Citation
HEALTH BELIEF MODEL IN AN INTERACTIVE AGE

Material Information

Title:
HEALTH BELIEF MODEL IN AN INTERACTIVE AGE
Copyright Date:
2008

Subjects

Subjects / Keywords:
Cultural studies ( jstor )
Culture contact ( jstor )
Health information ( jstor )
Internet ( jstor )
Pews ( jstor )
Research facilities ( jstor )
Research studies ( jstor )
Vaccinations ( jstor )
We they distinction ( jstor )
West Nile virus ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright the author. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
5/4/2002
Resource Identifier:
52205228 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

HEALTH BELIEF MODEL IN AN INTERACTIVE AGE By KAYE TRAMMELL A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS IN MASS COMMUNICATION UNIVERSITY OF FLORIDA 2002

PAGE 2

Copyright 2002 by Kaye Trammell

PAGE 3

ACKNOWLEDGMENTS There were many people who helped in making this thesis possible; however the most notable are my advisor and my husband. I would like to extend my gratitude to Dr. Mary Ann Ferguson, my advisor, for guiding me through this long research journey. Through it all, she remained positive and patient. She let me discover and come to understand the principles of quality research on my own, without letting me go astray. I would also like to thank my husband, Mark R. Trammell, Jr., who supported me and selflessly went without dinner from July 2001 – May 2002 as I worked on my thesis. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS...................................................................................................3 LIST OF TABLES............................................................................................................vii LIST OF FIGURES.............................................................................................................x ABSTRACT.......................................................................................................................xi CHAPTERS 1 AN INTERACTIVE AGE................................................................................................1 2 REVIEW OF LITERATURE...........................................................................................4 Health Belief Model........................................................................................................4 History of the Health Belief Model..........................................................................7 Preventive Health Behavior Research......................................................................8 Assumptions of the Health Belief Model...............................................................10 Complementary Theories.......................................................................................10 Limitations of the Health Belief Model.................................................................11 Alternate Theories to the Health Belief Model.............................................................13 Social Learning......................................................................................................13 Gainand Loss-Framed Messages.........................................................................13 Cultural Types...............................................................................................................14 The Internet as a Health Information Source................................................................17 Internet Source Credibility.....................................................................................18 Barriers to Using the Internet as a Health Information Source..............................20 Typical Internet Health-Seekers.............................................................................21 Allocentric Web surfers...................................................................................24 Idiocentric Web surfers....................................................................................27 Internet Compared to Other Media...............................................................................27 Research Gaps...............................................................................................................29 3 METHODS.....................................................................................................................31 Variables.......................................................................................................................31 Independent Variables............................................................................................32 Dependent Variable................................................................................................34 iv

PAGE 5

Message.........................................................................................................................34 Design...........................................................................................................................37 Source Credibility.........................................................................................................39 Culture...........................................................................................................................40 Demographics...............................................................................................................40 Recall and Retention.....................................................................................................40 Procedures.....................................................................................................................41 4 RESULTS.......................................................................................................................42 Participants....................................................................................................................42 Credibility.....................................................................................................................45 Family Integrity Scale...................................................................................................47 Message.........................................................................................................................60 Analysis for Hypotheses...............................................................................................64 Post-hoc Analysis..........................................................................................................75 Recall.....................................................................................................................75 Trustworthiness......................................................................................................79 Ethnicity In the Culture Factors.............................................................................89 Frequency of Family Communication..........................................................................95 Internet Use...................................................................................................................99 5 DISCUSSION...............................................................................................................101 Summary.....................................................................................................................101 Limitations..................................................................................................................102 Measurement........................................................................................................102 Design..................................................................................................................103 Scale Limitations..................................................................................................103 Subject Limitations..............................................................................................104 Terrorist Attacks...................................................................................................105 Anthrax Over West Nile Virus.............................................................................106 Conclusions.................................................................................................................107 Culture..................................................................................................................108 Source Credibility................................................................................................109 Health Belief Model.............................................................................................110 Internet as an Information Source........................................................................110 Future Research..........................................................................................................112 APPENDICES A ATTRIBUTES DEFINING INDIVIDUALISM AND COLLECTIVISM AND THEIR ANTECEDENTS AND CONSEQUENCES FROM TRIANDIS, McCUSKER, & HUI (1990)...................................................................................................................116 B INFORMED CONSENT FOR PROTOCOL HEALTH BELIEF MODEL IN AN INTERACTIVE AGE......................................................................................................118 v

PAGE 6

C SOURCE CREDIBITLIY SCALE...........................................................................119 D FAMILY INTEGRITY FACTOR............................................................................120 E CULTURE MESSAGES...........................................................................................122 F DEMOGRAPHIC AND INTERNET USE SURVEY..............................................126 G INFORMATION RETENTION AND COMPREHENSION SURVEY..................128 H DEBRIEF STATEMENT.........................................................................................130 LIST OF REFERENCES.................................................................................................131 BIOGRAPHICAL SKETCH...........................................................................................137 vi

PAGE 7

LIST OF TABLES Table page 3-1 Research Design............................................................................................................38 4-1 Self Reported Race of Participants................................................................................43 4-2 Participants’ Communication with a Member of the Family When at School..............44 4-3 Participants’ Communication with a Member of the Family When at Home...............45 4-4 Means, Standard Deviations, Factor Loadings, and Cronbach Alpha for Source Credibility—Shands at the University of Florida......................................................46 4-5 Factor Pattern Loadings on Original Seven-Factor Solution, Principal Axis Factor Oblimin Rotation.......................................................................................................49 4-6 Pattern Factor Loadings Oblimin Roatation, Means, Standard Deviations, and Cronbach Alpha for Family Integrity Scale...............................................................51 4-7 Factor Scores for Five-Factor Culture Solution............................................................55 4-8 Items Loading on Same Factor in Pancultural Factor Analysis of Individualism-Collectivism from Triandis (1986)............................................................................57 4-9 Self-Reported Culture for All Factors for Strong Culture Participants.........................60 4-10 Manipulation Check.....................................................................................................61 4-11 Behavior Elicited from Participants After Message Exposure.....................................62 4-12 Participants’ Retention Demonstrated Through Two True/False Questions................63 4-13 Participants’ Retention Demonstrated Through Responses to Six Unaided Recall Questions...................................................................................................................64 4-14 Effects of Culturally Matched Message for Factor 1 Dependence/Reliance Factor on Desire to be Contacted..........................................................................................66 4-15 Effects of Culturally Matched Level of Interactivity for Factor 1 Dependence/Reliance Factor on Desire to be Contacted...........................................67 vii

PAGE 8

4-16 Effects of Culturally Matched Message for Factor 2 Group vs. Self Factor on Desire to be Contacted..........................................................................................................68 4-17 Effects of Culturally Matched Level of Interactivity for the Factor 2 Group vs. Self Factor on Desire to be Contacted...............................................................................69 4-18 Effects of Culturally Matched Message for Factor 3 Closeness to Ingroups Factor on Desire to be Contacted..........................................................................................70 4-19 Effects of Culturally Matched Level of Interactivity for the Factor 3 Closeness to Ingroups Factor on Desire to be Contacted...............................................................71 4-20 Effects of Culturally Matched Message for Factor 4 Family Integrity Factor on Desire to be Contacted...............................................................................................72 4-21 Effects of Culturally Matched Level of Interactivity for Factor 4 Family Integrity Factor on Desire to be Contacted...............................................................................73 4-22 Effects of Culturally Matched Message for Factor 5 Community vs. Individual Identification Factor on Desire to be Contacted........................................................74 4-23 Effects of Culturally Matched Level of Interactivity for Factor 5 Community vs. Individual Identification Factor on Desire to be Contacted.......................................75 4-24 Means, Standard Deviation and F Test for Effects of Message Matching in Closeness to Ingroups Factor on Unaided Recall......................................................76 4-25 Means, Standard Deviation and F Test for Effects of Message Matching in Closeness to Ingroups Factor on Unaided Recall......................................................79 4-26 Chi Square on Effects of Trustworthiness for Factor 2 Group vs. Self on the Desire to be Contacted..........................................................................................................82 4-27 Chi Square on Effects of Trustworthiness for Factor 3 Closeness to Ingroups on the Desire to be Contacted...............................................................................................85 4-28 Chi Square on Effects of Trustworthiness for Factor 5 Community vs. Individual Identification on the Desire to be Contacted.............................................................88 4-29 Means, Standard Deviation and F Test for Effects of Ethnicity in Dependence/Reliance Factor.....................................................................................91 4-30 Means, Standard Deviation and F Test for Effects of Ethnicity in Family Integrity Factor.........................................................................................................................93 4-31 Means, Standard Deviation and F Test for Effects of Ethnicity in Community vs. Individual Factor........................................................................................................95 viii

PAGE 9

4-32 Correlation Between Culture Factors and Communication with Family Member While at School..........................................................................................................98 4-33 Correlation Between Culture Factors and Communication with Family Member While at Home...........................................................................................................99 4-34 Correlation Between Culture Factors and Frequency of Internet Usage.....................100 ix

PAGE 10

LIST OF FIGURES Figure page 4-1. Unaided Recall: Effects of Message Matching with Closeness to Ingroups Factor.....76 4-2. Unaided Recall: Effects of Trustworthiness and Level of Interactivity Matching for the Family Integrity Culture Factor...........................................................................78 4-3. Effects of Low Trustworthiness for Factor 2 Group vs. Self on the Desire to be Contacted...................................................................................................................80 4-4. Effects of High Trustworthiness for Factor 2 Group vs. Self on the Desire to be Contacted...................................................................................................................81 4-5. Effects of Low Trustworthiness for Factor 3 Closeness to Ingroups on the Desire to be Contacted..............................................................................................................83 4-6. Effects of High Trustworthiness for Factor 3 Closeness to Ingroups on the Desire to be Contacted..............................................................................................................84 4-7. Effects of Low Trustworthiness for Factor 5 Community vs. Individual Identification on the Desire to be Contacted.............................................................86 4-8. Effects of High Trustworthiness for Factor 5 Community vs. Individual Identification on the Desire to be Contacted.............................................................87 4-9. Ethnicity of Factor 1 Dependence/Reliance..................................................................90 4-10. Ethnicity in Factor 4 Family Integrity.........................................................................92 4-11. Ethnicity in Factor 5 Community vs. Individual Identification..................................94 4-12. Ethnicity in Frequency of Family Communication When at School...........................96 4-13. Ethnicity in Frequency of Family Communication When at Home............................97 x

PAGE 11

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts in Mass Communication HEALTH BELIEF MODEL IN AN INTERACTIVE AGE By Kaye Trammell May 2002 Chair: Dr. Mary Ann Ferguson Department: Journalism and Communications As more health information seekers turn to the Internet to get answers to their health questions, it becomes evident that research is needed to test traditional health communication models in the new medium. Using the constructs of the Health Belief Model (HBM), this study attempted to test the model’s relevance in a new Interactive Age. Additionally, this study reviewed the relevance of culture, an often missing aspect of Health Belief Model research, to determine if participants are more likely to engage in preventive health behavior based on messages culturally tailored or if exposed to a message in a culturally matched level of interactivity. Participants in this study (n = 204) were recruited from journalism, business, and computer classes at the University of Florida. Participants completed a series of surveys in which they indicated 1) Shands at the University of Florida source credibility, 2) their cultural tendencies, 3) demographic information, and a 4) retention quiz. Participants were exposed to a message about West Nile Virus. The message was written to target xi

PAGE 12

specific cultures and delivered in either a brochure (static) or through a Web page (interactive). Five culture factors emerged from the data. The strongest cultural scores (40/20/40) for each factor (n = 163) were tested to see if there was an interaction between a matched message or matched level of interactivity, resulting in behavior. The first hypothesis in this study proposed that participants who receive culturally matched messages are more likely to adopt the suggested PHB. This hypothesis was tested through a crosstab chi square and produced no significant results. The second hypothesis proposed that participants who receive messages in a culturally matched level of interactivity are more likely to adopt the suggested PHB. This hypothesis was tested through a crosstab chi square and produced no significant results. While neither of the hypotheses were supported in this study, source credibility and recall came to play an important role. Participants who scored source credibility Trustworthiness low, but received a matched level of interactivity in the Family Integrity factor, had a remarkably high retention score. In fact, the score was higher than those who felt the source was highly Trustworthy. Additionally, there was a main effect in the Closeness from Ingroups factor in which a matched message correlated with high recall. xii

PAGE 13

CHAPTER 1 AN INTERACTIVE AGE Throughout time, new technologies have affected the way people communicate. Communication helps builds relationships, convey information, and persuade others. One important type of communication that both conveys information and persuades is health communication. Health communication often is centered on the goal to inform about and move people to participate in healthy behaviors and lifestyles. Heath communication messages can be effective in face-to-face communication (Reagan & Collins, 1987; Johnson et al., 1992), print (Hofstetter et al., 1992), television, and even radio (Izenburg & Lieberman, 1998b; Hofstetter et al., 1992; Eastin, 2001). Recent research finds that people now are turning to the Internet to find information about medical diseases and conditions (Napoli, 2001). In the past decade, the use of the Internet has grown substantially (Pew Research Center, 2000a; Pew Research Center, 2001). The Internet has become known as the information superhighway, where a world of information can be found at the user’s fingertips. The Internet has grown to be a dynamic encyclopedia of knowledge and information where the information seeker is in the driver’s seat. It is different from traditional communication vehicles in that users, rather than message producers, set the course of information discovery. For the first time, information tailored to the user’s needs and desires is available instantaneously. This idea of specialized information in the form of “what you want, when you want it,” has taken root in other communication arenas as well. Cable 1

PAGE 14

2 television is now replete with small niche channels. Channels specializing in food, home and gardens, sports, or health are now available and beamed straight into living rooms. The Internet is an increasingly popular method for patients to seek information about health. Nearly one quarter of all Internet search engine requests revolve around health-information seeking queries (Ferguson, 1997). According to a recent Pew Internet and American Life Project study (Pew Research Center, 2000a), more than 52 million Americans, or 55% of Americans with Internet access, have used the Web to investigate health matters. The U. S. Census Bureau estimates that the number of people globally using the Internet to seek health care information will rise from 95.4 million in 1999 to more than 350 million by 2003 (Harris & Associates, 1999). People are not just turning to the Web for general library-type resources: they also are using the information they find to make specific health decisions, including self-diagnosis, whether or not to visit a doctor, and what specific facility to visit (Pew Research Center, 2000a). The research shows that people use the Internet to find information on staying healthy as well as to obtain more information about existing conditions. Other now-traditional vehicles of communication have been studied in regard to the extent of their effectiveness in delivering health messages (Hawkins et al., 1987; Napoli, 2001; Hofstetter et al., 1992; Witte et al., 1993; Izenburg & Lieberman, 1998b; Hofstetter et al., 1992). In addition to the effectiveness of media for health messages, culture is another variable. Steers et al. (1996) investigated this premise based on the belief that the relationship between health behaviors and beliefs differs for people of different ethnicity. Now that more people are turning to the Internet as a first source of health information, it is evident that study of message effectiveness must be done to explore this new

PAGE 15

3 international medium. The research in this thesis seeks to investigate message delivery vehicles and cultural responses to the message. It examines different approaches to a health message and its delivery in order to determine the most effective means of conveying certain health messages to specific cultures. This research is important because it tests whether people subscribe to their traditional culture-types while on the Internet or if they abandon their cultural ways for an “Internet culture.”

PAGE 16

CHAPTER 2 REVIEW OF LITERATURE This literature review will discuss the Health Belief Model (HBM; Rosenstock, 1974; Janz & Becker, 1984), the body of research on the HBM, the theoretical assumptions and limitations, and theories complementary to the HBM. Additionally, this chapter will introduce the culture types used in this study and examine Internet research. Health Belief Model The HBM seeks to explain the link between exposure to persuasive health messages and behavior. Based on psychological and behavioral theory, the model hinges on the individual’s desire to be healthy, the personal value he or she places on a particular health goal, and what he or she thinks is the likelihood of achieving that goal (Janz & Becker, 1984). This model is well-researched and accepted as being effective in promoting preventive health behaviors (Janz & Becker, 1984). For the purposes of this study, the HBM and its constructs will be used to predict intended behavior change. The study will also test for the variations in effectiveness due to culturally targeted messages and level of interactivity. The model has five constructs: perceived severity, perceived susceptibility, benefits, barriers, and cues to action. To understand the theory, which is basically an additive one, one can think of these constructs as components of an arithmetic formula, where adding the degree of perceived susceptibility, severity, and cues and subtracting the difference between benefits and barriers should equal the likelihood of taking the recommended 4

PAGE 17

5 health behavior (Rosenstock, 1974). Many social science research studies have been conducted on the HBM, furthering and elaborating it. Perceived severity refers to the receiver’s beliefs about the seriousness of the health problem about which you are communicating. Severity becomes important to the individual as he or she faces the possibility of contracting a disease or the dangers of leaving it untreated (Janz & Becker, 1984). Research finds that messages describing and concentrating on the severity of a health problem are most effective when used for dire illnesses and diseases, such as AIDS (Povinelli et al., 1996). Perceived susceptibility refers to one’s beliefs about the likelihood of contracting or developing a health problem. The more likely the person believes he or she is to develop a health problem, the more effective the message will be in persuading its recipient to the desired health action or behavior. According to Janz and Becker (1984), “this dimension refers to one’s subjective perception of the risk of contracting a condition” (p.2). The benefits in the HBM refer to the perceived benefits an individual would enjoy if he or she participated in the positive health behavior. This construct lends itself well to preventative health messages, urging consumers to brush or floss their teeth regularly (Kegeles, 1963a). Research finds that the benefits variable is effective in predicting health behaviors when combined with other variables, especially susceptibility (Rosenstock, 1974). Barriers refers to the perception of potential barriers one will face when attempting to perform the healthy behavior. For example, it may be necessary to mention that certain health behaviors can only be performed by visiting a doctor, for which a person would have to make an appointment, take time off from work, have insurance or other means of

PAGE 18

6 payment, etc. Barriers that stand in the way of the positive health behavior are an important consideration for health communicators. Because the message recipient is likely to do a “cost-benefit” analysis (Janz & Becker, 1984), weighing the benefits over the barrier, communicators are charged with removing or diluting barriers as much as possible. Cues to action refer to various messages the individual receives interpersonally and through the media. For example, cues to action can come in many forms, from public service announcements to news media reports. Additionally, cues to action can come directly from the person’s physician as he or she makes health recommendations. While it is difficult to point out a direct causal relationship between message exposure and desired health behavior, cues to actions are considered very important because they offer reminders to the patient. When another variable is combined with cues to action, individuals are more likely to engage in health behaviors (Janz & Becker, 1984). HBM studies can be categorized into three health behavior categories: preventive-health behavior (PHB), sick-role behaviors (SRB), and clinic visits (Janz & Becker, 1984). PHB studies seek to explore the early motivations surrounding health screenings and other behaviors meant to discover or prevent disease or illness. PHB studies have shown that the roles of benefits and susceptibility are key in motivating such behaviors. SRB refers to the actions one takes after the diagnosis of a disease or condition. The SRB dimension also can refer to the actions taken in hopes of stalling development of an existing disease or condition. Lastly, clinic visits have been studied as behavior using the HBM.

