|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 ORGANZATIONAL BLOGS AND CANCER MESSAGES: EXPERIMENTAL TESTING OF RECIPIENT AND MESSAGE FACTORS IN INFLUENCING ATTITUDES, INTENTION AND BEHAVIOR By PAULA RAUSCH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERS ITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010
2 2010 Paula Rausch
3 T o my family and friends : Without your love, support, confidence, and encouragement I could not hav e fulfilled my dream. You have braved the challenges of the last four years with me, and I am honored to share the success of this accomplishment with you
4 ACKNOWLEDGMENTS The past four years have been a journey I never expected to take, a dream I did know I had, a peak I never saw on my horizon. The expedition has been arduous and bumpy at times, one that left me both breathless and out of breath, one that complicated things greatly and yet clarified what was important one that taught me so much a nd also how much I still have to learn. Among my most valuable realization s is that I have not braved these challenge s alone, and t his dissertation would not have been possible with out the love and support of many people. I especially wish to thank Dr. De bbie Treise, my dissertation chair, mentor, and friend. She has truly encouraged and supported me throughout my graduate career, both personally a nd profession ally. Her energy, enthusiasm, easy laugh and remarkable spirit has allowed me not only to survive the demanding enterprise that earning a doctoral de gre e is but to thrive and achieve things that only a few years ago I w ould not have considered She has been cons tantly at my side throug h the ups and downs and m y life has been greatly enriched as a result of knowing her. Dr. Michael Weigold also deserves special thanks He spent countless hours helping me t ransform my dissertation vision into reality and guiding me through the complicated procedures that allowed my experiment to be completed successfully. He forced me to think things through and to keep pushing until I got it right, boost ing both my confidence and my sense of accomplishment. His excepti onal sense of humor and willingness to go above and beyond to help me succeed underscore the person he is, and he has contributed immeasurably to my graduate career and to my life. I also want to thank my other commit tee members. Dr. Virginia Dodd provided invaluable expertis e, recognized the value of my experience, encouraged me, and
5 generously provided opportunities to expand my perspectives and skills. Dr. Youjin C hoi and her keen insight challenged me to dig deep a nd consider alternative explanations This effort also would not have been possible without the infinite love and support of my family. My pare nts, Marian and Don Rausch, have accompanied me on each and every journey I have taken throughout my life belie ving in me and cheering me on every step of the way My mom ha s l istened endlessly, paying unexpected attention while I chattered on about my dissertation research, asking insightful questions posing intriguing possibilities, and jus t being interested in this endeavor tha t has engulfed my world She is my advocate, my confidante, and my lifelong friend My dad is the rock on which my lif e rests solid, dependable and a lways there for me He believe s in my abilities and lets me know i n so many ways how important I am to him My sister, Lynn, has been a sympathetic ear, provid ing the calm I needed throughout this process and help ing me keep things in perspective I am who I am because of them, and it is their love and encouragement th at allowed me to follo w and achieve my dre ams. I have also been inordinately lucky to have n umerous friends who stuck with me throughout this process, inspiring and uplifting me with laughter, distract ion, and fun Kathy DeWees Karen Zibolski, and Patrice Stefans have been h appily and inextricably root e d in my life a nd I cherish each of them and every minute of our time together Numerous others have contributed something special to my graduate career Jody Hedge and her smiling face made even the worst days be tter, and her assistance and unmatched ded ication were vit al to my completing this task. Kristin Meyer, Tom Mueller, Mike Clayton, Heather Edwards, Meredith Wang, and Richard Waters have all made the experience one I will always treasure. To a ll of these people an d to the m any more I can t possibly name I extend my love and deepest thanks.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF FIGURES ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 Need for Current Research ................................ ................................ ..................... 1 2 Role of Current Study ................................ ................................ ............................. 17 2 LITERATURE REVIEW ................................ ................................ .......................... 20 Cancer and Communication ................................ ................................ ................... 20 Blogs ................................ ................................ ................................ ................ 24 Blogs and Health ................................ ................................ .............................. 31 Persuasion Research ................................ ................................ .............................. 38 Involvement ................................ ................................ ................................ ...... 42 Source Credibility ................................ ................................ ............................. 43 Message Quality ................................ ................................ ............................... 47 Elaboration Likelihood Model ................................ ................................ ........... 48 Patient Empowerment ................................ ................................ ............................. 60 Health Information Seeking ................................ ................................ .............. 62 Assessing Empowerment ................................ ................................ ................. 64 Hypotheses an d Research Questions ................................ ................................ ..... 67 3 METHODOLOGY ................................ ................................ ................................ .... 70 Independent Variables ................................ ................................ ............................ 70 Involvement ................................ ................................ ................................ ...... 70 Message Quality ................................ ................................ ............................... 71 Source Credibility ................................ ................................ ............................. 71 Abilit y ................................ ................................ ................................ ................ 72 Dependent Measures ................................ ................................ .............................. 72 Attitudes ................................ ................................ ................................ ........... 72 Consumer Empowerment ................................ ................................ ................. 73 Other Dependent Measures ................................ ................................ ............. 75 Stimuli ................................ ................................ ................................ ..................... 76 Pretesting ................................ ................................ ................................ ................ 80 Participants ................................ ................................ ................................ ............. 84 Experimental Procedure ................................ ................................ ......................... 85
7 4 RESULTS ................................ ................................ ................................ ............... 91 The Sample ................................ ................................ ................................ ............ 91 Independent Variable Manipulation Checks ................................ ............................ 92 Hypothesis Testing ................................ ................................ ................................ 95 Attitude toward the Treatment Message ................................ ........................... 96 Attitude toward Positron Treatment ................................ ................................ .. 98 Research Question Exploration ................................ ................................ .............. 99 Attitude toward the Organization ................................ ................................ ...... 99 Attitude toward the Blog ................................ ................................ ................. 100 Empowerment ................................ ................................ ................................ 101 Standardized measures ................................ ................................ ........... 101 Information seeking items ................................ ................................ ........ 101 Other Intention and Behavioral Measures ................................ ...................... 103 5 DISCUSSION ................................ ................................ ................................ ....... 106 Involvement ................................ ................................ ................................ .......... 109 Source Credibility ................................ ................................ ................................ .. 118 Argument Strength ................................ ................................ ................................ 123 Gender Difference s ................................ ................................ ............................... 126 Implications ................................ ................................ ................................ ........... 128 For Theory ................................ ................................ ................................ ...... 128 For Practice ................................ ................................ ................................ .... 129 Limitations ................................ ................................ ................................ ............. 136 Future Research ................................ ................................ ................................ ... 138 Conclusion ................................ ................................ ................................ ............ 139 APPENDIX A COVER STORY MANIPULATIONS ................................ ................................ ...... 141 B BLOG MANIPULATIONS ................................ ................................ ...................... 143 LIST OF REFERENCES ................................ ................................ ............................. 145 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 173
8 LIST OF TABLES Table page 4 1 Participant demographic info rmation ................................ ................................ 104
9 LIST OF FIGURES Figure page 2 1 Simplified graphic representation of the ELM developed by Petty and Cacioppo (1981, 1986) ................................ ................................ ....................... 69
10 Abstract o f Dissertation Presented to the Graduate School of the University o f Florida in Partial Fulfillment o f t he Requirements for the Degree of Doctor o f Philosophy ORGANZATIONAL BLOGS AND CANCER MESSAGES: E XPERIMENTAL TESTING OF RECIPIENT AND MESSAGE FACTORS IN INFLUENCING ATTITUDES, INTENTION AND BEHAVIOR By Paula Rausch August 2010 Chair: Debbie Treise Major: Mass Communication Patients with cancer are among those having the greatest need and desire f or informatio n (Shim, Kelly, & Hornik, 2006), yet c onsiderable dissatisfaction remains regarding the amount and type of information that health professionals pro vide to those living with the disease (Arora et al.; IOM, 2007a). As a result, people consult o ther sources, with a growing number tur ning to social media platforms; however little research has examined organizational e health messages in blogs (Kim & Chung, 2007). Using the Elaboration Likelihood Model as a framework, this study sought to extend u nderstanding in this area by investigating experimentally the message processing that occurs among consumers of novel cancer treatment messages disseminated through a blog sponsored by a health organization and their effects on attitudes, behavioral intent ion s, and a specific action. Results of the experiment conducted among 152 undergraduate students show some support in a cancer trea tment context for hypotheses predic ted by ELM research in other contexts; however, the results are complex and not always as anticipated. Research questions consider ed the value of organizational blogs as a channel of health
11 information, whet her this two way venue provides to consumers the mu tual benefit of empowerment and the effects that involvement, argument strength and so urce credibility have on screening and other health behavioral intentions, and an information seeking action. To some extent, involvement, source credibility and argument strength each did influence some of the dependent attitude measures tested here. Amon g all subjects, source credibility influenced attitudes toward the message, positron treatment, and the organization as well as overall likelihood of seeking additional information about the cancer treatment and of taking an information seeking action Ar gument strength had a stronge r influence, affecting all four attitudinal measures (including toward the blog), as well as the probability of seeking additiona l information and of discontinuing a harmful behavior The results have both theoretical and pract ical implications, and provide empirical support for speculat ion that theories of persuasion such as the ELM that have often been overlooked with respect to health behaviors provide an additional avenue for targeting messages that can be beneficial for bot h organizations and consumers.
12 CHAPTER 1 INTRODUCTION Need for Current Research Advances in medicine have been unprecedented in recent decades, bringing with disease s that once meant certain death. What has been slower is the recognition by health organizations and providers that caring for patients with illnesses such as cancer involves much more than treating just their bodies; it is also essential to minister to their ps ychosocia l needs in order to ensure that the quality of t heir lives is the best it can be Among the se i s the necessity for patients to have access to reliable, accurate, and unders tandable health information they can use to manage their health and illness es, and to help them make informed decisions about care and treatment (Institute of Medicine [IOM], 2007a). Evidence suggests that barriers preventing patients from accessing needed health inform ation results in decreased well being on a variety of levels ( Arora, Johnson, Gustafson, McTavish, Hawkins, & Pingree, 2002). Patients with cancer, the second leading cause of death amo ng U.S. adults, are among those having the greatest need and desire for information (Shim, Kelly, & Hornik, 2006), in part because o f the potentially life threatening nature of their disease, th e specialized testing and treatments associated with each type of cancer, and the rapidly changing and highly complex nature of many of the associated therapies. In a report about cancer, which more than 40% of all Americans will be diagnosed with at some point in their lives, the IOM went as far as to state that neglecting to address these non p. 3). Despite this acknowledgement, considerable dissatisfaction
13 remains regarding the amount and type of information that health professionals provide to those living with cancer about the specifics of their individual disease, the best therapies to tr eat it, and the most effective ways to deal with the potential side effects of these treatments and to cope with the illness more broad ly (Arora et al.; IOM, 2007a). As a result, many cancer patients and their families take on the added responsibility of finding the information they want or need elsewhere, a growing trend in all areas of health. Studies routinely show that mass media are the primary means by which a majority of people obtain their health information (Dutta Bergman, 2004; IOM, health related belief Casarett, & Davis, 2006, p. 291). Obtaining health information is also associated with improved knowledge about cancer, and increased likel ihood of undertakin g healthy preventive lifestyles and obtaining some cancer screenings (Shim et al., 2006). The Internet has become a particularly important source of health and cancer information. A survey found that 80% of the 147 million A mericans who use the Internet have done so at one time to look for health or medical information, with more than half of those saying it had some effect on how they cared for themselves or someone else (Fox, 2006 2008 ). There is widespread agreement, including by hea lth care professionals and the federal government, that the Internet can provide tremendous benefits to people searching for health information (Science Panel, 1999). One example vel, which is
14 (U.S. Department of Health and Human Services, 2000; IOM, 2004). A nation al health literacy assessment of more than 19,000 adults showed that average health literacy health information from the Internet than among those who received this amount of inf ormation from any other communication source, including health professionals (Kutner, Greenberg, Jin, & Paulsen, 2006) Studies also have shown behavior change is related to obtaining health information from the Internet, with those using it more frequentl y for this purpose more likely to positively modify their health behaviors (Ayers & Kronenfeld, 2007). These days, health care organizations and companies both legitimate and not routinely have World Wide Web sites on which they promote their services, wi th cancer treatments becoming a more common topic, as evidenced by a Google search in November 2008 that returned more than 21 million results. These results include sources of health information considered to be highly credible, such as the National Cance r Institute ( NCI), which topped the list of resu lts as well as a plethora of organizations have paid for space to sell their products and services. The ubiquity of information avail able through the Internet and the accompanying disparate quality found there, including messages promoting health claims often not backed by scientific evidence, can be problematic, with no regulation over the data posted online and often little disclosure of the qualifications of those distributing the information or promoting the services (Berland, Elliot t, Morales, Algazy, Kravitz, & Broder, 2001; Haard, Slater, & Long, 2004; Sandberg, 2005; Vivian, 2009). As health care becomes increasingly
15 consumer dri ven, with more patients becoming involved in their own care, having a significantly greater say in their treatments and more often bypassing traditional sources of information and obtaining it directly from Internet sources (Haard et al.; Hesse, Moser, Ru tten, & Kreps, 2006; Petersen, Heesacker, Schwartz, & Marsh, 2000), so too will the associated problems grow. This trend drives home the imperative for legitimate health organizations to distribute accurate, trustworthy, and comprehensible information to p otential health consumers. Weblogs, or blogs, are one such message type. They have become so widely used in hat has hit the public and private health groups, are increasingly using them to engage, and connect and establish relationships with their consumers and publics Blogs were init ially seen mainly in the journalism and political realms (Hewitt, 2005); however, use of these online diaries has been exploding on all topics in the last few years, including those related to health. In addition to providing information, blogs create a ve nue for conversation and dialogue with the authors and others who are reading and commenting on the messages posted there, which is becoming increasingly important not only to consumers but to the business world (Technorati; Flynn, 2006). Said Richard Edel man, president and chief executive o fficer of Edelman, the world's largest independent public relations firm: (Technorati). Since September 2004, Edelman has been writing a blog spo nsored by his company, which has 3
16 relations, marketing and advertising contexts, and maj or corporations, including Wa lm art, General Electric, and Microsoft are using blogs to influence public opinion through internal and external blogs and bloggers (Barbaro, 2006; Flynn). In addition, the personally relevant information, social support and interactivity that blogs foster has been found to be valuable among patients, particularly those with often life changing diseases suc h as cancer (Science Panel). Considerable research has been conducted concerning the success of programs and campaigns aimed at preventing certain types of cancer by promoting healthy lifestyles and behavior changes, including eating adequate amounts of fruit and vegetables, ceasing to smoke, maintaining a healthy weight, increasing activity and exercise, and undertaking preventive screenings (McGuire, 199 9). In addition, such campaigns often are developed based on theories of behavior change such as the Health Belief Model, the Theory of Reasoned Action, and the Stages of Change Model that identify the beliefs and attitudes that need to be influenced, with little, if any, consideration of how messages might be able to induce the desired outcomes (Slater, 2006) To date, these health communication campaigns have overall had relatively low levels of success in changing behavior, both in prompting people to ha lt unhealthy behaviors or to adopt constructive ones ( Slater, 2006 ). Similarly, l ittle research has examined the persuasive influence that messages from health care organizations have on potential patients, despite inc reased recognition that they are more actively involved as consumers in choosing their health providers, care and treatment, and where they receive care (McCullough & Dodge, 2002). Further, no research could be found
17 investigating the persuasive influence that cancer treatment information dist ributed by health organizations may have on health consumers diagnosed with the disease and who may be seeking non traditional, novel, and/or alternative therapies, or who just want to find out about all the ir possibl e options, including those their health providers may not be aware of and/or have dis cussed with them. R esearch examining the use of blogs for disseminating health information, including cancer, is also limited (Kim & Chung, 2007), and studies investigating the persuasive influence of organizati onal blogs about any health related subject, including about cancer treatments could not be found. Role of Current Study Thus, the purpose o f this research is to better understand the persuasive effects of online organizational e health messages that cons umers may receive about a cancer treatment in order to gain insight that will be beneficial to health care organizations, message creators, and cancer patients More specifically, this study seeks to extend existing literature by focusing attention on a pe rsuasive theory of message processing and the resultant message effects by investigating experimentally the roles that personal involvement argument strength and source credibility play in determining the extent to which consumers elaborate on cancer tre atment mes sages disseminated through a blog sponsored by a health organization The Elaboration Likelihood Model (ELM), a dual processing theory of persuasion found to be predictive in other consumer contexts, is used as a framework to examine how the thre e variables described might the route central or peripheral they use to process this type of health information This study seeks to investigate the influence these persuasive messages have on various attitudes of consumers and on their behavioral intention to seek additional information about a
18 cancer treatment and to participate in cancer screening According to the ELM, individuals for whom messages do not have high personal relevance are no t motivated to think extensively about them and will tend to process information based on peripheral features, such as the credibility of the source who delivers the communication, rather than through the central route, which focuses on the cognitive proce sses involved in deeply thinking about issue relevant arguments (Petty & Cacioppo, 1986). Further, although both routes can result in persuasion, experimental evidence has found the central route to have advantages, including longevity of attitude and grea ter likelihood of behavior change (Petty & C acioppo, 1981, 1986 ). This study also will expand examination of these three variables to dependent measures that have not been investigated previously through the lens of the ELM, including attitude toward a hea lth organization and a channel of heal th communication, namely blogs. Not only will this study extend knowledge about the ELM into a new domain, increasing knowledge about its usefulness in another context which other researchers have suggested would be b eneficial (McCullough & Dodge, 2002), the understanding gained from this study also can help health organizations target their marketing and public relations messages to better meet the information needs of their various constituents. Previous research has identified both the importance of developing in Treiman, Bright, & Hesse, 2005), and the benefits that come from developing health communications that more directly target spe cific audiences and address these needs (Kreuter, Sugg Skinner, Holt, Clark, Haire Joshu, Fu et al., 2005). These advantages i nclude increased feelings of patient control (sbring, & Nrvnen, 2004) and enhanced
19 persuasion that come from matching messages unique message processing styles (Briol & Petty, 2006; Kotler, Roberto, & Lee, 2002). reinforcing the need for and benefits of add itional research in this area. Another goal of this dissertation is to broaden understanding of health eir many publics cancer consumers by examining one potential benefit: empowerment. The clear that such per sonal dispositional variables as well as situational variables such as distraction that will not be tested here may play a role in the persuasive influence of messages.
20 CHAPTER 2 LITERATURE REVIEW Cancer and Communication Cancer is a substantial threat to the health of Americans. Cancers of all types are t he second leading cause of mortality among adults in the U.S., accounting for 22.8% of all deaths in the U.S. in 2005, or an estimated 559,312 people (American Cancer Society, 2008). In 2008, more than 1.4 million Americans were expected to be diagnosed with cancer, with more than 565,000 people projected to die of the disease the same y ear (National Cancer Institute 2008). Cancer deaths killed more people than the next five most common causes put together: cerebrovascular diseases, chronic lower respiratory (Kung, Hoyert, Xu, & Murphy, 2008). In addition, while the incidence of heart disease and some other common causes of death have dropped in half or more since 1950, cancer deaths have dropped only slightly, from 193.9 to 183.8 deaths per 100,000 people in 2005 (ACS). Survival rates have been increasing over the past three decades, but even among those who do not die from it, cancer is a significant risk in American society, with one of every two men and women having a probability of developing some kind of cancer over their lifetimes (NCI). More than 11 million living U.S. adults have been diagnosed with cancer, including those with active disease and those who are cu red (NCI). As a resu lt, cancer is a common topic about which people seek to obtain 2005 survey found that 52 .3% of the nearly 5,600 respondents in 2005 had searched for information about cance r, an estimated 105.3 million U.S. adults, up from 47.2% in the
21 most recent search for cancer information, a far gr eater percentage sought this information first from the Internet (41.2%, or an estimated 49. 2 million U.S. adults), compared to about one in four ( 25.4% ) seeking it first from health care providers (NCI, 2005). In addition, the survey found that 44.6% looked for information about a specific cancer and 7.9% abo ut treatments and cures. Other research examining cancer information needs also have shown treatment information is frequently sought, with it being the most common topic of interest during certain phases of cancer care (Rutten, Arora, Bakos, Azi z, & Rowland, 2005). The HINTS survey also showed a significant percentage of people (46.1%) were strongly or somewhat concerned about the quality of the cancer information they received from all sources, and that the vast majority of respondents (89.2%) f ound the information they received from the Internet very or somewhat useful (NCI 2005). A federally appointed Science Panel (1999) report also indicated that the potentia enhancing health, minimizing total burden of illness, and optimizing relationships between beneficial because it has the ability to provide improved access to individual information, broader choices for users, potentially improved anonymity, greater on demand access a nd support, greater interaction and social support, and enhanced dissemination and currency of information than is possible with the one way communication typical of
22 traditional forms of media (Science Panel). In addition, The Panel found Internet forms of health communication could serve several functions, including disseminating information; enabling informed decision making; promoting healthy behaviors, self care, peer information exchange, and emotional support; and managing the demand for health servic es (Science Panel). health, its negative qualities must also be considered. For example, the quality, accuracy, and understandability of health information available on the Internet varies greatly, and studies show that deficiencies are likely common, with some online health content misleading at best and harmful at worst (Berland et. al, 2001; Molassiotis & Xu, 2004; Tatsioni et al., 2003). This is particularly troubling in light of a growing body of research suggesting that people are increasingly using online information to influence their health decisions, yet many are not questioning its reliability, believing that it is trustworthy and of good quality (Molassiotis & Xu; Morris & Avorn, 2003). In addition, Cline and Noland (2002) indicated the Internet can promote risky behaviors and can replacement of face to face interaction wit h computer Thus, cancer consumers have a need for and would benefit from access to useful, understandable and accurate information that may shed light on their conditions and/or aid them in their decision making about care and t reatment (Rutten et al., 2005), especially through convenient, easily accessible routes such as the Internet. While the overall Internet audience increased modestly by about 5% between July 2007 and July 2008, the use of online health information sites swe lled more than four times that,
23 growing 21% (comScore, 2008, September). A 2008 study by comScore, an online marketing analysis firm, found there were 189.1 million unique visitors online in July 2008, with more than 69 million visiting health information sites, including WebMD (which saw growth of 3% to 17.2 million unique visitors), Everyday Health (a 63% increase to 14.7 million), Revolution Health Network (a 182% increase to 11.3 million), and AOL Health (an 88% increase to 11.1 million). During the sam e period, the National Institutes of Health Web site (nih.gov) saw a significant audience decline (14%), as did MSN Health, which dropped 12.0% (comScore). Cancer was second only to pregnancy among the health conditions for which most people sought informa tion, with the number of search queries tracked in February 2008 for each totaling 7.7 million and 8.8 million respectively (comScore, 2008, April). The next closest search was for flu information at a far distant third with 1.8 mill ion searches. The Inter net not only provides opportunities for finding needed health information, but it is increasingly being used as a venue for health advertising. Display ad views on the health information sites tracked by comScore (2008, September) totaled more than 1.5 bil lion, which were seen by an estimated 54 million pe ople. Networking and support have also become important component s of the Internet as evidenced by th e significant growth of Web 2.0 social software platforms that provide more dynamic and inte ractive venu es than did the initial World Wide Web, and facilitate sharing, participation, and conversation ( Giustini, 2006) The type of communication made possible by blogs, wikis, and other social networking capita has characterized the Internet 4 ).
