Public relations camapgins in 2002 korean presidential election; functional analysis of political disourse and media effects of agenda setting and favorability

PUBLIC RELATIONS CAMPAIGNS IN 2002 KOREAN PRESIDENTIAL ELECTION: FUNCTIONAL ANALYSIS OF POLITICAL DISCOURSE AND MEDIA EFFECTS OF AGENDA SETTING AND FAVORABILITY A Dissertation Presented to The Faculty of the Graduate School University of Missouri - Columbia In Partial Fulfillment of the Requirements for the Degree Doctor o f Philosophy by CHEOLHAN L E E Dr, Wayne Wanta, Dissertation Supervisor MAY 2004 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 3137725 INFORMATION TO USERS The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. UMI UMI Microform 3137725 Copyright 2004 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The undersigned, appointed by the Dean o f the Graduate School, have examined the dissertation entitled PUBLIC RELATIONS CAMPAIGNS IN 2002 KOREAN PRESIDENTIAL ELECTION: FUNCTIONAL ANALYSIS OF POLITICAL DISCOURSE AND MEDIA EFFECTS OF AGENDA SETTING AND FAVORABILITY presented by Cheolhan Lee a candidate for the degree of Doctor o f Philosophy and hereby certify that in their opinion it is worthy o f acceptance. fi .. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOW LEDGEM ENTS First and foremost, I would like to acknowledge my advisor, Dr. Wayne Wanta. He is an incredible mentor. He was always available when I needed him. It was a valuable experience to work with him as his research assistant. I benefited tremendously from his knowledge and passion for research. Without his support, I could not reach this far. Thank God. Wanta is my advisor. This study would have not been possible without Dr. Logan’s encouragement and valuable advice. Dr. Logan has given me thoughtful feedback and has thought me the importance o f professional credibility and conceptual clarity. Dr. Logan is my first mentor who helped me when I faced the most difficult time during my doctoral student period. Dr. Cropp has been much more than my dissertation committee member. He has encouraged me to think critically and inspired me to question about the implications o f this study. He provided guidance and helpful questions from the early stages o f this study. Dr. Osterlind was instrumental in this study. His suggestions made this research stay in the course. He was always available whenever I needed his knowledge on statistics. I would like to express my deepest gratitude to Dr. Benoit. His Functional Theory o f Political Discourse influences me tremendously both academically and personally. I will keep pursing his theory throughout my life. Even though I am not a politician, I start questioning me, “What I am doing right now is ii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. functional?” He also shared his so much valuable time with me whenever I needed him. I cannot thank him enough. He is my teacher, mentor, and friend. I have to mention that I received a great amount o f support from JongHyuk Lee and Yun Jung, Choi. We worked together for the AEJMC and ICA conference papers and shared joy. Finally, my special thanks go to the Missouri School o f Journalism that supported me for three years. It has been a great experience to study in Columbia Missouri. Ill Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. PUBLIC RELATIONS CAMAPGINS IN 2002 KOREAN PRESIDENTIAL ELECTION; FUNCTIONAL ANALYSIS OF POLITICAL DISOURSE AND MEDIA EFFECTS OF AGENDA SETTING AND FAVORABILITY Cheolhan Lee Dr. Wayne Wanta, Dissertation Supervisor ABSTRACT This study examined the nature and effects of candidates’ political discourse and media effects on the audience. Based on the television spots, television debates, and press releases, the researcher found that candidates emphasized more on positive comments than negative ones. Candidates also addressed more policy than character. These findings contradicted that political campaign messages are mostly negative and not substantial. However, media frequently reported negative comments o f candidates, which might give the impression that political discourse is mostly negative. This study also examined the media effects on voters. Based on the comparison o f content analysis o f three major Korean newspapers to opinion poll results, this study supported traditional wisdom that media set the public agenda, or media influenced people what to think about, not what to think. However, the highly-involved and politically knowledgeable voters were not influenced by media’s agenda setting ability. They were more influenced by candidates’ campaign messages. Results also showed that not only media set the public agenda but also media’s net favorable (a sum o f positive coverage o f a candidate and negative coverage of the competing candidate) influenced candidates’ stance iv Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. in tracking polls. The implication o f this study is that while media has an ability to set the public agenda, media has stronger effects on audience than previously believed when the net favorability is considered. Media’s favorable coverage explained a substantial amount o f candidates’ fluctuation o f support rates throughout the election. This study contributed to understanding o f media effects by considering the influence candidates’ campaign efforts. V Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF TABLES Table Page 1. Most important vote determination in 2002 presidential election:Policy or Character.................................................................................................... 9 2. Functions o f Political Discourse by Medium............................................68 3. Candidates’ Use of Functions by Medium................................................ 71 4. Forms o f Policy Topics and Incumbency.................................................. 73 5. Functions o f Topics by Functions................................................. 6. Topics o f Utterances in Presidential Discourse....................................... 78 7. Functions o f Campaign and Media’s Use o f Campaign Discourse 8. Topics o f Campaign and Media’s Use o f Campaign Discourse..................................................................................................... 80 9. Differences in Horse-race Coverage During the Election........................81 10. Issue Priorities of Candidates across Media............................................. 83 11. Policy priorities among Candidates, Media, and Voters.........................86 12. Media Coverage of Candidates and the Public’s Evaluations of the Candidates........................................................ 76 79 89 13. The Supporting Rages o f Candidates in Tracking Polls and Net Favorability o f Media..................................................................................91 14. Testing Media Favorability and Support Rates by Single Regression Analysis........................................................ VI Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 91 LIST OF ILLUSTRATIONS Figure Page 1. The Effects o f Functions and Topics o f Candidates Discourse on Media and the Public....................................... 7 2. Explanations o f Image Restoration Strategies............................................ 14 3. First Level Agenda Setting Process............................................................. 44 4. An Agenda of Cognitive Attributes at the Second Level Agenda Setting................................................. 5. The Mechanism o f Affective Second Level Agenda Setting 47 ...........48 vu Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS ACKNOWLEDGEMENTS.....................................................................................ii ABSTRACT........................................ iv LIST OF TABLES ..................................................................................... vi ix LIST OF ILLUSTRATIONS........................................................ Chapter 1. INTRODUCTION.......................................................................................1 Study Overview and Justification Purposes o f the Study Backgrounds o f Korean Presidential Election 2. LITERATURE REVIEW ............................ The Functional Theory o f Political Discourse Presidential Television Spots Presidential Television Debates Press Releases and Media Agenda News Values News Values in Politics The Public Reaction to Negative News News Framing and Framing Effects Agenda Setting Theory Source Agenda Setting During the Election Campaign Audience Involvement in the Agendas Setting Process Characteristics o f Issues and Agenda Setting Agenda Setting Effects on Internet Forum Participants Agenda Setting at the Second Level The Effects o f Media’s Favorable Coverage o f Candidates 3. HYPOTHESES AND RESEARCH QUESTIONS... 10 ........... .52 4. METHODLOGY ............................................................................ 57 Coding procedure for presidential candidate discourse Coding procedure for newspaper coverage Poll data and Internet Forum Discussions Limitation o f using cross-sectional approach Intercoder reliability V lll Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5. RESULTS .............................................................. Functional Theory o f Political Discourse Media Use o f Candidates’ Discourse Media Coverage of Horse-race during the Election Agenda Setting Media Favorability 67 6. DISCUSSION................................................................................................92 REFERENCE LIST.............................................................................................. 112 APPENDIX EXAMPLES OF CANDIDATE DISCOURSE.........................................124 VITA............................................................................................................................125 IX Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CH APTER 1; INTRODUCTION Study Overview and Justification It is a new idea for both Korean presidential candidates and the public that the candidates’ messages are mostly conveyed by mass media. There has been strong criticism in every Korean presidential election that the candidates spend too much money, thus burdening Korean economy. As a response, in the 2002 presidential election, the candidates were prohibited from organizing mass rallies (which comprise the bulk o f campaign cost) for the first time. Instead, the candidates were allowed to use mass media with the media cost subsidized by the government. This election saw frequent use o f televised primary and general debates, television spots, and numerous endorsement speeches. While the importance o f the media’s coverage o f the candidates remains the same, the candidates’ media campaigns bear much more importance than any previous elections. The fact that presidential campaign messages are mostly delivered by media invokes several concerns. One o f the concerns is that candidates accentuate the negative or primarily attack competing candidates in their use of mass media. Most political columnists o f the major Korean newspapers worry about this phenomenon, arguing that it will cause voters’ cynicism towards politics and eventually alienate voters (Ansolabehre & Iyengar, 1995). 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Nonetheless, attacks seem to be inevitable because elections are inherently competitive. However, Korean scholars seem to worry about excessive attacks at the cost o f providing concrete information about the candidates such as policy positions in controversial issues or future plans (Yang, 2003). Another big concern o f media politics is that candidates place more emphasis on images than issues. Critics again worry that voters have to choose a presidential candidate without substantial information about his policy positions (Shin & Lee, 2003). In fact a candidate’s character, for example, is needed information. However, candidates cannot provide all the information about their policies nor can the public process all the information. Thus, a candidate’s character works as a broad guide for voters. U.S. media polls have shown that the public does not expect candidates to focus more on character than policy positions (see Benoit, 2003). However, Korean media polls showed that voters believed character is more important than policy when they decide their vote (See Table 1). The Korean voters’ preference to character comments might affect the application o f Functional Theory of Political Discourse. The functions and topics continue to be the subject o f empirical research; however, systematic analysis o f Korean political discourse is scarce. Even though campaign analysts of media unanimously argued that candidates relied too much on negative appeals in their political discourse, it cannot be concluded that the political messages are mainly negative and image-oriented without much systematic research (Shin & Lee, 2003). In order to understand functions and topics o f political discourse the Functional Theory o f Political Discourse, which is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. successfully applied to various types of political messages such as advertisements, debates, keynote addresses, and candidates’ webpages (Benoit 1999), is the right choice. This study adopts the functional analysis to explore the nature of the Korean presidential candidates’ message during the 2002 election. Media coverage o f presidential candidates is one o f the most important parts o f an election campaign. As Walter Lippmann (1922) stated, “Universally it is admitted that the press is the chief means o f contact with the unseen environment (p.4).” This statement is true considering that few voters actually meet the candidate personally; most voters know about the candidates only through various media sources. This is especially the case for Korea because the official election campaign period lasts only six weeks. Media’s ability to transfer the agenda to the public is known as agendasetting effect. Since McCombs and Shaw (1972) initiated the agenda setting study, it has become well established that the media set the public’s agenda and not the other way around. Agenda setting during the presidential election should be linked to priming effects; the most important problems in a country should be solved by the highest position in the government— the president. The media’s coverage o f issues becomes the criteria upon which voters evaluate their candidates. Thus, it is critical for candidates to influence the media’s agenda during the election. Relatively few empirical studies on the effects o f media on voters exist in Korea. The possible reasons are the short history o f media-oriented elections and the lack o f poll data asking, “What are the most important issues in Korea?” This study will unveil the nature o f the candidates’ political discourse Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and media’s effects on voters during the election based on empirical data. The media’s ability to set public agenda is commonly accepted, despite the lack of empirical evidence. Because agenda-setting is not an automatic process and numerous factors influence the process, this study will examine how agendasetting effects work during a presidential election. Purpose o f the Study The purpose o f this study is threefold. First, it will study the functions and topics o f political discourse. There are concerns that Korean presidential candidates place too much emphasis on character rather than policy and rely too heavily on negative political discourse. However, few studies have actually analyzed the presidential candidates’ discourse. Thus, this study will empirically find the functions and topics o f Korean presidential candidates’ discourse and examine Benoit’s Functional Theory Political Discourse that has been successfully applied to the U.S. presidential discourse. The researcher assumes that television spots, televised debates, and news releases together represent the candidates’ discourse. Second, the effects of candidates’ discourse on the media will be analyzed. In a media-oriented election, it is essential for candidates to convey their own policy positions and character to the public through the mass media. Previous studies on election coverage show that newspapers emphasize the negative aspects of campaigns and also focus on horse-race aspects o f campaign at the cost of policy (issue) coverage. In a sense, the functions o f the candidates’ discourse Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. can be understood as framing efforts used to influence media content. This study will analyze how mass media cover the candidates’ presidential discourse and determine what frames are predominantly used by the media during the election. Third, this study will examine the agenda setting effects using Gallup poll results. Recently, candidates have the ability to directly set the public agenda regardless o f media’s influence on the voters. The rationale is that media emphasize non-substantial issues too frequently, and as a result, fail to inform the public (Lichter & Noyes, 1995). This study will find who sets the public agenda during the 2002 Korean presidential election. Agenda setting is one o f the most established effects in the field; agendasetting effects are inherently connected to priming effects, or the process in which the media attend to some issues and people evaluate election candidates based on the issues that media emphasized during the election (Iyengar & Kinder, 1987). People choose a president knowing that he will be responsible for important issues society faces. Moreover, recent studies broaden the agenda-setting process by adding new dimensions. In line with traditional agenda-setting studies, the second-level agenda-setting hypothesis posits a more powerful and subtle impact on the audience, which increases the importance o f the candidates’ agenda setting efforts. Finally, the researcher will examine the media’s framing effects on the public’ online discussion. Framing effects can be understood in various ways; however, this study posits that framing effect is media content’s influence on voters attitudes and behavior (not the cognitive level upon which most early Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. agenda setting studies were conducted). This study will examine whether the media’s favorable coverage o f a candidate will directly lead to the public’ support o f the candidates. Previous framing effect studies report that the media’s net favorable coverage of a candidate is statistically significantly related to the candidate’s stance on tracking poll results (e.g., Shaw, 1999; Shelly & Hwang, 1991). Based on these findings, this study will examine whether this finding is applicable to 2002 Korean presidential election. Significance of the Study The unique point o f this study is that the focus is on the candidates’ public relations campaigns. Candidates exert every effort to favorably influence the public and media during the election. This study examines the composition o f Korean presidential candidates’ political discourse and explores how these candidates’ agenda-setting efforts influence the mass media and voters. Thus, the main questions o f this study examine the functions and topics o f candidates’ political discourse and explore the framing and agenda setting effects among the candidate discourse, mass media, and the public (See Figure 1). Finally, the researcher has to point out that presidential election campaign in Korea is often called Daesun Hongbo or Presidential Public Relations. The election campaign is usually regarded as public information campaign such as encouraging people to participate in voting. For this reason, this study uses public relations campaigns to indicate presidential election campaign efforts. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 1 The Effects o f Function and Topic o f Candidates Discourse on Media and the Public The Public Campaign Discourse (2)- ( 1) VaWnce Attitude Knowledge I T Mass Media (3) (4) (1) Media framing effects: What functions and topics o f candidates’ discourse are frequently selected by journalists? (2) Source influence on public: Does a candidate’s agenda set the public agenda? (3) Media content’s effects (framing effects and second-level agenda setting effects): Does media’s favorable coverage o f a candidate lead to the public’s favorable discussion about the candidates (second-level agenda setting)? Does media’s favorable coverage influence the candidates’ standing in the tracking polls (framing effects)? (4) Agenda setting effects: Does the mass media set the public agenda? Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Backgrounds of Korean Presidential Election in 2002 In 2002, the presidential debates bear more importance than ever before. The 2002 presidential election has been termed the nation’s first so-called media election since the candidates were allowed to use mass media— the cost was mostly subsidized by the government, which paid for televised primary and general debates, television spots, and numerous ten minutes speeches (some of them were not available). Furthermore, the candidates were prohibited from organizing mass rallies (Song, 2003). This election was an important clash in two significant aspects. First, the weakest link in the Korean democratic process was the political parties. In the past, political parties were essentially entourages centered on political leaders. In every election, new parties were made by the leading candidates (Kim, 2002). As a result, political parties have no continuity. In the past, people voted for candidates’ personalities and essentially in regional blocs (Steinberg, 2002). In the 2002 presidential election, however, no new parties were created for the election (Steinberg, 2002). The candidates for president were nominated through the primary, which means that the presidential candidates received support from established political parties (Kim, 2002). In addition, clear policy issues such as North Korea, Korea-the United States relationships concerning Status o f Forces Agreement (SOFA) and social welfare distinguished the political parties. For example, the Millennium Democratic Party (MDP) has fostered President Kim’s “Sunshine Policy” o f accommodation with the North for which President Kim won the Nobel Peace Prize (Ryu, 2002). As a candidate o f MDP, 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Roh Moo-Hyun has endorsed and promised to continue “Sunshine Policy” when elected. On the other hand, Lee Hoi-Chang, the candidate o f Grand National Party (GNP), who lost to Kim Dae-Jung in the last election, proposes more cautious and tempered dealings with the North (Ryu, 2002). Roh embraced a more approach whereas Lee preferred the conservative ideology on every policy. In this 2002 Korean presidential election, Lee of the GNP candidate and Kwon Young-Ghil o f the Democratic Labor Party (DLP) challenged the incumbent party candidate, Roh from MDP (Choi, 2002). In Korea, the president cannot run for the second term; thus, incumbent Kim Dae-Jung could not join the race. In the United States, Wattenberg (1991) argued that candidate-centered politics have increased over time as the political parties lost power. It appears that somewhat the opposite has occurred in Korea, as the rise o f stable political parties has attenuated to some extent the importance o f individual candidates and increased the importance o f the policy differences. