xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001966453
007 cr mnu|||uuuuu
008 081021s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002232
Kemp, Deena G.
Source credibility and public information campaigns :
b the effect of audience evaluations of organizational sponsors on message acceptance
h [electronic resource] /
by Deena G. Kemp.
[Tampa, Fla.] :
University of South Florida,
ABSTRACT: This study establishes a link between research on organizational source credibility and the effects of public information campaigns. Research has established that source credibility is one factor audiences evaluate when responding to messages and that credible information sources enhance message acceptance, while untrustworthy sources can interfere with desired message effects. Although source credibility studies have typically focused on the person delivering a message, recent studies indicate that audience perceptions of the organization sponsoring a message has a direct effect on message acceptance as well. Additionally, a few studies indicate that non-profit sources of health information are viewed as more credible, while such messages presented by for-profit organizations are less effective. This study uses an experimental procedure to investigate the relationship between organizational status, source credibility, and two possible effects of public service messages, information seeking and behavioral intent. Based on previous findings, the study hypothesized that the non-profit source would berated as more credible and that as the audiences' perception of source credibility increases so would their willingness to seek additional information or perform the advocated behaviors. Findings indicate, however, that organizational status does not have a significant effect on perceptions of source credibility. Nor does it significantly influence message evaluation, information seeking, or behavioral intent. As predicted, there was a positive correlation between source credibility, message credibility, problem recognition, personal relevance, information seeking, and behavioral intent. The results also indicate that information seeking positively predicts behavioral intent.
Thesis (M.A.)--University of South Florida, 2007.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 103 pages.
Adviser: Derina Holtzhausen, Ph.D.
Public service announcements.
x Mass Communications
t USF Electronic Theses and Dissertations.
Source Credibility and Public Informati on Campaigns: The Effect of Audience Evaluations of Organizational Sponsors on Message Acceptance by Deena G. Kemp A thesis submitted in partial fulfillment of the requirements for the degree of Master of Arts School of Mass Communications College of Arts and Sciences University of South Florida Major Professor: Derina Holtzhausen, Ph.D. Scott Liu, Ph.D. Kelly Page Werder, Ph.D. Date of Approval: November 27, 2007 Keywords: public service announcements, message evaluation, problem recognition, personal involvement, information seeking, behavioral intent Copyright 2007, Deena G. Kemp
DEDICATION This thesis is dedicated to my ne wborn daughter, Arianna, who was born one month before I defended my study. She wonÂ’t remember me holding her in one arm as I typed my concluding chapter with the othe r, but, someday, I hope this manuscript reminds her that through Christ nothing is impossible.
ACKNOWLEDGMENTS There are several people I must thank, fo r without their help completing this thesis would not have been possible. To my thesis chair, Dr. Holtzhausen, th ank you for being patient with me in the beginning when I changed my topic several times and for helping me clarify my thinking and writing as I developed this study. You we re always available to help and reassured me often through the process. If it were not fo r your support, especia lly toward the end, I truly would not have finished. To my thesis committee members, Dr. Liu and Dr. Werder, I appreciate the hard questions, helpful information, and constructiv e criticism, which ultimately made this final draft better. To my fellow graduate student and friend Andrea, your support, determination, and friendly competition kept me going when I wanted to give up. I especially want to thank my husband, An war, who spent his birthday helping me print 500 brochures and questionnaires, a nd my cousin-in-law Tammi, who spent a weekend helping me fold questionnaires, and la bel and stuff envelopes. I must also thank my fellow graduate students Titi and Molly fo r helping me administer the questionnaires and Dr. Rick Wilber for allowing me to us e his students as the studyÂ’s participants. Finally, to my mother, who was there to read my drafts, ask questions, and lend a comforting ear whenever I felt stuck, thank you for always believing in me.
i TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES v LIST OF ACRONYMS vi ABSTRACT vii CHAPTER 1: INTRODUCTION 1 CHAPTER 2: LITERATURE REVIEW 7 The Effects of Public Service Messages 7 Source Credibility and Message Acceptance 9 The Organization as Source 10 Non-profit versus Commercial Sponsors 13 Measuring Sponsor Credibility 17 Audience Perceptions of Message Sponsors 15 Sponsor Credibility and PSA Effect s: Information Seeking versus Behavio r al Intent 19 CHAPTER 3: THEORETI CAL FRAMEWORKS 21 Source Credibility and Information Processing 22 Situational Theory 23 Involvement and Third-Person Perception 24 Theory of Reasoned Action 26 CHAPTER 4: RESEARCH QUESTIONS AND HYPOTHESES 29 Organizational Status, Source Credibility, and Message Credibility 32 Source Credibility, Problem Recogni tion, and Personal Involvement 33 Source Credibility, Information S eeking, and Behavioral Intent 33 Information Seeking versus Behavioral Intent 35
ii CHAPTER 5: METHODOLOGY 36 Study Participants 36 Stimulus Materials 37 Manipulation Check 38 Measurement Apparatus 39 Procedure 42 Data Analysis 43 CHAPTER 6: RESULTS 44 Descriptive Statistics 45 Reliability 49 Research Question 1 and Hypothesis 1 51 Research Question 2 and Hypothesis 2 53 Research Question 3 and Hypothesis 3 56 Research Question 4 and Hypothesis 4 57 Research Question 5 and Hypothesis 5 59 Research Question 6 61 CHAPTER 7: DISCUSSION 63 Organizational Status, Source Credibility, and Message Credibility 63 Problem Recognition and Personal Relevance 66 Information Seeking and Behavioral Intent 68 CHAPTER 8: CONCLUSIONS 73 Study Limitations 74 Suggestions for Future Research 75 Implications for Practice 77 REFERENCES 79 APPENDICES 85 Appendix A.1: Participant Directions 86 Appendix B.1: Subtle NP Treatment 87 Appendix B.2: Subtle FP Treatment 88 Appendix B.3: Detailed NP Treatment 89
iii Appendix B.4: Detailed FP Treatment 90 Appendix B.5: Control Treatment 91 Appendix C.1: Manipulation Check 92 Appendix D.1: Measurement Instrument 93 Appendix E.1: Data Coding Sheet 99 Appendix F.1: Correlations 102
iv LIST OF TABLES Table 1. Experimental Treatments 36 Table 2. Categorical Demographics 44 Table 3. Distribution of Partic ipants to Treatments 45 Table 4. Descriptive Statistics 46 Table 5. CronbachÂ’s alphas 50 Table 6. Means and Standard Deviations for Indices 50 Table 7. Means and Standard Deviations for Source Credibility 52 Table 8. Means and Standard Deviations for Message Credibility 53 Table 9. Means and Standard Deviations for Problem Recognition 55 Table 10. Means and Standard Deviations for Personal Relevance 55 Table 11. Means and Standard Deviations for Information Seeking 57 Table 12. Means and Standard Devia tions for Behavioral Intent 59 Table 13. Means for Information Seeking and Behavioral Intent 60 Table 14. Regression Model for Information-Seeking Behavior 60 Table 15. Regression Model fo r Behavioral Intent 61 Table 16. Five-Factor Regression M odel for Behavioral Intent 62
v LIST OF FIGURES Figure 1. Relationships among source credibility message credibility, message acceptance, and information seeking 30 Figure 2. Relationships among source credib ility, message credibility, message acceptance, and behavioral intent 31 Figure 3. Proposed Model of Behavioral Intention 73
vi LIST OF ACRONYMS ANOVA Analysis of Variance CRM Cause Related Marketing DCM Dual Credibility Model DTC Direct-to-Consumer FPP First-Person Perception HPV Human Papilloma Virus PSA Public Service Announcements RLS Restless Leg Syndrome STP Situational Theory of Publics TPE Third-Person Effect TPP Third-Person Perception TRA Theory of Reasoned Action
vii Source Credibility and Public Informatio n Campaigns: The Effect of Audience Evaluations of Organizational Sponsors on Message Acceptance Deena G. Kemp ABSTRACT This study establishes a link between research on organizational source credibility and the effects of public information campaigns Research has established that source credibility is one factor audiences evaluate when re sponding to messages and that credible information sources enhance messa ge acceptance, while untrustworthy sources can interfere with desired message effects. Although source credibility studies have typically focused on the person delivering a message, recent st udies indicate that audience perceptions of the organization s ponsoring a message has a direct effect on message acceptance as well. Additionally, a few studies indicate that non-profit sources of health information are viewed as more cr edible, while such messages presented by forprofit organizations are less effective. This study uses an experimental procedure to investigate the relationship between organi zational status, source credibility, and two possible effects of public serv ice messages, information seeking and behavioral intent. Based on previous findings, the study hypothesi zed that the non-pr ofit source would be rated as more credible and that as the audiencesÂ’ perception of source credibility increases so would their willingness to s eek additional information or perform the advocated behaviors. Findings indicate, however that organizational status does not have a significant effect on perceptions of s ource credibility. Nor does it significantly influence message evaluation, information seek ing, or behavioral in tent. As predicted, there was a positive correlation between source credibility, message credibility, problem recognition, personal relevance, information se eking, and behavioral intent. The results also indicate that information seeking positively predicts behavioral intent.
1 CHAPTER 1: INTRODUCTION Public service messages are considered altrui stic promotional material in that they Â“address problems assumed to be of general concern to citizens at largeÂ…attempt to increase public awareness of such problem s and their possible solutions, and in many instances also try to influen ce public beliefs, attitudes, and behaviorsÂ” related to these issues (O'Keefe & Reid, 1990, p. 67). Unlike commercial advertising, these messages do not sell or promote a product or service. A lthough they are a form of issue advertising, they also differ from other issue advertisements like institutional and advocacy advertisements because they neither tout th e image of a company nor bolster the sociopolitical perspective of an organization. The term public service announcements or advertisements (PSAs) is commonly used to refer to public information campaigns, because most rely on television advertisemen ts for dissemination. Â“Other widely used channels and modes are radio spots, newspa per publicity, and pamphletsÂ”(Atkin, 2001, p. 26). PSAs are an integral part of health promotion campaigns (Andsager, Austin, & Pinkleton, 2001). Â“Over the past half-centur y, thousands of mass media campaigns have disseminated messages about dozens of different health topics to the U.S. populationÂ” (Atkin, 2001, p. 1). Campaigns to prevent smoking, reduce drunk driving, and encourage healthy eating habits represent the historical and typical uses of PSAs. In todayÂ’s health care environment, which stresses proactive h ealth behavior and the active involvement of
2 the healthcare consumer, health messages extend beyond promoting socially desirable behaviors to warning audiences about th eir risks for certain health conditions. Government agencies and health associat ions are the typical sponsors of health information campaigns. Â“Most campaigns have very limited monetary resourcesÂ” (Atkin, 2001, p. 27) and rely on Â“gratis placement in broadcast and print mediaÂ” (OÂ’Keefe, 1990, p. 67). Neither radio nor televi sion stations are now required by law to donate a specific amount of time to PSAs. But, as part of thei r mandate to prove they are operating in the public interest, broadcast stati ons have continued to provide free spots for public health messages (U.S. Department of Health a nd Human Services, 1992). There is no such incentive for print media, but some newspape rs and magazines provide free space as well (Atkin, 2001). However, competition for such spots is intense and PSAs are Â“ordinarily relegated to status behind re gular paid ads or commercials and are often apt to appear only as space and time become availableÂ” (O'K eefe & Reid, 1990, p. 68). In recent years, as access to free media placement has dimi nished significantly, Â“governmental and association sponsors of health campaigns have frequently relied on paid ads to gain more frequent and favorable cove rageÂ” (Atkin, 2001, p. 2). At the same time, the role of for-prof it groups as sponsors of public information campaigns is increasing. Liesse (1990) repor ted that the cause related marketing (CRM) efforts of commercial firms make it difficult to determine the difference between public service and corporate promotion. Â“CRM aligns brands with social causesÂ” and positions companies on a Â“social responsibility pl atformÂ” (Deshpande & Hitchon, 2002, p. 905) Corporations may fund the information campai gns of non-profit organizations or produce and disseminate their own public service me ssages. Â“A growing number of government
3 agencies are turning to corporate sponsors as a way to get across their public service messagesÂ” (Meyers, 1989, p. 22). As competition increases for donated media space, corporate funding may be essential for disseminating public service messages. Yet, there is some concern that CRM activities may do more than improve a companyÂ’s image and that some companies us e the guise of public service to increase profits (Liesse, 1990). This is a real con cern for health communication campaigns as pharmaceutical companies recently began sponsoring disease awareness campaigns about diseases directly linked to the companiesÂ’ product lines. Unlike direct-to-consumer (DTC) advertisements, which encourage consum ers with certain health conditions to ask their doctors about prescrib ing a specific brand of a drug, these messages encourage audiences to learn more about diagnosing a nd treating diseases they may or may not have. Though companies may not refer to th ese messages as public service campaigns, these Â“public service-type me ssagesÂ” (Liesse, 1990, p. 28) are indistinguishable from PSAs because they address a health issue w ithout linking it to a product or service. Disease mongering is the term used to de scribe marketing efforts designed to expand the market for products by convincing people they are sick and need medical intervention (Moynihan, Heath, & Henry, 2002). Fo r example, to expand the market for Viagra, Pfizer developed dis ease education messages encouraging men to talk to their doctors about erectile dysfunction. In 2003, GlaxoSmithKline launched a campaign to promote awareness about restless leg syndr ome (RLS). More recently, Merck funded a disease awareness campaign about the conn ection between the human papilloma virus (HPV) and cervical cancer only months before receiving approval to distribute the first HPV vaccine. While many may debate the ethi cs of such campaigns, a greater public
4 health concern is how these messages affect h ealth education, particularly when genuine health risks are involved. Do corporate sponsors as sources of pub lic service messages enhance or inhibit the effectiveness of health information cam paigns? This study seeks to answer this question by examining the relationship between audiencesÂ’ perceptions of corporate versus non-profit sponsors of a health message and their responses to the message. Lynn, Wyatt, Gaines, Pearce and Bergh (1978) argue that the audiencesÂ’ image of PSA sources is central to the issue of PSA effectiveness. This study uses concepts from the source credibility literatu re to investigate perceptions of the source and the message. In general, source effects have been well documented. However, research that looks spec ifically at the eff ects of organizational sources is just developing, a nd while source credibility resear ch is well established for specific messages such as news reports and consumer advertisements, it is an understudied area when it comes to public service announcements. More specifically, only two studies (Lynn, 1973; Lynn, Wyatt, Gain es, Pearce, & Bergh, 1978) look at the effect of organizational spons ors on audiencesÂ’ responses to PSAs. This study establishes a link between the developing area of organi zational source credibility research and research on the effects of public service messages. The experimental design for this study builds on research models that test organizational source credibility effects, prim arily from the area of consumer advertising. Because public service announcements advocate social issues rather than consumer behavior, the message effects c oncepts of brand attitude a nd purchase intentions from consumer marketing research do not apply. Thus, this article begins with a discussion of
5 the literature about the uses and effects of public information campaigns. Two main effects, information seeking and behavioral intention, have been offered as desirable outcomes of PSAs. The literature review continues with an overview of established source credibility concepts and key organizational credibility findings. Following the literature review, the theories that form the theoretical base for the study including two theories that may prove useful for understanding message acceptance for PSAs, the situational theory of publics (STP) and the theory of reasoned action (TORA), are discussed. The pr oblem recognition and person al involvement variables from the situational theory were used to operationalize message acceptance; however, there are strong parallels between these vari ables and the TORA attitude and subjective norms variables. Based on these parallels, th e study also suggests that the situational theory can be extended to include behavior al intention, and thus like the theory of reasoned action may also pr edict actual behavior. The study utilized a 2x2 experimental de sign plus a control group. Research participants were students in an intro ductory mass communications course at the University of South Florida. The treatments were health messages sponsored by a corporate health organization versus a non-profit health agen cy. Students in the control group viewed a health message without a sponsor. Based on current organizational credib ility findings, the research hypotheses argued that the corporate sponsor would be viewed as less credible as would messages attributed to the corporate sponsor. These lo wer credibility estima tes are expected to result in lower estimates of information seeking and intention to perform advocated behaviors. The studyÂ’s results show, however, th at organizational status does not have a
6 significant effect on perceptions of source credibility. Nor doe s it significantly influence message evaluation, information seeking, or be havioral intent. As predicted, there was a positive correlation between source credibilit y, message credibility, problem recognition, personal relevance, information seeking, and be havioral intent. The findings also indicate that information seeking positively predicts behavioral intent. The final sections of the paper discuss thes e results and their implications in light of existing knowledge about source credibility and public information campaigns, and provide suggestions for future research.
