.3... . 2......213 . gt 2;... .1. x . Ami}. . _ . ,ufifl... x F... . 1.: . . a ‘ s. 34?. 3.9.. .. E a. .34 . c. . v3“ fillet. mg in ti. air. .I It I: , mnraufirav. can? 1 In «a J.- .:.Jugmy. 3w .“mFEwNHma :i . fmsfiflww «a...» 1.»... i» negubmumm i.» I. .r I \ ‘n . i’ "as? i t ‘1. ,1}:- :._lq. ....F..... I 1. .3: 4 .ul . I In! . '. N. A ‘ a. x $154.." L : Fffi .. :3 >4- H. is: p. 4...! r? .: I: a . ITS: .353 flHllilllllllIllllllHUNlllllllllllllllllllllllllllllillllli 31293018 This is to certify that the dissertation entitled The. LH“$(om 9-9 KKOwEKS: 1"- R"! "F Percuveé set-F —\Lmto\e&e(- TM News Putin» wayke‘t During om (ctd’wncwa 16.. presented by CkeuMé— Y: Wu r-K has been accepted towards fulfillment of the requirements for ?L\- D. degree in Whit,“ Major professor Date J0 1J1: I“ quqfi MS U is an Affirmative Action/ Equal Opportunity Institution 0-12771 'v4 - ‘vw w _v‘.-v_ __.. ,__._. LIBRARY Michigan State University 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 we WWW“ THE ILLUSION OF KNOWING: THE ROLE OF PERCEIVED SELF -KNOWLEDGE IN NEWS MEDIA IMPACT DURING AN ELECTION CAMPAIGN BY Cheong-Yi Park A DISSERTATION Submitted to Michigan State University In partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Communication 1999 ABSTRACT THE ILLUSION OF KNOWING: THE ROLE OF PERCEIVED SELF-KNOWLEDGE IN NEWS MEDIA IMPACT DURING AN ELECTION CAMPAIGN BY Cheong-Yi Park This study has attempted to describe and explain the totality of news media impact by proposing the Illusion of Knowing model which consists of five elements: media exposure, perceived-self knowledge, personal and social involvement and factual knowledge. The Illusion of Knowing model is rooted in the assumption that news public generally monitors the world via news media and selectively comprehend and stores some of the information imparted by them. In this model, the nature of news media impact is seen as both process and consequence. The surveillance and information functions of news media are presumed to be respectively contributive parts of public opinion process at a societal perceptional level and learning consequence at an individual’s cogitive level. The powerful impact of news media on learning is least expected because even if they make a signficiant contribution in monitoring stage of news information, they give it away to the personal experience and predispositions in its storing stage. Interestingly, the role of news media in public opinion process is presumed to be enormous, based on the same reason. The discrepancy between perceived self-knowledge formed in monitoring stage and factual knowledge originating from storing stage might be the source of apparent news media power: The Illusion of Knowing. Supposed that people are uninformed but misperceive themselves as well informed. This illusion might lead them to being inert and remaining with the information “accidentally acquired” from news media. Keeping it in mind that the every piece of the “accidentally acquired” information is already compromised by news framing, we could see why news media are effective to stir public opinion but less capable of imparting information. Ultimately, this speculation has been generally confirmed by the survey data collected in Michigan during the 1998’s gubernatorial election. The survey was conducted in the city of East Lansing in Michigan from October 26th to November 2nd, 1998. The research population consists of a representative sample of the adult population aged 18 years and older selected by random sampling method. Two hundred twenty six of 500 sampled households participated in the survey and 212 questionnaires were completed. In terms of mode of surveying, a composite version of mailing and face-to-face contact was employed. The survey data supported that as news media exposure increased, the illusion of knowing phenomenon increased—the gap between perceived self- knowledge and factual knowledge. Also, people socially involved misperceived their knowledge level more than did people personally involved. As expected in this study, news media exposure indirectly influenced factual knowledge via perceived self-knowledge and social and personal involvement. THIS DISSERTATION IS DEDICARED TO MY DEAR PARENTS: aae&qee iv ACKNOWLEDGEMENTS I would like to acknowledge those persons who assisted me in completing this dissertation project and at the same time helping me remain sane. My primary thanks and gratitude go to my committee: Dr. Atkin (Chair), Dr. Dearing Dr. Salmon, and Dr. Tamborini. Without their help, I would never have dreamed of completing this dissertation. More importantly, I would like to thank them for being examples of professionalism and scholarly spirit that I hope to emulate throughout my career. Back to Korea, there are persons whose encouragement and support never failed me: my parents, mentors, and friends. Words can not express the gratitude that I have toward my parents, Pan-Seok Park and Hi-Soon Lee. Also, I am particularly grateful to Dr. U-Ryong Kim and Dr. In-Suk Chung for their professional advice and emotional supports. Another thanks go to my friends, especially Young-Jae Shin, J ung—Suk Lee, Young-Soon Yun, and Byong-Kuk You who hold special places in my heart. Finally, I thank God for the opportunities He has blessed me with. I also thank Him for the strength and endurance that I have so greatly needed. TABLE OF CONTENTS LIST OF TABLE LIST OF FIGURES CHAPTER 1 NEWS MEDIA AND PUBIC 1.1 Introduction 1.2 Literature Review 1.2] Research Traditions of News Media Study 1.22 The Nature of News Public: Passive but Active CHAPTER 2 WHILE YOU ARE FEELING INFORMED: THE ROLE OF PERCEIVED SELF-KNOWLEDGE IN NEWS MEDIA IMPACT 2.1 Understanding of Perceived Self-knowledge 2.11 Two Independent Memory Processes Model: Monitoring and Retrieval 2.12 Perceived Self-knowledge: Familiarity of Judgement Operator 2.2 Main Constructs in the Model of “Illusion of Knowing ” 2.21 Media Exposure 2.22 Perceived Self-knowledge 2.23 Social Involvement and Personal Involvement 2.24 Factual knowledge 2.3 The Predictions from the Model of “Illusion of Knowing ” 2.31 Perceived Self-knowledge 2.32 Being Involved 2.33 Acquisition of Knowledge 2.34 Illusion of Knowing CHAPTER 3 METHOD 3.1 Survey and Sampling 3.2 Measurement 3.21 Main Variables 3.22 Control Variables 3.3 Statistical analysis 3.31 Hypothesis Testing : Multiple Regression 3.32 Model Testing: Path Analysis vi CHAPTER 4 RESULTS 4.1 The Representativeness of Sample 4.2 The Validity and Reliability of Measurement 4.3 Hypotheses 4.4 The Illusion of Knowing Model: Consistent with Data CHAPTER 5 DISCUSSION: ILLSUION OF KNOWING 5.1 Theoretical Significance of the Study 5.2 The Implication of the Empirical Findings 5.3 Limitations 5.4 Conclusion APPENDIX: SURVEY QUESTIONNAIRE BIBLIOGRAPHY vii TABLE 1 TABLE 2 TABLE 3 TABLE 4 TABLE 5 TABLE 6 TABLE 7 TABLE 8 TABLE 9 TABLE 10 TABLE 1 1 TABLE 12 TABLE 13 TABLE 14 LIST OF TABLES Representaiveness of the Sample Validity of Media Exposure Validity of Perceived self-knowledge Validity of Personal and Social Involvement Validity of Factual Knowledge The Effects of Media Exposure on Perceived self-knowledge The Effects of Perceived self-knowledge on Social Involvement The Effects of Perceived self-knowledge on Personal Involvement The Case of Heavy Media Users: The Upper 25 % of the sample The Case of Heavy Media Users: The Upper 10 % of the sample Social Involvement Among Heavy Media Users Personal Involvement Among Heavy Media Users The Case of Non-regular Media Users: The Lower 25 % of the sample The Case of Non-regular Media Users: The Lower 10 % of the sample viii TABLE 15 TABLE 16 TABLE 1 7 TABLE 18 TABLE 19 TABLE 20 TABLE 21 TABLE 22 Personal Involvement Among Non-regular Media users Social Involvement Among Non-regular Media users Social Involvement and Personal Involvement The Effects of Social and Personal Involvement on Factual Knowledge Socio-Demographic Characteristics of the Illusioned Persons Impact of Media Exposure on the Illusion of Knowing The Interaction Between Media Exposure, Involvement and Illusion of Knowing Assessment of the Goodness-of-Fit of the Model ix FIGURE 1: FIGURE 2: FIGURE 3: FIGURE 4: FIGURE 5: FIGURE 6: LIST OF FIGURES Analogy from the Structure of Directory Files in Computers Relationship Between Social and Personal Involvement Exposure is Power Being Involved Illusion of Knowing The Path Coefficients for the Model CHAPTER 1 News Media and The Pubic 1. Introduction The state of research on news media effects is one of the most intriguing to social scientists. The pervasiveness of TV and newspapers and their virtual monopoly over the presentation of news information must suggest that what the media say and how they say it have enormous social and political consequences. Nevertheless, the scholarly literature has been much better at refuting and circumscribing media impact than at supporting it (see McQuail, 1987, for the details). The early research on media power from approximately 1900s to 1930s proposed the hypodermic needle model which represented the conventional wisdom of the time, postulating a direct, immediate, and powerful media effects. This kind of model was rooted in the notion of audience passivity, where exposure to news information was equated to absorption in veridical form. The presumed power of media was tempered considerably by social psychology, from 1940s through 1960s, which emphasized audience motivation as the key intervening variable between media content and personal opinion. The focus was given to the dynamic of perception, underscoring audience autonomy and the effects of predisposition and social role. The attitude towards the minimal effect model was nicely conveyed by Berelson ’s utterance (1959) about the withering of communication research. However, the minimal effect model has waned somewhat in the light of several advancing scholarships such as agenda-setting and spiral of silence during the last three decades. The communication scholars studying the impact of watching or reading reported news have shifted their attention from media effects on attitudes and behaviors to on cognition and back to on affects and attitudes, and from short-term effects to long-term effects. Now, media effects turned out to be fugitive rather than minimal. The inevitably inconclusive and fugitive characteristics could be ascribed to the complexity of phenomena involved with the interaction between individuality, resulting from personal experience and predisposition, and media cognitive effects to do with knowledge and opinion. As Graber put it: “People who are exposed to the mass media already possess a fund of knowledge and attitudes which they bring to bear on new information. Since we do not know precisely what this information is, nor the rules by which it is combined with incoming information, we can not pinpoint the exact contribution which mass media make to the individual’s cognitions, feelings, and actions”(l980, p11). To add to the complexity, the initial role of mass media has been merely assumed to impart information or opinion to audiences. However, the research about the assassination of President Kennedy in 1963 gave us some insight about another plausible role of news media. They found that people were very well aware of the assassination (Greenberg, 1964) but in terms of the actual knowledge relevant to the assassination, their knowledge was considerably lower than expected (Spitzer & Denzin, 1965). News media might not directly increase factual knowledgel but perceived self-knowledge, self-impression of one’ own knowledge. Furthermore, perceived self-knowledge might be differently associated with one’s social perception of problems and with one’s personal concerns, in accordance with self-awareness in relation to media issue. The social concern might be the point where news media vastly exercise while the personal concern might be mostly played with individual experience, values, and interests. Along with this line of thinking, the author proposes the “ Illusion of Knowing” model, on the basis of the assumption of passive but active news public. The model’s overall dedication to the understanding of news media impact is made on the basis of the speculation that public can play actively in a certain situation even if they tend to passively respond to what the media say. If they get personally involved, they will comprehend and seek out news information as well as systematically estimate their knowledge level. If they get socially involved, they will remain with the “accidentally acquired” information from news media and estimate their knowledge level by the familiarity of portrayed issues for a “heuristic cue”. The model predicts some major relations. 1) As media exposure increases, perceived self-knowledge will increase. 2) Perceived self-knowledge will be associated with social concern (social involvement) than personal concern (personal involvement) will be. 3) Personal involvement will be the stronger predictor of factual knowledge than social involvement will be. Ultimately, the model delineates why and when media exposure increases the gap between perceived self- ¥ 1 Half of the viewers of newscast have only a superficial recall of about half of the stories broadcast (Robinson, & Levy, 1986). knowledge and factual knowledge—the illusion of knowing phenomenon. This study has started with proposing perceived self-knowledge and social and personal involvement as mediating variables between media exposure and knowledge acquisition and ended up with explaining the illusion of knowing phenomenon. It is worth seeing the illusion of knowing phenomenon from the viewpoint of the roles of news media in public opinion process as well as in learning. The consequence of news media’s increasing the illusion of knowing phenomenon is more likely to be associated with the context of public opinion while the process is more likely to be related to the context of learning. Supposed that the public might be uninformed but misperceive themselves as well informed, what can be happening? The illusion of knowing might lead them to being inert, being assertive but not rational, and being manipulated by news media but believing in that they are too intelligent to be swayed by news media. It delineates why news media are effective to stir public opinion but less capable of imparting information. The illusion of knowing gears up media’s exerting power. In the present study, this speculation has been developed in empirical questions and theoretical explanation. First of all, the author has reviewed the literature on news media impact from the vieWpoint of passive but active news public in order to help the general understanding of the illusion of knowing phenomenon. Secondly, a newly constructed model, or the illusion of knowing model has been proposed. Thirdly, the hypotheses deducted from the model have been tested with the survey data. Finally, the findings and implications have been discussed. 1.2. Literature Review The core of news media effects study is given to the assumption on news audiences. In earliest versions of effects research, news media effects on the bare individuals of mass society were held to be fairly direct and unmediated. As media effects research had developed the assumption about audiences has become increasingly sophisticated, moving away from the simplicities of mass audiences. Audiences have turned out not to simply consist of a mass of separated individuals; rather they are made up of complex and interacting social groups. Now, the audiences are presumed to interact with news media content in order to create its social meaning. The conceptual development of news audiences is well reflected through the lenses of mass audiences of spiral silence hypothesis, news consumers of uses and gratification approach, and news public of agenda-setting model. 12] E II l'l' [Ii M 1' SI I ! I-Ill MIISN Bll' The media’s agenda-setting role2 has drawn much attention from the two distinctive perspectives since the early 70’s: McCombs & Shaw (1972)’s static appraisal of media’s agenda-setting role in public opinion formation and Cobb and Elder (1972)’s dynamic comprehension of media’s agenda-setting role in political process which has been inherited by Dearing and Roger (1997). These two models 2 According to Dearing and Rogers(1997), the credit for the fore runners of agenda- setting research should be given to Lippman(1922), Park(1940), Almond(1960), and Cohen(1963). are rooted in different assumptions about news audiences. In McCombs and Shaw’s agenda-setting model, news audiences are assumed to be passive news public while Dearing and Roger’s agenda-setting model identifies news audiences as mostly inactive but occasionally active news public. McCombs and Show’s model regards news audiences as only being aware of the social and consensual meaning of news which news media implant. Unfortunately, the news audiences don’t look like being involved with the social process of creating consensual meaning of news. They still behave like a mass, freeing from individual characteristics. Contrast to it, news audiences in Dearing and Rogers’ model attempt to influence public policy for their own benefits as well as to weight news media content over their personally relevant experience. That is, they behave collectively for common goals, which means they are more like a public rather than a mass. The difference is clearly shown in their working mechanism. We]; McCombs and Shaw (1972) have introduced the term of agenda setting in mass communication by finding the almost perfect rank-order correlation between media agenda and public agenda in the 1968 Presidential campaign. The McCombs and Show’s traditional idea that news media set agenda among the public through the mechanism of salience has been extended to the priming and framing effects of news media content. The extensions respectively result from the applications of mechanism of salience to cognition effects and discourse analysis. News media are found to prime audiences to give media agenda more weight in their overall evaluation of political candidates (Iyengar & Kinder, 1987). Several studies (Erbring, Goldenberg, & Miller, 198; Kronsnick & Brannon, 1993; Pan & Kosicki, 1997) have showed the possible linkage between agenda-setting hypothesis and cognitive priming theory as to analyze the effects of news media’s agenda on people’s evaluations and opinions of political leaders. That is, people pay more attention to media agenda and use them when evaluating political leaders or social problems. Finally, McCombs and Show (1993, 1997) think of priming as the extension of agenda-setting model. To be in accord, Willnat (1997) argues that the integration of agenda-setting and priming will allow a better understanding of how mass media not only tell us “ what to think about”, but also “ what to think”. On the other hand, the concepts of frame and framing incoherently have operated in diverse disciplines (Pan & Kosicki, 1993). However, in terms of media function, framing is regarded as the way that news media influence the public’s socio-cognition by selecting particular parts of reality while ignoring other aspects (Entman, 1993; Gamson & Modigliani, 1987; Giltin, 1980; Pan & Kosicki, 1993, Tuchman, 1978). McCombs argues that “ agenda-setting is a theory about the transfer of salience of the elements in the mass media’s pictures of the world to the elements in the pictures in our heads... The second level of agenda-setting further suggests that the media also tell us how to think about some objects (1997;1-3)”. That is, salience of a specific way of seeing a given reality is in the spotlight at a second level of agenda-setting metaphor so that media framing turns out to be another type of agenda-setting effects. - : On the contrary to the extensions of agenda-setting model which focus on the semantic relation between media agenda and public agenda, Dearing and Rogers (1997) have inherited Cobb and Elder(197l)’s idea of agenda building and developed the full-fledged model for the dynamic process of media’s agenda-setting role in a real world. Dearing and Rogers’s model (1997) is definitely rooted in the concept of process and in the appreciation of the role of media in social change. They regard media as one of the subsystems of social system. This model sets off with the notion of gate keeping. The public interest groups (Pertschuk, 1987), the trigger events themselves (Rogers, Dearing, and Chang, 1991) and influential celebrities( Kingdon, 1984 ; Chang, 1989), the policy makers(Berkowitz, 1992) and polling (Dearing, 1989; Trumbo, 1995) compete with one another in order to influence media agenda. If any issue or event gets through the gate of media attention, it turns out to be media agenda within the influences of real world indicators and personal experiences (Erbringer, goldenberg, and Miller, 1980). The media agenda set public agenda as shown in McCombs and Show’s paradigm. The public agenda have exerted their own power on policy agenda (Linsky, 1986; Whiteman, 1986). Meanwhile, the media agenda and the policy agenda have formed a reciprocal relationship in terms of the influence (Rogers, Dearing, and Chang, 1991; Trumbo, 1995). Ultimately, social change can be seen through the lens of agenda-setting process. These two perspectives are complementary rather than competitive. McCombs and Show’s paradigm brings up to literary interpretation and psychological explanation in the agenda-setting model while Dearing and Rogers’ model sheds light on how a society is working in relation to media. In general, agenda-setting research has some difficulty in sorting out the interaction effects between individuals’ personal experience and news media content even if recent research on agenda—setting function emphasizes the role of personal experience. The reason may lie in that agenda-setting research little clarifies when and how differently the personal experience and social collective experience work on the media-related behaviors of news public. SnimlflilenmndMamndienm A quarter of century has passed since Noelle-Neumann (1974, 1977) formulated the spiral of silence hypothesis that has been recognized as the cross- level explanation of public opinion formation (McLeod & Pan, 1991; Salmon & Glynn, 199?). With receiving critiques and recommendations as provocative as the model itself( Katz, 1982; McLeod, 1985; Salmon & Kline, 1985; Price & Allen, 1990), the spiral of silence research has focused on the accuracy of perception of public opinion and willingness to speak out along with several contingent variables such as issue obtrusiveness, society type, information credibility, involvement and self-efficacy (Glynn & McLeod, 1984; Lasorsa, 1991; Lin & Salwen, 1997; Major, 1997; Willnat, 1995). The contingent variables unknowingly question the assumption about the human’s social nature in spiral of silence hypothesis. What kind of situation sets people free from the fear of isolation is the unspoken common factor of the contingent variables. The more you are involved, the more your society is individualistic, the more the issue is obtrusive to you and the more you are self- confident, the more you free from the fear of isolation. The freedom of the fear of isolation finally let them speak out loudly, disengaging the spiral of silence. In this sense, it is clear that the phenomenon of spiral of silence should be colored by the assumption of passive mass audiences who are strongly influenced by the others’ opinion and lack of autonomy. Based on the assumption of passive mass audiences, the spiral of silence model has showed that the willingness of each individual to express his or her opinion is affected by his or her perception of what others think of an issue, ultimately changing the global environment of opinion. The tendency of one to speak up and the other to remain silent starts off a spiraling process which increasingly establishes one opinion as the prevailing one. It is by using their own “quasi-statistical sense” that individuals filled with fear of isolation try to estimate and forecast climate of opinion which ubiquitous and consonant mass media create. The individuals’ quasi-statistical sense is presumed to mirror media power. However, the accuracy of quasi-statistical sense is seemingly inconsistent in the spiraling process. Individuals are supposed to misperceive one louder voice as majority opinion during the first stage of spiraling process3, which means that their quasi-statistical organ should function poorly. However, when the misperceived majority opinion measures up to actual majority opinion at the later stages, their ’ According to Taylor (1982) and Shamir (1988) misperception is necessary to turn a minority group into a majority. 