<%BANNER%>

Interplay of Framing Tactic, Framing Domain, and Source Credibility in Direct to Consumer Hormone Replacement Therapy Ad...

Permanent Link: http://ufdc.ufl.edu/UFE0041165/00001

Material Information

Title: Interplay of Framing Tactic, Framing Domain, and Source Credibility in Direct to Consumer Hormone Replacement Therapy Advertising An Integration of Prospect Theory and Language Expectancy Theory
Physical Description: 1 online resource (119 p.)
Language: english
Creator: Kim, Kenneth
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: advertising, credibility, dtc, framing, language, persuasion, prospect, source
Journalism and Communications -- Dissertations, Academic -- UF
Genre: Mass Communication thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy INTERPLAY OF FRAMING TACTIC, FRAMING DOMAIN, AND SOURCE CREDIBILITY IN DIRECT-TO-CONSUMER HORMONE REPLACEMENT THERAPY ADVERTISING: AN INTEGRATION OF PROSPECT THEORY AND LANGUAGE EXPECTANCY THEORY By Kenneth E. Kim May 2010 Chair: Debbie Treise Major: Mass Communication The present study attempts to explore the interactive effects among the gain-loss framing domain, the attribute-goal framing tactic, and message source credibility on the persuasive outcomes associated with Direct-to-Consumer (DTC) hormone replacement therapy advertising. An experiment was designed with a 2 (framing tactic: attribute framing versus goal framing) times 2 (framing domain: gain framing versus loss framing) times 2 (source credibility: high versus low) between-subjects design, exploring the interactive effects of framing tactic and framing domain on the consumer?s attitude toward hormone replacement therapy and DTC ad-promoted behavior intentions. Women, aged 45-65 were recruited for the study samples. The data obtained revealed that loss framing was affected by the level of source credibility such that the loss framing impact on women?s attitudes toward a HRT and intentions to choose it decreased with a low credible source, while the gain framing impact was not affected by message source credibility. This is consistent with the postulates in the language expectancy theory suggesting that low credible sources are limited in using message strategies such that low source credibility with positive (gain) message strategies may have as much persuasive impact as high source credibility. The study findings also suggest that the relative advantage of DTC drug advertising framing loss may be dissipate when a loss-framed advertisement was sponsored by a low credible source.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kenneth Kim.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Treise, Deborah M.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041165:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041165/00001

Material Information

Title: Interplay of Framing Tactic, Framing Domain, and Source Credibility in Direct to Consumer Hormone Replacement Therapy Advertising An Integration of Prospect Theory and Language Expectancy Theory
Physical Description: 1 online resource (119 p.)
Language: english
Creator: Kim, Kenneth
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: advertising, credibility, dtc, framing, language, persuasion, prospect, source
Journalism and Communications -- Dissertations, Academic -- UF
Genre: Mass Communication thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy INTERPLAY OF FRAMING TACTIC, FRAMING DOMAIN, AND SOURCE CREDIBILITY IN DIRECT-TO-CONSUMER HORMONE REPLACEMENT THERAPY ADVERTISING: AN INTEGRATION OF PROSPECT THEORY AND LANGUAGE EXPECTANCY THEORY By Kenneth E. Kim May 2010 Chair: Debbie Treise Major: Mass Communication The present study attempts to explore the interactive effects among the gain-loss framing domain, the attribute-goal framing tactic, and message source credibility on the persuasive outcomes associated with Direct-to-Consumer (DTC) hormone replacement therapy advertising. An experiment was designed with a 2 (framing tactic: attribute framing versus goal framing) times 2 (framing domain: gain framing versus loss framing) times 2 (source credibility: high versus low) between-subjects design, exploring the interactive effects of framing tactic and framing domain on the consumer?s attitude toward hormone replacement therapy and DTC ad-promoted behavior intentions. Women, aged 45-65 were recruited for the study samples. The data obtained revealed that loss framing was affected by the level of source credibility such that the loss framing impact on women?s attitudes toward a HRT and intentions to choose it decreased with a low credible source, while the gain framing impact was not affected by message source credibility. This is consistent with the postulates in the language expectancy theory suggesting that low credible sources are limited in using message strategies such that low source credibility with positive (gain) message strategies may have as much persuasive impact as high source credibility. The study findings also suggest that the relative advantage of DTC drug advertising framing loss may be dissipate when a loss-framed advertisement was sponsored by a low credible source.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kenneth Kim.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Treise, Deborah M.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041165:00001


This item has the following downloads:


Full Text

PAGE 1

1 en-US INTERPLAY OF FRAMING TACTIC, FRAMING DOMAIN, AND SOURCE CREDIBILITY IN DIRECT TO CONSUMER HORMONE REPLACEMENT THERAPY ADVERTISING: en-US AN INTEGRATION OF PROSPECT THEORY AND LANGUAGE EXPECTANCY THEORY en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US By en-US en-US KENNETH E. KIM en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY en-US en-US UNIVERSITY OF FLORIDA en-US en-US 2010 en-US

PAGE 2

2 en-US en-US en-US en-US 2010 Kenneth E. Kim en-US en-US en-US

PAGE 3

3 en-US en-US en-US en-US To my beloved family en-US en-US en-US en-US

PAGE 4

4 en-US ACKNOWLEDGMENTS en-US This dissertation simply would not have been possible without the support and encouragement of many people. In particular, I would like to thank my advisor, Dr. Debbie Treise, for her invaluable advice and direction through my entire graduate career. Dr. Marilyn Roberts and Dr. Mike Weigold provided indispensible advice and support during my degree process. Dr. Virginal Dodd inspired me an excellent research topic in media and consumer health and assisted me in thinking about theory and practice. en-US I must express my thanks to my valued colleagues, mentors, and friends. I was so lucky to work with Jinseong and Dr. Jorge Villegas and they deserve considerable thanks for their valuable advice, brilliant input, and years of friendship. I also wish to thank Dr. John Sutherland and Dr. Cynthia Morton who passed on their teaching skills to me. Jody Hedge is a highly respectable friend at the UF and I owe my sincere gratitude to her valuable support and encouragement. en-US Last but not least I would like to express my special thanks to my beloved wife, Hye Eun my daughters, Cayla, and Kaylee for their sacrifice and support during my graduate process My father, Dr. Young Jin Kim also lavishes me with affection throughout my whole life. I love you all. en-US My dissertation is dedicated to my mo m, who passed away in 1997. I miss her deeply. en-US

PAGE 5

5 en-US TABLE OF CONTENTS en-US page en-US ACKNOWLEDGMENTS ........................................................................................................... 4en-US en-US TABLE OF CONTENTS ............................................................................................................ 5en-US en-US LIST OF TABLES ...................................................................................................................... 8en-US en-US LIST OF FIGURES .................................................................................................................... 9en-US en-US ABSTRACT ............................................................................................................................. 10en-US en-US CHAPTER en-US 1 en-US INTRODUCTION ............................................................................................................. 12en-US en-US Framing Effect in Judgment and Decision Making ............................................................. 12en-US en-US Limitations of Current Framing Literature .......................................................................... 14en-US en-US Research Purpose ............................................................................................................... 17en-US en-US Experiment Context: Direct to Consumer Hormone Replacement Therapy Advertising ...... 19en-US en-US Overview of Study ............................................................................................................. 19en-US en-US 2 en-US LITERATURE REVIEW ................................................................................................... 20en-US en-US Framing: Prospect Theory Perspective ............................................................................... 20en-US en-US Confounding Framing Effects ............................................................................................ 25en-US en-US Attribute Framing and Goal Framing .................................................................................. 26en-US en-US Other Moderators of Framing Effects in Health and Marketing Decisions .......................... 31en-US en-US Message Source Credibility ................................................................................................ 34en-US en-US Mechanisms of Message Source Impact ............................................................................. 35en-US en-US Gain-Loss Framing Effect and Source Credibility .............................................................. 36en-US en-US Source as a Moderator of Framing Effect: Language Expectancy Perspective ..................... 38en-US en-US Hypotheses and Research Question .................................................................................... 40en-US en-US 3 en-US METHODOLOGY ............................................................................................................. 44en-US en-US Design ................................................................................................................................ 44en-US en-US Participants ........................................................................................................................ 45en-US en-US Research Stimuli and Drug Category Selection .................................................................. 45en-US en-US Procedure ........................................................................................................................... 46en-US en-US Independent Variables ........................................................................................................ 46en-US en-US Message Framing Manipulations ................................................................................. 46en-US en-US Framing Manipulation Check Items ............................................................................. 47en-US en-US Source Credibility Manipulations ................................................................................ 48en-US en-US Pretest ................................................................................................................................ 49en-US

PAGE 6

6 en-US Dependent Measures .......................................................................................................... 49en-US en-US Attitude toward Hormone Replacement Therapy ......................................................... 49en-US en-US DTC ad-promoted Behavior Intentions ........................................................................ 49en-US en-US Blocking Variables ............................................................................................................. 50en-US en-US Attitude toward General Hormone Medications ........................................................... 51en-US en-US Personal Involvement .................................................................................................. 51en-US en-US Knowledge about Hormone Replacement Therapy ...................................................... 51en-US en-US Perceived Risk of Side Effects ..................................................................................... 51en-US en-US Previous Experience with Hormone Replacement Therapy .......................................... 52en-US en-US 4 en-US RESULTS .......................................................................................................................... 53en-US en-US Demographics .................................................................................................................... 53en-US en-US Group Size .................................................................................................................. 53en-US en-US Responden ....................................................................................... 53en-US en-US Manipulation Checks for Independent Variables ................................................................ 55en-US en-US Gain-Loss Framing Domain ........................................................................................ 55en-US en-US Attribute-Goal Framing Tactic .................................................................................... 56en-US en-US Source Credibility ....................................................................................................... 58en-US en-US Reliability Check for Dependent Variables and Blocking Variables .................................... 59en-US en-US Attitude toward Hormone Replacement Therapy ......................................................... 59en-US en-US Intention for Information-Seeking ............................................................................... 59en-US en-US Intention to Choose Hormone Replacement Therapy ................................................... 59en-US en-US General Attitude toward Hormonal Drugs ................................................................... 60en-US en-US Knowledge about Hormone Replacement Therapy ...................................................... 60en-US en-US Involvement ................................................................................................................ 60en-US en-US Perceived-risk of Side Effect ....................................................................................... 60en-US en-US Correlations among Variables ............................................................................................. 61en-US en-US Hypotheses Testing ............................................................................................................ 62en-US en-US Hypotheses 1a~d ......................................................................................................... 64en-US en-US ANCOVA Contrast Analyses with Framing Domain and Framing Tactic ............. 66en-US en-US Hypotheses 2a~d ......................................................................................................... 67en-US en-US ANCOVA Contrast Analyses with Gain-loss Framing Domain and Source Credibility ......................................................................................................... 69en-US en-US Research Question ....................................................................................................... 72en-US en-US Additional Testing .............................................................................................................. 73en-US en-US ANCOVA Contrast Analyses with Gain-loss Framing Domain, Attribute-goal Framing Tactic, and Source Credibility .................................................................... 73en-US en-US MANOVA Results ...................................................................................................... 74en-US en-US Summary of the Results ..................................................................................................... 75en-US en-US 5 en-US DISCUSSION .................................................................................................................... 77en-US en-US Summary of Study ............................................................................................................. 77en-US en-US Evaluation of Findings ....................................................................................................... 81en-US en-US Hypotheses 1a~d ......................................................................................................... 81en-US en-US Hypotheses 2a~d ......................................................................................................... 83en-US

PAGE 7

7 en-US Implication for Framing Theory .................................................................................. 86en-US en-US Implication for Message Framing and Source Credibility. ........................................... 87en-US en-US Implication for Drug Advertising Practices .................................................................. 88en-US en-US Limitations ......................................................................................................................... 89en-US en-US Suggestions for Future Research ........................................................................................ 90en-US en-US APPENDIX en-US A en-US GOAL GAIN-FRAMED AD WITH HIGH SOURCE CREDIBILITY ............................... 93en-US en-US B en-US G OAL LOSS-FRAMED AD WITH HIGH SOURCE CREDIBILITY ............................... 94en-US en-US C en-US ATTRIBUTE LOSS-FRAMED AD WITH HIGH SOURCE CREDIBILITY .................... 95en-US en-US D en-US ATTRIBUTE GAIN-FRAMED AD WITH HIGH SOURCE CREDIBILITY .................... 96en-US en-US E en-US ATTRIBUTE GAIN-FRAMED AD WITH LOW SOURCE CREDIBILITY ..................... 97en-US en-US F en-US Goal GAIN-FRAMED AD WITH LOW SOURCE CREDIBILITY ................................... 98en-US en-US G en-US ATTRIBUTE LOSS-FRAMED AD WITH LOW SOURCE CREDIBILITY ..................... 99en-US en-US H en-US Goal LOSS-FRAMED AD WITH LOW SOURCE CREDIBILITY ................................ 100en-US en-US I en-US MAIN STUDY QUESTIONNAIRE ................................................................................. 101en-US en-US J en-US DEBRIEFING STATEMENT .......................................................................................... 110en-US en-US LIST OF REFERENCES ........................................................................................................ 111en-US en-US BIOGRAPHICAL SKETCH ................................................................................................... 119en-US en-US en-US

PAGE 8

8 en-US LIST OF TABLES Table page en-US en-US 3-1 Ad message framing manipulations .......................................................................... 48en-US en-US 4-1 Number of participants by experimental groups ........................................................ 54en-US en-US 4..................................................................................... 54en-US en-US 4.................................................................................... 54en-US en-US 4-4 One-way ANOVA results for the gain-loss framing domain manipulation check ...... 56en-US en-US 4-5 One-way ANOVA results for the attribute-goal framing tactic manipulation check ....................................................................................................................... 57en-US en-US 4-6 One-way ANOVA results for the source credibility manipulation check ................... 58en-US en-US 4-7 Full-factorial ANOVA results for the source credibility manipulation check ............. 59en-US en-US 4-8 Correlations among variables ................................................................................... 62en-US en-US 4-9 MANCOVA tests for faming domain, framing tactic, and source credibility ............. 65en-US en-US 4-10 Results of between-subject test ................................................................................. 68en-US en-US 4-11 Means and standard deviations for framing domain and source credibility for attitude toward HRT ................................................................................................ 70en-US en-US 4-12 Means and standard deviations for framing domain and source credibility for Intention to choose HRT ........................................................................................... 72en-US en-US 4-13 ANCOVA contrast tests for loss framing-low credibility condition ............................ 73en-US en-US 4-14 ANCOVA contrast tests for gain framing-low credibility condition ........................... 73en-US en-US 4-15 MANOVA tests for faming domain, framing tactic, and source credibility ................ 75en-US en-US en-US

PAGE 9

9 en-US LIST OF FIGURES Figure page en-US 2-1 Expected interaction effect of gain-loss framing domain and attribute-goal framing tactic on attitude toward HRT ........................................................................... 42en-US en-US 2-2 Expected interaction effect of gain-loss framing domain and attribute-goal framing tactic on DTC ad-promoted intentions .............................................................. 42en-US en-US 2-3 Expected interaction effect of gain-loss framing domain and source credibility on attitude toward HRT ................................................................................................ 43en-US en-US 2-4 Expected interaction effect of gain-loss framing domain and source credibility on DTC ad-promoted intentions ..................................................................................... 43en-US en-US 4-1 Interaction effect of gain-loss framing domain and source credibility on attitude toward HRT ................................................................................................................... 70en-US en-US 4-2 Interaction effect of gain-loss framing domain and source credibility on intention to choose HRT. .............................................................................................................. 72en-US en-US en-US

PAGE 10

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy INTERPLAY OF FRAMING TACTIC, FRAMING DOMAIN, AND SOURCE CREDIBILITY IN DIRECTTO -CONSUMER HORMONE REPLACEMENT THERAPY ADVERTISING: AN INTEGRATION OF PROSPECT THEORY AND LANGUAGE EXPECTANCY THEORY By Kenneth E. Kim May 2010 Chair: Debbie Treise Major: Mass Communication The present study attempts to explore the interactive effects among the gain-loss framing domain, the attribute-goal framing tactic, and message source credibility on the persuasive outcomes associated with Directto -Consumer (DTC) hormone replacement therapy advertising. An experiment was designed with a 2 (framing tactic: attribute framing versus goal framing) 2 (framing domain: gain framing versus loss framing) 2 (source credibility: high versus low) between-subjects design, exploring the interactive effects of framing tactic and framing domain -promoted behavior intentions. Women, aged 45-65 were recruited for the study samples. The data obtained revealed that loss framing was affected by the level of source credibility such that the loss framing impact on attitudes toward a HRT and intentions to choose it decreased with a low credible source, while the gain framing impact was not affected by message source credibility. This is consistent with the postulates in the language expectancy theory suggesting that low credible sources are limited in using message strategies such that low source credibility with positive (gain) message strategies may have as much persuasive impact as high source

PAGE 11

credibility. The study findings also suggest that the relative advantage of a loss-framed ad may dissipate when the loss-framed ad is endorsed by a low credible source.

PAGE 12

12 en-US CHAPTER 1 en-US INTRODUCTION en-US Framing Effect in Judgment and Decision Making en-US In cognitive and decision science, there is a large body of literature reporting that one or a combination of communication variables (e.g., message itself, message source, and message recipient) can affect behavioral changes (see Perloff, 2003; Dillard, & Pfau, 2002; Kahneman & Tversky, 2000). Of these communication elements, the concept of framing as psychological processing of persuasive messages, has received a great deal of attention in the areas of health decision-making (Salovey & Williams-Piehota, 2004; Williams, Clarke, & Borland, 2001; Block & Keller, 1995) and marketing communications (Shiv, Britton, & Payne, 2004; Grewal, Gotlieb, & Marmorstein, 1994; Chang, 2007). Despite the confusion about its conceptual and operational definitions among scholars in different fields (see Scheufele & Tewksbury, 2007; Maule & Villejoubert, 2007), framing scholars generally agree that message framing refers to the ways in which a choice problem is phrased, and framing effect is concerned with how individuals mental representations, partly built by message framing, affect behavioral outcomes such as preference, judgment, choice (Kahneman, 2003; Tversky & Kahneman, 1981), attitude (Smith, 1996; Smith & Petty, 1996) and behavior intention (Meyerowitz & Chaiken, 1987; Jones, Sinclair & Courneya, 2003). en-US The concept of framing is essential to prospect theory, a behavioral decision theory developed by Kahnamen and Tversky (1979). Prospect theory describes how an individual judgment can be affected by seemingly different presentations of a choice problem in the risky or uncertain decision context According to prospect theory, relative value (e.g., gain versus loss) of outcomes is considered more important than objective evaluation of the end status of outcome in a risky decision-making process. Further, when a risky or uncertain choice option is presented

PAGE 13

13 en-US with loss outcomes, people tend to take the risky (uncertain) option to avoid the losses, while people are more likely to stay away from the option when presented with gain outcomes. Prospect theory postulates that negative feelings or perceptions such as displeasure, loss aversion, anxiety provoked by loss outcomes are more distinct than commensurate positive feelings such as pleasure or perceived benefit evoked by gain outcomes, and lead people to choose loss-framed choice option (Tversky & Kahneman, 1986; Kahneman, 2005). en-US The gain versus loss framing (presentation) domain is employed as a typical framing manipulation within the prospect theory framework. A gain-framed message emphasizes the positive prospects that result from performing the choice behavior (e.g., the chance of contracting cervical cancer decreases by 80 percent if a female receives the human papillomavirus vaccination), while a loss-framed message accentuates the negative outcomes caused by a failure to comply with the choice option (e.g., the chance of contracting cervical cancer increases by 20 percent unless you receive the human papillomavirus vaccination now). en-US Si nce Kahnamen and Tversky (1979) and Tversky and Kahneman (1981; 1984; 1986) first applied the framing concept as one of the decision biases (e.g., heuristics such as prejudice, source cues, etc.) to the areas of economic and medical policy decisions, a large body of evidence has developed in support of the central notion of the framing effect; people preference or judgment are more likely to be impinged by loss fram ed -information than the commensurate gain framed-information under uncertain or risky decision making activities (Kahnamean & Tversky, 2000). This framing effect has been observed in various disciplines, including health decision making (Salovey, Schneider, & Apanovitch, 2002; Rothman, Bartels, Wlaschin, & Salovey, 2006; Scott & Curbow, 2006; Meyerowitz and Chaiken, 1987); consumer goods purchase decisions (Berger & Smith, 1998; Smith, 1996; Chang 2007; Smith & Wortzel, 1997)

PAGE 14

14 en-US political decisions (Boettcher, 2004; Brewer & Gross, 2005) and public policy decisions (N an, 2007; Shah, Kwak, Chmierbach, & Zubric, 2004). en-US Limitations of Current Framing Literature en-US Despite the pervasiveness of the gain-loss framing effects on persuasion and choice behavior in diverse disciplines, many have criticized framing literature for its failure to reach a consensus that supports the robustness of loss advantage (Fagley & Miller, 1987; Zhang & Buda, 1999; Frisch, 1993; Salovey et al., 2002; Levin, Schneider, & Gaeth, 1998). A number of studies that demonstrate gain framing is more effective than loss framing, or that there is no difference between two, have been published (Martear, 1989; Levin, 1987; Levin, Chapman, & Johnson, 1988; Salovey et al., 2002). en-US To resolve this confusion, framing scholars have identified moderators of the framing effect. For example, gain-loss message framing interacts with other message-level variables, such as the framing presentation order (Buda, 2003) and different levels of behavioral outcome (e.g., individual versus societal level outcomes ; Shah et al, 2004) and contextual features relevant to the choice problems or objects, such as the temporal proximity of an event (e.g., whether outcome events occur in the near or distant future ; McElroy & Mascari, 2007), hypothetical versus real decision context (Kuhberger, Schulte-Mecklenbeck, & Perner, 2002), the type of product category (Chang, 2007; Smith 1996), and the type of health promotio n behaviors (e.g., detection versus prevention behaviors) (see Salovey et al, 2002; Rothman et al, 2006). en-US Framing effects are also influenced by individual variables, such as issue involvement (Maheswaran & Meyers-Levy, 1990), the need for cognition (Smith & Levin, 1996; Zhang & Buda, 1999), perceived health risk factors (Scott et al, 2006), perception of ambivalence on choice problems (Broemer, 2002) and education level (Smith, 1994).