7 CHAPTER 2: LITERATURE REVIEW As with any communication campaign, pub lic information campaigns can have direct effects at the cognitive, affective, a nd behavioral levels. C ognitive effects involve developing issue awareness, promoting knowledge gains and skill acquisition. These campaign outcomes tend to be easier to achie ve. Affective responses include changed beliefs, values, and attitudes, increased perceptions of i nvolvement with the issue and behavioral intention. Compared to cognitions, a ffective responses are harder to obtain. As with other messages, behavioral outcomes are usually the desired effect for public service messages. Â“Behaviors can range from minor act ions to major practices; the latter is the gold standard that is most difficult to change and maintainÂ” (Atkin, 2001, p. 15). The Effects of Public Service Messages Early research about information campai gns suggests they have limited effects on attitudes and behavior and are likely to fail (Hyman & Sheatsley, 1947). Recent research shows that most public service campaigns ha ve limited direct eff ects on behavior. In a meta-analysis of 48 campaigns, Snyder (2001) found that most resulted in a 5 to 10% change in behavior. However, Mendelsohn (1973) argues that campaigns can succeed if communicators focus their objectives on what media messages can be expected to achieveÂ—significant increases in knowledge and awareness. Researchers have supported this view of public information campaigns (Borzekowski & Poussain t, 1999; Ledingham, 1993). Similarly, Grunig and Ipes (1983) sugge st that public communication campaigns
8 serve an agenda-setting function and accomplis h Â“little more than putting a problem on a personÂ’s personal agendaÂ” (p. 38). Based on their argument that the purpose of communication campaigns is to increase per ceptions that an issue is problematic and personally relevant to members of an audience, it is reasonable to expect that PSAs that inform target audiences of pot ential health risks can be eff ective. This agenda-setting function should not be overlooked because it can have an indirect effect on the development of health attitudes and behaviors. Two message effects that may lead to behavioral outcomes include information seeking at the cognitive level and behavioral intention at the affective level. Information seeking is an active communication behavior that involves the planned scanning of the environment for messages about a specifi c topic (Clarke & Kline, 1974). Awareness messages should facilitate information seeking by prompting Â“active seeking from elaborated information sources such as we b sites, hotline operators, books, counselors, parents, and opinion leadersÂ” (Atkin, 2001, p. 17). Information seeking can impact the change process indirectly and may eventually lead to behavioral outcomes by providing access to more extensive information that incorporates multiple appeals, elaborate evidence and detailed instru ctions. People communicating ac tively about a situation are more likely to engage in behavior to do something about it (Grunig, 1989). Behavioral intention can be described as a predisposition to respond in a given manner (Atkin, 2001). Several theories suggest that the intention to perform a behavior is the proximal determinant of volitional behavi or (Ajzen, 1991; Fishbein & Ajzen, 1975; Fishbein & Cappella, 2006). Â“Intentions are as sumed to capture the motivational factors that influence a behavior; they are indi cations of how hard people are willing
9 to try, of how much of an effort they are planning to exert, in order to perform the behaviorÂ” (Ajzen, 1991, p. 181). Th erefore, campaigns that can foster behavioral intentions in the audience are better able to achieve behavioral responses. Such messages may do so by modeling the desi red behavior, presenting reasons why the behavior is beneficial providing incentives, and giving in structions about how to carry out the action, particularly for complex behaviors. Source Credibility and Message Acceptance Â“Source effects refers to perceptions of sources that make them more or less influential,Â” (Miller & Levine, 1996, p. 262) Describing step six of the RASMICE1 procedure for constructing a persuasive communication campaign, Mc Guire (2001) stated that among other things, constructing the messa ge involves selecting a source that has Â“the greatest potential for eliciting the out putÂ…needed to achieve the desired health behaviorsÂ” (p. 23). The effect of the source on message acceptance is one of the oldest lines of communication research (Self, 1996). It is widely accepted that communication effectiveness is based, in pa rt, on who delivers the message. Credibility has long been regarded as an important characteristic to increase the persuasive power of a message source. WebsterÂ’s dictionary defines credibility as the quality or power of inspiring belief. In one of the earliest source credibility studies, Hovland and Weiss (1951) conc luded Â“the effect of an unt rustworthy communicator is to interfere with the acceptance of the materialÂ” (p. 647). Si milarly, Fishbein and Ajzen (1975) stated that source credibility aff ects the probability of message acceptance. 1 RASMICE stands for 1. Reviewing the realities, 2. Axiological analysis, 3. Survey the sociocultural situation 4. Mapping the mental matrix, 5. Teasing out the target themes, 6. Constructing the communication, and 7. Evaluating effectiveness (McGuire, 1984a).
10 Researchers have established that individua ls are more likely to accept messages from highly credible sources, while message accep tance is less likely to occur with low credibility sources. The Organization as Source Many credibility studies treat the spokes person featured in a message as the source (Atkin & Block, 1983; Swartz, 1984; Ho mer & Kahle, 1990; Perse, Nathanson & McLeod, 1996; Yoon, Kim & Kim, 1998). Thes e studies focus on spokesperson or endorser credibility by examining attractiveness, (i.e. familiarity, similarity, and liking); expertise; and believability. Spokesperson studie s often examine the use of celebrities to endorse messages. Some studies have begun to investigate the role of the message sponsor as the source (Goldberg & Hartwick, 1990; Goldsmith, Lafferty, & Newell, 2000a, 2000b; Haley, 1996; Hammond, 1986; Lafferty & Go ldsmith, 1999; Lafferty, Goldsmith, & Newell, 2002; Newell & Goldsmith, 2001; Re id, Soley, & Vanden Bergh, 1981). Stern (1994) proposed a revised communication m odel for advertising that recognizes the multidimensionality of the source element. A ccording to Stern, the Â“without-textÂ” source of an advertisement is dual, reflecting the existence of a financial source and a creative source, both of which are distinguished from the Â“within-textÂ” s ourceÂ—the persona or communicator presented in the advertisement. The financial source is referred to as the sponsor. The sponsorÂ’s Â“communicative res ponsibilities include commissioning the ad, paying for it, approving it, and be ing held legally liable for what is in the text. It is the sponsorÂ’s name that permeates advertis ingÂ” (p. 8). Message sponsors are often organizations that pay for or initia te message production and dissemination.
11 With the exception of early studies by Lynn (1973; Lynn, Wyatt, Gaines, Pearce, & Bergh, 1978), studies on source effects fo r PSAs have focused on the spokespersons presented in advertisements rath er than examining the role of the sponsor as the source of a message. Lynn and colleagues (1978) identif ied several PSA source types: the private firm or profit making organization, the char itable organization, th e non-profit institution, the governmental agency, and the Advertising Council. Using factor analysis, these five organizational sources were collapsed to th ree factors: the commercial source, the noncommercial source, which included charitable organizations, non-profit institutions, and governmental agencies, and the Advertising Co uncil. This study did not investigate how participantsÂ’ perceptions of source credibi lity varied across the three source groupings. Lynn (1973) found that experimental groups could not distinguish between the Advertising Council and a tr aditional advertiser. Lynn et al. (1978) sought to investigate whether di fferent sources produced distinguishable evaluative and behavioral responses to PSA messages. The findings indicated that messages attributed to comme rcial sources had the second highest message evaluation scores of the thr ee types of organizational sour ces studied. Source attribution did not Â“adequately explain variability in behavioral responsesÂ” (p. 720). However, both the message evaluation and behavioral res ponse hypotheses were va guely described, as were methods used to measure these res ponses. Because behavior was defined as participants indicating a commitment to issu es presented in the PSA and the study did not systematically monitor such responses, it reveal ed little about behavior al responses. It is difficult to draw conclusions about the rela tionship between sponsor type and message effects based on the result of this one study.
12 Sponsor credibility is studied most in the areas of advertising and consumer marketing. This line of research, which is also more recent, offers a more definitive understanding of the relationship between orga nizational sponsors and message effects. Â“As with spokesperson credibility, companies wi th positive reputations would seem to be in a better position to get consumers to be lieve their advertising claimsÂ” (Goldberg & Hartwick, 1990, p. 173). The terms company reputation, corporate im age, advertiser credibility, attitude toward the advertiser, and corpor ate credibility have been used interchangeably to refer to the credibility of the orga nizational source. Goldsmith, Lafferty, and Newell (2000a) argued that credibility is only one component, albeit a cr itical component, of corporate reputation because reputation is the overall impression of the company. Thus, the terms advertiser or corporate credibility are used fr equently to refer to organizational sources in credibility studies. Advertiser credibility has been defined as the pe rceived truthfulness or honesty of the sponsor of an ad (MacKenzie & Lutz, 1989). Similarly, corporate credibility refers to stakeholder percep tions of a companyÂ’s trustworthiness and expertiseÂ—the believability of its intentions and communications at a particular moment in time (Goldsmith, Lafferty, & Newell, 2000b). In one of the first studies to systemat ically manipulate advertiser credibility, Goldberg and Hartwick (1990) hypothesized th at companies with positive reputations would be in a better position to get consumers to believe their advertising claims. They found participants in the negativ e sponsor reputation treatment rated the credib ility of the advertisement presented more strongly nega tive than did participants in the positive reputation treatment.
13 Lafferty, Goldsmith and NewellÂ’s (2002) dual credibility model (DCM) is a response to SternÂ’s (1994) argum ent that Â“it is necessary to investigate credibility as a bundle of effects flowing from differen t source componentsÂ…Â” (p. 12). The model explains the main and interacting effects of corporate credibility and endorser credibility on advertisement outcomes. DCM claims that corporate credibility is positively and directly related to attitude-toward-the-a d, attitude-toward-the-brand, and purchase intentions. These claims are supported by a study that tested the model (Lafferty, Goldsmith, & Newell, 2002) and ea rlier studies that investigated the main and interacting effects of corporate credibility and endorse r credibility on attitude-toward-the ad and purchase intentions (Goldsmith, Laffert y, & Newell, 2000a, 2000b). Lafferty and Goldsmith (1999) found that corporate credibil ity is independent of endorser credibility, is positively related to attitude-toward-thead, and appears to have a greater impact on attitude-toward-the-brand a nd on purchase intentions than endorser credibility. Non-profit versus Commercial Sponsors The results of several studies indica te government and non-profit sources are perceived to be more credible than fo r-profit organizational sources (Haley, 1996; Hammond, 1986; Lynn et al, 1978). In the area of health communication, major medical institutions and physicians t ypically are viewed as more credible sources of health information (Christensen, Ascione, & Ba gozzi, 1997; Cline & Engel, 1991; DuttaBergman, 2003; Frewer, Howard, Hedderle y, & Shepherd, 1996). Reid, Soley, and Vanden Bergh (1981) found that participants vi ew advocacy advertisements that present a commercially sponsored point of view more negatively than advertisements sponsored by a noncommercial source or no source. Their re sults also indicated that subjects Â“are
14 strongly disinclined to respond to a request a dvocated by a commercial source, but not to the same request advocated by a non-commerc ial sourceÂ” (p. 315). The authors suggested that commercial sources are viewed as less ob jective and having something to gain. This conclusion is similar to Hovland, Janis, and KelleyÂ’s (1953) hypothesis that when a person is perceived as having a definite inte ntion to persuade others, the likelihood is increased that he or she will be perceived as having something to gain and as less trustworthy. Based on this same hypothesis, Hammond (1986) investigated the credibility of organizations that advertise a bout health issues to determine if differences exist in the credibility of an organization when it is pe rceived as having something to gain from the advertisement. The study was designed to examine the effect when for-profit organizations use health information to suppor t their advertising clai ms. It differentiated between organizations that conduc t social advertising solely for corporate public relations reasons and organizations that make a dire ct profit from persuading customers through social advertising to adopt an advocated hea lth behavior. The results indicated that the non-profit and combination non-pr ofit/for-profit sources were perceived as significantly more credible than a for-profit source alone Although no significant relationship between source credibility and message acceptance wa s found, source credibility did have an affect on behavioral intention. The non-profit or combination sources were more effective in producing an intention on the part of the re spondent to change his or her behavior. Hammond (1986), Reid et al (1981), and Hovland, Jani s, and KelleyÂ’s (1953) findings suggest that certain organizational sponsors may not be the best sources of health messages if they cause audiences to respond passively, or even discount the
15 message. Haley (1996) argued th at an organizational sponsor may be a viewed as a credible source for some issues and not credible in relation to other issues. Measuring Sponsor Credibility It is often confusing to understand and de fine source credibility because of Â“the many operationalizations that appear in th e literatureÂ” (Ohanian, 1990, p. 41). Reviewing eight studies of endorser credibility in the advertising, marketing, and speech communication fields, Ohanian identified 16 di fferent dimensions that were used to measure credibility. These included: trustw orthiness, expertness, dynamism, objectivity, safety, qualification, competence, attractiveness, lik eability, evaluative, potency, activity, authoritativeness, character, be lievability, and sociability. E ach scale combined seven or less of these dimensions. Perhaps this multiplicity of dimensions reflects the agreement among theorists that credibility is a multidim ensional construct and the disagreement or uncertainty about what those dimensions are. Ohanian noted that only one of these scales was assessed for reliability and validity. The issue of measuring sour ce credibility is confounded more when the source is no longer defined as the Â“with-in textÂ” communi cator. For instance, in the field of mass communications the source concept represents the medium that disseminates a message such as a newspaper, radio or television st ation, magazine or even a website. In the present study, the source is con ceptualized as the organizatio n that sponsors the message. In an exploratory study of the organiza tion as source, Haley (1996) found that consumers evaluate organizational spons ors on a number of factors including recognizability, quality of product or servic e, history of pro-social involvement, congruency with personal values logical association with the issue, personal investment
16 in the issue, and intent. Caruana (1997) found similar factors contribut e to perceptions of corporate reputation. Supporting OhanianÂ’s ( 1990) claims that three credibility dimensions are enduring, Haley concluded that this variety of elements confirms, rather than adds to, the three primary credibility dimensions: expertise, trustworthiness, and attractiveness. Similarly, McCroskey and Y oung (1981) argued that despite attempts to find new dimensions with which to measure s ource credibility, particularly for different types of sources, the construct ha s been Â“amply definedÂ” (p.34). However, contrary to McCrosky and YoungÂ’ s perspective that existing credibility scales provide adequate measures, requiring on ly minor modifications for different types of sources, Haley (1996) suggested that ther e are unique aspects of the organizational credibility construct and that a separate scal e is required to measur e it. Likewise, other researchers have concluded that not all di mensions of scales intended to measure endorser credibility apply to organizati onal credibility. For instance, Newell and Goldsmith (2001) argued that the dimensions of attractiveness and likeability, while suitable for a persona, Â“would not charac terize corporate credibilityÂ” (p. 235). Several scales have been used in studie s that include the advertiser credibility construct (Goldberg & Hartwick, 1990; LaBarbera, 1982; MacKenzie & Lutz, 1989; Muehling, 1987; Settle & Golden, 1974). Agr eeing with HaleyÂ’s ( 1996) critique that organization credibility research has been ha mpered by the lack of a validated scale, Newell and Goldsmith (2001) proposed a scale to measure perceived corporate credibility based on the dimensions of expertise and trustworthiness from Hovland, Janis, and KelleyÂ’s (1953) source-credib ility model. Hovland and coll eagues defined expertise as the extent to which a communicat or is perceived capable of making correct assertions and
17 trustworthiness as the degree to which the audience perceives assertions made by the communicator to be valid. Through a six-pha se process, Newell and Goldsmith (2001) validated an eight-item scale with four item s to measure the expertise factor and four items to measure the trustwor thiness/truthfulness factor. Audience Perceptions of Message Sponsors Despite the use of objective measures to assess credibility, researchers agree that credibility is not an inherent trait of the source, but rather a perception of the receiver. Many researchers define and ex amine credibility as the recei verÂ’s response to the source (Berlo, Lemert, & Mertz, 1969; Gunther, 1992) As such, organizati onal credibility is situational, dependent upon ch aracteristics of the communica tion context. Some factors that may influence audience interpretations of the sponsorÂ’s credibility include knowledge of the source, existing attitudes to ward the source, and evaluation of the sourceÂ’s intent. In order for receivers to ev aluate sponsor credibility, they must first identify the organizational source of the message. The pr ominence of source iden tification may affect audience evaluations of source credibility. In several organizational credibility studies, participants were provided with descriptions of the organi zation that not only identified the sponsor but also induced perceptions of high or low credibility (Lafferty, Goldsmith & Newell, 2002; Lafferty & Goldsmith, 1999; Goldberg & Hartwick, 1990). Some endorser credibility researchers have focuse d on the relationship between the timing of identification for high versus low credibility sources and source effects (Homer & Kahle, 1990; Sternthal, Dholakia, & Leavitt, 1978; Ward & McGinnies, 1974). The findings of these studies indicate that th e persuasability of low credib ility sources may be increased
18 by delaying identification. For hi gh credibility sources, identif ication before the message appears to increase the sources persuasivene ss. However, delayed identification of high credibility sources does not appear to have a significant effect. Sternthal, Phillips, and Dholakia (1978) suggested that in some instances delayed id entification of high credibility sources may reduce persuasive ness whereas in other instances a high credibility source is equally persuasive before or after the message. Reid, Soley, and Vanden Berg (1981) found that open identificat ion of a commercial source for advocacy advertisements was related to low perceptions of the message. Thus, when identified, a commercial source had less of a persuasive effect. An important factor in the effectiveness of communication is th e attitude of the audience toward the communicat or (Hovland and Weiss, 1951). The receiversÂ’ level of awareness of the message sponsor may trigge r various reactions to the source based on their existing attitudes toward the organization. Â“Just as firms have a multitude of publics, they also have an array of reputations as each public often considers a different set of attributes. Moreover, even if the same attr ibute is considered by different publics it may be given a different weightingÂ” (Caruana 1997, p. 110). These attitudes may be based on direct experiences, such us purchasing produc ts or utilizing services, or indirect experiences, such as news reports and the pr aise or complaints of others, with the organization. These attitudes may also be tr ansferred from experiences with similar organizations or impressions about the categ ory of organizations to which it belongs. Receivers may also determine the credibility of an organization and the believability of the message based on their evaluation of the sponso rÂ’s intent. One scale used to measure the credibility of media organizations (Gaziano & McGrath, 1986) asked
19 participants to indicate whet her the organization was concer ned about the public interest or about making profits. This question is also applicable to for-prof it organizations that promote health information campaigns. Hovl and, Janis, and Kelle y (1953) indicate a source is seen as less trustworthy when the audience perceives the source has something to gain. Walster, Aronson, and Abrahams ( 1966) found that the persuasiveness of a low credibility source was increased when the source advocat ed a position incongruent to its own interests. Sponsor Credibility and PSA Effects: Inform ation Seeking versus Behavioral Intent Much of the source credibility research focuses on the effect the source has on eliciting the behaviors advocat ed by a message. In advertising and consumer marketing, the desired message effect is purchase beha vior. Message acceptance has been shown to mediate the relationship between source credib ility and behavioral intentions. That is, source credibility has a positive and direct effect on message acceptance. Greater message acceptance is positively related to increased purchase intentions (Lafferty & Goldsmith, 1999; Goldsmith, Lafferty & Newell, 2000a, 2000b). Message acceptance has been measured as attitude-toward-the-ad, whether positive or negative, and as ad credibility, the extent to which a consumer perceives the claims made about a given brand to be truthful (Mackenzie & Lutz, 1989). Similar concepts can be applied to message acceptance for public service announcements. Therefore, source credibility can be seen as having a positive and direct effect on favorable or unfavorable responses to the PSA, or perceptions that claims made about a given issue are valid. Studying PSA perceptions, Lynn (1973) used the term message evaluation and measured it using ag reement/disagreement. Lynn and colleagues
20 (1978) suggest that different types of s ponsors may produce distinguishable evaluative and behavioral responses to PSA messages. As with product adve rtisements, greater message acceptance may also be expected to affect behavioral outcomes for PSAs. These outcomes may be changes in social or health -related behaviors. Ho wever, debates about the effectiveness of PSAs suggest that they are not effective at mo tivating behavior but are more likely to motivate information seeking (Atkin, 2001; Grunig & Ipes, 1983; Mendelsohn, 1973). Whether PSAs are more eff ective at motivating overt behavior or information seeking behavior has not been studied empirically.
