. (2004). When the social self is threatened: Shame, physiology, and health. Journal of Personality, 72, 1191– 1216. https://doi.org/10.1111/j.1467-6494.2004.00295.x Dillard, J. P., & Lijiang Shen. (2005). On the nature of reactance and its role in persuasive health communication. Communication Monographs, 72, 144–168. https://doi.org/10.1080/03637750500111815 Duncan, D. T., Hansen, A. R., Wang, W., Yan, F., & Zhang, J. (2015). Change in misperception of child’s body weight among parents of American preschool children. Childhood Obesity, 11, 384–393. https://doi.org/10.1089/chi.2014.0104 Ebbeling, C. B., Pawlak, D. B., & Ludwig, D. S. (2002). Childhood obesity: Publichealth crisis, common sense cure. The Lancet, 360, 473–482. Ellis, A. (1962). Reason and emotion in psychotherapy. Oxford, UK: Lyle Stuart. Eyler, A. A., & Brownson, R. C. (2016). The power of policy to improve health. In S. Moreland-Russel & R. C. Brownson (Eds), Prevention, policy, and public health. 159 New York, NY: Oxford University Press. Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behavior Research Methods, 41, 1149–1160. https://doi.org/10.3758/BRM.41.4.1149 Featherstone, B. (2003). Family Life and family support: A feminist analysis. Basingstoke, UK: Palgrave Macmillan. Feigenson, N., & Park, J. (2006). Emotions and attributions of legal responsibility and blame: A research review. Law and Human Behavior, 30, 143-161. Feiring, C., Taska, L., & Lewis, M. (1996). A process model for understanding adaptation to sexual abuse: The role of shame in defining stigmatization. Child Abuse & Neglect, 20, 767–782. Fincham, F. D., & Bradbury, T. N. (1987). The assessment of marital quality: A reevaluation. Journal of Marriage and the Family, 49, 797–809. Fincham, F. D., & Jaspars, J. M. (1980). Attribution of responsibility: From man the scientist to man as lawyer. Advances in Experimental Social Psychology, 13, 81– 138. Finkelstein, E. A., Trogdon, J. G., Cohen, J. W., & Dietz, W. (2009). Annual medical spending attributable to obesity: Payer-and service-specific estimates. Health Affairs, 28, w822–w831. https://doi.org/10.1377/hlthaff.28.5.w822 Fleishman, J. L., & Payne, B. L. (1980). Ethical dilemmas and the education of policymakers. Hastings on Hudson, NY: The Hastings Center. Fogelholm, M., Nuutinen, O., Pasanen, M., Myöhänen, E., & Säätelä, T. (1999). Parent– child relationship of physical activity patterns and obesity. International Journal of Obesity, 23(12), 1262-1268. Fragale, A. R., Rosen, B., Xu, C., & Merideth, I. (2009). The higher they are, the harder they fall: The effects of wrongdoer status on observer punishment recommendations and intentionality attributions. Organizational Behavior and Human Decision Processes, 108, 53–65. https://doi.org/10.1016/j.obhdp.2008.05.002 Frieden, T. R., Dietz, W., & Collins, J. (2010). Reducing childhood obesity through policy change: Acting now to prevent obesity. Health Affairs, 29, 357–363. Friedlander, M. L., Heatherington, L., & Marrs, A. L. (2000). Responding to blame in family therapy: A constructionist/narrative perspective. American Journal of Family Therapy, 28, 133–146. 160 Furedi, F. (2013). Childhood at risk: How children became so precious. In F. Furedi (Eds), Moral crusades in an age of mistrust: The Jimmy Savile scandal (pp. 40– 52). London, UK: Palgrave Pivot. https://doi.org/10.1057/9781137338020_4 García-Cabrero, B., Sandoval-Hernández, A., & Martínez, M. G. P. (2017). Affective and cognitive processes as determinants of civic participation in Latin American countries. In B. García-Cabrero, A. Sandoval-Hernández, E. Treviño-Villareal, S.D. Ferráns, & M. G. P. Martínez (Eds), Civics and citizenship: Theoretical models and experiences in Latin America (pp. 129–153). Rotterdam, The Netherlands: SensePublishers. https://doi.org/10.1007/978-94-6351-068-4_7 Garnefski, N., & Kraaij, V. (2007). The cognitive emotion regulation questionnaire. European Journal of Psychological Assessment, 23, 141–149. Gerber, G. L., Cronin, J. M., & Steigman, H. J. (2004). Attributions of blame in sexual assault to perpetrators and victims of both genders. Journal of Applied Social Psychology, 34, 2149–2165. Gerstenberg, T., & Lagnado, D. A. (2010). Spreading the blame: The allocation of responsibility amongst multiple agents. Cognition, 115, 166–171. Gibson, R., & Zillmann, D. (2000). Reading between the photographs: The influence of incidental pictorial information on issue perception. Journalism & Mass Communication Quarterly, 77, 355–366. Gilbert, M. (2006). Who’s to blame? Collective moral responsibility and its implications for group members. Midwest Studies in Philosophy, 30, 94–114. Gilmore, J., Meeks, L., & Domke, D. (2013). Why do (we think) they hate us: AntiAmericanism, patriotic messages, and attributions of blame. International Journal of Communication, 7, 