PAGE 15

15 en-US Some criticisms about these mixed findings may point out different dependent variables measured as framing outcomes (see Levin et al., 1998). Although the original work on framing effect considered the relationship between framing and choice or preference in hypothetical situations, subsequent studies have taken diverse persuasion variables, including attitude toward choice problems, intentions to perform the actions, cognitive responses, and risk perceptions as outcomes of framing effect, which might lead to the controversy on the framing effect (see Maule & Villejoubert, 2007). en-US With respect to framing outcomes, however, Kahneman, one of pioneers in the decision sition on the framing outcome variables and the large body of empirical findings that include the aforementioned diverse outcome variables, it seems the mixed findings may not be attributable to the selection of dependent variables. en-US Recent work directs attention to the difference in framing manipulations as a possible explanation for the reversed effect (Krishnamurthy, Carter & Blair, 2001; Levin et al, 2002) While the original framing research based on prospect theory focused on the difference in outcome presentation of a choice problem highlighting a gain or loss domain, many scholars have defined loosely that the framing manipulation has been fragmented in message framing literature (Scheufele & Tewksbury, 2007; Krishnamurthy et al, 2001) Specifically, some research focuses on the attributes of a choice problem or choice object in a positive or negative way (e.g., 80 % of the users of this vaccination do not experience side effects versus 20% of the users of this vaccination experience side effects), other research deals with different outcomes or goals of a choice problem or object in a gain (positive) or loss

PAGE 16

16 en-US (negative) sense (e.g., taking this vaccination will prevent you from contracting the disease and other infections versus by not taking this vaccination, you may suffer from the disease and other infections) as framing manipulations. Each approach implies an equivalent expected value of the choice options, but the difference only exists in the selection of framing tactic (whether to focus on attribute or goal of a choice problem) and framing domain (gain versus loss) Recent framing literature has reported empirical evidence that framing effects are moderated by the selection of this framing tactic (Ferguson & Gallagher, 2007 ; Levin et al., 2002). This conclusion and its evidence deserve attention because they not only provide theoretical implications for developing robust framing hypotheses dependent on both framing tactic (attribute framing versus goal framing) and domain (gain fr aming versus loss framing), but they may also have important practical implications for developing communication strategy in medical marketing areas such as direct to consumer (DTC) pharmaceutical advertising. en-US The second limitation is that while individual characteristics have received significant attention as mediating or moderating variables in framing literature, only a handful of studies have illuminated the role of message source credibility in the relationship between framing and persuasion outcomes. Given that the source factor is considered an important contextual factor for persuasion and judgment (Perloff, 2003), which may play a part with framing, particularly in medical marketing communications, the source factor deserves attention. More importantly, previous research that examined the source credibility as a moderator of message framing effect fails to provide a sound theoretical framework that accounts for how the source factor interacts with framing. en-US Finally, little research has been published on the framing effect based on prospect theory in the context of DTC advertising. DTC advertising has unique characteristics that are attractive

PAGE 17

17 en-US to framing researchers because it intersects the health and advertising domains. It is possible to observe various judgment and persuasion variables relevant to health-behavioral changes, including intention to seek more information on the treatment and disease, intention to consult with health care providers to get help, intention to choose the treatment option and attitude toward the treatment option (Huh, DeLorme, & Reid, 2005). Understanding how framing DTC advertising works on these variables with source credibility may help DTC drug marketers and health professionals develop effective account strategies. en-US Research Purpose en-US The purpose of the present study is to address the limitations discussed above in the framing literature. It attempts to investigate the interaction between framing tactic and framing domain in the context of DTC advertising and explore the moderating role of source credibility in the relationship between framing and framing outcomes such as persuasion and decision making. Specifically, the study explores the interactive effect of the gain-loss framing domain with the attribute-goal framing tactic and source credibility with respect to DTC ad -promoted behavioral intentions such as intention for information search, intention to consult a doctor, intention to choose the drug and attitude toward the drug category (Huh et al., 2005). Few studies have investigated the interactive effect of framing tactic and framing domain in the context of prescription drug decision making that involves risky or uncertain elements (e.g., hormone replacement medicine; emergency contraceptive; medical abortion pills). en-US To develop the hypotheses, this project integrates two theoretical perspectives: prospect theory (Kahneman & Tversky, 1979; Tversky & Kahnemen, 1981) and language expectancy theory (Burgoon 1995; Burgoon, Denning, & Robers, 2002) The former is adopted to account more susceptible to loss outcome presentation than commensurate

PAGE 18

18 en-US gain and the latter provides an explanation of how the credibility of a message source affects the gain-loss framing effect on persuasion. en-US Among the many frameworks in persuasion and decision literature, this study employs language expectancy theory in order to understand the moderating role of source credibility in the relationship between message framing and behavioral intentions and attitude. According to this theory, the selection of message tactics is governed by social and cultural norms, and the persuasiveness of a message is affected by the relationship between the expectations people have for the message strategies and the characteristics of the message source (Burgoon, 1995; 1996) Specifically, low-credibility sources have limited freedom to choose message or language tactics relative to high-credibility sources (Burgoon et al, 2002) The theory posits that negative information focusing on loss, fear, or risks is regarded as inappropriate for low-credibility sources, and limited persuasion takes place in this condition (Burgoon, 1995) Accordingly, it is predicted that the loss (negative) framing effect diminish es with low source credibility. Integration of prospect theory with language prospect theory will contribute to our understating of how message source operates with gain-loss framing. en-US Finally, few studies have observed framing effects in the context of DTC drug advertising. As discussed above, the impact of framing DTC drug advertising is not confined to judgment on drug choice, and it has implications for other health-related consumer behavioral decisions (e.g., help-seeking behaviors such as information search and communication with doctors) as well as attitude toward prescription drugs. This framing and source credibility in DTC advertising message structure will contribute to understanding DTC drug marketing strategy for facilitating DTC ad -promoted behaviors.

PAGE 19

19 en-US Experiment Context: Direct to Consumer Hormone Replacement Therapy Advertising en-US Hormone replacement therapy is chosen for this study for the following reasons. First, it has been generally recognized as a risky or uncertain health decision choice for postmenopausal women (McIntosh & Blalock, 2005; Andrist, 1998). Second, empirical evidence has shown that consumers generally perceive hormonal medicines (e.g., steroid, hormonal contraceptives) as riskier than other forms of medication (see nnor, 2005). Third, recent studies have reported a great deal of new information publicized from mass media that counters or supports decisions to initiate or continue the hormone treatment (Wathen, 2006; Burger, 2006). Finally, there is little research in the framing and DTC drug advertising literature focusing on the hormone replacement therapy decision. en-US Overview of Study en-US An experiment is designed with a 2 (framing tactic: attribute versus goal framing) 2 (framing domain: gain versus loss framing) 2 (source credibility: high versus low) betweensubjects design, exploring the interactive effects of framing tactic, framing domain and source credibility on the DTC ad -promoted behavior int en tions and attitude toward hormone replacement therapy. en-US

PAGE 20

20 en-US CHAPTER 2 en-US LITERATURE REVIEW en-US This literature review first examines prospect theory (Kahneman & Tversky, 1978), with which the underlying process of the framing effect under uncertainty and risk is illuminated. Prospect theory focuses on gain-loss framing, loss aversion, and risk attitude (risk-seeking versus risk-averse) as major concepts (Tversky & Kahneman, 1986). Next, a discussion on the mixed findings on framing effects is provided. Third, the conceptual difference between attribute framing and goal framing, and the differential effects of the two are discussed (Levin et al., 1998). Fourth, the message source as a moderator between framing and its outcomes is discussed relative to the language expectancy theory (Burgoon et al., 2002). Last, hypotheses are proposed on the interactive effects of gain-loss framing and attribute-goal framing, and gain-loss framing and source credibility on the attitude toward the DTC ad drug category (hormone replaceme nt therapy, HRT ) and intentions for additional information seeking about the drugs, consulting with doctors for help, and drug choice en-US Framing: Prospect Theory Perspective en-US A frame is defined as ncies and Tversky (1979), the pioneers of the concept in decision science, posited frame as being an ntal construct that is established both by internal and external processes (e.g., experience, norm, knowledge, information structure, etc.), the term has been loosely defined to refer to either the message presentation domain (gain versus loss) or the mental reference point (see Maule & Villejoubert, 2007; Kahneman, 2000). Despite this confusion in conceptual definitions of the term, there is general agreement that framing is referred to as the methods in which choice problems or decision objects are presented and how these influence the likelihood

PAGE 21

21 en-US of decision making, including positive or negative evaluations of decision objects, behavioral intentions, and judgment, as well as actual choice (Kahnemen, 1991; Kahneman, 2003 ; ). en-US Prospect theory provides a framework for understanding how people make judgments under uncertain and risky contexts. The prospect theory was first proposed to challenge expected utility theory, the normative view that dominated the early risky decision paradigm in which reason-based judgments were believed to co -making process (Hastie & Dawes, 2001; Kahneman, 1991). Under this paradigm, people are assumed to perform consistent choice behaviors regardless of how the choice prospect is presented to them (Tversky & Kahneman, 1986). A reason for this claim is the assumption that underlying beliefs are never changing and consistent with their choices and preferences (Newell, Lagnado, & Shanks 2007). Prospect theory rejects this claim of invariance (Kahneman, 1991) and claims instead that human beliefs are often inconsistent with actual choices because people are susceptible to contextual situations or subjective reference points relative to choice outcomes (Tversky & Kahneman, 1986). In other words, Kahneman and Tversky argue that the relative value of outcomes represented in the mind is more important (e.g., gain versus loss) in the decision process than is the objective evaluation of the final status of outcome. Kahneman and Tversky emphasize the subjective value function in decision-making and account for this process by observing that people are prone to make choices that contradict their belief systems with the notion of psychological heuristics, such as loss aversion and status quo, with which a risky choice is made or abandoned (Kahneman, 1991; Newell et al., 2007). en-US Prospect theory suggests that framing is a key determinant that guides individual behavioral choices and describes framing as the status of deviation from psychological and

PAGE 22

22 en-US subjective reference points with respect to behavioral outcomes, and the deviations are considered in terms of relative gains or relative losses (Kahneman & Tversky, 1984; Hogarth, 1987). Original framing studies manipulated the choice options in terms of gain versus loss outcomes in either sure options or probability options (e.g., 200 (400) people will be saved (die) versus a 1/3 probability that 600 people will be saved (nobody will die). Later, a group of scholars dubbed this framing manipulation (sure versus gamble option) risk-choice framing (Levin et al., 1998). A general finding from this original research and subsequent replications in the framing literature is that loss framing weighs more in risk choice than does gain framing (see Kahneman & Tversky, 2000). The following is a most frequently cited example of framing effect in the context of a medical policy decision (Tversky & Kahaneman, 1981): en-US Decision Problem: Asian Disease en-US en-US Gain Framing Domain: en-US If Program A is adopted, 200 people will be saved (72 percent). en-US en-US If Program B is adopted, there is a one-third probability that 600 people will be saved, and a two-third probability that no one will be saved (28 percent). en-US en-US Loss Framing Domain: en-US If Program C is adopted, 400 people will die (22 percent). en-US en-US If Program D is adopted, there is a one-third probability that nobody will die, and a twothirds probability that 600 people will die (78 percent). en-US en-US In the example above, the statement that is composed of probabilities (e.g., two-thirds probability that all 600 people will die) involves the element of risk or uncertainty. Most respondents chose program A, the sure option in gain domain. According to the prospect theory, when gain outcomes are presented for risky options (Program B), people tend to orient toward the status quo (Kahneman, 2003). In other words, among people prone to consider the potential losses as well as the gains caused by taking the choice option, the losses (two-thirds probability

PAGE 23

23 en-US that all people will die) are more likely to be outweighed by the gains (a one-third probability that 600 people will be saved) (Tversky & Kahneman, 1981). Therefore, people tend to stay away from the risky choice as opposed to taking the option to jeopardize themselves in potential losses (Kahneman, 2003) s mind, and people seize the sure gain rather than the risky option presented with probabilities (Salovey et al., 2002). In contrast, when given the option between programs C and D, people chose option D, the risky option, instead of the sure loss. The prospect theory holds that when loss framing is offered, people are willing to accept the risky option to minimize or eliminate the potential losses presented to them. en-US To account for the prevalence of loss framing over gain framing, Kahneman and Tversky adopt concepts, such as reference points, subjective value function, risk perspectives (risk-taking versus risk-aversion), and loss aversion (Newell et al., 2007). A reference point exists when a person evaluates the outcomes of a choice problem, and the person evaluates the outcomes in terms of gains or losses relative to that reference point (Kahaneman, 2003). Prospect theory suggests that evaluation of choice problems is sensitive to the changes of the outcome presentations (Tversky & Kahneman, 1986), thus framing effects within the prospect theory arise through manipulation of this reference point (Newell et al., 2007). en-US People tend to evaluate semantically equivalent behavioral options as having either a gain or loss prospect relative to the reference point, and this mental operation, evoked by message framing, motivates different risk perspectives; either risk-aversion in the gain-framing domain, or risk-taking in the loss-framing domain (Kahneman, 1991). The prospect theory holds that loss framing is likely to motivate risk-seeking, which leads to a preference for risky or uncertain

PAGE 24

24 en-US options, while gain framing tends to provoke risk-aversion, which hinders the choices people are willing to consider (Kahneman & Tversky, 2000). en-US The value function describes the s-shaped curve (Kahneman & Tversky, 1978). The curve is concave in the domain of gain and convex in the domain of loss (Kahneman, 1991 ; Kahneman 2003). The value function is much steeper in the loss domain than in the gain domain which indicates that negative feelings or displeasure provoked by the loss prospect are stronger than the positive feelings or pleasure evoked by a commensurate gain prospect. As such, the prospect theory assumes loss-aversion as a key heuristic to explain why losses are more influential than commensurate gains (Newell et al., 2007). en-US Although the prospect theory was originally developed from the domain of choice and decision-making (see Kahneman & Tversky, 2002), as Kahneman recognized in his recent article, as the psychological principle claimed in the prospect theory, a relative respect to framing effects has been applied to more diverse social science disciplines (e.g., health psychology, consumer behavior, communication, medical decision, political science), and a large body of evidence has demonstrated that gain-loss framing effects are found in persuasiveness, including risk perceptions, evaluation of framing objects or issues (favorable versus unfavorable), behavioral intentions, and cognitive responses (O Connor, Pennie, & Dales, 1996; Grewal et al., 1994; Arora, Stoner, & Arora, 2006; Apanovitch, McCarthy, & Salovey, 2003; Shen & Dillard, 2007). For example, Venkatraman, Aloysius, & Davis (2006) found that the decision-framing effects on behavioral intentions are mediated by both perceived riskiness and perceived ambiguity. Notably, a significant number of studies on health decision have reported that gain-loss framing effects are found on risk perception associated with disease or medical choices (see Ferguson, Leaviss,

PAGE 25

25 en-US Townsend, Fleming, & Lowe, 2005), behavioral intentions (Chang, 2007), and attitude (Sibley, James, & Kirkwood, 2006), as well as choice (see Rothman et al., 2006; Salovey et al., 2002) en-US Confounding Framing Effects en-US A number of studies have tested the gain-loss framing hypothesis in psychology and behavioral literature. Although a large body of evidence has been presented to support the prospect theory, confounded and contradictory or null results also have been reported. For tested the framing hypothesis in the context of hormonal male contraceptive use, perceived as a risky option compared to other methods of contraception. They found no significant effect on intention between gain and loss framing. en-US Several reasons exist for these mixed findings. First, the conflicting results may be attributable to inconsistent operational definitions of framing. The original framing manipulation contains the numeric information with probabilities (see the Asian Disease example presented above). However, a number of framing studies, particularly in health decisions, have used the stimuli that defined the gain versus loss distinction in terms of qualitative scenarios (e.g., if you decide to get HIV tested, you may feel the peace ; see Rothman et al., 2006). In addition, some research took the simple positive versus negative distinction in terms of characteristics or attributes of decision objects (e.g., beef consists of 25% fat versus 75% lean ; Levin, 1988). en-US William, Clarke, & Borland (2001) distinguish the framing manipulation into behavior framing and statistical framing, and explore the potential differences in their manipulation. These researchers argue that the conflicting findings of framing effects reported in the health domain may be attributed to the inconsistency of framing manipulation. According to them, behavior framing emphasizes the gain or loss outcomes as a function of complying or not complying with the behavior without the statistical (probability) information (e.g., when you get/do not get the vaccine, you are doing your best/missing an opportunity to prevent flu). In the meantime,

PAGE 26

26 en-US statistical framing focuses on the statistical information of different outcomes as a function of either selecting or not selecting the choice behavior (e.g., of 100 people who opted for medical surgery, were diseased but survived). The majority of early framing studies that tested prospect theory contain the outcome probability information (e.g., disease study, gamble). This is because early research on the prospect theory compares the choices between sure gain and probable gain, and between sure loss and probable loss. In risky situations, such as with a gamble, the message presented with probabilities is considered a risky or uncertain option compared to the sure option. Williams et al. (2001) point out the majority of framing research on health behaviors has used the behavior framing manipulation. Based on this distinction, Williams et al. compared the two formats of framing in the context of a detection-promotion health behavior (e.g., HIV screening, breast self-examination). Because no evidence of the statistical framing effects was found for intention, perceived susceptibility, self-efficacy, perceived early detection risk, or anxiety about cancer, only partial support regarding perceived susceptibility and behavioral change was reported for behavior framing such that a loss-framed health message led to greater perceived susceptibility and positive change in the performance of breast selfexaminations. en-US Attribute Framing and Goal Framing en-US Recently, significant studies focused especially on medical and health decisions have presented evidence that a different tactic of framing (e.g., framing the attributes of choice problems versus framing the goals or outcomes of choice problems) interacts with the framing domain (e.g., gain versus loss or positive versus negative) (Ferguson & Gallagher, 2007). For example, Ferguson et al. (2007) examined the interactive effect of the framing tactic (attribute versus goal framing) with the framing domain (positive versus negative framing) on behavioral intentions relative to decisions on whether or not to receive the flu vaccination. They found that