21 CHAPTER 3: THEORETICAL FRAMEWORK This study examines how audience percepti ons of the credibili ty of organization sources affect responses to messages these organizations sponsor. The study applies existing knowledge of source cr edibility as an information processing cue to investigate the relationship between the source and messa ge effects. Unlike most organizational credibility studies, which examine consumer advertising, this study looks at a health information message. As such, message acceptance is discussed in terms of problem recognition and personal involvement, which are expected to lead to message effects, namely information seeking. Prior research involving public service messages indicate these variables are likely outcomes for in formation campaigns. Health information campaigns often seek to modify behavior as we ll. Thus investigating behavioral intent as a message effect is also appropriate. This chapter discusses the theoretical founda tions that were used to formulate the research questions and hypotheses described in Chapter 4. It begins with a brief overview of information processing theory as it appl ies to organizational source credibility. It continues with a discussion of the situati onal theory of publics (Grunig & Hunt, 1984), which predicts a relationship between probl em recognition, personal involvement, and information seeking. The study incorporates situational theory to investigate the relationship between source credibility and th eses variables. The third-person effect theory is used in conjunction with GrunigÂ’s s ituational theory to examine perceptions of
22 involvement as an outcome of message exposure. The chapter concludes with a discussion of the theory of reasoned action (Fishbein & Ajzen, 1975), which argues that behavioral intent is a strong predictor of actual behavior. This theory provides justification for the inclusion of behavior ques tions in message effects studies in general. Source Credibility and Information Processing Source credibility studies ar e often based on information processing theories such as the Elaboration Likelihood Model (ELM) of Persuasion (Petty & Cacioppo, 1986) and the heuristic/systematic processing model of persuasion (Chaiken, Liberman, & Eagly, 1989), which position source characteristics as a cue that audiences use to evaluate messages. Systematic or central processing is active and involves careful examination of the message and the arguments presented. Heuris tic or peripheral processing is simplistic and focuses on non-content aspects of the message. As a non-content element, source credibility often is considered a peripheral cue leading to low elaboration or passive/heu ristic processing of a message. In some instances, source credibility may function as a central cue and can trigger high elaboration or systematic pr ocessing of both the message and the source. Mackenzie and Lutz (1989) indicate that advert iser credibility is more of a central processing cue rather than a peripheral processing cue. This im plies that knowledge of an organizational sponsor will lead respondents to examine caref ully the credibility of the source and the validity of the message. Lafferty and Goldsmith (1999) argue that this active processing makes advertiser credibility more influen tial on evaluations of and responses to the message. Although the current study does not test information-processing theory, it incorporates concepts from organizational cr edibility models based on the tenet that
23 organizational credibility is a central processi ng cue that leads individuals to evaluate a message systematically. Situational Theory of Publics Grunig and Ipes (1983) argue that communi cation campaigns function to increase perceptions that an issue is problematic a nd personally relevant to members of an audience. This links the study of PSA e ffects to GrunigÂ’s (Grunig & Hunt, 1984) situational theory of publics (STP). Gr unig used problem recognition, level of involvement, and constraint recognition as i ndependent variables to predict whether a public will engage in information-seeking or information-processing behavior. The independent variables of problem recognition an d level of involvement are of particular importance to this study. Although the situationa l theory is intended to help understand the communication behaviors of publics by meas uring how they percei ve situations and also is used to differentiate between activ e and passive publics, these two independent variables also represent message acceptance for PSA campaigns according to Grunig and IpesÂ’ (1983) rationale. Situational theory also includes constraint r ecognition as its third independent variable. However, constraint recognition is not considered in this study. Although Grunig and Ipes discussed the ro le PSAs play in reducing constraint recognition, message acceptance is conceptualized in this study based on the agendasetting functionÂ—that an issue is a problem that is personally relevant. Thus, the constraint recognition variable does not fit th is studyÂ’s definition of message acceptance. Additionally, Grunig and Hunt (1984) proposed the variable to account for factors that hinder activist groups from forming.
24 Problem recognition is the extent to whic h individuals perceive that a situation has consequence for them and detect a problem in the situation. Involvement is the extent to which a problem or situation has personal relevance to an individual (Grunig & Hunt, 1984). Effective communication campaigns should increase individualsÂ’ problem recognition and perceptions of persona l relevance (Grunig & Ipes, 1983). The situational theory predicts that higher levels of problem recognition and involvement lead individuals to both proce ss and seek information. Unlike information processing which involves passive, unplanne d discovery of messages, information seeking is active and involves planned sca nning of the environment for messages about a specific topic. Public service announcemen ts can be viewed as an informative communication strategy. Informative strategi es (Zaltman & Duncan, 1977) are effective in creating problem recognition and work best when immediate behavioral changes are not required. Atkin (2000) refers to thes e as awareness messages, which, among other things, can be used to create recognition of a topic or practice, and to encourage further information seeking about the topic. Â“A ke y role of awareness messages is to arouse interest or concern and to motivate further exploration of the subj ect. In particular, messages should include elements designed to prompt active seekingÂ” (Atkin, 2000, p. 56). From this perspective, information seek ing is the desired be havioral outcome of communication campaigns. Greater message acceptance through increased problem recognition and involvement should lead to greater information seeking. Involvement and Third-Person Perception If public information campaigns, including public service announcements, function to increase perceptions of personal involvement, then it is ideal to measure
25 involvement as one form of message acceptan ce by determining if th e audience views the issue as personally relevant. Third-person effect (TPE) theory (Davison, 1983), posits that people perceive media messages to have greater effect on others than on themselves. This is known as third-person perception. G unther and Thorson (1992) were the first to study third-person perception using produc t advertisements and public service announcements. They found that public servi ce announcements produced larger positive effect estimates for self and others than both neutral and emotiona l product ads. This suggests that the direction of third-person perception is influenced by consideration of the messageÂ’s intent (Gunther & Thorson, 1992, p. 592). PSAs typically are viewed as altruistic messages. Therefore, it may be c onsidered desirable to be persuaded by such messages. There is some evidence that socially desirable messages produce smaller estimates of third-person percepti on (Eveland & McLeod, 1999; Lambe & McLeod, 2005). Other studies have found that public service messages, unlike most messages, produce greater perceptions of impact on self (Duck & Mullin, 1995; Duck, Terry, & Hogg, 1995). TPE studies ask participants to estimate media effects on themselves, referred to as first-person perception (FPP), and various groups of others. Higher TPP scores indicate that an individual does not see an issue as pers onally relevant, whereas higher FPP scores (often referred to as reversed TPP) indicate that the indi vidual views the issue as personally relevant. Use of TPP and FPP to measure involvement differs significantly from the standard approach used in situa tional theory studies, which typically ask respondents to indicate the degree to which they feel concerne d with or connected to an issue. While this may indicate level of invol vement, it does not position that involvement
26 as an outcome of the message. Asking participan ts to indicate the effect of the message on themselves and others allows the involvement variable to be posit ioned as a factor of message acceptance. Although it is known that some message s produce higher TPP scores, while others like PSAÂ’s may lower the TPP estim ate, few studies have looked at the relationship between source credibility a nd TPP estimates. White (1997) points to perceived source bias as a predictor of th e third-person effect. Studying defamation in news stories, Cohen, Mutz, Price, and Gunthe r (1988) found that as pe rceptions of source bias increase, estimates of persuasive impact on others versus self also increase. The relationship between TPP and perceived source bias for organizational sponsors of PSAs has not been studied. Theory of Reasoned Action Message effects studies frequently ask participants to indicate their intention to perform advocated behaviors. Hammond (1986) found a significant relationship between organizational credibility and behavioral intention. Although media effects studies often are criticized for measuring intent to act (i.e. purchase intention) rather than actual behavior, actual behavior is often difficult to measure. According to the theory of reasoned action (Fishbein & Ajzen, 1975), behavior al intention is a powerful predictor of actual behavior. Thus, measuring behavioral intent is suffici ent. The theory of reasoned action (TRA) posits that intentions are the im mediate antecedents to behavior and provide valid indications of how hard people are willing to try to pe rform a behavior. The greater an intention, the more likely it is the given behavior will be performed.
27 Behavioral intention is defined as a function of attitude and normative beliefs. Thus, messages can affect behavioral intenti on and the resulting be havior by influencing attitudes towards the behavior and beliefs a bout subjective norms regarding the behavior. The attitude variable encompasses personal feelings, whether negative or positive, about performing the behavior. Subjective norms refer to impressions of social pressures, that is what important others think about performing the behavior and whether or not important others would perform the behavior. As atti tudes toward the behavior become more positive and beliefs about subjective norms beco me stronger, the greater an individualÂ’s intention to perform the behavi or and the more likely the be havior will be performed. Despite agreement among some research ers that public information campaigns are more successful at fulfilling knowledge object ives rather than attitude or behavior objectives, this assumption has not been test ed. Furthermore, in addition to awareness objectives, public service announcements often include behavioral ob jectives, such as encouraging motorists to wear seatbelts or to abstain from drinking and driving. There is no evidence that such messages produce in formation seeking but not behavioral outcomes or that they are more effective at producing one but not the other. Therefore, it is important to include the behavioral intent variable. Doing so, adds another dimension to the studyÂ’s purposeÂ—testing the hypothesi s that public information campaigns are more effective at producing awareness be haviors rather than actual behavior. Although the situational theo ry (Grunig & Hunt, 1984) predicts that greater problem recognition and personal involvement lead to information seeking, it does not state that behavior is not a possible outcome of the two variables. Indeed, some parallels can be drawn between the att itude and subjective norms vari ables from the theory of
28 reasoned action and the situ ational theoryÂ’s personal involvement and problem recognition variables. Personal involvement in STP, like attitude in TRA, involves personal evaluations of an issue. Though problem recognition also involves personal assessment, like subjective norms, it may also take into consideration the relationship between the issue and referent others. In a refinement of the situ ational theory, Grunig (1997) discusses the external co mponents of problem recognition. Numerous health communication studies have tested TRA and the theory has been extended to create new theo ries such as the theory of planned behavior (Ajzen, 1985) and the integrative model (Fis hbein, 2000; Fishbein & Cappella 2006). Rather than test TRA, this study posits that problem recogni tion and personal involvement may not only lead to information seeking but may also have an effect on behavioral intention similar to the relationship between att itudes, subjective norms and behavioral intention as demonstrated by the theory of reasoned action.
29 CHAPTER 4: RESEARCH QUESTIONS AND HYPOTHESES This study was designed to examine the eff ect of different organizational sponsors on responses to public service messages. Exis ting research suggests a direct relationship between sponsor credibility ratings, percep tions of message cred ibility, and message effects. However, the relationship among these variables has not been examined adequately using public service announcements. According to Grunig and Ipes (1983), problem recognition and perceptions of personal relevance represent message accep tance for public information campaigns. Applying situational theory, which predic ts that greater problem recognition and perceptions of personal involvement lead to information seeking, information seeking behavior can be viewed as one possible effect of public service messages. Figure 1 shows the hypothesized relationships among source credibility, message credibility, message acceptanceÂ—represented by problem recogni tion and personal relevanceÂ—and information seeking.
30 Figure 1. Relationships among source credibility, message credibility, message acceptance, and information seeking. Based on the proposition, stated earlier, that greater problem recognition and perceptions of personal involvement may lead to behavior, this study also examines the effect of source credibility on behavioral outcomes. Source effect studies often ask participants to indicate their intent to perf orm a behavior, typically a purchase behavior. Problem Recognition Personal Involvement Message Credibility Source Credibility Information Seeking
31 According to the theory of r easoned action, intent to act is a valid indicator of actual behavior. Although some researchers argue that information campaigns are successful at motivating knowledge behavior rather than ac tual behavior, this has not been tested. Figure 2 shows the hypothesized relationship among source credibility, message credibility, message acceptanceÂ—represente d by problem recognition and personal relevanceÂ—and behavioral intent. Figure 2. Relationships among sou rce credibility, mess age credibility, message acceptance, and behavioral intent. Problem Recognition Personal Involvement Message Credibility Source Credibility Behavioral Intent
32 The following research questions and hypothe ses further explain the variables and relationships represente d in the above figures. Organizational Status, Source Cred ibility, and Message Credibility RQ1: Are audiences aware of th e for-profit or not-for-profit status of PSA message sponsors and how does this aw areness affect perceptions of source credibility and message credibility? Organizational credibility studies provid e evidence that for-profit sponsors are rated more negatively than non-profit sponsor s. The prominence of source identification may also impact audience evaluations of cred ibility. Rather than identify the corporate source by a logo alone, Lafferty, Goldsmith and Newell (2002; Lafferty & Goldsmith, 1999; Goldberg & Hartwick, 1990) provided partic ipants with a background sketch of the organization. In this study, a detailed desc ription was included on two stimuli as a means of openly identifying the sour ce as a corporate or non-prof it sponsor. Reid, Soley, and Vanden Bergh (1981) found that openly id entifying a commercial source produced negative evaluations of an advocacy advertisement. However, the same message attributed to a non-commercial source or no source was evaluated more positively. Previous organizational credibility studies found that source credibility was directly related to evaluations of message credibility. Thus, H1a: For-profit sponsors will be rated as less credible th an not-for-profit sponsors. H1b : Participants in the detailed-iden tification for-profit treatment group will report the lowest source credibility scores. H1c: There will be a positive and direct correlation between participantsÂ’ estimates of sponsor cred ibility and their estimates of message credibility.