10 quasi-statistical organ functions appropriately by sensing the magnitude of the majority opinion. The inconsistency of the quasi-statistical sense can be ascribed to either it is actually on and off, or the information sources for climate of opinion are in the variation with the stages of spiraling process. The unsolved part keeps scholars in questioning when news media content as an information source has an influence on the quasi-statistical sense. 11 I G l'fi I' I II “Uses and gratifications” model has been started with a functionalist approach to mass communication from 1940s and 1950s (Katz, Hass, and Gruevitch, 1973)4 and developed into a psychological communication model during 1970s (Rubin, 1992). In the early 1970s, Katz and his colleagues examined audience motivations in the context of self-and—society relationships, developing typologies of the uses people made of the media to gratify their own psychological and social needs (Katz et al., 1973). Since then, audiences have been supposed to initiate the mass communication process by selecting specific media contents. The conceptualization of the active and goal-oriented audiences has taken out media effects from the uses and gratifications approach. It is consistent with Katz (1959)’s suggestion that a media message cannot influence a person who has no use ‘ The best known examples as the early functionalist studies were Cantril’s (1942) analysis of quiz programs and Berelson’s(1954) study of what the newspaper means to its readers. Lasswell (1948) suggested that media perform four functions: surveillance of the environment, social integration, the transmission of cultural heritages. Write (1960), added entertainment function. 11 for it. Finally, the concept of media users or consumers replaces that of mass audiences in uses and gratification research tradition. In 1980’s, some scholars have attempted to demonstrate the possibility of media effects in the context of uses and gratifications approach 5(Rubin & Windahl, 1986; Palmgreen & Rayburn, 1982). Especially, Palmgreen and his colleagues have demonstrated the process of how media gratification obtained are transferred into media effects, by studying on TV news viewing. According to them, the expectations about the gratification attributes possessed by a media object are related to the gratifications actually obtained from media consumption (Rayburn & Palmgreen, 1984). In turn the media gratification obtained provides feedback to influence future gratifications sought (Palmgreen & Rayburn, 1979, Palmgreen et al, 1980). Finally, the transactional process produces media effects (Palmgreen and Rayburn, 1985; Wenner, 1982; Wenner, 1986)“. In terms of news media effects in uses and gratifications approach, the key variable, or the gratification has been hypothesized as consisting of surveillance7, 5 Even if the author addresses Use and Gratification approach in relation to news media impact, entertainment programs such as quiz show and soap opera have been the main focus of uses and gratification studies. ‘ For example, a person might believe that NBC Evening News’ broadcast of the 96’ Olympics is entertaining and she might feel positively toward this attribute. Her expectancy-value judgments would yield a generalized orientation to seek the gratification from the broadcasting and a positive attitude toward the news show. Her satisfaction with the news show would depend on what gratification she seeks from the news show. That is, the gratification obtained depends on the type of gratification sought. For example, if she sought information gratification rather than entertainment gratification but she evaluated the NBC Evening News show as entertaining, the NBC Evening News show would yield less the gratification for her. 7 Surveillance gratification means individuals use news media for help in obtaining information (Blumer, 1979; Gantz, 1978) and for keeping track of current events (McLeod &Becker, 1974; Palmgreen et al., 1980; Wenner, 1986). 12 entertainment“, and communicatory utility’. The gratifications sought, supposed to cause selective perception and differential attention to media, result from individual predispositions while gratifications obtained lead up to media effects. That is, through the mechanism of gratifications, personal factors facilitate or inhibit media effects. Although it is compelling that the role of personal factors is adopted in the mass media effects, the personal factors still fail to be embodied in the association with media-related social phenomena. Because socially shared experience or expectation is to be immerged into personal factors in uses and gratification research, it has often disregarded the seemingly irrational behavior of the public as explained in spiral of silence model or collective response of the public as expected in agenda-setting model. 1221' III [M Bll"E 'llili The three-research tradition shows how crucial the assumption of news public has been in terms of describing media-related phenomena. The researchers of agenda-setting and spiral of silence have claimed the power of news media with positioning news public in somewhat passive role while media gratificationists have been the most outspoken in championing the intervention of personal context over the power exertion of news media, based on the assumption of active news ' Entertainment gratification reflects concerns with excitement of elections races(Blumer and McQuail, 1969; Gantz, 1978; McLeod & Becker, 1974, Wenner, 1977, Palmgreen, Wenner, & Rayburn, 1980; Wenner, 1986). It is identified by Wenner(1977). l3 consumers. The two antagonistic sets of assumptions about the way that individuals interact with news media have resulted in encouraging the theoretical division. I_ ”S...” . . ... ,. . . “Hr-h. I .. H. . '3 in» a” _ ..- The passive/ active dichotomy could be derived from Descartes and Hume. Descartes sharpened and took as the focal point of his philosophy the Platonic notion that ideas are the most important aspect of human existence, by concluding that one was certain of one’s existence if one was aware of thinking. Descartes proposed that imagination is mind contemplating body while intellection is mind contemplating mind. That is, we do act as if our minds operate by a different set of rules than does our body. For example, when we say, “ I think I will go to the football game tomorrow”, we generally believe we have stated an intention over which we have voluntary control. When we make the statement, “ I have a toothache”, we are making a statement of fact over which we believe we have no control. We can change our decision about going to the football game at any time, but we cannot change perception (imaging) of the objective quality of having a toothache. For Descartes, intellection is both the foundation of human existence and the manner by which the world is understood. However, Hume proposed that sensation is the origin of most human action and therefore studying the nature of sensation10 provided the most certain understanding human behavior. As for Hume, body or object is placed in a central ’ Communicatory Utility/ interpersonal utility gratification is derived from the perception of getting information as useful in discussion with others(Atkin, 1972; McLeod & Becker, 1974; Palmgreen et al., 1980; Wenner, 1986) . ‘0 As for Hume, impression is information from the senses that may be initiated by some external object or may occur by the influence of an external object. Thus, 14 place in his epistemology where mind becomes, in part, a product of the object rather than a process independent of it. It is the seemingly inconclusive foundation of passive/active dichotomy of human behaviors whether mind is imposed from the outside by the characteristics of objects that impinge on the senses, or mind is a condition of human existence that orders sense data more than it is ordered by them. Eventually, with the rise of interest in human cognitive facility in 1980s, the old mind-body problem evolved into whether cognition is a self-generating activity of the human agents or it is the product of a reinforcement sequence. The focus moves from mind-body (behavior—cognitive) dichotomy to viewing of human cognition as active, and responsible or as passive, and powerless in essence controlled by external environment. Again, there seems to be no exit for the old type of the dichotomy when we see the matter from the viewpoint of the locus of control. However, the contention between them can be negotiated with respect to ‘human reaction to’ the influence of environment. There is noting like a living organism without having its environment". The environment plays certain roles in the shaping of human reactions, regardless of whether we assume that human mind operate actively or passively. Rather, the matter is how significantly the human mind reacts to the environment. Hume’s conception of impression begins with sensation and not with an object. Sensation is said to be equal to imagination for Descartes. 15 Human beings have reacted to their environment in the ways of whether being passively influenced by it or aggressively processing it. In some cases, human beings impose their own needs and interests on their environment and aggressively challenge it. They act for more than the simple gaining of gratification from the environment by bending it to their own needs. They are capable of even changing their environment when they feel like being deprived of their own interest. However, in the other cases, they merely dance around the environment. If they perceive the environment as not influential on their own personal life and interest, and thereby they see it from the viewpoint of the generalized others”, they appreciate it and remain as spectators rather than being aggressive. That is, it is the case that the self-interest oriented aggressive conduct appears when the generalized others-oriented passive conduct breaks down. Their tendency to efficiently respond to their environment let them both passive and aggressive to it. N l l' . . l | I' '3 In the context of news media impacts, human beings’ passive and active reaction to their environment serves as the latent predictors of different types of news media impact: perceptual, cognitive and behavioral outcomes. " “ All living organisms are bound up in a general social environment or situation, in a complex of social interrelations and interactions upon which their continued existence depends (Mead, 1934, p. 228).” ‘1 According to Mead(1934), the generalized other means the organized community or social group which gives to an individual his unit of self. '3 Since human mind is defined as reacting to its environment, the pure autonomy of human mind, or self-generating activity is denied and reactionary active response to the world is adopted for the term of activity. 16 First of all, human beings are seen as news consumers when they are exposed to news media content. As news consumers, they assimilate new ideas from mass media but news public doesn’t emerge until the news consumers become concerned about a social problem and see it from the viewpoint of the clash of contradictory opinions. That is, the members of news public“ are defined as being aware of their opinion position on controversial issues and psychologically united by indirect communication through news media between individuals and their perceived interest groups. The news pubic is formed through media activity of an organized group designated to maximize the interests of its own members. In general, the members of the news public are more likely to passively react to media issues so that they perceive the media issues in the same way as the media portray. They are whirling around the media. However, the members of the news public are not hopelessly passive but they play as the purposeful, reasoning actors within the influential spiral of media only when they have perceived the media issues as personally important and significant. That is, they select from media the information that fills their needs. Furthermore, they can seek out the information from the diverse sources in accordance with their own needs. They also aggressively attempt to change media agenda. Finally, in terms of news media impact, the following speculations are given. When news public is active, we could find the cognitive and behavioral effects of media such as selective attention, information-seeking intention, the increment of " The conception of news public is based on Blumer’s conception of public. According to Blumer(1957), the conception of the public as internally divided into factions is essential. 17 the actual knowledge, and readiness to behavior. When the news public is passive, there could be the powerful impact of news media on their perception. l8 CHAPTER 2 While You are Feeling Informed: The Role of Perceived self-knowledge in News Media Impacts Despite assertions about the importance of news media in imparting information to audiences, a substantial body of empirical findings suggests this is not the case. Public knowledge about some of the most publicized issues is often low (Bennett, 1988) and the correlation between newspaper reading and public affairs knowledge is, at best, moderate (McLeod et al., 1979; Robinson and Davis, 1990). In addition, prior political knowledge (Price & Zaller, 1993) and cognitive skill (Neuman, Just, & Crigler, 1992) turn out to nullify the impacts of newspaper exposure on knowledge acquisition. News media exposure might not directly increase factual knowledge but enhance perceived self-knowledge, or the self-impression of one’s own knowledge. Taking consideration that individuals’ information seeking about public issues is limited by their feeling adequately informed (Graber, 1988; Popkins, 1991), the effects of media exposure on knowledge acquisition would be mediated by the extent to which they are well informed about them. Thus, the self-impression about the one’s knowledge plays an important role in information processing (Schachter, 1983; Park, Gardner, & Thukral, 1988). The findings that news media have little influence on political learning may be ascribed to the lack of appreciation of perceived self-knowledge. In a learning process, perceived self-knowledge is mediating the impact of news media on knowledge acquisition. The mediating role of perceived self-knowledge results in 19 the illusion of knowing phenomenon, contingent with the public’s way of involvement with media issues. That is, people socially involved with issues portrayed in news media might not be willing to take cognitive efforts to comprehend the issues and thereby learn little from news media but misperceive themselves as well informed, due to the publicity. They will be illusioned in terms of their own knowledge level. The role of perceived self-knowledge is the very core in the consequence and process of the illusion of knowing model. 2. 1. Understanding of Perceived self-knowledge In psychology, the role of perceived self-knowledge has been studied in a wide variety of information processing tasks, including insight problem solving (Metcalfe, 1986), recognition (Scheachter, 1983), and choice task (Park, Gardner, & Thukral, 1988). Some findings indicate that there is often a discrepancy between actual knowledge and perceived self-knowledge (Lichtenstein & Fischoff, 1977; Schatacher, 1983). Most of these studies have been rooted in the assumption that perceived self-knowledge might originate from the capability of recall as earlier researchers suggested (Hart, 1965, 1967; Brown & McNeil], 1966): it might be simply based on the strength of memory (Schacter, 1983). However, it seems to be based on something besides retrieval memory. The ability of learning and change is not simply a function of actual knowledge; it depends. on perceived self-knowledge (Bransford 1979; Glenberg et al., 1982). Perceived self-knowledge is associated with how much people are familiar with the asked questions rather than how much they actually know the answers (Reder & 20 Ritter, 1992). Perceived self-knowledge seems to result from monitoring on whether or not a certain bite of factual knowledge is, without computing every bite of information. WWW There is high probability that people’s ability to monitor and judge their own knowledge which, in turn, influences perceived self-knowledge might be different from their ability to retrieve the knowledge which, in turn, leads to factual knowledge. This probability is founded on the two independent memory processes framework in terms of the relationship between judgement and memory (see Nelson & Naren, 1990, for review). The speculation is much clearer when the analogy is drawn from the organization of information in a computer. As seen in Figure 1, the names of files stored on a computer disk are normally listed in a separate directory file. When people want to store a specific file on the disk, they first consult the directory to see which part of the directory the file is well fitted into. If they want to check whether a specific file exists on the disk, they need only consult the directory to see whether it contains the name of the file, without having to access the content of the file itself. Consulting a directory is the first step of storing and retrieving information. This analogy implies two things: two independent memory processes (the directory file and the content file) and the immediacy of judgement operators (the directory) over specific information content (the content of the file). 21 C: My Document I Familiarity I Figure 1: Analogy from the Structure of Directory files in Computers Put it differently, when news information is available, it is encoded into a representation in working memory. During its stay in working memory, the news information serves as input for two independent processes; information is transformed and written into long-term memory traces (the content file) and information is operated on to produce judgement conclusion in working memory (the directory file). When people want to retrieve the specific content, they briefly tap in working memory and then search long-term memory: people’s retrieval capability is performed (e.g. going to the content file). When people want to store information or to retrieve if the information is available, they consult the directory in working memory: people’s monitoring ability is performed (e.g. going to the directory file). 22 ., ._W ”fun”, ,.....| o.-:. During such monitoring performance (e.g. consulting a directory), the information might be used as input for a judgement operator. The information might be placed within a directory in such a way that the information that is more relevant to a judgement operator has more advantages in subsequent monitoring performance. If the information is the higher relevant to the judgement operator, it can be recognized or recalled later better than those of the others because the information stored in long-term memory is filtered through the lens of the judgement operator. Perceived self-knowledge is based on the familiarity of a judgement operator that is induced by the frequency of the exposure to a certain object (Koriat & Lieblich, 1977; Reder & Ritter; 1992). That is, 1) familiarity functions as influencing the judgement of the level of factual knowledge during monitoring performance (i.e. an operator); 2) the level of familiarity is in variation with the frequency of presentation of a certain object. 2.2. Main Variables in the “Illusion of Knowing” Model In the “Illusion of Knowing” model, five variables are discussed in terms of how and when passive and active news consumption occurs: media exposure—it is compared to media attention, perceived self-knowledge, social involvement, ' personal involvement, and factual knowledge. This model is the product of an attempt to integrate the research traditions about news media on the basis of the assumption of passive and active news public. 23 WW The mixture of media exposure and media attention” has been widely used as the independent variable to predict a wide range of presumed media effects without the consideration about their difference. Based on the argument that media attention is the better predictor in the studies on the impact of media on public knowledge" (Chaffee & Scheuder, 1986; McLeod & McDonald, 1985), the mixture of media exposure and media attention seems to gain the legitimacy as the measure of the independent variable for the study of media impact. However, there is a clear possibility that media exposure and media attention are differently related to individuals’ message receptivity, considering that they are founded in different assumptions about audience activity. Media exposure to news has been the focus of sociological media research tradition" such as agenda-setting theory, spiral of silence model, and third-person effects model. The sociological media research regards the audiences of news media as essentially passive and manipulated, with the notion of media as acting agents in a society. " Usually, in terms of media exposure, respondents are asked the amount of the time spent generally watching and / or reading news media. The measure of media attention is about the amount of attention to specific news content. “ Unlike McLeod and Chaffee, Bartels(1993) argues that the nonfindings of the relationship between media exposure and political knowledge are attributed in part to the measurement error. '7 Recently, this research tradition accepts the intervening effects of interpersonal communication, individuals’ experience and knowledge, and the subjective perception of a real world, leading to the conceptualization of the less passive news public (Lasora & Wanta, 1990; Pan & Kosicki, 1997). 