24 According to one Internet analyst, m communities rooted in sharing exp eriences and advice, rather than simply being one way information resources for the consumer 2008, September, 3). One venue cancer consumers may use to seek this information, conversation, and support is by going directly to the Web site of a n organization providing health services. As a result, it would be mutually beneficial to both groups in the consumer health organization relationship to enhance understanding of persuasive organizational messages. Blogs Blogs were the first of the Web 2. 0 social technologies ( Giustini, 2006), and they have become one of the fastest growing segments of the Internet. Discrepancies abound about the number of weblogs, which are sometimes described as online diaries or personal journals, as well as how many ne w ones are being created each day, but there is widespread consensus about their explosive growth (Technorati, 2008). A 2008 survey of 17,000 Internet users in 29 countries found that 77% worldwide read blogs, and 45% (184 million) have started a blog, wit As a collective, the blogosphere rivals any mass media in terms of reach, time spent and In the U.S., 60.3 million people have read a blog, with 23% doing so on a daily basis and 43% on a weekly basis. About 26.4 million have started a blog (Universal McCann). The Pew Internet and American Life Project reports that 42% of Internet users, about one in three adults, have read a blog (Smith, 2008), up from 25% in 2005 (Pew Internet, 2005). In 2008, about 11% said they read blogs on a typical day, and 12% said they had created or worked on one of their own (Smith). The Washington Post reported the Web contains
25 112.5 million blogs (Halzak, 2008), and acc report, blog tracking company Technorati found that 900,000 blog entries are posted each day, with 79% of people blogging as a form of self expression, followed closely by those who do so because they want to s hare expertise and experience (73%). Although the use of blogs, both writing and reading them, has exploded in the last several years among the general public, data suggests that organizations overall have been slower to adopt blogs as part of their profes sional marketing and public relations strategies (Edelman & Intelliseek, 2005; Porter, Trammell, Chung, & Kim, 2007). al.). In a white paper guide to the blogoshere for marketers and companies, public relations giant Edelman and Intelliseek (2005) blogging phenomenon has drastically altered the landscape and challenged traditional tenets about the control of messaging by corporations, the media, the government, marke ters and company stakeholders business or organization that fails to grasp the Blogs, blogging, and the blogosphere (the so called online wor ld of blogs) have ezner, & Farrell, 2004, 33). Various descriptions of and ideas about weblogs have been advanced, however, for the purposes of this dissertation, blogs will be technically defined by Doctorow et al. (2002) as follows: A blog is a web page that contains bri ef, discrete hunks of information called posts. These posts are arranged in reverse chronological order (the most recent
26 posts come first). Each post is uniquely identified by an anchor tag, and it is marked with a permanent link that can be referred to by others who wish to link to it (p. 1, italics in original). are now forums for millions o f people discussing virtually any subject under the sun. In with topics as diverse as Existentialism, coastal preppies, body art, and Russian hackers (p. 37). Ty pically, blogs are created by individuals or sometimes small groups, but media organizations, politicians, schools and universities, and even government agencies have jumped on the bandwagon in recent years Typically, blog posts are informal and are publi shed frequently and nearly immediately on the Web without oversight by an external editor, and thus they have been described as being akin to a conversation or dialogue between the writer and reader (Hourihan, 2002). Although Doctorow et al. (2002) suggest blog posts are brief, which they descri 13), these authors also point out that blog writers, unlike those involved in more traditional forms of public writing, are unconstrained by word or space limitations. Others suggest b log posts are limited only by the time and talents of their writers. Blogs are often free to start and maintain, and require almost no technical knowledge or special equipment except a computer and access to the Internet (Blood 2002). In addition, compare d with typical Web sites, time plays a significant role in blogs, as indicated by the permanent links to older archived entries and the reverse chronological order of their postings (which some suggest is not
27 mandatory), providing both a sense of immediacy and an expectation of fre quent updates (Hourihan, 2002). Most blogs also share several other structural elements and common formats, largely the result of free, widely available, easy to use blogging templates. These include titles and subtitles for the b log itself, usually set off at the top of the site, as well as those that identify the topic of each post similar to the headlines and subheads in print publications. Most blogs have a main channel, which contains its most recent posts in reverse date orde r, and one to two sidebars to its left and/or right, where can be listed biographies of the writer(s), links to other blogs or web sites, calendars or other ways to reach older archived posts, or a variety of other items the primary author wishes to displa y. Each post is composed of text, which is the narrative that makes up the entry, and may contain links to news articles, events, or other items on another site that are generally relevant to the topic of the post or the blog. The post text is usually fol lowed by a time stamp (stating the time the posting was added), byline (providing the name or primary author), a permalink (providing a permanent copy of the post in an archived page off the main blog site), and interactive feedback mechanisms, such as comments forms or suggest links. These elements allow readers to share their own thoughts as an a ddendum to the original entry and to which t he primary author can respond or read t he comments posted by others (Doctorow et al., 2002). Comments are not mandatory unthinkable without the comments and the community of readers that comments make
28 2004). About 85% of blogs worldwide contain commenting systems (Technorati, 2008). Authors also frequently include photographs, images, music, or other audio or video files on the blog sites or in the posts. Advertisements are not uncommon, particularly on adds banner ads to blog pages (Doctorow et al, 2002), and with the advent of programs in which blog owners can enroll to have text and image advertisements placed on their sites for which they are pai d on a per click basis. The text to be included in a post is typed into a blogware tem plate, and with a single click it is automatically published on the blog site in conjunction with its associated elements. The n ewest post, or entry, is published on top, and as this happens, the older entries sink downward on the blog main, or home, page until they eventually are shuffled off to their permanent archived storage pages (Winer, 2003). Although the permalinks, comments, and other components have become standa rd, particularly since the 1999 advent of easy to use, automated templates and oth er blogging tools (Blood, 2000), two aspects of blogs have come to be widely considered fundamental facets, not only of their online identity but also of their uniqueness amo ng web content. web pages on similar topics, content, views, or opinions, with the home page entries of eight in 10 active blogs containing at least one link to an exte rnal site. Although the origins of blogs are rooted in politics and journalism and despite their continued magnitude in t hese areas, the vast majority about 90% of these links are not to traditional news sites, indicating that bloggers are using the techno logy for much more
29 than obtaining news (Henning). Second, the blog is an unedited personal portrayal of the author, and thus the traits, qualities, and personas they project through their writing are of primary importance (Winer, 2003; Blood, 2002). Along with the comment forms and other feedback mechanisms on blogs, these key traits establish the social connections and interactions that typify blogging and provide the foundation upon which are built a myriad of virtual communities uniting people of similar interests and shared experiences (Hewitt, 2005; Hourihan, 2002; Winer). There tends to be disagreement about when blogs debuted on the Web. Some, including Blood (2002), Winer (2001), and other pioneers of the form suggest the first blogs started in the e arly 1990s, while others, like Hewitt (2005), suggest it was much later in 1999. By and large, Jorn Barger, editor of one of the original weblogs is credited 1997 (Blood, 2000; Hewitt ). That was changed in mi d It's the links, 2006 ) by another blogger named Merholz, swiftly followed by the shortened form blog (Blood, 2000; It's the links ). Also generally agreed upon is that the first real growth spurt of blogging b egan in 1999 with the lau nch of free, automated blog software. A s econd surge started after the September 11, 2001, terrorist attacks and at the start of the war phenomenon by th e mainstream media. Further expansion is often attributed to the use of blogs as political tools in the 2004 presidential election. called A list blogs, such as Boing Boing and Engadget, have millions of readers that can rival circulation levels of mainstream media outlets, many of which have launched blogs themselves in order to
30 increase readership, and to better connect and open dialogue with readers (Smith, 2005). Similarly, politicians and political parties are using them to converse with voters, with both 2008 presidential campaigns hosting blogs, and more than one in four voters (27%) saying they read blogs about politics and the election. Prominent scientific journals like Nature also have blogs t o provide their readers an opportunity to comment on published stories in a variety of fields. In the business domain, blogs are increasingly being promoted as tools for conversing directly and interacting with new and existing customers, building and main taining relationships, shaping public opinion, gaining trust of prospective clients, building awareness and loyalty to brands, and putting a human face on corporate America (Barbaro, 2006; Flynn, 2006; Kirkpatrick, 2005; an electronic communications powerhouse that is likely to have greater impact on business communications and corporate reputations than e mail, instant messaging, and traditional market oriented w p. 4). About 5,000 corporate blogs ar e being hosted (Halzak, 2008). Although corporate blogs posted publicly on the surface web encompass a fairly small percentage of the blogosphere, a 2005 survey found a growing number are using this platform for public relations and marketing purposes, and indicated that companies would drive the next wave of blogs and other social media tools for promoting communication with both internal and external audiences (Guidewire Group, 2005). Chief executive of Marriott International, Bill Marriott, began writing a blog as a way to communicate, and people who started on his blog clicked through to book more than $5 million in hotel reservations (Halzak). Companies are also communicating more indirectly with consumers through blogs, and
31 8). For example, in an effort to polish its tarnished reputation, Walm art, one of authors with pro Wa lm art s tatements and information to use on their sites (Barbaro). Research has shown that blogs provide a more effective vehicle for establishing and including during times of cris is, than are more traditional public communication methods such as newsletters (Kelleher & Miller, 2006; Sweetser & Metzgar, 2007). The advantages of blogs come with potential headaches stemming from the loss of control that can accompany the unrestricted discussion for which blogs and bloggers are fabulous for relationships ( Kirkpatrick, 2005). Both personal and corporate bloggers have been positively impacted by their blogs, including an increased network of friends and other contacts, better awareness about the industry in which they are working, and new opportunities that otherwise would not have been available (Technorati, 2008). Blogs and Health Similarly, in the h ealth realm, doctors and nurses, hospitals and health organizations, and patient advocacy groups are increasingly launching blogs. connection between health professionals and the communicating substantial, evidence voices to a wider range of professionals who can reach broader audiences (Lagu, Kaufman, Asch, & Armstrong, 2008, p. 1642 and 1645). The peer reviewed scholarly
32 topic category. Large governmental organizations are also using blogs as a way to provide information to and get feedback from members of the public, including about health. One is these is the U.S. Department of Health and Human Services (HHS), the government's principal agency for protecting the health of Ameri cans and providing human services, which has a $707.7 billion annual budget; nearly 65,000 employees; and includes more than 300 institutes, centers and programs (HHS, 2008). HHS issues (HHS, n.d., 1), such as the cost of health care, mental health, Medicare, HIV/AIDS, and various health emergencies. Other health care organizations have r ealized the potential of the two way communication afforded by blogs, and as a result into healthcare as an efficient and effective communication Thielst, 2007), with public blogs being maintained by healthcare administrato rs, leaders, and consultants, including Paul Levy, president and CEO of Beth Israel Deaconess Medical Center in Boston (runningahospital.blogspot.com). Medical facilities are also integrating blogs written by patients, which provide them a way to keep frie nds and family informed about their progress and to receive return greetings, while also affording health professionals an opportunity to gain more in depth insights into the personal experiences and mental, emotional, and physical states of patients than are afforded by a typical medical appointment (Hillan, 2003).
33 Individuals are also blogging independently about their diseases and conditions, including about such diverse topics as brain injury, eating disorders, menopause, diabetes, and manic depression with a recent survey finding that that about 18% of bloggers write about health issues (Technorati, 2008). In launching a publication about blogs and the drug industry, market research firm Research and Markets (2006) wrote: cost of medical care, healthcare blogs will become more popular and influential as consumers assume greater responsibility for their own bloggers on various health topics would be valuable for improving health and health care (Health 2.0, 2007; Oransky, 2005). For example, writing in the medical journal Lancet Oncology Oransky stated his charge was to create a one stop, authoritative cancer blog for everyone affected by c ancer where patients, basic scientists, clinical researchers, and practicing oncologists could all blog on one site that would function as oncologists treat their patients. It would also be a place for surveillance of side effects directly from patients, and could be built on a site that is a collectio (Oransky, p. 839). Ma jor mass media outlets also have launched blogs specifically about health and medicine, suc h as USA Today's A Better Life, which lists more than 30 medical categories including cancer; the New York Times editors shares his experiences coping with prostate cancer; and a CNN correspondent, Dr. Sanjay Gupta, and his staff share news on health and medical
34 trends. The American Cancer Society also provides members of its online Cancer Survivors Network an opportunity to start a blog. And a national conference, Health 2.0, has grown up around targeting health care using Web 2.0 technologies such as blogs, wikis, mash ups, video, and other user generated content. The reach of these blogs differs greatly, but The Health Care Blog (2005) which The Wall Street Journal cites as Crowd, 2005, 2), has expanded to become a group blog and grown in readership from a few each day to more than 80,000 visits in October 2008 Despite the profound growth in this activi ty and its acknowledged potential in the health realm, empirical studies about health and medical blogs and bloggers is relatively limited, and scholarly research regarding cancer related blogs and those who author and read them are lacking ( Kim & Chung, 2 007). When articles about health and blogs do appear in scholarly or trade publications, they are more likely to be discussed in commentary or news items rather than as the topics of original research. Discussion of blogs in relation to health is much more common in newspapers, and among the prominent ones to do so is T he Wall Street Journal, in which Landro (2005) wrote that several advantages weblogs possess in regard to medical matters. These included that health blogs offer a venue for self expression that can be therapeutic; afford a more personal alternative to traditional Web content; allow patients to chronicle their experiences and share them with others; provide li nks to the latest medical news and studies; function as an advocacy forum; and act as communities for concern, support, and reassurance (Landro).
35 Other media accounts about health related blogs have identified similar creator benefits, with some accounts suggesting the therapeutic, or cathartic, effect stems not challenges, but from helping others to overcome fear and uncertainty, and better deal with their own situatio ns (Oransky, 2005; McNiel, 2006). Discussing these emotions in the kinds of revealing and intimate postings common to many blogs, especially those related to health issues, affords an outlet for feelings as well as distance from them, which may provide cla rity aided by the thoughts of others who weigh in with comment s (Snider, 2003). Media reports are rarer about the benefits derived specifically by readers of health blogs, but these readers also seem to reap some benefits from them. These include that the y provide a forum not otherwise available on the Web for obtaining information about stories of others, as well as consolation and strength from the knowledge that they are not alone in the struggles they may face, helpful tips about how others have dealt with their setbacks, and possibly confidence needed to persist (Harmon, 2003; Koenig, 2006; McNiel; Oransky). The virtually unlimited connections possible can create a sense of belonging to a larger social group for both blog authors and readers (Hillan, 2003), and allow those with even the rarest diseases to locate others worldwide who can provide them information about treatments, doctors, and just generally about livi ng life with the condition (Health 2.0, 2007). These online journals also allow those who are isolated due to emotional, geographic, financial, or physical barriers overcome these
36 limitations, and provide venues for frank discussion with anonymity typicall y not available in real world settings (Hillan). As is the case with all Web content, there are also some disadvantages associated with health blogs, including that the experiences contained in their postings may not be representative of those of average patients; that there may be issues of legitimacy and accuracy of the information posted, including that the stories and even the people who create them and/or the credentials they might cite may be fictitious; and that it is potentially risky to revea l personal information publicly (Harmon, 2003; Landro 2005 suddenly, especially when written by someone with a deadly disease such as cancer (Oransky, 2005). Blogs written by health profe privacy by revealing information that is confidential and/or identifiable (Lagu et al., 2008), and people may be more easily confused when they have so much health information, especially when it may be conflicting (Heal th 2.0, 2007). And because the virtual world lacks the non verbal cues that are fundamental in face to face interactions, misinterpretation can occur (Hillan 2003). Blogs also have been cited as a form of online social support for cancer patients (Hillan, 2003; Ornasky 2005 ). Scientific analyses of online social support have identified some of the same benefits that have been discussed anecdotally in the popul ar press : gathering information, both factual and based on the experiences of others; providing so cial contact and support; affording a safe atmosphere in which to express questions and concerns, and which increases the ability to discuss personal problems than in person venues allow (Kummervold et al., 2001). In addition, it has
37 been suggested that th e online environment may bring together socially dissimilar to (2005) suggests t hat blogs function as a social space in which self disclosures are made in order to gain readers that will develop into a community of similarly oriented, trusting, and sharing individuals. Hevern (2004) also suggests that blogs function as extended social networks in which their authors position themselves not only by the mere act of posting entries, but through the divers ity of content and the varied ev en oppositional voices that are used. Using an online survey, Kaye (2005) investigated the motivations of people who access blogs. The study, which resulted in 3,747 responses, was primarily interested in political and news blogs, and three of six motivations involved descriptions of current events, news or politics (information seeking and media checking, political surveillance, and convenience). However, the other three primary reasons for reading blogs have information to use in discussions with others, and fulfilling various emotional needs issues and current events viewpoints and interacting with like th the last being the most common predictive motivator (Kaye, p. 84). Rare are peer reviewed journal articles about cancer related blogs. However, in an exploratory study, Kim & Chung (2007) used cluster analysis of survey data to provide insight about th e demographics of cancer bloggers (mainly patients and their
38 family members and friends) and their use and perceptions of cancer blogs. The team found that that highly educated Caucasian females were the primary group who were involved with cancer blogs, a nd that these bloggers on average were older (57 years), results that differed from previous blog studies. These users fell into three clusters. The largest group, consisting of 42.5% of the 113 survey respondents, included long time cancer bloggers who al so commonly sought information through more traditional channels. They more frequently used cancer blogs to express their own opinions and they were the group that received the strongest empowe rment through blogs. The second largest group (33.6%) included frequent information seekers who were more likely to be novice bloggers. They felt the need to compile information and were more likely to use blogs to seek cancer knowledge. The smallest cluster (23.9%) consisted of highly motivated individuals who sought health information and were found to be the most likely to change their behaviors while using cancer blogs. They read blogs to increase their knowledge about cancer and to get timely updated information. Despite these differences, encouraging others and s haring their personal stories about cancer were the main reasons users in all three groups turned to blogs (Kim & Chung). T he authors concluded and their companions to engage in meaningful conversation and that sharing personal experiences via blogs may help patients better cope with their cancer related health Persuasion Research Research into various aspects of the persuasion process and elements that m ay enhance the influence of persuasive communications are centuries old, yet a universally agreed upon definition of the concept is lacking (Miller, 2002; Stiff & Mongeau, 2003).
39 There is widespread agreement, however, that persuasion is an intentional beh avior whose goal is to shape emotional, cognitive, and /or behavioral outcomes, with these three components sometimes described as the tripartite model (Bagozzi, 1978 ; Stiff & Mongeau). Persuasion is ubiquitous in almost all types of communication (Trenho lm, 1989). However, a distinction must be, and typically is, made between the concepts of persuasion and coercion. In the former, the receiver has the ability and choice to reject the message while he or she cannot in the latter (Stiff & Mongeau), a differ entiation that is particularly vital in matters of health. Numerous definitions of persuasion have been developed over the past several decades, with many explicitly articulating this difference through the inclusion of specific language, including words a nd phrases such as noncoercive (Trenholm, 1989) freedom (O'Keefe, 1990), free choice (Perloff, 1993) autonomous individuals (Woodward & Denton, 1992), and interactive process between sender and recipient (O'Donnell & Kable, 1982) In the health context e xamined here, persuasion will be regarded as to induce a change in the belief, attitude, or behavior of another person or group of persons through the transmission of a message in a context in which the persuadee has The long history of studies aimed at determining how messages can be made more persuasive has resulted in investigation of several aspects of the communication process best laid out by t he well known model that asks: Who (the source) says what (the message) in which channel (the medium/media of dissemination) to whom (the recipient or audience) with what effect
40 (which here would relate to extent of persuasiveness or influence )? (McGuire, 2000; Petty & Cacioppo, 1981; Rogers, 1994). Attitudes and changes in recipients resulting from persuasive communications toward these factors comprise the basis of much of the research and theory in the field, with behavior considered to a lesser extent (Stiff & Mongeau, 2003). An attitude is a feeling toward some individual or object that may prompt action at a later time (Rogers, 1994). The emotional or affective component of the tripartite attitude model involves the positive, negative, or neutral fee lings one has about an individual or object; the cognitive element concerns the beliefs about it; and the behavioral aspect encompasses the tendencies one has t o act or not with respect to it (Bagozzi, 1978; Stiff & Mongeau). The basis of persuasion studie s is the assumption that attitudes are able to influence action The relationship is complicated; however, and a number of factors with respect to each aspect of the communication process have been shown to influence the extent of the impact a message has on recipients. Specifi c source, message, and receiver variables to be experimentally investigated in this study will be described in more detail below. However, McGuire (2000) suggests tha t source characteristics that can increase persuasive impact include attractiveness, credibility, and similarity with respect to various demographic factors such as gender or ethnicity. Message variables havi ng an effect include the number and format of th e arguments presented, as well as the way they are presented, for example, the specific language used, and the level of humor or abstractness (McGuire). The order of the arguments in a message, the level of fear they generate, their comprehensibility, and
41 whether they include an implicit recommendation or an opposing argument also can disseminate the communication can play a role in enhancing or reducing persuasion, including the format of the message (e.g. written, aural, or visual) and the environment iver or audience traits have varying impact in their ability to influence someone in different situations (McGuire, ). These factors include personality vari ables, such as anxiety level or self esteem, as well as ability, lifestyle, and demographics such as sex. This dissertation will experimentally manipulate the followin g three independent variables involvement, sourc e credib ility, and message quality or argument strength in order to investigate the roles they play with respect to the ir pe rsuasive influence of health messages about a novel cancer treatment included on organizational blogs. Attitudes toward a message ( for example, an advertisement or news article), toward an object (often a product or service), and toward a brand or an organization promoting the message and/or product are widely measured as dependent variables in persuasion research. Assessment can be e ither direct, i.e. through people directly reporting their attitude, or indirect, in which attitudes are gauged without a person knowing it (Petty & Cacioppo, 1981). However, most persuasion studies rely on the former because direct reports typically provi de better validity and reliability, and a greater ability to pinpoint small differences i n attitudes (Petty & Cacioppo). As a result, this research seeks to determine the influence that each of the three i ndependent variables has on several dependent o utco me measures. The first four of these dependent variables involve various a ttitudes, specifically those toward the blog
42 message about the treatment, the cancer treatment itself, the organization providing the treatment and sponsoring the blog, and the blog channel of communication To supplement the attitude measures and provide an indication of how closely they may be associated with behavioral intentions, several additional dependent variable s assess the future likelihood that one will become empowered to take action with respect to the cancer treatment discussed i n the experimental blog messages. E mpowerment, which has been shown to be beneficial with respect t o health and medical issues, will be d escribed in greater detail subsequently Other behaviorally related dependent measures include those assessing the likelihood of taking action with respect to the cancer itself, and to undertaking a specific action related to the cancer treatment organization. In this research, a ll but the last dependent variable are assessed through self report measures Involvement The receiver trait of involvement has been considered under incongruent definitions over the course of persuasion research and has similarly been shown to have disparate outcomes in studies of it (Stif f & Mongeau, 2003; Johnson and Eagly, 1989). In order to better understand these contradictions and illuminate the role ( s ) that involvement plays, Johnson and Eagly (1989) conducted a meta analysis of involvement research that subsequently defined the conc induced by an association between an activated attitude and some aspect of the self con three concepts of personal involvement that affect persuasion in different ways. This included those whose involvement was activated by enduring values (called value relevant involvement), in which cases high involvement subjects were less persuaded than low involvement subjects. Second, among those
43 who became involved as a result of their ability to obtain de sirable outcomes (outcome relevant involvement), high involvement subjects were more persuaded by strong arguments and less persuaded by weak arguments than those having low involvement Third, among those whose involvement was triggered by the impression they made on others (impression relevant involvement), high involvement subjects were slightly less persuaded than low involvement subjects. Involvement is a variable often used in studies as a way to vary the motivation level subjects have to process mess ages of interest, which can be done by controlling the personal importance or relevance about the topic (issue involvement) or how (Chaiken, 1987). When participants hav e enough ability to process the message and when their involvement is high, they tend to be more motivated to think more about the validity (Chaiken). Results in this vein are consistent with traditional models of how persuasive information is processed (Chaiken), as well as with the aspects involving thoughtful or effortful cognitive pro cessing in the more recent dual processing models, the Systemic Heuristic Model (Chaiken, 19 80) and the Elaboration Likelihood Model (Petty & Cacioppo, 1981, 1986), the latter of which will be described in greater detail subsequently. Source C redibility The characteristics of a human source providing a message (distinguished here from those of a medium or channel through which a message is received) are key to
44 several source cha racteristics have been investigated; however, credibility is among the experiments on persuasion conducted by Hovland and his colleagues, two dimensions surface frequently i n studies of source credibility and have been found to adequately repres ; Stiff & Mongeau). The first is expertise, also called perceived to be a source of val 1953, p. 21), this involves the extent to which the communicator is identified as being knowledgeable and informed about the topic of the message (Stiff & Mongeau). The second dimension is trustworthiness, wh tell the truth (Stiff & Mongeau). It is important to note that neither credibility in general nor either of its two facets are characteristics of a source; rather, they are perceptions by the message recipients. In addition, although these two separate but related constructs are not typically manipulated separately in experiments, it is possible for a source to be perceived as being high in one (e.g. expertise) but low in the other (e.g. trustworthiness) (Stiff & Mongeau ). Also, the actual characteristics audience members is t he magnitude of the persuasion e en sources are cast as either high or low in credibility, those judged to be low on the ratings scales are not low in absolute terms, they are simply low relative to those who have As is the case in this disse rtation, credibility is often effectively manipulated in
45 occupation, training and/or experience, such that sources described as having h igh educational attainment or signi ficant experience in a field, for example, are rated as credibility. These includ e how a source delivers a message, their speaking style, the rate at which they deliver a message the inclusion of pieces of evidence to support their points, the stance they advocate, whether they repeat words or have articulation difficulties, and even In a review that spanned five decades of research, Pornpitakpan (2004) found that source credibility has been shown in the vast majority of studies to influence changes in e variety of topics, with highly credible sources in general creating significantly more attitude change and improving behavioral compliance. As the previous discussion suggests, however, a significant list of factors can interact with source credibility, leading not only to varying magnitudes of influence resulting from different levels of source credibility, but also the opposite possibility that in some circumstances low credibility sources can be more persuasive (Pornpitakpan). The factors that can inte ract with source credibility have been mentioned previously here not be reiterated. However, findings with respect to the interaction of the three independent variables being manipulated in this study bear inclusion and will be addressed subsequently. The ability to successfully utilize sources to present health information is of significant interest and concern because it affects not only the impact of a messages but
46 al so perceptions of an organization more broadly Sources serve a variety of functions in presenting health messages including to attract attention, engage audiences, increase exposure, facilitate understanding, heighten emotional arousal, increase retentio n, elicit positive responses, and enhance message uptake and behavior change ( Atkin, 2001; Salmon & Atkin 2003 ). D epending on the desired outcome, organizations and campaigns utilize a variety of source s which would include public relations representatives and other traditional spokespeople who customarily speak on behalf of an organization ; an d who share similar characteristics with the audience and can heighten self efficacy (Atkin, 2001) It is important to note that the efficacy of a messenger is highly dep endent on the situation and the audience, i.e. a source that is influential in one campaign or among one group may not be similarly effective in or with another (Atkin Each of these source types has also been found to have drawbacks. For e xample, messages provided by a medical authority can be perceived as dull or overly complex, while survivor stories may be so compelling that they distract from the substance of the message (Atkin, 2001) In addition, although public relations spokespeople or representatives are widely used in organizational messages, they have been shown to be perceived as having low credibility p articularly with respect to trus tworthiness (Callison, 2001) Questionable practices related to in sufficient disclosure on vide o news re leases, paying journalists to preview or promote products soliciting bloggers to