701–721. Godlee, F., Smith, J., & Marcovitch, H. (2011). Wakefield’s article linking MMR vaccine and autism was fraudulent. BMJ, 342, c7452. https://doi.org/10.1136/bmj.c7452 Goodall, C. E., & Reed, P. (2013). Threat and efficacy uncertainty in news coverage about bed bugs as unique predictors of information seeking and avoidance: An extension of the EPPM. Health Communication, 28, 63–71. https://doi.org/10.1080/10410236.2012.689096 Grinberg, E. (2012, February 7). Georgia’s child obesity ads aim to create movement out of controversy. CNN, Retrieved from https://www.cnn.com/2012/02/07/health/atlanta-child-obesity-ads/index.html Gruber, K. J., & Haldeman, L. A. (2009). Using the family to combat childhood and adult obesity. Preventing Chronic Disease, 6(3): A106. 161 Gudjonsson, G. H., & Singh, K. K. (1989). The revised Gudjonsson blame attribution inventory. Personality and Individual Differences, 10, 67–70. https://doi.org/10.1016/0191-8869(89)90179-7 Gurin, P., Miller, A. H., & Gurin, G. (1980). Stratum identification and consciousness. Social Psychology Quarterly, 43, 30–47. Ha, S. and Ahn, J. (2011). Why are you sharing others’ tweets?: The impact of argument quality and source credibility on information sharing behavior. Paper presented at Proceedings of the International Conference on Information Systems, Shanghai, China. Retrieved from https://aisel.aisnet.org/icis2011/proceedings/humanbehavior/4/ Haddock, G., Maio, G. R., Arnold, K., & Huskinson, T. (2008). Should persuasion be affective or cognitive? The moderating effects of need for affect and need for cognition. Personality and Social Psychology Bulletin, 34, 769–778. Haidt, J. (2003). Elevation and the positive psychology of morality. In C.L. M. Keyes & J.Haidt (Eds.), Flourishing: Positive psychology and the life well-lived (pp. 275289). Washington DC: American Psychological Association. Hall, S., French, D. P., & Marteau, T. M. (2003). Causal attributions following serious unexpected negative events: A systematic review. Journal of Social and Clinical Psychology, 22, 515–536. Hameleers, M., Bos, L., & Vreese, C. H. de. (2017). “They Did It”: The effects of emotionalized blame attribution in populist communication. Communication Research, 44, 870–900. https://doi.org/10.1177/0093650216644026 Hammond, E. M., Berry, M. A., & Rodriguez, D. N. (2011). The influence of rape myth acceptance, sexual attitudes, and belief in a just world on attributions of responsibility in a date rape scenario. Legal and Criminological Psychology, 16, 242–252. https://doi.org/10.1348/135532510X499887 Han, J. C., Lawlor, D. A., & Kimm, S. Y. (2010). Childhood obesity. The Lancet, 375, 1737–1748. https://doi.org/10.1016/S0140-6736(10)60171-7 Harris, J. L., Pomeranz, J. L., Lobstein, T., & Brownell, K. D. (2009). A crisis in the marketplace: How food marketing contributes to childhood obesity and what can be done. Annual Review of Public Health, 30, 211–225. Harth, N. S., Leach, C. W., & Kessler, T. (2013). Guilt, anger, and pride about in-group environmental behaviour: Different emotions predict distinct intentions. Journal of Environmental Psychology, 34, 18–26. Harvey, M. D., & Rule, B. G. (1978). Moral evaluations and judgments of responsibility. 162 Personality and Social Psychology Bulletin, 4, 583–588. Hasell, A., & Weeks, B. E. (2016). Partisan provocation: The role of partisan news use and emotional responses in political information sharing in social media. Human Communication Research, 42, 641–661. https://doi.org/10.1111/hcre.12092 Hassan, Z. A., Schattner, P., & Mazza, D. (2006). Doing a pilot study: Why is it essential? Malaysian Family Physician, 1, 70–73. Haug, S., Meyer, C., Ulbricht, S., Gross, B., Rumpf, H.-J., & John, U. (2010). Need for cognition as a predictor and a moderator of outcome in a tailored letters smoking cessation intervention. Health Psychology, 29, 367–373. https://doi.org/10.1037/a0019450 Hawkins, K. W., & Linvill, D. L. (2010). Public health framing of news regarding childhood obesity in the United States. Health Communication, 25, 709–717. https://doi.org/10.1080/10410236.2010.521913 Henry, A. F., & Short, J. F. (1954). Suicide and homicide: Some economic, sociological and psychological aspects of aggression. New York, NY: Free Press. Herbert, A., Gerry, N. P., McQueen, M. B., Heid, I. M., Pfeufer, A., Illig, T., … Hu, F. B. (2006). A common genetic variant is associated with adult and childhood obesity. Science, 312, 279–283. Herek, G. M., Capitanio, J. P., & Widaman, K. F. (2003). Stigma, social risk, and health policy: Public attitudes toward HIV surveillance policies and the social construction of illness. Health Psychology, 22, 533-540. Hogue, T. E., & Peebles, J. (1997). The influence of remorse, intent and attitudes toward sex offenders on judgments of a rapist. Psychology, Crime and Law, 3, 249–259. Holder, H. D., & Treno, A. J. (1997). Media advocacy in community prevention: News as a means to advance policy change. Addiction, 92(Suppl 2), 189-199. Holton, A., Weberling, B., Clarke, C. E., & Smith, M. J. (2012). The blame frame: Media attribution of culpability about the MMR–autism vaccination scare. Health Communication, 27, 690–701. Hu, L., & 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, 1–55. https://doi.org/10.1080/10705519909540118 Huddy, L., Feldman, S., Taber, C., & Lahav, G. (2005). Threat, anxiety, and support of antiterrorism policies. American Journal of Political Science, 49, 593–608. 