PAGE 27

27 en-US when the perceived procedural risk of vaccination (e.g., pain, side effects, etc. of medical treatments) is high, negative (loss)-goal framing is more effective than positive (gain)-goal framing and positive-attribute framing has a more stringent impact than negative-attribute framing. en-US The interplay of attribute-goal framing tactics and a gain-loss framing domain was first proposed by Levin, Schneider, and Gaeth (1998), who suggested a model in which the framing tactic interacts with the framing domain on persuasiveness and behavioral outcomes. Levin et al. operationalization, in diverse decision contexts in which the framing literature produce d confusing results. Many supporting arguments for this claim have been presented, and a group of scholars attempted to identify different types of operational definitions (Levin, Gaeth, Schreiber, Lauriola, 2002; Williams et al., 2001; Maule et al., 2007) and highlighted the need to investigate the potential difference in framing effects because of different message framing tactics (Levin et al., 1998). In particular, Levin et al. (1998) conceptualized different tactics of framing effects based on the empirical findings and claims that framing effects are conditioned by these framing tactics : attribute framing and goal framing en-US Attribute framing focuses on the attributes of a decision problem in a gain (positive) or loss (negative) sense (Levin et al., 1998; Levin et al. 2002). Typical examples of attribute the evaluation of ground beef (meat is 80% 20% death rate). Another method of presenting attribute framing involves phrasing the success versus failure rate of choice options, such as medical treatment (Wilson, Kaplan, & Schneiderman, 1987; Levin et al., 1988) or the selection of preventive measures (e.g.,

PAGE 28

28 en-US contraceptive use ; Linville, Fischer, & Fischhoff, 1993). According to Levin et al. (1998), the dependent variable observed for attribute framing is not limited to actual choice; instead, the jection (of choice en-US A series of recent studies comparing attribute-framing effects with goal-framing effects have identified reliable interactive effects. These hold that gain or positive framing is weighed heavier than loss or negative framing when attribute framing, while loss or negative framing weighs more than gain or positive framing when goal framing (see Karen, 2007; Ferguson & Gallagher, 2007; Levin et al., 2002; Krishnamurthy, Carter, & Blair, 2001). en-US This conclusion causes the following question to arise. Why is gain (positive) framing more effective in the likelihood of decision making than loss (negative) framing when framing an attribute of choice problems? To answer this question, Levin and his colleagues (1998; 2002) proposed choice problem evaluation. The associative model is based on the assumption that the valance of information has considerable effect on cognitive processing (Levin and Gaeth, 1998), and thus positive information in attribute framing is prone to activate positive associations in memory, and vice versa with respect to negative information, and these internal associations affect choice problem evaluation or persuasion (Krishnamurthy et al., 2001). More importantly, when evaluating framing objects, a kind of cognitive heuristic that favors positive labels is likely to occur and, in turn, positive attribute framing results in more favorable evaluations of choice objects or issues than does negative attribute framing (Levin et al., 2002). en-US A group of scholars accounted for this positive framing prevalence in attribute framing by suggesting a priming process of positive and negative information (see Levin et al., 1998;

PAGE 29

29 en-US Krishnamurthy et al., 2001). Studies found that persuasive messages which prime positive associations tend to evoke positive evaluations as opposed to priming negative associations in the context of impression formation, and that the positive associations were more effective on the impression of message objects than were the negative associations (Levin et al., 1998). en-US Given the nature of attribute framing that focuses on the objects or characteristics, the underlying framing mechanism of prospect theory seems to be seldom applied to attribute framing effects because of the inapplicability of loss aversion. Loss aversion is an essential element in prospect theory, which leads to risky choice (see Kahnamen & Tversky, 2000); however, this loss aversion is absent because attribute framing is not concerned with evaluations of choice or behavioral outcomes. In other words, the loss aversion is associated with a kind of displeasure brought about by negative outcomes or consequences because of failing to choose the suggested options; however, the loss outcome information that provokes loss aversion does not exist in attribute framing. Levin et al., (1998) provide the supporting account for this notion and argue that prospect theory is not applied to describe attribute framing effects because of the caused by not considering the options. The potential negative outcomes evoked by attribute framing are different from goal manipulations in quality: the former simply involves objective unfavorable or undesirable characteristics of framing objects themselves (e.g., failure rates of a preventive medical treatment), while the latter involves more subjective negative outcomes from not considering the choice options (e.g., increased chance of contracting sexually transmitted diseases by failing to use condoms). Accordingly, it is expected that negative (loss) framing advantages may not be observable in attribute framing because of the absence of loss aversion, which leads to risk-taking motivation. A large body of evidence for the relationship between

PAGE 30

30 en-US attribute framing and evaluation of choice objects has been demonstrated for decades For example, Levin et al. (1988) found that positive attribute framing (75% lean) produced more positive evaluations of the beef quality than negative framing (25% fat). en-US At the same time, it may be observed that risky or uncertain elements are always absent in the case of attribute framing decision making. In effect, risky choice situations often exist, particularly in the case of controversial medical treatment decisions, such as use of hormonal medicines that have a relatively high potential of critical side effects, or new product purchase decisions that may have a high price risk (Chang, 2007). en-US A number of studies have found that attribute framing effects are moderated by diverse contextual and individual features, such as personal relevance (Krishnamurthy et al., 2001) and perceived risks associated with medical decisions (Ferguson et al., 2007). However, very little research has focused on source credibility as a moderator in the attribute framing effect. Given that attribute framing effects often have focused on persuasiveness as a dependent variable, including attitude, intentions, or judgment regarding framing objects (Levin et al., 1998; Ferguson, 2007), source credibility also affects the aforementioned variables, specifically, the source factor as a moderator of framing and outcomes deserves attention. en-US Keren (2007) revealed an interesting finding in which experimental subjects who evaluated negative (loss) attribute framing (25% fat were considered more trustworthy than positive (gain) attribute framing (75% lean), but positive attribute framing was more effective in influencing their choice of product. Despite the general assumption that high credibility perception leads to greater persuasiveness, this study showed an asymmetric relationship between choice and source assessment because of framing. However, this study has limitations. It lacks an understanding of how message framing interacts with source assessment, since it did

PAGE 31

31 en-US not manipulate the level of message source credibility as an independent variable. A prediction may be reasonable from this insight that positive attribute framing effects decrease with low source credibility. en-US Goal framing effects abound in the health decision literature (See Maheswaran & Meyers-levy, 1990; Meyerowitz & Chiken, 1987; Nan, 2007). As discussed above, goal framing is distinguished from attribute framing in that the former focuses on the behavioral outcomes, while the latter stresses the characteristics of the choice object itself. Goal framing is concerned with outcome presentation in either gain (positive) or loss (negative) outcomes, but the semantic meaning or goal is the same between the two. en-US A general finding with goal framing effects is that loss (negative) framing works better than gain (positive) framing in the context of risky or uncertain choice s, as predicted in the prospect theory (See Rothman et al., 2006; Salovey et al., 2002). Loss aversion effects because of goal framing are applicable because the prospect theory focuses on the relative evaluation of outcomes of choice behaviors, which is similar to goal framing manipulation. en-US Other Moderators of Framing Effects in Health and Marketing Decisions en-US Building on prospect theory and empirical evidence, scholars in the applied health domain make a useful distinction in developing framing hypotheses in the context of health and medical decisions: prevention versus detection (Rothman and Salovey, 1997; Salovey et al., 2002; Rothman et al., 2006). Salovey et al. (2002) proposed that a gain-framed message has a greater advantage than a loss-framed message when encouraging prevention health behaviors (e.g., promoting sunscreen use, exercise), and that a loss-framed message is more effect than a gain-framed message when promoting detection behaviors (e.g., HIV screening, breast selfexamination). According to these scholars, detection behaviors are considered uncertain or risky choices because people may experience uneasiness when engaging in the disease detection

PAGE 32

32 en-US process (Meyerowitz & Chaiken, 1987). Although significant empirical findings have supported this hypothesis, some contradictory findings have also emerged. As a result, framing researchers in the health domain have revealed various conditions that determine the relative advantages of loss or gain framing. en-US Chang (2007) proposes another useful framework for framing hypotheses regarding the effect of advertising healthcare products (e.g., mouthwash; dental gum). He classifies healthcare product types in terms of prevention versus detection functions and suggests that persuasive effects of framing vary according to such functions. Chang hypothesizes that gain framing is more influential on attitude and intention regarding the product purchase for preventi ve healthcare products, while similarly persuasive effects are observed by using loss framing for detection healthcare products. Chang found a significant moderating effect of healthcare product type (e.g., prevention versus detection) in the framing and persuasion relationship. He further revealed that the persuasive effects of the framing were moderated by the perceived newness of the healthcare products: the benefits of gain framing were found over loss framing in both new prevention and detection products. His rationale for these framing assumptions roots the prevention-detection distinction in the literature of message framing effects on health promotion (Rothman et al., 2006). en-US Based on empirical evidence in persuasion literature, a group of scholars assume that negative information receives more attention than positive information in high involvement conditions. One mechanism that explains the loss advantage in high involvement conditions is overweighing (Maheswaran & Meyers-Levy, 1990). When message quality is constant and only the framing valence is distinct, the framing effect is susceptible to the level of involvement. Specifically, the negative (loss) framed message is overweighed in high involvement conditions

PAGE 33

33 en-US (Meyerowitz & Chaiken, 1987). A negative message is more likely to be perceived as informative and receive extra weight (Fiske, 1980). en-US Individual involvement has received considerable attention in the gain-loss framing literature (Maheswaran & Meyers-Levy, 1990; Zhang & Buda, 1999; Wiliams, Clarke, & Borland, 2001). Maheswaran and Meyers-Levy (1990) attempted to explain the message framing effect in the context of persuasion. By combining issue involvement with message framing, they provided a theoretical explanation on the mixed results of the framing effect. They found that the negatively framed message (loss framing) elicited more favorable attitudes and intentions regarding the diagnostic blood test in high involvement conditions, while the benefits of gain framing were found in low involvement conditions. Their rationales for the results are based on the persuasion theories (Elaboration Likelihood Model, ELM) According to these researchers, high involvement makes people scrutinize the negative message and assign extra weight to the message T hey are more likely to overestimate the probability of the outcomes, thus leaning toward risk-seeking behavior. en-US Zhang and Buda (1999) examined the moderating role of the need for cognition in the advertising message framing effect. They found the main effect of the advertising message framing and its interactive effects with need for cognition on three measures, including the attractiveness of the product, perception of product performance, and willingness to purchase the product. Although the study makes theoretical advances in understanding the individual level of conditions moderating the framing effect by incorporating the prospect theory with the elaboration likelihood model, some problems exist with its internal validity. Zhang and Buda tested and support the hypothesis that a positively framed message has a higher mean than a negatively framed one, for all dependent variables. In their background and discussion, sufficient

PAGE 34

34 en-US theoretical accounts for this relationship do not exist, although the direction they posited seems to conflict with the same evidence reported in the framing literature (e.g., loss or negative framing is much more desirable). Moreover, it is not clear how framing operates for the perception measure (e.g., perceived attractiveness of the product) in the boundary of prospect theory. en-US Message Source Credibility en-US The level of source credibility has been explored as moderating the gain-loss framing effect. A handful of studies in the decision framing and persuasion literature ha ve addressed the interactive effect of source factors with gain-loss framing (Jones, Sinclair, & Courneya, 2003; Grewal, Gotlieb, & Marmorstein, 1994; Buda, 2003; Arora, Stoner, & Arora, 2006). However, these studies fail to provide a sound theoretical framework that accounts for how gain-loss framing impact strengthens or weakens by the difference in the level of source credibility. en-US Source impact on communication confirmsthe early work on the persuasion and behavioral change (Hovland et al., 1953; Kelman, 1961). Source credibility is generally manipulated by imposing some characteristics on a message source, such as trustworthiness, expertise, attractiveness, and power (McGuire, 1985). It is well documented in the literature of persuasion and behavioral change that the impact of source factors is likely to be intensified by the source characteristics, but the persuasiveness of message does not always vary in direct proportion to the source impact (McGuire, 1985). The impact of source characteristics also interacts with other message features such as the quality of argument, framing fear appeals, and message intensity and consumer variables such as involvement and the need for cognition (see Burgoon et al., 2002).

PAGE 35

35 en-US Mechanisms of Message Source Impact en-US Kelman (1961) conceptualizes the message sources and accounts for how the source characteristics can affect the persuasiveness of a message. In his propositions regarding processes of opinion and behavior change, Kelman suggests that the message source impact on persuasiveness can be explained by one or more of three mechanisms: internalization, identification, and compliance. First, the internalization process describes the tendency for an individual to comply with a behavioral option when the option presented conforms to his or her internal value or belief system (McGuire, 1985, p. 262) and find it important to acquire accurate external information from valid sources. Kelman (1961) suggests that the critical characteristics of the source for reinforcing the internalization are the credibility cues namely, expertise and trustworthiness. Expertise represents knowledge of truth and trustworthiness reflects sincere motivation or intention to inform others of the truth (Hovland, et al. 1953; Kelman & Hovland, 1953; Mills & Jellison, 1967). en-US The second source-effect mechanism image with a source that has attractive characteristics such as likeability, familiarity, and similarity (McGuire, 1985). In short, identification-based attitudinal changes occur when a recipient links his or her self-image to the attractive sources and is satisfied by supporting the -images in a positive way by equating their attitudes with socially admired sources (Kelmen, 1953; Kelman, 1961). en-US Finally, Kelmen proposes compliance as a mechanism of source impact. The notion of seek out rewards and avoid punishments from another person. Compliance-based opinions or behavioral changes arise when a source is

PAGE 36

36 en-US supervised by the influencing source; this suggests that people comply to the message arguments, but may not believe them (Kelman 1961). en-US Hovland and Weiss (1951) examined the media source effect on information acquisition and opinion change with the focus on trustworthiness. Their findings show that while there was no difference in the acquired amount of information from high and low levels of trustworthy sources, significant positive changes in opinion in the advocated direction were seen in the high trustworthiness conditions, but no such change was found in the low trustworthiness conditions. An additional important finding of their s likely to be cancelled out by any time delay. After all, the study suggests that the source impact (trustworthiness) is found in the immediate exposure to a message, but such effect is likely to diminish by the time delay. en-US Gain-Loss Framing Effect and Source Credibility en-US Jones et al. (2003) explored the interactive effects of framing and source credibility on cognitive and behavioral changes with respect to physical exercise. These researchers based the theoretical frameworks on prospect theory and ELM and tested the hypothesis that the positively framed (gain) health promotion message was most distinctive when combined with a high expert source. Their results fit with this hypothesis for intentions, actual exercises, and cognitive responses, and, as a result, the participants showed more favorable exercise intentions and behaviors and induced more elaborations in high expert-gain framing conditions than in other conditions (e.g., high expert-loss, low expert-gain). However, their data did not support the hypothesis for attitude, which indicates that the ELM, one of their theoretical foundations, may not be an appropriate application to account for the interactive effects.

PAGE 37

37 en-US In contrast, Buda (2003) reported null effects of the interaction between the source credibility and message framing. He examined the interactive effects of framing, source credibility, and message presentation order on proposed job evaluations. He measured the perceptions of job attractiveness and job performance and the willingness to take the job offer as dependent variables and found that there was no two-way interaction between the framing and source credibility, but such a relationship was only significant when the message order was combined. Specifically, the gain-framed message was more effective than the loss-framed message when presented first from the low credibility source and presented last from the high credibility source. Although this study did not include the judgment variable, it provides an important insight into the potential moderating role of other message features (e.g., multiple framing and framing tactic) for the relationship of framing effect and source variable. en-US Most recently, Arora et al. (2006) tested the interactive effects of framing and source credibility on exercise promotion. They found significant interactions between the two on attitude and intentions regarding physical activity. Specifically, loss framing was susceptible to source feature to such an extent that both the mean scores for attitude and the intention to engage in exercise were significantly lower in the low-expert/loss framing condition than other conditions (e.g., the high-expert/gain, low-expert/gain condition). Considering the findings from the aforementioned research, it is predicted that, while the gain-framing impact is constant in a positive direction on persuasiveness, regardless of source credibility, the loss framing impact is sensitive to the level of credibility. en-US In summary, the effects of message framing on persuasiveness or behavioral change are loss (negative); framing effects are expected to decrease with low source credibility, while gain

PAGE 38

38 en-US (positive) framing effects are expected to increase with high source credibility. Based on these findings, how are these interactions accounted for? en-US Source as a Moderator of Framing Effect: Language Expectancy Perspective en-US High credible sources, as discussed above, do not always induce greater positive evaluations and behaviors than do the low credible sources, and vice versa (McGuire, 1985). For example, soprocess, despite the use of highly credible spokesmen. A plausible reason for this is that the message receivers may notice the mudslinging intention of the source while processing the message, and thus, the overall credibility (trustworthiness) of the argument of the source is considered low (see Kelman & Hovland, 1953; McGinnies & Ward, 1980). In contrast, a communicator perceived as low in expertise has the potential to induce persuasiveness as much as high experts when his or her argument is congruent with general social norms associated with the issue but given as less aggressive or negative (Burgoon, Denning, & Roberts, 2002). The examples above imply that the selection of message features or tactics can be important in predicting the relative persuasiveness of the message with source characteristics (Burgoon, 1996; Burgoon et al., 2002). Language expectancy theory focuses on the role of expectations that people have for the use of socially acceptable or unacceptable language or message strategies (e.g., use of aggression, fear appeals, or level of language intensity) by message sources in persuasive attempts (Burgoon & Miller, 1985; Burgoon, 1989). en-US Language expectancy theory provides the present study with a theoretical framework to explain the interactive effect of message framing and source credibility. This theory suggests that the tacit rules of appropriate message strategies are influenced by social or cultural norms (Burgoon, 199 expectations and preferences associated with the language or message usage in persuasive communications (Burgoon, 1989). According to

PAGE 39

39 en-US Burgoon and his colleagues, the persuasive effect is limited when a message source negatively violates the expected uses of message or language tactics, whereas the effect is reinforced when the expectations of the message uses are congruous with the perceived source characteristics (Burgoon, 1989; Burgoon et al., 2002; Burgoon, 1990). Specifically, language expectancy theory assumes that the use of negative message tactics (e.g., loss framing, fear appeal, negative political advertising) is considered a negative violation of expectations for low-credible sources and, thus, attitude or behavioral changes are limited when negative information is presented with low source credibility (Burgoon, 1990). en-US Accordingly, language expectancy theory proposes that low-credibility sources have limited use of message tactics relative to high-credibility sources, so as to have the least benefit from the use of negative health message tactics (Burgoon, 1996). This account is especially relevant to understanding why the interactive effect of source credibility with framing can be more distinctive in the negative framing condition than in the positive (gain) framing condition particularly goal loss-framed message framing G oal-loss framing is likely to produce negative feelings such as displeasure, anxiety, and fear (at least moderate fear), and risk perceptions as fear appeal messages may bring consumers (see Salovey et al., 2002). Kahnenman, one of the founders of gain-loss framing effects, also recognized the presence of negative feelings, such as displeasure, as a result of processing loss-framed messages, which leads to loss aversion (Kahneman, 2003). en-US Although there have been few studies examining the relationship of gain-loss framing and source credibility based on language expectancy theory, several theories and empirical

PAGE 40

40 en-US arguments framed in lends some support of the idea that positive information is regarded as a general message format relative to negative information. Thus, the loss-framed message is more likely to be perceived as an irregular format that is salient to people than the gain-framed message. They also suggest the possible link between loss framing and fear arousal in the context of health communication, in that a loss-framed message may induce negative emotions such as fear or threat. In line with this suggestion, Williams et al. (2001) revealed that the loss-framed health promotion message for breast cancer screening yielded more perceived susceptibility to breast cancer than the gainframed message. Considering this empirical evidence and theoretical rationale, it is assumed that the use of loss framing by a low-credibility source in health decisions may indicate a violation of language expectations in persuasive attempts. Based on these accounts, the hypotheses 3a and 3b are formulated. en-US Hypotheses and Research Question en-US Overall, the literature on framing effects (Ferguson et al., 2007; Levin et al., 2002; Levin et al., 1998; Krishnamurthy et al., 2001) that observed potential for the interactive effect between framing tactic and framing domain in the context of health and medical marketing decisions suggests that gain (positive) framing is more influential than loss (negative) framing in attribute framing, while loss framing is more effective than gain framing in goal framing. Therefore, hypotheses 1a-d proposed as follows. en-US H1 a: The attribute gain-framed DTC advertisement will produce more positive attitudes toward hormone replacement therapy (HRT) than attribute loss-framed DTC advertisement and the goal loss-framed DTC advertisement will produce more positive attitudes toward HRT than goal gainframed DTC advertisement. en-US

PAGE 41

41 en-US H1b : The attribute gain-framed DTC advertisement will produce stronger intentions of seeking more information about HRT than attribute loss-framed DTC advertisement and the goal lossframed DTC advertisement will produce stronger intentions of seeking more information on HRT than goal gain-framed DTC advertisement. en-US en-US H1c: The attribute gain-fram ed DTC advertisement will produce stronger intentions of consulting a doctor about HRT than attribute loss-framed DTC advertisement and the goal lossframed DTC advertisement will produce stronger intentions of consulting a doctor about HRT than goal gain-framed DTC advertisement. en-US en-US H1d : The attribute gain-fram ed DTC advertisement will produce stronger intentions of choosing HRT than attribute loss-framed DTC advertisement and the goal loss-framed DTC advertis em ent will produce stronger intentions of choosing HRT than goal gain-framed DTC advertisement. en-US en-US According to language expectancy theory, the persuasive impact of the loss (negative) message tactic decreases with low source credibility. Hypothese s 2a-d are proposed to test if there is a two-way interaction effect between frame domain (gain versus loss) and source credibility (high versus low). en-US H2: There will be a two-way interaction effect between gain-loss framing domain and source credibility on attitude toward HRT and DTC ad -promoted intentions. Specifically, the effect of frame domain (gain versus loss framing) on attitude toward HRT (H2a), intention to seek more information (H2b), intention to consult a doctor for further help (H2c), and intention to choose HRT (H2d) would be less distinctive when an ad message is presented by a low-credibility rather than a high-credibility source. Notably, the gain-framing impact on the dependent variables would be constant, regardless of the level of the ad message source credibility. en-US en-US RQ : Is there a three-way interaction effect among framing tactic, framing domain, and source credibility?

