<%BANNER%>

Issue Advertising

Permanent Link: http://ufdc.ufl.edu/UFE0024176/00001

Material Information

Title: Issue Advertising Effects of Frame Valence and Advertising Message Involvement on Attitudes and Behavior
Physical Description: 1 online resource (85 p.)
Language: english
Creator: Parker, Anna
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: advertising, attitudes, behavior, corporate, framing, involvement, issues, responsibility, social
Journalism and Communications -- Dissertations, Academic -- UF
Genre: Advertising thesis, M.Adv.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Pairing a brand with a social issue has become a key tactic for building brand equity. Such marketing communication strategies seek to move brands from trade marking symbols, to corporate citizens imbued with altruistic qualities that serve as a point of differentiation in the marketplace. This research introduces issue advertising, the paid, mediated expression of companies' corporate social responsibility initiatives, as the experimental stimuli. The research aimed to discover the most effective executional techniques given differences in the communication context. The independent variables under study are message frame valence and Advertising Message Involvement (AMI); the dependent variables are attitudes toward the ad and brand, and behavior intention. Relationships between variables were predicted using prospect theory and the Elaboration Likelihood Model of persuasion as the theoretical framework. Previous research indicated that there is no universal effect of frame valence and AMI level, nor a predictable interaction effect of these variables. In some cases negative information may inspire people to act to avoid a loss, where in other cases positive information may lead to more positive evaluations of the message and therefore higher intent to act. AMI level may enhance the cue effect of message framing, or enhance desire to avoid a loss or achieve a gain. Results of the experiment were limited by the unsuccessful manipulation of AMI level, but analyses discovered a significant effect of message framing on attitudes toward the ad and brand. Exploration of the research questions discovered that subject's level of enduring involvement with the issue had a significant effect on attitudes toward the ad and behavior intention. Suggestions for improvement to the experimental design are presented.
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 Anna Parker.
Thesis: Thesis (M.Adv.)--University of Florida, 2009.
Local: Adviser: Morton, Cynthia R.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-11-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024176:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024176/00001

Material Information

Title: Issue Advertising Effects of Frame Valence and Advertising Message Involvement on Attitudes and Behavior
Physical Description: 1 online resource (85 p.)
Language: english
Creator: Parker, Anna
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: advertising, attitudes, behavior, corporate, framing, involvement, issues, responsibility, social
Journalism and Communications -- Dissertations, Academic -- UF
Genre: Advertising thesis, M.Adv.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Pairing a brand with a social issue has become a key tactic for building brand equity. Such marketing communication strategies seek to move brands from trade marking symbols, to corporate citizens imbued with altruistic qualities that serve as a point of differentiation in the marketplace. This research introduces issue advertising, the paid, mediated expression of companies' corporate social responsibility initiatives, as the experimental stimuli. The research aimed to discover the most effective executional techniques given differences in the communication context. The independent variables under study are message frame valence and Advertising Message Involvement (AMI); the dependent variables are attitudes toward the ad and brand, and behavior intention. Relationships between variables were predicted using prospect theory and the Elaboration Likelihood Model of persuasion as the theoretical framework. Previous research indicated that there is no universal effect of frame valence and AMI level, nor a predictable interaction effect of these variables. In some cases negative information may inspire people to act to avoid a loss, where in other cases positive information may lead to more positive evaluations of the message and therefore higher intent to act. AMI level may enhance the cue effect of message framing, or enhance desire to avoid a loss or achieve a gain. Results of the experiment were limited by the unsuccessful manipulation of AMI level, but analyses discovered a significant effect of message framing on attitudes toward the ad and brand. Exploration of the research questions discovered that subject's level of enduring involvement with the issue had a significant effect on attitudes toward the ad and behavior intention. Suggestions for improvement to the experimental design are presented.
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 Anna Parker.
Thesis: Thesis (M.Adv.)--University of Florida, 2009.
Local: Adviser: Morton, Cynthia R.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-11-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024176:00001


This item has the following downloads:


Full Text

PAGE 1

1 ISSUE ADVERTISING: EFFECTS OF FRAME VALENCE AND ADVERTISING MESSAGE INVOLVEMENT ON ATTITUDES AND BEHAVIOR By ANNA C. PARKER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ADVERTISING UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Anna C. Parker

PAGE 3

3 To my Mom for love, for life

PAGE 4

4 ACKNOWLEDGMENTS My masters thesis research would not have been possible without the incredible resources available at the University of Florida. My committee chair, Dr. Cynthia Morton; and committee members, Dr. Jorge Villegas and Dr. John Sutherland, guided my research and offered valuable insight to inform my exper imental design. I truly benefited from the depth of experience held by my committee and the Advertising Department faculty in the College of Journalism and Communications. The friendships I made with fellow masters students in the M.Adv. program define my experience at the University of Florida. I cannot imagine this chapter in my life without Brian Canning, Amanda Ehrlich, Maritza Garcia, Kris Mehaffey, Beau Powers, April Shapiro, Mari Luz Zapata Ramos and Matthew Meltzer. In many ways, we pursued t his degree as a team. We pushed each other to achieve more than we thought possible, and supported each other so that each of us could give our all. Above all, I thank my family for the space to grow into the person they had already given me credit for be ing.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 7 ABSTRACT .......................................................................................................................................... 8 CHAPTER 1 INTRODUCTION ....................................................................................................................... 10 Importance of Present Research ................................................................................................. 10 Contribution to Issue Advertising Research .............................................................................. 13 2 LITERATURE REVIEW ........................................................................................................... 16 Message Framing ........................................................................................................................ 16 Definitions and Operationalization ..................................................................................... 16 Frame Valence and Persuasion ........................................................................................... 20 Involvement ................................................................................................................................. 23 Definitions and Operationalization ..................................................................................... 23 Advertising Message Involvement ..................................................................................... 26 Involvement, Frame Valence, and Persuasion ................................................................... 27 Theoretical Framework ............................................................................................................... 31 3 METHOD .................................................................................................................................... 33 Design and Pr ocedure ................................................................................................................. 33 Manipulation of Independent Variables .................................................................................... 34 Measurement of Dependent Variables ....................................................................................... 35 Reliability .................................................................................................................................... 36 Validity ........................................................................................................................................ 37 Stimulus Development ................................................................................................................ 38 Control of Confounding Variables ............................................................................................. 41 Manipulation Checks .................................................................................................................. 42 4 RESULTS .................................................................................................................................... 44 Main Experiment ......................................................................................................................... 44 Scale Reliability Analyses ................................................................................................... 45 Attitude toward the ad .................................................................................................. 45 Attitude toward the brand ............................................................................................ 45 Behavior intention ........................................................................................................ 45 Manipulation Checks ........................................................................................................... 46 Hypotheses Tests ................................................................................................................. 47

PAGE 6

6 5 DISCUSSION .............................................................................................................................. 59 Evaluation of Hypotheses ........................................................................................................... 60 Evaluation of Research Question ............................................................................................... 62 Limitations and Future Research ................................................................................................ 63 Managerial Impli cations ............................................................................................................. 65 APPENDIX A EXHIBIT 1: NESTL ISSUE AD ............................................................................................. 67 B EXHIBIT 2: GREENPEACE NEGATIVE FRAME ISSUE AD ............................................ 68 C EXHIBIT 3: GREENPEACE POSITIVE FRAME ISSUE AD .............................................. 69 D INFORMED CONSENT ............................................................................................................ 70 E NEGATIVE FRAME ISSUE AD STIMULUS ........................................................................ 71 F POSITIVE FRAME ISSUE AD STIMULUS ........................................................................... 72 G QUESTIONNAIRE ..................................................................................................................... 73 REFERENCE LIST ............................................................................................................................ 81 BIOGRAPHICAL SKETCH ............................................................................................................. 85

PAGE 7

7 LIST OF TABLES Table page 3 1 Descriptive statistics for pretest of binge drinking consequences ....................................... 42 3 2 Descriptive statistics for pretest of campus organizations ................................................... 43 4 1 Summary of respondent demographics ................................................................................. 51 4 2 Summary of attitude toward the ad scale .............................................................................. 51 4 3 Summary of attitude toward the brand scale ........................................................................ 52 4 4 Behavior intention scale ......................................................................................................... 52 4 5 Enduring involvement scale .................................................................................................. 52 4 6 Advertising message involvement (AMI) scale ................................................................... 53 4 7 Manipulation check of message framing .............................................................................. 53 4 8 Manipu lation check of AMI level ......................................................................................... 53 4 9 Descriptive statistics for hypotheses tests H1, H2 ............................................................... 54 4 10 Main effect of AMI level ....................................................................................................... 54 4 11 Main effect of message framing ............................................................................................ 54 4 12 Descriptive statistics for hypotheses tests H3a, H3b, H4 .................................................... 54 4 13 Interaction effects of message framing and AMI level ........................................................ 55 4 14 Descriptive statistics for RQ1 ................................................................................................ 56 4 15 Main effect of enduring involvement level .......................................................................... 58

PAGE 8

8 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Advertising ISSUE ADVERTISING: EFFECTS OF FRAME VALENCE AND ADVERTISING MESSAGE INVOLVEMENT ON ATTITUDES AND BEHAVIOR By Anna C. Parker May 2009 Chair: Cynthia Morton Major: Advertising Pairing a brand with a social issue has become a key tactic for building brand equity. Such marketing communication strategies seek to move brands from trade marking symbols, to corporate citizens imbued with altruistic qualities that serve as a point of differentiation in the marketplace. This research introduces issu e advertising, the paid, mediated expression of companies corporate social responsibility initiatives, as the experimental stimuli. The research aimed to discover the most effective executional techniques given differences in the communication context. The independent variables under study are message frame valence and Advertising Message Involvement (AMI); the dependent variables are attitudes toward the ad and brand, and behavior intention. Relationships between variables were predicted using prospect theory and the Elaboration Likelihood Model of persuasion as the theoretical framework. Previous research indicated that there is no universal effect of frame valence and AMI level, nor a predictable interaction effect of these variables. In some cases negative information may inspire people to act to avoid a loss, where in other cases positive information may lead to more positive evaluations of the message and therefore higher intent to act. AMI level may enhance the cue effect of message framing, or e nhance desire to avoid a loss or achieve a gain. Results of the experiment were limited by the unsuccessful manipulation of AMI

PAGE 9

9 level, but analyses discovered a significant effect of message framing on attitudes toward the ad and brand. Exploration of th e research questions discovered that subjects level of enduring involvement with the issue had a significant effect on attitudes toward the ad and behavior intention. Suggestions for improvement to the experimental design are presented.

PAGE 10

10 CHAPTER 1 INTRODUCTION Importance of Present Research The incorporation of issue advertising as part of a companys marketing communications has emerged as a powerful trend as more and more companies seek to create a link between their brand image and a social issue (Bl oom, 2007; Cone, Inc., 2004, 2006; Gupta and Pirsch, 2006; Sego, 2002; Tenser, 2006). Issue advertising, the experimental stimuli for this research, is defined as advertising paid -for by a for -profit company linking that company or its brands to a social issue (Sego, 2002). Issue advertising is evidence of a larger trend towards increased corporate social responsibility (CSR), which includes activities ranging from cause related marketing to corporate philanthropy to sponsorship. According to Boston bas ed Cone, Inc., a strategy and communications agency, half of all companies have programs associated with social issues (Cone, Inc., 1999; Sego, 2002). In fact, companies spend hundreds of millions of dollars annually in contributions to social cause programs (Sego, 2002). The reason, it seems, is that it works. Hoeffler and Keller (2002) point out the important role of companies association with causes in building brand equity, and Nan and Heo (2007) note that these brands are generally perceived as be ing generous and altruistic (p.66). According to a published research report by Cone, Inc. (2006), 61% of Millenials (those born between 1979 and 2001) say that making a difference in the world is important to them. Perhaps more importantly, this grou p of 78 million consumers believes that companies need to do the same (Cone, 2006). Data from Cone, Inc.s 2006 Millenial Cause Study, subtitled The Millenial Generation: Pro -Social and Empowered to Change the World indicated the following: 83% will tru st a company more if it is socially/environmentally responsible 74% are more likely to pay attention to a companys message when they see that company has a deep commitment to a cause

PAGE 11

11 69% consider a companys social/environmental commitment when deciding w here to shop 89% are likely or very likely to switch from one brand to another (price and quality being equal) if the second brand is associated with a good cause 66% will consider a companys social/environmental commitment when deciding whether to recomm end its products or services. This trend is not only evidenced in younger generations, such as Millenials. Ross, Stutts, and Pattersons (1992) study on consumer perceptions of organizations that engage in cause related marketing found that half of the 225 adults surveyed could remember an issue ad, and half of the sample said that knowledge of a companys association with an issue prompted them to purchase from that company to support the issue (Sego, 2002). A 2008 Cone, Inc. study specifically noted t he effect of environmental messages from companies, finding that 65% of respondents said, Companies that communicate about the environment make me want to buy products from them (Cone, Inc., 2008, p3). Cone, Inc.s 2004 Corporate Citizenship Study revea led that 80 percent of Americans could name a company that stands out in their mind as a strong corporate citizen. (Cone, Inc., 2004, p2). This is an increase from only 49 percent in 2001. In addition, 72 percent of respondents believe it is acceptable for companies to include cause or issue related messages in their advertising, and 86 percent said they were likely to switch brands based on a brands association with a cause (Cone, Inc., 2004). The prevalence of CSR initiatives has created a vast net work of terms, activities, and ideas associated with the concept. For the purposes of effective research, it is necessary to further delineate issue advertising as different from other CSR activities such as cause related marketing (CRM), and as different from traditional profit -oriented advertising. Issue advertising is perhaps the most tangible evidence of a companys corporate social responsibility initiatives. By definition, advertising is a paid, mediated form of

