2007 LIBRARY Michigan State University This is to certify that the dissertation entitled A STUDY OF DEFENSIVE ATTRIBUTION: MEASURING THE EFFECTS OF MEDIA BIAS presented by JAMESON CHRISTOPHER LONTZ has been accepted towards fulfillment of the requirements for the PhD. degree in CounselingPsycholggL Major Professor’s Signature % /3/ ‘95”; Date MSU is an Affirmative Action/Equal Opportunity Institution .--u-—A—-—.- — - PLACE IN RETURN BOX to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6/07 p:/CIRCIDateDue.indd-p.1 A STUDY OF DEFENSIVE ATTRIBUTION: MEASURING THE EFFECTS OF MEDIA BIAS By Jameson Christopher Lontz A DISSERTATION Submitted to Michigan State University In partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling, Educational Psychology, and Special Education 2006 Abstract A STUDY OF DEFENSIVE ATTRIBUTION: MEASURING THE EFFECTS OF MEDIA BIAS By Jameson Christopher Lontz Numerous empirical publications maintain the sentiment that biased media affects consumers. The present study explored cognitions of participants by manipulating a news story in a way that influenced affect, and subsequent attributions. This study investigated the degree to which media significantly influences emotion and leads to defensive attribution. Biased media reports may clout perceived knowledge, thus influencing self-govemment, relationship variables, opinion of mental health services, and general understanding of imparted information. All such outcomes call for psychologists and other professionals to intervene within their respective policy, clinical, counseling, and consulting settings where the results of the present study are applicable. Participants consisted of 625 randomly assigned students solicited from undergraduate social sciences courses at three universities: l) A small, private, Jesuit university in the Northwest US; 2) A large, public university in the Midwest US; 3) A large, public university in the Southeast US In this web-based study, participants read a news story of a crime that was committed and were then asked to attribute degree of blame to characters in the report. That is, degree of defensive attribution was measured. A measure of defensive attribution was based on H.H. Kelley’s (1973) attribution theory. It was postulated that a news story which was subtly manipulated so as to have the victim appear less virtuous would result in a higher degree of the victim being blamed for the crime. Participants” emotional response to characters in the news story was associated with defensive attribution. Implications of findings as well as firture research directions are discussed. Copyrighted by Jameson Christopher Lontz 2006 DEDICATION I dedicate this project to my mom, Vicki. It is with her bequest moral lessons of honesty, dedication, and hard work as well as stubbornness to no end that I attain every goal. I dedicate this project to my wife, Angel. Without objection, Angel has encouraged and supported every emotion along this challenging journey as we pursue our goals. ACKNOWLEDGEMENTS Thank you to each member of my dissertation committee for their concerned investment of time in order to provide the honest feedback that helped complete this project. The expert information and constructive criticism provided by Dr. John Kosciulek was particularly motivating in the beginning stages of this study. Dr. Gloria Smith, my program adviser, has been a reliable source of information and support as well as a provider of futuristic perspectives that have compelled this research. I wish to thank Dr. Nancy Crewe for not only allowing me to interrupt numerous lunch hours, but also for lending her sharp eyes and creative ideas to improving the quality of this final project. As a mentor, practicum supervisor, and committee member, Dr. Robert Fabiano has maintained high expectations while candidly expressing constructive criticisms as well as gracious compliments. For making it possible to conduct dissertation research at universities in opposite corners of the country, I thank Dr. Ken Rice as well as Dr. Paul Hastings for their zealous support and encouragement. Dr. Mark Reckase and Dr. Kimberly Maier provided instruction and specific feedback on the data analyses for this project. I wish to thank them both for their invested time and concern. Chueh-an Hsieh also reviewed and provided helpful feedback on the final results for this project. In addition to those named, I am especially grateful to faculty, administrative assistants, mentors, and my student peers for their hard work toward my success as well as friendships that made a vigorous endeavor tolerable. Finally, I am indebted to all members of my family for encouraging me to remember humble beginnings in the pursuit of every lofty endeavor. vi TABLE OF CONTENTS Content Egg Chapter 1- Introduction 1 Statement of Problem 3 Purpose of the Study 5 Hypotheses 7 Definition of Terms 8 Chapter 2- Review of Literature 9 Attribution Theory 9 Origins of Defensive Attribution 12 Bias in Media 15 Pre-l 980 l 5 1981 to 1990 16 1991 to Present 17 Summary 26 Chapter 3- Method 28 Participants 28 Materials 30 Research Design 30 Website 31 Data Analysis 32 Protection of Human Participants 33 Chapter 4- Results 35 Results of Additional Analyses 37 Chapter 5- Discussion 49 Limitations 50 Implications of Results 52 Directions for Future Research 54 Conclusion 57 References 59 Appendices 65 Appendix A: Solicitation Email 65 Appendix B: NEWS] & NEWSZ 66 Appendix C: First page of web-based study 69 Appendix D: Informed Consent 70 74 Appendix E: The Webpages vii Chapter 1 Introduction Myriad results indicate effects of bias in media. Price, Tewksbury, and Powers (1997) eXplored the effects of journal articles on the thoughts and feelings of readers. Price et al. randomly assigned 278 student participants in two studies to read one of four versions of a journal article. The first study found that type of journal article read significantly influenced which details readers focused on. Type of article also affected how readers evaluated topical information. The second study concluded that journal articles could affect decision making in regards to public policy. McClure, Lalljee, and Jaspars (1991) studied the effect of extremity of an event on number of causal explanations given for that event. Twenty participants evaluated events involving real as well as fictitious characters. Three studies used both unstructured and structured questionnaires to measure variables. The first two studies showed that some extreme events were attributed a single cause and other events were attributed a conglomerate of less extreme causes. The third study found that participants negatively correlated the number of causes with the extremity of an event. That is, the more extreme an event, the fewer causal explanations were attributed to that event. Therefore, according to McClure et a1. (1991), the way in which information is presented affects how consumers explain an event Daschmann (2000) explored a similar vein. Daschmann hypothesized that poll data influences what voters believe about the distribution of Opinions less than media presentations of testimonials. That is, testimonials are more likely than poll data to impact voters’ perception of who will win an election. Daschmann asked 274 participants to read a newspaper story about poll results and voter statements. Participants’ perceived climate