PAGE 42

42 en-US en-US Figure 2-1. Expected interaction effect of gain-loss framing domain and attribute-goal framing tactic on attitude toward HRT en-US en-US Figure 2-2. Expected interaction effect of gain-loss framing domain and attribute-goal framing tactic on DTC ad-promoted intentions

PAGE 43

43 en-US en-US Figure 2-3. Expected interaction effect of gain-loss framing domain and source credibility on attitude toward HRT en-US en-US Figure 2-4. Expected interaction effect of gain-loss framing domain and source credibility on DTC ad-promoted intentions

PAGE 44

44 en-US CHAPTER 3 en-US METHODOLOGY en-US De sign en-US The objective of the present study is to explore a two-way interaction between attributegoal framing tactic and loss-gain framing domain, a two-way interaction between gain-loss framing domain and source credibility, and a three-way interaction among the three independent variables on DTC ad-promoted behavioral intentions (Huh, DeLorme, & Reid, 2005) associated with hormone replacement therapy and attitudes of the target consumers toward hormone replacement therapy. en-US In order to achieve this research goal, an experiment wa s designed with a 2 (framing tactic: attribute versus goal framing) 2 (frame domain: gain versus loss framing) 2 (source expertise: high versus low) between-subjects experimental design, producing eight experimental groups in total. Previous studies have revealed that the effects of message framing on intentions and attitudes are also influenced by pre-experimental factors such as knowledge, previous attitudes, product experience, involvement, and demographics (e.g., education, ag e) (see Smith & Wortzel, 1997; Smith, 1996 ; Maheswaran & Meyers-Levy, 1990). Therefore, despite random assignment in the sampling procedure, the current study performed a MANCOVA as a primary statistical method to analyze these interaction and main effects while controlling attitude toward general hormonal medications such as steroid and oral contraceptives, previous experience with the hormone replacement therapy, product knowledge, level of personal involvement perceived risk of side effects and age. In addition, separate ANCOVA contrast analyses were conducted to compare statistically significant mean differences among experiment groups.

PAGE 45

45 en-US Participants en-US This study recruit ed the target consumers of hormone replacement therapy. Research sample groups consist ed of women between the age of 45 ~ 65, who are the product target age segments in Florida Participant s were all from samples derived from respondent panels developed and utilized by professional sampling house associated with a professional market research institution located in central Florida. To randomly assign the participants to each experimental condition, an email address list of the participants was used. For this, a local professional market research company located in central Florida distributed the Web-based online experiment links to the experiment subjects and collected the data from June 15 th to July 20 th of 2009. en-US Research Stimuli and Drug Category Selection en-US Eight versions of directto -consumer (DTC) prescription drug advertisements for a fictitious hormone replacement therapy brand named Seroxmo were developed for the current study. As discussed in Chapter 1, hormone replacement therapy was chosen as the product category for this study for the following reasons. First, it is generally recognized as a risky medical decision for women suffering from menopausal symptoms according to recent survey results (see Genazzani, Schneider, Panay, Nijland, 2006; Hoffmann, Hammar, Kjellgren, LindhAstrand, & Brynhildsen, 2005). According to prospect theory, the gain-loss framing effect can be observed in the context of risky or uncertain decision-making. Second, empirical evidence has shown that the consumer perceives hormonal medications such as contraceptives, steroids, and hormone replacement therapy as riskier than other forms of medication (see onnor, Ferguson, and health literature focusing on hormone replacement therapy as a health decision choice in spite of its controversial nature regarding efficacy and potential side-effects.

PAGE 46

46 en-US Procedure Participants received an email with the contact script and study Web-link from a professional market research provider located in Gainesville, Florida. 1 Each participant was asked to use a personal computer connected to the Internet to visit the study Web site In the beginning of the online experiment package, the participants were told wa s to investigate how consumers understand a drug s advertising copy. The p ackage also specified the nature of confidentiality of their information and voluntary participation. Following the instructions, they first read a hormone replacement therapy ad vertisement and then were instructed to answer the study questionnaire. Participants received a debriefing statement immediately after completion of the experiment (see Appendix J). en-US Independent Variables en-US To manipulate attribute framed versus goal framed-advertising messages, gain (positive) versus loss (negative) framed advertising messages, and low versus high message source credibility, eight versions of a DTC print advertisement for a fictitious hormone replacement therapy brand were created. All versions of the advertisement feature the exactly same characteristics in terms of any visual and textural elements except the manipulations (see appendix A through H ). en-US Message Framing Manipulations en-US Table 3-1 shows the four different message framing manipulations for the current study. The combinations of framing tactic and framing domain manipulations were adapted from previous studies (e.g., Krishnamurthy et al., 2001; Ferguson et al., 2005). 1 All surveys were incented with cash. The research company did not reveal the actual amount. en-US

PAGE 47

47 en-US Framing tactic is classified into two dimensions; attribute versus goal framing. Attribute framing focuses on attributes of a product. Therefore, gain-attribute versus loss-attribute framed messages can be constructed through emphasizing either desirable attributes of a drug product versus undesirable attributes of the product or presence (absence) or absence (presence) of a desirable (undesirable) attribute of the product (see Salovey et al. 2002; Krishnamurthy et al., 2001). For this study, gain-attribute versus loss-attribute framing was achieved through emphasizing the relatively high level of efficacy of hormone replacement therapy versus the relatively low level of its inefficacy. en-US Goal framing emphasizes outcomes of the use of a drug product. In the DTC ad context, the drug product outcomes are associated with health benefits as a result of the drug use. Therefore, gain-goal versus loss-goal framed ad messages can be manipulated by focusing on either the presence of positive health benefits of a drug product or the absence of negative health outcomes of the drug use versus absence (presence) of positive health benefits (negative health outcomes) as a result of failing to use the drug (see Salovey et al, 2002). For this study, gaingoal framing was manipulated by focusing on the positive health benefits of the drug use and relative absence of negative health outcomes, while loss-goal framing was achieved through emphasizing negative health outcomes as a result of not taking hormone replacement therapy. en-US Framing Manipulation Check Items en-US Two 5-point Likert items ( 1 = strongly disagree 5 = strongly agree ) were adapted from Broemer (2002) to assess the ga in -loss framing manipulation. Participants were asked to indicate whether he main message of the ad was described in a positive to attribute versus goal framing manipulation, two 5-point Likert scales were developed based on the conceptual definition of each framing (see Levin, 1998; Krishnamurthy, 2001). Two items The ad focused on side effects and safety of the hormone replacement therapy and

PAGE 48

48 en-US consumer satisfaction of the drug efficacy The ad focused on relative health risk of suffering from menopausal symptoms as a result of taking or not taking hormone replacement therapy. en-US Table 3-1. Ad message framing manipulations en-US en-US Source Credibility Manipulations en-US Source credibility was manipulated by describing the source (spokesperson) in the advertisements as either being a medical doctor affiliated in U.S. Food and Drug Administration (FDA) for the high credibility condition o r a drug customer for the low credibility condition (see Buda & Zhang, 2000; Zhang & Buda, 1999; Jones et al., 2003). In order to ensure credibility of the high source, the following statement was added to the high source condition ; Seroxmo is the hormone replacement therapy approved by the FDA. For the low credibility condition, the Framing Tactic Framing Domain Attribute Frame Goal Frame Gain Frame 80% of menopausal w omen are satisfied with hormone replacement therapy in relieving moderate to severe menopausal symptoms, such as hot flashes and night sweats. Research shows that more than 80% of hormone therapy users do not experience side effects. If you are using horm one replacement therapy you are taking advantage of the best method of relief from moderate to severe menopausal symptoms, such as hot flashes and night sweats and ensuring your comfort. By taking hormone therapy, your risk of developing many of the sympto ms associated with menopause can decrease by 70% Loss Frame Only 20% of menopausal women are not satisfied with hormone replacement therapy in relieving moderate to severe menopausal symptoms, such as hot flashes and night sweats. Research shows that less than 20% of hormone therapy users experience side effects therapy you are failing to take advantage of the best method of relief from moderate to severe menopausal symptoms, such as hot flashes and night sweats and fa iling to ensure your comfort. By not taking hormone therapy, your risk of developing many of the symptoms associated with menopause may increase by 30%.

PAGE 49

49 en-US statement indicati quote and only some demographic information (e.g., name, age, residence place) of the spokesperson was highlighted. In advertising research, the manipulation of a low credible source in an advertisement has been achieved by selecting a non-expert general customer as the spokesperson of the ad (Zhang, 1999; Grewal, Gotlieb, & Marmorstein, 1994). Source credibility was measured by rating five semantic differential items, anchored on 7po int scales ranging from 1 to 7 adopted from Ajzen & Fishbein (1980): Not knowledgeable/knowledgeable not competent/competent not expert/expert not trained/trained and not experienced/experienced en-US Pretest en-US Prior to the main study, a pretest w as conducted to ensure the validity of framing manipulations developed. A total of 36 women participants aged 40 to 65 were recruited voluntarily from the staff in the College of Journalism and Communications at University of Florida and from among the parishioners at a church in Gainesville, Florida. They were first exposed to a manipulated advertisement and were asked to answer the framing manipulated items. en-US Dependent Measures en-US Attitude toward Hormone Replacement Therapy en-US Attitude toward hormone replacement therapy was measured on nine 7 -point semantic scales developed by Oliver (1980): foolish/wise safe/risky harmful/beneficial pleasant/unpleasant waste of time/ wise use of time good for me/bad for me, useful/useless worthless/valuable and ineffective/effective en-US DTC ad-promoted Behavior Intentions en-US According to Huh et al. (2005), DTC ad tends to promote three different types of behavior, which include communication with doctors, seeking more information about treatment

PAGE 50

50 en-US options, and talking with others. Based on the items to measure these behaviors, different types of DTC-promoted behavior intentions were developed for this study. First, intention to seek more information about hormone replacement therapy was measured by two different statements, anchored on 7-point scales modified from Huh, DeLorme, & Reid, 2005. Participants were instructed to rate each of the statements by placing a check mark in one of seven spaces from very unlikely to very likely. Please rate the probability that you would participate in a free educational program about hormone replacement therapy and efficacy of the advertised drug product if local health services offered the program Please rate the probability that you would go to other media sources to get more information about hormone replacement therapy. ntention to communicate with a doctor was measured using a single statement anchored on a 7-point scale from very unlikely to very likely modified from Huh et al. This item was measured with the following statement Please rate the probability that you would meet with a doctor i f local health services offered a free opportunity to consult doctors about hormone replacement therapy. Last, intention to choose hormone replacement therapy was me asured by two items anchored on 7-point scales from very unlikely very likely. These items were measured with the following statements Please rate the probability that you would choose hormone replacement therapy if you had menopausal symptoms ; and probability that you would recommend hormone replacement therapy to others if they had en-US Blocking Variables en-US Altho experimental factors (e.g., demographic variables subjective knowledge level, pre-existing beliefs and other psychological variables) may influence the dependent variables given the controversial nature of the use of hormone therapy (see Shah et al, 2004). To remove such

PAGE 51

51 en-US confounding effects, the study included six control variables as covariates; age, attitude toward hormonal medicines (e.g., steroids) in general, personal involvement, previous experience with hormone replacement therapy, the subjective knowledge level, and perceived risk of side effects. These measures have been considered to influence on attitude and intention associated with health-related objects and decisions (see Salovoy et al., 2002). en-US Attitude toward General Hormone Medications en-US Four 7-point semantic differential scale items adapted from Burton & Lichtenstein (1988) were used to measure attitude toward hormone medications such as oral contraceptives and steroids. These included: e medications is favorable/unfavorable ; bad/good ; harmful/beneficial ; attractive/unattractive en-US Personal Involvement en-US Five 7-point semantic differential scale items adapted from Mittal (1995 ) were used to involvement in a particular issue. These included: therapy is important/unimportant ; means a lot to me/means nothing to me ; matters to me/does not matter ; significant/insignificant ; of no concern/of concern to me en-US Knowledge about Hormone Replacement Therapy en-US Two 5-point Likert scale items (1 = Little or no knowledge 5 = A great deal of knowledge) adapted from Bloch, Ridgway, & Sherrell (1989) were selfHow do you rate your knowledge of hormone replacement therapy relative to other people? an d How do you rate your knowledge of hormone replacement therapy relative to most of your friends. en-US Perceived Risk of Side Effects en-US Two 7-point Likert scale were developed to measure perception of the potential side effects of hormone replacement therapy. When I think about the side effects of

PAGE 52

52 en-US hormone replacement therapy, I feel nauseous; the side effects of hormone replacement therapy could put my life in danger. en-US Previous Experience with Hormone Replacement Therapy en-US Previous experience was measured with a filter dichotomous question and open-ended question, which included If you answered how long have you used the hormone replacement therapy en-US

PAGE 53

53 en-US CHAPTER 4 en-US RESULTS en-US Th is chapter presents the results of the experiment described in Chapter 3. The chapter begins by presenting demographics of the experiment participants. Then the manipulation checks for the independent variables (frame domain, frame tactic, and source credibility) and reliability checks for variables measured with a multi-item scale, and correlation analyses among a list of variables follow. The results of hypothesis and research question testing are then presented, followed by a summary of key findings and additional data analyses. en-US Demographics en-US Group Size en-US A total of 311 female participants were assigned randomly to one of eight experimental groups. Among them, the responses from 18 participants who failed to complete the questionnaire were excluded; thus, 293 responses were used for the data analyses. en-US The number of participants in the gain and loss framing conditions was 152 and 141, respectively, and the number of participants in the high and low-source credibility conditions was 147 and 146 each. Exactly 145 participants were assigned to the attribute framing condition and completed the study, while the 148 in the goal framing condition completed the test (see Table 4-1). en-US Out of 2.1 million active panelists represented, the respondents were filtered by age and gender. The research company sent 13,000 invites via email and the response rate for this study was 7%. 2 The target market for hormone replacement therapy includes women aged 45 to 69. 2 The research company specializes in all consumer segments. en-US

PAGE 54

54 en-US Women between the ages of 45 and 65 who resided in the state of Florida were selected as participants in the present study The mean age of this sample population was 52.7 with a standard deviation of 6.3. About 99% of the participants (N = 267) were white, and 48% reported their education level as either college degree or higher (see Table 4-2). About 44% of participants reported their economic status as being either lower middle or working class (see Table 4-3). en-US Table 4-1. Number of participants by experimental groups High Credibility Low Credibilit y Total Attribute Framing Gain Framing 37 38 75 Loss Framing 35 35 70 Goal Framing Gain Framing 39 38 77 Loss Framing 36 35 71 Total 147 146 293 en-US en-US Table 4-2 Education Participants Percent High School Degree 149 50. 9 College Degree 110 37.5 Master Degree 30 10.2 Doctoral Degree 1 0.3 Missing 1 1.0 Total 293 100.0 en-US en-US Table 4-3 Economic Status Participants Percent Working Class 69 23.5 Lower Middle Class 60 20.5 Middle Class 125 42 .7 Upper Middle Class 36 12.3 Upper Class 2 0.7 Missing 1 0.3 Total 293 100.0 en-US en-US About 75% (N = 218) of the participants currently have menopausal symptoms. Some 32% (N = 93) of the participants had experience using hormone replacement therapy, and abou t

PAGE 55

55 en-US 8% (N = 20) of them were using hormone replacement therapy during the study. The average period of the hormone replacement therapy treatment was about one year and three months. en-US Manipulation Checks for Independent Variables en-US Gain-Loss Framing Domain A manipulation check on the gain-loss framing domain was conducted with data obtained from a pretest and primary experiment of present study. For a manipulation check of the framing domain, the gain-loss framing manipulation scales used by Broemer (2002) were employed. 3 The participants reported their interpretation of an ad message via the following twoitem, 5The main message of the ad was described in a positive tone The main message of the ad was described in Each item was measured separately for the gainloss framing domain. The two items were submitted to separate ANOVAs, with the framing domain manipulation input as the independent facto r. The ANOVA results from the pretest (N=36) showed statistically significant differences for item (a) [ F (1, 35 = 4.78 p < .05 M gain = 4.00 SD = .84, M loss = 3.17, SD = 1.38], and item (b) [ F (1, 35 = 4.78 p < .05, M gain = 3.94 SD = .94 M loss = 3.06 SD = 1.11]. In addition the manipulation check on framing domain from the main experiment of this study also revealed statistically significant mean differences for item (a) [ F (1, 288) = 12.42, p < .01, M gain = 4.03, SD = .85, M loss = 3.64 SD = 1.02] an d item (b) [ F (1,287) = 7.50 p < .01, M gain = 3.89 SD = .77 M loss = 3.60, SD = .97] (see Table 4-4). en-US 3 Within the framing literature, very few studies have reported the manipulation check for the gain-loss framing domain. Framing scholars observe that this lack of manipulation check scales and items demonstrates a weakness in the validity of study findings and produces mixed findings.

PAGE 56

56 en-US Table 4-4. One-way ANOVA results for the gain-loss framing domain manipulation check Gain Framing Loss Framing F df1 df2 p M SD M SD Domain Item (a) 4.03 .85 3. 64 1.02 12.42 1 288 .00 0 Domain Item (b) 3 .89 .77 3. 60 .97 7.50 1 2 87 .00 7 en-US NOTE. Item (b) was reverse coded. Measures are based on a 5-point likert scale, with 5 indicating strongly agree and 1 indicating strong disagree. en-US en-US In order to te st if the faming domain manipulation was the sole factor that affected the perception of gain-loss framing presentation and no other independent variables moderated the effects, framing domain was submitted to a 2 (framing domain: gain versus loss ) 2 (framing tactic: attribute versus goal ) 2 (source credibility: high versus low) ANOVA. The two framing domain items were internally consistent ( = .80), and therefore were averaged into a single framing domain scale ( M = 3.80 SD = .84 ). The main effect of framing domain was found significant [ F (1, 280) = 12.66 p < .01, p = .043], but the main effect of framing tactic also significantly affected the gain-loss framing manipulation [ F (1, 280) = 10.35, p < .01, p = .036] No other interaction effects were found significant. en-US Attribute-Goal Framing Tactic Because few studies have developed the manipulation check scales for the attribute-goal framing tactic and those have not reported the results of the manipulation in the framing literature, 4 a manipulation check on framing tactic was performed with data obtained through the pretest and main experiment of this study. The participants were asked two questions that were designed to assess their interpretation of the ad message in terms of its focus on either the drug 4 No manipulation check scal e or result for the framing tactic (attribute versus goal framing) ha s been Nan (2007) also points out the absence of manipulation check as being a limitation in message framing research.