PAGE 12

12 communications from an identifiable source, designed to persuade the receiver to take some action, now or in the future (Richards and Curran, 2002, p74). In other words, issue advertising must be mediated, e.g. appearing in print, on television or radio airwaves, etc. This is distinctly different from cause related marketing, which could simply be the development of a charitable program, advertised or not. To specify, cause related marketing is defined as a program designed to create a partnership between a sponsoring firm and a nonpr ofit cause to raise money through product sales (Gupta and Pirsch, 2006, p314). Perhaps one of the most recognized recent examples of CRM is Yoplaits Save Lids to Save Life program, in which 10 cents is donated to the Susan G. Komen Breast Cancer Foun dation for every yogurt lid mailed in by customers (Nan and Heo, 2007). Advertisements promoting a companys CRM program often prominently feature a product, while issue advertising is focused almost exclusively on the issue message, though the source is identifiable. Definition of issue advertising as paid placements by a for -profit company also differentiates it from public service announcements (PSAs), and other pro -bono message placements that originate from non -profit organizations or government agenc ies. Gupta and Pirschs (2006) definition of CRM highlights another important difference between issue advertising and CRM. While CRM usually indicates a companys association with a nonprofit organization, issue advertising may or may not include such associations. In fact, there are many examples of issue advertising for social issues that are not represented by a specific non -profit or other organization. Consider Exhibit 1. (Appendix A) as an example of a recent issue advertisement from Nestl, askin g individuals to give to children in need. Just as issue advertising is distinct from other forms of corporate social responsibility, it also differs from traditional advertising for products or services. Sciulli and Bebko (2005) note

PAGE 13

13 that traditional prof itoriented advertising is aimed at persuading an audience to purchase goods or services, while social cause, or issue advertisings goal is the betterment of society (p19). While the main purpose of profit -oriented advertising is financial gain, the hi ghest hope of issue advertising is behavioral change, and certainly increased awareness of the issue (Sciulli and Bebko, 2005). Of course, product marketers also are banking on the fact that advertising issues that achieve a goal of societal betterment will also improve their corporate image and, in turn, their bottom line (Sciulli and Bebko, 2005; Ross, Patterson and Stutts, 1992). Contribution to Issue Advertising Research Discussion of issue advertising and other forms of cause related marketing often turns to whether such initiatives, despite their altruistic overtones, are in fact executed for the sole purpose of making money (Neff, Bush, York and Zmuda, 2008). To quote Neff et al.s 2008 article in Advertising Age, of course it is, and, well, it s hould be (Neff et al., p2, 2008). The authors point is that any brand communication, when successful, should make money. The article goes on to suggest that recognizing a return on investment is particularly crucial for issue related advertising, other wise those advertising dollars may be better spent on conventional advertising (Neff et al., 2008). With that type of pressure it becomes important for research to investigate what executional techniques are most effective in different issue message situ ations. The uniqueness of each situation is formed by characteristics of the message itself, characteristics of the audience, and characteristics of the communications context (Tsai, 2007; Entman, 1993). This study addresses each of these areas as variables, in an attempt to answer the following research questions: What is the most effective way to persuade consumers to engage in an issue related behavior? and How does issue advertising affect consumers attitudes toward the ad, beliefs about the brand and attitude toward the brand?

PAGE 14

14 The present research focuses specifically on the effects of message frame valence (negative or positive) and advertising message involvement (AMI), a term coined by Laczniak and Muehling (1989, 1993a, 1993b) to describe a motivational state inducing message processing (1993b, p303). The conceptual definition of AMI sets useful parameters on the more general concept of involvement, defined by Zimbardo (1960) as an individuals concern for an issue as related to their perso nal values and needs. Here, AMI narrows that definition to a construct that deals specifically with a persons relationship with the advertising message. The concept of message frame valence (whether the message communicates the benefits to be gained by t aking the advocated position versus the consequences of not taking the advocated position) taken together with Advertising Message Involvement (AMI) are the independent variables for examining issue related advertising messages. Research has shown that bot h positive and negative frames are appropriate for, and have strengths in, different situations (Tsai, 2007; Ferguson and Gallagher, 2007; Shiv et al., 2004; Meyers Levy and Maheswaran, 2004; Zhang and Buda, 1999; Levin et al., 1998; Obermiller, 1995; Meye rowitz and Chaiken, 1987). Furthermore, the research suggests that subjects level of involvement is a moderator for what type of appeal will be more effective. Much of the previous research that incorporated variables related to involvement and framing focused on a persuasive message related to health concerns or product purchase, but few studies to date have paired a branded advertising message with an issue, as is becoming common in advertising practice today. This research brands an issue ad with an organization name and logo in order to make inferences not only about the effect of frame valence and AMI on attitudes towards the ad and behavior intention, but also on the resulting attitudes toward the brand.

PAGE 15

15 The theoretical framework for this research utilizes prospect theory for modeling hypothesized relationships between variables related to behavior and the Elaboration Likelihood Model of persuasion to investigate evaluations of the experimental stimulus, or attitude object. The ultimate goal of thi s research is the discovery of a causal relationship between variables that advertising practitioners can reference so as to maximize each dollar spent in the burgeoning domain of issue advertising. Chapter 2 presents a comprehensive review of the literatu re surrounding studies on message framing and involvement, including analysis of definitions and examples of successful operationalization and manipulation. A discussion of the observed relationships between message framing, advertising message involvement, and persuasion is also presented. Chapter 3 describes the experimental method, including descriptions of the experimental design, manipulation of the variables under study, materials and procedure, and manipulation checks. Chapter 4 presents the resear ch findings and Chapter 5 discusses practical and managerial implications of the research findings and suggests recommendations future research.

PAGE 16

16 CHAPTER 2 LITERATURE REVIEW Message Framing Definitions and Operationalization Entman (1993) offers a basic de finition of framing as a concept, referring to it simply as a way to describe the power of a communicating text (p51). Through study of the way a communicating text is framed, we begin to understand how the human consciousness is influenced during infor mation transfer. The broad nature of the concept of framing allows for application across disciplines and incorporation into the investigation of many research questions. Specifically, the effects of framing have been studied in the media (e.g. Chyi and McCombs, 2004; Iorio and Huxman, 1996), in product advertising (e.g. Tsai, 2007; Shiv et al., 2004; Levin, Gaeth, Schreiber and Lauriola, 2002, 1998; Zhang and Buda, 1999; Martin and Marshall 1999), and in health communications (e.g. Ferguson and Gallaghe r, 2007; Levin et al. 2002, 1998, Maheswaran and Meyers Levy, 1990; Meyerowitz and Chaiken, 1987). In addition, Obermiller (1995) and Davis (1995) investigated the effects of framing on communication of environmental issues including water and energy cons ervation, recycling, and solid waste reduction. The bulk of framing research operationalizes the concept in terms of positive and negative frames (e.g. Ferguson and Gallagher, 2007; Tsai, 2007; Shiv et al., 1997, 2004; Levin et al., 2002, 1998; Zhang and B uda, 1999; Obermiller, 1995; Maheswaran and Meyers Levy, 1990; Meyerowitz and Chaiken, 1987) including this study. In contrast to the use of weak arguments in product advertising, where brand image defines the selling proposition more so than any concrete attribute of the product, it is both common and appropriate for social issues to be presented in both a positive and negative frame due to the inherent valence of a social issue. For example,

PAGE 17

17 Exhibit 2. in the appendices presents an ad by Greenpeace stres sing a negative consequence of not addressing global warming. The ad reads, Winter. Youll miss it when its gone. In contrast, a separate ad by Greenpeace, Exhibit 3. in the appendices, presents the positive outcomes of conservation, the advocated beh avior. The ad reads Were not only saving money with smaller ads, were also saving a few trees. Operationalization of the concept of framing also carries multiple definitions of positive and negative and various methods for differentiation. Zhang and Buda (1999) define a positively framed persuasive message as that which emphasizes a brands advantages or the potential gains to consumers resulting from the purchase or use of the brand (p1) and a negatively framed persuasive message as that which accentuates the potential losses to consumers if the brand is not chosen or a wrong decision is made in choosing a brand (p1). In Shiv et al.s (2004) research, positively framed messages stressed the benefits of following some recommendation while negatively framed messages stressed the consequences of not following the recommendation (p204). Meyerowitz and Chaiken (1987) operationalized framing similarly, with negative frames stressing the consequences of not performing breast self -examinations an d positive frames stressing the benefits of the recommendation. Chatterjee, Heath, Milberg and France (2000) refer to a gain (positive) and loss (negative) frame in their research, using the example of a cancer treatment survival rate where a 90% survival rate is the gain, or positive frame and a 10% death rate is the loss, or negative frame. Tsai (2007) also defined message frames in terms of gain and loss, adding that the positive frame (focusing on advantages) seeks to maximize happiness while the negative frame (focusing on losses) seeks to minimize pain (p364). Obermillers (1995) study of

PAGE 18

18 environmental communication employed the term sick baby and well baby to describe positive and negative appeals. According to Obermiller (1995), sick baby appeals present the problem as severe, threatening, or otherwise important (p56) whereas well baby appeals stress the significance of individual action (p55). Levin et al., 2002 and Levin, Schneider, and Gaeth (1998) identified three main framing m ethods and corresponding effects that help explain the differences in the positive and negative frame definitions presented: attribute framing, goal framing, and risky choice framing. According to Levin et al. (2002) attribute framing involves selection of a key object attribute and presenting it in either positive or negative terms. A common example is the presentation of the lean verses fat percentage of a food product (Tsai, 2007; Levin et al. 2002, 1998). The attribute, fat, could be presented as 85% lean in the positive frame condition, or 15% fat in the negative frame condition. According to Levin et al. (2002) attribute framing should affect evaluations favorably when the attribute is framed positively. Goal framing describes a persuasive m essage that stresses the benefits of performing an act related to achievement of a goal in the positive condition, and the negative consequences of not performing the advocated behavior in the negative condition. A persuasive message employing this method is expected to have different effects on evaluation based on frame valence, though Levin et al. (1998) noted less support for this effect than for other framing methods (Levin et al. 2002). Examples of the goal framing method are common in research surr ounding issue advocacy, such as health (e.g. Ferguson and Gallagher, 2007; Levin et al. 2002, 1998, Maheswaran and Meyers Levy, 1990; Meyerowitz and Chaiken, 1987) and environmental (e.g. Davis, 1995; Obermiller, 1995). For example, Maheswaran and Meyers Levy (1990) presented subjects with messages that either stated the benefits of taking a

PAGE 19

19 diagnostic blood test as a preventative measure against heart disease and high cholesterol (positive frame) or the benefits lost by not taking the diagnostic blood tes t, and therefore not addressing ones cholesterol level (negative frame). Risky choice framing involves presenting subjects with a choice involving risk, where the potential outcomes are framed either positively or negatively, for example in terms of suc cess rate or failure rate (Levin et al. 2002). Choices involving saving lives are common manipulations of risky choice framing. According to Tversky and Kahneman (1981), predictions about human behavior are usually based on the assumption that humans act rationally when presented with a choice. But research has shown that humans perception of what is rational depends on the presentation context, or frame, of the decision problem. In fact, people often exhibit opposite attitudes toward risk (either bein g risk averse or risk taking) depending on whether the information in presented in a gain or loss frame, even though the choices are factually equivalent. Tversky and Kahneman (1981) presented 152 subjects with a decision problem, asking them to imagine t hat a mysterious foreign disease was going to kill 600 people. The researchers presented them with two sets of plans to address the problem with opposite (positive and negative) frame valences. The positive frame described Program A as saving 200 people, and Program B as a 1/3 probability that all 600 people would be saved, and a 2/3 probability that no one would be saved. In this study, the researchers found that subjects tended to exhibit risk averse behavior and choose Program A, saving 200 lives. The negative frame described Program C, in which 400 people would die and Program D, in which there is a 1/3 chance that no one would die, and a 2/3 chance that all 600 people would die. In this scenario, subjects exhibited risk taking behavior, with most opting to adopt Program D in the hopes that no people would die. On further consideration, Programs A and C are identical, as are B and D.

PAGE 20

20 But the framing of the choices influences most subjects to make opposite choices under opposing frame conditions (Tversky and Kahneman, 1981). Comparison of framing methods by Levin et al. (1998) found that frame valence effectiveness depend on whether framing was based on attributes or goals. Levin et al. (1998) found evidence for superiority of a positive frame when the framing method is based on attributes, and superiority of a negative frame when the framing method is based on goals. Clearly there is no universal effect of frame valence. As Tsai (2007) notes, the persuasive function of positive versus negative fr aming very much depends on what message is communicated, to whom, and under which conditions (p366) (See also Shiv et al., 2004). However, application of several theoretical models defined in the framing literature helps explain some of the observed cons istencies of framing effects. Frame Valence and Persuasion Framing as a process requires selection of specific aspects of a persuasive communication, so as to make those aspects more significant in influencing the audience in the desired direction. This may include promotion of a specific problem definition, causal interpretation, moral evaluation, and/or treatment recommendation (Entman, 1993, p52; Chyi and McCombs, 2004). In this way the message frame becomes a determinant of how people attend to and interpret the highlighted information, which in turn influences how they evaluate it and act on it (Entman, 1993). But as Tsai (2007) notes, message framing itself is simply an executional technique, a form of advertising message construction which in itself does not serve a persuasiveness enhancer (p366). The interaction of frame valence and other variables related to the communicator, the message, the audience, and the cultural context create the persuasive function of framing (Entman, 1993).