PAGE 19

7 For the purposes of this study, this body of research will concentrate solely on PHB studies. This study will investigate intent to participate in positive PHB due to message exposure. Later research will be needed to build on the findings of this study by investigating actual behaviors, such as clinic visits, resulting from such PHB messages as discussed in this study. History of the Health Belief Model The HBM (Rosenstock, 1974; Janz & Becker, 1984) was originally devised in the 1950s to explain how health educators could encourage preventive health behavior and health screenings for tuberculosis (Rosenstock, 1974). A group of U.S. Public Health Service social scientists developed the model to explain the trends of masses failing to participate in early health screening behaviors (Janz & Becker, 1984). Hochbaum performed the first HBM study, which dealt with X-rays to detect tuberculosis in 1952 (Hochbaum, 1958; Rosenstock, 1974). Hochbaum investigated the effects of perceived susceptibility and benefits of early detection for tuberculosis. Hockbaum reported that perceived susceptibility to the disease was the greatest motivation for obtaining health screenings (1958). This research laid the groundwork for two of the later five constructs that make up the HBM. The next formative HBM study concentrated on perceived susceptibility and severity, benefits, and barriers to preventive dental care (Kegeles, 1963a; Rosenstock, 1974). While Kegeles’ findings have been criticized because of the extremely small sample size, high drop-out rate, and the availability of free or inexpensive dental care for the study (Rosenstock, 1974), the investigation found a positive correlation between preventive dental visits and number of supportive HBM construct beliefs. A later study (Kegeles,

PAGE 20

8 1963b) found that perceived severity and benefits could not predict subsequent preventive dental care visits. A 1963 study (Rosenstock, 1974) found that a combination of HBM variables improved the prediction power of the model. For example, perceived benefits of a positive health behavior alone may not be enough to persuade a message recipient to action. Perceived benefits added to perceived susceptibility increase the prediction power in the HBM. Adding another variable to that same example, such as low barriers, increases the message’s effectiveness (Rosenstock, 1974; Janz & Becker, 1984; Steers et al., 1996). While the contradiction of these findings to those of Kegeles (1963b) was not formally addressed in the discussion of the research findings, some researchers attribute the success of the earlier studies to the presence of media information or patient education materials. This idea of external information later became the final construct of the HBM: cues to action. Since its formation, the HBM has enabled social scientists to identify the most effective ways to frame positive health messages. Using its five constructs, the HBM has been found effective in developing health messages concerning testicular cancer education (Wohl & Kane, 1997), breast cancer screening (Fulton et al., 1991), reduction in cholesterol intake (McIntosh, 1996), and AIDS (Bakker et al., 1997), among others. Preventive Health Behavior Research Preventive-health behavior research investigates the effectiveness of messages that encourage people to take proactive, preventative measures to safeguard health. Preventive health can come in the form of clinic visits, vaccinations, or taking vitamins. The purpose of a preventative health message is to inform and persuade someone what

PAGE 21

9 they can do to keep themselves from contracting the medical condition mentioned in the message. Breast self-examinations (BSE) are key screening functions that enable women to discover a possibly cancerous growth. Research regarding BSEs found that when health communicators explain the benefits of BSEs and the susceptibility of the cancerous tumors, women are significantly more likely to undertake BSEs (Hallal, 1982). In the past two decades, much research has been conducted to determine effective means of motivating at-risk people to be tested for HIV. One study found that homosexual and bisexual youths were most motivated for the screening by decreasing barriers to the testing and increasing cues to action (Povinelli et al., 1996). A folate and grain intake study was conducted among indigent pregnant, minority women. The study found that when the mothers-to-be were advised on the importance of folate in neonatal development, they increased their folate intake level (Kloeblen, 1999). This data supports the HBM, in that the perception of both the benefits of having a healthy baby and reducing the barriers (such as cost by distributing free vitamins) motivated women to engage in PHB. A 1996 study using the HBM found benefits, costs, and physician advice the most effective in predicting reductions in fat and cholesterol intake (McIntosh et al., 1996). Findings suggest that cues to action, especially from a respected source such as a physician, play an important role in encouraging PHB. Researchers thus suggest that to increase the immediacy, physicians should pair effective cues and advice with the delivery of cholesterol results.

PAGE 22

10 Assumptions of the Health Belief Model The HBM’s main assumption is that the people receiving the health message are motivated and interested in the message and its content (Rosenstock, 1974). 1) People who do not place value on their health are not predicted to engage in PHB, according to the model. 2) The individual must place merit or value on the message and 3) the goal expressed in the message, which determines his or her likeliness to engage in PHB. 4) An important element is that the health message must be meaningful and 5) persuasive. 6) The individual must believe that engaging in the PHB will allow him or her to reach the health outcome or goal (McIntosh et al., 1996). An additional assumption is that the PHB message reaches the recipient. For an effect to be found due directly to the message, the message must be properly transmitted and comprehended. Thus, if the message were to come in the form of a television public service announcement and the power were to go out when the spot airs, PHB cannot be related to the televised message because the message would not have been received. Message delivery can come in many forms, such as through comprehension of posters, fliers, television or radio public service announcements, billboards, and other media delivery methods. Complementary Theories Cognitive dissonance theory (Festinger, 1957) provides an additional framework of support for the HBM. The theory of cognitive dissonance posits that when people are faced with a belief that runs contrary to their action or behavior, mental dissonance arises. This dissonance becomes so unpleasant that a choice must be made in favor of one or the other in order to reduce the dissonance. Research indicates that most people chose to alter their beliefs to restore consistency with behaviors.

PAGE 23

11 According to the theory of cognitive dissonance, people seek out information in which they are interested (Festinger, 1957). As previously mentioned, the HBM assumes that recipients are interested in the message in order for it to be effective. It is easy to apply both of these theoretical frameworks to Internet information seeking because a majority of people who use the Internet will use search engines or other mechanisms to actively search for specific information, when interested in a topic (Pew Research Center, 2000a). This type of motivation means that most people will not seek information that does not benefit them or if there is not any motivation for cognition. Festinger (1957) finds this important because if there is no perceived future need for information, then the information seeker will disregard the opportunity to engage in the process of gaining knowledge. After the subject has sought and gained new information, dissonance occurs if two clusters of information clash (Festinger, 1957). However, if competing ideas existed and the information was sought in order to clear up confusion, the new information gained will validate one of the earlier beliefs. Applying this idea to health communication, a health-information-seeker would continue to seek out answers to validate pre-determined ideas about health behaviors. Thus, when exposed to conflicting health messages the health-information-seeker is likely to adopt the behavior that becomes validated. This validation occurs at the first exposure to one of the originally conflicting information. Limitations of the Health Belief Model One of most important limitations of the HBM is the assumption that all people who are exposed to preventive health messages have equal opportunities, resources, and barriers to act upon the health message (Rosenstock, 1974).

PAGE 24

12 Along these lines, researchers have found that while the HBM correctly predicts health behavior for European-Americans, it is not as successful for some other ethnic groups. In a study devised to ascertain preventive HIV behaviors in college students, Steers et al. (1996) found that the HBM model did not correctly predict PHB in Asian-American, African-American, or Hispanic-American populations. The body of research concluded that the HBM is effective in predicting European-American behavior because the culture is concentrated on the self, whereas members of the other minority cultures often associate themselves with their sub-culture (Steers et al., 1996) or ingroup. Thus, messages that promote health for the good of oneself are not as effective for these cultures as messages that would promote personal health for the good of the group. The researchers went on to say, “health beliefs that concern the individual self (e.g., susceptibility) may be more likely to influence the behavior of Euro-Americans than those of other groups, whereas health beliefs that concern a group may be more likely to influence other ethnic groups’ behaviors than Euro-Americans” (p. 108). These findings could be attributed to the lack of cultural appropriateness for minority groups as the researcher suggested, or that the messages simply didn’t resonate with the individuals. These cultural differences are well-known and accounted for in other information dissemination campaigns, such as the 2000 United States Census (Sutten, 2000). In a campaign to create effective action cues, the government devised a series of slogans targeted toward specific cultural groups. A Caucasian-targeted slogan of “This is your future. Don’t leave it blank” was changed to an African-American targeted slogan of “This is our future. Don’t leave it blank” or the Native American slogan of “Generations are counting on this. Don’t leave it blank” (Sutten, 2000). This campaign resulted in

PAGE 25

13 higher participation of minority non-civic-minded citizens in the 2000 national Census than in previous Censuses (Sutten, 2000). While decades of research have found that the HBM is quite effective in predicting health behaviors related to specific diseases and illnesses, it is less useful when applied to vague health behaviors, such as annual physical examinations (Janz & Becker, 1984). Without specific and immediate calls to action, the person exposed to the health message does not have the opportunity to receive immediate gratification. Alternate Theories to the Health Belief Model Social Learning The theory of social learning suggests that people learn behaviors directly from observing the behavior of others. Thus, social learning provides cues to action, so that one might say that people learn a PHB by seeing others exhibit the behavior. Lindsay and Rainey (1997) posit that for adolescents, the act of smoking predisposes them to become more accepting of drugs, resulting in later use. Other researchers find that cigarette smoking involves a number of skills that provide an opportunity to embrace the wrong behaviors with other substances (Flemings et al., 1989). For example, a young smoker becomes used to the harshness of tobacco smoke, the feeling of a cigarette in his or her hands, and the nausea associated with the first cigarette smoking experiences making the first use of marijuana a familiar experience. The researchers posit that these experiences provide a “gateway” into the use of more serious drugs. Gainand Loss-Framed Messages Gainand loss-framed messages are very similar to the HBM’s constructs of benefits and barriers/severity. Loss-framed messages are effective in conveying the ideas of PHB for screening or disease-detection results and gain-framed behaviors are most effective

PAGE 26

14 for promoting healthy prevention-type behaviors (Rothman et al., 1999). In a study devised to test the effectiveness of gainand loss-framed messages, researchers created a fictitious disease and supplemental patient education material (Rothman et al., 1999). Results indicated that the participants who were exposed to the loss-frame message, communicating the costs and dangers of not participating in the PHB, reported that they were more likely to exhibit PHB and be screened than those exposed to the gain-framed message (Rothman et al., 1999). Other message framing studies have investigated the ideas of negative or loss-framed messages. One study, based on prospect theory, investigated the idea that people will choose a risky (but certain) behavior and subsequent outcome over an uncertain behavior but possibly positive outcome (Rothman et al., 1993). Results indicated that negatively framed health messages cause concern about one’s present health status (Rothman et al., 1993). According to the authors’ interpretation of prospect theory, PHBs, such as disease screenings, are risky in that the patient could discover that he or she has the ailment; thus, negatively framed messages are more effective for promoting such PHBs as BSE. Cultural Types This study focuses on the differences between two well-researched cultural groups: collective and individualistic cultures. An individualistic culture is one in which people place high value on the self. Members of collective culture place high value on the group to which they belong. Individualistic cultures are common in Britain and British-influenced cultures such as the United States and Australia (Hofstede, 1980). Collective cultures are predominant in Africa, Latin America, Asia, Southern Italy; and with traditional Greek, Chinese, and Far Eastern minorities; and among people from these subcultures who now reside in the United States (Bontempo et al., 1990). A list of

PAGE 27

15 attributes (see Appendix A) was developed to identify the attitudinal differences between the two culture types (Trinadis, McCusker, & Hui, 1990). Research finds that people from individualistic cultures make decisions based on evaluations of personal goals (Bontempo et al., 1990). Those from collective cultures place the highest values on the ingroup, and tend to make decisions based on what is best for the group. Research finds that members of collective cultures treat others in the ingroup with more intimacy and have higher levels of self-disclosure than those from individualistic cultures (Triandis et al., 1990). Because of this, individualists have a lower ability to deal with intimate encounters than collectivists but are more skilled at entering and leaving ingroups than collectivists (Triandis et al., 1990). Within these two cultural types lie the “parallel variations” (Bontempo et al., 1990) of allocentric and idiocentric individuals. Allocentric people are more in line with the collective culture and “pay more attention to ingroups than to their personal goals” (p. 202); and idiocentric individuals are more closely aligned to individualistic ideals and “emphasize personal goals over loyalty to ingroups” (p. 202). According to Bontempo et al. (1990), idiocentric people 1. Have a higher need for achievement 2. Are more likely to report feeling lonely 3. Place high value on competition According to Henry and Stephens (1977), idiocentric people: 1. Weigh advantages and disadvantages to the ingroup when making important prosocial decisions 2. Are more likely to need health care because of the effects of high stress levels

PAGE 28

16 3. According to Bontempo et al. (1990), allocentric people: 4. Tend to have happier marriages 5. Deal with life changes/stress better because of the social support systems which are in place in their lives 6. Safeguard their health 7. Place less value on competition or on individual fame 8. Are persistent to achieve goals against adverse conditions when it is for the good of the ingroup 9. Highly value cooperation, equality, honesty, and self-sacrifice for the good of the ingroup 10. Respond automatically in favor of their ingroups when faced with making decisions that call upon beliefs and values In a study to test prosocial behaviors and behavioral intentions of idiocentric and allocentric individuals, researchers found that, regardless of the level of anonymity, allocentric individuals seem to make the same prosocial choices (Bontempo et al., 1990). Idiocentric individuals tended to stray from prosocial, ingroup behaviors as the subject’s identity became masked. When anonymous, results indicated that idiocentric individuals felt free to behave in a manner that was best suited for them, as opposed to being tied down by social pressures. This finding is in line with the idea that allocentric individuals do not choose what is best for themselves but consistently and automatically make decisions and act so as to benefit the ingroup (Bontempo et al., 1990). Idiocentric individuals will choose behaviors that are most beneficial to themselves, regardless of the benefits or costs to the ingroup.

PAGE 29

17 Some minorities feel at greater risk to public health concerns than Caucasians. Research finds that American collectivist minorities are more likely to feel susceptible and “at risk” for levels of lead in their water than Caucasians (Griffin & Dunwody, 2000). The study reviewed a public service health information campaign regarding lead in tap water in Minneapolis and investigated knowledge gap hypotheses. This could be due to the fact that Caucasians were more likely to report knowing the preventative measures to be taken to ensure safe drinking water, resulting in a lower feeling of perceived susceptibility. Both idiocentric and allocentric groups use the Internet as a resource for information and communication purposes. Later in this paper, the Internet usage habits of each group will be discussed in depth. However, first it is important to look at the Internet as a health information resource and understand in what ways the Internet is being used. The Internet as a Health Information Source The Internet is an increasingly popular source for information on a variety of topics, including information on health. A recent study estimated that in 1998 more than 800 hospitals hosted Web sites and the content of more than 25,000 Web sites concerned health (Izenberg & Lieberman, 1998a). According to Mittman and Cain (2001), the Internet is conducive to informing or educating the general public about its health because the Internet is 1) inexpensive, 2) easy to use, 3) democratic, meaning there is equal opportunity for diverse messages, 4) long distance, meaning health care information seekers can acquire information across large geographical distances, and 5) increasingly functional, meaning the technology is growing and adapting to allow for more sophisticated delivery of messages through new coding programs, such as Java.

PAGE 30

18 Age, education, and income also play a predictive role in the likelihood of someone to use the Web to research health issues. According to a recent U.S. Department of Commerce report (2002), people aged 55 or older are more likely to use the Internet as a resource for health information, there is a correlation of household Internet usage with the level of education reached by a member of the household, and people who earn more than $75,000 a year were more likely than any other income group to use the Internet to seek health information. Internet Source Credibility Internet source credibility is an important issue for Web surfers. Just as Mittman and Cain (2001) point out democracy as a benefit to seeking health care information, meaning that all groups with diverse messages have an opportunity to post messages on the Internet, that benefit also serves as a barrier because it can be difficult for the novice user to distinguish between a credible source and non-credible one. This credibility issue makes source branding more important on the Internet (Mittman & Cain, 2001). A 1998 study reported that only 48% of the medical resources reviewed were hosted by a credible source (Izenburg & Lieberman, 1998b). This alarming fact has not been lost on the public. Fears of finding misinformation and bad health advice on the Internet abound (Napoli, 2001). According to Mittman and Cain (2001), the Internet “at its best can equip consumers to lead healthier lifestyles, detect potential medical problems early, work more collaboratively with physicians to treat illness, and learn of effective treatments to which a local provider may not have access” (p. 57). Researchers also warn, “at its worst, however, the information can undermine health objectives” (p. 57). As information seekers uncover information, they must decipher the credibility of the source. For a message to have maximum effectiveness, the recipient must trust the

PAGE 31

19 credibility of the message and its source (Johnson & Meischke, 1993). To address the issue of credibility of health information online, many measures have been developed by health organizations. There are online health information rating/grading systems available through uninterested third parties (Mittman & Cain, 2001). Source expertise is found to increase credibility in the eyes of the online health seeker (Eastman, 2001). When online health information seekers can find information from a reputable health care organization, the organization’s branding lends credibility to the message (Mittman and Cain, 2001). Researchers argue that as health information flourishes on the Internet, consumers will seek out well-known brands for health information. Complementary reports indicate that people are able to distinguish between reputable health information sources on the Internet and those that are not (Widman & Tong, 1997), but more research must be done in this area to fully substantiate this claim. A recent exploration into the source credibility of health information on the Internet yielded results that indicated that regardless of the level of familiarity of the information, highly credible sources increased the effect of the health message (Eastin, 2001). The study tested three different levels of source credibility, varying from a doctor, to a widow of a diseased victim, to a high school student. While the hypothesis of the study was not supported, important findings resulted from the study. Eastin (2001) found that there was a difference between message recipients’ review of the information/effect of message if the topic was something with which they were not familiar and the information was attributed to a credible source. It appears that when learning something new from a source the recipient finds credible, the online health message has a greater effect on the recipient. Additionally, Eastin (2001) reported that the variance in credibility could result

PAGE 32

20 from a number of unexplored variables, such as the page layout and design or the dynamics or interactivity. Barriers to Using the Internet as a Health Information Source The main barriers to relying on health information delivered through the Internet are: 1) Source credibility, 2) Security of transmitting personal health data, and 3) Single message for diverse audience. Determining source credibility remains one of the biggest obstacles Internet health-information seekers face (Napoli, 2001). This is because Web page publishing is so simple and cheap that anyone can create a Web site. In order to combat this barrier, some health information Web sites have begun posting seals of approval indicating that the content on the has been reviewed and rated as credible (Aspden et al., 2001). Health information always has been regarded as confidential and private to the patient and the organization administering care. As Mittman and Cain (2001) point out, ensuring that electronic transmission of personal health data is done securely remains a priority to organizations and patients. The final barrier, the message itself, remains an important issue to communicators. The Internet reaches a limitless international audience, and the difficulties of crafting one single message that will be equally effective across each culture are challenging. Health communicators who attempt to tailor messages to specific audiences may find this form of mass communication difficult to effectively manage. Gathering health information over the Internet has definite benefits in that it can be anonymous. This increased privacy becomes key because security concerns are paramount when it comes to transmitting health information via the Internet. However, some advanced health information seekers may be interested in interacting with a message source. Some users may be interested in receiving personalized medical advice,

PAGE 33

21 social support, etc. There are a number of reasons why an online health information seeker would share personal medical information/data with others via the Internet. For health care institutions, the security of this information as it is being transmitted is paramount (Mittman & Cain, 2001). Organizations, such as hospitals with Web sites, are making strides to protect their health care consumers by implementing secure, password-protected databases and firewalls, and ensuring the confidentially of Internet messages transmitted to them (Mittman & Cain, 2001). While there is still room for improvement, the industry is setting standards for information security which protect the patient and the health care organization. Typical Internet Health-Seekers For the purpose of this study, a health information seeker is one who purposely seeks out information about a health condition, disease, or topic. The information seeker desires to fill a void of knowledge and typically has a variety of channels he or she may use to find answers to his or her questions. Mittman and Cain (2001) found security, information quality and information quantity to be important issues across demographic groups. Research concerning how people deal with gaps in knowledge shows that people turn to the most convenient or closest information source (Dervin, 1989). After obtaining initial health information, people typically continue to search for additional sources. For example, pregnant women may rely on information from their doctor, husband, mother, or on television and print sources (Daniels & Parrott, 1996; Kyrouz et al., 1998), but do not rely on a single source to fill in knowledge gaps (DePietro & Clark, 1984). This method allows the information seeker to assign a level of credibility to the information gained, as he or she can compare the various pieces of information against one another for validation.