PAGE 57

57 en-US efficacy, degree and types of side-effects) or outcomes (benefits) of drug use (e.g., health improvement, reducing symptoms of menopause). Two 5-point Likert scales were presented to The ad focused on side effects and safety of the hormone replacement therapy and consumer satisfaction of the drug efficacy The ad focused on relative health risk of suffering from menopausal symptoms as a result of taking or not taking hormone replacement therapy (reverse coded). Each item was measured separately for the attribute-goal framing tactic. en-US The ANOVA results from the pretest (N=36) showed statistically significant differences for item (a) [ F (1, 34 = 11 .99 p < .01, M attribute = 3.50, SD = .92 M goal = 2.28, SD = 1.18], and item (b) [ F (1, 34 = 3.44 p = .0 7 M attribute = 3.44 SD = 1.10 M goal = 2.77, SD = 1. 06]. Although the difference for the item (b) was marginally significant ( p = .07) in the pretest, perhaps this might be due to small sample size. en-US The results of the main study s manipulation check on framing tactic revealed statistically significant mean differences for the items (a) [ F (1, 289) = 24.87 p < .001, M attribute = 3.57 SD = 1.06 M goal = 2.93 SD = 1.11], and item (b) [ F (1, 288) = 18.03, p < .001, M attribute = 3. 11 SD = 1.14 M goal = 2.56 SD = 1.05] (see Table 4-5). Participants who were exposed to an attributeframed ad scored significantly higher on the two scale items than those who read a goal-framed ad message. en-US Table 4-5. One-way ANOVA results for the attribute-goal framing tactic manipulation check Attribute Framing Goal Framing F df1 df2 p M SD M SD Tactic Item (a) 3.57 1.06 2.93 1.11 24.87 1 289 .000 Tactic Item (b) 3.11 1.14 2.56 1.05 18.03 1 288 .000 en-US Note. Item (b) was reverse coded. Measures are based on a 5-point likert scale, with 5 indicating strongly agree and 1 indicating strong disagree. en-US

PAGE 58

58 en-US Source Credibility en-US A manipulation check on source credibility was conducted using the data obtained from the main study. Source credibility was measured by rating five semantic differential scales anchored on 7-point scales ranging from 1 to 7, an approach that was adopted from Ajzen & Fishbein (1980): Not knowledgeable/knowledgeable; not competent/competent; not expert/expert; not trained/trained; and not experienced/experienced. The five items were internally consistent ( = .93) and were therefore averaged into a single source credibility scale ( M = 4.48, SD = 1.49). For the manipulation check for source credibility, the single credibility scale was submitted to one-way ANOVA, treating source manipulation (high versus low) as the independent factor. As Table 4-6 shows, a significant mean difference was observed between two groups [ F (1, 284) = 43.41 p < .01, M high = 5.03, SD = 1.37, M low = 3.93 SD = 1.43]. en-US In order to test if the source credibility manipulation was the sole factor that affected the level of source credibility and no other independent variables moderated the effects, source credibility was submitted to a 2 (framing domain: gain versus loss ) 2 (framing tactic: attribute versus goal ) 2 (source credibility: high versus low) ANOVA. As Table 4-7 shows no other main or interaction effects were found except the main effect of source credibility [ F (1, 278) = 43.41 p < .01, p = .135]. Therefore, source credibility was successfully manipulated. en-US Table 4-6. One-way ANOVA results for the source credibility manipulation check High Credible Source Low Credible Source F df1 df2 p M SD M SD (a) No t knowledgeable /knowledgeable 5.23 1.44 4.60 1.56 12.62 1 2 91 .000 (b) N ot competent/ competent 5.20 1.51 4.11 1.7 0 33.00 1 2 90 .000 (c) N ot expert/expert 4.72 1.50 3.35 1.69 52.09 1 2 87 .000 (d) N ot trained/trained 4.82 1.61 3. 44 1. 76 48.60 1 2 90 .00 0 (e) N ot experienced/ experienced 5.08 1.56 4.15 1.60 24.90 1 290 .0 00 Average ( = 93 ) 5.03 1.37 3.93 1.43 43.41 1 2 84 000

PAGE 59

59 en-US Table 4-7. Full-factorial ANOVA results for the source credibility manipulation check Independent variable F df1 df2 p p Source Credibility 43.41 1 2 78 .00 13 5 Framing Domain 16 1 2 78 69 .00 1 Framing T actic 3 39 1 2 78 07 .0 1 2 Source Domain .0 2 1 2 78 .89 .000 Source Tactic 25 1 2 78 61 .00 1 Domain Tactic 24 1 2 78 62 .00 1 Source domain tactic 16 1 2 78 .6 9 .001 en-US en-US Reliability Check for Dependent Variables and Blocking Variables en-US Attitude toward Hormone Replacement Therapy en-US Attitude toward hormone replacement therapy was measured on nine 7 -point semantic scales developed by Oliver (1980): foolish/wise safe/risky harmful/beneficial pleasant/unpleasant waste of time/ wise use of time good for me/bad for me, useful/useless worthless/valuable and ineffective/effective The Cronbach en-US Intention for Information-Seeking en-US Participants were asked to answer two questions to measure intention to seek more information about hormone replacement therapy. These measures were modified from Huh, DeLorme, & Reid, 2005 and included: Please rate the probability that you would participate in a free educational program about hormone replacement therapy and efficacy of the advertised drug product if local health services offered the program ; and, Please rate the probability that you would go to other media sources to get more information about hormone replacement therapy. The Cronbach 71. en-US Intention to Choose Hormone Replacement Therapy en-US Intention to choose hormone replacement therapy was measured by two items, anchored on 7-point scales from very unlikely to very likely These items were measured with the following statements: Please rate the probability that you would choose hormone replacement therapy if

PAGE 60

60 en-US you had menopausal symptoms ; and Please rate the probability that you would recommend hormone replacement therapy to others if they had menopausal symptoms. The Cronbach alpha for these items was .92. en-US General Attitude toward Hormonal Drugs en-US Four 7-point semantic differential scale items adapted from Burton & Lichtenstein (1988) were used to measure attitudes toward hormone medications such as oral contraceptives and steroids. These included: My attitude toward hormone medications is favorable/unfavorable ; bad/good ; harmful/beneficial ; attractive/unattractive Cronbach 6. en-US Knowledge about Hormone Replacement Therapy en-US Two 5-point Likert scale items (1 = Little or no knowledge 5 = A great deal of knowledge) adapted from Bloch, Ridgway, & Sherrell (1989) were used to measure self-reported familiarity with a particular issue. These items included: How do you rate your knowledge of hormone replacement therapy relative to other people?; and, How do you rate your knowledge of hormone replacement therapy relative to most of your friends. Cro items was .93. en-US Involvement en-US Five 7-point semantic differential scale items adapted from Mittal (1995 ) were used to measure involvement in with particular issue. These items included: hormone replacement therapy is important/unimportant ; irrelevant/relevant ; worthless/ valuable Cro these items was .96. en-US Perceived-risk of Side Effect en-US Two 7-point Likert side effects of hormone replacement therapy. These items were: When I think about the side

PAGE 61

61 en-US effects of hormone replacement therapy, I feel nauseous; and, the side effects of hormone replacement therapy could put my life in danger. Cronbach en-US Correlations among Variables en-US Correlation analyses among a number of variables were conducted to examine the appropriateness of a multivariate analysis of covariance (MANCOVA) for hypotheses testing (see Table 4-8). Education level and economic status were not significantly correlated with all four dependent variables (attitude and intentions). An interesting finding was that age was negatively correlated with information-seeking intention ( r = -.13 p < .05) and intention to choose hormone replacement therapy (HRT) ( r = -.14, p < .05). Experience with hormone replacement therapy (year(s) of use) was positively related with all four dependent variables: attitude ( r = .32, p < .001; information-seeking intention ( r = .17, p < .01); intention to consult doctor ( r = .30, p < .001); and intention to choose the hormone replacement therapy ( r = .17, p < .01). This indicates that the greater the consumer experience with hormone therapy, the more positive their attitudes are toward hormone therapy and intentions. In addition, all four dependent variables are significantly correlated with one another (see Table 4-8). This confirms that a multivariate analysis of covariance was a more proper method for testing the hypotheses than separate ANCOVAs. en-US en-US en-US en-US en-US en-US

PAGE 62

62 en-US Table 4-8. Correlations among variables en-US en-US Hypotheses Testing en-US Hypotheses 1a ~d were proposed to te st a two-way interaction between the gain-loss framing domain and the attribute-goal framing tactic used in an ad message (e.g., an attributegain framed ad versus an attribute-loss framed ad and a goal-loss framed versus a goal-gain framed ad) Hypotheses 1a~d predicted that the attribute-gain framing combination would have greater persuasive effects than would the attribute-loss framing combination and the persuasive effects would be greater in the goal-loss framing combination than in the goal-gain framing combination. These hypotheses were developed based on the postulates offered in the prospect theory, which state that loss outcome (goal) presentation is more likely to be persuasive than a corresponding gain outcome presentation. Recent empirical findings on the differential effect of gain-loss framing depending on the attribute-goal framing tactic also served as a rationale for the Age Education Economic Status Experience Attitude Information seeking intention Intention for consulting doctor Intention for HRT choice Age r p 1 Education r p .05 .43 1 Economic Status r p .02 .72 .37 .00 1 HRT Experience r p .22 .00 .07 .21 .04 .46 1 Attitude r p .03 .65 .04 .45 .01 .89 .32 .00 1 Information seeking Intention r p .13 .03 .06 .28 .04 .50 .1 7 .00 .49 .00 1 Intention for Consulting Doctor r p .14 .02 .05 .42 .02 .68 .30 .00 .75 .00 .626 .00 1 Intention for HRT Choice r p .11 .06 .01 .83 .00 .97 .17 .00 .40 .00 .73 .00 .58 .00 1

PAGE 63

63 en-US prediction that attribute-gain presentation would be more effective than attribute-loss presentation. en-US Hypotheses 2a~d were generated to examine a two-way interaction between the gain-loss framing domain and message source credibility. These hypotheses predicted that the persuasiveness of loss framing would be significantly less distinctive in low credible source than high credible source condition, whereas the persuasiveness of gain framing would be constant regardless of the level of source credibility. This interaction was predicted based on the postulates included in the language expectancy theory, which state that the effectiveness of negative language (e.g., loss outcomes or negative attributes of a decision object) in persuasive communication is limited with low credible sources. en-US The persuasiveness of message framing and source credibility was examined in two dimensions: attitude toward hormone replacement therapy and DTC ad -promoted intentions. The DTC ad -promoted intentions were classified into intention to search for further information about hormone replacement therapy, intention to consult a doctor, and intention to choose hormone replacement therapy. en-US MANCOVA analyses were performed to test the hypotheses because the current study design contained more than one dependent variable, and six pre-experimental factors were included with three independent variables. Despite random assignment of the samples for the experiment, some potential confounding factors were considered uncontrollable in view of the controversial nature of the study topic, the use of hormone replacement therapy, and the nature of the quasi-experiment design (see Shah et al., 2002). For example, previous experience with hormone replacement therapy might influence intention to choose the treatment option or attitudes toward the drug. The correlation analyses also confirmed that the level of experience

PAGE 64

64 en-US with the hormone therapy was positively correlated with all three dependent variables (see Table 4-8). In addition to previous experience, product knowledge, attitude toward hormonal medications in general (e.g., steroid, contraceptives), involvement with hormone replacement therapy, perceived risk of the side effect were included as covariates in the analyses. In addition to the MANCOVA analysis for testing H1a~d through H2a~d, separate ANCOVA contrast analyses for each dependent variable were performed to further investigate the relative effectiveness of each combination of framing domain, framing tactic and source credibility to other combinations. en-US A research question was developed to explore a three-way interaction among framing domain, framing tactic, and source credibility on the dependent variables. Keren (2005) found that consumers tend to perceive an attribute-loss (negative) framed message as being more trustworthy than an attribute-gain (positive) message. This suggests the possibility that an attribute-loss framed message may be more persuasive than an attribute-gain framed message with a high-credible source. Alternatively, the attribute-loss framing impact might be as distinct as the attribute-gain framing impact, even in a low-credible source condition. The research question sought to explore these potential dynamics. en-US Hypotheses 1a~d en-US Hypotheses 1a~d were posited to test a two-way interaction effect between framing domain and tactic on attitude and DTC-promoted intentions. en-US H1 a : The attribute gain-framed DTC advertisement will produce more positive attitudes toward hormone replacement therapy (HRT) than attribute loss-framed DTC advertisement and the goal loss-framed DTC advertisement will produce more positive attitudes toward HRT than goal gain-framed DTC advertisement. en-US en-US H1b: The attribute gain-framed DTC advertisement will produce stronger intentions of seeking more information about HRT than attribute loss-framed DTC advertisement and

PAGE 65

65 en-US the goal loss-framed DTC advertisement will produce stronger intentions of seeking more information on HRT than goal gain-framed DTC advertisement. en-US en-US H1c: The attribute gain-fram ed DTC advertisement will produce stronger intentions of consulting a doctor about HRT than attribute loss-framed DTC advertisement and the goal loss-framed DTC advertisement will produce stronger intentions of consulting a doctor about HRT than goal gain-framed DTC advertisement. en-US en-US H1d: The attribute gain-fram ed DTC advertisement will produce stronger intentions of choosing HRT than attribute loss-framed DTC advertisement and the goal loss-framed DTC advertisement will produce stronger intentions of choosing HRT than goal gainframed DTC advertisement. en-US en-US The independent variables were gain-loss framing domain and attribute-goal framing tactic, and the six covariates were submitted to a two-way MANCOVA for the predicted interaction effect. The MANCOVA results (see Table 4-9) indicated that no significant interaction effect exist ed between framing domain and framing tactic on the attitude toward the HRT and DTC-promoted intentions [Wilks = .99, F (4, 249) = .45, p > .05, p 2 = .007] Therefore, hypotheses 1a~d were not supported by the data obtained from the present study. en-US Table 4-9. MANCOVA tests for faming domain, framing tactic, and source credibility Independent variable Wilks Lambda F H df Error df p h p 2 Frame Domain .96 4.80 4 249 .03 .04 Frame Tactic .98 2.60 4 249 .23 .02 Source Credibility .96 2.48 4 249 .04 .04 Product Knowledge .94 3.85 4 249 .00 .06 Involvement .74 21.78 4 249 .01 .26 Perceived Side effects .91 6.54 4 249 .00 .10 General Attitude .82 13.76 4 249 .00 .18 Experience .90 6.62 4 249 .00 .10 Age .95 3.56 4 249 .01 .05 Domain Tactic .99 .45 4 249 .77 .01 Doma in Source .96 2.37 4 249 .05 .04 Tactic Source .96 2.79 4 249 .02 .04 Domain Tactic Source .98 1.32 4 249 .26 .02

PAGE 66

66 en-US Since no significant interaction was found between the gain-loss framing domain and attribute-goal framing tactic, the main effects of framing domain and framing tactic were examined through additional testing. The results of MANCOVA (see Table 4-9) showed that a main effect of the gain-loss framing domain on attitude toward hormone replacement therapy and the three DTC ad -promoted intentions were significant [Wilks = .9 6, F (4, 249) = 4.80, p < .0 5, p 2 = .040], while no main effect of the attribute-goal framing tactic was found [Wilks = .9 8, F (4, 249) = 2.60, p > .0 5, p 2 = .022]. en-US Since the results of the multivariate test show a significant main effect of framing domain, the effect of framing domain on each dependent variable was examined. The Table 4-10 indicates that the effects of the gain-loss framing domain were significant on attitude toward hormone replacement therapy [ F (1, 266) = 5. 63, p < 05 M gain = 4.08 SD = 1.38, M loss = 4.36 SD = 1.24], intention for information-seeking [ F(1, 266) = 6.44, p < .05, M gain = 4.00 SD = 1. 92, M loss = 4.60, SD = 1.77], and intention to consult a doctor [ F (1, 266) = 3. 73 p = .055 M gain = 4.15 SD = 2.17, M loss = 4.73 SD = 2.06]. However, the main effect of the gain-loss framing domain on intention to choose the hormone therapy option was not significant ( p < .05). Overall, the MANCOVA model suggests that a loss-framed ad elicited a more positive attitude toward HRT and stronger DTC ad -promoted intentions than a gain-framed ad, regardless of framing tactic. en-US ANCOVA Contrast Analyses with Framing Domain and Framing Tactic en-US As an additional test, analyses of ANCOVA contrast using the framing combination (e.g., attribute-gain versus attribute-loss framing; goal-loss versus goal-gain framing) as the independent variable were performed. These provided further examination of the significant mean differences in each dependent variable between participants exposed to a gain-framed ad

PAGE 67

67 en-US and a loss-framed ad. Although no significant interaction was found between framing domain and tactic in the MANCOVA results, the ANOVA contrasts would provide an insight into which framing combination exerted a greater persuasive impact than other combinations. en-US At titude toward hormone replacement therapy; The ANCOVA contrasts revealed that an ad framing attribute-loss elicited a more positive attitude than did an ad framing attribute-gain [ F (1, 259)= 4.01 p < .05, M attribute-loss = 4.32, SD = 1.28 M attribute-gain = 4.03, SD = 1.61], which was a reversed result compared to the effect proposed in H1a. An ad framing goal-loss had a more positive attitude than an ad framing attribute-gain [ F (1, 259)= 5.35, p < .05 M goal-loss = 4.40 SD = 1.21 M attribute-gain = 4.03 SD = 1.61]. However, no significant difference was noted between a goal-loss framed ad and a goal-gain framed ad ( p > .05). en-US DTC ad-promoted intentions; ANCOVA contrasts revealed that there was no significant difference in intention for information-seeking, intention to consult a doctor, and intention to choose the hormone therapy option for participants who were exposed to ads framing attribute-loss and attribute-gain. Additionally, no significant difference was observed in the three DTC ad -promoted intentions between participants who read a goal loss-framed ad and a goal gain-framed ad. en-US Hypotheses 2a~d en-US Hypotheses 2a~d were proposed to test a two-way interaction between gain-loss framing domain and source credibility. en-US H2: There will be a two-way interaction effect between gain-loss framing domain and source credibility on attitude toward HRT and DTC ad -promoted intentions. Specifically, the effect of frame domain (gain versus loss framing) on attitude toward HRT (H2a), intention to seek more information (H2b), intention to consult a doctor for further help ( H2 c), and intention to choose HRT (H2d) would be less distinctive when an ad message is presented by a low-credibility rather than a high-credibility source. Notably, the gainframing impact on the dependent variables would be constant, regardless of the level of the ad message source credibility. en-US

PAGE 68

68 en-US As shown in Table 4-9, the MANCOVA results, with attitude toward HRT and the three DTC ad -promoted intentions as the dependent variables, revealed a significant interaction effect between framing domain and source credibility. [Wilks = .9 6, F (4, 249) = 2.37 p = .0 5, p 2 = .04]. As the multivariate interaction between frame domain and source credibility was significant, the interactive effect on each dependent variable was examined to test hypotheses 2a~d. The Table 4-10 shows that an interaction between framing domain and source credibility significantly influenced attitude toward HRT ( F = 3.62, p < .05) and intention to choose HRT ( F = 6.45, p < .05). en-US However, the interaction effects on intention to seek more information and intention to consult a doctor did not approach significance (p > .05). Thus, H2b and H2c were not supported. en-US Table 4-10. Results of between-subject test en-US Independent variable Dependent Variable d f Mean Square F P Frame Domain Attitude toward HRT 1 3.91 5.63 0.02 Intention for Information seeking 1 15.64 6.44 0. 01 Intention to consult doctor 1 11.02 3.73 0.05 Intention to choose HRT 1 2.92 1.93 1.66 Frame Tactic Attitude toward HRT 1 0.55 0.80 0.37 Intention for Information seeking 1 5.38 2.21 0.14 Intention to consult doctor 1 0.79 0.27 0.61 Intention to choose HRT 1 1.28 0.85 0.36 Source Credibility Attitude toward HRT 1 4. 90 7.04 0.01 Intention for Information seeking 1 0.31 0.13 0.72 Intention to consult doctor 1 2.74 0.93 0.34 Intention to choose HRT 1 7.16 4.74 0.03 Knowledge Attitude toward HRT 1 3.24 4.66 0.03 Intention for Information seeking 1 1.94 0.80 0.37 2 Intention to consult doctor 1 3.87 1.31 0.25 Intention to choose HRT 1 0.03 0.02 0.89 Involvement Attitude toward HRT 1 22.55 32.47 0.00 Intention for Information seeking 1 92.74 38.15 0.00 Intention to consult doctor 1 139.13 47.04 0.00 Inten tion to choose HRT 1 73.49 48.61 0.00 Perceived side effects Attitude toward HRT 1 15.945 22.92 0.00 Intention for Information seeking 1 0.11 0.05 0.83 Intention to consult doctor 1 0.05 0.02 0.00 Intention to choose HRT 1 7.40 4.89 0.90