47 endorse beleaguered companies through special favors and other problematic issues likely lend themselves to such views particularly when they surface in the medi a as being ethically flawed A ccording to Callison, the public associates public relations with bending the truth if necessary to make the client look positive or to facilitate a cover up (p. 122). Likewise, journalists (Sallot & Johnson, 2006 ) beginning public relations students (Saunders, 1993), executives of companies who hire PR agencies (Agencyfinder survey, 2004) and even PR educators and professionals themselves (Newsom, Ramsey, & Carrell, 1993) have raised concerns about the credibility and truthf ulness in the field. On the other hand, physicians have been found to be highly credible and trustworthy sources In a surve y examining sources of health information, Avery (2010 ) found that doctors were perceived as the most credible among six sources w hich also included scientists, and government agency officials. HINTS 2007 found that 68.5% o a lot of trust in the health or medical information they received from a doctor or other health professional, that the source of health message s can have significant influence on recipients. Message Q uality Two types of messages are often discussed in persuasion: Those that are based primarily on logic or rational appeals and those based on emotions or affective appeals. Although it is often difficult to make a distin ction between the effects of these messages because there could be overlap, the quality of a communication provided to an audience is one type of rational message appeal that has been investigated (Stiff & Mongeau,
48 nfluence from sound reasoning and the quality quality evidence increases influences is highly intuitive (Stiff & Mongeau, p. 129). Whether evidence is included in a messa ge, the amount of evidence that may be provided, and the extent of the quality of the information has all been shown to have differing levels of influence, but studies have generally found there is a positive association between attitude change and evidenc e (Stiff, 1986). Most of these studies have looked at message or argument quality through the structure of the ELM as is being done here (Pornpitakpan, 2004), and in line with its predictions, research has consistently shown that increased quality of evide nce included in a communication increases persuasion among people who are more motivated to process its content and have the ability to do so (Petty & Cacioppo, 1981, 1986; Stiff, 1986; Stiff & Mongeau). Elaboration Likelihood Model As noted previously, pe rsuasion has been extensively investigated; however, the view that nearly every independent variable studied increased persuasion in some situations, had no effect (Petty & Cacioppo, 1986, p. 2). Endeavoring to resolve this impediment, Petty and Cacioppo (1981) unified some of the major concepts involved in prior attitude change approaches that often were a pplicable only in certain circumstances into the Elaboration Likelihood Model (ELM), one of two commonly discussed dual process models of persuasion that have been advanced along with the Heuristic Systematic Model (HSM). These models suggest that people p rocess information through two disparate pathways, each of which mediates persuasion differently (Chaiken, 1980; Chaiken, Duckworth, &
49 Darke, 1999; Petty & Cacioppo, 1981, 1986). In addition to providing a framework for organizing existing persuasion theor ies, the ELM systematized the multitude of variables that could influence attitudes and delineated when the different processes of persuasion were likely to come into play ( Briol & Petty, 2006 ; Cacioppo & Petty, 1987). The ELM suggests the necessity for t wo on a daily basis precludes them from pondering and carefully judging all of it (Petty & Cacioppo, 1986, p. 5). As a result the extent to which they elaborate on any given issue or piece of information varies not only with each individ ual but also with each situation, and can be impacted in magnitude and direction (favorably or unfavorably) by a host of factors (Petty & Cacio ppo). Part of this comprehensive scrutiny involves relating the new information derived from a message to what one already knows about a topic, and the extent to which one elaborates in this manner ranges across a continuum from a great deal to very little Petty & Cacioppo lay out the guiding principles of the ELM in seven specific postulates, which will be dis cussed more generally here. The first processing pathway, called the central route, involves using a significant amount of active, rational thought issue being considered (Booth Butterfield & Welbourne, 2002; Petty & Cacioppo, 1981, 1986). The reasoning undertaken through the central route does not have to be objective, i.e. taken only from the inform ation being communicated, it also can be colored, or biased, by past experiences, knowledge, and feelings (Petty & Cacioppo, 1986; Booth Butterfield & Welbourne). This route integrates the message learning approaches of Hovland and colleagues; self persuas ion approaches, such as role
50 playing, mere thought, and the cognitive response approach; and combinatory approaches, which merge information into an overall impression, such as the Theory of Reasoned Action (Petty & Cacioppo, 1981). Similarly, this more co gnitive, analytical route is similar to the systematic processing described in the HSM (Chaiken, 1980). The second is a considerably less thoughtful pathway to attitude change, termed the peripheral route in the ELM and the heuristic process in the HSM. It focuses instead on peripheral persuasion cues, which are setting that are sufficient to produce an initial attitude change without any active thinking about the attributes of the issues or the object under co nsideration 1981, p. 256, italics in original). These factors assimilate attitude change approaches involving learned responses contemplated in models of conditioning and modeling; judgmental approaches, in which evaluations of all thin gs including attitudes and beliefs are relative to some comparison; and attributional approaches that relate to the inferences people make (including about attitudes) based on the behaviors they hear and see (Petty & Cacioppo). These situations involve agr eement based on simple cues, such as the title or trustworthiness of a source, rather than on thoughtful consideration of the information presented, and are unrelated to the merits of the message (Booth Butterfield & Welbourne, 2002). It is important to n ote that either route can result in the same attitude about an issue or object (Booth Butterfield & Welbourne, 2002). Affect is contemplated in the ELM, and its developers conducted research attempting to measure facial muscle movement and other electrophy siological manifestations that might provide physical
51 states should serve as persuasive arguments or help in assessing the cogency of unmotivated or unable to undertake this extensive cognitive elaboration, feelings and emotions instead serve as a peripheral cue (Petty & Cacioppo, 1986, p. 214). Pleasant affect would be expected to enhance attitudes while the opposite would occur in the case of negative emotions. It may help to provide a simple e xample to better distinguish the differences in the somewhere new for the first time, typically we put sig nificantly more rational thought into how we are going to find our way there than we do for a place we have driven to many times in the past (or even once before). We often obtain information to help us, perhaps by consulting a friend who has driven to the location or contacting the place we are heading, and/or checking a map or printing turn by turn directions off the Internet so we are less likely to get lost or make a wrong turn. However, if we have been somewhere before, our cognitive effort is signific antly less, especially if it is a place we travel to and Dairy Queen on a corner commonly traveled distance and not remember a single specific detail from the trip, making our way entirely by the cues we see, and not realizing we are home until we are parked in the garage that or people along the way. Under the ELM (and the HSM), motivation level and ability to process a communication play key roles in affecting both the extent and direction (positive or
52 negative) of a persuasive communica & Cacioppo, 1986, p. 8), and is often investigated through variables related to the message, including how personally r elevant or involving the topic is; the recipients, such as their need for cognition; and the context of the communication. Ability, which relates to a person having the necessary knowledge to understand a message and the capacity to process it, may be infl uenced by variables such as distraction, repetition, prior knowledge, and the complexity and/or comprehensibility of the message (Petty & relates to the extent to which an individual is likely to be motivated and able to think carefully about a message. If either or both of these factors are missing, i.e. a person is either unmotivated or unable to thoroughly elaborate on a message, the ELM proposes he or she will look for more superficial ways or short cuts to judge the message, such as the inclusion of an expert, a celebrity, or an attractive or likeable source; or the number of arguments cas es, attitude change occurs through the peripheral route, and as a result is relatively temporary, and typically cannot be used to predict behavior. If no such cues are present, Petty and Cacioppo suggest a recipient will either keep or regain his or her in itial attitude On the other hand, when someone is both motivated and able to process a message, elaboration will occur. To a large degree, the extent of the elaboration and there fore the persuasive influence will be determined by the quality of the inform ation
53 or strength of the arguments presented such that strong, compelling arguments will lead an individual to more intense scrutiny or elaboration. In such cases, the central route to persuasion is evoked, resulting in persuasion that is more enduring, b etter able to predict future behavior, and less likely to be successfully counter argued against (Petty & Cacioppo, 1986; Petty, Haug t vedt, & Smith, 1995). A number of studies have examined the effects that the v ariables described previously involvement, m essage qu ality, and source credibility have had on persuasion within the framework of the ELM, especially their interaction. For example, as a receiver becomes more involved in an issue targeted in a message, i.e. it becomes more personally relevant to the difference; however, this quality has been found to make a great deal of difference for low involved people, i.e. those who elaborate a great deal on a message, are more likely to consider little outside the quality or strength of the information, while those with low involvement are likely to be equally or even more swayed by specious arguments, at least when they are receive d in the presence of various indicators, such as a credible source. Such studies have been conducted in a variety of contexts. However, motivation and/or involvement in health and medical issues may be unique from other areas. For example, Braverman (2008 ) suggested that rather than being a situational factor, involvement in medical interventions may instead be an individual characteristic. Therefore, as this study concerns health, the discussion here will be limited to describing studies similarly geared so that more analogous comparisons can be made.
54 One such study was undertaken by M cCullough and Dodge (2002). They examined the role these three factors played in advertisements about general hospital services. This is one of few studies found that have s pecifically investigated persuasive health messages that are not prevention or detection oriented, instead focusing on the provision of health services, and thus is more closely related to the current study than most others to be discussed. In line with t he ELM, the researchers determined that high involvement consumers rated the hospital higher and were more persuaded by advertisements containing a high quality message, i.e. those that included messages about core services, such as the availability of spe cialists the range of services offered tha n were those in the low involvement group. However, rat ings of the hospital when a low core services, such as its comfortable at mosp here, quality multimedia systems and delicious meals, were no different among those in the hig h and low involvement groups. With respect to source credibility, the researchers hypothesized (contrary to the ELM) that high involvement consumers would be mor e persuaded by a physician spokesperson than a non physician, while low involvement consumers would be less influenced by a physician. They proposed this counter position because they felt it impossible for consumers to make a distinction between a health care provider and the hospital services, and thus judgments of the former (which it appears were not measured in their study), would be reflected in assessments of the latter among those who were highly involved in the message. In addition, they believed t his inseparabil ity carried over to low involvement consumers such that the physician source would elicit
55 negative emot ions, diluting the influence this high credibility spokesperson should have among this group. What they found ran contrary not only to the also to their own. They found that the highly involved participants rated t he hospital significantly more favorably when a non physician spokesperson was used, and low involvement subjects evaluated the hospitals almost identically regardless of the source who presented the message. To explain the reasons why the physician and non physician spokespeople were equally influential, the team proposed as possibilities that the information about the source was included at the end of the ad vertisement (source timing has been shown to make a difference in some circumstances), and that the same source picture and number of statements were used in each of the hospital advertisements tested, perhaps minimizing or nullifying the peripheral cues u sed by the low involvement group. In addition to proposing hypotheses that run counter to published research findings on the ELM (without documented evidence from alternative studies), this research could also be somewhat problematic from the perspective o f the involvement manipulation. Here, two disparate situations were used, with those in the high involvement group being told to imagine they had just been told by their doctors that they had a large cyst on their knees that needed surgery. The low involve ment group was told they owed a large fine for a video tape they rented and never returned (McCullough and Dodge, 2002) These scenarios were pretested to determine that they produced similar emotional states; however, evoking involvement through such diff erent situations has been found to be problematic. In addition, injecting emotion into the experiment could confound the results, particularly as these scenarios were pre tested
56 for18 different emotions, and finding consistencies across so many variables i s difficult at best. Despite these shortcomings, this study provides insight into the potential impacts of organizational persuasive messages related to health care services and a jumping off point for additional research. Another study closely related to the current research (although considerably more process influencing variables could predict which o f four decisional styles two involving the central processing route and two the peripheral route patients would use to make medical decisions about their cancer treatment (Petersen et al., 2000). Th e study used the ELM within social influence theory to investigate a number of variables, finding support for the hypotheses that oretell decisional style. T his predictive ability was associated with both cognitive and emotional motivation, the ability to process the information, and cognitive responding, which refers to the extent to which a person produces favorable or unfavorable responses about a message. The researchers said the study was among a series of investigations aimed at increasing understanding of the ways people make health decisions so these processes might be changed in order to enhance coping, wellbeing, and other heal th outcomes (Petersen et al.). The two studies described involved empirical testing of ELM constructs in treatment and service related contexts; however, most experimental testing of the model in the health rea lm has concerned prevention oriented messages. For example, Braverman (2008) investigated the effects that involvement, message channel (text vs. audio) and the type of message (informational vs. testimonial) had with respect to
57 weight management communica tions. In part, the results showed that those who were highly motivated to improve their diets or lose weight were more effectively persuaded by informational messages, which were considered to permit the greater deliberation t ral processing (Braverman ). The study also found that people who were unmotivated in this area, i.e. had low involvement in the weight management message, were more persuaded by testimonials, which was felt to be more likely akin to peripheral cues because of their storytelling nature, consistent with the peripheral route of persuasion. The channel results conflicted across three experiments, but in general, the audio format was more effective for testimonials, while written informational messages were more persuasive. The author cautioned that involvement was measured as an inherent quality in the issue rather than manipulated and therefore needed additional investigation, but suggested the results offered new possibilities for better targeting preventive h ealth communication messages. In another prevention oriented experiment based on the ELM, Igartua, Cheng, and Lopes (2003) investigated the level of involvement and the format of short AIDS films (musical vs. dialogue) to determine the persuasive effect of HIV prevention messages, with respect to AIDS campaign messages (p. 524). Although they found involvement was irrelevant, they unexpectedly identified that the lower quali ty entertainment oriented musical format resulted in a greater number of posi tive responses while the high quality dialogue format was better liked, which t he authors stated conflicted with the postulates of the ELM. However, they did indicate the study id entified two routes to persuasion consistent with the ELM, with the musical format activating a peripheral pathway o f
58 processing while the dialogue format stimulated a central pathway of processing. They concluded saying the study had clear implications fo r designing AIDS prevention materials that would be differently effective for audiences persuaded through each of these pathways. In examining the effects of messages promoting regular dental flossing, Updegraff, Sherman, Luyster, and Mann (2007) found th at argument quality mattered only when They suggested, in part, that health message tailoring is effective when it matches important recipient characteristics, leading the m to carefully contemplate the information persuasion. Rath er than testing its constructs empirically i n areas of health a dditional research discusses the ELM as a framework or th eoretical perspective that can be effectively used to increase the persuasiveness of health messages in a variety of preve ntion and other health contexts. These include nutrition (Wilson, 2007), increasing medication compliance (Horton, Minniti, Mireylees, & McEntegart, 2008), breast cancer awareness (Holt, Lee, & Wright, 2008), promotion of pre chemotherapeutic sperm donation (Achille, Rosberger, Robitaille, Lebel, Gouin, Bultz, et al., 20 06), disordered eating (Withers & Wertheim, 2004), alcohol counter a dvertising, such as public service announcements and warning labels (Agostinelli, & Grube, 2002), and health promotion and follow up for childhood cancer survivors (Absolom, Eiser, Greco, & Davies, 2004). Some other researchers have suggested integrating t oriented constructs with other more behaviorally focused theories in an attempt to increase the effectiveness of
59 persuasive health promotion campaigns, which often have limited effects, at least in the long term. Slater (2006) stated that communication interventions based solely on communication content, communication strategies, and even, in some cases, evaluation and other theories that help explain how campaign messages may be attended to and processed (p. 156). Another study combined the ELM and Prospect Theory with respect to exercise, widely promoted as a prevention behavior for a number of chronic health co nditions. Prospect Theory holds that whether a message is framed in terms of potential gains or credibility source and a positive/gain framed message led to elaboration (Jones, Sinclair, & Courneya, 2003). This elaboration resulted in the most positive exercise intentions both immediately and two weeks later, and increased reports of engaging in strenuous exercise during that period. In addition, the team suggested the re sults indicated that noncredible sources cause d people to fail to engage in elaboration. Examinations of the ELM in the health domain seem to have increased in recent years; however, the totality of these results indicate that much works remains to be done with respect to the role persuasion can and does play in a variety of health contexts. Thus, the purpose of the current study is to augment existing scholarship by investigating the three conventional independent constructs previously discussed in depth i n the ELM literature in order to determine their applicability in health communications In addition, this research seeks to expand this research t o dependent variables and/or outcomes not previously tested with respect to the ELM, inc luding
60 additional att itudes, the concept of empowerment and health behavior intentions and action Patient E mpowerment Empowerment has been a rather loosely defined concept encompassing a range of intertwined factors. As this study deals with issues of health, it will deal wi th the more narrowly defined concept ca lled patient empowerment, a description employed only to differentiate it from other types of empowerment and not intended to imply that all health consumers are patients, an important distinction. For the purposes of this research, patient empowerment is conceptualized as the ability to take control of or influence the situation in which one finds oneself ( Nyatanga & Dann 2002). It includes the processes one utilizes to gain this power ( Nyatanga & Dann ), which often involves gathering information and resources and using them in order to gain as much knowledge as possible in order to help deal with an illness (Bulsara, Styles, Ward, & Bulsa ra, 2006) and to promote informed decision making (Williams 2002). Discussion o f p atient empowerment has grown over the past several decades with changes in the health care s ystem, including increased patients rights, competition among health providers, and organization and government promotion of personal responsibility for health a nd wellness; as well as an explosion in easy access to health and medical information (Roberts, 1999). It requires the active involvement of patients in order to increase control over the various environments and situations in which they find themselves, a nd is often considered to even the power level between p atients and providers in health care relationships (Marks Murray, Evans, Willig, Woodall & Sykes, 2006; Roberts). Although this can have negative consequences (e.g. it can inappropriately remove resp onsibility from health prov iders in some circumstances), it is
61 widely recognized that patient empowerment provides a variety of advantages to health care consumers and is desirable (Marks et al ; Roberts). However, it is noteworthy that like the lack of con sensus about its definition and ways to measure or evaluate it health consumers have vastly disparate ideas about what it means to be empowered (Roberts). The benefits of patient empowerment have been found to be significant, particularly among those who are facing chronic and/or life threatening diseases. In general, patient empowerment improves health such that those without it are less healthy, which often relates to various sociodemographic variables such as race, ethnicity, and gender (Roberts, 1999). Surveys have likewise found that health consumers value involvement and empowerment, and recognize that their ability to be empowered depends both on their access to information, as well as to their ability to establish and maintain effective relationship s with their health providers (Magee, 2003). In the U.S., 91% of survey respondents said that health consumers today asked more questions, made more choices, and actively evaluated benefits and risks more than patients did a decade ago. In addition, 80% fe lt they needed to more actively man age their own health (Magee). In a qualitative investigation specifically among cancer patients, the desired outcome of empowerment was not related to increased survival, but instead to the desire to achieve a positive at titude, increased acceptance of the disease and determination to cope with it, and enhanced feelings of control over it (Bulsara, Ward, & Joske, 2004). Attempts to dev elop a tool to measure have been undertaken, with suggest ions that health professionals may use it to help
62 and associated issues. Among the common themes found in developing the 28 item patient empower ment measure were having s ufficient resources to handle their illness, and sufficient and relevant information related to the illness (Bulsa ra et al., 2006). Health Information Seeking Healthy People 2010, which outlines U.S. government priorities for improving health among Americ participation and decisionmaking, and appropriate, quality information and support 2000 p. 10 11 ). Obtaining information and using it to gain knowledge about health and disease has been show n to have a number of benefits including improved self efficacy, positive changes in health behavior (Rimal, Flora, & Schooler, 1999), and increased health literacy (Speros, 2005), the latter of which has b een shown to be linked to a number of personal and societal health, as well as decreased health care costs, improved ability to share in making decisions about care, and reduced use of health care services (IOM, 2004; Speros, 2005). Better informed patients tend to participate more in making decisions about their care and are more satisfied with treatments, which also correlate s to broad ranging functional improvements (R amanadhan & Viswanath, 2006). Despite these known benefits, h owever, patients consistently report that they are not receiving adequate amount s and types of information from their health care providers ( Chen & Siu, 2001 ) and so they must seek information f rom other sources The Internet is one of those sources, and although numerous studies have examined the quality of health information found there, the results have been conflicting
63 due to significant variation in the type, quality and methodology employ ed. Regardless, c ancer patients have been found to be especially active in searching for health information, especially through the Internet to supplement the information they receive from their providers ( Chen & Siu, 2001 ; Huang & Penson, 2008). Doing so has been foun d to have a number of benefits. For example, i n a study of men with prostate support provided feelings of empowerment and control over health conditions and d ecision making by enabling patients to actively seek information. It also minimized risk by decreasing reliance on medical professionals and limiting inhibitions, which promoted safe disclosure of sensitive information that may otherwise cause embarrassmen t or humiliation in face to face situations. Similarly, Sharf (1997) investigated the ways in which a community developed among participants in an online breast cancer discussion group. In addition to promoting the sharing of information and experiences, this list provided social and psychological support without the geographical and scheduling constraints associated with in person groups. Participation resulted in empowerment and the ability for patients to benefit from the knowledge of others in meetin g new challenges Additional advantages to seeking health information on the Internet include increased ability to cope with the disease (Damian & Tattersall, 1991), decreased anxiety and other mental health disturbances (Rainey, 1985), improved communication (Johnson & Adelstein, 1991) and enhanced better decision making and satisfaction (Luker, Beaver, Leinster, Owens, D egner, & Sloan, 1995 ; Cawley, Kostie, & C apello 1990).
64 Assessing Empowerment Just as the concept of empowerment has been defined differe ntly by professionals and is perceived uniquely by individual patients, it has similarly been measured in a variety of ways. Often such measurements involve assessing empowerment as a psychological characteristic or disposition related to perceptions about control of health in general or specific situations including assessments of locus of control and self efficacy Locus of control relates to a general cross situational belief about the extent to which individuals believe that they can control events tha t affect them. With respect to health issues, the levels of control range from believing that events are entirely behavior and actions to feeling these events are determined by powerful others, fate or chance ( Wallston, Wallston, & DeVellis, 1978). Self efficacy on the other hand typically relates to relatively defined situations and activities A key component of self efficacy whether he/she can successfully institute a positive change in behavior and resist regressing in that action over time for example undertaking an exercise or diet program ( Bandura, 1977; Prochaska DiClemente, & Norcross 1992). Observational learning, social experience and rewar ds play heavily i nto the concept As noted, both of these concepts relate to psychological beliefs, either innate and relatively constant or dependent on the situation. As conceptualized in t his study and by others, however, becoming empowered moves beyond these types of emotional states ( Bulsara et al. 2004). It assumes that taking charge of depends not only on the specific situation but on active participation in it, including by acquiring as much knowledge as possible in order to h elp deal with an illness and promote informed
65 decision making (Bulsara et al., 2006; Nyatanga & Dann, 2002 ; Roberts, 1999 ; Williams, 2002). A number of potential measures were considered to assess empowerment in this study including self efficacy and locus of control ; however, su ch dispositions which typically are examined as independent variables, did not capture the active, information seeking aspect of the concept as identified here. Past empirical research investigating actions associated with becoming empowered is somewhat r are Some studies have looked at empowerment as an outcome variable in assessment of educational materials targeted to specific health issues such as preoperative education for ort hopedic patients (e.g. Johansson, Nuutila, Virtanen, Katajisto, & Salanter 2005). More commonly, research has examined the effectiveness of interventions designed to increase empowerment among patients by increasing their ability to ask questions of their health providers (Roberts, 1999). In a meta analysis, Edwards, Davies and Edwards (2009) sought to identify external influences on the information exchange and shared decision making that occurs during consultations between providers and patients with empowerment among the latter group among the identified outcomes (along with disempowerment and non empowerment) see k and engage in information and to appraise it in advance of a consultation with a health provider. S uch operat ionalization s fail to investiga te potentially empowering act ivities outside the health setting and therefore may be too narrow (Roberts) What transpires inside action the health care setting, e.g. in patient provider interactions, is intimately tied to what occurs outside of it, e.g. i nformation gathering by health consumers
66 (Roberts, 1999). I n order to capture the active nature of empowerment and to broaden the scope of empirical examinations about it to one of these external situation s this study therefore operationalized the concept through a series of three dependent variables that attempt to assess the degree of empowerment to seek additional information that might occur as a result of being exposed to online health messa ges about a serious cancer risk. T hese measures involved t wo standardized measures found in the literature associated with control and health information seeking, and four additional informati on from various listed sources The standar dized scales employed were Health Information Orientation (Dutta Bergman, 2004) and Perceived Behavioral Control (Ajzen, 2006). HIO is conceptually defined as the extent to which an individual is willing to look for health information (Dutta Bergman). PBC is a measure proposed by Ajzen as one of the constructs of the Theory of Planned Behavior ( along with behavioral beliefs and normative belie fs), which 1). These measures were selected because they most closel y approximated the empowerment linked to active use of health information and resources ( sbring & Nrvnen, 2004; Bulsara et al., 2006; HHS, 2000; Magee, 2003; Nyatanga & Dann, 2002). Specificall y, this study attempts to examine health information seeking empowerment that might arise from exposure to organizationally sponsored blogs. Web, and fascinating insights i nto the sociology of bloggers can be divined from the
67 analysis of the structure and content of bl is to provide insight into the aspect of bl ogspace that deals with organizational blog s related to c ancer treatment by attempting to better understand the applicability of the ELM and some of its main constructs in this context. This was accomplished by empirically test ing a series of hypo theses and answer ing four research questions in order to extend theoretical research on persuasion and ELM into the new realm of cancer tre atment messages and to new attitudes and other dependent variables including empowerment, preventive health activities, and a specific behavior. The results of this research also have practical implic ations in assist ing organizations to better target messages to health consumers, who will mutually benefit. Hypotheses and Research Questions Previous research on the ELM would suggest the following hypotheses, which will be investigated in this study: H1 When a message about cancer treatment contains strong arguments, highly involved consumers will evaluate the message more favorably than will those who are low in involvement. H2. When a message about cancer treatment contains weak arguments, consumers who are low in involvement will evaluate the message more favorably than will those who are highly involved. H3. When a highly credible source accompanies a message about cancer treatment, consumers who are low in involvement will have a more favorable at titude toward the message attitudes toward the treatment message will be relatively unaffected by source credibility. H4. When a message about cancer treatment contains strong argu ments, highly involved consumers will evaluate the cancer treatment more favorably than will those who are low in involvement. H5. When a message about cancer treatment contains weak arguments, consumers who are low in involvement will evaluate the cancer treatment more favorably than will those who are highly involved.