PAGE 34

22 The HBM assumption that PHB message-seekers are interested in and actively seek health information is believed to hold true for the typical Internet health information seeker. Most people accessing health information on the Internet do so through direct connection to reputable medical Web sites or through Internet-wide searches on a particular health issue (Pew Research Center, 2000a; Pezza, 1990). Of the health information seekers, those who have an “uncertainty orientation,” meaning they actively seek information to clarify questions they have about themselves, use the Internet to seek new information. Those with a “certainty orientation” seek only information that reinforces their previously held beliefs. This is an element of cognitive dissonance (Festinger, 1954) where people shut out information that would cause a cognitive rift in existing beliefs. This is important because it suggests that those seeking health information on the Internet have a purpose and are actively pursuing the information, as opposed to just discovering it unexpectedly. Based on the data about the typical Internet health-seeker from the Pew Internet and American Life research (Pew Research Center, 2000a), people are turning to the Web to seek out information, as opposed to just falling upon it in cyberspace. This is an important distinction because it means that these health-seekers are already motivated and primed to be persuaded by the messages. Thus, the assumption of motivation in the HBM is met. Eighty-three percent of online health-seekers believe it is important to get information from the Internet that they cannot get from other sources (Pew Research Center, 2000a). Because health seeking can be done relatively anonymously online, patients are able to fully investigate a variety of health topics they may be embarrassed about discussing with a physician or other reputable resource.

PAGE 35

23 Health topics and Internet/technology user level vary among online health information seekers. Similar to conventional health information seeking, women are more likely to log on to the Internet to research a health topic (Pew Research Center, 2000a; Pew Research Center, 2000b). While the Pew Internet and American Life research revealed that most people are worried about the credibility of health information on the Internet (Pew Research Center, 2000a), other sources claim that novice users blindly accept the expertise of Internet messages (Jiang, 2000). Research shows that messages on the Internet must be treated differently than traditional print messages (Nielsen, 2000). Web usability guru Jakob Nielsen found that people do not treat traditional print copy (such as pamphlets, brochures, ads, books, etc.) as they do words on a computer screen. Reading text from computer monitors slows down the reading process by 25 percent (Nielsen, 2000). Additionally, the Internet has become known to provide instant information-gratification. Thus, users become impatient more quickly and are accustomed to quickly obtaining information and clicking in search of the next page of information. Nielsen’s research suggests that people scan through the text, looking for bold words or words they associate with themselves (2000). Thus, many effective Web-based messages are formatted in simple, short sentences, use bolding features for key words, and use many bulleted lists. Nielsen tested a series of message wording techniques to find the most effective form of written communication for the Internet (2000). Taking a public relations-type tourism brochure that included information about a town, its history, what to do when visiting, and lodging information, Nielsen re-arranged the wording of

PAGE 36

24 the text several times to find the most effective writing style. The research found that readers were often slowed by page length, layout, and exaggerating verbiage and descriptions. Nielsen assigned the original brochure text as a baseline with zero percent readability. Through a combination of changes to the text, Nielsen was able to improve the readability and reader comprehension level. When the message was condensed to concise text, displayed in a scalable layout, and edited to include neutral wording, readability increased by 124 percent (Nielsen, 2000). Nielsen’s research, although not based specifically on Internet health information, is important because it reveals the way online information seekers use the medium, how to maximize messages through word choice, and how best to present messages in the Web page design. Allocentric Web Surfers For the purposes of this study, the following cultures and their American counterparts are assumed “allocentric”: African, Asian, Hispanic, East Asian, and Middle Eastern (Triandis et al., 1986; Trinadis, McCusker, & Hui, 1990). It is assumed that due to the collectivist nature of these cultures and their tendency to associate themselves with the ingroup, these cultures are allocentric; this is not to say that all people from these cultures and subcultures are allocentric, but that for the most part they display allocentric tendencies. Differences in Internet use by culture and ethnicity is found in a recent report published by the Pew Internet and American Life Project where the Web-use habits of African-Americans were explored (Pew Research Center, 2000b). According to the report, most African-Americans are going online to search for information and learn something new. African-Americans are more apt to investigate major life issues (find a job, place a to live, etc.) online than Caucasians (Pew Research Center, 2000b).

PAGE 37

25 The Pew report speculates that the differences in Web-use habits between these two sub-populations occur for a variety of reasons. Caucasians have been using the Internet longer and are therefore more comfortable with the technology (Pew Research Center, 2000b). Because Caucasians have been on the Internet longer, it is a more integrated part of their life; subsequently, it can be inferred that Caucasians could be more likely to turn to the Internet for health information due to this higher comfort level (Pew Research Center, 2000b). The Pew research indicates that African-Americans tend to surf the Internet at work as opposed to Caucasians who access the Internet from home (Pew Research Center, 2000b). Home use and work use of the Internet can result in two different types surfing patterns (Pew Research Center, 2000b). For instance, the workplace is more likely to have a fast connection, which could explain why African-Americans report a higher consumption of online media and music. Caucasians are more likely to use the Internet for recreational gaming purposes, a function more easily performed at home rather than at work (Pew Research Center, 2000b). Additionally, Caucasians are more likely than African-Americans to go online after working hours (Pew Research Center, 2000b). Based on this Pew Internet data, it can be inferred that African-Americans would be less likely to investigate personal health issues as opposed to Caucasians who can seek health information without disruption in the comfort of their own homes, while African-Americans may have to do so at work. Consistent with the characteristics of a collective culture, African-Americans report using the Internet as a means to keep in contact with friends and extended family (Pew Research Center, 2000b). While both African-Americans and Caucasians enjoy electronic mail about equally, the reasoning behind the enjoyment differs. According to

PAGE 38

26 the Pew study, more than half of the African-Americans studied reported using electronic mail because it is quick and easy, while Caucasians report that the main reason to use electronic mail is the convenience of sending messages whenever they want. When looking for health information, African-American females are more likely than African-American males to use the Internet as a research tool (Pew Research Center, 2000b). This finding is consistent with other subcultures, such as European-Americans (Pew Research Center, 2000b). However, when comparing health information seeking between the two races, African-American women are significantly more likely than Caucasian women to use the Internet to find health information (Pew Research Center, 2000b). This finding could be representative of the economic disparity between the two groups, Caucasians in the United States report a higher household income than African-Americans (Pew Research Center, 2000b), and thus are more likely to have direct access to good health insurance and health care treatment. Asian-American Internet users are quite different than other American Web users (Pew Research Center, 2001). According to the Pew report, Asian-Americans have been online longer (in terms of years), they use the Web more in their daily life, they stay on the Web longer each time they access the Internet, and their gender usage patterns are quite different from their allocentric peers, such as African-Americans and Hispanics. Interestingly enough, while women are typically the online health information seekers in most American cultural groups (Pew Research Center, 2000a), this is not the case with Asian-Americans; in this culture group, Asian-American men are twice as more likely to seek out such health information on the Web (Pew Research Center, 2001).

PAGE 39

27 Idiocentric Web Surfers A growing number of Caucasians have been using the Internet for nearly a decade and thus are very familiar with the many uses of the technology (Pew Research Center, 2000b). Because this sub-population has been using the Internet for longer and is considered Internet veterans, its Web-use habits differ from those of novice users. As experience with the Internet grows, users are more likely to use it more often and are open to exploring new sites and information on the Internet (Pew Research Center, 2000b). Caucasians’ response to questions regarding Web usage indicate that the technology is well-integrated into their lives, and 73% of users reported that they would miss the Internet if they no longer could access it, as opposed to only 62% of African-American users who reported the same sentiment (Pew Research Center, 2000b). Due to this mainstreaming into their daily lives, Caucasians are much less skeptical of the Internet. Caucasians report divulging a greater amount of personal and financial information about themselves via the Internet to obtain objects or information (Pew Research Center, 2000b). African-Americans, on the other hand, are much less likely to input such information in exchange for access to information. In the Pew study, these differences are assumed based on experience with the Internet. However, other research finds these differences in desired delivery and level of personalization varies across ethnicity. Internet Compared to Other Media In the past decade, the Internet has grown exponentially. Since its inception, the Internet has been compared to various other media. Vehicles for mediated communication have different characteristics and benefits (Hawkins et al., 1987; Napoli, 2001). While early research notes where the Internet falls short in comparison to other

PAGE 40

28 communication mediums, this researcher proposes that in basic PHB communication, the Internet can excel over other media because of the perceived anonymity and the information motivation that drives Internet health information searches. Additionally, as technology grows, the possibilities of improving Internet-delivered messages can increase as well. The advantages of Internet-delivered health messages are that: 1. Messages can be delivered at any time, day or night. 2. The message recipient is in complete control of delivery time. 3. The message can be delivered anywhere, including specialized airport Internet terminals, across the globe in cyber-cafes, in public libraries, on personal digital assistants, Web-enabled cellular phones, and many more convenient locations. 4. Health care information seekers on the Internet are in charge of their own message destiny. While message recipients have many suggested options on health care Web sites, they are truly in control in that they may chart their own path in seeking the messages most appropriate to the information goal. 5. Research shows that Internet health information seekers are genuinely interested in the information and thus more susceptible to the persuasiveness in the message. 6. In interpersonal communication, the message carrier can have extremely high credibility. 7. Non-verbal cues can be interpreted more easily from high-context individuals. 8. Interpersonal communication in this form allows for a more emotional and 9. Personal experience. Print health messages carry a high degree of credibility and are often cited as sources for health information by the public (Hofstetter et al., 1992) as opposed to passive media. Passive media are those that one can take in inactively, without requiring much mental

PAGE 41

29 exertion. Assuming the reader has already decided to pay attention to the message, print messages force readers to expend more cognitive effort to understand the message due to the nature of reading (Chaiken & Eagly, 1974). Research Gaps The HBM is well-researched and supported (Rosenstock, 1974; Janz & Becker, 1984). Research finds that health messages can be powerfully and effectively disseminated to the masses through such mass media mediums as television (Chew, 1998). The Internet can be available to many more households than television broadcasts because it crosses country and region boundaries. According to Tardy and Hale (1998), there has been little research on the topic of the significance of health information gathering. While the researchers note that some studies have been done on this, they criticize colleagues for lumping information seeking with other social support interactions. The researchers investigated the importance health information-seeking has on social support and the idea that such motivations lead to the formation of new social networks (Tardy & Hale, 1998). This idea that people are looking to new channels, such as the Internet, to fulfill not only informational but social needs opens the door to a new field of study. Building on the diffusion of innovations theory (Rodgers, 1983), the researchers identified a pathway of how health information seekers build support systems. This pathway is made of information from many different sources and leads to a more informed person as all of the information is synthesized. Source credibility is another extremely important variable in regard to messages distributed on the Internet. While anecdotal evidence supports the idea that people are more willing to trust known medical institutions (such as hospital Web pages or

PAGE 42

30 WebMD.com), the effectiveness of Internet PHB messages has not been studied on sources generally low in credibility (such as a personal Web page). Based on the literature reviewed and the existing body of research, this researcher concludes that a study investigating the role of culture and its implications on interactively-delivered health messages is important. Research shows that select ethnic groups are more likely to prefer traditional means of communication for health messages. For instance, African-Americans are weary about Internet security and exchanging personalized information via the Internet (Pew Research Center, 2000b). Conversely, Caucasians are more comfortable with and prefer such tailored information (Pew Research Center, 2000b). Thus, this study will use two messages, one written for idiocentrics and one written for allocentrics and expose them to a health message in an attempt to persuade them to behavior. The messages will be delivered in an interactive (Web page) format and a static (brochure) format. The study will investigate if a matching message or a matching level of interactivity, based on assumptions made from previous research, hold true for the Health Belief Model in this Interactive Age. These assumptions about “matching” and compatibility are based on previous research regarding cultures (Triandis et al., 1986) and Internet use (Pew Research Center, 2000b). The following hypotheses are proposed: H1: Respondents who receive culturally matched messages are more likely to adopt the suggested PHB. H2: Respondents who receive messages in a culturally matched level of interactivity are more likely to adopt the suggested PHB.

PAGE 43

CHAPTER 3 METHODS By using the heavily researched Health Belief Model (HBM; Rosenstock, 1974; Janz & Becker, 1984) as a basis for this study, this body of research can investigate the power and effectiveness of Web-based health messages across different cultures. Additionally, it is important to review a culture’s reaction to interactivity in regard to personalizing health messages. A review of relevant HBM literature has not found a single piece of published research investigating the power of the Internet using the predictions of the HBM for online health-seekers. There is very little data on PHB message tailored for maximized effectiveness across culture. Variables Several conditions and assumptions were held to be true for this research study. Because the HBM has been widely investigated in the past three decades, the assumptions and predictions of the HBM were held to be true. In particular, the research finding that messages promoting the perceived benefits and perceived susceptibility of a health condition are most effective in producing a positive health behavior change for preventive health behavior (PHB) messages is held to be true. Message source was also held constant. All of the messages used in this study were presented as informational health pieces from a reputable academic health center in the Southeast United States. Participants of the study had basic knowledge of the teaching hospital from which the health messages were presented as having originated. This is important, considering that source credibility is such an important issue to online health31

PAGE 44

32 information seekers (Pew Research Center, 2000a). Thus, it was assumed that source credibility would be high enough for all participants to not interfere with the results of this study. The following hypotheses are proposed: H1: Respondents who receive culturally matched messages are more likely to adopt the suggested PHB. H2: Respondents who receive messages in a culturally matched level of interactivity are more likely to adopt the suggested PHB. Independent Variables As the research hypotheses point out, the two independent variables are level of interactivity and cultural message. Interactivity is broken down into two levels: interactive and static. Interactive (computer-mediated) preventive health messages are messages accessed through a computer connected to an Internet Web page. A “Web page” refers to information found on the Internet, formatted through hypertext markup language or other such Internet programming language. Internet Web pages are searchable, contain low-resolution reproductions of photographs, and often contain hyperlinks allowing the user to explore a term or topic more fully. Much resaerch has investigated how different ethnic groups are using the Internet in general and how they use it for health information seeking (Pew Research Center, 2000b; Pew Research Center, 2000a). Caucasians would be more likely to be comfortable with the interactively-delivered message. The message became interactive at the end of the text where participants had an opportunity to choose (“Yes, I want to be contacted” or “No, I don’t want to be contacted”) and submit their intended behavior electronically.

PAGE 45

33 Static preventive health messages involve no interactivity. This static preventive health message in this study is delivered through the form of a brochure. Based on earlier research of how different ethnic groups use the Internet for health information seeking, African-Americans and Hispanics would be more likely to be comfortable with the static message. Research shows that these ethics groups are not comfortable with submitting personal information through the Internet because of Internet security and trust issues (Pew Research Center, 2000a). At the end of the brochure, participants were able to check a box on the brochure regarded intended behavior (“Yes, I want to be contacted” or “No, I don’t want to be contacted”). Cultural message is broken down into two types: allocentric and idiocentric. Allocentric, or community-oriented collective cultures, place high value on the good of the group. Such cultures are altruistic and are often moved to action by a message that promotes the well being of the community. Historically, the HBM has not been effective in persuading this cultural group because most tested HBM messages are based on self-prosperity for health. The allocentric message centers on the dangers posed to the community, people like him or her, and the measures one could take to avoid contracting the disease. Self-oriented or idiocentric culture places value on the individual. Historically, the HBM has been proven most successful in persuading this cultural group to participate in positive health behaviors. The idiocentric message communicates the susceptibility to the recipient and identifies the reader directly through language using second person, denoting specificity (Parrot, 1995). By speaking directly to the message recipient and

PAGE 46

34 conveying the dangers of the disease, an idiocentric person is more likely to be persuaded to work toward the health goal. Dependent Variable The dependent variable for this study was intended behavior. A message was said to be “effective” when the respondent signs up to receive more information. Message Some health messages are better communicated through the constructs of the HBM. Health communicators traditionally find success in communicating goal-oriented behaviors, such as the importance of inoculations. This research study will use three HBM constructs (perceived susceptibility, perceived severity, and benefits) to create a health message. The health message topic for this study was West Nile Virus. West Nile Virus has become a growing health concern in recent summers (Rutledge, 2001). Since its introduction into the United States in 1999, reports of the virus have spread from New York City to the northeast states and later to Florida (Rutledge, 2001). The virus, mostly found in birds and horses, can be transmitted to humans by mosquitoes. West Nile Virus is dangerous if undetected because it can lead to encephalitis, or swelling of the brain. The virus was chosen as the health message subject for this research because perceived susceptibility to contracting the virus and perceived benefits of preventative measures can easily be communicated. The virus is not a condition that one cultural group would be more likely to contract and the perceived barriers to adopting the PHBs are low; also, the precautionary steps are easy to implement, as they do not require clinic visits or inhibit any normal behaviors. The message or stimulus used in this research discussed the West Nile Virus and indicated that University of Florida researchers were developing a human vaccine.

PAGE 47

35 Florida’s summer climate makes the state vulnerable to the disease because of the 1) many bodies of water such as swamps in which mosquitoes breed, 2) a large horse population, and 3) warm weather, which draws people outside during times of increased threat of contraction. There is a vaccine available for horses at this time, but not for humans. West Nile Virus is not discriminating, meaning that no cultural group is biologically or culturally pre-disposed for contraction, nor does preventive behavior compromise any cultural behaviors and/or beliefs. Symptoms of the virus include headache, fever, muscle spasms and mental degradation and can lead to encephalitis, i.e., swelling of the brain. Chances of contracting the virus can be reduced by not going outdoors during dusk when mosquito populations are high, wearing insect repellant with at least 30 percent DEET to ward off bug bites, reducing the presence of stagnant water where mosquitoes breed, and fixing loose windows and screens in the home to keep mosquitoes out. Prior to the time this study was conducted, there was a heightened media awareness of West Nile Virus. For the first time, the virus had been found in the state of Florida, and three people in the state were reported to have contracted the virus. This researcher felt this media attention aided in providing media cues to participants of the study. However, because of the media attention, the researcher had to take extra measures in order to ensure that the stimulus was indeed inducing behavior as opposed to previous messages. To do so, the preventive health message indicated that University of Florida researchers were developing a vaccine for humans. This ensured respondents’ first exposure to the information about the availability of such preventative measures which eliminated possible media cues to action that would influence their likelihood to sign up for the

PAGE 48

36 vaccination. This allowed this researcher the ability to measure actual behavior (signing up for the information), rather than behavioral intention (planning to wearing bug spray when outdoors). Immediately after exposure to the message, participants were given the opportunity to register for information about the West Nile Virus vaccine after it had been developed. At the end of the experiment, each participant was debriefed and told that University of Florida researchers were not working on such a vaccination, nor would they be contacted if one were developed for humans. Each participant was able to take the debrief flier home with them, which had contact information for the principle investigator, in case any questions or concerns arose after leaving the experiment. Using the basic principles of successful preventive health messages through the HBM, each message began by exposing the subject to a personal level of susceptibility. The first paragraph of each of the messages started with a series of questions that implied a level of risk and susceptibility if the subject answered “yes.” The messages then explained preventive measures, created a minimal level of barriers, and outlined the benefits to the PHB. Each message offered information about the vaccine in development and an opportunity to register to be contacted once it was completely developed. Based on previous literature, Parrott (1995) created a set of guidelines for producing quality health communication messages to increase the effectiveness of the message. The author suggests that the use of novel settings and media aid in effectively communicating the message. In the interactive condition, this research study used such novelty through a health-risk quiz that personalized the subject’s susceptibility to West Nile Virus. Both messages were written to incorporate spatial immediacy (Wiener & Mehrabian, 1968)

PAGE 49

37 through words such as “this” and “here” rather than “that” and “there,” temporal immediacy (Wiener & Mehrabian, 1968) through active verbs as opposed to passive language, and the deletion of qualifiers (Wiener & Mehrabian, 1968), or words such as “may” and “possibly” that leave room for doubt. Each message was accompanied with a form that respondents could fill out in order to sign up for information about the vaccine upon its completion. The form could be submitted electronically in the interactive group or by returning the filled out form to the study facilitator for the static group. Design Using a 2 X 2 factorial design (see Table 3-1), this study tested static and interactively delivered messages that were considered either a match or a non-match for the cultural group. The matched group received either a static or interactive message written for the highest effectiveness for that culture. For example, allocentric/collective messages were written emphasizing the ingroup and the benefits the ingroup would reap from adopting the proposed behaviors and receiving the vaccination. This set of messages was written to reach the community-oriented cultural group through the use of the third-person, a more formal writing style. The idiocentric group’s message emphasized individualism and the benefits one would incur from adopting the proposed behaviors of the message. This set of messages was written to reach the self-oriented cultural group by personally referring to the reader in the first and second person (words such as “I” and “you”). Each culture group also presented a non-matching message.

