Cause Related Marketing: Understanding Celebrity, Consumer, and Product Fit Relationships by Whitney Skinner Nunnelley A dissertation submitted to the Graduate Faculty of Auburn University in partial fulfillment of the requirements for the Degree of Doctor of Philosophy Auburn, Alabama December 8, 2012 Keywords: Celebrity endorsement, charitable cause, self-concept, congruence Copyright 2012 by Whitney Skinner Nunnelley Approved by Carol Warfield, Co-Chair, Professor & Head, Consumer Affairs Veena Chattaraman, Co-Chair, Associate Professor, Consumer Affairs Lenda Jo Connell, Professor, Consumer Affairs Tsailu Liu, Associate Professor, Industrial Design ii Abstract The use of celebrity product endorsers in charitable cause marketing campaigns is increasing as charitable causes seek to establish market share among for-profit alternatives. Cause related marketing (CRM) studies show consumers (CONSR) give greater attention and financial support to campaigns when they perceive fit among the celebrities (CE), the endorsed charitable causes (ECC), and the endorsed products (EP). Previous CRM studies have considered these individual fit relationships but have not examined the holistic effects of fit between (a) CONSR-CE; (b) CONSR-ECC; (c) CONSR-EP; (d) CE-ECC; and (e) CE-EP on the success of CRM campaigns. This study proposed a conceptual model to explain how these relationships influence consumer attitudes and purchase intentions for Celebrity Endorsed Charitable Cause Products (CECCP). Specifically, it was predicted that consumer attitudes toward CECCP would be positively influenced by greater fit between: CE and EP (H1); CE and CONSR (H3, H4); and CE and ECC (H5). Furthermore, CONSR-EP fit (H2) and CONSR-ECC fit (H6) are expected to produce moderating effects on consumer attitude. Positive consumer attitude is predicted to positively influence purchase intention for a CECCP (H7, H8). Predicted effects were tested using a 2 (CE-ECC Fit: Low/High) x 2 (CE-CONSR Fit: Low/High) x 2 (CE-EP Fit: Low/High) between-subjects factorial experiment. Female millennials (ages 19 to 28) participated in online surveys that considered one of eight experimental conditions. The 411 subjects were provided a vignette outlining a female celebrity endorsement of a charitable cause product available for purchase. The celebrities, causes, and products employed were chosen based on the results of two pretests. iii Manipulation checks on the celebrity fit variables CE-EP fit, CE-CONSR fit, and CE- ECC fit were not confirmed. In light of this result, simple linear regression and ANCOVA were used to test the conceptual model hypotheses with the measured rather than the manipulated variables. Contrary to the hypotheses, the results suggest consumer attitudes toward CECCP were not influenced by the hypothesized celebrity fit relationships (CE-EP fit, CE-CONSR fit, and CE-ECC fit). Consumer fit relationships (CONSR-EP fit and CONSR-ECC fit) were hypothesized to moderate the influence of CE-EP fit and CE-ECC fit on consumer attitudes in the model, and this was partially supported. It was also found that that the consumer fit variables (CONSR-EP fit and CONSR-ECC fit) had a significant direct influence on consumer attitude and purchase intent. Next, the relative influence of consumer attitude toward an endorsed product and consumer attitude toward an endorsed charitable cause on purchase intention was measured using multiple regression analysis. Results indicated that consumer attitude toward an endorsed product had a greater influence on consumers? purchase intention for a CECCP than consumer attitude toward an endorsed charitable cause. Following the testing of the hypotheses, additional analyses were conducted to examine relationships not hypothesized in the original model, to explain the potential influence of unexplored fit relationships on consumer attitude and purchase intention. The results of regression analysis of celebrity fit factors on purchase intention revealed CE-EP fit, CE-CONSR fit, and CE-ECC fit were significant in influencing purchase intention directly. Although the results revealed CE-ECC fit to be negatively correlated with PI, together, the fit variables help explain 29 percent of the variance in purchase intention. However, Consumer fit factors had greater influence on purchase intention by explaining 39.6 percent of the variance. These results iv confirm their importance in directly explaining purchase intentions toward celebrity endorsed charitable cause products. Although there were limitations in this study, the findings offer implications that augment the literature concerning celebrity endorsers, CRM, Theory of Planned Behavior (TPB), and the Theory of Consuming Behavior. By building on the significant and non-significant results of this study, a revised conceptual model is proposed that posits the influence of three groups of factors -- consumer fit factors (CONSR-EP fit and CONSR-ECC fit), celebrity fit factors (CE-EP fit and CE-ECC fit) and celebrity-consumer fit factor on purchase intentions toward CECCP. The results also present insights into the theoretical framework used to explain the interaction between celebrity endorsers, symbolic goods, and consumers. The adapted Model of Consuming Behavior illustrates the role of the celebrity endorsers among a mix of consumer and symbolic goods. Female millennial response in this study is congruent with Cone, Inc (2010a) studies that suggest millennials are a powerful consumer group with awareness of marketing techniques. Specifically, those with previous donations to charitable causes may be more responsive to CRM campaigns. With additional analysis, the causal relationships between celebrity endorsers, consumers, and symbolic goods can aid marketers in identifying the proper fit between the endorser, the cause, the product, and the target market. v Acknowledgments I began the words of acknowledgment in my Master?s Thesis by offering gratitude to those who joined me and influenced me on my educational career adventure. Many of those individuals have continued with me and I would like to extend my gratitude to all of those who have made this dissertation possible. To my co-advisors, Dr. Veena Chattaraman and Dr. Carol Warfield, thank you for your generosity over the past six years. I will forever be thankful for the amount of time and effort you spent mentoring me and inspiring me to be my best. In my times of hardship and times of joy, your care, support, and positive encouragement meant all the difference. To my committee member, Dr. Lenda Jo Connell, thank you for your generosity of spirit and encouraging words. To my committee member and former Thesis chair, Tsai ?Lu? Liu, thank you for guiding me by example these past nine years. I have appreciated your insight, advice, and encouragement regarding my career and life. And to my former Thesis committee member, Bret Smith, thank you for being approachable these past ten years and for giving me your time and consideration. I would especially like to thank my parents and family. I?m certain I could not have completed this adventure without you. Without this adventure, I might not have met the love of my life, my husband, Michael. To him, I am most thankful for sacrificing his own dreams so that I could have mine. Your unconditional love and support sustain me. But most importantly, should I ever forget, your edits to this paper are invaluable. vi Table of Contents Abstract ......................................................................................................................................... ii Acknowledgments......................................................................................................................... v List of Tables ................................................................................................................................ x List of Figures ............................................................................................................................. xii Style Manual .............................................................................................................................. xiii Chapter 1: Introduction ............................................................................................................... 1 The Evolution of Cause Related Marketing .................................................................... 1 The Role of the Celebrity Endorser ................................................................................ 3 Purpose .............................................................................................................................. 5 Problem Statement ............................................................................................................ 5 Significance of Study ........................................................................................................ 6 Definition of Key Terms ................................................................................................... 7 Chapter 2: Literature Review ........................................................................................................ 9 The Increasing Prevalence of Celebrity Product Endorsement ........................................ 9 The Attributes of Successful Celebrity Endorsers .......................................................... 10 The Relationship Between the Celebrity and the Product .............................................. 12 The Role of Consumer Motivations ................................................................................ 14 An Evolving Understanding of Charitable Causes as Brands ........................................ 15 vii Conceptual Framework ................................................................................................... 17 The Theory of Symbolic Consumption........................................................................... 17 Conceptual Model .......................................................................................................... 22 Hypothesis Development ................................................................................................ 24 Celebrity endorser and endorsed product fit (CE-EP) .............................................. 24 Match-up hypothesis (MUH) .............................................................................. 24 Consumer and endorsed product fit (CONSR-EP) ................................................... 27 Self-concept ........................................................................................................ 27 Celebrity endorser and consumer fit (CE-CONSR) .................................................. 30 Parasocial relationships ....................................................................................... 31 Celebrity endorser and endorsed charitable cause fit (CE-ECC) .............................. 34 Consumer and endorsed charitable cause fit (C-ECC) ............................................. 36 Attitude and purchase intention ................................................................................ 37 Theory of planned behavior ................................................................................ 38 Charity brand literature ....................................................................................... 39 Chapter 3: Methodology ........................................................................................................... 40 Research Design.............................................................................................................. 40 Stimulus Development .................................................................................................... 41 Pretest 1 ..................................................................................................................... 41 Instrument ................................................................................................................. 42 Sampling and data collection procedure ................................................................... 45 Data Analysis ............................................................................................................ 46 Celebrity endorser ? consumer fit (CE-CONSR) ..................................................... 46 viii Pretest 2 ..................................................................................................................... 51 Instrument ................................................................................................................. 51 Sampling and data collection procedure ................................................................... 52 Data analysis ............................................................................................................. 53 Main Study ...................................................................................................................... 57 Instrument ................................................................................................................. 57 Sampling procedure and characteristics.................................................................... 58 Dependent measures ................................................................................................. 59 Manipulation check ................................................................................................... 61 Celebrity endorser ? endorsed product fit (CE-EP), celebrity endorser ? endorsed charitable cause fit (CE-ECC). ..................... 61 Celebrity endorser ? consumer fit (CE-CONSR) ............................................... 61 Moderating variables ................................................................................................ 63 Consumer? endorsed product fit (CONSR-EP), consumer ? endorsed charitable cause fit (CONSR-ECC) ........................... 63 Demographic items ................................................................................................... 64 Chapter 4: Results ..................................................................................................................... 66 Main Study ...................................................................................................................... 66 Sample description .................................................................................................... 66 Scale reliability ......................................................................................................... 68 Manipulation checks ................................................................................................. 69 Hypothesis testing ..................................................................................................... 71 Additional analysis.................................................................................................... 74 Chapter 5: Discussion and Conclusions .................................................................................... 79 ix Consumer Fit Variables .................................................................................................. 79 Consumer ? endorsed product fit (CONSR-EP), consumer ? endorsed charitable cause fit (CONSR-ECC) ................................. 79 Celebrity Fit Variables .................................................................................................... 81 Celebrity endorser ? endorsed product fit (CE-EP), celebrity endorser ? endorsed charitable cause fit (CE-ECC) ............................ 82 Celebrity endorser ? consumer fit (CE-CONSR) ..................................................... 83 Consumer Attitude and Purchase Intentions ................................................................... 84 Discussion of Non-Hypothesized Significant Findings .................................................. 85 Limitations ...................................................................................................................... 86 Theoretical and Managerial Implications ....................................................................... 87 Suggestions for Future Research .................................................................................... 91 References ................................................................................................................................. 93 Appendices ............................................................................................................................... 101 A. Defined Abbreviations ............................................................................................. 101 B. Pretest Information Letter ........................................................................................ 103 C. Pretest 1 Online Survey Questions ........................................................................... 106 D. Pretest 2 Online Survey Questions........................................................................... 114 E. Main Study Stimuli................................................................................................... 119 F. Main Study Information Letter ................................................................................. 121 G. Main Study Scenario Example ................................................................................. 124 H. Main Study Online Survey Questions ...................................................................... 126 I. Demographic Survey Questions ................................................................................ 135 J. Main Study Debriefing Page and Optional Email Submission ................................ 138 x List of Tables Table 1 Definitions of Key Terms Used Throughout Text ........................................................... 7 Table 2 Explanation of Variables Used in Adapted Model of Consuming Behavior ................. 20 Table 3 Pretest 1 Sample Characteristics and Frequency Distribution ....................................... 45 Table 4 Celebrity Endorser Mean Scores ................................................................................... 47 Table 5 Paired T-Tests: Celebrity Relevance ............................................................................. 48 Table 6 Charitable Causes Mean Scores ..................................................................................... 49 Table 7 Paired T-Tests: Charitable Cause Relevance ................................................................ 50 Table 8 Pretest 2 Sample Characteristics and Frequency Distribution ....................................... 53 Table 9 Celebrity Endorser & Charitable Cause Fit Criteria ...................................................... 53 Table 10 Paired T-Tests: Celebrity Endorser ? Charitable Cause .............................................. 55 Table 11 Celebrity Endorser ? Product Mean Scores ................................................................. 56 Table 12 Manipulated Variable Conditions for Factorial Design ............................................... 57 Table 13 Consumer Attitude and Purchase Intention Measurement Items Included in Main Study ............................................................................................... 60 Table 14 Celebrity Endorser Fit Measurement Items Included in Main Study .......................... 62 Table 15 Consumer Fit Measurement Items Included in Main Study ........................................ 64 Table 16 Demographic and Donation Behavior Measurement Items Included in Main Study ............................................................................................... 65 Table 17 Main Study Sample Characteristics and Frequency Distribution ................................ 67 Table 18 Main Study Scale Reliability ....................................................................................... 68 xi Table 19 Main Study Manipulation Check ................................................................................. 69 Table 20 Descriptive Statistics for Manipulation Check Measure for CE-CONSR Fit ............. 70 Table 21 Supported/Not Supported Hypotheses ......................................................................... 71 Table 22 Defined Abbreviations .............................................................................................. 102 Table 23 Main Study Stimuli by Condition .............................................................................. 120 xii List of Figures Figure 1 Grubb and Grathwohl 1967 Model of Consuming Behavior ...................................... 19 Figure 2 Adapted Model of Consuming Behavior ...................................................................... 21 Figure 3 Consumer response to celebrity endorsement of charitable cause product conceptual model ........................................................................................................ 23 Figure 4 Theory of Planned Behavior (Ajzen, 1991) ............................................................... 38 Figure 5 Research process ........................................................................................................ 41 Figure 6 Modified conceptual model for future research ........................................................... 