33 Source Credibility, Problem Recognition, and Personal Involvement RQ2: Is message acceptance different for for-p rofit message sponsors than for non-profit message sponsors? This study tested the precep t that public communicati on campaigns are primarily successful at placing problems on an individu alÂ’s personal agenda (Grunig & Ipes, 1983). Based on this perspective, message acceptance is viewed here as pe rceptions of problem recognition and personal relevance or invol vement. In this study, the first-person perception and third-person perception scales were used to measure involvement. Higher FPP scores would indicate great er estimates of personal impact. The magnitude of source credibility estimates as it related to great er awareness of the source was expected to produce greater message acceptance. H2a: There will be a direct, positive correla tion between participantÂ’s estimates of sponsor credibility and es timates of problem recognition. H2b : There will be a positive correlation be tween higher source credibility ratings and estimates of personal relevance. H2c : Participants in the detailed for-profit treatment will report the least problem recognition. H2d : Participants in the detailed for-profit treatment will report the lowest levels of personal relevance. Source Credibility, Information Seeking, and Behavioral Intent RQ3: Is there a difference in information se eking for for-profit message sponsors than for non-profit message sponsors?
34 Public communication campaigns may trigge r indirect behavior al outcomes, such as information seeking. The situational theo ry of publics predicts that when problem recognition and level of involvement are high information seeking behavior will occur. H3a : Higher levels of problem recognition wi ll correlate with greater intention to seek information. H3b : Higher levels of involvement will correlate with greater intention to seek information. H3c : Participants in the detailed for-prof it treatment will report lesser intent to seek information. RQ4: Is there a difference in be havioral intent for for-prof it message sponsors than for non-profit message sponsors? Rather than assume that information seek ing is the only effect of public service announcements and that problem recognition and personal involvement do not lead to behavior, this study also investigated partic ipantsÂ’ intentions to perform the health behaviors advocated by the message. The sour ce credibility model pr esented in Figure 2 indicates a possible relationship between pr oblem recognition, pers onal involvement, and behavioral intent. H4a : Higher levels of problem recognition wi ll correlate with greater behavioral intention. H4b : Higher levels of involvement will correlate with greater behavioral intention H4c : Participants in the detailed for-prof it treatment will report lower behavioral intent.
35 Information Seeking versus Behavioral Intent RQ5: Are there differences in participantsÂ’ wi llingness to engage in information seeking and their intent to perform behaviors advocated by the message? H5a : Participants across all treatment groups will repor t greater intention to carryout information seeking behavior than expressing intent to perform advocated health behaviors. RQ6: Which of the factors examined in this st udy contribute most to behavioral intent? The theory of reasoned action identifies attitude toward the action and subjective norms as predictors of behavioral intent. This study suggests that there are parallels between these two variables and the pers onal involvement and problem recognition variables from situational th eory. In general, this study hypothesizes that the sponsors organizational status influen ces behavioral outcomes. Sin ce overt behavioral responses are the often the desired effect for public serv ice messages, identifying the factors that are most likely to produce behavior has important implications for message effectiveness.
36 CHAPTER 5: METHODOLOGY This experiment used a 2X2 factorial design with a control group based on the manipulation of the independe nt variables organizational source and level of source identification (Table 1). The first indepe ndent variable, organizational source, was operationalized using a for-profit pharmaceu tical company versus a non-profit health agency as the sources. The second independent variable, level of source identification, was operationalized as subtle identificati on, use of each organizationÂ’s logo alone, and detailed identification, use of the logo along with a brief descripti on of the organization and references to the organization in the body of the brochure. The detailed identification concept is based on Reid, Soley, and Va nden BerghÂ’s (1981) discussion of open identification of organizational sponsors and is intended to increase the participantsÂ’ level of awareness of the for-profit (FP) or non-profit (NP) sponsor. Table 1. Experimental Treatments For-profit Non-profit Subtle Id Subtle FP Subtle NP Detail Id Detailed FP Detailed NP Study Participants The studyÂ’s participants were 381 unde rgraduate students enrolled in an introductory mass communications course at the University of South Florida. The course is required of students entering mass communi cations courses but is also open to the general student body as an el ective course option fulfilli ng the universityÂ’s general
37 education requirements. Thus, the sample was not randomly selected. However, the fact that the course is open to all students makes it more likely th at the sample is diverse in terms of student majors. Stimulus Materials The stimulus for each treatment was a tri-fold brochure (see Appendix B). The brochureÂ’s copy discussed the risks of and complications associated with the human papilloma virus (HPV). This topic represen ts a recent health awareness campaign that was sponsored by both for-profit and not-for-p rofit health organizations. The stimulus material was adapted from brochures crea ted by the Centers for Disease Control for message testing. Key phrases from actual messages produced by a pharmaceutical company and a non-profit company were included to reflect the message tones used in the information campaigns of both groups. In terms of content and layout, each br ochure was exactly the same with the exception of the source identification. The back panel of each brochure displayed the logo of the pharmaceutical company, for the two FP treatment groups, or the logo of the non-profit health ag ency for the two NP treatment groups. The pharmaceutical company Merck was chosen as the for-profit source and the American Cancer Society was selected as the non-profit source. For the subtle iden tification treatment groups, the logo appeared smaller than the logo used in the detailed identificati on treatment. The detailed identification stimuli also included a small l ogo on the front panel and a short description of the organization on the back panel. Each description was approximately 20 words and appeared below the organizationÂ’s logo in the same font and size. Th e descriptions were written objectively using a neutral tone. In addition, the organizations name appeared
38 several times throughout the body of the pamphl et. The message for the control group did not contain logos or any othe r organizational identifiers. Manipulation Check A manipulation check was conducted to c onfirm that the subtle identification treatments and the detailed identification treatments actually differed in terms of level of source awareness. The manipulation check meas ured the degree to which subjects were aware of the sponsor and the sponsorÂ’s stat us as a for-profit or non-profit organization under each treatment condition. Students from an introductory public relations class and an upper level public relations research class participated in the manipulation check (N = 75). The students viewed one of the five s timuli and then responded to four items that measured their level of awareness of the source (see Appendix C). The participants were asked to indicate the level to which they recalled the organizationÂ’s name, and to rate how easily they were able to identify the sponsor. A oneway analysis of variance (ANOVA) showed that differences in the level of source identification had a significant effect on both level of recall (F(4,58) = 2.66, p = .041) and ease of identification (F(4,60) = 2.482, p = .053). Participants reported higher levels of recall in the detailed forprofit (M = 4.07) and non-prof it treatment (M = 3.29) groups than did those in the subtle for-profit (M = 2.78) and non-profit groups (M = 3.08) and the control group (M = 2.08). Ease of identifi cation was also greater for participants in the for-profit (M = 4) and nonprofit treatment (M = 3.08) groups than for those in the subtle for-profit (M = 2.58) and non-profit groups (M = 2.93) and the control group (M = 2.27).
39 Measurement Apparatus The measurement instrument for this expe riment (see Appendix C) consisted of a 39-item questionnaire that included questions about the participants Â’ perception of source credibility, their per ception of message cred ibility, their level of problem recognition, their estimations of the messageÂ’s influen ce on themselves and on specified groups of others, their intention to s eek information, and their inte ntion to perform behaviors advocated in the message. The questionnaire so licited demographic information as well. Source Credibility. A nine-item, five-point Likert scale anchored by the terms strongly agree/strongly disagree was used to measure source cr edibility. Participants were also able to indicate if they had no opinion about source credibility. Three items measured trustworthiness (e.g. The organiza tion can be trusted to provide factual information) and three items measured expe rtise (e.g. The organiza tion is qualified to provide information about this issue). Hovl and, Janis, and Kelley (1953) operationalized source credibility using the dimensions of trustworthiness and expertise and numerous studies have illustrated the endurance of thes e dimensions to measure credibility. Newell and Goldsmith (2001) validated an eight-ite m, Likert scale to measure corporate credibility based on the dimensions of trustwor thiness/truthfulness and expertise. There is also evidence that consideration of the s ourceÂ’s motive affects evaluations of source credibility. (Hovland, Janis, & Kelley, 1953; Walster, Ar onson, & Abrahams, 1966) Thus, the final three items measured participantsÂ’ perceptions of the sourceÂ’s intent (e.g. The organization is concerned with making profits /the publicÂ’s well being). These items are based on Gaziano and McGrathÂ’s (1986) study of the perceived cr edibility of news organizations.
40 Message Credibility ParticipantsÂ’ perceptions of the message was measured by a six-item, five-point semantic differential sc ale anchored by the terms boring/interesting, unprofessional/professional, misleading/accu rate, dull/exciting, deceptive/truthful and overemphasizes/underplays. The first five ite ms are based on Goldberg and HartwickÂ’s (1990) measure of ad credibility. Gol dberg and Hartwick used the items misleading/sincere and deceptive/honest to measure ad credibility. The remaining three items were used Â“to provide a larger set to minimize the focus on the ad credibility itemsÂ” (p. 176). The final item, overemphasizes/underp lays, was added here as a third message credibility dimension. Problem Recognition Three five-point Likert items on a scale that ranges from strongly agree to strongly disagree were us ed to measure problem recognition (e.g. HPV is a serious health problem). Participants also had a no opinion option. These items are based on Grunig and HuntÂ’s (1984) conceptua lization of problem recognition as defined by situational theory. Personal Relevance. Personal relevance was measur ed using three five-point Likert items anchored by the terms less relevant/more relevant Participants also had a no opinion option. These items are based on Davi sonÂ’s (1983) third-pe rson effect theory. The items replicate questions asked by Gunter and Thorson (1992) in a TPP study involving public service announcements where pa rticipants were asked to indicate how the message affected the relevance of the issue for themselves and other groups of people. Information Seeking. The main dependent variable in this study, information seeking, will be measured based on the typi cal method used in st udies that test the
41 situational theory. Four fivepoint Likert items with end po ints not like/very likely asked participants to indicate their intention to actively seek information in various forms related to the information seek ing cues presented in the message (e.g. calling a toll-free number to request an information kit). Participants also had a no opinion option. Behavioral Intent. The instrument also included que stions to measure behavioral intention. Participants indicated their intention to perf orm each of the three behaviors advocated in the brochure. These were measur ed using five-point Li kert items with end points definitely do/definitely do not, defin itely will/definitely will not, and definitely false/definitely true. The items are base d on one method commonly used to assess behavioral intention in studies that test the theory of reasoned action (Madden, Ellen, & Ajzen, 1992). Source Identification. The four source-identificati on questions used in the manipulation check were repeated in the main experiment instrument to test the participantsÂ’ level of awareness of the source. Prior Knowledge Grunig and HuntÂ’s (1984) situati onal theory indicates that there are differences in the communi cation behavior of individuals who have prior awareness of or involvement with an issue and those who do not. This concept is also addressed by the Elaboration Likelihood Model (Petty & Cacioppo, 1986), and at least one study has examined the relationship between levels of involvement and the effect of high versus low credibility sources (Gotlieb & Sarel, 1991). Although this study did not manipulate the awareness/involvement variable, three qu estions were asked to identify possible differences in responses based on whether or not participants had pr ior awareness of the issue. Two five-point Likert items with the end-points nothing/a lot and not at
42 all/frequently measured participants level of awareness of the issue and exposure to messages about the issue. One categorical-lev el item was included to determine which participants have personal experience with the issue. Demographic Information. In addition to the previous items, participants were asked three demographic questions which yiel ded categorical-level data. Two questions asked participants to indicate their academic rank and academic discipline at the college level. Because of the diversity of students en rolled in the class, these questions allowed for determining differences in responses among students at different academic levels or from different colleges within the university to be identified. Participants also indicated their gender, which allowed for the final samp le of female respondents to be identified. Age was collected as ratio-level data and collapsed to categorical data. Procedure The experiment was conducted during th e first session of the classÂ’ weekly meeting. The researcher explained the purpose of the exercise and the survey process to the students. The participants were told that this was a masterÂ’s thesis study seeking to gauge college studentsÂ’ attitudes toward health messages. Students were randomly assigned to treatment groups. Each participan t received a packet containing one version of the stimulus brochure and a questionna ire booklet. Both th e envelope and the questionnaire included an identifying numb er that corresponded with the treatment material included in the packet. Students we re unaware that the survey packets they received were different from mo st of their fellow classmates. Directions for completing the process app eared on the outside of the stimulus packets (See Appendix A). The di rections listed the contents of the envelope and step-by-
43 step instructions for participa ting in the study. These directions were read to participants before they viewed the stimulus material. First, participants were instructed to remove the brochure leaving the questionna ire booklet inside. Particip ants were not given a set amount of time to read the brochure. Next, th e directions instructed them to return the brochure to the envelope and remove the questionnaire. Without referring to the brochure, participants were to proceed to an swer the question items. Instructions at the top of the questionnaire brie fly stated the purpose of the items included and asked participants to answer them as honestly as possible. U pon completing the questionnaire, participants were instructed to return the booklet to the en velope, reseal the package and refrain from communicating with others in the room until all questionnaires were collected. Data Analysis Data analyses for this study were performed using SPSS 15.0 for Windows. A p< .05 significance was used as the basis for rejecting the null hypot hesis for all tests performed. One-way analyses of variance ( ANOVA) tests were used to identify the between groups and within group differences fo r each of the items. Linear regressions were used to test the hypothe ses that predict correlations between variables. Finally, multiple regressions were used to analyze the relationship between source credibility, message credibility, problem recognition, person al relevance, and in formation seeking, and the relationship between source credibil ity, message credibility, problem recognition, personal relevance, and behavioral intention.
44 CHAPTER 6: RESULTS The study focused on the responses of female students to a womenÂ’s health issue, therefore, the final sample excluded the res ponses of male students. The final sample yielded 222 female respondents. As expected, a majority of respondents (n = 89) were in their sophomore year of college. The mean age was 19.4. Table 2 summarizes the demographic characteris tics of the sample. Table 2. Categorical Demographics n% Academic rank Freshman 7232.4 Sophomore 8940.1 Junior 5323.9 Senior 83.6 Other 00.0 College Arts & Sciences 13560.8 Business 7132.0 Education 41.8 Honors 52.3 Medicine 20.9 Nursing 31.4 Missing 20.9 Age 17 20.9 18 6931.1 19 7433.3 20 4721.2 21 219.5 22 41.8 23 31.4 26 10.5 53 10.5
45 Table 3 shows the distribution of particip ants among the four treatments and the control group. Table 3. Distribution of Participants to Treatments N % Control 43 19.4 Subtle NP 44 19.8 Subtle P 50 22.5 Detailed NP 42 18.9 Detailed P 43 19.4 Descriptive Statistics In addition to items intended to coll ect demographic data about the study participants, the research instrument included items to measure participantsÂ’ level of prior knowledge, awareness of the source, percep tions of source cred ibility and message credibility, estimates of problem recogniti on, personal relevance, information seeking, and behavioral intention. Table 4 presents the means and standard deviations for each of the items used to measure these eight constructs. Results of the items used to measure prior knowledge indicate that participants were somewhat knowledgeable about the issue (M = 3.43) and were exposed to a moderate amount of information on the topic (M = 3.15). The source credibility construct included items that measured trustworthiness, expertise, and intent. In general, participan ts indicated agreement that the organizational sources were trustworthy, knowledgeable, and concerned about the pub lic. Of the nine source credibility items, study participants mo st strongly agreed that the organization was concerned with the publicÂ’s well being (M = 4.46). The participants reported low levels of source recall (M = 2. 34) and ease identifying the source (M = 2.43).