of opinion, personal opinion, and voting intentions were measured. Testimonials influenced perceived climate of opinion and personal opinion, but did not affect voting intentions. The 2004 presidential election illuminates another example of how the modality by which information is disseminated may affect consumers. Leading up to the election, many television stations broadcast animated maps that reflect red and blue states. However, Maria Price, Ph.D., George Washington University associate professor and geography department chair, indicates that a county—by-county map makes the country appear more homogenous in its opinion than a state-by-state one (Bailey, 2005). This could be misleading because similarities between American voters may outnumber differences (Law, 2005). ‘ Willnat, He, Takeshita, and Lopez—Escobar (2002) investigated the degree to which 1,968 Asian and European student participants believe that United States media affects their culture. Participants were asked about their level of US. media consumption as well as how much they believe exposure to US. Media influences them and others. Most Asian participants reported believing they are positively influenced, in general, by US. media. Most European participants reported believing, in general, that they are negatively influenced by US. media. All participants reported that violence in US. media is a negative influence on the cultural values of themselves and others. Most European participants reported holding the belief that US. media influences other cultures more than European cultures. Most Asian participants held the conviction that US. media influences the cultural values of Asians to a higher degree than other cultures. IJ Overall, most participants in Willnat et al’s (2002) study indicated the perception that violence in US. media is a stronger influence on others than on themselves. Statement of Problem While myriad results do indicate effects of bias in media, the literature does not speak directly to implications for the practice of psychology. The problem is that biased media reports may clout perceived knowledge, thus influencing self-government, relationship variables, opinion of mental health services, and general understanding of imparted information. All such outcomes call for psychologists and other professionals to intervene within their respective policy, clinical, counseling, and consulting settings where the results of the present study are applicable. Participants were randomly assigned students solicited from undergraduate social sciences courses at 1) A small (approximately 6,100 students in 2004), private, Jesuit, university in the Northwest US; 2) A large (approximately 44,800 students in 2004), public, AAU (Association of the 62 leading research universities in America) university in the Midwest US; and, 3) A large (approximately 48,000 students in 2004), public, AAU university in the Southeast US. An email explaining the study was sent to instructors of the courses being solicited. The email asked approving instructors to forward the solicitation email to their class lists. There was a web-link in this email by which participants then accessed participation in the study. The email described how participants visit a website and complete questionnaires for a study that is being conducted. Sampling from these geographically diverse populations provides greater external validity. More geographically diverse samples may provide racial diversity between '2) participant samples, thus leading to more multicultural implications of results. Results of a study completed by Lontz (2004) produced face validity and construct validity for the measures used. Lontz hypothesized that changes in the verbiage of a news story would influence participants’ emotions and predict defensive attribution. The questionnaires created by Lontz appeared valid, and hypotheses were supported. That is, Lontz (2004) found that a biased media report predicts an increased degree of defensive attribution. In other words, participants were more likely to blame the victim of a crime when that victim was subtly made to appear less virtuous. The present study employed a web-based methodology. Participants visited an intemet website and then completed questionnaires which produced data for this study. Findings from intemet methodology are consistent with findings from traditional methodology (Gosling, Vazire, Srivastava, & John, 2004). Online participants have the freedom to withdraw at any time without the pressure to conform that may be present in person. Additional benefits of online methodology include: Increased ability to gather participants beyond traditional reaches, and thus increase sample sizes; increased efficiency in data collection; removal of the data entry process; low expense; and, increased diversity (e. g., gender) of samples. Further, in contrast to other assertions (e.g., Kraut, Patterson, Lundmark, Kiesler, Mukophadhyay, & Scherlis, 1998), intemet users are found to be no different in adjustment and depression from non-intemet-users. Internet methodology also acts as a motivating agent for eliciting the participation of frequent users of the intemet (Gosling et a1., 2004). Internet-based assessment can be far-reaching, enabling researchers to gather data from people with disabilities who might otherwise be discouraged from participation. Further, those individuals who do not speak English as a first language or whose cultural diversity precludes them from participating now have an alternative for inclusion (Naglieri, Drasgow, Schmit, Handler, Prifitera, Margolis, & Velasquez, 2004). By taking special precaution to consider its limitations, online methodology has great potential in the world of research. Purpose of the Study The present study hypothesized that bias in wording influences perception, and that these thoughts lead to an emotional reaction by consumers. This study investigated the degree to which media significantly influences emotion and leads to defensive attribution, that is, blaming of the victim. A measure of defensive attribution is driven by Kelley’s (1973) articulation of attribution theory. The most salient implication of the proposed study is that results, when communicated to consumers, may produce awareness of one’s attributions and awareness of the source of attributions (i.e., thoughts-9feelings-)attributions). That is, how does one explain events, and what is the source of one’s explanatory style? Further, results will add to mental health literature that addresses some of the stigma related to mental health services. Results can inform both media and laypeople about the influence of biased media. Imparted information can lead to reduced defensive attribution by consumers. In practice, psychologists, psychiatrists, and other mental health providers can help guide clients toward awareness of their own attributions. For example, therapists can help clients identify problem areas, such as how attributions about self and others could be more positive and informed so as to identify means of change. It has been postulated that the fashion by which information is imparted has consequences of astounding effect on consumers (e.g., Rule & Ferguson, 1986). Perceived characteristics of the person who conveys information are also influential (Dholakia & Stemthal, 1977; Rule & Ferguson, 1986). Specific wording also