<%BANNER%>

Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2011-05-31.

Permanent Link: http://ufdc.ufl.edu/UFE0022799/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2011-05-31.
Physical Description: Book
Language: english
Creator: Park, Jin
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

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

Notes

Statement of Responsibility: by Jin Park.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Weigold, Michael F.
Electronic Access: INACCESSIBLE UNTIL 2011-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0022799:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022799/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2011-05-31.
Physical Description: Book
Language: english
Creator: Park, Jin
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

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

Notes

Statement of Responsibility: by Jin Park.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Weigold, Michael F.
Electronic Access: INACCESSIBLE UNTIL 2011-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0022799:00001


This item has the following downloads:


Full Text

PAGE 1

1 EFFECTS OF CONSUMER MOOD STATES ON PROCESSING OF DISEASE INFORMA TI ON IN DIRECT TO CONSUMER ANTIDEPRESSANT ADVERTISING AND PERCEIVED FUTURE RISK OF DEPRESSION By JIN SEONG PARK A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE U NIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR PHILOSOPHY UNIVERSITY OF FLORIDA 200 9

PAGE 2

2 2009 Jin Seong Park

PAGE 3

3 To Hearan

PAGE 4

4 ACKNOWLEDG MENTS I thank my committee me mbers for their insights and advice. I especially thank my advisor and committee chair, Dr. Weigold. For many years Dr. Weigold has been a great mentor, teacher, role model, and motivator. I also thank Dr. Sutherland for teaching me how to live and teach a s a graduate student Dr. Treise for making the graduate program a great learning environment, and Dr. Chang Hoan Cho for leaving the footsteps for all Korean graduate students to follow in Without these people I would never have become who I am. I thank my family and my friends at the University of Florida for supporting and inspiring me over the last four years. Most of all, I extend very special thanks to my wife and life partner, Hearan Kim. Her patience, advice, kindness, and love inspire me to grow a s a teacher, researcher, and person.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURES ................................ ................................ ................................ ......................... 9 ABSTRACT ................................ ................................ ................................ ................................ ... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 11 History of Direct to Consumer Drug Advertising ................................ ................................ .. 11 Effects of DTC Advertising ................................ ................................ ................................ .... 13 Limitations of the Current Literature ................................ ................................ ...................... 15 Research Purpose ................................ ................................ ................................ .................... 18 Importance of Mood State ................................ ................................ ................................ ...... 22 Study Overview ................................ ................................ ................................ ...................... 23 Special Status of DTC Antidepressant Advertising ................................ ................................ 24 2 LITERATURE REVIEW ................................ ................................ ................................ ....... 26 Risk Perception ................................ ................................ ................................ ....................... 26 Mood State, Information Processing, and Judgment ................................ .............................. 28 Mood As Information ................................ ................................ ................................ ............. 33 Mood As Prime ................................ ................................ ................................ ....................... 37 Synthesizing the Two Perspectives ................................ ................................ ........................ 42 Perceived Diagnosticity of Inter nally Retrieved Life Experiences ................................ ........ 43 Hypotheses ................................ ................................ ................................ .............................. 46 3 METHODOLOGY ................................ ................................ ................................ ................. 53 Design ................................ ................................ ................................ ................................ ..... 53 Participants ................................ ................................ ................................ ............................. 54 Procedure ................................ ................................ ................................ ................................ 54 Independent Variables ................................ ................................ ................................ ............ 57 Mood State ................................ ................................ ................................ ....................... 57 Diagnosticity ................................ ................................ ................................ .................... 59 Opportunity ................................ ................................ ................................ ...................... 60 Stimulus ................................ ................................ ................................ ................................ .. 61 Development of the Stimulus ................................ ................................ ................................ 62 Pilot Study 1 ................................ ................................ ................................ ........................... 63 Pilot Study 2 ................................ ................................ ................................ ........................... 64 Dependent Variables ................................ ................................ ................................ ............... 65

PAGE 6

6 Perceived Future Risk of Depression ................................ ................................ .............. 65 Intentions to Seek Professional Help ................................ ................................ ............... 66 4 RESULTS ................................ ................................ ................................ ............................... 70 Manipulation Ch ecks ................................ ................................ ................................ .............. 70 Mood State ................................ ................................ ................................ ....................... 70 Perceived Diagnosticity of Discomforting Life Experiences ................................ .......... 72 Opportunity for Risk and Intention Estimation ................................ ............................... 73 Correlations Among Variables ................................ ................................ ............................... 74 Testing Hypotheses 1a and 1b ................................ ................................ ................................ 74 MANOVA Results for H1a and H1b ................................ ................................ .............. 75 Mixed MANOVA Results for H1a and H1b ................................ ................................ ... 76 Analyses of Simple Effects ................................ ................................ ............................. 77 Summary of the Test of H1a and H1b ................................ ................................ ............. 80 Testing Hypotheses H2a and H2b ................................ ................................ .......................... 81 MANOVA Results for H2a and H2b ................................ ................................ .............. 81 Mixed MANOVA R esults For H2a and H2b ................................ ................................ .. 82 An alyses of Simple Effects ................................ ................................ ............................. 83 Summary of the Tests of H2a and H2b ................................ ................................ ........... 85 Testing Hypotheses 3a and 3b ................................ ................................ ................................ 86 Test of H3a ................................ ................................ ................................ ...................... 87 Test of H3b ................................ ................................ ................................ ...................... 88 Additional Data Analyses ................................ ................................ ................................ 89 Summary of the Results ................................ ................................ ................................ .......... 92 5 DISCUSSION ................................ ................................ ................................ ....................... 102 Summary of Findings ................................ ................................ ................................ ........... 102 H1a and H1b ................................ ................................ ................................ .................. 102 H2a and H2b ................................ ................................ ................................ .................. 103 H3a and H3b ................................ ................................ ................................ .................. 103 Discussion of Findings ................................ ................................ ................................ ......... 104 Advertising Theory ................................ ................................ ................................ ........ 104 Advertising Practice ................................ ................................ ................................ ...... 106 Consumer Health ................................ ................................ ................................ ........... 113 Limitations of the Study ................................ ................................ ................................ ....... 116 Suggestions for Future Research ................................ ................................ .......................... 117 APPENDIX A INSTRUMENTAL MANIPULATION OF MOOD ................................ ............................ 121 Script for Happiness Inducing Procedure ................................ ................................ ............. 122 Life Event One ................................ ................................ ................................ .............. 123 Life Event Two ................................ ................................ ................................ .............. 124 Life Event Three ................................ ................................ ................................ ............ 125

PAGE 7

7 Script for Sadness Inducing Procedure ................................ ................................ ................ 127 Life Event One ................................ ................................ ................................ .............. 128 Life Event Two ................................ ................................ ................................ .............. 129 Life Event Three ................................ ................................ ................................ ............ 130 B INSTRUMENTAL MANIPULATION OF PERCEIVED DIAGNOSTICITY .................. 132 High Diagnosticity Versi on of the Antidepressant Advertisement ................................ ...... 132 Low D iagnosticity Version of the Antidepressant Advertisement ................................ ....... 133 C INSTRUMENTAL MANI PULATION OF OPPORTUNITY ................................ ............. 134 Instruction for Participants Under the Low Opportunity Condition ................................ .... 134 Instruction for Participants Un der the High Opportunity Condition ................................ .... 134 D QUESTIONNAIRE AND THE INSTRUMENTAL MANIPULATION OF OPPORTUNITY ................................ ................................ ................................ ................... 135 High Opportunity Ver sion ................................ ................................ ................................ .... 135 Low Opportunity Version ................................ ................................ ................................ .... 139 E SAMPLES OF ANTIDEPRESANT ADS ................................ ................................ ............ 143 Zoloft Ad ................................ ................................ ................................ .............................. 143 Paxil Ad 1 ................................ ................................ ................................ ............................. 144 Paxil Ad 2 ................................ ................................ ................................ ............................. 145 F P h RMAS GU IDING PRINCIPLES FOR DIRECT TO CONSUMER ADVERTISING .. 146 REFERENCES ................................ ................................ ................................ ............................ 148 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ........... 0

PAGE 8

8 LIST OF TABLES Table page 3 1. Fact or analysis of risk perception ................................ ................................ ........................... 68 3 2. Factor ana lysis of help seeking intention ................................ ................................ ............... 68 4 1. One way ANOVA results for the mood manipulation check ................................ ................. 94 4 2. Full factorial ANOVA results for the mood manipulation check ................................ .......... 94 4 3. One way ANOVA results for the diagnosticity manipulation check ................................ ..... 94 4 4. Full factorial ANOVA results for the diagnosticit y manipulation check ............................... 94 4 5. One way ANOVA results for the opportunity manipulation check ................................ ....... 95 4 6. Correlations among variables ................................ ................................ ................................ 95 4 7. MANOVA results for risk perception and help seeking intention when opportunity was low ................................ ................................ ................................ ................................ ...... 95 4 8. Mixed MANOVA results for r isk perception and help seeking intention when opportunity was low ................................ ................................ ................................ ........... 95 4 9. Group means and standard deviations for risk perception when happy mood and no mood manipulation conditions were comb ined ................................ ................................ 96 4 10. Group means and standard deviations for help seeking intention when happy mood and no mood manipulation conditions were combined ................................ ..................... 96 4 11. Group means and standard deviations for risk perception when happy mood and no mood manipulation conditions were not combined ................................ ........................... 96 4 12. Group means and standard deviatio ns for help seeking intention when happy mood and no mood manipulation conditions were not combined ................................ ............... 97 4 13. MANOVA results for risk perception and help seeking intention when opportunity was high ................................ ................................ ................................ ............................. 97 4 14. Mixed MANOVA results for risk perception and help seeking intention when opportunity was high ................................ ................................ ................................ .......... 97 4 15. MANOVA resu lts for risk perception and help seeking intention: All cases included ....... 97 4 16. ANOVA results for risk perception and help seeking intention: All cases included ........... 98

PAGE 9

9 LIST OF FIGURES Figure page 2 1. Expected ANOVA results for risk perception when opportunity was low ............................ 51 2 2. Expe cted ANOVA results for help seeking intention when opportunity was low ................. 51 2 3. Expected ANOVA results for risk perception when opportunity was low ............................ 52 2 4. Expected ANOVA results for help seeking intention when opportunity was high ............... 52 3 1. Eigenvalue plot for scree test for risk perception ................................ ................................ ... 69 3 2. Eigenvalue plot for scree test for help seeking intention ................................ ....................... 69 4 1. Observed ANOVA results for risk perception when opportunity was low ............................ 99 4 2. Observed ANOVA results for help seeking intention when opportunity was low ................ 99 4 3. Observed ANOVA results for risk perception when oppor tunity was high ......................... 100 4 4. Observed ANOVA results for help seeking intention when opportunity was high ............. 100 4 5. Mediation analysis when opportunity was low ................................ ................................ .... 101 4 6. Mediation ana lysis when opportunity was high ................................ ................................ ... 101

PAGE 10

10 Abstract of Dissertation Presented to the Graduate School of the Universi ty of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EFFECTS OF CONSUMER MOOD STATES ON PROCESSING OF DISEASE INFORMAT I ON IN DIRECT TO CONSUMER ANTIDEPRESSANT ADVERTISING AND PERCEIVED FUTURE RISK OF DEPRESSION By Jin Seong Park May 2009 Chair: Michael F. Weigold Major: Mass Communication The purpose of this research is to show that consumers mood states affect the way they process health information from consumer targeted drug advertisements as well as form perceptions of the future risk of diseases and intentions to seek professional help. The results generally supported this perspective, revealing that compared to those in happy moods, individuals undergoing sad moods at the time of the stu dy tended to overrate their future risk of clinical depression and report stronger intentions to seek professional help regarding depression. When consumers had low opportunity for estimating the future risk and help seeking intentions, the effects of mood s were significant regardless of the presence or absence of the information regarding depression self diagnosis in the antidepressant advertisement. When consumers had high opportunity for risk and intention estimation, the effects of moods were differenti al depending on whether the advertisement presented self diagnosis information or not. Findings of this study will be discussed for their implications for the theory and practice of consumer directed drug advertising.

PAGE 11

11 CHAPTER 1 INTRODUCTION History of Direct to Consumer Drug Advertising Over the past decade and a half, direct to consumer (DTC) pharmaceutical advertising was one of the fastest growing categories of advertising (Davis 2000). In 1992, only 17 prescription drugs were advertised directly to consumers, a number that increased to 79 five years later (Sasich 1999). The U.S. expenditure on DTC advertising increased from $25 million in 1988 (Morgan & Levy 1998) to $1.07 billion in 1996, and to $2.7 billion in 2002 (US General Accounting Office 2003). By 2000, DTC advertising accounted for 15% of the total promotional budget in the pharmaceutical industry (Brichacek & Sellers 2001). As a result, DTC advertising constituted 2.5% of the total advertising expenditure in the U.S. market in 2000 ( Na tional Institute for Health Care Management Research and Education Foundation, 2001), becoming the fourth largest consumer advertising category (Blankenhorn, Duckwitz, & Kerr, 2001). DTC prescription drug advertising is defined as any advertisement dev eloped by the pharmaceutical industry including radio, print, and/or television of prescription medication that targets the consumers/patients (Allison Ottey, Ruffin, Allison, & Ottey, 2003, p.121). According to Calfee (2002), the Food and Drug Administra tion (FDA) initially expressed the position that DTC advertising is not inherently in violation of FDA regulations (Terzian, 1999). Nevertheless, worried about DTC advertising s potential adverse influences on the public, the FDA initiated a moratorium on DTC ads in 1982 (Calfee, 2002). In 1985 the FDA lifted the moratorium because of fears it might conflict with the First Amendment despite the administration s previous concerns that DTC advertising is not necessarily in the public interest (Terzian, 1999 ). The DTC explosion occurred after the FDA released guideline s about the requirements for a fair and balanced representation of the drug in

PAGE 12

12 DTC advertising in 1997 and 1999 (Calfee, 2002; Kravitz & Wilkes, 2000). Currently, New Zealand and the U.S. are the only two economically advanced nations that permit DTC advertising (Coney, 2002; Hoek & Gendall, 2002). Promoting prescription drugs directly to consumers has aroused controversy in the U.S. Opponents argue that DTC advertising does not provide fair an d balanced information about the health benefits and risks of a drug (Bell, Wilkes, & Kravitz, 2000; Coney, 2002) They also point out that DTC advertising may substantially increase health care costs (Findlay, 2001), and adversely affect the doctor patien t relationship (Bell, Kravitz, & Wilkes, 1999; Mintzes, Barer, Kravitz, Kazanjian, Bassett, Lexchin, Evans, Pan, & Marion, 2002 ). Elliott (2003), a physician and bioethics professor, argues the pharmaceutical industry promotes newly constructed disease cat egories in an effort to increase drug company profits, a practice that may not increase public health. In line with these critical viewpoints, in the race for the 2004 Democratic presidential nomination, candidate Howard Dean called for a ban on DTC advert ising as a way to reduce prescription drug costs. Two other candidates, Richard Gephardt and John Edwards, offered similar proposals (Teinowitz 2003). In contrast, proponents argue the content of DTC advertising balances benefit and risk information and t herefore can educate consumers about diseases and treatments (Calfee, 2002). They further argue DTC advertising can increase awareness and encourage treatment of stigmatized and under diagnosed illnesses such as hypercholesterolemia and clinical depression (Calfee, 2002; Holmer, 2002). Proponents also point out DTC advertising encourages consumers to search for more information about health conditions and treatments ( Allison Ottey et al., 2003; Perri & Dickson, 1988), increases compliance with doctor ordere d treatments (Donohue Berndt, Rosenthal, Epstein, & Frank, 2004 ) and enhances doctor patient interactions (Holmer,

PAGE 13

13 1999). In fact, when Deans pledge to ban DTC advertising received media coverage, Dan Jaffe, executive vice president of the Association fo r National Advertisers, called the plan a prescription for disaster and added that [DTC] advertising often provides consumers [with] extremely valuable information that can save lives, often avoids serious health problems and in so doing often lowers he alth costs (Teinowitz, 2003, p.1). Effects of DTC Advertising Since the 1990s, empirical studies have focused on either the content of DTC advertising or its impact on consumer behavior. Content based studies investigated the nature of information presen ted in DTC advertising, usually in terms of its potential for educating consumers about diseases and treatments. For example, Bell et al. (2000), Kravitz and Wilkes (2000) and Roth (1996) content analyzed DTC advertisements and conclude d they had limited v alue as a source of health information because they d id not carry sufficient information about risk factors, the drugs mechanism of action, and their success in treating the disease and alternative treatments (Bell et al., 2000). Main, Argo and Huhmann ( 2005) point ed out DTC advertising depended more heavily on emotional than rational appeals, even in comparison with advertising for over the counter (OTC) drugs or dietary supplements. Some have argued that in recent years medical consumers have become mo re likely to seek detailed medical information and to participate in decisions that affect their health (Wolfe 2002). Simultaneously, researchers have shown an interest in exploring DTC advertisings impact on consumer attitudes and behaviors about health and medicine. For example, Sumpradt, Fors and McCormick (2002) found having positive attitudes toward DTC advertising and consumer characteristics, such as having chronic medical conditions, predict ed consumers willingness to discuss the advertised drug s with doctors. Beltramini (2006) found the perceived believability of

PAGE 14

14 DTC information and consumers comprehension of such information predicted their plans to consult doctors about health issues and request prescriptions for drugs seen in the ads. Resea rch also suggests when doctors refuse to prescribe requested drugs consumers are dissatisfied and may insist on the prescriptions (Bell et al., 1999; Mehta & Purvis, 2003). Kravitz et al. (2005) revealed consumers requests for drugs following exposure to DTC advertising increased the likelihood that doctors would prescribe the drugs. Herzenstein, Misra and Posavac (2005) found favorable attitudes toward DTC advertising resulted in less searching for information about the advertised drugs and an increased likelihood that the requested drugs were prescribed. In summary, research based on individual level data indicates exposure to DTC advertising is positively associated with consumer intentions to discuss health issues with doctors and request prescriptions for specific drugs. In turn, research conducted at an industry level indicates DTC advertising may not only predict consumers requests for specific drugs but also lead to the market expansion of a drug category, a construct represented by an increase in the number of visits to doctor s offices to discuss the disease the drug category is designed to treat, the number of diagnoses of the disease, and prescriptions written for the drug category. For example, combining nationally representative data from Na tional Ambulatory Medical Care Surveys (NAMCS) and TNS Media Intelligence/Competitive Media Reporting (CMR), Zachry, Shepherd, Hinich, Wilson, Brown, and Lawson (2002) revealed the DTC expenditure for antilipemics significantly predict ed the number of diag noses of hyperlipidemia prescriptions for antilipemics in general, and prescriptions for Zocor. Antilipemics is a drug class used for hyperlipidemia, a disease characterized by a high level of lipids in the bloodstream. Similarly,

PAGE 15

15 the expenditure for alle rgy medicine advertising was positively associated with the number of prescriptions for Claritin, and allergy medicine brand. Based on the same sources of data, Iizuka and Jin (2005) found between 1995 and 2000 the pharmaceutical industrys expenditure on DTC advertising predict ed an overall increase in visits to doctors offices by consumers. The association was constant across demographic groups, but became stronger after 1997, the year the FDA released a draft of the industry guideline. Donohue, Berndt, Rosenthal, Epstein, and Frank (2004) further revealed that the DTC expenditure for antidepressants was positively associated with an increase in consumers requests for antidepressant treatments after being diagnosed with depression. Donohue and Berndt ( 2004) similarly found that the DTC expenditure for antidepressants predicted the number of prescriptions for antidepressants. Differentiating from an increase in the requests for specific drugs, they call ed this expansion of a drug category a treatment ex panding effect (Donohue & Berndt, 2004, p.124) of DTC advertising. Although these studies were based on a survey design, and therefore did not establish cause and effect relationships, the findings indicate DTC advertising may expand a drug market Limit ations of the Current Literature Despite a large body of research, the current DTC literature is limited in scope for a number of reasons. First, r esearch that explores the effects of DTC advertising focus es on how the overall exposure to and attitudes tow ards DTC advertising explain a limited range of variables such as visits to doctors offices, requests for specific drugs, and adherence to treatment guideline s These are important variables with implications for marketing strategy and consumer health, an d therefore deserve attention. However, the literature neither accounts for the psychological processes through which exposure to DTC advertising produces such

PAGE 16

16 consequences nor specifies the content elements in DTC advertising that are likely to produce th em. Second, even though the literature abounds in studies that reveal the themes and content elements that characterize DTC advertising (Bell, Wilkes, & Kravitz, 2000; Huh & Cude, 2004; Kaphingst, Dejong, Rudd, & Daltroy, 2004; Roth, 1996), few studies ha ve illuminated how consumers process such information. This further limits the scope of the DTC literature, because most likely it is consumers processing of information in DTC advertising, not their exposure to DTC advertising per se, that produces cogni tive and behavioral effects. Further, the ultimate rationale for content focused research is the possibility that exposure to certain content elements will affect consumers in some ways. Therefore, employing information processing perspectives will contrib ute to the DTC literature by complementing content focused studies. Third, the concept of risk perception is missing from the literature, although research in health behavior and social cognition indicates consumers risk perceptions of diseases may under lie DTC advertising s behavioral influences. Consumers risk assessments are influenced by information they receive about the judgment domain (Kahneman & Tversky, 1972; Menon, Block, & Ramanathan, 2002), suggesting DTC advertising may influence risk percep tions of diseases by providing information on the diseases. Currently few studies have focused on DTC advertising s impact on risk perceptions of diseases. One possible exception is An (2007) who found a positive association between college students rec all of DTC antidepressant advertisements and the perceived prevalence of clinical depression in the US. An and Jin (2005) similarly found associations between consumers self reported attention to DTC television advertising in general and perceptions of th e prevalence of overactive bladder condition and erectile dysfunction in the US. However, the se studies did not

PAGE 17

17 test how consumer recall of advertised pharmaceutical brands and attention to DTC advertising in general predicted their own perceived risk of d iseases. Further, because the studies were conducted as surveys, neither established causal relationships or specified the message elements that might have contributed to DTC advertising s effects on prevalence or risk perceptions. In addition, the studies did not control potentially significant extraneous influences, such as interpersonal experiences with depression. I nterpersonal experience s have been shown to be important predictors of the perceived social reality of various social phenomena (Higgins and King, 1981; Shrum and Bischak, 2001; Wyer and Shrull, 1989) The fourth and last limitation is that the concept of mood state has not been incorporated into the current literature. Consumers mood state affects their cognitions of health issues (Salovey & Birnbaum, 1989). Moods also affect consumers risk assessments (Johnson & Tversky, 1978; Nygren, Isen, Taylor, & Dulin, 1996), and the way they process judgment relevant information (Forgas, 1995), suggesting consumer mood state may interact with exposur e to information in DTC advertising to determine risk perceptions. Once formed, risk perceptions of diseases can influence engagement in remedial and preventive behaviors, such as undergoing a screening test for breast cancer and consulting doctors to disc uss health issues (Block & Keller, 1998; Irwin, Valdiserri, & Holmberg, 1996; Raghubir & Menon 1998; Siegel, Raveis, & Gorey, 1998). Therefore, consumer mood state and exposure to information in DTC advertising may result in changes in health behavior thro ugh influencing consumer risk perceptions of health issues. Therefore, the concept of consumer mood state should be incorporated into the literature on the effects of DTC advertising. Visits to doctors offices to discuss a health problem and a screening test for a disease is logically considered a prerequisite of and antecedent to a diagnosis of the disease and

PAGE 18

18 prescription for a drug category designed to treat the disease, which are two indicators of the market expansion of a drug class (Donohue & Berndt 2004; Donohue et al., 2004; Iizuka & Jin, 2005; Zachry et al., 2002). Therefore, research on the combined effects of consumer mood state and exposure to specific content elements in DTC advertising on the perceived risk of a health problem and intentions to seek professional help to discuss a disease will add a much needed psychological account for why DTC advertising may lead to the market expansion of a drug class. Research Purpose The current project is designed to address the limitations described ab ove with respect to the DTC literature It does so by examining potential psychological process es that may be important to the effects of exposure to information in an DTC antidepressant advertisement on consumers perceived future risk of depression and i ntentions to seek professional help to discuss depression Among many perspectives in social cognition potentially applicable to the overall research purpose, this project employed the mood as information (Schwarz & Clore, 1983) and mood as frame (1981) t heories. The two perspectives are used to explore how consumer perceptions of the future risk of depression and help seeking intentions are affected by the three way interaction of consumer mood state perceived diagnosticity, defined as consumer perceptio ns of the degree that a number of discomforting life experiences indicate clinical depression and opportunity for deliberative risk and intention estimation. The role of mood state deserves attention because moods influence consumers probability and ris k assessments (Johnson & Tversky, 1978; Nygren et al., 1996), and processing of judgment relevant information (Forgas, 1995). Risk perception deserves attention because

PAGE 19

19 research suggests it can influence the likelihood of preventive and remedial health beh aviors, including visiting doctors offices to discuss health issues. Understanding what motivates consumers to initiate a discussion with doctors about diseases, drug treatment s and screening tests may be beneficial for DTC marketers and health professi onals who intend to motivate consumers to be responsible for their health. Knowing what triggers consumer remedial actions will be especially beneficial for marketers who promote drugs for relatively under treated diseases, as pharmaceutical companies most heavily advertise drugs for such diseases to spur a market expanding effect (Iizuke, 2004). Help seeking intentions are an important construct for DTC marketers because consumers need to seek professional help if they are to receive a prescription (Donoh ue & Berndt, 2004; Donohue et al., 2004; Iizuka & Jin, 2005; Zachry et al., 2002). In terms of consumer health, intentions to discuss a health problem deserve attention because detecting a disease is necessary for medical intervention and treatment The p erceived diagnosticity of internally retrieved life experiences is also a significant factor that determines consumer risk perception s of diseases. The content of DTC advertisements for antidepressants, such as Prozac (produced by Elli Lilly) Zoloft ( Pfiz er) Paxil CR (GlaxoSmithKlein, GSK) and Effexor XR (Wyeth) tend s to focus on information about the symptoms of depression including sleeplessness, physical exhaustion, sluggishness, and hopelessness However the ad content may or may not offer a diagnos tic guideline that pinpoints how consumers may form accurate perceptions of the degree that their experiences of the symptoms actually indicate clinical depression. Such a guideline will be required for one to properly interpret symptom information and mak e an accurate self diagnosis of depression