PAGE 50

38 Table 3-1. Research Design Variable – Person Culture Idiocentric | Allocentric | Message Idiocentric Allocentric | Idiocentric Allocentric ________________________________________________________________________ Variable 3 – | Level Interactive Match Unmatch | Unmatch Unmatch Static Unmatch Unmatch | Unmatch Match ________________________________________________________________________ The match and unmatch messages for each group (allocentric and idiocentric) were tested against the two levels of interactivity (static and interactive). The groups were tested on message effectiveness across these levels of interactivity. The static message was presented as a brochure on West Nile Virus, which appeared to be produced by Shands at the University of Florida. Research shows that groups who traditionally thought of as allocentric are apprehensive in interactive conditions (Pew Research Center, 2000a). Thus, the static condition would be a matched level of interactivity with allocentrics. The interactive message was presented as a Web page which appeared to be produced by Shands at the University of Florida. Research shows that groups who are traditionally thought of as idiocentric prefer interactive conditions and tailored messages based on personal data (Pew Research Center, 2000a). Thus, the interactive condition would be a matched level of interactivity with idiocentrics. Based on this, each participant had the opportunity to match or not match in two conditions: message and level of interactivity. It is possible for a participant to match in both condition, match in

PAGE 51

39 only one condition, or not match in any condition. The content of each of the idiocentric and allocentric messages was identical. Source Credibility Marketing data collected by the Shands at the University of Florida prior to this study aided in the identification and selection of this institution as a credible source of health information. The hospital’s market research found Shands at the University of Florida’s branding efforts resulted in 1) moderate association of quality health care without prompting, meaning that participants named Shands at the University of Florida first when asked to name a high-quality hospital, 2) high association of quality health care with prompting, meaning that participants chose Shands at the University of Florida when given choices of hospitals, 3) participants’ perception of academic medical centers or teaching hospitals is positive, and 4) the Shands HealthCare and the University of Florida branding efforts are nearly equally powerful. Each subject responded to a series of semantic differential questions weighing the authority of Shands at the University of Florida and its health care providers at the hospital. A modified version of McCroskey’s Source Credibility scale (McCroskey, 1966) measured the perceived credibility of Shands at the University of Florida. The instrument measures the subject’s perception of the source’s authority and character. This instrument is an often-used measure, as it has been used in more than 100 other studies to measure source credibility since its publication in 1966 (Rubin et al., 1994). The reliability of the variables measured by the scale varies from .85 to .98 alpha scores (Rubin et al., 1994). The validity of the scale is limited, as different studies have investigated a variety of constructs of source credibility (Rubin et al., 1994).

PAGE 52

40 Culture Culture was tested through the Family Integrity factor, a variation on a scale developed to measure individualism-collectivism, known as INDCOL (Triandis et al., 1986). The survey asks respondents to indicate their level of agreement using a 7-point Likert scale ranging from very strongly agree to very strongly disagree. The responses are coded to represent individual or collective orientation. The Family Integrity factor measures self-reliance, family integrity, interdependence, and distance from in-groups. The scale has proven reliable in previous research (Triandis et al., 1986; Triandis et al., 1990). Previous research suggests that the instrument finds collectivists rate higher on interdependence than individualists and lower on distance from in-groups (Triandis et al., 1990). Demographics Each participant was asked to complete a basic demographic and Internet use questionnaire and report their gender, race, age, year in college, reasons for Internet use, frequency of Internet use, and frequency of contact with a family member. The frequency of contact with family was measured in two levels: when at home and when at school. Recall and Retention Participants were tested as to comprehension and retention of information from the stimulus message. This measure included a manipulation check to ensure that the message, which the respondents were exposed to, appealed to the self (idiocentric) or others like the respondent (allocentric).

PAGE 53

41 Procedures All participants were briefed on their rights and responsibilities prior to participating in the study. Participants signed a consent form indicating that they agreed to participate in the study (see Appendix B). Participants first scored the source credibility for Shands at the University of Florida (see Appendix C). Then, participants completed the Family Integrity factor, which revealed their culture type (see Appendix D). Participants then were exposed to the preventive health message/stimulus (see Appendix E). After exposure to the message (presented in a matched or non-matched form across two levels of interactivity), participants were given the opportunity to sign up for the information on the vaccination after it has been developed. Next, participants were asked to fill out a basic questionnaire asking demographic and Internet usage information (see Appendix F). After completing the questionnaire, participants were measured for information retention concerning the stimulus (see Appendix G). Finally, participants were immediately debriefed as to the general purpose of the study (see Appendix H). The debriefing informed respondents that University of Florida researchers were not working on a vaccination for West Nile Virus nor would the respondents who wanted information about such a vaccination be contacted if one were to become available. Participants were given a debriefing flier to take home with them (see Appendix H) which included information on contacting the researcher if further questions about the study arose.

PAGE 54

CHAPTER 4 RESULTS Data for this study were collected between October 8 and October 30, 2001, from students at the University of Florida. This was less than 30 days after the terrorist attacks of Sept. 11. This study was an experiment. Participants Participants were recruited from University of Florida undergraduate and graduate communication and public relations courses, and computer network and software courses in the Computer and Information Science and Engineering Department at the University of Florida. Each course from which this researcher recruited the 204 participants was chosen because of the diverse make-up of the students. Several of the courses from which the researcher recruited had a diverse student body, including computer science, business, journalism, and engineering students. Participants self-reported their race (see Table 4-1), which was: 60.8% Caucasian, 9.3% African-American, 11.8% Hispanic, 9.3% Asian, and 6.4% Middle-Easterner; 2.5% of the participants selected “other.” Subject’s self-reported their age, which ranged from 18 to 50 years, with 73.8% of the participants were between the ages of 18 and 21. Participants in this study were both undergraduate (84.8%) and graduate (15.2%) students. 42

PAGE 55

43 Table 4-1. Self Reported Race of Participants ___________________________________ Race Percent N ___________________________________ Caucasian 60.8% 124 Hispanic 11.8 24 African-American 9.3 19 Asian 9.3 19 Middle-Eastern 6.4 13 Other 2.5 5 ___________________________________ 100% 204 Males represented 29.4% of the participants and females represented 70.6% of the subject. Communication with family while at school ranged from: several times a day for 14.2% of the participants, at least once a day for 22.1% of the participants, several times a week for 28.4% of the participants, at least once a week for 24.0% of the participants, at least once a month for 6.9% of the participants, rarely for 3.4% of the participants, and to never communicating with anyone in the family for 1% of the participants (see Table 4-2). Additionally, participants were asked about the frequency of contact with a family member when at home. Participants indicated that while at home they communicated with their family: several times a day (76%), at least once a day (7.8%), several times a week (1.5%), at least once a week (3.9%), at least once a month (1%), rarely (1.5%), and never communicate with anyone in the family ( n = 8.3%) (see Table 4-3).

PAGE 56

44 Table 4-2. Participants’ Communication with a Member of the Family When at School _________________________________________________________________ Percent N _________________________________________________________________ I don’t communicate with anyone in my family 1.0% 2 Rarely 3.4 7 At least once a month 6.9 14 At least once a week 24 49 Several times a week 28.4 58 At least once a day 22.1 45 Several times a day 14.2 29 _________________________________________________________________ Total 100% 204

PAGE 57

45 Table 4-3. Participants’ Communication with a Member of the Family When at Home _________________________________________________________________ Percent N _________________________________________________________________ I don’t communicate with anyone in my family 8.3% 17 Rarely 1.5 3 At least once a month 1 2 At least once a week 3.9 8 Several times a week 1.5 3 At least once a day 7.8 16 Several times a day 76 155 _________________________________________________________________ Total 100 204 _________________________________________________________________ Credibility Each participant was tested as to whether or not Shands at the University of Florida was thought to be a credible health information source. This test of credibility was important because Shands at the University of Florida was the attributed source of the health message used in this study. Participants completed a modified version of McCroskey’s source credibility scale. The 12-item scale (see Table 4-4) featured semantic differential items to gauge the subject’s perceptions of Shands at the University of Florida. Each item was weighed on a 7-point scale. The 12-item source credibility index, using the scale as a one-dimension

PAGE 58

46 construct, yielded a Cronbach’s alpha of .89. This is consistent with prior research using this scale. An initial factor analysis (principal axis factoring, oblimin rotation) revealed a three-factor solution through the scree plot method. Weighted factor scores were use to create three measures representing Liking, Expertise, and Trustworthiness, again consistent with most source credibility research. Table 4-4. Means, Standard Deviations, Factor Loadings, and Cronbach Alpha for Source Credibility—Shands at the University of Florida ________________________________________________________________________ Item a Factor Loading Mean SD Item if Alpha Deleted ________________________________________________________________________ Liking Shands Nice .77 5.5 1.1 .87 Shands Unselfish .73 5.0 1.3 .88 * Shands (Im) Moral .63 5.3 1.3 .87 * Shands (Un)Pleasant .61 5.1 1.3 .88 Shands Friendly .59 5.2 1.2 .88 * Shands (Dis)Honest .40 5.4 1.3 .87 Expertise Shands Qualified -.97 6.0 1.9 .88 Shands Informed -.60 5.7 1.3 .88 Shands Expert -.50 5.8 1.1 .87

PAGE 59

47 Table 4-4. Continued ________________________________________________________________________ Item Factor Loading Mean SD Item if Alpha Deleted ________________________________________________________________________ Trustworthinesss * Shands (Un)Intelligent -.90 5.9 1.5 .87 *Shands (Not) Valuable -.75 6.0 1.3 .87 * Shands (Un)Reliable -.35 5.5 1.5 .88 Reliability Coefficients 12 items Alpha = .87 Standardized item alpha = .89 * Items were reverse coded for analysis a Semantic differential where high numbers indicate positive association with credibility on a 7-point scale ________________________________________________________________________ The separate Liking factor for Shands at the University of Florida yielded a Cronbach’s alpha of .84. The Expertise factor for Shands at the University of Florida yielded a Cronbach’s alpha of .78. The Trustworthiness factor for Shands at the University of Florida, as measured by the scale, yielded a Cronbach’s alpha of .80. Family Integrity Scale A 22-question family integrity survey was used to measure the degree of idiocentrism or allocentrism for each subject. The entire Family Integrity Scale was used in its

PAGE 60

48 original form. One additional item was added to the traditional scale due to the Sept. 11, 2001 terrorist attacks at the time that the research data were collected. The decision to add this variable was due to the close proximity of data collection to the time of the terrorist attacks. This researcher foresaw a potential problem in accurately measuring nationalism. The data were collected one month after the terrorist attacks in New York, Pennsylvania and Washington, D.C. Because questions on the Family Integrity Scale attempted to gauge the ties to one’s community, there was concern that the emotional state of the nation would impact the findings. An item was needed to offset the scale item “it does not matter to me how my country is viewed in the eyes of other nations.” The new item was “it doesn’t matter to me how my community is viewed in the eyes of other communities” and it was intended to replace the country item, if necessary. An initial factor analysis (principal axis factoring, oblimin rotation) of the family integrity items suggested a seven-factor scale; however, several items did not load as expected, didn’t appear to “belong” with any of the seven factors, or did not load at all (see Table 4-5).

PAGE 61

49 Table 4-5. Factor Pattern Loadings on Original Seven-Factor Solution, Principal Axis Factor Oblimin Rotation FACTOR Item 1 2 3 4 5 6 7 It doesn’t matter to me how my community is viewed in the eyes of other communities. .64 -.43 -.45 I tend to do my own thing, and others in my family do the same. .61 When faced with a difficult personal problem, it is better to decide what to do yourself, rather than follow the advice of others. .53 Individuals should be judged on their own merits, not on the company they keep. .44 One does better work working alone than in a group. .39 One should live one’s life independently of others as much as possible. a .39 .-34 It does not matter to me how my country is viewed in the eyes of other nations. .37 .-35 What happens to me is my own doing. .31 If the child won the Nobel Prize, the parents should not feel honored in any way. .59 .36 *Children should live at home with their parents until they get married. .43 -.54 .31 Children should not feel honored even if the father were highly praised and given an award by a government official for his contributions and services to the community. .39 It is important to me that I perform better than others on a task. .46 *In most cases, to cooperate with someone whose ability is lower than yours is not as desirable as doing the thing on your own. .42 If the group is slowing me down, it is better to leave it and work alone. .40 *Ageing parents should live at home with their children. .48

PAGE 62

50 Table 4-5. Continued I would rather struggle through a personal problem by myself than discuss it with my friends. a .63 -.31 .-39 *I enjoy talking to my neighbors everyday. a *I like to live close to my good friends. a *What I look for in a job is a friendly group of coworkers. a .40 *I can count on my relatives for help if I find myself in any kind of trouble. a The most important thing in my life is to make myself happy. a .33 *I would help within my means if a relative told me that (s)he is in financial difficulty. a * Item reverse coded a Item removed in subsequent factor analysis. Next, the following eight items were deleted from factor analysis: “one should live one’s life independently of others as much as possible,” “I would rather struggle through a personal problem by myself than discuss it with my friends,” “I would help within my means, if a relative told me that s(he)is in financial difficulty,” “I like to live close to my good friends,” “the most important thing in my life is to make myself happy,” “what I look for in a job is a friendly group of coworkers,” “I enjoy talking to my neighbors everyday,” and “I can count on my relatives for help if I find myself in any kind of trouble.” These items were deleted because they did not load or the items did not appear to be testing the same culture constructs. The remaining 14 items were then submitted to another factor analysis that produced a five-factor solution. The five-factor solution is interpreted as Dependence/Reliance, Group vs. Self, Closeness to Ingroups, Family

PAGE 63

51 Integrity, and Community vs. Individual Identification ideals (see Table 4-6). The factor matrix shows the five-factor solution scores (see Table 4-7). Table 4-6. Pattern Factor Loadings Oblimin Roatation, Means, Standard Deviations, and Cronbach Alpha for Family Integrity Scale ______________________________________________________________________ Factor 1 Dependence/Reliance Item Factor Loading Mean a SD Item if Alpha Deleted I tend to do my own thing, and others in my family do the same. .63 4.0 1.7 .47 Individuals should be judged on their own merits, not on the company they keep. .57 2.8 1.6 .52 When faced with a difficult personal problem, it is better to decide what to do yourself, rather than follow the advice of others. .53 4.0 1.5 .58 Reliability Coefficients 3 items Alpha = .62 Standardized item alpha = .62 a 7-point Likert scale, high numbers indicate allocentric, low numbers indicate idiocentric ______________________________________________________________________

PAGE 64

52 Table 4-6. Continued Factor 2 Group vs. Self Item Factor Loadings Mean a SD Item if Alpha Deleted ______________________________________________________________________ It is important to me that I perform better than others on a task. .54 3.0 1.4 .51 * In most cases, to cooperate with someone whose ability is lower than yours is not as desirable as doing the thing on your own. .52 3.8 1.7 .52 If the group is slowing me down, it is better to leave it and work alone. .51 3.3 1.7 .50 What happens to me is my own doing. .40 3.0 1.4 .54 One does better work alone than in a group. .34 4.2 1.7 .55 * Indicates reverse coding Reliability Coefficients 5 items Alpha = .58 Standardized item alpha = .58 a 7-point Likert scale, high numbers indicate allocentric, low numbers indicate idiocentric

PAGE 65

53 Table 4-6. Continued Factor 3 Closeness to Ingroups Item Factor Loadings Mean a SD Item if Alpha Deleted ________________________________________________________________________ Even if a child won the Nobel prize the parents should not feel honored in any way. .69 6.4 1.0 n/a Children should not feel honored even if the father were praised and given an award by government. .58 6.1 1.3 n/a Reliability Coefficients 2 items Alpha = .55 Standardized item alpha = .56 a 7-point Likert scale, high numbers indicate allocentric, low numbers indicate idiocentric _______________________________________________________________________ Factor 4 Family Integrity Item Factor Loadings Mean a SD Item if Alpha Deleted _______________________________________________________________________ * Children should live at home with their parents until they get married. .89 4.3 1.7 n/a * Aging parents should live at home with their children. .47 2.4 1.8 n/a Reliability Coefficients 2 items Alpha = .62 Standardized item alpha = .62 * Indicates reverse coding a 7-point Likert scale, high numbers indicate allocentric, low numbers indicate idiocentric

PAGE 66

54 Table 4-6. Continued Factor 5 Community vs. Individual Identification Item Factor Loadings Mean SD Item if Alpha Deleted _______________________________________________________________________ It doesn’t matter to me how my community is viewed in the eyes of other communities. -.82 5.3 1.6 n/a It does not matter to me how my country is viewed in the eyes of other nations. -.64 5.2 1.4 n/a Reliability Coefficients 2 items Alpha = .68 Standardized item alpha = .68 7-point Likert scale, High numbers indicate allocentric, low numbers indicate idiocentric _______________________________________________________________________

PAGE 67

55 Table 4-7. Factor Scores for Five-Factor Culture Solution FACTOR Item 1 2 3 4 5 I tend to do my own thing, and others in my family do the same. .37 .11 -.00 .00 -.04 Individuals should be judged on their own merits, not on the company they keep. .26 .02 -.04 .02 -.01 When faced with a difficult personal problem, it is better to decide what to do yourself, rather than follow the advice of others. .23 -.01 .05 .00 -.05 It is important to me that I perform better than others on a task .003 .28 -.00 -.03 .00 *In most cases, to cooperate with someone whose ability is lower than yours is not as desirable as doing the thing on your own. -.01 .28 .07 .02 .00 If the group is slowing me down, it is better to leave it and work alone. .05 .27 -.04 -.01 .00 What happens to me is my own doing. .02 .18 .01 -.01 -.03 One does better work alone than in a group. .11 .17 -.02 .02 -.01 Even if a child won the Nobel Prize, the parents should not feel honored in any way. -.05 .04 .52 -.03 -.02 Children should not feel honored even if the father were highly praised and given an award by a government official for his contributions and services to the community. .03 -.02 .34 .01 -.00 *Aging parents should live at home with their children. -.06 .07 .05 .78 .07 *Children should live at home with their parents until they get married. .22 -.13 -.09 .17 -.02 It does not matter to me how my country is viewed in the eyes of other nations. .03 .02 .-02 -.05 -.25 It doesn’t matter to me how my community is viewed in the eyes of other communities. .05 -.01 .10 .06 -.70 * Indicates item reversed coded

PAGE 68

56 The factor loadings found in this research differed from the original Triandis (Triandis et al., 1986; Triandis et al., 1990) studies and factors. The data and factor loadings from Triandis’ original four factors showed low factor loadings (see Table 4-8). Additionally, previous research surveyed nationals within their own countries, whereas this study investigated American sub-cultures of other nationalities. When factor groupings contained similarly grouped items as found in the Triandis (Triandis et al., 1986; Triandis et al., 1990) findings, the original factor names were used. When factor items were not consistent with earlier findings, this researcher created intuitive labels for each factor. Rather than rely on Triandis’ low-loading factors, this researcher used the five separate factor indexes created from factor scores.

