‘ « llllllllllllllzlllllllllllllllllllllllllHllllllllllll gs, \ 301575 5732 LIBRARY Michigan State University This is to certify that the dissertation entitled SUPERFICIAL IMAGERY VERSUS SUBSTANTIVE MATERIAL: THE PROCESSING OF EACH IN POLITICAL ADVERTISEMENTS presented by Janet Katherine McKeon has been accepted towards fulfillment of the requirements for Ph.D. degree in Jommnnination Major professor Date 344/7461 ”47 MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771 PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE - ssp‘élsffié’ssl’ *- WT FEBOMZZHIO i ' m g 5‘ M71 DATE DUE DATE DUE “if 1 a @3031. 1/0! mas-nu SUPERFICIAL IMAGERY VERSUS SUBSTANT IVE MATERIAL: THE PROCESSING OF EACH IN POLITICAL ADVERTISEMENTS By Janet Katherine McKeon A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Communication I997 ABSTRACT SUPERFICIAL IMAGERY VERSUS SUBSTANT IVE MATERIAL: THE PROCESSING OF EACH IN POLITICAL ADVERTISEMENTS By Janet Katherine McKeon Candidates have become dependent on political advertisements as a means to present themselves to potential voters. Political advertisements allow audience members to learn about the substantive material as well as the superficial imagery that the candidate presents. But what type of information is processed more by audience members -- substantive material or superficial imagery? This study investigates the processing of superficial imagery and substantive material in political advertisements and attempts to extend the Heuristic-Systematic Model, a cognitive processing theory, to include other concepts of involvement. The processing of superficial imagery and substantive material in both sofi and hard message advertisements presented through the radio and television mediums are examined. Superficial imagery refers to the personal characteristics of the candidate, and substantive material refers to the issues that the candidate presents. Political involvement is examined as a predictor of processing superficial imagery and substantive material. Results indicate that with higher levels of political involvement, more substantive material is processed, but that low involvement individuals do not necessarily process more superficial imagery. In general, both low and high involvement individuals process equal amounts of information between soft and hard message conditions. Further, more superficial imagery is processed from television than from radio advertisements. To my parents, Shirley and James McKeon, and my husband, Bruce Lillie, for whose continual love, support, and understanding I am forever gratefiil. iii ACKNOWLEDGMENTS There are many people I would like to thank for their contribution to the final stages of my graduate education. First I would like to thank Dr. Charles K. Atkin, my knowing and wise advisor and chair, who continually offered his insight, knowledge, and time on numerous drafis. Truly, without his efforts, this publication would not have been possible. I would also like to thank my other committee members, Dr. Erwin P. Bettinghaus and Dr. Charles T. Salmon whose comments became invaluable contributions to this project. Your time and effort is appreciated. In addition, I would like to thank Dr. Kim Witte, not only for her contribution as a committee member, but also for her guidance as my research team advisor for my first two years of my doctoral program. Further, I am very thankful for all the special friendships I have made through the graduate school experience. Most of all, I would like to acknowledge the encouragement and support of my parents, Shirley and Jim McKeon, who have always been there for me in all my endeavors. Also, I would like to thank a very special person who I met in a graduate course a few years ago -- my wonderful husband Bruce. My greatest thanks go to him for his understanding, love, patience, insight, stress-reducing shoulder massages, and most importantly -- his good sense of humor. iv TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES CHAPTER I INTRODUCTION CHAPTER 2 POLITICAL ADVERTISEMENTS The Influence of Radio and Television Advertisements on Television Varying the Audio and Video Channels Superficial Imagery and Substantive Material Involvement CHAPTER 3 COGNITIVE PROCESSING Introduction Capacity Model of Attention Elaboration Likelihood Model Heuristic-Systematic Model Summary of Research Hypotheses CHAPTER 4 METHODS Overview Phase I: Pretest of Stimulus Materials Phase 2: Data Collection CHAPTER 5 RESULTS Overview Profile of Research Participants Profile of Political Candidate Hypotheses for All Advertisements Hypotheses for Sofi Message and Hard Message Advertisements Hypotheses for Television and Radio Advertisements vii viii 00 12 IS 18 18 I9 22 24 3O 32 32 33 36 50 50 50 SI 53 58 62 Research Question Regarding Candidate Evaluation Additional Analyses CHAPTER 6 DISCUSSION Overview Political Involvement and Recall with All Advertisements Political Involvement and Recall with Soft and Hard Messages Recall with Television and Radio Advertisements Candidate Evaluation Males and Females Implications of the Study Limitations Directions for Future Research APPENDIX A: Text for Advertisements APPENDIX B: Round 1 - Pretest of Stimulus Materials APPENDIX C: Round 2 - Sofi and Hard Messages APPENDIX D: Measurement Instrument APPENDIX E: Coding Categories for Thought-Listing Items APPENDIX F: Factor Loadings for Political Involvement APPENDIX G: Factor Loadings for Substantive Material APPENDIX H: Factor Loadings for Total Recall LIST OF REFERENCES 65 67 72 72 72 74 76 77 78 8O 83 85 88 89 90 98 I O3 106 l 07 I O8 109 LIST OF TABLES Table l - Means and T-test Results for Compassion, Caring, Hearted, Rational, Sympathy, and Soft Table 2 - Descriptive Data on Semantic Differential Scales Table 3 - Means for Total Recall, Superficial Imagery, and Substantive Material Table 4 - Means and T-tests for the Two Groups of Semantic Differential Scales Table 5 - Means and T-tests for Did Not Vote and Voted on Recall of Total Information, Recall of Superficial Imagery, Recall of Substantive Material, and Political Involvement Table 6 - Means and T-tests for Males and Females on Recall of Total Information, Recall of Superficial Imagery, Recall of Substantive Material, and Political Involvement vii 42 52 59 66 68 70 LIST OF FIGURES Figure l - The Soft Messages - Hard Messages Continuum viii I3 Chapter 1 INTRODUCTION When candidates present their positions on issues in political advertisements, they also convey information about their image regardless if this image information is intended or not. Given that elements of both image and issues are presented in advertisements, questions concerning what audience members recall from such advertisements are important. When watching political advertisements on television or listening to them over the radio, what is recalled more -- the superficial imagery, such as personal characteristics of the candidate, or substantive material, such as the issues presented by the candidate? Additionally, through what types of messages is superficial imagery and substantive material recalled, soft messages, backed by the candidate’s softer rationale of caring and understanding, or hard messages, backed by the candidate’s harder rationale of good policy and tough decision making? Although previous research indicates that message variables as well as individual differences influence audience members’ recall of political advertisements, researchers do not know exactly which factors determine whether audience members will cognitively process the issues (substantive material) that the candidate presents or the personal characteristics (superficial imagery) that the candidate possesses. Previous research has examined similar aspects of the substantive material versus superficial imagery question (often termed “issues versus image”) with a variety of theories. Some of these theories (‘0 include, uses and gratifications approach (Garramone, I983, I985), attraction and expectancy theory, rational appraisal model, ethology (Lanzetta, Sullivan, Masters & McHugo, 1985), schema theory, (Biocca, 1991), mood congruency and assimilation- contrast theory (Basil, Schooler & Reeves, 1991). But, these studies still do not illuminate which factors determine whether audience members will cognitively process the substantive material that the candidate presents or the superficial imagery that the candidate represents. Recent history has also indicated that a candidate’s image can become an issue in a campaign. In the 1996 presidential election, while candidate Dole was trying to make alliances with voters by stating that he was the bridge to the past, the incumbent, President Clinton, was making the claim that he was building the bridge to the fixture (Gibbs & Duffy, 1996). Dole’s image of being old was confirmed by many people as he continued his promise of being the bridge to the past, whereas Clinton developed the image of being able to lead the country into the next century. The images of these two candidates grew strong in the minds of the voters and soon became campaign issues. Although there are cases of a candidate’s image becoming a campaign issue, this study attempts to separate image from issue and addresses the question of what audience members cognitively process more of, superficial imagery or substantive material, and through which type of advertisements, soft messages or hard messages, this processing most occurs. Additionally, this study investigates the degree to which individuals’ political involvement motivates them to engage in the processing of superficial imagery and substantive material. To begin the investigation, this study compares the audience’s recall of superficial imagery and substantive material as presented by the candidate in political b) advertisements. Recall for soft and hard messages presented through radio and television channels is examined. In order to make valid comparisons, political advertisements for this study have been constructed so that soft and hard message advertisements present both superficial imagery and substantive material. In the soft message advertisements, the candidate presents a caring, compassionate, and understanding image, and the issues are packaged with a “caring” rationale. In the hard message advertisements, the candidate presents a tough decision-maker image, and the issues are packaged with a “good policy” rationale. Additionally, more traditional, and seemingly softer, verbs (e.g., support, favor) are used in the soft messages and more action-oriented verbs (e.g., advocate, enact legislation) are used in the hard messages. Based on the differences in image, rationale, and verbs between the soft and hard messages, soft message advertisements reflect superficial imagery and hard messages reflect substantive material. Soft messages have been developed to accentuate the superficial imagery of the candidate, given the “caring” image, “caring” rationale and use of traditional, softer verbs. Hard message advertisements have been developed to accentuate the substantive material presented by the candidate, given the “tough decision-maker” image, “good policy” rationale and the use of the action-oriented verbs. Thus, the soft message advertisements essentially convey more superficial imagery of the candidate and the hard message advertisements convey more substantive material presented by the candidate. For a complete review of the messages, please see Appendix A. Superficial imagery is defined as the personal characteristics (i.e., caring, friendly, sincere) that audience members perceive the candidate to have. The processing of superficial imagery is indirectly measured by how many personal characteristics are recalled by individuals exposed to political advertisements. In this study, processing of the candidate’s superficial imagery is indicated by a cluster of traits (characteristics) that reflect the candidate as a politician as well as a person (Nimmo & Savage, 1976; Shyles, 1984). These characteristics can reflect the candidate’s personality, presentation style, physical appearance, as well as how the candidate packages the issues that he or she presents (i.e., “caring” rationale or “good policy” rationale). Such traits can be derived fi'om the candidate’s nonverbal communication (e.g., gestures, physical stance) as well as verbal communication (e.g., articulate speech, content of speech). Superficial imagery is measured based on the number of these such personal characteristics recalled about the candidate. Substantive material is defined as the issues that the candidate presents. The processing of substantive material is indirectly measured by how many issues are recalled by individuals exposed to political advertisements. The processing of the candidate’s substantive material in this study is indicated by the issues or policies advocated by the candidate. Substantive material can be derived from candidate’s verbal communication (e.g., what the candidate says in terms of issues or policies). Substantive material is measured based on the amount of correct information recalled about the issues and policies presented, such as what ideas the candidate supports or what the candidate proposes to do. Theoretical guidance for this study is mainly derived from one cognitive processing theory, Heuristic-Systematic Model, and is applied to predict audience members’ recall of superficial imagery and substantive material presented in advertisements. Applying the general tenets of this theory to this political setting will yield a better understanding as to why some individuals cognitively process more superficial imagery of the candidate, while others cognitively process more substantive material presented by the candidate. In order to begin research on the recall of superficial imagery versus substantive material, Chapter 2 reviews political advertisements, superficial imagery and substantive material, as well as the concept of involvement. In Chapter 3, the Heuristic-Systematic Model, along with two other cognitive processing theories, Capacity Model of Attention and Elaboration-Likelihood Model, are examined and hypotheses are derived. A thorough description of the methods implemented for conducting the study, such as procedures and measures, is presented in Chapter 4. Chapter 5 presents the results of the analyses. The final chapter, Chapter 6, concludes with a discussion of the findings and offers implications for the findings discussed. Chapter 2 POLITICAL ADVERTISEMENTS The Influence of Radio and Television The first major historical event illustrating that images portrayed on television can be more powerful than the words spoken was the first “Great Debate” of 1960. Those who watched on television the first Kennedy-Nixon debate of 1960 thought that Kennedy had won the debate, however, those who had tuned into the debate from their radios believed that Nixon had won (Jamieson & Birdsell, I988; Wykoff, 1968). Both the radio and television audiences, not only Ieamed how well the two candidates could perform in a debate, but they also formed images of each of the candidate’s ability and character (Katz & Feldman, 1962). The difference between viewers’ cognitively processing the substantive material versus the superficial imagery lies with the powerfirl images that can be viewed via television because “viewers are more disposed to respond to the impression created by a televised message than its substance, by the pictures it conveys rather than the words” (Jamieson & Birdsell, 1988, p. 183). Viewers of the Kennedy-Nixon debate saw Nixon as a pale, thin, and shifty-eyed man, who was wearing a suit that was too big, and who shifted his weight too often while standing at the podium. Even his make-up failed to cover up his five o’clock shadow (Jamieson & Birdsell, 1988). Although all of these factors have reasons which do not indicate specific character flaws (fiom a recent hospital stay he had lost weight, leaving him pale, thin, thus wearing a suit too big; he kept glancing at a wall clock off stage resulting in shifty-eye movements; and he had injured his knee while getting out of his limousine causing him to favor one knee over the other), they were detrimental to the impression left with viewers concerning his overall character. In comparison, Kennedy looked sharper, giving the impression that he was more firm, in control, and could stand up to Khrushchev, the current leader of the USSR (Wykoff, 1968). To the television viewers, the television images told all, Kennedy clearly triumphed over Nixon. However, for those listening to the debate on the radio, they had a completely different impression. From the Southern Governors’ Conference which he had been assigned to cover, Earl Mazo (1962) wrote about the reaction in the room firll of governors, “Before the encounter on radio was half finished, every Kennedy partisan in the room was disparaging the idea of a fine, upstanding young man like Senator Kennedy having to clash verbally with a crusty old professional debater like Vice President Nixon. But the attitude changed immediately when the magic lantern of television came on,” (p. 6). Based on the radio presentation, Nixon’s deep resonant voice conveyed more command, conviction, and determination to those governors listening than the high- pitched voice with the Boston-Harvard accent that Kennedy’s voice possessed (Mazo, 1962). There were clear differences of opinion between television viewers and those listening from their radios as to which candidate prevailed in the first debate. By comparing the impressions of the “Great Debate” between the radio audience and the television audience, it is clear that television transforms political candidates into personal images (Wykoff, 1968). Over the years, the use of television as a political presentation tool has increased dramatically. Today, viewers are exposed to candidates through a variety of televised formats, including town hall meetings, debates, infomercials, news stories, interviews, talk shows, entertainment shows, and political advertisements. From these formats, political advertisements are the best choice for investigating the superficial imagery versus substantive material question because their structure of presentation allows for testable comparisons to determine similarities and differences. Many of the other formats do not provide such a clean process of comparison as they are not controlled presentations produced by the candidate. More specifically, positive political advertisements offer the best means of comparison given that they present the candidate and the candidate’s ideas, they do not present the opponent’s position, and they do not counter-argue the opponent’s positions. Additionally, positive political advertisements share many similar qualities such as length, source presenter, product (the candidate), and message content. Given the similarities among positive advertisements in general, constructing positive political advertisements for this study offers a good means through which to examine the cognitive processing of superficial imagery and substantive material. Advertisements on Television Impressive increases in campaign costs that are devoted to televised political advertising suggest that political advertising on television has an increasingly important role in the American political system (Shyles, 1986). Starting with some of the very first advertisements in 1952, a look at the political messages that have aired emphasizes the wide variety of possible styles. Dwight Eisenhower and Adlai Stevenson were the presidential candidates in 1952, the first year that political television commercials aired (Thorson, Christ & Caywood, 1991). One of Eisenhower’s televised political messages was a cartoon. That cartoon showed circus animals parading in a single row with an Eisenhower banner that read “We like Ike,” and as the animals marched along they were singing “You like Ike. I like Ike. Everybody likes Ike.” Later, some criticism was made of the advertisement in that the real role “was selling the President like toothpaste” (Mayar, 1958, p. 302). Eisenhower also had a series of advertisements appropriately called “Eisenhower Answers America.” Paid for by Citizens for Eisenhower, these advertisements showed a typical citizen asking a question of President Eisenhower. An example of one such advertisement is illustrated in Diamond and Bates (1992): a middle-aged woman asks Eisenhower, “The Democrats have made mistakes, but aren’t their intentions good?” Eisenhower responds, “Well, if the driver of your school bus runs into a truck, hits a lamppost, drives into a ditch, you don’t say his intentions are good; you get a new bus driver” (p. 56). Although Eisenhower probably would have won a second term regardless of the advertisements, his campaign raised the first major questions about politics, advertising, and television, such as: Do advertisements ignore issues and content in order to present images and emotion? and Does the best candidate win, or the most telegenic performer? 