75 xiii Style manual or journal used: Publication Manual of the American Psychological Association (7th edition) Computer software used: Adobe Photoshop CS5, Microsoft Word, and SPSS 17 for Windows 1 Chapter 1: Introduction This study examined how Cause Related Marketing (CRM) by celebrities influences consumer response to celebrity endorsed charitable products and influences the image of both the endorsing celebrities and the charities themselves. CRM relies upon charitable causes or otherwise worthwhile projects as a marketing strategy for promoting consumers? purchases of products or services (Fromherz, 2006). This study was designed to explore whether consumers are more likely to have positive attitudes and purchase intentions for celebrity endorsed cause products when fit is established among celebrity, consumer, cause, and product. The Evolution of Cause Related Marketing The earliest example of CRM discussed in the literature relates to the 1983 American Express campaign associated with the restoration of the Statue of Liberty in New York; American Express advertised that it would donate money to the restoration project each time consumers conducted a transaction with its credit cards or initiated a new credit card membership. The American Express campaign raised over one million dollars for the restoration project and was a forerunner to the increasing use and sophistication of CRM strategies (Josephson, 1984). Today, CRM is pervasive in the corporate world. Examples of prominent corporate-cause associations include Target with St. Jude Children?s Research Hospital; McDonalds with the Ronald McDonald House Charities; Fuse Drinks (among numerous other companies) with 2 the Susan G. Komen Foundation; and the Dallas Cowboys with the Salvation Army?s national Red Kettle Kickoff (Jones, 2010 a, b, c, d). Some CRM strategies create an association between a charity and a topically relevant product. For example, Nabisco has adopted a CRM campaign that links its animal cookies with support for the World Wildlife Fund (Strahilevitz, & Myers, 1998). Research suggests that CRM campaigns have the potential to be beneficial for both the corporation and the associated cause. Young adult consumers are interested in well-designed products that have an effective marketing campaign linked to an attractive cause (Rosenthal, 2010). According to the 2010 Cone Nonprofit Marketing Trend Tracker, 59 percent of all Americans are more likely to purchase a product if it is associated with a charitable cause (Cone, Inc., 2010b). Further, the Cone study found that when a corporate-sponsored cause becomes familiar to consumers, 56 percent of consumers are more willing to support it financially (Cone, Inc., 2010b). Charitable causes may also benefit from brand logos and other marketing elements arising from the CRM campaign that could allow them to differentiate themselves from the competition. Colors, logos, and symbols are considered important in charity advertising by 79 percent of young adults between the ages of 18 and 24 (Cone, Inc., 2010b; Jekielek & Brown, 2005). Building upon CRM techniques, charities sometimes even position themselves as a product brand. John Grounds, the Director of Communications for the National Society for the Prevention of Cruelty to Children explained that the idea of referring to charities as brands is relatively new to most people. However, charities utilize essential branding strategies in their every-day communications, product offerings, and fundraising techniques (Jekielek & Brown, 3 2005). By logical extension, charities are capable of ?extending? their brand to products without the need to utilize a separate established (for-profit) brand for marketing purposes. Feed Projects, a public nonprofit charity, is a charity that has wholly embraced the merger of charity and brand, offering trendy tote bags for sale that, in turn, benefit charities such as the World Food Program and UNICEF (The Feed Foundation, 2011). Feed Projects has eliminated the typical charity-brand linkage by establishing itself (the charity) as the brand and selling its products through numerous channels, including Whole Foods, Amazon.com, or Ralph Lauren. Regardless of the merchant, the product and associated message remains consistent (Finch, 2010). By selling its products at multiple retailers and leveraging various product branding strategies including celebrity endorsement, Feed Projects has succeeded in appealing to many consumer groups and achieved projected revenue of $4.5 million in 2010 (Fenn, 2010). The Role of the Celebrity Endorser As CRM techniques have grown and evolved, so has the use of celebrity endorsements in CRM campaigns. St. Jude?s Children?s Research Hospital has a long record of celebrities who appear on its behalf, like Jennifer Anniston and Robin Williams. Feed Projects also boasts endorsement by several celebrities, including Toby McGuire, Marcia Cross, Moby, and founder Lauren Bush (coolspotters.com, 2011a). These endorsements may be traditional?the celebrity appears in an advertisement for a charity or an associated product?or informal?the celebrity mentions a charity in an interview or is photographed using an associated product in public. Celebrity endorsements can be made by ?any individual who enjoys public recognition and who uses this recognition on behalf of a consumer good by appearing with it in an advertisement? (McCracken, 1989, p. 310). As a marketing strategy, companies invest large sums of money to align the product, the company, and the celebrity in hopes of generating 4 sufficient sales to achieve a profitable outcome (Edrogan, 1999). Characteristics of celebrities that affect consumer attitudes and purchase intentions include attractiveness, credibility, trustworthiness, and the extent to which a celebrity is perceived to be an expert relative to the product they endorse (Friedman & Friedman, 1979; Miller & Baseheart, 1969; Ohanian, 1990). Ideal product endorsement occurs when the celebrity and the product endorsed are a good fit (Kahle & Homer, 1985). The effectiveness (or ineffectiveness) of a celebrity product endorsement is related to how consumers view the products they buy. Grubb and Grathwohl (1967) examined goods as symbols and the consumption of these symbols as a way of reinforcing the self-concept and communicating the self-concept to others. Similarly, McCracken (1989) asserted that consumers fill their lives with the products and items that fit, complete, and fulfill their needs. Consumers may fit their lifestyles to that of a celebrity by adopting the celebrity?s attitudes and behaviors (McCracken, 1989). Celebrities who endorse charities are often seen as taking part in elite-challenging political action such as charity work because the celebrity demands a higher standard of life and applies this to their charity work (Inglehart, 1977). Those who seek to enhance their own self- concept are prone to be influenced by postmaterialism, defined as the values, self-expression, and desire to belong to a community (Giacalone, Jurkiewicz & Deckop, 2008). George Clooney, who works on behalf of the people of Darfur, is an example of a postmaterialist who appeals to the UN Security Council for aid and support (de Waal, 2008). Consumers who adopt a celebrity?s postmaterialist world view may be particularly prone to consume celebrity-endorsed cause products because such products are congruent with the desire to belong to a community and provide enhanced opportunities to reinforce and communicate self-concept. 5 Purpose The purpose of this study was to examine the fit relationships among consumers, causes, products, and celebrities who endorse charitable causes and products, and the influence of these fit relationships on consumer attitudes and purchase intent towards celebrity endorsed cause products. This study examined the attitude of consumers toward the product and charitable cause separately in order to understand their relative influence on consumers? purchase intentions for celebrity endorsed cause product. Fit relationships between celebrities and products were examined using a Match-Up Hypothesis (MUH) framework (Baker & Churchill, 1977; Kahle & Homer, 1985; Kamins, 1990; Kamins & Gupta, 1994). Celebrity endorsers were selected for equal attractiveness in an attempt to control for this factor?s influence on consumer perceived fit with a product and/or a charitable cause. Fit between a celebrity and a consumer was examined using a measure of celebrity attitude that examined consumers? parasocial relationships with celebrities by identifying the intensity of the consumers? relationship with the celebrity. Grubb and Grathwohl?s (1967) Model of Consuming Behavior contributed to an understanding of how self-concept measures establish the fit among the consumer, charitable cause, and endorsed product. Problem Statement Previous CRM studies have focused on the traditional relationship between a nonprofit and a brand, the relationship between a nonprofit and its own licensed products has not been explored. For example, Feed Projects feature the World Food Program and/or UNICEF on the products marketed, and Invisible Children features Invisible Children on its t-shirts. Prior CRM studies that examined charity products have not examined the symbolic nature of the product and 6 the charity, separate from their effects on the interaction with consumers and celebrities. As distinct symbolic goods, charities and charitable cause products should be examined individually for possible interaction with both consumers and celebrities. No CRM studies have examined holistically the fit relationships between: a) consumers and celebrity endorsers, b) consumers and endorsed charitable causes, c) consumers and endorsed products, d) celebrity endorsers and endorsed charitable causes, and e) celebrity endorsers and endorsed products and their influence on consumer attitudes and purchase intent toward celebrity endorsed cause products. A theoretical framework that supports the examination of these fit relationships is Rogers (1951) Theory of Self-Enhancement, which posits that individuals use symbolic goods and symbols to communicate to others as a way to enhance their own self-concept. Significance of Study Studying the relationships among celebrities, consumers, charitable causes and products will contribute to the CRM literature and have potential actionable marketing implications. Additionally, the conceptual model expands the Grubb and Grathwohl (1967) Model of Consuming Behavior by considering multiple relationships with a symbolic good or service within a single model. Previous celebrity endorsement literature has not applied this model to express the relationships between consumers and goods, thus the present study will fill this gap in CRM literature. This study will also expand the area of parasocial literature that uses the CAS scale to examine the intensity of celebrity worship among consumers. By isolating the Intense-Personal sub-scale?which is believed to be the most relevant to consumer-celebrity relationships? additional understanding of female millennials? parasocial relationship with celebrities will be gained (North & Sheridan, 2009). Formal studies have not previously examined the millennial 7 consumers? relative attitude toward celebrity endorsed products or a celebrity endorsed charitable cause. Finally, this study will contribute to the literature concerning the Theory of Planned Behavior (Ajzen, 1991), adding to the understanding of how female millennials? attitudes influence purchase intention behavior for a celebrity endorsed charitable cause product. Definition of key terms can be found in Table 1. Table 1 Definitions of Key Terms Used Throughout Text Brand: A brand is a label that carries meanings and associations. Brands are built by the utilization of brand tools including public relations (PR), advertising, events, sponsorships, spokespersons, and social causes, among others, to help differentiate the branded product or service among competitors (Kotler 2003). Celebrity Endorser : A celebrity endorser is ?any individual who enjoys public recognition and who uses this recognition on behalf of a consumer good by appearing with it in an advertisement? (McCracken, 1989, p. 310). Celebrity Endorser - Consumer Fit: Fit between the celebrity and consumer is established when the consumer identifies with and has positive attitude for the celebrity. Based on the concept of parasocial relationships, celebrities achieve a high degree of fit with the consumer when the consumer has positive intense/personal celebrity worship (Hung, Chan, & Tse, 2011). Celebrity Endorser ? Endorsed Product Fit: Based on the Match-Up-Hypothesis, a high degree of fit between an attractive celebrity and a product will be achieved when the product is congruent with the celebrity?s image and when the celebrity is perceived to hold values similar to the targeted consumer market (Kahle & Homer, 1985; Kamins & Gupta, 1994). Celebrity Endorser ? Endorsed Charitable Cause Fit: Based on the Match-Up-Hypothesis, celebrities will achieve cause fit in a similar manner to product fit; that is, a high degree of fit between an attractive celebrity and a charitable cause will be achieved when a charitable cause has symbolic traits and values that are congruent with the celebrity?s image. Charity: Charity is defined as ?what we give to alleviate the need, suffering, and sorrow of others, whether we know them or not? (Bremmer, R.H., 1994, p10). Consumer- Endorsed Charitable Cause Fit: Fit between the consumer and cause refers to how the consumer?s image and values align with a charitable cause?s image and values. 8 Table 1 continued Consumer- Endorsed Product Fit: Fit between the consumer and product refers to how congruent a consumer?s product evaluations (i.e., attitude formation and purchase intention) are with the consumers? self-image (Birdwell, 1968). Perceived Fit: The degree to which one variable is perceived to be congruent with another variable. Self-Concept: Self-Concept is defined as an individual?s perceptions of oneself (Sirgy et al., 1997). Self- Enhancement: Occurs when the individual receives responses and recognition from influential persons about him or herself which are perceived to strengthen the self-concept (Rogers, 1951). Self-Image: Born out of the concept of brand image, this reflects the individual?s physical being as well as evaluations and a descriptive definition of the individual and the individual?s actions. Self-Image influences the goods and services the individual consumes, which is why it has been studied extensively in regard to consumer behavior (Birdwell, 1968). 9 Chapter 2: Literature Review The relevant literature in support of a conceptual model that applies the symbolic consumption behavior model is discussed below. Following that, a model is proposed to explain celebrity-consumer-product relationships that influence attitudes and purchase intentions for Celebrity Endorsed Charitable Cause Products (CECCP). Supporting literature is presented for hypotheses proposed in the conceptual model. The Increasing Prevalence of Celebrity Product Endorsement The use of famous, attention-getting stars is an important strategy in product marketing and reaching the target consumer. Marketing communication strategies are designed to differentiate elements such as product design, packaging, and branding in the eyes of consumers; the use of celebrities is considered an effective way to achieve the desired differentiation (Erdogan, 1999). Celebrities bring a certain appeal to advertisements that increase the chances of enhancing the consumers? attentiveness to the advertisement (Spielman, 1981). Accordingly, celebrity endorsement has been found to be effective in making consumers more likely to recall both the advertisement and the brand-name (Friedman & Friedman, 1979). McCracken (1989) defined a celebrity endorser as ?any individual who enjoys public recognition and who uses this recognition on behalf of a consumer good by appearing with it in an advertisement? (p. 310). Though obviously the use of celebrities by companies dates back well before the advent of television or even radio, perhaps the most influential endorsement deal in recent history was Michael Jackson?s endorsement of Pepsi beverages, which began in November of 1984. It 10 shattered previous celebrity endorsement deals, both in terms of the size of Jackson?s contract and the return on that investment for Pepsi (Herrera, 2009). Multiple well-publicized incidents occurred which seemed to be harmful to the images of Jackson and/or Pepsi (i.e., an ad campaign taping resulting in Jackson?s hair catching fire and Jackson?s child molestations charges). Yet despite, or perhaps even because of these incidents, Pepsi sales increased eight percent during the time Jackson endorsed it (Erdogan, 1999; Gabor, Thorton, & Wienner, 1987). The success of the Jackson deal set the stage for Pepsi?s continued use of celebrities, particularly from the music industry, to endorse its products (Herrera, 2009). Lionel Richie, Madonna, the Spice Girls, Britney Spears, and Beyonc? have all been recruited by Pepsi for their ability to generate sales with Pepsi?s target market (Erdogan, 1999; Herrera, 2009). This marketing strategy continues to reach consumers today. Erdogan?s (1999) extensive literature review, citing data collected by Howard (1979), Motavalli (1988), and Shimp (1997), noted an increase in the use of celebrities from 1979 through 1997. Recent examples of prominent and successful celebrity-product endorsement relationships include the Black Eyed Peas for Target stores; Thandi Newton for Olay beauty products; and Catherine Zeta-Jones and Luke Wilson for competing mobile phone carriers T-Mobile and AT&T, respectively (Beer, 2010; Herrera, 2009; LaVallee, 2009; Target, 2009; T-Mobile USA, 2005). The Attributes of Successful Celebrity Endorsers Numerous theories have been offered that attempt to identify the most important criteria for a successful endorser-product relationship. The research in this area generally identifies trustworthiness, credibility, expertise, and attractiveness as qualities that can stimulate consumer attitude change. It should be noted, however, that the definitions applied to these terms may vary in significant ways from one study to the next, making it difficult to compare the findings of two 11 studies. Similarly, some studies employ specialized terms that may be related to or even overlap with one or more of the four categories identified above. Several studies emphasize the closely-related concepts of trustworthiness and source credibility. The dimension of trustworthiness is reflected in the audiences? degree of confidence in and acceptance of the celebrity?s message (Hovland, Irving, & Harold, 1953). Source credibility, according to Ohanian (1990), is ?a communicator?s positive characteristics that affect the receiver?s acceptance of a message? (p. 41). Miller and Baseheart?s (1969) study was an early effort to examine trustworthiness and the effect that persuasive messages containing opinionated versus non-opinionated statements have on attitude change. Their work indicated that when the source?s (endorser?s) trustworthiness was low, an opinionated message was ineffective in producing attitude change irrespective of the content of the message. Expertise was defined by Friedman and Friedman (1979) as knowledge obtained from training, experience, or study. They emphasized that expertise is not a trait of the endorser that is applicable to all product areas, but instead is product specific. Their research found that expertise has the ability to produce a more positive overall attitude, stronger intent to purchase, and enhanced credibility for the endorser (Friedman & Friedman, 1979). As defined by Pornpitakpan (2003), expertise is evaluated based on the extent to which the celebrity?s message is perceived to be based upon valid assertions. Pornpitakpan asserted that the dimensions of trustworthiness and expertise are the two criteria that relate most to a celebrity endorsers? credibility. The attractiveness of the endorser has been found to be an important factor in the success of a marketing campaign. Friedman and Friedman (1979) found that where attitude change toward a product is desired, the use of a generally attractive and likable celebrity endorser offers an advantage. However, attractiveness, as well as the other factors discussed above, is 12 inextricably connected to the relevance of the product to the endorser. Friedman and Friedman (1979) point out that celebrity-product match-up should be carefully considered because not all product and celebrity endorsements have the potential to produce enhanced trust in the celebrity endorser, stronger consumer purchase intention, or an overall positive consumer attitude for the endorser. Pornpitakpan (2003) emphasized the importance of the attractiveness of the endorser to the specific market audience for the product. Accordingly, Pornpitakpan employed a theory of ?source attractiveness? that identified four traits communicated by the celebrity in a message that can be evaluated. These traits have been found to correlate with the effectiveness of the advertisement in selling a product or service: familiarity, likability, similarity and attractiveness. This model is not unlike the Match-Up-Hypothesis (MUH) by Baker and Churchill (1977), which emphasized the importance of matching a celebrity?s attractiveness to a product that is congruent with the celebrity?s image. The Relationship Between the Celebrity and the Product The fit between product and celebrity is believed to be a determining factor for an effective celebrity endorsed ad campaign (Kahle & Homer, 1985). According to McCracken (1986), celebrities are selected to endorse products and services because the celebrities portray certain meanings. These meanings are gathered from their age, gender, culture, lifestyle, and other psychographic information. Collectively, the gathered meanings contribute to a celebrity?s status and reputation. The success or failure of a celebrity endorsement campaign depends not only on how trustworthy or well-liked a celebrity is, but how well the ?meanings? portrayed by the celebrity translate to the product to be endorsed (McCracken, 1986). Many researchers have proposed methods to ensure congruency, or fit, between celebrities? endorsements with products. 