PAGE 20

20 Some antidepressant ads stipulate that a particular health condition should last each day for two weeks to be considered a symptom of clinical depression (e.g., These are some symptoms of depression. They must occur each day for at least two weeks and interfere with your daily life, as stated in a Zoloft advertisement). This diagnostic guideline, almost identical in nature to the diagnostic guideline for depression that American Psychiatric Association (APA) pr esents in its Diagnostic and Statistical Manual of Mental Disorders IV ( 1994 ) typically does not appear i n advertisements for other antidepressant brands This is important because t he guideline requiring that symptoms be present nearly every day for two weeks is likely to result in fewer self diagnoses of depression than a guideline that permits the inference that one is depressed based on observation of the symptoms for a day or a few days. Therefore, exposure to the ad that states a range of discomfort ing life experiences indicate clinical depression will increase risk perception, while a version that reduces the diagnosticity of such life experiences will reduce risk perceptions. To develop hypotheses for the study, t wo contrasting perspectives on the effects of moods on risk estimation, the mood as information (Schwartz & Clore, 1983) and mood as prime (Bower, 1981; Wyer & Carlston, 1979) theories are reviewed and compared The two theories present distinct pathways through which affective states infl uence judgments, often leading to differential, and incompatible, predictions of the effects of mood state on judgments Then the literature further point s out that the environment in which risk and intentions are estimated, particularly how much opportun ity consumers have for deliberative estimation, is expected to determine the relative applicability of the two theories. Opportunity therefore is hypothesized to trigger a particular pathway through which temporary affective state s influence information pr ocessing and risk estimation. The review further illuminates that once a particular

PAGE 21

21 pathway is activated it would determine the pattern in which affective states and the perceived diagnosticity of depression related life experiences would have interactive influences on the perceived risk of depression and help seeking intentions An experiment is conducted to illuminat e the interactive effects of mood state, perceived diagnosticity and opportunity for risk and intention estimation on risk perception s and help seeking intention s From a health education perspective, a key objective for evaluating health information is to evaluate the extent such information helps consumers to make informed decisions about health issues (Peters, Lipkus, & Diefenbach, 2006). Most proponents of DTC advertising argue it advances this goal because it provides additional health information that consumers may not find elsewhere However, only a limited number of studies have actually examined how people process information in DTC advertising and form attitudes and beliefs about health issues One exception is Davis (2000) who explored how consume r s process risk information in DTC advertising David (2000) found that consumers evaluated a drug as safer and more appealing when the r isk statement in the advertisement for the drug was incomplete rather than complete, suggesting consumers may improperly interpret the number of risks presented as an indicator of the drug s performance in terms of safety. Because few studies have used an information processing approach to study the effects of DTC ads, this study may offer new insights about the mechanisms through which DTC advertising can affect consumer perceptions and knowledge about diseases, drugs, and treatments. In particular, this studys focus on exploring the way consumers process information about depression and form risk perceptions and intention s to seek professional help may contribute to understanding how pharmaceutical marketers can utilize DTC advertising for the

PAGE 22

22 m arket exp ansion of a drug category, because intention s to seek professional help for a disease are an antecedent of the market expansion of a drug class designed to treat the disease. The studys focus on risk perception s and help seeking intention s will contribut e to understanding DTC advertisings implications for consumer health, especially when one considers depression is a seriously under diagnosed and under treated disease (Holmer, 2002), and depression is more likely to be addressed if an increasing number o f consumers visit doctors offices to discuss the disease. The project may further act as a springboard for a discussion about how to lead people to become more informed consumers of information in DTC advertising. Importance of Mood State The inferences required of consumers in processing DTC advertisements make theories of affect and cognition particularly relevant for analyzing the effects of exposure to an DTC antidepressant advertisement on risk perception s and help seeking intentions, becaus e moods m ay directly influence risk perception s (Tversky & Johnson, 1978) or indirectly affect risk perception s through increasing the accessibility of past experiences of negative health conditions in consumer memory (Salovey & Birnbaum, 1989) Further, because mo ods are reported to influence health cognition, particularly consumers subjective experiences with and report ing of negative health conditions ( Croyle & Uretsky, 1987 ; Pettit, Kline, Gencoz & Gencoz, 2001; Salovey & Birnbaum, 1989), mood state may influen ce the degree to which ad audiences perceive the symptoms presented in DTC ads to be self relevant. In addition, Schwarz, Strack, Kommer, and Wagner (1987) found t he role of positive and negative mood state as an input for social cognition was greater for a judgment domain where emotional experiences have more relevance, such as subjective judgments of well being, whereas domain specific information had more effects on consumer evaluations of specific life domains such as job performance This means mood state may have more effects on consumer

PAGE 23

23 perceptions of depression than on perceptions of other diseases. Compared to other diseases with more tangible physical manifestations, the future risk of depression is a judgment domain where emotional experiences h ave more relevance The symptoms of some disease categories, such as allergy, arthritis, asthma, and overactive bladder are more physically tangible than those of depression, because they are based on clearly observable physical manifestations. Moods will have a relatively limited role in a person s judgment of whether s/he is experiencing such tangible and self evident physical symptoms as shortness of breath (a symptom of asthma) or rashes in skin (a symptom of allergy). In comparison, the symptoms of de pression involve either relatively intangible and non salient physical symptoms, such as low physical energy, loss of appetite, changes in weight and insomnia, or experiences interlinked with a person s affective states, such as feelings of worthlessness a nd guilt, depressed mood, diminished interest in life activities, irritation, and suicidal ideation (APA, 1994). Therefore, compared to other diseases such as asthma, allergy, or arthritis, moods will be more relevant to judging one s own experiences of de pression symptoms. Therefore, mood state will have stronger effects on the perceived future risk of depression and intention s to seek professional help to discuss depression. Study Overview An experiment with a 2 (mood state: sad versus happy combined wit h no mood manipulation ) 2 (perceived diagnosticity: low versus high) 2 (opportunity for risk and intention estimation: low versus high) between subjects design with a non factorial control group is used to collect data for this study, for a total of th irteen experimental groups. In the experiment, subjects are either induced to have sad or happy moods or do not receive a mood induction treatment. Then they are exposed to an antidepressant advertisement that includes

PAGE 24

24 either low or high diagnosticity inf ormation. Then participants report their perceived future risk of depression and help seeking intentions under a low or high opportunity condition. Under the low opportunity condition, mood state will directly influ ence risk perceptions and help seeking intentions without intermediary cognitive processes (Forgas, 1995), and therefore without interacting with perceived diagnosticity. Under the high opportunity condition, the effects of mood state on risk perceptions and help seeking intentions will occur t hrough intermediary cognitive processes (Forgas, 1995). Therefore, the effects of mood state on risk perceptions and help seeking intentions will vary depending on whether the ad contained low or high diagnosticity information. Special Status of DTC An tidepressant Advertising Of many categories of disease and medicine, this study focuses on depression and anti depressants for a number of reasons. First, anti depressants, such as Paxil CR and Zoloft, are among the most heavily advertised prescription dru gs in the DTC market (Rosenthal Berndt, Donohue, Frank, & Epstein 2002). In addition depression and manic depression are the most common form s of severe mood disorders (Altshuler, Hendrick, & Burt, 1998; Weiss & Lonnquist, 1997). Second, depression is a largely under diagnosed disease category (Holmer, 2002). The under diagnosis and under treatment of depression suggest that depression is a stigmatized disease (Elliott, 2003; Holmer, 1999, 2002). From a health policy perspective, DTC advertising may be able to play an important role in reducing the under treatment problem of depression. Third, symptoms of depression, such as loneliness, feeling s of social isolation inability to concentrate, and difficulty with thinking and making decisions (APA, 1994), suggest that DTC antidepressant advertising targets a potentially vulnerable group of consumers. Hollon (2004) points out special attention is required regarding the effects of DTC advertising aimed at

PAGE 25

25 consumers suffering psychiatric and neurological illn esses, because their decisional capacity may be impaired. Kravitz, Epstein, Feldman, Franz, Azari Wilkes, Hinton, and Franks (2005) consider DTC advertising a controversial ethical issue for similar reasons. In conclusion, DTC advertisings impact on risk perception and treatment related behaviors deserves more attention when it is designed to treat an under treated disease and targeted toward a psychologically vulnerable group of consumers, including potential users of antidepressants.

PAGE 26

26 CHAPTER 2 LITERAT URE REVIEW This review of the literature first examines the effects of risk perception on health behavior, especially consumer engagement in preventive and remedial actions to address health issues. Next, two theories of the effects of mood state on social judgments, the mood as information (Schwarz & Clore, 1983) and mood as prime (Bower, 1981) perspectives, are reviewed. The two theories are conceptualized as illuminating the two pathways through which mood state affects consumer judgments. Third, the two theories are compared, with a focus on the factors that determine the relative applicability of the two principles in a given situation. Last, hypotheses (H1a H3b) are proposed on the interactive effects of mood state, perceived diagnosticity of interna lly retrieved life experiences, and opportunity for risk and intention estimation on consumers perceived future risk of depression and intentions to seek professional help to discuss depression. Risk Perception Disease risk perc eption is defined as the p erceived likelihood of being diagnosed with a disease in the future (Rosenstock, 1990) A t a time when consumers are increasingly expected to be responsible for their health (Loroz & Lichtenstein, 2004) p erceived risk of diseas e is considered important in health behavior research because of its potentia l to produce changes in health related behavior. Research generally finds that high levels of disease risk perception predict consumer engagement in preventive and remedial behaviors. For example, perceived risk of cancer is positively associated with screening for colorectal (Blalock, DeVellis, Afifi, & Sandler, 1990), breast (Lipkus et al., 1996) and cervical (Seow, Wong, Smith, & Lee, 1995) cancer. Similarly, Croyle and Lerman (1999) found people s intenti ons to receive testing for genetic susceptibility

PAGE 27

27 to cancer are strongly influenced by overrated perceptions of cancer risk. Even though a number of motivational (e.g., motives for testing) and affective (e.g., anxiety about a disease) factors may moderate the effects of risk perceptions on health behavioral intentions (Shiloh & Ilan, 2005), a sizable body of studies conceptualize risk perception as a motivator of changes in health behavior (Block & Keller, 1998; Lipkus, Biradavolu, Fenn, Keller, & Rimer, 2 001; Menon, Block, & Ramanathan, 2002; Robinson, Rigel, & Amonette, 1998). Because of the role of risk perceptions in motivating health behaviors, researchers have explored how consumers make risk assessments, with a focus on revealing the psychological pr ocesses through which consumers generate estimates for the risk of undergoing stressful and disturbing life experiences in the future, including diseases Ideally, risk assessments, given the potential to impact health behaviors should be made from inform ation that is relevant to the particular judgment domain in consideration and diagnostic of the risk of the particular health issue. However, research frequently finds judgments, including probability and risk assessments, are influenced by inputs other th an factual and objective information, through mechanisms that are largely unconscious and automatic and therefore uncontrollable. As Epstein (1994) stated: There is no dearth of evidence in everyday life that people apprehend reality in two fundamentally d ifferent ways, one variously labeled intuitive, automatic, natural, nonverbal, narrative, and experiential, and the other analytical, deliberative, verbal, and rational (p.710). Whereas experts risk perceptions are based on the deliberative thinking mode and guided by the use of statistical rules of probability, non experts risk assessments may depend on more intuitive and automatic ways of knowing (Reventlow, Hvas, & Tulinius, 2001). To this effect, Wilson and Brekke (1994) argued judgment making is oft en biased due to mental contamination, or the process whereby a person has an unwanted judgment, emotion, or behavior because of mental processing that is unconscious or uncontrollable (p. 10). In a similar

PAGE 28

28 vein, Kraus, Torbjorn, and Slovic (1992) labe led the biased ways laypeople assess chemical risks as intuitive toxicology. Under the category of intuitive and unconscious influences are metacognitive experiences such as perceptual fluency (Jacoby & Dallas, 1981) and availability (Tversky & Kahneman, 1973) and representativeness (Kahneman & Tversky, 1972) heuristics, and the influence of affect (Bower, 1981; Schwarz & Clore, 1983) on judgments. Particularly relevant to this research several researchers (Constans & Mathews, 1993; Gasper & Clore, 1998; Nygren et al., 1996) have investigated the effects of consumers affective states on the way they process judgment related information a nd make judgments, including risk assessments. More important, risk assessments generally involve probability judgment s of emotionally disturbing experiences, and they rarely occur in an affectively neutral context (Johnson & Tversky, 1983). Therefore, the literature on affect and cognition is considered appropriate for the overall purpose of the current project, as it is designed to explore how consumers in various mood states form risk perceptions of depression in the face of an antidepressant advertisement presenting information about the symptoms and accurate diagnosis of depression. Mood State, Information Processing, and Judgment A prerequisite for a discussion of the literature on mood and judgment is an understanding of how the four conceptually similar constructs of affect, feeling, mood, and emotion are differentiated. No general agreement exists regarding the dis tinctions among the four concepts (Fiedler & Forgas, 1988; Frijda, 1986). However, Forgas (1992, 1995) presents a reasonable working system of classification. The system defines affect and affective states as general terms to refer to moods and emotions. M oods, in turn, are low intensity, diffuse and relatively enduring affective states without a salient antecedent cause and therefore little cognitive content (e.g. feeling good or feeling bad), whereas emotions are more intense, short lived and usually

PAGE 29

29 h ave a definite cause and clear cognitive content, (Forgas, 1992, p.230) such as feeling fear, worry, or anger about a target object. Experimentally induced affective states, therefore, are generally referred to as moods. This project focuses on the effect s of mood on consumers processing of medical information, risk perceptions, and intentions to seek professional help. The effects of mood on cognition and judgments are well documented in the literatures on product evaluation, response to advertising, ris k and probability estimation, and health related cognition and decision making. In the context of consumer research, Isen, Clark, Shalker, and Karp (1978) found positive moods led to favorable evaluations of products Batra and Ray (1986) and Edell and Bur ke (1987) also revealed advertisement induced positive moods led to more positive brand evaluations. Similarly, Alpert and Alpert (1990) showed positive moods led to stronger purchase intentions. Mathur and Chattopadhyay (1991) demonstrated positive consum er moods induced by the programming that surrounded an advertisement led to more positive responses to the ad, while negative moods led to more negative responses. Overall, this body of research indicates that mood state leads consumers to evaluate produc ts and form purchase intentions in mood congruent ways. In other words, consumers in positive moods respond to advertisements or evaluate products more positively than those undergoing negative moods. Further, this tendency has been observed whether the mo ods are induced by procedure gift through advertising content elements, or via the program that surrounds an advertisement. Recent studies showed the effects of mood state on cognition and product evaluation were moderated by a number of factors, such as conscious monitoring of one s feelings (Pham, Cohen, Pracejus, & Hughes, 2001) or the order of mood induction in relation to exposure to product attribute information (Yeung & Wyer, 2004). For example, Yeung and Wyer (2004) revealed

PAGE 30

30 consumers experimenta lly induced moods directly influenced product evaluation regardless of the information they subsequently received about product attributes or the criteria they used to evaluate the product. However, this effect was observed only when the moods were induced before, rather than after, consumers generated an affective response to the product s appearance. The indicated experimentally manipulated moods would be confused in consumers mind with the affective state generated by a stimulus only if the manipulated moods existed at the moment of making affective responses to the stimulus. In addition, a group of researchers suggested the effects of positive moods might be relatively homogeneous, whereas the effects of negative moods would be more heterogeneous. For example, Raghunathan and Pham (1999) found when making job selection decisions, individuals in sad moods favored options that involved high risk and high reward, whereas anxious individuals favored low risk/low reward options. Although Raghunathan and Pham s (1999) study revealed the heterogeneous effects of the different types of negative moods, this study focuses on sad moods, and does not incorporate the heterogeneous effects of negative moods. A large body of literature from social cognition and risk c ommunication indicates affective states influence judgments of the likelihood of future events, such as one s perceived probability of being affected by crimes, natural disasters, terrorisms, and so on. The effects are mood congruent, in that positive mood s lead people to overrate the likelihood of positive future events and underestimate the probabilities of negative events, whereas the reverse is true for those in negative moods. For example, Wright and Bower (1992) and Constans and Mathews (1993) report ed mood state had mood congruent effects on the perceived likelihood of future events. Nygren et al.

PAGE 31

31 (1996) found that in a gambling situation, people in a positive mood overestimated the chances of winning relative to the chances of losing. Gasper and Clo re (1998) reported state anxiety, which was temporarily induced by a mood induction procedure, and trait anxiety, which was permanently rooted in a person s personality traits, interacted to impact the perceived risk of personally relevant (e.g., conflict with parents, theft, and getting embarrassed in public) and relatively impersonal negative events (e.g., police violence and proliferation of AIDS). Using a survey design, Constans (2001) also revealed state and trait anxiety predicted undergraduate studen ts estimation of the risk of poor academic performance. Reflecting the importance of affective states on consumer risk perceptions across various domains, Slovic, Finucane, Peters, and MacGregor (2004) concluded intuitive feelings were the dominant metho d by which individuals evaluate risk, labeling the perspective as risk as feelings. The literature on the effects of mood state on probability and risk assessments is relevant to the current project, because risk perception of a disease, defined as the p erceived likelihood of developing a disease in the future (Rosenstock, 1990), can be logically conceptualized as one particular type of probability judgment Because consumer assessments of the likelihood of diseases are one form of risk perception, moods should also have mood congruent effects on the perceived risk of negative health conditions. In fact, Johnson and Tversky (1978) revealed negative moods le d individuals to overrate the frequencies of deaths due to leukemia and cancer. Salovey and Birnbaum (1989, study 3) reported the perceived risk of developing negative health conditions (e.g., high blood pressure) in the future was higher among participants in experimentally induced negative moods than those in positive moods. Using a survey design, Lipk us et al. (2000) revealed worries about

PAGE 32

32 developing breast cancer were positively associated with women s higher breast cancer risk perceptions. These studies fit in with the large body of research (Dunn & Schweitzer, 2005; Gasper & Clore, 2000; Schwarz & Clore, 1983) that illuminates negative feelings generate a pessimisti c point of view regarding various life domains, even when the procedure that actually elicited such feelings may not be logically related to the judgment in question. For example, an indi vidual s thinking about a sad event at the moment most likely will not determine the actual probability of the person s developing high blood pressure in the future. However, in Salovey and Birmbaum s (1989) study, participants in experimentally induced sa d moods reported higher risk perceptions of various negative health conditions than those in happy moods. Why do positive and negative moods affect social judgments, even when the judgment domain in question does not logically pertain to the event that actually elicited the moods? Two well established frameworks, named mood as information (Schwarz & Clore, 1983) and mood as prime (Bower, 1981; Wyer & Carlston, 1979) perspectives, present different accounts for the effects of mood state. The mood as information perspective posits that when people mak e social judgments, affective states function as an information input that directly influences the judgments (Schwartz & Clore, 1983), independent of information from other sources that normatively have m ore relevance to the judgment domain (Gorn, Goldberg, & Basu, 1993; Plam, 1998). In contrast, the mood as prime perspective (Bower, 1981) posits that moods affect social judgments through their impact on cognitive processes, by making mood congruent constr ucts more accessible in one s mind. In general, both approaches suggest moods affect judgments in a mood congruent

PAGE 33

33 way. Therefore, sad moods tend to generate a pessimistic outlook of the future, whereas happy moods lead to an optimistic outlook. Mood As Information The mood as information perspective emerged in a series of studies conducted by Schwartz and Clore (1983). In their studies, participants momentary negative feelings, induced by asking them to write about a recent life event that had made them feel bad, lowered their evaluations of happiness and satisfaction with their life in general. Participants that wrote about a pleasant life event evaluated their lives more positively. The control group, who did not write about a life event, who reported being as happy as participants who wrote about a pleasant event, evaluated their lives as positively as those who received a positive mood induction procedure. The effects were assumed to occur because people tended to misattribute their momentary feeling s to more general aspects of their lives, and as a result incorrectly refer to such feelings when making social judgments. In other words, in addition to the actual quality of their everyday life experiences, feelings became an informational input for part icipants evaluations of their lives in general In other words, participants in transient positive or negative moods mistook their experimentally induced affective states as their affective reaction to the object being judged, such as the question about t heir satisfaction with life in general (Schwarz & Clore, 1983). S upport for this interpretation came from findings that the differences in judgments between participants in positive and negative moods tended to disappear after the actual source of the fee lings was made salient. For example, Dunn and Schweitzer, (2005), Schwarz and Clore (1983), and Siemer and Reisenzein, (1998) reported the effects of negative feelings disappear ed when the actual source of the feelings was made salient. In other words, moo d state lost its value as information when participants were made aware of the actual source of the ir affective state

PAGE 34

34 Interestingly judgments of people who received a positive mood induction procedure may be unaffected by the salience of the actual sourc e of the moods (Schwarz & Clore, 1983), suggesting the informational value of experimentally induced positive and negative moods may be asymmetrical. This asymmetry may occur because the value of a positive or negative induced mood state as information for judgment making may depend on how much the induced moods differ from the average mood state that surrounds most people s ordinary life experiences. As Schwarz and Clore (1983) explained, the life evaluations of participants in positive induced moods did n ot drop after the mood was attributed to an experimental procedure, because experimentally manipulated happy moods did not significantly differ from the mood states of the non factorial control group. The mood as information framework has been frequently employed in research on social cognition and consumer judgments. For example, Schwarz, Strack, Kommer, and Wagner (1987) reported that moods influenced participants evaluation of their life in general, but their evaluations of specific life domains were m ore affected by domain specific information. This suggested the informational value of positive and negative mood s was high in relatively global, diffuse, and abstract judgment domains, such as evaluation of life in general, satisfaction with social intera ction in general, but low in concrete and specific judgment domains, such as evaluation of one s own job performance for a one week period or satisfaction with one s relationship with co workers. In addition, Gorn et al. (1993) reported people experiencin g happy moods tended to evaluate products more positively than people in sad moods, unaffected by information presented about the products, but this effect disappeared when the actual source of the moods was made salient. Source salience generally lowered the evaluations of participants in the positive mood

PAGE 35

35 condition, while the evaluations of those in the negative induced mood did not change significantly. This result may appear to contradict Schwarz and Clore (1983), in which source salience generally affe cted the judgments of participants in negative, rather than positive, induced moods. However, the asymmetry of the informational values of positive and negative induced moods occurred in Schwarz and Clore (1983) because the experimentally induced positive moods were similar to the average mood state of the control group. Therefore, one may judge that the findings of Gorn et al. (1993) did not contradict Schwarz and Clore (1983) as long as the average mood state of the participants before they received a mo od induction procedure was similar to the experimentally induced sad, rather than happy, moods. However, this possibility cannot be further explored because Gorn et al. (1993) did not include a non factorial control group in the study. Dunn and Schweitzer (2005) also found positive feelings increase d and negative feelings decreased participants perceived trustworthiness of co workers, and this effect became non significant when subject paid attention to the actual source of the affect. This indicated the differences in judgments between people who received a happy versus sad mood procedure disappeared when participants were made aware that the source of the ir moods had no relation to the issue being judged. Whether the positive or negative mood generally l oses its informational value of moods depends on how much the induced moods differ from the average mood state of the participants that do not receive a mood induction procedure. These findings clearly support the perspective that moods directly influence judgments by means of serving as an information input for judgment making.