PAGE 69

69 en-US Table 4-10. Continued en-US en-US ANCOVA Contrast Analyses with Gain-loss Framing Domain and Source Credibility en-US ANCOVA contrast analyses were conducted to examine the mean differences in attitude toward HRT (H2 a) and intention to choose HRT (H2d) between participants in the loss (gain) framing with a high credible source condition, and the loss (gain) framing with a low credible source condition. Attitude toward hormonal drug s Attitude toward HRT 1 17.68 25.41 0.00 Intention for Information seeking 1 2.79 1.15 0.29 Intention to consult doctor 1 12.79 4.31 0.04 Intention to choose HRT 1 61.66 40.78 0.00 Age Attitude toward HRT 1 3.83 25.42 0.02 Intention for Information seeking 1 1.36 1.15 0.46 Intention to consult doctor 1 0.24 4.31 0.78 Intention to choose HRT 1 1.52 40.78 0.32 Experience Attitude toward HRT 1 13.37 19.22 0.00 Intention for Information seeking 1 0.33 0. 13 0.71 Intention to consult doctor 1 2.63 0.89 0.35 Intention to choose HRT 1 18.40 12.17 0.00 Domain Tactic Attitude toward HRT 1 0.28 0.41 0.53 Intention for Information seeking 1 0.00 0.00 0.98 Intention to consult doctor 1 0.02 0.01 0.93 I ntention to choose HRT 1 2.28 1.51 0.22 Domain Source Attitude toward HRT 1 2.52 3.62 0.06 Intention for Information seeking 1 0.00 0.00 0.97 Intention to consult doctor 1 2.18 0.74 0.39 Intention to choose HRT 1 9.76 6.45 0.01 Tactic Source Atti tude toward HRT 1 3.91 5.63 0.02 Intention for Information seeking 1 15.64 6.44 0. 01 Intention to consult doctor 1 11.02 3.73 0.05 Intention to choose HRT 1 2.92 1.93 1.66 Domain Tactic Source Attitude toward HRT 1 1.42 2.04 .132 Intention for In formation seeking 1 10.50 4.32 0.14 Intention to consult doctor 1 1.39 0.47 0.63 Intention to choose HRT 1 6.86 4.54 0.12

PAGE 70

70 en-US Attitude toward Hormone Replacement Therapy; The ANCOVA contrast results revealed that a loss-framed ad from a low credible source ( M loss-credibility low = 3.94, SD = 1.11) induced a less positive attitude toward hormone replacement therapy [ F = 10.06 df = (1, 259), p < .01] than did a loss-framed ad from a high credible source ( M loss-credibility high = 4.79, SD = 1.22). However, the contrasts showed no mean difference in attitude toward HRT between the participants in a gain-framed ad compared to a high credible source and a gain-framed ad from a low credible source [ F (1, 259) = .36, p = .55 M gain-credibility high = 4.23 SD = 1.29 M gain-credibility low = 3.95, SD = 1.46]. These contrast results confirmed the MANCOVA results in support of H2a (see Table 4-11). Therefore, H2a was supported. en-US Table 4-11. Means and standard deviations for framing domain and source credibility for attitude toward HRT en-US en-US en-US en-US en-US en-US Figure 4-1 shows the interaction effect between gain-loss framing domain and source credibility on attitude toward HRT. en-US en-US Figure 4-1. Interaction effect of gainlo ss framing domain and source credibility on attitude toward HRT Framing Domain Loss Framing Gain Framing Source Credibility High n = 71 M = 4.79 SD = 1.22 n = 76 M = 4 23 SD = 1.2 9 Low n = 70 M = 3.94 SD = 1.11 n = 76 M = 3. 9 5 SD = 1 46

PAGE 71

71 en-US Intention to Seek Information ; The ANCOVA contrast results showed that no mean difference existed in intention for information-seeking between participants in a loss-framed ad from a low credible source and a loss-framed ad from a high credible source [ F = 0.31, df = (1, 264) p > .05, M loss-credibility low = 4.37, SD = 1.78, M loss-credibility high = 4.8 4 SD = 1.74]. The contrasts also revealed no mean difference in intention to seek information between participants in a gain-framed ad from a high credible source and a gain-framed ad from a low credible source. en-US Intention to Consult a Doctor; A mean difference in intention to consult a doctor between a loss-framed message from a high credible source and that from a low credible source was not observed from the obtained data [ F = 0.31, df = (1, 264) p > .05 M loss-credibility high = 4.78, SD = 2.13 M loss-credibility low = 4.68, SD = 1.99]. en-US Intention to Choose HRT ; As shown in Table 4-12, a loss-framed ad message from a high credible source ( M loss-credibility high = 4.93, SD = 1.80) produced stronger intention to choose HRT [ F = 11 .60 df = (1, 262), p < .01] than did a loss-framed ad message from a low credible source ( M loss-credibility low = 3.61, SD =1.87). However, the contrasts showed no mean difference in intention to choose HRT between the participants in a gain-framed ad from a high credible source and a gain-framed ad from a low credible source [ F (1, 262) = .12 p = .73 M gain-credibility high = 4.09 SD = 2.01 M gain-credibility low = 3.78, SD = 2.17]. These contrast results confirmed the MANCOVA results in support of H2d. Therefore, H2d was supported. en-US en-US en-US en-US en-US

PAGE 72

72 en-US Table 4-12. Means and standard deviations for framing domain and source credibility for Intention to choose HRT en-US en-US en-US en-US en-US en-US en-US en-US en-US Figure 4-2 shows the interaction effect between gain-loss framing domain and source credibility on intention to choose HRT. en-US en-US Figure 4-2. Interaction effect of gain-loss framing domain and source credibility on intention to choose HRT. en-US en-US Research Question en-US The research inquiry was to explore a three-way interaction effect among the gain-loss framing domain, the attribute-goal framing tactic, and the source credibility on the four dependent variables. The results of MANCOVA (see Table 4-9) did not demonstrate any p = .26). en-US Framing Domain Loss Framing Gain Framing Source Credibility High n = 71 M = 4.93 SD = 1.80 n = 76 M = 4 09 SD = 2 02 Low n = 70 M = 3.61 SD = 1.87 n = 76 M = 3. 78 SD = 2.18

PAGE 73

73 en-US Table 4-13. ANCOVA contrast tests for loss framing-low credibility condition Contrast conditions (Loss framing low credibility versus :) Loss framing hi gh credibility Gain framing high credibility Gain framing low credibility Attitude toward HRT .470 (.148) p =.002 .044 (.146) p =.764 .042 .146 p =.772 Intention to choose HRT .745 .219 p =.001 .116 .218 p = .594 .190 .216 p = .380 en-US NOTE: One tail contrast tests, entries in each cell refer to contrast estimate, standard errors, and p-value en-US en-US Table 4-14. ANCOVA contrast tests for gain framing-low credibility condition Contrast conditions ( Gain framing low credibility versus :) Loss framing high credibility Loss framing low credibility Gain framing high credibility Attitude toward HRT 512 (.14 7 ) p =.00 1 .0 42 (.14 6 ) p =. 772 .0 86 (.14 4 ) p =. 549 Intention to choose HRT 555 .21 6 p =.0 11 .1 90 .21 6 p = 380 074 .21 3 p = 730 en-US NOTE: One tail contrast tests, entries in each cell refer to contrast estimate, standard errors, and p-value en-US en-US Additional Testing en-US ANCOVA Contrast Analyses with Gain-loss Framing Domain, Attribute-goal Framing Tactic, and Source Credibility en-US The contrasts showed that a goal-loss framed ad message with a high credible source induced a more positive attitude toward hormone therapy than a goal-loss with a low credible source [ F = 11.77, df = (1, 255) p < .01 M goal -oss-credibility high = 4.98 SD = 1.18 M goal -loss-credibility low = 3.80, SD = .92]. In addition, no mean difference in attitude toward hormone therapy was observed between a goal-gain framed ad message with a high credible source and that with a low credible source ( F = 1.99, p > .05). These contrast results further support H2a.

PAGE 74

74 en-US In addition, if an ad message framing goal-loss was from a high credible source, it produced strong intention for information-seeking compared to a message framing goal-loss from a low credible source [ F = 4.79, df = (1, 260) p < .05 M goal-loss-credibility high = 5.12 SD = 1.58 M goal-loss-credibility low = 3.79, SD = 1.91]. Meanwhile, no mean difference in intention for information-seeking was observed between a goal-gain framed ad message with a high credible source and that with a low credible source ( F = .04, p > .05). en-US No mean difference in intention to consult a doctor was observed between a goal-loss framed ad message with a high credible source and that with a low credible source [ F = 0.28, df = (1, 259) p > .05, M goal -loss-credibility high = 4.88, SD = 2.21, M goal -loss-credibility low = 4.48, SD = 2.12] In addition, no mean difference in intention to consult a doctor was observed between a goalgain framed ad message with a high credible source and that with a low credible source ( F = .02, p > .05). en-US The ANCOVA contrast analyses showed a significant mean difference in intention to choose hormone replacement therapy between a goal-loss framed ad message with a high credible source and that with a low credible source [ F = 19.44, df = (1, 258), p < .01, M goal-losscredibility high = 5.19 SD = 1.70 M goal-loss-credibility low = 3.08, SD = 1.67]. Meanwhile, no mean difference in intention to choose hormone therapy was observed between a goal-gain framed ad message with a high credible source and that with a low credible source ( F = .28, p > .05). These data further support H2d. en-US MANOVA Results en-US A MANOVA was performed to investigate whether the present study findings were significantly different when all covariates were removed. As shown in the Table 4-15, the MANOVA results were not significantly different from the MANCOVA results. No significant interaction effect exist ed between framing domain and framing tactic on the attitude toward the

PAGE 75

75 en-US HRT and DTC-promoted intentions [Wilks = .99, F (4, 273) = .26, p > .05, p 2 = .004] However, there was a significant two-way interaction effect of framing domain and source credibility on the dependent variables [Wilks = .97, F (4, 273) = .25, p < .05, p 2 = .035]. In addition, no significant three-way interaction was found from the MANOVA results [Wilks = .99, F (4, 273) = .80, p > .05, p 2 = .012] Therefore, the MANOVA revealed similar results to the MANCOVA indicating that the blocking variables were successfully controlled in the experiment. en-US Table 4-15. MANOVA tests for faming domain, framing tactic, and source credibility Independent variable Wilks Lambda F H df Error df p h p 2 Frame Domain .97 1.98 4 273 .10 .03 Frame Tactic .99 .71 4 273 .59 .0 1 Source Credibility .93 4.60 4 273 .00 .0 7 Domain Tactic .99 .26 4 273 .90 .00 Domain Source .96 2.50 4 273 .04 .04 Tactic Source .97 2. 2 9 4 273 .0 6 .03 Domain Tactic Source .9 9 80 4 2 73 .53 .01 en-US en-US Summary of the Results en-US H1 a: The attribute gain-framed DTC advertisement will produce more positive attitudes toward hormone replacement therapy (HRT) than attribute loss-framed DTC advertisement and the goal loss-framed DTC advertisement will produce more positive attitudes toward HRT than goal gain-framed DTC advertisement (Not supported) en-US H1b : The attribute gain-framed DTC advertisement will produce stronger intentions of seeking more information about HRT than attribute loss-framed DTC advertisement and the goal loss-framed DTC advertisement will produce stronger intentions of se eking more information on HRT than goal gain-framed DTC advertisement (Not supported)

PAGE 76

76 en-US H1c : The attribute gain-fram ed DTC advertisement will produce stronger intentions of consulting a doctor about HRT than attribute loss-framed DTC advertisement and the goal lossframed DTC advertisement will produce stronger intentions of consulting a doctor about HRT than goal gain-framed DTC advertisement (Not supported) en-US H1d : The attribute gain-fram ed DTC advertisement will produce stronger intentions of choosing HRT than attribute loss-framed DTC advertisement and the goal loss-framed DTC advertisement will produce stronger intentions of choosing HRT than goal gain-framed DTC advertisement (Not supported) en-US H2: There will be a two-way interaction effect between gain-loss framing domain and source credibility on attitude toward HRT and DTC-promoted intentions. Specifically, the effect of frame domain (gain versus loss framing) on attitude toward HRT (2a) intention to seek more information (2b) intention to consult a doctor for further help (2c) and intention to choose HRT (2d) would be less distinctive when an ad message is presented by a low credible rather than a high credible source. Notably, the gain framing impact on the dependent variables would be constant, regardless of the level of the ad message source credibility (The obtained data found a two-way interaction between framing domain and source credibility: H2a and H2d were supported, and H2b and H2c were not supported) en-US RQ : Is there a three-way interaction effect among framing tactic, framing domain, and source credibility? ( The obtained data did not find a three-way interaction among framing domain, framing tactic, and source credibility).

PAGE 77

77 en-US CHAPTER 5 en-US DISCUSSION en-US This chapter presents an evaluation of each hypothesis in terms of the study findings reported in Chapter 4. The chapter begins with a summary of the study and findings, followed by a discussion of the implications for framing theory, advertising theory, DTC drug advertising practice, health marketing practice. Then the chapter will conclude by presenting the limitations of the study and suggestions for future research. en-US Summary of Study en-US Gain-loss framing manipulation has been a central element in the investigation into the effects of message framing on persuasion and decision making in the context of risky or uncertain choices. Prospect theory is a theoretical framework that posits the relative advantage of a loss-framed persuasive message over the corresponding gain-framed persuasive message. However, a large body of research in the framing literature has failed to support the idea that the impact generated by loss framing is more significant than that of the corresponding gain framing. The present study hoped to provide an understanding of the reason for the mixed findings in the framing literature. It investigated the potential interaction between the gain-loss framing domain and the attribute-goal framing domain in the context of DTC drug advertising and explored the moderating impact of source credibility in the relationship between framing and persuasion. The three independent variables of this study: framing domain, framing tactic, and source credibility had not been previously linked within one single study design Moreover, little research has investigated the interactive effect of framing tactic and framing domain in the context of prescription drug decision making that involves uncertain elements. en-US This research attempted to propose and test framing hypotheses (H1a~d ) in the context of DTC prescription drug advertising by relating another dimension of framing manipulation in

PAGE 78

78 en-US persuasive messages, namely the attribute-goal framing to the traditional message framing manipulation (gain versus loss framing) described in prospect theory. In the creative practice of advertising, attribute-framing messages can be created by emphasizing product attributes (e.g., components of drugs, level of consumer satisfaction about drug efficacy, or types or degree of side effects), while goal framing messages can be devised by accentuating the product outcomes functional or emotional benefits (e.g., health benefits as a result of drug use). en-US In addition, the present research relates the study of message framing to an important element of the advertising message; specifically, source credibility. Despite the fact that message source characteristics are considered one of the most important factors (e.g., message characteristics; message recipient) in persuasive communication and cognitive message processing aside from the message itself (Perloff, 2003; Fisk, 1991), only a handful of studies have attempted to investigate the relationship between message framing and message source credibility in the persuasion and decision making literature. Therefore, another critical focus of the present study was to explore the potential interaction of message framing domain (gain-loss framing) and source credibility (source expertise) in DTC drug advertising messages (H 2a~d). In persuasive communications such as advertising, the impact of the message source cue is considered particularly important in maintaining, enhancing, or even changing attitudes, behavioral intentions, or even actual choice. en-US Hormone replacement therapy was chosen as the product category. Selection of the drug category was based on the postulates enunciated in the prospect theory, which assume that the gain-loss framing effect can be observed when choice objects or behaviors involve risky or uncertain elements. Hormone replacement therapy is generally recognized as a risky or uncertain health decision choice for postmenopausal women (McIntosh & Blalock, 2005; Andrist, 1998),

PAGE 79

79 en-US and empirical evidence has shown that consumers generally perceive hormonal medicines as studies have investigated the interactive effect of the framing tactic and framing domain in the context of prescription drug or health care product decision making where risky or uncertain s are involved (e.g., concern about side effects from the use of hormonal medications, such as steroids or perceived uncertainty about the efficacy of prescription drugs, such as medical abortion pills and HPV vaccinations). en-US To develop the hypotheses and research question, this study attempted to integrate two theoretical perspectives in communication and decision science: the prospect theory (Kahneman & Tversky, 1979; Tversky & Kahnemen, 1981) and the language expectancy theory (Burgoon 1995; Burgoon, Denning, & Robers, 2002). Prospect theory was adopted to account for why outcome presentation (goal-loss framing combination) than to a commensurate gain outcome presentation (goal-gain framing combination) (H1a~d). The language expectancy theory was proposed to provide a rationale for the prediction that the message framing effect would be more distinctive in the high credible source condition than the low credible source (H2a~d). According to language expectancy theory, the persuasive effect is limited when a message source negatively violates the expected uses of message or language tactics, whereas the effect is reinforced when the expectations of the message uses are congruous with the perceived source characteristics (Burgoon, 1989; Burgoon, 1990). Language expectancy theory postulates that the use of negative languages is not appropriate for low credible message sources (Burgoon et al., 2002). Thus, the use of loss framing as a negative message strategy is considered a violation of expectations for low-credible sources, and thus, attitude or behavioral changes are likely to be limited when

PAGE 80

80 en-US negative information is presented with low source credibility. The prediction based on these assumptions was tested in the present study. en-US In addition, a series of empirical research findings that reported an interaction between framing domain (gain versus loss framing manipulation) and framing tactic (attribute versus goal fr aming manipulation) were described, which provided the rationale for the hypotheses (H1a~d) of the present study. Another important focus of the present study was the exploration of a threeway interaction between framing domain, faming tactic, and message source credibility in DTC drug advertising. The research question attempted to address the relative advantages of framing combinations (e.g., attribute-gain framing versus goal-loss framing; goal-gain framing versus attribute-loss framing) across the level of message source credibility ). According to Keren (2005), people tend to value a message presenting negative attributes of a product and consider it trustworthy. Therefore, an attribute loss (negative) framed message could be as effective as an attribute gain (positive) framed message if the message source was perceived as a credible source. The research question was to explore these message framing and source dynamics in the context of risky drug decision making. en-US Only a handful of studies have illuminated the role of message source credibility in the relationship between message framing and persuasion outcomes. Given that the source factor is considered an important contextual factor for persuasion and judgment (Perloff, 2003), it was anticipated that message source could play a part in framing, particularly in drug marketing communications. en-US en-US

PAGE 81

81 en-US Evaluation of Findings en-US Hypotheses 1a~d en-US Two types of framing were the focus of the current project: gain-loss framing and attribute-goal framing. The author dubbed tactic. Hypotheses 1a~d were to examine a two-way interaction between gain-loss framing domain and attribute-goal framing tactic. Drawing on recent work on the potential interactive effect between the two, this study tested a set of framing hypotheses in the context of hormone replacement DTC advertising, there would be an interaction between framing domain and frame tactic. Specifically, hypotheses 1a~d predicted that attribute-goal framing would be more influential in attitude toward hormone replacement therapy, intention for seeking more information on the drug product and efficacy, and intention for choosing hormone replacement therapy These hypotheses were based on the interaction between framing domain and framing tactic. However, the study data did not find a significant two-way interaction effect. Moreover, the experiment did not produce any predicted difference in the persuasiveness of the gain or loss framing across the two framing tactics. Therefore hypotheses 1a~d were not supported by current study data. en-US The present study data show a strong main effect of framing domain (gain versus loss framing) that loss framing was more effective than gain framing regardless of framing tactic (attribute versus goal framing). Specifically, the obtained data show that the attribute-loss framing effect on attitude toward the drug product and DTC ad -promoted behavior intentions were more pervasive than attribute-gain framing effect, which is quite contradictory with the prediction of the hypothesis 1a as well as th Krishnamurthy et al, 2001; Levin et al, 1998). Interestingly, among four combinations of framing domain and tactic

PAGE 82

82 en-US (attribute-gain; attribute-loss, goal-gain and goal-loss framing), the attribute loss framing had the highest mean score for all dependent variables. en-US Overall, the obtained data revealed the loss framing advantage but the pattern of the interaction between framing domain and framing tactic proposed and empirically validated in some framing research (e.g., Levin et al., 1998) was not observed in the current study One plausible reason for the attribute-loss framing advantage over the attribute-gain might be that message source) depending on the attribute-goal framing domain (Keren, 2007). In other words, people tend to have more trust in the attribute loss (negative)-frame message than attribute-gain (positive)-framed message (Levin & Gaeth, 1988; Keren, 2007). In persuasion context, the selfdisclosure of some undesirable aspects in a message may suggest that the message is trustworthy. Given that the trustworthiness of a message or message source is an important element in persuasion (McGuire, 1985) and the attitude and intentions were the key dependent variables of the present study, the potential positive assessments of the attribute loss-framed ad might lead to the strong persuasive impact of attribute-loss framing over the rest of framing combinations. en-US Another plausible reason for the overall loss framing advantages in this study might be that the study topic, hormone replacement therapy, was perceived as a risky choice among the study subjects (women age of 45~65) that loss aversion induced by loss framing might lead to risk taking (e.g., positive attitude toward the drug or intentions for choosing the drug option) as articulated in prospect theory regardless of the framing tactic. Most previous studies investigating an interaction between gain-loss framing and attribute-goal framing did not consider these contexts where risk-taking occurs enunciated in prospect theory. Therefore, it is speculated that the study participants recruited from the HRT target groups were well aware of