13 Lee and Thorson (2008) attributed incongruity to the mismatch of the image of the celebrity with the product. This prevents the consumer from forming any meaningful link between the celebrity endorser and the product. However, when the celebrity and the product are congruent, consumers have enhanced recall and affect toward the celebrity and brand/product, resulting in enhanced consumer evaluations and sales (Misra & Beatty, 1990). Mittelstaedt, Riesz and Burns (2000) confirmed in their study that for a successful endorsement to occur there must be an appropriate match between endorser and product. Unfortunately, incongruity between product and celebrity endorsers is difficult to predict, and often does not come to light until after a campaign has failed. For example, Bill Cosby was a successful endorser for Jell-O, but was unsuccessful in achieving a congruent celebrity-product relationship with E. F Hutton financial services (Mittelstaedt et al, 2000). Conversely, Michael Jordan is an example of celebrity spokesperson who was been able to succeed with a wide range of endorsement deals for a variety of products, including food products by Coke, Gatorade, Wheaties, and McDonald?s; shoes and apparel by Nike and Hanes; and communications services by WorldCom (Erdogan, Baker, & Tagg, 2001). That said, the literature suggests that some categories of products are more likely than others to be ?congruent? with the images of celebrities, irrespective of the particular celebrity chosen; thus such products are generally likely to benefit from celebrity endorsement. McCracken (1989) asserted that consumers are constantly looking for goods that provide meaning to them as a way to ?furnish? the self. In other words, consumers look to products that fit them. Petty, Cacioppo, and Schumman (1983) stated that when a product holds personal relevance to a consumer, it can increase a person?s motivation to form an opinion regarding the product. Lee and Thorson (2008) hypothesized that high product involvement moderates 14 celebrity product endorsement congruity due to the consumers? willingness to already be highly involved with the product category. Friedman and Friedman (1979) have stated that celebrity endorsers are an appropriate match for products with a high level of consumer perceived social or psychological risk. For example, consumers are more likely to be influenced by celebrity endorsement where their consumption choices will be judged by a group of peers (i.e., peer group acceptance). For their study, Friedman and Friedman chose three products that would challenge the participants? levels of perceived risk: (1) costume jewelry for social/psychological risk; (2) a vacuum cleaner for financial, performance, and or physical risk; and (3) a box of cookies for generally low levels of perceived risk in all areas (Friedman & Friedman, 1979). The study found that consumer preference for the costume jewelry was most subject to influence by celebrity endorsement. Friedman and Friedman asserted that consumers internalize or try to adopt the behaviors and attitudes of the endorsers of products they use. Internalization is especially relevant to consumer?s decisions as to luxury products such as jewelry. Everyday products, such as toilet paper or toothpicks, are better suited for a typical person endorsement rather than a celebrity endorsement (Friedman & Friedman, 1979). The Role of Consumer Motivations A successful CRM campaign must consider the motivations and values of a product?s target consumers, specifically their tendencies toward socially responsible actions. Giacalone et al. (2008) asserted that consumers may be divided into two groups: materialists and postmaterialists. Materialists are those who possess values of the industrial world including economic rewards and a focus on the self. A postmaterialist, in contrast, possesses values of considering future generations, having a greater sense of belonging to a community, and making 15 socially ethical decisions (Giacalone et al., 2008). In Giacalone et al?s (2008) study, the authors found that individuals with a low sense of materialism and an increase towards postmaterialism had a greater sensitivity toward corporate socially responsible actions. An expression of postmaterialist values is found in the movement calling for Fair Trade (FT) practices. Consumers in this movement, in turn, express a preference for purchasing FT products. In the European coffee market, FT coffee has been the only product to have sales growth and growth in its market segment (Cailleba & Casteran, 2010). Cailleba and Casteran found that consumers of FT coffee were likely to be what Giacolone et al. (2008) defined as postmaterialists. These are consumers who choose to consume FT coffee because of the ethical reward and the contribution towards future generations that FT consumption symbolizes to them. An Evolving Understanding of Charitable Causes as Brands Charity has been defined as ?what we give to alleviate the need, suffering, and sorrow of others, whether we know them or not? (Bremmer, 1994, p. 10). Bremmer?s book ?Giving? provides a historical account of charity and philanthropy, beginning with charity?s foundations in acts of direct kindness to those in need under the tenets of Christianity, Judaism and Islam. However, in today?s society, charity in the United States is most commonly demonstrated by writing and sending a check to the public charity the consumer has chosen to support (Bremmer, 1994). According to the Urban Institute?s National Center for Charitable Statistics (NCCS), in 2006 there were 1.5 million nonprofit organizations registered with the Internal Revenue Service (Urban Institute?s National Center for Charitable Statistics, 2008). The NCCS divides nonprofits into three groups of 501(c)(3) IRS tax designations: public charities, private foundations, and other tax-exempt organizations. Most of the public charities?nearly one million?are affiliated 16 with the arts, education, health care, or religious organizations. Examples of well-known public charities include National Public Radio, Teach for America, American Red Cross, and Fellowship of Christian Athletes. These public charities receive donations totaling approximately 98.8 million dollars per year in cash, goods, and services (Barton & Hall, 2010). By comparison, the entire market for retail clothing in the U.S. has a combined annual revenue of 150 billion dollars (www.hoovers.com, 2011). Charities and social causes are an important part of modern society. As both charities and causes have gained a better understanding of their potential supporters, they increasingly have been exploring branding and product marketing techniques in an effort to gain attention and financial support for their missions. This has inevitably led to a proliferation of logos and cause- related products that compete for consumer attention. John Grounds, the Director of Communications for National Society for the Prevention of Cruelty to Children, authored an editorial in the International Journal of Nonprofit & Voluntary Sector Marketing (IJNVSM) discussing charity in terms of branding (Jekielek & Brown, 2005). A brand is defined by Philip Kotler as a label that carries meanings and associations (2003). Brands are built by the utilization of brand tools including public relations, advertising, events, sponsorships, spokespersons, and social causes, among others, to help differentiate the branded product or service among competitors (Kotler, 2003). Charities and causes are increasingly utilizing branding and marketing strategies. Fair Trade USA offers a certification process and a logo that can be placed on certified products for purchase. According to a study by Lightspeed Research, 50 percent of respondents said a logo, such as the Fair Trade Certification logo, would influence their purchase (Mortimore, 2010). Rosenthal (2010) argued that well-designed products with cause linkages that employ effective 17 marketing strategies are catching brand-savvy consumers? attention because the products are perceived as cool. Products that Rosenthal identified as gaining market share are the FEED bags?cloth totes that that consumers can purchase to feed a child in Africa for one year?and Invisible Children T-shirts?which identify the wearer as part of the movement to help rebuild war-torn schools in Uganda and end child soldier servitude (Rosenthal, 2010). According to Rosenthal (2010), charities are beginning to understand young adults better with respect to marketing goods and services to them. Rosenthal asserted that young adults prefer to see their charitable activities as an invisible act rather than something to be publicly advertised by posting to Twitter or Facebook or wearing a charity event t-shirt (Rosenthal, 2010). That said, whether consumer motivations for charitable giving are motivated by social concerns or personal values, marketing concepts still have a major role in helping a charity gain support. Conceptual Framework This paper seeks to explore the relationships among charitable causes, celebrity endorsers, products, and consumers. Grubb and Grathwohl?s (1967) Model of Consuming Behavior has been explored to identify each of the relationships among the self, symbolic goods, and the responses from significant others. The result is a conceptual model that looks at the individual fit relationships formed among the celebrity endorsers, consumers, endorsed products, and endorsed charitable causes that influence consumer attitudes and purchase intentions toward celebrity endorsed charitable cause products. The Theory of Symbolic Consumption Grubb and Grathwohl (1967) were among the first to look at the interaction between consumers and the ?symbolic consumption? of products. Their work built upon Hall and 18 Lindsey?s (1957) research on Self Theory to help formulate a better understanding of the consumers? role in consumption. Self-Theory argues that an individual?s perceptions, feelings, and attitudes allow the individual to form a concept of ?self? as an object, or image. This self- image will affect an individual?s decisions and behavior, which will, in turn, result in additional feedback that further develops his or her concept of self. When interacting with others, Roger?s (1951) Theory of Self-Enhancement explains that an individual?s concept of the self is influenced by responses from influential persons such as teachers, peers, or significant others. The responses and recognition the individuals receive from influential persons about themselves can strengthen their self-concepts (Grubb & Grathwohl, 1967). The Theory of Symbolic Consumption (Grubb & Grathwohl, 1967) applies the framework of Self Theory (Hall & Lindsey, 1957) to consumption behavior, treating goods as symbols and the consumption of these symbols as a way for individuals to communicate the self- concept. Individuals perceive that others, such as family, peers, and individuals of authority, are expressing or enhancing their images by possessing or consuming certain goods; this process causes the goods to take on symbolic meanings. As a result, individuals may purchase goods that are publically recognized specifically for the recognition and self-enhancement that the good will conveys to the individual?s peers (Grubb & Grathwohl, 1967). The Grubb and Grathwohl 1967 Model of Consuming Behavior consists of six postulates, which can be summarized as follows: (1) an individual has a self-concept; (2) this self-concept has value to the individual; (3) the individual?s behavior will be directed towards self-concept enhancement; (4) influential persons, such as teachers, parents, peers, and significant others, have a significant role in the formation of the individual?s self-concept; (5) individuals use goods as symbols to communicate between influential persons and themselves; and (6) the interaction 19 process facilitated through the symbolic consumption of goods does, in fact, affect the individual?s self-concept (Grubb and Grathwohl, 1967). Figure 1. Grubb and Grathwohl 1967 Model of Consuming Behavior Figure 1, Grubb and Grathwohl (1967), illustrates each of these postulates. Individual A represents postulates 1 and 2. Behavior begins when individual A consumes symbol X; the symbolic meaning may be tied to the store where the purchase made, the brand line the product comes from, and/or the specific product itself. By purchasing the symbol, an individual acts in accordance with postulate 3. Line ?a? represents the intrinsic value of the symbol to the individual; that is, it represents how the purchase of the good directly and internally enhances the individual?s self-concept. The dotted lines represent three ways the consumption of the symbol enhance or affirm the concept of self extrinsically. Line ?b? represents how the good is also an extrinsic expression of self-enhancement, an affirmation of the individual?s self-concept. Line ?c? represents the extrinsic value of the symbol as a way of communicating with influential persons represented by audience B. This type of interaction is intended to enhance the individuals? self-concept through the audiences? responses (postulate 5). Line ?d? represents the several processes that follow the 20 communication with the audience that results in further enhancement of the individual?s self- concept. One result of communication with the audience is described as the intra-action process, defined as the process of internal communication. The communication has extrinsic value to the individual because it has the potential to produce desired responses from the audience. The individual may receive recognition from the audience for his or her purchase, creating further enhancement of the self-concept. The audience may also respond to the communication by themselves consuming the symbol (audience consumption is also represented by line c). The reciprocal consumption provides self-enhancement opportunities to both the individual and the audience, with each extracting value from the symbol as a result of the initial communication (Grubb & Grathwohl, 1967). The Grubb and Grathwohl Model of Consuming Behavior (1967) can be applied to examine the relationships between celebrity endorsers, their consumption and/or endorsement of charitable causes or cause products, and the target consumers. Table 2 outlines the variables applied in such a relationship and the revised relationship is modeled in Figure 2. Table 2 Explanation of Variables Used in Adapted Model of Consuming Behavior Individual A Symbol X Audience B Celebrity Endorser Charitable Cause Cause Product Consumers 21 Figure 2. Adapted Model of Consuming Behavior Like the consumer in the original model, the celebrity endorser (CE) has a self-concept that he or she values. The desire to affirm, express, and enhance that self-concept motivates the CE to publically endorse a charitable cause or product. The endorsement may take the form of the CE donating to a charity in a public way, purchasing the associated product and using it in public, or entering into a more formal endorsement of the charitable cause or product with an associated marketing campaign. Regardless of the form of the endorsement, it is effectively a form of consumption by the celebrity. By endorsing a charitable cause or product, the celebrity receives intrinsic value, or internally expressed and experienced self-enhancement by affirming his or her self-concept (line a). Likewise, the charitable cause or product can be enhanced, as the symbolic value of the product increases by being associated with the status the CE brings to it (line b). The endorsement of the charitable cause or cause product is also an intra-action, or internal communication between the CE and consumers (line d). The endorsement allows the CE to associate himself or herself with the endorsed charitable cause or endorsed product in the eyes Celebrity Endorser Endorsed Charitable Cause/ Endorsed Product Consumer d c b a 22 of the consumers (that is, the symbolic meanings that the celebrities bring to the endorsed charitable cause or endorsed product), enhancing the CE?s self-concept. This communication leads to responses by consumers that also have value to the CE. The dotted line d represents an extrinsic enhancement opportunity where the CE seeks the consumer?s response for self- enhancement purposes and the consumer responds to the CE based on the CE?s behavior. One response the celebrity desires is simply additional goodwill and affirmation from consumers as an audience as a result of being associated with a charitable cause. Of course, the CE is primarily interested in (and receives additional enhancement of his or her self-image from) the consumer responding by interacting with the charitable cause or associated product; that is, he or she hopes that the endorsement leads consumers to make direct donations to the charitable cause or buy the associated product. Dotted line c represents the extrinsic value experienced between the consumer and the CE?s endorsed charitable cause or endorsed product. The CE?s behavior has the outward effects of enhancing the value the endorsed charitable cause or endorsed product possesses through the consumer?s interpretation of the CE?s use of the charitable cause or product. Conceptual Model The proposed conceptual model in Figure 3 adapts construct relationships based on Grubb and Grathwohl?s (1967) Model of Consuming Behavior (see Figure 1). Grubb and Grathwohl?s (1967) variables: Individual, Symbol and Audience are replaced by Celebrity Endorsers, Endorsed Charitable Causes and Endorsed Products, and Consumers respectively (see Figure 2). The proposed conceptual model poses that fit between the CE and the variables Endorsed Charitable Causes (ECC), Endorsed Products (EP), and Consumers (CONSR) will influence consumer attitude toward a celebrity endorsed product and celebrity endorsed 23 charitable cause. These attitudes, in turn, influence purchase intent for celebrity endorsed charitable cause products (CECCP). Consumer ? endorsed product fit (CONSR-EP fit) is conceptualized as a moderator of the effect celebrity endorser-endorsed product fit (CE-EP fit) has on consumer attitudes toward an endorsed product. Similarly, consumer ? endorsed charitable cause fit (CONSR-ECC fit) is conceptualized as a moderator of the effect celebrity endorser ? endorsed charitable cause fit (CE-ECC fit) has on consumer attitudes toward an endorsed charitable cause. In the following hypothesis development section, theoretical and empirical support will be given to explain the relationships proposed in the conceptual model (Figure 3). Figure 3. Consumer response to celebrity endorsement of charitable cause product conceptual model Purchase Intent H2 Consumer ? Endorsed Product Fit H5 Celebrity Endorser ? Endorsed Charitable Cause Fit H6 Consumer ? Endorsed Charitable Cause Fit H7 Consumer attitude toward endorsed product H8 Consumer attitude toward endorsed charitable cause H1 Celebrity Endorser ? Endorsed Product Fit H4 H3 Celebrity Endorser ? Consumer Fit 24 Hypothesis Development The objective of the present study is to understand the effects of celebrity endorser fit relationships. Specifically, the study seeks to better understand what effect celebrity endorser ? endorsed product fit (CE-EP fit), celebrity endorser ? consumer fit (CE-CONSR fit) and celebrity endorser ? endorsed charitable cause fit (CE-ECC fit) relationships have on consumers? attitudes toward endorsed products and endorsed charitable causes and their subsequent impact on purchase intentions. Furthermore, this study will examine the moderating roles of consumer ? endorsed product fit and consumer ? endorsed charitable cause fit relationships on consumer attitudes. To attain these research objectives, hypotheses are developed based on the literature supporting the conceptual model relationships. Celebrity endorser and endorsed product fit (CE-EP fit). Based on the following literature, consumer perceptions of a celebrity?s attractiveness are one of the important factors impacting consumers? perceptions of CE-EP fit. Therefore, in establishing a fit relationship between celebrity endorser and endorsed product, the celebrity should be chosen based on attractiveness attributes. The Match-up hypothesis (MUH) provides the foundation for the CE- EP fit hypothesis. Match-up hypothesis (MUH). First noted by Baker and Churchill (1977), endorser attractiveness has an influence on advertisement evaluations. Thus the match-up hypothesis (MUH), as proposed by Baker and Churchill (1977) is a criterion by which marketers can carefully select the right celebrity endorser for the right product based on attractiveness and/or other attributes. Baker and Churchill (1977) manipulated attractiveness (attractive vs. unattractive model) and product type (coffee vs. cologne/perfume/aftershave lotion) in a 2x2 randomized factorial block study by gender. The authors hypothesized that when the models are 25 considered physically attractive they will: "have a stronger, positive impact on the ratings provided by both males and females when the advertised product has romantic overtones than when the advertised product does not have such overtones" (Baker & Churchill, 1977, p. 542). The results supported this hypothesis. Specifically, female models had a stronger, positive impact on products ratings. The female models also connected on a more personal level to the subjects when the product was familiar (Cologne/perfume/ aftershave lotion vs. coffee). The product match-up hypothesis is further supported by the finding that males responded negatively to attractiveness when the product was coffee but positively when the product was cologne/perfume/aftershave lotion. The Baker and Churchill (1977) research indicated that males tended to respond more favorably when the model was female regardless of product; however, there is little difference in female responses to either male or female models. The results indicated that attractiveness is an appropriate factor to consider or control for in spokesperson-product match-up. In a study from Kahle and Homer (1985), the authors used the MUH to support their manipulation of celebrity-source physical attractiveness, celebrity source likability, and participant product involvement. According to the authors, their study demonstrated that disposable razors are a physical attractiveness enhancer and when paired with an attractive celebrity endorser will convey a positive message to the person viewing the advertisement. Petty et al. (1983) used disposable razors in their study but failed to measure for endorser physical attractiveness when evaluating for advertisement effectiveness (Kahle & Homer, 1985). Instead, Petty and colleagues (1983) concluded that a celebrity?s notability (i.e., likability) was a more potent factor in consumer attitude formation for an advertised product. Kahle and Homer?s (1985) study was designed to replicate Petty et al.?s (1983) study but adjusting for celebrity 26 source effects. Kahle and Homer (1985) predicted that MUH would provide a main effect for attractiveness because of the proper match-up between the celebrity and the disposable razor. The results confirmed that MUH would lead to a main effect for attractiveness. Kamins (1990) also applied MUH to evaluate the use of celebrity endorser physical attractiveness to enhance advertisement evaluations. Kamins first identified celebrities based on attractiveness and familiarity in a pretest. Tom Selleck was chosen as physically attractive while Telly Savalas was chosen as physically unattractive. When