PAGE 36

36 A number of researchers further explored conditions that enhance or reduce the informative function of mood state. For example, Petty, Schumann, Richman, and Strathman (1993) repor ted people increasingly used extraneous affect as information in low elaboration conditions, particularly when motivation for information processing and need for cognition were low. In a similar vein, Siemer and Reisenzein (1998) revealed the direct effect s of feelings were stronger under conditions that reduced the consumer s opportunity for deliberate judgment making or information processing, such as time pressure and cognitive load from competing task demands, implying the mood as information framework represented a heuristic mental process more likely to be activated under conditions that reduce opportunity for deliberative information processing. Siemer and Reisenzein (1998) interpreted the finding as meaning the effects of mood on evaluative judgmen ts should be enhanced under such circumstances [that reduce opportunity for deliberative judgment making] (p.786). To be more accurate, however, what the finding really implied was that the direct, non cognitively mediated effects of mood on judgments, as conceptualized by the mood as information perspective, were reduced under conditions of low cognitive elaboration. Indeed, Petty et al. (1993) found that moods could still influence judgments under high elaboration conditions through their impact on the v alence of participants cognitive responses to information about the object being judged, suggesting moods influence judgments simultaneously through direct and cognitively mediated pathways. In summary, the mood as information literature suggests that mo ods operate as information which informs subsequent judgments Mood as information has relatively greater impact to the extent that the source of the mood is less salient (Schwarz & Clore, 1983; Gorn, et al., 1993), and participants are less motivated for and capable of cognitive information

PAGE 37

37 processing (Petty et al., 1993) and their opportunity for deliberative information processing is temporarily lowered by the existence of time pressure or competing task demands (Siemer & Reisenzein, 1998). Further, Fo rgas (1995) proposed the mood as information perspective predominantly applies when participants need to simplify the judgment because they are either unmotivated to process information or incapable of engaging in elaborate processing due to circumstantial factors that reduce their cognitive information processing capacity. As Forgas (1995) put it, the mood as information account is most likely to predict mood congruency in circumstances in which quick, simple, heuristic processing is adopted by a judge in response to contextual requirements (p.44). Siemer & Reisenzein (1998) showed moods directly affected participants judgment of satisfaction with their lives, and this effect occurred more prominently under the condition of time pressure, which limited o pportunity for deliberative judgment making. When applied to this study when consumers estimate risk in an environment that temporarily reduces their opportunity for deliberative information processing, the mood as information perspective would suggest m oods will have main and un moderated effects on the perceived risk of developing clinical depression in the future and intentions to seek professional help regarding depression. These effects will be observed whether consumers are exposed to the low or hi gh diagnosticity information, that is, whether the advertisement contains the APA diagnostic guideline or not. Mood As Prime The mood as information perspective is not the only model of how affect can influence judgments and decisions. Petty et al. (1993) suggested affective states might influence judgments through an alternative, cognitively mediated pathway even under the condition of high source salience. The mood as prime perspective accounts for this alternative pathway. T he mood as

PAGE 38

38 prime framework ( Isen et al., 1978; Bower, 1981) posits that moods influence judgments by priming mood congruent constructs. Once primed, these constructs with mood consistent affective implications become more accessible in memory (Isen et al., 1978). Because the high acc essibility of a construct in memory increases the likelihood that it will be retrieved and used in judgment making (Sherman & Corty, 1984; Shrum & O Guinn, 1993), the perspective suggests that moods tend to influence judgments in a mood congruent manner. I n other words, positive moods engender a positive outlook of the future by activating positive thoughts in one s mind, while negative moods increase the accessibility of thoughts with negative affective connotations. The notion that moods prime mood congr uent constructs in consumer memory comes from Bower s (1981) associative network theory : The semantic network approach supposes that each distinct emotion ... has a specific node or unit in memory that collects together many other aspects of the emotion tha t are connected to it by associative pointers ... Each emotion unit is also linked with propositions describing events from one s life during which that emotion was aroused ... These emotion nodes can be activated by many stimuli by physiological or symb olic verbal means. When activated above a threshold, the emotion unit transmits excitation to those nodes that produce the pattern of autonomic arousal and expressive behavior commonly assigned to that emotion ... Activation of an emotion node also spreads activation throughout the memory structures to which it is connected, creating subthreshold excitation at those event nodes ... Thus, excitation [of] the sadness node ... will maintain activation of that emotion and thus influence later memories retrieved (p.135). This process is set in motion in a largely non controlled, automatic manner (Forgas, 1995). In addition, although Forgas (1992, 1995) differentiated the definitions of emotion and mood, Bower (1981) apparently did not make a clear distinction be tween the two terms, as the reported studies employed a procedure that resembled a technique researchers typically use to manipulate particular moods in participants. In summary, the mood as prime framework suggests that moods influence judgments in a mood congruent way through priming and thus enhanc e the accessibility of mood congruent

PAGE 39

39 materials in consumer memory. Inducing a sad mood in a consumer s mind, for example, will prime constructs in memory (e.g., rainy day, poor academic performance, loss of si gnificant others, and illnesses ) connected to feelings of sadness. These primed constructs are then more likely to be used in formulating judgments than are un primed constructs, which are relatively unassociated with sad moods. Therefore, other conditions being equal, people in sad moods will have a more pessimistic outlook of a judgment domain than those in happy moods (Bower, 1981). Whereas the mood as information framework is generally more applicable when consumers have little substantial and detailed information to process or have low opportunity for deliberative information processing (Clore, Schwarz, & Conway, 1994), the mood as prime framework better accounts for the effects of mood in the presence of substantial judgment related information and un der conditions of high cognitive elaboration and constructive judgment making (Forgas, 1995). Therefore, when consumers are exposed to substantial information regarding a judgment domain under conditions of high cognitive elaboration, the mood as prime fra mework will generate more accurate predictions than the mood as information account. In addition, even though source salience will remove the direct, non cognitively mediated effects of mood state on judgments, the mood as prime view would suggest it may still be possible that the cognitively mediated effects of mood state still remain significant, because it may take higher levels of cognitive elaboration to correct for the biasing effects of mood primed, mood congruent constructs than the direct biasing effects of the mood as information. A number of researchers have found support for the mood as prime framework, demonstrating that moods enhance the relative accessibility of mood congruent over non mood

PAGE 40

40 congruent information in memory. In an early study, Bower, Monteiro, and Gilligan (1978) reported, after memorizing a list of words with varying affective valences, participants recalled mood congruent words prior to mood incongruent words, suggesting experimentally induced moods made mood congruent words more accessible. Isen et al. (1978) found individuals in a positive mood not only evaluated the products they own more favorably, but also recalled positively valenced words better than neutral or negative words from a list of vocabulary they had initially memorized. Similarly, Bower, Gilligan, and Monteiro (1981) revealed, following exposure to a narrative story, participants displayed an enhanced recall of the elements congruent with the affect they were experiencing at the moment of the recall test. Risk ind (1983) found recall latencies for mood congruent personal life experience memories were shorter than mood incongruent memories. This body of research overall points out moods enhance the accessibility of mood congruent information in memory. Because mo ods enhance the accessibility of mood congruent constructs in memory, positive moods lead to an optimistic outlook of one s life experiences, whereas negative moods bring a pessimistic outlook. For example, Forgas, Bower, and Kratz (1984) found participant s in a positive mood judge d their own social interaction more positively than those in other mood states, and this effect weakened in judgments of other s behavior. Forgas, Bower, and Moylan (1990) further reported participants in a happy mood tended to at tribute their own success to stable and internal causes and failure to external causes, and this pattern became less salient when participants judged others success and failure. In contrast, participants in a sad mood internally attributed their own failu re and externally attributed their own success, whereas a reverse pattern emerged in judgments of others success and failure.

PAGE 41

41 Forgas et al. (1990) suggested these effects were observed because happy moods primed positive and self appraising information w hereas sad moods made negative and self critical information more accessible in memory. Apparently, research overall suggests mood primed constructs have stronger mood congruent effects on judgments about one s own life experiences, rather than others exp eriences. However, using a state wide sample, Mayer, Gaschke, Braverman, and Evans (1992) found the effects of mood state were observed in judgments regarding non personal, generalized social events such as perceived chances of an atomic war and a spur in the state wide divorce rate. The mood as prime framework suggests that moods influence the perceived risk of depression in a mood congruent way because it primes and thus enhances the accessibility of mood congruent materials. Therefore, if participants do not receive information about clinical depression, the mood as prime framework would suggest that participants in sad moods, compared to those in happy moods, would overrate the risk of depression and have stronger help seeking intentions. This effect w ill occur because sad moods will activate other constructs associated with the feeling of sadness, such as various negative affective states (e.g., anxiety, regret, nervousness, and so on ) and life experiences with negative affective connotations (e.g., ma rital conflict, loss of spouse, sleeplessness, physical exhaustion, growing old, and so on ), and these constructs will be used for risk estimation. This account is especially relevant to this study because studies indicate that moods affect health cogniti on. For example, Croyle and Utretsky (1987) reported induced negative moods led people to generate a more negative evaluation of their health status and report more physical symptoms. Salovey and Birnbaum (1989) found people in a sad mood reported more phy sical symptoms, such as aches and pains, than those in a happy mood. Pettit et al. (2001) further

PAGE 42

42 reported positive moods reduced people s perceived experiences of negative health conditions, such as sinus, flu, and sore throat. Therefore, sad moods will p rime negative mental and physical experiences associated with a negative mood, and therefore will affect risk perceptions, resulting in the same effects predicted by the mood as information perspective. On the other hand, when participants receive informa tion about the symptoms of clinical depression and the APA diagnostic guideline, consumers may take the information into account to form risk perceptions of depression. The extent to which such information is processed and used for forming a judgment, howe ver, can differ due to a number of factors. For example, under a condition that allows extensive information processing, such as the absence of time pressure, Forgas (1995) suggested that the mood as prime framework would apply. Therefore, participants wou ld process the information relatively extensively to construct a risk estimate, and their risk perceptions would reflect both the influences of their mood state and the presence/absence of the diagnostic guideline. In comparison, under a condition that red uces information processing capacity, Forgas (1995) suggested that the mood as information principle would predominate. Therefore, moods will have direct effects on risk perceptions, relatively independent of the external information consumers refer to. Sy nthesizing the Two Perspectives In summary, the mood as information framework focuses on the direct, non cognitively mediated effects of mood state on judgment, whereas the mood as prime perspective focuses on the cognitive processes that mediate the effec ts of mood. Despite overall differences in the way the two theories conceptualize the mechanism of the effects, the literature indicates the two perspectives can most accurately be viewed as presenting two complementary accounts of mood effects on judgment s. The literature illuminates the two theories relative applicability and

PAGE 43

43 predictive validity vary due to differing conditions under which participants are exposed to persuasive communication, process information, and make judgments. Researchers (Dunn & Schweitzer, 2005; Fedorikhin & Cole, 2004; Petty et al.,1993; Pham et al., 2001; Schwarz & Clore, 1983; Siemer & Reisenzein, 1998) have pinpointed the conditions that determine each perspective s relative applicability. For example, Pham et al. (2001) and Verplanken, Hofstee, and Janssen (1998) revealed affective based evaluations of everyday stimuli occurred faster than cognition based evaluations. Pham et al. (2001) and Siemer and Reisenzein (1998) found the mood as information heuristic was more heavily used for judgments under time pressure. In turn, the very finding that affect based evaluation occurred faster than cognition based evaluation implies that the latter may apply more when participants have sufficient time to process information and make jud gments. Petty et al. (1993) indeed revealed that among participants with relatively higher need for cognition, mood lost its direct effects on judgments but retained its capability to influence judgments through cognitive processes. Perceived Diagnosticit y of Internally Retrieved Life Experiences When opportunity for risk and intention estimations is high, the mood as prime perspective suggests that mood state affects risk estimation through cognitive processes. Therefore, sad moods lead to higher risk per ceptions by making discomforting life experiences, including common symptoms of depression, more accessible in the consumer memory. In comparison, happy moods will make one s own experiences with negative affective connotations less retrievable. However, accessibility is not the only cognitive process that determines risk estimation. Once a negative life experience is retrieved, consumers may differ in their perception of how much it actually indicates they are clinically depressed. Following Raghubir & Me non (2005),

PAGE 44

44 the perceived degree that one s own experience of a negative life event indicates clinical depression is named in this study perceived diagnosticity of internally retrieved life experiences, or perceived diagnosticity in short. In summary, a negative life experience should be retrieved and perceived to be diagnostic of depression before it leads to higher risk perceptions and stronger help seeking intentions. How do consumers form perceptions of diagnosticity? Because consumers typically do not have sufficient knowledge about clinical depression to internalize their own diagnostic standards, when an external information source provides a diagnostic guideline, they will likely use it to interpret how much their retrieved negative life experie nces indicate clinical depression. Raghubir and Menon (2005) reported that exposure to an external information source influenced consumers perceived diagnosticity of life experiences. This study uses the content of an antidepressant advertisement to pro duce variances in the perceived diagnosticity of internally retrieved life experiences. Of many content elements that frequently appear in DTC antidepressant advertising, such as risk information and mechanisms of drug effects, this study focuses on the Am erican Psychiatric Association s diagnostic guideline intended for an appropriate application of the information on the symptoms of depression. The content of DTC advertising for many antidepressants, such as Prozac (produced by Elli Lilly) Zoloft ( Pfize r) Paxil CR (GlaxoSmithKlein, GSK) and Effexor XR (Wyeth) tends to focus on information about the symptoms of depression including sleeplessness, physical exhaustion, sluggishness, and hopelessness However the ad content may or may not offer a diagnost ic guideline that pinpoints how consumers can properly interpret symptom information in order to make an accurate self diagnosis of depression For example, some guideline s stipulate

PAGE 45

45 that a particular health condition should last each day for two weeks to be considered a symptom of clinical depression (e.g., These are some symptoms of depression. They must occur each day for at least two weeks and interfere with your daily life, as stated in a Zoloft advertisement). This diagnostic guideline, almost iden tical in nature to the diagnostic guideline for depression that American Psychiatric Association (APA) presents in its Diagnostic and Statistical Manual of Mental Disorders IV ( 1994 ) may or may not appear i n antidepressant advertisements for other leading brands This is important because t he guideline requiring that symptoms be present nearly every day for two weeks is likely to result in fewer self diagnoses of depression than a guideline that permits the inference that one is depressed based on observa tion of the symptoms for a day or a few days. In other words, the presence of the APA diagnostic guideline is important because it influences the perceived diagnosticity of consumers internally retrieved past experiences of possible symptoms of depressio n. In particular, if consumers receive the instruction that a particular life experience should persist two weeks before it can be accurately interpreted as a symptom of depression, they will be less likely to infer that having a life experience resembling a depression symptom for a day or a few days would indicate they might be clinically depressed. In summary, the presence of the diagnostic guideline is expected to reduce the perceived diagnosticity of the participants internally retrieved experiences o f the symptoms of depression, while the absence of the guideline has the opposite effects. By varying the presence/absence of the APA guideline in the advertisement, one can manipulate the participants perceived diagnosticity of past experiences as indica tors of depression. Once a symptom related past experience is perceived as a diagnostic input, it is likely to be perceived as an indicator of clinical depression and therefore increase the participant s risk perceptions of depression and

PAGE 46

46 help seeking inte ntions. If a past experience is perceived as less diagnostic, it is less likely to be referred to as a signal for clinical depression. Hypotheses Overall, the literature on the mood as information and mood as prime frameworks, especially a body of literat ure (Forgas,1992, 1995; Petty et al., 1993; Siemer & Reisenzein, 1998) that compares the relative applicability of the two complementary perspectives in varying situations, suggests that an environment reducing one s opportunity for deliberative risk and i ntention estimation is more conducive to the activation of a mood as information process. In contrast, an environment encouraging deliberative estimation is more likely to trigger a mood as prime process. In the experiment, participants mood state (sad v ersus happy), opportunity for risk and intention estimation (low versus high), and the perceived diagnosticity of life experiences (low versus high) will be manipulated. Then the perceived risk of depression and help seeking intentions will be measured. Wh en consumers do not have high opportunity for deliberation, the mood as information framework, rather than mood as prime, will predominate, producing only the main effects of mood state on risk perceptions without interacting with perceived diagnosticity. Therefore, sad mood individuals will perceive the risk of depression to be higher than happy mood individuals, regardless of the level of perceived diagnosticity. H1a : When opportunity is low, sad mood participants will report higher perceived future risk of depression than happy mood participants, whether perceived diagnosticity is low or high. Because risk perceptions are a major determinant of behavioral intentions for preventive and remedial actions, mood state will have similar effects on help seeking intentions. H1b : When opportunity is low, sad mood participants will have stronger help seeking intentions, whether perceived diagnosticity is low or high.

PAGE 47

47 H1a and H1b suggest that under the low opportunity condition, there will not be an interaction be tween mood state and perceived diagnosticity. Figure 2 1 and Figure 2 2 represent the expected results for risk perceptions and help seeking intentions under the low opportunity condition. These results will be observed if H1a and H1b are supported. When consumers have high opportunity for deliberation, the mood as prime perspective will be more applicable, leading to a significant interaction of mood state and perceived diagnosticity. Under the high opportunity condition, participants will be able to eng age in substantive information processing to form risk perceptions. Forgas (1995) suggested that when participants did not have to simplify their judgment due to the conditions that allowed extensive information processing, the mood as prime principle woul d dominate. Siemer and Reisenzein (1998) also reported that the effects of mood as information applied more prominently when participants made judgments under time pressure. Similarly, Petty et al. (1993) found that among participants with relatively low need for cognition, the mood as information principle dominated. Among participants with high need for cognition, mood states influenced judgments through influencing the thoughts generated about the judgment domain. To the extent that need for cognition can be considered as a referent to a person s trait cognitive capacity, the findings of Petty et al. (1993) may be considered as implying that the mood as prime perspective would apply better under the conditions that allow extensive information processing such as the high opportunity condition. In the current study, when participants have opportunity for deliberative risk estimation, they will engage in cognitive processing of information they received from the advertisement, whether perceived diagnostic ity is low or high. Because moods will make mood congruent experiences more accessible in memory, sad mood participants will perceive the symptoms

PAGE 48

48 presented in the advertisement to be more reflective of their life experiences than happy mood participants. Further, participants who receive high diagnosticity information will perceive their own internally retrieved experiences of the depression symptoms to be indicative of clinical depression. On the other hand, participants who receive low diagnosticity information, namely the APA type diagnostic guideline, will be less likely to perceive their experiences of depression symptoms to be indicative of depression. This will be the case because participants awareness of the guideline will make it less likely to judge that a person may be depressed based on observation of the symptoms for a day or two. In other words, participants under the low diagnosticity condition will perceive their experiences with the symptoms presented in the advertisement as less indic ative of their being clinically depressed, because, by definition, it is much harder to experience the symptoms each day for two consecutive weeks than to simply experience them for some time in their recent life. However, it is emphasized that the presen ce of an APA type guideline will have more effects on the perceived diagnosticity among sad mood participants than among happy mood participants. Compared to sad mood participants, happy subjects will initially have reported a substantially lower level of experiences of depression symptoms. If one rarely experienced a particular discomforting life event, chances will be already low that his/her experience of the event would be perceived to be indicative of depression. Therefore, when a symptom is perceived to be rarely experienced, exposure to the APA guideline would have limited capacity to further reduce the perceived degree that the experience indicates depression. H2a : When opportunity is high, exposure to the low diagnosticity information will reduce p erceived risk of depression significantly more among sad mood participants than among happy mood participants.

PAGE 49

49 Because risk perception is a major determinant of behavioral intentions for preventive and remedial actions, the following hypothesis is genera ted. H2b : When opportunity is high, exposure to the low diagnosticity information will reduce help seeking intentions significantly more among sad mood participants than among happy mood participants. Figure 2 3 and Figure 2 4 represent the expected res ults for risk perceptions and help seeking intentions under the high opportunity condition. These results will be observed if H2a and H2b are supported. H 1a and H1b suggested that under the low opportunity condition, mood state and perceived diagnosticit y would not interact to influence risk perceptions and intentions to seek professional help. Therefore, mood state would directly influence risk perceptions independent of the level of perceived diagnosticity. This was hypothesized to occur because under t he low opportunity condition, participants information processing capacity would be significantly reduced and become less extensive, and the mood as information principle would dominate. Therefore, participants under the low opportunity condition would di rectly refer to their current mood state to form an estimate of future depression risk in a simplified way (Forgas, 1995; Siemer & Reisenzein, 1998). Further, studies that reveal the association between risk perceptions and changes in health behavior or b ehavioral intentions (Block & Keller, 1998; Lipkus et al., 2001; Menon et al., 2002; Robinson et al., 1998) share the assumption that variations in risk perceptions cause changes in behavior, not the other way round. Therefore, the effects of mood manipula tion and perceived diagnosticity on help seeking intentions under the low opportunity condition will be moderated by their impact on the perceived future risk of depression. H3a : When opportunity is low, the effects of mood state on help seeking intention s will be mediated by the perceived future risk of depression.

PAGE 50

50 H2a and H2b indicated that under the high opportunity condition, there would be an interaction between mood state and perceived diagnosticity on risk perceptions and intentions to seek professi onal help. This was expected to occur because under the high opportunity condition, participants would engage in substantive information processing to form risk perceptions, and the mood as prime perspective would operate. Because risk perceptions are conc eptualized as a determinant of changes in health behavior or behavioral intentions (Block & Keller, 1998; Lipkus et al., 2001; Menon et al., 2002; Robinson et al., 1998) the following hypothesis is generated about the role of risk perception as a mediator of the interactive effects of mood state and perceived diagnosticity on help seeking intentions under the high opportunity condition. H3b : When opportunity is high, the interactive effects of mood state and perceived diagnosticity on help seeking intentio ns will be mediated by the perceived future risk of depression

PAGE 51

51 Figure 2 1 Expected ANOVA results for risk perception when opportunity was low Figure 2 2. Expec ted ANOVA results for help seeking intention when opportunity was low Low Diagnosticity High Diagnosticity Help Seeking Intentio n Happy Mood No Manipulation Sad Mood Low Diagnosticity High Diagnosticity Perceived Risk Happy Mood No Manipulation Sad Mood

PAGE 52

52 Figure 2 3 Expected ANOVA results for risk perception when opportunity was low Figure 2 4. Expected ANO VA results for help seeking intention when opportunity was high Low Diagnosticity High Diagnosticity Help Seeking Intention Happy Mood No Manipulation Sad Mood Low Diagnosticity High Diagnosticity Help Seeking Intention Happy Mo od No Manipulation Sad Mood

PAGE 53

53 CHAPTER 3 METHODOLOGY Design The objective of this study is to explore the effects of mood state, opportunity for risk estimation, and perceived diagnosticity of internally retrieved disease related life experiences on the consumer s perceived future risk of depression and intentions to seek professional help to discuss depression. To achieve this research purpose, an experiment was conducted with a 2 (evoked mood: sad versus happy combined with no mood manipulation) 2 (diagnosticity: low versus high) 2 (opportunity: low versus high) between subjects experimental design with a non factorial control group, producing thirteen experimental groups in total. Students enrolled in introductory a dvertising classes were recruited as participants. It is emphasized that the no mood manipulation condition is distinct from the non factorial control group. Participants in the former condition did not receive mood manipulation, but were exposed to the di agnosticity and opportunity manipulations. The control group did not receive any manipulation but completed the dependent measures. Because a moderate correlation was expected between risk perceptions and help seeking intentions, multivariate analysis of variance (MANOVA) was used as a primary statistical method of testing the three manipulated variables effects on the linear combination of the two dependent variables. If the three independent variables had significantly different effects on risk percepti ons and help seeking intentions, separate analyses of variance (ANOVAs) would be used to test how risk perceptions and help seeking intentions were influenced respectively. Last, analyses of simple effects were conducted to test if statistically significan t mean differences were observed among experimental groups as predicted by the six hypotheses.

PAGE 54

54 Participants Undergraduate students ( N = 269) enrolled in introductory advertising classes at the University of Florida participated in the study in return for c ourse credit. The class curricula did not include topics that might have sensitized subjects to the true purpose of the study. Though less preferable than a random sample from the general population, college students were considered as an appropriate sampl e for this study, considering depression is increasingly observed among young adults ( Kessler, Avenevoli, & Merikangas, 2001 ). However, homogeneity of the sample might yield results that differ from those observed in the general population. Approximately 68 percent ( n = 184) of the participants were females. Participants ranged in age from 17 to 29 ( M = 20.15, SD = 1.80), and included non Hispanic whites ( n = 172), Hispanics ( n = 40), African Americans ( n = 26), and Asian Americans ( n = 18). A total of 22 participants (8.20%) reported that they had previously been diagnosed as clinically depressed. A number of participants also had vicarious experiences with depression. For example, a total of 161 participants (59.9%) reported that their close others, such as family members, close relatives, or friends, had suffered from depression. Further, 155 participants (57.6%) were aware that their close others had sought professional help to treat depression, and 157 (58.4%) knew of close others who had taken antidepr essant medication. Only 24 percent ( n = 65) reported none of these vicarious experiences. A total of 15 participants ( n = 15) reported they had previously been diagnosed as having attention deficit hyperactivity disorder (ADHD). Procedure Subjects signed up for participation in class and were invited to a computer laboratory where the experiment was conducted. Sessions were run with groups of 15 to 24. Upon arrival at the laboratory, the informed consent was secured, and the participants were randomly assi gned to

PAGE 55

55 one of thirteen experimental groups. Then participants were instructed that the purpose of the research was to explore how people represent their autobiographical memories (disguised as Study One), and how students understand consumer targeted anti depressant advertisements (disguised as Study Two) Constructs were manipulated and measured using computer software named MediaLab. The software offers functions vital for experimental studies, such as presenting experimental stimuli in a controlled manne r, recording response times, and measuring variables. Participants who received mood manipulation proceeded in the order of mood induction, mood manipulation check, exposure to the advertisement, diagnosticity manipulation check, opportunity manipulatio n, opportunity manipulation check, and measurement of dependent variables. Participants who did not receive mood manipulation but received the other two manipulations, named the no mood manipulation participants, first completed the mood scale, and proceed ed in the same order of events. Participants who did not receive any manipulation, named the non factorial control group, first completed the mood scale and then the dependent measures. Both the no mood manipulation and control groups received additional f iller tasks after they completed the dependent measures, so that all the thirteen experimental groups spent approximately the same amount of time in each session. The mood induction procedure and mood manipulation check were disguised as Study One. Part icipants were informed in writing that their responses would be used to explore how people construct autobiographical memories. Exposure to the DTC antidepressant advertisement and the measurement of dependent variables were disguised as Study Two, which was ostensibly designed to explore college students evaluation of an early draft of an antidepressant advertisement. Much research on the effects of mood on social cognition uses similar study

PAGE 56

56 procedures to alleviate participants suspicion about the tru e purpose of research (Schwarz & Clore, 1983; Raghunathan & Pham, 1999; Siemer & Reisenzein, 1998). Two trained experimenters administered sessions. Each session consisted of the mood induction procedure, disguised as Study One, and exposure to the stim ulus and completion of dependent measures, disguised as Study Two. The two experimenters were ostensibly in charge of the two separate studies. To give the impression that the procedures represented two separate studies, two informed consents were sec ured. For the same purpose, the experimenters told participants that although we conduct two separate studies, we run them in one session for the sake of convenience. Further, after completing the mood induction procedure, participants were instructed in writing that this is the end of Study One. To proceed to Study Two, please click on Continue. Detailed instruction of the two procedures was presented only in writing, because participants in different experimental groups proceeded in different orders o f events. During each session, the primary experimenter ensured that participants sat apart from each other and did not talk with each other, s o that they complete d the instrument independently in an orderly environment To help differentiate clinical dep ression from short periods of sadness, participants were informed in writing that clinical depression is defined in this project as a form of medical illness that may require doctor s intervention for treatment. The instruction did not include symptoms o f depression, because information about depression symptoms and a guideline about how to interpret them for the self diagnosis of depression would be used to manipulate diagnosticity. In summary, participants were randomly assigned to one of thirteen exper imental groups determined by induced mood state (sadness versus happiness versus no mood manipulation),

PAGE 57

57 perceived diagnosticity of recalled life experiences (low versus high diagnosticity), and opportunity for risk and intention estimation (low versus hig h opportunity) in addition to a non factorial control group that did not receive any manipulation but only completed dependent variable measures. Each session took about 25 minutes to complete, regardless of experimental groups. After the dependent variabl es were measured, participants were debriefed and thanked. Independent Variables The three manipulated variables of this study included mood state, diagnosticity of internally retrieved depression related life events, and opportunity for risk and intentio n estimation. Mood state was manipulated by requiring participants to write about life events that had evoked very sad or happy moods in the past. Diagnosticity was manipulated by exposing participants to a fictitious antidepressant advertisement that redu ced or increased the perceived degree that a list of discomforting life experiences indicated clinical depression. Opportunity was manipulated by requiring participants to either estimate and report their risk perceptions and help seeking intentions as fas t as possible or take as much time as they needed to estimate and report them. Mood S tate Sad and happy moods were manipulated using the typical mood induction technique suggested by Schwarz and Clore (1983). To manipulate sad moods, participants were r equired to describe past life events that had made them very sad in the past. To manipulate happy moods, participants wrote about happy life events. In the mood manipulation procedure, disguised as a study about college students autobiographical memories participants first received the following general instruction. People experience many types of life events. Study One is designed to build a life event inventory and explore how people represent their autobiographical memories. For that

PAGE 58

58 purpose, you will be asked to describe three life events that made you sad [happy]. Click on Continue to proceed. The instruction had two versions. The sad mood version was directed to participants in the sad mood condition, whereas the happy mood version was delivered t o those in the happy mood condition. Otherwise, the two versions of general instruction had the same wording. After the general instruction, participants were instructed on the ma nn er in which they were encouraged to describe the events. In particular, fo llowing the suggestion of Dunn and Schweitzer (2005), participants were requested to write about the sadness (versus happiness) inducing life events as realistically as possible. Please describe a life event that made you very sad [happy] as realistically as possible, such that a person reading the description would become sad [happy] just from hearing about the situation. Please spend five minutes for this situation. After five minutes, the screen will automatically proceed to the next step. Similar to th e general instruction, the script was adjusted to the particular mood state (sadness versus happiness) that it was designed to manipulate. A manipulation check follow ed the mood induction procedure. A pretest, named Pilot Study One, was conducted prior to the main study to ensure that the mood manipulation was effective. Participants in the no mood manipulation condition did not undergo a mood induction procedure. Instead, they only reported their current moods, received diagnosticity and opportunity manipu lations, and completed the dependent measures. Then they received a directed writing procedure as a filler task. Similar to the mood induction procedure, the filler task required participants to write about three life events that made them feel very sad or happy in the past. However, the task was conducted after all the measures, including the mood check scale, were completed. Therefore, the writing filler task could not affect the participants responses to the mood scale and the dependent measures in any way. The task was conducted for the sole

PAGE 59

59 purpose of making no mood manipulation participants spend as much time as participants in the sad mood or happy mood condition did. The no mood manipulation condition was differentiated from the non factorial contr ol group, in which participants did not receive any manipulation but simply completed risk perception and help seeking intention measures. The control group also received a directed writing procedure as a filler task after they completed dependent measures Therefore, all thirteen experimental groups spent approximately the same amount of time in each session. Diagnosticity Diagnosticity was defined as the degree that participants perceived their own experiences of sleeplessness, feeling low in physical ene rgy, depressed mood, and difficulty making decisions indicated clinical depression. The diagnosticity manipulation procedure was disguised as Study Two, which was ostensibly separate from Study One, or the mood induction procedure. T he manipulation was introduced as a study of consumer responses to advertising. Participants were instructed that the study was an advertising copy test, designed to explore how consumers responded to an early version of a print advertisement for an antidepressant recently l aunched in the market. Diagnosticity was manipulated by exposing participants to two discrete versions of the antidepressant advertisement (Appendix B ). The top half of the ad listed the four life experiences typically considered as common symptoms of dep ression, including low energy, depressed mood, sleep problems, and difficulty making decisions. Then, at the start of the bottom half, the high diagnosticity version indicated that these life experiences were symptoms of depression. These [listed experienc es] are symptoms of depression. Some may say it s just in your head. But depression is a real disease with real medical causes. While the cause is unknown, depression may be related to an imbalance of chemicals in the brain. Clinical studies show Serexa CR can effectively correct this imbalance and relieve symptoms of depression.