10 (Diamond and Bates, 1992). The question of what viewers recall in terms of superficial imagery or substantive material continues to be very important as use and variety of television for political advertisements increases. Critics have often claimed that the content in televised political advertisements works with the inherent characteristics of television to highlight the candidate’s superficial imagery at the expense of gaining information on the candidate’s substantive material (Geiger & Reeves, 1991). The discrepancy between learning about substantive material versus learning about superficial imagery lies with the powerfirl images that can be viewed via television because Viewers are more disposed to respond to the impression created by a televised message than its substance, by the pictures it conveys rather than the words” (Jamieson & Birdsell, 1988, p. 183). It has been noted that personalities of the candidates are becoming more important as party identification is becoming less important (Lanzetta, Sullivan, Masters & McHugo, 1985). Television may tend to present emotional responses to viewers thus facilitating the shift from substantive material to superficial imagery. During a political campaign, viewers see the candidates on television daily. Given its nature, television provides in- depth information on political candidate’s facial images at the same time it presents the verbal messages. These nonverbal cues are extremely important as they are known to communicate emotion as well as provide information for viewers to construct trait evaluations (Lanzetta, Sullivan, Masters & McHugo, 1985). However, in some cases, previous research has found that both candidate qualifications and issue stands compose the message content that is most widely Ieamed 11 among viewers (Atkin, Bowen, Nayrnan, & Sheinkopf, 1973). Information gained on the substantive material presented in a political advertisement is most likely to be gained through the audio channel of the advertisement; impressions of the candidate’s superficial imagery are more likely to be created from both the audio and video channels (Garramone, Steele & Pinkleton, 1991). More specifically, it has been found that when political candidates are shown on television, viewers tend to use the pictures to judge the candidate on personality traits (Graber, 1987). It has also been found that viewers primarily focus on image formation (Glass, 1985) and use nonverbal information as a key factor in assessing a candidate (Rosenburg & McCafferty, 1987). This television medium is unique in that it offers the opportunity for candidates to create messages primarily of content while displaying images of the candidate through the combined use of audio and video channels. mm Audio M Video Chan_ne_l_§ As previously mentioned, audience’s impressions of a candidate’s image can be derived fiom both the audio and video channels. Given this, the four advertisements created for this study will differ in terms of their audio and video channels. While the content and amount of information, in terms of the issues presented in the advertisements will be the same, the candidate’s rationale of the issues differs, thus varying the audio channel. This occurs, such that in the soft message advertisements the candidate packages the issues with the general notion of “caring,” and that in the hard message advertisements the candidate packages the issues with the general notion of “good policy.” The “caring” rationale is composed of phrases such as “I understand the value of strong l2 families,” “I have compassion for those who are terminally ill,” and “I care about you and your community.” The “good policy” rationale is composed of phrases such as “I believe in the constitutional right of personal choice,” “it’s a practical means of maintaining the family unit,” and “I’m the candidate who can make tough decisions and solve tough problems.” Further, the video channel will also vary, such that one set of advertisements will be presented over the television channel (using both audio and video channels) while the other set of advertisements will be listened to over the radio channel (using only the audio channel). In addition to the “caring” rationale which accentuates the superficial imagery (verbaIIY), other image characteristics can be derived from the visual image of the candidate. Thus, these advertisements will have superficial imagery depicted through both the verbal packaging of the issues as well as the video channel offered through television. Superficial Imagery and Substantive Material The concepts of superficial imagery and substantive material can be firrther understood when considering their place on a continuum anchored by soft and hard messages. At the soft message end, there exists the most discernible soft image characteristics (e.g., kind, caring, smart, good looking). At the hard message end, there exists the most discernible ideas or issues proposed (e.g., advocates physician-assisted suicide, wants to enact legislation of benefits for domestic partners). Although containing aspects of both superficial imagery and substantive material, the advertisements created for this study can be separated into two sets, such that one set with the “caring” rationale more closely approaches the soft message end, while the other set with the “good policy” rationale more closely approaches the hard message end. The soft message advertisements are considered to approach the soft message end because the “caring” rationale along with the softer verbs invokes more image impressions about the candidate as it plays down the issues. The hard message advertisements are considered to approach the hard message end because the “good policy” rationale along with the action- oriented verbs is intended to work in conjunction with the issues to more strongly present the issues advocated by the candidate. The placement of the two sets of advertisements on the soft message-hard message continuum is presented in Figure 1. Caring Good Policy Rationale Rationale l l T l Soft Messages Hard Messages Figure l - The Soft Messages - Hard Messages Continuum Given the experimental manipulation considerations concerning this study, the advertisements have been created so that they are as close as they possibly can be to the soft and hard message ends of the continuum. Although these sets of messages are not as extreme as the continuum indicates, they are as different as possible considering that the issues, sentence structure, and the amount of information presented in the advertisements are the same. In order to achieve messages that represented the extreme ends of the l4 continuum, a number of message aspects would have to be manipulated. Given internal validity considerations, only the image and rationales were manipulated so that research findings pertaining to the differences between soft and hard messages would be valid. When comparing hard message advertisements with soft message advertisements, it is noted that emphasis placed on personality and ability in soft message advertisements is intended to arouse affective responses that elicit positive feelings fiom the audience members to the candidates (Geiger & Reeves, 1991). For hard message advertisements, messages are to be understood and not felt; they evoke more rational responses, such as judging candidate on their positions (F ishbein & Ajzen, 1981). From this it seems reasonable to assume that more positive feelings are elicited to the candidate when audience members recall aspects of soft message advertisements opposed to hard message advertisements. However, previous research has illustrated that hard message advertisements produced significantly more positive evaluations of the political candidate than did soft message advertisements (Kaid & Sanders, 1978). Other studies have also found that hard message advertisements elicit more positive evaluations of the candidates along with more elaborate representations of the candidates compared to soft message advertisements (Conover, I981; Geiger & Reeves, 1991). One explanation for this is that positive evaluations might be given due to the mere presentation of issues presented in the hard message advertisements (Geiger & Reeves, 1991). Given that the same issues are presented in both sets of advertisements (hard and soft messages), it is reasonable to expect that the soft message advertisements will result in more positive evaluations 15 concerning the candidate than will the hard message advertisements. In the present work, the evaluations of the candidates as well as recall of superficial imagery and substantive material will be investigated. Involvement Researchers have often used the concept of involvement when attempting to understand the conditions under which individuals are persuaded by others (Johnson & Eagly, 1989). Even though there has been considerable agreement among researchers that the concept of involvement is important, there has been considerable disagreement across research domains as to what involvement is (Salmon, 1986). Over many research studies, involvement has been considered in many different perspectives. Given that involvement has been studied in numerous ways, it is necessary for researchers to specify what kind of involvement is being studied in order to best understand the results. Some of the perspectives of involvement discuss involvement as an internal state (Pedersen, 1978; Tan, 1980), as a response to a stimulus (Petty and Cacioppo, 1979; Petty and Cacioppo, 1981b), and as a personality trait (Kassarjian, 1980). Within each of these perspectives there are a variety of ways that involvement is conceptualized. For example, the concept of ego-involvement was developed and brought into persuasion literature as a major role in a theory of attitude change, the social judgment-involvement approach (Hovland, Harvey & Sherif, I957; Sherif, Sherif& Nebergall, 1965; Sherif & Hovland, 1961). An individual’s self-concept is the basis for the research on ego-involvement. More specifically, ego-involvement is composed of the ego-involved attitudes constituting an individual’s self-picture (Salmon, 1986). Social judgrnent-involving theorists proposed 16 that attitudes provide an internal flame of reference for reacting to and judging stimuli related to the attitude (Sherif, Sherif & Nebergall, 1965; Sherif & Hovland, I961; Sherif & Sherif, 1967). Further, as suggested by Sherif and Cantril (1947), attitudes are consistent with one’s ego, therefore, an individual will become personally involved when such attitudes are situationally evoked. Given this perspective of involvement, ego-involvement is considered to be an internal state and is based upon a subject-object interaction principle (Salmon, 1986). Further, another perspective considers involvement as a condition that has “fixture consequences” for the individual (Salmon, 1986). For this perspective, involvement is being used as a stimulus rather than an internal condition of the individual. Studies using involvement as such usually experimentally manipulate involvement such that they present a topic of personal relevance or importance to the individual (Petty & Cacioppo, 1981b; Petty, Cacioppo, & Goldman, 1981). This is a different type of involvement than what is suggested by Sherif and Cantril (1947). Although involvement is still examined as an interaction between the individual and the stimulus, in the “future consequences” use it is defined as the importance of an issue opposed to the individual’s interest in the issue (Salmon, 1986). Political Involvement. For this study an individual’s political involvement is assessed. The type of involvement used in this study to measure political involvement is most closely associated with the concept of ego-involvement. Political involvement is constructed with two dimensions in mind: behavioral and psychological. Items tapping the behavioral dimension of political involvement inquire about one’s participation in political 17 activities such as volunteering on a campaign, reading newspaper stories about the campaign or candidate, or discussing political candidates with fiiends in hopes of persuading them to vote in favor of one candidate over the other. Items tapping the psychological dimension of political involvement inquire about one’s attitude or interest toward political activities such as having concern for the outcome of the campaign or believing that an understanding of party issues is worthwhile. Chapter 3 COGNITIVE PROCESSING Introduction This section presents three cognitive processing theories that are used as a basis for the predictions concerning the recall of substantive material and superficial imagery presented in the advertisements. Of the three, one main theory, Heuristic-Systematic Model, is used to generate hypotheses concerning political involvement and the recall of superficial imagery and substantive material. The other two theories (Capacity Model of Attention and Elaboration Likelihood Model) serve as additional theoretical support for assumptions made about political involvement and the processing of information presented in the advertisements. The Heuristic-Systematic Model (HSM) discusses heuristic and systematic processing in terms of processing “content cues” and “noncontent cues.” This study adapts these general information processing and message terms to fit the purpose of this study which is to determine differences in recall of substantive material information and superficial imagery information presented in political advertisements. Given how the advertisements were constructed for this study, the assertion is made that the concept of candidates’ substantive material presented in the advertisements theoretically corresponds to the HSM’s term “content cues,” and that the concept of candidates’ superficial imagery presented in the advertisements theoretically corresponds to the HSM’s term “noncontent cues.” l8 19 Although aspects of both substantive material and superficial imagery are in the advertisements, the soft message advertisements are designed to produce more impressions of the candidate’s image due to the use of softer verbs and a “caring” rationale. Whereas, the hard message advertisements are designed to produce more impressions conceming the issues presented by the candidate due to the use of action- oriented verbs and the “good policy” rationale. Each of the guiding theories is presented individually with their specific terms, however, the hypotheses are derived from the general tenets of the HSM and incorporate the political concepts of substantive material (as relating to “content cues”) and superficial imagery (as relating to “noncontent cues”). Capacig Model of Attention (CMA) Introduced by Kahneman in 1973, the CMA offers a parallel processing view for how individuals attend to objects and acts. This model suggests that individuals can process different stimuli simultaneously as opposed to focusing on one stimulus and leaving the other ignored or secondarily attended to. Kahneman maintained that within the bounds of a fixed upper limit, the amount of processing capacity an individual has available is variable and is a function of task difficulty. Therefore, individuals processing tasks of greater difficulty exert more effort, and as a result, have more available capacity compared to individuals who are engaged in processing simple tasks. Kahneman, Peavler and Onuska’s (1968) study (examination of short term memory and more difficult addition tasks with monetary rewards for good performance) showed that processing capacity expands as a function of the difficulty of task, and in some cases as a firnction of involvement. 20 Kahneman (1973) suggested that individuals allocate their capacity between primary and secondary processing tasks. Such that, the amount of capacity used for processing the primary task approaches total capacity allowing for only a little “spare” capacity for secondary information processing when there are higher levels of primary task demand. Further, the amount of capacity utilized for processing the primary task is considerably less than the total capacity allowing for more “spare” capacity for the individual to process secondary information simultaneously when there are lower levels of task demand. Both processing capacity and attention are directed toward the primary cues. After the primary cues have received their allocation of capacity, the remainder processing capacity is the available for processing secondary cues (Kahneman, 1973). Given that the capacity model assumes that the total amount of attention which can be distributed at any given time is limited, there are two key observations to keep in mind. First, variations in arousal levels of individuals are affected by variations in the task difficulty. Second, individuals’ ability to process several mental activities simultaneously depends on the effort each of the activities requires when performed in isolation. At low levels of primary task demand, secondary tasks can be processed in conjunction with the primary task as allowed by the spare capacity. At high levels of primary task, the individual must select which single task to focus on as the level of spare capacity is completely diminished (Kahneman, 1973). Different mental activities require different amounts of attention allocation on an individuals’ limited capacity. Easy tasks require little effort, whereas a difficult task requires more effort. An activity can fail, according to the model, because the demands on 21 attention exceed the available capacity, or because the allocation policy distributes the available capacity to other activities. Further, an activity can also fail if the input of relevant information is insufficient to attend to. Involvement is a key factor in the CMA in terms of how messages are cognitively processed. Individuals are not motivated to process many persuasive message cues at low levels of involvement. Therefore, the capacity for processing the primary task, whether the processing is toward content or noncontent cues, is minimal allowing for considerable “spare” capacity for secondary cues. With low involvement, individuals are not motivated to process either content or noncontent cues to a great extent, both types will receive minimal attention, subsequently having minimal influence on attitudes (Stiff, 1986; 1994). At moderate levels of involvement individuals are capable of processing both primary and secondary cues simultaneously as the total capacity demand has not exceeded available capacity. Compared to low involvement, these cues receive more attention and become more effective at influencing attitudes (Stiff, 1986; I994). Lastly, at levels of high involvement, individuals are forced to choose between processing content cues or noncontent cues because the capacity demanded for primary cue processing approaches total capacity. It is predicted that such individuals will select to process content cues opposed to noncontent cues due to their increased utility for making accurate decisions. Such highly involved individuals will be influenced by the content cues not the noncontent cues (Stiff, 1986; 1994). Ebflgion Likelihood Model (ELM) Petty and Cacioppo (1981a) presented a model of persuasion that assumes that individuals differ in their “elaboration” of persuasive messages, and that these differences are determined by whether individuals take the central or peripheral route to persuasion. “Central route” to persuasion was termed as the careful scrutiny of the content of the message. From this they posited that attitude change is a firnction of message content and elaboration. Additionally, they termed “peripheral route” to persuasion as the association of message recommendations with positive or negative cues in the environment. Attitude change through the peripheral route to persuasion is based on an individual’s assessment of these noncontent cues. The “central route” to persuasion occurs when elaboration is relatively high. Such individuals engage in extensive issue-relevant thinking, such that they carefirlly examine the content of the message, the arguments provided and consider other issue-relevant material, such as additional arguments from memory or self-created arguments derived from their exposure to the message. The “peripheral route” to persuasion occurs when elaboration is relatively low. Individuals pursuing the peripheral route typically use simple decision rules when considering the persuasive message. Rather than engaging in issue- relevant thinking, such individuals may rely on whether or not they liked the message source, or whether they thought the source was credible or trustworthy (Petty & Cacioppo, 1986). The more that individuals engage in issue-relevant thinking, the more they are elaborating, thus, the more they are evaluating the content of the message and assessing to b) the arguments presented in the message. The Elaboration Likelihood Model posits that persuasion can occur at any point on the elaboration continuum, but as the elaboration varies, the persuasion occurring will as well (Petty & Cacioppo, I986). The differences in the persuasion process are a result of which route to persuasion is taken. As with the CMA, involvement is also a key factor for processing information in the ELM. An individual’s involvement in the topic plays a key role with how much elaboration an individual is likely to give, therefore determining whether the central or peripheral route is taken. For studies examining this model, involvement is understood as how personally relevant the topic is to the receiver (i.e., a condition that has firture consequences for a receiver) (Salmon, I986). The theory posits that the more personally relevant the topic is to the receiver, the more motivation that receiver will have in order to engage in issue-relevant thinking. One study tested the notion of involvement by examining the effects of communicator expertise and argument strength on the persuasive effectiveness (Petty, Cacioppo, & Goldman, 1981). Message topic as well as the degree of involvement were varied, thus the receivers who were less involved with the topic would have relatively little elaboration, and the receivers who were more involved with the topic would have relatively high elaboration. The results indicated that individuals in the high involvement condition were significantly influenced by the strong arguments than the weak arguments. However, these same individuals were not significantly affected by the expertise portrayed by the message communicator. For these such individuals persuasion was obtained through the central route. Conversely, individuals in the low involvement condition were less influenced by the strength of the message argument, and were more 24 influenced by the expertise of the message source. For these such individuals persuasion was obtained through the peripheral route. Heuristic-Systematic Model (HSM) Chaiken (1980) examined heuristic and systematic information processing, although these processes are conceptually different, they both assess the validity of the conclusion in the message. With regard to the systematic view, message receivers not only actively attempt to understand and evaluate persuasive arguments, but they also determine the validity of the persuasive arguments in relation to the conclusion of the message. With systematic information processing, considerable cognitive effort must be used in order to accomplish this task; systematic processing demands uses cognitive capacity to assess content cues. With regard to the heuristic view, message receivers rely on the most accessible information (e.g., general attitude toward the source) in order to determine whether or not to accept the conclusion of a message. This information which is most accessible consists of noncontent cues, such as message or source characteristics (e.g., too lengthy or poor speaker). Contrary to the effort exerted for systematic processing, heuristic processing requires substantially less effort, demanding little cognitive capacity. Even though these are two distinct ways of processing information, there exists a concurrent processing assumption, such that systematic and heuristic processing can exert both independent and interdependem effects on attitude judgments (Eagly & Chaiken, 1993). Systematic processing focuses on detailed processing of the persuasive content of the message, whereas, heuristic processing focuses on accessible information to judge 25 acceptance of the message. Even though the primary focus for systematic processing is on the content of the message, it is possible that secondary attention is also given to noncontent cues. For individuals processing systematically, message factors will always have greater impact on persuasion on the message then source factors. For individuals processing the persuasive message heuristically, source factors have a greater impact on persuasion of the message than messages factors. For heuristic processing simple rules and schematics can mediate individuals’ attitudes (Eagly & Chaiken, 1993). The theory posits that systematic processing should be constrained more by situational demands (e.g., not enough time to process the message), and by individual differences (e.g., previous knowledge may facilitate processing), than heuristic processing as these constraints have the potential to reduce individuals’ ability for processing detailed information. However, heuristic processing is not effected by either of these variables. As with the other models, involvement is also a key factor in the cognitive processes associated with the HSM. In Chaiken’s (1980) work, involvement pertaining to “firture consequences” was investigated as a moderating variable with systematic and heuristic processing. In general, when considering individuals’ involvement with the message topic, under conditions of low involvement individuals utilized heuristic strategies of processing the persuasive messages, whereas, under conditions of high involvement individuals utilized systematic strategies of processing persuasive messages (Chaiken, 1980). More specifically, when high versus low consequences were manipulated, individuals with high consequences had greater opinion change in responses to messages backed by six arguments and were unaffected by the likability of the message source. 