46 Of the six items used to measure message evaluation, three measured participantsÂ’ perceptions of message credibil ity. Responses to these items i ndicate that the participants viewed the message as truthful (M = 4.24), professional (M = 4.23) and accurate (M = 4.09). In response to the problem recognition items, study participants almost strongly agreed that HPV was a serious health problems (M = 4.48), with serious complications (M = 4.66), that people should be concerne d about (M = 4.66). In terms of personal relevance, they agreed that the message made the issue relevant for them (M = 3.96), but also indicated that it was more relevant to their fellow classmates (M = 4.12) and other college students (M = 4.15). The results of the information seeki ng items show that participants were somewhat unlikely to speak to a health care professional abou t HPV (M = 2.70), less likely to visit a website (M = 2.19) and unlik ely to pick up a pamphlet from a student health center (M = 1.93) or cal l a toll free number to get more information (M = 1.37). In terms of behavioral in tent, participants were like ly to get vaccinated against HPV (M = 3.80), less likely to ge t tested for it (3.47), and so mewhat likely to tell a friend about the disease (M = 3.02). Table 4. Descriptive Statistics Construct Item Mean Std. Deviation Prior Knowledge How much you knew about HPV prior to reading this brochure. 3.43 1.00
47 How frequently you have come across information about HPV in the last 12 months. 3.15 1.13 Source Credibility The organization is qualified to provide information about HPV. 4.20 0.78 The organization can be trusted to provide factual information about HPV. 4.15 0.81 The organization is concerned with the publicÂ’s well being. 4.46 0.76 The organization is not an expert on HPV. (recoded) 3.54 1.11 The organization cannot be trusted to present reliable information about HPV. (recoded) 4.03 1.01 The organization is concerned with making profits. (recoded) 3.88 .98 I believe the organization provides unbiased information about HPV. 3.84 1.09 I believe the organization is knowledgeable about HPV. 4.24 0.80 I believe the organization has something to gain from publishing this information. (recoded) 2.84 1.19 Message credibility Boring/Interesting 3.46 1.04 Unprofessional/Professional 4.23 0.82 Misleading/Accurate 4.09 0.93 Dull/Exciting 2.90 0.84 Deceptive/Truthful 4.24 0.88 Overemphasizes/Downplays 2.96 0.69 Problem Recognition HPV is a serious health problem. 4.48 0.74 People should be concerned about the risks of HPV. 4.66 0.58
48 HPV can have serious complications. 4.66 0.63 Personal Relevance Has this brochure made the issue of HPV more relevant or less relevant for you? 3.98 0.85 Do you think this brochure made the issue of HPV more relevant or less relevant for other students in the class? 4.12 0.80 Do you think this brochure would make the issue of HPV more rele vant or less relevant for college students in general? 4.15 0.79 Information Seeking I will visit the web site to learn more about HPV. 2.19 1.20 I will call the toll-free number to request the HPV 1.37 0.68 I will ask a health professional about HPV risks. 2.70 1.43 I will pick up a pamphlet about HPV from the student hea lth center. 1.93 1.16 Behavioral Intent I intend to tell a friend about HPV. 3.02 1.32 I intend to get tested for HPV. 3.47 1.46 I intend to get vaccinated against HPV. 3.80 1.36 Source Identification I recall the name of the organization that sponsored this brochure 2.34 1.49 It was easy to identify the organization that sponsored this information 2.43 1.40
49 Reliability Prior to testing the hypotheses, Cronbach's alpha was used to assess the internal consistency of the multiple-item indices for source credibility, message credibility, problem recognition, personal relevance, info rmation seeking behavi or, and behavioral intent respectively. Reversed items were tr ansformed before performing the reliability analysis. The results of the an alyses are shown in Table 5. Good reliability estimates are coefficients of .70 or higher, while values between .80 and 1.00 indi cate high reliability (Stacks, 2002). Three items were developed to measure each of the three dimensions of source credibilityÂ—trust, expertise, and concern. None of these produced a reliability coefficient of .70 or higher. Next, these nine source cred ibility measures were submitted to factor analysis. The analysis produced three initial factors. Although each factor met the criteria for a measurement dimension, with a minimum of two items loading at greater than .60 on one factor and not greater than .40 on a ny other factor (Stack s, 2002), only the items that loaded on Factor 1 were used to create a single index to m easure source credibility. This factor contained four items with at le ast one item to represent each of the three dimensions of source credibility used in this study: trustÂ—Â“the organization can be trusted to provide factual information a bout HPVÂ”; expertiseÂ—Â“the organization is qualified to provide information about HP V,Â” and Â“I believe the organization is knowledgeable about HPVÂ”; and concernÂ— Â“the organization is concerned about the publicÂ’s well being.Â” These four item s produced a coefficient alpha of .828. Six items were used to measure partic ipantsÂ’ perceptions of the message. Only two of these items were intended to m easure message credibility. However, an
50 exploratory factor analysis showed that three items loaded on the credibility factor: professional/unprofessional, misleading/accura te, and deceptive/truthful. These three items produced an alpha of .843. The three items included to test pr oblem recognition produced a reliability coefficient of .777. The three items included to test personal relevance produced an alpha of .803. The four items used to measure inform ation seeking behavior produced an alpha of .762. The three items used to measure be havioral intent pr oduced a coefficient alpha .706. Based on the results of the reliability tests, the items were collapsed into indices for the six constructs: sour ce credibility, message cred ibility, problem recognition, personal relevance, information seeking, and be havioral intent. Tabl e 6 reports the means and standard deviations for each inde x, from the highest to the lowest. Table 5. CronbachÂ’s alphas Variables N of items Source Credibility 0.83 4 Message Credibility 0.84 3 Problem Recognition 0.78 3 Personal Relevance 0.80 3 Information Seeking 0.76 4 Behavioral Intent 0.70 3 Table 6. Descriptive Statistics for Construct Indices Mean Std. Deviation Problem Recognition 4.60 0.54 Source Credibility 4.28 0.64 Message Credibility 4.18 0.76 Personal Relevance 4.09 0.68 Behavioral Intent 3.44 1.09 Information Seeking 2.03 0.88
51 Research Question 1 and Hypothesis 1 Research Question 1 asked whether audien ces are aware of the for-profit or notfor-profit status of message sponsors and what affect that awarene ss has on perceptions of source credibility and message credibility. Hypothesis 1a Hypothesis 1a predicted that forprofit sponsors would be rated as less credible sources than not-for-profit sponsors. A one-way analysis of variance (ANOVA) was conducted to test this hypothe sis. Although the differences in source credibility scores for the non-profit versus fo r-profit treatments we re as predicted, the between group differences were not significant (F (1,151) = 3.155; p = .078). However, the mean source credibility scores for the non-profit treatment groups (M = 4.38) were higher than the mean source credibility scores for the for-profit treatment groups (M = 4.20). A cursory analysis of the mean scores of the five groups on this index showed the mean source credibility scores for both the subtle-id non-profit (M = 4.36) and detailed-id non-profit ( M = 4.4) were higher than t hose of the control group ( M = 4.26). The mean scores of the subtle-id for-pro fit and detailed-id for-profit ( M = 4.15, M = 4.25) were lower than both non-profit treatment gr oups and the control group. Nonetheless, hypothesis 1a was not supported. Hypothesis 1b Results from the same ANOVA were used to address hypothesis 1b. Hypothesis 1b predicted that participants in the detailed-id for-p rofit treatment would report the lowest source credibility scores ( M = 4.25). However, participants in the subtle-id for-profit treatment reported the lowest source credibility scores ( M = 4.15).
52 Thus, hypothesis 1b was not supported. Table 7 reports the source credibility means and standard deviations for each treatment group, from the highest to the lowest. Table 7. Means and Standard Deviations for Source Credibility Treatment Group N Mean Std. Deviation Detailed NP 38 4.40 0.55 Subtle NP 36 4.36 0.71 Control 34 4.26 0.64 Detailed FP 39 4.25 0.70 Subtle FP 40 4.15 0.58 Hypothesis 1c Hypothesis 1c predicted a direct and positive relationship between participantsÂ’ estimates of source credibility and their estimates of message credibility. PearsonÂ’s correlation analysis was used to test this hypothesis. (See Appendix F for the results of a comprehensive correlation analysis between indices). There was a statistically significant but moderate correlation between s ource credibility and message credibility ( r = .490, p = .000). This hypothesis was also tested using linear regression analysis, which supported the statistical signifi cance but moderate effect of the relationship. Message credibility, the dependent variable, was regr essed on source credibility, the independent variable. The regression indi cated that 23.6% of message credibility is explained by source credibility, R2 = .24, Adj. R2 = .236, F(1, 183) = 57.793, p = .000. Message credibility also produced a st atistically significant contribu tion to the prediction equation, = .490, t(183) = 7.602, p = .000. To further explore the implications of this hypothesis an ANOVA was conducted to compare participantsÂ’ evaluation of messa ge credibility across the five treatment groups. Although the differences in message cred ibility scores for th e non-profit versus for-profit treatments were as predicte d, the between group differences were not
53 statistically significan t (F (4, 214) = .928; p = .173). However, a cursory analysis of the mean scores of the five groups on this inde x showed a similar pattern as for source credibility. The mean message credibility scores for the s ubtle-id non-profit (M = 4.36), detailed-id non-profit ( M = 4.18) and the control group ( M = 4.26) were higher than those of the subtle-id for-profit (M= 4.01) and detailed-id for-profit ( M = 4.10). Thus, hypothesis 1c was supported. Table 8 reports th e message credibility means and standard deviations for each treatment group, from the highest to the lowest. Table 8. Means and Standard Devi ations for Message Credibility Treatment Group N Mean Std. Deviation Subtle NP 43 4.36 0.92 Control 41 4.30 0.67 Detailed NP 42 4.18 0.61 Detailed FP 43 4.10 0.69 Subtle FP 50 4.01 0.85 Research Question 2 and Hypothesis 2 Research Question 2 asked if message acceptance was different for for-profit message sponsors than for non-profit sponsors. Message acceptance was defined in terms of estimates of problem recognition and levels of involvement or personal relevance. Hypothesis 2a Hypothesis 2a predicted that part icipantsÂ’ perception of source credibility would affect their levels of probl em recognition. PearsonÂ’s correlation analysis showed that although the relationship between source credibility and problem recognition was statistically significant, the correlation was weak ( r = .272, p = .000). To further test this hypothesis and to determine to which ex tent problem recognition affected source credibility, linear regression analysis was conducted. The re gression indicated that only 6.9% of problem recognition was ex plained by source credibility, R2 = .074, Adj. R2
54 = .069, F(1, 182) = 14.543, p = .000. Source cred ibility also produced a statistically significant contribution to the prediction equation, = .272, t(182) = 3.813, p = .000. Thus, hypothesis 2a was supported. Hypothesis 2b Hypothesis 2b predicted a dire ct and positive relationship between participantsÂ’ estimates of source cr edibility and their estimates of personal relevance. PearsonÂ’s correlation analysis indi cated there was a statistically significant but moderate correlation between source credibility and personal relevance ( r = .437, p = .000). This hypothesis was also tested using linear regression analysis, which supported the statistical significance but moderate effect of the relationship. Pe rsonal relevance, the dependent variable, was regressed on source credibility, the independent variable. The regression indicated that 18.7% of personal re levance is explained by source credibility, R2 = .191, Adj. R2 = .187, F(1, 167) = 39.520, p = .000. Source credibility also produced a statistically significant cont ribution to the pr ediction equation, = .437, t(167) = 6.287, p = .000. Thus, hypothesis 2b was supported. Hypothesis 2c Hypothesis 2c predicted that part icipants in the detailed-id forprofit treatment would report th e least problem recognition ( M = 4.57). A one-way analysis of variance (ANOVA) was conducted to test this hypothesis. The between group differences were not statistical ly significant (F(4,213) = 1.166, p = .327). Additionally, participants in the su btle-id for-profit treatment reporte d the lowest problem recognition scores ( M = 4.47). Thus, hypothesis 2c was not su pported. Table 9 reports the problem recognition means and standard deviations fo r each treatment group, from the highest to the lowest.
55 Table 9. Means and Standard Deviations for Problem Recognition Treatment Group N Mean Std. Deviation Control 43 4.67 0.46 Subtle NP 43 4.67 0.54 Detailed NP 42 4.64 0.48 Detailed FP 42 4.57 0.56 Subtle FP 48 4.47 0.63 Hypothesis 2d Hypothesis 2d predicted that partic ipants in the detailed-id forprofit treatment would report the lowest levels of personal relevance. A one-way analysis of variance (ANOVA) was conducted to te st this hypothesis. The between group differences were not statistical ly significant (F(4,195) = .696, p = .596). Additionally, participants in the contro l group reported the lowest personal relevance scores ( M = 3.96). Participants in the subtle-id for-profit treatment reported lower personal relevance scores ( M = 4.04) than those in the deta iled-id for-profit treatment ( M = 4.14). The detailed-id non-profit group reported personal relevance scores ( M = 4.20) that were higher than the for-profit treatments and the control group. Ho wever, personal relevance scores for the subtle-id non-profit group ( M = 4.12) were only higher th an the control group and the subtle-id for-profit group. Thus, hypothesis 2d was not supported. Table 10 reports the personal relevance means and standard deviations for each treatment group, from the highest to the lowest. Table 10. Means and Standard Deviations for Personal Relevance Treatment Group N Mean Std. Deviation Detailed NP 38 4.20 0.70 Detailed FP 39 4.14 0.65 Subtle NP 40 4.12 0.59 Subtle FP 45 4.04 0.66 Control 38 3.96 0.82
56 Research Question 3 and Hypothesis 3 Research Question 3 asked if there were differences in the information seeking behavior of participants expos ed to the not-for-profit treatme nts than for those exposed to the for-profit treatments. Hypothesis 3a. Hypothesis 3a predicted a positive relationship between levels of problem recognition and intention to seek info rmation. Correlation analysis was used to test this hypothesis. Although the relationship was significant the correlation was weak ( r = .218, p = .002). This hypothesis was further expl ored using linear regression analysis, which supported the statistical significance but m oderate effect of the relationship. Intent to seek information, the dependent variable was regressed on problem recognition, the independent variable. The regr ession indicated that 4.3% of intent to seek information was explained by problem recognition, R2 = .047, Adj. R2 = .043, F(1, 207) = 10.300, p = .002. Problem recognition also produced a st atistically significant contribution to the prediction equation, = .218, t(207) = 3.209, p = .002. Thus, hypothesis 3a was supported. Hypothesis 3b Hypothesis 3b predicted a positive re lationship between levels of personal relevance and intenti on to seek information. Correlation analysis was used to test this hypothesis. Although th e relationship was statisticall y significant, the correlation was weak ( r = .312, p = .000). To further explore this re lationship regression analysis was conducted between intent to seek informa tion, the dependent variable, and personal relevance, the independent variable. The regre ssion indicated that 9.3% of intent to seek information was explained by personal relevance, R2 = .097, Adj. R2 = .093, F(1, 190) = 20.525, p = .000. Personal relevance also produced a statistically significant
57 contribution to the prediction equation, = .312, t(190) = 4.530, p = .000. Thus, hypothesis 3b was supported. Hypothesis 3c Hypothesis 3c predicted that partic ipants in the detailed for-profit treatment would report the leas t intention to seek informa tion. A one-way analysis of variance (ANOVA) was conducted to test this hypothesis. The between group differences were not statistically si gnificant (F(4 ,208) = 1.048, p = .384). Additionally, participants in the control group reported the lowe st intention to seek information ( M = 1.82). Participants in the subtle-id for-profit treatment reported lower information seeking scores ( M = 1.99) than those in the detailed-id for-profit treatment ( M = 2.07). Both the subtle-id non-profit and detail ed-id non-profit groups reported information seeking scores ( M = 2.11, M = 2.17) that were higher than the for-profit treatmen ts and the control group. Thus, hypothesis 3c was not supported. Table 11 reports the information seeking means and standard deviations for each treatment group, from the highest to the lowest mean. Table 11. Means and Standard Deviations for Information Seeking Treatment Group N Mean Std. Deviation Detailed NP 42 2.17 0.86 Subtle NP 44 2.11 0.97 Detailed FP 39 2.07 0.89 Subtle FP 46 1.99 0.79 Control 42 1.82 0.87 Research Question 4 and Hypothesis 4 Research Question 4 asked if there were di fferences in the behavioral intent for participants exposed to the not-for-profit treatments than for those exposed to the forprofit treatments.