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CH APTER 2* LITERATURE REVIEW The Functional Theory o f Political Discourse Functional Theory begins with the assumption that “a voter chooses to vote for the candidate who appears preferable to him or her, on whatever criteria are most salient to that voter” (Benoit, Balney, & Pier, 1998, p.4). A candidate does not need win every vote; nor must a candidate completely perfect: A candidate must simply appear preferable to other candidates for enough voters to win the election. Popkin (1994) wrote that “each campaign tries hard to make its side look better and the other side worse” (p. 232). Accordingly, political campaign discourse has three essential functions for a candidate to achieve the desired end: (1) acclaims, or remarks that enhance the candidate’s qualifications as an office-holder (positive utterances, self-praise), (2) attacks, or utterances that reduce the opponent’s qualifications as an office-holder (negative utterances, criticism), and (3) defenses, or comments that refute attacks (Benoit, Blaney, & Pier, 1998; Benoit 1999). These three functions allow voters to perform a simple form o f cost-benefit analysis. Acclaims increase a candidate’s perceived benefits, attacks increase an opponent’s perceived costs, and defenses can reduce a candidate’s perceived costs. Any one o f these strategies has the potential to persuade voters to vote for a candidate or vote against his or her opponents (Benoit, 1999). 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Note that Functional Theory does not assert that all voters explicitly calculate costs and benefits; only that political messages using these functions tend to influence attitudes by making the contenders appear more or less desirable to voters. Functional Theory argues that acclaims should, in general, be the most common function because they possess no drawbacks. Attacks, in contrast, are expected to be somewhat less common than acclaims because voters report that they do not like mud-slingling (Merritt, 1984; Stewart, 1975). Defenses should be the least common function because they possess three potential drawbacks. They make candidates appear reactive rather than proactive: Candidates do not initiate an issue when they use defense; instead, they respond to an opponent’s attack. Second, defenses usually take a candidate “off-message” because attacks are more likely to concern a candidate’s areas o f weakness. Third, a candidate must identify an attack to attempt to counteract it with a defense. However, if voters have not been exposed to an attack, or have forgotten it, the act o f refuting that attack may inform or remind voters o f a candidate’s alleged weakness. So, acclaims should be more common than attacks and defenses should be the least common function. Functional Theory (e.g., Benoit, 1999) declares that these three functions occur on two topics: policy and character. Similarly, Pomper (1975) acknowledged that many voters “change their partisan choice from one election to the next, and these change are most closely related to their positions on the issues and their assessment o f the abilities o f the candidates” (p. 10). Policy comments can concern past deeds, future plans, and general goals; character comments may 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. address personal qualities, leadership abilities, and ideals. The appendix illustrates acclaims and attacks on each form o f policy and character. In addition, the candidate’s status influences the functions o f political discourse; incumbent party candidates acclaimed more than they attacked, whereas challenger party candidate attacked more than they acclaimed (Benoit, Blaney, & Pier, 1998; Benoit & Brazel, 2002). For example, in the 2002 Korean presidential race, Lee used acclaiming when he promised that he would create one million job opportunities for women. Surely creating many jobs would be a desirable outcome. Attacks are negative utterances that emphasize an opponent’s disadvantages. Lee attacked his opponent by arguing that Roh would continue the current education policy which has failed to provide quality education. This is a reason for voters to consider Roh less desirable, which would increase Lee’s net favorability. Defenses respond to (refute) alleged weaknesses. Lee attacked Roh by arguing that the economic policies of Roh’s party were responsible for inflation. Roh aired a television spot which defended his party’s economic policies, arguing that the inflation rate was less than that caused by Lee’s party in the past. If Roh can refute attacks on his economic policy, that should improve his overall favorability. 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Defensive Strategies for Candidates Benoit (1995) suggested that the purpose o f the candidates’ defense utterances is, in general, a goal-directed activity. It focuses on one particular goal in discourse: restoring or protecting one’s reputation. Because a politician’s face, image, or reputation is so important, when politicians believe it is threatened, they are motivated to take action to alleviate this concern. Basically, an attack on one’s image is composed o f two factors: (1) an act occurred which is undesirable, and (2) the actor is responsible for that action (Benoit, 1995). Based on the Benoit’s Image Restoration Theory, five defense strategies can be drawn as shown below Figure 2. 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2 Explanation o f Image Restoration Strategies Denial Simple denial: I did not perform act Shifting the blame: Act performed by another Evading Responsibility Provocation: I responded to act o f another Defeasibility: Lack o f information or ability Accident: Act was a mishap Good intention: Meant well in act Reducing Offensiveness o f Event Bolstering: Stress good traits Minimization: Act not serious Differentiation: Act less offensive Transcendence: More important considerations Attack accuser: Reduce credibility o f accuser Compensation: Reimburse victim Corrective Action: Plan to solve or prevent problem Mortification: Apologize for act Source: Benoit (1997, p. 179). 14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Together, these three functions work like an informal form of cost-benefit analysis for voters: acclaims increase a candidate’s benefits, attacks increase an opponent’s costs, and defenses minimize or eliminate a candidate’s alleged costs (Benoit, Blaney, & Pier, 1998). In recent years, more American voters have said that policy, rather than character, is a more important determinant o f their vote for president (see Benoit, 2003). Given this preference o f most voters, Functional Theory predicts that policy utterances will be more common than character comments. Note that if enough voters shifted their preference. Functional Theory would predict that candidates would respond by emphasizing character more than policy. Using data from presidential primary TV spots from 1952-2000, presidential primary debates from 1948-2000, nomination acceptance addresses from 1952-2000, general TV spots from 1952-2000, and general debates from 1960 and 1976-2000, Benoit (2003) reported that in each one o f these message forms the candidate who discussed policy more than his opponent was significantly more likely to win the election. Presidential Television Spots Television spots are a vital component o f the modem presidential campaign in the United States. This is the case for Korea, too. Korean people have experienced the political ads twice, starting with the 1997 presidential election campaign. Even though the presidential spot is still a new concept, it is important in several ways. First, candidates spend huge amounts of money on political advertising. Ansolabehere and Iyengar (1995) reported that “The 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. amounts o f money spent on political advertising are staggering: Hundreds of millions of dollars are poured into what has become the main means o f political communication in the United Stats” (p. 3). In Korea, the candidate spent about a billion won, or ten-thousand U.S. dollars for each television ad (Chosunilbo, 2002). The cost o f ad makes candidates select their messages very carefully; thus, the presidential ads clearly and most succinctly show their issue positions and personal qualities. Because candidates do not choose their advertising message lightly, the ads contribute to the understanding of functions and topics of political campaign. Second, voters learn substantial amounts o f information on the presidential candidates and their policy positions from political advertisements. Patterson and McClure (1972) reported that televised political ads contributed voters to increase the knowledge about candidates’ policy positions and personal qualifications. Contrary to general belief that voters receive most information on candidates from news, voters learned about candidates’ policy positions and personal qualifications more from the political ads than from new (Patterson & McClure, 1976; Kern, 1989). In a similar vein, Brians and Wattenberg (1996) reported that “Recalling political ads is more significantly related to knowledge o f candidates’ issue positions than is reading the newspaper or watching political news on television (p. 185). These results might come from the fact that political ads addressed candidates’ policy positions and character qualifications whereas the news heavily focused on the coverage o f the horse race. 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Third, empirical studies o f political advertising effects show that such ads are influential. McClure and Patterson (1974) reported that “Exposure to the political advertisements was consistently related to voter belief change” (p. 16). Joslyn (1981) found that there was a positive relationship between election outcomes and advertising expenditures. Thus, research demonstrates that political advertising can influence voter decisions. Functions: Positive ads versus Negative ads. Most research on presidential television advertising categorizes ads into positive and negative spots. Three important longitudinal studies o f presidential television ads show that presidential candidates rely more on positive advertisements than negative ones. Based on a content analysis of political advertisements aired from 1952 to 1996, Kaid and Johnston (2000) reported that 62 percent o f the ads were positive and 38 percent were negative. Kaid and Johnson (2000) reported no differential use o f negative spots according to the incumbent status. West (1997) studied the presidential spots from 1952 to 1996, reporting that 48 percent of the ads were positive and 54 percent were negative. Finally, Benoit (2001) analyzed the presidential spots from 1952 to 2000, reporting that candidates most used their television spots most frequently to acclaim: 60 percent o f the themes in thee message engaged in selfpraise, followed by attacks: 39 percent o f the themes engaged in attack of competing candidates. In addition, Benoit added one more function: defense. When attacked, a candidate can comment and refute what is addressed by an opponent. Garramone (1985) suggested that the rebuttal is fan effective tool for 17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. lowering the evaluation o f the candidate who initiated the attack. However, Benoit (1999) found that defense were seldom used (1%). With regard to the effects o f negative ads, there is a concern that candidates heavily rely on negative advertising, which causes voters’ cynicism against elections and eventually leads to low-voter turnout (Pfau & Kenski, 1990). However, it is argued that negative spots actually help voters to make their decisions by providing the different issue positions between candidates. West (1993) argued that the most substantive information actually came in negative spots. The reason is that negative advertising is more likely to have policyoriented content because campaigners need a clear reason to attack the opponent. Jamieson, Waldman, and Sherr (2000) examined the amount o f policy content present in televised presidential campaign ads from 1952 to 1996. They found that attack ads addressed substantial issues as compared to advocacy ads (Jamieson et al, 2000). Topics: Policy (Issue) versus Character (Image). Most research on televised political spots has reported a heavier emphasis on issues than on image. Benoit (2001) studied public opinion polls from 1976 to 2000, suggesting that candidates’ emphasis on policy is a response to voter preference. The poll results show that policy (issue) is a more important determinant o f their voters than character (image). On the contrary, Korean voters in the 2002 presidential election answer that character is a more important determinant o f their votes than policy except one poll result. Thus, it is possible that the candidates who 18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. emphasize character are more likely to win the election as compared to the candidates emphasizing policy. Table 1 Most important vote determinant in 2002 presidential election: Policy or character Date Policy Character Poll 09/24/02 39% 35% Joungang Daily Poll 10/11/02 40% 42% Joins.com Poll 11/24/02 37% 57% Donga Daily Poll 11/26/02 38% 42% Jungang Daily Poll Note: Respondents were allowed to answer, “Don’t know” and “unsure” responses. Sources: All polls retrieved from the electric archive from Korean Press Association, which is available at http://www.kinds.or.kr 19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Not only issue but also image is an important topic o f the presidential spots. In a sense, voters choose their leader based on his or her ability, not the issue positions. Shyles (1986) wrote that image advertising is concerned with the candidate’ projected personality traits and character attributes, such as compassion, empathy, integrity, strength, activity, and knowledge. Joslyn (1980) further explained that almost all campaign visuals emphasize image factors. Garramone (1986) suggested that image advertising focuses on the “sum o f the perceived personal and professional characteristics o f the candidate” (p.236). As early as 1958, there was a concern that the emphasis o f campaign shifted from issues to images (Rubin, 1968). Patterson and McClure (1973) argued that political advertising has not provided voters with meaningful information by emphasizing image material while ignoring political issues. However, the longitudinal studies on the presidential television spots have reported that issues outnumber image. Joslyn (1986) studied 156 television spots aired in the 1980s, reporting that 79.6 percent o f presidential ads addressed issues (and the rest image). Joslyn (1986) suggested that political spots should not be a poor source o f information as many other observers have claimed. West (1997) reported that 61 percent concerned issues and 32 percent personal qualities. Kaid and Johnston (2000) found that 66 percent of advertisements emphasized issues and 34 percent image. Adopting themes as the coding unit, Benoit (2001) found that the presidential candidates discussed policy (61%) more than character (39%) based on content analysis o f the presidential spots from 1952 to 2000. 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Presidential Television Debates Televised presidential debates, although a relatively new political event in South Korea, have emerged as an important part o f the presidential election campaign. In 1997 Dae-Jung Kim and Hoi-Chang Lee initiated the practice o f televised presidential debates in the general campaign and the debates are expected to continue in the future. Debates attracted voters’ attention: In 2002, two-thirds o f Korean people watched one o f three presidential debates (Korea Press Foundation, 2003). Presidential debates merit the scholarly attention for several reasons. First, presidential debates are very important campaign events because they provide voters with a chance to observe the performance o f presidential candidates in face-to-face confrontation discussing the same topics (Benoit & Harthcock, 1999; Benoit & Wells, 1996; Carlin, 1994). Jamieson (1987) states that one advantage o f political debates is that they offer an extensive view o f the candidates’ issue positions and character, which provides voters with the opportunity to directly compare and contrast the candidates at the same time. Because casting a vote means choosing which candidate appears preferable to others, voters benefit from the opportunity to compare and contrast the candidates. Presidential debates fit that purpose very well. Second, debates are unique in terms o f message length. Because debates often last more than an hour, they are more spontaneous than other message forms: Voters may obtain a somewhat less contrived impression o f the candidates from watching debates than the can get from other kinds o f campaign messages, 21 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. like television spots. While candidates do prepare for the debates, not every question from the panelists, moderators, or audience members can be anticipated; not every remark from an opponent can be anticipated. Furthermore, unlike speeches or television spots with scripts and teleprompters, candidates are not usually allowed to bring prepared notes to debates. Thus, viewers may get a somewhat more spontaneous and accurate view o f the candidate in debates (Benoit, Blaney, & Pier, 1998, p. 172). Thus, it appears that political debates offer more candid views of the candidates compared with other messages such as political ads or formal speeches. Furthermore, presidential debates can influence the result o f the election. According to the survey o f Korean Press Foundation (2003), 20% of voters answered that television debates are the most important source o f information on candidates. Some have suggested that debates do not affect election outcomes: Jamieson and Birdsell (1988) stressed, “debates don’t very often convert partisans on one side to the other” (p. 161). However, this does not mean they are incapable o f influencing the outcome o f a campaign. However, in 2002 11.8% o f voters answered that they decided their supporting candidate based on the performance of the debates and this many voters was enough to have decided the outcome of this election (Korea Press Foundation, 2003). In addition, debates can influence elections by persuading undecided voters (Carlin, 1994; Pfau & Kang, 1991). It was reported that about 20% o f Korean voters had not decided how to vote at the time o f the debate (Kim, 2002a); they could have been influenced by the debate. Thus, presidential debates clearly merit scholarly attention. 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. There are only few studies focused on the Korean presidential debates. Yang (2003) studied the effects o f presidential debates, reporting that watching a presidential debate changed the previous evaluations o f the candidates’ of character; however, no such effects were found for leadership ability or policy preferences. Based on the content analysis o f the 2002 primary debates, Song (2002) found that the MDP candidates mostly addressed character (37%), followed by political reforms (28%) and other issues such as women, social welfare, education, and science (25%), and GNP candidates discussed mostly political reforms (39%) and character (37%). In the cases o f United States, a series o f studies address the nature o f presidential debates (Benoit, Blaney, & Pier, 1998; Benoit & Brazeal, 2002; Benoit & Harthcock, 1999; Wells, 1999). Benoit, Pier, Brazeal, Klyukovski, and Aime (2002) summarized analyses o f the presidential debates o f 1960, 19762000, finding that acclaims, or positive utterances about the candidate himself were the most frequent (55%) followed by attacks or negative utterances against the competitors (35%), and defense or refutations o f attacks (10%). They found that the candidates’ statements in debates focused on policy (75%) more than character (25%). In addition, Benoit et al. (2002) reported that the challenger tended to attack more often than the incumbent whereas the incumbent was more likely to acclaim than the challenger. The time has come to extend Functional Theory to non-American debates to discover how presidential debates in other countries are similar or different from U.S. debates. 