measures for familiarity were evaluated statistically (extremely familiar ? extremely unfamiliar), both celebrities were considered equally familiar. A luxury car was chosen as an attractive product while a home computer was chosen as an unattractive product. A 2x2 factorial design and random assignment to the four treatment groups provided consistent results for using attractiveness as a factor in product evaluations using the MUH. The consumers? evaluations indicated that the use of the attractive Tom Selleck as celebrity endorser for a luxury car resulted in positive celebrity credibility and positive advertisement attitude. Endorsement of the luxury car by Telly Savalas resulted in lower celebrity credibility and advertisement attitude. The authors stated that these results were consistent with prior MUH study results that suggest attractive celebrities are perceived as more credible and produce a more favorable attitude than unattractive celebrities. Kamins and Gupta (1994) later used the MUH to study congruence, or matching, between the celebrity spokesperson and product type by manipulating the spokesperson and product type. The authors used a 2x2 factorial design to examine spokesperson versus product image congruence. Each of the sample participants was assigned to one of four experimental conditions. The authors examined whether celebrity spokesperson and product image congruence would lead to self-internalization (the target consumer would adopt the values of the 27 spokesperson) and identification (the target consumer would adopt a desire to identify with the spokesperson). For this study, the celebrity spokesperson was perceived to be significantly more believable for the product congruent with his image. The authors found internalization to be weak, although they believed that familiarity with the spokesperson was a mediating factor in determining overall perceived attractiveness and self-internalization. The authors? second hypothesis sought to confirm that when a spokesperson is congruent with a product he/she will be perceived as more attractive based on the concept of identification. The effects of celebrity and product congruence were measured on product attitude and purchase intention. The study confirmed that a spokespersons? perceived attractiveness and consumer identification with the spokesperson are based on positive consumer perceptions of spokesperson-product congruity and spokesperson-consumer values congruity. In other words, a high degree of celebrity and product congruence results in a more favorable attitude toward the product and greater purchase intent. Based on the above literature, the following hypothesis is proposed: H1: Level of fit (high/low) between the celebrity endorser and endorsed product will positively affect consumer attitude toward an endorsed product Consumer and endorsed product fit (CONSR-EP fit). The hypothesized CONSR-EP fit relationship is founded in self-concept literature that suggests there is a relationship between the endorsed product and the consumer. This relationship is hypothesized to be a moderator for the CE-EP fit effect on consumer attitude toward a celebrity endorsed product. The self-concept literature suggests that consumers prefer products that are congruent with the self. Self-concept. Birdwell (1968) was the first researcher to use self-concept theory to study how a person projects his or her image onto his or her choice of product. Birdwell?s study found a high degree of congruency between a person?s self-image and his/her purchases. It found that 28 perceptions of an automobile purchase were viewed as an extension of the self and that a person?s income was an enabling factor in his/her ability to make purchases. Dolich (1969) similarly studied congruence relationships and found that there was greater congruity between self-image and preferred brands than non-preferred brands across the studied brand categories. Additional studies of self-concept have examined the relationship between products as symbols and how congruent the product is with the self (that is, the consumer?s self-concept). Grubb and Hupp (1968) hypothesized that the formation of a consumer?s self-concept from responses of significant others, teachers, and peers influenced the self-concepts the consumer attributed to other consumers who purchased a similar branded product. They also hypothesized that consumers would perceive the self-concept of consumers of a competing product different from their own self-concept. The results of the study comparing the self-concepts of different brand car consumers confirmed that consumers hold stereotypes of consumers who purchase certain products. They also found strong evidence for understanding product symbolism in the consumption process. Allison et al. (1980) examined self-concept in terms of gender and sex-role self-concept. The study sought to examine the effect of gender specific or stereotyped products on product preference for the purpose of changing the way marketers think about marketing products to specific genders. The authors found that masculine and feminine image perceptions of products were both significantly influenced by the interaction between gender and product. The marketing implication of this research is that gender-role self-concept influences consumer perceptions of products and that, based upon current gender-role stereotyping, consumers are likely to prefer products typed to their gender. 29 Sirgy (1982) is credited with the development of the Self-Congruity Theory, which posits that consumers prefer a product-user, or stereotypical image of the product user, that has congruency with the users? self-concept. Building upon approaches to studying self-concept, Sirgy identified four basic approaches that deal with product image: ?(1) product image as it relates to the stereotypic image of the generalized product user; (2) product image in direct association with the self-concept; (3) sex-typed product image; and (4) differentiated product images? (p.288). Sirgy?s third statement regarding ?sex-typed product image? refers to the masculine or feminine symbolic attributes of a product. His fourth statement about product image differentiation supports Munson and Spivey?s (1980) research that people evaluate self- concept constructs differently. Munson and Spivey (1980) proposed two different comparative constructs to self-measurement. The expressive self-construct is made up of (1) the ideal self (I) ?the way one views one? self?and (2) the looking-glass self (LG) ? the way one perceives that one?s self is viewed by others. The product expressive self-construct is made up of (1) the self-perception (SP) of how one views oneself given a product preference and (2) the way one believes he/she is perceived by others given his/her product preference or looking glass product expressive self (OSP). The authors applied these four views of the self to construct a 2x2 factorial design for investigating whether a symbolic product (in this case, an automobile) versus a functional product (a tennis racket) would have any effect on the various self-constructs. The authors found support for their hypothesis that the ?LG? self rather than the ?I? self was more similar to product expressive self-constructs. For the LG consumer, the automobile versus the tennis racket provided greater value expression. Based on these findings it can be concluded that highly conspicuous, or socially consumed and symbolic products are better for establishing a consumer-product fit relationship. 30 Govers and Schoormans (2005) studied self-image to identify consumers? preference for products with ?personalities? that are congruent with their self-image. Using Aaker?s (1997) definition to describe brand personality, Govers and Schoormans applied human personality characteristics to the definition of product personality. Govers and Schoormans? (2005) findings suggest that self-image congruence occurs when product-personality is positively related to consumer preference for products. Elliott and Wattanasuwan (1998) considered the relationship between consumers? product preferences and their affiliations with various social groups, such as a peer group and belonging to a club. They found that affiliation with a social group was a reflection of consumers? preference for products that have symbolic content that is congruent with the social group. Their findings also suggest that the product-personality needs to be congruent with that of the product-user image reflected in the social group. In conclusion, self-concept studies consistently find consumer preference for products that are congruent with the self. Based on the literature above, the following hypothesis is proposed: H2: Level of perceived fit between the consumer and endorsed product will moderate the effect of celebrity endorser - endorsed product fit on consumer attitude toward an endorsed product Celebrity endorser and consumer fit (CE-CONSR). The following is a review of celebrity and consumer relationship formation. It is hypothesized that consumer attitude toward celebrity endorsed products and celebrity endorsed charitable causes is a function of the strength of the CE-CONSR relationship. CE-CONSR relationships are founded on the degree to which 31 the consumer forms an attitude towards the celebrity. Parasocial relationships can define one- sided relationships between consumers and celebrities. Parasocial relationships. Just as designers must understand their target market, so must marketers identify fit between the celebrity and the consumer market for which an endorsement is intended. Literature characterizing the relationship between celebrities and consumers helps to establish how consumers are likely to respond to celebrities. A parasocial relationship describes a one-way relationship in which a celebrity or performer, whether playing himself or a fictional character, is engaged with an audience (Horton & Wohl, 1956). Celebrities reveal their private lives either voluntarily or involuntarily to the audience (the consumers) through parasocial relationships. Consumers have the opportunity to respond emotionally to the celebrities? private life information, such as a break up, but cannot share the private life of the celebrity or see the celebrity respond emotionally to his/her private life information (De Backer et al., 2007). Celebrity-consumer fan relationship has been a topic of widespread interest subsequent to the rise of television mass-marketing in the 1950s (McCutcheon & Maltby, 2002). Consumers respond to celebrities in a variety of ways, from mild entertainment to mild pathological celebrity worship (Maltby et al., 2002). In studies by McCutcheon and Maltby (2002) in the United States and England, celebrity worship by the stereotypical fan of film-industry stars was described as ?foolish?, ?irresponsible, and ?submissive.? McCutcheon, Lange, & Houran (2002) identified personality traits that suggest celebrity worship using the Celebrity Attitude Scale (CAS). A higher CAS score suggests a stronger attraction to a given celebrity and a higher propensity to engage in celebrity worship (McCutcheon et al., 2002, McCarley & Escoto, 2003). Under the CAS, scores received by individuals are classified by the following categories: Entertainment/Social, Intense/Personal, 32 and Mild Pathological (Maltby et al., 2002). The Entertainment/Social level can be seen in affirmative responses to statements such as ?My friends and I like to discuss what my favorite celebrity has done.? Intense/Personal levels are exemplified by the statement ?I have frequent thoughts about my favorite celebrity, even when I don?t want to.? Mild Pathological is indicated by a positive response to the statement ?If I were lucky enough to meet my favorite celebrity, and s/he asked me to do something illegal as a favor, I would probably do it? (Maltby et al., 2002). In another CAS-type study, North and Sheridan (2009) allowed participants to evaluate a public figure of their choosing using a list of 34 items with a 5-point agree/disagree Likert-type scale. The nominees included actors, musicians, comedians, athletes, politicians, and those in the fashion/modeling business. The nominees were evaluated for physical attractiveness in a companion test utilizing participants that were not part of the main study. Not surprising to the authors, the number of attractive public figures outnumbered unattractive public figures by a ratio of 3:1. Attractiveness was found to positively correlate with intense-personal celebrity worship but to negatively correlate with the borderline pathological attitude. For the entertainment-social level of worship and motivation there was neither a positive or negative relationship with the attractiveness of the public figure. As a result, North and Sheridan asserted that that attractiveness was linked to some, but not all, of the ways public figures interact with people. Hung et al. (2011), studied celebrity endorsement effects using a celebrity-consumer relationship approach with a CAS-type scale. The authors hypothesized that a celebrity- consumer relationship is a behavioral antecedent for endorser effects and that such a relationship will have an effect on consumers? evaluations of a product and purchase intentions. The study 33 questions were designed to classify participants? parasocial relationships with celebrities as high or low under two constructs: entertainment/social and intense/personal. The authors stated that these constructs were chosen because these celebrity worship motivations are common among consumers who have desires to know more or be closer to their celebrity. The authors pretested and translated the CAS scale and had participants self-rate their relationship to a sports celebrity and an entertainment celebrity. Study participants evaluated brands endorsed by both celebrities and answered questions designed to measure purchase intention. The authors found that participants generally had a positive relationship with both celebrities and the products endorsed by both celebrities. The study established fit between the celebrity and consumer and found that fit had an effect on consumers? evaluation of future endorsements made by the celebrity. Chiou, Huang, and Chuang (2005) examined celebrity adoration for a popular music group among female adolescents in China. The authors hypothesized that attitude toward the act of purchasing celebrity-endorsed merchandise and purchase intention for the merchandise would be stronger for the celebrity adoration group as opposed to the non-adoration group. Adoration was measured in terms of agreement with statements such as ?you consider yourself a fan?, ?your friends believe you a fan?, ?being a fan is important to you?, and ?you collect photos and news about the music group.? The authors utilized the Theory of Planned Behavior created by Ajzen (1991) which is discussed in more detail below in the section concerning attitude and purchase intention. The Chiou and colleagues study (2005) confirmed that strong celebrity adoration had effects on the participants? attitude toward the act, which is the favorable evaluation of purchases of celebrity merchandise. 34 Based on the literature above, the following hypotheses are proposed: H3: Level (high/low) of fit between the celebrity endorser and consumer will positively affect consumer attitude toward an endorsed product H4: Level (high/low) of fit between the celebrity endorser and consumer will positively affect consumer attitude toward an endorsed charitable cause Celebrity endorser and endorsed charitable cause fit (CE-ECC fit). MUH studies have generally found that when there is high celebrity-product congruence there is likely to be a more favorable attitude toward the product and greater purchase intent toward the product (Kamins & Gupta, 1994). However, no study has specifically examined the relationship between the match-up of a celebrity endorser and an endorsed charitable cause. This study aims to address this gap by applying the MUH to the celebrity endorser and endorsed charitable cause relationship in order to identify how attitude and purchase intentions are affected. The following is a brief description of such relationships in business practice and builds the need for examining this fit relationship in the conceptual model. Celebrities and charities are commonly linked for the purpose of raising funds. CRM activities that have involved celebrities include Macy?s with their Passport Presents Glamorama, Celebrity Bowling Challenge for Sickle Cell Disease, and Creative Coalitions Hollywood studded Academy Award parties for arts advocacy (Macy?s Inc., 2010; Fofana, 2009: Jones, 2010e). Celebrities are also increasingly associating themselves with products that are linked to a charitable cause. For example, in 2008 Kiehls launched a partnership with actor Brad Pitt to endorse the newly-created product Keihl?s Aloe Vera Biodegradable Liquid Body Cleanser. Keihl?s promised to donate one-hundred percent of net profits of sales of the product to JPF Eco 35 Systems, which supports global environmental initiatives. (Naughton & Born, 2008). In turn, JPF Eco Systems designated Brad Pitt?s Make it Right organization in New Orleans as a beneficiary of its efforts. Make it Right is committed to building eco-friendly housing for those displaced by Hurricane Katrina (Bhatti, 2009). Roberta Weiss, Senior Vice President of Global Marketing Development for Kiehl's, stated that the use of celebrity-endorsers is only appropriate if it makes sense for the brand?s mission. She further stated that Kiehl?s timing and partnership with Pitt coincided with Kiehl?s desire to participate in environmental issues and that an association with Pitt presented a particular opportunity to appeal to younger consumers (Naughton & Born, 2008). Reebok has created a collection of shoes in cooperation with record label Quannum Records and several of their artists. Sales of the shoes ultimately benefit those suffering from the ongoing conflict in Darfur (Vasquez, 2008). In a related effort, Tom?s shoes and the One for One program, promises to donate a pair of shoes to those suffering from conflict for every pair of shoes purchased. The venture pairs Tom?s shoes with the non-governmental organization Giving Partners, which actually distributes the donated shoes. Many celebrities?including Kristen Bell, Keira Knightly, Kristen Stewart, Scarlett Johansson, and Sophia Bush?have flocked to support Tom?s efforts (coolspotters.com, 2011b). Feed Projects is another example of CRM that involves a not-for-profit, Feed Foundation, with sales of fashionable canvas and burlap bags. The proceeds go to feed children through the anti-hunger initiatives of the World Food Programme and UNICEF. Celebrities that have aligned themselves with this anti-hunger cause include founder Lauren Bush, Rachel Bilson, Toby McGuire, and Marcia Cross, all of whom can be seen in public with their bags (coolspotters.com, 2011a). 36 Celebrities are seen as possessing several desirable qualities such as trustworthiness, attractiveness, credibility, familiarity, importance, and relevance that draw attention to the products and causes with which they align themselves. The conceptual model examines fit between the celebrity endorser and endorsed product. The literature recognizes that products represent symbolic traits that can be interpreted by the consumer. This study also examines celebrity fit with the charitable cause. The conceptual model treats charitable causes in the same manner as other studies have treated products; both charitable causes and their associated products have symbolic traits to be interpreted by the consumer. Thus the following hypothesis is proposed: H5: Levels (high/low) of fit between the celebrity endorser and endorsed charitable cause will positively affect consumer attitude toward an endorsed charitable cause Consumer and endorsed charitable cause fit (CONSR-ECC fit). According to the 2010 Cone Cause Evaluation study (Cone Inc., 2010a), 41 percent of Americans surveyed said they had purchased a product because it was associated with a social cause. Additionally, 83 percent stated they wanted to see their retailers and service providers offer a product that would benefit a social cause from their consumption of that product. When asked how CRM would influence their purchase decisions, 85 percent stated generally that they would have a positive image of a product and/or company if it was supporting a cause. The 2010 Cone survey found that 18-24 year old consumers were the most likely group to purchase a good that benefits a cause, with mothers of young children placing as the next most likely group. Cone?s Vice President of Cause Branding, Julie Kivistik, has asserted that company-charity partnerships can provide opportunities to build relationships with a young generation, thereby creating loyal customers through adulthood (Panepento, 2006). 