26 Additionally, individuals with high consequences compared to individuals with low consequences spent more time reading the persuasive message and spent more time thinking about the persuasive arguments presented by the source opposed to the source’s characteristics. High consequences individuals were able to recall more arguments as well as generate more message-oriented thoughts opposed to thoughts concerning the source characteristics. Further, individuals with low involvement had greater opinion change with the likable message source and were not affected by the number of arguments used in the message. When personal relevance was manipulated, individuals with high personal relevance indicated slightly greater opinion change with persuasive messages comprised on five arguments (from an unlikable source) than with persuasive messages with one argument (from a likable source). However, individuals with low personal relevance indicated significantly greater opinion change when receiving messages with one argument from a likable source than messages with five arguments from an unlikable source. The HSM hypothesizes about the “personal relevance” or “future consequences” type of involvement. However, given that the concept of political involvement used in this study is most similarly associated with ego-involvement, whose effects have not been thoroughly tested with HSM or the other cognitive processing theories, it is necessary to state that testing political involvement with the HSM is expanding the use of the theory to other concepts of involvement. Hypotheses are formulated based on the general involvement predictions of the HSM even though the use of involvement is conceptually different fiom that of political involvement. Essentially, by using political involvement, the HSM will be tested with a different concept of involvement, if the data support the 27 hypotheses, then this will be a start concerning the application of this theory and possibly others to include ego-involvement. Theoretical predictions concerning the processing of content cues and noncontent cues are adapted from the HSM to predict the processing of substantive material and superficial imagery as recalled by audience members. Concerning political involvement and the HSM, high involvement individuals, theoretically, will be motivated to primarily focus on the content of the messages, thus, they will process the advertisements through systematic processing. For low political involvement individuals, they will process the advertisements through heuristic processing because source factors (noncontent cues or superficial imagery) have a greater impact on recall than message factors (content cues or substantive material). When applying the basis of the HSM to political advertisements, it is expected that, high involvement individuals, given that their main focus of processing is on substantive material (content cues), will process more substantive material fiom the advertisements than low involvement individuals. Further, it is expected that, low involvement individuals, given that their main focus of processing is on superficial imagery (noncontent cues), will process more superficial imagery than high involvement individuals. Therefore, for all advertisements, the following hypotheses are derived: H1: Individuals with higher levels of political involvement will recall more substantive material than individuals with lower levels of political involvement. 28 H2: Individuals with lower levels of political involvement will recall more superficial imagery than individuals with higher levels of political involvement. Additional hypotheses can be derived when considering the HSM with the theoretical concept of soft and hard messages. For high involvement individuals, it is expected that they would recall about the same amount of information fiom both soft and hard messages because their higher level of involvement would motivate them to systematically process as much information as possible, regardless of the message condition. High involvement individuals in both the hard and soft message conditions will have their primary focus on the content of the message (both message conditions have the same amount of issue information) and will also have secondary attention on the noncontent cues or superficial imagery. Thus, high involvement individuals will process both substantive material and superficial imagery from both soft and hard message conditions because of their high motivation. This explanation applies equally to both substantive material and superficial imagery so that the focus of this hypothesis concerns total information recalled. For the low involvement individuals, they should be more motivated to heuristically process noncontent cues or superficial imagery as they focus on the source factors more so than the message factors. As a result, low involvement individuals should process more total information from the soft messages, which have been constructed to have more superficial imagery, than from the hard messages, which have been constructed 29 to have more substantive material. Given this, there is a contingent interaction effect predicted, such that: H3a: For high political involvement individuals, recall of total information (substantive material and superficial imagery) will be about the same for the hard message condition and the soft message condition. H3b: For low political involvement individuals, recall of total information (substantive material and superficial imagery) will be greater in the soft message condition than in the hard message condition. When considering the HSM and the television and radio advertisement conditions additional hypotheses are also derived. Given that the television advertisements visually convey more superficial imagery, it is expected that more superficial imagery will be processed fi'om the television advertisements than from the radio advertisements. As there is less superficially imagery conveyed in the radio advertisements, individuals will give more attention to the substantive material presented. Therefore, it is expected that more substantive material will be processed fiom the radio advertisements than fiom the television advertisements. Given this, the following hypotheses are presented: H4a: Both high and low political involvement individuals will recall more superficial imagery from television advertisements than from radio advertisements. H4b: Both high and low political involvement individuals will recall more substantive material from radio advertisements than fiom television advertisements. Finally, given the different findings of candidate evaluations from previous studies, one research question concerning the evaluations of the candidates must be asked: RQI: In general, will there be more positive evaluations of the candidate resulting from the soft message advertisements or from the hard message advertisements? Summgy of Research Hypotheses The foregoing arguments suggest a foundation for cognitive processing of information in political advertisements in which one’s level of political involvement will determine one’s recall of superficial imagery and substantive material from the political advertisements. Additionally, political advertisements constructed in terms of hard messages (substantive material) and soft messages (superficial imagery) are tested independently as well as together (total information) to determine any differences in cognitive processing. Further, both these hard and soft messages are presented through the television and radio mediums in which processing of superficial imagery and substantive material are tested. In general it is predicted that individuals with high political involvement will process more substantive material and that individuals with low political involvement will process more superficial imagery. Additionally, it is predicted that there is a contingent interaction effect with message condition, such that while high political involvement individuals will process equal amounts of total information in hard messages as compared to soft messages, low political involvement individuals will process more total information from the soft messages as compared to hard messages. Lastly, when making comparisons between the television advertisements and the radio advertisements, it is predicted that, regardless of political involvement, individuals will process more superficial imagery from the television advertisements as compared to the radio advertisements and that individuals will process more substantive material from the radio advertisements as compared to the television advertisements. Chapter 4 METHODS cm This study attempts to examine to what degree individuals cognitively process superficial imagery and substantive material in political advertisements. The methods used in this 2 x 2 x 2 design (political involvement by message type by medium) provide the opportunity to examine the cognitive processing of high and low political involvement individuals for both hard and soft message conditions as well as for television and radio advertisement conditions. This study was conducted in two phases. In the first phase (issue and message pretesting), political issues were pretested in terms of their degree of personal importance. Of the twenty-five political issues initially pretested, five issues for which the degree of importance was fairly evenly distributed (7-point Likert scale between ‘Very important” to “very unimportant”) were selected for further pretesting. The goal of Phase 1 was to detennine which issues were important and then create appropriate messages which polarized the subjects in terms of their personal importance and favorability. The rationale for this polarization was to maximize the probability of achieving differences between subjects. Additionally, the researcher could also eliminate the possibility that the consistently important and favorable issues confounded with treatment effects produced in Phase 2 (data collection). Therefore, by having students’ favorability and personal 32 importance generally polarized between the ends of the two continuums of importance and favorability, the risk of having the specific issue presented in the advertisement diminish treatment impact (by either having the majority of students feel very strongly for or against the issue or having students feel neutral toward the issue) was also minimized. In Phase 2, 233 experimental participants either viewed or listened to one of four political advertisements and then were given a questionnaire to answer based on their reactions to the advertisement to which each participant was exposed. Phase I: Pretest of Stimulus MM Research participants. A total of 81 subjects participated in this phase of the study. In Round 1 (political issue selection) and Round 2 (political message assessment), subjects were students from a large midwestem university. In both Rounds 1 and 2 students were told that a local politician was interested in their reactions to the issues or messages presented. Procedures. In Round 1, subjects (n = 53) were told that a State House Representative was trying to assess which political issues were most important to students on campus. Subjects were asked to evaluate 25 political issues based on the degree of personal importance. This procedure took about five minutes in length. In Round 2, subjects (n = 28) were told that a local politician was interested in their reaction to actual political messages. Five political messages (varying in terms of hard and soft message quality) were presented to each subject. In addition to assessing individual importance and favorability toward each message, subjects also evaluated the 34 messages in terms of specific hard-soft message attributes. This procedure took about 12 minutes in length. Measurement instruments. For Round 1, the survey consisted of 25 political issues which were selected in terms of being either hot local topics (e.g., stricter divorce laws), typical political issues (e.g., comprehensive health care coverage), or specifically student- oriented concerns that could be made into local political issues (e.g., easier access to student financial aid). See Appendix B for a complete list. Students were asked to judge each issue in terms of how personally important that issue is to them. The degree of importance was assessed with a 7-point Likert type scale ( l = “very important” and 7 = ‘yery unimportant”). For Round 2, the survey consisted of five messages (as determined by a polarized distribution of importance assessed in Round 1). In addition to assessing personal importance and favorability toward each message, the messages needed to be examined in terms of their hard and soft message qualities. To do so, three concepts were developed into three sets of semantic differentials (i.e., hard - soft, warrnhearted - cold-hearted, and emotional - rational). The four variations of the messages that were assessed consisted of two manipulations of hard messages and two manipulations of soft messages. Ha_rd and soft messages - Round 2 only. After analyzing the results of Round 1, five issues were selected for message development. For each of these five issues, both hard and soft messages were constructed for firrther testing. Two types of soft messages were created for all five issues. For both variations of soft messages, a “caring” rationale was used by the candidate, such as “Because I care about the residents of the XXXXX 35 area, or because I understand the value of strong families.” In addition to the soft messages having the “caring” rationale, one variation of these messages also used softer verbs and helping verbs, such as “I want to look into ...” and “I want to see ...” The other variation of the “caring” rationale message used typical political advertisement verbs such as W and f_av_or. A firll review of the messages tested is presented in Appendix C. Similarly, the hard messages were also constructed to have two variations of messages. For both variations of the hard messages, a “good policy” rationale was used by the candidate, such as “Because preventative health care saves taxpayers money, and “. . .because such laws are a practical means of maintaining family units.” In addition to the “good policy” rationale, one of the hard message variations also used harder (more action-oriented verbs) verbs, such as, “I’m in favor of instituting...” and “I want to enact legislation...” The other variation of the “good policy” rationale message used typical political advertisement verbs such as sapm and f_av_or. Results. The results of Round 2 indicated that the hard and soft messages using the harder and softer verbs in addition to the “good policy” and “caring” rationale were more distinct from one another than the‘hard and soft messages that used only the “good policy” and “caring” rationale with the typical political advertisement verbs of s_upp9_r_t and fair); Therefore, the following results apply only to the set of messages that have the harder and softer verbs in combination with the to the “caring” and good policy” rationales. When analyzing the responses concerning the degree of importance and favorability, those messages which responses were most evenly distributed were the 36 messages concerning benefits for domestic partners, stricter divorce laws, and support for physician-assisted suicide. Additionally, the results of the semantic differentials indicated that there was only a significant difference between the hard message version and the soft message version concerning the topic of benefits for domestic partners [t = -2.61, df = 12, p = .023 (14)]. However, when considering only two of the three semantic differentials (soft-hard and emotional-rational), the soft and hard messages concerning domestic partners were still significantly different [t = -2.33, df = 12, p = .038 (14)], and the other two messages concerning divorce laws and physician-assisted suicide were closer to achieving a significant difference between their hard and soft message counterparts, for stricter divorce laws [t = -2.03, df = 12, p = .065 (14)]; and for physician-assisted suicide [t = -I .84, df = 12, p = .091 (14)]. However, when considering the small sample size, the results are very encouraging in that these messages appear to represent soft and hard dimensions. Phase 2: Dat_a Collection Researcapanicm. A total of 233 subjects from a large midwestem university participated in this study. At first, students from two undergraduate classes were offered the opportunity to participate in the study. However, due to the initial turnout rate of subjects to their designated participation time (about 60%), three additional classes were approached for their participation in the study. Therefore, a total of five different undergraduate communication classes offered to have their students participate in the study (a total of 380 students were given the opportunity to participate in the study). 