23 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Press Releases and Media Agenda. Although public relation practitioners contribute to public office candidates for in election period, only relatively few studies focus on how press releases influence media agenda. The reason might be that journalist, ideally speaking, are not supposed to be influenced by the press release. But, that is not always the case. Concerning the source’s ability to influence mass media in the context of political campaign, Gans (1979) stated that “although it takes two to tango, either sources or journalists can lead, but more often than not, sources do the leading” (p. 116). His statement well illustrated how the political news during presidential election is being covered, because during that particular period, whatever statement the candidates make is newsworthy. In 1982, Gandy coined the term, “information subsidy” to describe controlled access o f information, as little cost or effort to the person receiving the information (1982). Public relation professionals provide information subsidies to mass media to distribute information systematically on behalf of their clients (Turk, 1985). If practitioners can obtain media placement o f their subsidies, then they have successfully influenced the media agenda— a process that has come to be known as “agenda building”. The success of these efforts is reported in part by studies that estimate 40% to 50% or more o f daily newspaper content originates from news releases (Wilcox, Adult, & Agee, 1988). Turk (1988) examined press releases in 1984 and 1985, and found that public information offices for six Louisiana state agencies provided information to eight Louisiana daily newspapers. More specifically, she found that about half o f the information 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. provided by the public information officers was used in subsequently published stories and that most topics identified by public information officers as salient, were the topics given salience in media coverage. Public relations information subsidies form part o f the source competitive market, in which sources and news department enjoy a directly reciprocal relationship. Sources depend on the news media to carry their information to the public; and in turn, the media depend on sources to supply them with information (McManus, 1994). Related to the media coverage during election campaign, two previous studies focused the public relations practitioners’ ability to set the media agenda. Bolden tracked Texas gubernatorial campaign releases during election and reported that about 20% o f the press releases were used by the media. However, content analysis of media coverage o f the press releases during a period of political campaign, found that 26 news releases generated only 7.6 percent o f the stories (Kaid, 1987). Recent studies show that press releases influence media content. Tedesco (2001) content-analyzed 1,479 candidate press releases and 756 network news stories using key words in context frames during the 2000 presidential primaries. Candidates and media issue agendas were positively correlated, especially for the Republican candidates. Tedesco (2001) further examined the direction of influence by examining autocorrelations, which suggested the relationship between candidates and media is reciprocal. However, the process frames were significantly correlated only for Republican candidate John McCain and the 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. networks, which the author explains may demonstrate that McCain and the media had a "love-affair" during the primaries. News Values Many have attempted to define news; however, none have provided a perfect answer. Hall (1981) suggests, “news value is one o f the most opaque structures of meaning modem society. All trae joumalists are supposed to posses it; few can or are willing to identify and define it” (p. 234). Park (1967) argued that news as a special form o f knowledge that has certain implications for human behaviors and mobility. Stephens’ definition of news is “new information about a subject o f some public interest that is shared with some portion o f the public” (Stephens, 1997, p. 9). Breed (1956) proposes multiple dimensions for defining news. News is the report of a recent event judged by newsmen to be worthy of publication for the interest and/or information o f members of their audience, and has the following characteristics to a greater or less extent: recency, interestingness, accuracy, availability, simplicity, significance, prudence, objectivity and superficiality; it is frequently mediated by an association, increasingly interpretive and is joumalistically stylized in form.” (p. 447) An indirect way to define news is to study news values. Many scholars address vast amount o f literature on newsworthiness. Galtung and Ruge (1965) suggest that events tend to become news if they have twelve factors such as frequency, threshold, intensity, unambiguity, meaningfulness, consonance, predictability, unexpectedness, continuity, composition, relevance to elite nations, elite people or something negative. Rosenberg (1970) adds three more concepts 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. to Galtung and Ruge’s study, which are, the degree of importance of event, physical distance of the events, and the degree o f the predictability of the events. Based on the sociological approach, Gans (1979) suggests that news values are some events having characteristics such as ethnocentrism, altruistic democracy, responsible capitalism, moderatism, social order, and national leadership. Based on previous news values studies, Shoemaker, Danielian, and Brebdlinger (1991) produce the model to test the news values systemically. They argue that various news values can be reduced two broad dimensions: deviance and significance. Deviance includes novelty, oddity, unusualness, prominence, sensationalism, and conflict or controversy; significance consists of importance, impact, consequence, and interest. In addition. Shoemaker et al., (1991) add two sub-concepts—proximity and timeliness to the deviance dimension. News values in the politics There is a consensus that presidential election receives extensive news coverage because o f its news values (Lichter & Noyes, 1995; Patterson, 1993). The president is the highest governmental position upon which the most important decisions are made. An election campaign often can be described as a drama in which two heroes fight each other for the position. Among news values mentioned above, the negative coverage and human interest are most frequently discussed in the media coverage during election campaigns. In this light, Patterson and McClure (1976) suggest that media mainly focused on candidates’ styles, stragecraft, strategy, campaign money, and staffs instead o f focusing substantive issues. When joumalists mentioned issues in their election coverage, it 27 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. means usually gaffe, controversies, strategies, and other non-substantial issues. When joumalists focused on the substantial issues, they more focused on issues that are related to simple ones (where candidates’ positions are easily contrasted without further explanations) rather than complex ones (Patterson & McClure, 1976). The only justification for emphasizing these trivial and controversial issues is news value such as human interest. As a result o f media’s focusing on horse race and strategy coverage, it is reported by several studies that voters leara more from candidates’ campaign messages as compared to the news coverage of elections (Patterson & McClure, 1976; Kern, 1989). The reason is that the press focused more on horse race, campaign strategies, campaign money, and campaign staffs than on candidates’ issue standings. In addition, joumalists frequently cover the candidates’ gaffe, controversies, and other non-substantial issues in the name o f issue coverage (Feitag, 2000; Johnson, 1993; Jonhson, Boudreau, & Glowaki, 1996; Kenney & Simson, 1993; King, 1990). It is also reported that joumalists act like character cop when they cover the candidates, meaning that joumalists usually emphasize candidates’ character flaws (King, 1995). In sum, presidential election coverage tends to focus on the strategic game played by the candidates in order to achieve presidency. As an unexpected result, the process o f choosing leader is trivialized because o f media’s game approach to cover the election. 28 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The public's reaction to negative news Although voters often complain about the harshness o f negative information, they find negative information more memorable than information focusing on candidates’ strength (Basil, Schooler, & Reeves, 1991). Negative political information may be easy to remember because o f its uniqueness, making it more likely to be noticed and easily convince the audience (Lau, 1985). It is possible that negative information is more credible and more informative for voters because negative information is often unexpected. Based on psychological approach, it is argued that negative information is perceived to be more important than positive information because people are often more motivated to avoid costs rather than to achieve gains (Kahneman & Tvesky, 1979). On a similar note, Meyerowitz and Chaiken (1987) suggest that a negatively-framed message is more persuasive than a positively-framed one. These previous studies also show why political campaign practitioners are not hesitant to attack their opponents through advertising and press releases. Based on the assumption that people regard negative information more important than positive one, mass media are likely to put more emphasis on negative information in politics (Stephens, 1997; Benoit & Currie, 2001). With regard to news values, conflict stories are always good sources. Gant and Dimmik (2000) found that news tends to rely on “bad” news, such as crime, accidents, disasters, and conflicts in politics. 29 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. News Framing and Framing Effects Frames are considered schemes for both presenting and comprehending news; thus these two concepts o f framing can be categorized into media frames and audience frames. (Scheufele, 1999). The media framing studies tend to focus on the sources’ influence on media’s presentation o f news (e.g., Huang, 1996), whereas audienceframing studies emphasize the media effects on audiences’ interpreting and responding the news (e.g., Rhee, 1999). Gitlin (1980) defines that “media frames are persistent patterns of cognition, interpretation and presentation, o f selection, emphasis, and exclusion, by which symbol-handlers routinely organize discourse” (p. 7). Gitlin also (1980) suggests that framing serves to “organize the world both for joumalists who report it and for audience members who rely on their reports” (p. 7). Media frames function as working routines for joumalists that allow the joumalists to quickly identify and classify information and to package it for efficient rely to their audiences (Gitlin, 1980). In a similar vein, Entman and Rojecki (1993) suggest that framing can be understood in two different aspects: (1) individual frames as informationprocessing individuals and (2) media frames as attributes o f the news itself. With the combination of media frames and individual frames, the framing studies serve as “the bridge between larger social and cultural realms and everyday understandings of social interaction” (Friedland & Zhong, 1996, p. 13). Gamson and Modigliani (1987) suggest that media frame is “a central organizing idea or story line that provides meaning to an unfolding strip of 30 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. events” (p. 143). In a similar vein, Tuchman (1978, p. 193) defined media frames that “news frames organizes everyday reality and the news frame is part of everyday reality.” Entman (1993) stated that “to frame is to select some aspects o f a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation and/or treatment recommendation” (p. 52). The news framing can systemically influence how audience members o f the news understand the events. (Price, Tewksbury, & Powers, 1995). Scheufele (1999) suggests that framing study lacks a commonly shared theoretical model underlying framing research. The reason comes from the different approach applied to the way o f framing is operationalized. Studies of frames as dependent variables have examined the role o f various factors in influencing the creation or modification o f frames (e.g., Shoemaker & Reese, 1996). Studies in which frames serve as independent variables typically are more interested in the effects for framing. In that case, the question is related to how individuals are influenced by media and how they react and engage in political action or participation (Scheufele, 1999). Based on his review o f previous framing literature, Scheufele (1999, p. 108) suggests that future framing studies have to answer those questions as shown below. With respect to media frames dependent variables, the framing studies should ask, “What factors influence the way joumalists or other societal groups frame certain issues?” and “how do these processes work and, as result, what are the frames that joumalists use?” With 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. respect to media as independent variable, the framing research should ask, “How can the audience member play an active role in constructing meaning or resisting media frames.” According to his suggestions, the researcher will examine how functions and topics o f presidential candidates’ discourse influence media and audience members with operationalization o f framing as both independent and dependent variables. In terms of audience framing effects, Iyengar (1991) reported that framing affects publics’ opinions. Iyengar (1991) provided two types o f framing. The epidemic framing focuses on the presentation o f concrete examples and includes individual examples of a larger problem, whereas the thematic framing presents an issue in societal context. This framing does not focus on individual examples however provide the systematic causes o f poverty. These different news frames affect viewers’ different reactions to the poor. Iyengar (1991) found that participants who repeatedly received news stories o f poverty based on episodic frame are more likely to hold the poor themselves responsible for their poverty and see the cause for poverty as the individuals’ character problem (e.g., laziness) than those who receive news stories o f poverty based on thematic framing. Framing effects are also applied to the election campaign. Rhee (1997) reported that people who receive election stories adopting strategy frame (e.g., tactics, winning and losing, horse race) tend to describe the campaign in terms o f strategy-oriented process. Those who received issue-framed news such as inflation, tend to characterize the campaign using issue-oriented approaches. 32 Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission. Agenda Setting Theory The idea o f agenda setting starts with Cohen’s (1963) argument that the mass media may not be successful in telling people what to think, but the mass media are stunningly successful in telling people what to think about. His argument initiated the subsequent agenda setting studies. McCombs and Shaw’s (1972) Chapel Hill study coined the term agenda setting to show the influence of the mass media on the salience o f public issues among the general public. After reviewing more than two hundreds agenda setting studies. Bearing and Rogers (1996) defined agenda setting is “a process through which the mass media communicate the relative importance of various issues and events to the public.” Agendas are “issues or events that are viewed at a point in time as ranked in a hierarchy of importance (p. 4).” Dealing and Rogers (1996) proposes three main areas in the agenda setting studies: the media agenda, the public agenda, and the policy agenda. Most of the agenda setting studies have focused on the relationship between the media and public agenda. Since the seminal McCombs and Shaw’s Chapel Hill study, numerous agenda setting studies successfully replicated the result that the media agenda influenced the public agenda. Agenda Setting Theory: The concept and early studies Klapper (1960) found that in general mass communication did not possess hypodermic need or bullet type o f strong effects on audience and only functions as weak cause o f mediating effects. However, the idea that the mass media ordinarily does not have any effects did not seem reasonable (Severin & Tankard, 33 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2002). McCombs and Shaw (1972) applied a different approach to mass communication effects. They suggested that the mass media had their effects on people’s perceptions rather than on their attitudes and named it agenda setting. Based on this notion, McCombs and Shaw (1972) proposed agenda setting theory states that issues that receive prominent attention by the media become the prominent issues for news audience. Since then, considerable empirical evidence supports this theory. Opinions survey on various topics found that there was a high agreement between public interests and media content (Semetko, Blumler, Gurevithc, & Weaver, 1991; Weaver et al., 1981; McCombs, 1981). Experimental studies also support that media set the public’s agenda, not vice versa. Iyengar and Kinder (1987) report that media content influences the public’s issue salience and also upon those issues the political figures were evaluated. Semetko (1996) summarized that support for the agenda-setting theory has been reported using various methodological approaches. He concluded that agenda setting is a process led by the powerful news media (Semetko, 1996). In the same fashion, McCombs (1994) reviewed agenda setting research and stated that the news media set the public agenda to a considerable degree. Source Agenda Setting During the Election Campaign Unlike the public opinion polls on social issues, agenda setting process during the election period includes two or more candidates having ability to promote their own ideas while attacking each other’s agenda. During a political campaign, candidates play active roles in the agenda setting process. Using advertising, homepages, interviews, or political addresses, 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. presidential candidates can exercise direct control through his or her advertising, and indirectly influence the media by providing the press releases that contain speeches, policy statements, and even personal characteristics. Regarding the political advertising’s ability to influence the media agenda, Roberts and McCombs (1997) reported that candidates have the ability to influence voters directly through paid advertising and indirectly through their campaign efforts to influence the news coverage. In the same fashion, Boyle (2001) confirmed that source agenda setting during the election supported Roberts and McCombs’ finding that the candidates set the media’s agenda. With regard to the source’s ability to influence the mass media in the context o f political campaign, Gans (1979) stated that “although it takes two to tango, either sources or joumalists can lead, but more often than not, sources do the leading” (p. 116). His statement well explained the political news coverage o f the presidential election because what candidates mention is very newsworthy at that time. Several studies support that media is not the only constituent influencing the public’s agenda. Just et al. (1996) described that agenda setting is a process by which multiple actors constract shared meanings about the campaign. Beyond this assertion, Dalton, Beck, Huckfeldt, and Koetzle (1998) propose that the actual agenda setters o f the campaign are candidates, not mass media. They introduced the transactional model which posits that the even though candidates, media, and voters constrain each other, the candidates’ attempt to promote their messages to the voters are successful without much intervention of media. The 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. rationales o f this model come from these three factors. First, candidates’ campaign strategists pretest their messages with focus groups and opinion polls to determine which messages are likely to evoke a public response (Altschuler, 1982). Thus, the candidates’ messages are partially conditioned by the expectations o f voters. Second, the voters may have their won interests, however they may have to respond to the choices provided by the candidates. Third, even if the mass media have their own agenda, their reporting o f the campaign is constrained by the candidates’ actions (Dalton et al., 1998). Thus, transactional model of agenda setting posits that it is possible that strong correlations exist between media content and pubic interests, however such correlations are not necessarily evidence of media causation during the election (Dalton et a l, 1998). It is interesting to note that candidates are the most important agenda setters for the public, which contradicts most previous agenda setting research. Dalton, Beck, Huckfeldt, and Koetzle (1998) argue that during the election candidates set the voters’ agenda while newspapers do not play the dominant agenda-setting role. Two explanations are possible for the limited agenda setting ability o f newspapers during the election. First, the agenda setting of presidential elections is different from the general process o f agenda setting. Presidential elections are structured, highly visible, and institutionalized settings. During the presidential elections relatively well-defined sets o f acts (the candidates and their campaigns) are consciously attempting to shape the public agenda (Dalton et al., 1998). Through their daily campaign annoimcements, the campaigns are trying to define content. In addition, the candidates are systematically monitoring public 36 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. opinion and attempting to persuade the public. This obviously differs from the general process o f interest formation on issues such as crime where there are no dominant actors and a single structure for information (e.g., Ghanem 1997). Unlike this, it is difficult for the press to convince the public to focus attention on issues that are not being discussed by either candidate or that lack the potential to resonate with the public. For newspaper, it is difficult to cover the issues that candidates do not promote. Thus, the public might learn more about campaign issues discussed by candidates. Second, several studies found that voters learn more from candidates’ campaign messages as compared to the news coverage o f elections (Patterson & McClure, 1976; Kern, 1989). The reason is that the press focused more on horse race, campaign strategies, campaign money, and campaign staffs than on candidates’ issue standings. In addition, joumalists frequently cover the candidates’ gaffe, controversies, and other non-substantial issues in the name of issue coverage (Feitag, 2000; Johnson, 1993; Jonhson, Boudreau, & Glowaki, 1996; Kenney & Simson, 1993; King, 1990). It is also reported that joumalists act like character cops when they cover the candidates, meaning that joumalists usually emphasize candidates’ character flaws (King, 1995). In sum, presidential election coverage tends to focus on the strategic game played by the candidates in order to achieve presidency (Patterson, 1980). As a result, the candidates’ campaign messages such as political advertising, debates, and speeches provide voters with candidates’ issue standings, which might set the voter agenda without much interplay o f mass 37 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. media. In this vein, Lichter and Noyes (1995) argued that “voter knowledge does not increase from exposure to day-to-day TV coverage, and increases modestly with day-to-day newspaper reading. Voters do leam from TV coverage of live campaign events, such as convention speeches and debates” (p. 101). Based on this model, one can expect that candidates set the media and public agenda and the mass media do not create the campaign agenda and voter agenda during the election period. Audience Involvement in the Agenda Setting Process Agenda setting theory posits that the media have a strong effect on the salience transferred to issues by individuals. By selecting which news stories about the “world outside” as Lippmann (1922) described, the mass media suggest which issues are deserving o f public attention at the cost o f the issues that do not receive media’s attentions. Traditionally, agenda setting research generally compared aggregate trends in the public agenda to aggregate trends in the news agenda (McCombs, Danielian, & Wanta, 1995). Although most agenda setting research focus on the media’s influence on audience as a whole, relatively little attention is paid to the individual difference during the agenda setting process. The agenda setting process is not an automatic one where audiences do not involve. For example, audience factors can moderate agendas setting effect. Iyengar and Kinder (1987) argue that previous political knowledge moderates the agenda setting process. Based on the experimental studies focusing on different type o f audience, Iyengar and Kinder (1987) reported that the less knowledgeable were more susceptible to media-provided 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. cues about the importance o f various issues. They found that the magnitude o f the agenda setting effect was negatively related with greater interest and participation in politics. In a similar vein, Tsfati (2003) argues that if audiences are active and critical towards news, they may resist the agenda offered by the media. Thus, it is reasonable to assume that Internet Forum participants during the election are more knowledgeable about the candidates than the general public. In sum, agenda setting effects might be weaker for the Internet Forum participants who actively express their concerns as compared to the general public. The Characteristics o f Issues and Agenda Setting Based on the assumption that agenda setting does not take place always, Zucker (1978) suggests that the obtrusiveness o f the issue may be an important factor in whether or not agenda setting occurs. He posits that issues that the public experiences directly, like unemployment, are obtrusive issues; whereas issues that the public may not experience as directly, like pollution, are unobtrusive issues. Based on his study comparing thee obtrusive issues such as the cost o f living, unemployment, and crime with three issues that were unobtrusive issues such as pollution, drug abuse, and the energy crisis, Zucker (1978) found that agenda setting was stronger for unobtrusive issues as compared to obtrusive ones. The researcher suggests, with the same logic, the highly involved audience will take political issues as unobtrusive issue as compared to the general audience. In a similar vein. Weaver (1977) argues that individuals differ in their need for orientation and that his may determine whether or not agenda setting occurs. 