163 Hwang, H., Pan, Z., & Sun, Y. (2008). Influence of hostile media perception on willingness to engage in discursive activities: An examination of mediating role of media indignation. Media Psychology, 11, 76–97. Ibrahim, A., Ye, J., & Hoffner, C. (2008). Diffusion of news of the shuttle Columbia disaster: The role of emotional responses and motives for interpersonal communication. Communication Research Reports, 25, 91–101. Kingree, & Martie Thompson. (2000). Twelve-step groups, attributions of blame for personal sadness, psychological well-being, and the moderating role of gender1. Journal of Applied Social Psychology, 30, 499–517. https://doi.org/10.1111/j.1559-1816.2000.tb02493.x Janoff-Bulman, R. (1979). Characterological versus behavioral self-blame: Inquiries into depression and rape. Journal of Personality and Social Psychology, 37, 17981809. Jeong, S.-H. (2007). Effects of news about genetics and obesity on controllability attribution and helping behavior. Health Communication, 22, 221–228. Johnson, G., & Newmeyer, J. (1975). Pleasure, punishment and moral indignation. Sociology & Social Research, 59, 82-95. Jones-Smith, E. (2014). Theories of counseling and psychotherapy: An integrative approach. Thousand Oaks, CA: SAGE Publications. Jurkowski, J. M., Green Mills, L. L., Lawson, H. A., Bovenzi, M. C., Quartimon, R., & Davison, K. K. (2013). Engaging low-income parents in childhood obesity prevention from start to finish: A case study. Journal of Community Health, 38, 1–11. https://doi.org/10.1007/s10900-012-9573-9 Kalsher, M. J., Phoenix, G. M., Wogalter, M. S., & Braun, C. C. (1998). How do people attribute blame for burns sustained from hot coffee? The role of causal attributions. Proceedings of the Human Factors and Ergonomics Society, 42, 651655. Kersh, R., Stroup, D. F., & Taylor, W. C. (2011). Childhood obesity: A framework for policy approaches and ethical considerations. Preventing Chronic Disease, 8(5), A93. Key, M. S., Edlund, J. E., Sagarin, B. J., & Bizer, G. Y. (2009). Individual differences in susceptibility to mindlessness. Personality and Individual Differences, 46, 261– 264. Kleinke, C. L., Wallis, R., & Stalder, K. (1992). Evaluation of a rapist as a function of expressed intent and remorse. The Journal of Social Psychology, 132, 525–537. 164 Knaus, W. J. (2006a). Frustration tolerance training for children. In A. Ellis & M. E. Bernard (Eds.), Rational emotive behavioral approaches to childhood disorders (pp. 133–155). Boston, MA: Springer. Knaus, W. J. (2006b). The cognitive behavioral workbook for depression: A step-by-step program. Oakland, CA: New Harbinger Publications. Knobe, J. (2004). Intention, intentional action and moral considerations. Analysis, 64, 181–187. Knowler, W. C., Pettitt, D. J., Saad, M. F., & Bennett, P. H. (1990). Diabetes mellitus in the Pima Indians: Incidence, risk factors and pathogenesis. Diabetes/Metabolism Reviews, 6, 1–27. Koller, D., & Mielck, A. (2009). Regional and social differences concerning overweight, participation in health check-ups and vaccination. Analysis of data from a whole birth cohort of 6-year old children in a prosperous German city. BMC Public Health, 9, 43. https://doi.org/10.1186/1471-2458-9-43 Konishi, N., Oe, T., Shimizu, H., Tanaka, K., & Ohtsubo, Y. (2017). Perceived shared condemnation intensifies punitive moral emotions. Scientific Reports, 7, 1-9. https://doi.org/10.1038/s41598-017-07916-z Krinsky, C. (2016). The Ashgate research companion to moral panics. London, UK: Routledge. Kumanyika, S. K. (2008). Environmental influences on childhood obesity: Ethnic and cultural influences in context. Physiology & Behavior, 94, 61–70. Kuźbicka, K., & Rachoń, D. (2013). Bad eating habits as the main cause of obesity among children. Pediatr Endocrinol Diabetes Metab, 19, 106-110. Lagnado, D. A., & Channon, S. (2008). Judgments of cause and blame: The effects of intentionality and foreseeability. Cognition, 108, 754–770. Latner, J. D., & Stunkard, A. J. (2003). Getting worse: The stigmatization of obese children. Obesity, 11, 452–456. Latridis, T., & Fousiani, K. (2009). Effects of status and outcome on attributions and justworld beliefs: How the social distribution of success and failure may be rationalized. Journal of Experimental Social Psychology, 45, 415–420. https://doi.org/10.1016/j.jesp.2008.12.002 Lau, R. W. (2009). The contemporary culture of blame and the fetishization of the modernist mentality. Current Sociology, 57, 661–683. 