24 On the other hand, media attention to news media has been the independent variable in psychological media research approach18 such as uses and gratifications theory rooted in the assumptions of active audience and selectivity. It argues that from the functionalist viewpoint (Write, 1959), people bend media to their own needs more readily than media overpower them. That is, media attention is motivational state whereas media exposure is not. The point is that media exposure and media attention are different constructs, which means that they have excess meaning beyond their empirical operationaliztions and that they are the different types of media consumption behaviors. Media exposure and media attention are more than hours spent watching or reading and amount of attention given to specific news content. People’s media exposure depends on the availability of media in their external environment while their media attention results from the internal state of their mind. Some research findings indicate the likelihood that media attention and media exposure are different types of media consumption behaviors. Compared to unobtrusive issues, people pay more attention to news content about obtrusive issues (Erbring, Goldenberg, & Miller, 1980; Demers, Craff, Choi, Pessin, 1989; Iyengar & Kinder, 1987; Pan & Kosicki, 1997) with the implication of greater media impacts. However, the effects of media exposure decrease as the obtrusiveness of an issue increases (Atwater, Salwen, & Anderson, 1985; Zucker, 1978). The research findings are seemingly preposterous in the context of news media impacts but '8 Individuals’ needs, motives, and media-related gratification may be undermined by the their social and psychological antecedents(McLeod and Becker, 1981). That 25 innately logical if it is accepted that media exposure is on an awareness state while media attention is on a motivational state. Media attention concerns the interaction between a message and human mind (Grimes, & Meadowcroft, 1995), requiring a certain amount of cognitive efforts towards information. Considering the limited capacity of human memory, the selectivity is the nature of media attention (Levy and Winndahl, 1985; Lin, 1993; Rubin & Perse, 1987). The selectivity of media attention is influenced by the motivation of news consumers to dissolve environmental uncertainty as well as by news consumers’ prior knowledge, and personal bias (Anderson & Lorch, 1983; Atkin, 1973; Lang, et. al, 1993). News consumers as purposive actors select specific news content among other news contents (Donohew and Tipton, 1973; Bogart, 1981). News media attention postulates a goal-oriented and selective media use already mediated by news public’s predisposition. The mediation demarcates media attention in terms of subjective motivational asset. The subjective motivational property of media attention is presumed to lead up to cognitive efforts. On the other hand, media exposure is assumed as relatively nonselective and non-motivational. It can be defined as individuals’ non-purposive media use by focusing on more habitual patterns of either exposing themselves to a type of news medium viewing or reading a type of news content rather than individual messages. There is the distinction between general media exposure and specific exposure to a certain type of news content. Specific exposure turned out to be the stronger predictor of media-related outcomes (Einsidel et. al, 1984; McLeod & is, there is also the possibility of undermining the voluntarism of the news public. 26 McDonald; 1985) while general exposure showed the low relationship between media exposure and knowledge acquisition (Tan & Vaughn, 1976, Mun, 1987). In this present study, media exposure is construed as general exposure to each news medium. The conceptualization of media exposure apparently includes the meaning of habitualness: it is the habitualized reaction of news consumers to their own milieu. The media exposure of news audiences is much more likely to be circumscribed by the accessibility and availability. Media exposure may be an environmental existence that has the property of cause rather than that of outcome while media attention may be the outcome of the mediation by personal tendency and the overblown public opinion. Ultimately, media exposure is considered the original source of news media impact. WW: In the study of news media impact, most of research has focused on appraisal knowledge rather than impression knowledge (i.e. perceived self-knowledge). Appraisal knowledge can be defined as one evaluated with an imposed criterion, including factual knowledge/belief knowledge (Salmon, 1985), correct knowledge/erroneous knowledge (Y ows, et al., 1991), and information holding] correct knowledge (Clarke & Kline, 1974). Appraisal knowledge in general can be well represented by Salmon’s distinction between factual knowledge and belief knowledge (1985). The factual knowledge indicates discrete bits of information such as names, dates, and facts that 27 are measured as correct answers by researchers whereas belief knowledge is what respondents believe they know about a particular issue. Especially factual knowledge has been favored for media-related outcome in the studies of news media impacts (Chaffee, Zhao, & Leshner, 1994; Rhee & Cappella; 1997, Becker & Dunwoody, 1985 ; Wanta & Elliot, 1995). On the other hand, perceived self-knowledge focused on in the present study, is produced by human agents’ impression of their own knowledge (Holland, 1995; Mondak, 1995; Salwen & Driscall, 1995). Perceived self-knowledge can be differentiated from appraisal knowledge. First, perceived self-knowledge results form people’ ability of monitoring their own knowledge while appraisal knowledge is based on how well people can retrieve whatever they believe or know. Second, appraisal knowledge actually requires knowledge holdings while perceived self-knowledge doesn't. Perceived self- knowledge might be constructed without any concrete knowledge. Third, perceived self-knowledge can not be judged as either right or wrong. It only reflects the self- impression of one’s own knowledge. Perceived self-knowledge is defined as the knowledge which people perceive as well-known and familiar to themselves regardless of how much and what they actually know. Even if individuals hold no knowledge about a particular issue, they can feel informed, based on familiarity to it. 28 WWW Definitinnaflnmlxement Throughout a half of a century, involvement has been conceptualized and operationalized differently by different theorists and researchers (Early & Chaiken, 1993). The inconsistent research results due to the various definitions of involvement have led some scholars to attempt to typify involvement. Salmon (1986) categorized involvement as a personality trait, an internal state, salience of a stimulus, and a stimulus property. Johnson & Eagly (1989) suggested three types of involvement such as value-relevant involvement, outcome- relevant involvement, and impression-relevant involvement. A meta-analytic review of involvement literature has also revealed that involvement has been regarded either as cognitive activation or as a motivational state from the standpoint of the consequence of the interaction between a recipient and such an external stimulus as a message, an event, an issue, a product, or object. A listing of motivational terms for involvement includes salience (Roser, 1990) importance (Bybee, 1978; Perloff, 1984; Zimbardo, 1960), attitude extremity (Perloff, 1989; Lo, 1994), commitment to a topic (Chaffee & Miyo, 1983; Rothschild & Ray, 1974), and personal relevance (Flora & Maibach, 1990; Krugman, 1965). The involvement terms as cognitive activation are interest (Atkin et al., 1976; Kronsnick & Brannan, 1993; Zimmer, 1981), information seeking (Chaffee & McLeod, 1973; Flora & Maibach, 1990; Williams, Rice, & Rogers, 1988; Rothenbuhler, 1991), thinking (Chaffee & McLeod, 1973; Flora & Maibach, 1990; 29 Rothenbuhler, 1991), attention (Greenwald & Leavitt, 1984; Perse, 1990; Rothschild 8; Ray; 1974 ), and recognition (Greenwald & Leavitt, 1984; Perse, 1990). The present study starts by attempting to define the position of involvement more in relation to the broad set of motivational constructs rather than to the set of cognitive activation constructs. Motivational terms such as salience and perceived importance are necessary and sufficient components in explaining the impact of news media on peoples’ involvement while cognitive activation terms including attention and knowledge are sufficient but not necessary. That is, the presence of cognitive activation indicates there is some sort of involvement, but even without the presence of cognitive activation, there is the presence of involvement. For example, even if you pay some attention to an issue of global warming, your attention might be due to your intelligence, not your issue involvement. However, the presence of motivational characteristics always requires the presence of involvement. Without motivation, involvement seems not to exist. Thus, the motivational characteristics of involvement might be the more essential, compared to the cognitive characteristics of involvement. Along with this line of thinking, the author adjusts Johnson and Early’s definition that involvement is a motivational state induced by the association between the self and activated attitude (1989), as to the specific purpose of the present study. Involvement is stipulated as a motivational state activated by the evaluative summary about the relationships between the self and the target audience to whom the self is presented. This definition actually moves the locus of involvement away from human’s internal state and close to external stimulus. 30 Furthermore, the author conceptualizes different types of involvement from the viewpoint of the self-concept: social involvement and personal involvement. The different types of involvement may exert differing effects on audiences’ perception and behavior such as knowledge acquisition. The typology of the self-related involvement is guided within the framework of the self-presentation theory, especially as it is captured in the perspective of the self-concept. W W Since the seminal writings of Mead(1934) and Cooley(1902), the self-concept has been getting great attention from social psychologists as well as from micro sociologists. Micro sociologists called symbolic interactionists have shed light on the idea that the content and organization of the self (e.g. identity) reflects the content and organization of society (Blumer, 1969, 1973; Goffman, 1959, 1967; Kuhn & McPartland, 1954; Merton, 1957; Turner, 1978). Within social psychology, most of the self-concept research have focused on self-efficacyl9 (Bandra, 1977, 1981), self-esteem(Cialdini, Borden, Thorne, Walker, Freeman, & Sloan, 1976; Hales, 1980, Rosenberg, 197s) and sen-consistency” (Festinger, 1957; Rockeach, 1973 , Greenwald, 1980), resulting from the study of motivation and cognition. '9 The idea of self-efficacy is grounded in the notion of human agency (e.g. active self). Such terms as mastery (Adler, 1927), intrinsic motivation (Deci, 1975), intentionality(Giddens, 1979; Taylor, 1979), and internal locus of control (Rotter, 1966) can be equated with self-efficacy. 1° Within social psychology, the motivation for self-esteem and self-consistency has been derived from the set of cognitive consistency theories such as balance theory (Heider, 1958; Newcomb, 1968), and dissonance theory (Festinger, 1957). 31 While sociologists have attempted to describe how self interacts with society, psychologists to explain the variance within self. Finally, social cognition scholars in psychology have hypothesized the relationship between within-self variance and between-selves-in-society variance, presenting the self-presentation theory. Generally, self-presentations are meant to establish a particular image of some personal characteristics or relationship in the eyes of another person” (Jones & Pittman, 1982; Baumeister, 1982). However, the primary audience for self- presentation is also oneself (Schlenker, 1980; Snyder, Higgins & Strucky, 1983; Greenwald & Breckler, 1985). The core of the self-presentation perspective seems to be concerned with the orientation of the presentation. Who is the primary audience of the presented self is the crucial point to make which facet of the self become salient: presenting the self to the outer audience (society, and in-or-out groups), and the inner audience (oneself). According to Greenwald and Breckler (1985), ego is related to four facets of the self: the public facet of the self (the unspecified outer audience), the private facet of the self (the inner audience), the collective facet of the self (the specified outer audiences- e.g. reference groups), and the diffuse self (no identifiable audience). Greenwald and his colleagues (1982, 1984, 1985, 1989) explains that one presents 2' The credit for the term of self-presentation might be given to Goffman’s impression management. He states that“ when an individual appears in presence of others, there will usually be some reason for him to mobilize his activity so that it will convey an impression to others which it is interests to convey”(1959, p. 4). 32 these different facets of the one’s self in accordance with situations”, in order to enhance the self-esteem 23. W: In the present study, involvement typology is rooted in the context of which facet of the self is involved while the self consumes news media content. Primarily, social self and personal self“ are respectively oriented towards oneself and others. Characterized with the dependency on social milieu, social self is clearly aware of the outer audiences including a society itself and social groups which are reflected on by societal norms and collective experience. Social self is built up with collective memory, demarcated in terms of historical-social consciousness. On the other hand, personal self characterized as relying on personalized context is the more responsive to the inner audiences such as the self’s own values, experience and interests. Personal self is ground on personalized memory not socially shared. 2’ The self “shifts from moment to moment as a function of motivational state and situation cues”(Jones & Pittman, 1982, p233). 23 Unlike Greenwald and his associates’ explanation, the different facets of the self seem to be associated with different motivations. The compliance or conformity relationship of the self to unspecified outer audiences (a society) has been usually examined within the frameworks of the needs for social approval, achievement of reward and power, self-enhancement, and self-affirmation (French and Raven, 1959; Kelman, 1961; Kiesler and Kiesler, 1969; Mikular, 1984, Moscovici, 1976, Tedeschi & Norman, 1985; Abram and Hoggs, 1990). The social identification relationship of the self to specified outer audiences (in-group members) has been considered within the frameworks of the needs for affiliation, self-esteem maintenance, and self-categorization (Kelman, 1961; Tajifel & Turner, 1979; Wylie, 1979; Baumeister, 1982; Swann, 1885; Breckler & Greenwald, 1986; Turner, 1987; Price, 1989; Hogg & McGarty, 1990). The internalization relationship of the self to oneself has been studied within the frameworks of the needs for cognitive consistency or balance, and self-regulation(Festinger, 1957; Heider, 1958; McGuire, 1960; Kelman, 1961; Kelly, 1967; Tesser, 1978; Judd & Kulik, 1980). 33 When the self interact with its environment such as personalized context and social milieu via news media content, the types of involvement will come to light. One is social involvement and the other is personal involvement. First of all, social involvement focuses on the idea that people want to anchor themselves to a society, resulting in seeking out social approval from the outer audiences. The antecedent conditions of the social involvement are 1) the awareness of the self from the viewpoint of collective experience, values and interests, and 2) the summary evaluation of the influential relationship of social milieu to the self. Thus, when people can monitor climate of social opinion, and when people as social units concern the social importance of a given event or issue, this type of involvement will emerge. It is a such situation that an issue is conditioned with society-oriented needs of an issue receiver. Secondly, personal involvement is associated with people’s desire to be in control. Under the estimate that a given event or issue will produce a certain impact on one’s own personal life, people are motivated to personally contextualize their own values, interests and experiences, leading to calculating the personal importance of a given event or issue. The antecedents are 1) the awareness of the self from the viewpoint of personal experiences, values, and interests and 2) the summary evaluation of the identity of the inner self. When people can estimate the importance of expected outcome, and when people personally concern the importance of a given issue or event, personal involvement will emerge. The interaction between an issue and human mind, 2‘ Social self and personal self are respectively equate to public facet of self and private facet of self in Greendwald’s work(1985). 34 especially conditioned with the self-oriented needs of an issue receiver, leads up to personal involvement. World 1'; Social Involvement | 1 I Nations : I l’\\ . I l. J Social Groups : : I : r '1 | I Close Relations \VI : I I : Personal Involvement Personal Self I: Figure 2: Relationship between Social involvement and Personal involvement Axiom 1: social and personal involvement can be expanded and contracted across different levels of the self’s environment in accordance with the evaluation of the relationship between the self and an audience that are either a society or inner self Axiom 1 implies that when the self-evaluation changes, the type of involvement also changes accordingly. The volatility of involvement is expected. Figure 2 illustrates axiom 1, or the contextual nature of involvement. In the figure, each line stands for the self’s environment, providing the frame of the reference for the self-evaluation. For example, A is aware of himself as a cosmopolitan at the world level, as a US citizen at the national level, as a male, and a MSU student at the organizational level (reference group level), as a son, a friend, and a brother at the close relationship level, and finally as individuated self that differentiate the self individual from others within a given social context. The each level can be associated with both social involvement and personal involvement. 35 Generally, as the self environment is getting downsized from a world or nation level to a personal level, the self is more responsive to the inner audiences and vice versa. That is, if you feel personally more involved with an issue or event, the power of your social self is getting weaker. Thus, your decision relevant to the issue might be influenced by your personal experience and interest rather than the collective interest on the behalves of nation or social groups. 22' ! 'W' [E I I]: ll The exposure to news media ultimately leads to the acquisition of factual knowledge in this model. The role of media on the acquisition of factual knowledge has been portrayed as insignificant in several studies (Neuman, Just, & Crigler, 1992; Patterson, 1980; Price & Zaller, 1993). In these studies, factual knowledge is measured as recognition of specific factual knowledge. Recently, some researchers measured audience understanding by counting such features of open-ended responses as the number of arguments, time frames, and causes and implications the person brings into the group discussion (McLeod, Pan, & Rucinski, 1989; Rhee & Cappella, 1997). However, the newly suggested measure of knowledge called as audience understanding tends to assess the impact of audiences’ schema rather than that of media exposure. Persons who have developed complex schema may learn more from news media than persons with simple schema. Learning from news media seems to favor highly educated persons who have already developed more complex schema. 36 Because the interventions of education and prior knowledge are regarded as the non-designed influences from the outside of the system of the “illusion of knowing” model, in this present study, acquisition of factual knowledge is measured as recognition of factual knowledge that means the discrete bits of information such as names, dates, and facts about an issue or object. 2.3 The Predictions by the “ Illusion of Knowing” Model To illustrate the theoretical development that underlies this study, Figures 3 through 5 show the progression of stages in the study of media exposure on acquisition of knowledge, leading up to the model of“ illusion of knowing.” It underscores the process of how a public processes news information. Media Exposure Perceived self-knowledge ——> Figure 3: Exposure is Power Hypothesis 1: Greater media exposure is positively related to greater perceived self-knowledge. Within the framework of two independent information processes model, news media effects can be understood as related to creating or modifying a familiarity judgement operator in a directory file at which monitoring performance occurs. People who more frequently expose themselves to news media might 37 develop the stronger familiarity with what media say than ones who don’t. The familiarity is the indicator of perceived self-knowledge. Heavy media users perceive themselves as well informed about what media report on politics, regardless of how much they actually know about it. Mondak(1995) found that exposure to local newspapers does not enhance knowledge of national or international politics, but contributes to perceived self- knowledge regarding local political campaigns. Holland’s data (1995) also indicated that exposure to TV and newspaper has the stronger correlation with perceived self- knowledge than factual campaign knowledge. The upper-hand phenomenon of perceived self-knowledge in news media efiects can be also seen through the lenses of agenda-setting, spiral of silence and cultivation theories. Without monitoring on what you know about the world that you belong in, you might have a difficult in developing the perceptions of what is the most important problem facing the nation, whether or not you comply majority opinion, and how dreadful the world is. The best feature of news media might be to provide a person with monitoring the world and self, increasing perceived self- knowledge. W Media Exposure . Perceived self-knowledge __>Social Involvement Personal Involvement Figure 4: Being Involved 38 Hypothesis 2A: 2-1: Perceived self-knowledge is positively associated with social involvement. 