68 H6. When a highly credible source accompanies a message about cancer treatment, consumers who are low in involvement will have a more favorable attitude toward the cancer treatment than w hen a less credible source is used. Highly involved affected by source credibility The ELM has not previously been used as a framework to test as dependent measures either attitude towar d an organization attitude toward an organizational blog or consumer empowerment to seek health information or with respect to health behavioral intentions or actions of the kind discussed her e. T herefore hypotheses cannot intelligently be made and thus the following research questions are appropriate: RQ1: What effects do involvement level, argument strength, and source have on attitudes toward a cancer treatment organization? RQ2: What effects do involvement level, argument strength, and source have o n attitudes toward an organizational blog as a channel of cancer treatment communication? RQ3: What effects do involvement level, argument strength, and source have on personal feelings of empowerment to seek health information ? RQ4: What effects do invo lvement level, argument strength, and source have on behavioral intention to seek thyroid cancer screening? On intention to stop using ake informational material? As stated, predictions about the outcomes for the previously untested attitudinal and empowerment measures cannot be made, thus research questions were proposed. However, Figure 2 1 contains a simplified graphic representation of the ELM that attempts to more clearly depict how both the independent a nd dependent variables being investigated here may fit within the context of the Model
69 Figure 2 1. Simplifie d graphic representation of the ELM developed by Petty and Cacioppo (1981, 1986). The representation shows how the independent and dependent variables being investigated in this experiment may fit within the context of the Model. name of the dependent variable signifies that at least one hypothesis was proposed has not been previously tested within the ELM framework and so a research question was posed.
70 CHAPTER 3 METHODOLOGY This study tested the previously stated hypotheses and examine d the research q uestions posed using the Elaboration Likelihood Model of persua sion as a framework. A 2 (involvement: low vs. high ) x 2 (message strength : weak vs. strong ) x 2 (source credibility: low vs. high ) between subject post test only experimental design was utilized This experiment was conducted to identify the influe nce that each of these three independent variables had on a number of depende nt outcome variables that involved several attitudinal measures feelings of empowerment, and behavioral intention and action, all of which will be described subsequently. Indepen d ent V ariables Involvement I nvolvement in the cancer scenario i.e. the issue of getting cancer from cell phones that was described in the cover story news re lease, was assessed using a three item me asure. It asked respondents to rate the information on t he following seven point semantic differential scale: important/unimportant, means a lot to me/means nothing to me, not needed/needed (Lord & Burnkrant, 1993). The two cover story news release involvement manipulations can be seen in Appendix A Although n ot typical, an assessment of i nvolvement in the experimental blog message itself also was conducted using a scale previously employed by Lord, Lee & Sauer (1994, 1995). This manipulation was checked by asking participants to respond e reading the cancer treatment information included on the blog, I was point scale items. These items included the following three semantic different measures: very uninvolved/very involved, concentrating very
71 little/concentra ting very hard, and paying very little attention/paying a lot of attention. The fourth item asked respondents to rate on a seven point Likert type scale ranging from strongly agree to strongly disagree the statement: I carefully considered the informati on in the blog about the therapy. Message Quality A manipulation check of argument strength/ quality was measured Participants were asked l you LEARNED about positron treatment from the blog you read? They were asked to respond to the question by rating their extent of learning previousl Boyd, & Brown (1998). Source Credibility A manipulation check of s ource credibility was evaluated by t wo combined scales, expertise and trustworthiness, traits which have been most commonly discussed in the literature and measured with respect to credibility. The scales used here are modifie d from those developed by Ohania n (1990), and consist of two sets of bipolar adjectives (11 items total) rated on a seven point the blog is five trustworthiness bipolar adjectives were: undependable/dependable, dishonest/honest, unreliable/reliable, insincere/s incere and untrustworthy/trustworthy. The six adjective pairs measuring expertise were: not an expert/expert, inexperienced/experienced, unknowledgeable/knowledgeable, unqualified/qualified, unskilled/skille d, and untrained/trained. Ohania n also used a fi ve item attractiveness scale when rating celebrity endorsers, which was not used here because the sources in the experimental materi als were the same person, in the same background and pose, and having the same facial expr ession; thus, the ratings of
72 attra ctiveness would have been expected to be equally distributed across this random population regardless of other factors. In the literature, the separate measures of trustworthiness and expertise are most commonly assessed jointly, i.e. as an overall mean cr edibility score calculated by combining the two subscales, which is what will be done for the purposes of th is dissertation. However, examin ing these measures separately in future studies may be warranted to increase understand ing of health messages Abil ity In addition to involvement, which indicates the personal importance of a message, ano ther factor that determines the route used to ability to process a communication. In this study, this facto r was presumed to be h igh across the population This assumption was made because all p articipants had achieved above average education levels, such that they all were high school graduates and had completed at least some college Dependent M easures Attitudes In order to gauge attitude toward the treatment message, participants were asked to describe their perceptions about the quality of the information presented in the blog information on a three item seven point semantic diffe rential scale. The scale consisted of the followi ng bipolar phrases : very weak/very strong, not very convincing/very convincing, and not very powerful/very powerful. The measure had been previously used by G rhan Canli & Maheswaran (2000). Attitude toward positron treatment was assessed using a seven po int semantic differential scale previously used by Stayman and Batra (1991). Participants were asked
73 thyroid cancer from using a cell phone, using positron treatment would be items were useless/useful, not beneficial/beneficial, low quality/high quality, and worthless/valuable. Attitude toward the organization (Positron Treatment and Research Institute) was determined using a seven point semantic differenti al scale The scale included the following five bipolar adjectives: unde pendable/dependable, dishonest/ honest, unreliable/reliable, insincere/sincere, and untrustworthy/trustworthy. It was developed by Trifts and Haubl (2003). Attitude toward the blog chan nel was measured using a scale adapted from Homer (1995). Participants were asked to respond t ratin g each of seven qualities on a seven point Likert qualities were: believable, interesting, informative, well designed, easy to follow, attention getting, and useful. Consumer Empowerment For the purposes of this study, empowerment was conceptualized as the ability to take control of situation (Nyata nga & Dann, 2002). It includes the processes utilized to gain this power and using them in order to gain as much knowledge as possible in order to help deal with an illness and to promote informed decision making (Bulsa ra et al., 2006; Nyatanga & Dann, Wil liams, 2002). As a result, thi s concept was operationalized using three scales used to measure the active process of gat hering information and resources Health Information Orientation (HIO) was measured here using a seven item Likert type scale ranging f rom stro ngly agree to strongly disagree, with the responses
74 about health issues in general (Dutta Bergman, 2004). The to read and watch sto edicine, I try to get as much information as possible about its benefits hing I can about th The se cond standardized scale was Per ceived Behavioral Control (PBC), which Ajzen (2006) pr oposed as one of the constructs of the Theory of Planned Behavior (with behavioral bel iefs and normative beliefs), as a measure that for actual control and contribute to the prediction This four item scale, with two items measuring capability and two item s measuring controllability on seven point seman tic differential scales are meant to capture peo confidence that they are capable of performing the behavior under investigation As prescribed, the four items used in this study followed example statements provided by Ajzen and were developed based on the behavior of interest, which here was opera tionally defined as becoming knowledgeable about treatments for thyroid cancer. The four items and their associated bipolar adjectives used in this experiment were: impossible /p / d efinitely false; How much control do you believe you
75 have over becoming knowledgeable about treatments for thyroid cancer? control /c omplete c ontrol It is mostly up to me whether or not I become knowledgeable abou trongly agree /s trongly disagree The first two items related to capability and the last two to controllability. The remaining four empowerment i tems assessed behavioral intention related to seeking additional information specifically about positron cancer treatment for thyroid cancer from cell phones from each of four different sources. E ach statement was measured on seven point scales using the t wo bipolar adjective unlikely/likely and (G rossbart, Muehling & Kangun, 1986) The four statements nal information about proton therapy from ron Treatment and Research Institute, Other D ependent Measures Two additional behavioral i ntention items and one actual behavior also were assessed. The behavioral intention items related to likelihood of attend ing a health screening is considering offering thyroid cancer screening to its students for a modest cost. If they behavior item involved specific action to take or not a sheet of contact
76 information for the Positron Treatment and Research Institute organizatio n that provided the positron treatment and spons ored the experimental blog. Stimuli Involvement was manipulated through a cover story in the form of a news release detailing a fictitious research study by Harvard University cancer res earchers that found a link between cell phone use and thyroid cancer in college students. The cover story for the high involvement group stated that head and neck scans found that more than one third of university studen ts 18 22 who participated in the stu dy had small to moderate sized tumors, the majority of which were cancerous while t he low involvement story suggested the thyroid cancer was detected in a much smaller percentage ( 1 of 1,000 ) of college students The cover story was described to particip ants as an as yet to be released news release that had been sent by Harvard to cancer treatment organizations nationwide to allow them to prepare to respond to questions by the public afte r the news was released Participants also were told that one of tho se cancer treat ment org anizations contacted communication faculty researchers at the University of Florida asking fo r assistance in evaluating information contained on blogs the organizat ion create d in response to this news to determine how to best present information about its treatment to the public. This deception was necessary in order to a) help ensure participan ts were unaware that it concerned their persuasibility, b) adeq uately manipulate involvement in the issue of cancer in a college age popula tio n, which has traditionally been hard to engage in this topic because it is considered a disease that occurs among older people, and c) to minimize demand characteristics. The only other differences between the two cover story manipulations related to the s everity of the threat, e.g. the low involvement story stated the link
77 a phones (the bar height of w hich differed based on the low versus high involvement scenario), lighten ing strikes, and heart disease. The second set of stimuli included t he organizational blog manipulations Participants were each exposed to one of four diffe rent experimental blog messages about the cancer treatment formatted on commercially available blo gging software. The four different blog message manipulations were based on combinations of source credibility (high vs. low) and argument strength/quality (strong vs. weak), as described further below. The format of the blog messages were identical in ter ms of organization name, size and location of logo and source photo, blog title, format of text, number of arguments, etc., with the conditional differences noted below. Each blog contained the following items described previously as being typical of most blogs These included a header with the name of the treatment organization sponsoring the blog (Positron Treatment and Research Institute) and the logo of the blogging software used (Blogger) along with the email address that was created for the source of the post (JoyceMarshall.PTRI@gmail.com). The main channel contained a dated page long message, or post, traditional last line identifying the writer and the time the comments were posted. The main posting included an introduction which was the same for each of the four
78 conditions, related to recent news about cells phones causing thyroid cancer and indicat ing that positron trea for treat ing it. This was followed by a section listing five bulleted advantages to positron treatment. Depending on the condition, these advantages contained either five strong/high qualit y or five weak/low quality statements For the purposes of this study, strong statements were operationalized to be those that related to the technology associated with the positron treatment and its curative properties, while the weak arguments involved aesthetic aspects of the city or the Institute and other issues distantly related to the treatment itself, such as social support (McCullough & Dodge, 2002). Other ELM studies have operationalized argument strength/quality in different ways; however, the d istinction chosen here adhered to the operational definition used in the only other empirical test of organizational health messages that was found and so was felt to be the most appropriat e. The positron treatment was fictitious, with its name and the or ganization nam e fabricated solely for the purpose of this experiment. However, the information in the messages about it were true and were based on publicly available inform ation. T he real name of the treatment was not used in the experimental materials in order to prevent any potenti al familiarity or other issues. The final paragraph of the main post, which was the same for all conditions solicited comments and/or questions, providing a telephone numbers for those preferring to call.
79 The source manipulations, consisting of either a physician (high credibility), or a public relations spokesperson (low credi bility) source/author were cont ained in the limit confounding of the source variables as much as possible with respect to attractiveness, ethnicity, gender and other personal attributes, the same person and name (Joyce Marshall) were used for both the physician and PR spokesperson. However, in addition to the titles/descriptions accompanying the photos, the woman was dressed differently in each photo to provide approp riate cues to their different roles (lab coat and stethoscope for the physician and suit jacket for the public relations representative). This section included, f rom top to bottom, a head shot photograph of the source/author and two t ext box es, one above t he photo that included her name /title and a degree or certification ( Dr./ MD or APR [ Accreditation in Public Relations] ) and one below containing a position description ( cancer specialist o r public relations representative) ; a blogroll ink s, including for the Positron Treatment and Research Institute, the National Cancer Institute and Harvard University Cancer Research Institute ; and a blog archive listing the current post. Thus, the four blog conditions were: high credibility physician s ource with strong/core service relevant message, high credibility physician source with weak message/ non core service related message low credibility public relations representative with strong/core service message, and low credibility public relations re presentative with weak / non core service related message. The four blog manipulations can be seen in Appendix B
80 Pretesting The experimental materials were pre tested to improve the likelihood of developing an effective and credible cover story and succes sful manipulations of the indep endent variables (Aronson, Ellsworth, Carlsmith, & Gonzales, 1990). A total of 177 participants were involved in the pre testing which was done in groups ranging from 16 to 65 people P articipants were told the testing was p art of a project evaluating health messages, which College faculty had been asked to assist with by a cancer treatment organization. In all cases completion of the survey questionnaires was f ollowe d by an informal group discussion during which questions were posed to participants to elicit their thoughts or feelings ab out the cover story, source, arguments instructions, etc. depending on the pretesting being conducted. The discussion sessions were recorded tes were taken by the researcher which were used to make subsequent modifications. Four between subject pretests were conducted to test the involvement manipulation cover stories, with revisions made between each pretest in an effort to increase the mean differences in involvement level to the greatest extent possible and to decrease possible confounds. As noted previously the cover story manipulations were identical with the exception of the words o r phra ses intended to evoke either high or low levels of involvement in the issue (relative to each other) which was getting thyroid cancer from cell phones The first pretest included two conditions, one intending to stimulate high involvement and the other to induce low involvement with the distinctions re lated to a) afflicted with thyroid cancer, and c) the length of time victims had used their cell phones.
81 Specifically, the low involvement scenario stated that cancerous thyroid tumors were found in 1% of older adults 55 and older who had used cell p hones for a decade or more, while the high involvement condition stated that these tumors were found in one third of college age students 18 22 no matter how long they had used ce ll phones (i.e. even with short term use) Among the 25 participants, means for the high involvement condition (M = 24.23) were higher than for the low involvement condition (M = 23.50), with internal reliability of .76. H owever, means even among those in t he low group showed r elat ively high levels of involvement, which was unexpected as research has shown that adolescents often fail to feel vulnerable to negative health outcomes and underestimate the magnitude of the risks to which they are expose d (e.g see Johnson, McCaul, & Klein, 2002). As a result, change s were made to the cover stories and an additional c ondition was added for the second pretest The changes involved removal of r eferences to the le ngth of cell use as well as quoted comment s by a person in the vulnerable age group that related to worry/ lack of worry about the research findings described A second low inv olvement cond ition was added which substituted college age students 18 22 for the 55+ population Among the 24 subjects who parti cipated, the mean among the high involvement group (M = 23.50) was higher than among both low involvement groups (M = 20.25 for the 55+ population described in the manipulation and M = 20.38 for the college population). Internal reliability for the four items was .79. O nly a small mean difference between the two low involvement conditions was found ; t herefore, the condition involving the middle age group was eliminated in all future t esting to decrease potential confounding
82 An additional modification wa s ma de in advance of the third pretest in an attem pt to further differentiate mean involvement levels between the low and high involvement groups C larifying phrase s were added to the conditions to further describe the likelihood of the thyroid cancer cau sing occurrence. The high invol vement condition added that it was a serious health issue, while the low involvement story indicated the likelihood was extremely rare, roughly equivalent to getting hit by lightening. Thus pretest involved 17 subjects with means showing that those who received the low involvement cover story reported lower involvement levels ( M = 21.91 ) than did those who received the high involvement story ( M = 25.00 ) Internal reliability for the four item measure was .64, which increased to .82 with removal of the relevance item. Subsequently, several additional modifications were made to the two cover story manipulations for the fourth pretest In the low involvement manipulation the percentage of students found to have thyroid tumors w Adjectives were added to the lead sentence s of both cover stories to qualify the cell phone thyroid cancer link ( ). In addition, histogram s were added at the end of the story/botto m of the page for each manipulation The histograms depicted a three way comparison between thyroid cancer from cell phones getting struck by ligh tning and heart disease with the bar height for thyroid cancer varying depending on the low/high condition. Fourteen sub jects participated. Means on the four item involvement measure were higher among those who received the high involvement condition ( M = 23.0 ) than among the low involvement group ( M = 20.25 ) Internal reliability for the four items was .47, which
83 increased to .64 if the relevance item was deleted. Means amon g the three item measure (excluding relevance) were M = 17.33 and M = 14.25 respectively. The ANOVA s for both the three and four item involvement measures were no t significa nt among the relatively small pretest sample sizes ; however, the differences between the high and low involvement means increased between the first and last pretest approaching significance As a result, just o ne additional change was made to both cover story conditions before use in the final experimental data collection. T he histogram titles were changed to Lifetime p robab i lity of occurrences in order to more accurately reflect the likelihood of heart di sease risk. Within subjects p retesti ng also was conducted among 101 participants to determine characteristics of the arguments/statements that were to be used in the experimental blog. In total, 38 arguments were pretested for strength/quality in addition to their believability, comprehensi veness, complexity and familiarity ( Petty & Cacioppo, 1986 ) As noted previously, high strength arguments related to the curative properties of the treatment, while low strength statements involved elements that were more aesthetic (McCullough & Dodge, 200 2). All of the state ments were true. In the first of three argument pretesting session s each of 72 subjects evaluat ed half of the arguments for either strength or the other four traits in order to prevent fatigue and decrease the likelihood of indiscrimin ate responses. An average strength rating was calculated for each statement, which was used to identify the highest and lowest quality statements Means on the other four factors also were calculated, and the arguments having the 10 highest and 10 lowest m eans that also were roughly equivalent in the levels of each of the four criteria were identified for further pretesting (Petty & Cacioppo,
84 1986) A n additional 29 subjects rated these 20 arguments on four items : strength, persuasiveness, understandability and complexity. In order to prevent confounding of the results, the five strong and five weak statements/arguments used in the blog message were ascertained by taking the five highest quality and five weakest statements tha t also were roughly equivalent o n mean level s of understandability and complexity. Between subjects pretesting of the two sources (PR spokesperson and doctor) also was conducted among 25 subjects. It is worth noting that this source pretesting session and others conducted early on in th e pretesting process included only measures for a single independent variable Later pretesting included multiple variables and more closely followed the procedure to be utilized in the actual experimental testing in order to identify and resolve any issue s that may have surfaced as problematic. This process also explains the disparity between the number of subjects participating in the pretesting and the numbers among whom each variable was tested. Participants A total of 331 undergraduate students 18 ye ars and older from a large Southeastern university participated in the study to receive extra credit. Of these, 177 were involved in pre testing the experimental materials, and the remaining 154 participated in the main experiment. Students were selected t o participate because they are more readily accessible than other populations, and researchers have found similar responses between college students and non students. For example, in an experiment investigating variables similar to the ones proposed in thi s study with respect to hospital advertising, McCullough & Dodge (2002) found no significant differences in responses between college students and employees of a financial institution. In addition, because
85 young adults of this age are often difficult to en gage in health related issues as discussed previously, this experiment was devised to render students as stakeholders by identifying a health threat that was anticipated to evoke current interest among this population Experimental Procedure P articipants w ere recruited for the actual experiment through three specific classes. In order not to divulge the true purpose of the experiment and help ensure experimental realism individuals were told the extra credit involved assisting to evaluate health marketing messages. Two of the three classes were large introductory advertising lectures (different sections of the same course), and the other was a large lecture mass communication course. Both courses but particularly the latter draw major s from throughout the university. T he experimenter developed a detailed script for the data collection sessions that was read to each group and she personally conducted all of the sessions in order to limit to the greatest extent possible any potential confounding among th e various groups by different language, explanations, moderator, etc. The instructions were standardized so that all participants could come to a similar under standing of what the research was about and what their responsibilities would be, and to ensure t hat all questions participants had were answered before the st art of the experiment (Aronson et al. 1990). Because of a timed thought listing measure (described below) and the need to debrief participants at the end of the data collection due to the decep tion involved it was necessary for all of them to progress from section to section simultaneously. T herefore, the experimenter read all of the instructions aloud from the script and as printed in the data collection booklets With the exception of the tho ught
86 listing measure, participants were provided as much time as they needed to finish each section, and the testing did not proceed until it was verified that everyone had finished. In order to help ensure participants who worked more quickly stayed invol ved in the subject, all papers and other materials were required to be removed from the desk tops and put away (including cell phones, laptops, newspapers, notebooks, etc.) P articipants were told these materials could not be accessed during the entirety o f the session and were provided an opportunity to leave the session before it started if they were unable to comply Twelve sessions were held over the course of a week in two different classrooms having similar setups, with group sizes ranging from 8 to 21 Limited availability required the use of two classrooms, and the sessions were held at various times in order to accommodate as many students as possible who wished to receive extra credit. The entire procedure took about 60 minutes, including 45 minut es to describe the experiment, obtain consent and work through the booklets, and the remainder for debriefing and questions. The researcher noticed no indications of participant fatigue during the sessions (e.g. participants pulling out other materials or talking to others) and in all cases the participants involved completed the entire contents of both booklets, despite being told they could discontinue answering the questions at any point. In addition, although written comments on the debriefing forms wer e rare, there were none indicating the experiment was too long or uninteresting. Among the handful of ng. However, that was not the case across the bo ard. One subject who was disconten ted with the process
87 Subjects were randomly given a pre numbered packet of experimental materials. The materials were placed into the envelopes prior to being numbered, and the researcher had no knowledge of which condition was placed in to each envelope. The eight conditions were run simult aneously, with each session including as many of them as possible decreasing potential confounds. T he researcher also was blind to the condition that each participant received so that she could not intentionally (or otherwise) influence participants in ce rtain conditions to behave in desired ways. The order of the experimental procedure in each data collection session was the same: introduction to the process, consent signing and form collection random distribution of envelopes containing experimental mat erials, completion of questionnaire booklet 1, completion of questionnaire booklet 2, collection of envelopes containing experimental materials, debriefing, and collection of the signed debriefing form s, which allowed subjects an opportunity not to have th eir responses used. The experimental materials were containe d in two booklets. B ooklet 1 was 11 pages (including the cover) and contained the following s ubstantive items: a ) an overview of the study and instructions about the news release cover story b ) a cover story inv olvement manipulation ( low or high ), c ) a series of post cover story questions, including those that functioned as a manipulation check of cover story i nvolvement two recall questions related to information contained in the cover story an d measure s of fear and personal cancer risk ( the latter two of whic h were collected as pilot data and will not be discussed in this dissertation ), d ) an introduction to and instructions for the
88 experimental blog message, e ) the experimental blog message ab out t he novel cancer treatment, and e ) a thought listing measure & Cacioppo, 1981, p. 252). P articipants were asked to spend fiv e minutes recording in their own words everything that went through their minds as they read the cancer treatment blog without going back to examine it. More specifically, they were asked to list all the thoughts and feelings they had about anything as the y read the blog, and also were asked to identify whether each of these thoughts or feelings was positive, neutral, or negative by putting a +, 0, or after each statement (Cacioppo & Petty ) The resultant comments were scored and analyzed with respect to the total number of thoughts listed overall and the number of positive vs. negative thoughts. The ELM suggests those who are highly involved will generate more thoughts overall; thus, this was used as a manipulation check of involvement in the blog messag e. In addition, favorable, or pro attitudinal, thoughts result in enduring persuasion while negative, or counter attitudinal, thoughts may have a so being advocated in the me ssage (Petty & Cacioppo) Booklet 2 contained the final series of questions and written instruct ions for answering them Throughout the experiment, the researcher read the instructions aloud to participants, who could read along with read them in their boo klets. The booklet 2 questions included the dependent measures including those measuring attitudes about the message/information, the treatment, the organization, and the blog; as well as how empowered they felt, including to seek additional i nformation about the treatment, and
89 several behavioral intentions. It also included manipulation check items to test involvement in the blog message, source credibility, and message quality/strength Most measures employed i n this study had been utilized i n prior pub lished literature and had reliability measures well above minimum levels. These scales/measures were carefully selected for their high intercorrelations in order to improve reliability (Cook & Campbell, 1979). In addition, the random assignment of partici pants to the eight conditions eliminated many threats to internal validity ( Cook & Campbell 1979 ). Most other threats that randomization does not rule out (e.g. multiple testing occurrences, and compensatory equilibrium, compensatory rivalry, demoralizati on that result from some groups receiving less desirable) were not factors in this experiment as no differentially desirable treatments were included, and only a single testing occurrence was used. Only one participant acc epted an offer not to have his/her responses used, which was questions in the booklets (3) all of them answered at least 90%, thus differential mortality also was not an issue. Demographic and other baselin e information questions we re also included at the end of b ooklet 2. In addition to usual demographic information such as age, gender, year in school and major, other information was collected, including personal cellular telephone use, personal experience with cancer, familiarity with positron treatment and purpose of the study. Most of these items were collected via check off items; however, participants also were asked to w rite short answers to identify their relationship to family
90 members or close friends who had been diagnosed with cancer, their perceptions about purpose of the study a s as sessing their persuasibility. At the conclusion of the experiment, all participants were thoroughly debriefed about the fictitious nature of the mobile phone re search cited in the cover story. They also were provided an opportunity to ask questions abo ut the study (both in a groups setting and on a one on one bases). P rinted background information about cell phone use and brain cancer was described to participants and made readily available for anyone who was interested to pick up and take with them whe n they left the testing site. After all questions had been answered and the debriefing forms collected, the participants were thanked for their participation and dismissed. The entire procedure took about 60 minutes, including 45 minutes to describe the ex periment, obtain consent and work through the booklets, and the remainder for debriefing.