37 Students were told that a local politician was interested in their reactions to political advertisements. For all the classes, extra-credit was offered to the students who elected to participate in the study. Further, it was made clear that if a student did not want to participate in the study, he or she could complete an alternative extra-credit assignment. Procedures. Announcements concerning the opportunity to participate in this political advertisement study were made to the five undergraduate classes. Students were told of the general purpose of the study (that a local politician wanted their reaction to political advertisements) and were instructed that if they were interested they could sign up for a time to participate in the study. Sign-up sheets were brought to each class the day of the announcement to facilitate student sign up and were then posted in an easy-to- find location in order to allow students to sign up to participate if they had not already done so. Per each session of the study there was only a maximum of 13 students allowed (this maximum number was allowing for some “no shows” as the desired number per group was 10 - 12). Opportunities to participate were offered through many time slots over a two and a half week period so that students would have the opportunity to choose a time to participate that was convenient for them. The students were also instructed as to which room in the building they must go for the study and that they must arrive on time as the study would promptly start at the designated time. Each of the sessions (groups) for the study were randomly assigned to one of the four treatments. The initial random assignment of the sessions was disrupted on the second day of data collection due to a bomb threat in the building. As there was no way of knowing what time individuals, who were not the trainers of bomb-sniffing dogs, would 38 be allowed into the building again, all of the sessions for that day were canceled. The students who were scheduled for those sessions were telephoned that evening and (based on their preference) were either resigned up during the phone call or were asked to resign up (by visiting the posted sign-up sheets) if they were still interested in participating. Toward the very end of the data collection period, sessions were assigned to treatment groups based on the expected number of participants. This type of calculated assignment of treatment to sessions was executed so that the four treatment groups would be as comparable in size as possible. This method worked very well, three of the treatment groups received 58 participants each and the other treatment group received 59 participants. In total, there were 28 sessions, averaging about 8 subjects per session. As subjects arrived for each session, their names were checked with the sign-up sheet. Any students showing up to a session who were not already signed up were asked to wait, they were allowed to participate if all the slots during that session were not filled. Each session began with briefly informing subjects as to the purpose of the study. Each subject was given a consent form to complete. Consent forms stated that the student’s participation was completely voluntary and that by signing the consent form each student was indicating his or her interest to participate in the study. Subjects were also informed that their responses were completely anonymous and confidential and that the consent form would only be used to issue them their extra credit. Forms were collected when all subjects had finished reading and signing their form. Next, the general procedures of the study were briefly explained to the subjects. Subjects were told that they were to complete the first page of the questionnaire (and only 39 the first page until given firrther instructions), and then wait quietly until all others were finished with the first page (political involvement items). When all subjects had completed the first page, they were either presented with a television advertisement or a radio advertisement. The message was presented to the participants, and then after a short break (about a minute, during which the message was rewound), the message was presented a second time. Presenting the message a second time helped facilitate the recall of issues and also reflected the increased exposure to advertisements which occurs during elections. Subjects were then instructed to complete the rest of the questionnaire (general reactions to the political advertisement) and to once again wait quietly until all others had also finished. After all subjects had completed the entire questionnaire, questionnaires were collected and subjects were debriefed as to the purpose of the study. Stimulus materials. The stimulus materials consisted of four advertisements (two radio advertisements and two television advertisements). In total, two scripts were designed for the advertisements. The two radio advertisements were created from the soundtracks of the two television advertisements. For a complete reading of the advertisement scripts (soft and hard messages), please see Appendix A. After pretesting the messages as described in the Message Pretesting section, the hard message versions of the three selected issues (stricter divorce laws, physician-assisted suicide, and benefits for domestic partners) were combined to form one advertisement, while the soft versions of the same issues were combined to form another advertisement. Each hard and soft advertisement was then given beginning and concluding sentences. Although they were similar in structure, the differences in the beginning and concluding 4O sentences between the hard and soft advertisements were determined by the rationale of the advertisement, such that the soft message (“caring” rationale) received beginning and concluding sentences that indicated the candidate’s caring, while the hard message (“good policy” rationale) received beginning and concluding sentences that indicated the candidate was a tough decision-maker. A video consultant was hired for the filming of the television advertisements and also for creating the radio advertisements fiom the television soundtrack. An individual was selected to act as the candidate for these advertisements. The candidate was selected based on his physical characteristics; the candidate was fi'ee of any physical and verbal quirks that could pose as a distraction to either the viewing or listening audience. In order to minimize differences between the delivery of the hard and soft message, the actor was coached as to his delivery of the two messages. The candidate wore a typical candidate’s attire (dark suit, white shirt and a maroon-colored conservative tie). The candidate was filmed from his chest up and the background used in the filming was gray. After filming was complete, the radio advertisements were constructed from the television advertisements’ soundtracks and backup tapes for all for advertisements were created. Manipulation check. A manipulation check was conducted to determine if subjects’ perceptions of the soft and hard message advertisements correctly corresponded to the basis of their construction, and if the messages were significantly different. One thought-listing variable (called manipulation check) was coded based on to what degree the respondent thought that the candidate was either a “caring” individual or a “tough decision-maker.” When comparing this item between the soft and hard message groups, t- 41 test results indicate that there is a significant difference between perceptions of the messages as consistent with message development [t = 19.42, df = 231, p = .0001 (233)]. As designed, individuals presented with the soft messages perceived the candidate as caring [M = 4.25, sd = .96 (117)] and that individuals presented with the hard messages perceived the candidate as a tough decision-maker [M = 1.78, sd = .98 (116)]. Further, other items also served to determine the degree to which respondents perceived the hard and soft messages to be similar to a variety of soft-hard dimension items. The results of these t-tests are presented in Table l. 42 Table 1 Means and T-test Results for Compassion, Caring, Hearted, Rational, Smpathy, and Soft Item Mean ad Subjects t-value df p Compassionate 3.91 220 .000 Soft Message 4.87 1.34 117 Hard Message 4.19 1.25 105 Caring 3.76 217 .000 Soft Message 5.05 1.35 117 Hard Message 4.38 1.27 102 Warm-Hearted 3.43 219 .001 Soft Message 4.80 1.21 117 Hard Message 4.26 1.14 104 Rational‘ 0.76 225 .451 Soft Message 3.52 1.56 116 Hard Message 3.37 1.38 111 Sympathetic 3.68 225 .000 Soft Message 4.93 1.45 117 Hard Message 4.23 1.43 110 Soft 4.37 214 .000 Soft Message 4.61 1.31 113 Hard Message 3.83 1.33 103 Note: These items have all been scored a “7,” and their opposites uncompassionate, uncaring, cold-hearted, emotional, unsympathetic, and hard have all been coded a “l * There was not a significant difference between the hard and soft messages, and given that it may have been diflicult for respondents to judge the candidate on the emotional- rational semantic differential as they may not have seen that candidate as either of these, this result is not surprising. These results indicate that individuals in the soft message condition rated the candidate as more compassionate, caring, warm-hearted, sympathetic, and soft, whereas by comparison, in the hard message condition individuals rated the candidate as more uncompassionate, uncaring, cold-hearted, unsympathetic, and hard. Although the differences between the soft and hard message conditions are statistically significant, it is important to note that the low magnitude of difference between the messages (average difference between conditions is .59) may indicate that it is not a powerful manipulation. Measurement instrument. The measurement instrument consists of two sections. The first section (and also the first page) presents eighteen, 5-point Likert type questions, which were designed to measure political involvement. The first set of these political involvement items attempts to measure the respondent’s psychological (political) involvement, whereas the second set of the political involvement items attempts to measure the respondent’s behavioral (political) involvement. Subjects complete this first section of the questionnaire before they are presented the political advertisement. The second section of the measurement instrument consists of a variety of items attempting to measure the respondent’s recall of information presented in the advertisement, such as thought-listing and true/false questions, as well as opinion-oriented items, such as semantic differentials. Additionally, a variety of items attempt to assess the respondent’s party affiliation and past voting behavior as well as his or her general demographic characteristics. After subjects have been exposed to the political advertisement, they are asked to complete this second section of the questionnaire. In order to assure understandability of the questions and to determine the amount of space needed for the thought-listing items, the questionnaire was pretested (the soft message television advertisement was shown). The results of the pretest indicated that all 44 items were clear and that there was plenty of space provided for the thought listing task. A complete review of the measurement instrument is presented in Appendix D. Codingprocedures. There were three items on the questionnaire that required respondents to either describe or list their responses. Given the specific goal of this thought-listing section (to measure the recall of superficial imagery and substantive material), specific categorical responses were searched for. The responses that were sought included 1) a check to see if the manipulation worked (that the respondent would mention that was either “caring” individual or “tough decision-maker”) (called manipulation check), 2) a brief description of the issues presented (divided into three sections pertaining to the three issues) (called divorce, partner, suicide), 3) any additional listing of an issue-oriented response (e.g., “he’s a family-oriented guy”) (called additional issues), and 4) any listing of information which concerned the candidate’s image or characteristics (called superficial imagery). Even though there were three distinct issues per message (stricter divorce laws, physician-assisted suicide, and benefits for domestic partners), the rationale as well as the beginning and concluding sentences provided additional issue content which respondents could potentially recall. Thought-listing items were coded based on the categories of manipulation check, divorce, partner, suicide, and additional issues. Appendix E presents a detailed description of the coding scheme. Independent measures. Political involvement was measured with eighteen items, the first set of ten items inquired about psychological involvement (e.g., interest in politics, interest in following campaigns), while the second set of eight items inquired about behavioral involvement (e.g., volunteering time to work on a campaign, trying to persuade 45 others to vote for one candidate over the other). Based on their item content, all eighteen items appear to have face validity. Although the standardized item alpha was very good for the initial 18 items, or = .93, firrther tests were employed to determine the validity of the scale. The initial eighteen items were subjected to tests of internal consistency to determine which item’s errors were greater than sampling error. After several runs to determine the best fit of the items, four items were eliminated from the original eighteen (items eliminated were #6, #10, #13, and #14). The percent of items for which the error is greater than sampling error is five percent which fits into the acceptable range of error. There was some initial concern with eliminating items #13 and #14 (“I’ve donated money to a candidate or a campaign” and “I’ve distributed campaign materials for a candidate,” respectively) as these items were the most extreme in measuring political behaviors, and subsequently would help determine individuals with high political involvement. Given this concern, additional analyses were conducted to assess to what degree individuals in the high involvement group would be lost to the low involvement group when questions #13 and #14 were eliminated. However, due to the content of question #15 (“I’ve volunteered my time for a candidate to work on a campaign”), all but one individual remained in the high political involvement group compared to the analysis without questions #13 and #14. Therefore, it was determined that there should be no concern with eliminating items #13 and #14 in terms of losing behavioral political involvement content in the questionnaire. The standardized item alpha for this 14-item political involvement scale remains very good, or = .93. 46 Confirrnatory factor analysis was then performed on this l4-item political involvement scale to firrther determine internal consistency. Results of the factor analysis indicated that these items sufficiently measured the same construct. The factor loadings are presented in Appendix F. Given the factor loadings and the percentage of items for which error was greater than sampling error, it appears that the l4-item political involvement scale has internal consistency. When considering the two subscales of psychological political involvement and behavioral political involvement, reliabilities of each also appeared to be very good. For the new 8-item psychological subscale, standardized item alpha was very strong at, or = .91; and for the new 6-item behavioral involvement subscale, standardized item alpha was fairly strong at, or = .81. Additionally, these two subscales correlate well with one another, r = .76 (233), p = .0001. To establish high and low political involvement groups, the 14-item political involvement scale was dichotomized into high and low involvement based on a median- split [median = 2.714, sd = .713 (233)]. Those individuals whose involvement score was greater than 2.714 were grouped into the “high involvement” group (n = 118); those individuals whose involvement score was less than 2.714 were grouped into the “low involvement” group (n = 115). Dependent measures. Substantive material was measured by creating one scale composed of two main sets of items (thought listing and true/false questions) which measured the substantive material recalled by the respondent. Given the nature of substantive material, such that direct questions concerning the content could be asked in 47 addition to thought listing, multiple methods for measuring this recall were employed. When considering the thought-listing items, substantive material was coded based on the respondent’s recall of the issues that the candidate presented in the advertisement. Thoughts listed by subjects were coded based on the three issues (divorce, partner, and suicide) that were presented in the advertisements. The reliability of this substantive material scale composed of these three recalled issues is acceptable with the standardized item alpha or = .60. Given that this scale represents the accuracy of which the issues in the advertisements were recalled it is expected that the reliability for such a scale would be moderately low. Another thought-listing item which was added into the overall substantive material scale represents the number of additional issues (substantive material) recalled that were not previously coded based on the recall of the three issues. As this is a one-item measure, no reliability can be calculated. Based on the true/false items which asked specific questions about the issues presented in the advertisements, another scale was formed that measured the degree to which a respondent answered all questions correctly. Although the content for the advertisements was different (hard message versus soft message), thus creating different sets of correct answers for the true/false questions, one scale to compare all individuals was created. The correct answers (whether true or false) for the specific message were recoded in the same direction to create a score for those items. As there were two message groups, this was performed for both the soft and hard message groups independently. Combining these sets of independent scores created one scale that 48 accurately measured the correct information recalled based on the message that was presented. Given that this scale is composed of items that measure the accuracy of recalled substantive material from the advertisements, it is expected that the reliability for the scale will also be moderately low, as the reliability would depend more upon respondents’ correct recall of all three issues presented in the advertisements. Based on combining the four treatment groups into the categories of soft messages and hard messages, the standardized item alphas are as follows: Soft Messages, or = .45; Hard Messages, or = .53. Although these alphas are low, this is still a fair measure given that individuals may recall more from one issue than from another. In order to create one general measure (a combination of the thought-listing and true/false items) to measure the substantive material recalled, one scale was created by summing the true/false scale with the scores item the thought-listing items. This is the scale that is used to measure substantive material. The reliability of this general substantive material scale is or = .53. Although this reliability is only moderate, internal consistency tests indicate that all item errors were within sampling error. Additionally, exploratory factor analysis was then performed on these items to determine how well they were measuring the same construct. The factor loadings are presented in Appendix G. Given the nature of superficial imagery, such that it should be a pure measure of the individual’s impression of the candidate’s image and not one that is prompted by the questionnaire, there was only one available means of measuring respondent’s recall of superficial imagery. Superficial imagery was measured by asking the respondent to list characteristics of the candidate that he or she could remember. Such characteristics (e.g., 49 nice looking, smart, middle-aged) were coded based on how many were listed. Given that other items inquiring about superficial imagery may prompt the respondent (e.g., “How trustworthy is the candidate?”), measuring superficial imagery stands as a one-item measure, therefore, no reliability can be calculated. Total recall (combining all of the measures for superficial imagery and substantive material) is derived by adding together the superficial imagery measure to the substantive material scale. Given that this total recall measure is potentially combining theoretically opposite constructs (substantive material and superficial imagery), the reliability of this scale is expected to be fairly low, standardized item alpha or = .44. Internal consistency tests indicate that 26 percent of the items have errors which are greater than sarnpling error. However, as this total recall scale is designed to measure recall of both superficial imagery and substantive material, such error is expected and acceptable. Additionally, exploratory factor analysis was performed. Once again, given that total recall measures the recall of both substantive material and superficial imagery, it is not surprising that some of the items, specifically additional issues and superficial imagery, have low factor loadings. The factor loadings for total recall are presented in Appendix H. Chapter 5 RESULTS Overview This section begins with a general description of the profile of the participants and a profile of the perceptions of the actor who was used as the political candidate in the advertisements. This is followed by the findings and explanation of the findings for each of the hypotheses and the research question. For each of the hypotheses, the findings are initially presented with the entire population, but given that the nature of this study essentially deals with the United States political system, subsequent analyses are performed with using only the United States citizen population (N = 221). Additional analyses were also conducted on other variables, such as, those who voted in the last presidential election, as well as males and females. Profile of Research Participants The 233 subjects in this study consisted of 158 women and 75 men. Ninety-five percent of the participants were between 18 and 26 years of age, with an average age of 22 years (sd = 3.68). The remaining 5% of participants were between 27 and 50 years of age. Most of the participants were Caucasian (82%), 7% were Asian, 6% were Afiican American, 2% were Hispanic, 3% indicated “other,” and 2 subjects did not respond to that question. Of the 233 subjects, 43% were seniors, 41% were juniors, 14% were sophomores, 1% were fi'eshman, one subject (0.4 %) indicated graduate student, and two subjects did not respond to that question. 50 51 Ninety-five percent of the subjects are United States citizens, and 84% of them are registered to vote. F ifty-seven percent of the subjects voted in the last presidential election (November, 1996). Over a third of the participants stated that they are Democratic (39%), 31% are Republican, 16% are Independent, 3% are Libertarian, 9% indicated that they are “other,” and 2% did not answer the question. When asked about how they consider themselves politically, about half of the subjects considered themselves “moderate” (53%), 30% considered themselves “liberal,” 16% considered themselves “conservative,” and 1% did not answer the question. Profile of the Political Candidate To make sure that findings in the hypotheses are not due to any extreme evaluations of the candidate, it is important to looked at respondents’ evaluations of the candidate on specific items. This evaluation is based on several semantic differentials. In addition to the means and standard deviations, Table 2 also presents the percentage of participants who either expressed their neutrality on the item by (responding with a “4”) or left that item blank. 52 Table 2 Descriptive Data on Semantic Differential Scales Dimension M g % natrked 4 or left blag Trustworthy 4.48 1 .36 29.2 Qualified 3 .97 1.31 43.8 Honest 4.61 1.30 33.9 Sincere 4.60 1.52 17.6 Attractive 3 .78 1 .27 48.5 Friendly 4.47 1.30 26.6 Experienced 3 .65 1.27 45.1 Knowledgeable 4.32 1.23 36. 