2-2: Perceived self-knowledge is positively associated with personal involvement. As media fuel people’s feeling informed, their heightened sensitivity to their own knowledge leads to the relative fluctuation between two types of self-awareness state: social self and personal self 25. In some cases, people are more oriented towards the personal self such as their own value, experience, and interests while they more consciously reflect on the social self rooted in collective value and experience in other cases. In the present study on news media, the relative fluctuation between social self and personal self is expected to depend on which memory“ is sought out. Specifically, the induction of personal and social involvement is associated with 1) the ability of individuals to immediately access to their own personalized memory on issues portrayed by news media and 2) the capacity of media to place issues or events into individuals’ collective memory. First, personalized memory is a contributive part of the self-impression of one’s own knowledge as introduced earlier. The personal self will be salient when people expose to obtrusive issues such as unemployment and inflation that enter into people’s personal life on a regular basis (Weaver, Graber, McCombs, & Eyal, 1981; Zucker, 1978). In this case, the self-impression about the amount of knowledge 25 The matter is about which types of second-level judgement operators have more strength with the first-level judgement operator, or familiarity. 2‘ For example, individuals’ memory of Gulf War in 1991 consists of two parts: collective memory and personalized memory. Personal experience such as having relatives and friends serving in the Gulf were stored in personalized memory while 39 holding (perceived self-knowledge) is positively associated with personal involvement because perceived self-knowledge is partly rooted in personal experience and interest. Secondly, news media can be said to construct the news public’s collective memory that is the house of their social self, or the antecedent condition of social involvement. Media tend to sensitize the socially shared knowledge and experience, or collective memory by promoting problems or issues to the status of social and political issues (McLeod, Becker, & Byrnes, 1974; MacKuen, 1981). Media exposure primes the importance of collective perception, not the importance of personal concerns (Mutz, 1992). Media may provide the social self with collective memory that is the basis of social involvement. Thus, the perceived self-knowledge associated with the content of news media is expected to show the positive association with social involvement. Ultimately, the relationships between social and personal involvement and perceived self-knowledge can be predicted because perceived self-knowledge is defined as originating from the two sources: collective memory provided by news media in a modern society and personal memory. media usually served for constructing collective memory such as positioning Gulf War as the most high-tech war in 20t century. 40 Hypothesis 2B: 2-3: Perceived self-knowledge is more strongly associated with social involvement rather than personal involvement within heavy regular media users. 24: Perceived self-knowledge is more strongly associated with personal involvement rather than social involvement within non-regular media users. As seen in Figure 4, media exposure increases perceived self-knowledge which, in turn, perceived self-knowledge will have impacts on personal involvement and social involvement. In the interpretation in the relative strengths of perceived self-knowledge to social involvement and personal involvement, it is important whether social self or personal self is salient. Since media content is oriented towards collective memory, it is expected that perceived self-knowledge have the stronger relationship with social involvement than personal involvement within heavy media users. When media report a war, it is portrayed in the light of the contrast and comparison of the past wars and other socially shared experience. That is, media tap immediately the news public’s collective memory rather than their personalized memory so that there may be no requirement of cognitive efforts for the monitoring of collective memory. In terms of cost-effectiveness, the initial monitoring will give its priority on collective memory only if heavy media exposure is guaranteed. Contrast to it, non-regular media users are expected to be more influenced by their personal involvement because they tend to rely on their personal experience rather than media for information sources. Personalized memory is presumed to be more easily accessible sources of information (Ball-Rockeach & DeFleur, 1976; Perloff, 1985) to non-regular media users. 41 2232. i! '! "Ii [K II Media ___> Perceived __> Societal Exposure Self-knowledge Invo vement Factual Knowledge Personal j Involvement Figure 5: Illusion of Knowing Hypothesis 3-1: Social involvement and personal involvement are positively related. As discussed in the section for conceptualization of the main variables, social involvement and personal involvement have motivational state in common, leading to their co-variation. Depending on the characteristics of external stimuli such as issue type, they are either negatively or positively related to one another. For example, in the case of the issue of banning smoking in bars and pubs, smokers might have the stronger personal involvement than social involvement while non-smokers might have the stronger social involvement rather than personal involvement. Since the bars and pubs are considered as only public places for smokers, the smokers’ self-interest would be high at the risk. Their social consciousness would be overshadowed by personal self-interest so that the social involvement and personal involvement might be negatively related. However, non- smokers would see the banning as one of the social outcomes of an anti-smoking 42 public campaign. The social involvement is prior to personal involvement. In that case, there might be positive relationship between social and personal involvement because there is no conflict between social self and personal self. Compared to the case of the smokers’ perception of the banning, election is more likely to be social matter crossing all population in a recent modern society, leading to the priority of social involvement to social involvement. That is, the personal self reflects on social self. As strongly as people perceive voting as social responsibility, they could transfer the social responsibility into personal matter. Thus, as social involvement increases in an election campaign, personal involvement will increase in an election. Hypothesis 3-2: Personal involvement is the stronger predictor of factual knowledge than is social involvement. Considering the self-interest nature27 of human beings, it is expected that personal involvement enhances learning rather than does social involvement. When people perceive the personal relevance to an issue, they engage in more issue- relevant thinking (Chaiken, 1980; Petty & Cacioppo, 1979). The more people are personally involved with an issue, the more systematically they process it. Due to their self-orientation, people will pay more attention to personally relevant media issues than socially important media issues, resulting in increasing acquisition of knowledge about the personally relevant media issues (Atkin & Gantz, 1975; Erbring, Goldenberg, & Miller, 1980; Robinson and Davis, 1990). ’7 Self-interest means that individuals do not directly care about the welfare of others. 43 On the other hand, the people socially involved with an issue portrayed in news media might be satisfied with the information “accidentally acquired” from news media since they might not strongly register the expected subjective utility of information; their level of the desire to comprehend might be low; ultimately they might prove to hold less knowledge than in the other case. 2 3 ' Ill . [H w' If media exposure has a direct influence on perceived self-knowledge, but not on factual knowledge, there should exist the gap between them. The discrepancy between perceived self-knowledge and factual knowledge may be the source of news media power. The biased estimation of knowledge level will have an influence on the public’s perception about news media power. When people might be uninformed but misperceive themselves as well informed, what can be happening? This illusion might lead them to being inert, being assertive but not rational, and being manipulated by news media but believing in that they are too intelligent to be swayed by news media. They might perceive themselves as invincible to news media impact. The invincible self might play a role in people’s devaluating news media impact. Compared to the invincible self, the perceived vulnerable self appears when people underestimate their knowledge level. In this context, the illusion of knowing means that people misperceive themselves as less informed than they are actually informed, leading to imaging themselves as vulnerable to news media impact. The vulnerable self might play a role in people’s exaggerating news media impact. 44 Either the exaggeration or devaluation of news media impact might be the forces in the scene of public opinion process. Research Question 1: What are the social demographic differences among the persons who either underestimate or overestimate their knowledge level? It could be interesting to compare the overestimation cases with underestimation cases of the illusion of knowing phenomenon in terms of socio- demographic characteristics: sex, education, income, and age. Especially, there might be some differences between grammar school graduates and college graduates in terms of the estimation of knowledge level. Hypothesis 4-1: As media exposure increases, the illusion of knowing will increase. News media are more likely to increase the gap between perceived self- knowledge and factual knowledge. The impact of news media might be more significant and substantial on the direct consequence (i.e. perceived self-knowledge) rather than the indirect final outcome (i.e. factual knowledge), especially considering another mediating process between perceived self-knowledge and factual knowledge. The level of perceived self-knowledge is the function of news media exposure while the level of factual knowledge is the combination function of news media exposure, perceived self-knowledge and the type and level of involvement. That is, as new media exposure increase, perceived self-knowledge 45 may proportionally increase but factual knowledge may not proportionally increase. This tendency may result in increasing the gap between perceived self-knowledge and factual knowledge. Hypothesis 4-2: The interaction between media aposure and social involvement has a significant effect on the illusion of knowing more than does the interaction between media exposure and personal involvement. Considering the self-interest nature of human beings, the illusion of knowing phenomenon is expected to be prevalent in such situations that the public is more socially involved with an event or issue portrayed by news media rather than personally involved with it. If the public gets personally involved with any issue, they will be more active to comprehend the portrayed issue in news media and thereby they will have the higher level of factual knowledge. Thus, the gap between perceived self-knowledge and factual knowledge will be less than in the case of social involvement. If the public gets socially involved with any issue, they will be inert in terms of comprehending the news information about it, and thereby they will have the lower level of factual knowledge, but higher perceived self-knowledge. That is, social involvement may be the stronger predictor of illusion of knowing than is personal involvement. 46 CHAPTER 3 Method The “Illusion of Knowing” model introduced in chapter 2 is applied to the explanation of the process of news media impacts in the context of an election campaign. The 1998’s Michigan gubernatorial election, held on November 3", serves as the applied case of the “Illusion of Knowing” model so that each variable in the model is modified within the electoral campaign context (e.g. factual knowledge means actual political knowledge). Ultimately, this chapter discusses the details about how to conduct a survey for the purpose of the present study. 3. 1. Survey and Sampling The survey28 has been conducted in the city of East Lansing in Michigan from October 26th to November 2nd, 1998. The research population consists of a representative sample of the adult population aged 18 years and older in that city selected by the random sampling method. Five hundred households were drawn from the 18779 residential addresses listed on1998’s East Lansing telephone directory; the selection of the respondents within a sample of households was based on the availability. First, the author used the computer phone directory program in the main library at Michigan State University in order to access to the addresses of households in the city of East Lansing. Second, the 18779 residential addresses of 1' Prior to the actual survey, the pilot survey was conducted for testing the reliability of and supporting the criterion-related validity of the measurement scales. The 60 undergraduate 47 East Lansing were transformed into data format and downloaded to an Excel spreadsheet. Third, the address data in Excel were transferred to the SPSS program for random sampling. Finally, the SPSS program randomly selected 500 households from the 18779 residential address. From the 500 sampled households, 226 participated in the survey and 212 questionnaires were completed. Fourteen questionnaires were eliminated because of their incompleteness. In terms of mode of surveying, a composite version of mailing and face-to-face contact was employed. During the period of the field dates, available adults at the households randomly selected were asked, in person, to complete survey questionnaires at any convenient time for them on that day and leave it either the doorsteps or mail boxes of their houses; the next day, the questionnaires left on the doorsteps or the mail boxes were gathered. If there was no questionnaire left outside the houses, research assistants posted on the front doors the messages to solicit the survey participation again. However, if there still was no questionnaire outside the next day, the research assistants attempted to solicit survey participation from next door neighbors. This data-collecting mode was devised to increase response rates by personalizing an unknown researcher and maximizing respondents’ convenience. In terms of the implementation of the survey, the internal validity and external validity of the data set were secured by the following. A) The period of field dates was less than seven days in order to prevent history threat. Thus, the students from Department of Communication at Michigan State University was administered the measurement scales on October, 13‘” for the sake of a pilot survey. 48 verification of when the questionnaires had been completed was checked. B) The representativeness of the sample was checked with the US census data of the City of East Lansing. 3. 2. Measurement First of all, the survey questionnaire consists of two sections: One is measuring the main variables and the other socio-demographic characteristics. Secondly, for the validity of the measurement scales, each construct has at least three indicators that is verified by internal consistency theorem. W W: Media exposure is construed as general exposure to a certain news content. It is the passive reaction of news consumers to their own milieu. In the present study, respondents are asked to indicate how regularly they get the news from a certain type of news medium on 1-to-7 point scale (1= Never and 7=Very regularly). The scores from the three medium-related questions are summed and divided by 3 to produce a composite score for media exposure with higher scores indicating more heavy media exposure to news media than lower scores. The three sources of the news are local TV programs, newspapers, and radio. mm: It is defined as the knowledge people perceive as familiar to themselves regardless of how much they actually know. The self- impression of one’s own knowledge is constructed from the viewpoints of personal 49 self and social self. Thus, the operationalization consists of two parts: other- oriented and self-oriented impressions of one’s own knowledge. Respondent are asked to indicate on a 1-to-7 point scale how much they feel informed of 1) this gubernatorial election, 2) the candidates running for governor, 3) the issues in this election, 4) other Michigan residents’ opinion about the candidates, and 5) other Michigan residents’ opinion about the issues in the election. The composite scores from 5 questions about the news of the election range from 1 to 7 with 1 indicating the lowest level of perceived self-knowledge and 7 indicating the highest level of perceived self-knowledge. Wu is defined as a social facet of motivational state activated by evaluation of the relationships between self and society. This measure is a 4-item, 7 point semantic differential scale that is summated and divided by the number of items. The possible scores range from 1 to 7. Respondents are asked to evaluate 1998’s gubernatorial election in Michigan in terms of significance, importance, relevance, and concern at a societal level. Wm: It is defined as a personal facet of motivational state activated by evaluation of the relationships between self and society. This measure is a 4-item, 7 point semantic differential scale with possible scores ranging from 1 to 7. Respondents are asked to evaluate 1998’s gubernatorial election in Michigan in terms of significance, importance, relevance, and concern at a personal level. 2’ The measures of social involvement and personal involvement come from the modification of Zaichokosky(1985)’s personal involvement inventory 50 W: The factual knowledge is measured as how many the public acquires the discrete bits of information about this gubernatorial election. Respondents are asked 14 questions relevant candidates’ issue positions and personal details. The scores range from the lowest score 1 to the highest score 14. mm It is defined as the knowledge gap between perceived self-knowledge or the self-impression of one’s own knowledge and factual knowledge. Thus, the discrepancy between them is used for a measure of the illusion of knowing. It is obtained by multiplying the scores of perceived self- knowledge ranging from 1 to 7 by 2.14 and then subtracting the scores of factual knowledge ranging from 1 to 14 from the newly constructed scores of perceived self- knowledge. Widener: There are two types of control variables: One set is about the variables relevant to perceived self-knowledge and factual knowledge such as education, campaign experience and interpersonal communication sources. The other is about the socio-demographic variables influencing social and personal involvement such as party identification, race, sex, age, income, and education. Sex: Respondents are asked if they are males or females. This study is interested in the possible influence of sex difference on social and personal involvement. Age: It is measured by directly asking respondents what year they were born. This study expects that the older a person is, the more he or she is involved personally and socially in the election. 51 WThe income level is categorized by eight groups: S 0-8 8,999; 8 9,000-517,999; S 18,000 -$ 26,999; 527,000-835,999 $ 36,000-853,999; $54,000- $71,999; $72,000$89,999; over $ 90,000. The income level is an indirect indicator of social economic status of an individual so that this study suspects the possible influence of SES on the level of social and personal involvement. mu is directly measured by asking respondents what are their racial identities. The racial identities are categorized as Caucasian, African-American, Hispanics, Asian-American, American Indian and others. There might be the difi'erence between racial groups in terms of the levels of personal and social involvement. WW: Political party identification has been regarded as important for understanding political behaviors such as voting, political participation and evaluation of political leaders. Party identification is supposed to have an influence on social and personal involvement. The categories of party identification are Democrats, Republicans, Independents, and others. Education; Education is measured by asking respondents how many years of formal education they have received. The score ranged from 0 to 20. 0 means no formal education and 20 stands for more than 4 year graduate study. The level of education is also considered to have influence on both the level of involvement and the level of factual knowledge. There might be the tendency of the educated persons to be more socially and personally involved in electoral participation. Also, they are said to develop the complex schema which make them better understand political knowledge. The better understanding might have influence on factual knowledge. 52 W“ is measured by modifying the variable of “age”. Value 1 is assigned to less than 21 year olds from 18 year olds, value 2 to less than 24 year olds, value 3 to less than 28 year olds, and value 4 to over 28 years olds. This scale was made on the consideration that John Engler has been a three- time candidate for Michigan gubernatorial election. That is, as for the first voters who ranged from 18 year olds to 21 year olds, any information about Engler might be new while Engler might be quite well known to 28 year olds. Over 28 year olds might know how to process election information as well as information about Engler better than anyone under 28 year old. Campaign experience looks to be directly related to perceived self-knowledge. The familiarity of certain knowledge developed through cumulative exposure to election campaign might make the experienced voters feel more informed than others. WWW Interpersonal communication has been considered one of the significant information sources that compete with news media. Also, it might be one element of the personal context of individuals. In order to measure the effects of interpersonal communication as an informational source, this study asks respondents to indicate on the scale of 1 to 7 points how regularly they get the news from friends and neighbors. 