PAGE 60

60 In contrast, the low diagnosticity version suggested at the start of the bottom half that consumers needed to be cautious in concluding that the listed life experiences actual ly indicated clinical depression. To that effect, they were given the following instruction. These [listed experiences] are symptoms of depression only if they last nearly every day for two weeks. Some may say it s just in your head. But depression is a real disease with real medical causes. While the cause is unknown, depression may be related to an imbalance of chemicals in the brain. Clinical studies show Serexa CR can effectively correct this imbalance and relieve symptoms of depression. The APA guid eline, or the instruction that the life experiences should occur nearly every day for two weeks to be considered as symptoms of depression, was presented in a light blue color. The procedure ended with a manipulation check. The rationale for this manipula tion was that the APA guideline requiring that the listed problems be present nearly every day for two weeks to be considered as indicative of depression was likely to reduce consumer perceptions of the degree that their own experiences of the problems ind icated they were clinically depressed. In contrast, simply stating that the problems are depression symptoms, as the low diagnosticity copy did, would permit the inference that one is depressed based on observation of the experiences for a day or a few day s. A pretest, named Pilot Study Two, was conducted to ensure that the manipulation was effective. Opportunity After exposure to the advertisement, risk perceptions and help seeking intentions were measured. Opportunity was manipulated by giving participant s two discrete instructions for completing the measures for risk perceptions and help seeking intentions. Before they received the dependent variable measures, participants in the low opportunity condition were instructed they needed to complete the measur es as fast as possible. Read this instruction VERY carefully. What follows is a questionnaire on your life and depression. In the real world, consumers often make quick judgments while they are busy.

PAGE 61

61 To make this study as realistic as possible, please co mplete the following three questions AS FAST AS YOU CAN. Then participants in the low opportunity condition were reminded of the instruction when they received and completed dependent measures. For example, to complete the first measurement item for risk perceptions, they were given the following instruction. In you thinking, what are the chances that you will suffer from clinical depression in the near future? AS FAST AS YOU CAN, report in percentage between 0% and 100%. Then click Continue to proceed In contrast, participants in the high opportunity condition were encouraged to take as much time as they needed for deliberation. The following is the instruction presented before they received dependent variable measures. Read this instruction VERY car efully. What follows is a questionnaire on your life and depression. Researchers point out that having accurate ideas about a disease is important for preventing or treating the disease. So please take AS MUCH TIME AS YOU NEED to deliberate sufficiently. T hen when the participants in the high opportunity condition received and completed the measures for risk perceptions and help seeking intentions, they were further reminded. In you thinking, what are the chances that you will suffer from clinical depressio n in the near future? Take AS MUCH TIME AS YOU NEED, and report in percentage between 0% and 100%. Then click Continue to proceed. To check the opportunity manipulation, the amount of time participants took to answer each of the risk perception and inten tion measurement items was automatically recorded by Medialab, the software used for this study. Mean differences in response time across the two opportunity conditions were used to check the manipulation. Stimulus To manipulate the perceived diagnosticity of depression related life experiences, two versions of a DTC print advertisement for a fictitious antidepressant brand named Serexa CR were created (Appendix B ). Both versions of the advertisement included a visual illustration of a

PAGE 62

62 dark, cloudy sky, wit h a ray of sunshine streaming through the cloud onto the sea. Around the sunray, four common symptoms of depression were listed, including low energy, depression, sleep problems, and difficulty making decisions. This top half section remained the same acro ss all experimental groups. The bottom half included body copy with the diagnosticity manipulation. The high diagnosticity version stated that the listed negative life experiences were common symptoms of depression. In contrast, the low diagnosticity versi on emphasized that the experiences were considered as symptoms of depression only if they lasted nearly every day for two weeks. The remaining copy was equivalent across the two versions. It included further information about the advertised drug, such as s ide effects, safety warning, and mechanism of action. The copy and layout of the advertisement were designed to resemble real antidepressant advertisements. Development of the Stimulus A pretest was conducted to select depression symptoms to be included i n the advertisement. The purpose was to determine a list of potential symptoms of clinical depression frequently experienced by undergraduates. The rationale was that diagnosticity would be more effectively manipulated if frequently, rather than infrequent ly, experienced discomforting life events were presented as potential symptoms of depression. Exposure to the APA guideline was expected to reduce the degree that a participant perceived his/her experience of a discomforting life event, such as sleep prob lems or low physical energy, to be diagnostic of depression. This is because the guideline would inform participants that an experience should last nearly every day for two weeks to be considered diagnostic of depression, and therefore would prohibit the i nference that one is depressed based on observation of the experience for a day or a few days.

PAGE 63

63 If one rarely experienced a particular discomforting life event, chances will be already low that his/her experience of the event would be perceived to be indic ative of depression. Therefore, when a rarely experienced symptom is presented, exposure to the APA guideline would have limited capacity to further reduce the perceived degree that the experience indicates depression. A total of 29 symptoms were compiled from two widely accepted measures of depression with established validity and reliability (APA 1994; Zung 1965). The four frequent symptoms thus selected were sleep disorder, feeling tired for no reason, difficulty with thinking and making decisions, and d epressed feelings. These four symptoms were presented in the advertisement as sleep problems, low energy, difficulty making decisions, and depressed, considering how these selected symptoms were represented in actual DTC antidepressant advertising campaigns. Pilot Study 1 A pilot study was conducted to ensure that the mood induction procedure manipulated current mood state. To induce happiness or sadness, 72 participants completed a directed writing task designed to create sad or happy moods. Simil ar to the mood induction procedure typically used to manipulate affective states (Lerner & Keltner, 2001; Schwarz & Clore, 1983; Strack, Schwarz, & Gschneidinger, 1985), the writing task required participants to describe three life events that made them ve ry happy (versus sad) in the past as realistically as possible. Then participants reported their current moods on the following three item, seven point semantic differential scale: item (a) 1 = gloomy, 7 = joyful; item (b) 1 = sad, 7 = happy; and item ( c) 1 = upset, 7 = elated. The three check items were internally consistent ( = .94) and therefore were averaged into a single scale Collected d ata supported the prediction that the participants who wrote about happiness evoking life events would repo rt a happier mood state on the average [ t (70) = 6.59, p

PAGE 64

64 < .001 M sad = 2.40, SD = .90, M happy = 4.01, SD = 1.14]. The mood induction procedure was equally effective for male [ t (24) = 5.15, p < .001] and female [ t (44) = 4.56, p < .001] participants. This fi nding was confirmed by analysis of variance (ANOVA), which showed the mood gender interaction was not significant [ F (1, 68) = .51, p > .05)]. Males and females also did not differ in their reported moods [ t (70) = .96, p > .05]. Pilot Study 2 Pilot stud y 2 was designed to ensure that exposure to the APA guideline would reduce the perceived diagnosticity of ones own experience of common depression symptoms. If the manipulation was successful, participants only exposed to depression symptoms should percei ve their own experiences of the symptoms to be more indicative of clinical depression, compared to those exposed to the symptoms and the APA guideline. In other words, exposure to the APA guideline would reduce the perceived diagnosticity of ones own reca lled experiences of the symptoms presented in the advertisement. A total of 42 participants were recruited from an introductory advertising class, and were assigned to either the low diagnosticity (with the APA guideline) or the high diagnosticity (witho ut the APA guideline) condition. Participants were first exposed to an antidepressant advertisement that contained four depression symptoms, including low energy, depressed mood, sleep problems, and difficulty making decisions. Except of the diagnosticity manipulation, the ad was equal across the two diagnosticity conditions. Then participants reported on a seven point scale (1 = never, 7 = nearly every day) how often they experienced the symptoms during the last two weeks. Then, for each symptom, the parti cipants reported on a seven point scale (1 = not indicative at all, 7 = very indicative) the degree that their reported experience of the symptom was indicative of clinical depression (Raghubir & Menon, 2005).

PAGE 65

65 Responses to the four check items were inter nally consistent ( = .95). Therefore, the items were averaged into a single scale. The diagnosticity manipulation was successful. The perceived diagnosticity of ones own experiences of depression related symptoms was higher among the participants who did not receive the APA guideline than among those who received the APA guideline [ t (46) = 2.20, p < .05, M low diagnosticity = 2.88, SD = 2.02, M high diagnosticity = 4.15, SD = 2.00]. ANOVA revealed the manipulation was equally effective among male and female participants, because the gender diagnosticity interaction was not significant [ F (1, 43) = 1.80, p > .05]. Dependent Variables Perceived Future Risk of Depression A three item measure of perceived future risk of depression was used, as suggested by Le vy, Shea, Williams, Quistberg, and Armstrong (2006). Participants first answered in percentage the following single item question: item (a), In your thinking, what are the chances that you will suffer from clinical depression in the near future? Please re port in percentage between 0% (no chance of depression) and 100% (definitely will develop depression). Similar measures of perceived absolute risk have been frequently applied in the literature (Loroz & Lichtenstein, 2004; Raghubir & Menon, 2001; Levy, Sh ea, Williams, Quistberg, & Armstrong, 2006 for a review). The second item asked participants to respond on a seven point scale (1 = very low, 7 = very high) to the following statement: item (b), In your thinking, your risk of suffering from clinical depre ssion in the near future will be __ _____ Last, by checking a seven point scale (1 = much lower, 7 = much higher), participants answered the following question: item (c), In your thinking, compared to people of your age, your risk of developing clinical depression in the near future will be ________.

PAGE 66

66 Reponses to the three items were Z transformed and then summated into a single index score representing the participant s perceived future risk of clinical depression ( M = .00, SD = 2.67). After Z transfor mation, the three items were internally consistent ( = .82). To check the dimensionality of the three Z transformed risk measurement items, they were factor analyzed using a principal axis factoring extraction method with Varimax rotation. One factor was extracted. The factor accounted for 74% of the varia nce, with an Eigenvalue of 2.22. Table 3 1 shows factor loadings. The lowest factor loading was .66, which was for item (c). The scree plot shows that the three Z transformed risk perception items were summarized into one factor. The high internal consiste ncy ( = .82) further confirmed that only one underlying dimension existed for the three Z transformed risk perception items (Figure 3 1 ). Intentions to Seek Professional Help To measure intentions to seek professional help to discuss depression, participants reported on a three item, seven point scale (1 = strongly disagree, 7 = strongly agree) their agreement with the following three statements: item (a), If the University Health Services offered a free screening day for depression, I would intend to partic ipate; item (b), If the University Health Services offered a free educational program about depression, I would intend to participate; and item (c), If the University Health Services offered an opportunity to consult doctors about depression, I would i ntend to participate. This scale was modified and expanded from the single item scale used by Raghubir and Menon (2005). Responses to the three items were averaged into a single index score ( M = 2.63, SD = 1.01). To explore the underlying factor structur e of the three intention measurement items, a principal axis factor analysis with Varimax rotation was conducted for the three items. One factor was extracted, and it accounted for 75.56% of the variance, with an Eigenvalue of 2.27. Table 3 2 shows factor loadings. The lowest factor loading was .74, which was for item (a). The

PAGE 67

67 scree plot shows that the three measurement items fell into one factor (Figure 3 2). The high internal consistency revealed in Pilot study 2 ( = .92) as well as the main study ( = 84) further confirmed the finding that only one dimension existed for the three measurement items.

PAGE 68

68 Table 3 1 Factor analysis of risk perception ( = .82) Item Factor loadings Cronbach s if deleted Item (a): Risk in percentage .82 .73 Item (b): Non comparative risk .87 .70 Item (c): Relative risk .66 .83 Table 3 2. Factor analysis of help seeking intention ( = .84) Item Factor loadings Cronbach s if deleted Item (a): Risk in percentage .74 .81 Item (b): Non comparative ris k .82 .76 Item (c): Relative risk .83 .75

PAGE 69

69 Figure 3 1. Eigenvalue plot for scree test for risk perception Figure 3 2. Eigenvalue plot for scree test for help seeking int ention 0 1 2 3 Eigenvalue Factor Number 1 2 3 0 1 2 3 Eigenvalue Factor Number 1 2 3

PAGE 70

70 CHAPTER 4 RESULTS This chapter reports the results of the study described in Chapter 3. The chapter begins by presenting the manipulation checks for the three independent variables. Then correlations among a list of variables are presented as a preliminary analysis. The results of hypothesis testing are then reported, followed by a summary of additional data analyses to explore unexpected but theoretically meaningful relationships among variables. The chapter concludes with a summary of res earch findings. Manipulation Checks Mood State Following mood induction, participants reported their current mood state on the following three item, seven point semantic differential scale: item (a), 1 = gloomy, 7 = joyful; item (b), 1 = sad, 7 = happy ; and item (c), 1 = upset, 7 = elated. The no mood manipulation group did not receive a mood induction procedure but reported their current mood state. The three items were internally consistent ( = .95), and therefore were averaged into a single mood scale ( M = 4.38, SD = 1.39). For manipulation check, the single mood scale was submitted to one way ANOVA, treating mood manipulation as the sole independent factor. As Table 4 1 shows, significant mean differences existed among the sad mood, happy mood, and no mood manipulation conditions [ F (2, 247) = 52.62, p < .01, M sad = 3.28, SD = 1.35, M happy = 4.96, SD = 1.10, M no mood manipulation = 4.84, SD = 1.08]. Table 4 1 also shows the ANOVA results for the three measurement items that constitute the single m ood scale ( = .95). Bonferonni post hoc test revealed that participants in the sad mood condition differed significantly in the average mood state from those in the happy mood ( M sad = 3.28, SD = 1.35,

PAGE 71

71 M happy = 4.96, SD = 1.01, p < .01) and the no mood man ipulation ( M sad = 3.28, SD = 1.35, M no mood manipulation = 4.84, SD = 1.06, p < .01) conditions. This suggested that sad mood participants would significantly differ in the perceived future risk of depression and help seeking intentions from those in the t wo other mood conditions. In contrast, participants in the happy mood and no mood manipulation conditions did not significantly differ in their mood state ( M happy = 4.96, SD = 1.01, M no mood manipulation = 4.84, SD = 1.06, p > .05), as typically reported in prior research on mood and social cognition (Schwarz & Clore, 1983). This suggested the two mood conditions would not lead to significant mean differences in risk perceptions and help seeking intentions. Further, in using MANOVA for hypothesis testing, the happy mood and no mood manipulation conditions were combined into a combined mood condition. The sad mood and combined mood participants were significantly different in their current mood states ( M sad = 3.28, SD = 1.35, M combined mood = 4.90, SD = 1. 09, p < .01). If the mood manipulation was successful, it would be the only factor of the three independent variables that significantly affected participants current mood state, and the effects would not be moderated by other independent variables. To t est these requirements, the current mood state was submitted to a 2 (mood state: sad versus combined mood) 2 (diagnosticity: low versus high) 2 (opportunity: low versus high) ANOVA. As Table 4 2 shows, the mood manipulation was the sole independent fac tor that significantly influenced the current mood state [ F (1, 242) = 101.16, p < .01, p = .295]. No other independent variables or their interactions were significant. Therefore, the mood manipulation was successful.

PAGE 72

72 Perceived Diagnosticity of Discomfor ting Life Experiences On a four item, seven point (1 = never, 7 = nearly everyday) scale, participants reported how often they underwent each of the following four discomforting life experiences for the last two weeks: low energy, feeling depressed, sleep problems, and difficulty making decisions. The experiences are typically considered as common symptoms of depression. Then participants indicated on a four item, seven point (1 = not at all likely, 7 = very likely) scale the likelihood that their own repor ted experience of low energy, depressed mood, sleep problems, and difficulty making decisions would indicate they were clinically depressed. The four items were internally consistent ( = .88), and therefore were averaged into a single scale representing p erceived diagnosticity ( M = 3.02, SD = 1.56). Perceived diagnosticity was submitted to one way ANOVA, entering the diagnosticity manipulation as the sole independent factor. As Table 4 3 shows, the mean difference was significant between the low and high diagnosticity conditions [ F (1,248) = 7.35, p < .01, M low diagnosticity = 2.79, SD = 1.52, M high diagnosticity = 3.31, SD = 1.55]. Table 4 3 also shows the ANOVA results for the four items that constituted the diagnosticity scale Further, to test if the d iagnosticity manipulation was the sole factor of the three independent variables that affected perceived diagnosticity and no other independent variables moderated the effects, perceived diagnosticity was submitted to a 2 (mood state: sad versus combined m ood) 2 (diagnosticity: low versus high) 2 (opportunity: low versus high) ANOVA. As Table 4 4 shows, the diagnosticity manipulation was the sole independent factor that significantly affected perceived diagnosticity [ F (1, 242) = 6.86, p < .01, p = .028]. No other independent variables or their interactions had significant effects.

PAGE 73

73 Opportunity for Risk and Intention Estimation The opportunity manipulation was checked by measuring the amount of time participants spent responding to the three ri sk perception measurement items and the three help seeking intention measurement items. Medialab recorded response times by millisecond, and the unit was later transformed into seconds. The response times for the three risk perception measurement items wer e submitted to separate ANOVAs, with the opportunity manipulation entered as the sole independent factor. As Table 4 5 shows, statistically significant mean differences in response time were observed for item (a) [ F (1, 248) = 57.84, p < .01, M high opportun ity = 22.73, SD = 10.77, M low opportunity = 14.75, SD = 4.55], item (b) [ F (1, 248) = 26.24, p < .01, M high opportunity = 12.82, SD = 4.82, M low opportunity = 9.86, SD = 4.28], and item (c) [ F (1, 248) = 17.24, p < .01, M high opportunity = 10.65, SD = 5.25, M low opportunity = 8.28, SD = 3.61 ]. However, it was noticeable that the mean difference was greatest for item (a), and then dropped as participants proceeded to item (b) and item (c). Especially, the response times of the high opportunity participants wer e gradually reduced and became closer to the response times of the low opportunity participants. The response times for the three intention measurement items were also submitted to separate ANOVAs. As Table 4 5 shows, the two opportunity conditions led to significant mean differences on item (a ) [ F (1, 248) = 9.89, p < .01, M low opportunity = 9.50, SD = 3.54, M high opportunity = 10.97, SD = 3.80] and item (b) [ F (1, 248) = 5.83, p < .02, M low opportunity = 7.35, SD = 2.61, M high opportunity = 8.44, SD = 4.27] However, the mean difference on item (c) was not significant [ F (1, 248) = .00, p > .05, M low opportunity = 7.43, SD = 3.55, M high opportunity = 7.41, SD = 4.22 ]. It was observed that the mean difference dropped as participants proceeded from item (a) to item (b) and item (c). Especially, the response times of the high opportunity participants became gradually reduced. This implied that in this study moods might have similar effects on risk

PAGE 74

74 perceptions and help seeking intentions across the two opportunity groups, because variances in opportunity were hypothesized to cause the mood state to affect social judgments differentially. Correlations Among Variables Before hypotheses were tested, correlations among a number of variables were explored (Table 4 6) Age or gender was not significantly related with any other variable. Vicarious experience of clinical depression was positively related with perceived future risk of depression ( r = .23, p < .01) and help seeking intentions ( r = .23, p < .01). This conf irmed Park and Grows (2008) finding that the more vicarious experience of clinical depression one has, the higher risk perceptions and the stronger help seeking intentions they tended to report. In addition, risk perceptions and help seeking intentions we re also positively correlated ( r = .40, p < .01 ). This confirmed that it was appropriate to use MANOVA, rather than separate ANOVAs, as a primary method of testing H1a through H2b. Testing Hypotheses 1a and 1b Hypotheses 1a and 1b were designed to examine the effects of mood state and perceived diagnosticity on risk perceptions and help seeking intentions when the opportunity for risk and intention estimation was low. The two hypotheses predicted that under the low opportunity condition, sad mood participa nts would report higher risk perceptions (H1a) and stronger help seeking intentions (H1b) than happy mood participants, whether perceived diagnosticity was low or high. MANOVA was used as a primary method of testing the two hypotheses, because it is a de pendence technique that measures the differences for two or more metric variables based on a set of categorical (non metric) variables (Hair, Anderson, Tatham, & Black, 1998, p. 326). In dealing with multiple dependent variables, using MANOVA instead of s eparate ANOVAs reduces the probability of making a Type I error as well as increases the statistical models

PAGE 75

75 power to detect significant group differences, especially when the dependent variables are correlated (Hair et al., 1998). This further suggested M ANOVA should be used to test H1a and H1b, because risk perceptions and help seeking intentions were significantly correlated ( r = .40, p < .01, Table 4 6). H1a and H1b would be supported if the multivariate analysis indicated that when opportunity was low, mood had main effects on the linear combination of risk perceptions (H1a) and help seeking intentions (H1b), and the effects were not moderated by perceived diagnosticity. The two hypotheses would be further supported if analyses of simple effects reveale d, under the low opportunity condition, sad mood participants reported significantly higher risk perceptions (H1a) and stronger help seeking intentions (H1b) than happy mood participants, whether perceived diagnosticity was low or high. MANOVA Results for H1a and H1b To test H1a and H1b, risk perceptions and help seeking intentions for low opportunity participants were submitted to a 2 (mood state: sad versus combined mood) 2 (perceived diagnosticity: low versus high) MANOVA. Because the happy mood and no mood manipulation participants did not differ in their mood state, the two conditions were combined. In addition, H1a and H1b predicted the main effects of mood state when opportunity was low. Therefore, experimental groups under the high opportunity co ndition were deleted from this analysis. The multivariate results supported H1a and H1b. As Table 4 7 shows, mood state had significant main effects on the linear combination of risk perceptions and help seeking intentions [Wilks F (2, 119) = 5.59, p < .01, p 2 = .086]. Further in support of the two hypotheses, F (2, 119) = .35, p > .05 p 2 = .006] nor the mood F (2, 119) = .37, p > .05, p 2 = .006] was

PAGE 76

76 significant. Figure 4 1 and 4 2 show that the main effects of mood state on risk perceptions and help seeking intentions were approximately equivalent whether perceived diagnosticity was low or high. Homogeneity of variance covari ance matrices was not violated [Boxs M = 17.30, F (9, 51231.81) = 1.86, p > .05]. Mixed MANOVA Results for H1a and H1b The MANOVA results in support of H1a and H1b would be invalidated if mood state had significantly different effects on risk perceptions a nd help seeking intentions. If the effects were differential, conducting separate ANOVAs would be more appropriate than a MANOVA for testing the two hypotheses. To rule out this possibility, a three way mixed MANOVA was designed. The model included mood st ate (sad versus combined mood) and perceived diagnosticity (low versus high) as between subjects factors and treated risk perceptions and help seeking intentions as a within subjects factor. To make the scores on each of the two dependent variables compara ble, the scales for risk perceptions and help seeking intentions were Z transformed. Only groups under the low opportunity condition were included in the analysis. As Table 4 8 shows, the multivariate results indicted that the interaction of mood state an d the within subject factor was not statistically significant [Wilks = .98, F (1, 120) = 2.51, p > .05 p 2 = .020]. This ruled out the possibility that mood state might have had differential effects on risk perceptions and help seeking intentions and th erefore the MANOVA results should be invalidated. The within subjects factor also did not moderate the effects of either perceived diagnosticity or the mood state perceived diagnosticity interaction. Therefore, the previous MANOVA results in support of H 1a and H1b were valid, and follow up ANOVAs were not necessary. Boxs test revealed that the covariance matrices were homogeneous across groups [ M = 17.30, F (9, 51231.81) = 1.86, p > .05].