impacts consumers. McClure, Allen, and Walkey (2001) found that participants gave varied explanations as to why a building was destroyed during an earthquake based upon wording of the news report. That is, participant explanations depended on whether the building was reported to have been destroyed because of poor building design, lack of human preparedness, or the magnitude of the earthquake. There is cogent evidence that the impact on some individuals will be different from the impact on others. Specifically, children are more susceptible to media’s technique of ambiguous statements and underlying messages (Dittrnan, 2004). Violence in media may foster aggression, indifference to violence, or even antisocial behavior (Berkowitz, 1984). Scenarios in which aggression is somehow rewarded, for example, the bad guy gets away, teach children to be aggressive (Center for Communication, 1998). The fashion by which media presents information also impacts self-perceived knowledge. For example, a biased media report can affect one’s thoughts and lead to inaccurate assumptions regarding a particular political issue (e.g., Park, 2001; Daschmann, 2000; Price, Tewksbury, & Powers, 1997). Further, the agenda of a particular media source can sway opinion. Given that negative information makes up approximately 60% of what is stored in memory (P.B. Pedersen, personal communication, November 7, 2003), a negative interpretation of, for example, a political figure will elicit attributions accordingly. The agenda of a news reporter can mislead the 6 lay person on important issues. Rondeau (2003) implies a negative connotation in use of the term treatment when distinguishing it from counseling (Lontz, 2004). For the purposes of this study, it is appropriate to assume that programmed bias in media does influence consumers in some ways. Biased media is defined here as that media which has a planned agenda in its presentation as shown by a distinct difference in wording. An example of biased media is given by the news story for this study. The present problem is that biased media reports are manipulated so as to clout perceived knowledge, thus influencing self-government, relationship variables, opinion of mental health services, and general understanding of imparted information. The product of this manipulation calls for intervention by psychologists and other mental health professionals. Participants in this web-based study visited a website where they read a news story of a crime that was committed (Appendix B). Participants were then asked to attribute degree of blame to each of the characters in the news story (Appendix E). It was hypothesized that specific wording differences in the news story would impact emotional responses to the information and result in increased defensive attribution, or blaming the victim. 0 H]: A news story that is high in distinctiveness, consensus, and consistency will increase defensive attribution (u; - 112 i 0). 0 H2: There is a significant negative correlation between defensive attribution and approval of characters in a news story that differs in distinctiveness, consensus, and consistency (R< 0). Definition of Terms 1) Attribution: A causal explanation 2) Defensive Attribution: Blaming the victim a. Defensive attribution occurs when the attributor erroneously explains an event as being entirely due to (or the fault of) the person who is directly affected by the event. This phenomenon is an attribution error which fails to consider all of the situational details. The result is unjust blaming of the victim for something that is not entirely within the victim’s control. b. Defensive attribution is not blaming in the case of a clear cause-effect relationship between an individual’s actions and intended consequences. For example, blaming a person for getting high, after taking a drug with the intention of getting high, is not defensive attribution. Blaming the same person for overdosing is defensive attribution; overdosing is an uninvited adverse consequence. Automobile accidents, misfortune due to natural disasters, and crimes may often be attributed to individuals who do not have full control over the situation. Defensive attribution is a failure to consider all contributing factors that may lead up to an event. Defensive attribution is not describing what happened; it is an erroneous interpretation of why something happened. 3) Biased Media: That which has a planned agenda in its presentation as shown by a distinct difference in wording Chapter 2 Review of Literature The present study applied Kelley’s (1973) augmentation principle, which is otherwise known as blaming the victim with a values-based judgment. This study investigated the degree to which a bias in media significantly influences emotion and leads to defensive attribution. A measure of defensive attribution was based on Kelley’s (1973) articulation of attribution theory. The purposes of this investigation were to: 1) determine whether bias in media influences emotions, and 2) measure the degree to which emotions influenced by biased media lead to defensive attribution. The following is a concise review of relevant literature. A complete but parsimonious explanation of Kelley’s (1973) articulation of attribution theory is followed by additional relevant literature. Bias in media is explored chronologically in ascending order by type of media: 1) Pre-1980 written media, radio-, and television-broadcasts, 2) 1981 to 1990 written media, radio-, and television-broadcasts, and 3) 1991 to Present written media, radio-, and television-broadcasts. Given the contemporary emphasis of this literature review, the reader will become well informed as to how Kelley’s theory applies to bias in media. Following a synopsis of Kelley’s theory, the first two sections of the literature review (i.e., Pre-1980, and 1981 to 1990) combine literature from written, radio-, and television-broadcast media. The final section (1991 to Present) discusses each of these genres in turn. Attribution Theory Attribution theory is primarily concerned with exploring why things happen. The theory incorporates both the information used and what is done with this knowledge to determine causation. H.H. Kelley (1973) uses the term “psychological epistemology” in speaking to the process where a person both “knows his world” and “knows what he knows.” This process is regarded as true introspection of causal attributions. Kelley posits that “it is precisely common sense with which attribution theory is concerned” (p. 107). The following section is a succinct explanation of the key constructs in Kelley’s theory. Attribution theory (Kelley, 1973) posits that a three part interaction transpires when attribution, that is, a causal explanation, occurs in a given situation. This three part interaction is made up of Person (i.e., idiosyncratic variables), Entity (i.e., subject of the interaction), and Time (i.e., circumstantial factors). For example, a man (the person) might admire a painting (entity) in a gallery of several paintings (time). Qualities within the person, something about the painting, or the man’s surroundings, i.e., the gallery and other people present, determine why the man chooses that one painting to admire. Further, and more complex, collaboration of all three, i.e., PersonXEntityXTime, may provide an explanation. Within this triadic perspective, the sub-constructs