165 Laufer, D., Gillespie, K., McBride, B., & Gonzalez, S. (2005). The role of severity in consumer attributions of blame: Defensive attributions in product-harm crises in Mexico. Journal of International Consumer Marketing, 17(2–3), 33–50. Laufer, D., Silver, D. H., & Meyer, T. (2005). Exploring differences between older and younger consumers in attributions of blame for product harm crises. Academy of Marketing Science Review, 7, 1-15. Leask, J., Hooker, C., & King, C. (2010). Media coverage of health issues and how to work more effectively with journalists: A qualitative study. BMC Public Health, 10, 535. https://doi.org/10.1186/1471-2458-10-535 Leiserowitz, A. (2006). Climate change risk perception and policy preferences: The role of affect, imagery, and values. Climatic Change, 77, 45–72. Lerner, J. S., Goldberg, J. H., & Tetlock, P. E. (1998). Sober second thought: The effects of accountability, anger, and authoritarianism on attributions of responsibility. Personality and Social Psychology Bulletin, 24, 563–574. Leskela, J., Dieperink, M., & Thuras, P. (2002). Shame and posttraumatic stress disorder. Journal of Traumatic Stress, 15, 223–226. https://doi.org/10.1023/A:1015255311837 Li, M., Chapman, S., Agho, K., & Eastman, C. J. (2007). Can even minimal news coverage influence consumer health-related behaviour? A case study of iodized salt sales, Australia. Health Education Research, 23, 543–548. Lin, L.-C., Huang, P.-H., & Weng, L.-J. (2017). Selecting path models in SEM: A comparison of model selection criteria. Structural Equation Modeling: A Multidisciplinary Journal, 24, 855–869. Locke, A. E., Kahali, B., Berndt, S. I., Justice, A. E., Pers, T. H., Day, F. R., … Yang, J. (2015). Genetic studies of body mass index yield new insights for obesity biology. Nature, 518, 197-206. Loehlin, J. C. (1998). Latent variable models: An introduction to factor, path, and structural analysis. Hillsdale, NJ: Lawrence Erlbaum Associates Publishers. Lussier, Y., Sabourin, S., & Wright, J. (1993). On causality, responsibility, and blame in marriage: Validity of the entailment model. Journal of Family Psychology, 7, 322-332. MacCallum, R. C., Zhang, S., Preacher, K. J., & Rucker, D. D. (2002). On the practice of dichotomization of quantitative variables. Psychological Methods, 7, 19-40. Madden, M. E., & Janoff-Bulman, R. (1981). Blame, control, and marital satisfaction: 166 Wives’ attributions for conflict in marriage. Journal of Marriage and the Family, 43 663–674. Maio, G. R., & Esses, V. M. (2001). The need for affect: Individual differences in the motivation to approach or avoid emotions. Journal of Personality, 69, 583–614. Major, L. H. (2009). Break it to Me Harshly: The effects of intersecting news frames in lung cancer and obesity coverage. Journal of Health Communication, 14, 174– 188. https://doi.org/10.1080/10810730802659939 Malcarne, V. L., Compas, B. E., Epping-Jordan, J. E., & Howell, D. C. (1995). Cognitive factors in adjustment to cancer: Attributions of self-blame and perceptions of control. Journal of Behavioral Medicine, 18, 401–417. Malle, B. F., Guglielmo, S., & Monroe, A. E. (2013). Moral, cognitive and social: The nature of blame (Vol. xviii). Hove, UK: Psychology Press. Malle, Bertram F., Guglielmo, S., & Monroe, A. E. (2012). Moral, cognitive, and social: The nature of blame. Social Thinking and Interpersonal Behaviour, 313–331. Malle, Bertram F., Guglielmo, S., & Monroe, A. E. (2014). A theory of blame. Psychological Inquiry, 25, 147–186. Malle, Bertram F., & Nelson, S. E. (2003). Judging mens rea: The tension between folk concepts and legal concepts of intentionality. Behavioral Sciences & the Law, 21, 563–580. Mandel, D. R. (2009). Predicting blame assignment in a case of negligent harm. Mind & Society, 9, 5–17. https://doi.org/10.1007/s11299-009-0064-3 Mantler, J., Schellenberg, E. G., & Page, J. S. (2003). Attributions for serious illness: Are controllability, responsibility and blame different constructs? Canadian Journal of Behavioural Science, 35, 142-152. Mariz, L. S., Enders, B. C., Santos, V. E. P., Tourinho, F. S. V., & Vieira, C. E. N. K. (2015). Causes of infantile-juvenile obesity: Reflections based on the theory of Hannah Arendt. Texto & Contexto-Enfermagem, 24, 891-897. https://doi.org/10.1590/0104-07072015002660014 Marschall, D., Sanftner, J., & Tangney, J. P. (1994). The state shame and guilt scale. Fairfax, VA: George Mason University Martin, B. A. S., Lang, B., & Wong, S. (2003). Conclusion explicitness in advertising: The moderating role of need for cognition (NFC) and argument quality (AQ) on persuasion. Journal of Advertising, 32(4), 57–65. 