58 Hypothesis 4a Hypothesis 4a predicted that high er levels of problem recognition would correlate with greater be havioral intent. Correlation anal ysis was used to test this hypothesis. Although the relationship was si gnificant, the correlation was weak ( r = .356, p = .000). Subsequently, behavioral intent, the dependent variable, was regressed on problem recognition, the independent variable The regression indicated that 12.2% of behavioral intent was expl ained by problem recognition, R2 = .127, Adj. R2 = .122, F(1, 197) = 28.622, p = .000. Problem recognition al so produced a statis tically significant contribution to the prediction equation, = .356, t(197) = 5.350, p = .000. Thus, hypothesis 4a was supported. Hypothesis 4b. Hypothesis 4b predicted that higher levels of personal relevance would correlate with greater be havioral intent. Correlation anal ysis was used to test this hypothesis. There was a statistically signi ficant but moderate correlation between personal relevance and behavioral intent, r = .422, p = .000. Subsequently, behavioral intent, the dependent variable, was regre ssed on personal relevance, the independent variable. The regression indi cated that 17.8% of behavior al intent was explained by personal relevance, R2 = .178, Adj. R2 = .174, F(1, 182) = 39.429, p = .000. Personal relevance also produced a statistically signi ficant contribution to the prediction equation, = .42222, t(182) = 6.279, p = .000. Thus, hypothesis 4b was supported. Hypothesis 4c. Hypothesis 4c predicted that part icipants in the detailed-id forprofit treatment would report the lowest levels or behavioral intent A one-way analysis of variance (ANOVA) was conducted to te st this hypothesis. The between group differences were not statistical ly significant (F(4,198) = .448, p = .384). Additionally, participants in the cont rol group reported the lowe st behavioral intent ( M = 3.33).
59 Participants in the subtle-id for-profit treatm ent reported lower behavioral intent scores ( M = 3.43) than those in the deta iled-id for-profit treatment ( M = 3.46). The detailed-id non-profit group reported behavioral intent scores ( M = 3.63) that were higher than the for-profit treatments and the control group. Howe ver, behavioral scores for the subtle-id non-profit group ( M = 3.38) was only higher than th e control group. Thus, hypothesis 4c was not supported. Table 12 reports the behavior al intent means and standard deviations for each treatment group, from the highest mean to the lowest. Table 12. Means and Standard Deviations for Behavioral Intent Treatment Group N Mean Std. Deviation Detailed NP 39 3.63 1.11 Detailed FP 38 3.46 1.05 Subtle FP 42 3.43 1.09 Subtle NP 43 3.38 1.21 Control 41 3.33 1.03 Research Question 5 and Hypothesis 5 Research Question 5 asked if there were differences in participantsÂ’ willingness to engage in information seeki ng and their intent to perfor m behaviors advocated by the message. Hypothesis 5a. Hypothesis 5a predicted that pa rticipants across all treatment groups would report greater inte ntion to seek information th an to perform the advocated health behaviors. There was a significant but moderate correlation between information seeking and behavioral intent, r = .541, p = .000. However, a comparison of means between groups in terms of information s eeking and behavioral intent showed that participants in all treatment groups reported lower levels of information seeking (M = 2.03) than levels of behavioral intent (M = 3.44). This indi cates that in this study the
60 participants were more likely to behave as was requested, than to seek more information. Thus, hypothesis 5a was not supported. Table 13 reports the means for information seeking and behavioral intent for each treatment group. Table 13. Means for Information Seeking and Behavioral Intent Treatment Means for Information Seeking Means for Behavioral Intent Control 1.82 3.33 Subtle NP 2.11 3.38 Detailed NP 2.17 3.63 Subtle FP 1.99 3.43 Detailed FP 2.07 3.46 To further explore the research questi on, two linear regression analyses were conducted. In the first analys is the dependent variable information seeking, was regressed on the independent variables of s ource credibility, message credibility, problem recognition, and personal relevance, as sugge sted in Figure 1. The regression indicated that 8.7% of information seeking was e xplained by these four variables, R2 = .110, Adj. R2 = .087, F(4, 156) = 4.834, p = .001. Of the fo ur independent variables only personal relevance produced a statistically signifi cant contribution to the prediction equation, = .233, t(156) = 2.705, p = .008. The regression model for information seeking is presented in Table 14. Table 14. Regression Model for Information-Seeking Behavior Predictor t(149) p Source Credibility .029 0.309 .758 Message Credibility .040 0.458 .647 Problem Recognition .134 1.661 .099 Personal Relevance .233 2.705 .008
61 In the second analysis the dependent vari able, behavioral inte nt, was regressed on the same independent variables of source credibility, message credibility, problem recognition, and personal relevance, as sugge sted in Figure 2. The regression indicated that these four variables explaine d 20.3% of behavioral intent, R2 = .224, Adj. R2 = .2037, F(4, 149) = 10.759, p = .000. Of the four inde pendent variables, two variables produced a statistically significant contribution to the prediction equation. Personal relevance contributed more to the prediction equation, = .273, t (149) = 3.309, p = .001, than did problem recognition, = .174, t(149) = 2.237, p = .027. The regression model for behavioral intent is presented in Table 15. Table 15. Regression Mode l for Behavioral Intent Predictor t(149) p Source Credibility .115 1.275 .204 Message Credibility .081 0.968 .335 Problem Recognition .174 2.237 .027 Personal Relevance .273 3.309 .001 Research Question 6 Research Question 6 asked which of the f actors examined in the study contribute most to behavioral intent. This research question was answered using linear regression analysis. Information seeking was added to the predictor variable s shown in Table 15. The regression indicated that 37.3 % of the variance in behavior al intent is explained by these five variables, R2 = .394, Adj. R2 = .373, F(5, 145) = 18.865, p = .000. These results indicate that information seeking accounted for an additional 17% of explained variance in behavioral intent. Of the five independent variab les, two variables produced a statistically significant cont ribution to the prediction e quation, personal relevance,
62 = .150, t(145) = 1.977, p = .050 and information seeking, = .458, t (145) = 6.592, p = .000. Information seeking made the greates t contribution to the prediction equation. With the addition of information seeking, problem recognition no longer made a unique contribution to the prediction equation. Thus of the factors examined in this study, information seeking contributes most to behavioral intent. Table 16 presents the regression model for behavioral inte nt based on these five factors. Table 16. Five-Factor Regression Model for Behavioral Intent Predictor t(145) p Source Credibility .101 1.255 .212 Message Credibility .084 1.123 .263 Problem Recognition .085 1.199 .233 Personal Relevance .150 1.977 .050 Information Seeking .458 6.592 .000
63 CHAPTER 7: DISCUSSION In recent years, the organizational so urces of public service messages have diversified. Whereas, PSA sponsors were t ypically government agencies and non-profit organizations, for-profit organizations are s ponsoring public information campaigns more frequently, particularly in the area of health communication. For instan ce, in an official statement about its operations, the pharmaceutical company Merck states that it Â“publishes unbiased health information as a non-profit servic eÂ” (Merck, 2007). The increasing cost of producing effective info rmation campaigns and decreasing access to dissemination channels have necessitated such changes. Organizational Status, Source Cred ibility, and Message Credibility Although public service messages produced by for-profit organizations do not promote the companiesÂ’ services, they often a dvocate health issues cl osely related to the companiesÂ’ product lines. In the professional arena, there has been some speculation about the potential profit mo tive of companies sponsoring such campaigns. In terms of source credibility, the results of several studies suggest that for-prof it sources are viewed as less credible than non-profit sponsors becau se they are seen as having something to gain from the message (Haley, 1996; Ha mmond, 1986; Reid, Soley, & Vanden Bergh, 1981). Contrary to these findings, the resu lts of this study indicate no significant differences in the perceived credibility of the commercial source versus the non-profit source. Although the source credibility means of both for-profit source treatments were
64 lower than the source credibi lity means of both the non-profit treatments and the control group, in general participants reported high source credibility ratings across all groups. This finding suggests that a udiences do not evaluate the cr edibility of an organization based on its for-profit or non-profit status. Thus, in previous studies where for-pr ofit sources and messages attributed to these sources were found to be significantly less credible, the conc lusion that audiences perceive for-profit sources as less objective and having something to gain may not be an accurate analysis of the process underlying credibility evaluations While evaluation of intent is likely a valid measure of source cred ibility, it seems unlikely that this evaluation is made based on the sponsorÂ’s status as a for-profit or non-prof it organization alone. Based on an earlier conclusion that open identification of a commercial source related to greater perception that the s ource had something to gain (Reid, Soley & Vanden Bergh, 1981), this study included level of source identification as its second independent variable. The result s indicate that more detailed or obvious references to the source is significantly related to greater recognition of the sponsor and the sponsorÂ’s status as a for-profit or non-profit source. The anticipated result was th at as participants became more aware of the sponsor and the s ponsorÂ’s organizational status, credibility ratings for the non-profit source would increas e while perceived credibility of the forprofit source would decrease. However, in ad dition to the finding that there were no statistically significant differen ces in credibility ratings be tween the treatment groups, the results also show that the means of both the detailed-id non-profit and detailed-id forprofit treatments were higher than their subtle -id counterparts. Add itionally, participants indicated strongest agreement that the source, in general, was concerned with the publics
65 well being. These results provide further support to the conclusion that the source is not evaluated based on its status as a fo r-profit or non-profit organization. Particularly interesting is that 79% of participants in the control group, which had no sponsor, responded to the source credibility items, rather than indicating no opinion. Perhaps audiences assume there is an organiza tional source but pay l ittle attention to the sponsor, thereby relying on other cues, such as content cues to evalua te the source. Some studies indicate that public service messages are viewed pos itively and assumed to have an altruistic intent (Duck & Mullin, 1995; Duck, Terry, & Hogg, 1995; Gunther & Thorson, 1992). This perception of PSAs as a me ssage category may be transferred to the organizations that sponsor public service messages. Thus, if a message appears to function in the best interest of the audience, then the source of the message may also be viewed as publishing the information in the audienceÂ’s best interest. Alternatively, greater awareness of the source may lead the audience to evaluate it on individual factors, such as familiarity or reputation. Messages in the for-profit treatments of this study were attributed to the pharmaceutical company Merck. A recent article published by the Ha rris Interactive reported that 60% of adults think pharmaceutical companies generally do a good job compared to 39% who think the industry generally does a bad job (Harri s Interactive, August 8, 2007). Thus, highcredibility ratings of the fo r-profit source in this study ma y be related to increased awareness that the organi zational sponsor is a ph armaceutical company. However, according to the poll, the pharmaceutical industry has a 21 point positive rating, placing it near th e bottom of the list of industr ies that received a positive rating and ranking higher than only six of the 21 industries rated in the poll. And
66 although perceptions of the industry have increas ed over the last four years, they are still significantly lower than 10 years ago. The fact that Merck is a pharmaceutical company may not contribute as much to perceptions of the sponsorÂ’s credibility as might perceptions of MerckÂ’ s reputation itself. No significant differences were found in the credibility ra tings of messages attributed to non-profit spons ors versus for-profit sponsor s. There was a significant positive correlation between source credibility and message credibility. Participants reported high message credibility ratings acro ss all groups. However, participants in the for-profit treatments reported lower message cr edibility ratings than did participants in the non-profit treatments and the control group. This is not surprising, given the similar source credibility ratings for each group. These finding indicates that regardless of the for-profit/non-pro fit status of the sponsor, the audienceÂ’s perception of the source does affect its perc eption of the message. This finding supports previous organization credibility studies that showed a direct relationship between source credibility and me ssage evaluation or attitude-toward-the-ad (Goldberg & Hartwick, 1990; Goldsmith, Laff erty, & Newell, 2000a, 2000b; Lafferty & Goldsmith, 1999; Lafferty, Goldsmith, & Newell 2002) with the exce ption of one study that found no significant relationship between source credibility and message evaluation (Hammond, 1986). Problem Recognition and Personal Relevance Of the three types of campaign messa gesÂ—awareness, instructional, and persuasiveÂ—most campaign messages can be classified as awareness messages. These messages aim to create recognition of a topi c and impart new information. Mendelsohn
67 (1973) argued that campaigns can succeed if they focus on increasing knowledge and awareness. The current study did not test for knowledge gains as a result of exposure to the message. For the purpose of this study, message acceptance was operationalized as estimates of problem recognition and personal relevance, two indepe ndent variables from the situational theory of publics. Atkin ( 2001) states that awareness messages should convey the impression that the health problem is important. This study examined the relationship betw een the organizational status of the source and message acceptance. That is whether for-profit and non-profit sources produced different levels of problem rec ognition and personal rele vance. Although there was a significant correlation between source cr edibility and problem recognition, neither organizational status nor level of identificati on produced significant differences in levels of problem recognition. Likewise neither organizational status nor level of identification produced significant differences in levels of pe rsonal relevance. Just as source credibility ratings were high across all treatment groups so were levels of problem recognition. In general, personal relevance estimates were moderately high. The results do not support the studyÂ’s hypothesis that for-profit sponsors produce lower estimates of problem recognition. Yet, they do support claims that public service messages, regardless of the status of the organizations that produce th em, are successful at placing problems on an individualÂ’s agenda (Grunig & Ipes, 1983). Across all groups, estimates of problem rec ognition were higher than estimates of personal relevance. While all participants agr eed that the message made the issue relevant to them, they showed strong agreement that the issue was a problem. This suggests that participants were more willing to admit that HPV is a serious health problem, but less
68 willing to admit that it is relevant to them. The estimates of personal relevance or firstperson perception scores were lower than th e third-person perception scores. This finding supports the third-person effect hypothesis th at people are less likely to admit greater impact on self than others even for soci ally desirable messages (Davison, 1983; Duck & Mullin, 1995; Duck, Terry, & Hogg, 1995). Howeve r, the first-person estimates were not significantly lower than the third-person estimates and indicated that participants felt the issue was personally relevant. This finding is consistent with previous studies that found smaller differences between firstand thir d-person estimates for public service messages (Duck, Terry, & Hogg, 1995; Gunther & Thorson, 1992). Information Seeking and Behavioral Intent According to Atkin (2001) awaren ess messages should encourage further information seeking. Although MendelsohnÂ’s (1 973) argument that public information campaigns can succeed in increasing knowledge and awareness was intended to counter Hyman and SheatsleyÂ’s (1947) pe ssimistic view of information campaigns, he essentially agreed with their argument that they have limited effects on behavior. Based on this agreement that campaigns are more successful at producing knowledge effects rather than behavioral effects, it was hypothe sized that participants across all treatments would report greater levels of information seeking effect s than behavioral effects. Although it was not possible to measure actual behavior in res ponse to the message, th e theory of reasoned action posits that intention to act is a valid indicator of actual behavi or (Fishbein & Ajzen, 1975). Comparing the estimates of information s eeking to the estimates of behavioral intention found in this study calls into quest ion the argument that public service messages
69 are more successful at prompting awarenesstype behavior than they are at inducing actual behavior (Borzekowski & Poussain t, 1999; Ledingham, 1993; Mendelsohn, 1973). Neither the information seeking nor the behavioral intent estimates fell in the Â“likelyÂ” to Â“very likelyÂ” range of the scale. However, participants across all groups indicated that they were somewhat likely to perform the be haviors advocated in the message, while they disagreed that they would seek additional in formation. Thus, the pa rticipants indicated greater behavioral intent than intent to seek informati on. Although this finding did not support the studyÂ’s hypothesis that participan ts would be more likely to carry out information seeking behavior, it has signifi cant implications for public service message effects. One possible explanation for the low estimat es of information seeking may be the channel used to deliver the public service message in this study. Unlike television PSAs and other such media which ha ve the broadest reach but can only deliver a superficial amount of information (Atkin, 2001), pamphlets are able to deliver depth of information. Thus, participants may not see the need to seek more information based on a perception that the brochure provided suffici ent information on the issue. Neither organizational status nor leve l of identification produced significant differences in information seeking or behavior al intent. However, source credibility was positively related to information seeking and behavioral intent. The relationship between source credibility and behavioral effects has be en well established in the literature. These findings support the results of other organizatio nal credibility studies that have looked at purchase intentions (Goldberg & Hartwic k, 1990; Goldsmith, Lafferty, & Newell, 2000a, 2000b; Lafferty & Goldsmith, 1999; Lafferty, Goldsmith, & Newell, 2002) and health
70 behaviors (Hammond, 1986). Lynn et al. (1978) looked at behavior al intent as a result of source attribution for a public service announcement. However, that study did not examine source credibility and the results in dicated Thus, the pres ent study represents initial evidence of a relationship between sour ce credibility and be havioral effects for public service messages. The situational theory predicts that in creasing an individualÂ’ s level of problem recognition and personal involvement will resu lt in greater information seeking (Grunig & Hunt, 1984). The results of this study, wh ich show that both problem recognition and personal involvement were pos itively related to informa tion seeking, provide further support for the theory. However, as predic ted in this study, problem recognition and personal involvement were also positively rela ted to behavioral inte nt. This study argued that there were parallels between the pr oblem recognition and personal involvement variables from situational theory and the at titude and social norms variables from the theory of reasoned action. The findings pr ovide initial support for this claim. What the study did not predict, but wa s a significant finding, was that problem recognition and personal relevance correlated mo re strongly with beha vioral intent than they did with information seeking. Additiona lly, regression analysis showed that these two variables also predicted mo re of the variance in behavior al intent than they did in information seeking. The higher behavioral intent estimates, along with the stronger correlation between problem recognition, persona l relevance, and behavioral intent, suggests that public service messages can succ essfully encourage behavior if they are able to increase audiencesÂ’ perceptions of problem recognition and personal relevance.