91 CHAPTER 4 RESULTS The Sample In total, 154 undergraduate university students participated in the experiment. They represented 41 majors and included those in the ir first (15.8%, n=24) through fifth years of college (n=3). Two sets of responses were removed, one because the participant requested it in writing as allowed on her debriefing signature form, and the other because the subject wrote that she knew the cove r story was false after having been diagnosed with thyroid cancer four months earlier. As a result, the data set used for a majority of the analyse s consisted Just under two thirds were women (64.5%, n=98). The majority were whi te (61.2%, n=93), approximately equal to that of the university undergraduate population overall (62.5%), and their average age was 19.9 years. As is the case in the university population as a whole, Hispanics accounted for the second largest ethnic group in this study, followed by blacks/African Americans. Those majoring in advertising comprised 19.7% (n=30) of the total, while other majors each consisted of less than 10% (most were less than 5%). A more complete demographic breakdown is shown in Table 4 1 All respondents reported using a cell phone regularly. The mean number of times cell phones were used per day for calling people was M = 8.24, SD 7.06, wit h a median of 6.0 and a range 1 -50. On average, participants spent 2.28 hours per day calling peo ple (SD 2.87), with the median 1.5 hours and the daily range from less than one hour to 24 hours. A majority (89.5%, n=136) indicated they had previously read a blog about any topic, with more than half (56.6%, n=86) reporting they had read a blog whose pr imary purpose was to talk about health or medical issues.
92 More than three quarters of subjects reported prior experience with cancer. Four (2.6%) had been personally diagnosed with the disease (including the subject who had thyroid cancer) dating back to four months, and one, three and 10 years earlier. However, no participant reported ever receiving radiation treatment (the treatment described on the experimental blog was a type of radiation). In addition, 76.3% (n=116) reported that a member of their fam ily or a close friend had been diagnosed with any type of cancer. Subjects were asked to list the people they knew who had been diagnosed with cancer, and the number of relationships listed (e.g. grandfather, best friend, next door neighbor, etc.) ranged f rom one (40.9%, n=56) to six (1.5%, n=2). Where possible, these relationships were further classified for analysis into immediate family members (which were operationally defined in this study as parents, step parents, or siblings). Among the 139 responses where the type of relationship could be ascertained (some participants instead listed types of cancer rather than relationship), one in five (20.86%, n=29) identified an immediate family member, with the vast majority of these subjects (86.2%, n=25 of 29) indicating they had had one immediate family member with cancer, while four (13.8%) reported tw o such close relationships. None of the subjects correctly identified the purpose of the experiment. In addition, as expected, mean fam iliarity ratings were low on the two item, seven point fami liarity scale (14 points total). The mean ratings for t he fictional positron treatment were M = 2.59, SD 1.800, and for the Positron Treatment and Research Institute were M = 2.58, SD 1.71 Independent Variable Manipulati on C hecks Manipulation checks were conducted on each of the three independent variables: involvement in the cover story, source credibility, and argument strength. As described
93 below, the ANOVAs for these checks all found the expected main effects and were successful. Unexpected effects also were found as noted subsequently Internal reliability analyses also were conducted for the multi item measures used to assess measu res were greater than .80, and all item total correlations were greater than 0.30, which is above the level needed to provide good support for internal consistency reliability (Morgan, Leech, Gloeckner, & Barrett, 2004). A qualitative thought listing measu re was used, in part, as a manipulation check of involvement in the experimental blog messag e about the cancer treatment. The involvement manipulation check was conducted immediately after participants received the news release cover story manipulation and before participants had been exposed to the source or to the arguments. The ANOVA revealed the expected simple in the cover story, F(1, 142) = 25.75, p < .001, eta = .39. The means showed that participants who read the high involvement news release were more involved in the issue (M = 18.48) than were subjects who received the low involvement release (M = 15.86). No other main or interaction effects were significant. T he ANOVA for the 11 item overall source credibility manipulation check (consisting of trustworthiness and expertise) revealed an expected main effect of source, F(1,144) = 12.21, p = .001, eta = .28. Means showed that participants rated the PR source as si gnificantly lower in credibility (M = 43.89) than the physician (M = 50.68). An unanticipated effect of argument strength also was also found, F (1,144) = 37.19, p < .001, eta = .45. Participants exposed to the strong arguments rated the
94 source as signific antly more credible (M = 53.21) than did those who received the weak arguments (M = 41.36). No other significant effects were found. As expected, separate ANOVAs for the split credibility measure, i.e. for the five item trustworthiness measure and for the six item expertise measure likewise found effects based on source and argument strength; thus, they w ill not be detailed separately. As noted, extensive pretesting of the arguments was conducted, which yielded significant mean differences between the five strong and five weak arguments/statements used in the blog. Therefore, verification of the argument strength manipulation was assessed with a single item rated on a seven point scale ranging from you feel you learned about significant main effect of argument strength, F (1,144) = 122.06, p < .001, confirmed that participants who read the blog containing the strong argumen ts reported they learned significantly more about the treatment (M = 4.42) than did subjects who received the weak arguments (M = 1.88). No other main or interaction effects were found. A thought listing activity was used, in part, as a manipulation check of involvement in the blog message about the treatment. The ANOVA for the total number of thoughts and/or feelings listed about the treatment message found no significant main effect of involvement, such that involvement in the issue had no significant inf luence on the extent to which the blog message about the treatment generated thoughts or feelings among the participants. No other significant main or interaction effects of source credibility and argument strength were found.
95 With respect to the number of positive thoughts generated after reading the blog, the ANOVA found a simple main effect of involvement on positive thoughts, F(1,144) = 8.67, p = .004, eta = .24. The means showed the unexpected finding that lowly involved participants generated more fav orable thoughts and feelings (M = 2.14) than did those who were highly involved (M = 1.47). A second main effect of argument strength on number of positive thoughts also was found, F (1,144) = 19.01, p < .001, eta = .34. The means showed that participants who read the strong arguments generated significantly more positive thoughts and feelings (M = 2.31) than did those exposed to the weak arguments (M = 1.31). No other significant effects were found. The ANOVA for the number of negative thoughts generated f ound a simple main effect of argument strength, F (1,144) = 16.31, p < .001, eta = .32. The means showed subjects who received the weak arguments generated more negative thoughts (M = 3.37) than those reading the strong statements (M = 2.09). Hypothesis T esting organization and blog, and for the empowerment measures devised for this study (which related to obtaining additional information about the cancer treatment from various source s) were all .90 or above. Inter item correlations for the attitudinal measures toward the message, the treatment and the organization all were higher than .30. The alpha for the seven item standardized Health Information Orientation (HIO) empowerment measu re was .83, with two correlations falling below .30. The four item Perceived Behavioral Control (PBC) measure showed low internal reliability at .48. It appears this is likely related to two of the four items being reversed in the questionnaire, which was not the case with most of the other items, including two pages of questions
96 leading up to the PCB measure. For example, several people rated as maximally Analyses of the dependent variables including the sociodemographic information (gender, race, age, daily cell phone use, and family history of cancer) as covariates found no effects except in one case, verifying that the findings were due to manipulations of the independent variables rather than to individual differences. The one case where this was not confirmed was with respect to the effect gender had on attitude toward the treatme nt, F (1,141) = 6.51, p = .012, eta = .21, a small to medium effect. A t test showed that males rated the treatment significantly more favorably (M = 21.74) than did their female counterparts (M = 19.18). Attitude toward the Treatment Message The ANOVA fo r the dependent measure assessing attitude toward the treatment message found an expected effect of argument strength, F (1,143) = 117.49, p <.001, eta = .67, a much larger than typical effect. Means showed that subjects who received strong arguments rated the message more favorably (M = 13.17) than did those who were exposed to the weak message (M = 6.63). A second effect of source also was found, F (1,143) = 9.00, p = .003, eta = .24. Participants assigned the doctor source responded more positively to th e message (M = 10.80) than did those assigned the PR source (M = 8.99). As expected, a significant two way interaction was found between source and involvement on attitude toward the treatment message, F (1,143) = 7.47, p = .007, eta = .22. Among low invo lvement participants, a simple main effect of source was found,
97 F (1, 74) = 11.22, p = .001, eta = .36. For subjects low in involvement, the message associated with the doctor was rated more favorably (M = 12.19) than the message associated with the PR so urce (M = 8.70). Among highly involved subjects, there was no difference in perceptions of the treatment message as a function of source. message. The first proposed that hig hly involved consumers who receive a message about a cancer treatment containing strong arguments will evaluate the message more favorably than will those low in involvement. This was not supported. Both high and low involvement groups rated the strong mes sages similarly (M = 13.20 and 13.14 respectively). A second related hypothesis posited that a weak treatment message would lead low versus high involvement consumers to evaluate the message more favorably Means showed when the treatment message contained weak arguments, low involvement participants did rate the message more favorably than did those who were highly involved. However, the difference was not statistically significant (p = .09); therefore, H2 was only partially borne out. A third hypothesis st ated that low involvement subjects will rate a message more favorably when it is associated with a high as compared to a low credibility source, while high involvement subjects will be unaffected by source credibility. As predicted, low involvement partici pants who received the high credibility source treatment message rated it more favorably than did those receiving a message from a low credibility source. In addition, attitudes toward the treatment message among highly involved subjects were unaffected by the source. Thus, H3 was fully supported.
98 Attitude toward Positron Treatment For the measure assessing overall attitude toward positron treatment, the ANOVA showed two significant main effects. The first was related to source credibility, F (1,143) = 5.71 p = .018, eta = .19, a small to medium effect. Means showed that subjects who received the blog containing the high credibility physician source reported more favorable attitudes toward the treatment (M = 21.26) than did those exposed to the low credibil ity PR spokesperson (M = 19.14). Attitude toward the treatment was also influenced by argument strength, F(1,143) = 60.03, p < .001, eta = .54, a much larger than typical effect. Means showed those exposed to the strong arguments had more favorable attitud es toward the treatment (M = 23.63) than did those who received the weak arguments (M = 16.77). No other effects were found. The second set of hypotheses addressed attitude toward the positron treatment, with H4 proposing that a message about the treatmen t containing strong arguments would lead highly involved consumers to evaluate the treatment more favorably than will those who are low in involvement. This study found no interaction effects between involvement and argument strength; thus H4 was not prove n. H5 asserted that weak arguments in a cancer treatment message will result in low involvement consumers evaluating the treatment more favorably than will those who are high in involvement. And while the means showed that the less involved subjects did te nd to rate the treatment more favorably than did more involved subjects (M = 17.78 and 15.65 respectively), the difference was not statistically significant (p = .13), providing only partial support for this hypothesis. The final hypothesis dealt with the influence of source credibility on attitude toward the treatment. It was proposed that low involvement consumers will have more
99 favorable attitudes toward the treatment when the source is high as compared to low in credibility. Attitudes of high involvem ent subjects were expected to be unaffected by the source. Across all subjects, source did have a significant effect on attitude toward the positron treatment; however, this did not depend t on involvement level. Th erefore H6 was not supported. Research Qu estion Exploration Four research questions also explored attitudes that have not previously been examined within the framework of the ELM. Specifically these questions addressed how involvement, source credibility, and argument strength influenced attitu des toward a cancer treatment organization and toward an organizational blog. These questions also addressed feelings of empowerment, behavioral intention concerning cancer screening and mobile telephone use; and a behavior involving taking a specific acti on related to obtaining additional information about the treatment and/or the organization. Attitude toward the Organization RQ1 considered the effects that involvement level, argument strength, and source had on attitudes toward a cancer treatment organiz ation called the Positron Treatment and Research Institute. The ANOVA identified two significant main effects. There first was an effect of source, F (1,143) = 6.74, p = .010, eta .21, a medium effect. Means showed that subjects viewing the doctor source h ad a more positive attitude toward the organization (M = 23.38) than did those who saw the PR source (M = 21.12). A second significant main effect of argumen t strength also was found, F (1, 143) = 39.98, p < .001, eta = .47, a large effect. Means showed tha t those who read strong arguments reported significantly more favorable attitudes toward the organization (M = 25.00) than did those who read weak arguments (M = 19.49).
100 The analysis also revealed a significant two way interaction between involvement level and source, F (1,143) = 10.31, p = .002, eta .26, a medium or typical effect. A simple main effect of source was found for those low in involvement, F (1,74) = 17.23, p<.001. Among low involvement subjects, the high credibility physician source created mo re favorable attitudes toward the organization (M = 25.2) than did the low credibility PR source (M = 20.1). No simple source effect was found for high involvement subjects. No other two or three way interactions were found. Attitude toward the Blog The s econd research question explored the influence of involvement level, argument strength, and source on attitudes toward an organizational blog as a channel of cancer treatment communication. The analysis showed that argument strength had an effect on attitu de toward the blog, F (1,143) = 75.87, p < .01, eta = .59, a much larger than typical effect. Means showed that those receiving the strong arguments were significantly more likely to rate the blog favorably (M = 39.73) than did those receiving weak argumen ts (M = 27.27). In addition, involvement level had a marginally significant effect on attitude toward the blog, F (1,143) = 3.83, p = .05, eta = .16, a small effect. Means showed that high involvement subjects rated the blog less favorably (M = 32.10) than did low involvement subjects (M = 34.90). A significant two way interaction on attitude toward the blog was found for involvement level and source, F (1,143) = 6.37, p = .013, eta = .21. A simple main effect of source was found for those low in involveme nt, F (1, 74) = 7.74, p < .01, eta = .27, a medium effect. For low involvement subjects, the high credibility source created more positive attitudes toward the blog (M = 38.03) than did the low credibility source (M =
101 32.43). Among those who were highly in volved, the simple effect was insignificant (M = 30.83 for the doctor and M = 32.54 for the PR source). Empowerment RQ3 contemplated the influence of involvement level, argument strength, and source on personal feelings of empowerment. Empowerment was ass essed using two standardized measures that related to health information Health Information Orientation (HIO) and Perceiv ed Behavioral Control (PBC). In addition, a series of items operationalized to measure behavioral intention to seek additional inform ation about positron treatment from several sources were a lso used to measure empowerment Standardized m easures ANOVA showed no significant main or interaction effects with respect to the standardized HIO measure. In addition, there was minimal variance in the average responses to the seven items in this measure, with the grand mean M = 35.0 and the upper and lower bounds 33.8 and 36.2. On the overall four item PBC measure, ANOVA identified a single main effect of involvement, F (1,144) = 5.01, p=.027, et a = .18, a small to medium effect. The means showed that highly involved individuals perceived they had slightly more beh avioral control (measured with two items each of capability and controllability) (M = 24.37) than did those with low involvement (M = 2 2.97). No other significant main or interaction effects were found. Information seeking items Analyses were conducted of each of four empowerment questions relating to intention to seek additional information about positron treatment from four different so personal physicians, the Positron Treatment and Research Institute, and from the blog
102 author, i.e. the source of the information. In addition, an overall empowerment score was comp uted for each participant by summing the scores on these four measures. The ANOVA for the likelihood of seeking additional information about positron treatment from any source found a simple main effect of source, F (1,143) = 5.48, p = .021, eta = .19, a small to medium effect, with the means showing that those who were exposed to the high credibility source were more likely to say they would seek additional information (M = 13.15) than were those who received the low credibility source (M = 12.22). No ot her significant effects were found. On the likelihood of participants asking their personal physician about the treatment, a main effect of argument strength was found, F (1,143) = 5.09, p = .026, eta = .18. The means showed the strong arguments resulted in a greater likelihood of 11.52). No other significant effects were found. Two main effects were found with respect to the likelihood participants would contact the Posit ron Treatment and Research Institute about its treatment. A main effect of source was found, F (1,143) = 4.48, p = .036, eta = .17. The means showed that the high credibility doctor source resulted in a greater probability of contacting the organization (M = 9.82) than did the low credibility PR source (M = 8.44). A second main effect of argument strength also was found, F (1,143) = 4.35, p = .039, eta = .17. The means showed that those who received the strong arguments were more likely to say they would co ntact the organization (M = 9.81) than were those who got the weak arguments (M = 8.45). No other significant effects were found.
103 There were no significant effects for the measure asking the extent to which subjects would write comments or questions to the blog author. The grand mean for the measure was M = 7.58. Finally, on the summed empowerment measure, a main effect of source was found, F (1,143) = 5.11, p = .025, eta = .18. The means showed the doctor source resulted in greater information seeking emp owerment (M = 43.36) than did the PR source (M = 39.59). Other Intention and Behavioral Measures The final research question explored the effects that involvement level, argument strength, and source have on behavioral intention and behavior. The behavior al intentions measured likelihood of seek ing thyroid cancer screening and ceasing to use informational ma terial also was examined. There was a significant main effect of i nvolvement on the likelihood of obtaining screening for thyroid cancer, F (1,144) = 10.70, p = .001, eta = .26, a typical effect. The means showed that those receiving the high involvement cover story were more likely to say they would be screened for thyr oid cancer (M = 10.42) than did those receiving the low involvement release (M = 8.60). A significant main effect of argument strength on likelihood of giving up cell u se also was found, F (1,143) = 10.20, p = .002, eta = .26. The means showed that those w ho read the weak arguments were more likely to give up using their cell phones if diagnosed with thyroid cancer (M = 8.76) than were those who received the strong arguments (M = 6.71). On the behavior activity, 31.6% (n=48) of subjects overall kept the co ntact information sheet about the fictitious Positron Treatment and Research Institute identified in the experiment. To investigate whether high vs. low involvement, high vs.
104 low source credibility and strong vs. weak arguments had effects on whether or no t subjects took the contact sheet, a chi square statistic was used. The Pearson chi square results show that high vs. low source credibility had a significant difference on whether part the high credibility source are more likely than expected to take the contact form than those viewing the low credibility source. Phi, which indicates the strength of the associ ation between the two variables is .178, considered a small effect size (Morgan et al., 2004). Chi square analyses revealed there were no differences on taking the form based on either involvement level or argument s trength. Table 4 1. Participant d emogr aphic information Variable N % Age (years) M = 19.93 Range: 18 28 152 100 Gender Female 98 64.5 Male 53 34.9 Race White 93 61.2 Black or African America 13 8.6 Hispanic 29 19.1 Asian 7 4.6 Other 10 6.6 Year in school 1 or freshman 24 15.8 2 or sophomore 46 30.3 3 or junior 44 28.9 4 or senior 35 23.0 5 3 2.0 Major Advertising 30 19.7 All 40 others 122 <10.0 each Cell phone use for calling Times a day 151 M = 8.24 Range: 1 to 50
105 Hours a day 152 M = 2.28 Range: 0 to 24 Read any blog? No 16 10.5 Yes 136 89.5 Read health/medical blog? No 65 42.8 Yes 86 56.6 Family/friend w/cancer? No 36 23.7 Yes 116 76.3 Immediate family w/cancer No 110 72.4 Yes 29 27.6 *Among 139 w / identified f amily relationships
106 CHAPTER 5 DISCUSSION This research sought to examine empirically the applicability in a health context of a theory of persuasion that has been widely tested and found to be predictive in other consumer areas, particularly with respect to attitudes toward products, brands, and advertising messages. Health has traditionally not been identified as consumerist; how ever, the trend toward doing so began about a decade ago in response to fears that managed care was resulting in limits on freedom and control in health decisions by patients and health providers (Schneider & Hall, 2009). The turn to consumer driven health focuses on stronger consumer choice and putting decision making in the hands of individual patients, rather than by health plans, employers, and health care providers (Robinson & Ginsburg, 2009). Similarly, patient focused or consumer centered health care and communication have been frequent topics of debate and discussion, most recently with regard to health reform, with a considerable body of research suggesting such approaches improve key health outcomes (Epstein, 2000; Stewart, et al., 2000). Unprecede nted access to information, especially through the convenient, cheap, and limitless world of the Web and Internet, has fueled the trend toward mor e health consumerism. This ubiquity of information amplifies the needs by both health organizations and consum ers to be able to fight the significant message clutter and enable the former to more effectively reach their target audiences and the latter to obtain the accurate, timely, and understandable information they desire. As a result, this e xperiment examined the applicability of one approach which recognizes the magnitude of communications to which people are exposed and establishes factors that help predict the way they are likely to process the information
107 given they are unable to carefully scrutinize it al l. S pecifically, it investigates the Elaboration Likelihood Model of persuasion with respect to messages about a novel cancer treatment appearing on organizational blogs in order to determine how and when each of its two processing routes come into play. I n order to do so, it tests six hypotheses that would be predicted by past research investigating the ELM in othe r consumer contexts, and investigates four research questions exploring variables that previously have not been examined with in this framework. T hese research questions intended to consider the value of organizational blogs as a channel of health information, whether this two way venue provides to consumers the mutual benefit of empowerment with respect to information gathering/seeking, and the ef fect s that message and personal factors have on both behavioral intentions and action R esults of the var ious analyses show some support in the cancer treatment contex t investigated here for the hypotheses predicted by other ELM research ; however, the resu lt s are complex and not always as anticipated. H1 and H2 proposed that involvement and argument strength would interact to jointly affect attitude toward the message about the treatment; however, this was only partially borne out among lowly involved parti cipants and not to a statistically significant degree. Argument stre ngth did have a robust influence on attitude toward the treatment message among participants overall, but unexpected was that this influence was not dependent on of in volvement As predicted by the ELM, H3 accurately hypothesized that involvement level and source would interact to jointly affect attitude s toward the message about the treatment. H4 and H5 also found only partial support for the effect of argument strengt h on attitude toward the treatment, and only among lowly
108 involved subjects and not to a significant extent. H6 considered i nvolveme nt level and source credibility, which had no joint effect on ratings of favorability toward the cancer treatment as predicte d by the ELM This model suggests that highly involved or motivated subjects who also have the ability to process a communication (which was presumed here based on educational attainment and prior testing of the comprehensibility and understandability of t he messages) will cognitively elaborate o n a message and work to understand it, which is associated with central route processing ( Petty & Cacioppo, 1981). If these highly involved subjects perceive the arguments in the message to be strong and compelling, their scrutiny will cause them to think favorable thoughts that lead to enduring attitude change and/or persuasion. Numerous studies have shown these positive thoughts translate to favorable attitudes about brands, products, and messages, particularly adv ertising. Alternatively, those who are low in involvement/motivation will not elaborate on or analyze the message (or will consider them only slightly or to a much lesser degree) instead looking to other non issue relevant persuasive cues such as source c redibility. While this peripheral route to persuasion can result in attitude change, it is relatively temporary, typically lasting only as long as the cue (s) & Cacioppo, p. 263) Prior research has not investigated within the ELM f ramework attitudes toward the blog as a channel of health communication or toward a health organization providing a cancer treatment which were therefore posed for examination in this study through research questions. However, surmis ing from the prior exp lanation of the ELM and past research the presence of strong arguments could be expected to result among highly
109 involved subjects in elaboration and favorable attitudes toward the blog channel and the health organization, while among those low in motivati on a highly credible source would instead act as the predominant attitu dinal factor These ELM tenets did hold among lowly involved subjects in this study such that those who received the highly credible physician source held significantly more favorable attitudes toward both the organization and the blog than did those who viewed the PR representative. However, argument strength had a strong significant effect among all participants but not differentially among highly involved subjects as might have been predicted. To some extent, involvement (which here is a situational recipient factor ) and the message attributes of source credibility and argument strength each did influence some of the dependent attitude measures tested here. Among all subjects, source credibility influenced three of the four attitudinal measures (toward the message, positron treatment, and the organization), as well as overall likelihood of seeking additional information about the cancer treatment and of actually taking a specific info rmation seeking action. Argument strength had an even stronger influence, affecting all four attitudinal measures (including toward the blog), as well as the probability of seeking additional information from some sources and of discontinuing a harmful beh avior, which here was using a cell phone after being diagnosed with thyroid cancer Involvement p. 667 ) and in determining the routes individu als use to process information (Petty & Cacioppo, 1981, 1986) For the most part here involvement in the issue presented here which measures personal importance or
110 r elevance and motivati on to process the information with respect to marketing/public relations messages about a novel cancer treatment. This result seems to reinforce findings in a preventive health context related t o AIDS campaign messages that involvement may not be as sign ificant a predictor of attitudes in health contexts as it has been found to be in other consumer areas (Igartua, Cheng, & Lopes, 2003). Among all participants, involvement level s had significant effects on the empowerment measure of Perceived Behavioral C o ntrol and on the behavioral intention to obtain cancer screening, as well as a marginal influence on attitude toward the blog channel As would be predicted by the ELM, among low motivated individuals, involvement also jointly influenced with source credib message, the organization, and the blog. However, no similar combined expected influence of argument strength was found among those who were highly motivated. This may relate at least in part to difficulties explica ting t he concept of involvement and of manipulating and measuring it. As discussed previously, involvement has received a significant amount of investigation, particularly as it relates to consumers and products, and to theories of persuasion such as the ELM In this study, involvement was operationally defined to reflect personal importance and meaning; however, various studies have operationalized the concept differently (Braverman, 2008). Thus, d espite this host of research, involvement is a complex concept th at is still not well understood in many cont exts, including in health and medicine The notion of health is similarly multifaceted, complicating the issue further. This suggests that a great deal of investigation empirical and otherwise is required to unr avel the threads that intertwine the complicated pattern these concepts jointly create.