1 Competent 4.49 1 .61 37.8 Intelligent 4.68 1.05 37.4 Note: All items shown were coded toward the higher end of the scale “7” and their opposites (e.g., untrustworthy, unqualified, etc.) were coded at the lower end of the scale ..1 .” For several items (those with 30% or more respondents who marked “4” or left item blank), respondents were not very involved in the task of evaluating the candidate. As a result, such responses of “4” or leaving item blank impacted the overall mean for those items. However, in general, it is important to note that these items did not have many extreme evaluations of the candidate as most of the responses were normally distributed around the mean. Therefore, based on these results, it can be assumed that the overall findings will not be influenced by any extreme evaluations of the candidate. Additionally, t-tests were conducted to determine that there were no significant differences between high and low political involvement groups as well as between soft message and hard message advertisement groups on these items. When comparing the high and low political involvement groups no significant differences were found. When comparing the soft and hard message advertisement groups, there were significant differences found with the semantic differential items of fiiendly [t = 2.93, df = 222, p = .004 (224)] and experienced [t = -2.25, df = 210, p = .025 (212)]. For the item of fiiendly, evaluations in the soft message condition [M = 4.71, sd = 1.24 (115)] were higher than for the hard message condition [M = 4.21, sd = 1.32 (109)]. For the item of experienced, evaluations in the hard message condition [M = 3.85, sd = 1.32 (102)] were higher than for the soft message condition [M = 3.46, sd = 1.19 (110)]. In general, individuals in both the soft and hard message conditions thought that the candidate was more fiiendly than unfiiendly and also thought that the candidate was less experienced than experienced. When looking at each item specifically, there is good reason for the significant difference. For the item of fiiendly, it seems reasonable that individuals viewing the soft message (in which the candidate presents a “caring” rationale) would be more likely to judge the candidate as fiiendly than in the hard message condition in which the candidate presents a “tough decision-maker” rationale. Similarly, for the item of experienced, it seems reasonable that individuals viewing the hard message would be more likely to judge the candidate as more experienced as he offers a “tough decision-maker” rationale opposed to a “caring” rationale. In general, these data indicate that the candidate’s characteristics did not provide reason for crucial differences between relevant groups. Hypotheses for All Advertisements mag. Hypothesis I predicted that individuals with higher levels of political involvement would process more substantive material than individuals with lower 54 levels of political involvement. To examine this prediction, 3 t-test was performed to determine if there were significant differences in recall of substantive material between high and low political involvement individuals. Although the high involvement group [M = 11.30, sd = 2.78 (118)] processed slightly more substantive material than the low involvement group [M = 11.07, sd = 2.82 (115)], there was not a significant difference [t = -.63, df = 231, p = .532 ( 233)]. Therefore, these data fiom the overall sample do not support this hypothesis. For the subset of subjects who indicated that they were United States citizens, the results differ from the overall sample. Again, high involvement individuals [M = 11.61, sd = 2.35 (1 13)] processed slightly more substantive material than low involvement individuals [M = 11.12, sd = 2.76 (108)]. Even though the difference doubled as it increased fiom .23 of a point (overall sample) to .49 (United States population), there was still not a significant difference [t = -1.43, df= 219, p = .154 (221)]. In addition, recall of substantive material was separated to test for differences between the high and low involvement groups on the thought-listing measures as well as on the true/false items. For the thought-listing items, although high involvement individuals recalled more substantive material [M = 9.10, sd = 2.30 (113)] than the low involvement individuals [M = 8.62, sd = 2.71 (108)], there was not a significant difference [t = -1.41, df = 219, p = .159 (221 )1. Similarly, for the true/false substantive items, although the high involvement individuals processed slightly more [M = 2.51, sd = .223 (113)] than the low involvement individuals [M = 2.50, sd = .238 (108)], there was not a significant difference [t = -.48, df= 219, p = .632 (221)]. 55 Additionally, analysis of covariance, which analyzed the effects of political involvement on recall of substantive material after removing sex, was performed. The results indicated that although the effects of political involvement on recall of substantive material increased, the effect was not significant F (1, 220) = 2.77; p = .098. Further analyses were conducted by trichotomizing political involvement (bottom third < 2.429, top third > 2.929). These high involvement individuals [M = 11.85, sd = 2.35 (69)] processed significantly more substantive material than the low involvement individuals [M = 10.87, sd = 2.72 (81)], with [t = -2.37, df= 148, p = .019 (150)]. Recall of substantive material was also separated into thought-listing and true/false measures to test for differences. For the thought-listing items, high involvement individuals recalled significantly more substantive material [M = 9.33, sd = 2.32 (69)] than the low involvement individuals [M = 8.37, sd = 2.67 (81)], with [t = -2.36, df= 148, p = .02 (150)]. However, for the true/false items, although the high involvement group [M = 2.52, sd = .203 (69)] recalled more than the low involvement group [M = 2.50, sd =.245 (81)], there was not a significant difference [t = -.52, df = 148, p = .607]. Finally, correlational analyses were conducted to describe the degree of relationship. Political involvement correlated significantly with recall of substantive material [r = .15 (221), at p = .03], however it is important to mention that this is a weak correlation. Additionally, separate correlations for political involvement and recall of substantive material were performed for males and females. When correlating political involvement with recall of substantive material, there was not a significant correlation for males [r = .09 (70), at p = .471], however, there was a significant correlation for females 56 [r = .21 (151), at p = .01 1]. When firrther considering the separate measures of substantive material with females, there was a significant correlation with the thought- listing items [r = .20 (151), at p = .013]. In general, the data support the hypothesis. When using the high and low involvement individuals based on trichotomizing political involvement, both the general recall of substantive material as well as the thought-listing measures produced significant results. Hypothesis 2. Hypothesis 2 predicted that individuals with lower levels of political involvement would process more superficial imagery than individuals with higher levels of political involvement. To examine this prediction, a t-test was performed to determine if there were significant differences in recall of superficial imagery between high and low involvement individuals. Although the difference was not significant [t = -.39, df = 231, p = .693 (233)], opposite results occurred as the high involvement group [M = 3.03, sd = 2.30 (118)] processed slightly more superficial imagery than the low involvement group [M = 2.90, sd = 2.38 (115)]. Therefore, these data from the overall sample do not support this hypothesis. Supplemental analyses were performed using only the subjects who indicated that they were United States citizens. However, when testing Hypothesis 2, the results did not improve. High involvement individuals [M = 3.04, sd = 2.30 (113)] still processed slightly more superficial imagery than the low involvement group [M = 2.97, sd = 2.39 (108)], however, there was still not a significant difference [t = -.20, df = 219, p = .841 (221)]. Additionally, individuals in the television condition were analyzed separately 57 because the radio condition could be diminishing the relationship between political involvement and recall of superficial imagery. For the television condition, the t-tests between the high and low involvement groups revealed that for superficial imagery, the high involvement group still processed more [M = 4.62, sd = 2.03 (50)] than the low involvement group [M = 4.30, sd = 2.12 (61 ), but the difference was not significant [t = -.82, df= 106, p = .412 (111)]. Additionally, analysis of covariance, which analyzed the effects of political involvement on recall of superficial imagery after removing sex, was performed. The results indicated that although the effects of political involvement on recall of substantive material increased, the effect was still not significant F (1, 220) = .327; p = .568. Further analyses were conducted with the trichotomized political involvement. As the high involvement individuals [M = 3.01, sd = 2.40 (69)] processed about the same amount as the low involvement individuals [M = 3.05, sd = 2.49 (81)], the difference was not significant [t = .09, df= 146, p = .931 (150)]. Additionally, individuals in the television group were further considered with the trichotomized political involvement. T-tests between the high and low involvement groups revealed that the high involvement group [M = 4.74, sd = 2.19 (31)] did recall more than the low involvement group [M = 4.59, sd = 2.11 (44), but the results were not significant [t = -.30, df= 63, p = .77]. Finally, correlational analyses were conducted to describe the degree of relationship. Political involvement failed to significantly correlate with recall of superficial imagery [r = -.03 (221), at p = .634]. Separate correlations for political involvement and 58 recall of superficial imagery were performed for males and females. When correlating political involvement with recall of superficial imagery, there was not a significant correlation for males [r = -.04 (70), at p = .761], likewise, there was not a significant correlation for females [r = .01 (151), at p = .864]. In general, these data do not support the hypothesis. Overall, the results of these two hypotheses indicate that there is not a definite . relationship between political involvement and the amount of superficial imagery or substantive material recalled from the advertisements. There is general support for Hypothesis 1 when considering trichotomized political involvement with the United States population. Additionally, when considering only females, the significant positive correlation suggests that with higher levels of political involvement more substantive material is recalled. However, there is no support for Hypothesis 2. Hypotheses for Soft Message and Hard Message Advertisemen_t§ Hypotheses 3a_and 3b. To initially test the predicted contingent interaction of Hypotheses 3a and 3b, analysis of variance (ANOVA), which analyzed the effects of political involvement and message condition on recall of total information, was performed. The main effect for the soft and hard message conditions was not significant, F (1, 232) = .05; p = .818. Additionally, the main effect for political involvement was also not significant, E (I, 23 2) = .52; p = .474. Further, there were no two-way interactions detected between political involvement and message condition, E (1, 232) = 1.142; p = .286. Thus, the basic prediction of a contingent interaction is not supported. Means for 59 high and low political involvement groups across the hard and soft message conditions are presented in the left column of Table 3. Table 3 Me_ans for Total Recall, Superficial Imagery and Substantive Material. For Entire Population W Soft Hard High Political Inv. 14.54 14.14 Low Political Inv. 13.70 14.31 For United States Citizen Population Total Recall Soft Hard High Political Inv. 14.78 14.53 Low Political Inv. 13.83 14.39 Recall of Imager_'y Soft Hard 3.13 2.94 2.77 3.06 Recall of Imagegg Soft Hard 3.15 2.93 2.88 3.08 Recall of Material Soft Hard 11.41 1 1.21 10.92 11.25 Recall of Material Soft Hard 11.63 11.59 10.96 1 1.31 When performing the same calculations with the United States citizen population, similar results were found. The AN OVA revealed that there were no significant differences between the soft and hard message conditions, E (1, 220) = .10; p = .746, or for political involvement, E (1, 220) = 1.42; p = .235, on recall of total information. Nor were there any two-way interactions detected between political involvement and message condition, E (1, 220) = .79; p = .376. 60 As Hypothesis 3a offers a prediction for individuals in the high involvement group, high involvement individuals were selected for additional examination. T-tests were performed on just the high political involvement individuals to determine if there were significant differences on their recall of total information between the soft and hard message groups. As predicted, there were no significant differences in total recall between the two message groups [t = .60, df = 116, p = .55 (118)]. The results indicated that high political involvement individuals in the hard message group [M = 14.14, sd = 3.67, (63)] process about the same amount of total information as high political involvement individuals in the soft message group [M = 14.54, sd = 3.48, (55)]. Additionally, the results of the t-test with the United States population also indicated that there were no significant differences between high involvement individuals in the hard message group and high involvement individuals in the soft message group [t = .42, df = 111, p = . 675 (113)]. Therefore, even given the United States citizen population, high involvement individuals presented with the hard message [M = 14.53, sd = 3.24, (60)] process about the same amount of total information as high involvement individuals when presented soft messages [M = 14.78, sd = 3.17, (53)]. As Hypothesis 3b offers a prediction for individuals in the low involvement group only, low involvement individuals were selected for additional examination. T-tests were performed on just the low political involvement individuals to determine if there were significant differences between their recall of total information between the soft and hard message groups. There were no significant differences in total recall between the two message groups [t = -.91, df= 113, p = .366 (115)]. The results indicated that low 61 political involvement individuals in the hard message group [M = 14.31, sd = 3.55, (53)] processed slightly more total information than low political involvement individuals in the soft message group [M = 13.70, sd = 3.69, (62)]. Although the results are not significant, the amount of total recall is in the opposite direction as predicted. Additionally, the results of the t-test with only the United States population also indicated that there were no significant differences between low involvement individuals in the soft message group and low involvement individuals in the hard message group [t = -.81, df = 106, p = . 419 (108)]. Even when considering the United States citizen population, low involvement individuals presented with soft messages [M = 13.83, sd = 3.60, (59)] process less total information (although not significant) than low involvement individuals when presented hard messages [M = 14.39, sd = 3.35, (49)]. As with the previous calculations, these results are also in the opposite direction. Additional analyses were performed to determine if there were any differences in recall for the components of total recall (superficial imagery and substantive material). Item means, of the entire population and the United States citizen population, for total recall as well as superficial imagery and substantive material based on soft and hard message conditions and high and low involvement groups have been presented in Table 3. For both recall of superficial imagery and recall of substantive material, as well as recall of total information, there were no significant differences between the soft and hard message conditions for either the entire population or the United States citizen population. Further, analyses were conducted by separating the components of substantive material. The t-test on the thought-listing measure of substantive material did not yield significant ————‘ 62 results. For the thought-listing measure, low involvement individuals in the hard message condition [M = 8.76, sd = 2.36 (49)] processed more than individuals in the soft message condition [M = 8.51, sd = 2.99 (59)], with [t = -.48, df= 106, p = .633 (108)]. However, for the true/false substantive items, low involvement individuals recalled significantly more from the hard message condition [M = 2.55, sd = .233 (49)] than individuals in the soft message condition [M = 2.45, sd = .235 (59)], with [t = -2.21, df= 106, p = .029 (108)]. Overall, the predicted contingent interaction is not supported, but the component of 3a is supported. The results of two hypotheses indicate that high involvement individuals will process about equal amounts of total information when comparing soft and hard messages, thus these data support Hypothesis 3a. However, for low involvement individuals, it appears as though they process slightly more information when presented with hard messages then when presented with soft messages. For this relationship, the differences are significant for only the true/false substantive material items. These results for Hypothesis 3b are in the opposite direction as what was predicted and may be a result of the low political involvement individuals responding more to the “tough decision- maker” rationale (and thus recalling more) than the “caring” rationale. This will be discussed firrther in the Discussion section. Hymtheses for Television and Radio Advertisements Hymthesis 4a. Hypothesis 4a predicted that both high and low involvement individuals would recall more superficial imagery from television messages than from the radio messages. To examine this prediction, a t-test was performed to determine if there were significant differences between the television and radio advertisement conditions on 63 recall of superficial imagery. Individuals in the television condition [M = 4.35, sd = 2.12 (117)] recalled more superficial imagery than individuals in the radio condition [M = 1.57, sd = 1.60 (116)]. When testing the difference between these means, the results indicate that there is a significant difference between the treatment groups of television and radio [t = 11.30, df = 231, p = .0001 (233)]. Therefore, these data are consistent with the prediction that more superficial imagery will be recalled from television advertisements than fi'om radio advertisements. For the United States citizen population, the t-test results are very similar. Individuals in the television condition [M = 4.44, sd = 2.07 (111)] recalled more superficial imagery than individuals in the radio condition [M = 1.55, sd = 1.56 (110)]. When testing the difference between these means, the results indicated that there is a significant difference between the treatment groups of television and radio [t = 1 1.68, df = 219, p =.0001(221)]. Hypothesis 4b. Hypothesis 4b predicted that both high and low political involvement groups will process more substantive material from the radio advertisements than from the television advertisements. To examine this prediction, a t-test was conducted to determine if there were significant differences between the television and radio advertisement conditions on recall of substantive material. As predicted, individuals presented with the radio advertisements [M = 11.43, sd = 3.06 (116)] recalled more substantive material than individuals presented with television advertisements [M= 10.95, sd = 2.50 (1 17)]. However, when testing the difference between these means, the results indicate that there is not a significant difference between the treatment groups of radio and 64 television [t = -l.31, df = 231, p = .192 (233)]. Therefore, these data do not support the prediction that more substantive material will be recalled from the radio advertisements than from the television advertisements. For the United States citizen population, the t-test results are very similar. Individuals in the radio condition [M = 11.64, sd = 2.69 (110)] recalled more substantive material than individuals in the television condition [M = 11.10, sd = 2.41 (111)]. However, when testing the difference between these means, the results indicate that there is not a significant difference between the treatment groups of radio and television [t = -I .58, df = 219, p = .1 16 (221)]. Additionally, the thought-listing and true/false items for substantive material were considered. However, consistent with the previous findings, there were no significant differences for either the thought-listing or the true/false measures. For the thought-listing measures, although the radio group [M = 9.15, sd = 2.64 (110)] processed more substantive material than the television group [M = 8.58, sd = 2.36 (111)], the difference was not quite significant [t = -1.72, df= 219, p = .087 (221). For the true/false measure, the television group processed slightly more [M = 2.52, sd = .227 (111)] than the radio group [M = 2.49, sd =.233 (110)], and the results were not significant [t = 1.16, df= 219, p = .245 (221)]. Overall, the results of these hypotheses indicate that there is significant difference on the amount of superficial imagery recalled between the television and radio advertisement groups, but that there is not a significant difference on the amount of substantive material recalled between the radio and television advertisement groups. 65 Research Question Regarding Candidate Evaluation. Research question 1. The research question inquired as to which message condition, soft or hard, would result in more positive evaluations of the candidate. In general, for all the evaluation items, respondents in the soft and hard message conditions evaluated the candidate similarly. Ten semantic differentials were grouped into two groups in terms of their correspondence with the expected evaluation outcome for either the soft messages or hard messages as based on their face validity. The items which correspond to the soft message condition are trustworthy, honest, sincere, attractive, and fiiendly. The items which correspond to the hard message condition are qualified, experienced, knowledgeable, competent, and intelligent. The results of comparing the evaluations of the soft and hard message groups based on these two semantic differential groups is presented in Table 4. 