39 Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission. Weaver demonstrates that the higher the need for orientation, the more susceptible the individual is to mass media agenda setting effects. Need for orientation is composed o f two factors: the relevance o f the information and the degree o f uncertainty concerning the subject o f the messages. The greater the uncertainty concerning the issue for the audience, the greater the need for orientation they need. This uncertainty will lead audience to rely on the mass media (Severin & Tankard, 2003). Agenda Setting Effects On Internet Forum Participants One o f the contributions in this study is to consider the audience factor in the process o f agenda setting. The characteristics o f Internet Forum participants should be different from that o f the general public, which might mediate the agenda setting effects. Previous studies show that online users, like the general public, express and exchange their opinions, and form and change their public opinion through diverse computer-mediated communications such as e-mails, bulletin boards and chat rooms. Ogan (1993), through the study o f the Turkish Electronic Mail List (TEL), suggested that online users have a good sense o f a community, a nation and the world by exchanging information and sharing experience. One o f the TEL members described their online activities as this: People come and go, say a few witty things, argue politics, ask about help or information, share some news or an article they have recently read, or make annoimcements about something they are involved in (p. 187). Tanner’s (2001) case study o f a Chilean electronic bulletin board on the issue the arrest o f ex-dictator Augusto Pinochet demonstrates the online users’ 40 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. powerful influence on social change. Online users in the study were sensitive to the media coverage, exchanged their opinions, and finally formed an agreed public opinion that influenced the general public and public policies. The differences between online users and the general public are usually explained by the demographic characteristics. Online users have been recognized to be likely to be upscale in occupation, education, and income in comparison with the general public (James, Wotring, & Forrest, 1995; Hills & Hughes, 1998). A demographic report on Prodigy (1990) showed that 65% o f Internet users have college degrees, 67% work as a professional or in management, and have a median income o f $73,000 per year (compared to the national average of $32,100). A 2001 survey (Belden Continuing Marketing Study) also confirmed these demographical characteristics o f online users in terms o f gender, age, educational level, occupation and income. This audience profile analysis showed that online users were different from average voters; thus, online discussants cannot be regarded as representative of the general public. Personality o f online users has been discussed mainly in the studies on the online use of political information. In general, online users were reported to be more politically interested and active than the general public (Johnson & Kaye, 1998; Johnson & Kaye, 2000). They also tend to report high levels o f selfefficacy, the belief that one has the power to manage prospective situations, and they even think they can influence political process (Kaye & Johnson, 2002). Over all, the general online users are more likely to be active participants who are confident o f what they believe than the general public. 41 Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission. Wanta (2000) suggests that the unique characteristics o f computermediated communication could be influential on the spiral o f silence process in online discussions. In his study on a chat room, Wanta (2000) described the special environment that online users were faced with: Chat room discussions are clearly taking place in a non-intimidating setting. The specificity o f the chat room communication allows individuals to be less susceptible to peer pressure and, by extension, to be less likely to accept the dominant opinions than in interpersonal communication situation where peer pressure to confirm is quite strong. Therefore, a more open discussion o f personal opinions and attitudes may be expected when the user is expressing positions in a virtual community (p. 7). Related to this study, Roberts, Wanta, and Dzwo (2002) examined the agenda setting process and the role it may play in the electronic bulletin boards (EBB). Using the frequency of EBB discussions o f each issue as the surrogate for the public agenda, they found that media coverage o f all issues but the abortion issue, which is very obtrusive in nature, provide individuals with information they can use in their EBB specific-issue discussion. This study shows that the online discussion is also influenced and constricted by the mass media’s information providing roles. Agenda Setting at the Second-Level While the agenda-setting metaphor states that the media my not tell us what to think but are successful in telling us what to think about (Cohen, 1962), the metaphor for second-level agenda setting states that the media tell us how to think about some objects (McCombs, Llamas, Lopez-Escobar, & Rey, 1996). 42 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The second level agenda setting is based on the notion that an issue may include several subissues. For example, in a national election campaign, economy issue receives intensive attentions, and the candidates and media discuss several subissues o f economy. The economy issue would include balancing the budget, reducing the national debates, income tax reform, interest rates, and creating jobs. In this case, McCombs et al. (1997) suggest that agenda setting also take place at the subissue levels. In other words, if mass media focus on the tax reform among other sub-issues o f the economy, the public will perceive that the tax reform is the most important economic issue. 43 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 3 First-level (traditional) Agenda Setting Process Public Agenda Media Agenda Object “► Object — First-level agenda-setting effects Source: McCombs & Shaw, 1972. Media Agenda Priorities Public Agenda Priorities 1. Foreign policy 1. Foreign policy 2. Law and order 2. Law and order 3. Civil rights 3. Civil rights 4. Fiscal policy 4. Fiscal policy 5. Public welfare 5. Public welfare 44 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This second level agenda setting is important during the election because candidates might bring different attributes (subissues) o f the economy. For example, one candidate addresses the economy problem as job shortage; while, the other candidate emphasizes the tax cut. If the publics link the economy issue to job shortage, the candidates addressing these attributes will have a chance to be perceived as a better performer o f economy. See Figure 3. Clearly, the second level agenda setting includes framing approach. Entman (1993) defined that “To frame is to select some aspects o f a perceived reality and make them more salient in a communication text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation and/or treatment recommendation for the item described. Gamson and Modigliani (1989) stated that “a frame is a central organizing idea for making sense of relevant events and suggesting what is at issues.” Tankard, Hendrickson, Silberman, Bliss, and Ghanem (1991) defined that media framing as “the central organizing idea for news content that supplies a context and suggests what the issue is through the use o f selection, emphasis, exclusion, and elaboration.” Whether framing influences the second level agenda setting or vice versa is out of this study’s scope. However, the way issues (or problems) are framed by media or candidates slants public opinions significantly (Entamn, 1993). There are several studies supporting the second level o f agenda setting hypothesis. McCombs, Llamas, Escobar-Lopez and Rey (1997) found support for a second level o f agenda setting during the 1996 Spanish general election on two attribute dimensions - substantive and affective descriptions. Substantive 45 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. attributes dealt with information about qualities o f the candidates: experience with foreign affairs, for example. Affective attributes dealt with positive, neutral or negative comments about candidates such as morality, personal qualities, or leadership (See Figure 4 and 5). 46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4 An Agenda o f Cognitive Attributes at the Second Level o f Agenda Setting Media Agenda Attribute Public Agenda ---------------------- ► Attributes (subissues) (subissues) Topics: Impact on opinions, Impact on behavior An Agenda o f Attributes at the Second Level o f Agenda Setting. An issue has several attributes (Weaver et ah, 1981). For example the issue o f economics has: 1. Unemployment 2. Inflation 3. Taxes 4. State o f the economy 5. Government spending and size 47 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5 The mechanism o f Affective Second Level Agenda Setting Public Agenda Media Agenda -► Attributes Attribute — (1) positive coverage positive evaluation (2) negative coverage negative evaluation An example o f second-level agenda setting effects at affective dimension Positive coverage o f Favorable Evaluations of John McCain on moral issues John McCain on moral issues George W. Bush on leadership George W. Bush on leadership Foreign nations those foreign nations 48 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Based on the 2000 Republican presidential primary in New Hampshire, Golan and Wanta (2001) examined the second level agenda setting hypothesis. They found that John McCain was frequently described as “strong on morality,” which linked to the publics’ favorite evaluations o f McCain’s morality as leader over Bush. In the same manner, Bush is frequently covered by media as strong on tax cut, which gave readers salience cues that Bush is strong on tax cut. The rationale of this effect comes from the fact that newspapers coverage discussing a candidate’s stance on tax cut, for example, will give readers salience cues that the candidate is strong on tax cut. In other words, readers will learn about the candidate’s issue emphasis appeared on the media and link the candidate to that issue. The affective second-level agenda setting effect was found in the international news coverage and its influence on public opinions about those countries. Previous studies showed that international news coverage has a direct influence on U.S. public opinion. For example, a study by Sal wen and Matera (1989) found correlations between foreign news coverage and public opinion that suggested that international news coverage does indeed have an agenda setting effect. Wanta and Hu (1993) examined the agenda setting impact of international news and found a strong agenda setting impact of international news stories on American public opinion, especially on conflict related stories and concrete presentations. McNelly and Izcaray (1992) found that news exposure significantly related to positive feelings towards countries and to perceptions of those countries as successful. Semetko, Brzinski, Weaver and Willnat (1992) 49 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. found that attention to foreign affairs news was a better predictor of positive perceptions o f nations than simple exposure to newspapers. Several recent studies supported the affective dimension o f second-level agenda setting effects. Kiousis, Bantimaroudis and Ban (1999) examined the second level o f agenda setting through two experiments that manipulated media portrayals o f candidate personality and qualification traits. They found subjects’ impressions o f candidate personality traits mirrored media portrayals o f those traits. However, media portrayals o f personality traits did not affect a candidate's overall salience. Results also indicate that candidate qualifications influenced affective perceptions of politicians. Shah, Domke, and Wackman (1996) examined the relationships among media frames, individual interpretations of issues and voter decision-making. They found media frames and issue interpretations substantially influence the type o f decision-making strategy that voters use. The Effects o f M edia’s Favorable Coverage o f Candidates This study posits that effects o f media’s content on readers are regarded as framing effects. Several studies supported framing effects on the public during the election. Domke, Fan, Fibison, Shah, Smith, and Watts (1998) examined whether the quantity o f positive and negative news coverage o f the candidates is related to the public’s preference o f either Bill Clinton or Dole. This study found that a candidate’s level o f support is closely related to the media’s net favorable coverage (positive coverage o f a candidate plus negative coverage of opponent). Using a computer content analysis o f news coverage and tracking poll data, 50 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Domke et al., (1998) reported that news coverage alone explain the substantial amount o f variance for the change o f the candidates’ support rates. Shaw (1999) also examines the effects o f news media favorability on candidates in 1996 U.S. presidential election. He found that media coverage of campaigns affected the tracking poll results. The movements o f candidate’s stance in the public opinion polls are significantly correlated to media’s favorable and rmfavorable coverage o f the candidates’ campaigns. In sum, these communication scholars argue that media has direct effects on the public’s attitudes and behavior, which once was negated by Klapper (1960). In other words, media effects are not so minimal when researchers control demographic information o f the public. 51 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CH APTER 3: HYPOTHESES AND RESEARCH QUESTIONS Based on previous functional analysis studies, framing and agenda setting effects, and media’s news coverage during election periods, I present three sets of hypotheses and research questions. Functional Theory o f Political Discourse Hypotheses and Research Questions RQl. What is the proportion o f functions (acclaims, attacks, defenses) in the presidential campaign .discourse? RQ2. What is the proportion o f policy and character utterances in the presidential campaign discourse? RQ3. What is the proportion o f the forms o f policy (past deeds, future plans, and general goals) and character (personal qualities, leadership abilities, and ideals) in the presidential campaign discourse? This study also tests two hypotheses. Previous studies suggest that incumbents use a greater proportion o f acclaims than challengers, who in turn use a larger percentage of attacks than incumbents. HI. The incumbent party candidate will acclaim more than the challenging party candidate in his political discourse, whereas the challenging party candidate will attack more than the incumbent party candidate. 52 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Previous studies also have found that incumbents particularly and consistently differ from challengers on use o f past deeds: H2. Incumbent political discourse acclaims more than it attacks on past deeds, while challenger political discourse attacks more than it acclaims on past deeds. Functional Theory does not make an explicit prediction on the relationship between discourse topic and election outcome. However, a basic presupposition o f Functional Theory is that “a voter chooses to vote for the candidate who appears preferable to him or her, based on whatever criteria are most salient to that voter” (Benoit, 2003). H3. The presidential candidate who discusses character more than his opponent is likely to win the election. The second set contains hypotheses that are theoretically connected with framing effects and source agenda setting. Miller, Andsager, and Riechert (1988) showed that issues that candidates emphasized in press releases were different from the news stories written on the candidates’ issue positions. Journalists frame news stories by what they include or omit, which provides the “big picture” o f the candidates election campaigns for the public. Based on Benoit and Currie’s (2001) work, I expect that newspapers place more emphasis on attack and defense than on acclaim. This finding is consistent with previous findings that media 53 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. heavily rely on “game frame,” or “present public life as a contest among scheming political leaders” (Cappella & Jamieson, 1997, p. 7). H4. The proportion o f attacks and defenses will be higher in media converge than in the candidates’ political discourse; whereas, the proportion of acclaims will be lower in media coverage than in the candidates’ political discourse. Based on news values studies which report that character is intrinsically interesting and functions as “character cop” for candidates, I expect that media more heavily emphasize character than the candidates do in their political discourse. H5. The proportion o f character comments in media coverage will be significantly higher than the actual proportion o f policy and character comments in their political discourse. Finally, the third set o f research questions are posited based on media’s effect on the public. Most agenda setting studies report that media set the public agenda during the election; however, a couple o f recent studies argue that candidates’ ability to setting the public agenda exceeds media’s agenda setting ability (Roberts & McCombs, 1994; Dalton et al., 1998). The reason for this weakened media effect is that media do not provide substantial information. Instead, they focus heavily on “horse race” coverage during the election. Thus, I expect that horse- race, or who is ahead and who is behind, is the most dominant 54 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. theme in the election. As a result, candidates’ campaigns set the agenda for the public more strongly than the media. In addition, this study uniquely focuses on the audience which is often neglected in most agenda-setting research. I expect that Internet Forum participants, individuals who are highly involved in politics, are more susceptible to agenda setting effects than the general voters. The rationale is that agenda setting is basically social learning; thus, Internet Forum participants follow the news coverage o f election and are very knowledgeable about the hotly debated issues during the election. RQ3. Does media’s coverage o f horse-race aspects of campaigns dominate coverage o f candidates’ policy and character? RQ4: (agenda setting effects) Between the candidates’ campaign and media, who sets the public agenda during the Korean presidential election? H6: Highly involved voters are more susceptible to media’s agenda setting effects than the general voters. Previous studies of media’s framing effects show that media emphasis on the issues may also affect voters’ evaluation o f the candidates. In a similar vein, the second level agenda-setting hypothesis suggests that positive media coverage will lead to favorable evaluation o f a candidate among voters. Thus, I expect that positive coverage o f candidates by newspapers will lead to positive evaluations of the candidates. From this perspective, the fifth research question is drawn as shown below. 55 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. RQ5 (second-level agenda setting): Does the public discuss the candidates positively (negatively) as media report the candidates positively (negatively)? RQ6 (framing effects): Do the media’s favorable coverage o f candidates influence the candidates’ stance in the tracking polls? 56 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CH APTER 4; M ETHODOLOGY This study focuses on two main aspects: the nature and effects of campaign discourse and media coverage. For examining the nature o f campaign discourse and the media’s use of candidates’ political discourse, the Functional Theory o f Political Discourse is adopted. For studying the framing and agenda setting effects, this study combines the content analysis o f media, Internet Forum discussions, and poll data. The main research goals o f this were to: (1) Analyze the functions and topics o f candidates’ political discourse and compare it with media coverage. (2) Examine the agenda setting effects on the general publics and highly involved publics. (3) Explore the media contents’ effects on the publics. Does the media’s favorable coverage o f the campaign lead to favorable discussions (second-level agenda setting)? Does media’s favorable coverage influence the candidates standing in the tracking polls (framing effects)? To address the first research goal, I examine television spots (including endorsement speeches which are very similar to TV ads), news releases, and television debates. Each candidate aired six television advertisements and eight 57 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. endorsement speeches. The candidates were participated all the three presidential debates which were aired on December 3*^^, December 10*’’, and December 16*’’. The whole population was included for the functional analysis o f Korean presidential discourse. After that, I assess how journalists used the candidates’ discourse in the news stories. In other words, I compare the actual functions and topics of political discourse addressed by candidates to media’s use o f candidates’ discourse in the news stories. This comparison will show what types o f frames (e.g., acclaims vs. attacks, policy vs. character) o f candidates are mostly salient in the news stories. To address the second goal, I will compare the campaign agenda, media agenda, and voter agendas. By using rank-order correlations, the results will show who sets the voter agendas during the 2002 Korean presidential election. To address the third goal, I will use poll results and online discussions. Using correlations between media’s coverage and the publics’ response to candidates, this study will find the framing effect and second-level agenda setting effects. Coding Procedures for Presidential Candidates ’ Discourse The coding procedure had three steps. First, the candidates’ political discourse such as television advertisements, televised debates, and news releases will be utilized into themes. The coding unit for the candidate discourse is theme, or utterances that address a coherent idea. Berelson (1952) defined a theme as “an assertion about a subject” (p. 18). The length o f theme may vary from a phrase to 58 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. several sentences because the candidates’ discourse is often colloquial and enthymematic. The Functional Theory o f Political Discourse posits that each theme has only three functions. Acclaims portray the candidate or the candidate’s party favorably; attacks portray the opposing candidate or opposing candidate’s party unfavorably; defenses explicitly respond to a prior attack on the candidate or the candidate’s party (Benoit, Pier & Blaney, 1997). Most o f the utterances in political advertisements belong to one o f the three functions listed above. Some utterances that did not fit into any o f these funetions were omitted. For instance, a few themes that discuss the past events without praising the candidate or disparaging the opponent were excluded. Second, the topics o f acclaims and attacks were divided into either policy or character. If a theme discussing governmental action and problems pertaining to such action, it is coded as policy. If a theme addresses properties, abilities, or attributes of a candidate (or party), it is coded as character. Third, policy themes will be classified as past deeds, future plans, and ideals; while, character themes will be classified as personal qualities, leadership abilities, and ideals (The examples will be provided in the Appendix). Finally, each theme will be coded based on following issues: eeonomy, political reform, social stability, North Korea and U.S. relationship, social welfare, administration reform, and education. Coding Procedure fo r Newspaper Coverage o f Presidential Candidates 59 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. News coverage analysis has two steps. First, I selected three major newspapers, Chosun Daily, Jungang Daily, and Hankyereoh Daily. News items were collected by using KINDS (Korean Integrated News Database System) database with the keywords, “Lee Hoi Chaing or Roh, Moo-hyun or presidential election.” Based on stratified random sampling methods, 600 news stores were selected. Then, each news story containing quotation from the candidates and public relations practitioners will be chosen and analyzed by the functional analysis o f political discourse method, fri other words, the actual candidates’ presidential discourse will be compared to media’s use o f the candidates’ discourse in terms o f functions and topics. After this procedure, every news story will be again utilized as themes, the method that was used to analyze the presidential discourse. However, the coding procedure for news items had several changes. First, the newspaper coverage o f the candidates was coded into three categories: policy, character, and horserace. Because previous studies o f election coverage report that horse race is the dominant theme in the media, this study includes the horse race category. Horse race coverage includes public support, expectations, organizations and finances, mood-of-the country, voter profiles, campaign strategies and political endorsements (Johnson, 1993; King 1984). In other words, if the media’s campaign coverage focuses on the aspects such as winning and losing, strategy and logistics, and opinion polls, this coverage will be coded in the horse race category. This study identified five elements o f horse-race coverage as shown below (Johnson, 1993; King, 1990). 