71 These findings also provide initial jus tification for an expansion of GrunigÂ’s (Grunig & Hunt, 1984) situational theory. Currently the theory only predicts the conditions whereby information processing and information seeking will occur. However, the correlation between two of the theories independent variables, problem recognition and personal involvement, and be havioral intent, suggests th at the theory can go beyond predicting awareness behavior to predicting actual behavior through behavioral intent. Given that the theory is often used to de termine which segments of a population will become active, adding a behavioral dimension can add value to the theories predictive ability. Atkin (2001) indicated that information seeking was an important effect of awareness messages because it had strong potential to produce behavioral outcomes. Grunig (1989) also stated that information seeking was rela ted to behavioral outcomes. Neither author referenced empirical evidence of these claims. One of the most significant findings for this study is that, of the factor s examined, information seeking has the most effect on behavioral intent. There was a si gnificant relationship between information seeking and behavioral intent and the correlation of these tw o variables was the strongest of all the correlations found. Additionally, in formation seeking predicted almost 17% of the variance in behavioral inte nt. This finding provides an ev en greater argument for the expansion of the situational th eory to incorporate behavioral intent. Figure 3. proposes a model of the theory based on a revision that incorporates behavioral intent.
72 Figure 3. Proposed Model of Behavioral Intention Problem Recognition Personal Involvement Information Seeking Behavioral Intent
73 CHAPTER 8: CONCLUSIONS This study sought to investigate the re lationship between message sponsors and message effects for public information campaigns. It asked whether for-profit organizations as sponsors of health informa tion enhance or inhibit the effectiveness of health messages. The findings indicate that for-profit sources can be as successful as nonprofit sources in increasing audiencesÂ’ percep tions of problem rec ognition and personal involvement for a health issue. Additionally, while audiences may rate for-profit sources as somewhat less credible than non-profit sour ces, there were no si gnificant differences in the levels of information seeking or be havioral intention generated by messages from either type of sponsor. Thus, it appears that for-profit sources do not inhibit the effectiveness of public service messages. These findings do not support previous studi es that have attributed lower source credibility and message credibility ratings of for-profit sources to audience evaluations of intent, namely that the for-profit source ha s something to gain. Although an organization may gain from a public service message, if th e information presented is deemed accurate and unbiased, any evaluation of profit motive may not impact overall source evaluations. Alternatively, high message credibility ratings may lead to higher source credibility estimates and the potential profit motive may be overlooked. In other words, a credible message may create the impression of an altrui stic source. This may be why participants reported higher levels of agreement that th e source was concerned with the publicÂ’s well
74 being, but were less apt to concede that th e source had something to gain or was concerned with making a profit. It may also ex plain why participants in the control group reported high source credibility ratings, even though the message they were exposed to contained no sponsor information. Perhaps the studyÂ’s most significant findi ng is that problem recognition, personal involvement, and information seeking all positively predict be havioral intention, suggesting additional routes that can be used to maximize public service message effects. Study Limitations Although this study presents significant findings that link so urce credibility research to studies on the effects of public information campaigns, it also had several limitations that prevent gene ralization of the findings. A common criticism of academ ic studies is the use of college students as the study participants. This study used a conveni ence sample of undergraduate students. Because the sample was not randomly selected the results cannot be taken as representative of a larger population. Add itionally, the researcher noted that some students responded to the experime ntal exercise as they woul d a graded assignment. Even though there were no right answ ers, student responses may have been altered by their perceptions of what th e answer should be. The study only used one message channel as the stimulus. Thus, certain results may be more closely related to the channel selected rather than a result of public service messages in general. Additionally, the se lected topic is popular. Although students reported moderate levels of awareness about the issue, it is still likely that their responses are affected by their existing levels of knowledge. Public se rvice messages often seek to
75 create levels of awareness about problems that are not well known. Thus, responses to effective items in this study may have differe d if the topic was less salient. Perhaps the audience is more critical of the source for issues that are not as well known. The study used two organizations as th e sources that are likely familiar and viewed favorably by participants. As a result the lack of significant differences in the source credibility ratings may be due to actua l perceptions of these organizations, than they are due to differences in the for-prof it or non-profit status of each. Also, only two organizational categories, for-profit and nonprofit were considered. Thus, this study cannot comment on or compare the effects of other organizational cat egories, such as governmental sponsors. Although this study used established scal es to measure problem recognition, information seeking, and behavioral intent, this study did not use validated scales to measure source credibility and personal rele vance. As a result, comparing responses on these items to the findings of other studies will be limited. Suggestions for Further Research Based on the findings presented and the li mitations discussed, several areas for further research can be identified. Few cred ibility studies or PSA effects studies use participants who are not college students. In order to lear n more about the relationship between source credibility and message effects, future studies should seek to incorporate more diverse samples. As noted, it is possible that the informati on seeking findings were a result of the message channel selected for the study. Becau se there are few orga nizational credibility studies that address public service messages, there is great opportuni ty to explore this
76 topic through the variety of message channe ls by which public information campaigns are disseminated. Broadcast spots are the mo st common dissemination method for PSAs; therefore, research budgets a llowing, television or radio ad s are a logical channel to incorporate in future studies. In addition to examining source credibility effects for different PSA formats, future studies should also consider differing leve ls of issue salience. The stimulus in this study presented a somewhat salient issue. Ma nipulating issue salien ce as an independent factor along with source type may reveal how different sources are perceived for high versus low salient issues. In marketing and advertising research, st udies have begun to compare the effects of corporate sponsors versus spokespersons on message effects. This study did not include a spokesperson. However, PSAs typi cally use spokespersons to deliver the message. Comparing the role of sponsor credibil ity to spokesperson credibility will reveal more about the effects of organizational s ponsors on message outcomes. Future studies should include message spokesperson as a variable. This study manipulated organizational status and level of source awareness as the independent variables. Credibility studies ar e concerned with the effect that differing perceptions of credibility have on message effe cts. Therefore, it is desirable to manipulate participantsÂ’ perceptions of credibility. In addition to ma nipulating the status of the sponsoring organization, future research may also manipulate credibility for the different sponsors as well. Doing so is likely to reveal more about the relationship between source credibility and the message effects variables addressed in this study.
77 One such relationship that warrants furt her study is the posit ive correlation found between source credibility and information seeking. Unlike the relationship between source credibility and behavioral intent, this relationship has not been studied before and more investigation is necessary be fore any conclusions can be drawn. As noted, the relationship between the situ ational theory variables and behavioral intent is a major finding in this study. Tes ting the relationship between the situational theory variables and behavioral intent will allow for an expansion of the theory. In order for the inclusion of behavioral intent to tr uly expand the theory, th e role of constraint recognition, which was excluded from this study, must also be investigated. The theory of planned behavior, the theory or reasoned actions successor, includes the variable behavioral control to account fo r the role that self-efficacy plays in behavior. In some ways, recognizing the constraint s that prevent one from perf orming an action is akin to recognizing the level of cont rol one has over that action. Implications for Practice Reid, Soley and Vanden Bergh (1981) sugge sted that commercial sponsors should avoid open identification for advocacy advertisin g. The results of this study indicate that increased awareness of the corporate s ponsor through open identification will not necessarily have negative impli cations. In fact, the positive cr edibility ratings suggest that for-profit sponsors can benefit from open iden tification when it comes to public service messages. The perceived altruistic nature of the message may foster greater perceptions of credibility for the organizations. Spons oring PSAs may not only help for-profit organizations build brand awareness, but also contribute to their phi lanthropic efforts and foster goodwill with key publics.
78 Hammond (1986) suggested that commerc ial sponsors partner with non-profit organizations to increase their credibility a nd improve message effectiveness. For-profit organizations often use front organizations or create a foundation for the purpose of disseminating public service messages. This stud yÂ’s findings indicate that neither strategy is necessary. The changing nature of PSA dissemination has involved corporate sponsors in the process. The results indicate that it is feasible for non-profit organizations and government agencies to turn to corporat e sponsors to address PSA production and dissemination problems. In terms of health messages, these sponsors appear to be as effective as non-profit sponsors at motivating information seeking and behavioral intent. It seems that the issue is not whether cor porate sponsor inhibit these message effects, rather the issue is how to increase thes e responses regardless of message source. Participants in this study reported high levels of pr oblem recognition and personal relevance but only moderate levels of behavi oral intent and low levels of information seeking. This indicates a challenge for me ssage developmentÂ—to create high enough levels of problem recognition and personal releva nce that can translate into greater intent to seek information and perform advocated behaviors.
79 REFERENCES Ajzen, I. (1985). From intentions to actions: a theory of planned behavior. In J. Kuhland & J. Beckman (Eds.), Action-control: from cognitions to behavior (pp. 11-39). Heidelberg: Springer. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50 (179-211). Andsager, J. L., Austin, E. W., & Pinklet on, B. E. (2001). Questioning the value of realism: young adults' processing of messa ges in alcohol-related public service announcements and advertising. Journal of Communication, 51 (1), 121-142. Atkin, C. K. (2000). Theory and Prinicples of Media Health Campaigns. In R. E. Rice & C. K. Atkin (Eds.), Public Communicaiton Campaigns (3rd ed., pp. 49-68). Thousand Oaks: Sage Publications, Inc. Atkin, C. K. (2001). Impact of Public Service Adve rtising: Research Evidence and Effective Strategies Menlo Park, CA: The Henry J. Kaiser Family Foundation. Berlo, D. K., Lemert, J. B., & Mertz, R. J. (1969). Dimensions for Evaluating the Acceptability of Message Sources. The Public Opinion Quarterly, 33 (4), 563-576. Borzekowski, D. L. G., & Poussaint, A. F. (1999). Public service announcement perceptions: A quantitative examination of anti-violence messages. American Journal of Preventive Medicine, 17 (3), 181-188. Caruana, A. (1997). Corporate reputation: concept and measurement. Journal of Product & Brand Management, 6 (2), 109-118. Chaiken, S., Liberman, A., & Eagly, A. H. (1989). Heuristic and systematic information processing within and beyond the persuasion context. In J. S. Uleman & J. A. Bargh (Eds.), Unintended Thought New York: Guilford Press. Christensen, T. P., Ascione, F. J., & Bago zzi, R. P. (1997). Understanding How Elderly Patients Process Drug Informa tion: A Test of a Theory of Information Processing. Pharmaceutical Research, 14 (11), 1589-1596. Clarke, P., & Kline, F. G. (1974). Mass Media Effects Reconsidered: Some New Strategies for Communication Research. Communication Research, 1 224-270.
80 Cline, R. J., & Engel, J. L. (1991). Colle ge student's perceptions of sources of information about AIDS. Journal of American College Health, 40 (2), 55-63. Cohen, J., Mutz, D., Price, V., & Gunther, A. (1988). Perceived Impact of Defamation: An Experiment on Third-Person Effects The Public Opinion Quarterly, 52 (2), 161-173. Davison, W. P. (1983). The ThirdPerson Effect in Communication The Public Opinion Quarterly, 47 (1), 1-15. Deshpande, S., & Hitchon, J. C. (2002). Caus e-related marketing ads in the light of negative news. Journalism and Mass Communication Quarterly, 79 (4), 905. Duck, J. M., & Mullin, B. A. (1995). Th e Perceived Impact of the Mass Media: Reconsidering the Third Person Effect. European Journal of Social Psychology, 25 (1), 77-93. Duck, J. M., Terry, D. J., & Hogg, M. A. (1995). The Perceived Influence of AIDS Advertising: Third Person Effects in the Context of Positive Media Content. Basic and Applied Social Psychology, 17 (3), 305-325. Dutta-Bergman, M. (2003). Trusted Online Sources of Health Information: Differences in Demographics, Health Beliefs, and Health-Information Orientation. Journal of Medical Internet Research, 5 (3). Eveland, W. P., & McLeod, D. M. (1999). The E ffect of Social Desi rability on Perceived Media Iimpact: Implications for Third-Person Perceptions. International Journal of Public Opinion Research, 11 (4), 315-333 Fishbein, M. (2000). The role of theory in HIV prevention. AIDS Care, 12 (3), 273-278. Fishbein, M., & Ajzen, I. (1975). Belief, Attitude, Inten tion and Behavior: An Introduction to Theory and Research Reading, Mass: Addison-Wesley. Fishbein, M., & Cappella, J. N. (2006). The Role of Theory in Developing Effective Health Communications. Journal of Communication, 56 S1-S17. Frewer, L. J., Howard, C., Hedderley, D., & Sh epherd, R. (1996). What Determines Trust in Information About Food-Related Risks? Underlying Psychological Constructs. Risk Analysis, 16 (4), 473-486. Gaziano, C., & McGrath, K. (1986). M easuring the Concept of Credibility. Journalism Quarterly, 63 (3), 451-462. Goldberg, M. E., & Hartwic k, J. (1990). The Effects of Advertiser Reputation and Extremity of Advertising Claim on Advertising Effectiveness. The Journal of Consumer Research, 17 (2), 172-179.