PAGE 21

21 Som e researchers and advertising practitioners support a simple cue effect as an explanation for understanding the effects of frame valence. A cue, or priming, effect suggests that because a message is framed positively, it will engender more positive evalua tion of the attitude object, thereby enhancing persuasion. Conversely, a negatively framed message may arouse negative associations and lead to negative evaluation of the attitude object, thereby decreasing persuasion (Shiv et al., 1997, 2004; Zhang and Bu da, 1999; Smith and Petty, 1996; Maheswaran and Myers -Levy, 1990). Still, much research supports the effectiveness of negative framing in some contexts (Shiv et al., 1997, 2004; Martin and Marshall, 1999; Smith and Petty, 1996; Obermiller, 1995; Maheswara n and Meyers -Levy, 1990; Meyerowitz and Chaiken, 1987). The theory of negativity bias may help explain this. According to Smith and Petty (1996), negatively framed information may be more heavily weighted when incorporated into an audiences evaluations simply because of its novelty. As Meyerowitz and Chaiken (1987) note, negative information becomes particularly salient because most people generally perceive the world around them to be positive. Prospect theory is an important theoretical model used b y researchers to discuss the effects of frame valence on behavior specifically. Prospect theory is commonly applied in discussion of the risky choice and goal framing methods (Levin et al, 1998), where frame valence is described in terms of gains and loss es. Smith and Petty (1996) suggest that prospect theory is a modification of the expected utility model, which equates the utility of a risky choice to the expected utility of the outcome of that choice. According to prospect theory, peoples responses t o losses are typically more extreme than responses to gains. In other words, there is more displeasure associated with losing something than pleasure associated with gaining something of the same value. Therefore people are often shown to be risk averse in choices

PAGE 22

22 involving gains, and risk seeking in choices involving loss (Ferguson and Gallagher, 2007; Zhang and Buda, 1999; Levin et al. 1998; Smith and Petty, 1996, p454; Maheswaran and Meyers Levy, 1990; Meyerowitz and Chaiken, 1987). Meyerowitz and C haiken (1987) illustrated this phenomenon in their research of frame valences effect on womens intention to perform breast self -examinations and their attitude towards the behavior. The choice to do breast self -exams was seen as a risk, because of the p ossibility of discovering a cancerous lump. Negative framing stressed the consequences of not performing the exam and was more effective than positively framed messages that stressed the benefits of the exam. In this way subjects were less risk averse in order to avoid a loss, than to achieve a gain. As Levin et al. (1998) note, when people realize what they stand to lose by not acting, they are more motivated to act (See also Ferguson and Gallagher, 2007, p668). Zhang and Buda (1999) assert that adve rtising execution variables rarely have a universal effect, as evidenced by the inconsistent effect of frame valence in the research discussed. Perhaps more important than understanding the effect of frame valence is the identification and study of other variables that enhance or assuage the effect. The construct of involvement and similar constructs related to processing level have been shown to play a significant role in moderating the effects of frame valence and other variables in many of the studies r eviewed. In fact, consideration of a subjects level of involvement also helps explain why positively framed messages may have a persuasive advantage in some scenarios, while negatively framed messages are more effective in others. Similar to the concept of framing, involvement has been defined and operationalized in different ways and warrants a discussion of the dominant themes in the literature, particularly those with bearing on the present research.

PAGE 23

23 Involvement Definitions and Operationalization The s tudy of involvement and its impact on attitude change became popular in the mid 1900s as researchers addressed the relationship between attitude change and the degree to which a persuasive communication is related to the person to be influenced (Johnson and Eagly, 1989, Zimbardo 1960). As Tsai (2007) and Shiv et al. (2004) point out, research must account for fundamentally determining characteristics of an audience, such as involvement level, when analyzing the effects of frame valence (Tsai, 2007, p364). Zhang and Buda (1999) reiterate a basic definition of involvement as an internal state of arousal with intensity, direction and persistence properties (Zhang and Buda, 1999, p2; Andrews and Durvasula, 1991). In this way involvement is distinguished from the concept of arousal which lacks the intensity and direction components (Andrews and Durvasula, 1991). Some research defines involvement simply as the level of personal relevance of the attitude object (Petty and Cacioppo, 1979, Petty et al., 1983, P etty and Cacioppo, 1990). Zimbardo (1960) describes involvement as the importance of the cognitive elements of the situation (p87). Johnson and Eagly (1989) sought to bring order to the myriad of involvement definitions with a classification system, similar to the way Levin et al. (1998) developed a typology of framing methods. Because involvement is an important cause variable in this research, a brief overview of the conceptual definitions and classification are presented. A meta analysis of involvem ent studies conducted in the 1970s and 1980s resulted in the possible classification of the concept of involvement as belonging to one of three distinct categories: valuerelevant involvement, outcome relevant involvement, and impression relevant involve ment (Johnson and Eagly, 1989).

PAGE 24

24 Value relevant involvement refers to a messages connection to attitudes that are linked to important values, or aspects of the self that are especially important and enduring (Johnson and Eagly, 1989, p290). The value relevant concept of involvement can be described using the sub -concept of latitudes of acceptance, rejection and noncomitment (Johnson and Eagly, 1989, Zimbardo 1960). The latitude of acceptance is comprised of a persons own position and others that he or she finds acceptable (Johnson and Eagly, 1989, p291). The latitude of rejection is comprised of positions that a person does not accept, and the latitude of noncomitment is comprised of positions that are neither acceptable nor unacceptable (Johnso n and Eagly, 1989, p291). Within this theoretical tradition value relevant involvement is operationalized by classifying subjects as either high or low -involvement depending on the widths of their latitudes. Impression relevant involvement (also referre d to as response involvement) is defined as, the individuals concern with the consequences of his response or with the instrumental meaning of his opinion (Zimbardo, 1960, p87). In this research tradition impression or response involvement is typical ly manipulated through varying the consequences of the subjects reported attitudes. For example, telling subjects that their attitude or evaluation of an issue would reveal a lot, or a little, about themselves. In later studies involvement was also man ipulated by telling subjects either that they would or would not be interviewed after reading the message, which therefore encouraged them to commit more/less resources to processing (Johnson and Eagly, 1989). Finally, outcome relevant involvement refers t o the relevance of an issue to currently important goals or outcomes (Johnson and Eagly, 1989, p292). To illustrate, outcome relevant involvement could be manipulated by using a message that advocates policy change at ones own

PAGE 25

25 university in the near future, versus advocating a policy change at a distant university that will not occur for many years (Johnson and Eagly, 1989). Presumably in the latter scenarios, outcome relevant involvement would be lower than in the former, due to the distal and tempora l proximity of the goal. While this classification system is helpful in understanding the various roots of, and manipulations of involvement, other researchers assert that it over -classifies the concept and argue that value relevant involvement and outcome relevant involvement are in fact overlapping concepts (Petty and Cacioppo, 1990). Petty and Cacioppo (1979) prefer the terms issue involvement and response involvement (See also Zimbardo, 1960). Issue involvement is defined as the extent to which t he attitudinal issue under consideration is of personal importance (Petty and Cacioppo, 1979, p1915). This view holds that it is not important whether the relevant concept is related to a value, a goal, a person, or an object. Instead, the critical asp ect of issue involvement is that the topic of the message is perceived as important to the self (Petty and Cacioppo, 1990, p368). In other words, issue involvement assumes that a message related to my values, goals, or attitudes is more highly involvin g that messages related to non personal values, goals, or attitudes. In contrast, the focus of response involvement is on the immediate situational reward (Petty and Cacioppo, 1979, p 1916). Zimbardo (1960) notes that response involvement relates the co ncept to the subjects ability to achieve a goal. In other words, the subject is concerned with the consequences of his or her response (p 87). The range of definitions for involvement incurs a variety of methods for manipulation in an experimental cont ext, from memorization of ad message elements to influencing subjects expectations that a product will or will not be available to them. Distraction methods that inhibit subjects opportunity to process a message, such as memorization of a short or long digit

PAGE 26

26 number, has also been employed as a manipulation of involvement (Shiv et al., 2004; Andrews and Durvasula, 1991). In general, the objective of the techniques used to manipulate involvement to isolate a variable known to influence advertising process ing, so that theoretical relationships between variables can be examined (Laczniak and Muehling, 1993). The prominent use of involvement -driven theoretical frameworks highlights the importance of successful manipulation (Andrews and Durvasula, 1991, p194). Advertising Message Involvement Laczniak and Muehling (1989, 1993a, 1993b) also recognized the need to further narrow the conceptual focus of involvement and suggested classification of involvement within a specific domain, such as product class invol vement, or advertising message involvement (AMI). Here we begin to narrow the definition of involvement from encompassing a persons relationship with the communication context overall, to a definition focused on the subjects relationship with the messag e itself. Laczniak and Muehling (1993b) define AMI as a motivational state induced during message processing (p 303). Therefore subjects in the high involvement, or high AMI, condition are expected to pay closer attention to message claims, concentrate more on the examining and interpreting the claims, and in general allocate more mental effort to these activities than their counterparts in the low AMI condition (See also Laczniak and Muehling, 1989, Table 1, p 30 for examples of published definitions of message involvement). Laczniak and Muehling (1993a) suggest three methods of operationalizing AMI: 1) an attention construct dealing with the direction and intensity of subjects focal attention, 2) a personal or situational construct dealing with the me ssage and personal relevance, and 3) an elaborative process construct based on subjects cognitive elaboration (p60). The attention construct is manipulated by directing subjects to pay attention to specific ad elements, thereby

PAGE 27

27 limiting the attention paid to other elements. Manipulation of the attention construct is also possible by directing attention at other stimuli while the subjects view the experimental stimuli. Manipulation of a messages personal and/or situational relevance can be done either b y telling subjects that they will be able to choose a product after viewing the ads or by asking them to role play in a way that encourages subjects to process the message as if it were personally important to them. When manipulating an advertising mess age in terms of levels of processing or elaboration, researchers often ask subjects to memorize advertising content with the knowledge that they will be tested (high involvement) or to simply evaluate the ad and not expect any tests of retention (low invol vement) (Laczniak and Muheling, 1993a, p6061). Because the present research is concerned with ad attitude formation, as well as brand attitude formation and behavior intention as a result of viewing an issue advertisement, AMI serves as the preferred l ens used to examine the more general concept of involvement throughout subsequent discussion. Involvement, Frame Valence, and Persuasion Involvement is seen as a key variable in the study of how attitudes are formed and how resulting behavior is influen ced (Kokkinaki and Lunt, 1999). Several of the theoretical models discussed have been applied to studies of frame valence and involvement. It is interesting to note that differences in the application of theory, and differences in the operationalization o f frame valence and involvement lead to different findings regarding the effectiveness of frame valence on persuasion. Meyerowitz and Chaiken (1987) based their prediction for the performance of frame valence on prospect theory, predicting that a pamphlet promoting compliance with the behavior of breast self -examination would be more effective if it contained persuasive arguments stressing the negative consequences of nonadherence rather than arguments stressing the positive

PAGE 28

28 consequences of performing the examination. Underlying this hypothesis is the assumption that performing a breast self -exam is a risk -seeking behavior, whereas not performing the exam is a risk averse choice. In this way, when arguments stressing the importance of breast self -exams ar e framed to emphasize the positive consequences of adherence, women may encode these arguments as relative gains. Because risk aversion is expected to dominate in this situation, women should presumably adopt the risk averse option of avoiding doing the e xams. But exposure to arguments that emphasize the potential losses inherent in nonadherence may shift the reference point from one of relative optimism regarding health status to one of some doubt. In this loss domain they may engage in doing the breas t self -exams to alleviate doubts about their health (Meyerowitz and Chaiken, 1987, p501). Levin et al. (1998) support Meyerowitz and Chaikens (1987) results, noting that when framing is based on goals involving gains and losses, there is an advantage for negatively framed messages because people are motivated to avoid loss. Tsai (2007) also notes that in a situation of high personal relevance concentrate more on trying to minimize losses (i.e. being risk averse for losses), whereas in the situation of low personal relevance, subjects are more susceptible to maximizing gains (i.e. being risk -seeking for gains). This application of prospect theory supported Tsais (2007) prediction that high involvement consumers would respond more favorably to negatively framed messages, while low involvement consumers would respond more favorably to positively framed messages. Shiv et al. (1997, 2004) noted important differences in the effectiveness of message framing when involvement is manipulated in terms of opportuni ty, or motivation. Opportunityrelated manipulation often involves decreasing the subjects ability to allot cognitive resources (i.e. engage in cognitive elaboration) to the message, for example by having them remember a long digit number while reading t he ad. Motivation -related manipulation usually involves

PAGE 29

29 increasing the relevance of the message, for example by telling subjects that a product will soon be available in their market, or that they will have to make a decision regarding the product. When Shiv et al. (1997, 2004) apply this theory of elaboration, they find that negative framing is more effective when cognitive elaboration is high, and when involvement is manipulated using motivation -related factors, and that positive framing is more effecti ve when using opportunityrelated factors. Obermiller (1995) posits that concern for the issue is important for an advertising message to be effective. In a situation issue salience is low, the message must increase concern. But when issue salience is hi gh, the message must avoid making the subject feel helpless, and instead increase the belief that adherence to the advocated behavior will help address the issue. Laczniak and Muehling (1993b) also recognize that subjects pre -existing motivational state, or involvement with the domain under study (such as a product class) has bearing on subjects message processing and subsequent formation of attitudes. In addition, Levin et al (2002) noted that differences in their study of red meat consumption and Meyer owitz and Chaikens (1987) breast self examination study may be attributable to differences in the prior salience of each issue. Maheswaran and Meyers Levy (1990) used the Elaboration Likelihood Model of persuasion to predict that people in the low involvement condition are unlikely to pay close attention to message arguments, or integrate their beliefs into an overall attitude. Instead, people in the low involvement condition use peripheral cues in the message context, such as whether the message is asso ciated with positive or negative cues. In this way frame valence would vary directly with evaluations, meaning a negative frame would engender negative evaluations and a positive frame would engender positive evaluations for low involvement subjects.