PAGE 69

57 Table 4-8. Items Loading on Same Factor in Pancultural Factor Analysis of Individualism-Collectivism from Triandis (1986) _______________________________________________________________________ Factor 1 Self Reliance with Hedonism When faced with a difficult personal problem, it is better to decide what to do yourself, rather than follow the advice of others. .57 If the group is slowing me down, it is better to leave it and work alone. .50 I would rather struggle through a personal problem by myself than discuss it with my friends. .41 I tend to do my own thing, and others in my family do the same. .39 The most important thing in my life is to make myself happy. .35 One does better work working alone than in a group. .43 What happens to me is my own doing. .34 Factor 2 Closeness to Ingroups Children should not feel honored even if the father were highly praised and given an award by a government official for his contributions and services to the community. .73 If the child won the Nobel Prize, the parents should not feel honored in any way. .65 In most cases, to cooperate with someone whose ability is lower than yours is not as desirable as doing the thing on your own. .32 Factor 3 Family Integrity Ageing parents should live at home with their children. .58 Children should live at home with their parents until they get married. .58 It is important to me that I perform better than others on a task. .48 One should live one’s life independently of others as much as possible. .31

PAGE 70

58 Table 4-8. Continued Factor 4 Dependence/Reliance and Sociability I like to live close to my good friends. .43 I would help within my means, if a relative told me that s(he) is in financial difficulty. .41 Individuals should be judged on their own merits, not on the company they keep. .40 _______________________________________________________________________ The Dependence/Reliance construct was created through factor scores where three items were primary. The items included: “I tend to do my own thing, and others in my family do the same,” “when faced with a difficult problem, it is better to decide what to do yourself, rather than follow the advice of others;” and “individuals should be judged on their own merits, not on the company they keep.” The standardized item alpha score if the index included only those items would have been .62. The Group vs. Self factor gauged participants’ dependence on themselves vs. the group. The five primary items included: “it is important to me that I perform better than others on a task;” “in most cases, to cooperate with someone whose ability is lower than yours is not as desirable as doing the thing on your own;” “if the group is slowing me down, it is better to leave it and work alone;” “what happens to me is my own doing;” and “one does better work alone than in a group.” The standardized item alpha score if the index included only those items would have been .58. The Closeness to Ingroups factor was created using two primary items: “even if a child won the Nobel Prize, the parents should not feel honored in any way” and “children should not feel honored even if the father were highly praised and given an award by a

PAGE 71

59 government official for his contributions and services to the community.” The standardized item alpha score if the index included only those items would have been .56. The Family Integrity construct contained two primary items: “aging parents should live at home with their children” and “children should live at home with their parents until they get married.” The standardized item alpha score if the index included only those items would have been .62. The final factor, Community vs. Individual Identification, was composed of two primary items: “it does not matter to me how my country is viewed in the eyes of other nations” and “it doesn’t matter to me how my community is viewed in the eyes of other communities.” The standardized item alpha score if the index included only those items would have been .68. Because culture type is central to the hypotheses in this study, this researcher needed to access the “strongest” cultural participants. Participants’ self-reported culture scores for each factor was separated into a 40/20/40 to determine the high and low scorers. This method was used in order to test participants who scored more strongly toward a culture, rather than use an arbitrary line to separate idiocentrics and allocentrics. The top 40% scorers for each culture factor were used to test the hypotheses as the strongest allocentrics. The bottom 40% scorers for each culture factor score were used to test the hypotheses as the strongest idocentrics. Each of the five factors for these “strongest” cultural participants were analyzed separately. It was important to analyze each culture factor score separately because of the differences in each factor. Based on this, it was possible for a single participant to measure allocentric on some culture factors and idiocentric on other culture factors. Because allocentrism and idiocentrism are

PAGE 72

60 dimensions of culture, not considered concrete culture types, this was acceptable (Triandis et al., 1986). Finally, using this method allowed for equal representation of allocentric and idiocentric participants (see Table 4-9). Table 4-9. Self-Reported Culture for All Factors for Strong Culture Participants _________________________________________ Percent N _________________________________________ Strongest Idiocentric 49.7% 81 Strongest Allocentric 50.3 82 _________________________________________ Total 100% 163 _________________________________________ Message All participants were randomly assigned to conditions either with allocentric (51.0%) or idiocentric (49.0%) messages and in either an interactive (50.0%) or static (50.0%) format. A manipulation check at the end of the questionnaire found that those who were allocentric and received the allocentric message identified that the message as expressing dangers for people like them; those who were idiocentric and received the idiocentric message indicated that it was expressed dangers to them directly. The manipulation check read: “This message was intended for: me or others like me.” The participants who circled both “me” and “others like me were put into a “both” category for data analysis. The manipulation check was performed through a crosstab chi square analysis and reached statistical significance ( p < .01), as seen in Table 4-10.

PAGE 73

61 Table 4-10. Manipulation Check _________________________________________ Participants’ Culture _________________________________________ Allocentric Idiocentric Total N The message emphasized risk to: Me 36.5% 72.0% 53.9% 110 Others 52.9 25.0 39.2 80 Both 10.6 3.0 6.9 14 _________________________________________ Total 100% 100% 100% 204 Chi Square (1, 1) = 26.26, p < .01 _________________________________________ After exposure to the messages, 64.7% of the participants indicated that they wanted to be contacted about a vaccination, as seen in Table 4-11.

PAGE 74

62 Table 4-11. Behavior Elicited from Participants After Message Exposure Question: Do you want to be contacted with information about a vaccine? _________________________________________ Percent N _________________________________________ Yes 64.7% 132 No 35.3 72 _________________________________________ Total 100% 204 _________________________________________ Next, study participants completed a short eight-item retention quiz and demographic questions. In this quiz, retention was measured through two true/false questions and six unaided recall questions. The true/false questions asked respondents if West Nile Virus was food poisoning (false) and if 100 people in Florida had contracted West Nile Virus since 1999 (false). All participants responded to the true/false questions (see Table 4-12), giving either one correct response (16.7%) or two correct responses (83.3%). Participants also demonstrated message retention (see Table 4-13) through their responses to six unaided recall questions scoring: 0 correct responses (0.0%), one correct response (0.5%), two correct responses (1.0%), three correct responses (5.4%), four correct responses (20.6%), five correct responses (39.7%), or six correct responses (32.8%). These retention questions asked participants how West Nile Virus is

PAGE 75

63 transmitted, in which state was it first reported, after many days can symptoms occur, and to list three ways one can avoid contracting West Nile Virus. Table 4-12. Participants’ Retention Demonstrated Through Two True/False Questions _______________________________________________ Percent N 0 Correct Responses 0.0% 0 1 Correct Response 16.7 34 2 Correct Responses 83.3 170 _______________________________________________ 100% 204 _______________________________________________

PAGE 76

64 Table 4-13. Participants’ Retention Demonstrated Through Responses to Six Unaided Recall Questions ________________________________________________ Percent N 0 Correct Responses 0% 0 1 Correct Responses .5 1 2 Correct Response 1.0 2 3 Correct Responses 5.4 11 4 Correct Responses 20.6 42 5 Correct Responses 39.7 81 6 Correct Responses 32.8 67 ________________________________________________ 100% 204 Analysis for hypotheses H1: Participants who receive culturally matched messages are more likely to adopt the suggested PHB. H2: Participants who receive messages in a culturally matched level of interactivity are more likely to adopt the suggested PHB. The self-reported culture scores for each of the five factors were compared to the message type and level of interactivity to which the subject was exposed. This comparison yielded a new variable of either match or unmatched with the message and with the level of interactivity for each subject. Participants were identified as either idiocentric or allocentric for each of the five culture factors. If an idiocentric subject received an idiocentric message, a new variable was created to indicate a message match. If an idiocentric subject received an allocentric message, a new variable was created to

PAGE 77

65 indicate an unmatched message. This analysis process was also done to determine culture compatibility for the level of interactivity. An idiocentric was considered matched when exposed to a message in the interactive level, while allocentrics were considered matched when exposed to a static message. This cultural compatibility determination was made based on previous research as to the most effective way to communicate with different culture and ethnic groups (Pew Research Center, 2000b; Triandis et al., 1986). Finally, a chi square analysis tested whether or not a message or level of interactivity match would yield the desired behavior to test the hypotheses. Each hypothesis was tested five times, once for each culture factor. Because the dimensions of culture were so different, it became important to measure each factor separately. Merging the five, very different cultural constructs, would have been an inaccurate measure of the participants’ culture. Additionally, the idiocentric/allocentric labels are meant to be dimensions of culture and not exclusively individualistic or collective. As seen in Tables 4-14 and 4-15, the self-reported Dependence/Reliance culture factor participants were exposed 52.1% of the time to a culturally matched message and 47.9% of the time to a culturally unmatched message. Of those who desired to be contacted, 61.7% received a culturally matched message and 61.5% received a culturally unmatched message. For interactivity, 50.3% of the participants were culturally matched and 49.7% of the participants were culturally unmatched. Of those who desired to be contacted, 64.6% received a culturally matched level of interactivity and 64.2% received a culturally unmatched level of interactivity. A crosstabs chi square analysis revealed that behavior was not a significant result for participants who received a culturally matched message vs. those who did not (Chi

PAGE 78

66 Square (1, 1) = 54, p < .23) for the Dependence/Reliance factor. A crosstabs chi square analysis revealed that behavior was not a significant result for those who received a culturally matched level of interactivity vs. those who did not (Chi Square (1, 1) = .00, p < .48) for the Dependence/Reliance factor. Table 4-14. Effects of Culturally Matched Message for Factor 1 Dependence/Reliance Factor on Desire to be Contacted _________________________________________ Message Match Status _________________________________________ Match Unmatch Total N Do you want to be contacted? Yes 61.7% 61.5% 64.4% 105 No 39.2 38.5 35.6 58 _________________________________________ Total 100% 100% 100% 163 Chi Square (1, 1) = .54 , p < .23 _________________________________________

PAGE 79

67 Table 4-15. Effects of Culturally Matched Level of Interactivity for Factor 1 Dependence/Reliance Factor on Desire to be Contacted _________________________________________ Level of Interactivity Match Status _________________________________________ Match Unmatch Total N Do you want to be contacted? Yes 64.6% 64.2% 64.4% 105 No 35.4 35.8 35.6 58 ________________________________________ Total 100% 100% 100% 163 Chi Square (1, 1) = .00, p < . 48 _________________________________________ As seen in Tables 4-16 and 4-17, 53.4% of the Group vs. Self culture factor participants were exposed to culturally matched messages and 43.6% were exposed to culturally unmatched messages. Of those who desired to be contacted, 63.2% received a culturally matched message and 64.5% received a culturally unmatched message. For the level of interactivity in this factor, 50.3% of the respondents were culturally matched vs. 49.7% who were culturally unmatched to the level of interactivity. Of those who desired to be contacted, 63.4% received a culturally matched level of interactivity and 62.4% received a culturally unmatched level of interactivity. A crosstabs chi square analysis revealed that behavior was not a significant result for participants who received a culturally matched message vs. those who did not (Chi

PAGE 80

68 Square (1, 1) = .03, p < .43) for the Group vs. Self factor. A crosstabs chi square analysis revealed that behavior was not a significant result for those who received a culturally matched level of interactivity vs. those who did not (Chi Square (1, 1) = .01, p < .46) for the Group vs. Self factor. Table 4-16. Effects of Culturally Matched Message for Factor 2 Group vs. Self Factor on Desire to be Contacted _________________________________________ Message Match Status _________________________________________ Match Unmatch Total N Do you want to be contacted? Yes 63.2% 64.5% 63.8% 104 No 36.8 35.5 36.2 59 ________________________________________ Total 100% 100% 100% 163 Chi Square (1, 1) = .03 p < .43 _________________________________________

PAGE 81

69 Table 4-17. Effects of Culturally Matched Level of Interactivity for the Factor 2 Group vs. Self Factor on Desire to be Contacted _________________________________________ Level of Interactivity Match Status _________________________________________ Match Unmatch Total N Do you want to be contacted? Yes 63.4% 64.2% 63.8% 104 No 36.6 38.8 36.2 59 ________________________________________ Total 100% 100% 100% 163 Chi Square (1, 1) = .01, p < .46 _________________________________________ As seen in Tables 4-18 and 4-19, in the Closeness to Ingroups culture, 48.5% of the participants were exposed to culturally matched messages vs. 51.5% of the participants who were exposed to culturally unmatched messages. Of those who desired to be contacted, 62.0% received a culturally matched message and 64.3% received a culturally unmatched message. For this level of interactivity of this factor, there were 54.0% of the participants who were culturally matched with the level of interactivity vs. 46.0% of the participants who were culturally unmatched with the level of interactivity. Of those who desired to be contacted, 60.2% received a culturally matched level of interactivity and 67.7% received a culturally unmatched level of interactivity. A crosstabs chi square analysis revealed that behavior was not a significant result for participants who received a culturally matched message vs. those who did not (Chi

PAGE 82

70 Square (1, 1) = 09, p < .38) for the Closeness to Ingroups factor. A crosstabs chi square analysis revealed that behavior was not a significant result for those who received a culturally matched level of interactivity vs. those who did not (Chi Square (1, 1) = .72, p < .20) for the Closeness to Ingroups factor. Table 4-18. Effects of Culturally Matched Message for Factor 3 Closeness to Ingroups Factor on Desire to be Contacted _________________________________________ Message Match Status _________________________________________ Match Unmatch Total N Do you want to be contacted? Yes 62.0% 64.3% 63.2% 103 No 38.0 35.7 36.8 60 ________________________________________ Total 100% 100% 100% 163 Chi Square (1, 1) = .09, p < .38 _________________________________________

PAGE 83

71 Table 4-19. Effects of Culturally Matched Level of Interactivity for the Factor 3 Closeness to Ingroups Factor on Desire to be Contacted _________________________________________ Level of Interactivity Match Status _________________________________________ Match Unmatch Total N Do you want to be contacted? Yes 60.2% 67.7% 63.2% 103 No 39.8 33.3 36.8 60 ________________________________________ Total 100% 100% 100% 163 Chi Square (1, 1) = .72, p < .20 _________________________________________ As seen in Tables 4-20 and 4-21, the 48.5% of the Family Integrity culture factor participants were exposed to culturally matched messages vs. 51.5 % of the participants who were exposed to culturally unmatched messages. Of those who desired to be contacted, 65.8% received a culturally matched message and 66.7% received a culturally unmatched message. For the level of interactivity for this factor, 44.8% of the participants were exposed to a culturally matched level of interactivity vs. 55.2% of participants who were exposed to culturally unmatched level of interactivity. Of those who desired to be contacted, 69.9% received a culturally matched level of interactivity and 63.3% received a culturally unmatched level of interactivity. A crosstabs chi square analysis revealed that behavior was not a significant result for participants who received a culturally matched message vs. those who did not (Chi

PAGE 84

72 Square (1, 1) = 01, p < .45) for the Family Integrity factor. A crosstabs chi square analysis revealed that behavior was not a significant result for those who received a culturally matched level of interactivity vs. those who did not (Chi Square (1, 1) = .77, p < .19) for the Family Integrity factor. Table 4-20. Effects of Culturally Matched Message for Factor 4 Family Integrity Factor on Desire to be Contacted _________________________________________ Message Match Status _________________________________________ Match Unmatch Total N Do you want to be contacted? Yes 65.8% 66.7% 66.3% 108 No 34.2 33.3 33.7 55 ________________________________________ Total 100% 100% 100% 163 Chi Square (1, 1) = .01, p < .45 _________________________________________

PAGE 85

73 Table 4-21. Effects of Culturally Matched Level of Interactivity for Factor 4 Family Integrity Factor on Desire to be Contacted _________________________________________ Level of Interactivity Match Status _________________________________________ Match Unmatch Total N Do you want to be contacted? Yes 69.9% 63.3% 66.3% 108 No 30.1 36.7 33.7 55 ________________________________________ Total 100% 100% 100% 163 Chi Square (1, 1) = .77, p < .19 _________________________________________ As seen in Tables 4-22 and 4-23, the 51.5% of the Community vs. Individual Identification culture factor participants were exposed to culturally matched messages vs. 48.5% who were exposed to culturally unmatched messages. Of those who desired to be contacted, 59.5% received a culturally matched message and 70.9% received a culturally unmatched message. For the level of interactivity for this factor, 51.5% of the participants were exposed to culturally matched level of interactivity vs. 48.5% who were exposed to culturally unmatched level of interactivity. Of those who desired to be contacted, 70.2% received a culturally matched level of interactivity and 59.5% received a culturally unmatched level of interactivity. A crosstabs chi square analysis revealed that behavior approached a significant result for participants who received a culturally matched message vs. those who did not (Chi

PAGE 86

74 Square (1, 1) = 2.31, p < .06) for the Community vs. Individual Identification factor. A crosstabs chi square analysis revealed that behavior approached a significant result for those who received a culturally matched level of interactivity vs. those who did not (Chi Square (1, 1) = 2.07, p < .08) for the Community vs. Individual Identification factor. Table 4-22. Effects of Culturally Matched Message for Factor 5 Community vs. Individual Identification Factor on Desire to be Contacted _________________________________________ Message Match Status _________________________________________ Match Unmatch Total N Do you want to be contacted? Yes 59.5% 70.9% 65.0% 106 No 40.5 29.1 35.0 57 ________________________________________ Total 100% 100% 100% 163 Chi Square (1, 1) = 2.31, p < .06 _________________________________________

PAGE 87

75 Table 4-23. Effects of Culturally Matched Level of Interactivity for Factor 5 Community vs. Individual Identification Factor on Desire to be Contacted _________________________________________ Level of Interactivity Match Status _________________________________________ Match Unmatch Total N Do you want to be contacted? Yes 70.2% 59.5% 65.0% 106 No 29.8 40.5 35.0 57 ________________________________________ Total 100% 100% 100% 163 Chi Square (1, 1) = 2. 07, p < .08 _________________________________________ No hypotheses in this study were supported. Post-hoc Analysis Although none of the hypotheses in this research study were supported, post-hoc analyses showed significant relationships between variables, culture, and ethnicity. Recall A main effect on recall occurred in the Closeness to Ingroups culture factor in the message condition. When the message matched the participants’ self-reported culture in this factor, recall was greater (see Figure 4-1). In this factor, participants in the matched message condition had a mean recall of 4.9 ( SD = 1.01), while participants in the unmatched message condition had a mean recall of 4.5 ( SD = 1.29) ( F = 4.89, p < .03) (see Table 4-24).