81 Goldsmith, R. E., Lafferty, B. A., & Newell, S. J. (2000a). The impact of corporate credibility and celebrity credibility on consumer reaction to advertisements and brands. Journal of Advertising, 29 (3), 43. Goldsmith, R. E., Lafferty, B. A., & Newell, S. J. (2000b). The Influence of Corporate Credibility on Consumer Attitudes and Purchase Intent Corporate Reputation Review,, 3 (4), 304-318. Gotlieb, J., & Sarel, D. (1991). Comparative Advertising Effectiveness: The Role of Involvement and Source Credibility. Journal of Advertising, 20 (1), 38-45. Grunig, J. E. (1989). Sierra St udy Shows Who Become Activists. Public Relations Review, 15 (3), 3-24. Grunig, J. E. (1997). A situational theory of publics: Conceptual history, recent challenges and new research. In D. Mo ss, T. McManus & D. Vercic (Eds.), Public relations research: An international perspective (pp. 3Â–48). Boston: International Thomson Business. Grunig, J. E., & Hunt, T. (1984). Managing public relations New York: Holt, Rinehart and Winston. Grunig, J. E., & Ipes, D. A. (1983). The an atomy of a campaign against drunk driving. Public Relations Review, 9 (2), 36-52. Gunther, A. C. (1992). Biased Press or Bias ed Public? Attitudes Toward Media Coverage of Social Groups The Public Opinion Quarterly, 56 (2), 147-167. Gunther, A. C., & Thorson, E. (1992). Pe rceived Persuasive Effects of Product Commercials and Public Service Announ cements: Third-Person Effects in New Domains. Communication Research, vol, 19 (5), 574-596. Haley, E. (1996). Exploring the construct of organization as source: Consumer's understandings of organizational sp onsorship of advocacy advertising. Journal of Advertising, 25 (2), 19. Hammond, S. L. (1986). Health Advertising: Th e Credibility of Orga nizational Sources. Communication Yearbook Harris Interactive (August 8, 2007) Some Substantial Changes in How Americans View Different Industries [Electronic Version]. The Harris Poll #79. Retrieved 11/09/2007 from http://www.harrisinteractive.com/ harris_poll/printerfriend/index.asp Homer, P. M., & Kahle, L. R. (1990). Source expertise, time of s ource identification, and involvement in persuasion: an el aborative processing perspective. Journal of Advertising
82 Hovland, C. I., & Weiss, W. (1951). Th e Influence of Source Credibility on Communication Effectiveness. The Public Opinion Quarterly, 15 (4), 635-650. Hyman, H. H., & Sheatsley, P. B. (1947). Some Reasons Why Information Campaigns Fail. The Public Opinion Quarterly, 11 (3), 412-423. LaBarbera. (1982). Overcoming No-Reputa tion Liability Through Documentation and Advertising Regulation. Journal of Marketing Research, 19 (2), 223-228. Lafferty, B. A., & Goldsmith, R. E. (1999). Corporate CredibilityÂ’s Role in ConsumersÂ’ Attitudes and Purchase Intentions When a High versus a Low Credibility Endorser Is Used in the Ad. Journal of Business Research 44 (2), 109-116. Lafferty, B. A., Goldsmith, R. E., & Newell, S. J. (2002). The dual credibility model: The influence of corporate and endorser cr edibility on att itudes and purchase intentions. Journal of Marketing Theory and Practice, 10 (3), 1-12. Lambe, J. L., & McLeod, D. M. (2005) Understanding Third-Person Perception Processes: Predicting Perceived Impact on Self and Others for Multiple Expressive Contexts. Journal of Communication, 55 (2), 277-291. Ledingham, J. A. (1993). The kindness of st rangers: Predictor va riables in a public information campaign. Public Relations Review, 19 (4), 367-384. Liesse, J. (1990, October 8, 1990). Line betw een public service, paid ads blurs. Advertising Age, 28. Lynn, J. (1973). Perception of Public Service Advertising: Source, Message and Receiver Effects. Journalism Quarterly, 50 673-679, 689. Lynn, J., Wyatt, R., Gaines, J., Pearce, R ., & Bergh, B. V. (1978). How Source Affects Response to Public Service Announcements. Journalism Quarterly, 55 (4), 716720. MacKenzie, S. B., & Lutz, R. J. (1989). An Empirical Examination of the Structural Antecedents of Attitude toward the Ad in an Advertising Pretesting Context. Journal of Marketing, 53 (2), 48-65. Madden, T. J., Ellen, P. S., & Ajzen, I. (1992). A comparison of the theory of planned behavior and the theo ry of reasoned action. Personality and Social Psychology Bulletin, 18 (1), 3-9. McCroskey, J. C., & Young, T. J. (1981). Ethos and Credibility: The Construct and It's Measurement after Three Decades. Central States Speech Journal, 32 24-34. McGuire, W. J. (2001). Input and Output Va riables Currently Promising for Constructing Persuasive Campaigns. In R. E. Rice & C. K. Atkin (Eds.), Public
83 Communication Campaigns (3rd ed., pp. 22-48). Thousand Oaks: Sage Publications, Inc. Mendelsohn, H. (1973). Some Reasons W hy Information Campaigns Can Succeed. The Public Opinion Quarterly, 37 (1), 50-61. Meyers, J. (1989). Government groups look beyond PSAs. Advertising Age, 22. Miller, M. D., & Levine, T. R. (1996). Pers uasion. In M. B. Salwen & D. W. Stacks (Eds.), An Integrated Approach to Communication Theory and Research (pp. 261-276). Mahwah, New Jersey: Lawrence Erlbaum Associates, Publishers. Moynihan, R., Heath, I., & Henry, D. (2002) Selling sickness: the pharmaceutical industry and disease mongering. BMJ, 324 (7342), 886-890. Muehling, D. D. (1987). An investigation of factors underlying attitude-towardadvertising-in-general. Journal of Advertising Newell, S. J., & Goldsmith, R. E. (2001). The development of a scale to measure perceived corporate credibility. Journal of Business Research, 52 (3), 235-247. O'Keefe, G. J., & Reid, K. (Eds.). (1990). The Uses and Effects of Public Service Advertising (Vol. 2): Lawrence Erlbaum Associates. Ohanian, R. (1990). Construction and Validat ion of a Scale to Measure Celebrity Endorser's Perceived Expertise, Trus tworthiness, and Attractiveness. Journal of Advertising, 19 (3), 39-52. Petty, R. E., & Cacioppo, J. T. (1986). Communication and Persuasion: Central and Peripheral Routes to Attitude Change New York: Springer-Verlag. Reid, L. N., Soley, L. C., & Vanden Bergh, B. G. (1981). Does source affect response to direct advocacy print advertisements? Journal of Business Research, 9 (3), 309019. Self, C. C. (1996). Credibility. In M. B. Salwen & D. W. Stacks (Eds.), Integrated Approach to Communicati on Theory and Research (pp. 421-442). Mahwah, New Jersey: Lawrence Erlbaum Associates, Publishers. Settle, R. B., & Golden, L. L. (1974). Attr ibution Theory and Advertiser Credibility. Journal of Marketing Research, 11 181-185. Stern, B. B. (1994). A revised communication model for advertising: Multiple dimensions of the source, the message, and the recipient. Journal of Advertising, 23 (2), 5.
84 Sternthal, B., Dholakia, R., & Leavitt, C. (1978). The Persuasive Effect of Source Credibility: Tests of Cognitive Response. The Journal of Consumer Research, 4 (4), 252-260. Sternthal, B., Phillips, L. W ., & Dholakia, R. (1978). The Persuasive Effect of Source Credibility: A Situational Analysis The Public Opinion Quarterly, 42 (3), 285-314. U.S. Department of Health and Human Services (1992). Making health communication programs work: A planner's guide (NIH Publication No. 92-1493) Bethseda, MD: National Cancer Institute. Walster, E., Aronson, E., & Abrahams, D. (1966) On increasing the persuasiveness of a low prestige communicator. Journal of Experimental Social Psychology Ward, C. D., & McGinnies, E. (1974). Persuasi ve effects of early and late mention of credible and noncredible sources. Journal of Psychology, 86 17-23. White, H. A. (1997). Considering Interac ting Factors in the Third-Person Effect: Argument Strength a nd Social Distance. Journalism and Mass Communication Quarterly, 74 (3), 557-564 Zaltman, G., & Duncan, R. (1977). Strategies for planned change New York: Wiley.
86 Appendix A.1: Participant Directions Directions: Read and follow all of the instructions below. Do not open the envelope until instructed. 1. This packet contains a brochure and a que stionnaire booklet. When instructed, open the envelope and remove only the brochure. Leave the questionnaire booklet inside. 2. Take as long as you need to read the enti re brochure. Then, return the brochure to the envelope and remove the questionnaire. Do not refer to the brochure again 3. Read the instructions for each section carefully. After answering all questions, return the booklet to the enve lope and reseal the envelope. The packet will be collected when everyone is finished
87 Appendix B.1: Subtle NP Treatment
88 Appendix B.2: Subtle FP Treatment
89 Appendix B.3: Detailed NP Treatment
90 Appendix B.4: Detailed FP Treatment
91 Appendix B.5: Control Treatment
92 Appendix C.1: Manipulation Check Instructions: The following questions pertain to the h ealth message you just read. Please answer as honestly as possible. 1. On the following scale, where 1 represents strongly disagree and 5 represents strongly agree, please indicate your level of agreement with the following statement by marking the appropriate box: I recall the name of the organization that sponsored this brochure. Strongly Strongly Disagree Agree DonÂ’t Know 2. On the following scale, where 1 represents strongly disagree and 5 represents strongly agree, please indicate your level of agreement with the following statement by marking the appropriate box: It was easy to identify the organizat ion that sponsored this information. Strongly Strongly Disagree Agree DonÂ’t Know 3. Do you recall if the organization that sponsored this information was a Government Agency Non-profit Organization Health Care Facility Pharmaceutical Company I donÂ’t recall There was no sponsor 4. Please write the organizationÂ’s name in the space provided below. __________________________________________________________________________
93 Appendix D.1: Measurement Instrument Instructions This booklet contains questi ons about your impression of the health message you just read. Please answer as hones tly as possible. There are no right or wrong answers. Your responses will remain completely conf idential. Thank you for completing the questionnaire. Section I: Please respond by marking the appropriate box. 1. On the following scale, where 1 represents No thing and 5 represents A lot, please indicate how much you knew about HPV prior to reading this brochure. 1 2 3 4 5 Nothing A lot No opinion 2. On the following scale, where 1 represents Not at all and 5 represents Frequently, please indicate how frequently you have come across in formation about HPV in the last 12 months. 1 2 3 4 5 Not all Frequently No opinion 3. Do you know anyone who has been diagnosed with HPV? Yes No Unsure Section II: The following questions ask your opinion of the organization that produced the brochure. On the followi ng scales, where 1 represents Strongly Disagree and 5 represents Strongly Agree, pl ease indicate your level of agreement with the following statements. 4. The organization is qualified to provide information about HPV. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion
945. The organization can be trusted to pr ovide factual information about HPV. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion 6. The organization is concerned with the publicÂ’s well being. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion 7. The organization is not an expert on HPV. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion 8. The organization cannot be trusted to pr esent reliable information about HPV. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion 9. The organization is concerned with making profits. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion 10. I believe the organization provides unbiased information about HPV. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion 11. I believe the organization is knowledgeable about HPV. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion
9512. I believe the organization has something to gain from publishing this information. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion Section III: Finish the statement below by chec king the box that best represents your opinion for each of the items listed. Compared to most brochures I have seen about health issues, I found this brochure 13. Boring Interesting 14. Unprofessional Professional 15. Misleading Accurate 16. Dull Exciting 17. Deceptive Truthful 18. Overemphasizes Downplays Section IV: On the following scales, where 1 represents Strongly Disagree and 5 represents Strongly Agree, please indicate your level of agreement with the following statements. 19. HPV is a serious health problem. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion 20. People should be concerned about the risks of HPV. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion
9621. HPV can have serious complications 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion Section V: On the following scales, where 1 represents Less Relevant and 5 represents More Relevant, please indicate your response to the following questions. 22. Has this brochure made the issue of HPV more relevant or less relevant for you? 1 2 3 4 5 Less More Relevant Relevant No opinion 23. Do you think this brochure made the issue of HPV more relevant or less relevant for other students in the class? 1 2 3 4 5 Less More Relevant Relevant No opinion 24. Do you think this brochure would make the issue of HPV more relevant or less relevant for college students in general? 1 2 3 4 5 Less More Relevant Relevant No opinion Section VI: On the following scales, where 1 represents Not Likely and 5 represents Very Likely, please indicate ho w likely is it that you will perform the actions described as a resu lt of reading the brochure. 25. I will visit the web site to learn more about HPV. 1 2 3 4 5 Not Very Likely Likely No opinion
9726. I will call the toll-free number to request the HPV information kit. 1 2 3 4 5 Not Very Likely Likely No opinion 27. I will ask a health professional about HPV risks. 1 2 3 4 5 Not Very Likely Likely No opinion 28. I will pick up a pamphlet about HPV from the student health center. 1 2 3 4 5 Not Very Likely Likely No opinion Section VII: On the following scales, where 1 represents Not Likely and 5 represents Very Likely, please indicate ho w likely is it that you will perform the actions described as a resu lt of reading the brochure. 29. I intend to tell a friend about HPV. 1 2 3 4 5 Definitely Definitely Will Will Not No opinion 30. I intend to get tested for HPV. 1 2 3 4 5 Definitely Definitely Do Do Not No opinion 31. I intend to get vaccinated against HPV. 1 2 3 4 5 Definitely Definitely True False No opinion
98Section IIX: Please respond by marking the appropriate box. On the following scales, where 1 represents Str ongly Disagree and 5 represents Strongly Agree, please indicate your level of agreement with the following statements 32. I recall the name of the organizati on that sponsored this brochure. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion 33. It was easy to identify the organization that sponsored this information. 1 2 3 4 5 Strongly Strongly Disagree Agree No opinion 34. Do you recall if the organization that sponsored this information was a Government Agency Non-profit organization Health Care Facility Pharmaceutical Company I donÂ’t recall No sponsor 35. Please write the organizationÂ’s name in the space provided below ___________________________________________________________________________ Section IX: The following questions will help us understand your answers. Please respond by marking the appropriate box. 36. Please indicate your academic rank: Freshman Sophomore Junior Senior Other _________________________________ 37. What college are you in? Arts/Sciences Business Education Engineering Honors College Medicine Nursing Public Health Visual/Performing Arts 38. What is your gender? Female Male 39. What is your age? _______________________ Thank you for participating in this study!
99 Appendix E.1: Data Coding Sheet Section I: Prior Knowledge Var. 1 knowledge Scale: nothing/a lot How much you knew about HPV prior to reading this brochure. Var. 2 exposure Scale: not at all/frequently How frequently you have come across information about HPV in the last 12 months. Var. 3 diagnosis Categorical 1=yes; 2=no; 3=unsure; 9=no response Do you know anyone who has been diagnosed with HPV? Section II: Source Credibility Var. 4 expert_qualified Scale: SD/SA The organization is qualified to provide information about HPV. Var. 5 trust_factual Scale: SD/SA The organization can be tr usted to provide factual information about HPV. Var. 6 concern_well being Scale: SD/SA The organization is concer ned with the publicÂ’s well being. Var. 7 expert_not Scale: SD/SA The organization is not an expert on HPV. Var. 8 trust_not reliable Scale: SD/SA The organization cannot be tr usted to present reliable information about HPV. Var. 9 concern_profit Scale: SD/SA The organization is concer ned with making profits. Var. 10 trust_unbiased Scale: SD/SA I believe the organization pr ovides unbiased information about HPV. Var. 11 expert_knowledge Scale: SD/SA I believe the organization is knowledgeable about HPV. Var. 12 concern_gain Scale: SD/SA I believe the organization has something to gain from publishing this information.
100Section III: Message Credibility Var. 13 boring_interesting Scale: boring/interesting I found this brochure boring/interesting Var. 14 unprof_professional Scale: unprofessional/professional I found this brochure unpr ofessional/professional Var. 15 mislead_accurate Scale: misleading/accurate I found this brochure misleading/accurate Var. 16 dull_exciting Scale: dull/exciting I found this brochur e dull/exciting Var. 17 deceptive_truthful Scale: deceptive/truthful I found this brochure deceptive/truthful Var. 18 over_down Scale: overemphasizes/downplays I found this brochure overemphasizes/downplays Section IV: Problem Recognition Var. 19 problem_serious Scale: SD/SA HPV is a serious health problem. Var. 20 problem_concern Scale: SD/SA People should be concerned about the risks of HPV. Var. 21 problem_complications Scale: SD/SA HPV can have serious complications. Section V: Personal Relevance Var. 22 relevant_self Scale: LR/MR Has this brochure made the issue of HPV more relevant or less relevant for you? Var. 23 relevant_class Scale: LR/MR Do you think this brochure made the issue of HPV more relevant or less releva nt for other students in the class? Var. 24 relevant_other Scale: LR/MR Do you think this brochure would make the issue of HPV more relevant or less relevant for college students in general?
101Section VI: Information Seeking Var. 25 info_web Scale: not likely/very likely I will visit the web site to learn more about HPV. Var. 26 info_toll free Scale: not likely/very likely I will call the toll-free number to request the HPV information kit. Var. 27 info_doctor Scale: not likely/very likely I will ask a health professional about HPV risks. Var. 28 info_clinic Scale: not likely/very likely I will pick up a pamphlet about HPV from the student health center. Section VII: Behavioral Intent Var. 29 tell someone Scale: definitely will/will not I intend to tell a friend about HPV. Var. 30 tested Scale: definitely do/do not I intend to get tested for HPV. Var. 31 vaccinated Scale: definitely true/false I intend to get vaccinated against HPV. Section IIX: Source Identification Var. 32 name_recall Scale: SD/SA I recall the name of the organization that sponsored this brochure Var. 33 identify_easy Scale: SD/SA It was easy to identify the organization that sponsored this information Var. 34 org_status Categorical 1= gov.; 2=non-profit; 3 = health facility; 4=pharmaceutical; 5= no recall 6 = no sponsor; 9 = no response Do you recall if the organization that sponsored this information was a Government Agency/Non-profit organization/Health Care Facility/Pharmaceutical Company/I donÂ’t recall/No sponsor Var. 35 name Categorical 1=correct; 2=incorrect 9 = no response Please write the organizationÂ’s name in the space provided
102Section IX: Demographic Information Var. 36 academic rank categorical Please indicate your academic rank: Var. 37 college categorical What college are you in? Var. 38 age ratio What is your age?
103 Appendix F: Correlations Source Credibility Message Credibility Problem Recognition Personal Relevance Information Seeking Behavioral Intent Source Credibility Pearson Co rrelation 1.490.272.437.170.317 Sig. (2 tailed) .000.000.000.022.000 N 187185184169180174 Message Credibility Pearson Co rrelation .4901.190.306.102.239 Sig. (2 tailed) .000 .005.000.142.001 N 185219216199211200 Problem Recognition Pearson Co rrelation .272.1901.325.218.356 Sig. (2 tailed) .000.005 .000.002.000 N 184216218198209199 Personal Relevance Pearson Co rrelation .437.306.3251.312.422 Sig. (2 tailed) .000.000.000 .000.000 N 169199198200192184 Information Seeking Pearson Correlation .188.8.131.5221.541 Sig. (2 tailed) .022.142.002.000 .000 N 180211209192213199 Behavioral Intent Pearson Corre lation .317.239.356.422.5411 Sig. (2 tailed) .000.001.000.000.000 N 174200199184199203