53 3.3 Statistical Analysis 3311! II '-I I' 'Mll'lB . ! I . Multiple regression analysis"’0 is employed for testing the hypotheses. This study attempts to examine the statistical significance of each regression model and the strength of association between variables in each model. In terms of examining the statistical significance of each model, there are two tests involved. First of all, F-test is used for testing that amount of variation explained by each regression model doesn’t occur by chance (i.e., R2 is grater than 0). R2 stands for the explanation power of a model. Secondly, t-test is employed for testing that the regression coefficients in a model differed significantly from 0 ( i.e., B is greater or less than 0). Regression coefficients provide a means of measuring the change in the dependent variable for a given change in an independent variable assuming the other independent variables are bold constant. R2 is the sum of regression coefficients times their respective simple correlation coefficient (Hair, J r., Anderson, Tatham, 1984). On the other hand, the strengths of the associations between variables can not be directly compared to one another due to the different units of measurement of the independent variables. Thus, the standardized regression coefficients of the independent variables in each model (beta coefficients, b) are used for the comparison of the relative strengths of the association between the independent variables and the dependent variable. 3° In terms of selection method, this study entered all independent variables into analysis in a single step. 54 However, beta coefficients are used only as a guide to the relative importance of the independent variable included in an equation, and only over the range for a set of sample data (Hair, J r., Anderson, Tatham, 1984). That is, beta coefficients will be meaningless with different sample data from an original data set which means, beta coefficients can not be used for hypothesis-testing. W In this research, the following equations are drawn in the hypotheses of this present study. H1: Y We...“ ,dnknowledg, = Constant + [31( media exposure) + [32 (education) + B3 (campaign experience) + B4( interpersonal communication) + error H 2-1: Y mm involvemem = Constant + [31( perceived self-knowledge) + [32 (age) + B3 (sex) + B4( race) + [35(party identification) + [36 ( income) + [37 ( education)+ error H 2-2: Y perm... involvement = Constant + [M perceived self-knowledge) + [32 (age) + B3 (sex) + [34( race) + [35(party identification) + B6 ( income) + [37( education)+ error H 2-3 : R2 of the social model is bigger than R2 of the personal model within heavy media users. Social Model: Y social involvement within heavy media users = Constant + [M perceived self-knowledge) + error Personal Model: Y pm“... ...”...mm. with... .....y "...“. um, = Constant + [31( perceived self-knowledge) + error 55 H 2-4: R2 of the personal model is bigger than R2 of the social model Within non-regular users. Social Model: Y social involvement within non-regular media users = Constant + [31( perceived self-knowledge) + error Personal Model: Y personal involvement within non-regular media users = Constant + [31( perceived self-knowledge) + error H 3-1: Y pen“... inmmmm = Constant + [3,( Social Involvement) + [32 (age) + B3 ( 88X) + [34( race) + [35(party identification) + Bo (income) + [39( education)+ error H 3-2: [31( personal involvement) > [32 (social involvementf’1 Y non... knowledge = Constant + error + [31( personal involvement) + [32 ( social involvement) H 4-1: Y the “...,ion of knowing = Constant + [31( media exposure) + [32 (age) + [33( education) + [34( sex) + [35( income) + error 3' Accoring to Neter, J and W. Wasserman( 1974), when tests about regression coefficients are desired that do not involved testing whether one or several b 2 equal zero, extra sums of squares cannot be used and the general linear test approach requires separate fittings of the full and reduced models. In this study, the full model (Y = C + [3. personal + [32 social + e) and the reduced model (Y = C +[3t (personal + social) + e where [3 , denotes the common coefficient for [3. and [32 under Ho: [31 = [32) are used for the following F test statistic: [SSEO’CdHCCd model)-SSE(full model)/df reduced model‘df fun model] / [SSEa'lIII “10(1er df fun model]- *SSE means error sum of squares. 3' This hypothesis is tested with General Linear Model procedure that can provide analysis of variance for one dependent variable by one or more factors and] or variables. GLM is useful for testing the interaction effects between independent variables. 56 H 4-232: Y .... mum" Orknowlng= Constant + ( media exposure) + error (media exposure X social involvement) + (media exposure X personal involvement) 332M “-1 |° 'E II! I' Since the study propose the model decomposing the causal relations between media exposure, perceived self-knowledge, social and personal involvement, and factual knowledge, it is necessary to employ path analysis technique. Path model generally requires over-identification and recusiveness to examine the system of the structural equations (Asher, 1976). The concept of identification refers to the relationship between the number of unknowns contained within the system and the number of linearly independent equations. If there are more equations than unknown, it is called overidentification. The overidentification situation produces a finite set of solutions for the determination of path coefficients with the property that the solutions generated will vary according to the equations used in their model. The recursiveness means that all variables in a model are measured. If the model is recursive, there should be no feedback loop and bi-direction. That is, the correlations between the error terms and endogenous variables are zero by definition, so that the unknown relationships produced by the correlations between them are eliminated, which means the equations are solvable. 57 The illusion of knowing model is overidentified with 3 linear correlations but not recursive. There is one feed back loop between socialinvolvement and personal involvement. It is problematic in terms of solving the equations in the model. Thus, the author restricts the reciprocal relationship between social involvement and personal involvement into unidirectional influence from social involvement to personal involvement. The justification of the unidirectional influence lies in the hypothesis that among heavy media users, media exposure is mure likely to have influence on social involvement rather than on personal involvement—this model ultimately focuses on the decomposition of the effects of media exposure on political learning, which means that the behavior of heavy media users is centered rather than is that of non-regular users. W: By making comparison between the structural equations in the model, the understanding of the effects of media exposure in political learning will be provided. First of all, the direct influences only in the model are shown in the following equations. Xl( Media Exposure) = Exogenous variable X2 (Perceived self-knowledge)= P21 X1+ E2 X3 (Social Involvement) = P32 X2 + E3 X4 (Personal Involvement) =P42 X2 + P43 X3 +E4 X5 (Factual Knowledge) = P54X4 + P53 X3 + P52 X2 + E5 58 The following structural equations are generated in the system of the “illusion of knowing model” by examining the indirect and direct effects of media exposure on factual knowledge. X1( Media Exposure) = Exogenous variable X2 (Perceived self-knowledge)= P21 X3 (Social Involvement) = P32 P21 X4 (Personal Involvement) =P42 P21+ P43 P21 X5 (Factual Knowledge) =P52 P21+ P53 P32P21+ P54 P421)” + P54 P43 P32 P21 Ultimately, factual knowledge can be decomposed by four indirect sources. The direct relationship between media exposure and perceived self-knowledge is a key credential to all of four indirect sources. That is, the previous findings of the low correlation between media exposure and factual knowledge might be attributed to the lack of understanding the relationship between media exposure and the moderating variable, perceived self-knowledge. 59 CHAPTER 4 Results In this chapter, the results of data analysis are presented. First, the demographic characteristics of the sample are discussed. Next, the hypotheses driven by the illusion of knowing model are tested with regression; all hypotheses are supported in full or in part. Finally, the illusion of knowing model itself is tested with path; the data are consistent with the model. 4. 1. Sample Two hundred twenty six of the five hundred samples (about 45 %) participated in the survey. Two hundred twelve of the collected survey questionnaires were completed. Fourteen questionnaires were eliminated because of their incompleteness. Unfortunately, the sample tended to under-represent younger persons, less educated, and less affluent persons, compared with 1990 US. Census data for East Lansing. Respondents were reported with a mean age of 40 years, 16.26 years of formal schooling completed and a mean income of approximately 5 35,000. The age difference between the respondents and the East Lansing population (research population) was likely due to the fact that the half of the population are students in Michigan State University and this survey seems to fail in putting the student respondents in the sample. Yet, compared with Michigan census data, the proportions of age groups in the sample were representative. Considering that this research meant to generalize its findings to the whole adult population of Michigan 60 as well as that of East Lansing, the under-sampling of student population can not be problematic. Furthermore, there was less danger in this under-representation of the some age groups, because the purpose of this study was on the explanation of the relationship between media exposure, perceived self-knowledge, and factual knowledge. WWII]: Sex Sample City of East Lansing State of Michigan Male 49.8 48.51 48.55 Female 50.2 51.49 51.45 Total 100 % 100 % 100 % Race-Individual Sample City of East Lansing State of Michigan Caucasian 85.8 84.59 83.45 African American 3.9 6.93 13 .90 Hispanic 2.5 2.50 2.17 Asian American 2.5 6.99 1.13 American Indian 1.5 0.34 0.60 Other 3.9 1.15 0.93 Total 100 % 100 % 100 °/o 61 Age Range Sample City of East Lansing * State of Michigan 18 to 24 years 29 41. 6 10. 81 25 to 44 years 31.3 36 32. 07 45 to 54 years 19.7 8. 1 10. 20 55 to 60 years 5.6 2. 6 4. 23 61 to 64 years 5.1 2. 6 4. 32 Over 65 9.6 8 11. 93 Total 100 0/o 100 °/o 100 % Education Level Sample City of East Lansing State of Michigan Grammar school 0.5 ** 3.38 23.22 High school 5.8 6.45 32.30 College 54.9 50.59 38.04 Graduate School 38.8 39.57 6.43 Total 100 % 100 % 100 0/o Note : * : In the population of the city of East Lansing, dormitory residents of Michigan State University were eliminated. **: Due to illiteracy, some respondents whose education attainment belongs to the category of grammar school were not counted. 62 Income City of East Lansing State of Michigan Under $9,999 21.74 15.58 $10,000 to $14,999 11.42 8.58 $15,000 to $24,999 17.34 16.41 $25,000 to $34,999 12.89 15.34 $35,000 to $49,999 12.40 18.66 $50,000 to $74,999 11.40 16.26 $75,000 to $99,999 5.80 5.41 Over $100,000 7.02 7.52 Total 100 % 100 "/o Income Sample Income Sample Under $8,900 16.0 $36,000 to $53,999 19.3 $9,000 to $17,999 7.0 $54,000 to $71,999 13.4 $18,000 to $26,999 8.6 $72,000 to $89,999 5.3 $27,000 to $35,999 8.6 Over $90,000 21.4 Total 100 % 63 4.2. The Validity and Reliability of Measurement In terms of assessment of measurement, validity and reliability are the keys. Reliability means the consistency in assessing responses while validity is about the extent to what a measure is supposed to measure. The reliabilities of the scales that are employed in this study are estimated with Cronbach’s alpha, one of the internal consistency methods. The validity of each scale is assessed in terms of face validity and construct validity. Especially, for the sake of construct validity, each scale is assessed with internal consistency theorem and parallelism from perspective of the classical theory of measurement model. I 2] II I' I'll [II I The Cronbach’s alphas of media exposure, perceived self-knowledge, personal involvement, social involvement, and factual knowledge were respectively .36, .87, 94, .96, and .78. While the reliabilities of the other scales were acceptable, the reliability of media exposure seemed to be problematic (.36). The previous studies have reported the constant low reliability of media exposure. The author speculates that the low reliability of media exposure might be due to its multi-dimensionality -the items to measure media exposure are heterogeneous in their relation to one another. For example, the persons who used heavily TV news programs can not heavily rely on newspapers for the news information. People’s allotting hours for news information would be naturally limited by the hours spent for other activities so that TV and newspapers may compete to each other in order to get people’s limited time. High TV users may prove to be low newspapers users. The low reliability of media exposure might be ascribed to the heterogeneous in the relationships between the items. If it is the case, alpha does not provide an optimal estimate of reliability because alpha is a maximum likelihood estimator of a parameter. The more the inter-item correlations diverge from one another, the more the value given by alpha underestimates the true reliability (Zaller & Carmines, 1980)”. l 22 II l'l'l I II I First, the content of each indicator of the scales was examined in terms of the appropriateness of the operationalization. All indicators of each construct were reviewed if they reflect the full domain of the construct“. Second, every scale was assessed with the internal consistency theorem for a uni-dimensional construct and parallelism for a dual dimensional construct, from the perspective of the classical theory of measurement. The confirmatory factor analysis was used for estimating the internal consistency theorem and the parallelism. 33 Thus, the author turns to coefficient theta (6 ). Unlike alpha, theta properly assesses the reliability of multi-dimensional scales. However, the theta coefficient was .38. It was still low, implying the possibility of the measure of media exposure fluctuating widely because of random disturbance (error), regardless of its dimensionality. 3‘ For example, among the indicators of media exposure, web site exposure was originally considered one of the indicators. However, web site exposure was dropped out from the scale of media exposure because of the following reasons: 1) web is not mass medium in that the number of web users has not reached a critical mass. 2) the nature of web usage is less habitual but more selective. 65 The internal consistency theorem means that if two measures are alternate indicators of the same underlying construct, then the correlation between them will be the product of their factor loadings. Assessment of the internal consistency concerns 1) the degree to which each indicator is caused by its respective construct and 2) whether or not the obtained correlations are within sampling error of their predicted values (the product of respective factoring loadings). On the other hand, the parallelism theorem implies that the correlation between indicators of two separate constructs is the product of their factor loadings and the correlation between the constructs. Assessment of the parallelism is about if the item correlations with the outside factor are flat. That is, the obtained correlations between the indicators of one construct and the indicators of the other construct are within sampling error of their predicted values. W: As seen in Table 2, the strongest indicator to explain media exposure was exposure to newspaper (its squared factor loading is .40) while the most weak indicator was exposure to Radio (its squared factor loading was .07). In terms of the goodness-of-fit model, the differences between the obtained and predicted correlations were clearly small—no more than .01. The data of measure of media exposure were consistent with the measurement model for the construct validity. The measure of media exposure showed low reliability and high validity. 66 thainedfluelatiuns Factor TV Radio Newspaper Loading TV .13* .36 Radio .09 .07* .26 Newspaper .23 .17 .40* .63 Note: * : Diagonals are communalities. Wm Factor TV Radio Newspaper Loading TV .36 Radio .09 .26 Newspaper .22 .16 .63 E l I | . l- I. I I I I' I Factor TV Radio Newspaper Loading TV .36 Radio .00 .26 Newspaper .01 .01 .63 67 W: The squared factor loadings of the five indicators were . 69, .65, .62, .40, and 48. All five indicators are quite good to explain the construct of perceived self-knowledge. Among them, the indicator to ask “how well they feel informed of the election in general” turned out to be the strongest (.69) one to represent the true score of the construct, perceived self-knowledge. As the Table 3 reveals, the errors (the differences between obtained and predicted correlations) were small except that of the obtained and predicted correlations between the indicator 4 and 5 (.28). Thus, a confidence interval was drawn around the correlation between the indicator 4 and 5 in order to decide if the error .28 is substantial enough to reject the null hypothesis of no error discrepancy. The discrepancy between the lowest and highest points of the interval was .14. The error .28 was over the range of the confidence interval (.14). However, considering the sample size of 200 and no problematic relation of the other indicators to either indicator 4 or 5, .28 error was regarded as trivial. 68 III I . I C I I. Factor 11 I2 13 14 15 Loading 11 .69* .83 I2 .79 .65* .80 I3 .72 .73 .62* .79 14 .44 .44 .37 .40* .63 15 .48 .41 .51 .72 .48* .69 Note: * : Diagonals are communalities. Feeling informed of I]: 142 I5: This governor election in general Other Michigan residents’ opinion about the candidates : The issues in this election : The candidates running for governor Other Michigan residents’ opinion about the issues 69 Factor I, I2 13 14 Is Loading I; .83 I2 .66 .80 I3 .66 .63 .79 L .52 .50 .50 .63 Is .57 .54 .55 .43 .69 E [Q] I . l- I. I I I I. I Factor I; 12 I3 I, Is Loading I; .83 I2 .13 .80 I3 .06 .10 .79 I, -.08 -.06 -.13 .63 Is -.09 -.09 .03 .28 .69 70 WW1: As reported in Table 4, the squared factor loadings for 4 indicators for personal involvement and another 4 indicators for social involvement were sufficiently high. There was no weak indicator among them. In the present study, involvement was categorized into personal and social involvement. They were considered as separate constructs. The examination of the factor structure confirms the dual dimensional model— Note that a high cross-factor loading can not be problem in confirmatory factor analysis. As Table 4 indicated, there was no indicator of personal involvement that had more than .10 error discrepancy between the obtained and predicted correlations among themselves(i.e. the upper part of the lower triangle in the error matrix). Also the indicators of social involvement showed the error range from .02 to .06 between the obtained and predicted correlations among themselves ( i.e. the lower part of the lower triangle in the error matrix). Especially, the obtained and predicted correlation matrices between the indicators of social and personal involvement revealed the error discrepancies between them were safely placed within the sampling error of zero (i.e. Italics in the error matrix)-the evidence of parallelism. Based on the above reasons, the author ultimately advocates that personal and social involvement were separated constructs. 71 ill I . I C I I. Factor Loading“ P1 P2 P3 P4 8, S2 S3 84 F1 F2 P1 .76* .87 .66 P2 .85 .78* .88 .69 P3 .75 .77 .83* .91 .63 P4 .75 .75 .90 .81* .90 .64 81 .64 .63 .56 .56 .84* .67 .92 S2 .60 .64 .56 .56 .92 .88* .66 .94 S3 .60 .61 .60 .63 .81 .84 .84* .68 .92 S4 .57 .64 .59 .61 .80 .81 .88 .80* .68 .90 Note: * : Diagonals are communcalities (squared factor loadings). **: The between factor 1 and factor 2 is .73. P. : Importance S. : Importance P2: Concern S2: Concern P3: Relevance S3: Relevance P4: Significance S4: Significance 72 Factor Loading P1 P2 P3 P4 Si 52 33 54 F1 F2 P. .87 .66 P2 .76 .88 .69 P3 .79 .80 .91 .63 P4 .78 .79 .81 .90 .64 S. .58 .59 .61 .60 .67 .92 S2 .60 .60 .62 .61 .86 .66 .94 S3 .58 .59 .61 .60 .85 .86 .68 .92 So .57 .58 .60 .59 .83 .85 .84 .68 .90 P1 P2 P3 P4 81 S2 S3 8.. Factor Loading P1 .87 .66 P2 .08 .88 .69 P3 -.04 -.02 .91 .63 P4 -.03 -.04 .09 .90 .64 S; .06 .04 -. 05 -.04 .67 .92 S2 .00 .04 -. 06 -. 05 .06 .66 .94 S3 .02 .02 -. 01 .03 -.04 -.02 .68 .92 S4 .00 .06 -. 01 .02 -.03 -.04 .05 .68 .90 73 W: According to Table 5, the questions about the personal characters of the gubernatorial candidates turned out to be the most weak indicators (K1 and K2 )for measuring factual knowledge while the questions about the pledged specific policy plan (K8, K10, K12, and K13) gained the higher place in terms of representation of respondents’ factual knowledge. The error matrix also illustrated that all of 14 indicators well measured the unidimensionality of the construct. Even the biggest error (.13) was not enough to damage the one factor model. 74 * : Diagonals are communalities. 75 Oil . IC II. KKKKKKKKKKKKKK 1 2 3 4 s o 7 s 9 10 11 12 13 I4 * K1 .09 * K2 .12 .09 * K3 .28 .16 .12 'k K, .07 .19 .19 .10 * Kg .22 .16 .15 .17 .28 * K6 .09 .10 .18 .22 .19 .12 * K; .14 .17 .14 .05 .31 .09 .23 '1: K3 .14 .14 .14 .23 .33 .20 .34 .40 * Kg .09 .21 .05 .02 .19 .07 .29 .37 .18 * Km .11 .09 .21 .05 .33 .25 .27 .33 .25 .31 * Kn .07 .08 .07 .14 .20 .10 .11 .21 .12 .28 .11 * K12 .13 .20 .14 .25 .29 .17 .35 .44 .33 .33 .23 .37 '1: K13 .10 .15 .21 .15 .27 .25 .28 .36 .28 .42 .12 .29 .31 * K, .22 .05 .15 .15 .22 .16 .23 .31 .20 .22 .24 .26 .27 .22 Note: Factor Loading: KI K2 K3 K4 K5 K6 K7 Ks K9 K10 K11 K12 K13 K14 .30 .31 .35 .32 .53 .35 .48 .63 .43 .55 .33 .61 .56 .47 E I. I H: I I. K K K K K K K K K K K K K K 1 2 3 4 s 6 7 s 9 lo 11 12 l3 l4 Kl K2 .09 K3 .11 .11 K4 .10 .10 .11 K5 .16 .16 .19 .17 K5 .11 .11 .12 .11 .19 K7 .14 .15 .17 .15 .25 .17 Kg .19 .20 .22 .20 .33 .22 .30 K9 .13 .13 .15 .14 .23 .15 .21 .27 Km .17 .17 .19 .18 .30 .19 .26 .35 .24 K11 .10 .10 .10 .11 .17 .12 .16 .21 .14 .18 K]; .18 .19 .21 .20 .32 .12 .29 .38 .28 .34 .20 K13 .17 .17 .20 .18 .30 .20 .27 .35 .24 .31 .18 .34 K14 .14 .15 .16 .15 .25 .16 .23 .30 .20 .29 .16 .29 .26 76 K K K K K K K K K K K K K K 1 2 3 4 5 6 7 8 9 10 11 12 13 14 K1 K2 .03 K3 .17 .05 * K. .03 .09 .08 * K3 .06 .00 .04 .00 'k 'k K, .02 .01 .06 .11 .00 * 'lr * K, .00 .02 .03 .10 .06 .08 'k * * * * K3 .05 .06 .08 .03 .10 .02 .04 * 'I: * 'k 'k K, .04 .08 .10 .12 .04 .08 .08 .10 * 'k * * Km .06 .08 .02 .13 .03 .06 .01 .02 .01 * * * * 'k * K" .03 .02 .03 .03 .03 .02 .05 .00 .02 .10 * sir * K12 .05 .01 .01 .05 .03 .05 .06 .06 .05 .01 .03 * * 'lr * * 'k * * K13 .07 .02 .01 .03 .03 .08 .05 .01 .01 .11 .06 .05 * 'k * * * K14 .08 .10 .01 .00 .03 .00 .00 .01 .00 .07 .08 .03 .01 Note: * : Negative Sign 77 4.3. Hypotheses Hypothesis 1: Greater media exposure is positively related to greater perceived self-knowledge. As the unstandardized regression coefficient in Table 6 demonstrated, media exposure had a significant effect on perceived self-knowledge( [3 = .438 , p<.001): as much as people were exposed to news media, they felt more informed about the 98’s gubernatorial election. Hypothesis 1 was supported. The standardized regression coefficient for media exposure was .432 (p<. 