PAGE 77

77 Analyses of Simple Effects As a second test of H1a and H1b, analys es of simple effects were conducted. Similar to the multivariate analysis, the happy mood and no mood manipulation conditions were combined. H1a and H1b would be supported if, when opportunity was low, sad mood participants reported higher risk perceptions and stronger help seeking intentions than combined mood participants. The two hypotheses would be further supported if the significant mean differences between sad mood and combined mood participants were observed whether perceived diagnosticity was low o r high. Risk perceptions. Table 4 9 summarizes the group means and standard deviations of risk perception. Levenes test showed that variances were equivalent across the groups [ F (8, 260) = .68, p > .05]. Simple effects tests revealed that when opportunity was low and perceived diagnosticity was high, sad mood participants reported higher risk perceptions than combined mood participants [ t (260) = 3.00, p < .01, M sad mood&high diagnosticity = 1.88, SD = 2.35, M combined mood&high diagnosticity = .03, SD = 2.0 3 ], supporting H1a. The mean difference was insignificant when perceived diagnosticity was low [ t (260) = 1.69, p > .05, M sad mood&low diagnosticity = 1.15, SD = 2.34, M combined mood&low diagnosticity = .03, SD = 2.36 ]. Therefore, mood state predicted risk perceptions only for high diagnosticity participants. This failed to support H1a, because the hypothesis predicted that sad moods would cause higher risk perceptions for both the levels of perceived diagnosticity. However, the effects of mood on risk perce ptions under the low diagnosticity condition were marginally significant ( p < .10). Therefore, the analyses of simple effects did not completely fail to support H1a. Analyses of simple effects also revealed that perceived diagnosticity had no significant effects on risk perceptions for participants in either the sad mood [ t (260) = .98, p > .05, M sad mood&high diagnosticity = 1.88, SD = 2.35 M sad mood&low diagnosticity = 1.15, SD = 2.34 ] or the combined

PAGE 78

78 mood [ t (260) = .01, p > .05, M combined mood&high diag nosticity = .03, SD = 2.30, M combined mood&low diagnosticity = .03, SD = 2.36 ] condition. This supported the predictions made in H1a. Planned contrasts overall revealed that when opportunity was low, mood state was the only factor that produced a signific ant mean difference in the perceived future risk of depression. However, the effects of mood state were significant only when perceived diagnosticity was high. The effects were only marginally significant ( p < .10) when diagnosticity was low. Therefore, th e results only moderately supported H1a. Help seeking intentions. Planned contrasts were also used to explore mean differences in help seeking intentions. Table 4 10 summarizes the group means and standard deviations of help seeking intentions when the hap py mood and no mood manipulation conditions were combined. Levenes test showed that variances were not equivalent across groups [F(8, 260) = 2.48, p < .05]. Planned contrasts revealed that the sad mood condition did not lead to significantly higher help s eeking intentions whether diagnosticity was high [ t (27.39) = 1.38, p > .05, M sad mood&high diagnosticity = 2.95, SD = 1.24, M combined mood&high diagnosticity = 2.56, SD = .68 ] or low [ t (43.19) = .64, p > .05, M sad mood&low diagnosticity = 2.75, SD = .94, M combined mood&low diagnosticity = 2.58, SD = 1.06 ], failing to support H1b. Because planned contrasts supported H1a but failed to support H1b, one may conclude that the effects of mood state were stronger on risk perception than on help seeking intentions. However, the reduction of the effects of mood state was not considered statistically significant, because the previous mixed MANOVA results revealed that the effects of mood state were equivalent on risk perceptions and help seeking intentions [Wilks = .98, F (1, 120) = 2.51, p > .05 p 2 = .020]

PAGE 79

79 The effects of perceived diagnosticity were also not significant on help seeking intentions whether participant were under the sad mood [ t (38.79) = .60, p > .05, M sad mood&high diagnosticity = 2.95, SD = 1.24, M sad mood&low diagnosticity = 2.75, SD = .94 ] or the combined mood condition [ t (59.27) = .07, p > .05, M combined mood&high diagnosticity = 2.56, SD = .68, M combined mood&low diagnosticity = 2.58, SD = 1.06 ], in support of the predictions made in H1b. Over all, planned contrasts failed to support H1b. Simple effects when the happy mood and no mood manipulation conditions were not combined. Table 4 11 and 4 12 summarize the group means and standard deviations of risk perceptions and help seeking intentions wh en the happy mood and no mood manipulation conditions were separate. Levenes test showed that variances were equivalent across groups for risk perceptions [ F (12, 256) = .62, p > .05], but not for help seeking intentions [ F (12, 256) = 1.91, p < .05]. The r esults largely confirmed the previous analyses of simple effects conducted after the happy mood and no mood manipulation conditions were combined. Further it was noticeable that no significant mean differences in risk perceptions and help seeking intentio ns were observed between participants under the happy mood and no mood manipulation conditions. The mean difference in risk perceptions was not significant whether diagnosticity was high [ t (256) = .32, p > .05, M happy mood&high diagnosticity = .10, SD = 2.65, M no mood manipulation&high diagnosticity = .13, SD = 2.04 ] or low [ t (256) = .12, p > .05, M happy mood&low diagnosticity = .02, SD = 2.12, M no mood manipulation&low diagnosticity = .08, SD = 2.64 ]. The mean difference in intentions was also insignif icant whether diagnosticity was high [ t (36.42) = .25, p > .05, M happy mood&high diagnosticity = 2.53, SD = .77, M no mood manipulation&high diagnosticity = 2.59, SD = .63 ] or low [ t (31.95) = .32, p > .05, M happy mood&low diagnosticity = 2.63, SD = .93, M no mood manipulation&low diagnosticity = 2.52, SD = 1.21 ].

PAGE 80

80 These results confirmed the expectation that participants under the happy mood and no mood manipulation conditions would not significantly differ in their risk perceptions and help seeking intentions because they did not differ in their mood states. Summary of the Test of H1a and H1b The MANOVA results supported H1a and H1b, as the main effects of mood state were significant [Wilks = .91, F (2, 119) = 5.59, p < .01, p 2 = .086], and neither perceived diagnosticity [Wilks = .99, F (2, 119) = .35, p > .05 p 2 = .006] nor the mood diagnosticity interaction [Wilks = .99, F (2, 119) = .37, p > .05, p 2 = 006] had significant effects on the linear combination of risk perceptions and help seeking inten tions. The mixed MANOVA results indicated that the effects of mood state were equivalent on risk perceptions and help seeking intentions [Wilks = .98, F (1, 120) = 2.51, p > .05 p 2 = .020]. Because MANOVA was a primary method of testing H1a and H1b, it was concluded that the two hypotheses were supported. Planned contrasts moderately supported H1a but failed to support H1b. A significant mean difference in risk perceptions was observed between sad mood and combined mood participants when diagnosticity w as high [ t (260) = 3.00, p < .01, M sad mood&high diagnosticity = 1.88, SD = 2.35, M combined mood&high diagnosticity = .03, SD = 2.03 ], supporting H1a. However, the difference was only marginally significant when diagnosticity was low [ t (260) = 1.69, p < .10 M sad mood&low diagnosticity = 1.15, SD = 2.34, M combined mood&low diagnosticity = .03, SD = 2.36 ], failing to support H1a. No significant mean differences were observed for help seeking intentions, failing to support H1b. The reason why H1a was only mode rately supported and H2a was not supported the opportunity manipulation was suboptimal. When opportunity was low, moods were hypothesized to directly affect judgments without undergoing a cognitive route because time pressure would prevent the

PAGE 81

81 cognitive pr ocess from interfering with the mood states direct effects on judgments. It was judged that the time pressure produced by requiring the subjects to report as fast as possible was not sufficient for the direct, informative effects of moods to be observed. Instead, a more powerful method of manipulating low opportunity would be forcing the low opportunity subjects to report within a short time frame. Testing Hypotheses H2a and H2b Predictions made in hypotheses 2a and 2b were confined to the experimental groups under the high opportunity condition. When opportunity was high, it was expected that exposure to low diagnosticity information would reduce risk perceptions (H2a) and help seeking intentions (H2b) significantly more among participants in the sad m ood condition than among those in the happy mood condition. H2a and H2b would be supported if the MANOVA results indicated that when opportunity was high, the mood state perceived diagnosticity interaction had significant effects on the linear combinati on of the perceived future risk of depression and help seeking intentions to discuss depression, and the interaction had equivalent effects on the two dependent variables. H2a and H2b would be further supported if planned contrasts revealed that exposure t o low diagnosticity information significantly reduced risk perceptions and help seeking intentions among sad mood participants, whereas the effects would be reduced to insignificance among happy mood participants. MANOVA Results for H2a and H2b To test H2a and H2b, risk perceptions and help seeking intentions for participants under the high opportunity condition were submitted to a 2 (mood state: sad mood versus combined mood) 2 (perceived diagnosticity: high versus low) MANOVA. The happy mood and no mood manipulation conditions were combined. Only experimental groups under the high

PAGE 82

82 opportunity condition were included in the analysis, because H2a and H2b predicted the effects of mood state and perceived diagnosticity when opportunity was high. The multiva riate results supported H2a and H2b. As Table 4 13 shows, the mood state perceived diagnosticity interaction was significant for the linear combination of risk perceptions and help seeking intentions [Wilks = .92, F (2,121) = 5.50, p < .01, p = .083]. In addition, mood state [Wilks = .87, F (2,121) = 9.26, p < .01, p = .133] and perceived diagnosticity [Wilks = .92, F (2,121) = 5.59, p < .01, p = .085] showed significant main effects. Variance covariance matrices were homogeneous across groups [Bo xs M = 10.94, F (9, 52927.27) =1.18, p > .05]. As Figure 4 3 and 4 4 show, the effects of perceived diagnosticity on risk perceptions and help seeking intentions were stronger among sad mood participants than among combined mood participants. The interacti on occurred in a pattern predicted in H2a and H2b. Exposure to low diagnosticity information significantly reduced risk perceptions (Figure 4 3) and help seeking intentions (Figure 4 4) among participants in the sad mood condition, whereas it did not have significant effects under the combined mood condition. Mixed MANOVA R esults For H2a and H2b To rule out the possibility that the mood state perceived diagnosticity interaction might have had significantly different effects on risk perceptions and help se eking intentions, a three way mixed MANOVA was conducted. The model entered mood (sad versus combined mood) and perceived diagnosticity (low versus high) as between subjects factors and treated risk perceptions and help seeking intentions as a within subje cts factor. To make the scores on the two dependent variables comparable, the single measures for risk perceptions and help seeking

PAGE 83

83 intentions were Z transformed. The analysis only included groups under the high opportunity condition. As Table 4 14 shows, the multivariate results revealed that the mood state perceived diagnosticity within subject factor interaction was not significant [Wilks = 1.00, F (1, 122) = .03, p > .05 p 2 = .000]. This suggested that the mood state perceived diagnosticity interaction equally affected risk perceptions and help seeking intentions. Further, the within subject factor did not moderate the effects of moo d state or diagnosticity. Therefore, the previous MANOVA results in support of H2a and H2b were valid. Boxs test revealed that the covariance matrices were equal across groups [ M = 10.95, F (9, 52927.27) = 1.18, p > .05]. Analyses of Simple Effects Planne d contrasts were used to test if the group mean differences in risk perceptions and help seeking intentions confirmed the MANOVA results in support of H1a and H2b. The happy mood and no mood manipulation conditions were combined. Risk perceptions. Table 4 10 summarizes the group means and standard deviations of risk perceptions when the happy mood and no mood manipulation conditions were combined. Levenes test showed that variances were equivalent across groups [ F (8, 260) = .68, p > .05]. W hen opportunity was high and the induced mood state was sadness, participants who received the high diagnosticity information reported higher mean risk perception than those who received the low opportunity information [ t (260) = 3.03, p < .01, M sad mood&high diagnosticity = 1.72, SD = 2.46, M sad mood&low diagnosticity = .48, SD = 2.77 ]. This supported the prediction made in H2a that exposure to the low diagnosticity information would significantly reduce risk perceptions among participants in the sad mood condition.

PAGE 84

84 In c ontrast, the mean difference in risk perception between the high and low diagnosticity conditions was insignificant among participants in the combined mood condition [ t (260) = .45, p > .05, M combined mood&high diagnosticity = 1.17, SD = 2.26, M combined m ood&low diagnosticity = .93, SD = 2.16 ]. This supported the prediction made in H2a that exposure to the low diagnosticity information would not significantly reduce risk perceptions among happy mood participants. As a result, the mean difference in risk p erception between sad mood participants and combined mood participants was significant when perceived diagnosticity was high [ t (260) = 4.76, p < .01, M sad mood&high diagnosticity = 1.72, SD = 2.46, M combined mood&high diagnosticity = 1.17, SD = 2.26 ], but insignificant when diagnosticity was low [ t (260) = .69, p > .05, M sad mood&low diagnosticity = .48, SD = 2.77, M combined mood&low diagnosticity = .93, SD = 2.16 ]. Help seeking intentions. Similar results were observed for help seeking intentions (Table 4 10). When opportunity was high and sadness was induced, participants who received the high diagnosticity information reported stronger help seeking intentions than those assigned to the low diagnosticity information [ t (35.63) = 3.46, p < .01, M sad mood&h igh diagnosticity = 3.35, SD = .85, M sad mood&low diagnosticity = 2.39, SD = .97 ]. The mean difference between the high and low diagnosticity conditions became insignificant among participants in the combined mood condition [ t (76.34) = .40, p > .05, M combi ned -mood&high diagnosticity = 2.38, SD = .98, M combined mood&low diagnosticity = 2.30, SD = .87 ]. These results supported the prediction that the effects of perceived diagnosticity on help seeking intentions would be stronger among participants in the sa d mood condition than among those in the combined mood condition. As a result, the mean difference in intentions between participants in the sad mood condition and the combined mood condition was significant when perceived diagnosticity was high [ t (58.73) = 4.2, p < .01, M sad mood&high diagnosticity = 3.35, SD = .85

PAGE 85

85 M combined mood&high diagnosticity = 2.38, SD = .98 ], but insignificant when diagnosticity was low [ t (31.75) = .32, p > .05, M sad mood&low diagnosticity = 2.39, SD = .97, M combined mood&low diagn osticity = 2.30, SD = .87 ]. This further supported H2b. Therefore, planned contrasts strongly supported H2a and H2b. Simple effects when the happy mood and no mood manipulation conditions were not combined. Table 4 11 and 4 12 summarize the group means a nd standard deviations of risk perceptions and help seeking intentions when the happy mood and no mood manipulation conditions were separate. Levenes test showed that variances were equivalent across groups for risk perception [F(12, 256) = .50, p > .05] and help seeking intentions [F(12, 256) = 1.54, p > .05] The mean difference in risk perceptions between the happy mood and no mood manipulation participants was not significant whether diagnosticity was high [ t (256) = 1.15, p > .05, M happy mood&high dia gnosticity = .74, SD = 2.45, M no mood manipulation&high diagnosticity = 1.62, SD = 2.00] or low [ t (256) = .30, p > .05, M happy mood&low diagnosticity = 1.04, SD = 2.20, M no mood manipulation&low diagnosticity = .82, SD = 2.18]. Similarly, the happy mo od and no mood manipulation conditions did not lead to a significant mean difference in help seeking intentions whether diagnosticity was high [ t (36.05) = .32, p > .05, M happy mood&high diagnosticity = 2.43, SD = 1.09, M no mood manipulation&high diagnostic ity = 2.33, SD = .87] or low [t(39.74) = .23, p > .05, M happy mood&low diagnosticity = 2.27, SD = .85, M no mood manipulation&low diagnosticity = 2.33, SD = .92]. The results supported the expectation that the two mood conditions would not lead to signific ant mean differences in risk perceptions and help seeking intentions. Summary of the Tests of H2a and H2b The MANOVA results supported H2a and H2b, because the mood state perceived diagnosticity interaction was significant for the linear combination of risk perceptions and help

PAGE 86

86 seeking intentions [Wilks = .92, F (2,121) = 5.50, p < .01, p = .083] The effects of diagnosticity on risk perceptions and help seeking intentions were stronger under the sad mood condition than under the combined mood condition (Figure 4 3 and 4 4 ). The result of mixed MANOVA revealed that the mood diagnosticity interaction equally affected risk perceptions and help seeking intentions [Wilks = 1. 00 F (1, 122) = 03 p > 05 p = 000]. Analyses of simple effects also revealed that when the induced mood was sadness, expos ure to the low diagnosticity information significantly reduced risk perceptions [ t ( 260 ) = 3. 03 p < 01, M sad mood&high diagnosticity = 1. 72, SD = 2.46, M sad mood&low diagnosticity = .48, SD = 2.77] and help seeking intentions [ t ( 35.63 ) = 3. 46 p < 01, M sad mood&high diagnosticity = 3. 35, SD = .85, M sad mood&low diagnosticity = 2.39, SD = .97 ] In contrast, under the combined mood condition, exposure to the low diagnosticity information did not significantly lower risk perceptions [ t ( 260 ) = 45 p > 05, M combined -mood&high diagnosticity = 1. 17, SD = 2.26, M combined mood&low diagnosticity = .93, SD = 2.16] or help seeking intentions [ t ( 76.34 ) = 40 p > 05, M combined mood&high diagnosticity = 2. 38, SD = .98, M combined mood&low diagnosticity = 2.30, SD = .87 ] .. The results of MANOVA and analyses of simple effects strongly supported H2a and H2b. Testing Hypotheses 3a and 3b H3a and H3b predicted that risk perceptions would mediate the effects of mood on help seeking intention s When opportunity was low, H3a predicted that risk perceptions would mediate the main effects of mood state on help seeking intentions. When opportunity was high, H3b predicted that risk perception s would mediate the effects of the mood diagnosticity interaction on help seeking in tention s Baron and Kenny (1986) suggested that to conclude significant mediation effects occurred, one needs to confirm the following four requirements: requirement (a), the predictor is

PAGE 87

87 significantly related to the outcome; requirement (b), the predictor is significantly related to the mediator; requirement (c), the mediator is significantly related to the outcome after the predictor is controlled; and requirement (d), after the mediator is controlled, the predictor should be significantly less related to the outcome than before the mediator is controlled. Test of H3a Hypothesis H3a predicted that when opportunity was low, the perceived future risk of depression would mediate the effects of mood state on help seeking intentions. The hypothesis conceptuali zed manipulated mood as the predictor, help seeking intentions as the outcome, and the perceived future risk of depression as the mediator. To test H3a, three regression models were built, as suggested by Baron and Kenny (1986). Similar to the multivariate tests of H1a and H1b, the regression analyses included participants under the low opportunity condition, because H3a predicted the mediating role of risk perceptions under the low opportunity condition. Also similar to the previous analyses, the happy moo d and no mood manipulation conditions were combined. Figure 4 5 summarizes the results of the four step mediation analysis. In the first regression model, help seeking intentions were regressed on mood state. The coefficient for mood state was negative, su ggesting that participants in the combined mood condition tended to report lower help seeking intentions. However, the relationship was not statistically significant ( B = .29, p > .05). Therefore, Requirement (a) was not satisfied. In the second regressio n model, perceived future risk of depression was regressed on mood state. The coefficient for mood state was significant ( B = 1.51, p < .01), revealing that sad moods resulted in higher risk perceptions. Therefore, Requirement (b) was satisfied. In the t hird regression model, help seeking intentions were regressed on risk perceptions. Mood state was entered as a control variable. Risk perceptions were significantly related with

PAGE 88

88 intentions ( B = .18, p < .01). Therefore, Requirement (c) was satisfied. To te st Requirement (d), the coefficient for mood state in the first regression model ( B = .29, p > .05) was compared with the coefficient in the third model ( B = .02, p > .05). The comparison revealed that the size of mood states relationship with help seek ing intentions dropped as risk perception was controlled. Sobels (1982) test, a method of testing the indirect effects of an independent variable on the dependent variable, showed that the drop was statistically significant ( Z = 2.90, p < .01). Although R equirements (b), (c), and (d) were satisfied, the rejection of Requirement (a) suggested there were no significant effects of mood state on help seeking intentions to be mediated. Therefore, the four step analysis failed to support H3a. Test of H3b Figure 4 6 summarizes the results of the four step mediation analysis conducted to test H3b. Hypothesis 3b conceptualized the mood state perceived diagnosticity interaction as the predictor, help seeking intentions as the outcome variable, and the perceived fu ture risk of depression as the mediator. To test H3a, three regression models were built. Similar to the multivariate and univariate tests of H2a and H2b, only responses made under the high opportunity condition were included in the analyses. In the first regression model, help seeking intentions were regressed on mood, diagnosticity, and the mood diagnosticity interaction. The interaction was significantly correlated with intentions ( B = .88, p < .05), satisfying Requirement (a). In the second regressi on model, risk perceptions were regressed on mood, diagnosticity, and the mood diagnosticity interaction. The coefficient for the interaction term was significant ( B = 2.43, p < .01), satisfying Requirement (b). In the third model, help seeking intenti ons were regressed on risk perceptions, entering mood state, perceived diagnosticity, and the mood state perceived diagnosticity interaction as

PAGE 89

89 control variables. The coefficient for risk perception was significant ( B = .11, p < .01). Therefore, Requirem ent (c) was satisfied. To test Requirement (d), the coefficient for the mood diagnosticity interaction in the first regression model ( B = .88, p < .05) was compared with the coefficient in the third model ( B = .62, p > .05). The comparison showed that the interaction effects on help seeking intentions were reduced as risk perceptions were controlled. Sobels test showed that the reduction was statistically significant ( Z = 16.49, p < .01). The four requirements for establishing mediation effects were a ll confirmed. Therefore, H3b was supported. In addition, because the predictors (e.g., mood state) relationship with the outcome (e.g., help seeking intentions) became statistically insignificant after controlling the mediator (e.g., risk perceptions), it was concluded that a complete, rather than partial, mediation occurred. Additional Data Analyses Hypotheses 1a through 2b implied that the interaction of mood state and perceived diagnosticity would significantly affect perceived future risk of depress ion and help seeking intentions to discuss depression when opportunity was high (H2a and H2b), whereas the interaction effects would be reduced and become insignificant when opportunity was low. The four hypotheses were supported by the MANOVA results, and received moderate support from analyses of simple effects. These results indicated that mood, diagnosticity, and opportunity would have three way interaction effects on risk perceptions and help seeking intentions. MANOVA results. To test the three way in teraction, risk perceptions and help seeking intentions were submitted to a 2 (mood state: sad versus combined mood) 2 (diagnosticity: high versus low) 2 (opportunity: high versus low) MANOVA. Again, the happy mood and no mood manipulation conditions w ere combined, as the two mood conditions did not lead to significantly different mood states. This analysis included all experimental groups, except for the

PAGE 90

90 non factorial control group that did not receive any manipulation. The assumption of homogeneous va riance covariance matrices was satisfied [Boxs M = 30.66, F (21, 89626.61) = 1.42, p > .05]. Table 4 15 summarizes the multivariate results. The three way interaction did not have significant effects on the linear combination of risk perceptions and help seeking intentions [Wilks = .99, F (2, 241) = 1.31, p > .05, p = .011] This implied that even though the effects of the mood state perceived diagnosticity were statistically significant when opportunity was high and insignificant when opportunity was low, the reduction of the interaction effects was not statistically significant. ANOVA results. Follow up ANOVAs were conducted to check if the three way interaction of mood, diagnosticity, and opportunity significantly affected risk perceptions and help seeking intentions respec tively. Levenes test revealed that variances were homogeneous for risk perceptions [ F (7, 242) = .27, p > .05] and help seeking intentions [ F (7, 242) = 1.61, p > .05] Table 4 16 summarizes the ANOVA results. The univariate results revealed that the three way interaction did not have significant effects on risk perceptions [ F (1, 242) = 1.85, p > .05, p = .008] or help seeking intentions [ F ( 1 242 ) = 1. 73 p > 05, p = .007] The ANOVA results confirmed that multivariate results, suggesting that opportu nity did not significantly determine the effects of the mood state perceived diagnosticity interaction. Comparison with the non factorial group. Analyses of simple effects revealed that exposure to an antidepressant ad could lead to higher risk perceptio ns. For example, when opportunity was high, the group that received a sadness inducing procedure and the high diagnosticity information reported higher risk perceptions than the non factorial control group [ t ( 256 ) = 3.64 p < 01 M sad mood&high diagnostic ity&high opportunity = 1. 72, SD = 2.46 M control = .93 SD

PAGE 91

91 = 3. 00] defined as the participants who received no manipulations but completed dependent measures. In contrast, the mean risk perception of sad mood participants who received the low diagnostici ty information were not significantly higher than that of the non factorial control group [ t ( 256 ) = .58 p > 05 M sad mood&low diagnosticity&high opportunity = 48, SD = 2.77 M control = .93 SD = 3. 00] When opportunity was low, sad mood participants reported higher risk perceptions than the non factorial control group, whether diagnosticity was high [ t ( 256 ) = 3.72 p < 01 M sad mood&high diagnosticity&low opportunity = 1. 89, SD = 2.35 M control = .93 SD = 3. 00] or low [ t ( 256 ) = 2.70 p < 01 M sad mood&low diagnosticity&low opportunity = 1. 15, SD = 2.35 M control = .93 SD = 3. 00]. Therefore, exposure to DTC advertising may result in the market expansion of a drug class by presenting information on the symptoms of a disease without a guideline abou t how such information should be interpreted. Given the powerful effects of sad moods in increasing risk perception, one may argue that the market expansion of a drug category could also occur because drug advertising campaigns effectively put consumers in to negative mood states and induce high risk perceptions of diseases. Comparison of males and females. It was noticeable that male and female participants did not differ in their risk perceptions [ F ( 1,267 ) = .73 p > .05, M male = 20 SD = 2. 71 M female = .10 SD = 2.52 ] and help seeking intentions [ F ( 1,267 ) = .46 p > .05, M male = 2. 56 SD = 1. 00 M female = 2.65 SD = 1.01 ] replicating the findings of Park and Grow (2008). Epidemiological studies show that the lifetime risk of depression is approximately 13 percent for men and 20 to 25 percent for women (Kessler et al., 1993, 1994; NCS, 2003). Therefore, it was possible that the male participants overrated their future risk of depression or the female participants underrated their risk.