111 This study attempted to move existing literature forward in that endeavor by exploring personal feelings of involvement in a potentially life threatening health issue. Statistically significant differences in mean involvement level were obtained through the final cover story manipulations. However, even the participants who received the low involvement condition (which stated the probability of getting thyroid cancer fro m cell phones was less likely than getting struck by lightning), had a relatively high mean level of involvement. This was unexpected as research shows that people, adolescents in particular, often fail to feel vulnerable to negative health outcomes and un derestimate the magnitude of the risks to which they are exposed (e.g. see Johnson, McCaul, & Klein, 2002; Weinstein, 1984). Informal discussions with participants during pretesting suggest this unanticipated high involvement level may have been at least p artially related to the saturation of mobile phones in society (100% in this study and at least 83% of adults nationwide [Rainie, 2010]) and to the absolute reliance on these devices, particularly by those in this college age cohort and/or this age group o verall (Rainie & Keeter, 2006; Rausch & Edwards, 2008). Comments written on the qualitative thought listing portion of the experimental questionnaire revealed this also to be the case among at least some of these subjects. Another likely factor relates to the cover story identifying that cancerous tumors were found in college students, not in middle age or older people where they are more expected. Health promotion programs and materials commonly suggest ways to avoid the likelihood of cancer by decreasing risk and/or adopting healthy behaviors, e.g. quitting smoking to decrease lung cancer risk. Here however, the cancer risk is relatively immediate, occurring in young adulthood, while the onset of many other
112 cancers, including lung cancer associated with sm oking, typically is much more gradual and does not manifest itself until late middle or older age. Research has shown that feelings of risk susceptibility and vulnerability do predict health behavior (e.g. Weinstein, 1984); however, young adults may not p erceive these slow, ongoing risks as applicable or pertinent so early in their lives, which may help explain why involvement in th is Also potentially associated with the une xpectedly high involvement rates across all participants was that despite the relatively young average age of the sample, more than three quarters indicated they had experienced a close encounter with cancer through a family member or friend. In addition, four participants (of 153) had been personally diagnosed with cancer at some po int in their lives. Although this absolute number is small, this amounts to 2.6% of the sample, substantially above the national mean prevalence for this age group. National Can cer Institute data show about 11.4 million people alive in 2006 had been diagnosed with some form of cancer sometime in their lives (NCI SEER, n.d.). The estimated prevalence among young adults i s low, ra nging from 0.16% among those 10 to 19 year old s to 0.27% among 20 to 29 year olds. Thus, the prevalence among the pool of subjects in the experiment under discussion was at least 10 times higher than the national average. The reasons for this are a mystery, particularly since cancer incidence in Florida, which is the home state for about 82% of the students at the university where the experiment was conducted is lower than nationally (NCI s tate, n.d). Further complicating the concept of involvement in this study was that although analyses of the manipulat ion check of involvement in the cover story revealed a
113 statistically significant difference, much of that difference may have been eroded after participants read the blog containing the positron treatment message. The thought listing activity for total num ber of thoughts unexpectedly showed that involvement level had no significant individual or joint influences with source credibility and argument strength on the total number of thoughts generated as would have been predicted by the ELM Involvement level did play a significant role in influencing the number of positive thoughts; however, lowly involved participants unexpectedly generated more favorable thoughts than their highly involved counterparts. In addition, the joint differences with source credibil ity and argument strength that would have been predicted by the ELM did not occur in th e context investigated in is study. A three item rating assessing tion to the message on the blog likewise showed that that level of involvement in the issue of getting thyroid from cell phones also was unrelated to involvement in the treatment message i tself. Only argument strength regardless of involvement level or source affected how much subjects perceived they attende d to the treatment message. Argument strength will be discussed further subsequently; however, it is notable that only the message factor of argument strength showed significant influences on the number of qualitative positive and negative thoughts generat ed overall, such that those receiving the strong arguments generated more positive and fewer negative thoughts than did those receiving the weak arguments. The reasons for these unanticipated results related to the blog message can only be surmised, but gi ven the overall high levels of involvement may relate to disappointment or displeasure with the content of the information about the cancer
114 treatment Evidence to support this supposition stems from relatively moderate mean ratings on measures gauging exte nt of perceived learning from the blog and from attitudes toward the treatment message (M = 4.42 of 7 and M = 13.12 of 21 respectively even among those receiving the strong arguments and resulting in the highest ratings). These mean s were considerably lowe r than those obtained in pretesting, where t he five strong arguments used in the blog received considerably high er ratings of strength and p ersuasibility (M = 5.81 ) T his could have been due to the within groups testing in which subjects rate d both strong and weak arguments together which may have made the strong seem to be more powerful i n comparison to the we ak ones In addition, moderate assessments of believability (one of seven items used to assess treatment message attitudes) also may have contribute d to possible dissatisfaction with the messages (GM = 4.43). Thus, f uture studies should take these factors into consideration. A high level of i nvolvement on its own also was associated with significantly higher perceived behavioral control ; however, pote these items (described previously) make it difficult to discuss any real significance in these findings. Involvement alone also played a role in influencing intention to participate in a cancer screening. Al though not specifically predicted by the ELM, this finding is expected given that behavior research and theories show that personal relevance and motivation often are linked to and/or predictive of a variety of positive health be havio r s. Involvement also had a marginal effect (p=.05) on attitude toward the blog; however, it was opposite what might have been presume d based on past persuasion research Subjects who were highly involved rated the blog significantly less favorably
115 than did those of low involve ment. This finding is particularly surprising in light of past research which has shown that individuals who are highly involved in their health tend to engage in seeking health information from active oriented channels (Dutta Bergman, 2004). Such channels include print sources of information such as newspapers and magazines, as well as the Internet. Prior studies have not specifically examined attitude s toward a blog with respect to these recipient and message factors and so cannot provide any insight. How ever, it is possible that highly motivated individuals in this study may have perceived the blog to be a less appropriate, less active, and/or less biased venue for obtaining important cancer information. This may stem from the fact that blogs are much mor e informal than many other forms of Internet communication and traditionally have been written by individuals sharing specific point s of view and strong opinions. The moderate overall believability of the blog may lend credence to this notion. The findings might also have to do with the arguments on the blog, as even the true and strongest ones were rated as being only moderately persuasive, which will be discussed in greater detail subsequently. These results might also point to the fact that involvement i n a n issue, defined by Chaiken (1980 ) as personal importance or relevance about a topic and measured here and elsewhere as such, may not adequately embody the concept of motiv ation with respect to serious and perhaps other health issues. This is at least p art ially supported in this study by the fact that analyses of the initial standardized four item reliability in conjunction with the other items (importance, meaning and ne ed), thus the relevance item was removed from the measure for the actual experimental data
116 collection. Involvement in the issue of getting thyroid cancer from cell phones as s (1989 ) concept of outcome relevant involvement, which relates to people becoming involved as a result of their ability to obtain desirable outcomes which here likely related to successfully treating thyroid cancer caused by cell phones. As they predict with thi s type of involvement, high involvement subjects were more persuaded than low involvement subjects by strong arguments about the cancer treatment and less persuaded by weak arguments. However, the desire for favorable outcomes even in life threatening situ ations may not provide enough motivation to change habitual behaviors in the short term, as evidenced by the relatively small percentage who said they would give up using a cell phone even if they were diagnosed with cance r specifically caused by it. This research confirms that involvement is complex. This may be especially true with respect to health issues, as was the case here. Braverman (2008) suggests that in factor ( p. 668), which she tested by measurin g rather than manipulating involvement a s was done in this experiment inherent feelings of personal involvement and need for cognition on three topics. As a result, it is difficult to make comparisons between the studies. However, together they raise the possibility that involvement in health issues may rely upon or relate to, at least partially, more innate personal characteristics, or beliefs or feelings that arise from individual experience. For example, Kid influenced their being actively involved in caring for themselves, such that those with high perceived control felt active involvement in their own care was more positive and
117 n ecessary than did tho se having low perceived control. The proposal that health involvement is a personal trait also relates more closely to the concept put forth by Johnson & Eagly (1989) of value relevant involvement, which is involvement activated by end uring values. D ata collected in this study on health information orientation (HIO) and perceived behavioral control (PBC) that were used as standardized dependent measures of empowerment also may be (and have been in other studies ) employed as independent variables gauging natural personal traits that may act as stand ins for the manipulated variable of involvement. Here, when re analyzed as independent variables in place of involvement, the ANOVAs showed that neither HIO nor PBC had any significant main or interaction effects on any of the four dependent attitudinal measures However, this may relate to othe r factors, including when the ratings on these measures were collected in th e course of the experiment, possible misreading of some PBC items (described previously) and relat ively small differences in mean HIO scores among all participants. G reat care was taken to identify through the literature an involvement measure that had high internal reliability in a number of studies and best operationalized the concept of issue involvement as conceptualized in this study. However, it is possible that did not end up to be the case, which may be a limitation of this study. Future research should investigate experimentally other variables that can be used as a stan d in for motivation, including those that have already been tested, such as need for cognition, as well as those that have not. T his experiment reiterates past findings that more research is needed to better understand the concept of involvement and how t o best measure it particular ly with
118 respect to health and cancer communication more broadly. This could include examining these concepts to determine whether these finding s are borne out with llomavirus (H PV); in p opulations of varying ages; in relation to both short and long term health risks ; and with respect to involvement in other health contexts where it is situationally manipulated as well as measured as an innate personal trait. Source Credibility I n the context examined here, feelings of involvement did not always exhibit a joint influence with source credibility; however, this message cue did emerge as playing a role in several of the attitudes among those for whom the cancer issue was of lower imp ortance Specifically, attitudes toward the treatment message, the organization, and perceptions about the trustworthiness and expertise of the source providing the messa ge, with perceptions of high source credibility resulting in more favorable attitudes. These source characteristics had no influence on attitudes among those who were highly involved in the issue. As predicted by research in other contexts, this set of fin dings indicates that processing of cancer treatment information also followed the tenets of the ELM. However, these results conflict with those conducted in another health context. McCullough and Dodge (2002) found no differences in attitudes toward an adv ertised hospital among low involvement subjects based on the spokesperson used, while a non physician spokesperson was more influential than a doctor among highly involved subjects. As a result, the researchers suggested that hospital advertising need not be reliant on using physicians, who often have negative attitudes about advertisi ng, as sources of information.
119 In addition to the persuasiveness of source credibility in conjunction with recipient involvement, the type of source also influenced alone part three of four measures, as well as with respect to health information seeking empowerment, and a related behavior. Specifically, the physician source was more persuasive among subjects overall, such that those who received th e message from this source reacted more positively to the message, the treatment, and the organization, which falls in line with what would be expected based on prior source credibility research. Those who viewed the physician also felt more empowered to s eek additional health information, which might be expected, although empirical evidence could not be fou nd confirming this. The implications of this finding are especially relevant in public relations, which seeks to establish and maintain mutually benefic ial relationship s with its various constituencies (Grunig, 1993; Hon & Grunig, 1999; Ledingham, 2003), as well as with respect to most nontraditional media channels, where the lines between marketing and PR have blurred and organizations are increasingly communicating directly with consumers (Scott, 2009). Also notable for health organizations is that subjects in this study who received the treatment message from the high credibility physician source were more likely to physically take a document citing c ontact information about the organization providing the treatment tha n were those who viewed the low credibility PR source This behavior of taking (or not) this contact information which would be desired by an organization provides a more explicit way t the treatment and/or the organization and its staff and facilities than are indications gleaned from behavioral intention measures such as the ones measured in this study.
120 Qualitative comments in th is study as well as broader research on source credibility suggest those who received the doctor spokesperson may have felt less anxious about either getting thyroid cancer from cell phones and/or more comfortable about the treatment and/or the organizatio n, while the less credible PR spokesperson might have caused more anxiety about these things, which may have resulted in the latter being less apt to desire the information a nd/or take the contact document at least in the short term. Much past research on persuasion and source credibility would suggest the influence of a high credibility source would lead to a greater propensity for taking the document. However, a behavioral action of this kind has not been described in prior persuasion research, so i t is worth noting that these findings do not fall in line with the previously mentioned ELM based experiment on hospital advertising McCulloug h and Dodge (2002) found that highly involved subjects rated the hospital significantly more favorably when the y recei ved the ads featuring a non physician spokesperson (specifically a woman identified as Mrs. Wells pictured holding a baby ) than from a doctor (identifi ed as Dr. Wells), prompting the authors to speculate that subjects may have perceived the non physician t o be more candid. However, that was likely not the case here, as ratings of trustworthiness alone (without expertise) showed that overall participants felt the doctor source was significantly more trustworthy than the PR spokesperson who functioned as the non doctor source. Although this does not provide direct insight in to the findings of this study the hospital advertising experiment raise s questions about the use of general public or non professional sources in persuasive health communications. Testimo nials from such
121 sources have been found to be more persuasive among lowly involved consumers, with conjecture that the reason is because a non professional person is easy to identify with and that such messages are likely to appeal more to emotions than to rational thinking (Braverman, 2008). As a result, these elements function as peripheral cues for those who elaborate little on the information. Future research should investigate experimentally the use of non professional spokespeople, such as a patient w ho has been diagnosed with the disease under investigation in conjunction with professional sources in order to better identify the health contexts and channels in which each source would be most effectively utilized. For example, Braverman found that aud io testimonials were more persuasive than those in written format. Although blogs are typically a much less formal channel of communication than are other traditional written forms of media (including Web sites) and may seem to lend themselves to the testi be the case, at least on written blogs and where there is a desire to promote elaboration on messages a nd central processing. The current research investigated professional sources and their credibility as the peripheral cue; however, there is some evidence to suggest that other cognitive shortcuts, e.g. likeability and/or attractiveness, may have also had some influence. As described previously, both sources were identical with the exception of their titles and clothing in order to ensure that only the identified cue was involved and to minimize the influence of other attributes. S uch factors should have been equally distributed across both sources and in all eight condition s due to the randomization used in this study; however, some participants did address in their qualitative thought listing commen ts
122 that the source looked mean, stern, or unhappy, and some suggested she should have been smiling. Surprising was that such co mments were not articulated during source pretesting discussio ns among 25 participants. The differences between the photos used in pretesting and those used in the actual materials were seemingly minor such that the pretesting materials included larger, c olor photos only without text (the only text was the title and description th at accompanied it) versus smaller, black and white photos on the blog, which were accompanied by the message text and other content previously described In addition, participant s involved in pretesting did not receive the cover story perceptions and/or expectations. It is also important to mention th at additional characteristics, such as gender and/or race, their feelings about their expertise. This was not found to be the case in this study i.e. there were no signif icant differences in ratings of the Caucasian expertise based on gender or race of the participants H owever, this may differ depending on the context and the population suggesting more direct comparisons of such source characteristics are needed in organization al health communications. As predicted by the ELM and other persuasion research, source credibility was shown to be highly influential both alone and in conjunction with involvement. How ever, the results here conflicted with other si milar research in the context of health. What is clear is that getting a better handle on these source issues is crucial, particularly with respect to which traits or elements might function as cues, in which situations, with
123 respect to what sources and i n what channels Therefore, additional empirical re search is needed in this area. Argument Strength Argument strength surfaced as the most influential independent variable in the serious health context analyzed here, irrespective of involvement or relevanc e or th e credibility of the source Among all participants, s trong arguments shaped attitudes toward the message, the positron treatment, the or ganization and the blog, as well as or and from the treatment organization, and intention to give up using a cell phone. On all four of the attitude measures these effects were significantly larger than typical However, the ANOVAs showed none of the interaction effects with involvement le vel that would have be en predicted or might have been extended from past ELM research. This model predicts that those who are highly involved in the issue will be motivated to scrutinize the message, and so will be more influenced by strong arguments than will those who do n ot elaborate on the information. The ELM also would predict those receiving the low involvement scenario would be relatively oblivious to and unconcerned with the arguments, instead basing their attitudes on easy shortcut s including s ou rce factors of the type already discussed and other cues such as the total number of arguments (the idea that more arguments is perceived as better) However, this was clearly not the case here. Some insight into this occurrence can be gained through the comments on the qualitative thought listing activities. A number of subjects who received the low involvement cover story and the weak arguments on the blog commented about the lack of subs tantial information about the treatment itself, indicating it inst ead contained
124 frivolous messages about the location and aesthetics of the treatment facility. Not surprising was that such comments were more common among those who received the high involvement/weak argument condition. Anecdotally, many of these comments across those in both involvement groups expressed considerable anger and frustration that this was the case. Strong arguments also led to much higher levels of believab ility with respect to the blog. T ogether, these quantitative and qualitative findings se em to indicate that most individuals are to some extent motivated to examine arguments relating to a serious and relative ly immediate health threat of the kind described here. Further, this may provide ev idence that those who are typically unmotivated to t hink deeply about a message may be pushed to do so in these kinds of situations Past ELM research has similarly found that a highly credible source can lead to such unexpected inducements among highly motivated subjects in certain scenarios (Heesacker, Pe tty, & Ca cioppo 1983). These inferences seem reasonable in light of the results here ; however, other explanations also merit discussion. First, the fact that this was such an extremely serious and relevant health issue among participants overall as descri bed previously may have triggered perceptions that the weak arguments seemed thoughtless and flippant in comparison even among those of low involvement who typically would pay little attention to the actual content of the messages. Qualitative comments to this effect were reported, and although this would typically not be the case under the ELM among lowly involved or motivated individuals there seems to be a distinct difference with respect to messages dealing with serious health issues such as cancer i ndicating a critical need for f urther research to better understand these results
125 Such sensitivities to argument quality of the kind that surfaced here could potentially have significant adverse effects on health organizations and also could be detriment al to health consumers. This is of particular concern in light of ELM research showing that weak arguments provided to highly motivated individuals could cause them to mentally rehearse unfavorable thoughts, resulting in their resisting, rejecting, or even opposing the message or point of view being advocated, creating a boomerang effect that persists and will be difficult to be change d by other efforts (Petty & Cac ioppo, 1981). This study suggests that to some extent this also might be the case among some lowly involved individuals with respect to cancer information and possibly other health communications. The intention to give up using a cell phone if diagnosed with the thyroid cancer also was highly related to argument strength; however, in the opposite direction than might have been expected. Those who read the weak arguments were significantly more likely to say they would give up using their cell phones if diagnosed with thyroid cancer than were those who received the strong arguments, run ning counter to prior research. T he reason for t his could be because the strong arguments, which contained information about the curative properties of the treatment, provided reassurance that getting thyroid cancer was n ot a death sentence and could be treated with f ew side effects. This reassurance was verified anecdotally in some of the comments cited in the thought listing exercise. On the other hand, the weak arguments likely provided no such reassurance, as they related to the facility and other issues, such as t he quality of the area and center, and companionship with other patients
126 Also of note is that a lthough argument strength had much larger than typical effect s on all four attitude measures and on perceptions of learning overall the arguments were rated as only mo derately strong, persuasive and believable. This was the case although they were true and had been rated during pretesting to be much stronger on average This is likely due at least in part to a desire for the inclusion of cure rates associated w ith the treatment and other statistics, as identified in some of the qualitative comments. This suggests that, despite overarching misunderstanding of statistical data and the easy abil ity to distort such facts, consumers expect such information to be pres ented with respect to health matters, particularly major co nditions like cancer. S uch statistics are difficult to quantify in broad based health communications because of the vastly differing nature, types, and staging of cancer and disparate patient profi les ; however finding ways to incorporate such information seems to be important to cancer information consumers Doing so might enhance favorable attitudes about treatments and the organization providing them. It also might improve perceptions of believ ab ility, as well as feelings of empowerment among health consumers particularly among central processors who elaborate on such information This lends further support for the presumption that such perceptions are highly subjective and as such, health messag es require significant formative research in order to adequately address these issues. Gender Differences The majority of the socio demographic variables, including gender, race, age, daily cell phone use, and experience with cancer had no distinct influe nce on any of the attitudes toward the treatment, such that males viewed the treatment significantly more
127 favorably than did females. This difference may have been the result of chance alone; however, other research may shed light on this unexpected finding For example, although women tend to be greater seekers of health information in general than are men (NCI, 2005), studies have found that health information needs vary by gen der. Service were less likely than males to inquire about specific treatment information, with their information needs instead relating to more general cancer information as we ll as to screening and social support. Seale, Ziebland, & Charteris Black (2006) found similar gender differences concerns focusing on treatment information medical personnel and procedures, wh ile women were more concerned with broader issues, including the impacts of illness, and emotional support. Seale (2005) also found gender differences with respect to the kinds of things men and women take into consideration whe n making decisions about can cer treatment, finding evidence that popular cancer Web sites on the Internet reinforce these distinctions. These findings suggest that women may be less involved in obtaining information about treatments for cancer, which in turn could diminish their atti tudes about a specific cancer treatment, as was the case here with positron treatment. However, other studies have found that treatment information is among the most commonly sought by cancer patients regardless of gender or other socio demographic factors ( e.g. Metz et al. 2003). Therefore, gender difference s deserve great er exploration with respect to cancer and other disease related treatment information and warrant consideration by those involved in developing health information
128 Implications F or T heory This study su ggest information based on involvement level via the central and peripheral routes do have applicability with respect to a persuasive communication about a serious or pot entially life threatening health issue. First, it provides direct evidence in a health context supporting past research in other areas that peripheral cues such as source credibility have a significant effect on attitudes toward a persuasive message among people who are less involved in the issue under discussion. In addition, it extends past ELM research by showing support for the fact that such cues also can positively influence attitudes that people with low involvement levels have toward a health organi zation and to an organizational blog as a c hannel of health communication. This study also provides indirect support for ELM predicted central processing among those who are highly involved in a serious health issue. As is the norm in this type of experime ntal research, the analyses of variance on the dependent variables were computed using the manipulated involvement level (high and low) and significant differences in mean involvement levels between these two groups were verified based on the manipulation check measure. However, it can be argued from a theoretical standpoint that overall the participants in this study were all highly involved, having mean involvement levels on the high end of the measurement scale As a result, the data were re analyzed us ing a median split of reported level of involvement in the issue (which often functions as the break point for high and low involvement level in empirical studies of the ELM ) based on the three item involvement manipulation measure (importan ce, meaning and necessity with median = 17.5). These
129 results showed that argument strength and involvement did interact to predict attitude toward the treatment message and toward the treatment itself. The means of the simple effects test among only highly involved subjects (based on the median split) showed they reacted much more favorably toward both when they received the strong arguments than when they received the weak as would be predicted by the ELM. Because the involvement scores were so high overa ll and argument strength surfaced as such a robust fa ctor among all participants, this joint effect of involvement and argument strength also was present to a lesser extent among the lowly involved. Regardless, these results provide evidence that the dual processing ELM performs as expected and indicate s it has the ability to effectively predict the factors that lead to persuasion in this health context; therefore more research is warra nted to explore the use of the M odel with respect to cancer messages an d those about other health issues. As suggested by Slater (2006) and others, examining the ELM in conjunction with other theories that focus more heavily on health behaviors would be especially beneficial with respect to health issues. For Practice The re sults of this ex periment provide several lessons that might help health message developers increase the effectiveness of the communication s they have with their constituencies. First, this study supports with empirical evidence the theory that people proce ss health information differently and elaborate on messages to varying degrees. In doing so, it verified that the strength or quality of the information being presented and the source used to do so can help better target health messages to the specific way s people process this type of information depe nding on how personally important the information is perceived to be or how motivated a person is to process the
130 information. This supports n umerous other studies showing that targeted health i nformation is mor e persuasive (sbring, & Nrvnen, 2004; Briol & Petty, 2006; Kotler, Roberto, & Lee, 2002 ) I t further suggest s these three linked factors as well as the ELM more broadly can help guide the targ eting and development of health messages, particularly in ways that will help promote elaboration on them and central processing Doing so will better ensure recipients of these messages counter with fewer arguments against them and that the position advoc ated will last for a longer time, i.e. will be more likely to result enduring persuasion (Petty & Cacioppo 1981, 1986). In addition, r egardless of how involved one is in a health issue, it appears from this study that there may be a decline in the feelin gs of personal importance or motivation once people are actually exposed to a persuasive message Here, subjects appeared to pay le ss attention to and/ or concentrate d to a lesser degree on the actual treatment message than might have been anticipated based on the broader involvement in the issue that preceded their coming in contact with the info rmation. This tendency may be more pronounced among members of the general public who typically do not analyze a single message as was done in this experiment but instead may be seeking, digesting, analyzing, and processing multiple resources during their information gathering efforts or may instead be coming in contact with information serendipitously. It may be particularly true among those unlike the college stu dent participant s in this study, who do not have experience handling many message s about a subject including elderly and disadvantaged populations This phenomenon also may be manifested to an even greater degree with respect to health messages. Research shows searches for health
131 information often are conducted by family members or friends of those afflic ted with a disease or disorder specific messages despite high personal relevance in the he alth issue more broadly. Likewise, patients diagnosed with a serious or life threatening disease such as cancer may be less motivated to process particular messages, especially in the early stages of the disease when they are more apt to ga ther a plethora of information (Squires et al., 2005 ) In conjunction with this apparent decline in motivation to process information, the ability to do so also plays a significant role and must be taken into consideration, espec ially in the early stag es after a diagnosis when health consumers may be in shock or disbelief It is important for health message developers to recognize t h is potential attenuation in motivation to process health messages, and to note that it may be mitigated by the strength of the arguments used as was the case here. This indicates that creating strong or high quality messages with respect to health issues is imperative Noteworthy is that strength and quality (which might include beli perceptions rather than any inherent factors in the arguments themselves, such as truth. One way health organizations might increase the quality, strength, and believability of health information is by utilizi ng where possible understandable, well explained quantitative data and statistics, especially related to prognostic indicators and treatment cure rates. These findings confirm the crucial need for testing of health information before it is broadly utilized in order to ensure recipients perceive that these messages contain strong high quality arguments, which are also c omprehensible and believable.