167 Mazzocco, P. J., Alicke, M. D., & Davis, T. L. (2004). On the robustness of outcome bias: No constraint by prior culpability. Basic and Applied Social Psychology, 26, 131–146. Meier, B. P., & Robinson, M. D. (2004). Does quick to blame mean quick to anger? The role of agreeableness in dissociating blame and anger. Personality and Social Psychology Bulletin, 30, 856–867. Messaris, P., & Abraham, L. (2001). The role of images in framing news stories. In S.D. Reese, O.H. Gandy Jr, & A.E. Grant (Eds), Framing public life: Perspectives on media and our understanding of the social world (pp. 231-242). London, UK: Routledge. Meyer, C. B., & Taylor, S. E. (1986). Adjustment to rape. Journal of Personality and Social Psychology, 50, 1226-1234. Mikolajczak, C. (2009). Availability of indium and gallium. Chicago, IL: Indium Corporation. Mikula, G. (2003). Testing an attribution-of-blame model of judgments of injustice. European Journal of Social Psychology, 33, 793–811. Mulder, L. B., & Aquino, K. (2013). The role of moral identity in the aftermath of dishonesty. Organizational Behavior and Human Decision Processes, 121, 219– 230. Mulford, C., Lee, M. Y., & Sapp, S. C. (1996). Victim-blaming and society-blaming scales for social problems. Journal of Applied Social Psychology, 26, 1324–1336. Mullinix, K. J., Leeper, T. J., Druckman, J. N., & Freese, J. (2015). The generalizability of survey experiments. Journal of Experimental Political Science, 2, 109–138. https://doi.org/10.1017/XPS.2015.19 Murphy, J., & Coleman, J. (2013). Philosophy of law: An introduction to jurisprudence (1st ed). Boulder, CO: Routledge. Muthén, L. K., & Muthén, B. O. (2015). Mplus User’s Guide (7th ed.). Los Angeles, CA: Muthén & Muthén. 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, 224–247. Nelkin, D. (1991). AIDS and the news media. The Milbank Quarterly, 69, 293–307. Nelkin, D., & Gilman, S. L. (1988). Placing blame for devastating disease. Social 168 Research, 55, 361–378. Nelson, T. E., & Oxley, Z. M. (1999). Issue framing effects on belief importance and opinion. The Journal of Politics, 61, 1040–1067. https://doi.org/10.2307/2647553 Nir, L. (2011). Motivated reasoning and public opinion perception. Public Opinion Quarterly, 75, 504–532. Obermann, M.-L. (2011). Moral disengagement among bystanders to school bullying. Journal of School Violence, 10, 239–257. O’Connell, A. A., & McCoach, D. B. (2008). Multilevel modeling of educational data. Charlotte, NC: Information Age Publishing. Ohbuchi, E., Hanaizumi, H., & Hock, L. A. (2004). Barcode readers using the camera device in mobile phones. Paper presented at 2004 International Conference on Cyberworlds, Tokyo, Japan. https://doi.org/10.1109/CW.2004.23 Ohtsubo, Y. (2007). Perceived intentionality intensifies blameworthiness of negative behaviors: Blame-praise asymmetry in intensification effect. Japanese Psychological Research, 49, 100–110. O’Keefe, D. J. (2002). Persuasion: Theory and research. Thousand Oaks, CA: Sage. Oliver, M. B., & Raney, A. A. (2011). Entertainment as pleasurable and meaningful: Identifying hedonic and eudaimonic motivations for entertainment consumption. Journal of Communication, 61, 984–1004. Olthof, T., Ferguson, T. J., & Luiten, A. (1989). Personal responsibility antecedents of anger and blame reactions in children. Child Development, 60, 1328–1336. O’Reilly, M. (2014). Blame and accountability in family therapy: Making sense of therapeutic spaces discursively. Qualitative Psychology, 1, 163-177. Ortony, A., Clore, G. L., & Collins, A. (1990). The cognitive structure of emotions. New York, NY: Cambridge University Press. Pepitone, A. (1976). Toward a normative and comparative biocultural social psychology. Journal of Personality and Social Psychology, 34, 641-653. Perryman, M., & Sidoti, K. (2015). Ethical considerations in the treatment of childhood obesity. Medicolegal and Bioethics, 17. https://doi.org/10.2147/MB.S63710 Pickard, H. (2013). Irrational blame. Analysis, 73, 613-626. https://doi.org/10.1093/analys/ant075 169 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, 145–153. https://doi.org/10.1136/tc.10.2.145 Pizarro, D. A., Uhlmann, E., & Bloom, P. (2003). Causal deviance and the attribution of moral responsibility. Journal of Experimental Social Psychology, 39, 653–660. Qi, Q., Chu, A. Y., Kang, J. H., Jensen, M. K., Curhan, G. C., Pasquale, L. R., … Qi, L. (2012). Sugar-sweetened beverages and genetic risk of obesity. New England Journal of Medicine, 367, 1387–1396. https://doi.org/10.1056/NEJMoa1203039 Quigley, B. M., & Tedeschi, J. T. (1996). Mediating effects of blame attributions on feelings of anger. Personality and Social Psychology Bulletin, 22, 1280–1288. https://doi.org/10.1177/01461672962212008 Reed, A., Aquino, K., & Levy, E. (2007). Moral identity and judgments of charitable behaviors. Journal of Marketing, 71, 178–193. Renaud, E. F. (2008). The