PAGE 92

92 Summary of the Res ults This chapter reported the results of the study described in Chapter 3. The study was designed to explore how consumer mood state and perceived diagnosticity, defined as the degree that consumers perceived a list of discomforting life events to be indi cative of clinical depression, affect the perceived future risk of depression and intentions to seek professional help to discuss depression. Opportunity, defined as how much constraint consumers had in estimating risk perceptions and help seeking intentio ns, was conceptualized as a factor that determines how mood state and diagnosticity affect the two dependent variables. In particular, when consumers had low opportunity for risk and intention estimation, it was predicted that sad moods would lead to high er risk perceptions (H1a) and stronger help seeking intentions (H1b), whether perceived diagnosticity was low or high. The MANOVA results supported H1a and H1b, showing when opportunity was low, mood state was the only factor that significantly affected ri sk perceptions and intentions. However, the two hypotheses received weak support from planned contrasts, because, when opportunity and diagnosticity were both low, sad moods did not lead to higher risk perceptions. Further, when opportunity was low, sad mo ods did not result in stronger help seeking intentions than happy moods. When consumers had high opportunity for risk and intention estimation, it was hypothesized that exposure to the diagnosticity reducing information would reduce risk perceptions (H2a) and help seeking intentions (H2b) significantly more among sad mood participants than among happy mood participants. The MANOVA results strongly supported H2a and H2b. The two hypotheses also received strong support from analyses of simple effects. When op portunity was high, exposure to the low diagnosticity information significantly reduced risk perceptions and help seeking intentions among sadness induced participants, whereas the

PAGE 93

93 exposure did not have significant effects among those in the happy mood and no mood manipulation conditions. Further, risk perceptions were conceptualized as a mediator of the effects of mood state and perceived diagnosticity on help seeking intentions. Therefore, when opportunity was low, risk perceptions were hypothesized to m ediate the main effects of mood state on help seeking intentions (H3a). Application of Baron and Kenny (1986) s four step analysis of mediation failed to support this hypothesis, because mood state did not significantly affect help seeking intentions and t herefore there were no main effects to be mediated. When high opportunity was given, risk perceptions were hypothesized to mediate the effects of the mood diagnosticity interaction on help seeking intentions (H3b). This hypothesis was supported. Risk pe rceptions completely mediated the effects of the mood diagnosticity interaction on intentions, because, after risk perceptions were controlled, the interaction was no longer significantly related with help seeking intention.

PAGE 94

94 Table 4 1. One way ANOVA results for the mood manipulation check Sad mood Happy mood No manipulation F df1 df2 p M SD M SD M SD (a) G loomy j oyful 3.25 1.41 4.98 1.22 4.89 1.17 49.45 2 247 .00 (b) S ad h appy 3.28 1.58 5.09 1.20 5.01 1.29 47.46 2 247 .00 (c) U pse t e lated 3.32 1.27 4.81 1.11 4.61 1.03 42.26 2 247 .00 A verage ( = .95) 3.28 1.35 4.96 1.01 4.84 1.08 52.62 2 247 .00 For each item, a 7 point scale was used (1 = gloomy, sad, upset, 7 = joyful, happy, elated). Table 4 2. Full factorial ANOVA results for the mood manipulation check Independent variable F df1 df2 p p Mood 101.16 1 242 .00 .295 Diagnosticity .55 1 242 .46 .002 Opportunity .55 1 242 .46 .002 Mood diagnosticity .00 1 242 .96 .000 Mood opportunity .00 1 242 .99 .000 Diagnosticity opportunity .96 1 242 .33 .004 Mood diagnosticity oppor tunity .28 1 242 .60 .001 Table 4 3. One way ANOVA results for the diagnosticity manipulation check Low diagnosticity High diagnosticity F df1 df2 p M SD M SD L ow energy 2.77 1.75 2.97 1.66 .85 1 248 .40 D epressed 2.58 1.57 3.16 1.74 7.47 1 248 .01 S leep problems 2.91 1.93 3.59 2.06 7.31 1 248 .01 D ifficulty making D ecisions 2.88 1.72 3.53 1.90 7.98 1 248 .01 A verage ( = .88) 2.79 1.52 3.31 1.55 1 248 .01 For each item, a 7 point scale was used to report the likelihood that the symptom indicated clinical depression (1 = not at likely, 7 = very likely). Table 4 4. Full factorial ANOVA results for the diagnosticity manipulation check Independent variable F df1 df2 p p Mood .14 1 242 .81 .000 Diagnosticity 16.35 1 242 .01 .028 Opportunity 2.28 1 242 .33 .004 Mood diagnosticity .10 1 242 .84 .000 Mood opportunity .27 1 242 .74 .000 Diagnosticity opportun ity 1.02 1 242 .51 .002 Mood diagnosticity opportunity 1.22 1 242 .48 .002

PAGE 95

95 Table 4 5. One way ANOVA results for the opportunity manipulation check Low opportunity High opportunity F df1 df2 p M SD M SD Time: risk 1 14.75 4.55 22.73 10.77 57. 84 1 248 .000 Time: risk 2 9.86 4.28 12.82 4.82 26.24 1 248 .000 Time: risk 3 8.28 3.61 10.65 5.25 17.23 1 248 .000 Time: intention 1 9.50 3.54 10.97 3.80 9.89 1 248 .002 Time: intention 2 7.35 2.61 8.44 4.27 5.82 1 248 .017 Time: intention 3 7.43 3.5 5 7.41 4.22 .001 1 248 .972 Table 4 6. Correlations among variables Age Gender Vicarious experience Risk perception Help seeking Intention Age r p 1 Gender r p .26 .00 1 Vicarious experience r p .04 .54 .08 .22 1 Risk perception r p .05 42 .05 .39 .23 .00 1 Help seeking intention r p .04 .53 .04 .50 .23 .00 .40 .00 1 Table 4 7. MANOVA results for risk perception and help seeking intention when opportunity was low Independent variable Wilks Lambda F H df Error df p p Mood .91 5.59 2 119 .01 .086 Diagnosticity .99 .35 2 119 .71 .006 Mood d iagnosticity .99 .37 2 119 .69 .006 Table 4 8. Mixed MANOVA results for risk perception and help seeking intention when opportunity was low Effect Wilks Lambda F H df Error df p p ws .96 5.25 1 120 .02 .042 ws m ood .98 2.51 1 120 .12 .020 ws d iagnosticity 1.00 .06 1 120 .81 .000 ws m ood d iagnosticity 1.00 .03 1 120 .86 .000 ws is a within subjects factor including risk perception and help seeking intention.

PAGE 96

96 T able 4 9. Group means and standard deviations for risk perception when happy mood and no mood manipulation conditions were combined Sad Combined m ood Low diagnosticy High diagnosticity Low diagnosticity High diagnosticity Low opportunity n = 20 M = 1.15 SD = 2.34 n = 22 M = 1.88 SD = 2.35 n = 37 M = .03 SD = 2.36 n = 45 M = .03 SD = 2.30 High opportunity n = 19 M = .48 SD = 2.77 n = 26 M = 1.72 SD = 2.46 n = 42 M = .93 SD = 2.16 n = 39 M = 1.17 SD = 2.26 Table 4 10. Group means and standard deviations for help seeking intention when happy mood and no mood manipulation conditions were combined Sad Combined mood Low diagnostic ity High diagnosticy Low diagnosticity High diagnosticity Low opportunity n = 20 M = 2.75 SD = .94 n = 22 M = 2.95 SD = 1.24 n = 37 M = 2.58 SD = 1.06 n = 45 M = 2.56 SD = .68 High opportunity n = 19 M = 2.39 SD = .97 n = 26 M = 3.35 SD = .85 n = 42 M = 2.30 SD = .87 n = 39 M = 2.38 SD = .98 Table 4 11. Group means and standard deviations for risk perception when happy mood and no moo d manipulation conditions were not combined Sad Happy No manipulation Low diagnos ti city High diagnos ticity Low diagno s ticity High diagnos ticity Low diagnos ticity High diagnos ticity Low opportunity n = 20 M = 1.15 SD = 2.34 n = 22 M = 1.88 SD = 2.35 n = 19 M = .02 SD = 2.12 n = 20 M = .10 SD = 2.65 n = 18 M = .08 SD = 2.64 n = 25 M = .13 SD = 2.04 High opportunity n = 19 M = .48 SD = 2.77 n = 26 M = 1.72 SD = 2.46 n = 21 M = 1.0 4 SD = 2.20 n = 20 M = .74 SD = 2.45 n = 21 M = .82 SD = 2.18 n = 19 M = 1.62 SD = 2.00

PAGE 97

97 Table 4 12. Group means and standard deviations for help seeking intention when happy mood and no mood manipulation conditions were not combined Table 4 13. MANOVA results for risk perception and help seeking intent ion when opportunity was high Independent variable Wilks Lambda F H df Error df p p Mood .87 9.26 2 121 .000 .133 Diagnosticity .92 5.59 2 121 .000 .085 Mood Diagnosticity .92 5.50 2 121 .005 .083 Table 4 14. Mixed MANOVA results for risk percepti on and help seeking intention when opportunity was high Effect Wilks Lambda F H df Error df p p ws 1.00 .37 1 122 .55 .003 ws Mood 1.00 .39 1 122 .54 .003 ws Diagnosticity 1.00 .47 1 122 .50 .004 ws Mood Diagnosticity 1.00 .03 1 122 .86 .000 ws is a within subjects factor including risk perception and help seeking intention. Table 4 15. MANOVA results for risk perception and help seeking intention: All cases included Independent variable Wilks Lambda F H df Error df p p Mood .90 13.82 2 241 .00 .103 Diagnosticity .97 3.92 2 241 .02 .032 Opportunity .96 5.00 2 241 .01 .040 Mood d iagnosticity .97 4.12 2 241 .02 .033 Mood o pportunity 1.00 .46 2 241 .63 .004 Diagnosticity o pportunity .99 1.52 2 241 .22 .012 Mood d iagnosticity o pportunity .99 1.31 2 241 .27 .011 Sad Happy No Manipulation Low DIAG High DIAG Low DIAG High DIAG Low DIAG High DIAG Low Opportunity n = 20 M = 2.75 SD = .94 n = 22 M = 2.95 SD = 1.24 n = 19 M = 2.63 SD = .93 n = 20 M = 2.53 SD = .77 n = 18 M = 2.52 SD = 1.21 n = 25 M = 2.59 SD = .63 High Opportunity n = 19 M = 2.39 SD = .97 n = 26 M = 3.34 SD = .85 n = 21 M = 2.27 SD = .85 n = 20 M = 2.43 SD = 1.09 n = 21 M = 2 .33 SD = .92 n = 19 M = 2.33 SD = .87

PAGE 98

98 Table 4 16. ANOVA r esults for risk perception and help seeking intention: All cases included Independent variable Dependent variable F df1 df2 p p Mood Risk perception 25.48 1 242 .00 .095 Help seeking intention 10.34 1 242 .00 .041 Diagnosticity Risk perception 4.56 1 242 .03 .018 Help seeking intention 6.07 1 242 .01 .024 Opportunity Risk perception 9.94 1 242 .00 .039 Help seeking int ention .72 1 242 .40 .003 Mood d iagnosticity Risk perception 6.39 1 242 .01 .026 Help seeking intention 4.78 1 242 .03 .019 Mood o pportunity Risk perception .08 1 242 .77 .000 Help seeking intention .92 1 242 .34 .004 Diagnosticity o pportunity Risk perception .96 1 242 .33 .004 Help seeking intention 2.90 1 242 .09 .012 Mood d iagnosticity o pportunity Risk perception 1.85 1 242 .18 .008 Help seeking intention 1.73 1 242 .19 .007

PAGE 99

99 Figure 4 1. O bserved ANOVA results for risk perception when opportunity was low Figure 4 2. Observed ANOVA results for help seeking intention when opportunity was low Low Diagnosticity High Diagnosticity Risk Perception Combined Mood Sad Mood 1.88 1.15 .03 .03 Lo w Diagnosticity High Diagnosticity Risk Perception Combined Mood Sad Mood 2.95 2.75 2.56 2.58

PAGE 100

100 Figure 4 3. Observe d ANOVA results for risk perception when opportunity was high Figure 4 4. Observed ANOVA results for help seeking intention when opportunity was high Low Diagnosticity High Diagnosticity Risk Perception Combined Mood Sad Mood 1.72 .48 1.17 .93 Low Diagnosticity High Diagnosticity Risk Perception Combined Mood Sad Mood 3.35 2.39 2. 38 2.30

PAGE 101

101 Figure 4 5. Mediation ana lysis when opportunity was low Note: *Coefficient when risk perception was not controlled **Coefficient when risk perception was controlled Figure 4 6. Mediation analysis when opportunity was high Note: *Coefficient when risk per ception was not controlled **Coefficient when risk perception was controlled Mood state Risk perception Help seeking intention *B = .29 *p > .05 B = 1.51 p < .01 B = .18 p < .01 **B = .02 **p > .05 Mood state Diagnosticity Risk perception Help seeking intention *B = .88 *p < .05 B = 2.43 p < .01 B = .11 p < .01 **B = .62 **p > .05

PAGE 102

102 CHAPTER 5 DISCUSSION This chapter presents the theoretical and practical implications of the results of the study described in Chapter 4. First, a summary of the research fi ndings is presented. Then the chapter proceeds to the findings implications for advertising theory and practice and consumer health. The chapter will conclude by presenting the limitations of the study and suggestions for future research. Summary of Findi ngs While consumers perceived future risk of a disease and intentions to engagement in preventive and remedial actions are multiply determined, this study focused on how consumers process health information from a direct to consumer prescription drug adve rtisement and form perception s of the future risk of depression and intention s to seek profession al help to discuss depression. In particular, because moods influence consumers subjective experiences with negative health conditions ( Croyle & Uretsky, 1987 ; Pettit, Kline, Gencoz & Gencoz, 2001; Salovey & Birnbaum, 1989) this study explored the possibility that consumers current moods might influence the way they process information from a drug advertisement and further determine their perceived future ris k of a disease. H1a and H1b Hypotheses 1a and 1b predicted that when opportunity was low, sad mood participants would report higher risk perceptions (H1a) and stronger help seeking intentions (H1b) than happy mood participants, regardless of the level of perceived diagnosticity. The MANOVA results supported H1a and H1b. When opportunity was low, mood state had significant main effects on risk perceptions and help seeking intentions. Neither perceived diagnosticity nor the mood state perceived diagnostici ty interaction was significant. The mixed MANOVA results

PAGE 103

103 indicated that mood state equally affected the two dependent variables. Analyses of simple effects supported H1a but failed to support H1b. H2a and H2b Hypotheses 2a and 2b posited that when opportun ity was high, consumer mood state and perceived diagnosticity would have interactive effects on risk perceptions (H2a) and help seeking intentions (H2b). Specifically, exposure to the low diagnosticity information would reduce risk perceptions and help see king intentions significantly more when the induced mood was sadness, than when happiness was induced. The multivariate results supported these two hypotheses by showing that when opportunity was high, the mood diagnosticity interaction significantly aff ected the linear combination of risk perceptions and help seeking intentions. Simple effects tests showed that risk perceptions and help seeking intentions were lower in the low versus high diagnosticity conditions only for sad mood participants. Diagnost icity was not a predictor of risk perceptions or intentions for happy mood participants. Therefore, H2a and H2b were strongly supported. H3a and H3b H3a predicted that when opportunity was low, consumer mood state s effects on help seeking intentions woul d be mediated by risk perception s Baron and Kenny s (1986) four step mediation analysis was conducted, entering mood state as the predictor, risk perceptions as the mediator, and help seeking intentions as the outcome variable. The results failed to supp ort H3a. Mood state, the hypothesized predictor, was significantly related with risk perceptions, the hypothesized mediator. Risk perceptions significantly predicted help seeking intentions, the hypothesized outcome. Further, when intentions, the hypothesi zed outcome variable, were regressed on mood state, the coefficient for mood state significantly dropped when risk perception was controlled, compared to when risk perceptions were not

PAGE 104

104 controlled. Although these requirements for mediation were satisfied, m ood state was not significantly associated with intentions, suggesting there were no significant effects of the predictor to be mediated in the first place. H3b posited that when opportunity was high, the interactive effects of mood state and perceived dia gnosticity on help seeking intentions would be mediated by risk perceptions. Baron and Kenny s (1986) four step analysis was conducted, entering the mood diagnosticity interaction as the predictor, risk perceptions as the mediator, and help seeking inten tions as the outcome variable. The results strongly supported H3b, because the four requirements of mediation effects were all satisfied. In addition, risk perceptions completely mediated the mood diagnosticity interaction s effects on help seeking inten tions, because the interaction was no longer significantly related with intentions when risk perceptions were controlled. Discussion of Findings Advertising T heory Despite a large body of research, the current literature on DTC advertising generally focuse s on drug advertising s impact on the way consumers seek health information and interact with doctors to get specific drugs, focusing on variables such as consumer awareness and attitudes regarding DTC advertising, visits to doctors offices, and requests for specific drugs. E x ploring the role of advertising in constructing consumer perceptions of diseases is important, because consumer decisions about health behavior are based on how health issues are perceived Despite its importance, t his research area r emains largely unexplored in the advertising literature. This study was designed to contribute to the DTC advertising literature by illuminating the social cognitive effects of DTC advertising. The study revealed that exposure to drug advertising could in fluence consumer perceptions of the future risk of a disease, and the effects were stronger when consumers were in sad moods and the opportunity for risk estimation was limited.

PAGE 105

105 This indicated that DTC advertising might influence the way consumers perceive the social reality of diseases. Further, although much research exists on the effects of DTC advertising, the use of survey methodologies has limited scholars efforts to make causal inferences. For example, Bell et al. (1999) reported that exposure to DTC advertising was associated with visits to doctors offices and consumer insistence on the prescription of specific drugs, but they could not conclude DTC advertising caused consumers to seek professional help. In addition, the relationships reported in the literature are often not based on the application of theories, suggesting the observed relationships might have been spurious or occurred by chance. The current study overcame these shortcomings by producing psychological accounts for why exposure to DTC advertising might lead consumers to visit doctors offices and under what circumstances such associations would likely be observed. Research on the effects of DTC advertising should make use of theories to better organize and expand knowledge, and sti mulate and guide further research (Infante & Wormack, 1993). In addition, the extant literature on DTC advertising often does not reveal how consumers actually process information, and how content elements may account for the cognitive and attitudinal effe cts of a drug advertising campaign. Due to this neglect, the literature does not produce insights for drug advertising practitioners. The current study addressed this limitation. For example, the study revealed that excluding the guideline about how inform ation about the symptoms of a disease should be interpreted might result in raising consumer awareness of the future risk of a disease. The effects were strengthened when consumers were in sad moods. This may inform creative strategists and media planners that, if increasing awareness of the disease s risk is the primary objective of a drug advertising campaign, diagnosticity information needs to

PAGE 106

106 be deleted and creative elements and media planning strategy should be utilized to induce sad moods. Last, this study incorporated the concept of consumer mood state into the DTC drug advertising literature. This is an important contribution, because experiencing a disease and thinking about treatment options likely entail negative affective states. Processing info rmation in a negative affective state is expected to be typical, rather than atypical, of consumers who seek information on depression, cancer, attention disorder, sexually transmitted diseases, and many other health issues. The concept of mood state, ther efore, should not be left out in the DTC literature. However, the construct has not received much attention. This study revealed that consumers in sad moods processed information and formed risk perceptions and help seeking intentions differently from the way happy consumers did. Future researchers are strongly encouraged to apply and extend this perspective and find out how other mood states, such as anxiety and anger, may affect consumer perceptions of health issues. Advertising P ractice The findings of this study also have significant implications for building effective strategy for a drug advertising campaign Especially, the findings add insights into whether a drug advertising campaign should focus on category or brand expansion. The findings also hav e implications for building effective creative and media strategy. This section begins with the findings implications for category versus brand expansion. Category v ersus b rand e xpansion. Putsis and Dhar (2001) suggested brand promotions might result in e xpanding a new product category in addition to leading consumers to switch brands This perspective appears to receive empirical support, because research based on industry data indicated DTC advertising might lead to the market expansion of a drug catego ry, characterized by increasing numbers of consumer visits to doctor s offices to discuss the disease

PAGE 107

107 the drug category is designed to treat, doctors diagnoses of the disease, and prescriptions written for the drug category (Donohue & Berndt, 2004 ; Zachry et. al., 2002). The findings of this study present a psychological explanation for why DTC advertising could cause the market expansion of a drug class. In particular, this study revealed that exposure to the high diagnosticity information in a DTC drug advertisement, such as an antidepressant advertisement that does not include the APA guideline about when discomforting life experiences might or might not indicate clinical depression, tended to increase consumer perceptions of the future risk of depress ion. This tendency was observable especially among participants who had sufficient opportunity for risk estimation and were in sad, rather than happy, moods. The results in support of H3b further suggested that once formed, high risk perceptions could lead to stronger intentions to seek professional help to discuss depression, a likely antecedent of the market expansion of the antidepressant drug category. This tendency was observed when participants had high opportunity for risk estimation. The results in support of H3b suggested that when a drug ad carries high diagnosticity information and targets consumers who are undergoing discomforting life events or sad moods, it could lead to the market expansion of a drug class by increasing the perceived risk of t he problem. This interpretation receives empirical support from prior research. Research on health behavior shows consumers risk assessment of a health problem may produce attitudinal and behavioral changes, such as engaging in preventive and remedial be haviors (Block & Keller, 1998; Raghubir & Menon 1998), including consultation with doctors. Therefore, changes in risk perception triggered by DTC advertising may drive consumers to visit doctors offices to discuss depression and/or request antidepressant treatment.

PAGE 108

108 However, the current study suggests this interpretation should be appreciated with caution, because the effects of mood state on help seeking intentions and the mediating role of risk perceptions were not established when opportunity was low, rejecting H3a. Caution is further needed, because simple effects tests revealed that help seeking intentions of the non factorial control group were not lower than those of the sad mood participants who received the high diagnosticity information, whether opportunity was high or low. This study revealed that DTC advertising might lead to the market expansion of a drug class. Then the key decision for an advertising strategist to make will be, between category expansion and brand expansion, which should be p rioritized. Arens, Weigold, & Arens (2008) suggested that when a product category is introduced to the market, a limited number of consumers know about the product and its benefits, and therefore it is more important to trigger a primary demand, defined as consumer demand for the product category. The relative emphasis may shift to selective demand, defined as demand for a specific brand, as the product enters the growth or maturity stage. This perspective suggests that when launching a campaign for a drug that treats a relatively unfamiliar disease, pharmaceutical companies may consider using marketing tools to make consumers more aware of the symptoms and risk of the relevant disease, because this may lead to the market expansion of the drug category. Thi s may be even more important than emphasizing the drug s competitive advantages c onsidering that a limited market size is one of the major reasons that the introduction of a new product fails. As consumers become more familiar with the disease, more empha sis may be placed on points of difference between brands in a drug category.

PAGE 109

109 This cycle appears to characterize DTC advertising campaigns for antidepressants. Campaigns for early market leaders, such as Prozac and Zoloft, convey ed the message that clinica l depression is common while campaigns for brands launched at a later stage emphasized points of difference For example, Paxil CR was positioned as a social anxiety disorder treatment while Wellbutrin XL was positioned as the first antidepressant withou t sexual side effects. Further, unlike Prozac and Zoloft, a ll DTC advertisements for Effexor XR and Wellbutrin XL feature female models, implying the two brands are targeted at women. Creative e xecution. Building an effective creative strategy is a require ment for a great advertising campaign (Arens et al., 2008). Given that DTC drug advertising could lead to the market expansion of a drug category, one may wonder how practitioners could build creative strategy to trigger the process. The results of this st udy suggest a range of potential strategies. First, the results of this study in support of H2a suggested that including high diagnosticity information could increase risk perceptions. Therefore, presenting uncomfortable but frequently experienced problem s and emphasizing they are typical symptoms of a disease could lead to higher risk perceptions. The results in support of H3b further suggested that once formed, higher risk perceptions would lead to stronger intentions to deal with the particular health p roblem. This may lead to the market expansion of the drug. The results also revealed that of the three independent factors explored in this study, mood state was the strongest determinant of risk perceptions and help seeking intentions. The results in su pport H1a revealed that when opportunity was low, mood state explained 8.6 percent of the variances in risk perceptions. When opportunity was high, moods explained 10.5 percent of the variances in risk perceptions and 4.1 percent of the variances in help s eeking intentions. This

PAGE 110

110 indicates that it is important for advertising practitioners to induce particular mood states that maximize the effects of a drug advertising campaign. The study induced sad moods by making participants write about life events with negative affective connotations. Applying the psychological literature on mood and social cognition to consumer research, Gardner (1985) suggested that marketing practitioners had largely four commercially applicable approaches to influencing consumer mood states, including service procedures and interactions, point of purchase stimuli, the content of the advertising campaign, and the context of the advertising campaign. Combining Gardner s (1985) perspective with the results of this study produces a numbe r of implications for practitioners. For example, if raising consumer awareness of the risk of a disease is the core objective of a drug advertising campaign, practitioners may utilize the content elements of the ad to temporarily induce negative mood stat es. Several anti depressant advertising campaigns indeed have placed emphasis on visual and textual elements that are apparently capable of inducing sad moods. For example, a Zoloft advertisement (Appendix E Zoloft Ad) depicts a figure resembling an egg with a downhearted look under a waning moon against a pitch dark backdrop. A Paxil ad (Appendix E Paxil Ad 1) depicts a woman with an empty, worried look in the midst of anonymous people. Another Paxil ad (Appendix E Paxil Ad 2) displays a list of depr ession symptoms as if they are visually separating a woman with an empty or worried look from her loved ones. This is no surprise, given that one of advertising s basic goals is to raise awareness of an uncomfortable situation and present the product as a means to bring a desired situation. The results of this study accounted for why the approach might be effective for drug advertising. The approach is effective because it puts consumers into negative mood states, and as a result

PAGE 111

111 increases their perceived risk of the problem. The results also suggested that inducing a negative mood state through advertising appeals should be used primarily when the goal of the ad campaign is to induce consumers to form high risk perceptions of the health problem. Media p lan ning. Given that media costs constitute approximately 80 percent of the advertising budget (Kelley & Jugenheimer, 2008), it is crucial to the success of an advertising campaign to deliberate sufficiently on how the selection and purchase of the media space and time may serve the objectives of the campaign. The results of this study and the literature on the social cognitive effects of mood state have implications for making effective media plans. Garder (1985) suggested that consumer moods might be manipul ated by the content as well as context of an advertising campaign. The context is defined as the surroundings in which consumers encounter and process an advertising stimulus. The media context of an ad influences consumers affective states, and further d etermines how they respond to the ad. For example, Goldberg and Gorn (1987) found the affective valence of the media content in which the ad is placed affects consumer processing of the ad. In particular, they found that watching a happy, rather than sad TV program led to more positive mood state greater perceived effectiveness of the ad, more positive cognitive responses to the ad, and better recall of the content of the ad. Goldberg and Gorn s (1987) study suggests that the media strategy for a drug ad vertising campaign may affect consumers moods by determining the media context of an advertising campaign, and further affects how they respond to the ad. The results of this study revealed that, once formed, consumers mood states affected their percepti ons of the future risk of a disease and intentions to seek professional help. Therefore, media planners need to develop plans to place a drug ad in a media context appropriate to the purpose of the campaign.