37 Yurchisin, Kwon, and Marcketti (2009) considered the topic of ethical consumption of cause-related products. Ethical consumption has been defined as ?the conscious and deliberate decision to make certain consumption choices due to personal moral beliefs and values? (Crane & Matten, 2004, p. 290). Beginning in 2005, rubber charity bracelets?led by the Lance Armstrong Foundation?s yellow bracelets that supported cancer research?became a very successful cause-related fashion product that suggested consumers were motivated by personal moral beliefs and values to purchase ethically based products. The authors found that consumers who had purchased a rubber charity bracelet were highly influenced by fashion and celebrities. The authors suggested that future development of cause-related products should meet consumers? interests in fashion and use celebrities to help market the product (Yurchisin et al., 2009). A fit relationship between the consumer and the endorsed charitable cause is initiated based on the relationship formed between the celebrity endorser and the charitable cause. The celebrity endorser?s relationship with the charitable cause initiates communication between the consumer and the charitable cause. In reference to the literature concerning the symbolic nature of charitable causes, this study seeks to establish fit based on the perceived strength of the relationship between the consumer and the charitable cause. Thus the following hypothesis is proposed: H6: Level of perceived fit between the consumer and endorsed charitable cause will moderate the effect of celebrity ? endorsed charitable cause fit on attitude toward an endorsed charitable cause. Attitude and purchase intention. This study considered how the attitude and behavior of the consumer relates to the consumer?s response to a celebrity?s endorsement of a charitable 38 cause product. The Theory of Planned Behavior (TPB) (Ajzen, 1991) was used to understand consumer attitude and purchase intention. TPB, along with brand literature, helps to define the consumer consumption process and attitude formation process. This study hypothesized that charities associated with a product will act in a similar manner to brands in affecting consumer beliefs about an associated product. As such, brand literature is reviewed in discussing the predicted impact of a charity association on consumer attitude and purchase intention. Theory of planned behavior. The Theory of Planned Behavior (Ajzen, 1991), has three variables: attitude, subjective norm, and perceived behavioral control. Francis and colleagues (2004), described the variables as follows: (1) whether a person is in favor of doing or performing the intended behavior (attitude); (2) whether a person feels social pressure to do or perform the intended behavior (subjective norm); and (3) how much control the person feels over doing or performing the intended behavior (perceived behavioral control). Each of these variables is a predictor of an intended behavior (see Figure 4). Figure 4. Theory of Planned Behavior (Ajzen, 1991) Perceived behavioral control is a variable that relates to constraints on an individual?s ability to perform a behavior and differentiates TPB from Theory of Reasoned Action (Fishbein & Ajzen, 1975). Such constraints include time, money, and or other factors such as parental or 39 familial influences (Ajzen, 1985, 1991; Chiou, et al, 2005). Subjective norms relate to social group or reference group influences which can have an effect on an individual?s motivation to comply with a socially acceptable behavior (Gotschi et al., 2010). Charity brand literature. It is hypothesized, based upon a review of brand literature, that charities associated with a product are similar to brands in influencing a consumer?s beliefs and ideas about the product. According to brand literature, attitude about a product brand can be a determinant of consumer response to the product (Keller, 1993). Moreover, celebrity endorsers embody symbolic meanings and thus have brand-like qualities (McCracken, 1989; Kotler, 2003). Assuming that consumers will form beliefs and attitudes about the celebrity endorser as a brand, any links the celebrity endorser forms to the endorsed product will be subject to beliefs and attitudes about the celebrity endorser. Likewise, any links the celebrity endorser forms to the endorsed charitable cause will be subject to the formation of beliefs and attitudes about the celebrity endorser. According to the TPB model, the formation of beliefs necessitates that a behavior will have a certain outcome. Purchase intention is an example of a certain outcome. Fishbein and Ajzen (1975) hypothesized that as an individual forms a belief, formed attitude is directly proportional to the possibility that the behavior will produce an outcome. Using the structure of Ajzen?s TPB model (1991), we propose that the formation of attitude will have an effect on purchase intention for a celebrity endorsed charitable cause product (CECCP). Thus, this study hypothesizes the following: H7: Consumer attitudes toward an endorsed product will positively influence purchase intent toward a Celebrity Endorsed Charitable Cause Product (CECCP) H8: Consumer attitudes toward a charitable cause will positively influence purchase intent toward a Celebrity Endorsed Charitable Cause Product (CECCP) 40 Chapter 3: Methodology This paper utilized an experimental design to examine the relationships hypothesized previously in Chapter 2. Two pretests were conducted to select the independent variable stimuli and assess the reliability of the scales. The design of the pretests, as well as the main study to test the hypotheses, are discussed below. A sample of female millennials, from a research firm?s national database, was recruited for each phase of the study. Research Design Two pretests were conducted to select two celebrities, products, and charities for manipulating the levels of the independent variables in the main study. The purpose of each pretest in the research design is outlined in Figure 5. This allowed for a 2 (low vs. high CE-EP Fit) x 2 (low vs. high CE-CONSR Fit) x 2 (low vs. high CE-ECC Fit) between-subjects factorial experiment. Each of the fit variables was manipulated by scenarios, which proposed that a celebrity had been chosen to endorse a given product that was associated with a given charitable cause, resulting in a total of eight hypothetical scenarios, using all possible combinations of the fit variables. Each volunteer participating in the main study was randomly assigned one of the eight experimental conditions. Participants were informed (1) that a certain celebrity had endorsed a specified product that was available for sale; (2) that the product was, in turn, associated with a specified charity; and (3) that purchase of the product would benefit the charity in a manner described. 41 Pretest 1: Celebrity and Charitable Cause Selection Pretest Purpose: (1) To narrow a list of celebrities from an initial list of 20 based on familiarity, attractiveness, relevance, and likability. Celebrities that lack high familiarity, likability or attractiveness will be discarded. Of those, celebrities with especially high or low relevance to the consumer will be used in Pretest 2. (2) To narrow a list of 20 charitable causes based on relevance, familiarity, and importance. Only charities that are equal in familiarity and have especially high or low relevance and importance to the consumer will be used in Pretest 2. Pretest 2: Congruency Pretest Purpose: To identify levels of congruency (fit) between each of the celebrities selected in Pretest 1 and (1) each of the selected charitable causes and (2) each of the product categories. Consumer perceived fit at high/low levels will be used as stimuli for the main study. The second Pretest will result in the selection of celebrities, causes, and products that meet the following: 1) high and low celebrity product fit; 2) high and low celebrity consumer fit; 3) high and low celebrity cause fit. Main Study Experimental Methodology: Online Survey Purpose: To examine the hypothesized relationships. Figure 5. Research process Stimulus Development Pretest 1. The first pretest was used to evaluate a list of celebrity endorsers, some of whom were selected for further analysis in Pretest 2. In addition, Pretest 1 evaluated a list of charitable causes, again in order to select a narrowed group for use in Pretest 2. An online survey was administered through a U.S. research and consulting company to a female population 42 between the ages of 19-24. The use of a female sample was preferred because females are more highly involved in the product categories to be used in the study than a male sample, thus offering greater opportunities for identifying fit among the studied variables (Lee & Thorson, 2008). Millennials, those born between 1982 and 2000 are a significant demographic group and are the target audience for this study (Howe & Strauss, 2000). According to a 2010 Cone Evolution Study, millennials specifically between the ages of 19-24 have a passion for supporting causes and tailoring their shopping habits to causes they can support (Cone, Inc., 2010a). Therefore, this study includes females between the ages of 19-24. The survey asked the sample to evaluate twenty female celebrities and nineteen charitable causes in the manner described below. Real female celebrities rather than fictitious celebrities were selected for this study to maintain ecological validity and because of the real-life nature of the questions in Pretest 1 (Lee & Thorson, 2008). The primary goals of Pretest 1 were (1) to identify celebrities that exhibit equal levels of familiarity, likeability, and attractiveness but disparate (high and low) levels of relevance to the sample and (2) to identify charitable causes that are most salient to the sample group by measuring for familiarity, relevance, and importance. Pretest 1 was administered after receiving approval from the University?s Institutional Review Board for Human Subjects Research (IRB) (Protocol #11-216 EX 1107). Instrument. In the construction of the Pretest 1 instrument, the researcher first compiled a list of twenty female celebrities using celebrity-oriented press publications such as People Magazine and US Magazine. Observations from the 68th Annual Golden Globe? Awards broadcast on NBC held on January 16, 2011, and the 53rd Grammy Awards broadcast on CBS on February 13, 2011, also aided in the selection of female celebrities for the list. The list of twenty female celebrities included celebrities who are actors, singers, and professional athletes. 43 In considering the pool of celebrities, the researcher attempted to select celebrities anticipated to be highly rated by study participants for familiarity, likability, and attractiveness but varied in rated relevance. The twenty female celebrities were evaluated as followed: (1) how familiar each celebrity was to the sample; (2) how attractive each celebrity is to the sample; (3) how relevant the participants considered each celebrity to her own self-image; and (4) whether (and how much) the participants like each celebrity. Becker-Olsen, Cudmore, and Hill (2006) used familiarity and likability scales in their study for well-liked and highly-familiar firms such as Home Depot and Revlon and had Cronbach?s ? =.96 for liking, ? =.96 for familiarity. The Becker-Olson study also evaluated relevance for charitable causes. The Pretest 1 scales for evaluating the celebrities were modeled upon those employed by the Becker-Olson study for evaluating consumer opinions of corporations. The survey instrument also included a measure for attractiveness in order to control for this factor. The survey asked participants to rate each of the celebrities on a seven-point scale for familiarity, relevance and likability (Becker-Olsen et al., 2006). To evaluate familiarity, participants were asked whether each celebrity was familiar to them (1=unfamiliar/7=familiar). Likeability was evaluated by asking participants to rate their personal impression of the celebrity (1=negative/7=positive). Participants were also asked whether they found the celebrity relevant to their self-image (1=not relevant/7=relevant) and whether they found the celebrity attractive (1=extremely physically unattractive/ 7 = extremely physically attractive). To develop manipulations for the eight experimental conditions in the main study, based on the results of the Pretest 1, the researcher considered celebrities that rated similarly-high for familiarity (participants are generally familiar with the celebrities), likability (participants generally like the celebrity) and attractiveness (participants generally find the celebrities 44 attractive). Celebrities that exhibited varying levels of high/low for relevance (how relevant the celebrity is to your self-image) were considered with the above factors. Of those that met these criteria, a group of four celebrities were chosen that were rated by participants as highly relevant (High CE-CONSR fit) while a second group of four were chosen that were rated low in relevance (Low CE-CONSR fit). The Pretest 1 survey also included a list of nineteen charitable causes that were gathered by the researcher using CharityNavigator.com. Charity Navigator is a leading website designed to allow consumers to evaluate the financial health of over 5,500 charities. The charities selected for the list were chosen for anticipated importance, relevance and familiarity based after reviewing Charity Navigator?s various evaluations of the charities. For example, Charity Navigator?s list of the top ten most requested charities and the top ten ?charities worth watching? contributed primarily to the list. Pretest 1 participants were asked to (1) rate how familiar each charity was to them and (2) rate how important and relevant they believed each charity is. The same seven-point measure for familiarity and relevance that was used to evaluate celebrities was employed for the list of charitable causes (1=unfamiliar/7=familiar, 1=not relevant/7=relevant) (Becker-Olsen et al., 2006). A similar seven-point scale was used to evaluate each of the charitable causes for importance (1=not important/7=very important) (Becker-Olsen et al., 2006). The objective was to identify the charities that participants found to be especially high and low in relevance but equal in importance and familiarity for Pretest 2. Similar to the selection of celebrities, eight charities were chosen that represented equal groups of High CONSR-ECC fit and Low CONSR- ECC fit. 45 Sampling and data collection procedure. A U.S. research and consulting company was used to distribute the pretest to a national sample. To recruit the sample, the research and consulting company sent an email message to potential participants with an information letter (see Appendix A for the information letter) informing them of the study, incentive, and how to participate. After reading the information letter, participants could click on the link embedded in the email to be taken to the online survey (see Appendix B for the online survey). A total of 57 females participated in Pretest 1. Of those that participated, 13 surveys were eliminated because of missing data and eight additional surveys were eliminated because the participant was over the age of 24. A final total of 36 surveys were evaluated to determine which celebrities and causes would be used in Pretest 2 (see Table 3). Of the 36 participants, 52.9 percent were 22-24 years old and while not a requirement to participate in the study, 80 percent said that they were currently enrolled in either a two or four year institution. Over 80 percent of the participants were from Texas and 75 percent stated they were of White/Caucasian ethnicity. Table 3 Pretest 1 Sample Characteristics and Frequency Distribution Sample Characteristics n % Sample Characteristics n % Age (N=36) State of Residence (N=36) 19 3 8.3% Alabama 1 2.8% 20 1 2.8% Arkansas 1 2.8% 21 13 36.1% Florida 1 2.8% 22 2 5.6% Iowa 1 2.8% 23 6 16.7% Texas 30 83.3% 24 11 30.6% Virginia 2 5.6% Higher Education: Institution Type (N=36) Ethnicity (N=36) White/Caucasian 27 75.0% Less than 2 year 1 2.8% African American 3 8.3% 2 year 1 2.8% Hispanic 2 5.6% 4 year 33 91.7% Asian 1 2.8% Native American 1 2.8% Higher Education: Current Enrollment (N=36) Other 2 5.6% Yes, Full time 29 80.6% Yes, Part time 1 2.8% Not enrolled 6 16.7% 46 Data analysis. Descriptive statistical analysis was performed to calculate the mean value for each of the constructs: attractiveness, familiarity, likability, and relevance with respect to each celebrity and familiarity, importance, and relevance with respect to causes. Celebrity endorser ? consumer fit (CE-CF). For celebrities, narrowing the list of 20 down to eight (four for each group of high/low) was done based on controlling for attractiveness (all celebrities had to be considered attractive). The scale for attractiveness had a Cronbach?s ? of .92, demonstrating high reliability. The mean scores for familiarity and likability did not provide additional insight into high and low CE groups. Hence, the factors of familiarity and likability were eliminated from consideration. The scale for relevance was also found to have high reliability (? = .954). Table 4 provides a visual explanation of how the celebrities qualified for consideration in Pretest 2. 47 Table 4 Celebrity Endorser Mean Scores Attractive Familiarity Likability Relevance Blake Lively 5.39 Taylor Swift 6.56 Anne Hathaway 5.83 Anne Hathaway 3.75 Anne Hathaway 5.31 Anne Hathaway 6.5 Emma Watson 5.44 Emma Watson 3.33 Emma Watson 5.14 Katy Perry 6.31 Taylor Swift 5.28 Taylor Swift 3.17 Taylor Swift 5.03 Kelly Clarkson 6.28 Jennifer Hudson 4.97 Blake Lively 3.08 Katy Perry 4.92 Lindsay Lohan 6.28 Katy Perry 4.83 Katy Perry 2.97 Leona Lewis 4.82 Raven-Symon? 5.92 Kelly Clarkson 4.78 Jennifer Hudson 2.92 Jennifer Hudson 4.83 Emma Watson 5.86 Raven-Symon? 4.58 Kelly Clarkson 2.72 Maria Sharapova 4.69 Jennifer Hudson 5.5 Ellen Page 4.58 Raven-Symon? 2.61 Danica Patrick 4.64 Blake Lively 5.08 Blake Lively 4.47 Ellen Page 2.58 Hayley Williams 4.61 Leona Lewis 4.75 Hayley Williams 4.44 Lea Michele 2.42 Ellen Page 4.58 Ellen Page 4.5 Leona Lewis 4.33 Hayley Williams 2.33 Lea Michele 4.56 Danica Patrick 4.22 Lea Michele 4.25 Leona Lewis 2.25 Gina Carano 4.47 JoJo Levesque 4.03 Maria Sharapova 4.19 Maria Sharapova 2.17 Michele Wie 4.33 Lea Michele 4.03 Mia Wasikowska 4.14 Lindsay Lohan 2.11 JoJo Levesque 4.31 Maria Sharapova 3.56 JoJo Levesque 4.14 JoJo Levesque 2.08 Lindsay Vonn 4.25 Hayley Williams 3.44 Lindsay Vonn 4.08 Danica Patrick 2.06 Mia Wasikowska 4.25 Michele Wie 2.78 Danica Patrick 4.03 Michele Wie 2.06 Kelly Clarkson 4.19 Mia Wasikowska 2.61 Michele Wie 4.03 Lindsay Vonn 2.06 Raven-Symon? 4.11 Lindsay Vonn 2.53 Gina Carano 3.81 Mia Wasikowska 1.94 Lindsay Lohan 3.22 Gina Carano 1.81 Lindsay Lohan 2.19 Gina Carano 1.81 Note: Highlighted cells represent celebrities who met criteria for selection for Pretest 2 Based on the mean scores for celebrity relevance and attractiveness, Anne Hathaway (AH), Blake Lively (BL), Emma Watson (EW), and Taylor Swift (TS) were chosen to represent the high group of celebrities for use in Pretest 2; and Danica Patrick (DP), JoJo Levesque (JL), Leona Lewis (LL), and Maria Sharapova (MS) were chosen to represent the low group of celebrities for use in Pretest 2. Paired sample t-tests (see Table 5) on celebrity relevance confirmed that there was significant difference between the high and low groupings of celebrities, with minimal differences within each group. 48 Table 5 Paired T-Tests: Celebrity Relevance High Group Low Group AH TS EW BL JL DP MS LL High Group AH -2.249 -1.570 -2.523 -4.537 -4.852 -4.433 -4.453 .031 .125 .016 .000 .000 .000 .000 TS 2.249 .505 -.415 -3.529 -3.413 -2.598 -2.566 .031 .616 .681 .001 .002 .006 .015 EW 1.570 -.505 -.851 -3.663 -3.930 -3.829 -3.499 .125 .616 .401 .001 .000 .001 .001 BL 2.523 .415 .851 -3.137 -3.317 -2.868 -2.712 .016 .681 .401 .003 .002 .007 .010 Low Group JL 4.537 3.529 3.663 3.137 -.128 .301 .734 .000 .001 .001 .003 .899 .765 .468 DP 4.852 3.413 3.930 3.317 .128 .780 .879 .000 .002 .000 .002 .899 .441 .385 MS 4.433 2.598 3.829 2.868 -.301 -.780 .393 .000 .006 .001 .007 .765 .441 .697 LL 4.453 2.566 3.499 2.712 -.734 -.879 -.393 .000 .015 .001 .010 .468 .385 .697 Note: Highlighted cells represent significant differences at the .05 significance level Charitable causes were first evaluated based on their mean scores. Of the 19 charities surveyed, 9 charities were eliminated from further review because they failed to be familiar to the sample. Only charities that rated highly in familiarity were then evaluated for high and low degrees of relevance and importance. The scale for familiarity had a Cronbach?s ? of .858. Table 6 provides a visual explanation of how the charities were selected for Pretest 2. 