of Distinctiveness, Consensus, and Consistency determine to what one attributes the outcome of a situation. Distinctiveness measures how closely associated one’s attribution is to a stimulus. For example, a measure of distinctiveness would determine how closely one’s explanation of why a house was destroyed during a hurricane is related to the force of the hurricane (i.e., the stimulus). Distinctiveness is also a matter of selectivity, which leads to attributions of preventability. For example, was the house exclusively destroyed, or were all houses in the area destroyed? Consensus, on the other hand, is a phenomenon 10 indicating that one is more confident about his or her attributions when that decision is supported by others. That is, when others make the same attribution, or when the individual believes others will make the same attribution, there is more confidence in one’s causal explanation. Consensus may also be fostered by expert opinion (or some model perceived as an expert). For example, a forensic psychologist in a criminal law trial may be viewed as an expert by jurors. A measure of consistency determines the frequency with which distinctiveness and consensus of information occurs over time. For instance, it may be known (italics added) by most people that all houses in the direct path of a hurricane are usually destroyed. Degrees of distinctiveness, consensus, and consistency vary depending on how information is presented. According to Kelley’s theory, validity in attribution requires simultaneous consideration of distinctiveness, consensus, and consistency. That is to say, in order to accurately measure reasons for an individual’s interest in a painting, for example, the researcher will analyze person, entity, time, distinctiveness. consensus, and consistency as potential influences. Additional concepts in the theory include the discounting principle and the augmentation principle. The discounting principle states that the cause given for an effect may be discounted with the presence of other feasible causes. For example, a crime against some antagonist may be unjustifiably attributed to his or her personal qualities. This example introduces the augmentation principle in which action is more likely attributed to the actor when “constraints, costs, sacrifices, or risks” (Kelley, 1973) are involved in the action. For example, if a prostitute is reportedly assaulted, consumers of this news may attribute the crime to the actions and decisions of the prostitute. Personal variables and wrong place at the wrong time factors are often ambiguous and therefore difficult for the consumer to ascertain. This can result in defensive attribution by the lay consumer. In sum, Kelley’s attribution theory carries the theme that causal inferences reflect mindful (and sometimes mindless) consideration of multiple causes for an effect. However, Kelley writes about certain attribution errors that may occur. For instance, actors in a given situation may attribute their actions to situational factors and the actions of others to unique and specific personal factors (Ross, 1977, as cited in Weiten & Lloyd, 2003). This may be due to the predication by some individuals that victims deserve misfortune if undesirable characteristics are present or poor choices are made by the victim (Lerner, 1980). This phenomenon also serves individuals in that it enables denial of one’s own potential victimization (Lambert, Burroughs, & Nguyen, 1999). Another theme of Kelley’s writings is gleaned with a quote once recorded by the Roman known as Virgil, “Happy is he who has been able to perceive the causes of things” (p. 127). Origins of Defensive Attribution According to Wilhelm Wundt, psychologists are to study the psychological processes by which we experience and observe the external world (Hothersall, 2004, p. 123). Although it may seem commonplace, an explanation of defensive attribution and the psychological processes that motivate this phenomenon is useful here. To articulate an exact origin of defensive attribution would be overly bold, given the various cultural factors that influence one’s explanatory style. l—Iowever, some underlying processes are worth exploring. In particular, the concepts of defensiveness and locus of control are meaningful here. This dissection of defensive attribution includes a definition followed by complexities that contribute to the concept. Weiten and Lloyd (2003) discuss defensive attribution in terms of, “a tendency to blame victims for their misfortune so that one feels less likely to be blamed in a similar way” (p. 159). Examples include blaming victims of domestic violence or even rape for their apparent lack of motivation to either leave or avoid such harmful situations. Breaking down defensive attribution to defensiveness alone reveals an attempt by the individual to shield oneself from pain. This may include reacting to save face or when feeling threatened interpersonally, such as when feeling belittled (p. 197). Aside from constructs that may provoke a defensive response, there may be further underlying factors. Carl Rogers articulates defensiveness in terms of the reaction to a situation where self-concept is threatened (Weiten & Lloyd, 2003). When there is incongruence between one’s self-concept and how others view that person, the result is anxiety for that person. For example, the individual who is reacting defensively will go so far as to “. . .ignore, deny, and twist reality” (p. 48) to maintain their view of self and to avoid the inevitable anxiety that is produced by acknowledging lack of personal virtue. Engebretson, Clark, Niaura, Philliips, Albrecht, and Tilkemeier (1999) found that patients who left a cardiac rehabilitation program prematurely were more likely to exhibit higher levels of anxiety and defensiveness, to be younger, and to perceive themselves as healthier. Kraemer, Salisbury, and Spielman (1998) found that defensiveness, among other traits, was not predictive of treatment completion among juveniles in a sex offender treatment program. Rogers, along with Abraham Maslow, analyzes defensiveness even deeper by emphasizing that the roots of congruence lie in one’s sense of personal worth. 13 Personal worth may be affected by locus of control, which may be affected by one’s self esteem (W eiten & Lloyd, 2003). Blatier (2000) describes a positive correlation between locus of control and type of prison sentence. Blatier found that those prisoners who served their sentence at a site outside of the actual prison had a more internal locus of control compared to those prisoners who served their sentence inside of the prison. That is, the study found that prisoners who served their sentence outside of the actual prison were more likely to see themselves in control of their crimes. In other words, rather than blaming the victim, judicial system, or some other entity for their being in prison, prisoners who served their sentence outside of the actual prison took more responsibility for their crimes. Hamilton and Akhter (2002) analyze locus of control in terms of effort and ability. However, Hamilton and Akhter assert that ability is uncontrollable and effort is controllable. Therefore, when an individual defensively attributes, they are judging the effort of the victim (i.e., whether the victim tried hard enough) without considering ability (i.e., whether the victim really can