001); those for interpersonal communication, voting experience and education were respectively .170, .236, and - .021. Among the control variables, the beta coefficients for interpersonal communication (p=.009) and campaign experience (p=.002) proved to be statistically significant. The model including 3 control variables explained 28 % of the total variance of perceived self-knowledge (F (4, 192)=18.606 p<.001) while only the media exposure explained 18.6 0/o of the total variance. Unlike other studies that have more emphasized the roles of the control variables, interpersonal communication, education and campaign experience, rather than media exposure in terms of political learning, perceived self-knowledge turned out to have the stronger relationship with media exposure than with the other variables. 78 Unstandardized Standardized Coefficients Coefficient Model ' Standard Sig. R’ B Error Media Exposure .438 .066 .420 .000 Control Variables Education -.012 .042 -.021 .771 Campaign Experience -.250 .078 .236 .002 Interpersonal Communication .131 .050 .170 .009 . 279 " Note: a. Dependent Variable: Perceived self-knowledge b. P< .001 79 Hypothesis 2-1: Perceived self-knowledge is positively associated with social involvement. As shown in Table 7, the effects of perceived self-knowledge on social involvement was statistically significant at 99 % confidence level ([3 =.431, p<.001): hypothesis 2-1 was supported. As people were getting to feel informed about the election, they were socially involved with it. The model with 6 socio-demographic control variables and perceived self- knowledge accounted for 28 °/o of the total variance (F(7, 170)=9.267, p<.001). Perceived self-knowledge alone was responsible for 19 % of the total variance. Among the 5 control variables, sex was statistically significant. That is, females (b=.129, p=.053), were relatively more socially involved with the 98’s gubernatorial election, compared to males. The influence of control variables was minimal and insignificant, compared to that of perceived self-knowledge (b=.454, p<.001), so that perceived self-knowledge can be considered as the main impact on social involvement in this model. 80 Unstandardized Standardized Coefficients Coefficient Model ' Standard Sig. R’ B Error Perceived Self-knowledge . 436 .067 .454 .000 Control Variables Education -.039 .044 -.066 .373 Incomes -.022 .042 -.041 .595 Ages -.0030 .006 -.041 .594 Sex(Female) .338 .173 .129 .053 RaceOVhites) .294 .247 .079 .237 Party Identification .31 2 .194 .110 .110 .276 " Note: a. Dependent Variable: Social Involvement b. P < .001 81 Hypothesis 2-2: Perceived self-knowledge is positively associated with personal involvement. As shown in Table 8, the effects of perceived self-knowledge on personal involvement were statistically significant at 95 % confidence level ([3=.576, p<.001): hypothesis 2-2 was supported. The more people felt informed about the election, the more they were personally involved with it. The model with 6 socio-demographic control variables and perceived self- knowledge explained 40 °/o of the total variance (F (7, 171),=16.331 p<.001). Perceived self-knowledge alone accounted for 33 °/o of the total variance. Among the 6 control variables, sex, age and party identification gained a marginal statistical significance. That is, females (b=.118, p=.052), Democrats and Republicans (b=.125, p=.045), and younger persons (b=.131, p=.063) were relatively more personally involved with the 98’s gubernatorial election than were males, Independents and apolitical persons, and older persons. However, the standardized beta coefficients of control variables were quite low, compared to that of perceived self-knowledge ( b=. 515, p<.001). Thus, the perceived self-knowledge can be acknowledged as the main force on personal involvement in this model. In addition, it is worth of noticing that political party identification and age contributed to the personal involvement in the case of electoral campaign. 82 Unstandardized Coefficients Standard Standardized Sig. R’ Model ‘ B Error Coefficient Perceived Self-knowledge .574 .072 .515 .000 Control Variables Education -.014 .046 - .020 .768 Incomes .029 .045 .046 .518 Ages -.01 1 .006 -.131 .063 Sex(Female) .358 .183 .1 18 .052 Race(Whites) -.001 .262 -.002 .968 Party Identification .413 .205 .125 .045 . 401 " Note: a. Dependent Variable: Personal Involvement b. P < .001 83 Hypothesis 2-3: Perceived self-knowledge is more strongly associated with social involvement rather than personal involvement within heavy regular media users. In order to block the effects of news media in the association between perceived self-knowledge and social involvement, media users with more than 5.6 points out of 7 points in the item of media exposure were selected as regular media users. Upper 25 °/o of the sample was categorized as regular media users. The comparison of two simple regression models reveals that perceived self- knowledge was more likely to have influence on personal involvement (R2=. 255, F (1, 53)=18.146, p<.001) rather than on social involvement (Rz=.110, F (1,53) = 6.571, p=.013 ) among regular media users. The personal involvement model accounted for 10 °/o more than did social involvement model. The data were not consistent with the hypothesis. In addition, the author couldn’t do F-test for whether the .145 difference of R2 was statistically significant because the residuals of the two models were dependent—the correlation between the residuals was .892. That is, social involvement and personal involvement were more likely to be highly related among regular media users. If the residuals were independent, F-test would have been done by dividing the error sum of squares in the social model by the error sum of squares in the personal model. Furthermore, this finding, even if not testable, substantially violated a real world-based speculation of the author. Regular media users must be more socially involved with an event or issue portrayed by news media, considering that that 84 media tend to promote problems or issues to the status of social and political issues (McLeod, Becker, & Byrnes, 1974; MacKuen, 1981) by priming the social importance, not the importance of personal concerns (Mutz, 1992). Thus, the author redefined the upper 10 °/o of the sample as heavy regular media users, and reanalyzed it. The comparison of two simple regression models indicated an opposite finding to that of the previous analysis with the upper 25 % of the sample. Perceived self-knowledge was more likely to have influence on social involvement R2=. 285, F (1, 24)=9.566, p=.005) rather than on personal involvement (R2=.179, F (1,23) = 5.012, p=.035 ). However, F-test for the difference of R2 could not be performed because of the same reason with the upper 25 % sample case—the residuals of the two model was correlated (.554). The fact that the correlation between residuals of the two models reduced from .894 to .554, gave some impression that social involvement and personal involvement were more likely to respectively form distinctive relationship with perceived self-knowledge among the upper 10 % of the sample rather than the upper 25 % of the sample. Thus, the author decided to categorize the upper 10 °/o of the sample as heavy media users for analysis of the impact of the control variables. Furthermore, the models including the 6 socio-demographic variables and perceived self-knowledge failed to explain the variance in both social involvement (R2= .316, F (7, 15)= .989, p=.475) and personal involvement( R2 = .374, F (7, 14)=1.197, p=.365), as seen in Table 11 and 12. Notice that R square in each model is inflated by the number of the socio—demographic variables. Adjusted R squares 85 were respectively .044 and .062. The result showed that perceived self-knowledge and 6 socio-demographic models explained virtually nothing in the total variance of social and personal involvement. However, the failure of the both models can be partially ascribed to the small sample size. If the substantial number of the samples had been provided, the result of the social involvement model would have been different, considering that perceived self-knowledge ( [32.780, p=.032) alone turned out to be statistically significant in affecting heavy media users’ social involvement. 86 Unstandardized Coefficients Model Standardized Sig. R’ B Std. Coefficients Error Perceived Self-knowledge " . 316 .123 .332 .005 .110 Perceived Self-knowledge b . 613 .144 .505 .000 .225 13111112; Wale Unstandardized Coefficients Model B Std. Standardized Sig. R2 Error Coefficients Perceived Self-knowledge ' . 835 .270 .534 .005 .285 Perceived Self-knowledge b . 682 .305 .423 .035 .179 a. Dependent Variable: Social Involvement b. Dependent Variable: Personal Involvement 87 Unstandardized Coefficients Standardized Model ' Standard Coefficient Sig. R2 B Error Perceived Self-knowledge .780 .329 .475 .032 Control Variables Education -.053 .186 -.069 .778 Incomes -.156 .178 -.205 .396 Ages -.01 1 .017 -.156 .533 Sex(Female) .159 .629 .056 .804 Race(Whites) -.552 .990 -.131 .585 Party Identification .264 .959 .062 .585 .316 b' ‘ Note: a. Dependent Variable : Social Involvement b. P = .475 c. Adjusted R2 = .044 88 Unstandardized Coefficients Standardized Model‘ Standard Coefficient Sig. R2 B Error Perceived Self-knowledge .629 .339 .404 .085 Control Variables Education .206 .196 .250 .314 Incomes .130 .200 .162 .527 Ages -.018 .019 -.226 .364 Sex(Female) .108 .676 .036 .875 Race(Whites) -.233 1.037 -.053 .825 Party Identification -.407 1.209 -.077 .741 .365 b: ° Note: a. Dependent Variable: Personal Involvement b. P = .362 c. Adjusted R’ = -.062 89 Hypothesis 2-4: Perceived self-knowledge is more strongly associated with personal involvement rather than social involvement within non-regular media users. Non-regular media users who obtained less than 4.0 in the media exposure scores consist of about 25 % of the sample. As they were more feeling informed, they showed the tendency to be involved more personally in the election rather than socially. The perceived self-knowledge played a more active role in affecting non- regular media users’ personal involvement (R2= .362, F(1, 67)=38.046, p<.001) rather than their social involvement (Rz=.231, F(1, 67)=20.116, p<.001). The personal Involvement model explained 13 % more than did the social involvement model. The hypothesis was consistent with the data. However, I can’t do F-test for whether the difference was statistically significant or not, because of the dependency between the residuals of the two models—the correlation between residuals of the two model was .893. Furthermore, the lower 10 0/o of the sample were also analyzed in order to correspond to the analysis of heavy media users. The finding of the lower 10 % of the sample was consistent with that of the lower 25 % of the sample. The perceived self-knowledge had an influence on non-regular media users’ personal involvement (R2= .503, F(1, 25)=25.294, p<.001) more than their social involvement (R2=.364, F(1, 25)=14.320, p<.001). In addition, the residuals of the two models were not independent: r= .851. The .04 difference between residual correlations of the two models implied that there was a least difference between lower 10 °/o and 25 0/o of 90 the sample in terms of the relationships of perceived self-knowledge with social and personal involvement. In order to be consistent with the analysis of heavy media users, the lower 10 % of the sample were chosen for analyzing the impact of control variables on social and personal involvement. Tabla}; WW Unstandardized Standardized Model Coefficients Coefficients Sig. R2 B Std. Error Perceived Self-knowledge ' .588 .131 .481 .000 . 231 Perceived Self-knowledge " .797 .129 .602 .000 .361 Iahleli: WWW Unstandardized Model Coefficients Standardized Sig. R2 Std. Coefficients B Error Perceived Self-knowledge ' . 812 .215 .603 .001 . 364 Perceived Self-knowledge b . 944 .188 .709 .000 .503 Note: a. Dependent Variable: Social Involvement b. Dependent Variable: Personal Involvement 91 As Table 15 indicated, the model including 6 socio-demographic control variables and perceived self-knowledge variable explained 67 % of the total variance in personal involvement (R 2 = .778, Adjusted R 2 = .675, F (7, 15)=7.527, p=.001). Perceived self-knowledge ([3: .76, p=.002) also indicated the statistical significance in terms of explaining the variance in personal involvement of the non regular media users. Among the control variables, education significantly affected the non-regular media users’ personal involvement. That is, the more they were educated, the less they were personally involved with the election (b=-.384, p=.012). The less educated non-regular media users might be the dead zone of political campaigns. On the hand, the model including 6 socio-demographic control variables and perceived self-knowledge variable explained 36 % of the total variance in social involvement (R ’ =.565, Adjusted R ’ = .362, F(7,15)=2.784, p=.046). As seen in Table 16, none of the control variables were statistically significant to explain the variance in social involvement of non-regular media users. Perceived self- knowledge( [3=. 66, p= 043) alone proved to marginally obtain statistical significance. 92 Unstandardized Standardized Coefficients Coefficient Model ' Standard Sig. R2 B Error Perceived Self-knowledge .764 .201 .605 .002 Control Variables Education -.356 .124 - .384 .012 Incomes .129 .09 .153 .254 Ages -.0084 .011 -.108 .447 Sex(Female) -.082 .591 .022 .891 Race(Whites) 1.149 .848 .246 .195 Party Identification .764 .532 .199 .172 .778 " Note: a. Dependent Variable : Personal Involvement b. P = .001 93 a. Dependent Variable: Social Involvement b. P = .046 c. Adjusted R’ = -.362 94 Unstandardized Coefficients Model ' Standard Standardized Sig. R’ B Error Coefficient Perceived Self-knowledge .668 .303 .493 .043 Control Variables Education -.252 .186 -.254 .196 Incomes -.045 .163 -.049 .788 Ages .0068 .016 .080 .683 Sex(Female) -.372 .888 -.081 .720 Race (Whites) 1.779 1.274 .356 .183 Party Identification .629 .800 .153 .565 .565 9 ° Note: Hypothesis 3-1: As social involvement increase, personal involvement will increase. As seen in Table 17, social involvement had strong effects on personal involvement ([3=.709, p<.001). As people were getting more socially involved in the election, they were also tending to get more personally involved in it. Among the control variables, age, income and party identification were found to have influence on personal involvement. The more people aged, the less they got personally involved in the election ([3: -.169, p=.008). The richer people were, the more they were personally involved with the election ( [3: .071, p=.078). Democrats and Republicans were personally more involved with the election than Independents and apolitical persons ([3: .325 p=.084). 95 Unstandardized Standardized Coefficients Coefficient Model ' Standard Sig. R2 B Error Social Involvement .709 .066 .61 1 .000 Control Variables Education -.005 .042 -.008 .892 Incomes .071 .040 .1 12 .078 Ages -.014 .005 -.169 .008 Sex(Female) .162 .167 .053 .355 Race (Whites) -.192 .238 -.045 .41 9 Party Identification .325 .187 .098 .084 .525 ” Note: a. Dependent variable: Personal Involvement b. P<.001 96 Hypothesis 3-2: Personal involvement is the stronger predictor of factual knowledge rather than is social involvement. As the Table 18 indicated, personal involvement ([3=.770, p<.001) had a statistically significant influence on the level of factual knowledge while social involvement proved to be insignificant to explain the level of factual knowledge ([3= -.l37, p=.533) : hypothesis 3-1 was partially supported. The model explained about 12 °/o of the total variance in factual knowledge (F (2, 254)=13.560, p<.001). Islam IheEffectuLSociaUnmlxemenLandfiermaUnmlyemem onEactualJinewlcdgc Unstandardized Standardized Coefficients Coefficient Model ' Standard Sig. R2 B Error Social Involvement -.137 .220 -.057 .533 Personal Involvement .770 .185 .379 .000 . 123 " Note: a. Dependent Variable: Factual Knowledge b. P<.001 97 Research Question 1: What are the characteristics of the people that either overestimate or underestimate their own knowledge level? Among 212 samples, 91 persons overestimated their knowledge level and only 8 persons underestimated it. Generally, the illusion of knowing phenomenon defended on persons’ overestimation of their knowledge level. In terms of socio-demographic characteristics, males tended to correctly estimate their knowledge level more than did female. The differences of age and income seemed not to have an influence on people’s estimation of their knowledge level. However, it was interesting that the persons who attained no more than high school education had an inclination to the misperception of their knowledge level more than did the persons who completed more than high school education. Especially, the high school graduates’ misperception of their knowledge level was ascribed to the overestimation of it rather than the underestimation of it. W 5']! I'll I.I. [II III. 12 Correct Sex Underestimating Overestimating Estimation Total Case Case Case Case Male 5 39 64 100 Female 3 48 53 104 Missing 0 4 4 3 Total 8 91 1 1 3 212 98 Correct Age Range Underestimating Overestimating Estimation Total Case Case Case Case. 18 to 24 years 4 29 22 55 25 to 44 years 0 26 39 65 45 to 54 years 2 18 19 39 55 to 60 years 1 3 5 9 61 to 64 years 0 8 0 8 Over 65 0 0 25 25 Missing 1 7 3 1 1 Total 8 91 1 13 212 Correct Education Level Underestimating Overestimating Estimation Total Case Case Case Case Grammar school 0 1 0 1 High school 0 9 3 12 College 7 47 57 1 1 1 Graduate School 1 33 51 85 Missing 0 1 2 3 Total 8 91 1 13 212 99 Correct Income Underestimating Overestimating Estimation Total Case Case Case Case Under $8,900 2 14 13 29 $9,000 to $17,999 1 6 7 14 $18,000 to $ 26,999 0 7 11 18 $ 27,000 to $ 35,999 1 8 8 17 $36,000 to $53,999 0 12 13 35 $54,000 to $71,999 0 6 19 25 $72,000 to $89,999 0 7 4 11 Over $ 90,000 1 20 19 40 Missing 1 7 14 22 Total 8 91 1 13 212 100 Hypothesis 4-1: As media exposure increases, the illusion of knowing will increase. The whole model including media exposure and 4 control variables explained about 7 % of the total variance in illusion of knowing (F (5, 174)=2.757, p=.020). As Table 20 indicated, media exposure and the illusion of knowing had a statistically significant relation ([3=. 509, p=.000. The strength between them was quite strong. Among control variables, education was significant ([3= -.236, p=.035) while income, sex, and age were not. The more people were educated, the less they were illusioned regarding their knowledge level. In terms of the relative strengths of media exposure and education, media exposure (b= .218) could be interpreted as having slightly more influence on illusion of knowing than education (b= -.173) within this studied data set. 101 II I I [M I. E II [III . iii . Unstandardized Coefficients Model Standard Standardized Sig. R2 Beta Error Coefficients Media Exposure .509 .184 .218 .006 Control Variables Education -.236 .111 -.173 .035 Income -.0027 .1 07 -.022 .766 Sex .137 .442 .023 .752 Age -.0049 .014 -.030 .727 .073' Note: a. : p<.05 102 Hypothesis 4-2: The interaction between media exposure and social involvement has a significant effect on the illusion of knowing while doesn ’t the interaction between media exposure and personal involvement. Finally, the underlying main hypothesis in this study was tested that the influence of media exposure on the illusion of knowing is interacting with type of involvement. General Linear Model procedure was used for testing this interaction effect. AS seen in Table 21, the model including the main effect of media exposure, and the interaction effect between media exposure and social and personal involvement explained about 7 "/o of the total variance in the illusion of knowing (F(3, 200)=5.044, p=.002). The main effects of media exposure (F(1, 200)=3.054, p=.082) turned out to be marginally significant in this model. The interaction effects between media exposure and social involvement (F(1, 200)=7.124, p=.008) proved to be statistically significant while the interaction effects between media exposure and personal involvement (F(1, 200)=.026, p=.871) didn’t. When people, exposed to news media, were socially involved with media issues, they misestimated their knowledge level more than when they, exposed to news media, were personally involved with the media issues. The media exposure associated with social concern had a stronger influence on the illusion of knowing. 103 . gll‘rl' Sum of Mean Squares Squares DF F-Value Sig. Model 173.366 57.589 3 6.205 .000 Intercept 24.830 24.830 1 2.666 .104 Media Exposure 28.438 28.438 1 3.054 .082 Media Exposure X Social Involvement 66.346 66.346 1 7.124 .008 Media Exposure X Personal Involvement .245 .245 1 .026 .871 Error 1862.640 9.313 200 Total 2036.006 204 104 4.4 The Illusion of Knowing Model: Consistent with Data LAIASssssmenLofPathfleflicicnts The path coefficients for this model were presented in Figure 6. The path coefficients were corrected for attenuation, considering measurement errors. Unfortunately, the path coefficients for the relationships of personal involvement and social involvement to factual knowledge seemed to be smaller. When the 68 % confidence intervals35 for these path coefficients were drawn, the path coefficients proved to barely obtain the statistical significance. The upper and lower endpoints of the path coefficient for the relationship between social involvement and factual knowledge were respectively -.06 and -.31. The confidence interval of the path coefficient for the relationship between personal involvement and factual knowledge ranged from .28 to .01. In terms of the relative strengths among variables, media exposure had a strong effect on perceived self-knowledge (P2j=.85). The perceived self-knowledge was the source of the direct effects on factual knowledge (Ps2=.58), on personal involvement (P42 =. 37 ) and on social involvement ( P32=.58). As seen in Figure 6, the perceived self-knowledge was more positively associated with social involvement than with personal involvement. However, the stronger relationship between social involvement and perceived self-knowledge should be interpreted with the recognition that the influential direction from social involvement to personal involvement (P43=.52) was hypothetically set up in the context of news media impacts. 