PAGE 83

83 en-US the risky elements concerning the procedure of hormone replacement drug taking and its efficacy, which might lead to the strong loss advantage no matter what framing tactic was combined. Although the results were not statistically significant, the data appeared that goal-loss framing was more effective than goal-gain framing, which is congruent with the gain-loss framing framework based on prospect theory en-US Salovey et al. (2002) provide an insight into the different impact of gain versus lo ss framing depending on the type of a choice object. Based on empirical findings, they propose and prove a framing postulate that loss-framed medical (or health care) messages have more persuasive than the correspondent gain-framed messages when a choice object involves elements of risk in trying the medical option, while gain-framed messages have more persuasive effects than the correspondent loss-framed messages when no element of risk is perceived As discussed earlier in this chapter, recent survey studies and mass media reports have found that hormone replacement therapy is perceived as a risky option among the target consumers. en-US Hypotheses 2a~d en-US Hypotheses 2a~d tested a two-way interaction between gain-loss framing domain and source credibility such that the effect of framing domain on attitude toward HRT (2a), intention to seek more information (2b), intention to consult with health care providers for further help (2c), and intention to choose HRT (2d) would be less distinctive when an ad message is presented by a low credible rather than a high credible source. Notably, the gain framing impact on the dependent variables would be constant, regardless of the level of the ad message source credibility. en-US The present experiment produced a difference in the persuasiveness of the frames across the two source credibility conditions. The persuasiveness of the loss-framed message was reduced under the conditions of low credible-source. As predicted, source credibility and framing

PAGE 84

84 en-US domain (gain versus loss framing) interacted on measures of attitude (H2a) and intention to choose HRT (H2d), but the obtained data did approach significant differences for H2b (intention for information-seeking) and H2c (intention to consult a doctor). The se results imply that loss framing is affected by the level or source credibility such that the loss framing impact decreases with low credible source, while the gain framing impact in not affected as much as loss framing by source credibility. This is consistent with the postulates in the language expectancy theory suggesting that low or non-credible sources are limited in using message strategies such that low source credibility with positive (gain) message strategies may have as much persuasive impact as high source credibility, but using negative (loss) strategies may lower persuasiveness of low source credibility (Burgoon, Denning, & Roberts, 2002). en-US Further, the study results were somewhat consistent with previous study findings on message framing and source credibility in that message framing impact is likely to decrease in low or non-credible source conditions. For example, Jones, Sinclair, & Courneya (2003) revealed that a gain-framed message from a credible source was the most influential in promoting exercise intention that other faming and source credibility conditions. en-US However, Jones et al. study findings are not congruent with the present study results in terms of the direction of the interaction between source credibility and message framing. Jones et al. (2003) also found a significant two-way interaction between message framing and source credibility such that there was no difference in exercise intention between the negative (loss) framing with low credibility condition and negative (loss) with high credibility condition, whil e the positive (gain) framing with high credibility source produced significantly more positive intentions than the positive framing with low credibility source condition.

PAGE 85

85 en-US Although there has been little research examining the relationship of gain-loss framing and source credibility based on language expectancy theory, some empirical evidence supports the interaction observed in the present study For instance, Zhang et al. (1999), argue that that positive information is regarded as a general message format relative to negative information. Thus, the loss-framed message is more likely to be perceived as an irregular format that it is salient to people than the gain-framed message. They also suggest the possible link between loss framing and fear arousal in the context of health communication, in that the loss-framed message may induce negative emotions such as fear or threat. In line with this suggestion, Williams et al. (2001), revealed that the loss-framed health message promoting breast cancer screening yielded more perceived susceptibility to breast cancer than the gain-framed message. This account is especially relevant to understanding why the interactive effect of source credibility with framing can be more distinctive in the negative framing condition, particularly loss-goal framing, than the positive (gain) framing condition, because loss-goal framing is likely to produce negative feelings such as displeasure, anxiety, and fear (at least moderate fear) and risk perceptions as fear appeal messages may bring to consumers (see Salovey et al., 2002). Kahnenman, one of the founders of gain-loss framing effects also recognized the presence of negative feelings, such as displeasure, as a result of processing loss-framed messages, which leads to loss aversion (Kahneman, 2003). This empirical evidence adds to our understanding for the present study assumption and findings that the use of loss framing by a low-credible source could indicate a violation of language

PAGE 86

86 en-US expectations in persuasive attempts and lead to low persuasiveness. The data obtained also indicted a significant main effect of the source credibility indicating message source would be an important factor in the persuasiveness of drug advertising. en-US Implication for Framing Theory en-US It is noteworthy that most previous studies focused on the relative impact of gain-loss framing but did not consider the risky or uncertain choice context where the relative (goal) loss framing advantage is likely to be found. The present research considered a controversial and attempted to test the framing hypotheses articulated in prospect theory. The choice of the drug category was based on the recent empirical survey results and media reports that medical decision-making regarding hormone therapy among menopausal women and even health providers is y and side effects. en-US Although this study failed to test the framing hypotheses, it attempted to introduce the notion of attribute-goal framing tactic to the study of the gain-loss framing domain effect, thus suggesting that the persuasiveness of a loss-framed message is greater than a gain-framed message when the messages are combined with goal framing tactic. In contrast, the persuasiveness of a gain-framed message is greater than a loss-framed message when attribute framing tactic is used in the messages. en-US Research on attribute-goal framing has not been as active in the advertising literature or in that related to health care marketing communications. This research highlights the possibility that a persuasive message can be framed either by emphasizing th

PAGE 87

87 en-US Implication for Message Framing and Sour ce Credibility. en-US Based on the approaches of the prospect and language expectancy theories, the study anticipated that a loss-framed message in a high credible condition would have a more persuasive effect than a gain-framed message in a low credible condition. Basically, the present study hoped to reveal that the advantage of goal-loss framing in a risky decision context would be affected by message source. Although no three-way interaction between framing domain, framing tactic, and source credibility was found, the ANCOVA contrasts with the combination of the three factors as an independent variable provide an insight into the effect of source credibility on the message framing. The contrasts showed that a goal-loss framed ad message with a high credible source induced a more positive attitude toward hormone therapy and intention to choose hormone replacement therapy than goal-loss framed ad with a low credible source, which implies that the persuasiveness of goal-loss framing is likely to decrease with a low credible source. In contrast, no mean difference in attitude toward hormone therapy was observed between a goal-gain framed ad message with a high credible source and that with a low credible source. en-US These findings suggest that DTC drug advertising messages presented by a non-credible source can benefit by using a gain framing strategy, while the effectiveness of loss framing may likely be reduced with a non-credible source. Specifically, an advertisement promoting drugs involving risky elements (e.g., hormonal medications) should use a high credible source with the loss framing strategy. According to the data obtained, the attribute-loss framed messages had the most persuasive effect than any other framing conditions. en-US Despite similar results in the findings from the Jones et al. (2003) study, the interpretation of the present data contradicts what Jones et al. argue regarding an interaction of gain-loss framing and source credibility. They concluded from the perspective of ELM that for health-

PAGE 88

88 en-US related persuasive messages, gain (benefit) framing will have greater effects than loss (risk) framing because people are not likely to elaborate on loss-framed messages relative to gainframed messages in a low credible source condition. But their conclusion did not consider what the prospect theory postulates concerning the relative advantage of loss framing in the risky context. The choice context for the present study included the risky or uncertain element (choosing hormone replacement therapy among the target consumers) that it was expected that loss-framed messages in a high credible source might be the most effective. The findings from this study have important implications for drug advertisers and health care product marketers efficacy and side effects, and to increase their interest in a positive way and thus realize sales. en-US Implication for Drug Advertising Practices en-US Advertising materials designed to enhance drug choice are often presented to the consumers wi thout a great deal of understanding of theory. Literature that is focused on message framing may provide guidance for maximizing the effectiveness of DTC drug advertising messages. Drug products that are new, infrequently used (e.g., emergency contraceptives), and uncertain relative to the efficacy and side effects (e.g., hormone repl acement therapy, HPV vaccines, and medical abortion pills) may be better promoted by focusing on the negative consequences of rejecting the drug, or identifying in the ads the negative attributes of choosing a different method of treatment. For instance, messages in commercials promoting the use of an emergency contraceptive or PSA caused by getting a surgical abortion should focus on the negative (loss) outcomes (e.g., infection) of not choosing the option promoted or of choosing the surgical option rather than concentrating on the benefits of the emergency contraceptive. Also, when a new prescription drug category is introduced to a market, the findings of this study suggest that a loss-framed ad

PAGE 89

89 en-US will be more effective, because consumers will likely have limited information about the drug and they will be unsure of their own decision to use the drug. For the same reason, a loss-framed ad will be more effective even for an existing drug category when it is introduced a new group of target consumers with limited first-hand or vicarious experiences with the drug category. en-US Limitations en-US Of all the limitations acknowledged in the present experimental design, the ecological validity of manipulated stimuli may be first addressed. Given the nature of drug commercial messages, some would doubt whether attribute loss-framed ad messages that emphasize only the negative attributes of a drug are constructed in real-world advertising. In other words, any positive aspects of a drug would be clearly depicted along with its negative aspects in the realworld ad messages (e.g., two-sided message strategy). However, it is also true that the attribute loss-framed ad messages indirectly imply the positive aspects of a product promoted. For instance, a message that states 20% of the drug users experienced the side effect also mean s that 80% of the users did not experience the side effect at all en-US The current study data failed to support the hypotheses proposing a two-way interaction between attribute-goal framing tactic and gain-loss framing domain. This may have been due to a lack of clear distinction in operationalization between attribute framing and goal framing Despite the significant mean difference in perceptions between the two tactics, the distinction in the framing manipulation check of the current study was not large. Indeed, the solid manipulation check scales have not been developed yet in the framing literature. More sophisticated manipulation scales for message framing need to be developed. en-US Another limitation that should be recognized is associated with the quasi-experimental design used for the present study. A Web-based experiment was designed to accommodate the target consumers study participation. Given the limitation of study funds and in order to broadly

PAGE 90

90 en-US reach the target groups of hormone replacement therapy, the Web-based experiment was the best method for this study. However, although there was random assignment of the samples, it has a limitation in control over the potential confounding factors that might influence the study results. For instance, the current study did not control other social cognitive factors, such as the level of self efficacy, social stigma, or perceived future risk of developing a health problem. In addition given the study subjects demographic segments, women aged 45-65, some participants might find it difficult to fully understand the online experimental procedure. Therefore, future research should consider these potential outside factors that may affect the causal relationship between message framing and persuasive outcomes. en-US Another potential limitation can be also recognized in the characteristics of the study samples. Subjects in this study might not reflect the general population because their geographic background was concentrated in the Florida areas, where people share many of the same demographic (e.g., retired people) and psychographic characteristics (e.g., different lifestyle and value). In addition, the study participants were recruited from th e research company panels and thus, a limitation in establishing the external validity with the present data is acknowledged en-US Finally, the present study measured persuasion variables as dependent measures because the roles of advertising are associated with persuasion more so than with actual choice (purchase) of products. However, the framing effect based on the prospect theory directly relates to the actual choice or judgment. Given the present study samples comprises th e target consumers, the actual choice might have been a more valid measure as suggested in the prospect theory rather than the persuasion variables: intentions and attitudes en-US Suggestions for Future Research en-US A problem of the attribute-goal framing approach is that no sound theoretical framework has yet been developed to explain why gain framing can be more effective than loss framing in a

PAGE 91

91 en-US message emphasizing attributes of a product The present study also failed to find the interactive effect between framing domain and framing tactic on attitudes and intentions regarding HRT This study found that loss-framed messages were more persuasive regardless of the attribute-goal framing tactic One possible reason for this result could be because the study subjects (women aged 45-65) were well aware of the health risk of menopause and were motivated to take the uncertain option (hormone therapy) to avoid potential health risks even in attribute loss-framed message conditions containing unwanted information about the promoted drug Therefore, future research should consider a variety of contextual factors including different types of drug and consumer segments. en-US Gain-loss framing manipulation can be developed in various ways (Salovey, Schneider, & Apanovitch, 2002). The typical manipulation of gain-loss framing is that gain-framed messages focus on attaining a desirable outcome as a result of complying with a choice option and loss-framed messages emphasize attaining an undesirable outcome caused by not complying with a choice option. As Salovey and his colleague suggest, a framing study can be designed with different framing manipulations such that a gain-framed advertisement fo cusing on the benefit of either attaining a desirable outcome or avoiding an undesirable one, and a loss framedmessage emphasizing failure to attain a desirable outcome or attaining an undesirable one. Development of more systematic framing manipulation techniques would contribute to establishing robust framing hypotheses and our understanding of the mixed findings in the framing literature. en-US To the best of author s knowledge, the present study represents the first empirical test of the effects of multiple framing and source credibility in the context of DTC drug advertising. The author encourage future investigators to take this research initiative further by applying it to a

PAGE 92

92 en-US wider arrange of media options, consumer segments, and media contexts, so that the DTC drug advertising literature contributes to specifying a set of solid guidelines on which drug marketers can perform on. en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US

PAGE 93

93 en-US APPENDIX A en-US GOAL GAIN-FRAMED AD WITH HIGH SOURCE CREDIBILITY en-US en-US

PAGE 94

94 en-US APPENDIX B en-US GOAL LOSS-FRAMED AD WITH HIGH SOURCE CREDIBILITY en-US

PAGE 95

95 en-US APPENDIX C en-US ATTRIBUTE LOSS-FRAMED AD WITH HIGH SOURCE CREDIBILITY en-US en-US

PAGE 96

96 en-US APPENDIX D en-US ATTRIBUTE GAIN-FRAMED AD WITH HIGH SOURCE CREDIBILITY en-US en-US en-US

PAGE 97

97 en-US APPENDIX E en-US ATTRIBUTE GAIN-FRAMED AD WITH LOW SOURCE CREDIBILITY en-US en-US

PAGE 98

98 en-US APPENDIX F en-US GOAL GAIN-FRAMED AD WITH LOW SOURCE CREDIBILITY en-US en-US

PAGE 99

99 en-US APPENDIX G en-US ATTRIBUTE LOSS-FRAMED AD WITH LOW SOURCE CREDIBILITY en-US en-US

PAGE 100

100 en-US APPENDIX H en-US GOAL LOSS-FRAMED AD WITH LOW SOURCE CREDIBILITY en-US en-US

PAGE 101

101 en-US APPENDIX I en-US MAIN STUDY QUESTIONNAIRE en-US en-US

PAGE 102

102 en-US en-US Drug Advertisement Copy Test Your participation in this survey is voluntary and anonymous The main purpose of this experimental study is to evaluat drug advertisement copy. Please note that the questionnaire is not designed to judge your opinions about this issue, so there is no right or wrong answer to the following questions. There is no penalty for not participating. The survey is anonymous you cannot be linked to your responses in any way. Questions regarding this study should be directed to Dr. Treise (352) 392-6557. en-US en-US 1. The following questions are about the information of the ad you just read. Please response to the following questions on each of the scales below. For instance, if you agree very strongly, you would mark the 5. If you disagree very strongly, you would mark the 1. en-US en-US a. The ad focused on side effects and safety of the hormone replacement therapy and consumer satisfaction of the drug efficacy. Strongly disagree Strongly agree 1 2 3 4 5 en-US en-US b. The ad focused on relative health risk of suffering from menopausal symptoms as a result of taking or not taking hormone replacement therapy. Strongly disagree Strongly agree 1 2 3 4 5 en-US en-US c. The main message of the ad was described in a negative tone. Strongly disagree Strongly agree 1 2 3 4 5 en-US

PAGE 103

103 en-US d. The main message of the ad was described in a positive tone. Strongly disagree Strongly agree 1 2 3 4 5 en-US en-US en-US 2. The following questions are about the ad source (spokesperson). Please give your response to the following questions on each of the scales below. en-US 2-a. The source (spokesperson) of the ad was Not knowledgeable Knowledgeable 1 2 3 4 5 6 7 en-US 2-b Not Competent Competent 1 2 3 4 5 6 7 en-US 2Not expert Expert 1 2 3 4 5 6 7 en-US 2Not trained Trained 1 2 3 4 5 6 7 en-US 2Not experienced Experienced 1 2 3 4 5 6 7 en-US en-US en-US 3. Please rate your feeling towards hormone replacement therapy for each statement below. en-US 3Foolish Wise 1 2 3 4 5 6 7

PAGE 104

104 en-US 3Safe Risky 1 2 3 4 5 6 7 en-US en-US 3Harmful Beneficial 1 2 3 4 5 6 7 en-US en-US 3P l easant Unpleasant 1 2 3 4 5 6 7 en-US en-US 3-e. Hormone replaceme Waste of time Wise use of time 1 2 3 4 5 6 7 en-US en-US 3Good for me Bad for me 1 2 3 4 5 6 7 en-US en-US 3Useful Useless 1 2 3 4 5 6 7 en-US en-US 3-h. Hormone replacement ther Worthless Valuable 1 2 3 4 5 6 7 en-US en-US 3Ineffective Effective 1 2 3 4 5 6 7

PAGE 105

105 en-US 4. The next questions ask about the likelihood you seek more information about hormone replacement therapy. Please give your response to the following questions on each of the scales below. Not at all Very Likely 4 a. Please rate the probability that you would participate in a free educational program about hormone replacement therapy and efficacy of the advertised drug product if local health services offered the program. 1 2 3 4 5 6 7 4 b. Please rate the probability that you would go to other media sources to get more information about hormone replacement therapy. 1 2 3 4 5 6 7 en-US en-US 5. Please rate the probability that you would meet with a doctor if local health services offered a free opportunity to consult doctors about hormone replacement therapy. Not at all Very likely 1 2 3 4 5 6 7 en-US en-US 6-a. Please rate the probability that you would take hormone replacement therapy if you had menopausal symptoms. Not at all Very likely 1 2 3 4 5 6 7 en-US en-US 6-b. Please rate the probability that you would recommend hormone replacement therapy to others if they had menopausal symptoms. Not at all Very likely 1 2 3 4 5 6 7 en-US en-US en-US en-US en-US en-US en-US

PAGE 106

106 en-US 7. The following questions ask you about your knowledge of hormone replacement therapy. Please give your response to the following questions on each of the scales below. Little or no knowledge A great deal of knowledge 7 a. How do you rate yo ur knowledge of hormone replacement therapy relative to other people? 1 2 3 4 5 7 b. How do you rat e your knowledge of hormone replacement therapy relative to most of your friends? 1 2 3 4 5 en-US en-US 8. The following statements are about how important hormone replacement therapy is to you. Please give your response to the following questions on each of the scales below. en-US en-US 8Unimportant Important 1 2 3 4 5 6 7 en-US en-US 8Irrelevant R elevant 1 2 3 4 5 6 7 en-US en-US 8Worthless Valuable 1 2 3 4 5 6 7 en-US en-US en-US en-US

PAGE 107

107 en-US en-US 9. The following statements are your thoughts regarding the risks of potential side effects that hormone replacement drugs may cause Please give your response to the following statements on each of the scales below. en-US en-US 10-a. When I think about the side effects of hormone replacement therapy, I feel nauseous. Strongly disagree Strongly agree 1 2 3 4 5 6 7 en-US en-US 10-b. The side effects of hormone replacement therapy could put my life in danger. Strongly disagree Strongly agree 1 2 3 4 5 6 7 en-US en-US en-US 10. The following statements are about your general thoughts about hormone medications such as steroids and oral contraceptives. Please give your respon se to the following statements on each of the scales below. en-US 10Unfavorable Favorable en-US en-US 10Bad Good 1 2 3 4 5 6 7 en-US 10Harmful Beneficial 1 2 3 4 5 6 7 en-US en-US 10-d. I believe taking hormone medication in general Attractive Unattractive 1 2 3 4 5 6 7 en-US en-US

PAGE 108

108 en-US Background Information en-US en-US 1. Have you ever been on hormone replacement therapy? en-US Yes _______ No _______ en-US en-US herapy? Please write in the number: en-US en-US en-US en-US _________year(s) _________ month(s) en-US en-US en-US 3. Are you taking the hormone replacement therapy now? en-US en-US Yes _______ No _______ en-US en-US en-US 4. Have you ever suffered from menopausal symptoms such as hot flashes, night sweat, vaginal symptoms, etc.? en-US en-US Yes _______ No _______ en-US en-US 5. Has anyone among your family members, relatives or close friend taken hormone replacement therapy? en-US en-US Yes _______ No _______ en-US en-US 6. How old are you? Please write in the number: en-US en-US ________ years old en-US en-US en-US en-US