66 Table 4 Means and T-tests for the Two Groups of Smantic Differential Scales. 1th man a M t_-_vzu_uq at p Trustworthy .85 207 .399 Soft Message 4.58 1.39 107 Hard Message 4.42 1.31 102 Honest -I.23 201 .221 Soft Message 4.51 1.29 104 Hard Message 4.74 1.35 99 Sincere -.62 215 .534 Soft Message 4.55 1.63 111 Hard Message 4.68 1.43 106 Attractive 1.77 190 .078 Soft Message 3.96 1.22 98 Hard Message 3.64 1.29 94 Friendly 2.63 210 .009 Soft Message 4.72 1.25 1 10 Hard Message 4.25 1.32 102 Total . 93 230 .356 Soft Messages 4. 48 l. 05 l I 7 Hard Messages 4. 35 I. 03 l [5 Qualified -.69 198 .492 Soft Message 3.89 1.25 104 Hard Message 4.02 1.35 96 Experienced -2.45 198 .015 Soft Message 3.45 1.22 105 Hard Message 3.88 1.30 95 Knowledgeable -.30 203 .763 Soft Message 4.30 1.81 106 Hard Message 4.35 1.27 99 Competent -1.28 202 .201 Soft Message 4.41 1.15 105 Hard Message 4.62 1.15 99 Intelligent 1.49 207 .138 Soft Message 4.82 1.01 109 Hard Message 4.60 1.09 100 Total -. 80 226 .423 Soft Messages 4. 18 .88 1 16 Hard Messages 4. 28 1.07 l 12 _ 67 Although these results are not significant, they do offer usefirl information. For the items corresponding to soft messages, individuals in the soft message condition did evaluate the candidate more positively than individuals in the hard message condition. Similarly, for the items corresponding to hard messages, individuals in the hard message condition did evaluate the candidate more positively than individuals in the soft message condition. However, when considering all of the items independently, individuals in the hard message group evaluated the candidate more positively on more items than individuals in the soft message group. Additionally, when considering the items independently, only two items had significant differences (fiiendly and experienced, as mentioned in the candidate profile section). These results indicate that individuals in the soft message condition evaluated the candidate as significantly more fiicndly than individuals in the hard message condition, and that individuals in the hard message condition evaluated the candidate as significantly more experienced than individuals in the soft message condition. Additionflmalyses Additional analyses were conducted to firrther understand individuals’ recall of substantive material and superficial imagery. More specifically, differences between those who had voted in the last presidential election versus those who had not, and those who are female versus those who are male were analyzed. Voted versus did not vote. When considering possible differences between those who voted in the last presidential election and those who did not, only for political involvement was there a significant difference. T-tests were performed between those 68 who voted in the last presidential election and those who did not (United States citizens only) on the variables of total recall, recall of substantive material, recall of superficial imagery, and political involvement. Means and results of the t-tests are presented in Table 5. Table 5 Means and T-tests for Did Not Vote and Voted on Recall of Toad Information, Refill of Superficial Imageg, Recall of Substantive Material and Political Involvement. Item Mean M Subjects t-value df p Total Recall -.82 203 .414 Did Not Vote 14.16 3.07 88 Voted 14.52 3.52 133 Recall of Imagery I -1.01 211 .314 Did Not Vote 2.82 2.03 88 Voted 3.13 2.52 133 Recall of Material -.16 177 .876 Did Not Vote 11.34 2.68 88 Voted 11.39 2.49 133 Political Inv. -6.62 208 .0001 Did Not Vote 2.39 .59 88 Voted 2.96 . 70 133 Note: T-test results are based on unequal cell sizes. Also, the total recall and recall of substantive material measures are calculated based on summed scores, whereas, the political involvement scale is calculated based on averaged scores. The only significant difference between those who voted and those who did not vote was for the variable of political involvement. In general, these results are consistent with logical expectations, such that the more politically involved one is, the more likely he 69 or she will be to vote in major political elections. However, it would also be expected that because those who voted are more politically involved that they also should process more substantive material than those who are less involved. One explanation for why this is not the case, may be due to the laboratory nature of the study, such that subjects were presented the advertisement twice and they were prompted to pay attention to the advertisement. Macs versus feraalg. When considering the differences between males and females in terms of political involvement and recall, the results are very interesting. T- tests were conducted to determine if there were any initial differences between males and females for recall of total information, recall of superficial imagery, recall of substantive material, and political involvement (United States citizen population). There were significant differences between males and females on three of the four variables. The results are presented in Table 6. 70 Table 6 Means and T-tests for Males and Fanales on Recall of Tomnfomaiom Recall of Superficial Imagery, Recall of Substaative Material and Political Involvement Item Mean M Subjects t-value M p Total Recall -2.72 152 .007 Males 13.53 3.00 70 Females 14.77 3.43 151 Recall of Imagery -2.41 157 .017 Males 2.49 2.05 70 Females 3.25 2.43 151 Recall of Material -1.31 138 .194 Males 11.04 2.51 70 ' Females 11.52 2.58 151 Political Inv. 2.99 125 .003 Males 2.95 .75 7O Females 2.63 . 68 151 Note: T-test results are based on unequal cell sizes. Also, the total recall and recall of substantive material measures are calculated based on summed scores, whereas, the political involvement scale is calculated based on averaged scores. These results indicate that females recall more superficial imagery and total information (based on superficial imagery) than males, but that males are more politically involved than females. To firrther examine recall of substantive material, t-tests for the thought-listing and true/false measures were performed. For the thought-listing measures, although the females recalled more [M = 9.03, sd = 2.53 (151)] than the males [M = 8.50, sd = 2.46 (70)], there was not a significant difference [t = -1.49, df = 138, p = .140 (221)]. For the true/false measure, the males actually recalled slightly more [M = 2.54, sd = .23 ‘+‘ 71 (70)] than the females [M = 2.49, sd = .231 (151)], but there was not a significant difference [t = 1.68, df= 138, p = .096 (221)]. What makes these results interesting is that although males are more politically involved, females recall more total information, and specifically, more superficial information, than males. Such findings might be a result of females being more interested in the characteristics of the candidate compared to males. Thoughts on the differences are explored further in Chapter 6. Chapter 6 DISCUSSION 0_v_§r_vie_w The discussion of the results is composed of several sections. To begin, the general patterns of findings with political involvement and recall for all advertisements, soft and hard messages, and television and radio advertisements are discussed. Then possible explanations for the differences between males and females are explored. This is followed by implications for the results of this study. Limitations for this study are then discussed. Finally, directions for firture research are presented. Pplitical Involvement and Recall with All Advertisements In general, it was predicted that high involvement individuals, based on systematic processing, would process more substantive material compared to low involvement individuals and that low involvement individuals, based on heuristic processing, would process more superficial imagery compared to high involvement individuals. Generally, when using trichotomized political involvement, the data support the prediction that high political involvement individuals process more substantive material compared to low political involvement individuals. Significant results are produced between these high and low involvement groups for both the general recall of substantive material and the thought-listing measures of substantive material. However, there was no support for the prediction that low political involvement individuals process more superficial imagery compared to high political involvement individuals. 72 As based on the HSM, the prediction that high involvement individuals should process more substantive material is expected. However, the results for the prediction that low involvement individuals should process more superficial imagery than high involvement individuals were contrary to the theoretical basis of the HSM. The results were in the opposite direction as predicted, such that high involvement individuals processed more superficial imagery than low involvement individuals, however, they were not significant. One explanation for why the results were not in the expected direction may be due to the internal validity considerations of the study. Although the soft and hard messages differed in terms of their rationales, the minor changes between these two message conditions may account for the lack of a significant differences in the expected direction for low political involvement and recall of superficial imagery. When considering the t-tests for the key semantic differentials concerning the manipulation check, although the results indicated that the messages significantly differed from one another, the means still clustered around the neutral value (“4”) which may additionally indicate that the messages were not that different from each other. Given additional messages, in which one message would offer more visuals than the other, a significant correlation between recall of superficial imagery and low political involvement would probably result. When considering the applicability of the HSM to political involvement, based on these results, there is only minimal support. Certainly additional similar examinations must be conducted in order to determine the appropriateness of applying the HSM to political involvement. 74 Political Involvement and Recall with Soft and Hard Messagas In general, a contingent interaction was predicted, such that high political involvement individuals would be motivated to engage in systematic processing, focusing primarily on the content cues, and as a result they would process about the same amount of information in the soft message condition as in hard message condition. Additionally, it was predicted that low involvement individuals would be motivated to engage in heuristic processing, focusing primarily on the source factors, and as a result they would recall more information from the soft message condition than fi'om the hard message condition. High involvement individuals presented with hard messages did process the same amount of information as high involvement individuals presented with soft messages. Given that the message conditions did not vary as to the amount of information presented, it was expected that high involvement individuals would recall about the same amount of information fiom both soft and hard messages. However, low involvement individuals presented with soft messages did not recall more total information than low involvement individuals presented with hard messages. In fact, low involvement individuals in the hard message condition actually recalled slightly more information than low involvement individuals in the soft message condition. The prediction was based on the theoretical concept that low involvement individuals will recall more superficial imagery than substantive material, and therefore, they would also recall more information from the soft message (superficial imagery message) than from the hard message (substantive material message). However, the opposite result occurred. 75 One possible explanation for this opposite effect may be due to the different rationales presented in the messages. In the hard message condition the rationales that were presented (“tough decision-maker”) were probably less traditional and less expected than the rationales in the soft message condition (“caring” rationales), thus, individuals in the hard message condition paid attention because the less traditional rationales increased their attention. As a result of this increased attention for the low involvement individuals, more total information was recalled. If both sets of rationales had been traditional, the results for high involvement individuals would probably still be the same as these individuals, regardless of the message, would be motivated to process information. However, for the low involvement individuals, results may be different by using more traditional rationale, such that low involvement individuals would be motivated to process more information from the soft messages than from the hard messages. When considering the processing of information fiom the soft and hard messages as based on general level of political involvement, the applicability of the HSM is somewhat mixed. High involvement individuals were motivated to process information regardless of the message condition, however, so too were the low involvement individuals. Once again, results inconsistent with the HSM for these predictions may be a result of the rationales used in the messages. Essentially, because of the less traditional rationales used in the hard message, individuals in the low involvement group were motivated beyond the prediction of the theory to process the information presented in the messages. To further understand the applicability of the HSM to predictions about the 76 processing of high and low involvement individuals when presented with hard and soft messages, additional examinations must be made. Refill with Television and Radio Advertisements In general it was predicted that regardless of political involvement, individuals would process more superficial imagery fi'om the televised advertisements than fi'om the radio advertisements and that individuals would process more substantive material fiom the radio advertisements than fiom the television advertisements. Only the prediction for recall of more superficial imagery for the television condition was confirmed. When considering the prediction of recall of substantive material, it is important to note that the results were in the expected direction, such that more substantive material was processed from radio advertisements than from the television advertisements. One explanation for the lack of significant difference between the radio and television conditions on recall of substantive material may be a combination of presenting the messages twice to the subjects and having the issues be fairly novel. It is possible that with the second presentation of the advertisement, individuals gave more attention to the issues presented (and not necessarily the imagery), because of the novelty of the issues. The prediction derived from the HSM, that more superficial imagery will be processed from television than radio, was confirmed. However, the prediction concerning the processing of more substantive material from the radio message than the television message was not confirmed. Once again, although the HSM applies to the prediction of recalling more superficial imagery from television than radio, given some of the constraints of the study, it is unwarranted to assume that the HSM does not apply to the processing of 77 more substantive material fiom the radio messages until further investigations are conducted. When considering all the hypotheses, it is important to discuss the potential concern with the low standardized item alphas for the total recall measure as well as one of the components of that measure, the true/false items. When considering the true/false items (Hard Message, or = .53 and Soft Messages, or = . 45), one explanation for the low alphas may be that individuals may be recalling some of the issues more correctly than others. Such a possibility would serve to minimize the reliability of that measure as there would be a lack of consistency between the degree of issues correctly recalled. Further, the total recall measure also has a low reliability (or = .44). This is a result of the low reliability of the true/false items as well as the addition of the superficial imagery measure to the substantive material measures. In order to calculate total recall all substantive material and superficial imagery items must be combined into one measure, by doing so the alpha will naturally be low as opposite constructs are summed together. Given the nature of this study, these measures even with their low alphas had to be used. As a result, these low alphas may have served to minimize the effects which were predicted. Candidate Evaluation On items which would be expected to correspond to soft messages (trustworthy, honest, sincere, attractive, and fiiendly), individuals in the soft message condition did evaluate the candidate more positively than individuals in the hard message condition. Additionally, on items which would be expected to correspond to hard messages (qualify, experienced, knowledgeable, competent, and intelligent), individuals in the hard message 78 condition did evaluate the candidate more positively than individuals in soft message condition. Although these results are not significant, in terms of positive evaluations it is important to note that individuals in the soft message condition will be more likely to evaluate the candidate more positively on such softer image qualities, whereas, individuals in the hard message condition will be more likely to evaluate the candidate more positively on such harder image qualities. Although the soft and hard message groups did not significantly differ on many of the evaluation items (as they should not have), the results for the items of “friendly” and “experienced” are very interesting. Individuals in the soft message condition indicated that the candidate was more fiiendly than individuals in the hard message condition which was probably a result of the “caring” rationale. Individuals in the hard message condition indicated that the candidate was more experienced than individuals in the soft message condition which was probably a result of the “good policy” rationale. In general, such findings indicate that in the soft message conditions evaluations will be more likely based on superficial imagery, whereas, in the hard message conditions evaluations will be more likely based on substantive material. Males and Females Although there were no predictions in terms of political involvement or recall of information concerning males and females, supplemental analyses on males and females indicate that there are several differences. First, the findings indicate that there are significant differences in the level of political involvement between males and females. In - general, males were more politically involved than females. Males were more likely to 79 have an interest in politics, follow politics and political events, have a good understanding of politics, as well as try to understand more about politics by listening to talk radio and watching political TV programs. Second, the findings indicate that there are significant differences in the amount and type of information recalled fiom political advertisements. In general, females recalled more information fiom the political advertisements than males. The results indicated that females recalled more total information and more superficial imagery from the advertisements than males. However, when analyzing the data firrther, it is evident that the greater amount of superficial imagery processed by females compared to males also accounts for the significant difference in total processing. Therefore, it is more accurate to reveal that females process more superficial imagery than males. Further, it was also found that with females, as political involvement increases so does their processing of substantive material. However, when comparing males and females on recall of substantive material the difference was not significant. Although the subject population is composed of two-thirds females and one-third males, the overall study findings are only effected somewhat as the female-to-male ratio does not necessarily account for a lot of the variance. These results are interesting because the HSM would predict that those who are more politically involved should recall more from the political advertisements. One possible explanation for these results may be twofold. First, given that politics in general is a male-dominated arena (that there are more males than females in office), males more than the females may be more apt to be interested in politics or at a minimum feel socially 8O pressured to be interested in politics and therefore indicate such on the survey. Second, because females have a lower level of political involvement, they may be less motivated to process the substantive material and may be more motivated to process superficially imagery. Mtions of the Study This study makes some useful contributions to the theoretical literature in regards to achieving a better understanding of the relationship between the cognitive processing of political advertisements and political involvement, and to applied political campaigns in regards to the effect of soft and hard messages as well as using television versus radio. From this study, patterns suggesting that high political involvement leads to more processing of information are evident. As the HSM was tested with the concept of an ego-involvement concept (political involvement), such findings offer a start for the extension of the cognitive processing theories into the political literature by offering firrther explanation for what type of information and by whom the information is processed. As predicted by the HSM, the findings indicate that high levels of political involvement lead to greater cognitive processing of substantive material presented in political advertisements. However, for low involvement individuals, the data generally fail to support the prediction. In addition this unsupported prediction being a possible result due to the minor differences between the soft and hard messages, this finding may be derived from the tough decision-maker image and “good policy” rationale (in that they are nontraditional and may serve to stimulate attention and processing) more so, than the lack of applicability of these cognitive processing theories to the political arena. 