60 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (1) Public support: Public support is concerned with who is ahead and who is behind according to published public opinion polls. (2) Organizational and financial strength: This category indicates the news stories that assess how successful the candidate is in raising money and having a solid staff to get out the vote. (3) Endorsements: Endorsements by key party officials and opinion leaders who help bring in more volunteers and donations as well as heighten the perception that the candidate has broad public support. (4) Campaign performance: This category includes how well the candidate fared in events such as straw polls and debates in addition to performance in the key-note speech, advertisements, and media interview. (5) Voter profile: This category reports the types o f voters (e.g., 20s, bluecollar workers) who support or against the candidates. Voter profile is different from public support in that voter profile addresses the sub­ population of voters. After detecting the news coverage o f horse race aspects o f election, this study will examine how newspapers cover candidates’ issue priorities. In order to compare newspapers’ opinion polls, the issue category used for newspaper coveragewere adopted. To deteet the media’s agenda during the election, each theme o f the news story were coded for the following: (a) candidate policy, with 61 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the categories: economy, political reform, social stability. North Korea, social welfare, agriculture, education, and regional conflicts (b) the nature o f affective attribute frame, coded as positive, neutral, or negative. This study originally planned to test character (image) agenda setting effects on voters; however, no opinion poll was conducted on this topic. Thus, a candidate’s character agenda could not be examined. In order to collect public opinion poll data, the researcher used Korean Press Association websites in which newspapers’ tracking poll were retrieved. Poll data about the election were collected to obtain the supporting rates for the two candidates. Every data provided the exact supporting rates o f the candidates. The following are the dates when the poll were conducted by media: Oct. 27, Oct. 30, Nov. 3, Nov. 6, Nov. 10, Nov. 14, Nov. 17, Nov. 22, Nov. 25, and Nov. 28. The polling institutions include Joongang Dialy, Kookmin Daily & Yoido Research, Hankook Daily & Media Research, Hankyoreh Daily, Donga Daily & Korea Research, Chosun Daily & Gallup Korea. Poll Data and Internet Forum Discussions The researcher expects that the candidates standing in the tracking polls are caused by media’s net favorability o f candidate A (positive coverage of candidate A plus negative coverage o f candidate B). To examine the link between media’s content and the tracking poll results, ten opinion polls that appeared in the major Korean newspapers is used. To test the agenda setting effects, Gallup poll and Korean Research poll data are compared to the agendas o f Korean presidential candidates and 62 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. prestigious newspapers such as Chosun daily, Jungang daily, and Hankyeoreh daily. The Gallup poll conducted on December 7* asked, “What are the most important issues for the 2002 presidential election?” The poll results will be compared with the content analysis o f candidates’ political discourse and newspaper coverage at one time. This study posits that Internet Forum discussion participants are highly involved in the presidential elections. Thus they are less influenced by agenda setting effects. For this purpose, the frequency o f Internet discussions o f each issue and image will serve as the surrogate for agendas o f high-involved voters, and the content analysis will be compared to the content o f newspapers and the candidates’ messages. In order to analyze the Internet Forum participants’ discourse, online forum messages will be sampled from three portal sites, which are, Naver, Chosun Ilbo, and Yahoo. Online messages that contained the name o f the candidates in their title: “Rho Moo-hyun” and “Lee Hoi Chang,” will be sampled and analyzed. For this study, one whole online posting message is used as the unit o f analysis. Initially, the researcher tried to utilize online messages as themes, but had to utilize the whole message as the coding unit because o f low coder reliability. In addition, most o f the messages were concentrated on one theme; thus, the researcher believed that adopting a whole message as the coding unit serves this study better. Rarely online users discussed more than one theme with diverse perspectives. 63 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Finally, this study will examine the links between newspaper coverage o f Korean presidential election and voter perceptions o f candidates in terms of second-level agenda setting. According to second level agenda setting hypothesis at the affective dimension, positive coverage of the candidates will lead to positive evaluations o f the candidates; while, negative coverage o f the candidates will lead to negative evaluations o f the candidates. In order to test this hypothesis, content analysis o f news coverage will be compared to Internet Forum messages during the election. These tests will examine whether positive coverage o f the candidates on policy and character influenced whether the voters were more likely to link the attributes to a candidate. Unlike typical agenda setting studies which examine two-time periods to detect the effects, this study will use chi-squares to examine the link between mass media coverage and poll results. The reason is that the poll was conducted only once. This exploratory study will show trends based on the logic of the second-level agenda setting, while chance cannot be ruled out as a cause o f the results. Limitation o f Cross-Sectional Approach in Agenda Setting Effects To examine the agenda setting effects, cross-lagged correlations (Spearman’s Roh) will be used. The primary advantage o f cross-lagged correlations is that they provide direct evidence on the agenda-setting relationship across time; a major requisite for asserting that one agenda causes or influences another. The logic behind cross-lagged correlations is that if one variable really is the cause o f another, the correlation o f X (the cause) at time one with Y (the 64 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. effect) at time two should be stronger than the opposite cross-lag, Y the effect at time one with the cause X at time two (McCombs & Shaw, 1972). Another advantage of cross-lagged analysis is the ability to test the alternative hypothesis of agenda setting, which is the public setting the media agenda. However, this study adopts a one-time cross-sectional analysis instead o f two time frames to detect agenda setting effects. The reason is that Korean presidential campaign periods last only six weeks Considering that agenda setting needs at least one-month time lag to transfer the media agenda to voters (Winter & Eyal, 1981), the researcher has to assume that either candidates or media set the voter agenda. In fact, a majority of agenda setting studies successfully show that media set the agenda, not the other way around. In addition, it is reasonable to assume that the publics will be influenced by media and campaigns because the presidential candidates were nominated by each party just fifty days before the election day. In Korea, voters do not have much time to set the agenda for the media and campaigns. Intercoder reliability Two coders content-analyzed all three Korean presidential spots and debates. Both coders analyzed a subset o f 10 percent o f the sample to check the intercoder reliability using Cohen’s kappa. With regard to political spots, Cohen’s kappa (1960) was .77 for coding theses as acclaims, attacks, and defenses; .72 for classifying themes as policy or character and .86 for classifying policy themes as past deeds, future plans, or general goals and .69 for coding character themes as personal qualities, leadership ability, or ideas. 65 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Intecorder reliability for debates was .87 for coding themes as acclaims, attacks, or defenses, .82 for classifying themes as policy or character, and .72 for classifying policy themes as past deeds, future plans, or general goals; and .86 for coding character themes as personal qualities, leadership ability, or ideals. Two coders also checked the reliability for news releases, news coverage, and online discussions. Approximately 10% o f each item was tested; .88 for classifying news releases as policy or character; .81 for coding character themes acclaim, attack, and defense. With regard to news coverage, .71 for categorizing agendas; .68 for coding tones; 79 for detecting horse race. The reliability coefficient o f Internet Forum discussion was .69 for analyzing agendas, and .82 for detecting tones. According to Landis and Koch (1977) values o f kappa from .61-. 80 reflect substantial agreement and values from .81-1.0 indicate “almost perfect” agreement (p. 165). Thus, the kappa scores yielded in this study showed from substantial to almost perfect reliability. 66 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CH A PTER 5; RESULTS In this chapter, I report separately on the findings o f Functional Theory o f Political Discourse, media coverage o f presidential campaign, framing effects and agenda setting effects. The content analysis findings are structured by the four sets of research hypotheses. The interpretations o f findings are presented in the discussion chapter. The first research question, concerning the proportion o f acclaims, attacks, and defenses in the presidential discourse, will be answered for each medium: television spots (including endorsement speeches), television debates, and press releases. Functions Television spots. Candidates used more acclaims (79%) than attacks (20%) in the presidential spots. The one-way chi-square showed that there was significant difference o f using functions (x^ =7.0, df=l, p< .01). Acclaims outnumbered attacks. For example, this spot illustrates an acclaim: “Rob: When 1 am president, 1 will guarantee that women can keep their jobs without worrying about their babies by providing enough day nurseries.” 67 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2 Functions o f Political Discourse by Medium Medium Acclaims Attacks Defenses TV spots 191 (79%) 48 (20%) 1 (.4%) TV debates 121 (54%) 74 (33%) 30(13%) Press Releases 186 (58%) 102 (32%) 35 (11%) Total 498 224 66 In Lee’s spots, Roh was criticized for his scandalous connections with Kim’s government. Lee asserted that there is an urgent need for power transfer because o f a series o f corruption scandals that have occurred under President Kim Dae-jung and Roh’s party. Defenses were used once in the television spots. Defenses reminded voters of his weakness and made him discuss weakness; thus, this function was not frequently used. However, Roh in his ads used defense strategy once. For example, Roh responded Lee’s attack on his political reform plans: “Lee is calling on voters to judge President Kim’s scandal-plagued governments, but the public wants to eliminate old-fashioned politician Lee who are involved with draft scandal and illegal fund-raising allegation.” In Roh’s response, he defused the charge by attacking his accuser. There are many ways to defend oneself against attacks (Benoit, 1995); however, this study does not discuss the defense strategies. 68 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Television debates. Like presidential spots, acclaims were the most common function (54%), followed by attacks (33%), and defenses (13%) were least frequent. The difference o f using these functions was also statistically significant (X ^ = 1 2 .3 ,d ^ 2 , E< .01). For example, Roh praised himself that he would continue the President Kim’s “Sunshine Policy” which contributed to bringing peace on the Korean peninsula. Lee also boasted his future plan: “I will make a million of new jobs for women.” Both o f these utterances are clear examples o f acclaims, because they assume that peace and new jobs are desirable to voters. See Table 2 for these data. Attacks were also quite common in the debates, compromising 33% o f the utterances. The MDP candidate attacked Lee’s policy by labeling Lee as a hard­ liner who is willing to risk another war on the Korean peninsula. Defenses accounted for 13% o f the themes in these debates. When Lee was attacked as a supporter o f chaebol, the family-owned big business, he refuted this attack. He mentioned that he actually proposed the reform o f chaebol. Although less common than attacks or acclaims, defenses were used repeatedly. Press Releases. The Korean presidential candidates also addressed acclaims (58%) more than attacks (32%) and defenses (11%). A one-way chisquare test showed that the candidates used the function differently. Acclaims were most frequently used as compared to attacks and defenses. For example, Roh pledged to eradicate social discrimination against the uneducated in his press releases. This is an example o f acclaiming Roh’s ideal and implicitly praised his success as politician even though he did not graduate from college. Roh in 69 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. another press releases attacked Lee by stating that hi misled the public. Roh stated that even Lee knew that Seoul would continue to serve as an economic capital regardless o f relocation o f capital city to Chung-Chung area. Lee also acclaimed his medical reform plan. He promised that he could unburden the individuals’ medical bills by subsidizing the old and poor people. Lee also praised his leadership by stating that his leadership was based on law and principle. Compared to presidential spots, defenses are frequently addressed. Lee answered Roh’s initial attack on his North Korea policy by stating, “it is a simply lie that two Koreas would face a severe military tension if I were elected.” Comparison across the Media. The each candidate’s presidential campaign messages were more frequent than attack, followed by defenses. This ordering makes sense that candidates should show their strengths to the voters. Attacking the competing candidate is also beneficial if the voters accept the attacks; however, it is dangerous that many Korean voters professed they did not like mudslinging (Lee, 2002). The candidates’ use o f three media is slightly different. Television advertisements were used most positively; while, defenses were more frequently used in television debates. Defenses made the candidate off-message, and look reactive rather than proactive. Thus, it is best for the candidate to avoid using defense utterances. However, the candidates who were usually well advised by political consultant or public relations practitioners surrendered to use defenses when they were attacked in front o f their faces. 70 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. HI posited that Roh, the incumbent party candidate, would acclaim more than a challenging party candidate, Lee, whereas Lee would attack more than Roh. This hypothesis will be tested by each medium. Presidential Spot. The first hypothesis was supported. There was a significant difference in acclaims versus attacks for an incumbent versus the challenger (x^ =6.23, df=l, p< .01). Table 3 Candidates ’ use o f functions by medium Medium Acclaims Attacks Defenses TV spots (Roh) 105 (89%) 11 (10%) 1 (.1%) TV spots (Lee) 86 (70%) 37 (30%) TV debates (Roh) 53 (56%) 27 (28%) 15(16%) TV debates (Lee) 68 (52%) 47 (36%) 15(11%) Press Release (Roh) 100 (64%) 42 (27%) 15 (9%) Press Releases (Lee) 86 (52%) 60 (36%) 20(12%) 71 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4 Forms o f Policy Topics and Incumbency Policy Past Deeds* Future Plans General Goals TV spots (Roh) 3 24 3 22 2 TV spot (Lee) 2 22 10 1 35 3 TV debates (Roh) 9 6 13 4 9 1 TV debates (Lee) 5 26 24 1 10 7 PR (Roh) 13 8 23 14 27 6 PR (Lee) 6 18 32 10 17 19 Topics This section answers the second research question about the proportions of policy and character discourse by comparing these topics across the four media represented in this study. Then, the three forms o f policy will be broken out to answer the third research question. Finally, the third hypothesis will be tested. Television spots. The candidates devoted 55% o f their discourse to comment policy topics, and 45% to comment character. The candidates devoted 21% of their policy utterances to past deeds, 49% to general goals, and 30% to future plans. For instance, this ad attacked on the basis o f past deeds: “Roh promised when his party took the regime, there would be no more government 73 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. As shown above Table 4, Roh used acclaims (89%) significantly more than did challengers (70%), while challengers tended to attack (30%) more than incumbent party candidate (11%) in the presidential spots. This hypothesis was supported (x^ =15.78, df=l, p< .01). However, this hypothesis was not supported in the presidential debates =1.87, df=2, 2>.05). Both Roh and Lee used approximately same amount of acclaims, attacks, and defenses. This finding might come fi-om the fact that the candidates’ discourse was regulated and influenced by modulator. HI was supported in the press releases (x^ =5.26, df=2, g< .10). Roh in the press releases used more self-praise (64%) than Lee did (52%), whereas Lee focused on attacks (36%) as compared to Roh did (27%). The second hypothesis predicted that the incumbent party candidate was significantly more likely to acclaim on past deeds, while challengers were more likely to attack than to acclaim on past deeds. This hypothesis was supported for all the campaign discourse: presidential spots (x^ =15.78, dfi=l, p< .01). , television debates (x^=15.78, df=l, p< .01), and press releases (x^ =15.78, df=l, P< .01). Result revealed that it was easy for a challenger to attack on the incumbent’s past deeds. Table 5 shows that Roh acclaimed significantly more likely to acclaim than to attack on past deeds, Lee were more likely to attack than acclaim on past deeds. 72 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. corruption. But his party was involved in forty-seven political scandals.” Another spot articulated general goals: “When I am president, I will guarantee that women can keep their jobs without worrying about their babies by providing enough day nurseries.” These excerpts illustrate the common policy utterances. See Table 4. Most character comments concerned ideals (51%). Others addressed personal qualities (33%) and leadership (16%). For example, this advertisement praised candidate Roh’s personal qualities: “I am a self-made man from a poor family.” Roh also praises his ideal: “I am the only candidate who will lead two Koreas to be unified. Lee attacked Roh’s ideal by stating that Korean people could not live by Roh’s radical ideology. Thus, these examples illustrate television spots’ use o f character. Television Debates. The candidates in the television debates spent 63% of the comments on policy and 37% on character. Note further that both candidates emphasized policy in the debates (Lee, 64%; Roh, 60%). For example, Lee explained his science policy: “I will use more than 3% o f GDP for research and development projects:” Next, I analyzed policy and character utterances into three subforms. The most common form o f policy was past deeds (46%), followed by future plans (36%), and general goals (18%). For example, Roh praised his party’s past deeds: “MDP’s excellent treatments o f the economic crisis saved the nation.” With regard to future plans, Roh promised that he would strictly apply tax regulations to prohibit conglomerate owners from transferring their wealth to their children without paying proper inheritance taxes. Lee discussed his general goals: “I will 74 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. expand the welfare system to benefit the disabled, the elderly and children.” In sum, the candidates developed policy claims in a variety o f ways. In character, discussion o f personal qualities (36%) and ideals (42%) were most common, followed by leadership ability (22%). For example, Roh attacked Lee’s leadership that he was the chairperson o f the ruling party in 1997 when the country was suffering from an economic crisis. The fact that Lee was in charge during such times casts doubt on Lee’s leadership ability. Lee attacked Roh’s ideals by mentioning, “We cannot live by Roh’s racial ideology.” When Lee mentioned his personal qualities, he praised himself by stating that he was the man o f principle. Thus, the candidates employed all three forms o f character utterances. Press Releases. Utterances ,in the press releases spent 58% of their messages on policy and 33% on character. Policy utterances were almost equally divided on past deeds (26%), future plans (38%), and general goals (34%). When candidates address character, they tend to focus on personal qualities (42%) and ideals (41%). Leadership abilities (17%) were not as frequently used as these two sub-character forms. For example, Roh praised his leadership by mentioning that his negotiation skills as administer saved the bankruptcy o f Samsung Auto Company. Lee also praised his personal qualities by emphasizing that he was never involved in any political corruption scandals. Roh attacked Lee: “the old politician did not meet the need o f new politics.” Like television spots and television debates, press releases used all three forms o f character utterances. 