attachment characteristics of combat veterans with PTSD. Traumatology, 14(3), 1-12. Reynolds, S. J., & Ceranic, T. L. (2007). The effects of moral judgment and moral identity on moral behavior: An empirical examination of the moral individual. Journal of Applied Psychology, 92, 1610-1624. Roberts, J. V., & Stalans, L. J. (2004). Restorative sentencing: Exploring the views of the public. Social Justice Research, 17, 315–334. Roese, N. J. (1994). The functional basis of counterfactual thinking. Journal of Personality and Social Psychology, 66, 805-818. Romer, D., Jamieson, K. H., Riegner, C., Emori, M., & Rouson, B. (1997). Blame discourse versus realistic conflict as explanations of ethnic tension in urban neighborhoods. Political Communication, 14, 273–291. Rosenbaum, J. E., & Johnson, B. K. (2016). Who’s afraid of spoilers? Need for cognition, need for affect, and narrative selection and enjoyment. Psychology of Popular Media Culture, 5, 273-289. Rowland, M. (2012). Blame or no-blame? Themes in media discourse about recent emergencies in Canberra. Social Alternatives, 31(3), 28-32. Rozin, P., Lowery, L., Imada, S., & Haidt, J. (1999). The CAD triad hypothesis: A mapping between three moral emotions (contempt, anger, disgust) and three moral codes (community, autonomy, divinity). Journal of Personality and Social Psychology, 76, 574–586. 170 Rule, B. G., & Nesdale, A. R. (1976). Emotional arousal and aggressive behavior. Psychological Bulletin, 83, 851-863. Runciman, W. B., Roughead, E. E., Semple, S. J., & Adams, R. J. (2003). Adverse drug events and medication errors in Australia. International Journal for Quality in Health Care, 15(suppl_1), i49–i59. Russell, C., Cameron, E., Socha, T., & McNinch, H. (2014). “Fatties cause global warming”: Fat pedagogy and environmental education. Canadian Journal of Environmental Education (CJEE), 18, 27–45. Saguy, A. C., Gruys, K., & Gong, S. (2010). Social problem construction and national context: News reporting on “overweight” and “obesity” in the United States and France. Social Problems, 57, 586–610. https://doi.org/10.1525/sp.2010.57.4.586 Salerno, J. M., & McCauley, M. R. (2009). Mock jurors’ judgments about opposing scientific experts: Do cross-examination, deliberation and need for cognition matter? American Journal of Forensic Psychology, 27, 37–60. Sargent, M. J. (2004). Less thought, more punishment: Need for cognition predicts support for punitive responses to crime. Personality and Social Psychology Bulletin, 30, 1485–1493. Scanlon, T. M. (2013). Interpreting blame. In J. Coates & Tognazzini (Eds.), Blame. New York, NY: Oxford University Press. Schichor, D. (1995). Punishment for profit: Private prisons/public concerns (1st ed). Thousand Oaks, CA: SAGE Publications, Inc. Schiller, N. G., Crystal, S., & Lewellen, D. (1994). Risky business: The cultural construction of AIDS risk groups. Social Science & Medicine, 38, 1337–1346. Schimel, J., Greenberg, J., & Martens, A. (2003). Evidence that projection of a feared trait can serve a defensive function. Personality and Social Psychology Bulletin, 29, 969–979. Shahtahmasebi, D. (2016, October 6). The war in syria: Who is actually to blame? Retrieved from http://theantimedia.org/syria-war-blame/ Shapiro, S. (1987). Self-mutilation and self-blame in incest victims. American Journal of Psychotherapy, 41, 46-54 Shaver, K. G., & Drown, D. (1986). On causality, responsibility, and self-blame: A theoretical note. Journal of Personality and Social Psychology, 50, 697–702. https://doi.org/10.1037/0022-3514.50.4.697 171 Sheikh, S., & McNamara, M. E. (2014). Insights from self-blame and victim blaming. Psychological Inquiry, 25, 241–244. Sher, G. (2005). In praise of blame. New York, NY: Oxford University Press. Shichor, D. (2005). The meaning and nature of punishment. Long Grove, IL: Waveland Press. Shoemaker, D. (2013). Blame and punishment. In D. J. Coates, & N.A. Tognazzini (Ed.), Blame: Its nature and norms (pp.100–118). New York, NY: Oxford University Press. Shultz, T. R. (1990). A rule based model of judging harm-doing. Presented at the Program of the Twelfth Annual Conference of the Cognitive Science Society, 2528 July 1990, Cambridge, Massachusetts: Psychology Press. Shultz, T. R., & Wells, D. (1985). Judging the intentionality of action-outcomes. Developmental Psychology, 21, 83-89. Shultz, T. R., Wright, K., & Schleifer, M. (1986). Assignment of moral responsibility and punishment. Child Development, 57, 177–184. Shultz, T., Schleifer, M., & Altman, I. (1981). Judgments of causation, responsibility, and punishment in cases of harm-doing. Canadian Journal of Behavioural Science, 13, 238–253. https://doi.org/10.1037/h0081183 Silventoinen, K., & Kaprio, J. (2009). Genetics of tracking of body mass index from birth to late middle age: Evidence from twin and family studies. Obesity Facts, 2, 196– 