49 Table 6 Charitable Causes Mean Scores Familiarity Importance Relevance Susan G. Komen for the Cure 6.17 American Red Cross 6.64 Susan G. Komen for the Cure 5.0 American Red Cross 6.14 American Cancer Society 6.53 American Red Cross 4.86 American Cancer Society 5.83 Feed the Children 6.33 American Cancer Society 4.78 American Society for the Prevention of Cruelty to Animals 5.08 Susan G. Komen for the Cure 6.31 American Society for the Prevention of Cruelty to Animals 4.47 The Humane Society of the United States 5.06 Invisible Children 5.97 Feed the Children 4.22 Feed the Children 4.92 American Society for the Prevention of Cruelty to Animals 5.94 Invisible Children 4.22 Fellowship of Christian Athletes 4.81 The Humane Society of the United States 5.86 The Humane Society of the United States 4.08 Invisible Children 4.33 Friends of the World Food Program 5.67 Doctors Without Borders, USA 3.97 Doctors Without Borders, USA 4.31 Doctors Without Borders, USA 5.64 World Vision 3.92 VH1 Save The Music Foundation 4.03 Paralyzed Veterans of America 5.5 Rainforest Foundation US 3.83 World Vision 3.28 Disabled American Veterans Charitable Service Trust 5.22 Best Buddies International 3.69 Rainforest Foundation US 3.22 Rainforest Foundation US 5.08 VH1 Save The Music Foundation 3.67 Disabled American Veterans Charitable Service Trust 2.97 World Vision 4.97 Fellowship of Christian Athletes 3.64 Compassion International 2.78 Marine Corps Heritage Foundation 4.94 Disabled American Veterans Charitable Service Trust 3.61 Best Buddies International 2.78 Compassion International 4.69 Compassion International 3.58 Paralyzed Veterans of America 2.61 New York Restoration Project 4.67 Friends of the World Food Program 3.58 New York Restoration Project 2.61 VH1 Save The Music Foundation 4.67 Marine Corps Heritage Foundation 3.56 Friends of the World Food Program 2.5 Best Buddies International 4.67 Paralyzed Veterans of America 3.47 Marine Corps Heritage Foundation 2.31 Fellowship of Christian Athletes 4.56 New York Restoration Project 3.28 Note: Highlighted cells represent charitable causes that met criteria for selection for Pretest 2 Based on the results, it was determined that the consumers considered all charities important and therefore, importance was not a usable construct to determine high and low groups. Of the remaining 10 charities, the highest and lowest scoring charities on the factor 50 relevance were selected for further analysis. The scale for charitable cause relevance had Cronbach?s ? of .962 providing reliability to the measures used. Based on the relevance means, the American Red Cross (ARC), American Cancer Society (ACS), Susan G. Komen for the Cure (SGK), and Association for the Prevention of Cruelty to Animals (ASPCA) were chosen to represent the high group for use in Pretest 2; and Doctors Without Borders (DWD), Fellowship of Christian Athletes (FCA), The Humane Society of the US (HS), and VH1 Save the Music Foundation (VH1) were chosen to represent the low group for use in Pretest 2. Paired sample t-tests (see Table7) were conducted on charitable cause relevance to determine whether significant differences existed between the high and low groups. The results confirmed that there is significant difference between the high and low groupings of most charitable causes, with minimal differences within each group. Table 7 Paired T-Tests: Charitable Cause Relevance High Group Low Group SGK ARC ACS ASPCA HS DWB VH1 FCA High Group SGK -.896 -1.244 -2.108 -2.686 -3.527 -4.357 -3.476 .377 .222 .042 .011 .001 .000 .001 ARC .896 -.502 -1.904 -2.606 -3.358 -4.248 -4.024 .377 .619 .065 .013 .001 .000 .000 ACS 1.244 .502 -1.418 2.292 -2.987 -3.768 -3.254 .222 .619 .148 .028 .005 .001 .003 ASPCA 2.108 1.904 1.418 -1.667 -2.201 -3.245 -2.118 .042 .065 .148 .104 .034 .003 .041 Low Group HS 2.686 2.606 2.292 1.667 -.426 -2.548 -1.057 .011 .013 .028 .104 .672 .014 .298 DWB 3.527 3.358 2.987 2.201 .426 -1.429 -1.000 .001 .001 .005 .034 .672 .162 .324 VH1 4.357 4.248 3.768 3.245 2.584 1.429 -.077 .000 .000 .001 .003 .014 .162 .939 FCA 3.476 4.024 3.254 2.118 1.057 1.000 .077 .001 .000 .003 .041 .298 .324 .939 Note: Highlighted cells represent significant differences at the .05 significance level 51 Pretest 2. Pretest 2 aimed to identify charitable causes and product categories that demonstrate high and low fit with the Celebrities chosen from Pretest 1. A second online survey was administered by the same U.S. research and consulting company to a sample of females that fit the previous criteria for participation in Pretest 1. The online survey asked questions in an effort to identify different levels (high and low) of congruency between the selected celebrities and both (1) various charitable causes and (2) various product categories. As an outcome, Pretest 2 aimed to: (1) identify charitable causes that demonstrate high and low fit with the celebrities selected for the high/low celebrity-consumer fit conditions and (2) identify products that demonstrate high and low fit with the celebrities selected for the high/low celebrity-consumer fit conditions. Pretest 2 was administered after receiving approval from the University?s Institutional Review Board for Human Subjects Research (IRB) (Protocol # 11-216 EX 1107). Instrument. The goal of Pretest 2 was to match each celebrity (chosen from Pretest 1) with a charitable cause that, in the opinion of the survey participants, had a high and low level of congruency with the celebrity. The survey utilized Kamins and Gupta?s (1994) seven-point congruent/incongruent scale; survey participants were asked to evaluate the congruity between each of the potential celebrity endorsers and the endorsed charitable causes (1=incongruent/7=congruent). The survey also asked participants to evaluate the congruency of a list of eight product categories with each celebrity (chosen from Pretest 1). The researcher reviewed CECCP currently available at shoppingblog.com, a charity shopping website. After reviewing the range of products that are typically endorsed by celebrities, including those endorsed for a charitable cause, the researcher was able to form product categories. The product categories determined were: jewelry, bags (including handbags and totes), beauty products, home d?cor accessories, 52 diet and weight loss products, cleaning products, consumer packaged foods (including perishable and non-perishable food items packaged for sale), and clothing (such as branded t-shirts). These product categories represent a large range of endorsable products that fit the main study. The symbolic nature of these products is important to this study because they represent items that are commonly purchased items the target market is likely to make. The aim of this section was to identify product categories that represent both the high and low levels of congruency with each celebrity. The Kamins and Gupta?s (1994) seven-point congruent/incongruent scale was used for this test as well. Sampling and data collection procedure. The U.S. research and consulting company used in Pretest 1 was used to distribute Pretest 2 to a national sample. To recruit the between subjects sample, the research and consulting company sent an email to potential participants with an information letter (see Appendix B for the information letter) informing them of the study, incentive, and how to participate. After reading the information letter, participants could click on the link embedded in the email to be taken to the online survey (see Appendix D for online survey). As in Pretest 1, the recruited sample for Pretest 2 was female between the ages of 19-24. A total of 64 females participated in Pretest 2. Of those that participated, 31 surveys were eliminated because of missing data or were considerably over the age of 24. Seven surveys were retained for this survey where the participant stated their age was 25. Although the information letter specified the desired age to participate was 19-24, according to Howe and Strauss (2000), millennials today could be as old as 29. Therefore, these participants were included in the data analysis. A final total of 33 surveys were evaluated to determine which celebrities and charitable causes would be used in Pretest 2. The average age of Pretest 2 participants was 22 (see Table 8). 53 Table 8 Pretest 2 Sample Characteristics and Frequency Distribution Sample Characteristics n % Sample Characteristics n % Age (N=33) State of Residence (N=33) 19 6 18.2% Alabama 2 6.1% 20 3 9.1% California 1 3.0% 21 4 12.1% District of Columbia 1 3.0% 22 5 15.2% Maryland 1 3.0% 23 4 12.1% Massachusetts 1 3.0% 24 4 12.1% Nevada 1 3.0% 25 7 21.2% Ohio 3 9.1% Oklahoma 1 3.0% Ethnicity (N=33) Pennsylvania 1 3.0% White/Caucasian 28 84.8% Texas 21 63.6% Asian 3 9.1% Other 2 6.1% Higher Education: Institution Type (N=33) Higher Education: Current Enrollment (N=33) Less than 2 year 1 3.0% Yes, Full time 29 63.6% 4 year 32 97.0% Yes, Part time 1 6.1% Not Enrolled 6 30.3% Data analysis. In order to arrive at the outcome of two celebrities that represent the High/Low CE-CONSR fit conditions and the High/Low CE-ECC fit conditions, selection criteria was set and followed (see Table 9). Table 9 Celebrity Endorser & Charitable Cause Fit Criteria Celebrity Level of Fit Charitable Cause High CE High Fit High ECC High CE Low Fit Low ECC Low CE High Fit Low ECC Low CE Low Fit High ECC 54 The high group charities were evaluated individually to see which celebrity demonstrated the highest level of fit. In all instances, Anne Hathaway (AH) was perceived to have the highest level of fit. AH was then evaluated in each of the low group charities to determine the causes with which she had a low level of fit. AH had low levels of fit for low group charities FCA [M=2.55, SD=1.543] and VH1 [M=3.82, SD=1.740]. Low group celebrities were then evaluated for high fit with FCA and VH1. Within FCA, Maria Sharapova (MS) [M=5.09, SD=1.990] and Danica Patrick (DP) [M=5.06, SD=1.784] were the only low group celebrities perceived to have the highest levels of fit. Within VH1, Leona Lewis (LL) [M=5.73, SD=1.875] and JoJo Levesque (JL) [M=5.12, SD=2.118], were the only low group celebrities perceived to have the highest levels of fit. MS, DP, LL, and JL were then evaluated to see what level of fit they demonstrated among the high group charities. MS did not demonstrate the lowest levels of fit with any of the high group charities. Because disparate levels of fit are desired, MS was eliminated. With MS eliminated due to moderate levels of fit with each of the high group charities, DP demonstrated low levels of fit [M=3.91, SD=1.702] with ASPCA. JL and LL both demonstrated the lowest levels of fit with each of the four high group charities. LL had the lowest level of fit with the ARC and the greatest disparity with AH. Hence, the pairings of AH and DP with ASPCA and FCA and the pairing of AH and LL with ARC and VH1 were reserved for further evaluation. A set of paired sample t-tests revealed a significant difference in the scores for AH [M=5.12, SD=1.474] and DP [M=3.91, SD=1.702] within the high group charity, ASPCA [t(32)=3.409, p=.002]. There was also significant difference in the scores for AH [M=2.55, SD=1.543] and DP [M=5.06, SD=1.784] within the low group charity, FCA; [t(32)=-6.713, p = 0.000]. These results confirm that there are disparate levels of fit between AH and DP for both the high/ low group charity, FCA and ASPCA. The paired sample t-tests also revealed significant 55 difference in the scores for AH [M=4.85, SD=1.460] and LL [M=3.24, SD=1.501] within the high group charity, ARC; [t(32)= 5.025, p=.000]. For the low group charity, VH1, there was significant difference in the scores for AH [M=3.82, SD=1.740] and LL [M=5.73, SD=1.875]; [t(32)=-4.893, p=.000]. The results of these paired t-tests also suggest disparate levels of fit between AH and LL for the high/low group charity, ARC and VH1 (see Table 10). Table 10 Paired T-Tests: Celebrity Endorser-Charitable Cause Celebrity Endorser-Endorsed Charitable Cause Fit Option 1 ECC CE-High CE-Low SD p ASPCA (high) AH (high) 5.12 DP (low) 3.91 3.409 .002 FCA (low) AH (high) 2.55 DP (low) 5.06 -6.713 .000 Celebrity Endorser-Endorsed Charitable Cause Fit Option 2 ECC CE-High CE-Low SD p ARC (high) AH (high) 4.85 LL (low) 3.24 5.025 .000 VH1 (low) AH (high) 3.82 LL (low) 5.73 -4.893 .000 Note: ECC=Endorsed Charitable Cause, CE=Celebrity Endorser The results of the paired sample t-tests provide evidence that either pairing would sufficiently meet the criteria for CE-CF and CE-CCF main study stimuli selection. Before determining which pair to use, the three celebrities, AH, DP, and LL?s levels of congruency among the product categories were evaluated and compared. The mean ordinal scores for perceived congruence for each of the three selected celebrity endorsers provided the basis for analysis (see Table 11). Contrary to the celebrity endorser-endorsed charitable cause evaluations where contrasting levels of congruence were desired, similar levels of congruence for the product categories were needed to determine high/low groups. Similarities, rather than the highest or lowest means, were set as evaluating criteria. 56 Table 11 Celebrity Endorser-Endorsed Product Mean Scores AH DP LL Bags 5.09 2.97 4.33 Beauty 5.45 3.24 4.58 Clothing 4.85 3.91 4.42 Cleaning 1.91 2.15 2.06 Food 2.33 3.45 2.15 Diet 3.36 3.09 3.12 Home 3.09 2.88 2.94 Jewelry 5.12 2.79 4.42 Note: Highlighted cells represent products selected for celebrity-endorsed product fit conditions The levels of congruency across all product categories for DP and LL were compared to those of AH for any commonalities. DP had very low means and low perceived congruency with all product categories. Conversely, LL had common levels of congruence with AH in the clothing (high) and food (low) product categories. Since these product categories provided the best and worst fit among the high/low celebrity endorsers, they were selected for use in the main study. From the results of the congruency analysis, Table 12 shows the two celebrity endorsers, Anne Hathaway (AH) and Leona Lewis (LL), the two endorsed charitable causes, the American Red Cross (ARC) and VH1 Save the Music Foundation (VH1), and the two endorsed product categories, clothing and food in formulating the manipulations of the three fit variables. For the main study stimuli, the author chose a winter scarf with a signature color evocative of the endorsed charitable cause to represent the clothing category (high) and chocolate covered pretzels in a ?limited edition? tin (color and logo coordinated to charitable cause) to represent the food category (low). Reference appendix E for images of stimuli used in each of the main study scenarios. 57 Table 12 Manipulated Variable Conditions for Factorial Design Questionnaire Version Celebrity-Consumer Fit Stimuli Celebrity-Charitable Cause Fit Stimuli Celebrity-Product Fit Stimuli 1 AH ? High CE-CONSR ARC ? High CE-ECC Clothing ? High CE-EP 2 AH ? High CE-CONSR ARC ? High CE-ECC Food ? Low CE-EP 3 AH ? High CE-CONSR VH1 ? Low CE-ECC Clothing ? High CE-EP 4 AH ? High CE-CONSR VH1 ? Low CE-ECC Food ? Low CE-EP 5 LL ? Low CE-CONSR VH1 ? High CE-ECC Food ? Low CE-EP 6 LL ? Low CE-CONSR VH1 ? High CE-ECC Clothing ? High CE-EP 7 LL ? Low CE-CONSR ARC ? Low CE-ECC Food ? Low CE-EP 8 LL ? Low CE-CONSR ARC ? Low CE-ECC Clothing ? High CE-EP Note: CE=Celebrity Endorser, CONSR=Consumer, ECC=Endorsed Charitable Cause, EP=Endorsed Product Main Study The purpose of the main study is to test the hypothesized relationships between celebrities, charitable causes, and charity-related products in order to understand the influence celebrity fit relationships have on consumer attitude and purchase intention of a Celebrity Endorsed Charitable Cause Product (CECCP). Instrument. The stimuli chosen as result of Pretests 1 and 2 were used in a 2x2x2 between-subject factorial design experiment. Brief ?press releases? (see Appendix E) were drafted announcing the celebrity and charitable cause partnership proposed in each scenario. Questionnaires were generated for each of the eight experimental conditions. All eight versions of the questionnaire included identical questions (other than the stimuli) presented in a written scenario manipulating the independent variables: high /low celebrity endorser-endorsed product fit (CE-EP fit), high /low celebrity endorser-consumer fit (CE-CONSR), and high /low celebrity 58 endorser-endorsed charitable cause fit (CE-ECC). Each participant was randomly assigned to one of the eight conditions. Participants were informed how they as consumers could help fund the endorsed charitable cause by the purchase of each endorsed product. The main study utilized an online questionnaire to collect participants? data. Each questionnaire contained: (1) a scenario that corresponded to one of the eight experimental conditions (2) measurements for consumer attitude and purchase intention, (3) manipulation check measures for CE-CONSR fit, CE-EP fit, and CE-ECC fit, (4) measurements for the moderating variables CONSR-EP fit and CONSR-ECC fit, and (5) demographic items. Sampling procedure and characteristics. Female participants were recruited by the same U.S. research and consulting company used to recruit the samples for Pretest 1 and Pretest 2. A total of 437 participants were recruited for the Main Study. According to a G power analysis with alpha = .05, power = .80, and a moderate effect size, the minimum number of participants needed for this experiment was 240 or 30 participants per experimental condition (Cohen, Cohen, West, & Aiken, 2003). In accordance with the policies and procedures established by the Office of Human Subjects Research, female participants were invited to participate by an email invitation explaining that their knowledge of celebrities and brands can serve as a useful tool in identifying trends in product endorsements. The participants were able to begin the survey by clicking on a hyperlink in the recruiting message pointing to the online survey. At the end of the survey, participants were provided with a debriefing page thanking them for their time and acknowledging that the ?press release? was fictitious in an attempt to capture their true response to the scenario. Participants were provided with an incentive for participating if following the debriefing page they submitted their email addresses for one of ten $25 cash prizes. Ten 59 participants were notified by email at the end of the survey period that they were randomly selected to receive a cash prize. Of the 437 survey responses, 26 responses were eliminated due to an indicated age greater than 30 (three responses), male gender (four responses), or because more than 20 percent of the questions remained unanswered (nineteen responses). The final useable data set was comprised of 411 survey responses. Dependent measures. The dependent measures in this study were consumer attitude toward an endorsed product, consumer attitude toward an endorsed charitable cause, and consumer purchase intention for a CECCP. A direct measure of consumers? attitude towards the endorsed product was taken using Osgood, Suci, and Tannenbaum?s (1957) semantic differential scale. This method of identifying a direct measurement of attitude is one of the most widely used measurements of attitude (Fishbein & Ajzen, 1975). Kamins and Gupta (1994) adapted the Osgood scale to consider four items on a seven-point scale (1=bad/7=good, 1=Unpleasant/7=Pleasant, 1= Disagreeable/7=Agreeable and 1=Unsatisfactory/7=Satisfactory). The scale was found to have a standardized alpha coefficient of 0.92 (Kamins & Gupta, 1994). This version of the scale was utilized to measure consumer perceived attitude toward the endorsed product by asking participants to ?describe the degree to which the product is related to one or the other end of the scale? (Osgood et al., 1957). This study measured consumer attitude toward an endorsed charitable cause by using the same four-item scale listed in Table 13 but with adapted wording to reflect charitable cause instead of product. 60 Table 13 Consumer Attitude and Purchase Intention Measurement Items Included in Main Study Variable/Construct Item Consumer Attitude toward an Endorsed Product (Osgood et al.,1957) semantic differential scale Reliability: ? = 0.92. Describe the degree to which the [product] is related to one or the other end of the scale. ? 1=Bad/7=Good, ? 1=Unpleasant/7=Pleasant, ? 1=Disagreeable/7=Agreeable ? 1=Unsatisfactory/7=Satisfactory Consumer Attitude toward an Endorsed Charitable Cause (Osgood et al.,1957) semantic differential scale Reliability: ? = 0.92 Describe the degree to which the [charity] is related to one or the other end of the scale. ? 1=Bad/7=Good, ? 1=Unpleasant/7=Pleasant, ? 1=Disagreeable/7=Agreeable ? 1=Unsatisfactory/7=Satisfactory Purchase Intention Measurement (PI) (Bower & Landreth, 2001; Kamins & Gupta, 1994; Ohanian, 1990) Reliability: ? = 0.90 ? How likely you are to inquire about the [Celebrity-Endorsed Cause Product]? ? How likely are you to consider purchasing [Celebrity- Endorsed Cause Product]? ? How likely are you to actually purchase for personal use the [Celebrity-Endorsed Cause Product]? According to Fishbein and Ajzen (1975), intentions are a result of attitudes and normative influences. While purchase intention is not an exact predictor of actual purchase, it is the closest predictor of the behavioral intent available to most researchers (Francis, et al., 2004). The purchase intention measure used in this study was built upon likely-to-purchase measurements used by Kamins and Gupta (1994), Ohanian (1990), and Bower and Landreth (2001) (See Table 13). These measures asked the participant (1) ?How likely you are to inquire [as to the chosen cause product stimuli]?? (2) ?How likely are you to consider purchasing [the chosen cause product stimuli]??, and (3) ?How likely are you to actually purchase for personal use [the chosen cause product stimuli]?? Responses to each of these questions were rated on a seven-point scale (1=very unlikely/7=very likely). Ohanian (1990) measured for correlations with attractiveness 61 and found the purchase intention item to be significant at the p > .05 level. Bower and Landreth (2001), found the scale to have a Cronbach alpha of .90. Manipulation check. The goal of the manipulation checks was to examine whether the manipulation of the three independent variables into high and low levels was successful or not. Manipulation checks for CE-EP fit and CE-ECC fit were intended to identify the effectiveness of the use of the MUH for the selection of the variables. To check for how well the variables celebrity endorser, endorsed product, and endorsed charitable cause were manipulated in the eight scenarios, participants were asked if they agreed or disagreed with statements about the CE, EP, and ECC. Celebrity endorser-endorsed product fit (CE-EP fit), celebrity endorser-endorsed charitable cause fit (CE-ECC fit). Using Govers and Mugge?s (2004) scale, participants evaluated each statement on a seven-point scale (1=strongly disagree/7=strongly agree). The scale was found to have a Cronbach alpha coefficient of .94 (Govers & Mugge, 2004). The statements adapted for use in this study were as follows: (1) ?[Celebrity] does (not) identify herself with the [product/cause]?; (2) ?This [product/cause] does (not) match with [celebrity?s] personality?; and (3) ?This [product/cause] is (in-)consistent with the way [celebrity endorser] sees herself.? Celebrity endorser-consumer fit (CE-CONSR fit). This study used the revised Celebrity Attitude Scale (CAS) by Maltby et al. (2002) to evaluate the manipulation of CE-CONSR fit. The revised scale has been used successfully in other studies of celebrity worship with reported Cronbach's alpha coefficients for the CAS of .92, .86, and .89 (Ashe & McCutcheon, 2001; Maltby et al., 2002; and McCutcheon et al., 2003). This study only used six of the eleven Intense-Personal questions from the scale to identify whether or not consumers had a strong fit 62 with the celebrity endorser. The six questions are located in Table 14. The other subscale dimensions and five items from the scale were eliminated because they were not relevant to the purpose of this study. The scale uses a seven-point scale (1=strongly disagree/7=strongly agree). Table 14 Celebrity Endorser Fit Measurement Items Included in Main Study Variable/Construct Item Celebrity Endorser-Consumer Fit (CE-CONSR fit) Manipulation Check Intense/Personal Items from Celebrity Attitude Scale (McCutcheon & Maltby, 2002) Reliability: ?=.92, ?=.86, and ?=.89 ? If I were to meet [celebrity] in person, he/she would already somehow know that I am his/her biggest fan. ? I share with [celebrity] a special bond that cannot be described in words. ? When something good happens to [celebrity] I feel like it happened to me. ? The successes of [celebrity] are my successes also. ? When something bad happens to [celebrity] I feel like it happened to me. ? If someone gave me several thousand dollars to do with as I please, I would consider spending it on a personal possession (like a napkin or paper plate) once used by [celebrity]. Celebrity Endorser-Product Fit (CE-PF), Celebrity Endorser- Cause Fit (CE-CCF) (adapted from Govers & Mugge, 2004) Manipulation Check Reliability: ? = .94 ? [Celebrity] does (not) identify herself with the [product/charity]. ? This [product/charity] (does not match) with [celebrity?s] personality. ? This [product/charity] is (in-) consistent with the way [celebrity] sees herself. Moderating variables. It was hypothesized that CONSR-EP fit and CONSR-ECC fit moderate the effects of celebrity fit constructs on consumer attitude toward an endorsed product and consumer attitude toward an endorsed charitable cause. The moderating relationships were studied because of differential consumer involvement with products and charitable causes and their effects on the dependent variables. 