PAGE 30

30 Pett y and Cacioppos (1986) ELM suggests that brand attitudes (Attb) are formed either via central or peripheral route processing, and that level of involvement determines the route, or level of elaboration (See also Laczniak and Muehling, 1993b, p302). The E LM asserts that when subjects form attitudes after viewing an advertising message under high involvement conditions, the stimulus has a stronger impact than for low involvement subjects, because subjects employ central route processing, as opposed to peripheral route processing. The cognitive response model is an important precursor to development of the Elaboration Likelihood Model of persuasion. The ELM considers a variety of factors that influence peoples motivation and ability (components of involveme nt) to consider a persuasive message and form judgments. The model further suggests that there are two routes to persuasion, central route involving high elaboration, or high involvement, and peripheral route involving low elaboration, or low involv ement (Petty and Cacioppo, 1984). According to Petty et al. (1983), peripheral route processing is characterized by reliance on heuristic cues, rather than careful consideration of the pros and cons presented in an advertising message. The attitudes form ed via peripheral route processing have been shown to be less strongly held and less predictive of behavior. Central route processing is characterized by careful consideration of the information contained in the advertising message, and therefore the resulting attitudes are thought to be more enduring, and a better predictor of future behavior. Petty et al. (1983) therefore suggest that because low involvement incurs peripheral route processing where peoples attitudes are influenced by executional cues, message framing will have a stronger effect. In other words, people are less likely to expend cognitive resources to process a message and instead employ superficial analyses, resulting on an evaluation of the attitude object that is not related to the object itself but peripheral cues (Petty and Cacioppo,

PAGE 31

31 1984, p673). Conversely, under high involvement and central route processing people will allot more cognitive resources to message processing, resulting in enduring attitudes and a link between the a ttitude and the attitude object (such as the advertised issue). In this case peoples attitudes are influenced by message content, so message framing will have less of an effect. Theoretical Framework Prospect theory was used to predict relationships b etween frame valence (positive or negative) and low and high AMI levels on resulting behavior intention, and the Elaboration Likelihood Model of persuasion (ELM) was used to predict relationships between frame valence (positive and negative), and low and h igh AMI levels and the resulting evaluations of the ad and brand. The combination of these theoretical models allow more specific predictions to be made about the relationship between independent and dependent variables. Specifically, prospect theory is u tilized to model hypotheses related to a consumer choice; the behavior intention dependent measure, and the Elaboration Likelihood Model of persuasion is used to model hypotheses related to evaluations of the attitude object (experimental stimuli) and resu lting attitudes toward the ad and brand. This theoretical framework aided in evaluating the interactions between frame valence and AMI, as well as predicting the effectiveness of frame valence for subjects in high and low involvement. Based on the review of the literature and the proposed theoretical framework, the following questions and hypotheses are advanced to guide this research investigation. H1 : High advertising message involvement (AMI) will result in a) more favorable attitudes toward the issue a d and b) more favorable attitudes toward the brand than low advertising message involvement (AMI). H2 : Positively framed issue advertising messages will result in a) more favorable attitudes toward the ad and b) more favorable attitudes toward the brand t han negatively framed advertising messages.

PAGE 32

32 H3a : Negatively framed issue advertising messages will be more effective in influencing intent to behave for subjects in the high involvement condition. H3b : Positively framed issue advertising messages will be more effective in influencing intent to behave for subjects in the low involvement condition. H4 : Overall, message framing will have a stronger effect on influencing intent to behave for subjects in the low advertising message involvement condition versus the high advertising message involvement condition. RQ1 : What are the effects of subjects self reported level of enduring involvement with the issue on attitude toward the ad, attitude toward the brand, and behavioral intention?

PAGE 33

33 CHAPTER 3 METHOD Design and Procedure A 2 (Frame valence; negative/positive) x 2 (AMI level; high/low) post test only experimental design was employed to test the relationships between independent and dependent variables. The independent variables for the investigation are the a dvertising message frame valence (positive versus negative frame) and advertising message involvement level (high versus low involvement). The dependent variables are attitude toward the ad (Attad), attitude toward the brand (Attb), and behavioral intenti on. The stimulus portion of the experiment is composed of reading instructions that place subject in the high or low AMI condition, and an ad with copy that places subjects in either the negative or positive frame condition. The negatively and positively framed print ad stimuli both contained the same headline and call to action. The headline read Think Before You Drink. and the call to action read: Please visit www.health.ufl.edu for more reasons to Think Before You Drink. The imagery and brand are also the same for both issue ads; the there was a simple white to gray gradient background, and an image of three hands, each holding a beer and toasting each other. The organization, or brand in the ad, appeared across the bottom with the UF | University of Florida logo in the lower left and the Health Science Center name to the right. The body copy differentiated the negative and positively framed issue ads from each other. The negatively framed issue ad read as follows Raise your glass and cheers the consequences of binge drinking: Tasting your dinner twice Losing control of when and where you pass out Waking up with a hangover Consuming more than 5 drinks in a night can make you look bad and feel even worse.

PAGE 34

34 The positively framed issue ad read as follows Raise your glass and cheers the benefits of drinking in moderation: Only tasting your dinner once Staying in control of when and where you go to bed Waking up without a hangover Consuming less than 5 drinks in a night can make you lo ok good and feel even better. The experiment lasted 15 20 minutes. Subjects returned their completed questionnaires to the moderator of the experiment. Manipulation of Independent Variables Two independent variables were manipulated through the sti mulus: Advertising Message Involvement (AMI): Low and high Subject were primed for a low or high advertising message involvement condition through the reading instructions in the questionnaire. The high involvement reading instructions stated the follow ing as the study background: Binge drinking is defined as having more than 5 drinks in one night. The student government at the University of Florida is launching a campaign that focuses on the issue of binge drinking during the Spring 2009 semester. St udents will soon be asked to vote on campaign messages to determine which will be implemented. Please take your time reviewing the following proposed campaign advertisement. Your evaluation of the message is very valuable in helping the University of Fl orida develop an effective campaign. The low involvement reading instructions stated the following as the study background: Binge drinking is defined as having more than 5 drinks in one night. The state of Florida is considering development of a campaign that focuses on the issue of binge drinking. The statewide campaign would be launched on college campuses sometime in 2011. Please review the following proposed campaign advertisement Frame valence: Positive and negative The study population viewed an is sue ad executed with either a positive or negative frame. The positive frame stressed the benefits of attending to the issue presented in the ad, and asked subjects to visit a Web site related to the

PAGE 35

35 advertising message as a measure of behavioral intentio n. The negative frame stressed the consequences of not attending to the issue presented in the ad, and also asked subjects to visit a Web site related to the advertising message as a measure of behavioral intention. The information provided in the positi vely and negatively framed ads was identical, with changes only to the phrasing of each statement to present it as a gain, or loss. Enduring issue involvement. Subjects also rated their personal level of enduring involvement with the issue. This self assessment was used to categorize subjects as high, medium, or low issue involvement and this variable was analyzed post hoc, to look for any moderating effects. Measurement of Dependent Variables Three dependent variables were measured: Attitude towards the ad : This variable describes the subjects overall evaluations of the ad in terms of favorability and liking (Hoyer and MacInnis, 2007) and was measured using three 7 point semantic differential scales rating the ad as good/bad, favorable/unfavorable, and p leasant/unpleasant. Attitude towards the brand: This variable describes subjects evaluation of the brand, represented by inclusion of a logo with a brand name on the stimulus ad. Attitudes toward the brand were measured using three 7 point semantic diff erential items rating the brand as good/bad, favorable/unfavorable, and satisfactory/unsatisfactory. Goldsmith, Lafferty and Newell (2001) employed this scale to specifically measure attitude towards a company by asking respondents to indicate their respon se the following statement on each of the three dimensions: My overall impression of the University of Florida Health Science Center is (good/bad, favorable/unfavorable, and satisfactory/unsatisfactory).

PAGE 36

36 Behavior intention: This variable describes subjects likelihood of adhering to the advocated behavior presented in the issue ad (visiting a Web site to learn more about the issue). Behavioral intention was measured using three 7 point semantic differential items anchored by very unlikely/very likely. Sub jects were asked to indicate how likely (since viewing the ad) they are to visit the Web site soon, to visit the Web site in the future, and to what degree the ad made them more or less likely to visit the site in the future (Not at all likely/Very likely) (Maheswaran and Meyers -Levy, 1990). Measurements of the dependent variables was compared and analyzed to test the hypotheses, explore the research question and form conclusions. Reliability Reliability refers to the repeatability of the experiment. I n a highly reliable experiment, the experiment would yield the same data or results each time (Babbie, 2007). Use of published scales as the instrument and pretesting of the stimulus helped ensure reliability. The enduring involvement scale is a 10 item m easure adapted from Zaichowskys (1985) 22item Personal Involvement Inventory. This abbreviated version of the scale was employed by Zaichowsky in 1994 with reported stability scores between .73 and .84. The scale used to test the AMI manipulation was e mployed by Ahluwalia, Unnava, and Burnkrant (2001), who reported an alpha of .85 for the scale. The scale used to test the framing manipulation was successfully employed by Maheswaran and Meyers Levy (1990) and Shiv, Britton, and Payne (2004). The scale measuring attitude towards the ad has a reported Chronbachs alpha of.93, and the scale measuring attitude towards the brand has a reported Chronbachs alpha of .91 (Tsai, 2007). The scale measuring behavioral intention was successfully employed and publi shed by Maheswaran and Meyers Levy (1990). The questions on the questionnaire are not detailed or personal in nature, and therefore should encourage reliable responses.

PAGE 37

37 Validity Validity refers to the extent to which the chosen measurement is able to mea sure the concepts to which it is applied (Babbie, 2007). Careful creation of the questionnaire and use of accepted, reliable scales to measure concepts helped ensure the validity of this experiments findings. Goldsmith, Lafferty and Newell (2001) employe d the scale measuring brand attitudes and stated that the scales items along with those measuring other constructs were examined via principle axis factor analysis with oblique rotation. All items were described as loading as expected which provides evidence for convergent and discriminant validity (Bruner, Hensel and James, 19982001). Ahluwalia, Unnava, and Burnkrant (2001) used the advertising message involvement scale and found that those who scored higher in the involvement scale also scored higher on a recognition memory index, measuring accurate response to statements made about the ads. Babbie (2007) noted four types of validity: face validity, criterionrelated validity, construct validity and content validity. The use of accepted and publish ed scales for measurement of the dependent variables should ensure face validity. Criterion validity refers to the ability of an external test to confirm the findings. Construct validity refers to whether related variables exhibit similar responses to the stimulus (Babbie, 2007). For example, attitude toward the ad and attitude toward the brand are expected to vary directly, because favorable evaluations of an ad often lead to favorable evaluations of the brand (Hoyer and MacInnis, 2007). Finally, conte nt validity refers to whether the instrument measures the range of meanings represented by a concept (Babbie, 2007). Again, the use of tested scales for measurement of the dependent variables should ensure content validity. Experimental designs must also give consideration to both external and internal validity. External validity refers to the ability to generalize findings to the broader population (Maxwell

PAGE 38

38 and Delaney, 1999). This experiment uses a convenience sample of college students, and therefore caution must be used when generalizing findings to the more diverse adult population. The artificiality of the experimental setting could also lead to external invalidity, since subjects may not respond to ads the same way they would in a normal setting. Internal validity is the absence of alternative explanations for the results of the study (Babbie, 2007). Eight main sources of internal invalidity will be discussed briefly: history, maturation, testing/retesting, instrumentation, statistical regression selection biases, experimental mortality and demoralized control group subjects (Babbie, 2007; Maxwell and Delaney, 1999). Because the experiment took place at one time, and took less than 15 minutes, history and maturation should not have affected the study. In addition, there was no opportunity for changes in instrumentation. The testing environment could alter responses, especially if subjects know what the experiment is designed to study. The introduction to the questionnaire helped control for this by telling subjects that the researcher is simply interested in their responses to the ads, without naming the independent variables specifically. Statistical regression, or extreme scores, are in part controlled for by employing muti item measures and having a large enough sample size for each experimental condition. Random assignment is the best way to avoid selection biases, and this experiment was double -blind; the researcher was unaware of which experimental group subjects were in until the data w as analyzed. Experimental mortality could have existed if respondents choose not to participate, did not sign the informed consent form, or did not complete the questionnaire. There were no incomplete questionnaires or unsigned informed consent forms ret urned during the experiment. Stimulus Development A total of 105 subjects participated in three pretests conducted prior to the experiment. The pretests sought to evaluate specific components of the investigation, including the