try harder). Sigmund Freud’s psychoanalytic theory is even more microscopic in its dissection of the concept at hand. Freud’s theory (Weiten & Lloyd, 2003) outlines numerous defense mechanisms that are identified to protect an individual from experiencing anxiety and guilt. These defense mechanisms include: Rationalization (falsely justifying poor behavior); Repression (failing to acknowledge unconscious feelings); Projection (attributing interpersonal problems to others); Displacement (taking out emotion, often anger, on uninvolved individuals); Reaction formation (behavior that is profoundly opposite of one’s true feelings); Regression (a reversion to childlike behavior in response to some stressful event or pattern); and, Identification (forming an alliance, either made up or real, that brings with it a sense of increased self esteem). Defensiveness may also be used to protect others. Ingram and Hutchinson (1999) found that knowledge of stigma about one’s own disposition led mothers who were HIV (human immuno virus) positive to be more defensive of their children. That is, these mothers were increasingly likely to take steps that would prevent their children from contracting HIV because of personal experience with being stigmatized against for having the illness. In returning to a broad-based discussion, Schutz (1998) describes defensiveness in terms of being a self-presentation style. Further, Schutz sees this presentation style as a taxonomy or classification, rather than, more specifically, a trait. With a thorough understanding of attribution theory as well as some of the constructs that impact degree of defensive attributions, this review now turns to additional relevant literature. Bias in Media Pre-1980 Kaplan and Sharp (1974) found that participants perceived negatively portrayed characters in a written story as competent and dynamic, yet untrustworthy, compared to characters who unintentionally produced the same outcome. Advertising, the practice of persuasion, often utilizes admirable characters, for example, sports superstars, to sell a product. Dholakia and Stemthal (1977) found the credibility of a source to have an effect on attribution. Dholakia and Stemthal’s participants rated high credibility sources as “more trustworthy and expert” than low credibility sources; but, results of perceived attractiveness of the sources were not different between groups. Dholakia and Stemthal concluded with the suggestion that “attribution of behavior affects people’s feelings about an issue rather than their cognitions or behavioral intentions...” (p. 231). Eagly, Wood, and Chaiken (1978) contrasted the findings of Dholakia and Stemthal (1977). Eagly et al. (1978) found that participants perceived communicators as less biased when the participants’ expectations of the communicator were not confirmed. Confirmed expectations lead to the perception that communicators were insincere and manipulative. Additionally, Eagly and Chaiken (1975) found that communicators perceived as attractive were more persuasive compared to unattractive communicators. This finding was magnified when communicators advocated undesirable positions. Participants also assmned that communicators with attractive characteristics would more likely advocate desirable positions. 1981 to 1990 Rule and Ferguson (1986) discussed observers’ reactions to aggressive events depending on interpretations. Cognitive interpretations influence moral evaluations and thus attributions. In considering children as observers, not only can media encourage children to use aggression in solving interpersonal problems and produce indifference to violence, but it is also postulated that violent acts on television may cultivate antisocial behavior. Observers may be inclined to aggressive behavior, but this is not inevitable (Berkowitz, 1984). Gulotta (1983) continued in discussion of how attributions by observers of an event are more influenced by their perception of another’s actions rather than overt behavior of the actor. Further, level of anticipation of an event may have an effect on attribution. Unanticipated events may be more likely to elicit snap judgment causal explanations (Hastie, 1984). I6 In a study of comparative advertising, Gorn and Weinberg (1984) found that perception was most influential on participants’ ratings of brand similarity. Therefore, the nature of comparative advertising (i.e., brand X portrayed as better than brand Z) is effective in swaying practical attribution with these products. In another comparison, Gaddy and Tanjong’s (1986) review differentiated natural disasters from sociopolitical ones (e.g., crime waves) in terms of attribution. The bestowal of responsibility to the victims of a crime differed significantly from attributions in natural disaster situations. Additional evidence for bias in media is shown by Mulilis and Lippa (1990) who found that participants increased earthquake preparedness when exposed to media that was manipulated so as to have threat of an earthquake be perceived as more probable. 1991 to Present Written The most contemporary literature continues to voice problems associated with bias in media. However, much salience lies in what media brings the world. Fuller (1996) maintains the necessity (despite all faults) of media by citing Thomas Jefferson: “Were it left to me to decide whether we should have a government without newspapers, or newspapers without a government, I should not hesitate a moment to prefer the latter” (pp. 23-24). McClure’s (1998) discussion relates to Kelley’s (1973) discounting principle, where McClure proposes that people will discount attributions when another plausible explanation exists. On the other hand, participants will often “anchor” themselves to their initial attribution and essentially dismiss any alternatives. These findings supported McClure, Lalljee, and Jaspars’ (1991) finding that participants will increase the weightiness of a single cause instead of attributing additional causes in the case of extreme effects. However, this contradicted Kelley’s finding that extreme effects generate more causes. To reconcile this contradiction, replication of these studies would seem appropriate. McClure, Walkey, and Allen (1999) hypothesized there to be a difference between low risk-takers with an internal locus of control around a natural disaster in two populations (i.e., students and non-students). Results produced evidence that students perceive earthquake damage as more preventable than non-students, and that non-students indicate more “complex explanations.” However, both samples showed a positive correlation between internal locus of control and perceived ability to prevent earthquake damage. Further, distinctive damage (e.g., one building selectively destroyed in a block of several buildings) was rated by participants as more preventable than global damage (i.e., all buildings destroyed). A positive correlation between age and earthquake preparation was found. A negative correlation was found between age and risk-taking. Willnat, He, Takeshita, and Lopez-Escobar (2002) hypothesized that participants from ten major universities in eight countries (Japan, Indonesia, Hong Kong, China, Spain, Germany, Great Britain, and the Netherlands) would attribute effects of United States media as more influential on others than on themselves. Where most Asian participants indicated