202. Simons, D. J., & Chabris, C. F. (2012). Common (mis) beliefs about memory: A replication and comparison of telephone and Mechanical Turk survey methods. PloS One, 7(12), e51876. Skitka, L. J., Bauman, C. W., Aramovich, N. P., & Morgan, G. S. (2006). Confrontational and preventative policy responses to terrorism: Anger wants a fight and fear wants" them" to go away. Basic and Applied Social Psychology, 28, 375–384. Skorupski, J. (1993). The definition of morality. Royal Institute of Philosophy Supplements, 35, 121–144. Slothuus, R. (2008). More than weighting cognitive importance: A dual-process model of issue framing effects. Political Psychology, 29, 1–28. Small, D. A., Loewenstein, G., & Slovic, P. (2007). Sympathy and callousness: The impact of deliberative thought on donations to identifiable and statistical victims. 172 Organizational Behavior and Human Decision Processes, 102, 143–153. Smith, A. (2013). Moral blame and moral protest. In D. J. Coates & N. A. Tognazzini (Eds.), Blame: Its nature and norms (pp. 27–48). New York, NY: Oxford University Press. Smith, P. J., Humiston, S. G., Marcuse, E. K., Zhao, Z., Dorell, C. G., Howes, C., & Hibbs, B. (2011). Parental delay or refusal of vaccine doses, childhood vaccination coverage at 24 months of age, and the Health Belief Model. Public Health Reports, 126(Suppl 2), 135–146. Smith, T. W., & Brehm, S. S. (1981). Cognitive correlates of the Type A coronary-prone behavior pattern. Motivation and Emotion, 5, 215–223. Solomon, R. C. (2008). True to our feelings: What our emotions are really telling us. New York, NY: Oxford University Press. Spark. (2011, September 1). How has the childhood obesity rate changed in 30 years? Retrieved from https://sparkpe.org/blog/how-has-the-childhood-obesity-ratechanged/ Spruijt-Metz, D. (2011). Etiology, treatment and prevention of obesity in childhood and adolescence: A decade in review. Journal of Research on Adolescence, 21, 129– 152. https://doi.org/10.1111/j.1532-7795.2010.00719.x Strelan, P., & Lawani, A. (2010). Muslim and Westerner responses to terrorism: The influence of group identity on attitudes toward forgiveness and reconciliation. Peace and Conflict: Journal of Peace Psychology, 16, 59-79. Sun, Y. (2014). Rethinking public health: Promoting public engagement through a new discursive environment. American Journal of Public Health, 104, e6–e13. https://doi.org/10.2105/AJPH.2013.301638 Sun, Y., Krakow, M., John, K. K., Liu, M., & Weaver, J. (2016). Framing obesity: How news frames shape attributions and behavioral responses. Journal of Health Communication, 21, 139–147. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate analysis. Boston, MA: Allyn and Bacon Tanaka, J. S. (1993). Multifaceted conceptions of fit in structural equation models. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 10-39). Newbury Park, CA: Sage Publications Inc. Taylor, C., & Kleinke, C. L. (1992). Effects of severity of accident, history of drunk driving, intent, and remorse on judgments of a drunk driver. Journal of Applied 173 Social Psychology, 22, 1641–1655. Tennen, H., & Affleck, G. (1990). Blaming others for threatening events. Psychological Bulletin, 108, 209-232. Tetlock, P. E., Self, W. T., & Singh, R. (2010). The punitiveness paradox: When is external pressure exculpatory – And when a signal just to spread blame? Journal of Experimental Social Psychology, 46, 388–395. https://doi.org/10.1016/j.jesp.2009.11.013 Thompson, S. C. (1985). Finding positive meaning in a stressful event and coping. Basic and Applied Social Psychology, 6, 279–295. Tomai, E., & Forbus, K. (2007). Plenty of blame to go around: A qualitative approach to attribution of moral responsibility. Paper presented at Proceedings of Qualitative Reasoning Workshop 2007, Aberystwyth, U.K. Trangsrud, H. B. (2010). Examination of victim and perpetrator blame in date rape scenarios and exploration of ambivalent sexism subtypes as predictors of male and female rape myths among a sample of college students (Doctoral dissertation). Retrieved from https://mospace.umsystem.edu/xmlui/handle/10355/9125 Unnithan, N. P. (1994). The currents of lethal violence: An integrated model of suicide and homicide. Albany, NY: SUNY Press. U.S. Census Bureau. (2017). The nation’s median age continues to rise. Retrieved from https://www.census.gov/library/visualizations/2017/comm/median-age.html Vander Ven, T., & Vander Ven, M. (2003). Exploring patterns of mother-blaming in anorexia scholarship: A study in the sociology of knowledge. Human Studies, 26, 97–119. Verplanken, B. (1989). Involvement and need for cognition as moderators of beliefs— attitude—intention consistency. British Journal of Social Psychology, 28, 115– 122. Verplanken, B. (1991). Persuasive communication of risk information: A test of cue versus message processing effects in a field experiment. Personality and Social