PAGE 88

76 Unaided Recall: Effects of Message Matchingwith the Closeness to Ingroups FactorMessage Match Status for Closeness to IngroupsMessage UnmatchMessage MatchEstimated Marginal Means6.05.55.04.54.03.53.02.52.01.51.0.50.0 4.54.9 Figure 4-1. Unaided Recall: Effects of Message Matching with Closeness to Ingroups Factor Table 4-24. Means, Standard Deviation and F Test for Effects of Message Matching in Closeness to Ingroups Factor on Unaided Recall ____________________________________________ Mean SD Message Match 4.9 1.01 Message Unmatch 4.5 1.29 F test (1,1) = 4.54, p < .03 ____________________________________________ In addition, there was an interaction between the Family Integrity culture factor and the level of interactivity in relation to message recall. When participants rate Trustworthiness high for Shands at the University of Florida, there is a nearly equal

PAGE 89

77 amount of recall regardless of whether the message is matched or unmatched. However, when participants rate the Trustworthiness of Shands at the University of Florida low on source credibility, the level of interactivity match status is an important predictor of recall. Participants who do not trust Shands and who are exposed to a matched level of interactivity have a statistically higher recall than those who are matched (see Figure 4-2). Participants who rated Trustworthiness high for Shands at the University of Florida had a 4.9 mean in both the matched and unmatch level of interactivity conditions. Participants in the unmatched level of interactivity condition who rated Shands low in Trustworthiness had a mean of 4.3; participants in the matched level of interactivity who rated Shands high in Trustworthiness had a mean of 5.2 ( F = 3.39, p < .07) (See Table 4-25).

PAGE 90

78 Unaided Recall: Effects of Trustworthiness andLevel of Matching for Family Ingrity FactorTrustworthiness of Shands at UFLow CredibilityHigh CredibilityEstimated Marginal Means6.05.55.04.54.03.53.02.52.01.51.0.50.0 Match StatusUnmatchedMatched 5.24.94.34.9 Figure 4-2. Unaided Recall: Effects of Trustworthiness and Level of Interactivity Matching for the Family Integrity Culture Factor

PAGE 91

79 Table 4-25. Means, Standard Deviation and F Test for Effects of Message Matching in Closeness to Ingroups Factor on Unaided Recall ____________________________________________ Mean SD High Trustworthiness Level Match 4.9 .28 Level Unmatch 4.9 .27 Low Trustworthiness Level Match 4.72 .18 Level Unmatch 4.23 .25 F test (1, 1) = 3.38, p < .07 ____________________________________________ Trustworthiness One assumption of the Health Belief Model (HBM; Rosenstock, 1974; Janz & Becker, 1984) is that message recipients must find the source and the message trustworthy. Trustworthiness, the source credibility factor identifying the level of reliability and trust participant’s found in Shands at the University of Florida, became a key element in this research study. This research study found a greater result in intended behavior when trustworthiness is high through the matched message condition in the Group vs. Self culture factor (see Figures 4-3 and 4-4; Table 4-26).

PAGE 92

80 Low TrustworthinessMessage Match Status for Factor 2Unmatched LevelMatched MessageCount403020100 Intended BehaviorNo, Don't Contact MeYes, Contact Me 27241018 Figure 4-3. Effects of Low Trustworthiness for Factor 2 Group vs. Self on the Desire to be Contacted

PAGE 93

81 High TrustworthinessMessage Match Status for Factor 2Unmatched MessageMatched MessageCount403020100 Intended BehaviorNo, Don't Contact MeYes, Contact Me 22311714 Figure 4-4. Effects of High Trustworthiness for Factor 2 Group vs. Self on the Desire to be Contacted

PAGE 94

82 Table 4-26. Chi Square on Effects of Trustworthiness for Factor 2 Group vs. Self on the Desire to be Contacted _____________________________________________________________________ Do you want to be contacted? Yes No Total N Low Trustworthiness Message Match 47.1% 64.3% 53.2% 42 Message Unmatch 52.9 35.7 46.8 37 _____________________________________________________________________ 100% 100% 100% 79 Chi Square (1,1) = 2.15, p < .07 _____________________________________________________________________ Do you want to be contacted? Yes No Total N High Trustworthiness Message Match 58.5% 45.2% 53.6% 45 Message Unmatch 41.5 54.8 46.4 39 _____________________________________________________________________ 100% 100% 100% 84 Chi Square (1,1) = 1.40, p < .12

PAGE 95

83 In the Closeness to Ingroups factor, Trustworthiness becomes an important factor as participants in the unmatched level of interactivity condition overcome this barrier and desire to be contacted when Trustworthiness is high (see Figures 4-5 and 4-6; Table 4-27). Low TrustworthinessLevel of Interactivity Match Status for Factor 3Unmatched LevelMatched LevelCount3020100 Intended BehaviorNo, Don't Contact MeYes, Contact Me 2129924 Figure 4-5. Effects of Low Trustworthiness for Factor 3 Closeness to Ingroups on the Desire to be Contacted

PAGE 96

84 High TrustworthinessLevel of Interactivity Match Status for Factor 3Unmatched LevelMatched LevelCount3020100 Intended BehaviorNo, Don't Contact MeYes, Contact Me 29241611 Figure 4-6. Effects of High Trustworthiness for Factor 3 Closeness to Ingroups on the Desire to be Contacted

PAGE 97

85 Table 4-27. Chi Square on Effects of Trustworthiness for Factor 3 Closeness to Ingroups on the Desire to be Contacted _____________________________________________________________________ Do you want to be contacted? Yes No Total N Low Trustworthiness Level Match 58.0% 72.7% 63.9% 53 Level Unmatch 42.0 27.3 36.1 30 _____________________________________________________________________ 100% 100% 100% 83 Chi Square (1,1) = 1.87, p < .09 _____________________________________________________________________ Do you want to be contacted? Yes No Total N High Trustworthiness Level Match 45.3% 40.7% 43.8% 35 Level Unmatch 54.7 59.3 56.3 45 _____________________________________________________________________ 100% 100% 100% 80 Chi Square (1,1) = 1.50, p < .35

PAGE 98

86 In the Community vs. Individual Identification Factor, there was an increase in the desire to be contacted in the level of interactivity match status as Trustworthiness increased (see Figures 4-7 and 4-8; Table 4-28). Low Trustworthiness Level of Interactivity Match Status for Factor 5Unmatched LevelMatched LevelCount403020100 Intended BehaviorNo, Don't Contact MeYes, Contact Me 20281712 Figure 4-7. Effects of Low Trustworthiness for Factor 5 Community vs. Individual Identification on the Desire to be Contacted

PAGE 99

87 High TrustworthinessLevel of Interactivity Match Status for Factor 5Unmatched LevelMatched LevelCount403020100 Intended BehaviorNo, Don't Contact MeYes, Contact Me 27311513 Figure 4-8. Effects of High Trustworthiness for Factor 5 Community vs. Individual Identification on the Desire to be Contacted

PAGE 100

88 Table 4-28. Chi Square on Effects of Trustworthiness for Factor 5 Community vs. Individual Identification on the Desire to be Contacted _____________________________________________________________________ Do you want to be contacted? Yes No Total N Low Trustworthiness Level Match 58.3% 41.4% 51.9% 40 Level Unmatch 41.7 58.6 48.1 37 _____________________________________________________________________ 100% 100% 100% 77 Chi Square (1,1) = 2.08, p < .07 _____________________________________________________________________ Do you want to be contacted? Yes No Total N High Trustworthiness Level Match 53.4% 46.4% 51.2% 44 Level Unmatch 46.6 53.6 48.8 42 _____________________________________________________________________ 100% 100% 100% 86 Chi Square (1,1) = .37, p < .21

PAGE 101

89 Ethnicity In the Culture Factors This study based cultlural predictions on ethinicty and traditional cultural categorizations for the ethnic groups. Because of that, is is important to look how each ethnic group scored within each of the five culture factors in this study. Factor 1 measured Dependence/Reliance through items that identified one’s desire to do his or her own thing, be judged on own merits, and make own descisions. In this factor, Asian and Middle-Eastern participants rank extremly allocentric (see Figure 4-9; Table 4-29). This finding is statistically significant (F = 4.09, p < .01). These findings are consistent with cultural ethnitc data in previous studies.

PAGE 102

90 Ethnicity of ParticipantsOtherMiddle-EasternAsianHispanicAfrician-AmericanCaucasianCulture Factor 1 (Idiocentric to Allocentric)2.01.81.61.41.21.0 1.32.02.01.51.31.5 Figure 4-9. Ethnicity of Factor 1 Dependence/Reliance

PAGE 103

91 Table 4-29. Means, Standard Deviation and F Test for Effects of Ethnicity in Dependence/Reliance Factor ____________________________________________ Mean SD Caucasian 1.49 .51 African-American 1.25 .50 Hispanic 1.46 .52 Asian-American 2.00 .00 Middle-Eastern 2.00 .00 Other 1.25 .50 F test (1, 1) = 4.09, p < .01 ____________________________________________ Factor 4 measured Family Integirty through items in which participants rated their believes about continued cohabiation between children and parents. In this factor, Caucasions rate the most idiocentric and Asians and Middle-Easterns rate the most allocentric amoung ethnic groups (see Figure 4-10; Table 4-30). This finding is statistically significant ( F = 4.32, p < .01). This finding was expected and is consistent with previous cultural ethnic findings.

PAGE 104

92 Ethnicity of ParticipantsOtherMiddle-EasternAsianHispanicAfrician-AmericanCaucasianCulture Factor 4 (Idiocentric to Allocentric)2.01.81.61.41.21.0 1.82.02.01.71.81.4 Figure 4-10. Ethnicity in Factor 4 Family Integrity

PAGE 105

93 Table 4-30. Means, Standard Deviation and F Test for Effects of Ethnicity in Family Integrity Factor ____________________________________________ Mean SD Caucasian 1.40 .50 African-American 1.75 .50 Hispanic 1. 96 .48 Asian-American 2.00 .00 Middle-Eastern 2.00 .00 Other 1.75 .50 F test (1, 1) = 4.32, p < .01 ____________________________________________ Factor 5 measured Community vs. Individual Identification through items which gauged one’s sense of community and nationalism. In this factor, Africian-Americans rate the most allocentric and Asians and Middle-Easterns rate the most idiocentric amoung ethnic groups (see Figure 4-11; Table 4-31). This finding is statistically significant ( F = 2.39, p < .05). This finding was nearly opposite of what was expected. This finding is attributed to the effects of Sept. 11, 2001 and the subsequent barrier Middle-Easterns may have felt was necessary in order to connect with what other Americans were feeling in the wake of the tradgedy.

PAGE 106

94 Ethnicity of ParticipantsOtherMiddle-EasternAsianHispanicAfrician-AmericanCaucasianCulture Factor 5 (Idiocentric to Allocentric)2.22.01.81.61.41.21.0 1.81.11.21.52.01.5 Figure 4-11. Ethnicity in Factor 5 Community vs. Individual Identification

PAGE 107

95 Table 4-31. Means, Standard Deviation and F Test for Effects of Ethnicity in Community vs. Individual Factor ____________________________________________ Mean SD Caucasian 1.49 .51 African-American 2.00 .00 Hispanic 1. 46 .52 Asian-American 1.22 .44 Middle-Eastern 1.14 .39 Other 1.75 .50 F test (1, 1) = 2.39, p < .05 ____________________________________________ Frequency of Family Communication Triandis et al. (1986) make assumptions about the frequency of contact with family members based on the idiocentric/allocentric cultural tendancies. This study allowed the opportunity to see how often participants communicate with a member of their family when at school and at home. Caucausion participants indicated that they communicated with a family member more frequently than Asian and Middle-Eastern particiapants while at school (see Figure 4-12). This finding is statistically significant ( F = 5.46, p < .01). This finding is expected considering the cost a Middle-Eastern student may incur from frequent telephone communication vice that what the American students would incur.

PAGE 108

96 EthnicityOtherMiddle-EasternAsianHispanicAfrician-AmericanCaucasianFrequency of Family Communication (School)7.06.05.04.03.02.01.0 3.64.24.14.95.05.3 Figure 4-12. Ethnicity in Frequency of Family Communication When at School Caucausion participants indicated that they communicated with a family member more frequenlty than Asian and Middle-Eastern particiapants when at home (see Figure 4-13). This finding is statistically significant ( F = 18.64, p < .01). This finding is unexpected, as traditional cultural research finds Middle-Easterners as allocentric and in frequent communication with their family and ingroup (Triandis et al., 1986). This reseacher attributes this finding to the level of independence the Middle-Eastern students would have to have in order to travel to the United States for school.

PAGE 109

97 EthnicityOtherMiddle-EasternAsianHispanicAfrician-AmericanCaucasianFrequency of Family Communication (Home)7654321 744677 Figure 4-13. Ethnicity in Frequency of Family Communication When at Home There was a significant correlation between four of the five culture factors and participants’ communication with a family member while at school (see Table 4-32). There was a negative correlation between the Dependence/Reliance factor, the Family Integrity factor, and the Community vs. Individual factor with the frequency of family communication at school. This negative correlation is attributed to content of the three factors. Each of the negatively correlated factors encompasses family and community-orientated concepts. Thus, a person who rates high on these factors is more likely to communicate at a different rate than those who rate high on the Group vs. Self and Closeness to Ingroups factors.

PAGE 110

98 Table 4-32. Correlation Between Culture Factors and Communication with Family Member While at School ________________________________________________________________________ When I’m at School I Communicate with a Family Member ________________________________________________________________________ Pearson Significance N Factor 1 Dependence/Reliance -.15 .01 204 Factor 2 Group vs. Self .19 .01 204 Factor 3 Closeness to Ingroups .20 .01 204 Factor 4 Family Integrity -.16 .02 204 Factor 5 Community vs. Individual ID -.08 .13 204 There was a correlation between the five culture factors and participants’ communication with a family member when at home (see Table 4-33). As revealed earlier, the means for the Dependence/Reliance factor and the Family Integrity factor were split – with one item rating very idiocentric and one item rating very allocentric. This researcher asserts that this variance in each of the factors caused the negative correlation between the frequency of family communication at home and these two variables.

PAGE 111

99 Table 4-33. Correlation Between Culture Factors and Communication with Family Member While at Home ________________________________________________________________________ When I’m at Home I Communicate with a Family Member ________________________________________________________________________ Pearson Significance N Factor 1 Dependence/Reliance -.36 .01 204 Factor 2 Group vs. Self .13 .03 204 Factor 3 Closeness to Ingroups .13 .03 204 Factor 4 Family Integrity -.40 .01 204 Factor 5 Community vs. Individual ID .25 .01 204 Internet Use This research study based it’s assumptions about Internet usage on existing data from the Pew Research Center (2000a). This researcher applied the traditional idiocentric/allocentric culture label to the ethnic Internet usage data from the Pew study. There was no significant correlation between any of the five culture groups and how often each used the Internet (see Table 4-34).

PAGE 112

100 Table 4-34. Correlation Between Culture Factors and Frequency of Internet Usage ________________________________________________________________________ Internet Usage ________________________________________________________________________ Pearson Significance N Factor 1 Dependence/Reliance -.20 .38 204 Factor 2 Group vs. Self .03 .33 204 Factor 3 Closeness to Ingroups -.05 .26 204 Factor 4 Family Integrity .10 .07 204 Factor 5 Community vs. Individual ID .09 .08 204

PAGE 113

CHAPTER 5 DISCUSSION Summary Data from this research suggest that tailoring messages through either culturally targeted messages or levels of interactivity does not have a significant effect on the likelihood that receives would adopt the proposed preventive health behavior (PHB). The null hypothesis was rejected and there was no difference in the likelihood of adopting the PHB when culture as matched with the message or level of interactivity. Based on one study, it is not possible to tell if culture has no bearing in this Interactive Age or if the manipulation was not strong enough. The hypotheses were near significance in support of accepting the null for the Community vs. Individual Identification factor. That factor accessed participant sense of community and nationalism. The first hypothesis, which submitted that a message match could predict behavior, failed when more participants in the unmatched message condition desired to be contacted than those who received matched messages. As the post-hoc analysis revealed, participants’ cultural categorization for this variable nearly opposite of how existing research categorizes people based on ethnicity. This researcher submits that this “opposite effect” is due to the terrorist attacks. During the time this data was collected, there was an increased scrutiny in the American media regarding Middle-Easterners, which may have lead to them rating themselves as more idiocentric in this factor with less nationalism and sense of community. When testing the second hypothesis for the same factor, the data completely reversed in that more participants in 101

PAGE 114

102 the matched level of interactivity desired to be contacted than those in the unmatched. This finding is attributed to the role Trustworthiness plays in predicting intended behavior. A post-hoc analysis revealed that participants in a matched level of interactivity rated source Trustworthiness higher at a statistically significant level. Although source credibility is central to the Health Belief Model (HBM; Rosenstock, 1974; Janz & Becker, 1984), it was not an expected indicator in this research study. A one-way anova test found a main effect on recall in the Closeness to Ingroups culture factor. This factor included items that measured one’s closeness to the familial ingroup and honor issues associated with outside praise of a family member. If participants were culturally matched with the message in this factor, recall was significantly greater. When idiocentrism/allocentrism was defined by the Family Integrity factor, participants who rated Shands at the University of Florida low in Trustworthiness and were exposed to a matched level of interactivity had a statistically significant greater recall. In fact, this group reported a greater recall than even those who felt Shands at the University of Florida was highly Trustworthy. It appears that in this case, matched level of interactivity is key in predicting recall. This recall can later translate into a cue to action if exposed to like messages again. Limitations Measurement The data in this research study were reliable in terms of the criteria of internal consistency. This is evident based on the acceptable alpha levels in all of the culture factors and the source credibility factors. Construct validity is an issue in this study because the constructs found in this study varied from the Triandis et al. (1986) findings.

PAGE 115

103 Design There is limited measurement validity in this study. The validity is strengthened by the use of pre-existing and previously tested scales. However, this study found Triandis et al (1986) family integrity factor culture scale limited and that it yielded unreliable results. Because the hypotheses in this study were not supported, there are no effects to generalize. Thus there is no external validity to this study. Because of the small size of each experimental condition, this study has limited power in predicting behavior. A future replication of this study would benefit from more participants and a better representation of different ethnic groups. Scale Limitations The use of the Triandis et al. (1986) scale may have been a major limitation in this study. This study used the scale to gauge cultural variations within the United States; however, the scale originally was used on “pure” cultural participants. Originally, the scale was used to compare people from different cultures to one another, such as Chinese respondents living in China against Australian respondents living in Australia. On the other hand, this study attempted to gauge the level of idiocentrism or allocentrism across American sub-cultures, such as Asian-Americans, African-Americans, Hispanic-Americans, etc. Another limitation concerning the Family Integrity Factor dealt with the way Triandis et al. (1986), created culture factors. Triandis and his colleagues did not create factors based on the items that loaded highly with one another. Instead, the original scale combined items based on face value of what the item intended to investigate. Because of this analysis method, the scale becomes very difficult to rely upon when later data collected through the scale yield different results.

PAGE 116

104 Finally, the Triandis scale measured cultural variation nearly two decades ago. It is possible that people have changed since the scale was created. This nation has seen two wars and other cultural impacts since the scale was created. Based on the current data, this researcher attests that the scale is no longer valid; its effectiveness in gauging cultural differences should be further evaluated. A further examination of the Triandis scale among sub-cultures is needed to assess if this scale is properly used to measure culture among such a homogenous group. Subject Limitations University of Florida college students may not fully represent their respective cultures. The University of Florida is a major southern university with over 44,000 students, many of whom travel from outside of Gainesville to attend classes at the institution. While there is a range of socio-economic statuses in the student body, the average cost of attending the University of Florida is $19,170 for out-of-state undergraduates and $23,230 for out-of-state graduate students. The economic impact of tuition, transportation, and housing is great on any family. This different status may affect the cultural values and differentiate students from the rest of the general population, which limits the generalizability of this study. Exposure to the message was controlled, but there was no way to determine how many of the participants exposed to the messages would have sought out the information on their own. Additionally, there is no way to determine if those who would seek out the information would then be more willing to participate in the proposed PHB. In a real life setting, health information seekers would seek and then read only the pages that contained content covering self-selected health topics. Thus, the external validity of the findings in this study is limited.