105 Finally, personal involvement was positively related with factual knowledge (P54=.15) while social involvement was negatively related with factual knowledge(Ps3=-.18). The more people were socially involved in the election, the less they had factual knowledge about it. On the other hand, the more people were personally involved in the election, the more they had factual knowledge about it. That is, the effects of media exposure in political learning were mediated by perceived self—knowledge, social involvement, and personal involvement. Especially, the relationship between media exposure and perceived self-knowledge seemed to be credential. Media .47(.85) Perceived .49(.58) Societal -.11(-.18) Exposure Self-knowle anolvement .33 (.37) .43 (.58) Y .53( .52 Factual Knowledge Personal I Involvement .17(.15) Note: Parentheses mean corrected path coefficients. 3’ Compared to 95 % confidence interval test for statistical significance, 68 % confidence interval test is more sensitive to Type II error, having a stronger statistical power. Due to the small sample size, a stronger statistical power is desirable in this present study. 106 W The presented path coefficients could be considered substantial enough to continue the assessment of the fit of the model (see Figure 6). Table 20 presented the information needed for assessing the path analysis that includes the obtained correlations, the predicted correlation, and the error differences between the predicted and obtained correlations. There were two sources of the errors: one was the error between the obtained and predicted correlations of media exposure and social involvement (.06) and the other was the error between the obtained and predicted correlations of media exposure and personal involvement (-.01). However, even the largest error discrepancy didn’t look large enough to be considered statistically significant (Z= .035, p= .727). That is, as seen in Table 20, the error discrepancies between the obtained and expected correlations among the five variables were within sampling error of zero. The null hypothesis that there is no discrepancy between the expected and obtain values can not be rejected ( X2=.012, df=3, p= .989). The discrepancy was within sampling error of zero and the data are consistent with the proposed model. 107 Obtained Correlations: Media Perceived Personal Social Factual Exposure Self- Involvement Involvement Knowledge knowledge Media .47 .32 .30 .26 Exposure 1.00 Perceived .59 .49 .47 Self- .85 1.00 knowledge Personal .69 .34 Involvement .55 .65 1.00 Social .21 Involvement .52 .54 .72 1.00 Factual Knowledge .49 .58 .39 .24 1.00 Note: Lower triangle = Corrected correlations Upper triangle = Uncorrected Correlations Predicted Correlations: Media Perceived Personal Social Factual Exposure Self- Involvement Involvement Knowledge knowledge Media ‘ Exposure 1.00 Perceived Self- .85 1.00 knowledge Personal Involvement .55 .65 1.00 Social Involvement .46 .54 .72 1.00 Factual Knowledge .49 .58 .39 .24 1.00 108 Error Matrix (Obtained Correlation - Predicted Correlation): Media Perceived Personal Social Factual Exposure Self- Involvement Involvement Knowledge knowledge Media Exposure .00 Perceived Self- .00 .00 knowledge Personal Involvement -.01 .00 .00 Social Involvement .06 .00 .00 .00 Factual Knowledge .00 .00 .00 .00 . 00 109 CHAPTER 5 Discussion: Illusion of Knowing The purpose of this research has been to causally examine the role of perceived self-knowledge in the impact of news media on political learning, producing the Illusion of Knowing model. For the purpose of this study, the survey has been conducted during the 1998’ gubernatorial election campaign in Michigan. Regression analysis and path analysis were employed for testing the relationships deducted from the “illusion of knowing” model. The results generally confirmed the model. 5.1. The Theoretical Significance The study advances theoretically in the studies of news media impact on public opinion as well as political learning by differentiating monitoring and retrieval process in memory and by identifying the relationship between them. First, the essence of news media exposure is to awaken the public’s political schema by mobilizing their monitoring process, but not to directly influence the public’s political memory for retrieval process. The awakened political schema absorbs and structures incoming news information. In the process of structuring, some news information takes some advantages and others are disadvantaged in accordance to the familiarity of the information to political schema. Ultimately, the awakened political schema has influence on the retrieval capacity, resulting in showing the variation of recall of the political knowledge. The role of news media in 110 political learning should have been introduced as influencing the public’s monitoring process prior to their retrieval process. Second, the understanding of the connection between monitoring and retrieval processes leads to that of the opinion formation function of news media. The discrepancy between perceived self—knowledge and factual knowledge respectively formed in monitoring and retrieval process is the source of news media power: illusion of knowing. The biased estimation of one’s own knowledge level means to bring up the illusion of knowing phenomenon among the public. When the public might be uninformed but misperceive themselves as well informed, what can be happening? This illusion might lead them to being inert, being assertive but not rational, and being manipulated by news media but believing in that they are too intelligent to be swayed by news media. The invincible self might play a role in devaluating news media impact. Compared to the invincible self, the vulnerable self appears when people underestimate their knowledge level. In this context, the illusion of knowing means that people misperceive themselves as less informed than they are actually informed, leading to imaging themselves as vulnerable to news media impact. The vulnerable self might play a role in exaggerating news media impact. Either the exaggeration or devaluation of news media impact might be the forces in the scene of public opinion process. Thirdly, the illusion of knowing model may be well addressed in the association with third-person effects hypothesis. Lasora (1989) found that the discrepant perception between perceived third- and first-person media effects were 111 greater among those who perceived themselves as experts, without actually being experts, than among those with real political knowledge. Mutz (1989) discovered that the tendency to perceive others as more influenced by the mass media was most prominent among those who consider an issue very important. That is, the qualification of the self is a focal point in third-person effects hypothesis, as conditioned in the illusion of knowing model. The future research on the illusion of knowing model can be conducted in contingent with third-person effects hypothesis. 5.2. The Implications of Empirical Findings The illusion of knowing model shows and mobilizes the relationship between the learning and opinion formation functions of news media. To some extent, this study has been successful in providing rationalization of and empirical data for the model. Among 9 hypotheses deducted from the model, 6 hypotheses were supported; 2 hypotheses were unable to be tested due to the correlation between error terms; 1 hypotheses were partially supported. Finally, the illusion of knowing model itself proved to be consistent with the data when path analysis was performed. The author revisits some interesting empirical findings and theoretical suggestions of this study. First of all, media exposure indirectly influences political knowledge via perceived self-knowledge and involvement. The indirect paths between media exposure and political knowledge explained why the influential relationship between them has been often masqueraded in the null relationship. In 112 deed, the several indirect paths between them wane out, resulting in the low correlation between them. Secondly, as expected in this study, the relationship between media exposure and perceived self-knowledge was compelling—the correlation coefficient was .49 and the corrected correlation coefficient for the attenuation was .85. No one will dispute that the regression coefficient is sufficient enough to support the stronger relationship between media exposure and perceived self-knowledge than the relationship between media exposure and factual political knowledge (. 26). Thirdly, perceived self-knowledge was found to be more associated with social involvement than with personal involvement when the relationships within the illusion of knowing model are analyzed with path analysis. As a learning motivation, personal involvement may be more effective than may social involvement. Based on this finding, we might also project that people probably overestimate the level of their relevant knowledge to an issue or event when seeing it from the viewpoint of social self. Contrast to it, people may underestimate the level of their relevant knowledge to it when seeing from the viewpoint of personal self. Fourthly, social involvement was negatively related with the acquisition of factual knowledge while personal involvement was positively related with it in the “illusion of knowing” model. This finding implied that when an issue or event be proposed as a personal matter, the public’s active learning is expected. Fifthly, the differences between heavy media users and non media-regular users in the relationships of social and personal involvement to perceived self- knowledge are drawing the attention. Among heavy media users, perceived self- 113 knowledge was more associated with social involvement than with personal involvement. Non-media users tended to associate perceived self-knowledge with personal involvement more than with social involvement. Finally, the illusion of knowing phenomenon existed. The gap between perceived self-knowledge and factual knowledge increased as media exposure increased. Especially, it is interesting that people socially involved misperceive themselves as well informed more than did people personally involved. 5.3. Limitations There are a few considerations to this study that need to be acknowledged. First of all, the illusion-of-knowing phenomenon might be more sensible in the cases of the coverage of social and political issues than the cases of election campaign coverage. Keeping in mind that the gap between perceived self-knowledge and factual knowledge is ascribed to the function of the negative relationship between familiarity and accurate recall, the difference between social and political issue coverage and election coverage is clear. Social and political issues have developed throughout cumulative news coverage, which means that the public has been exposed to them for a long-term period. The accumulated exposure to social issues gives to the public great familiarity to the issues. Yet the complexity of the social issues requires the public to take cognitive efforts to develop well-organized schema. There was higher likelihood that most of the public feels informed but small of them have comprehended social and political issue coverage. Ultimately the gap between 114 perceived self-knowledge and factual knowledge can be maximized as long as the one-shot dramas of social or political events are portrayed as in conjunction with the long-lost reminders of social and political issues. Contrast to it, election coverage tends to be cycling but novel because most election is played with a different set of candidates. Thus, election coverage may bring “new” information about the candidates’ personal characters and their issue position. In addition, political campaigners fight over who will control the media agenda during the campaign period (Lang & Lang, 1983) by creating new frames of old and tedious issues. The novelty let the public be aware of and accurately estimate their own knowledge level. In other words, the gap between perceived self- knowledge and knowledge accumulation can not be big as much as that in the case of social and political issues. Thus, social and political coverage rather than an electoral coverage would have been a better choice for this study. Secondly, even if the semantic scales of social and personal involvement that adopted Zaichkowsky’s scale for personal involvement were successful measures, there still has been a room for improvement. In the case of social involvement, the measure could have focused on the perceived importance of social approval because social approval is presumed as an important force for social involvement. Also, the personal involvement could have improved when perceived self-interest was directly asked. 115 5.4. Conclusion Dimock and Popkin(1997) found that even if Americans spent more time on reading newspaper s and watching TV news than the public in any other European country, they were less politically knowledgeable. They blamed the less informed citizen phenomenon on American journalism culture. However, there might be another reason that can not be ruled out. In their research, they measured the knowledge relevant to international news. Considering the individualistic and capitalistic nature of the American public, international news is hard to be seen as personal matter. International news could be something important and relevant at the societal level but not at the personal level. That is, the findings of the higher amount of media exposure and the lower level of factual knowledge might have been ascribed to the American public’s less personal involvement with the international matter. In terms of the impact of news media, type of involvement is one of important contingent variables. The agenda-setting effects of news media are found to be lessened in the case of the obtrusive issues such as unemployment and inflation that enter into people’s personal life on a regular basis (Weaver, Graber, McCombs, & Eyal, 1981; Zucker, 1978). That is, when personal concern is salient, news media seem to loose their touch. This study indicates that it also holds true in the process of how news media create the illusion of knowing phenomenon. Considering that media tend to promote problems or issues to the status of social and political issues (McLeod, Becker, & Byrnes, 1974; MacKuen, 1981) by priming the social importance, not the 116 importance of personal concerns (Mutz, 1992), it is not unexpected that the impact of news media on the illusion of knowing would be amplified by social involvement, but evaporated by personal involvement. The findings on the impact of news media could respectively shed light on the roles of news media in media campaign and public opinion. First, the interaction effects between media exposure and social involvement on the illusion of knowing explain why media campaign messages focusing on social responsibility often fail to influence people’s understanding of the purposes of the campaigns while media campaign messages designed from the perspective of personal benefit and cost cope well with the skeptical audiences. In terms of media campaign, a message inducing personal concerns is expected to work well at least in individualistic and capitalistic culture. On the other hand, as earlier mentioned, the illusion of knowing might lead people to being inert, being assertive but not rational, and being manipulated by news media but believing in that they are too intelligent to be swayed by news media. The illusion of knowing makes media exert their hegemony. As long as human history concerns, the inert public is doomed to fall into a mass without the help of civic journalism that opens “moral discourse in public sphere”. 117 APPENDIX Survey Questionnaire 118 Dear Participant: You are being asked to participate in a study of public opinion about the 1998’s Michigan gubernatorial election. The general goal of the study is to develop an understanding of how the public thinks of the election for a governor. By participating in the study, you will help us to expand our knowledge of election campaign. The questions contained in this booklet are concerned with your opinions and feelings about the 1998’s Michigan gubernatorial election. Also, there are some questions about your media use. Please read each statement carefully, and then indicate your feelings and opinions about the election on the scale provided. Remember, there are no right or wrong answers. Your opinions and feelings are what matters. The data gathered in this study will be reported on a group basis. Your answers will be kept in strict confidence. If you have any questions about any of the questions, please contact me. Office Phone: 355-5148 Home Phone: 355-4117 E-mail: new Thank you, Cheong-Yi Park Principal Investigator CASE ID: 119 The first item is about news consumption behaviors. Answer each question as it relates to your general style of news consumption. 1. Please, indicate how regularly you get the news from each of the following types of news sources on a scale of 1 to 7 (1 means not at all and 7 means very regularly). TV Local News Program 1 2 3 4 5 6 7 Newspapers 1 2 3 4 5 6 7 Radio 1 2 3 4 5 6 7 Web sites 1 2 3 4 5 6 7 Neighbors and friends 1 2 3 4 5 6 7 2. How regularly do you get news about the election campaign for governor? Not at All 1 2 3 4 5 6 7 Very Regularly 3. How closely do you pay attention to the news about the campaign for governor? Not at All 1 2 3 4 5 6 7 Very Closely 4. Now, please indicate how closely you follow each type of news about the campaign for governor (1 is not at all closely and 7 is very closely). News about each candidate’s character 1 2 3 4 5 6 7 News about each candidate’s issue positions 1 2 3 4 5 6 7 News about political parties 1 2 3 4 5 6 7 News about election events 1 2 3 4 5 6 7 News about election polling 1 2 3 4 5 6 7 5. When did you start to closely follow the news about the governor’s campaign? June July August September Early October Last week Not yet 120 6. We would like you to think about this election campaign for governor. How well do you feel informed of the following topics relevant to this election? 1 means you don’t feel informed at all and 7 means you feel very well informed. This governor election in general 1 2 3 4 5 6 7 The candidates running for governor 1 2 3 4 5 6 7 The issues in this election 1 2 3 4 5 6 7 Other Michigan residents’ opinion about the candidates 1 2 3 4 5 6 7 Other Michigan residents’ opinion about the issues 1 2 3 4 5 6 7 7. Please judge this election campaign according to your personal opinion, as an individual having your own experiences and values. Unimportant 1 2 3 4 5 6 7 Important to Me Of no concern 1 2 3 4 5 6 7 Of concern to Me Irrelevant 1 2 3 4 5 6 7 Relevant to My life Insignificant 1 2 3 4 5 6 7 Significant to Me 8. Regardless of your own personal feeling. How do you, as a member of a society, judge this election campaign? Unimportant 1 2 3 4 5 6 7 Important to people in Michigan Of no concern 1 2 3 4 5 6 7 Of concern to our State \I Irrelevant 1 2 3 4 5 6 Relevant to Michigan residents Insignificant 1 2 3 4 5 6 7 Significant to the State future 121 10. 11. Please, evaluate the impact of news media on the Michigan residents’ opinions about this election for governor. The news media impact on the public opinion about the candidate Very Weak 1 2 3 4 5 6 7 Very Powerful The news media impact on the public opinion about the issues in the election Very Weak 1 2 3 4 5 6 7 Very Powerful The news media impact on the public’s voting choice Very Weak 1 2 3 4 5 6 7 Very Powerful Before making a final decision, people might need additional information about the candidates and issues. How much more information do you need about each of the following topics? On the scale, 7 means you need a great deal of information while 1 means you don’t need any more information at all. Character and personalities of the candidates 1 2 3 4 5 6 7 Issue positions of the candidates 1 2 3 4 5 6 7 Issues in the election 1 2 3 4 5 6 7 Events and happenings in the election 1 2 3 4 5 6 7 We would like to ask the some details about John Engler and Geoffrey Fieger. If you are not sure the answer, please just circle “don’t know” . Which one gulps vitamins by handfuls and has a personal trainer? Engler Fieger Don’t Know Which one is a practicing attorney who has won many big-ticket lawsuits? Engler Fieger Don’t Know 122 Which one attended parochial school as a youngster? Engler Fieger Don’t Know Whose wife was a corporate lawyer? Engler Fieger Don’t Know Which candidate gets the endorsement of Michigan Manufacturers Association? Engler Fieger Don’t Know Whose running mate was a teacher and high school principal? Engler Fieger Don’t Know 12. We would like to ask if you know about the proposals and issue positions of Fieger and Engler. If you are not sure the answer, please just circle “don’t know”. Which candidate is more likely to favor strict environmental protection? Engler Fieger Don’t Know Which candidate has a plan of mandatory drug testing of welfare recipients? Engler Fieger Don’t Know Which candidate is more likely to favor assisted-suicide? Engler Fieger Don’t Know Which candidate pledges to repeal the law that bans teacher strikes? Engler Fieger Don’t Know Which candidate plans to put prisoners to work in community service program? Engler Fieger Don’t Know 123 Which candidate is more likely to favor restriction on abortion? Engler Fieger Don’t Know Which candidate proposes to cut Michigan Income tax from 4.4 percent to 3.9 percent over five years? Engler Fieger Don’t Know Which candidate proposes to make it a crime for gang members to recruit teen-agers? Engler Fieger Don’t Know 124 Please, answer each of the following questions about yourself. 1. What year were you born? 2. You are: (circle one) Male Female 3. What is the highest grade of school or year of college you complete? (circle one) Grammar school 0 1 2 3 4 5 6 7 8 High School 9 10 11 12 College 13 14 15 16 Graduate 17 18 19 20+ 4. Which do you consider yourself? Caucasian African-American Hispanic Asian-American American Indian Other 5. With which political party do you most closely identify? Democrat Republican Independent Other 6. What was your approximate household income last year? S 0- $8,999 $ 9,000-$l7,999 $18,000-$26,999 $27,000-$35,999 $36,000-$53,999 $54,000-$71,999 $72,000-$89,999 Over $ 90,000 Thank you for your participation. 125 BIBLIOGRAPHY 126 BIBLIOGRAPHY Abram, D., & Hoggs, M. A. (1990). Social identity theory: Constructive and critical advances. London: Harvester Wheatsheaf , and New York: Springer-Verlag. Almond, G. A. (1950). The American people and foreign policy. New York: Harvourt Brace. Asher, H. B. (1976). Causal Modeling. Beverly Hills, CA: Sage. Atkin, Charles K. (1972). “Anticipated communication and mass media information-seeking.” Public Opinion Quarterly. 36, 189-199. Atkin, C.K.(1973). Instrumental utilities and information seeking. In P. Clarke(ed.), New Models for Communication Research (pp. 205-239). Beverly Hills, CA: Sage. Atkin, Charles K., Galloway, John, and Nayman, Oguz B. (1976). News media exposure, political Knowledge and campaign interest. Journalism Quarterly. 