Psychology Bulletin, 17, 188–193. Vitell, S. J., Bing, M. N., Davison, H. K., Ammeter, A. P., Garner, B. L., & Novicevic, M. M. (2009). Religiosity and moral identity: The mediating role of self-control. Journal of Business Ethics, 88, 601–613. Vogel, D. L., Werner-Wilson, R. J., Liang, K., Cutrona, C. E., Seeman, J. C., & Hackler, 174 A. H. (2008). The relationship of physiological arousal with demand and withdraw behavior: Examining the accuracy of the escape-conditioning hypothesis. Sex Roles, 59, 871-879. https://doi.org/10.1007/s11199-008-9490-7 Voth, J., & Sirois, F. M. (2009). The role of self-blame and responsibility in adjustment to inflammatory bowel disease. Rehabilitation Psychology, 54, 99-108. Wachs, F. L., Cooky, C., Messner, M. A., & Dworkin, S. L. (2012). Media frames and displacement of blame in the don imus/rutgers women’s basketball team incident: sincere fictions and frenetic inactivity. Critical Studies in Media Communication, 29, 421–438. https://doi.org/10.1080/15295036.2011.646282 Wakefield, A. J., Murch, S. H., Anthony, A., Linnell, J., Casson, D. M., Malik, M., … others. (1998). Ileal-lymphoid-nodular hyperplasia, non-specific colitis and pervasive developmental disorder in children. Lancet, 351, 637–641. Wallace, R. J. (1994). Responsibility and the moral sentiments. Cambridge, MA: Harvard University Press. Waller, B. N. (2005). Responsibility and health. Cambridge Quarterly of Healthcare Ethics, 14, 177–188. https://doi.org/10.1017/S0963180105050218 Warmke, B. (2011). Is forgiveness the deliberate refusal to punish? Journal of Moral Philosophy, 8, 613–620. Warner, R. H., VanDeursen, M. J., & Pope, A. R. (2012). Temporal distance as a determinant of just world strategy. European Journal of Social Psychology, 42, 276–284. Weber, H. (2004). Explorations in the social construction of anger. Motivation and Emotion, 28, 197–219. https://doi.org/10.1023/B:MOEM.0000032314.29291.d4 Weeks, B. E., Friedenberg, L. M., Southwell, B. G., & Slater, J. S. (2012). Behavioral consequences of conflict-oriented health news coverage: The 2009 mammography guideline controversy and online information seeking. Health Communication, 27, 158–166. Weinberg, N. (1994). Self-blame, other blame, and desire for revenge: Factors in recovery from bereavement. Death Studies, 18, 583–593. Weiner, B. (1995). Judgments of responsibility: A foundation for a theory of social conduct. New York, NY: Guilford Press. Weiner, B. (2006). Social motivation, justice, and the moral emotions: An attributional approach. New York, NY: Psychology Press. Weiner, B., Perry, R. P., & Magnusson, J. (1988). An attributional analysis of reactions to 175 stigmas. Journal of Personality and Social Psychology, 55, 738-748. WHO. (2017). WHO | Facts and figures on childhood obesity. Retrieved from http://www.who.int/end-childhood-obesity/facts/en/ Williams, C. W., Lees-Haley, P. R., & Price, J. R. (1996). The role of counterfactual thinking and causal attribution in Accident-Related judgments. Journal of Applied Social Psychology, 26, 2100–2112. Winterich, K. P., Aquino, K., Mittal, V., & Swartz, R. (2013). When moral identity symbolization motivates prosocial behavior: The role of recognition and moral identity internalization. Journal of Applied Psychology, 98, 759-770. Wringe, B. (2017). Rethinking expressive theories of punishment: Why denunciation is a better bet than communication or pure expression. Philosophical Studies, 174, 681–708. Wydo, M. R. (2003). Measuring anger in a prison population using the Anger Disorders Scale and the Personality Assessment Inventory (Doctoral dissertation). Retrieved from https://digitalcommons.pcom.edu/psychology_dissertations/150/ Xu, Z. X., & Ma, H. K. (2016). How can a deontological decision lead to moral behavior? The moderating role of moral identity. Journal of Business Ethics, 137, 537-549. Yang, Z. J., Kahlor, L. A., & Griffin, D. J. (2014). I share, therefore I am: A US-China comparison of college students’ motivations to share information about climate change. Human Communication Research, 40, 112–135. https://doi.org/10.1111/hcre.12018 Yen, S., & Siegler, I. C. (2003). Self-blame, social introversion, and male suicides: Prospective data from a longitudinal study. Archives of Suicide Research, 7, 17– 27. Young, R., Subramanian, R., & Hinnant, A. (2016). Stigmatizing images in obesity health campaign messages and healthy behavioral intentions. Health Education & Behavior, 43, 412–419. https://doi.org/10.1177/1090198115604624 Zaibert, L. (2009). The paradox of forgiveness. Journal of Moral Philosophy, 6, 365– 393. Zillmann, D., Knobloch, S., & Yu, H. (2001). Effects of photographs on the selective reading of news reports. Media Psychology, 3, 301–324. Zuk, C. V., & Zuk, G. H. (1989). The conflict cycle in the case of an adolescent in crisis. Contemporary Family Therapy, 11, 259–266.