132 The source factor of credibility a lso was highly influential alone as well as in s involvement in the health issue indicating that source of information is extremely important when dealing with a potentially life threatening medical issue like cancer treatment. As a result, organizations producing cancer communications should ensure m essages and information are presented by a credible spokesperson regardless of whether recipients are highly involve d or not Although this may seem to contradict the well documented finding that targeted messages are more effective, according to the ELM, those who are very motivated to elaborate on a message are relatively unaffected by the source of the inf ormation (as was verified here). In addition, sources perceived as highly credible by those who normally would not closely examine or think deeply abou t information have been shown to increase scrutiny of persuasive messages (Heesacker, Petty, & Cacioppo 1983). T hus employing a highly credible spokesperson as perceived by receivers who are unmotivated to process a message, will help ensure the highest level of persuasiveness among the greatest number of recipients The channel used to disseminate the communication can play a role in enhancing reason to believe lts that organizational blogs can be a valuable tactic for delivering targeted health and ca ncer information to audiences. This is the case provided the messages contain what recipients be lieve to highly credible sources and strong arguments as noted abov e The vast majority of participants nearly nine of 10 had read blogs, providing additional evidence this is a communication channel people of this age group ar e familiar and comfortable
133 reported considerably heavier reading of general blogs than the 40% to 49% found in the 18 29 age group nationally (Jones & Fox, 2009). This difference could be due, at least in part, to the growing inclusion of blogs in university courses, which several respondents alluded to in their questio nnaires. This generation also has been found to be the most likely to say the advent of blogs has been a positive change in the media landscape (Pew Research, 2009), indicating they may be more apt to make blogs a ro utine part of their media menu and may expect to be able to access through them the specific information, resources, and answers they are seeking. In addition, more than half of respondents stated they had read a blog specifically about health. Blogs usually do not spontaneously emerge as a res ult of the searches most typically done for health information through Google and other general search sites (Eysenbach & Khler, 2002; Fox & Jones 2009), indicating people in this group are specifically seeking such information from this channel. The res blog use also is higher than that of adults nationwide, with a national survey finding that health or medical issues on a blog specifically (Rainie, 2009). Despite their relatively young age and low risk of health issues, the 18 to 29 year olds in Generation Y are among the most likely of any age group to seek health information via the Interne t (72%), followed closely by 71% of 30 49 year old group ( Fox & Jones ). No data was gathered in this study about the reasons those who accessed health blogs did so or the kinds of informati on they were seeking, but these results indicate the channel could provide a worthwhile means for organizations to connect with those looking for health information.
134 The use of social networking sites for providing and gathering health information is curr ently not widespread with just over one thir d of adults having accessed these channels ; however, data show these sites are growing in popularity P rovided blogs contain credible sources and strong arguments this channel has the added benefit of empowering health consumers to want to obtain more information from a variety of sources, including their own health care providers and the organization sponsoring the blog. This lends support for the notion that this venue has the capability to engender mutual bene ficial health communications and relationship building. Relationships that are perceived as beneficial to individuals result in their enhanced satisfaction, cooperation, support and loyalty; decreased resistance, conflict, and complaints; and increased lik elihood of affecting behavior changes, and ultimately of achieving an uch additional research is warranted to further ascertai n the potential of and classify blogs and other social netwo rking platforms as health communication tools. Three additional concepts may be applicable to future investigations of organizational e health blog messages and so will also be briefly discussed. They are uses and gratifications; interactivity; and media involvement, including parasocial interactio n and parasocial relationships. Although not identified initially by its current term, uses and gratifications was among the earliest media study approaches to view audiences as active and gratification seeking r ather than as a passive group on which media had direct effects (Baran & Davis, 2003). Uses and gratifications, an approach used often in research related to the Internet and World Wide Web and employed by Kaye (2005) in investigating motivations of genera l weblog users, is defined as
135 p. 256). It presumes that users have certain motivations for accessing media, and that they are aware of these and actively seek out the types of media they need to satisfy these needs (Kaye ). Thus, it follows that readers of cancer blogs, including those who seek information about cancer treatments and/or the organizations and health professionals who provide them, a re likely to be deliberately pursuing these sites because the y fulfill some need or desire. The interactivity possible on blogs, where both senders and receivers of (McMillan, 19 99, p. 377) affords the two way communication channel that is so valuable in health communication. Numerous formal definitions have been advanced of interactivity, as have criteria for identifying it, however, Jensen (1998) measure of a med involvement, which is seen as particularly beneficial in promoting health, where informational intervent ions aimed at behavior change have generally had poor outcomes (Neuhauser & Kreps, 2003). By their very nature, blogs promote conversation or dialogue, and thus are likely to be highly interactive (McMillan). Similarly, Doctorow et who comment on blog posts have the ability to control, at least to some degree, subsequent commentary by the primary author as well as the overall messages other readers of the blog take fr om it.
136 Another construct that relies on media users as active participants is media involvement (as distinguished from the more general concept of involvement described Bocarnea, 2003, p. 45), whether they be characters being portrayed, or stars or celebrities themselves. Most of the research in this area has involved the repeated expos ure viewers have to characters and stars through television, and has spawned the two related notions of parasocial relationships o a media personality, for example as friends (Brown et al., p. 46). The latter concept in particular has been since demonstrated with a variety of types of celebrities and media, ough scholarly research in these areas has focused on involvement or identification with celebrities, these concepts may be likewise be relevant in relationships between readers of blogs and their creators /sources/authors for general blogs as well as those geared toward cancer. Anecdotal media accounts suggest that many users tend to read the same blogs repeatedly, and this was verified in a study by comScore (2005), and therefore thes e readers may become knowledgeable about the person who is writi ng the bl og and involved in their online live s. This is also evidenced by some comments posted to blog entries, particularly those providing advice, encouragement, reassurance, or support. Limitations There are a number of limitations to this experiment and the re cognition of these should help refine future research efforts Some of these have been discussed
137 previousl y including the relatively high overall involvement levels, the fact that even the strongest arguments were perceived as only moderately persuasive a nd believable despite their truth and accuracy, and that the photographs of the source made her seem unsympathetic and unapproachable to some participants which is particularly problematic on a two way communication channel such as a blog Other limitati ons include that the subjects received black and white hard paper copies of the blog s rather than viewing them in the online, color format in which they would typically be viewed U sing the easily recogniz able Google blogging tool Blogger to create the blo gs provide d a prompt or cue to participants that the fo rmat was a blog rather than a static World Wide Web page However, doing so also may have affected the credibility of the manip ulations if they were viewed as less legitimate or believable as a result. Longitudinal testing to determine how enduring were the persuasive effects also was not possible b ecause of the nature of the involvement manipulation and the ease of further investigation by participants of the connection between cell phones and thyroid cancer after they left the experimental setting. This is an important component in testing the ELM to determine whether the predicted benefit of central processing by those who elaborated on the messages to a greater extent did in fact result in attitudes and behavioral intentions that were m ore likely to remain over time. In addition, while an actual behavior measure was included, it is diffic ult to make any predictions about what taking or not the contact sheet for the treatment organization actually mig ht have meant with respect to behavior. Future research should devise alternative ways to assess behavior that requires actually performing some specific information seeking
138 action, such as texting to a number that purportedly will result in an organizatio n automatically sending additional information Wherever possible, significant steps were taken control the potential confounding variables in this study, and randomization of the manipulations and participants, and simultaneous inclusion of all conditions in each setting should have minimized any differences between the various groups. Despite these attempts, it is not possible to control all potentially confounding fact ors in an experimental setting. Finally, this study also utilized a convenience sample of university students, and although the manipulations were such that this group became a stakeholder public in the information, this precludes generalizations beyond the population here. However, this research provides key insights into the use of the ELM in a health context and provides direction for additional research in this area. Future R esearch Throughout the discussion of this research, suggestions have been made concerning potent ial areas for future research. These included those related to involve ment, source credibility, and argument strength, with respect to the dependent variables investigate d here as well as in others of interest. Testing in other populations, particularly those who are at various stages of cancer, and settings, and with respec t to other nontraditional media channels such as Facebook an d Twitter would be beneficial. Another area that should be further examined is the role that organizational blogs may be able to play in health communication. This study extended past pers uasion r esearch by looking at the mutual beneficiality of health messages not only to a sponsoring organization but also to those who read them, an area that in the past may have been considered to be in conflict with the goals of persuasion studies This benefit
139 to audiences has been foremost in public relations communications, but is becoming increasingly important in marketing and to organizations more broadly, especially with the advent of new technologies and platforms that provide users complete control over when, where, how and even whether they will receive the information organizations wish to provide. This research looked at one indicator of the quality of an organization public ability to empower readers to want to seek additional information and gain knowledge about the cancer treatment. Further research exploring this concept should be undertaken, as well as into other indicators of exchange relationships. These are trust, sati sfaction and commitment, or the degree to which parties believe that the relationship is worthwhile to continue. Additional empirical research testing constructs of the ELM and other successful models of persuasion in combination with behavior change theor ies also would be advantageous, for example the Health Belief Model elements of severity, susceptibility, barriers and benefits. Conclusion This research provides empirical support for speculation that theories and models of persuasion such as the ELM th at have often been overlooked in past research involving health behaviors provide an additional avenue for targeting messages that can be beneficial for both organizations and consumers. This is good news in light of research suggesting that many health ca mpaigns based solely on theories of health behavior change have been inadequate (Slater, 2006) The ELM, which can identify how messages are processed and under what conditions they might be successful, has clear implications for designing health informati on for individuals persuaded through
140 each of the two routes. This insight can help boost the effectiveness of health materials, both when the ELM is used alone as well as in conjunction with health behavior based constructs. In addition to providing suppo rt for use of the ELM in health campaigns and extending the dependent variables on which it has been tested, this research also extends the use of experimental methodology in the field of public relations. Experimentation is the most rigorous methodology; however, it is rarely used in public relations research (Boynton & Dougall, 2006). Advancing the use of experiments can provide the field with additional possibilities for evaluating and measuring the effect of their activities.
141 APPENDIX A
143 APPENDIX B B log Manipulations Manipulation 1 : High credibility source with weak arguments Manipulation 2: High credibility source with strong arguments Manipulation 3: Low credibility source with w eak arguments Manipulation 4: Low credibility source with strong arguments
148 LIST OF REFERENCES Absolom, K., Eiser, C., Greco, V., & Davies, H. (2004). Health promotion for survivors of childhood cancer: A minimal intervention [Electronic version ]. Patient Education and Counseling, 55 (3), 379 384. Achille, M., Rosberger, Z., Robitaille, R., Lebel, S., Gouin, J., Bultz, B., et al. (2006). Facilitators and obstacles to sperm banking in young men receiving gonadotoxic chemotherapy for cancer: The per spective of survivors and health care professionals [Electronic version]. Human Reproduction 21 (12), 3206 3216. p rocess. (2004, May 26). Business Wire Retrieved February 16, 2010, from http://findarticles.com/p/articles/mi_m0EIN/is_2004_May_26/ai_n6045405/?tag=c ontent;col1 Agostinelli, G., & Grube J. (2002) Alcohol cou nter advertising and the media: A review of recent r esearch [Electronic version] Alcohol Research and Health 26 (1), 15 21 Ajzen, I. (2006 ). Constructing a TpB q uestionnaire: Conceptual and methodological c onsiderations Retrieved August 9, 2009, from htt p://people.umass.edu/aizen/pdf/tpb.measurement.pdf American Cancer Society. (2008). Cancer facts & f igures 2008 Retrieved September 16, 2008, from http://www.cancer.org/docroot/stt/stt_0.asp Aronson, E., Ellsworth, P. C., Carlsmith, J. M., & Gonzales, M. H. ( 1990). Methods of research in social p sychology New York: McGraw Hill
149 Arora, N. K., Johnson, P., Gustafson, D. H., McTavish, F., Robert P. Hawkins, R. P., & Pingree, S. (2002). Barriers to information access, perceived health competence and psycho social health outcomes: Test of a mediation model in a breast cancer sample [Electronic version]. Patient Education and Counseling, 47 (1), 37 46 sbring, P., & Nrvnen, A. (2004, February). Patient power and control: A study of women with uncertain illn ess trajectories [Electronic version]. Qualitative Health Research 14 (2), 226 240. Atkin, C. (2001). Impact of public service advertising: Research evidence and effective strategies Retrieved October 23, 2008, from http://www.kff.org/entmedia/3153c index.cfm Avery, E. (2010). The role of source and the factors audiences rely on in evaluating credibility of health information [Electronic version] Public Relations Review 36 (1), 81 83. Ayers, S. L., & Kronenfeld, J. J. (2007) Chronic illness and health seeking information on the Internet [Electronic version]. health: An Interdisciplinary Journal for the Social Study of Health, Illness and Medicine, 11 (3): 327 347. Bagozzi, R. (1978, January). Th e construct validity of the affective, behavioral, and cognitive components of attitude by analysis of covariance structures [Electronic version]. Multivariate Behavioral Research 13 (1), 9 31 Bandura, A. (1977). Self efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84 (2 ), 191 215. Baran, S. J., & Davis, D. K. (2003). Mass communication theory: Foundations, ferment, and future Belmont, CA: Wadsworth/Thomson Learning.
150 Barbaro, M. (2006, March 7). Wal Mart enlists bloggers in P .R. campaign [Electronic version]. New York Times Berland, G. K., Elliott, M. N., Morales, L. S., Algazy, J. I., Kravitz, R. L., Broder, M. S., et al. (2001). Health information on the Internet: Accessibility, quality, and readability in English and Span ish [Electronic version]. The Journal of the American Medical Association ( 285 ) 20, 2612 2621. Blood, R. (2000, September 7). Weblogs: A history and perspective Retrieved November 13, 2008, from http://www.rebeccablood.net/essays/weblog_history.html Blood, R (2002). The weblog handbook: Practical advice on creating and maintaining your blog Cambridge, MA: Perseus Books Group. Blood, R. (2004). How blogging software reshapes the online community [Electronic version]. Communications of the ACM 47 (12), 53 55. Booth Butterfield, S. impact on persuasion, theory and research [Electronic version]. In J. P. Dillard, & M. Pfau (Eds .), The persuasion handbook: Developments in theory and practic e (pp. 155 174) Thousand Oaks, CA: Sag e Boynton, L. & Dougall, E. (2006). The methodological avoidance of experiments in public relations research [Electronic version] PRism, 4 (1), 1 14. Bra verman J. (2008 ). Testimonials versus informational persuasive messages: The moderating effect of delivery mode and personal involvement [Electronic version]. Communication Research 35 (5), 666 694.
151 Briol, P. & Petty, R. (2006 ). Fundamental processes le ading to attitude change: Implications for cancer prevention communications [Electronic version]. Journal of Communication 56 S81 S104. Broom, A. (2005). Virtually Healthy: The impact of Internet use on disease experience and the doctor patient relations hip [Electronic version]. Qualitative Health Research, 15 (3), 325 345. Brown, W. J., Basil, M.D., & Bocarnea, M. C. (2003). The influence of famous athletes on health beliefs and practices: Mark McGwire, child abuse prevention, and androstenedione [Elect ronic version] Journal of Health Communication, 8, 41 57 Bulsara, C. E. Styles, I., Ward, A. M., & Bulsara, M. K. (2006). The psychometrics of developing the Patient Empowerment Scale [Electronic version]. Journal of Psychosocial Oncology, 24 (2), 1 16. Bulsara, C., Ward, A., & Joske, D. (2004). Patient involvement in clinical nursing haematological cancer patients: Achieving a sense of empowerment by use of strategies to control illness [Electronic version]. Journal of Clinical Nursing 13 (2), 251 258. C acioppo J. T. & Petty R. E. (1981). Social psychological procedures for cognitive response assessment: The thought listing technique. In T. V. Merluzzi C. R. Glass & M. Genest (Eds.), Cognitive assessment (pp. 309 342). New York: Guilford Press.
152 Cacio ppo, J. T., & Petty, R. E. (1987). Stalking rudimentary processes of social influence: A psychophysiological approach. I n M. Zanna et al. (E ds.) Social Influence: The Ontario Symposium, Vol. 5 Hillsdale, NJ: Lawrence Erlbaum. Callison, C. (2001). Do PR pr actitioners have a PR problem? The effect of associating source with public relations and client negative news on audi ence perceptions of credibility [Electronic version]. Journal of Public Relations Research 13 219 234. Cawley, M., Kostie, J., & Capell o, C. (1990). Informational and psy chosocial needs of women choosing conservative surgery/primary radiation for breast cancer [Electronic version] Cancer Nursing, 13 90 94. Chaiken, S. (1980) Heuristic versus systematic information processing and the use of source versus message cues in persuasion [Electronic version] Journal of Personality and Social Psychology 39 752 766. Chaiken, S. (1987) The heuristic model of persuasion. In M. P. Zanna, J. M. Olson and C. P. Herman (eds.), Social influence: The O ntario Symposium (Volume 5, pp. 3 39), Hillsdale, NJ: Erlbaum Chaiken, S., Duckworth, K., & Darke, P. (1999). When parsimony fails [Electronic version]. Psychological Inquiry 10 (2), 118 123. Chen, X., & Siu, L. L. (2001). Impact of the media and the In ternet on oncology: Survey of cancer patients and o ncologists in Canada [Electronic version]. Journal of Clinical Oncology ,19 (23), 4291 4297.
153 Cline, R. J., & Noland, V. J. (2002, November). The dark side of the Internet: Risk promoting functions of interac tive health communication. Paper presented at the American Public Health Association conference in Philadelphia, PA. comScore. (2005, August). Behaviors of the b logosphere Retrieved March 30, 2006, from www.comscore. com comScore. (2008, April 10). Pregnancy and cancer top the list of the most commonly searched health conditions Retrieved November 30, 2008 from http://www.comscore.com/press/release. asp?press=2171 comScore. (2008, September 9). Online health information category grows at rate four times faster than total Internet Retrieved November 30, 2008, from http://www.comscore .com/press/release.asp?press=2436 Cook, T., & Campbell, D. (1979). Quasi experimentation: Design and analysis issues for f ield s ettings Boston: Houghton Mifflin Company. Damian, D., & Tattersall, M. H. (1991) Letters to patients: Improving communication in cancer care [Electronic version]. Lancet, 338 923 992. Doctorow, C., Dornfest, R., Johnson, S., Powers, S., Trott, B., & Trott, M. G. (2002). Essential blogging Drezner, D., & Farrell, H. (2004, November/Decembe r). Web of influence [Electronic version]. Foreign Policy 32 40. Dutta Bergman M. J. ( 2004). Primary sources of health information : Comparisons in the domain of health attitudes health cognitions, and health behaviors [Electronic version]. Health Commun ication, 16 (3), 273 288.
154 Retrieved November 18, 2009, from http://www.edelman.com/ image/insights/content/ISwp_TrustMEdia_FINAL.pdf Edwards, M., Davies, M., & Edwards, A. (2009). What are the external influences on information exchange and shared decision making in healthcare consultations: A meta synthesis of the literature [Electronic version]. Patient Education and Counseling, 75 (1), 37 52. Epstein, R. M. (2000). The science of patient centered care [Electronic version] Journal of Family Practice, 49 (9), 805 807. Eysenbach, G., & Khler C (2002). How do consumers search for and appr aise health information on the world wide web? Qualitative study using focus groups, usability tests, and in depth interviews [Electronic version]. British Medical Journal, 324 573 577. Fishman, J. M., Casarett, D., & Davis, L. (2006). Mass media and medi cine: When the most trusted media mislead [Electronic version]. Mayo Clinic Proceedings, 81 (3), 291 293. Flynn, N. (2006). Blog rules: A Business guide to managing policy, public relations, and legal issues New York: American Management Association. Fox, S. (2006, October 29). Online health search 2006 Retrieved November 18, 2009, from http://www.pewinternet.org/Reports/2006/Online Health Search 2006.aspx Fox, S. (2008, August 26 ). The engaged e patient population Retrieved November 18, 2009, from http://www.pewinternet.org/Reports/2008/The Engaged Epatient Population.aspx
155 Fox S. & Jones, S. (2009, June). The social life of health information Retrieved January 16, 2010, from http://www.pewinternet.org/Reports/2009/8 The Social Life of Health Information.aspx Giustini, D. (2006 ). How Web 2.0 is changing medicine [Electronic version]. British Medical Journal, 333 1283 1284. Grossbart, S. Muehling, D. D., & Kang un N. (1986). Verbal and visual references to competition in comparative advertising [Electronic version]. Journal of Advertising 15 (1), 10 23. Grunig, J. E. (1993). Image and substance: From symbolic to behavioral relationships [Electronic version] Pub lic Relations Review, 19 121 139. Guidewire Group. (2005, October). Blogging in the enterprise: Executive s ummary Retrieved November 25, 2008, from http://www.awarenessnetwo rks.com/resources/GW_Survey_BlogEnt.pdf Grhan Canli, Z., & Maheswaran, D. (2000). Cultural variations in country of origin effects [Electronic version] Journal of Marketing Research, 37 309 317. Haard, J., Slater, M., & Long, M. (2004, October). Scient ese and ambiguous citations in the selling of unproven medical treatments [Electronic version]. Health Communication 16 (4), 411 426 Halzak, S. (2008, August 25). Marketing moves to the blogosphere: Business model shifts to engage customers online [Electr onic version]. R etrieved December 2, 2008, from http://www.washingtonpost.com/wpdyn/content/article/ 2008/08/24/AR2008082401517.html
156 Harmon, A. (2003, August 25) Finding comfort in strangers with an online diet journal. Electronic version]. New York Times n.p. Health 2.0 [Electronic version]. (2007, September 8). Economist 16. Heesacker, M., Petty, R., & Cacioppo, J. (1983). Field dependence and attitude change: Source credibility can alter persuasion by affecting message relevant thinking [Electronic version] Journal of Personality 51 (4), 653 666. Henning, J. (2004). The blogging iceber g Retrieved December 2, 2008, from http://perseus.com/survey/resources/perseus_blogging_iceberg.pdf Hesse, B. W., Moser, R. P., Rutten, L. F., & Kreps, G. L. (2006). The Heal th Information National Trends Survey: Research from the baseline [Electronic version]. Journal of Health Communication, 11 (1), 7 16. Hevern, V. W. (2004). Threaded identity in cyberspace: Weblogs and positioning in the dialogical self [Electronic version ]. Identity 4(4), 321 335. Hewitt, H. (2005). Blog: Understanding the information reformation that's changing your world Nashville: Thomas Nelson. Hillan J. (2003). Phy sician use of patient centered w eblogs and online journals [Electronic version]. Clini cal Medicine & Research, 1 (4), 333 33 5. Holt, C. L., Lee, C. & Wright, K. (2008). A spiritually based approach to breast cancer awareness: Cognitive response analysis of communication effectiveness [Electronic version]. Health Communication, 23 (1), 13 22. Homer, P. (1995) Ad size as an indicator of perceived advertising costs and eff ort: The effects on memory and perceptions [Electronic version]. Journal of Advertising 24 (Winter), 1 12.