PAGE 39

39 manipulation of the indepe ndent variables, the research stimuli, and the measurement instrument. The issue of binge drinking was chosen as the experimental stimuli due to its relevance within the sample population, and the existence of binge drinking awareness campaigns on which t o model the experimental stimuli. The first pretest (N=16) helped develop the copy points for the experimental stimuli by gauging the credibility, truth and believability of ten consequences of binge drinking. The set of consequences was culled from a re view of binge drinking awareness ads from college campuses in the U.S., including campaigns developed at the University of Florida. The goal was to determine consequences of binge drinking perceived most significant. Ten binge drinking consequences were cu lled from an informal review of binge drinking awareness ads targeted to college students, including ads from the University of Connecticuts Remember Last Night campaign, the Youth Leadership Institutes After Too Many campaign, and ads designed by st udents at the University of Florida for the Universitys GatorWell Health Promotion Services. Participants were asked to rate the ten consequences presented in Table 3 1 using (3) 7 point scale items with endpoints Not at all credible/Very credible, No t true at all/Absolutely true, Not at all believable/Highly believable. The Chronbachs alpha for this scale was .77 .87 so the items were summed to create a scale. Having a hangover (M=5.97, SD=.79), Throwing up (M=5.34, SD=.85), and Passing out (M= 5.20, SD=1.20) were found to be the most credible/true/believable. The first pretest also tested three University of Florida organizations in terms of familiarity and credibility, in order to select one as the brand featured in the experimental stimuli. T hree campus organizations (GatorWell Health Promotion Services, University of Florida Health Science Center, Shands Health Care Center) were tested for familiarity and

PAGE 40

40 credibility. A relatively unfamiliar brand was desired, to control for existing brand a ttitudes. A strong credibility rating was also desired. The University of Florida Health Science Center was found to be least familiar (M=2.19, SD=1.18) and was also ranked the highest regarding credibility. This organization was therefore chosen as the brand in the stimulus. The first pretest also included a measure of enduring involvement with the binge drinking issue, to ensure the issue was not overly polarizing, in which case existing attitudes could mute any effect of the independent variables. Za ichowskys 10item enduring involvement scale (Cronbachs .73 .84), adapted from the 22item Personal Involvement Inventory was employed (Bruner and Hensel, 19982001). An acceptable mean was found (M=4.03, SD=.99), and scores varied from 2.20 to 5.50 (mean of 10 items), S2=.98. The second pretest (N=30) tested the frame valence manipulation and the perceived fit between the binge drinking issue and the campus organization (UF Health Science Center) selected as the brand for the ad. The sample was div ided into two experimental groups (n=15, n=15); one group received a positively framed ad as the stimulus and the other a negatively framed ad. Respondents responded to (2) 7-point scale items rating the degree to which they believed the ad stressed the positive implications of drinking in moderation, and the degree to which they believed the ad stressed the negative implications of binge drinking. Two independent samples t -tests were performed to ensure successful manipulation of the frame valence variab le. The first t test confirmed significantly different group means on the positive implications item (MD=3.13, SE=.89), t(28)=3.53, p=.00. The second t -test confirmed significantly different group means on the second item, stress of negative implications (MD= 2.47, SE=.74), t(28)= 3.33, p= .00. Respondents were also asked to rate the perceived fit of the issue and the Health Science Center using (3) 7-point items anchored by Does not

PAGE 41

41 belong/Belongs with, Does not go together/Goes together, Does not fit together/Fits together. The issue/brand pair achieved an acceptable mean fit rating (M=5.06, SD=1.46). The third pretest (N=60) tested the scales to be used in the final questionnaire, including measures of attitude towards the ad, attitude towards t he brand, behavior intention, enduring involvement, and manipulation checks for frame valence and AMI level. The sample was divided into four experimental groups (n=15; Low AMI/Negative frame, High AMI/Negative frame, Low AMI/Positive frame, High AMI, Posi tive frame). To test the framing manipulation, respondents indicated the extent to which they felt the advertising message stressed either the positive or negative implications of adhering or not adhering to the advocated behavior on a 7 point scale. A one -way ANOVA confirmed the framing manipulation for the degree respondents believed the ad stressed positive implications (M=3.37, SD=1.87), F(3, 56)=5.08, p=.00, and negative implications (M=5.43, SD=1.58), F(3, 56)=4.62, p=.01. To test the AMI manipulation respondents were asked to indicate the extent of their involvement with the advertising message on four 7 -point semantic differential items: Very uninvolved/Very involved, Concentrating very little/Concentrating very hard, Paying very little attention/Pa ying a lot of attention; and by indicating the extent to which they agreed/disagreed with the following statement: I carefully considered the advertising claims about binge drinking ( strongly disagree/strongly agree ). A one way ANOVA found the means betw een high and low involvement groups to vary in the direction expected, however the difference was not significant (M=4.06, SD=.43) F(3, 56)=.60, p=.61. Therefore the reading prime was strengthened for the main experiment. Control of Confounding Variable s Perhaps the most common confounding variables are existing attitudes or evaluations of the independent and dependent variables. Existing attitudes towards brands was controlled for by the pretests, which measured familiarity and credibility of three University of Florida

PAGE 42

42 organizations to be used as the brand. Existing attitudes towards the advertised issues could also confound the results. As stated, subjects rated their personal level of enduring involvement with the issue and this variable was coded and analyzed post -hoc to see if enduring involvement, as opposed to the manipulated AMI (situational) involvement had bearing on the dependent measures. Manipulation Checks There is a possibility for human error in generating positively and negatively fr amed ad copy. To ensure AMI was successfully manipulated, the post -exposure questionnaire included the same four 7 -point semantic differential items used in the pretest to test the involvement manipulation: very uninvolved/very involved, concentrating ve ry little/concentrating very hard, paying very little attention/paying a lot of attention; and the indication of the extent to which they agreed/disagreed with the following statement: I carefully considered the advertising claims about binge drinking, (st rongly disagree/strongly agree). As an additional measure of involvement level, subjects were asked to respond to an unaided recall question, asking them to write down the Web address listed in the ad. To ensure successful manipulation of the framing va riable, a scale item used in the pretest was again used to measure the extent to which subjects felt the advertising message stressed either the positive or negative implications of adhering or not adhering to the advocated behavior on a 7 -point scale. Table 3 1. Descriptive statistics for pretest of binge drinking consequences Mean Std. Deviation Aggression 4.92 .88 Throwing Up 5.34 .85 Embarrassing Yourself 4.84 1.05 Looking Unattractive 4.53 1.09

PAGE 43

43 Table 3 1. Continued Mean Std. Deviation Injuring Yourself 4.83 .84 Passing Out 5.20 1.20 Having a Hangover 5.97 .79 Impotence 4.30 .93 Loss of Memory 4.61 1.06 Promiscuity 4.94 .66 Table 3 2. Descriptive statistics for pretest of campus organizations Mean Std. Deviation GatorWell Health Promotion Services Not at all familiar/Extremely familiar 4.25 3.53 UF Health Science Center (HSC) Not at all familiar/Extremely familiar 2.19 1.36 Shands Health Care Center (SHCC) Not at all familiar/Extremely familiar 3.69 2.23

PAGE 44

44 CHAPTER 4 RESULTS Main Experiment Four hypotheses and one research question were tested to discover how the presentation of information as a benefit or consequence, and a persons level of situational involvement with the message, affect a persons attitudes toward the communication and hi s or her intent to act. A total of 204 subjects took part in the experiment comprising four experimental groups of n=51 each. Subjects were recruited from the undergraduate courses in the College of Journalism and Communications and the Warrington College of Business Administration at the University of Florida. Subjects received extra credit points at the discretion of their instructor for participation. Subjects were randomly assigned to one of the four experimental groups (negative frame/high -AMI; negativ e frame/low -AMI; positive frame/high -AMI; positive frame/low -AMI) via random distribution of the packets by the professor in the lecture. The subjects signed an informed consent form before beginning the experiment. Participants ranged in age from 17 to 28 years of age, M=20.26, SD=1.33. Forty-eight percent were male (n=98) and 52% (n=106) were female. Gender was dispersed throughout the four experimental group (Table 4 3). Most of the respondents were Juniors (39.2%, n=80) followed by Sophomores (33.3% n=68), then Seniors (24.5%, n=50) and finally Freshman (2.9%, n=6). The number of respondents above and below the drinking age of 21 was also calculated due to the stimulus issue. One hundred twenty-nine (36.8%) were age 21 or older, 75 (63.2%) were under 21.

PAGE 45

45 Scale Reliability Analyses Attitude toward the ad Attitude toward the ad was measured with 3 items on a 7 -point scale with endpoints of good/bad, favorable/unfavorable, and pleasant/unpleasant. The Cronbachs alpha for this construct is .76, whic h meets the acceptable reliability of .70 identified by Nunally (1978). Attitude toward the brand Attitude toward the brand was measured with 3 items on a 7 point scale with endpoints good/bad, favorable/unfavorable, unsatisfactory/satisfactory. The Cronbachs alpha for this construct is .86. Behavior intention Behavior intention was also measured using 3 7 -point items. The first asked respondents Since viewing the ad, how likely are you to visit the Web site listed in the ad soon?, the second asked Since viewing the ad, how likely are you to visit the Web site listed in the ad in the future? and finally To what degree did the ad influence your intention to visit the Web site in the future?. When the items were analyzed for reliability, the two it ems asking about intention to visit the Web site soon and in the future showed low correlation (r -.30). A review of the data indicated that subjects might have perceived the questions as mutually exclusive; they would either visit soon or in the future, but would not indicate a high intention to do both. Several reliability analyses were run and the scale reached highest Cronbachs alpha value when the item Since viewing the ad, how likely are you to visit the Web site listed in the ad in the future? wa s removed. The Cronbachs alpha of the revised scale is .76. Advertising Message Involvement This construct was measured as a manipulation check using a 4 item, 7 point scale. The respondents were asked to respond to the statement: When viewing the ad I was: The

PAGE 46

46 endpoints for three of the items are very uninvolved/very involved, concentrating very hard/concentrating very little, paying very little attention/paying a lot of attention. The fourth item was a semantic differential item asking students to respond to the statement, I carefully considered the advertisements claims about binge drinking, on a 7 -point scale with endpoints strongly agree/strongly disagree. The Cronbachs alpha of the AMI scale is .73. Enduring Involvement This concept was me asured as one of the last items in the questionnaire to discover if subjects level of enduring involvement with the issue of binge drinking had bearing on the dependent measures. Zaichowskys 10item enduring involvement scale, adapted from the 22item P ersonal Involvement Inventory was used. Respondents were asked to respond to the statement The issue of binge drinking on college campuses is: on 7 -point scales with endpoints unimportant/important, relevant/irrelevant, means nothing to me/means a lot t o me, worthless/valuable, interesting/boring, unexciting/exciting, unappealing/appealing, mundane/fascinating, needed/not needed, involving/uninvolving. The Cronbachs alpha of the Enduring Involvement scale is .84. Manipulation Checks Multivariate analy sis of variance was performed to check manipulation of message framing. The message framing manipulation was successful. Respondents who viewed the positively framed ad indicated higher agreement that the ad stressed positive implications (M=5.62, SD=1.24) than respondents who viewed the negatively framed ad (M=3.74, SD=2.07). Similarly, respondents who viewed the negative ad indicated higher agreement that the ad stressed negative implications (M=5.87, SD=1.43) than subjects who viewed the positively fram ed ad (M=4.66, SD=1.73), F(2, 201)=40.99, p=.00.

PAGE 47

47 A one -way ANOVA test was performed to check manipulation of AMI level. Manipulation of AMI level was not successful using the AMI index or the individual scale items as dependent variables. A comparison of means between the high versus low AMI condition showed no statistical significance between groups, suggesting that subjects in the high and low AMI groups reported similar levels of involvement (M=4.29 vs. M=4.24), concentration (M=3.93 vs. M=3.82), atten tion (M=4.54 vs. M=4.44), and consideration to message claims (M=5.23 vs. M=4.93). However, the means for each scale item were slightly higher for the high AMI group than for the low AMI group. The mean AMI index score was also higher for the high AMI gr oup, but the result did not approach significance. The mean values on the AMI index for high and low involvement groups were M=4.50, SD=1.03 and M=4.36, SD=1.20 respectively. Hypotheses Tests Two multivariate analyses of variance and a univariate analysis of variance were performed to test the hypotheses. The independent variables included were message frame valence (positive or negative) and AMI level (low or high). The dependent variables were attitude toward the ad and attitude toward the brand, or be havior intention depending on the hypothesis. Hypothesis H1 predicted that the high AMI condition would result in more favorable attitudes toward the ad and brand than the low AMI condition. Results of the multivariate analysis of variance do not suppo rt this hypothesis due to the unsuccessful AMI manipulation. Mean attitudes toward the ad were higher in the high AMI group (M=4.70, SD=1.24) than in the low AMI group (M=4.62, SD=1.13), but the results did not approach significance (F=.25 p=.62). Mean attitudes toward the brand were actually equivalent in the high AMI group (M=5.26, SD=1.15) and the low AMI group (M=5.26, SD=1.13). In other words, there was no