positive influences from US. media, most European participants indicated that US. media has a negative effect on their cultural values. All participants indicated that violence in US. media can have a negative effect on self and others; but, influence on cultures outside of the US. was perceived as more significant. Further, data from European participants suggested the perception that US. media affects cultural l8 values of other cultures more than US. cultures; data fi'om most Asian participants suggested the opposite, especially students from Hong Kong who perceived “very little differences” in US. media effect across cultures. Other opinion data includes Daschmann (2000) who found that opinions broadcast by media had a “considerably stronger” effect on poll results than raw poll data that was broadcast. However, none of the media analyzed in Daschmann’s study significantly affected voting intentions. In contrast, poll data may have been affected by another of Daschmann’s findings, that anti-govemment exemplars were seen as more credible. Biased media may impact attribution in myriad ways. Regarding the thoughts and feelings of participants, Price, Tewksbury, and Powers (1997) found that written news frames, which did not increase frequency of cognitive responses, affected attributions about the issue in the news frames. The same researchers found that variation in the news frames could affect public policy decisions. According to Price and Tewksbury (1997), important parts of a message will activate ideas leading to certain attributions. These “applicability effects” compel a person to apply the same thoughts and feelings in later decisions, resulting in attributions referred to as “accessibility effects.” That is, it seems as though people can develop an inflexible style of attribution. Further results suggested that news frames, which discuss tuition increases, stimulated somewhat universal thoughts and feelings in students. That is to say, the consensus phenomenon appears to have affected participants in this study. The point here is that even if the specifics of a news story do not necessarily apply to a consumer, the phenomenon of consensus (Kelley, 1973) does. James and Hensel (1991) conclude with their elaboration on “Attitude-Toward- the-Acf’ in which it was postulated that affective reactions (i.e., feelings) influence attitudes about products in advertising. Further, considering that negative information makes up approximately 60% of what is stored in memory (P.B. Pedersen, personal communication, November 7, 2003), negative advertising, or any negative information, plays a large role in decision making. The power of media to impress was also exemplified by Dashborough and Ashkanasy’s (2002) review article where they indicated that leadership is: 1) an emotional process modeled by the leaders who influence followers, and 2) a social interaction modus. In short, followers make decisions regarding perceived intentions of leaders. These decisions elicit emotion, and vice versa, which in turn elicits attributions. Dashborough and Ashkanasy pointed out, however, that the ability of followers to accurately interpret leaders’ intentions is subjective. In accordance with attribution theory, Gotlieb and Sarel (1991) found that when a highly credible source is used in comparative advertising, there is a positive effect on item preference (e.g., images of superstars to sell more sports drinks). On the other hand, when a less credible source, for example, a shifty car salesperson, is paired with a product, the source will be discredited. Shamir (1991) contrasted this by finding that such attribution comes from the group (i.e., consensus) and is not due to credibility of the source. Along a similar vein, McClure et a1. (2001) wanted to find whether biased media contributes to fatalism regarding natural disasters. That is, McClure et a1. wanted to know if people believe damage occurring from natural disasters is preventable. The results of four studies found that participants were less likely to prepare for earthquakes ”)0 when fatalistic attributions are made. For example, media can present information about an event in a way that makes the event appear inevitable. High consistency and consensus and high distinctiveness, in various conditions presented to participants in newspaper story form, lead to attributions where participants laid blame on building design, and held the belief that damage to buildings ruined during the earthquake was preventable. Along with these findings, there were differences between the age groups sampled. A younger population may be more susceptible to media bias. College student participants made attributions that were affected by both consensus and distinctiveness. Participants from the general public responded to only consensus information when making attributions. Further, all participants made attributions based on consensus alone. That is to say, all participants assumed that other participants would make the same attributions. Radio Cooper (1996) nostalgically discusses parental concerns about violence in broadcast media first appearing in 1933. At this time, the “na'i'vete” of children was thought to make them susceptible to violent programs, and parents protested shows such as “Ether Bogeyman.” In these times, parents were even more weary with television, as it produced “over excitement and nightmares” (p.25) in children. In 1954 the California Youth Authority (CYA) reported a rise in youth crime rates and laid blame on television crime programs. At this time, James Bennett (director of the Bureau of Prisons) spoke in advocacy of television programming, saying that only “unstable. . .rebellious. . .unhappy” (p. 29) children would be affected by such programming; but this contention remained a source of debate with the CYA. During the 19603, President Lyndon Johnson addressed the media with the following remarkable statement: There is no denying it. You, the broadcast industry, have the enormous power in your hands. You have the power to clarify. And you have the power to confuse. . .No law and no set of regulations and no words of mine can improve you or dictate your daily responsibility. All I mean to do- what I’m trying to do-is to remind you that where there’s great power there must also be great responsibility. This is true for broadcasters just as it’s true for Presidents, and seekers for the Presidency (C00per, 1996, p. 49). This statement, followed by the assassination of Robert Kennedy, ignited President Johnson’s appointment of the National Commission on the Causes and Prevention of Violence, whose purpose was to investigate violence in the United States (Cooper, 1996). Yanovitzky and Cappella (1999) found that attitudes toward political leaders in the 1996 Presidential election were only slightly influenced by political talk radio (PTR) shows. This study found that pre-existing attitudes about political leaders were more influential; although these results were not mediated by political knowledge. Therefore, Yanovitzky and Cappella concluded that PTR shows have little if any effect on political opinion. Rossler and Schenk (2000), however, found that, although media effects proved moderate in general, participants with a high need for persuasive messages were affected. Further, Rossler and Schenk found a prevalence of similar opinion among communication