157 Hon, L. C., & Grunig, J. E. (1999). Measuring relationships in public relations. Retrieved October 23, 2009, from http://www.instituteforpr.org/research_single/guidelines_measuring_relationships/ Horton, R., Minniti, A., Mireyl ees, S., & McEntegart, D. (2008, November). A randomized trial to determine the impact on compliance of a psychophysical peripheral cue based on the Elaboration Likelihood Model [Electronic version]. Contemporary Clinical Trials 29 (6), 823 828. Hourihan, DevCenter. Retrieved December 27, 2008 from www.oreillynet.com/lpt/a/2474 Hovland, C. I., Janis, I. L., & Kelley H. H. (1953) Communicat ion and persuasion New Haven, CT: Yale University Press. Huang, G. J. & Penson, D. F. (2008). Internet health resources and the cancer patient [Electronic version] Cancer Investigation 26(2), 202 207. Igartua J J Cheng L & Lopes, O. (2003). To th ink or not to think: Two pathways towards persuasion by short films on AIDS prevention [Electronic version] Journal of Health Communication ,8 513 528. Institute of Medicine. (2004, April). Health l iteracy: A prescription to end confusion Retrieved Janua ry 28, 2010 from http://www.iom.edu/Reports/2004/Health Literacy A Prescription to End Confusion.aspx I nstitute of Medicine. (2007a, October). Report brie f for providers Cancer care for the whole patient: Meeting psychosocial health needs Retrieved March 6 2010 from http://books.nap.edu/openbook.php?record_id=11993
158 Institute of Medicine. (2007b, Octo ber). Report brief for patients. Cancer care for the whole patient: Meeting psychosocial health needs Retrieved March 6, 2010 from http://www.iom.edu/~/media/Files/Report%20Files/2007/Cancer Care for the Whole Patient Meeting Psychosocial Health Needs/PsychosocialpatientsReportBriefFINAL3web.ash It's the links, stu pid [Electronic version] (2006, April 20). The Economist Retrieved November 21, 2008, from http://www.economist.com/surveys/ displaystory.cfm?story_id=6794172 Jensen, J. cept in media and communication studies [Electronic version] Nordicom Review, 19 (1):185 204. Johansson, K., Nuutila, L., Virtanen, H., Katajisto, J., & Salanter, S. (2005). Preoperative educa tion for orthopae dic patients: S ystematic review. Journal of Advanced Nursing 50 (2), 212 223. Johnson, B. T., & Eagly, A. H. (1989). Effects of involvement on persuasion: A meta analysis [Electronic version]. Psychological Bulletin, 106 (2), 290 314. Johnson I. A., & Ade lstein D. J (1991). The use of recorded interviews to enhance physician patient communication [Electronic version] Journal of Cancer Education, 6 (2), 99 102. Johnson, R. J., McCaul, K. D., & Klein, W. M. (2002). Risk involvement and risk perception amo ng adolescents and young adults [Electronic version] Journal of Behavioral Medicine, 25 (1), 67 82
159 Jones, L., Sinclair, R., & Courneya, K. (2003). The effects of source credibility and message framing on exercise intentions, behaviors, and attitudes: An integration of the Elaboration Likelihood Model and Prospect Theory [Electronic version] Journal of Applied Social Psychology 33 (1), 179 196. Jones, S., & Fox S. (2009, January 28). Gene rations online: Charts. Retrieved January 16, 2010, from http://www.slideshare.net/PewInternet/generations online in 2009 charts?type=powerpoint Kaye, B Atlantic Journal of Communication 13 (2), 73 95. Kelleher, T., & Miller, B. M. (2006). Organizational blogs and the human voice: Relational strategies and relational outcomes [Electroni c version] Journal of Computer Mediated Communication 11 395 414. Kidd L Hubbard G R & Kearney N (2009) Perceived control and involvement in self care in patients with colorectal cancer [Electronic version] Journal of Clinical Nur sing, 18 (16), 2292 2300. Kim, S., & Chung, D. (2007, October). Characteristics of cancer blog users [Electronic version] Journal of the Medical Library Association 95 (4), 445 450. Kirkpatrick, D. (2005, January 10). Why there's no escaping the blog. Fort une R etrieved December 2, 2008, from http://money.cnn.com/magazines/fortune/ fortune_archive/2005/01/10/8230982/index.htm Koenig, D. (2006, January). Log on and lose weight [Electronic version] Fitn ess, 62 63. Kotler P., Roberto, N., & Lee, N. (2002). Social marketing: Improving the quality of life (2nd ed.). Thousand Oaks, CA: Sage Publications.
160 Kreuter, M. W., Sugg Skinner, C., Holt, C. L., Clark, E. M., Haire Joshu, D., Fu, Q., et al (2005). Cul tural tailoring for mammography and fruit and vegetable intake among low income African American women in urban public health centers [Electronic version] Preventive Medicine: An International Journal Devoted to Practice and Theory 41 53 62. Kumar, R., Novak, J., Raghavan, P., & Tomkins, A. (2004). Structure and evolution of blogspace [Electronic version] Communications of the ACM, 47 (12), 35 39. Kummervold, P. E., Gammon, D., Bergvik, S., Johnsen, J. K., Hasvold, T., & Rosenvinge, J. H. (2002). Social support in a wired world [Electronic version]. Nordic Journal of Psychiatry, 56 (1), 59 65. Kung, H. C., Hoyert, D. L., Xu, J. Q., Murphy, S. L. (2008). Deaths: Final data for 2005. Na tional vital statistics reports Retrieved November 30, 2008, from http://www.cdc.gov/nchs/data/nvsr/nvsr56/nvsr56_10.pdf Kutner, M., Greenberg, E., Jin, Y., & Paulsen, C. (2006, September). The health literacy ional Assessment of Adult Literacy Retrieved October 22, 2009 from http://nces.ed.gov/pubs2006/2006483.pdf \ Lagu T., Kaufman E. J., Asch D. A., & Armstrong K. (2008). Content of weblogs writte n by health professionals [Electronic version] Journal of General Internal Medicine 23 1642 1646. Landro, L. (2005, May 4). Blogging from your sickbed [Electronic version]. The Wall Street Journal (Eastern Edition), n.p.
161 Ledingham, J. A. (2003). Explica ting relationship management as a general theory of pub lic relations [Electronic version ]. Journal of Public Relations Research, 15 (2), 181 198. Lord, K. R., & Burnkrant, R. E. (1993). Attention versus distraction: The interactive effect of program involve ment and attention devices on commercial processing [Electronic version]. Journal of Advertising 22 (1), 47 60. Lord, K. R., Lee, M., & Sauer, P. L. (1994). Program context antecedents of a tt itude toward radio c ommercials [Electronic version]. Journal of t he Academy of Marketing Science 22 (1), 3 15. Lord, K. R., Lee, M., & Sauer P. L. (1995). The combined influence hypothesis: Central and peripheral antecedents of attitude toward the ad [Electronic version] Journal of Advertising 24 (1), 73 85. Luker, K A., Beaver, K., Leinster, S. J., Owens, R. G., Degner, L. F., & Sloan, J. A. (1995). The information needs of women newly diagnosed with breast cancer [Electronic version] Journal of Advanced Nursing 22 134 141. Magee M. (2003, September 23). Physici an patient relationships, patient empowerment and the role of i nformation. PowerPoint presentation to th e Food and Drug Administration direct to consumer public h earing. Retrieved December 2, 2009, from http://www.fda.gov/cder/ddmac/p6magee/sld015.htm Marks, D. F., Murray, M., Evans, B., Willig, C., Woodall, C., & Sykes, C. M. (2006). Health psychology: Theory, research and practice (2 nd ed.). Thousand Oaks, CA: Sage.
162 McCullough, T., & Dodge, H. R. (2002). Understanding the role consumer involvement plays in the effectiveness of hospital advertising [Electronic version] Health Marketing Quarterly, 19 (3), 3 20. McGuire, W. (1999). Constructing social p sychology: Creative and critical a spects N ew York: Cambridge University Press. McGuire, W. J. (2000) Input and output variables currently promising for construction persuasive communications In R. E. Rice & C. K. Atkin (Eds.), Public communication campaign s, (2nd ed. pp. 22 48) Thousand Oaks, C A: Sage. McMillan, S. J. (1999). Health communication and the internet: Relations between interactive characteristics of the medium and site creators, conte nt, and purpose [Electronic version]. Health Communication, 11 (4B), 375 390. McNeil, M. (2006, Febru ary 18). Cancer patient blogs to shed light on treatment [Electronic version]. The Associated Press n.p. Metz, J. M., Devine P., DeNittis A., Jones, H., Hampshire, M Goldwein, J., & Whittington, R. (2003). A m ulti institutional study of Internet utiliza tion by radiation oncology patients [Electronic version] International Journal of Radiation Oncology Biology Physics 56 (4):1201 120 5. Miller, G. R. (2002). On being persuaded: Some basic distinction s. In J. P. Dillard & M. W. Pfaf (Eds.), The p ersu a sion handbook: Developments in t he ory and p ractice (pp. 3 16). Thousand Oaks, CA: Sage. Molassiotis, A., & Xu, M. (2004). Quality and safety issues of web based information about herbal medicines in the treatment of cancer [Electronic version]. Complementary T herapies in Medicine, 12 217 227.
163 Morgan, G. A., Leech, N. L., Gloeckner, G. W., Barrett, K. C. (2004). SPSS for introductory statistics: Use and Interpretation (2nd ed.). Mahwah, NJ: Lawrence Erlbaum. Morris, C. A., & Avorn, J. (2003). Internet marketing of herbal products [Electronic version]. JAMA/Journal of the American Medical Association, 290 (11), 1505 1509. National Cancer Institute. (2005 2007 ). Health Information National Trends Survey: How Americans find and use health information Retrieved Nov ember 30, 2009 from http://hints.cancer.gov/questions/index.jsp National Cancer Institute. (2008). Surveillance Epidemiology End Results SEER stat fact sheets: Cancer, all sites Retrieved Septem ber 9, 2008, from http://seer.cancer.gov/statfacts/html/all.html National Cancer Institute (n.d). St ate cancer profiles: Incidence rate report for Florida by c ounty Retrieved January 16, 2010 from http://statecancerprofiles.cancer.gov/incidencerates/index.php National Cancer Institute SEER cancer statistics r eview 1975 2006. (n.d). Table 2.18 United States cancer p revalence estimates. Retrieved January 16, 2010, from http://seer.cancer.gov/csr/1975_2006/browse_csr.php?section=2&page=sect_02 _table.18.htmlc Neuhau ser, L., & Kreps, G. L. (2003). Rethinking communication in the e health era [Electronic version]. Journal of Health Psychology, 8 (1), 7 23. Newsom, D. A., Ramsey, S. A.,& Carrell, B. J. (1993). Chameleon chasing II: A replication [Electronic version] Pub lic Relations Review, 19 33 47.
164 Nyatanga, L., & D ann, K. (2002). Empowerment in n ursing: The role of philosophical and psychological factors [ Electronic version] Nursing Philosophy 3 234 239. O'Donnell, V., & Kable, J. (1982). Persuasion: An interactiv e dependency approach New York: Random House. Ohanian, R. (1990). Construction and validation of a scale to measure celebrity orthiness, and attractiveness [Electronic version] Journal of Advertising 19 ( 3 ), 39 52. O'Keefe, D. J. (1990). Persuasion: Theory and research Newbury Park, NJ: Sage O'Keefe G. J., Boyd, H. H., & Brown, M. R. ( 1998). Who learns preventive health care information from where: Cross channel and repertoire comparisons [Electronic version] Hea lth Communication, 10 (1), 25 36. Oransky, I. (2005). Cancer blogs [Electronic version]. Lancet Oncology 6, 838 839. Perloff, R. M. (1993). The dynamics of persuasion Hill sdale, NJ: Lawrence Erlbaum Petersen, S., Heesacker, M., Schwartz, R., & Marsh, R. (2000, September). Predictors of decision making style among cancer patients: An empirical test of theory [Electronic version] Psychology & Health 15 (5), 663 675. Petty R. E. & Cacioppo J. T. ( 1981 ). Attitudes and persuasion: Classic and contemporary approaches Dubuque IA : William C. Brown Petty, R.E., & Cacioppo, J.T. (1986). Communication and persuasion: Central and peripheral routes to attitude c hange New York: Springer Verlag.
165 Petty, R. E., Haugtvedt, C. P., & Smith, S. M. (1995). Message ela boration as a determinant of attitude strength. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength: Antecedents and consequences Hillsdale, NJ: Lawrence Erlbaum Associates. Pew Internet & American Life Project. (2005, May 2). New data on blogs a nd blogging Retrieved December 2, 2008, from http://www.pewinternet.org/press_release.asp?r=104 Pew Research Center for People and the Press. (2009, December 21). Current decade rates as wor st in 50 years: Internet, cell phones are changes for the better Retrieved January 16, 2010, from http://people press.org/reports/pdf/573.pdf Pornpitakpan, C. (2004). The persuasiveness of source credibility: A critical review of five decades' e vidence [Electronic version] Journal of Applied Social Psychology 34 (2), 243 281. Porter, L. V., Trammell, K. S., Chung, D., & Kim, E. (2007). Blog power: Examining the effects of practitioner blog use on power in public relations [Electronic version] Public Relations Review, 33 (1), 92 95. Prochaska, J. O., DiClemente, C. C., & Norcross, J. C. (1992). In search of how people change: Applications to addictive behaviors. American Psychologist 47 1102 11 14 Rainey, L. C. (1985). Effects of preparatory patient education for radiation oncology patients [Electronic version]. Cancer, 56 1056 1061. Rainie L. (2009, October 7). The rise of the e patient Retrieved January 16, 2010, from http://www.pewinternet.org/Presentations/2009/40 The rise of the e patient.aspx
166 Rainie, L. (2010, January). Internet, broadband, and cell phone statistics Retrieved January 16, 2010, from http://www.pewinternet.org/Reports/2010/Internet broadband and cell phone statis tics.aspx Rainie, L., & Keeter, S. (2006, April 3). Americans and their cell phones. Retrieved January 23, 2010, from http://www.pewinternet.org/Reports/2006/Amer icans and their cell phones.aspx Rak, J. (2005). The digital queer: Weblogs and Internet identity [Electronic version]. Biography: An Interdisciplinary Quarterly 28 (1), 166 182. Ramanadhan, S., & Viswanath, K. (2006). Health and the information nonseeker : A profile [Electronic version] Health Communication, 20 (2), 131 139. Rausch, P., & Edwards, H. E. (2008). An assessment of injuries and the potential for injuries in a multi tasking generation of college students American Public Health Association Stu dent Research Award winner. Poster presented October 27 during national con vention in San Diego, CA Research and Markets: Healthcare blogs are expected to be used more as consumers take greater responsibility for their healthcare [Electronic version]. (20 06, March 27). Business Wire n.p. Rimal, R. N., Flora, J. A., & Schooler, C. (1999). Achieving improvements in overall health orientation [Electronic version] Communication Research, 26 (3), 322 348 Roberts K. J. (1999). Patient empowerment in the Unite d States: A critical commentary [Electronic version]. Health Expectations 2 (2), 82 92. Robinson, J., & Ginsburg, P. (2009). Consumer driven health care: Promise and performance [Electronic version]. Health Affairs 28 Supplement 1, W272 W281.
167 Rogers, E. M (1994). A history of communication s tudy: A biographical approach New York: Free Press. Rutten, L., Arora, N., Bakos, A., Azi z, N., & Rowland, J. (2005 ). Information needs and sources of information among cancer patients: A systematic review of research (1980 2003) [Electronic version] Patient Education & Counseling 57 (3), 250 261. Sallot, L. M., & Johnson, E. A. (2006). To contact or assessments of public relations subsidies and contact preferences [Electronic version] Public Relations Review, 32 83 86 Salmon C. T., & Atkin, C. (2003 ). Using media c ampaigns for health p romotion In T. L. Thompson, A. M. Dorsey K. I. Miller & R. Parrott (Eds.), Handbook of Health Communication (pp. 449 472) Mahwah, NJ : Lawrence Erlb aum. Sandberg, H. (2005). Information and communication in society [Electronic version]. Acta Paediatricia, 94 (Suppl 448), 38 39. [Electronic version] Public Relations Journal, 49, 8. Schneider, C., & Hall M. (2009). The patient life: Can consumers direct health care ? [Electronic version]. American Journal of Law & Medicine 35 (1), 7 65. Scott, D. M. (2009). The new rules of marketing and PR: How to use news releases, blogs, podcasting, viral marketing, an d online media to reach buyers directly Hobo ken, NJ: John Wiley and Sons.
168 Science panel on interactive communication and health. (April, 1999). Wired for health an d well being: The emergence of interactive health c ommunication [Electronic version]. Washi ngton, D.C.: U.S. Department of Health and Human Services, U .S. Government Printing Office. Seale, C. (2005). Portrayals of treatment decision making on popular breast and prostate cancer web sites [Electronic version] European Journal of Cancer Care 14 ( 2), 171 174. Seale, C., Ziebland, S., & Charteris Black, J. (2006). Gender, cancer experience and internet use: A comparative keyword analysis of interviews and online cancer support groups [Electronic version] Social Science & Medicine 62 (10), 2577 2590 Sharf, B. F. (1997). Communicating breast cancer on l ine: Support and empowerment on the i nternet [Electronic version]. Women and Health, 26 (1), 65 84. Shim, M., Kelly, B., & Hornik, R. (2006). Cancer information scanning and seeking behavior is associat ed with knowledge, lifestyle choices, and screening [Electronic version]. Journal of Health Communication, 11 (1), 157 172. Slater, M. (2006). Specification and misspecification of theoretical foundations and logic models for health communication campaigns [Electronic version] Health Communication 20 (2), 149 157. Smith, S. A. (2005). A newsroom's fortress walls collapse [Electronic version]. Nieman Reports 59 (3), 44 45.
169 Smith, A. (2008, July 22). New numbers for blogging and blog readership Retrieved No vember 25, 2008, from http://www.pewinternet.org/PPF/p/1494/pipcomments.asp Snider, M. (2003, September 15). The intimacy of blogs [Electronic version]. Maclean's, 116 (37), 40 41. Speros C. (2005). Health literacy: Concept analysis [Electronic version]. Journal of Advanced Nursing, 50 (6), 633 640. Squiers, L., Rutten, L., Treiman, K., Bright, M., & Hesse, B. (2005, October 2). Cancer patients' information needs across the cancer care con tinuum: Evidence from the Cancer Information Service [Electronic version] Journal of Health Communication 10 15 34. Stayman, D. M., & Batra, R. (1991). Encoding and retrieval of ad affect in memory [Electronic version] Journal of Marketing Research, 28 (2), 232 239 Stewart, M., Brown, J. B., Donner, A., Mcwhinney, I. R ., Oates, J., Weston, W. W., & Jordan, J. (2000). The impact of patient centered care on outcomes [Electronic version] Journal of Family Practice 49 ( 9), 796 804. Stiff, J. B. (1986). Cogn itive processing of persuasive message cues: A meta analytic review of the effects of supporting information on attitudes [Electronic version] Communication Monographs 53 (1), 75 89. Stiff, J. B., & Mongeau, P. A. (2003). Persuasive c ommunication ( 2 nd ed. ). New York: Guilford Press.
170 Sweetser, K. D., & Metzgar, E. (2007). Communicating during crisis: Use of blogs as a relationship management tool [Electronic version] Public Relations Review 33 340 342. Tatsioni, A., Gerasi, E., Charitidou, E., Simou, N., Mavreas, V., & Ioannidis, P. A. (2003). Important drug safety infor mation on the i nternet [Electronic version]. Drug Safety, 26 (7), 519 527. Technorati. (2008). State of the blogosphere 2008. Retrieved November 23, 2008, from http://www.technorati.com/blogging/state of the blogosphere/ The health care blog (2005, January 1). About THCB Retrieved December 4, 2008, from http://www.thehealthcareblog.com/ Thielst, C. B. (2007). We blogs: A communication tool [Electronic version]. Journal of Healthcare Management 52(5), 287 289. Trenholm, S. (1989) Persuasion and social influence Englewood Cliffs, NJ: Prentice Hal l. Trifts, V., & Haubl, G. (2003). Information availability and cons umer preference: Can online retailers benefit from providing access to competitor price information? [Electronic version]. Journal of Consumer Psychology, 13 (1/2), 149. Updegraff, J. A., Sherman, D. K., Luyster, F. S., & Mann, T. L. (2007).The effects of m essage quality and congruency on perceptions of tailored health communications [Electronic version] Journal of Experimental Social Psychology, 43 (2), 249 257.
171 U.S. Department of Health and Human Services. ( log: About this b log. Retrieved December 4, 2008, from http://secretarysblog.hhs.gov/about.html U.S. Department of Hea l th and Human Services. (2000, Nove mber). Healthy People 2010: Understanding and improving h ealth ( 2 nd ed. ). Retrieved November 19, 2008 from http://www.healthypeople.gov/Document/pdf/Volume1/11HealthCom.pdf U.S. Department of Health and Human Services. (2008, March). HH S: What we do. Retrieved December 4, 2008, from http://www.hhs.gov/about/whatwedo.html/ Universal McCann. (2008, March). Power to the people: Social media tracker Wave.3. Retrieved December 2, 2008, f rom http://www.universalmccann.com/Assets/2413%20 %20Wave%203%20complete%20document%20AW%203_20080418124523.pdf Vivian, J. (2009). T he media of mass communication ( 9 th ed. ). Boston, MA: Pearson. Wallston, K. A.,Wallston, B. S., & DeVellis, R. (1978) Development of the Multidimensional Health Locus of Control (MHLC) scales. Health Education Monographs 6 160 170. Weinstein, N. D. (19 84). Why it won't happen to me: Perceptions of risk factors and susceptibility [Electronic version]. Health Psychology,3 (5), 431 457. What the in crowd knows: From Hollywood to Wall Street, our guide to the blogs insiders read to stay current [Electronic version]. (2005, November 16). The Wall Street Journal.
172 Williams, T. (2002). Patient empowerment and ethical decision m aking [Electronic version] Dimensions of Critical Care Nursing 21 (3), 100 104. Wilson, B. J. (2007). Designing media messages about hea lth and nutrition: What strategies are most effective? [Electronic version]. Journal of Nutrition Education and Behavior, 39 (2), Suppl. 1, S13 S19. Winer, D. (2001, November 16). The history of weblogs Retrieved January 23, 2008, from http://www.userland.com/theHistoryOfWeblogs Winer, D. (2003, May 23). What makes a weblog a weblog ? Retrieved December 27, 2008, from http://blogs.la w.harvard.edu/whatMakesAWeblogAWeblog Withers, G., & Wertheim, E. (2004). Applying the Elaboration Likelihood Model of persuasion to a videotape based eating disorders primary prevention program for adolescent girls [Electronic version] Eating Disorders 12 (2), 103 124. Woodward, G. C., & Denton, R. E. (1992). Persuasion and influence in American life (2nd ed.) Prospect Heights, IL: Waveland
173 BIOGRAPHICAL SKETCH Paula Rausch received a Bachelor of Arts in communication, graduating with honors from t he University of North Florida in Jacksonville. She also holds an Associate of Science degree in nursing from Minneapolis (Minnesota) Community College. She spent seven years as a professional reporter covering government, politics, health, business, and o ther beats for three Florida newspapers, during which time she won several journalism awards. She worked for the University of Florida in communication related positions before pursuing her graduate studies at the UF College of Journalism and Communication s beginning in the f program in Fall 2006, jointly pursuing a Certificate in Public Health at the UF College of Public Health and Health Professions. During her graduate education, Rausch was involved in a num ber of international communication pursuits, including conducting research in Nicaragua and Ethiopia, and serving as a program assistant and instructor for UF study abroad programs in Italy, Greece and Spain, and France. She was awarded a health communicat ion fellowship at the National Cancer Institute (NCI) part of the National Institutes of Health, in Bethesda, Maryland, where she spent her final six months as a doctoral student. After she graduates with her Ph.D. in August 2010, Rausch will join the staff of the NCI Office of Communications and Education as a Communications Program Manager in the Communication s Planning and Coordination Branch