PAGE 48

48 main effect of AMI level on respondents evaluation of the ad or the brand (Wilks Labmda = .10, F(2, 199)=.15, p=.87). Descriptive and inferential statistics for analysis of AMI effect are presented in tables 4 11. and 412. Hypothesis H2 predicted that positively framed advertising messages would result in more favorable attitudes toward the ad and brand than negatively framed advertising messages. Results of the multivariate analysis of variance support this hypothesis, as shown in table 4 13. Subjects who viewed a positively framed ad reported more favorable attitudes toward the ad (M=4.94, SD=1.21) than subjects who viewed a negatively framed ad (M=4.38, SD 1.09). The results were statistically significant (F=11.90, p=.00). Similarly, subjects who viewed a positively framed ad also reported more favorable attitudes toward the brand (M=5. 44, SD=1.15) than subjects who viewed a negatively framed ad (M=5.08, SD 1.17) and this difference was statistically significant (F=5.16, p=.02). These group differences are also evidence of a main effect of frame valence on resulting attitudes toward the brand (Wilks Lambda = .94, F(2, 199)=6.53, p=.00). In other words, people who were presented with an issue ad listing the benefits of drinking in moderation evaluated the ad and the brand, the Health Science Center, more favorably than people who viewed an issue ad listing the consequences of binge drinking. Hypothesis H3a predicted that negatively framed advertising messages would result in higher behavior intention scores for subjec ts in the high AMI condition and H3b stated that positively framed advertising messages would result in higher behavior intention scores for subjects in the low AMI condition. Again, the unsuccessful AMI manipulation was a limitation. Results of the anal ysis of variance follow the prediction, but the results are not statistically significant as shown in table 4 15. Subjects in the high involvement condition who viewed a negative ad reported higher intention to seek out information from the Web site liste d in the ad

PAGE 49

49 (M=2.34, SD=1.31) than subjects who viewed a positive ad (M=2.14, SD=1.30). Subjects in the low involvement condition who viewed a positive ad reported higher intent to behave (M=2.38, SD=1.21) than subjects who viewed a negative ad (M=2.05, S D=1.09). There is no statistically significant interaction effect of effect of message frame valence and AMI level on the dependent measure of behavior intention (F=2.44, p=.12). Hypothesis H4 predicted that overall message framing would have a stronger effect on behavior intention for subjects in the low AMI condition than the high AMI condition. However, the two -way ANOVA did not show a significant interaction effect of message framing and AMI level (p=.12). The analysis of variance did show that the difference between the negative and positive frame behavior intention means for the low AMI subjects (M=2.05 2.38, SD= 1.09, 1.21) was larger than the difference between the negative and positive frame behavior intention means for the high AMI subjects ( M=2.34 2.14, SD=1.31, 1.30), as reported in table 4 15. A 3 x 2 x 2 (enduring involvement x AMI level x frame valence) factorial multivariate analysis was used to explore RQ1. The factorial analysis also allowed for examination of any interaction effects between the three factor variables. The enduring involvement measure was recoded as three levels based on mean scores; low (1 2.33), medium (2.34 4.67) and high (4.68 7). The analysis discovered a main effect of enduring involvement with the issue of b inge drinking on two dependent measures, attitude toward the ad (F=4.02, p=.02), and behavior intention (F=5.46, p=.01). In other words, higher levels of enduring involvement varied directly with subjects favorable evaluation of the ad and their reported behavior intention. Enduring involvement level did not have an effect on attitudes toward the brand (F=.45, p=.64). No significant interactions effects of enduring involvement level x AMI level, enduring involvement level x frame valence, or enduring involvement level x AMI level x frame valence were found.

PAGE 50

50 Descriptive statistics for the factorial analysis are presented in table 4 16. and inferential statistics for the main effect of enduring involvement level are presented in Table 4 17. In summary, the unsuccessful manipulation of AMI level precluded discovery of statistically significant relationships between AMI level and the dependent measures; attitude toward the ad, attitude toward the brand, and behavior intention. This limitation also prevented discovery of an interaction effect between AMI level and message framing. The successful manipulation of message frame valence as either positive or negative allowed for discovery of a causal relationship between frame valence and resulting attitudes towa rd the ad and brand, in that a positively framed ad engendered more favorable attitudes. Exploration of the research questions confirmed that a persons level of enduring involvement with an issue is a predictor of attitudes toward the ad, and behavior in tention; but does not have bearing on attitudes toward the brand. The interaction of enduring involvement level and message frame did not show any statistically significant relationship, and analysis of interaction effect of enduring involvement level and AMI level is limited by the unsuccessful manipulation of AMI level.

PAGE 51

51 Table 4 1. Summary of respondent demographics Total Respondents Low AMI/ Negative Frame High AMI/ Negative Frame Low AMI/ Positive Frame High AMI/ Positive Frame N % N % N % N % N % Gender Male Female Total 98 106 204 48% 52% 100% 29 22 51 14.2% 10.8% 25% 23 28 51 11.3% 13.7% 25% 18 33 51 8.8% 16.2% 25% 28 23 51 13.7% 11.3% 25% Age 17 18 19 20 21 22 23+ Total 1 4 52 72 54 14 7 204 .5% 2.0% 25.5% 35.3% 26.5% 6.9% 3.5% 100% 1 0 14 17 12 4 3 51 .5% .0% 6.9% 8.3% 5.9% 2% 1.5% 25% 0 0 10 20 15 2 4 51 .0% .0% 4.9% 9.8% 7.4% 1.0% 2.0% 25% 0 2 12 21 13 3 0 51 .0% 1% 5.9% 10.3% 6.4% 1.5% .0% 25% 0 2 16 14 14 5 0 51 .0% 1.0% 7.8% 6.9% 6.9% 2.5% .0% 25% Class Fr eshman Sophomore Junior Senior Total 6 68 80 50 204 2.9% 33.3% 39.2% 24.5% 100% 3 14 22 12 51 1.5% 6.9% 10.8% 5.9% 25% 0 14 26 11 51 .0% 6.9% 12.7% 5.4% 25% 1 18 23 9 51 .5% 8.8% 11.3% 4.4% 25% 2 22 9 18 51 1.0% 10.8% 4.4% 8.8% 25% Drinking Ag e Under 21 21+ Total 129 75 204 63.2% 36.8% 100% 32 19 51 15.7% 9.3% 25% 30 21 51 14.7% 10.3% 25% 35 16 51 17.2% 7.8% 25% 32 19 51 15.7% 9.3% 25% Table 4 2. Summary of attitude toward the ad scale In my opinion the ad is: N Mean St d. deviation Good/Bad 204 3.24 1.43 Favorable/Unfavorable 204 3.24 1.36 Pleasant/Unpleasant 204 3.54 1.52 Construct Statistics N items Mean Std. deviation 3 10.02 3.55 Cronbachs = .76

PAGE 52

52 Table 4 3. Summary of attitude toward the brand scale My o verall impression of the Health Science Center is: N Mean Std. deviation Good/Bad 204 5.34 1.36 Unfavorable/Favorable 204 5.19 1.28 Unsatisfactory/Satisfactory 204 5.21 2.13 Construct Statistics N items Mean Std. deviation 3 15.77 3.41 Cronbachs = .86 Table 4 4. Behavior intention scale N Mean Std. deviation Since viewing the ad, how likely are you to visit the Web site listed in the ad soon? 204 2.22 1.37 To what degree did the ad influence your intention to visit the Web site in the futu re? 204 2.24 1.37 Construct Statistics N items Mean Std. deviation 2 4.46 2.46 Cronbachs = .76 Table 4 5. Enduring involvement scale The issue of binge drinking on college campuses is: N Mean Std. deviation Unimportant/Important 204 5.70 1.38 Re levant/Irrelevant 204 5.77 1.54 Means nothing to me/Means a lot to me 204 4.15 1.57 Worthless/Valuable 204 5.39 1.27 Interesting/Boring 204 4.56 1.42 Unexciting/Exciting 204 3.83 1.35 Unappealing/Appealing 204 3.83 1.40 Mundane/Fascinating 204 3.79 1 .18 Needed/Not needed 204 5.52 1.60 Involving/Uninvolving 204 4.72 1.30 Construct Statistics N items Mean Std. deviation 10 47.30 9.09 Cronbachs = .84

PAGE 53

53 Table 4 6. Advertising message involvement (AMI) scale When viewing the ad I was: N Mean St d. deviation Very uninvolved/Very involved 204 4.26 1.48 Concentrating very hard/Concentrating very little 204 3.88 1.55 Paying very little attention/Paying a lot of attention 204 4.49 1.46 I carefully considered the advertisements claims about binge drinking (Semantic Differential) 204 5.08 1.50 Construct Statistics N items Mean Std. deviation 4 17.71 4.47 Cronbachs = .73 Table 4 7. Manipulation check of message framing Experimental Group N Mean Std. deviation Believe ad stressed positive i mplications of drinking in moderation Negative Frame Positive Frame Total 102 102 204 3.74 5.62 4.67 2.07 1.24 1.94 Believe ad stressed negative implications of binge drinking Negative Frame Positive Frame Total 102 102 204 5.87 4.66 5.26 1.43 1.73 1.69 R Squared = .236/.130 (Adjusted R Squared) = .232/.125 Wilks Lambda = .71, F=40.99, d.f.=2, p<.05 Table 4 8. Manipulation check of AMI level Experimental Group N Mean Std. deviation Advertising Message Involvement High AMI Low AMI Total 102 102 204 4.50 4.36 4.43 1.03 1.20 1.12 F = .80, d.f. = 1, p > .05

PAGE 54

54 Table 4 9. Descriptive statistics for hypotheses tests H1, H2 DV AMI Level Mean Std. deviation N Attitude toward the ad High Negative Frame Positive Frame Total 4.37 5.01 4.70 1.1 6 1.25 1.24 51 51 102 Attitude toward the ad Low Negative Frame Positive Frame Total 4.39 4.85 4.62 1.03 1.17 1.13 51 51 102 Attitude toward the brand High Negative Frame Positive Frame Total 5.02 5.50 5.26 1.25 1.00 1.15 51 51 102 Attitude toward the brand Low Negative Frame Positive Frame Total 5.14 5.38 5.26 1.08 1.17 1.13 51 51 102 R Squared = .059 Table 4 10. Main effect of AMI level Factor Dependent Variable Type III sum of squares Mean square F Sig. AMI Level Attitude towards the ad Attitu de towards the brand .340 .000 .340 .000 .254 .000 .615 1.00 Wilks Lambda = 1.00, F = .15, d.f. = 2, p > .05 Table 4 11. Main effect of message framing Factor Dependent Variable Type III sum of squares Mean square F Sig. Frame Valence Attitude toward s the ad Attitude towards the brand Behavior intention 15.93 6.60 .21 15.93 6.60 .21 11.90 5.16 .14 .00 .02 .71 Wilks Lambda = .94, F = 4.34, d.f. = 3, p < .05 Table 4 12. Descriptive statistics for hypotheses tests H3a, H3b, H4 DV AMI Level Mean St d. deviation N Behavior intention High Negative Frame Positive Frame Total 2.34 2.14 2.24 1.31 1.30 1.30 51 51 102 Behavior intention Low Negative Frame Positive Frame Total 2.05 2.38 2.22 1.09 1.21 1.16 51 51 102 R Squared = .013 (Adjusted R Squared = .002)

PAGE 55

55 Table 4 13. Interaction effects of message framing and AMI level Factor Dependent Variable Type III sum of squares Mean square F Sig. Frame Valence x AMI Level Attitude towards the ad Attitude towards the brand Behavior intention .46 .71 3.71 .46 .71 3.71 .34 .55 2.44 .56 .46 .12 Wilks Lambda = .98, F = 1.19, d.f. = 3, p > .05

PAGE 56

56 Table 4 14. Descriptive statistics for RQ1 DV Enduring Involvement Level AMI Level Frame Valence Mean Std. Deviation N Attitud e toward the ad Low Low Negative Positive Total ------0 0 0 High Negative Positive Total -4.17 4.17 -2.12 2.12 0 2 2 Medium Low Negative Positive Total 4.13 4.53 4.29 1.01 1.03 1.03 24 17 41 High Negative Positive Total 4.25 4. 74 4.47 .85 1.20 1.03 24 19 43 High Low Negative Positive Total 4.62 5.01 4.84 1.02 1.22 1.14 27 34 61 High Negative Positive Total 4.48 5.27 4.89 1.38 1.22 1.35 27 30 57 Attitude toward the brand Low Low Negative Positive Total ------0 0 0 High Negative Positive Total -5.67 5.67 -.94 .94 0 2 2 Medium Low Negative Positive Total 5.10 5.29 5.18 1.16 .93 1.06 24 17 41 High Negative Positive Total 4.83 5.46 5.11 1.20 .90 1.11 24 19 43 High Low Negative Positive Total 5.17 5.4 2 5.31 1.03 1.29 1.18 27 34 61 High Negative Positive Total 5.19 5.51 5.36 1.30 1.09 1.19 27 30 57 Behavior intention Low Low Negative Positive Total ------0 0 0 High Negative Positive Total -1.00 1.00 -.00 .00 0 2 2 Medium Low N egative Positive Total 1.83 1.97 1.89 .86 .93 .88 24 17 41

PAGE 57

57 Table 4 14. Continued DV Enduring Involvement Level AMI Level Frame Valence Mean Std. Deviation N High Negative Positive Total 2.02 1.89 1.97 1.15 1.21 1.16 24 19 43 High Low Negative Posi tive Total 2.24 2.59 2.43 1.25 1.29 1.28 27 34 61 High Negative Positive Total 2.63 2.37 2.49 1.40 1.35 1.37 27 30 57

PAGE 58

58 Table 4 15. Main effect of enduring involvement level Factor Dependent Variable Type III Sum of Squares Mean Square F Sig. Endu ring Involvement Level Attitude towards the ad Attitude towards the brand Behavior intention 10.57 1.17 16.10 5.29 .59 8.05 4.02 .59 .5.46 .02 .64 .01 Wilks Lambda = .91, F (6, 386) = 2.98, p < .05

PAGE 59

59 CHAPTER 5 DISCUSSION Issue Advertising has become an important part of many brands marketing communications, as companies seek to build a relationship with consumers that goes beyond the purchase environment. It is no longer only the role of non-profit organizations or government to communicate the impo rtance and implications of social issues to the public. Such messages can effectively be carried by brands backed by industry resources and the power of persuasive communication. The purpose of this research was to investigate the effectiveness of issue advertising; to discover what types of messages of most persuasive given different audience characteristics and communication contexts. The research manipulated two independent variables, message frame valence (negative or positive) and subjects level of Advertising Message Involvement (AMI). Subjects level of enduring involvement was also measured during the experiment and analyzed post hoc as an additional factor that may influence attitudes and behavior. The dependent variables measured were attitu de toward the ad, attitude toward the brand, and behavior intention. Measurement of attitudes toward the ad and brand allow for discovery of the way pairing a brand and an issue affect evaluation of the communication itself, and the brand. Measurements of attitude towards the ad and attitude towards the brand are of particular importance to this research, as it uses branded issue advertisements as the experimental stimuli, rather than health messages or unbranded product advertising messages. Measurement of behavior intention is an especially important variable given the issue ad stimulus, as the highest goal of issue awareness is action.