53(2), 231-237. Atwater, Tony, Salwen, Michael B., and Anderson, Ronald B. (1985). Media agenda-setting with environmental issues. Journalism Quarterly. 62, (2, Summer), 393-397. Ayres, Q. Whitfield & Whiteman, David(1984). Congressional Reapportionment in the 1980s: Types and Determinants of Policy Outcomes. Political Science Quarterly, 99(2), 303-314. Babrow, A. S., & Swanson, D. L. (1988). Disentangling antecedents of audience exposure levels: Extending expectancy-value analyses of gratification sought from television news. communication Monographs, 55, 1-21. Baumeister, Roy F. (1982). A self-presentational view of social phenomena. Psychological Bulletin. 91(1), 13 -26. 127 Becker, Lee B., and Dunwoody, Sharon. (1982). Media use, public affairs knowledge and voting in a local election. Journalism Quarterly. 59, (2, Summer), 212-218,255. Bennett, E.(1988). Know-nothings revisited: The meaning of political ignorance today, Social Science Quarterly.69, 476-90. Brasford, J. D. (1979). Human cognition: Learning, understanding and remembering. Belmont, CA: Wadsworth. Breckler, S. J., &Greenwald, A. G. (1986). Motivational facets of the self. In R. M. Sorrentino & E. T. Higgins (Eds). Handbook of Motivation and Cognition (pp.145-155), New York: Gilford Press. . Brown, R., &McNeil, D. (1966). The tip of the tongue phenomenon. Journal of Verbal Learning and Verbal Behavior, 5, 325-337. Bybee, Carl R. (1978). Testing a process model of involvement, Communication Research. 5(4), 413-435. Cantril, Hardly (1942). Professor quiz: a gratification study. In P. F. Lazarsfeld and Frank Stanton (eds.), Radio Research 1941. New Tork: Duell, Sloan, and Pearce Chaffee, Steven H., and McLeod, Jack M. (1973). Individual vs. social predictors of information seeking. Journalism Quarterly. 50(2), 237- 245. Chaffee, Steven H., and Miyo, Yuko. (1983). Selective exposure and the reinforcement hypothesis: An intergenerational panel study of the 1980 presidential campaign. Communication Research. 10(1), 3-36. Chaffee, Steven H., and Schleuder, Joan. (1986). Measurement and effects of attention to media news. Human Communication Research. 13(1), 76-107. Chaffee, Steven H., Zhao, Xinshu., & Leshner, Glenn. (1994). Political Knowledge and the campaign media of 1992. Communication Research. 21(3), 305-324. 128 Chaiken, S. (1980) Heuristic versus systematic information processing and the use of source versus message cues in persuation. Journal of Personality and Social Psychology, 39, 752-766. Chang, T. (1989). The impact of Presidential statements on editorial regarding U.S.-China policy, 1950-1984. Communication Research, 16(4): 486-509. Cobb, Roger W. & Elder, Charles D. (1971). The Politics of Agenda-Building: An Alternative Perspective for Modern Democratic Theory .The Journal of Politics, 33 (4), 892-915. Cohen, B. C. (1963). The press and foreign policy. Princeton, NJ: Princeton University Press. Cooley, C. H. (1902). Human nature and the social order. New York: Scrubner. Crigler, Ann N., Just, Marion, and N euman, W. Russell. (1994). Interpreting visual versus audio messages in television news. Journal of Communication. 44, (4, Autumn), 132-149. Dearing, James W. (1989). Setting the polling agenda for the issue of AIDS. Public Opinion Quarterly. 53, (3, Fall), 309-329. Dearing, J., & Rogers, E. (1997). Agenda-setting. Newsbury Park, CA: Sage. Demers, David Pearce, Craff, Dennis, Choi, Yang-Ho, and Pessin, Beth M. (1989). Issue obtrusiveness and the agenda-setting effects of national network news. Communication Research. 16, (6, December), 793-812. Donohew, Lewis, Tipton, Leonard, and Haney, Roger. (1978). Analysis of information-seeking strategies. Journalism Quarterly. 55, (1, Spring), 25-31. Einsiedel, Edna F., Salomone, Kandice L., & Schneider, Frederick P. (1984). Crime: effects of media exposure and personal experience on issue salience. Journalism Quarterly. 61(Spring), 131-136. 129 Entman, Robert M. (1993). Framing: Toward clarification of a fractured paradigm. Journal of Communication. 43(4), 51-58. Erbring, Lutz, Goldenberg, Edie N. & Miller, Arthur H(1980). Front-Page News and Real-World Cues: A New Look at Agenda-Setting by the Media. American Journal of Political Science, 24 (1), 6-49. Eveland, William P. J r., McLeod, Douglas, & Signorielli, Nancy (1995). Actual and perceived US. public opinion: the spiral of silence during the Persian Gulf war. International Journal of Public Opinion Research, 7, 91-109. Festinger, L. (1957). A theory of cognitive dissonance. Evanston: Row Peterson. Flora, June A., & Maibach, Edward W. (1990). Cognitive responses to AIDS information: the effects of issue involvement and message appeal. Communication Research. 17(6), 759-774. Gamson, W.A.,& Modigliani, A.(1989). Media discourse and public opinion on nuclear power: A constructionist approach. American Journal of Sociology, 95,1-37. Ghanem, S. (1997). Filling in the tapestry: The second level of agenda setting. In M. McCombs., D. Shaw., D. weaver(Eds.), Communication and democracy exploring the intellectual frontiers in agenda-setting theory. Lawrence Erlbaum: New J ersy. Giltin, T.(1980). The whole world is watching: Mass media in the making and unmaking of the new left. Berkeley, CA: University of California Press. Glenberg, A. M., & Bradley, M. M. (1984). Coactivation and comprehension: contribution of text variables to the illusion of knowing. Memory & Cognition, 12, 355-360. Graber, Doris. A. (1980). Mass media and American Politics. Washington: Congressional Quarterly. 130 Graber, D. A.(1988). Processing the News: How people tame the information tide (2nd ed.). New York; Longman. Greenwald, A. G. (1982). Ego-task analysis. In A.H. Hastrof & A. M. Isen (Eds.), Cognitive Social Psychology (pp. 109-147). New York: Elsvier/ North Holland. Greenwald, A. G., & Brecker, S. J. (1985). To whom is the self presented? In B. R. Schlenker (Ed.). The self and social life, pp. 126-145. Greenwald, A. G., & Leavitt, C. (1984). Audience involvement in advertising: Four levels. Journal of Consumer Research, 11, 581-592. Grimes, Tom., & Meadowcroft, Jeanne. (1995). Attention to television and some methods for is measurement. Communication Yearbook. 18, 133-161. Gruneberg, M. M. (1978). The feeling of knowing, memory blocks and memory aids. In M. M. Gruneberg & P. Morris (Eds.), Aspects of Memory (pp. 187-209). London: Methuen. Hart, J. T. (1965). Memory and the feeling of knowing experience. Journal of Educational Psychology, 56, 208-216. Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Hoggs , M. A., & McGarty, C. (1990). Self-categorization and socialidentity. In D. Abrams & M. A. Hoggs(eds.). Social identity theory: Constructive and critical advances(pp. 10-27). London: Harvester Wheatsheaf , and New York: Springer-Verlag. Holland, Barry A. (1995). The news and the 1992 presidential campaign: perceived vs. Actual political knowledge. J & MC Quarterly. 72, 786-798. 131 Iyengar, S., & Kinder, D. (1987). News that matters: Television and American opinion. University of Chicago Press. Johnson, Eric., & Russo, J. E. (1984). Product familiarity and learning news information, Journal of Consumer Research, 11, 542-550. Jones, E. E., &Pittmann, T. S. (1982). Towards a general theory of strategic self-presentation. In J. Suls (Ed.). Psychological perspective on the self. Vol. 1, pp. 231-262. Hillsdale, NJ: Erlbaum. Johnson, Blair T., & Eagly, Alice H. (1989). Effects of involvement on persuasion. a meta-analysis. Psychological Bulletin. 106(2), 290-314. Joseph, F. H. Jr., Anderson, R., & Tatham, R. (1990). Multivariate Data Analysis. New York: Macmillian, Inc. Judd, C. M., & Kulik, J. A. (1980). Schematic effects of social attitudes on information processing and recall. Journal of personality and social psychology, 38, 569-578. Katz, D. (1959). The functional approach to the study of attitudes. Public Opinion Quarterly. 24, 163-204. Katz, Elihu., Haas, Hadassah., and Gurevitch, Michael. (1973). One the use of the mass media for important things. American Sociological Review. 38(2), 164-181. Kelman, Herbert C. (1961). Processes of opinion change. The Public Opinion Quarterly. 25, 57-78. Kelly, H. H. (1967). Attitude theory in social psychology. In D. Levine (ED.), Nebraska Symposium on Motivation (Vol. 15, pp. 192-238). Lincoln: University of Nebraska Press. Kiesler, C. A., & Kiesler, S. B. (1969). Conformity. MA: Addison-Wesley. 132 Kingdon, John. (1984). Agends, Alternatives, and public policies. Harper Collins Publishers. Koirat, A., & Lieblich, I. (1977). A study of memory pointers. Act Psychologica, 41, 151-164. Krosnick, John A., & Brannon, Laura A. (1993). The impact of the Gulf War on the ingredients of presidential evaluations: multidimensional effects of political involvement. American Political Science Review. 87(4). 963-975. Krugman, H. E. (1965). The impact of television advertising: learning without involvement. Public Opinion Quarterly, 29, 349-283. Lasora, Dominic L. (1989). Real and Perceived effects of “Amerika”, Journalism Quartely, 66, 373-378. Lasorsa, Dominic L. (1991). Political outspokenness: factors working against the spiral of Silence. Journalism Quarterly. 68(1). 131-140. Lasorsa, Dominic L., and Wanta, Wayne. (1990). Effects of personal, interpersonal, and media experiences on issue salience. Journalism Quarterly. 67 (4), 804-813. Levy, Mark R., and Windahl, Seven. (1984). Audience activity and gratifications: A conceptual clarification and exploration. Communication Research. 11(1), 51-78. Lichtenstein, S., & Fischoff, B. (1977). Do those who know more also know more about how much they know? Organizational Behavior and Human Performance, 20, 158-183. Lin, Carolyn A., & Salwen, Michael B. (1997). Predicting the silence on a controversial public issue. The Howard Journal of Communication. 8, 129-141. 133 Linsky, M. (1986). How the press affects federal policy making. New York: Norton. Lippmann, W. (1922). Public Opinion. New York: Harcourt Brace. Lo, Ven-hwei. (1994). Media use, involvement, and knowledge of the Gulf War. Journalism Quarterly. 71(1), 43-54. Markus, H. Smith, J., & Moreland, R. L. (1985). Role of the self-concept in the social perception of others. Journal of Personality and Social Psychology, 49, 1494-1512. McGuire, W. J. (1960). Cognitive consistency and attitude change. Journal of Abnormal and Social Psychology, 60, 354-358. McCombs, M., Shaw, D., & Weaver, D. (1997). Communication and democracy exploring the intellectual frontiers in agenda-setting Theory. New Jersey: Lawrence Erlbaum. McCombs, M., & Shaw, D. L. (1972). ‘The agenda-setting function of mass media. Public Opinion Quarterly. 36, 176-187. McCombs, M., & Shaw, D. L. (1993). The evolution of agenda-setting research: Twenty-five years in the marketplace of ideas. Journal of Communication. 43(2), 58-67. Manford H. Kuhn, & Thomas S. McPartland (1954), An Empirical Investigation of Self-Attitudes, American Sociological Review, 19 (1)954), pp. 68-76. Major, Ann Marie (1997). Pluralistic ignorance and the climate of opinion in a real- time disaster prediction. International Journal of Public Opinion Research. 9(2), 170-190. McLeod, Jack M., Becker, Lee B., and Byrnes, James E. (1974). Another look at the agenda-setting function of the press. Communication Research, 1(1), 131-166. 134 McLeod, Jack M., Bybee, Carl R., and Durall, Jean A. (1982). Evaluating media performance by gratifications sought and received. Journalism Quarterly. 59(1), 3-12, 59. McLeod, J. M., Bybee, C. R., & Durall, J. A. (1979). The 1976 presidential debates and the equivalence of informed political participation. Communication Research. 6, 463-487. McLeod, Jack M., & McDonald, Daniel G. (1985). Beyond simple exposure: Media orientations and their impact on political processes. Communication Research, 12(1), 3-33 McLeod, Jack M., & Pan, Zhongdang. (1991). Multilevel analysis in mass communication research. Communication research. 18(2),140-173. McQuail, D. (1987). Mass communication theory: An introduction. London: Sage. Metcalfe, J.(1986). Feeling of knowing in memory and problem solving. Journal of Experimental Psychology, 20, 158-183. Merton, Robert K. (1957). Priorities in Scientific Discovery: A Chapter in the Sociology of Science. American Sociological Review, Vol. 22, No. 6. (Dec), pp.635-659. Mikular, G. (1984). Persoal relationships: remarks on the current state of research. European Journal of social psychology, 69, 372-378. Mondak, Jeffery J. (1995). Newspapers and political awareness. American Journal of Political Science, 39, 513-527. Moscovici, S. (1976). Social influence and social change. London: Academic Press. Mutz, Diana C. (1989). The Influence of Perception of Media Influence , International Journal of Public Opinion Research, 1, 3-24. 135 Mutz, Diana C. (1992). Contextualizing personal experience: the role of mass media. The Journal of Politics. 56(3), 689-714. Nelson, T. O., & Narens, L. (1990) Metamemory: A theoretical framework and new findings. In G. Bower(ed.), The Psychology of Learning and Motivation: advances in research and theory (V. 26, pp. 125-173). San Diego, CA: Academic Press. Neuman, J. (1976). Patterns of recall among television news viewers. Public Opinion Quarterly, 40(1), 115-123. Noelle-Neumann, Elisabeth. (1974). The spiral of silence: A theory of public opinion. Journal of Communication. 24, (2,), 43-51. Noelle-Neumann, E.(1984). The spiral of Silence: Public opinion-our social skin. Chicago, IL: University of Chicago Press. Palmgreen, P. (1984). Uses and gratifications: A theoretical perspective. Communication Yearbook, 8, 20-55. Palmgree, P., & Clarke, P. (1977). Agenda setting with local and national issues. Communication Research, 4, 435-452. Palmgreen, P., & Rayburn, J. D., II. (1979). Uses and gratifications and exposure to public television: A discrepancy approach. Communication Research, 6, 155-179. Palmgreen, Philip, & Rayburn II, J. D. (1985). A comparison of gratification models of media satisfaction. Communication Monographs. 52, 335-346. Palmgreen, Philip, Wenner, Lawrence A., and Rayburn, J. D., II. (1980). Relations between gratifications sought and obtained: A study of television news. Communication Research. 7, (2), 161-192. Pan, Zhongdang., & Kosicki, Gerald M. (1993). Framing analysis: An approach to news discourse. Political Communication. 10(1), 55-75. 136 Pan, Zhongdang., & Kosicki, Gerald M. (1997). Priming and media impact on the evaluations of the president’s performance. Communication Research. 24(1). Park, C. Whan., Gardner, Meryl P., & Thukral, Vinod K. (1988). Self-perceived knowledge: some effects on information processing for a choice task. American Journal of Psychology. 101, 401-424. Patterson, T. E. (1980). The mass media election: How Americans choose their president. New York: Praeger. Perloff, Richard M. (1984). Political involvement: a critique and a process-oriented reformulation. Critical Studies in Mass Communication. 1,146-160. Perloff, Richard M. (1989). Ego-involvement and the third person effect of televised news coverage. Communication Research. 16(2), 236-262. Perry, David K. (1990). News reading, knowledge about, and attitudes toward foreign countries. Journalism quarterly. 67, 353-358. Perse, Elizabeth M. (1990). Involvement with local television news. Human Communication Research. 16(4), 556-581. Pertschunk, M. (1987). The role of public interest groups insetting the public agenda for the 90’s. Journal of Consumer Affairs, 21(2). 171-182. Petty, R. E., & Cacioppo, J. T. (1979). Issue Involvement can increase or decrease persuasion by enhancing message-relevant cognitive responses. Journal of Personality and Social Psychology, 46, 69-81. Pratkanis, A. J. & Greenwald, A. J.(1989). Attitude structure and function. Hillsdale, NJ: Erlbaum. Pratkanis, A. J. (1989). The cognitive representation of attitudes. In A. R. Pratkanis, S. J. Breckler, & A. G. Greenwald(Eds.), Attitude structure and function (pp. 71-89). Hillsdale, NJ: Erlbaum. 137 Price, Vincent. (1989). Social identification and public opinion: effects of communicating group conflict. Public Opinion Quarterly. 53,197-224. Price, vincent & Allen, Scott(1990). Opinion spirals, silent and otherwise: applying small-group research to public opinion phenomena. Communication Research, 17, 369-392. Rayburn II, J. D., & Palmgreen, Philip. (1984). Merging uses and gratifications and expectancy-value theory. Communication Research. 11(4), 537-562. Reder, Lynne M., & Ritter, Frank E. (1992). What determines initial feeling of knowing? Familiarity with question terms, not with the answer. Journal of Experimental Psychology: Learning, Memory, and Cognition. 3, 435-451. Rhee, June W., & Cappela, Joseph N. (1997). The role of political sophistication in learning from news (Meaning schema development). Communication Research. 24(3), 197—233. Robinson, John P., and Davis, Dennis K. (1990). Television news and the informed public: All information-processing approach. Journal of Communication. 40, (3, Summer), 106-119. Robinson, J. P, & Levy, M. R. (1986). The main source: Learning from television news. Beverly Hills, CA: Sage. Rokeach, M. (1973). The nature of human values. New York: Free Press. Roger, Connie. (1990). Involvement, attention, and perceptions of message relevance in the response to persuasive appears. Communication Research. 17(5), 571-600. Rogers, E., Dearing, J., & Chang, S. (1991). AIDS in the 1980s: The agenda-setting process for a public issue. Journalism Monographs 126. Rosengren, K. E.(1974). Uses and gratifications: A paradigm outlined. In J. G. Blumler and E. Katz (Eds.), The uses of Mass Communication. pp. 269-286. Beverly Hills, CA: Sage. 138 Rothenuhler, Eric W. (1991). The process of community involvement. Communication Monographs, 58, 63-78. Rothschild, Michael L. & Ray, Michael L. (1974). Involvement and political advertising effect: an exploratory experiment. Communication Research. 1(3), 264-285. Rubin, Alan M. (1984). Ritualized and instrumental television viewing. Journal of Communication. 34 (3), 67-77. Rubin, Alan M. (1977). Television usage, attitudes and viewing behavior of children and adolescents. Journal of Broadcasting. 21(3), 355-369. Rubin, Alan M., & Perse, Elizabeth M. (1987). Audience activity and television news gratifications. Communication Research. 14(1), 58-84. Rubin, Alan M., and Windhal, Stven. (1986). The uses and dependency model of mass communication. Critical Studies in Mass Communication. 3 (2), 184-199. Salmon, Charles, T. (1986). Perspectives on involvement in consumer and communication research. In Brenda Dervin and Melvin J. Voigt Norwood(eds.), Progress in Communication Sciences( PP. 243-268 ). NJ: Ablex.. Salmon, C. T. & Kline, F. G. (1985). The spiral of silence ten years later: an examination and evaluation. In K. Sanders, L. L. Kaid, and D. Nimmo (eds.), Political Communication Yearbook 1984. Carbondale: Southern Illinois University Press. Salwen, Michael B. (1988). Effect of accumulation of coverage on issue salience in agenda setting. Journalism Quarterly. 65, (1), 100-106,]30. Salwen, Michael B., & Driscoll, Paul D. (1995). Feeling informed? The ‘assurance function’ of the mass media. International Journal of Public Opinion Research. 7(3), 271-275. 139 Schachter, D.L.(1983). Feeling of knowing in episodic memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 9, 39-54. Schlenker, B. R. (1980). Impression management: The self-concept, social identity, and interpersonal relations. Monterey, CA: Brooks/Cole. Schleuder, J., McCombs, M., & Wanta, W(1991). Inside Agenda-setting Process: How Political Advertising and TV News Prime Viewers to think about issues And candidates. In Frank Biocca(Ed.)., Television and Political Advertising 1: Psychological Processes(pp. 263-310). Hillsdale, NJ: Lawrence Erlbaum. Snyder, C. R., Higgins, R. L., & Strucky, R. J. (1983). Excuses: The masquerade solution. New York: Wiley. Solomon, R. (1980). The opponent process theory of acquired motivation. American Psychologist, 35, 691-712. Spiro, R. J., & Sherif, C. W. (1975). Consistency and relativity in selective recall with differing ego-involvement. British journal of social and clinical psychology, 14, 351-361. Swann, W. B., Jr. (1983). Self-verification: bringing reality into the harmony with the self. In J. Suls & A. G. Greenwaldwds) Psychological perspectives. New York: Basic Books. Tajifel & Turner, R. H. (1979). An integrative theory of intergroup conflict. In W.G. Austin & S. Worchel (Eds.). The Social psychology of intergroup relations, pp. 33-47. Monterey, CA: Brooks/ Cole. Takeshita, T. & Mikami, S.(1995). How did mass media influence the voters’ choice in the 1993 general election in J apan?: A study of agenda-setting. Keio Communication Review. 17, 27-41. Tan, Alexis., & Vaughn, Percy. (1976). Mass media exposure, public affairs knowledge, and black militancy. Journalism Quarterly. 53(2), 271-279. 140 Taylor, D. Garth. (1982). Pluralistic ignorance and the spiral of silence: a formal analysis. Public Opinion Quarterly. 46, 311-335. Tesser, A. (1978). Self-generated attitude change. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 11, pp. 289-338). San Diego, CA: Academic Press. Tuchman, G.(1978). Making News: A study in the construction of reality. New York: Free Press. Trumbo, C. (1995). Longitudinal modeling of public issues with the agenda-setting process: the case of global warming. Journalism and Mass Communication Monographs 156. Wanta, Wayne., & Elliot, William R. (1995). Did the “magic” work? Knowledge of HIVIAIDS and the knowledge gap hypothesis. J & MC Quarterly. 72(2), 312-320. Weaver, D., Zhu, J., & Willnat, L. (1992). Information sources and agenda-setting: testing a theory of bridging. Paper presented to 47th annual conference of American Association for the Public Opinion, Florida, May, 1992. Wenner, L. A.(1982). Gratifications sought and obtained in program discrepancy: A study of network evening news programs and “60 Minutes”. Communication Research, 9, 539-560. Wenner, Lawrence A. (1986). Model specification and theoretical development in gratifications sought and obtained research: a comparison of discrepancy and traditional approaches. Communication Monographs. 53, 160-179. Willnat, Lars. (1995). Public Opinion and political outspokeness in pre-1997 Hong Kong: two tests 0 the spiral theory. Asian Journal of Communication. 5(2), 47-66. Zaichkowsky, J. E. (1986). Conceptualizing Involvement. Journal of Advertising, 15, 4-14. 141 Zeller, R. A., & Carmines, E. G. (1980). Measurement in the social science. Cambridge: Cabridge University Press. Zimbardo, Philip G. (1960). Involvement and communication discrepancy as determinants of opinion conformity. Journal of Abnormal and Social Psychology. 60(1), 86-94. Zimmer, Troy A. (1981). Media exposure to campaigns: Public anticipation and involvement in elections. Communication Research. 8(2), 189-204. Zinkhan, G. M., & Muderrisouglu, A. (1985). Involvement, familiarity, cognitive differentiation, and advertising recall: A test of convergent and discriminant validity, Proceedings: Association for consumer Research, 12, 356-361. 142