63 Consumer-endorsed product fit (CONSR-EP fit), consumer-endorsed charitable cause fit (CONSE-ECC fit). Govers and Schoorman?s (2005) scales for product-personality, product preference, and user-congruency were adapted for use in this study to establish fit between CONSR-ECC fit and CONSR-EP fit (see Table 15). Gover and Schoormans?s scale utilized questions that helped the individual consider the fit of these variables based on product personality and product preferences. Adapted items included: ?I identify myself with this [charity]?? on a seven-point scale that used end points that correlated to the question (1=Strongly Do Not Identify/7=Strongly Identify?. Additional items to establish fit included: ?Considering the types of people who prefer this [charity], do you identify with these people?? on a seven- point scale (1=strongly disagree/7=strongly agree). These items helped establish the consumers? understanding of how they perceived themselves in relationship with the variable. The original scale had three subscales each with reliable Cronbach alpha coefficients i.e., product-personality congruence, ? = 0.91; consumer preference, ?= 0.89; and user-congruence, ?= 0.95. 64 Table 15 Consumer Fit Measurement Items Included in Main Study Variable/Construct Item Consumer-Endorsed Product Fit (CONSE-EP fit) Consumer-Endorsed Charitable Cause Fit (CONSR-ECC fit) Moderating variable measures (adapted from Govers & Schoormans, 2005) Reliability: ? = 0.91, ?= 0.89, and ?= 0.95 Produc t Pre fer enc e ? Produc t Pe rsona lity Congru enc e ? I identify myself with this [product/charity]. ? This [product/charity] matches me. ? I think this [product/charity] is attractive. ? I think this is a good [product/charity]. Use r Congr ue nc e ? If you consider your own personality and compare it to the description of the [product/charity] to what extent are they dissimilar/similar? ? Considering the types of people who prefer this [product/charity], do you identify with these people? ? If you consider the types of people who prefer this [product/charity], are you like these people? ? If you consider the types of people who prefer this [product/charity], are they similar to the way you see yourself? Demographic items. The participants were asked demographic profiling questions at the conclusion of the main study survey questions. Age and gender were self-qualifying items provided in the information page before participants were assigned to one of the survey conditions. However, age and gender were verified in the demographic questions following the survey. Along with age and gender, ethnicity, higher education, family combined income, and charitable giving questions were included (see Table 16). 65 Table 16 Demographic and Donation Behavior Measurement Items Included in Main Study Variable/Construct Item Demographic Items Donation Behavior Items ? What is your age? ? What is your gender? ? What is your ethnicity? ? Are you currently attending school? ? What kind of college or university do you attend? ? What college, university, or technical school do you attend? ? What is the combined annual income of your family? ? How often do you donate to a charitable cause? ? How likely are you to make a donation of $25 or more in the next year to each of the following types (church or church- based organization, a school or school-based organization, A charitable or non-profit organization that is NOT church or school-based?) ? When deciding to give money to a charitable or non-profit organization that supports an issue or cause, how likely are you to give to?(An organization based in your city; An organization that serves people in your neighborhood; A national organization; An international organization) 66 Chapter 4: Results The purpose of the main study was to test the hypothesized relationships between celebrity endorsers, consumers, endorsed charitable causes, and charity-related endorsed products. The results of the manipulation checks, scale reliabilities, hypothesis testing, and sample demographics are presented in order to understand the influence celebrity fit relationships have on consumer attitude and purchase intention towards a CECCP. Main Study The main study was conducted after two pretests identified stimuli to be used in eight experimental conditions. A questionnaire was distributed to a national sample and collected over four days. The on-line survey had over 5,000 initial clicks but through self-qualifying, a total number of 437 surveys were collected. To qualify for the survey, the participant needed to be considered a millennial, be female, and complete at least 20 percent of the survey. The total number of qualified responses was 411. Sample description. The sample for the main study consisted of 411 females recruited by a U.S. research and consulting company. Of the 411 responses, a total of 405 participants stated their age was between 19-28 years old with 45.4 percent 23-24 years old. The sample was geographically diverse with California having the largest population of participants (13.2%), followed by Texas (7.6%), Florida (7.3%), and Pennsylvania (6.1%). With respect to ethnicity, a 67 majority of participants identified themselves as White/Caucasian (72.5%). The remainder of the participants identified themselves as African American (10.5%), Hispanic (7.8%), Asian (.2%), and Other (4.7%). The sample was well educated with over half the population (53.3%) indicating they were currently enrolled in a higher education institution either full-time (43.2%) or part-time (11.1%) while 45.7 percent of the population indicated they were not currently enrolled. However, 99.9 percent of the sample indicated post high school training at 4-year, 2- year, and technical training schools. Refer to Table 17 for frequency distributions. Table 17 Main Study Sample Characteristics and Frequency Distribution Sample Characteristics Sample Characteristics Age (N=405) n % State of Residence (N=409) n % n % 19 45 11.1% Alabama 13 3.1% New Jersey 8 1.9% 20 45 11.1% Arizona 6 1.5% New York 21 5.1% 21 68 16.8% Arkansas 1 .2% North Carolina 20 4.8% 22 57 14.1% California 54 13.1% North Dakota 1 .2% 23 96 23.7% Colorado 6 1.5% Ohio 16 3.9% 24 88 21.7% Connecticut 4 1.0% Oklahoma 1 .2% 25 2 .5% Delaware 3 .7% Oregon 11 2.7% 27 3 .7% Florida 30 7.3% Pennsylvania 25 6.1% 28 1 .2% Georgia 16 3.9% Rhode Island 1 .2% Ethnicity (N=408) Hawaii 2 .5% South Carolina 9 2.2% White/Caucasian 296 72.5% Idaho 3 .7% Tennessee 2 .5% African American 43 10.5% Illinois 19 4.6% Texas 31 7.5% Hispanic 32 7.8% Indiana 19 4.6% Utah 3 .7% Asian 1 .2% Iowa 3 .7% Vermont 1 .2% Other 19 4.7% Kansas 3 .7% Virginia 13 3.1% Higher Education: Kentucky 3 .7% Washington 7 1.7% Institution Type (N=406) Louisiana 7 1.7% West Virginia 2 .5% Less than 2 year 115 28.2% Maine 3 .7% Wisconsin 5 1.2% 2 year 93 22.9% Maryland 7 1.7% 4 year 198 48.8% Michigan 8 1.9% Higher Education: Minnesota 5 1.2% Current Enrollment (N=405) Mississippi 5 1.2% Yes, Full time 175 43.2% Missouri 7 1.7% Yes, Part time 45 11.1% Nevada 1 .2% Not enrolled 185 45.7% New Hampshire 4 1.0% 68 Scale reliability. Each of the scale items was evaluated for reliability. The results of each of the Cronbach alpha tests are reported in Table 18. The three-item scales for CE-ECC fit and CE-EP fit, adapted from Govers and Mugge (2004), did not have adequate reliability [? = 0.323 and ? = 0.335] respectively. When the third item, ?This [product/charity] is consistent with the way [celebrity] sees herself? was eliminated from the scale, the two-items for CE-EP fit and CE- ECC fit achieved reliability [? = 0.910 and ? = 0.873] respectively. The newly adapted CE-EP fit and CE-ECC fit scale items were computed into single variables. Table 18 Main Study Scale Reliability Variable N Mean Standard Deviation Cronbach Alpha Number of Items Celebrity Endorser-Consumer Fit (CE-CONSR) Manipulation Check 411 2.221 1.607 .977 6 Celebrity Endorser-Endorsed Product Fit (CE-EP) Manipulation Check 411 4.455 1.047 .335 3 Product Personality Scale Item Adjustment 411 4.652 1.552 .910 2 Celebrity Endorser-Endorsed Charitable Cause Fit (CE-ECC) Manipulation Check 411 4.622 1.048 .323 3 Product Personality Scale Item Adjustment 411 4.858 1.532 .873 2 Consumer-Endorsed Product Fit (CONSR-EP) Moderating variable measures 411 3.742 1.659 .963 8 Product Preference/Product Personality subscale 411 3.892 1.699 .913 4 User Congruence subscale 411 3.591 1.743 .972 4 Consumer-Endorsed Charitable Cause Fit (CONSE-ECC) Moderating variable measures 411 3.961 1.513 .952 8 Product Preference/Product Personality subscale 411 4.098 1.531 .873 4 User Congruence subscale 411 3.825 1.628 .971 4 Consumer Attitude toward Endorsed Product Dependent Measures 411 5.307 1.423 .971 4 Consumer Attitude toward Charitable Cause Dependent Measures 410 5.523 1.402 .978 4 Purchase Intention Measurement (PI) Dependent Measures 409 3.529 1.637 .950 3 69 Manipulation checks. Independent samples t-tests were used to determine the success of the manipulation of celebrity fit with the consumer, charity, and product. The results were inconsistent with the findings from the pretests. Levene?s test (Levene, 1960) for equality of variance for CE-EP fit [F=3.149, p=.077], CE-CONSR fit [F=1.070, p=.302], and CE-ECC fit [F=.763, p=.383] was greater than p=.05 suggesting that the variance between means were not significantly different. The outcomes, available in Table 19, suggest that the manipulations were not successful. Table 19 Main Study Manipulation Check Construct Condition N Mean Std. Dev. Levene's Test for Equality of Variances t-test for Equality of Means F Sig. t df Sig. (2- tailed) Mean Dif. Std. Error Dif. CE-EP High Fit 202 4.651 1.622 3.149 .077 -.014 409 .989 -.002 .153 Low Fit 209 4.653 1.486 CE- CONSR High Fit 208 2.343 1.606 1.070 .302 1.555 409 .121 .246 .158 Low Fit 203 2.097 1.602 CE-ECC High Fit 210 4.948 1.570 .763 .383 1.219 409 .224 .184 .151 Low Fit 201 4.764 1.486 In pretests, CE-EP fit and CE-ECC fit were established using Kamins and Gupta?s (1994) seven-point congruent/incongruent scale. These scales were not reused in the main study. Instead, an adaptation of Govers and Mugge?s (2004) scale was used to examine the established congruity (Pretest 2) between the CE and the EP and ECC using statements that examined the personality of the CE. This measure did not provide support for the manipulations. Specifically, the three item measure did not have reliability and was reduced to the first two items of the scale. These two items, both for CE-ECC and CE-EP, were not sufficient in confirming the effectiveness of the manipulations in the fit relationships. The failed manipulation check could 70 also be the result of using a within-subjects method for pretesting and a between-subjects method during the main study. Although the reported means were in the expected direction for the high and low conditions, the difference between the means was not statistically significant. In the pretest, consumers were exposed to all the stimuli, so they could form comparative evaluations. In the main study, consumers were not influenced by stimuli from other conditions. Rather, the consumer gave a general response to the stimuli since they did not have any other experiences or beliefs to compare it to. Pretest 1 established CE-CONSR fit based on the MUH literature. A seven-point scale for familiarity, relevance, likability, and attractiveness was used to select a high and low level of CE-C fit (Becker-Olsen et al., 2006). These scales were not reused during the main study. Instead, intense/personal items from the CAS by McCutcheon and Maltby (2002) were used based on literature that suggested that there was a correlation with the perceived attractiveness of the CE (established in Pretest 1) and the level of CE-CONSR fit. The mean response to each of the six questions suggests that a majority of the consumers indicated low ratings on the items. (see Table 20). This lead to elimination of the variability in the scale and resulted in inconclusive findings on the CE?s role in influencing consumer attitude and purchase intention. Table 20 Descriptive Statistics for Manipulation Check Measure for CE-CONSR Fit Item 1 Item 2 Item 3 Item 4 Item 5 Item 6 N 411 411 411 411 411 411 Mean 2.3917 2.2433 2.2482 2.1971 2.2311 2.0170 Median 2.0000 1.0000 1.0000 1.0000 1.0000 1.0000 Std. Deviation 1.70790 1.68325 1.71840 1.68714 1.73841 1.65048 71 Hypothesis testing. Manipulation checks on the manipulated variables CE-EP fit, CE- CONSR fit, and CE-ECC fit were not successful. Therefore, measured components of the manipulated variables were used in simple linear regression and ANCOVA to test the conceptual model hypotheses. In light of this change, the wording of H1, H3, H4, and H5 was changed to replace ?level of fit (high/low)? with ?perceived fit? and ?affect? with ?influence? (see Table 21). Table 21 Supported/Not Supported Hypotheses H1 Perceived Celebrity-Endorsed Product fit (CE-EP) will positively influence Consumer Attitude toward an endorsed product Not Supported H2 Perceived Consumer-Endorsed Product fit (CONSR-EP) will moderate the influence of CE-EP fit on Consumer Attitude toward an endorsed product Partially Supported H3 Perceived Celebrity-Consumer fit (CE-CONSR) will positively influence Consumer Attitude toward an endorsed product Not Supported H4 Perceived Celebrity-Consumer fit (CE-CONSR) will positively influence Consumer Attitude toward an endorsed charitable cause Not Supported H5 Perceived Celebrity-Endorsed Charitable Cause fit (CE-ECC) will positively influence Consumer Attitude toward an endorsed charitable cause Not Supported H6 Perceived Consumer-Endorsed Charitable Cause fit (CONSR- ECC) will moderate the influence of CE-ECC fit on Consumer Attitude toward an endorsed product Partially Supported H7 Perceived Consumer Attitudes toward an endorsed product will positively influence Purchase Intent toward a Celebrity Endorsed Charitable Cause Product (CECCP) Supported H8 Perceived Consumer Attitudes toward an endorsed charitable cause will positively influence Purchase Intent toward a Celebrity Endorsed Charitable Cause (CECCP) Not Supported Hypothesis 1 examined the influence of consumers? perceived fit between CE-EP fit on consumer attitude toward an endorsed product using simple linear regression. The results revealed a non-significant influence for CE-EP fit on consumer attitude toward an endorsed product [F(1,409)= .360, p=.719, b= .018, R2= 0 percent]; thus H1 was not supported. The beta 72 coefficient was positive resulting in a positive slope relationship. Although the slope is positive, the relationship is not significant. Specifically, the results suggest that CE-EP fit was not a significant predictor of consumer attitude toward an endorsed product in this study. Hypothesis 2 predicted that perceived CONSR-EP fit would moderate the influence CE- EP fit has on consumer attitude toward an endorsed product. The moderating relationship was tested using ANCOVA with CE-EP fit as a fixed factor and CONSR-EP fit as covariate. A preliminary analysis evaluating the homogeneity-of-regression (slopes) assumption indicated that the interaction between CONSR-EP fit and consumer attitude toward an endorsed product was not significant [F(1,407)=.286, p=.593]. The results of the ANCOVA revealed a significant influence for CONSR-EP on consumer attitude toward an endorsed product [F=105.849, p<.001]. However, the influence of CE-EP fit on consumer attitude toward an endorsed product was non-significant [F(1, 408)=.691, p=.406]. These results revealed that 20.6 percent of the variance in consumer attitude toward an endorsed product can be explained by CONSR-EP fit. Therefore, H2 was partially supported. H3 and H4 examined perceived fit between the celebrity endorser and consumer on consumer attitude toward the endorsed product and endorsed charitable cause respectively. It was hypothesized that perceived fit between CE-CONSR would positively influence consumer attitude toward an endorsed product (H3) and endorsed charitable cause (H4). The relationships were tested using simple linear regression. The results revealed a non-significant influence of CE-CONSR on consumer attitude toward an endorsed product [F(1,409)= -.287, p = .744, b= - .015, R2= 0 percent] and a non-significant influence on consumer attitude toward an endorsed charitable cause [F(1,408)= -1.318, p = .188, b= -.065, R2= .4 percent]. Thus, H3 and H4 were not supported. 73 In H5, perceived CE-ECC fit was predicted to positively influence consumer attitude toward an endorsed charitable cause. The results of the simple linear regression revealed a non- significant influence for CE-ECC on consumer attitude toward an endorsed charitable cause [F(1,408)= .812, p=.417, b= .040, R2= .2 percent]. Thus H5 was not supported. In H6, CONSR-ECC fit was predicted to moderate the relationship between CE-ECC fit and consumer attitude toward an endorsed charitable cause. The moderating relationship was tested using ANCOVA with CE-ECC fit as fixed factor and CONSR-ECC fit as covariate. The homogeneity-of-regression (slopes) assumption indicated that the interaction between CONSR- ECC fit and consumer attitude toward an endorsed charitable cause was not significant [F(1,408)=.392, p=.532] signifying the condition for ANCOVA was met. The results revealed a non-significant influence of CE-ECC fit on consumer attitude toward an endorsed charitable cause [F(1,407)=.371, p=.543]; however, there was a significant influence of CONSR-ECC fit on consumer attitude toward an endorsed charitable cause [F(1,407)=88.794, p<.001] and 18 percent of the variance explained in the model. Therefore, these results partially support H6. This study hypothesized that by examining consumer attitudes for an endorsed product separately from consumer attitudes toward an endorsed charitable cause would provide insight into consumer purchase intention for a celebrity endorsed charitable cause product (CECCP). Both H7 and H8 were first analyzed with simple linear regression. The results revealed a significant influence for consumer attitude toward an endorsed product on purchase intention [F(1,407)=5.270, p< .001, b= .273, R2= 7.4 percent]. Similarly, a significant influence for consumer attitude toward an endorsed charitable cause on purchase intention was found [F(1,407)=5.136, p< .001, b= .247, R2= 6.1 percent]. These results explained very little variance in purchase intention with little relative difference between the influence of consumer attitude 74 toward an endorsed product and endorsed charitable cause on purchase intention. Therefore, a multiple regression analysis was conducted to determine the collective influence on purchase intention. The results revealed a significant influence for consumer attitude toward an endorsed product on purchase intention [F(2,406)=2.895, p= .004, b= .197, R2= 8 percent]; however, the influence for consumer attitude toward an endorsed charitable cause on purchase intention was non-significant [F(2,406)=1.552, p= .112, b= .106, R2= 8 percent]. The results suggest that 8 percent of the variance in purchase intention can be explained by consumer attitude toward an endorsed product. Therefore, H7 is supported while H8 is not supported. Additional analysis. Several of the hypotheses of this study were not supported. However, in order to determine the full extent of influence of fit relationships on consumer attitude and purchase intention, additional analysis of potential unexplored relationships was undertaken in an attempt to better explain the data. The hypothesized influence of celebrity fit variables, CE-EP, CE-CONSR, and CE-ECC, on consumer attitudes toward an endorsed product and endorsed charitable cause was not supported in the results. Hence, additional analyses were performed to examine the potential influence of these predictor variables on purchase intentions. The results of the multiple regression analyses with CE-EP, CE-CONSR, and CE-ECC on purchase intention revealed significant influence for CE-EP on purchase intention [F(3,405)=55.485, p= .019, b=.142]; CE- CONSR on purchase intention [F(3,405)=55.485, p<.000, b=.518]; and CE-ECC on purchase intention [F(3,405)=55.485, p= .003, b= -.181]. These results indicate that together the celebrity fit variables explain 29.1 percent of the variance in PI. Although significant, the results revealed the beta coefficient for CE-ECC was negative. A negative slope relationship illustrates increases in CE-ECC decreases purchase intention. The slope relationships for CE-CONSR and CE-EP 75 were positive indicating increases in CE-CONSR and CE-EP increase purchase intention. These significant results present important relationships to examine further (see Figure 6). Figure 6. Modified conceptual model for future research In this study, CONSR-EP and CONSR-ECC fit were not hypothesized to have a direct influence on consumer attitudes, rather a moderating influence. In order to examine the direct influence of these variables on consumer attitude toward an endorsed product and endorsed charitable cause, simple linear regression analyses were performed. The results revealed a significant influence for CONSR-EP on consumer attitude toward an endorsed product [F(1,409)=10.267, p< .001, b= .453, R2= 20.5 percent]. Similarly, the results revealed a significant influence for CONSR-ECC on consumer attitude toward an endorsed charitable cause [F(1,408)=9.453, p< .001, b= .424, R2= 18 percent]. These results suggest that CONSR-EP and CONSR-ECC fit are useful as predictors, rather than moderators of consumer attitudes. Consumer Fit Variables Consumer?Endorsed Product Fit Consumer?Endorsed Charitable Cause Fit Celebrity Endorser-Consumer Fit Celebrity Fit Variables Celebrity Endorser?Endorsed Product Fit Celebrity Endorser?Endorsed Charitable Cause Fit Purchase Intention for a Celebrity Endorsed Charitable Cause Product 76 Additional examination of CONSR-EP/CONSR-ECC fit as possible predictors of purchase intention was done to see if