PAGE 109

109 en-US 7. What is your ethnic background? Please circle. en-US (1) Asian/Pacific Islander en-US (2) Black en-US (3) Hispanic en-US (4) Native American en-US (5) White en-US (6) Do not wish to disclose en-US en-US en-US 8. What is the highest degree of education you have earned? en-US en-US en-US en-US 10. Which economic class do you consider your family to be among the following five categories? en-US Working Class________ Lower Middle Class________ Middle Class_______ en-US Upper Middle Class________ Upper Class _________ en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US -Thank you-

PAGE 110

110 en-US APPENDIX J en-US DEBRIEFING STATEMENT en-US en-US Title: Interplay of Framing Tactic, Framing Domain, and Source Credibility in DTC Hormone Replacement Therapy Advertising: An Integration of the Prospect Theory and Language Expectancy Theory en-US en-US Investigator: Kenneth Kim (Doctoral Student, Department of Advertising at University of Florida) en-US en-US Thank you for your participation in this study. The purpose of the current study was to investigate the effect of the drug ad message on your attitude toward hormone replacement therapy, and your intentions to take the hormone therapy option. The ad message you viewed was phrased in an either positive or negative tone with focusing on either the product attributes or health benefits of hormone replacement therapy. For this, the fictitious brand named Seroxmo was created. The current study is hypothesizing that a negative ad message is more effective than a positive ad message when the ads emphasize the product benefits. In contrast, a positive message is more effective than a negative ad message when the messages stress the product attributes (For example, drug components, level of sideef fect etc.). This study is also testing the hypothesis that perception of the message source (spokesperson) credibility influences the ad message effects. Specifically, the researcher predicts that the negatively presented ad message effects decrease with the low credible source, while the positively presented message effect s increase with the high credible source. en-US en-US There was no deception used in the conduct of this study. However, should you feel discomfort or distress for any reason upon completion of this study, please contact the University of Florida Institutional Review Board at (352) 846-1494. If you should have any questions concerning your participation, please feel free to contact Kenneth Kim, Department of Advertising at University of Florida at (352) 275-8124 or kkim@jou.ufl.edu en-US en-US Your responses in this experiment are important and greatly appreciated. A report of this research should be ready for circulation by the end of August 2009. If you would like to receive a copy, please send an e-mail to Kenneth Kim expressing your interest in the results and providing your e-mail address. Thank you again for your participation. en-US en-US en-US en-US en-US en-US

PAGE 111

111 en-US LIST OF REFERENCES en-US Ajzen I., & Fishbein, M. (1980). Understanding attitude and predicting social behavior Englewood Cliffs, NJ: Prentice-Hall. en-US en-US Andrist, C. L. (1998). The impact of media attention, family history, politics and maturation on Health Care for Women International 19 243-260. en-US en-US Arora, R., Stoner, C., & Arora, A. (2006). Using framing and credibility to incorporate exercise Journal of Consumer Marketing 23(4), 199-207. en-US en-US Block, L. G., & Keller, P. A (1995). When to accentuate the negative: The effects of perceived efficacy and message framing on intentions to perform a health-related behavior. Journal of Marketing Research 32(2), 192-203. en-US en-US Berger, P. D., & Smith, G. E. (1998). The impact of prospect theory based framing tactics and advertising effectiveness. International Journal of Management Science 26(5), 593-609. en-US en-US Bloch, P., Ridgway, N. M., & Sherrell, D. L. (1989). Extending the concept of shopping: An investigation of browsing activity. Journal of the Academy of Marketing Science, 17, 1321. en-US en-US Boettcher, W. A. (2004). The prospects for prospect theory: an empirical evaluation of international relations applications of framing and loss aversion. Political Psychology 25(3), 331-336. en-US en-US Brewer, P. R., licy issues: Effects on content and quality Political Psychology 26(6), 929-948. en-US en-US Broemer, P. (2002). Relative effectiveness of differently framed health messages: the influence of ambivalence. European Journal of Social Psychology, 32, 685-703. en-US en-US Bruner, G. C., James, K. E., & Hensel, P. J. (2001). Marketing Scales Handbook: A compilation of multi-item measures. Chicago IL. American Marketing Association en-US en-US Buda, R. (2003). The interactive effect of message framing, presentation order, and source credibility on recruitment practices. International Journal of Management, 20 (2), 156163. en-US en-US Burger, H. G. (2006). Hormone therapy in the WHI era. Australian and New Zealand Journal of Obstetrics and Gynecology, 46 84-91. en-US en-US Burgoon, M. (1989). Message and persuasive effects. In J. Bradac (Ed.), Message effects in communication science (pp. 129-164). Newbury Park, CA: Sage. en-US

PAGE 112

112 en-US Burgoon, M. (1990). Language and social influence. In H. Giles & P. Robinson (Eds.), Handbook of language and social psychology (pp. 51-72). London: Wiley. en-US en-US Burgoon, M. (1995). Language expectance theory: Elaboration, explication, and extension. In C. R. Berger & M. Burgoon (Eds.), Communication and social influence processes (pp. 2952). East Lansing: Michigan State University Press. en-US en-US Burgoon, M. (1996). (Non)Compliance with disease prevention and control messages: Communication correlates and psychological predictors. J ournal of Health Psychology. 1, 279-296. en-US en-US Burgoon, M., Denning, V. P., & Roberts, L. (2002). Language expectancy theory. In J. p. Dillard, & M. Pfau (Eds.), The persuasion handbook: Developments in theory and practice (pp. 117-136). London: Sage. en-US en-US Burgoon, M., & Miller, G. R. (1985). An expectancy interpretation of language and persuasion. In H. Giles & R. N. St. Clair (Eds.), Recent advances in language communication and social psychology (pp. 199-229). London: Lawrence Erlbaum. en-US en-US Burton, S., & Lichtenstein, D. R. (1988). The effect of ad claims and ad context on attitude toward the advertisement. Journal of Advertising 17(1), 3-11. en-US en-US Chang, C. T. (2007). Health-care product advertising: The influences of message framing and perceived product characteristics. Psychology & Marketing, 24 (2), 143-169. en-US en-US Dillard J. P., & Pfau. M. (2002). The persuasion handbook: Developments in theory and practice London : Sage. en-US en-US Fagley, N. S., & Miller, P. M. (1987). The effects of decision framing on choice of risky vs. certain options. Organizational Behavior and Human Decision Processes 39, 264-277. en-US en-US Ferguson, E., & Gallagher, L. (2007). Message framing with respect to decisions about vaccination: The roles of frame valence, frame method and perceived risk. British Journal of Psychology 98, 667-680. en-US en-US Ferguson, E., Leaviss, J., Townsend, E., Fleming, P., & Lowe, K. C. (2005). Perceived safety of donor blood and blood substitutes for transfusion: the role of informational frame, patient groups and stress appraisals. Transfusion Medicine 15, 410-412. en-US en-US Fiske, S. T., & Taylor, S. E. (1991). S ocial cognition New York: McGraw-Hill. en-US en-US Frisch, D. (1993). Reasons for framing effects. Organizational Behavior and Human Decision Processes 54, 399-429. en-US

PAGE 113

113 en-US Genazzani A., Schneider, H., Panay, N., & Nijland, E. (2006). The European menopause survey 2005: Women's perceptions on the menopause and post menopausal hormone therapy. Gynecological Endocrinology 22(7). 369-375. en-US en-US Grewal, D., Gotlieb, G., & Marmorstein, H. (1994). The moderating effects of message framing and source credibility on the price-perceived risk relationship. Journal of Consumer Research, 21 (June), 145-153. en-US en-US Hastie, R., & Dawes, R. M. (2001). Rational choice in an uncertain world: the psychology of judgment and decision making London: Sage. en-US en-US Hoffmann, M., Hammar, M., Kjellgren, K., Lindh-Astrand, L., & Brynhildsen, J. (2005). Ch anges in women's attitudes towards and use of hormone therapy after HERS and WHI. Maturitas 52. 11-17. en-US en-US Hog ar th, R. M. (1987). Judgment and choice: The psychology of decision New York; Wiley. en-US en-US Hovland, C I., Janis, L. I., & Kelley, H. H. (1953). Communication and Persuasion New Haven CT : Yale University Press. en-US en-US Hovland, C I., & Mandell, W. (1957). Is there a 'law of primacy' in persuasion? In The Order of Presentation in Persuasion. Carl I. Hovland et al., New Haven, CT : Yale University Press. 1-22. en-US en-US Hovland, C. I., & Weiss, W. (1951). The influence of source credibility on en-US communication effectiveness. Public Opinion Quarterly vol. 15: p. 635-650. es-ES es-ES Huh, J., DeLorme, D. E., & Reid, L. N. (2005). Factors affecting trust in on line prescription drug informatio n and impact of trust on behavior following exposure to DTC advertising. Journal of Health Communication, 10 711 731. en-US en-US Jones, L., Sinclair, R., & Courneya, K. (2003). The effects of source credibility and message framing on exercise intentions, behaviors, and attitude: An integration of the elaboration likelihood model and prospect theory. Journal of Applied Social Psychology 33, 1, 179196. en-US en-US Kahneman, D. (2003). A perspective on judgment and choice: Mapping bounded rationality. American Psychologist 58, 9, 697-720. en-US en-US Kahneman, D., & Tversky, A. (1979). Prospect Theory: An analysis of decisions under risk. Econometrica, 47 263-291. en-US en-US Kahneman, D., & Tversky, A. (2000) Choices, values, and Frames. New York: Cambridge University Press. en-US

PAGE 114

114 en-US Kahneman, D., & Tversky, A. (1984). Choices, values, and frames. American Psychologist, 39 341-350. en-US en-US Kelman, H. C (1953). Attitude change as a function of response restriction. Human Relations, 6, 185-214. en-US en-US Kelman, H. C (1961). Processes of Opinion Change. T he public Opinion Quarterly, 2 5(1), 57-78. en-US en-US Kelman, H. C., & Hovland, C. I. (1953). "Reinstatement" of the communicator in delayed measurement of opinion change. Journal of Abnormal and Social Psychology, 48, 327335. en-US en-US Keren, G. (2007). Framing, intentions, and trust-choice incompatibility. Organizational Behavior and Human Decision Process 103, 238-255. en-US en-US Kuhberger, A. (1998). The influence of framing on risky decisions: A meta-analysis. Organizational Behavior and Human Decision Processes 75(1), 23-55. en-US en-US Kuhberger, A., Schulte-Mecklenbeck, M., & Perner, J. (2002). Framing decisions: Hypothetical and real. Organizational Behavior and Human Decision Processes 89, 1162-1175. en-US en-US Krishnamurthy, P., Carter, P., & Blair, E. (2001). Attribute framing and goal framing effects in health decisions. Organizational Behavior and Human Decision Processes 85, 382-399. en-US en-US Levin. I. P. (1987). Associative effects of information framing. Bulleting of the Psychonomic Society, 25 (2), 85-86. en-US en-US Levin. I. P., Chapman, D. P., & Johnson, R. D. (1988). How consumers are affected by the framing of attribute information before and after consuming the product. Journal of Consumer Research 15(3), 374-378. en-US en-US Levin, I. P., Gaeth, G. J., Schneider, S. L., & Lauriola, M. (2002). A new look at framing effects : Distribution of effect sizes, individual differences, and independence of types of effects. Organizational Behavior and Human Decision Processes 88, 411-429. en-US en-US Levin, I. P., Schneider, S. L., & Gaeth, G. J. (1998). All frames are not created equal: A typology and critical analysis of framing effects. Organizational Behavior and Human Decision Processes 76, 149-188. en-US en-US Linville, P. W., Fischer, G. W., & Fischhoff, B. (1993). AIDS risk perceptions and en-US decision biases. In J. B. Pryor & G. D. Reeder (Eds.), Th e social psychology of HIV infection (pp. 5-37). Hillsdale, NJ: Erlbaum. en-US en-US Maheswaran, D., & Meyers-Levy, J. (1990). The influence of message framing and issue involvement. Journal of Marketing Research, 27 (August) 361-367. en-US

PAGE 115

115 en-US Marteau, T. M. (1989). Framing of information: Its influence upon decisions of doctors and patients. British Journal of Social Psychology 28(1), 89-94. en-US en-US Maule, J., & Villejoubert G. (2007). What lies beneath: Reframing framing effects. Thinking & Reasoning 13(1), 25-44. en-US en-US McElroy, T., & Mascari, D. (2007). When is going to happen? How temporal distance influences processing for risky-choice framing tasks. Social Cognition 25(4), 495-517. en-US en-US McGinnies E., & Ward, C. D. (1980). Better liked than right: Trustworthiness and expertise as factors in credibility. Personality and Social Psychology Bulletin, 6, 467 472. Personality and Social Psychology Bulletin 6, 467-472. en-US en-US McGuire, W. J. (1985). Attitudes and attitude change. In G. Lindzey & E. Aronson (Eds.), The handbook of social psychology (Vol. 2, pp. 238-241). NY: Random House. en-US en-US McIntosh, J., & Blalock, S. (2005). Effects of media coverage of Women's Health Initiative study on attitudes and behavior of women receiving hormone replacement therapy. American Journal of Health-System Pharmacy, 62 69-74. en-US en-US McElroy, T., & Mascari, D. (2007). When is it going to happen? How temporal distance influences processing for risky-choice framing tasks. Social Cognition 25(4), 495-517. en-US en-US Mercer, J. (2005). Prospect theory and political science. An nual Review of Political Science 8, 121. en-US en-US Meyerowitz, B., & Chaiken, S. (1987). The effect of message framing on breast self-examination attitude, intentions, and behavior. Journal of Personality and Social Psychology, 52 500510. en-US en-US Mills, J., & Jelliso n, J. M. (1967). en-US Effect on opinion change of how desirable the communication is to the audience the communicator addressed.en-US Journal of Personality and Social Psychology, 5(4):459-63. en-US en-US Mittal, B. (1995). A comparative analysis of four scales of involvement. P sychology & Marketing 12, 663-682. en-US en-US Nan, X. (2007). Social distance, framing, and judgment: A construal level perspective. Human Communication Research 33, 489-514. en-US en-US Newell, B. R., Lagnado, D. A., & Shanks, D. R. (2007). Straight choices: The psychology of decision making. Hove: Psychology press. en-US en-US O'Connor, D. B., Ferguson, E., & O'Connor, R. C. (2005). Intentions to use hormonal male contraception: the role of message framing, attitudes, and stress appraisals. British Journal of Psychology 96, 351-369.

PAGE 116

116 en-US en-US O'Connor, A. M., Pennie, R. A., & Dales, R. E. (1996). Framing effects on expectations, decisions, and side effects experienced : The case of influenza immunization Journal of Clinical Epidemiology 49, 1271-1276. en-US en-US Oliver, R. L.(1980). A cognitive model of the antecedents and consequences of satisfaction decisions. J ournal of Marketing Research 17, 460-469. en-US en-US Perloff, R. M. (2003). The dynamics of persuasion : Communication and attitudes in the 21st cen tury Mahwah N J: Lawrence Erlbaum Associates. en-US en-US Petty, R. E., Priester, R, J., & Brinor R. (2002). Mass media attitude change: Implications of elaboration likelihood model of persuasion, in J. Bryan and D. Zillmann (Eds.), Media Effects: Advances in Theory and Researc h (pp. 453-488). Hillsdale, NJ: Lawrence Erlbaum Associates. en-US en-US Rothman, A. J., Bartels, R. D., Wlaschin, J., & Salovey, P. (2006). The strategic use of gainand loss-framed messages to promote healthy behavior: How theory can inform practice. Journal of Communication, 56 S202-S220. en-US en-US Rothman, A. J., & Salovey, P. (1997). Shaping perceptions to motivate healthy behavior: The role of message framing. Psychological Bulletin, 121 (1), 3-19. en-US en-US Salovey, P., Schneider, T. R., & Apanovitch, A. M. (2002). Message framing in the prevention and early detection of illness. In J. P. Dillard, & M. Pfau (Eds.), The persuasion handbook: Developments in theory and practice (pp. 391-406): Developments in theory and practice. London: Sage. en-US en-US Salovey, P., & Williams-Piehota, P. (2004). Field experiments in social psychology: Message framing and the promotion of health protective behaviors. The American Behavioral Scientist 47, 488-505. en-US en-US Scheufele, D. A., & Tewksbury, D. (2007). Framing, agenda-setting, and priming: The evolution of three media effects models. Journal of Communication, 57(1), 9-20. en-US en-US Sibley, C. G., Liu, J. H., & Kirkwood, S. (2006). Toward a social representations theory of attitude change: The effect of message framing on general and specific attitude toward equality and entitlement. New Zealand Journal of Psychology 35(1), 3-13. en-US en-US Scott, L. B., & Curbow, B. (2006). The effect of message frames and CVD risk factors on behavioral outcomes. American Journal of Health Behavior 30(6), 582-597. en-US en-US Shah, D. V., Kwak, N., Chmierbach, M., & Zubric, J. (2004). The interplay of news frames on cognitive complexity. Human communication Research, 30 (1), 102-120. en-US

PAGE 117

117 en-US Shen, L., & Dillard, J. P. (2007). The influence of behavioral inhibition/approach systems and message framing on the processing of persuasive health messages. Communication Research 34(4), 433-467. en-US en-US Shiv, B., Britton, J. A. E., & Payne, J. W. (2004). Does elaboration increase or decrease the effectiveness of negatively versus positively framed messages? Journal of Consumer Research, 31 199-208. en-US en-US Smith, G. E. (1996). Framing in advertising and the moderating impact of consumer education. Journal of Advertising Research, 36 (5), 49-64. en-US en-US Smith, S. M., & Petty, R. E. (1996). Message framing and persuasion: A message processing analysis. Personality and Social Psychology Bulletin 22(3), 257-268. en-US en-US Smith, S. M., & Levin, I. P. (1996). Need for cognition and choice framing effects. Journal of Behavioral Decision Making 9, 283-290. en-US en-US Smith G. E., & Wortzel L. H. (1997). Prior knowledge and the effect of suggested frames of reference in advertising. Psychology and Marketing 14 (2), 121-143. en-US en-US Spence, M. R., Elgen, K. K., & Harwell, T. S. (2003). Awareness, prior use, and intent to use emergency contraception among Montana women at the time of pregnancy testing. Maternal and Child Health Journal, 7 (3), 197-203. en-US en-US Tversky, A., & Kahneman, D. (1981). The framing of decisions and the rationality of choice. Science, 221 453-458. en-US en-US Tversky, A., & Kahneman, D. (1986). Rational choice and the framing of decisions. Journal of Business, 59, 251-278. en-US en-US Venkatraman, S., Aloysius, J. A., & Davis, F. D. (2006). Multiple prospect framing and decision behavior: The mediational roles of perceived riskiness and perceived ambiguity Organizational Behavior and Human Decision Processes 101(1), 59-73. en-US en-US Wathen, N. C. (2006). Health Information Seeking in Context: How Women Make Decisions Regarding Hormone Replacement Therapy. Journal of Health Communication, 11 477493. en-US en-US Williams, T., Clarke, V., & Borland, R. (2001). Effects of message framing on breast-cancerrelated beliefs and behaviors: The role of mediating factors. Journal of Applied Social Psychology, 31 (5), 925-950. en-US en-US Wilson, D. K., Kaplan, R. M., & Schneiderman, L. (1987). Framing of decisions and selections of alternatives in health care. Social Behavior 2, 51-59.

PAGE 118

118 en-US Zhang, Y., & Buda, R. (1999). Moderating effects of need for cognition on responses to positively versus negatively framed advertising message. Journal of Advertising 28 (2), 115. en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US en-US

PAGE 119

119 en-US BIOGRAPHICAL SKETCH en-US Kenneth E. Kim was born in Seoul, Korea. He attended Korea University in Seoul, Korea and graduated with a Bachelor of Arts in Journalism & Mass Communication and a Master of Arts in Mass Communication. Before Kenneth pursued his Ph.D. in mass communication at the University of Florida, he attended the communication graduate programs at Cornell University and Florida State University. His research path centers on cognitive effects of a variety of communication tactics with an emphasis on strategic message use and its impact on persuasiveness in pharmaceutical advertising Kenneth has been happily married to Hye Eun for 7 years and they have two lovely daughters: Cayla, age 5 and Kaylee, age 1. Kenneth joined the Department of Mass Communication & Communication Studies at the Towson University in the Fall of 2009 and teaches undergraduate advertising courses.