networks. Broadcasts of current events also suggest bias. In a story by National Public Radio (2003), it was suggested that some Arab networks (television and radio broadcasts) have influenced civilian opinion of US. conflict in Iraq by the ways in which news is reported. ’7’) _~ Television Gerbner (1998) tells the story of a child who once answered an inquiry about the value of Thomas Edison in our modem world by saying that without Mr. Edison, “. . .we would still be watching television by candlelight” (p. 75). Television provides a base of models with its characters. Whether a steamy love affair between a client and his therapist as in “Prince of Tides” (Streisand, 1991), or a highly misconstrued presentation of what mental health services entail as in “Anger Management” (Seqal, 2003), bias in media impacts consumers in potentially damaging ways. It is in the nature of movie making to manipulate the audience (Turner, 2005). Given that approximately 98% of homes in the United States have a television (Bushman & Anderson, 2001), the effects are far-reaching. Television characters model attitudes and behavior, including aggressiveness. In addition, television desensitizes viewers to violence, and, on the other end of this spectrum, increases fear of victimization. Further, by glamorizing violence on television, producers send the message that such violence is acceptable. Therefore, aggression could be learned by rewarding violence (Center for Communication, 1998). Still, consistent with a view later adopted by Freud, Aristotle insisted that the cathartic nature with which emotions are aroused by drama can be therapeutic (Hothersall, 2004, p. 28). Judgments about a television character influence interpretations of acceptable violence. If an admired character is involved in, for example, a gun battle, that character’s behavior will be rationalized more often than when a less admired character performs the same behaviors. To extend this point, Dillard, Plotnick, Godbold, Friemuth, and Edgar (1996) found public service announcements (PSAs) on the topic of AIDS/HIV Id 1,) to produce affect in participants that influenced the persuasiveness of the announcements. Further, believability of occurrences also influences emotional arousal. The realistic quality with which scenes are enacted on television may further influence attributions. Given the advance of technology in present day television graphics, believability is no longer the issue it once perhaps was. “Crime is a major staple” in news, and consumption, i.e., viewer interest, of such media influences what is broadcast. It has also been postulated that some television news stations target a younger audience with more brief stories and less sound bites. This brings fear to the idea that viewers will make attributions only by what they see. The quick pace of news stories does not provide opportunity for interpretation, and local news stations are the largest culprit of such practices. The bottom line is that a choice is made by station producers as to what will be aired, and this choice impacts consumers accordingly (Hamilton, 1998). Firmstone (2002) reported a nationally representative survey from the United Kingdom in which participants were asked to give their opinion after viewing several television clips of the Persian Gulf War. Results indicated that 70% of viewers approve of all the footage, but 57% of respondents do not approve of scenes where captured coalition forces are being exploited on Iraqi television. Some respondents felt that television stations were justified in airing scenes showing Iraqis in a bad light, but that scenes showing Iraqis surrendering might lower support for the war against Iraq. This leaves media with a decision of what to provide its viewers; and the decision varies among broadcasters. For example, WTN broadcasting company aired “more explicit material” (e.g., charred bodies) compared to other companies such as the BBC. Firmstone (2002) concluded that a combination of factors, i.e., context and genre within which violent acts occur, in television affect viewer interpretation. Jones (2002) speaks to bias in media reports by citing a January, 2001 San Francisco Chronicle news article with the headline: “Aggressive tendencies fostered in children by violent television shows and video games can be tempered if they cut back their viewing and playing, a new Stanford University study shows.” However, the news article apparently did not discuss all of the results. That is, kids who watched violent programs improved their behavior when cutting back television viewing, but not any more than kids who watched nonviolent programs. Gerbner (1998) illuminates other biases in television media. For example, White males are over represented by 150% of their true population pr0portion. Older persons are represented by only 20% of their proportion. Further, only 1.3% of characters in television are impoverished (p. 76). Park (2001) discusses the illusion ofknowing concept where perceived social approval compels individuals to accept the consensus of a group. This illusion of knowing, where actual and self-perceived knowledge are incongruent, may cause individuals to be inadvertently manipulated by media. The manipulation is inadvertent because individuals may believe they are too intelligent to be manipulated. Park’s results indicate a negative correlation between educational attainment and illusion of actual knowledge. Further, compared to consumption of written and radio media, television consumption leads to more inaccurate self-perceived knowledge. Consensus may be damaging. Social involvement in political issues leads to misperception of one’s knowledge more often than when involvement is not within a social consensus (Park, 2001) Summary This study is motivated by the compelling evidence that bias in media does exist. For purposes of this study, biased media is defined as that media which has a planned agenda in its presentation as shown by a distinct difference in wording. McClure et a1. (2001) concluded, “there is no evidence” that television and radio are more valid than written messages. No matter the type of media, bias can be damaging. An integral review of relevant literature does not, however, speak to measures of defensive attribution, that is, blaming the victim, regarding a crime that has occurred. The literature also has not documented full employment of web-based methodology, which further utilizes technology and thus increases external validity of results with its long-armed reach at samples. In one comparison, Gaddy and Tanjong’s (1986) review differentiated natural disasters from sociopolitical ones (e.g., crime waves) in terms of attribution. The bestowal of responsibility to the victims of a crime differed significantly from attributions in natural disaster situations. Additional evidence for bias in media exists as shown by Mulilis and Lippa’s (1990) finding that participants increased earthquake preparedness when exposed to media that was intentionally manipulated so as to have an earthquake be perceived as threatening. In another comparison, Hastie and Park (1986) made case for a distinction between online and memory-based judgments. More concluding in this respect, McCabe, Boyd, Couper, Crawford, and D’Arcy (2002) found that web-based survey methodology produced higher response rates than US. mail survey