PAGE 112

112 For example, consumers will process a drug adve rtisement in a negative affective state if the ad is inserted next to the programming or editorial content that evokes sad moods. Therefore, if the objective of the campaign is to increase the perceived future risk of a disease, the media planner should fi nd ways to place it around media time or space capable of triggering negative mood states, such as a news story about terrorism, a feature about people suffering from a natural disaster, a sadness evoking movie, or a special report about infamous crimes. I n contrast, if the campaign has a different key objective, such as enhancing consumer memory of the drug s benefits and competitive advantages, reducing the perceived probability of experiencing side effects, or enhancing self efficacy, the media planner s hould consider placing the ad around a media context conducive to happy moods. Many researchers (Goldberg & Gorn, 1987; Mathur & Chattopadhyay, 1991) revealed that a positive mood state led to a better recall of favorable information and higher perceived p robability of positive outcomes. Similarly, Salovey and Birnbaum (1989) reported happy moods led to increased self efficacy. In fact, this is why most research on the effects of mood state on advertising effectiveness suggested that ads should be placed in a happiness inducing media context to achieve optimum effects. In addition, consumers may experience periodic mood swings along the time of the day. For example, on the average, consumers may be in more negative mood states in the nighttime than in the da ytime. Therefore, if increasing risk perception is the key objective of a campaign, the media planner may place the campaign during late night television shows that may evoke sad moods. If enhancing self efficacy is the key objective, a different time peri od may be selected. Another possibility is that consumers who are chronically in sad moods may tend to view late night television shows. This implies, in addition to inserting commercials during late night shows to target this audience, media planners may track down the late night viewers media use

PAGE 113

113 patterns and place ads in other media vehicles used by the audience. This strategy will be especially effective if the disease category is extremely unfamiliar and therefore it is important to remind the audien ce repeatedly of the risk of the disease. Consumer H ealth The findings also have implications for health promotion. First, the results of this study may indicate that consumers do not always process health and disease information and make judgments in a rational manner. Neither are they aware of the psychological mechanism in which the communication campaign persuades them. Researchers found (Mathur & Chattopadhyay, 1991) that consumers typically did not think that the mood induction procedure might hav e influenced their cognitive responses to the ad, although the procedure significantly affected their responses. Therefore, given that an important objective of a health communication campaign would be helping consumers process information and make health decisions in an informed manner (Peters et al., 2006) the results of this study may lead one to question whether or how consumers need to be informed or educated about how the content and the context of a drug advertising campaign affect their thoughts an d decisions about health issues. In addition, if DTC advertising influences consumer perceptions of the future risk of depression, as the findings of this study suggested, what impact will the phenomenon have on consumer health? To that end, the following three hypothetical situations may be conceptualized. Situation 1 : C onsumers may largely have underrated risk perception, and exposure to the ad results in more realistic risk perception Situation 2 ; C onsumers may largely have overrated risk perception, an d exposure to the ad results in further inflated risk perception. Situation 3 : Consumers may have risk perceptions more or less evenly split around the realistic risk estimates presented by the epidemiological data, and exposure to the ad results in more r ealistic risk perception for some members and more inflated perception for other members.

PAGE 114

114 Increasing risk perceptions may benefit consumers in the first situation, whereas the second situation may not. The third presents a more complex picture. However, given that depression has been a largely under diagnosed and stigmatized disease, some may argue that even the second situation, in which consumers are led to have further inflated risk perceptions, may have positive behavioral consequences, such as encour aging depressed people to visit a doctor s office (Donohue et al., 2004). However, it may also have negative behavioral influences, such as leading consumers to make unnecessary visits to the doctor s office and therefore driving up health costs (Findlay, 2001). Future researchers may conduct a study to determine which of the following three situations holds true. Noticeably male and f emale participants were equal in their risk perceptions and help seeking intentions, re plicating Park and Grow (2008). Consi dering the lifetime risk of depression is approximately 13 percent for men and 20 to 25 percent for women (Kessler et al., 1993, 1994; NCS, 2003) the male participants could have overrated their future risk of depression or it was also possible that the female participants could have underrated their risk. However, one may argue that even if the second scenario were true, meaning even if a sizeable number of consumers were to initially have overrated perceptions of the risk of depression and DTC advertis ing further inflated their risk perception s DTC advertising could still be more beneficial than harmful to society. This perspective is based on the assumption that the social costs of misdiagnosing non depressive persons as clinically depressed and presc ribing antidepressant medication are not as substantial as the costs of failing to detect and medicate clinical depression patients. This perspective may be all the more convincing, considering depression is a largely under diagnosed and under treated dise ase and failing to treat it might result in serious consequences including social isolation and suicidal attempts. On the contrary,

PAGE 115

115 the consequences of misdiagnosing and mis medicating non depressive persons would only include unnecessary visits to the doc tors office and the risk of undergoing the prescribed drugs side effects. Therefore, it would make sense for society to put more emphasis on detecting and treating depression patients than on discouraging non depressive persons from seeking treatment. Th e findings of this study also have implications for the business ethics of pharmaceutical advertising. P ublic criticism is mount ing against DTC advertising as opponents contend that consumers are misled by DTC advertising. At FDA hearings they suggested t he possibility of placing a ban or moratorium on DTC advertising To defend DTC advertising, GlaxoSmithKline sent 8,000 sales representatives out to manage the public policy issue in its favor (Thomaselli 2006). T he pharmaceutical industry s professional association the Pharmaceutical Research Manufacturers of America (PhRMA) has been strong proponents of DTC advertising. To that end, in August 2005 PhRMA introduced codes of conduct aimed at providing advertisers with guideline for DTC advertising (PhRM A, 2005, Appendix F ) What are the potential implications of PhRMA s principles for the ethics of pharmaceutical advertising, especially regarding the research findings of this study? The preamble to the principles states that a strong empirical record d emonstrates that DTC communications about prescription medicines serve the public health by increasing awareness of diseases (PhRMA, 2005) Therefore, PhRMA encourages drug companies to promote disease awareness through DTC advertising (Principle 9 in App endix F ). However, none of the principles refers to the possibility that exposure to DTC advertising for a disease may result in leading consumes to have overrated perceptions of the risk of the disease. PhRMA (2005) also emphasizes DTC advertising should be designed to responsibly educate the consumer about that medicine, and, where appropriate, the condition for which it

PAGE 116

116 may be prescribed. Given risk perception s potentiality to produce behavioral consequences (Block & Keller, 1998; Irwin et al., 1996; Raghubir & Menon, 1998; Siegel et al., 1998), PhRMA may consider encouraging drug companies to include in DTC advertising information the risk of developing the relevant disease. Such an initiative will increase the educational value of DTC advertising, w ith higher potentiality to responsibly increase consumer s awareness about disease s and medicine and discourage them from overrating their vulnerability to the diseases Limitations of the Study The current study has a number of conceptual and methodolo gical limitations. First, this study was conducted in a laboratory. Some may argue because it was conducted in an artificial environment, the study lacks an ecological validity. However, that limitation is inherent in all non quasi experiments designed to control external influences and establish causal relationships. Second, participants in this study might not reflect the general population. Participants were undergraduate students enrolled in introductory advertising classes at the University of Florida and therefore were demographically, geographically, and psycho graphically more homogeneous than the general US population. Therefore, the results of this study may not apply to the general population. However, it will be unrealistic to argue that an ex periment needs to be conducted with a nationally representative sample of participants. A more realistic approach will be to conduct future studies to explore how the results may be replicated or varied across different segments of the population (e.g., yo ung versus senior people), research settings (e.g., laboratory experiment versus field experiment), and cultures (e.g., Asian Americans, Hispanic Americans, versus white Americans). This approach will enhance the external validity of the findings.

PAGE 117

117 Third, s ome may argue it would be more recommendable if the study design had included a post experiment demand characteristic check. Because this study involved deception, making participants believe the session consisted of two separate studies instead of one res earch project, a post experimental check could have been conducted to ensure that the participants did not guess the true purpose of the study. The literature on the effects of mood consists of studies that did not include a demand characteristic check (Ed ell & Burke, 1987; Fedorikhin & Cole, 2004) and those that included a check (Goldberg & Gorn, 1987; Mathur & Chattopadhyay, 1991). Those with a demand characteristic check found that no participants accurately guessed the predictions made in the studies. The last limitation of the study was the failure to find significant effects of the mood diagnosticity opportunity interaction on either risk perceptions or help seeking intentions, or their linear combination. This was a perplexing result, because the mood diagnosticity interaction was not significant when opportunity was low, as hypothesized in H1a and H1b, and significant when opportunity was high, as hypothesized in H2a and H2b. This might have occurred because the manipulation of opportunity was less than optimal. Especially, one of the three response time measures for intention showed the manipulation was not optimal. Since opportunity was hypothesized to determine the effects of the mood diagnosticity interaction, a more effective manipulation of the construct may have generated a significant three way interaction. Suggestions for F uture R esearch This study showed consumer mood state influenced the perceived future risk of a disease and intentions to seek professional help to discuss it. Mood state was manipulated by requiring participants to write about sad or happy life events. The findings would be more commercially applicable if mood state was manipulated in other ways. Gardner (1985) suggested particular

PAGE 118

118 mood states might be induced by var ying the content of the ad or the context surrounding the ad. Therefore, future researchers may expose participants to a happiness or sadness inducing media context, such as a magazine report or television show, and explore how consumers process informatio n from a drug advertisement. Another research idea would be to see if manipulated moods would affect consumer cognitions about aspects of the disease or drug other than risk perceptions or help seeking intentions. For example, a future study may be desig ned to explore if different moods lead to different expectations of the drug s success rate, the probability of experiencing side effects, the severity of the side effects, recall of the drug s competitive advantages, and so on In fact, Salovey and Birnbu am (1989) found that happy moods led to higher self efficacy regarding health issues. A future study may replicate this finding in the context of DTC drug advertising. In addition, this study does not take into account the perspective that various specif ic affective states might exist under the broad categories of negative and positive mood states. For example, Raghunathan, Pham, and Corfman (2006) and Raghunathan and Pham (1999) found that anxiety led to risk aversive behavior whereas sadness caused risk taking behavior, revealing that specific negative mood states tended to have heterogeneous effects on judgment and decision making process. Therefore, it will be worthwhile to explore if different types of positive and negative mental states would pro duce the same affects as were observed in this study. As discussed in the limitation section, this study failed to find significant effects of the thee way interaction of mood, diagnosticity, and opportunity. Because the manipulation checks revealed that mood and diagnosticity were successfully manipulated and opportunity manipulation was less than optimal, the failure is likely attributable to the way opportunity was

PAGE 119

119 manipulated. Therefore, it will be worthwhile to test if an alternative method of manipul ating opportunity, such as forcing high opportunity participants to spend a certain amount of time before responding and low opportunity participants to respond within a limited time frame, would generate significant three way interaction. In addition, th e mood and social cognition literature suggests that happy moods facilitate processing and recall of information with a positive affective connotation, whereas sad moods enhance recall of negative information. This may indicate the effects of informational ads with negatively originated motives versus transformational ads with positively originated motives may have differential effects on the way consumer think and feel about the disease and the drug. A future study may be designed to explore how the intera ction between mood and information versus transformational motives affects consumer responses to the ad vertisement Another future research area may emerge from the perspective that individuals have varying levels of information processing capacity, which may determine the way moods affect social judgments. For example, Petty et al. ( 1993 ) revealed that consumers with high need for cognition, a likely correlate of high information processing capacity, were less affected by induced mood states in judgment m aking than those with low need for cognition. Therefore, consumers with high processing capacity, an umbrella concept that could be operationally defined as high need for cognition, high numeracy, or high level of education among others, may be less influe nced by mood states in forming estimates of the future risk of diseases and help seeking intentions. This is even more probable when one considers that processing the APA diagnostic guideline and forming future risk perception s require numerical and probab ilistic thinking.

PAGE 120

120 In addition, mood states were temporarily manipulated in this study. Research (Schwartz & Clore, 1983) generally found that experimentally induced moods tended to be short lived. If induced moods are transient, how long will their effe cts on risk perceptions and help seeking intentions last? If consumer judgments are as transient as their determinants are, the practical implications of this study will be seriously limited. A future study may be conducted to explore if consumer risk perc eptions and help seeking intentions, once there are formed in an experimental setting, mold their future health behavior. Last, it will contribute to the DTC literature to explore how the effects of mood, diagnosticity, and opportunity may differ across v arious diseases. This study revealed the three independent variables interactive effects on consumer perceptions of depression. It will be worthwhile to explore whether the effects are replicated for other diseases, such as attention disorder, sleep probl ems, or seasonal allergy. A disease may have aspects that set it apart from other diseases. For example, restless leg syndrome likely differs from depression in the sense that the perceived risk of experiencing is lower, the symptoms are less ambiguous and less intense, and consumers are less knowledgeable about the disease. Then it will be worthwhile to explore if the three independent variables have the same effects on consumer perceptions of restless leg syndrome and help seeking intentions. It will be f urther worthwhile to explore how one can categorize diseases, and how the effects of drug advertising appeals may change depending on the category of diseases. Such research will contribute valuable insights to the practitioners in charge of building creat ive strategy for various drug campaigns.

PAGE 121

121 APPENDIX A INSTRUMENTAL MANIPUL ATION OF MOOD

PAGE 122

122 Script for Happiness Inducing Procedure Instruction: P eople experience many types of life events. This study is designed to build a life event inve ntory and explore how people represent their autobiographical memories. For that purpose, you will be asked to describe three life events that made you very happy. Please d escribe your life events as realistically as possible such that a person reading th e description would become happy just from hearing about the situation. Please spend five minutes for each situation.

PAGE 123

123 Life Event One Please d escribe a life event that made you very happy as realistically as possible such that a per son reading the description would become happy just from hearing about the situation (5 minutes)

PAGE 124

124 Life Event Two Please d escribe a life event that made you very happy as realistically as possible, such that a person reading the d escription would become happy just from hearing about the situation (5 minutes).

PAGE 125

125 Life Event Three Please d escribe a life event that made you very happy as realistically as possible, such that a person reading the description wou ld become happy just from hearing about the situation (5 minutes)

PAGE 126

126 The following is a questionnaire on your current mood state. For each of the following items, p lease check a scale that best represents the feelings you are currently unde rgoing. Gloomy Joyful Sad Happy Upset Elated

PAGE 127

127 S cript for Sadness Inducing Procedure P eople experience many types of life events. This study is designed to build a life event in ventory and explore how people represent their autobiographical memories. For that purpose, you will be asked to describe three life events that made you very sad. Please d escribe your life events as realistically as possible such that a person reading th e description would become sad just from hearing about the situation. Please spend five minutes for each situation.

PAGE 128

128 Life Event One Please d escribe a life event that made you very sad as realistically as possible such that a pers on reading the description would become sad just from hearing about the situation (5 minutes)

PAGE 129

129 Life Event Two Please d escribe a life event that made you very sad as realistically as possible, such that a person reading the descrip tion would become sad just from hearing about the situation (5 minutes).

PAGE 130

130 Life Event Three Please d escribe a life event that made you very sad as realistically as possible, such that a person reading the description would become sa d just from hearing about the situation (5 minutes)

PAGE 131

131 The following is a questionnaire on your current mood state. For each of the following items, p lease check a scale that best represents the feelings you are currently undergoin g. Gloomy Joyful Sad Happy Upset Elated

PAGE 132

132 APPENDIX B INSTRUMENTAL MANIPUL ATION OF PERCEIVED DIAGNOSTICITY H igh Diagnosticity Version of the Antidepressant Advertisement

PAGE 133

133 Low D iagnostic ity Version of the Antidepressant Advertisement

PAGE 134

134 APPENDIX C INSTRUMENTAL MANIPUL ATION OF OPPORTUNITY Instruction for Participants Under the Low Opportunity Condition Read this instruction VERY carefully. What follows is a questionnair e on your life and depression. In the real world, consumers often make quick judgments while they are busy. To make this study as realistic as possible, please complete the following three questions AS FAST AS YOU CAN Instruction for Participants Under the High Opportunity Condition Read this instruction VERY carefully. What follows is a questionnaire on your life and depression. Researchers point out that having accurate ideas about a disease is important for preventing or treating the disease. So pleas e take AS MUCH TIME AS YOU NEED to deliberate sufficiently.

PAGE 135

135 APPENDIX D QUESTIONNAIRE AND TH E INSTRUMENTAL MANIP ULATION OF OPPORTUNI TY High Opportunity Version The primary focus of this questionnaire is to explore your understanding of the advertisement you just saw. In answering the questions, try as best as you can to report your own thoughts about the ad. Please note that clinical depression is defined in this project as a form of medical illness that may require doctor s intervention for treatment. A 1) For the last two weeks, how often did you have the following feeling/experience? Never Nearly everyday Sleep problems A 2 ) According to the ad how likely is it that your reported experience with sleep problems would indicate s you are clinically depressed? Not at all Very likely B 1) For the last two weeks, how often did you have the following feeling/experience? Never Nearly everyday Difficulty making decisions B 2 ) According to the ad how likely is it that your reported experience with difficulty making decisions indicates you are clinically depre ssed? Not at all Very likely C 1 ) For the last two weeks, how often did you have the following feeling/experience? Never Nearly everyday Depressed mood

PAGE 136

136 C 2 ) According to the ad how likely is it that you r reported experience with depressed mood indicates you are clinically depressed ? Not at all Very likely D 1 ) For the last two weeks, how often did you have the following feeling/experience? Never Nearly everyday Low phy sical energy D 2 ) According to the ad how likely is it that your reported experience with low physical energy indicates you are clinically depressed? Not at all Very likely E 1) In you thinking, what are the chances that you will suffer from clinical depression in the near future? ____________% Take AS MUCH TIME AS YOU NEED and report in percentage between 0% and 100% Then click Continue to proceed. E 1 1) If you have ever been diagnosed of clinic al depression, please check here: ________________ F) In your thinking, your risk of suffering from clinical depression in the near future will be Take AS MUCH TIME AS YOU NEED and check the scale that best represents your perception Then click Contin ue to proceed. Very low Moderate Very high G) Compared to people of your age, your risk of suffering from depression in the near future will be Take AS MUCH TIME AS YOU NEED and check the scale that best represents your perception Then click Continue to proceed. Much lower Neither lower nor higher Much higher

PAGE 137

137 H) Please check the scale that best reflects your agreement with each of the following items. Take AS MUCH TIME AS YOU NEED and check the scale that best represents agreement. Very Strongly Disagree Strongly Disagree Disagree Neutral Agree Strongly Agree Very Strongly Agree If the University Health Services offered a free screening test for depression, I would intend to receive it. If the University Health Services offered a free educational program about depression, I would intend to participate. If the University Health Services offered an opportunity to consult doctors about depression, I would intend to particip ate. I 1 ) Has anyone among your family members, relatives or close friends ever suffered from depression? Yes______ No______ Dont know ______ I 2 ) Has anyone among your family members, relatives, or close friends ever sought prof essional help to deal with depression? Yes ______ No______ Dont know ______ I 3) Has anyone among your family members, relatives, or close friends taken antidepressant medication? Yes ______ No______ Dont know ______ I 4) Have you ever be en diagnosed with ADHD ( attention deficit hyperactivity disorder )? J 1) Please check your gender. Male_____ Female_____ J 2) How old are you? ______ years old J 3) What is your ethnic background? White, not Hispanic _____ Hispanic, of any race _____ Black, not Hispanic _____ Asian or Pacific Islander ________ American Indian, Eskimo, or Aleut _______ Other _____ J 4) Which class do you consider your family to be among the following five categories? Working Class________ Lower Middle C lass________ Middle Class_______ Upper Middle Class________ Upper Class _________

PAGE 138

138 Debriefing Statements Debriefing Statement for Session A: The procedure you completed in Session A might have caused you to feel happy or sad. R esearchers have demonstra ted that the moods created by this type of procedure typically disappear within a short period of time. Debriefing Statement for Session B: The ad presented in this study was developed for the purpose of research, and does not represent a real brand. How ever, the information in the ad accurately describes clinical depression.

PAGE 139

139 Low Opportunity Version The primary focus of this questionnaire is to explore your understanding of the advertisement you just saw. In answering the questi ons, try as best as you can to report your own thoughts about the ad. Please note that clinical depression is defined in this project as a form of medical illness that may require doctor s intervention for treatment. A 1) For the last two weeks, how of ten did you have the following feeling/experience? Never Nearly everyday Sleep problems A 2 ) According to the ad how likely is it that your reported experience with sleep problems would indicate s you are clinically depressed? Not at all Very likely B 1) For the last two weeks, how often did you have the following feeling/experience? Never Nearly everyday Difficulty making decisions B 2 ) According to the ad how likely is it that your reported experience with difficulty making decisions indicates you are clinically depre ssed? Not at all Very likely C 1 ) For the last two weeks, how often did you have the following feeling/experience? Never Nearly everyday Depressed mood C 2 ) According to the ad how likely is it that your reported experience with depressed mood indicates you are clinically depressed ?

PAGE 140

140 Not at all Very likely D 1 ) For the last two weeks, how often did you have the following feeling/experience? Never Nearly everyday Low physical energy D 2 ) According to the ad how likely is it that your reported experience with low physical energy indicates you are clinically depressed? Not at all Very likely E 1) In you thinking, what are t he chances that you will suffer from clinical depression in the near future? ____________% AS FAST AS YOU CAN, report in percentage between 0% and 100%. Then click Continue to proceed E 1 1) If you have ever been diagnosed of clinical depressi on, please check here: ________________ F) In your thinking, your risk of suffering from clinical depression in the near future will be AS FAST AS YOU CAN check the scale that best represents your perception Then click Continue to proceed. Very lo w Moderate Very high

PAGE 141

141 G) Compared to people of your age, your risk of suffering from depression in the near future will be AS FAST AS YOU CAN check the scale that best represents your perception Then click Continue to proce ed. Much lower Neither lower nor higher Much higher H) Please check the scale that best reflects your agreement with each of the following items. As FAST AS YOU CAN, check the scale that best represents your agreement. Then click co ntinue to proceed. Very Strongly Disagree Strongly Disagree Disagree Neutral Agree Strongly Agree Very Strongly Agree If the University Health Services offered a free screening test for depression, I would intend to receive it. If the University Health Services offered a free educational program about depression, I would intend to participate. If the University Health Services offered an opportunity to consult doctors about depression, I would intend to participate. I 1 ) Has anyone among your family members, relatives or close friends ever suffered from depression? Yes______ No______ Dont know ______ I 2 ) Has anyone among your family members, relatives, or close friends ever sought professional help to deal with depression? Yes ______ No______ Dont know ______ I 3) Has anyone among your family members, relatives, or close friends taken antidepressant medication? Yes ______ No______ Dont know ______ I 4) Have you ever been diagn osed with ADHD ( attention deficit hyperactivity disorder )? J 1) Please check your gender. Male_____ Female_____

PAGE 142

142 J 2) How old are you? ______ years old J 3) What is your ethnic background? White, not Hispanic _____ Hispanic, of any race _____ Black, not Hispanic _____ Asian or Pacific Islander ________ American Indian, Eskimo, or Aleut _______ Other _____ J 4) Which class do you consider your family to be among the following five categories? Working Class________ Lower Middle Class____ ____ Middle Class_______ Upper Middle Class________ Upper Class _________ Thank you very much for your participation Debriefing Statement for Session A: The procedure you completed in Session A might have caused you to feel happy or sad R esearchers have demonstrated that the moods created by this type of procedure typically disappear within a short period of time. Debriefing Statement for Session B: The ad presented in this study was developed for the purpose of research, and does not represent a real brand. However, the information in the ad accurately describes clinical depression.

PAGE 143

143 APPENDIX E SAMPLES OF ANTIDEPRE SANT ADS Zoloft Ad Paxil Ad 1

PAGE 144

144 Paxil Ad 1

PAGE 145

145 Paxil Ad 2 APPENDIX 6

PAGE 146

146 APPENDIX F P H RMA S GUIDING PRINCIPLES F OR DIRECT TO CONSUMER ADVERTISING To express the commitment of PhRMA members to deliver DTC communications that serve as valuable contributors to public health, PhRMA has established the foll owing voluntary guiding principles. Principle 1. These Principles are premised on the recognition that DTC advertising of prescription medicines can benefit the public health by increasing awareness about diseases, educating patients about treatment opti ons, motivating patients to contact their physicians and engage in a dialogue about health concerns, increasing the likelihood that patients will receive appropriate care for conditions that are frequently under diagnosed and under treated, and encouraging compliance with prescription drug treatment regimens. Principle 2. In accordance with FDA regulations, all DTC information should be accurate and not misleading, should make claims only when supported by substantial evidence, should reflect balance betw een risks and benefits, and should be consistent with FDA approved labeling. Principle 3. DTC television and print advertising which is designed to market a prescription drug should also be designed to responsibly educate the consumer about that medicine and, where appropriate, the condition for which it may be prescribed. Principle 4. DTC television and print advertising of prescription drugs should clearly indicate that the medicine is a prescription drug to distinguish such advertising from other adv ertising for non prescription products. Principle 5. DTC television and print advertising should foster responsible communications between patients and health care professionals to help patients achieve better health and a more complete appreciation of b oth the health benefits and the known risks associated with the medicine being advertised. Principle 6. In order to foster responsible communication between patients and health care professionals, companies should spend an appropriate amount of time to e ducate health professionals about a new medicine or a new therapeutic indication before commencing the first DTC advertising campaign. In determining what constitutes an appropriate time, companies should take into account the relative importance of inform ing patients of the availability of a new medicine, the complexity of the risk benefit profile of that new medicine and health care professionals knowledge of the condition being treated. Companies should continue to educate health care professionals as a dditional valid information about a new medicine is obtained from all reliable sources. Principle 7. Working with the FDA, companies should continue to responsibly alter or discontinue a DTC advertising campaign should new and reliable information indica te a serious previously unknown safety risk. Principle 8. Companies should submit all new DTC television advertisements to the FDA before releasing these advertisements for broadcast.