PAGE 117

105 Terrorist Attacks The data for this study were collected around the one-month anniversary of the worst terrorist attack against the United States in the nation’s history. The nation’s slow recovery from this tragedy may have affected the results of this study. During the data collection period, the media bombarded the public with information on the attacks, the rescue process, other terrorist threats, and the looming promise of an eminent war. Since the terrorist attacks, it appears the nation has become more unified. Immediately following the attacks, Congress set aside party lines and unanimously passed legislation to get the country on the road to recovery. A noticeable sense of nationalism and feeling of the “American community” emerged. While anecdotal, businesses and homes across the nation displayed flags. In fact, several national retailers sold out of American flags because the demand to purchase them so rapidly rose after the attacks. In fact, this researcher believes that the terrorist attacks did affect the results in that Middle-Eastern and Asian participants ranked much more idiocentric in the Community vs. Individual Identification factor which measure participants’ sense of community and nationalism. It became very important for these two culture groups in the United States to separate themselves from their nationality and connect with other Americans at this time. These groups also may have felt additional pressures to create barriers between their traditional communities in order to function without the threat of violence or harm against them. These ideas are supported by media reports of an increase of attacks against the ethnic groups within the United States, particularly the south.

PAGE 118

106 Anthrax Over West Nile Virus During the weeks the data were collected, one person was diagnosed with and died from exposure to anthrax in the state of Florida. The media reported that several others in the deceased’s office had been exposed to this deadly biological weapon. The exposure of people in Florida became the first of many throughout the Eastern United States, including reported exposure in the nation’s capital, Washington D.C. Many of the participants in this study were journalism school students, who may have paid more attention to the reports of anthrax in Florida and beyond. Additionally, the first cases of anthrax exposure were targeted at media outlets, outlets at which these students could one day work. While it is anecdotal evidence, several participants commented that they would have been more interested in a health message regarding the development of an anthrax vaccine, as opposed to one for West Nile Virus. While West Nile Virus and mosquito alerts continued during this time frame in the local media, these messages and media cues were over-shadowed by those related to the threat of anthrax. Prior to administering the health message stimulus in the experiment, this researcher was concerned as to the relevancy of West Nile Virus in light of recent events. In the days following the terrorist attacks, concern and relevancy of many public issues lessened. However, there were no viable alternatives to using West Nile Virus as the message stimulus. If the researcher chose another disease or virus, it would not have differed from the presented threat of the West Nile Virus. If this researcher had changed the health message stimulus to inform participants about a biological or chemical weapon, such as small pox or anthrax, the ability to replicate the findings of this study may have been limited. This is believed to be the case because a replication of such a study a year later may have produced entirely different findings when the perceived

PAGE 119

107 threat of anthrax was not as high. This researcher believes that in a time when the public was overly interested in and fearful of such health problems, the study’s findings would have yielded inaccurate results. Had this researcher changed the health topic/message stimulus to anthrax, it would have resulted in an inflated level of interest in the vaccination. The purpose of this study was not to test public interest in medical information, but the effects of message stimulus through interactive delivery. Conclusions This research study attempted to find a relationship between complex issues such as culture, source credibility and the level of interactivity. This jump into tying all of these concepts together resulted in finding limited significant relationship findings. Thus, future research must forge forward more slowly. Important lessons such as knowing one’s audience and culture when targeting a message became an important theme in the findings that did yield statistical significance. For example, the data from this study shows that health communicators should access the level of Trustworthiness of their message source and culture type prior to disseminating a message in order to ensure it’s impact is as full as possible. Additionally, while the purpose of a health message is typically to persuade one to undertake a particular behavior, health communicators should react positively to the discovery of ways to increase recall as demonstrated in this study. If a message recipient is able to recall the health information, the health communicator is assured that the message got through. Even though it may not inspire a message recipient to adopt the health goal, it can serve later as a cue to action. This finding supports that aspect of the HBM.

PAGE 120

108 Culture Culture played an integral role in this research study. This research study hypothesized that culture would be the deciding factor as to whether or not a health message was effective. While the hypothesis was not supported, other interesting effects emerged from the data. The Closeness to Ingroups culture factor displayed a main effect when culturally matched messages resulted in a significantly higher recall. This direct relationship is key in that health communicators can increase recall by simply phrasing the verbiage in a culturally matched way. One interesting relationship looked at the role the level of interactivity and Trustworthiness play in recall for the Family Integrity factor. This factor identified participants’ beliefs about continued cohabitation between parents and adult children. When Trustworthiness was high between the matched and unmatched level conditions, recall was nearly equal. Yet, when Trustworthiness is low, participants rely on the level of interactivity match in order to recall more. Thus, if health communicators can properly identify those who find them low in Trustworthiness prior to disseminating a message, and can match their level of interactivity, communicators can increase recall for this group. The breakdown of ethnicity across culture factors illustrated interesting differences between groups. In some of the culture factors, traditional cultural expectations were met. However, some factors resulted in completely opposite effects. For example, for the Dependence/Reliance culture factor and Family Integrity factor participants’ cultural responses were in line with what traditional suggests. Yet, in the Group vs. Self factor, which identified the desire to out-perform, not cooperate and desire to work alone, Caucasians ranked more allocentric than Middle-Eastern participants. While this finding

PAGE 121

109 was not statistically significant, it is probably due to the fact that many undergraduate classes rely on group projects and working together. This researcher submits that Middle-Eastern participants may have ranked surprisingly idiocentric due to their assimilation into this idiocentric American culture. Another finding contrary to traditional culture research found Asians the most idiocentric in the Closeness to Ingroups factor. However, this finding was not statistically significant. A particularly interesting breakdown of ethnicity in the Community vs. Individual Identification factor, which measured sense of community and nationalism, found Asians and Middle-Eastern participants the most idiocentric at a statistically significant level. Again, this researcher attributes this transposed image of traditional culture to the events of Sept. 11, 2001 and the use of the inaccurate Triandis culture scale. While culture did not necessarily play a predictive role in the affect of a HBM message, more than 60% of all the participants in the study desired to be contacted after exposure to the message. Additionally, while message recall was not as desired as actual behavior, this relationship remains an important finding. Health communicators are one step closer to reaching the target audience by knowing how they can influence recall. Source Credibility Source credibility played an extremely important role in the recall of messages received interactively. Additionally, recall becomes predictable when Trustworthiness is high and the level of interactivity is matched. This significant finding can help health communicators positively impact message recall if they determine and match culture when delivering a message. This effect is found in the Family Integrity culture factor. When participants rated Shands at the University of Florida Trustworthiness high, their recall level is similar, regardless of matched or unmatched level of interactivity.

PAGE 122

110 However, when participants rate Shands at the University of Florida Trustworthiness low, the matching level of interactivity is key in predicting recall. This is important because it says that matching level of interactivity is key for those who will find the source low on Trustworthiness in source credibility. Thus, if health communicators know a certain culture or demographic is more likely to rate Trustworthiness in source credibility low, then it would behoove the communicator to match the level of interactivity in message exposure. Health Belief Model The constructs of the HBM were used to create health communication messages in this study. Several of the assumptions central to the HBM have a high degree of importance. For instance, Trustworthiness in the source credibility can have a positive influence on retention and intended behavior. According to the findings of this study, the cross-cultural implications of the HBM remain limited in this Interactive Age. Neither of the hypotheses, which suggested a positive behavior-orientated outcome if the message appealed to culture, were supported. Thus, this researcher ascertains that messages built using traditional HBM constructs are limited in effectiveness. While a positive behavior outcome is not always produced, this study did show an important link between recall and culture and Trustworthiness. Even though this was not the intended result, health communicators are breaking through and getting the health message across to recipients. This increased retention about a health topic could become a media cue to action (a construct of the HBM) and pave the way for a future message to be effective. Internet as an Information Source Early research found that “people who have health concerns are more likely to consult a specialized print medium or health professional” (Hofstetter et al., 1992, p. 268). Now

PAGE 123

111 that more specialized information sources are available, this researcher finds it a natural transition that more people will turn to the Internet for health information. While current technology is limited and has not yet completely saturated the market, the coming decade will find a synthesis of: 1) The convenience of the Internet, 2) The human qualities of interpersonal communication, 3) The static record print offers, and 4) The drama and personalization of television and radio. Current research indicates that the Internet can function as a social-support system as well as a trusted information resource, filling in the gaps left through traditional interpersonal health communication with providers (Johnson et al., 1992). Much of the technology to bridge these gaps exists but is not yet widely used by the public. The addition of Web cameras (Web-cams) can provide patients with direct links to health care professionals and increase patients’ quality of care. When the patient has the power to speak directly to medical experts all over the world, the patient becomes empowered with knowledge and access to additional medical teams. By speaking directly with the physician through a Web-cam, the patient is able to receive immediate feedback, pick up on emotional cues, and interpret non-verbal communication. Additionally, the social interaction sought by many health information seekers can be provided easily through online support groups (Mittman & Cain, 2001). Researchers predict that future online health care messages will include anecdotal social support through patient testimonials (Mittman & Cain, 2001). With this trend in mind, this writer believes that health care institutions soon may offer these real-life patient success stories alongside disease descriptions, symptoms and treatment options.

PAGE 124

112 Not only is a vast population of people going online to find health information, these information-seekers typically share their findings with others (Aspden & Katz, 2001). This sharing could be a result of people looking for health information online for others (Pew Research Center, 2000a). This researcher speculates that this type of sharing is actually a result of acquiring information for others and as an opportunity to validate the credibility of the acquired information from the Internet. Since the advent of the Internet, it has grown into a powerful medium for message dissemination. Because of this, it is more important than ever to investigate the implications of this new information source and test the effectiveness of health messages on it. Many users are turning to the Internet for both sick-role information as well as PHB suggestions. More specifically, it is important to understand what the HBM would predict for encouraging PHB through the Internet across various cultures. Because health seeking can be done anonymously online, patients are able to investigate fully a variety of health topics they may be embarrassed about discussing with a physician or other reputable resource, and they may have questions they would not ask of a person. Yet, different cultures react in a variety of ways to the personalization of information. Additionally, it is more important than ever to word messages in neutral language to maximize the message’s effectiveness across many diverse cultures. If a culture is more community-centered, the reader will not respond to text that promotes the self-importance of a health issue. Future Research This research attempted to tie together three very different and well-researched worlds: the HBM, the effect of culture on the success of a message, and the Internet. Whether future research investigates the role of the HBM, how to induce positive PHB,

PAGE 125

113 what role culture plays in the use of the Internet, or how people use the Internet to find answers to their health questions, research in all of these areas must continue. Researchers must be have more information and become better equipped to make more accurate predictions about the roles of culture and Internet use in successful health messages. Even though the predicted results in this study were not significant, that finding in itself is a breakthrough for the field. While this study failed to show a link between culturally matched messages and behavior, it did show that the predictions for a non-Web world do not hold true in the new interactive age. As the HBM nears its third decade in health communication, messages, more research should be done to test its relevance in a digital world. One of the assumptions of the HBM is that message recipients are interested in the message. Because Internet-delivered messages are often actively sought, according to Pew research, it seems a natural that the HBM would extend to the Internet. Most online health information seekers are trying to expose themselves to specific health messages; this high level of interest in the health topic is the basis of the HBM’s traditional success. The world of the Internet could open new doors of possibility for communicators. It is important for future research to investigate which constructs are found to be more effective online and which constructs have limited success online. After the discovery of these HBM strengths and weaknesses, one should continue to further this body of knowledge and compare whether a particular construct is more effective in the traditional or online medium. Further, investigators should study which types of health messages are better communicated through these static and interactive media. For example, it

PAGE 126

114 would be interesting to find which medium health information seekers prefer when searching for complex disease information versus basic health and wellness tips. Future research regarding the implications of the HBM in an interactive age must be done. Future studies should carefully choose health topics that are both timely and relative to the study participants. Participants may have been more interested in the West Nile Virus prior to the terrorist attacks and bioterrorism threats occurring. One research study into this new field will not suffice to offer health communicators the “answer” to this emerging question. More research must be done to investigate which PHB messages are more appealing to different types of cultures, as well as how different cultures prefer to receive the messages. The field of Internet health still faces many big questions. Are some people more susceptible to Internet-delivered messages than others? Do cultures that would traditionally react to a PHB message react differently to the same message when it is delivered to them interactively? How likely are people to review Internet health messages, and how does this compare to traditional means of health information gathering? Because the Internet and like technologies offer the capability to tailor messages on an individual level, it is important that future research uncover exactly what makes a Web-based PHB message more attractive to consumers and how people differ in their preferences for these messages. In this study, this was attempted through appealing to one’s culture through message type and level of interactivity. Perhaps this is not the key. Future research should investigate the attractiveness and usability of a Web page to see if a certain “style” of page generates a predictable result. Researchers also should

PAGE 127

115 investigate the effectiveness of the truly tailored message – where the user enters information about himself and receives “targeted” information. An example of this is the ability to enter your own zip code on national news Web site and see your local weather and regional headlines. For health communication, this could mean offering a quiz in which the user provides lifestyle information and is then supplied with health messages based on his lifestyle needs and interest. This study only scratched the surface of needed online health research. While the findings of this research were not predicted, they remain an addition to the growing body of Internet health research on messages. Future research must continue to explore the relationship between culture and credibility with health messages.

PAGE 128

APPENDIX A ATTRIBUTES DEFINING INDIVIDUALISM AND COLLECTIVISM AND THEIR ANTECEDENTS AND CONSEQUENCES FROM TRIANDIS, MCCUSKER, & HUI (1990) Antecedents Attributes Consequences Individualism Affluence Emotional detachment from Socialization for self-reliance _______________________________________________________________________ Cultural complexity ungroup and independence Hunting/food gathering Personal goals have primacy Good skills when entering Upper social class over ingroup goals new groups Migration Behavior regulated by Loneliness Urbanism attitudes and Exposure to the mass cost-benefit analyses media Collectivism Unit of survival is food Family integrity Socialization for obedience ingroup Self defined in ingroup and duty Agriculture terms Sacrifice for ingroup Large families Behavior regulated by Cognition: Focus on ingroup norms on common elements Hierarchy and harmony with ingroup within ingroup members 116

PAGE 129

117 Ingroup is seen as Behavior: Intimate, saving homogeneous face, reflects Strong ingroup/outgroup hierarchy, social distinctions support, interdependence ________________________________________________________________________

PAGE 130

APPENDIX B INFORMED CONSENT FOR PROTOCOL HEALTH BELIEF MODEL IN AN INTERACTIVE AGE Purpose of the research study: The purpose of this study is to examine effectiveness of interactive messages. This research is being conducted for the master’s thesis of Kaye Trammell, a mass communication graduate student at the University of Florida. What you will be asked to do in the study: You will be asked to fill out a survey about Shands at the University of Florida, complete another survey indicating how you feel about the various statements presented, and read a health message. You do not have to answer any questions that make you feel uncomfortable. You will be allowed to take the debriefing paperwork home with you. Time required: 1 hour Risks and Benefits: There are no direct benefits to the researcher for this study. However, the field of mass communication will benefit from the knowledge gained by the study. The only possible benefits a participant in the study would receive would be extra credit, if his or her professor chooses to give extra credit. Compensation: The researcher will not compensate you in anyway for participation in this study. Confidentiality: Data will be anonymously collected. Voluntary participation: Your participation in this study is completely voluntary. There is no penalty for not participating. Right to withdraw from the study: You have the right to withdraw from the study at anytime without consequence. Whom to contact if you have questions about the study: Kaye Trammell, kaye@trammell.com or Dr. Mary Ann Ferguson, College of Journalism and Communications, P.O. Box 118400, Gainesville, FL 32611, ph. (352) 392-6660 Whom to contact about your rights as a research participant in the study: UFIRB Office, Box 112250, University of Florida, Gainesville, FL 32611-2250; ph 392-0433. Agreement: I have read the procedure described above. I voluntarily agree to participate in the procedure and I have received a copy of this description. Participant: ___________________________________________ Date: _________________ Principal Investigator: _ Kaye Trammell____ ________________ Date: __ Sept. 21, 2001 _____ 118

PAGE 131

APPENDIX C SOURCE CREDIBITLIY SCALE Instructions: On the scales below, please indicate your beliefs about Shands at the University of Florida and its doctors. Shands at UF is the academic medical center of UF. Circle the number between the adjectives which best represents your beliefs about Shands at UF and its doctors. Numbers ” and ” indicate a very strong feeling. Numbers ” and ” indicate a strong feeling. Numbers ” and ” indicate a weak feeling. Number ” indicates you are undecided or do not understand the adjectives themselves. Please work quickly. There are no right or wrong answers. Do you think Shands at UF and its doctors are . _______________________________________________ Reliable 1 2 3 4 5 6 7 Unreliable Uninformed 1 2 3 4 5 6 7 Informed Unqualified 1 2 3 4 5 6 7 Qualified Intelligent 1 2 3 4 5 6 7 Unintelligent Valuable 1 2 3 4 5 6 7 Worthless Inexpert 1 2 3 4 5 6 7 Expert Honest 1 2 3 4 5 6 7 Dishonest Unfriendly 1 2 3 4 5 6 7 Friendly Pleasant 1 2 3 4 5 6 7 Unpleasant Selfish 1 2 3 4 5 6 7 Unselfish Awful 1 2 3 4 5 6 7 Nice Moral 1 2 3 4 5 6 7 Immoral 119

PAGE 132

APPENDIX C FAMILY INTEGRITY FACTOR Read the statements below. Please indicate the level you agree or disagree with the statement. Selecting ” means that you very strongly agree with the statement. Selecting ” means that you very strongly disagree with the statement. Please read the questions quickly and respond with your first impression. 1. One should live one’s life independently of others as much as possible. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 2. I would rather struggle through a personal problem by myself than discuss it with my friends. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 3. I would help within my means, if a relative told me that s(he)is in financial difficulty. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 4. I like to live close to my good friends. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 5. The most important thing in my life is to make myself happy. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 6. It is important to me that I perform better than others on a task. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 7. I tend to do my own thing, and others in my family do the same. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 8. Aging parents should live at home with their children. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 9. What I look for in a job is a friendly group of coworkers. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 10. Children should live at home with their parents until they get married. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 11. One does better work alone than in a group. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 120

PAGE 133

121 12. Individuals should be judged on their own merits, not on the company they keep. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 13. When faced with a difficult personal problem, it is better to decide what to do yourself, rather than follow the advice of others. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 14. It does not matter to me how my country is viewed in the eyes of other nations. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 15. I enjoy talking to my neighbors everyday. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 16. I can count on my relatives for help if I find myself in any kind of trouble. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 17. What happens to me is my own doing. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 18. If the group is slowing me down, it is better to leave it and work alone. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 19. Even if a child won the Nobel Prize, the parents should not feel honored in any way. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 20. Children should not feel honored even if the father were highly praised and given an award by a government official for his contributions and services to the community. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 21. In most cases, to cooperate with someone whose ability is lower than yours is not as desirable as doing the thing on your own. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 22. It doesn’t matter to me how my community is viewed in the eyes of other communities. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree

PAGE 134

APPENDIX E CULTURE MESSAGES Idiocentric Message Do you enjoy being outdoors in the early evening? Do you rarely wear insect repellent when you're outdoors? Do you forget to dispose of stagnant water outside of your home to keep mosquitoes away? If you answered “yes” to any of these questions, you are at risk for West Nile Virus. A threat to your life The international virus has been found throughout Africa, Asia and the Middle East. After moving into the United States in 1999, the virus has spread from New York through the country. The virus has now taken a stronghold in the state. To date, cases of the West Nile Virus are reported in 10 Florida counties and three people have been infected this summer. The swamp-like climate of Florida makes you especially susceptible to contracting the West Nile Virus. Mosquitoes transmit this disease to you through bites. You can contract the virus when an infected mosquito bites you. Once you're infected, the virus leads to encephalitis, or swelling of the brain. If you don't treat the virus, you could die. Symptoms Symptoms generally occur about 15 days after you are bitten by a West Nile Virus infected mosquito. If you are infected with the West Nile Virus, you can experience: Fever Headache Rash Altered mental state or stupor Muscle convulsions, weakness or stiffness Vaccination University of Florida researchers are working on a vaccine for the virus. You may be able to use this vaccination as a preventive measure, protecting you from contracting the West Nile Virus. When the vaccine is available, Shands at UF is considering making it available to you free of charge. 122

PAGE 135

123 Protect yourself In addition to your free vaccination at Shands at UF, you can reduce your chances of contracting West Nile Virus by: Wearing insect repellent with at least 30% DEET Staying indoors at dusk when the mosquito population feeding is at its highest Wearing long sleeves and long pants when you go outside Disposing of stagnate water, possible mosquito breeding grounds Sign up today Sign up to receive information about the vaccine after its developed. To be contacted by Shands at UF to find out how you can receive your free vaccine, please check the box below. Yes, I’d like to be contacted to receive information about the vaccine. No, I don’t want to be contacted about the vaccine.