75 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The third research question concerns the allocation o f the three forms of policy and character utterances. Future plans and general goals are emphasized; while, past deeds were less frequently addressed. With regard to character, there is somewhat more disparity in the percentage o f remarks used in support o f each form o f character. Leadership abilities are the least common form of character in all three media (19%). Ideals are the most common form (43%), followed by personal qualities (38%). Table 5 Forms o f Topic by Functions (Number o f themes) Character Policy Past Future General Personal Leadership Deeds* Plans Goals Qualities Ability Ideals TV Spots 2 25 34 4 57 5 30 7 17 1 52 6 TV 14 32 37 5 18 3 17 19 15 6 19 9 19 26 55 24 44 25 22 18 13 4 33 5 33 70 109 35 115 32 62 41 40 10 98 19 Debates Press Releases Total 26% 36% 37% 38% 19% * Acclaims/ attacks 76 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 43% The third hypothesis predicted that presidential candidate who discusses character more than his opponent is likely to win the election. Benoit (2001) originally found that the candidates addressing more policy than the competitors would win the election because the U.S. voters answered that they would choose the president based on policy. Because Korean voters showed that they preferred character to policy comments, the third hypothesis in this study predicted that the candidate who discussed more character than his competing candidate would win the election. This hypothesis is tested by each medium. Results from television spots supported this third hypothesis (X^[df-1]=4.5, p<.05). In the presidential spots, the winner, Roh, emphasized character (52%) more than Lee who discussed his character (39%), whereas Lee focused more on his policies (60%) than Roh did (48%). This hypothesis was also supported by the candidates’ use o f press releases. Compared to Lee’s utterances in press releases, Roh emphasized his character comments (x^=2.73, df=l,/?< .10). However, this hypothesis was not supported by the results o f television debates. Both candidates in the debates addressed more on policy than character (X^= .04,df=l,p>.05). 77 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 6 Topic o f Utterances in Presidential Discourse Medium Policy Character Roh (TV spots) 57 (48%) 60 (52%) Lee (TV spots) 74 (60%) 49 (39%) Roh (TV debates) 58 (62%) 36 (48%) Lee (TV debates) 82 (63%) 48 (27%) Roh (PR) 98 (62%) 59 (38%) Lee (PR) 118(71%) 48 (29%) M edia’s Use o f Candidates ’ Discourse The first hypothesis predicted that the proportion o f attacks and defenses will be higher in media converge than in the candidates’ political discourse; whereas, the proportion o f acclaims will be lower in media coverage than in the candidates’ political discourse. The data confirmed this prediction [df=2\- l6A ,p<.Q \). When journalists select the comments from the candidates and PR practitioners, they tend to focus on attacks and defenses as compared to acclaims. 78 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 7 Functions o f Campaign and M ed ia ’s Use o f Campaign Discourse Acclaims Attacks Defenses Campaigns* 498 (63%) 224 (28%) 66 (8%) Newspapers 103 (49%) 75 (36%) 32(15%) * The sum o f television spots, television debates, and press releases In the candidates’ political discourse, attacks constituted 28% of the utterances but were 36% o f the descriptions in the news stories. Similarly, defenses were more common in the news stories (15%) than in the campaign discourse (8%). Conversely, acclaims, which constituted 63% o f the campaign discourse, were underrepresented in the news stories at 49%. This result shows that journalists’ use o f candidate discourse tends to over-report attacks and defenses while under-reporting acclaims. See table 7. Hypothesis 6 predicted that newspapers’ direct use o f candidate discourse would emphasize character more (and policy less) than the actual campaign discourse. However, this hypothesis was not supported ((X^[df=l]= .57, p>.05). There was no significant difference between topics discussed in the campaign discourse and in the newspapers’ direct use o f campaign discourse. See table 6. 79 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 8 Topics o f Campaign and M ed ia ’s Use o f Campaign Discourse Medium Policy Character Campaigns 487 (62%) 300 (38%) Newspapers 137 (65%) 73 (35%) M edia’s Coverage o f Horserace The third research question questioned whether media’s coverage o f horse race outnumbered policy and character. Previous U.S. studies report that horse-race coverage o f election outnumbers issue coverage (e.g., King, 1990, Patterson, 1991). Results show that even though Korean newspapers do not focus on horse-race coverage as much as U.S. media do, a substantial amount o f coverage was allotted to horse-race coverage. A total o f 22% media coverage (132 out o f 600 articles) discussed horse­ race aspects o f the election. 80 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 9 Differences in Horse-race Coverage During the Election Endorsement Voter Public Org.and Performance support Finances 48 24 17 23 20 36 18 13 18 15 Profile No. of Article % When newspapers report horse-race aspects o f campaigns, they mostly emphasize public support, followed by organizational and financial strength, and campaign performances. Voter profile and endorsement coverage were also discussed. With regard to research question 3, results showed that horse-race coverage o f election was not the dominant theme during the 2002 Korean presidential election. Election and Agenda Setting Effects The fourth research question discussed who set the public’s agenda between campaigns and media during the Korean presidential election. Table 10 presents the topics that appeared in candidates’ presidential discourse across media: television spots, television debates, and press releases. Roh campaigns mostly focused on political reforms and economy, whereas Lee campaigns emphasized economy and corruption. As shown below Table 10, 81 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Roh’s campaign issues were quite similar across the media, whereas Lee’s campaign emphasized issues differently from one another. In order to compare campaign’s influence on voters with media’s ability to set the public agenda, this study used the average issue priorities among television spots, television debates, and press releases. 82 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 10 Issue Priorities o f Candidates across Media Roh Ads Roh debates Roh PR Lee Ads Lee PR Total 11.7(5) Lee debates 17.6 (2) Political reform Economy 26.2 (1) 22.7(1) 26.6 (1) 14.8 (4) 19.9(1) 14.4 (3) 19.8 (2) 19.6 (2) 15.2 (3) 24.1(1) 21.8(1) 19.2 (2) Corruption 10.9 (4) 7.8 (7) 8.6 (5) 5.9 (6) 6.8 (7) 4.8 (6) 7.3 (6) Social Welfare North Korea Social Stability Education 18.2 (2) 14.8 (3) 13.3 (4) 15.6 (2) 8.3(5) 15.6 (3) 14.3 (4) 7.8 (6) 7.9 (6) 14.2 (3) 26.4 (1) 15.7(3) 19.5(2) 15.3 (3) 5.5 (7) 6.0 (5) 4.7 (7) 2.0 (9) 3.7 (9) 3.1(9) 4.2 (8) 3.9 (8) 5.8 (8) 2.8 (8) 5.4 (7) 6.5 (6) 4.6 (7) 4.7 (7) Agriculture 1.4 (9) 4.1 (9) 1.9 (9) 3.8 (8) 4.5 (8) 3.9 (8) 3.3 (9) Regional Conflicts 11.7(5) 10.0 (4) 8.5 (6) 14.2 (4) 13.0 (4) 11.6(5) 11.5(5) 83 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The key point for answering RQ4 is to assess the public’s interests. Candidates attempt to define the election in terms o f their preferred issues; however, their efforts are restricted by media coverage and public’s interests. The voters might want the candidates to discuss their concerns; however, the election revolves around the choices the candidates offered. The distribution o f the public’s issue interests and the issues presented in campaigns and media is displayed in Table 10. Results show that the correlations between media issues and public issues (r= .89, p<. 01) are stronger than those of campaign issues and public issues (r= .71, p<. 05). Newspaper coverage of election was closely linked to public interests and evidence that campaign interjects its own agendas in public’s agendas is limited. After controlling the campaign’s influence on public agenda, the correlations between the news media and the publics are statistically significant (r= .78, p<. 05). Thus, this study supports the traditional agenda setting effects, or the media have the ability to transfer their agendas to the public. The sixth hypothesis was set to study whether the audience types mediated the agenda-setting effects. The research assumes that Internet Forum discussants are more politically knowledgeable than general voters; thus, the discussants are more susceptible to the media’s agenda. Based on this notion, H6 predicted that the issue correlations between the general public and media are stronger than those o f between Internet Forum discussants and media. This hypothesis is not accepted. In fact, correlations 84 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. between the general public’s agenda and media agenda (r= .62, p<. 05) are stronger than those of between Internet discussants’ agenda and media (r= .48, g> .05). Instead, the issue interests o f Internet discussants are more closely correlated to campaigns than the media (r= .90, p<. 01). See Table 10 for each constituent’s issue priorities. 85 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 11 Policy Priorities among Candidates, Media, and Voters Campaigns Media Public (Poll data) Political Reform Economy 19.9(1) 17.2 (3) 12.5 (3) Internet Discussants 23.1 (1) 19.2 (2) 24.1 (1) 38.1(1) 13.2 (4) Corruption 7.3 (6) 15.1(2) 13.3 (2) 11.2(5) Social Welfare 14.3 (4) 11.2 (5) 8.1(4) 14.5 (3) North Korea 15.3 (3) 14.2 (4) 7.9 (5) 18.2 (2) Social Stability 4.2 (8) 5.7 (7) 6.1 (6) 4.3 (8) Education 4.7(7) 7.4 (6) 3(7 ) 9.0 (6) Agriculture 3.3 (9) .9 (9) 2.9 (8) 2.1 (9) Regional Conflicts 11.5(5) 4.1 (8) 2.6 (9) 4.4(7) Note: The number in the parenthesis is the rank. 86 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. RQ 5 (second-level agenda setting at affective dimension): Does the public discuss the candidates positively (negatively) as media cover the candidates positively (negatively)? RQ 6 (framing effects): Does the media’s net-favorable coverage of candidates influence the candidates’ stance in the tracking polls? The fifth and sixth research questions concern the effects of media content on audience. The media’s direct influence on voters were negated since the minimal effect research was established by Lazarsfeld, Berelson, and Gaudet (1944); however, a substantial number o f studies recently report that media have stronger and direct effects on publics than minimal effects perspectives. The fifth research question addressed the second-level agenda setting effect on voters at affective level. In other words, this study examined whether the way an issue is covered in media (positive coverage vs. negative coverage) affects the salience of that object on the public agenda. Results show that the positive coverage o f candidates lead to positive discussion for those candidates and negative coverage o f candidates lead to negative evaluations for those candidates. Using the online postings o f Internet Forum discussants during the election, this study examines whether the correlations between positive coverage o f a candidate and positive evaluations on the Internet Forum existed. Results show that correlations between positive coverage o f Lee and positive discussions about Lee are solely statistically significant (r= .76, g<. 05). 87 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. However, negative coverage o f Lee is not related to negative evaluations o f Lee (r =. 12, e >.05). For candidate Roh, the tone o f coverage did not affect Internet Forum participants in a statistically significant way. Newspapers’ positive coverage of Roh did not lead to positive discussions about Roh (r=. 46, p>. 05) and negative coverage o f Roh was not related to negative evaluations o f Roh (i= .26, p> .05). In sum, the results are mixed. Overall, less support was found in the analyses for affective attributes. The findings suggest second-level agenda setting was at work between positive coverage o f Lee and positive discussions about him. This means that positive media coverage did not always lead to positive public perceptions and negative coverage did not necessarily induce negative public perceptions. 88 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 12 Media Coverage o f Candidates and the Public’s Evaluations o f the Candidates Internet Forum Messages Media coverage Time Date Roh Lee Lee Roh Pos. Neg. Pos. Neg. Pos. Neg. Pos. Neg. 1 Oct.24-26 22 10 14 13 16 10 11 22 2 Oct.27-29 12 8 9 11 10 18 16 15 3 Oct.SO-Nov.l 11 10 19 16 8 22 20 13 4 Nov.3-5 17 10 13 14 18 24 25 13 5 Nov.7-9 14 7 13 16 15 16 29 34 6 Nov.11-13 12 8 11 15 13 14 26 13 7 Nov. 14-16 19 10 21 9 20 23 15 16 8 Nov. 19-21 12 13 22 14 8 16 14 20 9 Nov.22-24 21 16 25 8 20 13 32 17 10 Nov.25-27 19 17 18 12 13 18 31 9 89 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The sixth research question was posited based on the notions that news media coverage o f candidates provide stimulus for changes o f candidate support rates. The purpose o f this research question was to assess whether candidate support rates were explained by media coverage o f candidates in terms of favorability. Using support rates as dependent variables, a liner regression was used to determine if the tone change o f media coverage explained the public’s candidate support rates. Liner regression indicated that media favorability for Lee accounted for 24.0 % of the variances in Lee’s support rates. This study also found that media favorability for Rob explained 37.5% o f the variance in Rob’s support rate. Results showed that media’s favorable coverage for candidates alone explained a substantial amount o f candidates’ support rates. Thus, this study reports that media content affects the public’s response in terms o f tracking polls. See Table 13. 90 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 13 The Supporting Rates o f Candidates in Tracking Polls and Net Favorability o f Media Time *Media Net Favorabilit* Date Lee Supporting Rates (%)** Date of poll Roh Lee Roh conducted 1 Oct.24-26 35 24 Oct. 27 37.2 18.4 2 Oct.27-29 23 17 Oct. 30 34.8 20.2 3 Oct.31-Nov.2 27 29 Nov. 3 33.8 20.8 4 Nov.3-5 31 23 Nov. 6 35.9 21.8 5 Nov.7-9 30 20 Nov. 10 36.0 16.8 6 Nov.11-13 27 19 Nov. 14 36.8 21.1 7 Nov. 14-16 28 31 Nov. 17 35.6 17.7 8 Nov. 19-21 26 35 Nov. 22 36.1 22.5 9 Nov.22-24 29 39 Nov. 25 35.2 42.7 10 Nov.25-27 31 34 Nov.28 38.8 44.4 * Media Net Favorability= A candidate’s positive news stories plus the competitor’s negative news stories. ** Korean opinion polls typically ask, “Which candidate do you support?” instead o f the typical U.S. poll question, “Who would you vote for if the election were held today?” Table 14 Testing Media Favorability and Support Rates by Single Regression Analysis Favorability beta SE Lee .557 1.21 Roh .667 .353 Explained R. E .310 .224 <.05 .445 .375 <.01 91 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CH APTER 6: DISCUSSION This study’s two main goals were (1) to explore the nature o f candidates’ political discourse and journalists’ use o f candidates’ discourse and (2) to examine media content’s effects on both cognitive and behavioral levels. Since media have functioned major tools for presidential campaigns, the importance of understanding presidential campaign discourse has grown. However, relatively few studies provided evidences o f functions and effects o f political discourse. There are several assumptions made by Korean campaign analysts that candidates too much rely on negative campaign and emphasize too much on character rather than providing substantial information about their policy. This is the first study examining the function o f political discourse from the 2002 presidential election. Because o f few previous functional studies on Korean election, the results were discussed by comparing findings o f U.S. functional discourse studies. Implications o f the Study Functional Theory o f Presidential Discourse Presidential spots. In the presidential spots, acclaims (79%) were more common than attacks (20%). Defenses were used rarely (.4%) but did occur occasionally. This ordering o f three functions is consistent with American presidential spots (Benoit, 2001) and was predicted by Functional Theory of 92 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Political Discourse. The order o f acclaims and attacks is consistent with Tak, Kaid, and Lee’s (1997) study o f the first group o f Korean presidential advertisements (67% positive, 33% negative). Functional Theory posits that policy is more frequently used than character because of voters’ preference to policy discussions. However, in the 2002 Korean presidential commercials, character (45%) was discussed nearly as frequently as policy (55%). This finding contrasts with American presidential ads, which in 2000 followed the traditional American pattern o f emphasizing policy (70%) more than character (30%). The difference may be a response to voter preferences. As suggested above, more American voters say policy rather than character is the most important determinant o f their presidential vote and the presidential candidates’ discourse might confirm to these preferences. However, public opinion polls during the 2002 Korean presidential election did not reveal the same preferences among voters. Overall, Korean voters split roughly evenly between those who believe policy is most important and those who believe character is most important. However, a closer look at these data suggests that as the campaign progressed, character may have become more important to voters. The proportion o f voters who reported that character was the most important determinant o f their preferred candidate increased from 35% on September 4, to 42% on October 11, and then 57% on November 24. This could explain, in part, why Roh won the election. It would be an overstatement that Roh was the winner because his 93 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. emphasis on character was consistent with voter preferences. Unlike Lee who emphasized policy (72%) more than character (28%), Roh addressed more on character (65%) than policy (35%) in his television spots and endorsement speeches (which were regarded as TV spots because o f the similar format and length). It is true that a single factor carmot predict the election result; candidates, issues, and voters are too complex for that. However, Functional Theory predicts that the candidate who discusses the topics that are most important to most voters should have an advantage. It is possible that Roh gained some benefits among voters with his emphasis on character. Note that this means this possibility-that Roh may have gained an advantage because he stressed character more than policy-does not contradict Functional Theory. It predicted that American presidential candidates who stressed policy more than their opponents are significantly more likely to win elections (Benoit, 2003). But if character were most important to most Korean voters at the time they voted. Functional Theory would predict an advantage for the candidate who discussed character most. It would be interesting to locate public opinion poll data on the relative importance o f policy and character during the 1997 Korean election. The functions o f political advertisements are influenced by incumbency status. Roh, who represented the incumbent party, employed more claims and fewer attacks than his challenger-party opponent. This finding is consistent with previous research on American spots: the incumbent party candidate tends to 94 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. employ more acclaims and fewer attacks than the challenger (Benoit, Blaney, & Pier, 1998; Benoit, 2002). The incumbent parties have records that can be used to evaluate their performance. In this vein, Benoit et al. (2003) argued that Gore made a mistake by not campaigning harder on the Clinton-Gore record. In the 2002 Korean presidential election, Roh, the incumbent party candidate, acclaimed more often than he attacked on general goals and ideals, whereas Lee, the challenger, attacked more than he acclaimed on past deeds. This finding is different from the previous studies which suggest that the incumbents mostly rely on his or her past deeds for acclaiming (Benoit, Pier, & Blaney, 1998). The difference may come from the fact that, as noted earlier, Korean presidents are limited a single term in office (and unlike the U.S., there is no vice presidential running mate who can run on the administration’s record as Bush did in 1988 and Gore did in 2000); thus, the past deeds o f the President are not directly related to Roh, the incumbent party candidate. Nevertheless, the challenger in the 2002 Korean campaign used the incumbent party’s past deeds for attacking, which is consistent with the previous findings (Benoit, Pier, & Blaney, 1998). Results show that future plans and general goals were the most frequent forms o f policy and ideals and personal qualities the most common forms of character in Korea. The main difference appears to be that American ads are more likely to offer specific policy proposals than Korean ads. Television Debates. This, study foxmd that acclaims (54%) were the most common fimction in the 2002 presidential debates, followed by attacks (33%), and then defenses (13%). This finding is important because it challenges the 95 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. notion that political campaign discourse was mostly negative. The candidates for the Korean presidency did engage in attacks during the debate, but attacks accounted for fewer than four in ten comments. Furthermore, this finding is consistent with Korean presidential spots. This order was found in other campaign message forms in the United Sates such as acceptance addresses (Benoit, Wells, Pier, & Blaney, 1999) and presidential television spots from 1952 to 2000 (Benoit, 2001). This suggests that the functions o f political debates are employed similarly in different cultures. It is notable that Korean candidates use defenses more frequently than American candidates. This tendency might be related to “face saving” which is very important in Pacific Asian countries under the influence o f Chinese culture (Lowe, 1996). Even though the Korean candidates used substantial amount o f defenses, they mostly relied on denial strategies and used apology in the final phase. It should be more effective if the candidates are engaged in diverse defense strategies. The public wants to know why the candidates made mistakes and how they would compensate for those mistakes in detail. During the debates, Korean candidates most emphasized future plans and past deeds. Past deeds were not frequently discussed in presidential spots and press releases. The reason is that candidates did not necessarily want to discuss their past deeds; however, the moderator allotted time for discussion of candidates’ past deeds. When the candidates addressed their character, they addressed personal qualities, leadership abilities, and ideals almost equally. 