PAGE 60

6 0 Evaluation of Hypotheses Prospect theory was used to predict relationships between the independent variables and behavi or intention. Prospect theory is an important theoretical model used by researchers to discuss the effects of frame valence on audience choice or action. Prospect theory is commonly applied in discussion of goal framing methods (Levin et al, 1998), where frame valence is described in terms of gains and losses. According to prospect theory, peoples responses to losses are typically more extreme than responses to gains. In other words, there is more displeasure associated with losing something than pleasu re associated with gaining something of the same value. Research (e.g. Levin et al., 1998; Meyerowitz and Chaiken, 1987) has discovered that when framing is based on goals involving gains and losses, there is an advantage for negatively framed messages be cause people are motivated to avoid loss. Hypotheses H3a and H3b sought to test prospect theory within an issue advertising context, predicting that a negatively framed issue ad would be more persuasive in influencing behavior intention than a positively framed issue ad for subjects in the high AMI condition, and that a positively framed issue ad would be more effective for subjects in the low AMI condition. The expectation is that more highly involved people would be more willing to act to mitigate a lo ss while less involved people would be more influenced by achieving a gain. The test of hypotheses H2a and H3b did not yield a statistically significant result, though high AMI subjects in the negative frame condition did report higher intent to behave th an subjects in the positive frame condition and low AMI subjects in the positive frame condition reported higher intent to behave than those in the negative frame condition. Had this relationship been statistically significant it may have helped guide development of issue advertising copy based on audience involvement levels. Hypothesis H4 was tested to see if framing has a more pronounced effect on behavior intention for low AMI subjects than for high AMI subjects. This hypothesis was guided

PAGE 61

61 by the conce ptualization of message framing as an executional technique (Tsai, 2007) that subjects in the low involvement condition would give greater weight to as a heuristic cue. Again, no statistically significant difference in the means for each group was found, perhaps due to the limitation of a weak AMI level manipulation. The Elaboration Likelihood Model of persuasion (ELM) was used to predict subjects evaluations of the attitude object, the issue ad stimulus. A review of the literature suggests that the sim ple cue effect of framing has a strong influence on subjects attitude formation. A cue effect suggests that because a message is framed positively, it will engender more positive evaluation of the attitude object. Conversely, a negatively framed message m ay arouse negative associations and lead to negative evaluation of the attitude object. Maheswaran and Meyers Levy (1990) used the ELM to model their hypotheses regarding subjects propensity to take a diagnostic blood test, and predicted that people in t he low involvement condition are unlikely to pay close attention to message arguments, or integrate their beliefs into an overall attitude. Instead, people in the low involvement condition use peripheral cues in the message context, such as whether the me ssage is associated with positive or negative cues. In this way frame valence would vary directly with evaluations, meaning a negative frame would engender negative evaluations and a positive frame would engender positive evaluations for low involvement s ubjects. Hypothesis H2 predicted that positively framed advertising messages would result in more favorable attitudes toward the ad and brand and statistical evidence of this relationship was found. This follows the prediction of the Elaboration Likeliho od Model for effects of a simple cue effect. However, it is expected that subjects in a low involvement condition would be more susceptible to a heuristic cue than high involvement subjects. In this research subjects in the

PAGE 62

62 high AMI condition actually re porter greater variation in mean evaluations for positive versus negative frame than subjects in the low AMI condition. This may be due in part to the prediction of the ELM that attitudes formed via central route processing, under conditions of higher cog nitive elaboration, are more enduring than attitudes formed during peripheral route processing. Hypothesis H1 predicted that subjects in the high AMI condition would report more favorable attitudes toward the ad and brand than subjects in the low AMI condition. Tests of H1 did not find a significant effect of AMI level on attitudes toward the brand and ad. One reason for this may be the weak manipulation of the AMI level independent variable. The reading prime for the stimulus was intended to place subjects in either a high or low involvement condition. Subjects in the high AMI condition did report slightly higher levels of involvement in the manipulation check (M=4.50, SD=1.03 vs. M=4.36, SD=1.20) but the difference was not statistically significant. One reason for this may have been the experimental setting. All subjects were reminded at the start of the experiment to read the questions carefully. Multiple items were reverse -coded, and subjects needed to pay attention when marking their response on the scale. This may have primed all subjects to report higher levels of concentration and attention on the manipulation check questions. It is interesting to note that one of the scale items checking AMI manipulation that asks specifically about how car efully they considered message claims (rather than asking about their general state of mind when viewing the ad) did show a significant difference between high and low AMI groups (F=3.73, p=.01). Evaluation of Research Question Enduring involvement level proved to be an important variable to measure post -hoc. This research question assessed how subjects existing involvement with the binge drinking issue affected the dependent variables. A statistically significant effect of enduring involvement level

PAGE 63

63 w as found on attitudes toward the ad and behavior intention. In other words, subjects who care more about the issue of binge drinking and perceive it as personally relevant will respond more favorable to ads promoting the issue and will be more likely to ac t. It is interesting that this variable did not have an effect on evaluations of the brand. This finding brings to light a difference between direct to consumer advertising and issue advertising. In a consumer context, it is easier and more efficient to get existing purchases to buy more; than to gain new customers. But in an issue awareness or advocacy context, is it more valuable to communicate with the involved audience, or to get a less involved or uninvolved audience to participate? Limitations and Future Research The use of a convenience sample is a limitation of this research. All respondents were undergraduate students, making the research difficult to generalize to the general population. However, a strength of this research is the generalize abilility of the topic to other college and university campuses. Developing an effective binge drinking issue awareness campaign is a concern for many education institutions. The unsuccessful manipulation of AMI level was the main limitation of this stu dy. Subjects reported similar levels of involvement with the issue advertising message regardless of the reading prime. This result may point to the importance of exposure frequency or duration of the stimulus. Subjects were not required to read the pr ime or look at the issue ad for any set length of time, and only read the prime and viewed the ad once. This level of exposure may not have been sufficient. One idea for lengthening exposure would be to distribute the stimulus and reading prime before th e experiment, then collect the stimulus and distribute the questionnaire. Subjects may spend more time reading the prime and viewing the ad if they know they will not be able to turn back to it. Subjects AMI level could have also been manipulated further within the attentional construct. Subjects in the low AMI condition could have been given multiple ads

PAGE 64

64 to view, limiting the attention paid to the stimulus ad. Subjects in the high AMI condition would only view one or two ads, therefore increasing focus on the stimulus (Laczniak & Muehling, 1993). Selection of binge drinking as the stimulus issue may have also confounded the AMI level manipulation. This is a common and often-discussed issue on college campuses. The pretest conducted to develop the st imulus tested consequences of binge drinking that the sample population viewed as true, credible and believable. In this way the issue ad stimulus was not presenting new information, but reiterating what students already knew and believed. Presentation o f new information may have been more salient, and therefore a better motivator of behavior. Inclusion of a control group in the experimental design could have also strengthened this study. This way subjects in the low and high AMI conditions could not on ly be compared with each other, but to a baseline message involvement level with no reading prime. The design of the stimulus itself could also be improved. The University of Florida Health Science Center was chosen as the brand for the ad, but the readi ng instructions informed subjects that either the University of Floridas student government (high AMI prime) or the state of Florida (low AMI prime) were interested in developing the issue awareness campaign. Increasing the congruency between the reading prime and the issue ad stimulus could make the stimulus more credible. Overall, behavior intention mean scores were very low (M=2.23, SD=1.23) indicating most subjects had little intention of visiting the Web site to learn more about the advocated issue soon, or in the future. Behavior intention could have also been more congruent with the reading prime, asking students about their intention to vote online on potential issue ads (high AMI prime) or by asking students to visit a state of Florida Web site to review potential issue ads (low

PAGE 65

65 AMI prime). Conducting a similar experiment online might have allowed for a more realistic gauge of behavior intention. For example, a hyperlink could be included in the last part of the questionnaire and students would have the option of clicking to learn more about the student governments or state of Floridas campaign. Managerial Implications Several research findings have practical implication for practitioners incorporating issue advertising into their marketing m ix. First, frame valence has significant bearing on attitudes toward the ad and brand, but a negative frame may be more effecting in influencing behavior intention for high involvement subjects. Brands may have to weigh the importance of desired effects when creating issue ad messages. If attitudes toward the brand must be protected, then a negative frame may be a risk. If behavior is the ultimate goal, a negative frame may be more effective. It was also discovered that subjects level of enduring involvement had bearing on attitudes toward the ad and behavior intention, but not on attitudes toward the brand. It may be important for brands to bear in mind that just because an audience is involved with an issue; creating a message around that issue doe s not automatically earn a brand favorability with an audience. In other words, an issue ad alone may not be sufficient to build a consumer brand relationship. The findings of this research are also the product of an experimental setting. Tests of theor y in this context may have limited application in the field. Consumers generally see and perceive advertising within an environment saturated with advertising and marketing communications. A different effect of message framing may be found in the field. For example, Smith and Petty (1996) note the existence of netativly bias in interpreting communication. The researchers suggest that negatively framed information may have more bearing on audiences evaluations simply because it stands out from other mes sages. Meyerowitz and Chaiken (1987)

PAGE 66

66 also note that negative may be more salient because people are generally surrounded by positive information. The broad -based goal of this research was to discover the most effective way to present issue advertising m essage copy given differences in audience involvement level with the advertising message. In other words, if an audience is not interested in the message or does not care, what is the best way to increase interest? And if an audience is interested, what is the best way to capitalize on that to achieve objectives, whether attitude or behavior based? This research shows that message framing has an effect on attitudes toward the ad and brand, and that enduring involvement level affects attitudes toward the ad and behavior intention. Additional research may discover significant effects of AMI level on attitudes and behavior, and discover more about the way message frame valence and advertising message involvement interact.

PAGE 67

67 APPENDIX A EXHIBIT 1: NESTL ISSU E AD

PAGE 68

68 APPENDIX B EXHIBIT 2: GREENPEAC E NEGATIVE FRAME ISS UE AD

PAGE 69

69 APPENDIX C EXHIBIT 3: GREENPEAC E POSITIVE FRAME ISS UE AD

PAGE 70

70 APPENDIX D INFORMED CONSENT

PAGE 71

71 APPENDIX E NEGATIVE FRAME ISSUE AD STIMULUS

PAGE 72

72 APPENDIX F POSITIVE FRAME ISSUE AD STIMULUS

PAGE 73

73 APPENDIX G QUESTIONNAIRE Issue Advertising Experiment You have been selected to participate in an experiment testing the effects of different issue advertising executions on readers evaluations of the ad. Please review and sign the attached informed c onsent form if you choose to participate. This questionnaire contains 29 questions, and should take 15 minutes to complete. All responses are valuable. Your identity will remain anonymous and your responses confidential. You may withdraw from the expe riment at any time with no penalty. When you are finished please return the completed questionnaire and signed informed consent to the professor. If you need clarification on a question at any time during the experiment, please raise your hand. This re search would not be possible without your helpful input and evaluations. Thank you sincerely for your time, Anna C. Parker Researcher annacp@ufl.edu

PAGE 74

74 Study Background: Binge drinking is defined as having more than 5 drinks in one night. The student government at the University of Florida is launching a campaign that focuses on the issue of binge drinking during the Spring 2009 semester. Students will soon be asked to vote on campaign messages to determine which will be implemented. Please take your time reviewing the following proposed campaign advertisement. Your evaluation of the message is very valuable in helping the University of Florida develop an effective campaign.