81 When differences between the high and low involvement groups on processing were examined, predictions concerning the high involvement individuals were supported, however, the same is not true for the low involvement groups. The findings concerning low involvement were probably a result of some of the constraints imposed by the study. Therefore, before conclusions can be completely determined about the applicability of the HSM to political involvement, additional studies should be conducted. Further, when considering implications for the processing of more superficial imagery from the television messages than from the radio messages, the HSM’s predictions are confirmed. However, when considering the processing of more substantive material from the radio messages than from the television messages, even though the results were not significant, they were in the predicted direction, thus some reserved support can additionally be given to the HSM for its applicability. However, additional examinations for these predictions concerning the HSM should be performed. Additional implications exist for political consultants and campaign managers. Television, compared to radio, offers a more superior medium to have more information conveyed, especially if the desired information to be conveyed is superficial imagery in nature. Given this, consultants and campaign managers would be wise to have the good- looking and well-poised candidates present their ideas and image in television advertisements, and to have their less attractive and less poised candidates do more “blind” publicity, such as radio advertisements or television advertisements that show illustrations other than the candidate. However, it is important to mention that individuals receiving information on political candidates are not necessarily medium specific, that is, 82 individuals will eventually see and hear the candidate in many formats through a variety of different media. Therefore,.for the candidates who can convey positive images, the opportunities to showcase their positive image should be maximized, whereas for the candidates who fail to convey positive images, the opportunity to openly present images should be minimized. Implications for soft and hard message presentations are also evident. If the candidate wants to convey a general positive image, then political advertisements should be packaged in terms of soft messages, such as specifically presenting positive and general rationale. However, if the candidate wants to mainly convey that he/she has experience, then the political advertisements should be packaged in terms of hard messages, such as specifically presenting a more aggressive and action-oriented rationale. Such strategies would be very advantageous for the first-time candidate running for a lower level office position as these candidates often have very little experience, if any. Therefore, to construct hard messages that would convey experience, even if minimal exists, would help the candidate gain voters’ attention and even possibly their vote. Further, consultants and campaign managers may want to target the female audience with the advertisements that afford audience members the opportunity to process considerable superficial imagery which would result in greater processing. If the advertisements were done well and their evaluations were positive, then the result of the increased processing by the females may very well capture the female vote for that candidate. To attract the male vote to the candidate, consultants and campaign managers may want to consider the higher level of political involvement that males possess. Advertisements could be designed that are straight forward and serve to establish a camaraderie based on the mutual interest in politics and political events between the candidate and the male audience members. Advertisements, for example, could have the candidate advocating an American duty of being politically involved and understanding the positions of all candidates. Further, the advertisement could compliment those audience members who already are involved and encourage them to encourage others to become involved. By having the advertisement establish a camaraderie or liking between male audience members and the candidate, more information will be recalled from the advertisements. As with the female audience, if the information that is recalled is positive information, then the candidate will capture the male vote as well. Limitations There are some limitations of this study which must be acknowledged. First, limitations as a result of the laboratory nature of this study should be recognized. Given the laboratory setting of this study, subjects were prompted in advance to pay attention to the advertisement when presented to them. As a result, participants probably gave more attention to the advertisement in this environment then they may have in a more realistic environment such as having the advertisement presented to them amongst other advertisements while watching television at their home or while listening to the radio in their car. Additionally as the advertisement was presented twice, this may have also increased recall beyond realistic levels. Although presenting the advertisement a second time simulated increased exposure (similar to what occurs during the election campaign), this may actually have aided the less politically inclined subjects to pay more attention as 84 to compensate for the lack of inclination or interest in politics, thus artificially increasing the recall of information from the advertisements. However, had the advertisements not been presented a second time, it is very realistic to assume that participants would have had very little recall of the information presented. Given that during a political campaign when political advertisements are prevalent, many advertisements are presented and subsequently repeated, presenting the advertisement twice was a realistic decision for the design of the study even though it may also serve as a limitation. A second limitation which should be recognized is that there was only one way for which to measure the construct of superficial imagery. Given the nature of this construct, that it should be pure in form and thus its specific recall not prompted by the survey instrument, the only means for which to measure was to ask a very general question in terms of the characteristics remembered about the candidate. Although a variety of semantic differentials were also on the survey that inquired about the candidate’s “trustworthiness,” “intelligence,” “knowledge,” etc., such items would essentially prompt subjects to think of the candidate in those terms. Even though the questionnaire instructed the respondent to leave the item blank if he/she did not think of the candidate in “those terms,” the semantic differentials were unable to be used as a pure means of measuring superficial imagery. A third limitation of the study is that all three of the issues were fairly novel issues for a candidate to put forth. However, the issues had to be novel so that no prior impressions of candidates presenting these issues would be with the subjects and also so that no prior impressions of these issues as political issues would be that evident. As a 85 result of the novelty of the issues, greater recall of information in these advertisements may be more evident than recall of information fiom other more general advertisements. Additionally, results may have been different with familiar issues opposed to novel issues when considering the number of times messages were presented. However, given that there was variance in the general recall of the substantive material and superficial imagery, the novelty of the issues is only a minor limitation. A fourth limitation which should be mentioned is that the soft - hard message continuum may have varied in additional ways and that perhaps the messages did not exactly represent the soft - hard ends of the continuum. For example, the hard message condition, in addition to being more substantive, may also have been more complex than the soft message condition, and therefore this difference between messages may have affected some of the findings. Further, as observed in the corresponding semantic differentials during the manipulation check, the means were clustered around “4” which indicates that the soft and hard messages, although they were statistically different, may not be as different fi'om one another as they should be to represent opposing ends of the soft - hard message continuum. This potential inequivalency between message conditions, as well as the potential lack of correspondence to soft and hard messages, may account for some of the findings that were not significant. Directions for Future Researah This study offers many additional and interesting ideas for firture research. For all future studies examining the recall of superficial imagery and substantive material additional measures of these two constructs must be employed. Given the low reliabilities 86 of some of the substantive material measures and the total recall measure, more exact items as well as a greater quantity of items should be considered in order to increase the reliability. For superficial imagery, additional items to measure this tricky construct should be constructed so that there is more than one item fiom which to interpret findings. Additionally, more exact hard and soft messages should be constructed so that the message are equivalent in comparison and that they better correspond to the soft - hard ends of the continuum. In general, this research could be extended to include a more in-depth examination of novel political messages and rationale compared to traditional political messages and rationale in terms of the content and amount of cognitive processing. The results might indicate that there is more processing with the conditions of novel rationale and novel messages than with the traditional rationale and traditional message conditions. A second study that would further examine the impact of television on the recall of information from political advertisements would be beneficial in extending this research. To investigate differences between a variety of messages and rationales presented over the television medium would be helpfirl in determining when to use and not use television when it comes to recalling information. Additionally, such a study would also advance practical ideas (do’s and don’ts) for general marketing and campaign strategies. A third study that would more specifically investigate the differences between males and females in terms of their political involvement and cognitive processing of political advertisements would not only be very interesting, but would also advance the science of marketing political candidates to gain the male or female vote. 87 Finally, a fourth study that would extend the processing of image and issue processing to the vote outcome would be very beneficial. To investigate the influence of such processing with the election outcome would offer additional insight into the persuasion and political literature as well as serve as general guidelines for political consultants and campaign managers. This study, with its useful and interesting findings, contributes to both the communication and political literature. In general, the results of this study provide the start to an extension for the HSM into the political communication arena by examining political involvement (essentially ego-involvement). As a result, readers gain a better understanding of the relationship between cognitive processing and political involvement. Additional understanding is also provided for the relationships between cognitive processing and mediums through which to advertise. This study through its findings also encourages firture research to examine differences between males and females as well as the differences between noveland traditional messages. Although firrther studies examining the HSM, and other cognitive processing theories in general, with political involvement should be conducted, this study offers a promising start to the extension. APPENDICES APPENDIX A TEXT FOR ADVERTISEMENTS Soft Message (Superficial Imagegy) Hi, I’m Mark Stevens. I’m running for State Representative because I care about you and the communities in which we all live. I’m in favor of reforming divorce laws in order to keep more families together because I understand the value of strong families. I also favor providing health benefits for domestic partners because I care about the happiness and well-being of all citizens. And, because I have compassion for those who are terminally ill, 1 support the practice of physician-assisted suicide. I’m the candidate who cares about you and your community. Vote for me, Mark Stevens, the candidate who cares. (102 words) H_ard MessagafSubstantive an Hi, I’m Mark Stevens. I’m running for State Representative because I’m the candidate who can make tough decisions and solve tough problems. I’m in favor of instituting stricter divorce laws because it’s a practical means of maintaining the family unit. Also, I want to enact legislation of health benefits for domestic partners because married couples receive these same benefits. And, because I believe in the constitutional right of personal choice, I advocate the legalization of physician-assisted suicide. I’m the candidate who can make tough decisions and solve tough problems. Vote for me, Mark Stevens, the candidate who can handle the tough issues. (103 words) 88 APPENDIX B ROUND 1 - PRETEST OF STMULUS MATERIALS The State House Rqaresartative (MSU jurisdiction) is tryingto assess whidr issues are mos irrportantto MSU studarts. For eadr ofthe following issues. please circle the response which best indicates how irrportant that issue is to you. Please use the following scale: Very ImpomWIya I. Ilnpomnt(l)= 2. Somewhatlmportant(SI)- 3, Neutral (N)- 4. Samar“ Unhnportant(SU)- 5. (U) Unhapomnt - 6. and Very Unhnportant (VU) -7. Very Very Important Uhporut How Important Is the Issue of to you? VI I SI N 811 U VII 1. abortion rigrts proteaion l 2 3 4 5 6 7 2. dismmtling affirmative action 1 2 3 4 5 6 7 3. redudngtu'niar rats 1 2 3 4 5 6 7 4. strictu- college admission standards 1 2 3 4 5 6 7 5. irraeasing errploymart opportun'nies for college 1 2 3 4 5 6 7 grads 6. establishing more rights for dorrrestic partners 1 2 3 4 5 6 7 7. balancingthe State's budget 1 2 3 4 5 6 7 8. crrtt'mg m and local taxes I 2 3 4 5 6 7 9. reducing classroom size at MSU 1 2 3 4 5 6 7 10. mimal rigrts proteaion 1 2 3 4 5 6 7 ll. cartrolling the State's deficit 1 2 3 4 5 6 7 l2. legalia'ng physician-assisted suicide 1 2 3 4 5 6 7 l3. easier accesstostudartfinancial aid 1 2 3 4 5 6 7 14. students’ freedom to have kegger parties I 2 3 4 5 6 7 l5. gestasudenthealthcarecovcrage l 2 3 4 5 6 7 16. autornm'c mien of driver’s lioarse for 1 2 3 4 5 6 7 lira-time driving inpaired conviction 17. wider date rape sanmeing l 2 3 4 5 6 7 18. pamitt'mg more nudart cars on earrpus l 2 3 4 5 6 7 l9. increasing low-cod haldl care for AIDS patients 1 2 3 4 5 6 7 20. reducing the drinking age 1 2 3 4 5 6 7 21. wider divorce laws 1 2 3 4 5 6 7 22. softening panties for drug possession l 2 3 4 5 6 7 23. prohibiting srnok'mg on animus l 2 3 4 5 6 7 24. tougrcr polieia for fake ID possasion I 2 3 4 5 6 7 25. construa'mg more sate prisms 1 2 3 4 5 6 7 89 APPENDIX C ROUND 2 - SOFT AND HARD MESSAGES Soft Message and Softer Verbs A local politician is intersted in your reunion to the following political messages. After each. there are some questions corrcerningthat message Please answer cad) of the questions concerning eadr message. There aretwo messages on the front and three on the badt. Becauselcareahouttheresldentsofthe Lanahrgamlwantmseeamoncomprehendveheahhearecovcngeforevcrym This manage Is: (For cadr. place a chedtrnark that best refleas how you feel about the message) msmcere ‘_ a” M __ __ _ __ sincere hard __ __ _ __ _ _ __ soft warndrearted __ ‘_ _ __ _ _ _ cold-hearted loud _ __ __ _ __ _ __ quia emotional _ _ _ _ __ _ __ rational Hoar hnportant Is this Very Very Issue to you? (circle one) Inportant Inportant Neutral Uninportant Uninportant flow favorable are you Very Very toward this issue? (circle one) Favorable Favorable Neutral Unfavorable Unfavorable I‘mlnfavoroflooldngatstrlcterdlvorcelawalnordertolreeprnorefandlleatogether-mInderstanddrevalaeofstroag fuflllea. This nteaaage is: (For each. place a dreckrnark that heat refleds how you feel about the message) manicure ,_ _ __ __ __ __ _ sincae hard __ __ __ _ __ __ _ soft warrrdrearted _ -_ g_ __ _ __ __ cold-hearted loud __ ____ __ _ __ __ _ quid. emotional __ _.. __ _ _ _ __ rational Haw Import”! II this Very Vay bone to you? (circle one) Inportant Inportant Neutral Unirrportarrt Ilninportant flow favorable are you Very Very toward this hone? (circle one) Favorable Favorable Neural Unfavorable Unfavorable lmttosectheconstructlonofmorestateprbonssowehavefewercrflalsonthesmcohecaueIa-coaeernedahodthe aafetyofourcomdflea. This message ls: (Foreadelacea dredtmark thatbest reflects how you feel abouttherncaaage) msincae _ __ __ _ _ _ _ sincere hard __ _ ,_ _ _ __ ___ _ soft warnircarted __ __ __ __ _ -_ _ cold-hearted loud __ __ _ _ _ _ __ quid emotional __ _ _ __ __ _ __ ratrar' al How Important ls thh Very Very bane to you? (circle one) Inportant Inportant Neutral Unirrportant Uninportant How favorable am you Very Very toward Ms lane? (circle one) Favorable Favorable Neutral Unfavorable Unfavorable 9O 91 BecameIcareahonttheneedsofthoaewhoaretemrhuliyill.Isnpportthelegaflaadonoi’phyaiclan—asslstedauiclde. Tide message Is: (For eadr. place a dreckmark that ha reflects how you feel about the message) msarcere hard warndrearted loud emotional Howilnportantiathb issuetoyou?(circieone) How favorahle are you toward this lane? (circle one) Very Inponant Very Favorable lrrportarrt Favorable ._..__.._. Nartral Neutral Uninportant Unfavorable sincere soil mid-halted quid rational Unfavorable IwanttoloolrhrtoleglshdonofhealthandtaxheneflufordomeadcpartnenhecanaelcanahoutthehppheaandweI-hehg ofallcldaena. This message is: (For eadr. place a dreckmark that be: refleds how you feel about the message) msincere hard warnirearted loud emaional Howlrnportaatlalhb lasaetoyoa? (drcleone) How favorable are you toward tide lune? (circle one) 7*— Favorable ._..._. Irmortant Favorable ..— .— _.__. ...— Neutral Neutral “- _— Unirwonant Unfavorable sincae Unfavorable Hard Message and Harmer Verbs A local politician is otter-crud in your reaction to the following political messages. After each. there are some questions concerning that mange Please mswer each of the questions ccnceming eadr message There are two messages on the fruit and three on the bade. Thmk You!!! Because preventative health care saves taxpayers money, I want to start a more compreberulve health care coverage for everyone. This message ls: (For eadL place a dreckmark that best reflects how you feel about the message) msincere __ __ __ __ _ __ __ sincere hard __ _ _ _ __ _ _ soft wamiiearted __ _ _ __ _ __ _ cold-hearted loud H __ a.” __ M. M _ .._. quid emotional __ ._ _ __ _ _ __ ratrona' I How lmportult ls this Very Va-y Issue to you? (circle one) lnportant Important Nam! Uninportant Uninportant How favorable are you Very . Very toward this lssae? (circle one) Favorable Favorable Nana] Unfavorable Unfavorable I’mlnfavoroflnstltntlngstflcterdlvorcehwshrordertokeepmorefanflestogetherbecausesuchlawsareapracflcdme-s ofmdntaflngfanllynnlts. This Ilse-sage Is: ( For eadl. place a diedunark that beat refleaa how you feel about the message) moan _ __ __ _ __ _ _ sinca'e hard _ _ _ __ _ _ _ soft wamhearted _ __ fl__ __ _ __ __ cold-hearted loud ___ __, __ __ __ __ __ quiet emotional __ _ p__ __ __ _ _ mma' I How Important ls tlds Very Va'y Issue to you? (circle one) lnportant lnportant Neutral Uninportant Unirrportant How favorable are you Very Very toward this lssae? (circle one) Favorable Favorable Neural Unfavorable Unfavorable lwanttolnplesnenttheconstmctlonofmoreatateprisonssowehavefewercrlmlnalsonthestreetabecauselwanttosave Mnmdollars. This message Is: (For eadt. place a diedcmark that has reflects how you feel about the message) msincu'e __ __ ’_ _ __ _ _ sincere hard ” __ _ __ __ _ __ sell warniiearted _ __ _ _ _ _ _ cold-hearted loud M W. ___ ____ ____ ___ M quid emotional __ _ _ __ __ __ __ ratat' al How Important Is this Very Very lssae to you? (circle one) lnportant lnportant Neutral Uninportant Uninportmt How favorable are you Verv Very towaul an. Issue? (circle one) Favorable Favorable Neutral Unfavorable Unfavorable 93 MiWehMWflghtofwrmdMIdvmmkpm-MWMe This message is: (For adt. place a dieckmark that beg refleas how you feel about the message) rnaatcene hard warnilarted loud emotional __. iiowhport-tisdlis Very issnetoyoa? (circle one) lnportmt How favorable are you Very toward tHs hsne? (circle orse) Favorable lnportant Favorable Neural Neutral Unirrportant Unfavorable sincere sofl cold-hearted quid rational iwaattoenaetlegisiatloaofheall-adtaabenefitfordomesdepartnenbeeaase-anledcoaplesrecdvedsesesmebeaefls. ThismessageisuForeadelaceadiedrmarkthatbestrdleashmyou feelaboutthernasage) marncere hard warniiearted loud emaicnal _ How hWhthb Very lssaetoyon? (circleone) invert“ How favorable are you Very toward tfls issue? (circle one) Favorable .