96 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Press Releases. The research expected that press releases are mostly negative because the purpose o f press releases was to influence media’s coverage. It means that even though candidates attacked the competitors, they did not suffer backlash because media blamed their competitors (if attacks were adopted by journalists). In line with this assumption, Wicks and Souley (2003) found that 75% of news releases during the 2000 U.S. campaigns contained an attack on the opponent. However, Korean presidential candidates relied mostly on acclaims as compared to attacks and defenses. It might come from the fact that news releases were also easily accessed by voters because candidates uploaded their press releases on their official campaign websites. Press releases are usually regarded as textual information that is distributed only to journalists to be used as a basis for news stories. However, candidates’ press releases during the election functioned also as direct mail for the public. Note that 60.4% o f voters answered they visited candidates’ websites at least once during the election (Shin, 2003). It is likely that candidates regarded press releases one o f their official media to promote their messages as well as to influence journalists to write favorably for them and unfavorable to the competitors. The proportion o f functions and topics were consistent with presidential spots and debates. One exception is that candidates’ press releases addressed specific plans (future plans) more than any other medium did. Future plans are often complex and need to be elaborated. Because of no money and time restrictions for using press releases, candidates might see good 97 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. opportunities to address their future plans fully and also attacked the competitors’ future plans in detail. Overall, television spots, television debates, and press releases used most frequently acclaims, followed by attacks and defenses. This ordering o f functions in campaign discourse makes senses. Acclaims make a candidate appear desirable, so candidates use them frequently. Attacks are used for the purpose of hurting the opponents’ desirability. However candidates moderate the use o f attacks because voters do not prefer mudslinging (Merritt, 1984). Defenses are infrequent because they can take a candidate off-message to a topic that probably favors the opponent. Defenses also might make a candidate appear weak. In addition, it is possible that a voter did not hear (or had forgotten) an opponent’s attack; thus, refuting the initial attack may actually inform (or remind) voters o f a candidate’s potential weakness (Benoit, 1999). Thus, it is reasonable for acclaims to be most common in campaign rhetoric, followed by attacks and then, infrequently, by defenses. The predictions for incumbency (incumbents acclaim and defend more than challengers, who attack more than incumbents) were supported generally. The only exception was that Lee (a challenger party candidate) used defenses more frequently than Roh (the incumbent party candidate) did. This may be a function of the fact that Lee was more o f front-runner than Roh for most of election campaign period. Research on American debates foxmd that more attacks are targeted at the candidates who are higher in the polls (Benoit et al., 2002). 98 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Some scholars argue that modem politics campaign discourse lacks substance. Schutz (1995) suggested that political campaigns focused on the candidates’ personalities rather than on political issues. However, empirical studies on the nature o f political campaign discourses showed that, as has been found in the U.S., candidates addressed more policy than character in the Korean presidential debates. Results show that Roh, the incumbent party candidate, acclaimed more often than he attacked on the past deeds, whereas the challenger, Lee, attacked more than he acclaimed on the past deeds. This is interesting because Roh was the incumbent party candidate but not the President (the president of Korea may not run for re-election). This meant that the previous President’s past deeds was not necessarily Roh’s responsibilities. For the challenger, past deeds was good source o f attacks. Lee tried to link Roh with the scandals in Kim’s governments. Lee adopted past deeds to attack rather than to acclaim. On the other hand, Roh did not attack on Lee’s past deeds as much as Lee did even though Lee had substantial experience as statesman. The ordering o f the specific forms of policy and character utterances was fluid as shown in the previous studies o f presidential debates. In Korea, past deeds outnumbered future plans and general goals were the least frequently used. Even though candidates can easily acclaim with general goals as compared to future plans which demanded specific plans, they preferred to propose future plans. This might be related to the response to the media’s criticism that political candidates make unrealistic promises. In order for candidates to avoid this criticism, they seem to use more future plans than general goals. With regard to the forms of 99 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. character, the candidates relied on personal qualities and ideals rather than leadership ability across media. M edia’s Use o f Candidates ’ Discourse Based on previous studies on news values and election coverage, the researcher expected that the proportion o f attacks and defenses will be higher in media converge than in the candidates’ political discourse; whereas, the proportion o f acclaims will be lower in media coverage than in the candidates discourse. In order to study media’s use o f candidate discourse, this study attempts to analyze what types of political discourse get prominent attentions by journalists. Candidates routinely include in news stories during the election, however, journalists select only a few among candidates discourse based on their news value judgments. A comparison between candidate discourse and media’s citation o f candidates discourse support this hypothesis. Media overly used attacks and defenses as compared to acclaims, which was not consistent with the ordering o f candidates’ use o f political discourse. This finding is consistent with U.S. studies. For example, Reber and Benoit (2001) reported that attacks and defenses were more prominent in newspaper reports about those debates than in the debates themselves in 2000 presidential primary debates. Benoit and Currie (2001) also found that coverage o f Bush and Gore in the 2000 presidential debates followed the same pattern: the coverage accentuated negative and defensive utterances. There are several reasons why candidates’ negative comments are attractive to journalists. First, attacks are conflictual and conflicts add interest to 100 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. news stories (Benoit & Currie, 2001). Journalists often judge the newsworthiness of events by their degree o f negativity. For journalists, it is a no-lose situation when candidates attack one another. Even if the charges prove to be false, reporters can always rail against the candidates who initiated that attack for slandering his opponent and engaging in sleazy campaigning (Ansolabehere & Iyengar, 1995). In fact, it is necessary for candidates to appear preferable to other candidates for winning the election. Thus, candidates should find weaknesses of opponents and address them. However, the mass media emphasize too much on negative aspects of campaigns as compared to the actual candidates discourse, which might cause voter cynicism and even low-voter turnout (Ansolabehere & Iyengar, 1995). Sabato (1993) also argued that that people often appear to be irate about media’s negative coverage o f candidates. It is out o f scope to examine the effects o f attack journalism on voters; however, the media’s overreporting of candidates’ attacks were inaccurate at least. Korean voters believe that election campaigns negative (Yang, 2003). This belief might stem more from the news coverage o f the campaign than form the campaigns themselves. Defenses are also over-represented in media citation o f candidates’ comments. As in the case o f attacks, defenses emphasize conflict and show differences between the candidates. However, acclaims, which are neglected in media coverage, are also important to voters. Voters need to know the candidates’ strengths and their goals and proposal to make informed decisions about the relative strengths o f presidential candidates. 101 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. With regard to character coverage, Domke et al. (1997) found that “during the latter stages o f the campaign, character because the most prominent issue in the media coverage” (p. 737). Presidential campaigns are often held in the last stage o f the campaign. However, character was not over-represented in the news coverage o f candidates. The proportion o f policy and character utterances appeared in newspapers was no different form the proportion o f policy and character comments in candidates’ discourse. Who Sets the Public Agenda During the Election? The conventional wisdom in political research is that the media play a dominant role in setting the agenda o f elections. There is substantial empirical evidence that the issues presented by the media strongly influenced the public’s own issue priorities, not the other way around (e.g., McCombs & Shaw, 1972; McCombs, 1981; Shaw & McCombs, 1977; Semetko, Blumler, Gurevitch, & Weaver, 1991). However, several studies challenged the media’s ability to transfer issue saliencies to the public during the election. Instead, candidates took the most dominant role in agenda setting process. With regard to campaign’s influence on voters, Kern (1989) reported that “by a ration o f 4 to 1, Americans received the majority of their information about candidate positions on the issues from ads rather than the news” (p. 47). Different factors are at work here; however, a majority o f scholars pointed at media’s horse-race coverage o f election campaigns. Patterson (1980) argued, “in its coverage o f the presidential campaign, the press concentrates on the 102 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. strategic game played by the candidates in their pursuit o f the presidency, thereby de-emphasizing the questions o f national policy and leadership” (p. 21). Horse­ race coverage does not limit its scope to who is ahead and who is behind in the public opinion polls. Media’s coverage o f horse-race includes organizational and financial strength, endorsements, and candidates’ campaign performance. More than half o f all election coverage was allotted to horse-race in U.S. (Freitag, 2000; King, 1990; Russonello & Wolf, 1977). In other words, the horse-race coverage o f the presidential campaign was the dominant theme at the cost o f other substantial issue coverage. As a result, recent studies argue that candidates’ ability to setting the public agenda exceeds media’s agenda setting ability (Dalton et al., 1998; Robert& McCombs, 1994). However, the findings showed that Korean media did not focus more on horse race than candidates’ policy positions and character. Only 22% o f news stories included horse-race elements o f election. Compared to U.S. coverage o f horse race, Korean newspapers emphasized candidates’ issue positions and character. Previous U.S. studies on media coverage o f horse race showed that candidates’ stance in tracking polls and their financial strength were mostly emphasized. Korean newspapers also focused on candidates’ support rates, however, they put more emphasis on candidates’ performance in the debates, advertisements, and interviews with media. The media may have shown interest in candidates’ performances because these media-oriented political campaigns are newly introduced in Korean political systems. 103 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The researcher questioned who set the public agenda during the election. Results showed that the media agenda was highly correlated to the voter agenda. After controlling for the campaign’s agenda setting effects, the media agenda was significantly correlated to the public agenda. Thus, this study concludes that the media is capable o f raising the importance o f an issue in the public’s mind. High correlations did not mean that causation; however, this study assumed that the media set the public agenda. In other words, journalists determine which issues should be treated importantly and transfer their issue salience to the public. There is a reason to believe this assumption. First, Korean election campaign period lasted only two months. Thus, the public did not have enough time to set their agenda and influence public and media. Most previous agenda setting studies found that media set the public agenda, not the other way around. Based on this assumption, numerous agenda setting studies reported that media’s agenda setting ability based on high correlations between media and voters (e.g., Dalton et al., 1998; Ghanem, 1997). Audience Factor in Agenda-Setting Process One of the criticisms against agenda setting theory is that individual differences are often not considere. Weaver (1977) suggested that individuals differ in their need for orientation and that this might decide whether agenda setting took place or not. Need for orientation is based on two factors; the relevance o f the information and the degree o f uncertainty concerning the subject of the message. Weaver argued that these two factors were positively related to the strength o f agenda setting effects. Simply put, if voters who were highly 104 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. involved with the issues, they were more susceptible to media’s agenda setting effects. This argument was tested in this study during 2002 Korean presidential election. The researcher expected that highly involved voters are more susceptible to media’s agenda setting effects than the general voters. Highly involved voters were operationalized as Internet forum discussion participants during the election campaign. Because forum participants actually expressed their ideas and concerns about candidates, they were regarded as more active and knowledgeable than the general public. Results show that the correlations between media and general public are higher than those o f media and the high-involvement public. This finding is not consistent with Weaver’s argument. In fact, the correlations between campaigns and high-involvement voters are very high (r=. 90, p<. 01). One possible explanation is that highly involved voters exposed themselves to those news items that are in agreement with their existing attitudes and to avoid those news items that are not (Cotton, 1985). In other words, active audiences seem to resist media’s agenda setting effects. Instead they were more influenced by campaign agenda than media agenda. Thus, the researcher reported that media’s agenda setting effect is weaker for highly-involved voters than for the general public. Media Content’s Effects on the Public Lazarsfeld, Berelson, and Gaudet (1944) argued that the direct effects of campaigns on election outcomes were minimal. Klapper (1960, p.8) also stated that, “mass communication ordinarily does not serve as a necessary and sufficient cause o f audience effects but rather functions among and through a nexus of 105 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. mediating factors and forces.” Even though election campaigns do not possess direct effects on voters according to the limited effects approach, hundreds of millions o f dollars were spent on media campaign. For a country whose election campaign was mostly delivered through media in a short period o f time, the minimal effects approach is often doubted (Yang, 2003). Based on this notion, McCombs introduced second-level agenda setting and Shaw (1999) provided that media’s favorability o f the candidate would directly influence the fluctuation o f poll results. With regard to second-level agenda setting, McCombs— who originally posited that mass media had their effects on people’s perception rather than on their attitude or behavior— suggested that media have stronger effects than the traditional agenda setting. McCombs stated that media are also successful in influencing how to think and called it second-level agenda setting. Second level agenda setting suggested that attributes o f issues (sub-issues) that are receiving the media’s emphasis should also be the attributes that the public indicates are most important. In the context o f political campaigns, the mass media’s coverage o f candidates can shape voter perceptions with regard to the object’s attributes. Thus, the media’s attributes coverage o f candidates can have a strong influence on voter perceptions (Golan & Wanta, 2001). Second-level agenda setting takes place at two types o f attributes in media coverage: cognitive and affective attributes. Affective agenda setting suggested that tone o f media coverage influenced readers. In other words, positive (or negative) coverage would influence how positively (or negatively) individuals 106 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. view candidates. If a newspapers not only report a candidate’ issue stance on moral issues but do so in a positive light—highlighting positive aspects o f his stances-readers may link this positive coverage to the candidate again (Golan & Wanta, 2001). As a result, positive coverage o f a candidate will lead to positive evaluations o f that candidate, while negative coverage o f a candidate will lead to negative evaluations of that candidate. In sum, media transfer the tone of discussions o f the candidates to voters according to the second-level agenda setting at the affective dimension. Several studies found this effect (Ghanem, 1996; Lopez-Escobar, McCombs, Rey, 1998) This study tested the affective attribute agenda setting effect using onlinediscussion messages and media coverage. Results showed that positive coverage of Lee led to positive evaluations o f Lee; however, negative coverage o f Lee did not induce the negative discussions o f Lee. For Roh, neither positive nor negative coverage did not influence voters’ evaluations o f him. This study concludes that positive coverage did not always lead to positive public perceptions. In addition, negative coverage did not influence negative public perceptions. The researcher speculates that selective exposure played a role in the information processing o f newspapers. Compared to the general public, Internet discussion participants are more active to support or attack candidates; thus, the media’s tone might not influence individuals’ discussions. Shaw (1999) used different approach to show the media’s impact on voters’ behavior. He argued that the hypothesis o f minimal effects o f campaign came from the lack o f control for the effects o f the favorability o f news media 107 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. coverage. Media favorability means, according to Shaw, that the sum o f a candidate’s positive news coverage and the competitor’s negative news coverage. Shaw (1999) expected that the effects o f the favorailibty o f the news media’s coverage o f campaign could influence presidential election outcome. A few recent studies have argued that the media’s relatively favorable coverage of Bill Clinton played a particularly influential role in the 1992 presidential election (Patterson, 1993; Sabato, 1993). Secondly, current presidential candidates’ campaigns were mostly delivered through media, thus helping candidates reach more voters than either personally or through their surrogates. From this reason, Shaw argued that the candidates’ support rates changes should be, at least in part, influenced by media coverage o f candidates. This study examined whether the media favorability influenced the candidates’ stance in tracking polls. The media favorability is defined as the sum o f a candidate’s positive coverage and the competitor’s negative coverage. Results here showed that media’s net favorable coverage for a candidate explained 38% o f the variances in the presidential candidates’ stance in tracking polls. This finding showed that media coverage o f candidates did affect the vote in the 2002 Korean presidential election. The researcher did not offer an explanation for how candidates can receive favorable media coverage. Instead, the focus o f study was to examine the media content’s effects on voters at the behavioral level. These findings draw a several theoretical suggestions. First, although the presidential election outcomes are most frequently explained by individual’s political affiliations and the power o f money (Campbell, 1992; 108 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Campbell & Wink, 1990), this study showed that aggregate levels of candidate support can change because o f favorable media coverage. The researcher did not argue that media’s coverage is the most important factor to explain election outcomes. Instead, the researcher suggested that news media coverage of candidates could provide the stimulus for change o f support rates in the tracking polls. Benoit, Blaney, and Pier (2003) suggested that voters had no choice but to choose to vote for the candidate appearing preferable to them on whatever criteria. For this decision, voters conducted informal cost-benefit analysis in which benefit is candidates’ strengths and cost is candidates’ weakness. Results showed that people conducted their cost-benefit analysis based on evaluations of candidates presented by mass media. Limitations and Areas for Future Studies This dissertation is composed o f two main parts: (1) the content analysis o f presidential political discourse and (2) the examination o f media effects on the public. For the content analysis part, the researcher included the presidential spots including endorsement speeches, television debates, and press releases. However, there is no guarantee that these three media represent the candidates’ discourse perfectly. It is possible that other political speeches such as keynote speeches, newspaper ads, and party platforms represent the candidates’ discourse better. Thus, future studies should include more political discourse delivered by the campaigns. 