PAGE 147

147 Principle 9. DTC television and print advertising should include inf ormation about the availability of other options such as diet and lifestyle changes where appropriate for the advertised condition. Principle 10. DTC television advertising that identifies a product by name should clearly state the health conditions for which the medicine is approved and the major risks associated with the medicine being advertised. Principle 11. DTC television and print advertising should be designed to achieve a balanced presentation of both the benefits and the risks associated with the advertised prescription medicine. Specifically, risks and safety information in DTC television advertising should be presented in clear, understandable language, without distraction from the content, and in a manner that supports the responsible dialog ue between patients and health care professionals. Principle 12. All DTC advertising should respect the seriousness of the health conditions and the medicine being advertised. Principle 13. In terms of content and placement, DTC television and print ad vertisements should be targeted to avoid audiences that are not age appropriate for the messages involved. Principle 14. Companies are encouraged to promote health and disease awareness as part of their DTC advertising. Principle 15. Companies are enco uraged to include information in all DTC advertising, where feasible, about help for the uninsured and underinsured.

PAGE 148

148 REFERENCES Allison Ottey, S., Ruffin, K., Allison, K., & Ottey, C. C. (2003). Assessing the impact of direct to consume r advertisements on the AA patient: A multitude survey of patients during the office visit. Journal of the National Medical Association, 95(2), 120 131. Alpert, J. I., & Alpert, M. I. (1990). Music influences on mood and purchase intentions. Psychology & M arketing, 7(2), 109 133. Altshuler, L. L., Hendrick, V., & Burt, V. K. (1998). Mood Disorders. In L. A. Wallis (Ed.), Textbook of womens health. Philadelphia, PA: Lippincott Raven Publishers. American Psychiatric Association. (1994). Diagnostic and stati stical manual of mental disorders IV Washington, D.C. An, S. ( 2007 ). College students exposure to direct to consumer (DTC) antidepressants ads : Impact on perceived prevalence, treatment options and drug inquiry. Presented to the Health Communication Div ision, International Communication Association annual conference, San Francisco, CA. M ay 24 28, 2005. An, S., & Jin, H. S. (2005). The effects of direct to consumer prescription drug advertising on perception toward the stigmatized illnesses Presented to the Advertising Division, Association for Education in Journalism and Mass Communication annual conference, San Antonio, TX. August 10 13, 2005. Arens, Weigold, & Arens (2008). Contemporary advertising. New York: McGraw Hill Companies, Inc. Batra, R., & Ray, M. L. (1986). Affective responses mediating acceptance of advertising. Journal of Consumer Research, 13, 234 249. Bell, R. A., Kravitz M. S., & Wilkes R. L. ( 1999 ). Advertisement induced prescription drug requests: Patients anticipated react ions to a physician who refuses The Journal of Family Practice, 4 8 ( 6 ), 446 452 Bell, R. A., Wilkes, M. S., & Kravitz, R. L. (2000). The educational value of consumer targeted prescription drug print advertising. The Journal of Family Practice, 49(12), 1 092 1098. Beltramini, R. F. (2006). Consumer believability of information in direct to consumer (DTC) advertising of prescription drugs. Journal of Business Ethics, 63, 333 343. Blalock, S. J., DeVellis, B. E., Afifi, R. A., & Sandler, R. S. (1990). Risk perceptions and participation in colorectal cancer screening. Health Psychology, 9, 792 806. Blankenhorn, K., Duckwitz, N., & Kerr, M. (2001). Power to the people. Medical Marketing & Media, 36(8), 66 70.

PAGE 149

149 Block, L. G., & Keller, P. A. (1998). Beyond Prote ction Motivation: An Investigative Theory of Health Appeals. Journal of Applied Social Psychology, 28(17), 1584 1608. Brichacek, A., & Sellers, L. J. (2001). Flexing their budgets: Big pharma spend trends. Pharmaceutical Executive Retrieved November 30, from http://www.pharmaportal.com/articles/pe/pe0901_sellers.pdf Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129 148. Bower, G. H., Gilligan, S. G., & Monteir o, K. P. (1981). Selectivity of learning caused by affective states. Journal of Experimental Psychology: General, 110, 451 473. Bower, G. H., Monteiro, K. P., & Gilligan, S. G.. (1978). Emotional mood as a context for learning and recall. Journal of Verba l Learning and Verbal Behavior, 17, 573 585. Calfee, J. E. (2002). Public policy issues in direct to consumer advertising of prescription drugs. Journal of Public Policy & Marketing, 21(2), 174 193. Clore, G. F., Schwarz, N., & Conway, M. (1994). Affectiv e causes and consequences of social information processing. In R. S. Wyer & T. K. Srull (Eds.), Handbook of social cognition (2nd ed.), Hillsdale, NJ: Erlbaum, 323 417. Coney, S. (2002). Direct to consumer advertising of prescription pharmaceuticals: A co nsumer perspective from New Zealand. Journal of Public Policy and Marketing, 22(2), 213 223. Constans, J. I., & Mathews, A. M. (1993). Mood and the subjective risk of future events. Cognition and Emotion, 7, 545 560. Constans, J. I. (2001). Worry propensi ty and the perception of risk. Behavioral Research and Therapy, 39, 721 729. Croyle, R. T. & Lerman, C. (1999). Risk communication in genetic testing for cancer susceptibility. Journal of National Cancer Institute Monographs, 25, 59 66. Croyle, R. T., & Uretsky, M. B. (1987). Effects of mood on self appraisal of health status. Health Psychology, 6, 239 253. Davis, J. J. (2000). Riskier than we think? The relationship between risk statement completeness and perceptions of direct to consumer advertised pr escription drugs. Journal of Health Communication, 5, 349 369. Donohue, J. M., & Berndt, E. R. (2004). Effects of direct to consumer advertising on medication choice: The case of anti depressants. Journal of Public Policy & Marketing, 23(2), 115 127. Dono hue, J. M., Berndt, E. R., Rosenthal, M., Epstein, A. M., & Frank, R. G. (2004). Effects of pharmaceutical promotion on adherence to the treatment guidelines for depression. Medical Care, 42(12), 1176 1185.

PAGE 150

150 Dunn, J. R., & Schweitzer, M E. (2005). Feeling and believing: The influence of emotion on trust. Journal of Personality and Social Psychology, 88(5), 736 748. Elliott, C. (2003) Better than Well: American Medicine Meets the American Dream New York: W.W. Norton. Edell, J. A., & Burke, M. C. (1987). Th e power of affects in understanding advertising effects. Journal of Consumer Research, 14, 421 433. Epstein, S. (1994). Integration of the cognitive and the psychodynamic unconscious. American Psychologist, 49, 709 724. Fedorikhin, A., & Cole, C. A. (200 4). Mood effects on attitudes, perceived risk and choice: Moderators and mediators. Journal of Consumer Psychology, 14(1&2), 2 12. Fiedler, K., & Forgas, J. P. (Eds.). (1988). Affect, cognition and social judgments. Gttingen Germany: Hogrefe. Findlay, S. D. (2001). Direct to consumer promotion of prescription drugs: Economic implications for patients, payers, and providers. PharmacoEconomics, 19(2), 109 119. Fischhoff, B., Slovic, P., Lichtenstein, S., Read, S., & Combs, B. (1978). How safe is safe enou gh? A psychometric study of attitudes toward technological risks and benefits. Policy Sciences, 9, 127 152. Forgas, J. P. (1992). Affect in social judgments and decisions: A multi process model. In M. Zanna (Ed.), Advances in Experimental Social Psycholog y (Vol. 25, pp. 227 275). San Diego, CA: Academic Press. Forgas, J. P. (1995). Mood and judgment: The affect infusion model. Psychological Bulletin, 117(1), 39 66. Forgas, J. P., Bower, G. H., & Krantz, S. (1984). The influence of mood on perceptions of social interactions. Journal of Experimental Social Psychology, 20, 497 513. Forgas, J. P., Bower, G. H., & Moylan, S. J. (1990). Praise or blame? Moods effects on attributions for success or failure. Journal of Personality and Social Psychology, 59, 809 819. Frijda, N. H. (1986). The emotions. Cambridge, England: Cambridge University Press. Gasper, K., & Clore, G. L. (1998). The persistent use of negative affect by anxious individuals to estimate risk. Journal of Personality and Social Psychology, 74(5) 1350 1363. Gasper, K., & Clore, G. L. (2000). Do you have to pay attention to your feelings to be influenced by them? Personality and Social Psychology Bulletin, 26(6), 698 711.

PAGE 151

151 Gorn, G. J., Goldberg, M. E., & Basu, K. (1993). Mood, awareness and produ ct evaluation. Journal of Consumer Psychology, 2(30, 237 256. Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (2008). Multivariate data analysis. Upper Saddle River, NJ: Prentice Hall PTR. Herzenstein, M., Misra, S., & Posavac, S. S. (2005). H o w consumers attitudes toward direct to consumer advertising of prescription drugs influence ad effectiveness, and consumer and physician behavior. Marketing Letters, 15(4), 201 212. Higgins, E. T., & King, G. (1981). Accessibility and applicability: So me activation rules influencing judgment. Journal of Experimental Social Psychology, 31, 218 243. Holmer, A. F. (1999). Direct to consumer prescription drug advertising builds bridges between patients and physicians. Journal of American Medical Associati on, 281(4), 380 382. Holmer, A. F. (2002). Direct to consumer advertising Strengthening our health care system. New England Journal of Medicine, 346, 526 528. Hollon, M. F. (2004). Direct to consumer marketing of prescription drugs: A current perspecti ve for neurologists and psychiatrists CNS Drugs, 18(2), 69 77. Irwin, K. L., Valdiserri, R. O., & Holmberg, S. D. (1996). The Acceptability of Voluntary HIV Antibody Testing in the United States: A Decade of Lessons Learned. AIDS, 10(14), 1707 1717. Iizu ka, T. (2004). What explains the use of direct to consumer advertising of prescription drugs? The Journal of Industrial Economics, 52(3), 349 372. Iizuka, T., & Jin, G. Z. (2005). The effect of prescription drug advertising on doctor visits. Journal of Ec onomics & Management Strategy, 14(3), 701 727. Isen, A. M. (1984). Toward understanding the role of affect in cognition. In R. Wyer & T. Srull (Eds.), Handbook of social cognition (pp. 179 236). Hillsdale, NJ: Lawrence Erlbaum. Isen, A. M., Clark, M., Sh alker, T. E., & Karp L. (1978). Affect, accessibility of material in memory, and behavior: A cognitive loop? Journal of Personality and Social Psychology, 36(1), 1 12. Jacoby, L. L., & Dallas, M. (19081). On the relationship between autobiographical memor y and perceptual learning. Journal of Experimental Psychology: General, 110, 306 340. Jarvis, W. B. G. (2000). MediaLab 2000 [computer software]. New York, NY: Empirisoft. John, D. R. (1999). Consumer socialization of children: A retrospective look at tw enty five years of research. Journal of Consumer Research, 26, 183 213.

PAGE 152

152 Johnson, E. J., & Tversky, A. (1983). Affect, generalization, and the perception of risk. Journal of Personality and Social Psychology, 45(1), 20 31. Kahneman, D., & Tversky, A. (197 2), Subjective Probability: A Judgment of Representativeness Cognitive Psychology 3 (3), 430 454. Kalichman, S. C., & Cain, D. (2005), Perceptions of HIV/AIDS Prevalence and Risks for HIV/AIDS and Other Sexually Transmitted Infections: Preliminary Stud y of Intuitive Epidemiology Annals of Behavioral Medicine 29(2), 100 105. Keller, L. D., & Jugenheimer. D. W. (2008). Advertising media planning: A brand management approach. New York: M.E. Sharpe, Inc. Kessler, R. C., Avenevoli, S., & Merikangas, K. R (2001). Mood disorder in children and adolescents: An epidemiological perspective. Biological Psychiatry, 49, 1002 1014. Kraus, N., Malmfors, T., & Slovic, P. (1992). Intuitive toxicology: Expert and lay judgments of chemical risks. Risk Analysis, 12(2) 215 232. Kravitz, R. L., Epstein R. M. Feldman, M. D., Franz, C. E., Azari, R., & Wilkes, M.S., Hinton, L., & Franks, P. (2005). Influence of Patients Requests for Direct to Consumer Advertised Antidepressants: A Randomized Controlled Trial. Journal of the American Medical Association, 293( 16 ) 1995 2002. Kravitz, R. L., & Wilkes, M. S. (2000). Direct to consumer prescription drug advertising, 1989 1998: A content analysis of conditions, targets, inducements, and appeals. Journal of Family Practice, 4 9(4), 329 335. Lerner, J. S., & Keltner, D. (2001). Fear, anger, and risk. Journal of Personality and Social Psychology, 81(1), 146 159. Levy, A. G., Shea, J., Williams, S. V., Quistberg, A., & Armstrong, K. (2006). Measuring perceptions of breast cancer risk. Cancer Epidemiology Biomarkers & Prevention, 15(10), 1893 1898. Lipkus, I. M., Biradavolu, M., Fenn, K., Keller, P. A., & Rimer, B. A. (2001). Informing Women About Their Breast Cancer Risks: Truth and Consequences. Health Communication 13 (2), 20 5 226. Lipkus, I. M., Kuchibhatla, M., BcBride, C. M., Bosworth, H. B., Pollak, K. I., Siegler, I. C., & Rimer, B. K. (2000). Relationships among breast cancer perceived absolute risk, comparative risk, and worries. Cancer Epidemiology, Biomarkers & Preve ntion, 9, 973 975. Lipkus, I. M., Rimer, B. K., Lyna, P. R., Pradhan, A. A., Conaway, M., & Woods Powell, C. T. (1996). Colorectal screening patterns and perceptions of risk among African American users of a community health center. Journal of Community H ealth, 21, 409 427.

PAGE 153

153 Loroz, P. S., & Lichtenstein, D. R. (2004). The moderating role of perceived behavior outcome covariation on consumer estimates of health risk. Journal of Public Policy & Marketing, 23(1), 54 64. Main, K. J., Argo, J. J., & Huhmann, B A. (2004). Pharmaceutical advertising in the USA: Information or Influence? International Journal of Advertising, 23, 119 142. Mathur, M., & Chattopadhyay, A. (1991). The impact of moods generated by television programs on responses to advertising. Psy chology & Marketing, 8(1), 59 77. Mayer, J. D., Gaschke, Y. N., Braverman, D. L., & Evans, T. W. (1992). Mood congruent judgment is a general effect. Journal of Personality and Social Psychology, 63(1), 119 132. Mehta, A., & Purvis, S. (2003). Consumer r esponses to print prescription drug advertising. Journal of Advertising Research, 43(2), 194 206. Menon, G., Block, L. G., & Ramanathan, S. (2002), We re at as Much Risk as We Are Led to Believe Effects of Message Cues on Judgments of Health Risk, Jour nal of Consumer Research 28 (March), 533 549. Mintzes, B., Barer, M. L., Kravitz, R. L., Kazanjian, A., Bassett, K., Lexchin, J., Evans, R. G., Pan, R., & Marion, S. A. (2002). Influence of direct to consumer pharmaceutical advertising and patients requ ests on prescribing decisions: Two site cross sectional survey. British Medical Journal, 324(7332), 278 279. Morgan, C., & Levy, D. (1998). To their health. Brandweek, 39, 30 33. Nygren, T. E., Isen, A. M., Taylor, P. J., & Dulin, J. (1996). The influenc e of positive affect on the decision rule in risk situations: Focus on outcome (and especially avoidance of loss) rather than probability. Organizational Behavior and Human Decision Processes, 66(1), 59 72. Park, J & Grow, J. M. ( 2008 ). The social reali ty of depression: DTC advertising of antidepressant s and perceptions of the prevalence and lifetime risk of depression Journal of Business Ethics 79(4), 379 393 Park, J., & Lemanski J. L. (2007). Content elements of DTC advertising for antidepressants and perceived lifetime risk of clinical depression Proceeding s of the American Academy of Advertising Burlington, VT, April 12 15, 2007. Perri, M., & Dickson, W. M. (1988). Consumer reaction to a direct to consumer prescription drug advertising campai gn. Journal of Health Care Marketing, 8(2), 66 69. Peters, E., Lipkus, I., & Diefenbach, M. A. (2006). The function of affect in health communications and in the construction of health preferences. Journal of Communication, 56, S140 S162.

PAGE 154

154 Pettit, J. W., Kline, J. P., Gencoz, T., Gencoz, F., & Joiner, T. E. (2001). Are happy people healthier? The specific role of positive affect in predicting self reported health symptoms. Journal of Personality, 35(4), 521 536. Petty, R. E., Schumann, D. W., Richman, S. A., & Strathman, A. J. (1993). Positive mood and persuasion: Different roles for affect under high and low elaboration conditions. Journal of Personality and Social Psychology, 64(1), 5 20. Pham, M. T. (1998). Representativeness, relevance and the use o f feelings in decision making. Journal of Consumer Research, 25, 144 159. Pham, M. T., Cohen, J. B., Pracejus, J. W., & Hughes, G. D. (2001). Affect monitoring and the primacy of feelings in judgment. Journal of Consumer Research, 28, 167 18. Raghubir, P ., & Menon, G. (1998). AIDS and Me, Never the Twain Shall Meet: The Effects of Information Accessibility on Judgments of Risk and Advertising Effectiveness. Journal of Consumer Research, 25(June), 52 63. Raghubir, P., & Menon, G. (2005). Depressed or Just Blue?: The Persuasive Effects of a Self Diagnosis Inventory, Journal of Applied Social Psychology, 35 (12), 2535 2559. Raghunathan, R., & Pham, M. T. (1999). All negative moods are not equal: Motivational influences of anxiety and sadness on decision ma king. Organizational Behavior and Human Decision Processes, 79(1), 56 77. Raghunathan, R., & Pham, M. T. & Corfman, K. P. (2006). Informational properties of anxiety and sadness, and displaced coping. Journal of Consumer Research, 32(4), 596 601. Reve ntlow, S., Hvas, A. C., & Tulinius, C. (2001). In really great danger... The concept of risk in general practice. Scandinavian Journal of Primary Health, 19, 71 75. Riskind, J. H. (1983). Nonverbal expressions and the accessibility of life experience me mories: A congruency hypothesis. Social Cognition, 2, 62 86. Robinson, J. K., Rigel, D. S., & Amonette, R. A. (1998). What promotes skin self examination? Journal of American Academy of Dermatology, 38, 752 757. Rosenstock, I. M. (1990). The health belie f model: Expanding health behavior through experience. In K. Glanz, F. M. Lewins, & B. K. Rimer (Eds.), Health behavior and health education: Theory, research, and practice (pp 39 62). San Francisco: Jossey Bass Publications. Rosenthal, M. B., Berndt, E. R., Donohue, J. M., Frank, R. G., & Epstein, A. M. (2002). Promotion of prescription drugs to consumers. New England Journal of Medicine, 346(7), 498 505. Roth, M. S. (1996). Patterns in direct to consumer prescription drug print advertising and their pub lic policy implications. Journal of Public Policy & Marketing, 15(1), 63 65.

PAGE 155

155 Salovey, P., & Birnbaum, D. (1989). Influence of mood on health relevant cognitions. Journal of Personality and Social Psychology, 57, 539 551. Sasich, L. D. (1999). Peddling pil ls : The rise of direct to consumer prescription drug advertising and the dangers to consumers. Multinational Monitor, 20, 23 26. Schwarz, N. (2004). Metacognitive experiences in consumer judgment and decision making. Journal of Consumer Psychology, 14(4), 332 348. Schwarz, N., & Clore, G. (1983). Mood, misattribution, and judgment of well being : Informative and directive functions of affective states. Journal of Personality and Social Psychology, 45(3), 513 523. Schwarz, N., Strack F., Kommer, D., Wagner D. (1987). Soccer, rooms, and the quality of life: Mood effects on judgments of satisfaction with life in general and with specific domains. European Journal of Social Psychology, 17, 69 79. Seow, A., Wong, M. L., Smith, W. C. S., & Lee, H. P. (1995). Be liefs and attitudes as determinants of cervical cancer screening: A community based study in Singapore. Preventive Medicine, 24, 134 141. Sherman, S. J. and Corty E. ( 1984 ). Cognitive heuristics. I n R. S. Wyer & T. K. Srull (Eds.), Handbook of social cog nition, 1 ( pp. 189 286 ) Hillsdale, NJ : Lawrence Erlbaum. Shrum, L. J. & Bischak V. D. ( 2001 ) Mainstreaming, resonance, and impersonal impact: Testing moderators of the cultivation effect for estimates of crime risk Human Communication Research 27(2), 187 215. Shrum, L. J & O Guinn, T. C. (1993). Processes and effects in the construction of social reality: Construct accessibility as an explanatory variable Communication Research, 20(30), 436 471. Shiloh, S., & Ilan, S. (2005). To test or not to test? Moderators of the relationship between risk perceptions and interest in predictive genetic testing. Journal of Behavioral Medicine, 28(5), 467 479. Siegal, K., Raveis, V. H., & Gorey, E. (1998). Barriers and Pathways to Testing Among HIV Infected Women. A IDS Education and Prevention 10(2), 114 127. Siemer, M., & Reisenzein, R. (1998). Effects of mood on evaluative judgments: Influence of reduced processing capacity and mood salience. Cognition and Emotion, 12(6), 783 805. Slovic, P. (1987). Perception o f risk. Science, 236, 280 285. Slovic, P., Finucane, M. L., Peters, E., MacGregor, D. G. (2004). Risk as analysis and risk as f eelings: Some thoughts about affect, reason, and rationality. Risk Analysis, 24(2), 311 322.

PAGE 156

156 Sobel, M. E. (1982). Asymptotic c onfidence intervals for indirect effects in structural equation models. In S. Leinhardt (Ed.), Sociological Methodology (pp. 290 312). Washington DC: American Sociological Association. Strack, F., Schwarz, N., & Gschneidinger, E. (1985). Happiness and rem iniscing: The role of time perspective, affect, and mode of thinking. Journal of Personality and Social Psychology, 49(6), 1460 1469. Sumpradit, N., Fors, S. W., & McCormick, L. (2002). Consumers attitudes and behavior toward prescription drug advertisin g. American Journal of Health Behavior, 26(1), 68 75. Teinowitz, I. (2003). DTC ads become issue in race for democratic nomination. Advertising Age, 74(42), 8. Terzian, T. V. (1999). Direct to consumer prescription drug advertising. American Journal of L aw & Medicine, 25, 149 167. Tiedens, L. Z., & Linton, S. (2001). Judgment under emotional certainty and uncertainty: The effects of specific emotions on information processing. Journal of Personality and Social Psychology, 81, 973 988. Tversky, A. & Kahn eman D. (1973). Availability: A Heuristic for Judging F requency and Probability. Cognitive Psychology 5, 207 232. U.S. General Accounting Office. (2003). Prescription drugs: FDA oversight of direct to consumer drug advertising has limitations. Washington DC: Author. Verplanken, B., Hofstee, G., & Janssen, H. J. W. (1998). Accessibility of affective versus cognitive components of attitudes. European Journal of Social Psychology, 28, 23 35. Weiss, G. L., & Lonnquist. (1997). The sociology of health, heal ing, and illness. Upper Saddle River, NJ: Prentice Hall, Inc. Williams, S., & Voon, Y. W. W. (1999). The effects of mood on managerial risk perceptions: Exploring affect and the dimensions of risk. Journal of Social Psychology, 139, 268 287. Wilson, T. D. & Brekke, N. (1994). Mental contamination and mental correction: Unwanted influences on judgments and evaluations. Psychological Bulletin, 116(1), 117 142. Wolfe, S. M. (2002). Direct to consumer advertising: Education or emotion promotion? New England Journal of Medicine, 346, 524 526. Wright, W. F., & Bower, G. H. (1992). Mood effects on subjective probability assessments. Organizational Behavior and Human Decision Making Processes, 52, 276 291. Wyer, R. S., & Carlston, D. (1979). Social cognition, in ference, and attribution. Hillsdale, N.J.: Earlbaum.

PAGE 157

157 Wyer, R. S., & Srull, T. K. (1989). Memory and cognition in its social context. Hillside, NJ: Lawrence Erlbaum. Yeung, C. W. M., & Wyer Jr., R. S. (2004). Affect, appraisal, and consumer judgment. Jour nal of Consumer Research, 31, 412 424. Zachry III W. M., Shepherd, M. D., Hinich, M. J., Wilson, J. P., Brown, C. M., & Lawson, K. A. (2002). Relationship between direct to consumer advertising and physician diagnosing and prescribing. American Journal o f Health System Pharmacy, 59, 42 49. Zung, W. W. K. (1965). A Self Rating Depression Scale Archives of General Psychology, 12, 63 70.

PAGE 158

BIOGRAPHICAL SKETCH Jin Seong Park is a fifth year doctoral candidate specializing in advertising at the Universit y of Floridas College of Journalism and Communications. In August 2002, he received his masters degree in mass communications with emphasis on advertising from Marquette University in Milwaukee, Wisconsin. In February 2002, he was awarded his bachelors degree in journalism and mass communications from Korea University in Seoul, South Korea. As a graduate student, Jin Seong Park has presented a number of papers at conferences, including the annual conventions of the Association for Education in Journa lism and Mass Communication, American Academy of Advertising, and International Communication Association. After graduation, he plans to move to Philadelphia, Pennsylvania, where he will teach Introduction to Advertising, Advertising Media Planning, and Qu antitative Advertising Research as an assistant professor at Temple Universitys Department of Advertising.