— ____.. lnportsnt Favorable Neural Neutral Uninportant Unfavorable sincere soil cold-batted ratrm' al Verv Unfavorable 94 Soft Message and Traditional Verbs A local politician is interested in your reaction to the following political messages. After eadt. there are some qualms calcern'slg that message Please answer each of the questions concerning each message There are two messages on the front and three on the back. Thank You!!! Becauseicarcabodtberesidentaofthe[Mamisupponamoncomprebensivebeahbcarccovengeforeveryone Tids message is: (For eadl. place a checkmarkthat best reflects how you feel aboutthe message) msrncere _ __ _ __ __ _ __ sincere hard __ m ___ _ _ _ __ __ soil warnilesrted __ _ _ __ _ _ __ cold-hearted loud ,_ 7__ __ __ _ __ _ quiet emotional _ __ _ __ __ _ _ ratimal How important is this Very Very issue to you? (circle one) lnportant lnponant Neutral Uninportant Uninportant How favorable are you Very Very toward this issue? (circle one) Favorable Favorable Neutral Unfavorable Unfavorable ifavorstricterdivorceiawainordertokeepmorefandflestogetberbeeauseindenbdtbevdneofstrongfaralles. Tlls message is: (For each. places chedunarlrthatbed reflects how you feel abouthernaasge) lnarncerc __ ____ * _ __ __ _ sincere hard __ __ _ _ _ _ _ soil warnilarted __ __ _ _ _ _ _ cold-hearted loud _ _ __ _ _ _ _ quid emotional _ ___ ,_ __ ___ _ _ rational How W is this Very Very issue to you? (circle one) inportant lnportant Nannl Uninportant Unirrpcrtant ilow favorable are you Very Verv toward all. issue? (circle one) Favorable Favorable Naltral Unfavorable Unfavorable immfltbecomtrnedoaofmondateprhoassowebavefewercrhinflsontbesuceubecausel-concernedabodtbe safetyofoarcoa-a-rltles. Tbimessage is: (For eadl. place a dredcmarkthat best refleds how you feel aboutthe message) lnsincere __ __ __ __ __ _ __ sincere hard _ 7__ __ __ _ _ _ sofl. warndlesrted _ __ _ __ __ __ _ cold-halted loud +_ ,_ __ _ __ _ _ quid mm al _ _ _ __ __ __ __ mm' al How W is tfls Very Very issue to you? (circle one) lrrportant lrrponant Neutral Uninportant Unirrport-lt How favorable are you Vav Very toward t” issue? (circle one) Favorable Favorable Neutral Unfavorable Unfavorable 95 Becamelcarcabontdleneedsofdlosewhoarcterndnaiiyflliendonekgaflzingphysician-assistedsrricide. This message is: (For each. place a checkmarit that best reflects how you feel about the message) insincere hard warrrd'lesrted loud ernoticnal liowilaportaatistllis issnetoyon? (circleone) How favorableareyorl toward this lune? (circle one) Vay Inportmt Va? Favorable lnportant Favorable Neutral Neutral Uninportant Unfavorable sinca'e soil. cold-hearted quid rational Very Uninportant V63! Unfavorable marncere hard warnilearted loud emotional Howisnportantisthis issnetoyou?(circieone) Howfavorableareyou towardtlrlsissaeNcir-cleone) —_ _ Var important Va? Favorable Inportant Favorable .__ Neutral Neutral This message is: (For each. place a dleckmarlt that best refleds how you feel about the message) Uninportant Unfavorable ladvoeatelegisiatlonofheslthandtasburefitsfordomesdcpartnenbeanseicanaboutbebapphasandwefl-behgofafl sincere soil. cold-hearted quid rational Va? Uninportsnt Va? Unfavorable 96 Hard Message and TraditionaLVerbs A loal politician is irlterdned in your readion to the following political messages. After eadr. there are some questions concaning that message Please answer each of the questions conceming esdl message. There are two masages on the front and three on the back. Thank You!!! Because preventadve health care saves taxpayers money. i support a more comprehensive health care coverage for everyone. This message is: (For eadt. places dleckmarit that bed reflects how you feel aboutthe message) lnsalcere _ __ __ _ _ __ _ sincere hard _ V_ __ __ __ _ __ soil warnilesrted __ _ _ _ _ __ _ cold-hearted loud . _ ___ 7 _ g _ _ _ _ _ quid emotional __ _ _ _ _ _ __ ratialal liow hiportsnt is this Very Very issue to you? (circle one) Inportant inportant Neutral Uninportant Uninportant How favorable are you Very Very toward this issue? (circle one) Favorable Favorable Neutral Unfavorable Unfavorable Ifavoratricterdivorcehwsinordertokeepmonfanrfluwgetberbecaueucbhwsmapncfledmeauofm fanilyunlts. This message is: (For each. places d'leckmark that bed reflects how you feel abouthe message) rns'swae _ _ __ _ __ _ __ sincd'e hard __ __ _ _ _ _ __ soil. warni'lesrted _ __ g“ __ _ __ _ cold-hearted loud _ __ __ _ _ _ _ quid anotiml _ __ p_ _ __ __ _ ratialal How important is this Va'y Very Issue to your? (circle one) lnportant lnponant Neutral Uninportmt llnirlportant How favorable are you Very Very toward this hue? (circle one) Favorable Favorable Ndltral Unfavorable Unfavorable inppontbemndrnedonof-onsateprboumwehvekwercrhhsbontbembeeaueiw‘tosaveibtnretas dohrs. This message is: (For eadt. place a dredtmark that best reflects how you feel abomthemdsage) msrncerc _ _ __ _ _ __ _ sincrse hard __ w _ _ __ __ __ soil warnilesrted _ __ _ _ _ _ __ cold-hearted loud _ __ __ __ 7_ _ _ quid emotional __ __ __ _ _ _ _ ratialal How important is this Very Very issue to you? (circle one) lnportant important Neural Unirrportsnt Uninportna How favorable are you Very Very toward thb issue? (circle one) Favorable Favorable Neural Unfavorable Unfavorable 97 BeamihflevehmechomifigmofpmmchdceieMomkgafldnngkfle This message is: (For each. place a dleckmark that bed reileds how you feel about the message) insincere hard warndlearted loud emotional iiowisnportsntisthis issnetoyou?(circleone) iiow favorable are you toward this issue? (circle one) Very inportmt Very Favorable lrrportsnt Favorable Neutral Ndnral Uninportant Unfavorable sincere soil cold-hearted quid rational Very Unummn' Va? Unfavorable insincere hard warnilearted loud andional Howhrportntlsthis issnetoyonNcircleone) Howfavorableareyon towardthisisane?(circleone) ._— Very Important Vdv v Favorable lnportant Favorable Neutral Neutral 'l'bisrnessageis: (Foredeplacea diedtmarkthatbedrefleds how you feel aboutthemesssge) Uninportarlt Unfavorable ladvoeatelegisiationofhealtbandtasbuteflhfordomesdcpartnenbecauemanledconplurcedvedlesesarnebenems. Van! Uninportmt Va? Unfavorable APPENDIX D MEASUREMENT INSTRUMENT For each of the following questions. circle your response based on the options indicated below. Strongly Strongly Agree Agree Neutral Disagree Disagree 1. l have a lot of interest in politics. I 2 3 4 5 2. I @ about how the outcome of an election affects me. 1 2 3 4 5 3. l have a strong interest in following political mmflgls. l 2 3 4 5 4. I have a good understanding of politics. 1 2 3 4 5 5. Even during nonelection years I follow poli_tical eveng. l 2 3 4 5 6. I’m interested in following the political campaigns of l 2 3 4 5 presidential candidates. 7. l‘m interested in following the political campaigns of l 2 3 4 5 candidates for governor and sena_tor. 8. I’m interested in following the political campaigns of l 2 3 4 5 candidates for state legislature. 9. I believe that understanding mm issues is worthwhile. l 2 3 4 5 10. It’s important to me to know candidatcs’ msitions of l 2 3 4 5 issues. Very Fairly Often Often Sometimes Seldom Never l I. i read newsmpgr stories about campaigns or candidates. 1 2 3 4 5 12. i listen intently to political advertiscmtllts on television. I 2 3 4 5 13. I’ve donated money to a candidate or a campaign. I 2 3 4 5 l4. I've distributed campaign materials for a candidate. I 2 3 4 5 IS. l‘ve volunteered my time for a candidate to work on a l 2 3 4 5 campaign. l6. I’ve tried to oersugde others to vote for one candidate 1 2 3 4 5 over another candidate. 1?. I listen to talk radio to understand more about politics. I 2 3 4 5 18. I watch mlitical TV promms to understand more about 1 2 3 4 5 politics. STOP! WAIT FOR FURTHER INSTRUCTIONS. 98 99 19. Please describe what you consider the memorable aspects of the advertisement. 20. Please list all the ideas/issues the candidate mentioned that you can remember. 21. Please list all the characteristics of the candidate that you can remember. When you were viewing or listening to the message, to what degree did you think about the candidate in the following terms? Think of how strongly you felt about the candidate regarding the continuum of each item. Place a check mark on the line that best reflects how you feel. (If you did not think of the candidate in a specific term. then leave that item blank.) 22. Trustworthy Untrustworthy 23. Qualified Unqualified 24. Emotional Rational 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 41. 42. 100 Dishonest Sympathetic Sincere Close Casual Unattraetive Unfriendly Conservative Excitable Experienwd Knowledgeable Competent __ Strong Warm-Hcarted Intelligent Caring Compassionate Hard Honest Unsympathetic Insincere Distant Formal Attractive Friendly Liberal Calm lnexperienced Unknowledgcable Incompetent Weak Cold-Hcarted Unintelligent Uncaring Uncompassionate 803 Based on what the candidate actually said in the advertisement. please answer the following statements. Circle your response. 43. 44. 45. 46. 47. 48. The candidate favors divorce laws. The candidate believes that individuals have the constitutional right to personal choice. The candidate understands the value of strong families. The candidate wants to construct more state prisons to put criminals behind bars. The candidate favors legislation of health benefits for domestic partners. The candidate endorses physician-assisted suicide. True True True True True True False Unsure False Unaue False Unsure False Unsure False Unsure False Unsure 101 49. The reason the candidate supports domestic partner health benefits is 50. The candidate expressed his compassion for those who are terminally ill. because married couples receive these same benefits. 51. The candidate favors a tax cut for everyone. True False True False True False Unsure Unsure Unsure The following questions, please circle your response. Please circle only gm response for each question. 52. 53. 54. 56. 57. 58. 59. 61. 62. 63. Prior to viewing this message. did you know who the candidate was? YES NO NOT SURE This candidate was running for: US. CONGRESS US SENATE PRESIDENT STATE REPRESENTATIVE Would you vote for this candidate? YES NO NOT SURE 55. If YES. why? Did you vote in the last Presidential election? YES NO Are you registered to vote? YES NO What is your party affiliation? DEMOCRATIC REPUBLICAN INDEPENDENT REFORM What party affiliation did you grow up with? DEMOCRATIC REPUBLICAN INDEPENDENT REFORM . Politically spcaking. how do you consider yourself? LIBERAL MODERATE CONSERVATIVE How old were you on your last birthday? DON’T KNOW LIBERTARI AN OTHER LIBERTARIAN OTHER What is your ethnicity? AFRICAN AMERICAN ASIAN HISPANIC NATIVE AMERICAN WHIrE/CAUCASIAN OTHER (PLEASE SPECIFY) Are you a US. citizen? YES NO 102 64. What ycar in school are you? FRESHMAN SOPHOMORE JUNIOR SENIOR 65. What is your sex? MALE FEMALE 66. In which class are you receiving extra-credit for your participation? TC 275 COM 391 COM 340 OTHER Thank you for your participation! APPENDIX E CODING CATEGORIES FOR THOUGHT-LISTING ITEMS Manipulation Check The purpose of this category is to detennine if there were differing perceptions of the overall tone of the message, such that those respondents exposed to soft messages would perceive the candidate tO be caring in nature, and that those respondents exposed to hard messages would perceive the candidate to be a tough decision-maker. Based on the degree to which “caring” or “tough” were mentioned, phrased were coded as such: I = Tough was mentioned 2 = Inference of tough was mentioned 3 = Mention of neither caring or tough 4 = Inference of caring was mentioned 5 = Caring was mentioned Divorce The purpose of this category is to determine to what degree respondents recalled the details of the stricter divorce laws issue that was presented. The coding category is as follows: 0 = NO mention of the divorce issue 1 = Mention of the divorce issue, but in the wrong direction (such as “not favoring divorce laws”) 2 = Partial mention of divorce issue by use of words or fragments, may also use adjectives, but would not use verbs (such as “divorce” or “stricter divorce laws”) 3 = Complete mention of the divorce issue by use of verbs and/or rationale (such as “supports stricter divorce laws” or “stricter divorce laws because he believes in keeping families together”) 103 104 Partner The purpose of this category is to determine to what degree respondents recalled the details of the benefits for domestic partners issue that was presented. The coding category is as follows: 0 = NO mention of the benefits for domestic partners issue I = Mention of the benefits for domestic partners issue, but in the wrong direction (such as “not favoring partner benefits”) 2 = Partial mention of the benefits for domestic partners issue by use of words orfragments, may also use adjectives, but would not use verbs (such as “domestic partners” or “partner benefits”) 3 = Complete mention of the benefits for domestic partners issue by use of verbs and/or rationale (such as “supports benefits for domestic partners” or “benefits for domestic partners because he believes these couples should get the same benefits that married couples do”) Suicide The purpose of this category is to determine to what degree respondents recalled the details of the physician-assisted suicide issue that was presented. The coding category is as follows: 0 = NO mention of the physician-assisted suicide issue 1 = Mention of the physician-assisted suicide issue, but in the wrong direction (such as “not favoring physician-assisted suicide”) 2 = Partial mention of the physician-assisted suicide issue by use Of words or fragments, may also use adjectives, but would not use verbs (such as “physician suicide” or “suicide”) 3 = Complete mention of the physician-assisted suicide issue by use of verbs and/or rationale (such as “supports physician-assisted suicide” or “physician-assisted suicide because he has compassion for those who are terminally ill”) 105 Additional Issues The purpose of this category was to determine if there were issues that were recalled in addition to the three main issues presented in the advertisement. Additional issues could have been recalled fi'om the rationale presented in the advertisement, such as “family-oriented,” or “believes in rights.” These issues were not coded, there actual number mentioned by the respondent was tabulated. Superficial lmageg The purpose of this category was to determine the number of characteristics of the candidate that the respondent listed. These issues were not coded, there actual number mentioned by the respondent was tabulated. APPENDIX F FACTOR LOADINGS FOR POLITICAL INVOLVEMENT Item fictor l *Polinvl .85 *Polinv2 .58 *Polinv3 .84 *Polinv4 .71 *PolinvS .82 Polinv6 *Polinv7 .78 *Polinv8 .75 * Polinv9 . 5 5 Polinv l O *Pohnvll .74 *Pohnvlz .59 Pohnvl3 Pohnv14 *PothIS .46 *Pohnv16 .66 *Poflnv17 .57 *Pofinv18 .70 * Indicates that the item was used in the political involvement scale. 106 APPENDIX G FACTOR LOADINGS FOR SUBSTANTIVE MATERIAL Item Factor 1 Lactor 2 Divorce .74 .13 Partner .62 -.27 Suicide .77 .07 True/False .52 -.34 Additional Issues .21 .90 107 APPENDIX H FACTOR LOADINGS FOR TOTAL RECALL Item [actor I fictor 2 Divorce .74 -.07 Partner .63 . 10 Suicide .77 -.O8 True/False .52 .08 Additional Issues . I4 -.84 Superficial Imagery . 15 .82 Note: Additional Issues probably loads low onto this factor because few subjects mentioned additional issues. Superficial Imagery probably loads low on this factor as it is the only item measuring characteristics of the candidate. However, the low loadings of these two items do not pose a concern because the purpose of this scale is to measure total recall of information presented in the advertisement and therefore must consider all measures of recall. 108 LIST OF REFERENCES REFERENCES Atkin, C. K., Bowen, L., Nayman, O. B., & Sheinkopf, K. G. (1973). Quality versus quantity in televised political ads. Public Opin_ion Quarterly, 37, 209-224. Basil, M., Schooler, C., & Reeves, B. (1991). Positive and negative political advertising: Effectiveness of ads and perceptions Of candidates. In F. Biocca (ed) Television and Political Advertisi_ng, Volume I, (pp. 245-262). Hillsdale, NJ: Lawrence Erlbaum Associates. Biocca, F. (1991). Viewers’ mental models of political messages: Toward a theory of semantic processing of television. In F. Biocca (ed.) Television and Political Advertising, Volume 1, (pp. 27-89). Hillsdale, NJ: Lawrence Erlbaum Associates. Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source versus message cues in persuasion. Journal of Personalig and Social Psychology, 32, 752-766. Conover, P. J. (1981). Political cues and the perception of candidates. American Politics (hLarterly. 2, 427-448. Diamond, E. & Bates, S. (1992). The spot: The rise of political advertising on television (3rd ed.). Cambridge, MA: The MIT Press. Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Orlando: I-IBJ. F ishbein, M., & Ajzen, I. (1981). Attitudes and voting behavior: An application Of the theory of reasoned action. In G. M. Stephenson & J. M. Davis (Eds), Proggess in applied social pgchology (pp. 253-313). New York: John Wiley. Garramone, G. M. (1983). Issue versus image orientation and effects of political advertising. Corruppnication Research, _1_0, 59 - 76. Garramone, G. M. (1985). Motivation and political information processing: Extending the gratifications approach. In S. Kraus and R. M. Perloff (Eds), Mass media and political thought: An information-processing approach (pp. 201-219). Beverly Hills, CA: Sage. Garramone, G. M., Steele, M. E., & Pinkleton, B. (1991). The role of cognitive schemata in determining candidate characteristic effects. In F. Biocca (Ed.) Television and political advertising (Vol. I, pp. 311-328). Hillsdale, NJ: Lawrence Erlbaum Associates. 109 110 Geiger, S. F ., & Reeves, B. (1991). The effects of visual structure and content emphasis on the evaluation and memory for political candidates. In F. Biocca (Ed) Television and politieaflidvertisarg (Vol. 1, pp. 125-143). Hillsdale, NJ: Lawrence Erlbaum Associates. Gibbs, N., & Duffy, M. (1996, November). Two men, two visions. Time, pp. 30 - 36. Glass, D. P. (1985). Evaluating presidential candidates: Who focuses on their personal attributes? Public Opinion Quarterly 49 517-534. ’_9 Graber, D. A. (1987). Television news without pictures? Critical Studies in Mags Communfirtion. 5, 74-78. Hovland, C. 1., Harvey, O. J., & Sherif, M. (1957). Assimilation and contrast effects in reactions to communication and attitude change. Journal of Abnormal and Social Psychology, _5_§, 244-252. Jamieson, K. H. & Birdsell, D. S. ( 1988). Presidential debates: The challenga pf creating an informed e1eetoga_t§. New York: Oxford University Press. Johnson, B. T. & Eagly, A. H. ( 1989). Efi‘ects of involvement on persuasion: A meta-analysis. Paychological Bulletin. 106. 290-314. Kahneman, D. ( 1973). Attention and effort. Englewood Cliffs, NJ: Prentice Hall. Kahneman, D., Peavler, W. S., & Onuska, L. (1968). Effects of verbalization and incentive on the pupil response to mental activity. Canadian Journal of Psychology, 22, 186-196. Kaid, L. L., & Sanders, K. R. (1978). Political television commercials: An experimental study Of type and length. Commanieation Research, 5, 57-70. Kassarjiart, H. H. (1980). Low involvement: A second look, Proceedings of the Association for Consumer Research, 8, 31-34. Lanzetta, .l. T., Sullivan, D. G., Masters, R. D. McHugo, G. J. (1985). Emotional and cognitive responses to televised images Of political leaders. In S. Kraus and R. M. Perloff (Eds), Mass media and political thought: An informationaprocessing approach (pp. 85-116). Beverly Hills, CA: Sage. Mayar. M. (1958). Madison avenue U. S. A. New York: Harper & Brothers. Mazo, E. (1962). The Great Debates. Santa Barbara, CA: Center for Study of Democratic Institutions. 111 Nimmo, D., & Savage, R. L. ( I976). Qindidates a_r_rd their images. Pacific Palisades, CA: Goodyear. Pedersen, J. T. (1978). Political involvement and partisan change in presidential elections, Americaaloumal of Political Science. 22,, 18-30. Petty, R. E. & Cacioppo, J. T. (1979). Issue involvement can increase or decrease persuasion by enhancing message-relevant cognitive responses. J cum] of Personality and Social Psychology, 37, 1915-1926. Petty, R. E., & Cacioppo, J. T. (1981a). Attitudes and persuasion: Classic and contempom approaches. Dubuque, IA: Brown. Petty, R. E., & Cacioppo, J. T. (1981b). Issue involvement as a moderator of the effects of attitude on advertising content and context. In K. B. Monroe (Ed), Advances in consumer resaaaph (Vol. 8, pp. 20-24). Ann Arbor, MI: Association for Consumer Research. Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: Central and peripheral routes to attitude change. New York: Springer-Verlag. Petty, R. E., Cacioppo, J. T., & Goldman, R. (1981). Personal involvement as a determinant of argument-based persuasion. Journal of Personality and Social Psychology, 4_1, 847-855. Rosenburg, S. W., & McCafferty, P. (1987). The image and the vote: Manipulating voters’ preferences. Public Opinion marterly, fl, 31-47. Salmon, C. T. (1986). Perspectives on Involvement in Consumer and Communication Research. In B. Dervin and M. Voigt (eds) Progress in Communication Sciences, vol. 7, pp. 243-269. Norwood, N .J .: Ablex Publishing Company. Sherif, M., & Cantril, H. (1947). The pgchology of ego-involvements. New York: John “Wiley and Sons. Sherif, M., & Hovland, C. I. (1961). Social judgment: AssimilatioL and contrast effects in communication and attitude change. New Haven, CT: Yale University Press. Sherif, M. & Sherif, C. W. (1967). Attitude as the individual’s own categories: The social judgment-involvement approach to attitude and attitude change. In C. W. Sherif & M. Sherif (Eds), Attitude.ago-involvement, and change (pp. 105-139). New York: “Wiley. Sherif, C. W., Sherif, M., & Nebergall, R. E. (1965). Attitude and attitude change: The socig judgment-involvement approach. Philadelphia: Saunders 112 Shyles, L. (1984). Defining “images” of presidential candidates from televised political spot advertisements. PolitichBelavior. _6_,171-181 Shyles, L. (1986). The televised political spot: Its Structure, content, and role in the political system. In L. L. Kaid, D. Nimmo, & K. Sanders (Eds), New perspectives on political advertising (pp. 107 - I38). Carbondale, IL: Southern Illinois University Press. Stiff, J. B. (1986). Cognitive processing of persuasive message cues: A meta- analytic review of the effects of supporting information on attitudes. Comraumgtion Monogzaphs, 5;, 75-89. Stiff, J. B. (1994). Persuaaive communication. New York: The Guilford Press. Tan, A. S. (1980). Mass media use, issue knowledge, and political involvement, Public Opinion Quarterly, 53, 241-248. Thorson, E., Christ, W. G., & Caywood, C. ( I991). Selling candidates like tubes Of toothpaste: Is the comparison apt? In F. Biocca (Ed) Television and political advertising, (Vol. I, pp. 145-172). Hillsdale, NJ: Lawrence Erlbaum Associates. Wykoff, G. (1968). Image Candidates. New York: The Macmillan Company. I‘IICHI hurrah \