109 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The second part o f this dissertation focused on examining effects o f media on the public during the election. For this purpose, media content was compared to poll data. The opinion poll data provided the professional polling institutions is one o f the accurate data. However, the research questions were limited by the questionnaires o f opinion poll data. The following studies on this topic may consider conducting the survey that can examine the attribute agenda setting at the cognitive level, which will enrich the body o f media effect studies. Conclusion This study examined the nature and effects o f Korean presidential discourse. For these purposes, the researcher relied on three important conceptualizations in the political communication field— that is— Benoit’s the Functional Theory o f Political Discourse, McCombs’ first and second level agenda setting theory, and Shaw’s media favorability effects on the voters. Findings show that there are similarities and differences in applying these theories to Korean presidential election discourse. First, this study applies the Functional Theory o f Political Campaign Discourse to the 2002 Korean presidential campaigns using the incumbent party candidate, Roh Mu Hyun and the challenging party candidate, Lee Hoi Chang. The finding challenges the notion that Korean presidential campaign was mostly negative. The incumbent party candidate devoted more discourse to acclaiming than the challenger candidate. Despite the belief that modem campaigns lack in 110 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. substance, the candidates in the debates emphasized more on policy than on character. Second, the researcher examines the effects o f media during the election. As a theory o f media effects, the traditional (first-level) agenda setting approach might lack media’s direct influence on the public. By incorporating second level agenda setting theory and framing effects approach, this study found that media also influenced the attitudes and behavior o f voters. This study does not claim that media content alone explained the voting behavior or attitude towards the presidential candidates. Instead, it is found that the fluctuation o f opinion poll results can be substantially explained by media content. To establish this argument, a more concerted effort needs to be made in including experiments and surveys which can directly assess the changes in public opinion. Ill Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. REFERENCES Altschuler, B. (1982). Keeping a finger on the public pulse: Private polling and presidential election. Westport, CT: Greenwood Press. Ansolabehere, S., & lyengr, S. (1995). Going negative: How political advertisements shrink and polarize the electorate. New York: Free Press. Benoit, W. L. (2001). The functional approach to presidential television spots: Acclaiming, attacking, and defending 1952-2000. Communication Studies, 5 2 ,109-126. Benoit, W. L., & Brazeal, L. M. (2002). A functional analysis o f the 1988 Bush-Dukakis presidential debates. Argumentation and Advocacy, 38, 219-233. Benoit, W. L., Blaney, J. R., & Pier, P. M. (1998). Functions o f campaign ‘96: Acclaiming, attacking, and defending discourse. Westport, CT: Praeger. Benoit, W.L. (1995). Accounts, excuses, and apologies: A theory o f image restoration strategies. Albany: State University of New York Press. Benoit, W.L. (1997). Image restoration discourse and crisis communication. Public Relations Review, 23, 177-186. Benoit, W.L. (1999). Seeing spots: A functional analysis o f presidential television advertisements, 1952-1996. Westport, CT: Praeger. Benoit, W.L. (2003). Topic of presidential campaign discourse and election outcome. Western Journal o f Communication, 67, 97-113. Benoit, W.L., & Brazel, L.M. (2002). A functional analysis o f the 1988 Bush-Dukakis presidential debates. Argumentation and advocacy, 3 8 ,219-233. Benoit, W.L., & Hathcock, A. (1999). Functions of the Great Debates: Acclaims, attacks, and defenses in the 1960 presidential debates. Communication Monographs, 66, 341-357. 112 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Berelson, B. (1952). Content analysis fo r social sciences and humanities. Reading, MA: Addison-Wesley. Boyle, T. (2001). Intermedia agenda setting in the 1996 presidential election. Journalism & Mass Communication Quarterly, 78, 26-44. Brazeal, L.M., & Benoit, W.L. (2001). A functional analysis o f congressional television spots, 1986-2000. Communication Quarterly, 49, 437-453. Breed, W. (1955). Social control in the newsroom. Social Forces, May: 326-335. Cappella, J.N., & Jamieson, K.H. (1994). Spiral o f cynicism: The press and the public good. New York: Oxford University Press. Choi, Y. (2002, December 8). Presidential debates. Korea Times. Obtained on line from the http://www.kinds.or.kr. Cohen, J. (1960). A coefficient o f agreement for nominal scales. Educational and Psychological Measurement, 20, 37-46. Cohen, B.C. (1963). The press andforeign policy. Princeton, N.J.: Princeton University Press. Cotton, J.L. (1998). Cognitive dissonance in selective exposure. In D. Zillmann and J. Bryant, eds., Selective Exposure to Communication, pp. 11-33. Hillsdale, N.J.: Lawrence Erlaum. Dalton, R., Beck, P.A., Huckfeldt, R., & Koetzle, W. (1998). A test o f media-centered agenda setting: Newspaper content and public interests in a presidential election. Political Communication, 15, 463-481. Dearing, J.W., & Rogers, E.M. (1996). Communication concepts 6: Agenda-setting. Thousand Oaks, California: Sage. Domke, D., Fan, D.P., Fibison, M., Shah, D.V., Smith, S.S., & Watt, M.D. (1997). News media, candidates and issues, and public opinion o f the 1996 presidential campaign. Journalism & Mass Communication Quarterly, 75,718-738. 113 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Elliott, W.R., & Sothirajah, J. (1993). Post-debate analysis and media reliance: Influences on candidate image and voting probabilities. Journalism Quarterly, 70, 321-335. Entman, R.M. (1993). Framing: Towards clarification o f a fractured paradigm. Journal o f Communication, 43, 51-58. Entman, R.M., & Rojecki, A. (1993). Freezing out the public: Elite and media framing o f the U.S. anti-nuclear movement. Political Communication, 10, 155-173. Fleiss, J.L. (1981). Statistical methods fo r rates and proportions. New York: Wiley. Fouser, R.J. (2002, November 18). Election 2002 sets important precedents. Korea Heralds. Obtained on line from the http://www.kinds.or.kr. Freitag, A.R. (2000). An analysis: New Hampshire primary coverage. Newspaper Research Journal, 21, 94-106. Friedland, A., & Mengbai, Z. (1996). International television coverage o f Beijing Spring 1989: A comparative approach. Journalism & Mass Communication Monographs, 156, 1-61. Galtung, J., & Ruge, M.H. (1965). The structure of foreign news. Journal o f Peace Research, 2, 64-91. Gamson, W. (1992). Talking politics. New York: Cambridge University Press. Gamson, W., & Modigliani, A. (1987). The changing culture of affirmative action. In R.Bruangart (Ed.), Research in political sociology (Vol. 3, pp. 137-177). Greenwich, CT: JAI Press. Gans, H.J. (1979). Deciding what is news: A study o f CBS evening news, NBC nightly news, Newsweeks and Times. New York: Panthom Books. Garamone, G.M. (1985). Effects o f negative political advertising: The roles o f sponsor and rebuttal. Journal o f Broadcasting & Electronic Media, 29, 2, 147-159. 114 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ghanem S. (1997). Filling in the tapestry: The second level o f agenda setting. In M.McCombs, D.L. Shaw, and D. Weaver, eds., Communication and Democracy: Exploring the intellectual frontiers in agenda-setting theory, pp. 3-14. Mahwah, N.J.: Lawrence Erlbaum. Ghanem, S. (1996). Media coverage o f crime and public opinion: An exploration o f the second level agenda setting. Ph.D. dissertation: University o f Texas, Austin. Gitlin, T. (1980). Te whole world is watching: Mass media in the making & unmaking o f news left. Berkeley. University o f California Press. Golan, G. & Wanta, W. (2001). Second-level agenda setting in the New Hampshire Primary: A comparison o f coverage in three newspapers and public perceptions o f candidates. Journalism and Mass Communication Quarterly, 78, 247-259. Hall, S. (1981). The determinants o f photographs. In S.Cohen and J. Young (Eds.), The manufacturing o f news: A reader (p. 226-243). Beverly Hills, CA:Sage. Holtz-Bacha, C., Kaid, L.L., & Johnston, A. (1994). Political television advertising in Western democracies: a comparison o f campaign broadcasts in the United States, Germany, and France. Political Communication, 11, 67-81. Huang, K.S. (1996, May). A comparison between media frames and audience frames: The case o f the Hill-Thomas controversy. Paper presented at the annual convention o f the International Commimication Association, Chicago. Iyengar, S. (1991). Is anyone responsible? How television frames political issues. Chicago: University o f Chicago Press. Iyengar, S., & Kinder, D.R. (1986). News that matters: Television and American opinion. Chicago: University o f Chicago Press. James, M. C., Wotring, E., & Forrest, E. J. (1995). An exploratory study o f the perceived benefits o f electronic bulletin board use and their impact on other commimication. Journal o f Broadcasting and Electronic Media, 39, 30-50. 115 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Jamieon, K.H., & Birdsell, D.S. (1998). Presidential debates: The challenge o f creating an informed electorate. New York: Oxford University Press. Jamieson, K.H. (1996). Packaging the presidency: A history and criticism o f presidential campaign advertising. New York: Oxford University Press. Jamieson, K.H., & Waldman, P. (2002). The morning after: the effect o f the network call for Bush. Political Communication, 19, 113-119. Johnson, T. J. & Kaye, B. K. (1998). Cruising is believing? Comparing media and traditional sources on media credibility measures. Journalism and Mass Communication Quarterly, 75, 325-240. Johnson, T.J. (1993). Filing out the race form: How the media covered horse race in the 1988 primaries. Journalism Quarterly, 70, 300310. Johnson, T.J., Kaye, B.K. (2000). Using is believing: the influence o f reliance on the credibility o f online political information among politically interested internet users. Journalism & Mass Communication Quarterly, 77, 865-876. Johnston, A., & Kaid, L. L. (2001). Image ads and issue ads in U.S. presidential advertising: Using videostyle to explore stylistic difference in televised political ads from 1952 to 2000. Journal o f Communication, 281-300. Johnston, A., & Kaid, L.L. (2002). Image ads and issue ads in U.S. presidential advertising: using videostyle to explore stylistic differences in televised political ads from 1952 to 2000. Journal o f Communication, 52, 281-301. Joslyn, R.A. (1980). The content o f political spot ads. Journalism Quarterly, 57, 92-100. Just, M., Criegler, A., Alger, D., & Cook, T. (1996). Crosstalk: Citizens, candidates, and the media in a presidential campaign. Chicago: University o f Chicago Press. Kahneman, D., & Tversky, A. (1984). Choices, values, and frames. American Psychologists, 39, 341-350. 116 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Kaid, L.L. (1994). Political advertising in the 1992 campaign. In R.E. Denton (Ed.), The 1992 presidential campaign: A communication perspective (pp. 111-127). Westport, CT: Praeger. Kenney, K., & Simpson, C. (1993). Was coverage of the 1988 presidential race by Washington’s two major dailies biased.^ Journalism Quarterly, 70, 345-355. Kern, M. (1989). 30 second politics: Political advertising in the eighties. New York: Praeger. Kern, M. (2001). Disadvantage A1 Gore in Election 2000: Coverage o f issue and candidate attributes, including the candidate as campaigner, on newspaper and television news web sites. American Behavioral Scientist, 44, 2125-2139. Kim, K. (2002, December 18). Finally, the nation goes to polls. Korea Times. Obtained on line from the http;//www.kinds.or.kr. King, E.G. (1995). The flawed characters in the campaign: prestige newspaper assessments o f the 1992 presidential candidates' integrity and competence. Journalism & Mass Communication Quarterly, 72, 84-98. King, E.G. (1990). Thematic coverage o f the 1988 presidential primaries: A comparison o f USA Today and the New York Times. Journalism Quarterly, 67, 83-87. Kiousis, S., Bantimaroudis, P., & Ban, H. (1999). Candidate image attributes: Experiments in the substantive dimension o f second level agenda setting. Communication Research, 26, 414-415. Klapper, J.T. (1960). The effects o f mass communication. New York: Free Press. Korean Press Foundation. (2003). Datongryung sunkuwa media [Presidential elections and media]. Seoul: Hanul. Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrica, 33, 159-174. Lau, R.R. (1985). Two explanations for negativity effects in political behavior. American Journal o f Political Science, 29, 119-138. 117 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Lazarsfeld, P.F., Berelson, B., & Guadet, H. (1944). The p eo p le’s choice. New York: Columbia University Press. Lichter, S.R., & Noyes, R.E. (1995). Good intentions make bad news: Why Americans hate campaign journalism. Lanham, MD: Rowman. Lippmann, W. (1922) Public Opinion. New York: Free Press. Lowe, A. (1996). Face saving in business relationships: A basic social process. Helsingfors: Swedish School o f Economics and Business Administration Llamas, E.L., McCombs, M. Rey, F. (1997). Two levels of agenda setting among advertising and news in the 1995 Spanish elections. Political Communication, 15, 225-238. McCombs, M. E. (1981). Setting the agenda for agendasetting research: An assessment o f the priority ideas and problems. In G.C. Wilhoit and H.de Bock (Eds.), Mass Communication Review Yearbook, 2, 209-211. Beverly Hills, Calif.: Sage. McCombs, M., & Shaw, D. (1972). The agenda setting function o f mass media. Public Opinion Quarterly, 36, 176-187. McCombs, M., & Shaw, D. (1977). The emergence o f American political issues: The agenda-setting function o f the press. N.Y.: West Publishing Co. McCombs, M.F., Llamas, J.P., Lopez-Escobar, E., & Rey, F. (1997). Candidate images in Spanish elections: Second-level agendasetting effects. Journalism & Mass Communication Quarterly, 74, 703717. McNelly, J.T., & Izcaray, F. (1992). International News Exposure and Images o f Nations Journalism & Mass Communication Quarterly 63, 546-553. Merritt, S. (1984). Negative political advertising: Some empirical findings. Journal o f Advertising, 13, 27-38. Meyerowitz, B.E. & Chaiken, S. (1987). The effect of message framing on breast self-examination attitudes, intentions, and behavior. Journal o f Personality and Social Psychology, 52, 500-510. 118 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Miller, M.M., Andsager, J.L., & Riechert, B.P. (1988). Framing the candidates in presidential primaries: Issues and images in press releases and news coverage. Journalism & Mass Communication Quarterly, 75, 312-324. Ogan, C. (1993). Listserver communication during the Gulf War: What kind o f medium is the electronic bulletin board? Journal o f Broadcasting and Electronic Media, 37, 177-196. Park, R. (1967). News as a form o f knowledge. In R.H. Turner (Ed.), On social control and collective behavior (pp. 3252). Chicago: Chicago University Press. Patterson, T.E. (1980). The mass media election: How Americans choose their president. New York: Praeger. Patterson, T.E. (1993). Out o f order. New York: Alfred A. Knopf. Patterson, T.E. (1996). Bad news, period. Political Science and Politics. 29, 17-20. Patterson, T.E., & McClure, R.D. (1976). The unseeing eye: The myth o f television power in national politics. New York: Putnam. Pfau, M., & Kang, J.G. (1991). The impact o f relational messages on candidate influence in televised political debates: 1988 George Bush and Michael Dukakis presidential debate. Communication Studies, 42, 114-129. Pfau, M., & Kenski, H.C. (1990). Attack politics: Strategy and defense. New York: Praeger. Pomper, G.M. (1975). Voters ’ choice: Varieties o f American electoral behavior. New York: Dodd, Mead, & Company. Reber, H., & Benoit, W.L. Presidential debate stories accentuate the negative. Newspaper Research Journal, 22, 30-44. Rhee, J.W. (1997). Strategy and Issue Frames in Election Campaign Coverage: A Social Cognitive Accoxmt o f Framing. Journal o f Communication 47, 26-48. 119 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Robert, M., Wanta, W., & Dzwo, T-H. (2002). Agenda setting and issue salience on-line. Communication Research, 29, 452-466. Roberts, M., & McCombs, M. (1994). Agenda setting and political advertising: Origins of news agenda. Political Communication, 11, 249-262. Rosenberg, K.E. (1970). International news: Intra and extra media data. Acta Socilogica, 13, 96-109. Russonello, J.M., & Wolf, F. (1977). Newspaper coverage o f the 1976 and 1968 presidential campaigns. Journalism Quarterly, 360-364. Sabato, L.J. (1993). Is there an anti-republican, anti-conservative media tilt? Campaigns and Elections, 14, 16-20. Sallot, L.M., & Steinfatt, T.M., & Salwen, M.B. (1998). Journalists' and public relations practitioners' news values: perceptions and cross-perceptions. Journalism & Mass Communication Quarterly, 75, 366-378. Salwen, M.B., & Matera, F. (1989). Public Salience of Foreign Nations. Journalism Quarterly, 69 , 623-632. Scheufele, D.A. (1999). Framing as a theory o f media effects. Journal o f Communication, 49, 101-120. Semetko, H.A. (1996). The media. In L. LeDuc, R. Niemi, & P. Norris (Eds.), Comparing democracies (pp. 51-71). Newbury Park, CA: Sage. Semetko, H.A., & Canel, M. J. (1997). Agenda-senders versus agenda-setters: television in Spain's 1996 election campaign. Political Communication, 14, 459-480. Semetko, H.A., Blumler, J.G., Gurevitch, M., & Weaver, D. H. (1991). The formation o f campaign agendas. Hillsdale, NJ: Lawrence Erlbaum Associates. Semetko, H.A., Joanne, B., Weaver, D., & Willnat, L. (1992). TV News and U.S. Public Opinion About Foreign Countries: The Impact o f Exposure and Attention. International Journal o f Public Opinion Research, 4, 18-36. 120 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Severin, W. J. & Tankard, J. W. (2001). Communication theories: Origin, methods, and uses in the mass media. New York: Longman. Shah, D., Domke, D., Wackman, D.B. (1996). To Thine Own Self Be True: Values, Framing, and Voter Decision Making Strategies,” Communication Research 23, 509-560. Shelley, M.C., & Hwang, H-D. (1991). The mass media and public opinion polls in the 1988 presidential election. American Politics Quarterly, 19, 59-79. Shaw, D. R. (1999). The impact o f news media favorability and candidate events in presidential campaign. Political Communication, 16, 183-202. Shaw, D.R. (1994). Strong persuasion? The effects o f campaigns in U.S. presidential elections. Unpublished doctoral dissertation. University o f California, Los Angeles. Shin, H., & Lee, H.S. (2002). A Study o f Korean Presidential Election Coverage. Seoul: Korea Press Foundation. Shoemaker, P., Danielian, L.H., & Brendlinger, N. (1991). Deviant acts, risky business and U.S. interests: The newsworthiness o f world events. Journalism Quarterly, 68, 4, 781795. Shyles, L. (1986). The television political spot: Its structure, content, and role in the political system. In L. L. Kaid, D. Nimmo, & K. R. Sanders (Eds.) New perspectives on political advertising. Carbondale: Southern Illinois University Press. Stephens, M. (1997). A history o f news. Harcourt brace college publishers: Fort Worth. Stewart, C.J. (1975). Voter perceptions o f mudslinging in political communication. Central States Speech Journal, 26, 279286. Tak, J., Kaid, L., & Lee, S. (1997). A cross-cultural study o f political advertising in the United States and Korea. Communication Research, 24, 413-435. 121 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Tanner, E. (2001). Chilean conversations: Internet forum participants debate Augusto Pinochet’s detention. Journal o f communication, 51, 383-403. Tedeseo, J.C. (2001). Issue and Strategy Agenda-Setting in the 2000 Presidential Primaries. American Behavioral Scientist, 44, 2048-2067 Turk, J.V. (1985). Public relations’ influence on the news. Newspaper Research Journal, 7, 15-27. Turk, J.V. (1986). Public relations’ influence on the news. Newspaper Research Journal, 7, 15-27. Wanta, W. (2000). Chatrooms and the spiral o f silence: An examination o f online discussions during the final 1996 U.S. presidential debates. Paper presented to the political communication division at the International Communication Association annual convention, Acapulco, Mexico. Wanta, W., & Hu, Y-W. (1993). The Agenda-Setting Effects o f International News Coverage: An Examination of Differing News Frames. International Journal o f Public Opinion Research, 5 250-264. Weaver, D.H. (1977). Political issues and voter need for orientation. In D.L. Shaw & M.E. McCombs (Eds.), The emergence o f American political issues: The agenda-setting function o f the press (pp. 107-119). St. Paul, MN: West. Weaver, D.H. (1991). Issue salience and public opinion: Are there consequences o f agenda-setting? International Journal o f Public Opinion Research, 3, 53-68. Weaver, D. (1981). Media agenda-setting in a presidential election. New York; Praeger. Weaver, D. (1994). Media agenda setting and elections: voter involvement or alienation? Wicks, R.H., & Souley, B. (2003). Going negative: Candidate usage o f internet websites during the 200 presidential campaign. Journalism & Mass Communication Quarterly, 78 (?). 122 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Wilcox, D.L., Ault, P.H., & Agee, W.K. (1989). Public relations: Strategy and tactics (2"‘*ed.). New York: Harper & Row. Winter, J., & Eyal, C. (1981). Agenda setting for the civil rights issue. Public Opinion Quarterly, 45, 376-383. Yang, S.C. (2003). Presidential Election and media. Seoul: Korea Press Foundation. Zucker, H. (1978). The variable nature o f news media influence. In Communication Yearbook 2, (Ed.), Ruben, B. pp. 225-240. New Brunswick, N.J.: Transaction Books. 123 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix. Sample Acclaims and Attacks on Forms o f Policy and Character Policy Acclaim Past Deeds; Roh claimed that his party’s sunshine policy was the only choice which contributed to bring peace in Korean peninsula. Attack Past deeds: Lee emphasized the need for a power transfer by railing against a series o f corruption scandals that have occurred under President Kim Dae-jung and MDP. Acclaim Future Plan: Kwon promise he will create a heavy wealth tax imposing on the wealthy to provide better social welfare to the poor. Attack Future Plan: Lee dismissed Rob’s plan to move the capital city to, saying it demonstrated Rob’s high-handed approach to the issue and was an empty promise aimed at winning over voters in the Chungchong Provinces. Acclaim General Goal: Lee promised to increase governmental spending on social welfare. Atack General Goal: Lee’s hard-line North Korea policy would risk another war on the Korean peninsula (Roh). Character Acclaim Personal Qualities: “I was an adopted child o f the people because I was a self-made man from Roh, a self-made man from farming family” (Roh). Attack Personal Qualities: Roh labeled Lee as the incardination o f old politics by mentioning that he was suspicious o f numerous political scandals. Acclaim Leadership Ability: Lee praised that his leadership is based on law and principle. Attack Leadership Ability: Lee is the aristocratic and authoritarian leader (Roh). Acclaim Ideals: Roh stated that the role of society is final safety net for the poor. Attack Ideals: Roh is a radical leftist (Lee). 124 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. VITA Cheolhan Lee was bom in 1973, in Seoul, Korea. He holds the following degrees: B.A. in Mass Communication from Yonsei University; M.A. in Mass Communication from Yonsei University; M.A. in Public Communications from Syracuse University; Ph.D. in Journalism from the University MissouriColumbia. His primary research interest is mass persuasion, or advertising (selling products), public relations (selling ideas), and election campaign (selling candidates). During his stay in the U.S., he worked with Dr. Comstock, Dr. Benoit, and Dr. Wanta and published several articles in Joumal o f Human Subjectivity, Commimication Quarterly, Asian Journal o f Communication, and Joumalism and Mass Communication Quarterly. 125 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.