PAGE 136

124 Allocentric Message Do you and your loved ones enjoy being outdoors in the early evening? Do they rarely wear insect repellent when they’re outdoors? Does your family forget to dispose of stagnant water outside of the home to keep mosquitoes away? If you answered “yes” to any of these questions, then the people you care about are at risk for West Nile Virus. A threat to your community The international virus has been found throughout Africa, Asia and the Middle East. After moving into the United States in 1999, the virus has spread from New York through the country. The virus has now taken a stronghold in the state. To date, cases of the West Nile Virus are reported in 10 Florida counties and three people have been infected this summer. The swamp-like climate of Florida makes your loved ones especially susceptible to contracting the West Nile Virus. Mosquitoes transmit this disease to your community through bites. Your loved ones can contract the virus when an infected mosquito bites them. Once they’re infected, the virus leads to encephalitis, or swelling of the brain. If the virus isn’t treated, they could die. Symptoms Symptoms generally occur about 15 days after being bitten by a West Nile Virus infected mosquito. If someone you care about is infected with the West Nile Virus, he or she can experience: Fever Headache Rash Altered mental state or stupor Muscle convulsions, weakness or stiffness Vaccination University of Florida researchers are working on a vaccine for the virus. Your community may be able to use this vaccination as a preventive measure, protecting them from contracting the West Nile Virus. When the vaccine is available, Shands at UF is considering making it available to your community free of charge. Protect your loved ones In addition to the free vaccination at Shands at UF, you can reduce your love one’s chances of contracting West Nile Virus by: Wearing insect repellent with at least 30% DEET

PAGE 137

125 Staying indoors at dusk when the mosquito population feeding is at its highest Wearing long sleeves and long pants when outside Disposing of stagnate water, possible mosquito breeding grounds Sign up today Sign up to receive information about the vaccine after its developed. To be contacted by Shands at UF to find out how the people you care about can receive the free vaccine, please check the box below. Yes, I’d like to be contacted to receive information about the vaccine. No, I don’t want to be contacted about the vaccine.

PAGE 138

APPENDIX F DEMOGRAPHIC AND INTERNET USE SURVEY Instructions: Please complete the following phrases about yourself. I am: Male Female I am in my ________ year of college: 1 st 2 nd 3 rd 4 th 5 th or more I am ____ years old. Race: Caucasian African-American Hispanic Asian Middle-Eastern Other When I’m at school, I communicate with a member of my family: (check one) ____ Several times a day ____ At least once a day ____ Several times a week ____ At least once a week ____ At least once a month ____ Rarely ____ I don’t communicate with anyone in my family When I’m at home, I communicate with a member of my family: (check one) ____ Several times a day ____ At least once a day ____ Several times a week ____ At least once a week ____ At least once a month ____ Rarely ____ I don’t communicate with anyone in my family 126

PAGE 139

127 I use the Internet: (check all that apply) ____ For school ____ For communication (e-mail, instant message, chat) ____ For work ____ To play games ____ To research topics I’m interested in ____ To find out more information about my hobbies ____ I don’t use the Internet I’m on the Internet: (check one) ____ Several times a day ____ At least once a day ____ Several times a week ____ At least once a week ____ At least once a month ____ Rarely ____ I have never been on the Internet

PAGE 140

APPENDIX G INFORMATION RETENTION AND COMPREHENSION SURVEY Instructions: Please read the following statements and answer the questions to the best of your ability. Your responses will be kept anonymous. 1. West Nile Virus is a type of food poisoning. True False 2. How is West Nile Virus transmitted? _____________________________ 3. West Nile Virus was first reported in the United States in ______________. (name state) 4. I am in danger of contracting West Nile Virus. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 5. I think the West Nile Virus vaccination can protect my health. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 6. Since 1999, more than 100 people in Florida have contracted the West Nile Virus. True False 7. In the future, I will protect myself from West Nile Virus. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 8. Symptoms of West Nile Virus occur after _______ days. 9. List three ways you can avoid contracting the West Nile Virus: _____________________________________________________________________ _____________________________________________________________________ _____________________________________________________________________ 128

PAGE 141

129 10. After reading the information about West Nile Virus, I signed up to for the vaccine. True False 11. I plan on receiving the vaccination for West Nile Virus to protect me from infection. Very strongly agree 1 2 3 4 5 6 7 Very strongly disagree 12. The message about West Nile Virus emphasized a risk to: Me Others like me

PAGE 142

APPENDIX H DEBRIEF STATEMENT Thank you for participating in this research study. The purpose of this study was to investigate the relevance of culture when reviewing health communication messages. Today you learned about the West Nile Virus. There is not currently a human vaccine for West Nile Virus. University of Florida researchers are not developing a human vaccine. The information that you read about the vaccine was false. Your name and contact information was not gathered in this study. You will not be contacted by Shands at the University of Florida if a human vaccine is developed for West Nile Virus. You can protect yourself from West Nile Virus by: Wearing insect repellent with at least 30% DEET Staying indoors at dusk when the mosquito population feeding is at its highest Wearing long sleeves and long pants when you go outside Disposing of stagnant water, possible mosquito breeding grounds Thank you for participating in the study! If you have any questions about this study, please contact the Kaye Trammell, primary researcher, at (352) 392-6660 or via e-mail at kdt@ufl.edu. Kaye Trammell is a graduate student of Dr. Mary Ann Ferguson, phone (352) 392-6660, e-mail mferguson@jou.ufl.edu. 130

PAGE 143

LIST OF REFERENCES Aspden, P., & Katz, J. E. (2001). Assessments of quality of health care information and referrals to physicians: A nationwide survey. Chapter in The Internet and Health Communication: Experiences and Expectations , ed. Ronald E. Rice and James E. Katz. Thousand Oaks, CA: Sage Publications. Pp. 99-106. Bakker, A. B., Buunk, B. P., Siero, F. W., & Van Den Eijnden, R. J. J. M. (1997). Application of a modified health belief model to HIV preventive behavioral intentions among gay and bisexual men. Psychology and Health, 12, 481-492. Chaiken, S., & Eagley, A. H. (1976). Communication modality as a determinant of message persuasiveness and message comprehensibility. Journal of Personality and Social Psychology, 34, 605-614. Chew, F., Palmer, S., & Kim, S. (1998). Testing the influence of the health belief model and a television program on nutrition behavior. Health Communication, 10 (3), 227-245. Daniels, M. J., & Parrott, R. L. (1996). “Prenatal care from the women’s perspective: A thematic analysis of the newspaper media” chapter in Evaluating Women’s Health Messages , ed. Roxanne Louiselle Parrot and Celeste Michelle Condit. Thousand Oaks, CA: Sage Publications. Pp. 222-233. DePietro, R., & Clark, N. M. (1984). A sense-making approach to understanding adolescents’ selection of health information source. Health Education Quarterly, 11 (4), 419-430. Dervin, B. (1989). Audience as listener and learner, teacher and confidante: The sense-making approach. Chapter in Public communication campaigns (2 nd ed.), ed. R.E. Rice & C.K. Atkin. Newbury Park, CA: Sage. Pp. 67-86 Eastin, M. S. (2001). Credibility assessments of online health information: The effects of source expertise and knowledge of content. Journal of Computer-Mediated Communication, 6( 4). Retrieved August 2001 from the World Wide Web: http://www.ascusc.org/jcmc/vol6/issue4/eastin.html . Ferguson, T. (1997). Health online and the empowerment of the medical consumer. Joint Committee Journal on Quality Improvement, 23 (5), 251-257. Festinger, L. (1957). A theory of cognitive dissonance. Evanston, IL: Row Peterson. 131

PAGE 144

132 Flemings, R., Leventhal, H., Glynn, K., & Ershler, J. (1989). The role of cigarettes in the initiation and progress of early substance use. Addictive Behavior, 14, 261-272. Fulton, J. P., Buechner, J. S., Scott, H. D., DeBuono, B. A., Feldman, J. P., Smith, R. A., & Kovenock, D. (1991). A study guided by the health belief model of the predictors of breast cancer screening of women ages 40 and older. Public Health Reports, 106 (4), 410-421. Griffin, R. J., & Dunwoody, S. 2000. The relation of communication risk to judgement and preventive behavior related to lead in tap water. Health Communication, 12 (1), 81-107. Hallal, J. C. (1982). The relationship of health beliefs, locus of control, and self concept to the practice of breast self-examination in adult women. Nursing Research, 31, 137-142. Harris & Associates. (1999). Sixty million seek health info online in the US . Retrieved December 1999 from the World Wide Web: http://www.nau.ie.surveys . Hawkins, R. P., Gustafson, D. H., Chewning, B., Bosworth, K., & Day, P. M. (1987). Reaching hard-to-reach populations: Interactive computer programs as public information campaigns for adolescents. Journal of Communication, 37 (2), 8-28. Henry, J. P., & Stephens, P. M. (1977). Stress, health, and social environment. New York: Springer. Hochbaum, G.M. (1958). Public participation in medical screening programs: A sociopsychological study. PHS publication no. 572. Washington, D.C.: U.S. Government Printing Office. Hofstede, G. (1980). Culture’s consequences. Beverly Hills, CA: Sage. Hofstetter, C. R., Schultze, W. A., & Mulvihill, M. M. (1992). Communications media, public health, and public affairs: Exposure to multimedia community. Health Communication, 4( 4), 259-271. Izenberg, N., & Lieberman, D. A. (1998a). The Web, communication trends, and children’s health – Part 2: The Web and the practice of pediatrics. Clinical Pediatrics, 37 (4), 215-221. Izenberg, N., & Lieberman, D. A. (1998b). The Web, communication trends, and children’s health – part 3: The Web and health consumers. Clinical Pediatrics, 37 (5), 275-285.

PAGE 145

133 Janz, N. K., & Becker, M. H. (1984). The health belief model: A decade later. Health Education Quarterly, 11 (1), 1-47. Jiang, J. J., Klein, G., & Vedder, R. G. (2000). Persuasive expert systems: The influence of confidence and discrepancy. Computers in Human Behavior, 16 (2), 99-109. Johnson, J. D., & Meischke, H. (1992a). Differences in evaluations of communication channels for cancer-related information. Journal of Behavioral Medicine, 15 (5), 429-445. Johnson, J. D., & Meischke, H. (1992b). Mass media channels: Women’s evaluations for cancer-related information. Newspaper Research Journal, 13 (1-2), 146-159. Johnson, J. D., & Meischke, H. (1993). A comprehensive model of cancer-related information seeking applied to magazines. Human Communication Research, 19 (3), 343-367. Johnson, J. D., Meischke, H., Grau, J., & Johnson, S. (1992). Cancer-related channel selection. Health Communication 4 (3), 183-196. Kegeles, S. S. (1963a). Some motives for seeking preventive dental care. Journal of American Dentistry Association, 67, 90-98. Kegeles, S. S. (1963b). Why people seek dental care: A test of a conceptual formulation. Health Human Behavior, 4, 166-173. Kloeblen, A. S. (1999). Folate knowledge, intake from fortified grain products, and periconceptual supplementation patterns of a sample of low-income pregnant women according to the Heath Belief Model. Journal of the American Dietetic Association, 99 (1), 33-39. Kyrouz, E. M., Holt, M., Mittman, R., & Everett, W. (1998). Twenty-first century health care consumers. Menlo Park, CA: Institute for the Future. Lindsay, G. B., & Rainey, J. (1997). Psychological and pharmacologic explanations of nicotine’s “gateway drug” function. Journal of School Health, 67 (4), 123-127. McIntosh, W. A., Kubena, K. S., Jiang, H., Peterson Usery, C., & Karnei, K. (1996). An application of the health belief model to reductions in fat and cholesterol intake. Journal of Wellness Perspectives, 12 (2), 98-107. Mittman, R. & Cain, M. (2001). The future of the Internet in health care. Chapter in The Internet and Health Communication: Experiences and Expectations , ed. Ronald E. Rice and James E. Katz. Thousand Oaks, CA: Sage Publications. Pp. 47-74.

PAGE 146

134 Napoli, P. M. (2001). Consumer use of medical information from electronic and paper media: A literature review. Chapter in The Internet and health communication: Experiences and expectations, ed. Ronald E. Rice and James E. Katz. Thousand Oaks, CA: Sage. Pp. 79-98. Nielsen, J. (2000). Designing Web Usability . Indianapolis, IN: New Riders Publishing. Parrott, R. L. (1995). Motivation to attend to health messages: Presentation of content and linguistic considerations. Chapter in Designing Health Messages: Approaches from Communication Theory and Public Health Practice , ed. E. Maibach & R. L. Parrott. Thousand Oaks, CA: Sage Publications. Pew Research Center. (2000a). The online healthcare revolution: How the Web helps Americans take better care of themselves. Report published by the Pew Internet and the American Life Project: Washington, D.C. Retrieved June 2001 from the World Wide Web: http://www.pewinternet.org . Pew Research Center. (2000b). African-Americans and the Internet. Report published by the Pew Internet & American Life Project: Washington, DC. Retrieved August 2001 from the World Wide Web: http://www.pewinternet.org . Pew Research Center. (2001). Asian-Americans and the Internet: They young and the connected. Report published by the Pew Internet & American Life Project: Washington, DC. Retrieved February 2002 from the World Wide Web: http://www.pewinternet.org . Pezza, P. E. (1990). Orientation to uncertainty and information seeking about personal health. Health Education, 21 (2), 34-36, 49. Povinelli, M., Remafedi, G., & Tao, G. (1996). Trends and predictors of human immunodeficiency virus antibody testing by homosexual and bisexual adolescent males, 1989-1994. Archives of Pediatrics & Adolescent Medicine, 150 (1), 33-39. Reagan, J., & Collins, J. (1987). Sources for health care information in two small communities. Journalism Quarterly, 64 (2-3), 560-563, 676. Rodgers, C. R. (1983). Diffusion of innovations. New York: Free Press. Rosenstock, I. M. (1974). Historical origins of the health belief model. Health Education Monographs, 2, 328-335. Rothman, A. J., Martino, S. C., Bedell, B. T., Detweiler, J. B., & Salovey, P. (1999). The systematic influence of gainand loss-framed messages on interest in and use of different types of health behavior. Personality and Social Psychology Bulletin, 25 (11), 1355-1369.

PAGE 147

135 Rothman, A. J., Salovey, P., Antone, C., Keough, K., & Martin, C. D. (1993). The influence of message framing on intentions to perform health behaviors. Journal of Experimental Social Psychology, 29, 408-433. Rubin, R. B., Palmgreen, P., & Sypher, H. E. (1994). Communication Research Measures: A Sourcebook. New York: Guilford Press. Rutledge, R. C. (2001). West Nile Virus: Risk and prevention for horses in Florida. Fact sheet published by University of Florida Extension Institute of Food and Agricultural Sciences (Fact Sheet ENY-651). Steers, W. N., Elliot, E., Nemiro, J., Ditman, D., Oskamp, S. (1996). Health beliefs as predictors of HIV-preventive behavior and ethnic differences in prediction. The Journal of Social Psychology, 136 (1), 99-110. Sutten, K. (2000). Marketing to the masses. Marketing Research, 12 (2), 42-43. Tardy, R. W., & Hale, C. L. (1998). Getting “plugged in”: A network analysis of health-information seeking among “stay-at-home moms.” Communication Monographs, 65 , 336-357. Triandis, H. C., Bontempo, R., Betancourt, H., Bond, H., Leung, K., Brenes, A., Georgas, J., Hui, H. C., Marin, G., Setiadi, B., Sinha, J., Verma, J., Spagenberg, J., Touzard, H., & de Montmollin, G. (1986). The measurement of the etic aspects of individualism and collectivism across cultures. Australian Journal of Psychology, 38 , 257-26. Triandis, H. C., McCusker, C., & Hui, H. (1990). Multimethod probes of individualism and collectivism. Journal of Personality and Social Psychology, 59 (5), 1006-1020. U.S. Department of Commerce (2002). A nation online: How Americans are expanding their use of the Internet. Economics & Statistics Administration, National Telecommunication and Information Administration: Washington, D.C. Widman, W. R., & Tong, D. A. (1997). Requests for medical advice from patients and families to health care providers who publish on the World Wide Web. Archives of Internal Medicine, 157 (2), 209-212. Wiener, M., & Mehrabian, A. (1968). Language within language: Immediacy, a channel in verbal communication. New York: Appleton-Century-Crofts. Witte, K., Peterson, T. R., Vallabhan, S., Stephenson, M. T., Plugge, C. D., Givens, V. K., Todd, J. D., Becktold, M. G., Hyde, M. K., & Jarrett, R. (1993). Preventing tractor-related injuries and deaths in rural populations: Using a persuasive health

PAGE 148

136 message (PHM) framework in formative evaluation research. International Quarterly of Community Health Education, 13, 219-251. Wohl, R. E., & Kane, W. M. (1997). Teachers’ beliefs concerning teaching about testicular cancer and testicular self-examination. Journal of School Health, 67 (3), 106-111.

PAGE 149

BIOGRAPHICAL SKETCH Kaye Trammell grew up alongside her two sisters, twins Shelly and Shelby, and graduated from Junction City High School in 1994. She attended Kansas State University, Johnson County Community College, and University of Missouri at Kansas City before she joined the United States Navy. On Jan. 2, 1996, Kaye left Kansas to live the life of a Sailor. After boot camp, Kaye attended the Defense Information School to learn her Navy trade: journalism. After training, she reported to the USS John F. Kennedy (JFK) in Mayport, Fla. While there, Kaye met her future husband, Mark R. Trammell, Jr., who coincidentally was the first person she interviewed as an official Navy journalist. She spent her 21 st birthday in Paris and enjoyed European port visits. In July 1997, she transferred to Fleet Surveillance Support Command in Chesapeake, Va., where she served until she was honorably discharged in Jan. 2001. In May 2001, she received a direct commission as a Public Affairs Officer in the Naval Reserves. Kaye has been Web Editor at Shands HealthCare since Jan. 2001. She looks forward to a long life in academia as she works toward her doctorate and aspires to be a professor investigating the way technology continues to change communication. Kaye holds an A.S. (general studies) from Tidewater Community College in Chesapeake, Va., a B.S. (communication) from Old Dominion University in Norfolk, Va., and a M.A.M.C. from the University of Florida in Gainesville, Fla. 137