PAGE 75

75 Instructions: Read the following questions and statements carefully. Thi nk about the ad you just viewed. Indicate how you feel about each statement at this moment There are no right or wrong answers. Mark only one response by putting an X on the scale. The first three questions ask for your opinion of the advertisement. Please record your response to each question on a scale of 1-7. In my opinion the ad is: 1. Good :____:____:____:____:____:____:____: Bad 1 2 3 4 5 6 7 2. Favorable :____:____:____:____:____:____:____: Unf avorable 1 2 3 4 5 6 7 3. Pleasant :____:____:____:____:____:____:____: Unpleasant 1 2 3 4 5 6 7 The next three questions ask for your opinion of the University of Florida organization listed on the advertisement. Please record your response to each question on a scale of 1 -7. My overall impression of the Health Science Center is: 4. Good :____:____:____:____:____:____:____: Bad 1 2 3 4 5 6 7 5. Unfavorable :____:____:____:____:____:____:____: Favorable 1 2 3 4 5 6 7 6. Unsatisfactory :____:____:____:____:____:____:____: Satisfactory 1 2 3 4 5 6 7

PAGE 76

76 The next three questions ask about your intention to visit the Web site listed in the ad. Please read each statement and record your response by placi ng an X on the scale. 7. Since viewing the ad how likely are you to visit the Web site listed in the ad soon ? ____ Very likely ____ Likely ____ Somewhat likely ____ Neither likely nor unlikely ____ Somewhat unlikely ____ Unlikely ____ Very unlikely 8. Since viewing the ad, how likely are you to visit the Web site listed in the ad in the future ? ____ Very unlikely ____ Unlikely ____ Somewhat unlikely ____ Neither unlikely nor likely ____ Somewhat likely ____ Likely ____ Very likely 9. To what degr ee did the ad influence your intention to visit the Web site in the future? Influenced me very much :____:____:____:____:____:____:____: Did not influence me at all 1 2 3 4 5 6 7

PAGE 77

77 The next three questions ask about how you felt when you read the message, and how you approached reading it. Please record your response to each question on a scale of 1-7. When viewing the ad I was: 10. Very Uninvolved :____:____:____:___ _:____:____:____: Very Involved 1 2 3 4 5 6 7 11. Concentrating very hard :____:____:____:____:____:____:____: Concentrating very little 1 2 3 4 5 6 7 12. Paying very little attention :____:____:____:____:____:____:____: Paying a lot of attention 1 2 3 4 5 6 7 Please read t he following statement and record your response by placing an X on the scale. 13. I carefully considered the advertisements claims about binge drinking. ____ Strongly Agree ____ Agree ____ Somewhat Agree ____ Neither Agree nor Disagree ____ Somewhat Disagree ____ Disagree ____ Strongly Disagree The next question asks if you remember the Web site address listed in the ad. Write the address as you remember it, without turning back to the advertisement. 14. ___________________________________________________________ IMPORTANT : Do not turn back to the ad to answer question #14. It is not important whether you remember the Web site address; anything you remember is valuable.

PAGE 78

78 The next questions ask about the advertising message copy. Please record your response to by placing an X on the scale. 15. The advertising message stressed the positive implications of drinking in moderation. ____ Strongly Agree ____ Agree ____ Somewhat Agree ____ Neither Agree nor Disagree ____ Somewhat Disagree ____ Disagree ____ Strongly Disagree 16. The advertising message stressed the negative implications binge drinking. ____ Strongly Disagree ____ Disagree ____ Somewhat Disagree ____ Neither Disagree nor Agree ____ Somewhat Agree ____ Agree ____ Strongly Agree

PAGE 79

79 The next ten questions ask how you feel personally about the issue of binge drinking. Think about how you feel about the issue of binge drinking in gene ral, and how you would answer these questions on any given day, regardless of your participation in this experiment. Read the statement carefully and respond to each question by placing an X on the scale. The issue of binge drinking on college campus es is: 17. Unimportant :____:____:____:____:____:____:____: Important 1 2 3 4 5 6 7 18. Relevant :____:____:____:____:____:____:____: Irrelevant 1 2 3 4 5 6 7 19. Means nothing to me :____:____:____:____:____:____:____: Means a lot to me 1 2 3 4 5 6 7 20. Worthless :____:____:____:____:____:____:____: Valuable 1 2 3 4 5 6 7 21. Interesting :____:____:____:____:____:____:____: Boring 1 2 3 4 5 6 7 22. Unexciting :____:____:____:____:____:____:____: Exciting 1 2 3 4 5 6 7 23. Unappealing :____:____:____:____:____:____:____: Appealing 1 2 3 4 5 6 7 24. Mundane :____:____:____:____:____:____:____: Fascinating 1 2 3 4 5 6 7 25. Needed :____:____:____:____:____:____:____: Not Needed 1 2 3 4 5 6 7 26. Involving :____:____:____:____:____:____:____: Uninvolving 1 2 3 4 5 6 7

PAGE 80

80 The last three questions are about you. 27. Please indicate your gender with an X. ____ Male ____ Female 28. Please write your age in years. _________ 29. Please indicate your class level. ____ Freshma n ____ Sophomore ____ Junior ____ Senior ____ Graduate student (Masters or Ph.D) You have completed the experiment. Thank you again for your time and valuable opinions. Please return the testing materials, including this questionnaire and the sig ned informed consent form to the professor.

PAGE 81

81 REFERENCE LIST Andrews, J.C. & Durvasula, S. (1991). Suggestions for Manipulating and Measuring Involvement in Advertising Message Content. Advances in Consumer Research. 18(1991) 194201. Bloom, J. Agencies will have to steer marketers toward the big ideal. (Column). Advertising Age 78.40 (Oct. 8, 2007): 24. General Reference Center Gold Gale. University of Florida. 10 Nov. 2007 . Bruner, G.C., Hensel, P.J. & James (1998 2001). Marketing Scales Handbook: A Compilation of Multi Item Measures. American Marketing Association Cacioppo, J.T., & Petty, R.E. (1984). The Elaboration Likelihood Model of Persuasion. Advances in Consumer Resear ch. 11(1), 673675. Chatterjee, S., Heath, T.B., Milberg, S.J., & France, K.R. (2000). The Differential Processing of Price in Gains and Losses: The Effects of Frame and Need for Cognition. Journal of Behavioral Decision Making, (13) 61 75. Chyi, H.I. & McCombs, M. (2004). Media Salience and the Process of Framing: Coverage of the Columbine School Shootings. Journalism and Mass Communication Quarterly, 81(1) 22 35. Cone Inc. (1999). ConelRoper Cause Related Trends Report Boston, MA: Cone, Inc. Con e Inc. (2004). 2004 Cone Corporate Citizenship Study. Boston, MA. Retrieved May 10, 2008 from . Cone Inc. (2006). The 2006 Cone Millenial Cause Study. Boston, MA. Retrieved May 10, 2008, from . Cone Inc. (2008). 2008 Green Gap Survey Fact Sheet. Boston, MA. Retrieved May 10, 2008 from . Davis, J.J. (1995). The Effects of Message Framing on Response to Environmental Communications. Journalism & Mass Communication Quarterly, 72(2) 285299. Entman, R.M. (1 993). Framing: Toward Clarification of a Fractured Paradigm. Journal of Communication, 43(4) 51 58.

PAGE 82

82 Ferguson, E. & Gallagher, L. (2007). Message framing with respect to decisions about vaccination: The roles of frame valence, frame method and perceive d risk. British Journal of Psychology, 98(2007) 667 680. Gupta, S., & Pirsch, J. (2006). The company-cause -customer fit decision in cause related marketing. Journal of Consumer Marketing. 23(6) 314326. Hoeffler, S. & Keller, K.L. (2002). Building Bra nd Equity Through Corporate Societal Marketing. Journal of Public Policy & Marketing. 21(1) 78 89. Hoyer, W., & MacInnis, D. (2007). Consumer Behavior Boston, MA: Houghton Mifflin Company. Iorio, S.H. & Huxman, S.S. (1996). Media Coverage of Politic al Issues and the Framing of Personal Concerns. Journal of Communication, 46(4) 97 115. Johnson, B.T. & Eagly, A.H. (1989). Effects of Involvement on Persuasion: A Meta Analysis. Psychological Bulletin. 106(2), 290314. Johnson, B.T. & Eagly, A.H. (19 90). Involvement and Persuasion: Types, Traditions, and the Evidence. Psychological Bulletin. 107(3), 375384. Kokkinaki, F. & Lunt, P. (1999). The effect of advertising message involvement on brand attitude accessibility. Journal of Economic Psychology. 20(1999) 41 51. Laczniak, R.N. & Muehling, D.D. (1993a). The Relationship between Experimental Manipulations and Tests of Theory in an Advertising Message Involvement Context. Journal of Advertising, 22(3) 59 74. Laczniak, R.N. & Muehling, D.D. (1993b). Toward a Better Understanding of the Role of Advertising Message Involvement in Ad Processing. Psychology & Marketing. 10(4) 301319). Lacniak R.N., Muehling, D.D. & Grossbart, S. (1989). Manipulating Message Involvement in Advertising Research. Journal of Advertising. 18(2) 28 38. Levin, P.I., 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(2) 149188. Levin, P.I., Gaeth, G.J., Schreiber, J. & 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(1) 411429. Mah eswaran, D. & Meyers levy, J. (1990). The Influence of Message Framing and Issue Involvement. Journal of Marketing Research, 27(August 1990) 361367.

PAGE 83

83 Martin, B. & Marshall, R. (1999). The interaction of message framing and felt involvement in the conte xt of cell phone commercials. European Journal of Marketing, 33(1/2) 206218. Meyerowitz, B.E. & Chaiken, S. (1987). The Effect of Message Framing on Breast Self Examination Attitudes, Intentions, and Behavior. Journal of Personality and Social Psychol ogy, 52(3) 500510. Meyers Levy, J. & Maheswaran, D. (2004). Exploring Message Framing Outcomes When Systematic, Heuristic, or Both Types of Processing Occur. Journal of Consumer Psychology, 14(1&2) 159167. Nan, X., & Heo, K. (2007). Consumer Respons es to Corporate Social Responsibility (CSR) Initiatives. Journal of Advertising, 36(2), 6374. Neff, J., Bush, M., York, E.B., & Zmuda, N. (2008). Yes, there is an ROI for doing good. Advertising Age 79(21) 14 15. Obermiller, C. (1995). The Baby is S ick/The Baby is Well: A Test of Environmental Communication Appeals. Journal of Advertising, 24(2) 5570. Petty, R.E., & Cacioppo, J.T. (1979). Issue involvement Can Increase or Decrease Persuasion by Enhancing Message Relevant Cognitive Responses. Jo urnal of Personality and Social Pyschology, 37(10) 19151926. Petty, R.E., Cacioppo, J.T., & Schumann, D. (1983). Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement. Journal of Consumer Research, 10 (September 1983) 135146. Petty, R.E., & Cacioppo, J.G. (1984). The Effects of Involvement on Responses to Argument Quantity and Quality: Central and Peripheral Routes to Persuasion. Journal of Personality and Social Psychology, 46(1) 69 81. Petty, R.E., & Cacio ppo, J.T. (1990). Involvement and Persuasion: Tradition Versus Integration. Psychological Bulletin, 107(3) 367374. Richards, J.I. & Curran, C.M. (2002). Oracles on Advertising: Searching for a Definition. Journal of Advertising, (31(2) 63 77. Ross J.K., Patterson, L.T., & Stutts, M. (1992). Consumer Perceptions of Organizations That Use Cause Related Marketing. Journal of the Academy of Marketing Science, 20(1) 9397.

PAGE 84

84 Sciulli, L.M., & Bebko, C. (2005). Social Cause versus Profit Oriented Adve rtisements: An Analysis of Information Content and Emotional Appeals. Journal of Promotion Management, 11(2/3) 17 36. Sego, T. (2002).Consumers' Ethical Judgments of Issue Advertising. Advances in Consumer Research. 29, 80 85. Shiv, B., Edell, J., & P ayne, J.W. (1997). Factors Affecting the Impact of Negatively and Postively Framed Ad Messages. Journal of Consumer Research, 24(December 1997) 285294. Shiv, B., Edell Britton, J.A., & Payne, J.W. (2004). Does Elaboration Increase or Decrease the Effe ctiveness of Negatively versus Positively Framed Messages Journal of Consumer Research, 31(June 2004) 199208. Smith, S.M., & Petty, R.E. (1996). Message Framing and Persuasion: A Message Processing Analysis. Personality and Social Psychology Bulletin, 22(3) 257268. Tenser, J. The New Samaritans; Corporate social responsibility grows as marketers find that doing good is good business. Advertising Age 7724 (June 12, 2006): S 1 General Reference Center Gold Gale. University of Florida. 10 Nov. 2007 . Tsai, S. (2007). Message Framing Strategy for Brand Communication. Journal of Advertising Research, (September 2007) 364 377. Tversky, A. & Kahneman, D. (1981). The Framing of Decis ions and the Psychology of Choice. Science, 211(4481) January 30, 1981, 453458. Zhang, Y. and Buda, R. (1999). Moderating Effects of Need for Cognition on Responses to Positively versus Negatively Framed Advertising Messages. Journal of Advertising, 28(2) Summer 1999, 1 14. Zimbardo, P.G. (1960). Involvement and Communication Discrepancy as Determinants of Opinion Conformity. Journal of Abnormal and Social Psychology, 60(1) 86 94.

PAGE 85

BIOGRAPHICAL SKETCH Anna Parker was born in Baltimore, Maryland, a nd grew up in Goshen, Indiana. She attended Goshen College and graduated with a B.A. in communication. Anna engaged in many communication forms as a student, including print and broadcast journalism, and producing. After graduation, Anna worked as a pro ducer with Everblue Media in Goshen, Indiana. She went on to work in marketing and business development which nurtured her passion for marketing and communication strategy. Anna returned to school in 2007 to pursue a Master of Advertising at the Univers ity of Florida. While pursing her masters degree, Anna completed an internship with BBDO, Atlanta, where she worked in account management for REI (Recreational Equipment, Inc.). After completion of her masters degree, Anna will return to an agency and a pply her commitment to brand leadership.