methods. Further, McCabe et al. found that a more gender-mixed sample resulted from the web—based methodology in a college undergraduate population. These findings compel methodology for the present study which investigated the degree to which media significantly influences emotion and leads to defensive attribution. Finally, replication research is underutilized (Karr & Larson, 2005). Therefore, by replicating the methodology of Lontz (2004), this study adds to a body of literature while also working to substantiate previous empirical findings. Chapter 3 Participants Method Six-hundred twenty-five participants, with a mean age of 20.8 years, from undergraduate social sciences courses at three universities were randomly assigned to read one of the two versions of a news story about a crime. Participants were solicited from 1) A small (approximately 6,100 students in 2004), private, Jesuit, tmiversity in the Northwest US; 2) A large (approximately 44,800 students in 2004), public, AAU (Association of the 62 leading research universities in America) university in the Midwest US; and, 3) A large (approximately 48,000 students in 2004), public, AAU university in the Southeast US. A total of 650 participants took part in the study. Prior to analyses, data from 25 participants who submitted insufficient data were excluded from results. Table 1 Race/Ethnicity of Participants Identification Reported Frequency Asian or Asian American, including 42 Chinese, Japanese, Pacific Islanders, and others Black or African American 50 Hispanic or Latino, including Mexican 38 American, South American, Central American, and others White, Caucasian, Anglo, European 468 American (non-Hispanic) American Indian or Native American 2 Mixed: Parents are from two different 14 groups Middle Eastern 6 Race or ethnicity is not listed above 3 .__.._...._ _______j Table 2 Completed Education Education School Class Standing Present College Class Standing Reported Frequency Freshman 248 Sophomore 157 Junior 85 Senior 57 Graduated or Pursuing a Graduate Degree 73 Table 4 Reported Gender Male 191 Female 417 An email describing the study (Appendix A) was sent to instructors of the courses being solicited. The solicitation email asked approving instructors to forward the electronic letter to their class lists. There was a web-link in this email by which participants accessed participation in the study. The email described how participants visit a website and complete questionnaires for a study that is being conducted. ")0 Materials Following a comprehensive literature review, both the news stories and the data collection measures were created by the researcher of the present study. An analysis of H.H. Kelley’s (1973) articulation of attribution theory lead to the creation of a defensive attribution measure. After reading one of the two versions of the news story, participants completed two questionnaires (see Appendix E). The first questionnaire asked participants to rank each of the characters in the news story by level of responsibility for the crime that has been reported (i.e., most responsible for the crime receives a ranking of 1). The next questionnaire asked participants to rate each of the characters in the news story on a 6-point scale. The purpose of this rating is to identify the degree to which participants agree with the actions of each news story character. Wording of the news story differs in distinctiveness, consensus, and consistency. Participants were randomly assigned to read either NEWS] or NEWS2 (i.e., the version of the news story that is high in distinctiveness, consensus, and consistency). Research Design A post-test only control group design drives this study (Table 5). The fabricated news story in this study has two versions which differ in magnitude of distinctiveness, consensus, and consistency. The experimental group of participants read the version of the news story that is high in distinctiveness, consensus, and consistency. After reading one of the two versions, participants in each group were asked to make responsibility ratings and agreement ratings. The degree to which the victim of the crime is blamed for the crime, i.e., defensive attribution, was measured. Data consisted of the defensive attribution means for each group. A statistical comparison of means elicited results. It was hypothesized that a media report that is high in distinctiveness, consensus, and consistency will increase defensive attribution (u; - p2 7E 0). A correlation between responsibility and agreement ratings for characters in the news story was then calculated. Table 5 Model of the study design. A post-test only control group design. Group Time-) Group 1 Treatment Observation Group 2 —-- Observation Website When participants visited the study’s website, they were presented with informed consent (Appendix D) preceding any data collection. Once learning that they may withdraw at any time without penalty, and agreeing to participate, participants observed the following message: Attention: The following news story is from a large newspaper on the East Coast US. Ifyou wish to continue, please be aware that the contents of the news story may be disturbing. A debriefing statement immediately followed data collection. The debriefing statement reads: The news story you just read wasfabricated, not real, and any familiar characters or events are completely coincidental. If you feel personally distressed by participation in this stuafv, please do not hesitate to contact the researcher. Contact information for the primary researcher then followed. ')J H Data Analysis 0 H1: A news story that is high in distinctiveness, consensus, and consistency will increase defensive attribution (m - u; ¢ 0). 0 H2: There is a significant negative correlation between defensive attribution and approval of characters in a news story that differs in distinctiveness, consensus, and consistency (R< 0). After reading, participants were asked to rank each of the characters in the news story they read by the characters’ level of responsibility for the crime that was committed. The next questionnaire asked participants to identify their feelings toward, that is, agreement with the actions of, each of the characters in the news story (see Appendix B). Data were analyzed using a Mo-independent-samples t-test via SPSS. The mean scores of defensive attribution for the two randomly assigned groups were compared. This was done by assigning a value to the score of defensive attribution. That is to say: a score of l (i.e., most responsible for the crime) received a value of 6; a score of 2 received a value of 5; 3 received a value of 4; 4 a value of 3; 5 a value of 2; and, a score of 6 received a value of I when entering raw data for statistical analyses. For example, a participant who ranked the victim of the crime as being the 3rd most responsible for occurrence of the crime received a raw score attribution ranking of 4. The raw score from each participant was summed with the scores for all participants in the respective group, and a mean for each group was obtained. The means of the two groups were then compared using a two-independent-samples t-test. Additionally, it was hypothesized that there is a significant negative correlation between defensive attribution and approval of (i.e., agreement with) characters in a news story that is high in distinctiveness, consensus, and consistency (R