variance in purchase intention could be further explained. A multiple regression analysis was performed to evaluate the influence of CONSR-EP fit and CONSR-ECC fit on purchase intention. The results revealed a significant influence for CONSR- EP fit [F(2,406)=133,328, p<.001, b= .411]; and for CONSR-ECC fit [F(2,406)=133.328, p<.001, b=.282]. Together, the consumer fit variables, CONSR-EP and CONSR-ECC fit, explained 39.6 percent of the variance in PI, which reflects their importance in directly explaining purchase intentions toward celebrity endorsed charitable cause products. Next, questions asked during the demographic portion of the online questionnaire were computed as fixed factor variables to further understand the relationships and influence on consumer attitude and purchase intention. Participants answered questions regarding their age, ethnicity, state of residence, current college or university enrollment, and income range. Each demographic variable was analyzed using ANCOVA with the hypothesized independent variables in the model. The fixed factor of age was computed into a nominal variable using a median split. Four ANCOVA tests were performed on consumer attitude toward an endorsed product and consumer attitude toward an endorsed charitable cause using the original categorical fixed factors, CE-EP, CE-CONSR, and CE-ECC and the new fixed factor of Age. The Levene's test for equality of variances was significant for CE-CONSR and Age on consumer attitude toward an endorsed product [F(3,401)=3.036, p=.029]; significance was also found for CE- CONSR and Age on consumer attitude toward an endorsed charitable cause [F(3,401)=2.994, p=.031]. For these variables, the conditions between them are significantly different and cannot be measured using ANCOVA. However, the Levene's test for equality of variances indicated that the interaction between CE-EP and AGE [F(3,401)=1.015, p=.386] and the interaction between 77 CE-ECC and AGE [F(3,401)=1.380, p=.248] were not significant. Although the conditions for ANCOVA were met, there were no significant results with AGE as a fixed factor on either consumer attitude toward an endorsed product [F(1,403)=.341, p=.560] or consumer attitude toward an endorsed charitable cause [F(1,403)=1.431, p=.232]. The exact same procedure used for AGE was used to determine if ethnicity, state of residence, and current enrollment status would provide significant results on consumer attitudes. Levene's tests (Levene, 1960) for equality of variances were analyzed before proceeding with ANCOVA. The results revealed non-significant influence for the additional demographic variables. Therefore, demographic items age, ethnicity, income, state of residence, and enrollment status were not useful moderators or predictors of consumer attitudes or purchase intention. Participants were asked a series of questions regarding their frequency of donation to charitable causes and how likely they were to donate to a charitable cause. The frequency of donation was computed into a nominal fixed factor by grouping those who answered with a frequency of donation in one group and those who never donated in another. Donation frequency behavior was evaluated using ANCOVA, simple linear regression, and multiple regression analysis. The results revealed that the condition for ANCOVA for donation behavior as a covariate on the influence of consumer attitude toward an endorsed charitable cause on purchase intention was not met [F(36,353)=2.193, p<.000]. Levene's test for equality of variances revealed the same significant result for donation behavior as a covariate on the influence of consumer attitude toward an endorsed product on purchase intention [F(21,368)=3.001, p=.248]. This indicates that the conditions in donation behavior and of consumer attitude toward an endorsed charitable cause are not homogeneous but are very 78 different. Further analysis using simple linear regression revealed: 1) a significant influence for donation behavior on consumer attitude toward an endorsed charitable cause [F(1,388)=6.951, p=.009, b= .133, R2= 1.8 percent]; 2) a non-significant influence for donation behavior on consumer attitude toward an endorsed product [F(1,388)=3.435, p=.065, b= .094, R2= .9 percent]; and 3) significant influence for donation behavior on purchase intention [F(1,388)=9.215, p=.003, b= .152, R2= 2.3 percent]. Although these findings for the influence of donation behavior on consumer attitude toward an endorsed charitable cause and purchase intention are significant, donation behavior accounts for very little of the variance suggesting that donation behavior is only a small factor in the formation of positive consumer attitude and purchase intention. 79 Chapter 5: Discussion and Conclusions This chapter discusses the results from hypothesis testing in relation to the literature and the theoretical framework on which the current study was based. Additional analyses on unexplored relationships, not hypothesized in the original model, were performed to seek a better understanding of the influence of fit relationships on consumer attitude and purchase intention. The limitations that exist within this study, along with the theoretical and managerial implications of the findings, are also discussed in this chapter. Finally, the paper includes suggestions for future research that could build upon and extend the findings of the current study. Consumer Fit Variables Consumer fit variables were hypothesized to moderate the effect of consumer perception of celebrity fit on the formation of attitude and purchase intention for a CECCP. The moderating role of consumer fit was partially supported in this study. Additional analysis of direct influence of consumer fit variables revealed that consumer fit explains 39.6 percent of the variance in purchase intention. Compared to 29.1 percent of variance in purchase intention explained by celebrity fit variables, the analyses identified CONSR-EP fit and CONSR-ECC fit as the main predictors of purchase intention for a CECCP. These results, although divergent from the hypotheses presented, are significant in light of self-concept literature. Consumer-endorsed product fit (CONSR-EP), consumer- charitable cause fit (CONSR-ECC). The moderating role of consumer fit was proposed based on the adapted 80 Model of Consuming Behavior (Figure 2). For example, this study used a scenario of Anne Hathaway?s (AH) endorsement of the American Red Cross (ARC) and the promotion of an American Red Cross Signature Red Scarf to illustrate the adapted Model of Consuming Behavior. The scenario initiates communication between AH and the consumer through her endorsement of ARC and the Signature Scarf. According to the model, the consumers? response has the potential to enhance self-concept for the consumer and AH, as well as generate communication between the consumer and AH. These initial findings revealed that consumer perceived fit with the symbolic goods (ARC & Signature Scarf) is significant in influencing consumer attitudes towards the endorsed cause and product. It is unclear however, if the consumer fit hypotheses (H2 and H6) would have been fully supported had the manipulations of the CE fit been successful. The additional analyses performed to examine the influence of the consumer fit variables revealed significant direct relationships with consumer attitudes as well. CONSR-EP and CONSR-ECC explained 20.5 and 18 percent, respectively, of the variance in consumer attitudes. This finding is consistent with the conceptual model?s depiction of the Theory of Planned Behavior (Ajzen, 1991) where the formation of attitudes is one variable that predicts behavioral intent. These significant findings are also supported by the literature referenced in this paper that consumers prefer symbolic goods that are congruent with their self-concept (Birdwell, 1968; Elliott & Wattanasuwan, 1998; Govers & Schoormans, 2005; Grubb & Hupp, 1968; Sirgy, 1982). These findings are also consistent with the original Model of Consuming Behavior (Grubb & Grathwohl, 1967; Figure 1) and the Theory of Symbolic Consumption (Grubb & Grathwohl, 1967). The theory postulates that consumers have a self-concept, they value their 81 self-concept, and perform behaviors directed toward enhancing their self-concept. Intent to purchase a CECCP is a self-concept enhancing behavior. Consumer preference for symbolic goods that fit or are congruent with the consumers? self-concept is also consistent with the findings of prior empirical studies (Birdwell, 1968; Elliott & Wattanasuwan, 1998; Govers & Schoormans, 2005; Grubb & Hupp, 1968; Sirgy, 1982). This study specifically measured CONSR-EP fit and CONSR-ECC fit with scale items designed to identify the consumers? perceptions of endorsed product/endorsed charitable cause congruity across items relative to their own personality (Govers & Schoormans, 2005). The scale items are based on applying human personality characteristics to a product and measuring how congruent the product?s personality is to the consumer?s personality. Govers and Schoormans (2005) found self-image congruence occurs when product-personality is positively related to consumer preference for products. These results offer support for Govers and Schoorman?s (2005) findings by indicating that charitable causes can have personality characteristics that are desirable and match the consumers? self-image. No previous study has used the product-personality scale as an approach to understanding CONSR-ECC fit. Nor has the scale been used to understand the consumers? preference for purchasing a CECCP as it relates to his/her self-image. Celebrity Fit Variables As discussed in the previous chapter, the CE, the EP, and the ECC were the manipulated variables in pretests. Contrary to the hypotheses, the results of the current study suggest that consumer attitudes towards endorsed products and endorsed charitable causes are not influenced by the hypothesized celebrity fit relationships; instead these predictor variables directly explain 29.1 percent of the variance in purchase intention. 82 Celebrity endorser ?endorsed product fit (CE-EP), celebrity endorser ? endorsed charitable cause fit (CE-ECC). The independent t-tests revealed that the celebrity endorsers (Anne Hathaway and Leona Lewis); product categories (signature colored winter scarf and tin of chocolate covered pretzels); and charitable causes (American Red Cross and VHI Save the Music Foundation) were not successful in manipulating high and low levels of CE-EP fit and CE-ECC fit. Lee and Thorson (2008) suggested that a moderate level of mismatch (moderate incongruence rather than either high congruence or high incongruence) produces the most favorable purchase intention, with high congruency and incongruence being evaluated lower. Their findings are consistent with (or, at least, are not inconsistent with), the non-significant findings of this study, which employed high and low (extremes) of celebrity-product congruity but did not include an example of moderate congruency. It is possible that including moderate congruency would have created a greater contrast between low and moderate congruency, as well as high and moderate congruency. The Lee and Thorson (2008) study also found that a moderate level of mismatch between the celebrity endorser and product results in positive purchase intention when the endorsed product is inexpensive or low risk. The current study did not conduct pretests to measure the manipulated products for value (expensive vs. inexpensive) or financial risk (high vs. low). Consequently, it is impossible to determine whether the non- significant findings for both CE-EP fit and CE-ECC fit for the products in this study could be (to some degree) attributable to consumers? perceptions of these factors. Despite the unsuccessful manipulation checks, the CE-EP and CE-ECC fit variables were found to have a significant direct influence on purchase intention. Although CE-ECC fit was significant in influencing purchase intention, the result is an unexpected negative correlation. The literature search did not reveal any precedent for this unexpected finding. The negative beta 83 coefficient could be explained by consumers? perceptions that the celebrity endorsement was disingenuous. However, without additional studies, the interpretation of this significant finding is speculative. Celebrity endorser-consumer fit (CE-CONSR). The results from the independent t-test revealed that the celebrity ? consumer manipulation was not successful. Anne Hathaway and Leona Lewis were selected via pretests based on consumers? perceptions of equal attractiveness but disparate levels of relevance as measured in pretests. This study used the concept of parasocial relationships, or one sided relationships, to understand the relative influence of the consumers? perceived level of fit with the CE on their formation of attitude and purchase intention for the celebrity endorsed products and endorsed charitable causes. Previous studies have suggested that consumers tend to engage in celebrity worship (parasocial relationships) to a greater degree with celebrities with higher measures of attractiveness. For example, North and Sheridan (2009) found that perceived attractiveness positively correlated with the intense-personal items on the CAS. In the current study, both celebrities were identified as relatively ?attractive,? and thus it was predicted that study participants would exhibit high scores on the CAS for both celebrities. However, generally, speaking, consumer scores on the CAS were rather low [M=2.221, SD=1.607], tending towards the ?Strongly Disagree? direction of the scale. Thus, the study did not provide support for the manipulated CE-CONSR fit variable in the model, suggesting that the use of CAS measure as a manipulation check to verify perceived levels of congruency was a poor fit for this study. That said, the use of the CAS in this study proved somewhat useful in that, as discussed in more detail below, it did tend to confirm the findings of Hung et. al. (2011) that higher scores on the CAS are correlated with greater purchase intention. 84 In additional analysis of the celebrity-consumer fit variable, multiple-regression analysis found that CE-CONSR fit was significant in influencing purchase intention. These results may be better explained by the adapted Model of Consuming Behavior and the celebrity-consumer interaction than the consumers? tendency to worship either Anne Hathaway or Leona Lewis. The results that celebrity fit variables have an influence on consumers? purchase intent of a CECCP is relevant in consideration of future studies. Within the context of this study, celebrity fit variables explained less of the variance in purchase intention than consumer fit variables. This finding suggests that consumers were likely to purchase a celebrity endorsed charitable cause product based on their individual fit with the charitable cause and product. This suggests that the consumer valued his/her self-image congruency with the charitable cause product more than their congruency with the celebrity endorser. Consumer Attitude and Purchase Intention The dependent measures in this study were consumer attitude towards an endorsed product, consumer attitude towards an endorsed charitable cause, and consumer purchase intention for a CECCP. Attitudes toward endorsed product and endorsed charitable cause were measured as separate variables, rather than a single consumer attitude variable, to determine the relative influence of consumer attitude toward an endorsed product and consumer attitude toward an endorsed charitable cause on purchase intention. The results of multiple regression analysis suggest consumer attitude toward an endorsed product has more relative influence on purchase intention than consumer attitude toward an endorsed charitable cause. This paper referred to the Theory of Planned Behavior (Ajzen, 1991) in hypothesis development to explain the formation of consumer attitudes and purchase intent. In the multiple regression analysis, consumer attitude toward an endorsed product was the only variable 85 revealed to be a significant predictor of purchase intention. Possible explanations for these findings may relate to the positive consumer-product fit influence on consumer attitude toward an endorsed product. Govers and Shoormans (2005) found consumers? preference was positive when the products? personality was congruent with the consumers? personality. In the TPB, attitude is one of the factors that influence behavioral intent. While there was a positive influence from CONSR-ECC fit on consumer attitude toward an endorsed charitable cause, consumer attitude toward an endorsed charitable cause was not significant in predicting purchase intention. Yurchisin et. al., (2009) found attitude toward a CRM fashion product did not influence the predicted purchase behavior. However, Yurchisin et al. (2009) attribute the mass appeal for fashion related CRM products to consumer perceptions of the product?s consistency with fashion trends and popular celebrities. For consumer attitude toward an endorsed product, the products? association with fashion trends and popular celebrity endorsers may have been stronger than perceptions of the charitable causes? association to the celebrity endorsers. Discussion of Non-Hypothesized Significant Findings Additional analyses on unexplored relationships, not hypothesized in the original model, were performed to seek a better understanding of the influence of fit relationships on consumer attitude and purchase intention. Demographic questions provided additional data to be used in understanding the fit relationships and donation behavior was identified as having a significant influence on purchase intention. The study evaluated donation behavior and considered the possible relationships it had within this model. Donation behavior was found to significantly influence purchase intention. The study employed a sample of millennials (ages 19-28), a group identified in reviewed literature as being likely to purchase a good related to a charitable cause (Cone Inc., 2010a). The 86 positive relationship between donation behavior of female millennial consumers and purchase intention for a CECCP is consistent with literature concerning postmaterialism. Postmaterialists believe in making ethical and socially responsible decisions (Giacalone et al., 2008). One can surmise from this result that the female millennial consumer is likely to purchase a CECCP because she believes it to be an ethical and socially responsible decision. Limitations Despite its interesting hypothesized and non-hypothesized findings, this study?s results must be considered with respect to its limitations. First, consumer fit with the endorsed product and charitable cause were not manipulated in pretesting. Specifically, the products presented as stimuli were chosen by the author based on pretests that concluded clothing and food products had the best/least fit with a set of potential celebrity endorsers. The specific stimuli chosen?a winter scarf and a tin of gourmet chocolate-covered pretzels?appeared to be consistent with product endorsements found on shoppingblog.com. A third pretest, as used by Govers and Schoormans (2005), could have evaluated several variants of scarfs with personality characteristics to determine which scarf had the best fit with the consumer. Similarly, charitable causes were evaluated only for congruence with potential celebrity endorsers. A set of personality characteristics could have been applied to a wider set of charitable causes (congruent with the chosen celebrities) to determine if there was one that had a better fit with the consumer. Next, this study used a within-subjects methodology during pretesting and a between- subjects methodology for the main study. In pretesting, the within-subjects method provided examples of all the potential stimuli and the coordinating questions. Pretest 1 asked questions about attractiveness, whereas Pretest 2 asked participants to evaluate congruence. The main study employed a between-subjects method that allowed the participant to view only one 87 condition with the selected stimuli. Since only one set of stimuli was viewed, the consumer was not influenced by experiences and beliefs of other celebrity endorsers, endorsed charitable causes, or endorsed products. This could explain why the manipulations were supported in the within-subjects pretests but not in the between-subjects experiment. Lastly, the study?s utility in evaluating celebrity-consumer fit relationships may have been limited by the use of the scale addressing parasocial celebrity-consumer fan relationships and the CAS. This study identified female millennials in an attempt to understand if the proliferation of celebrity endorsements is a response to consumers who would score higher on the CAS scale. The scale items used from the intense/personal CAS subscale were revised to eliminate items that were not considered relevant or were considered too intense for the purpose of this study. The results of this measure suggest that most consumers indicated low ratings on all the items, thus eliminating variability in this measure. This resulted in a poor check of the manipulation of fit between the celebrity endorser and consumer, leading to inconclusive findings on the success of the manipulations. Theoretical and Managerial Implications This study examined a holistic model of the fit relationships between: a) consumers and celebrity endorsers, b) consumers and endorsed charitable causes, c) consumers and endorsed products, d) celebrity endorsers and endorsed charitable causes, and e) celebrity endorsers and endorsed products. These relationships were supported by prior empirical studies and theories. The results of this study offer theoretical contributions to the current body of knowledge for understanding fit relationships, attitude, and purchase intention for a CECCP. The most significant findings of this study, although not hypothesized in the conceptual model, are supported by the theoretical framework used in hypothesis development. Specifically, 88 the findings of this study revealed CONSR-EP fit and CONSR-ECC fit relationships to be the most significant predictor variables of purchase intention. The results extend the work of prior empirical self-concept studies (Birdwell, 1968; Elliott & Wattanasuwan, 1998; Govers & Schoormans, 2005; Grubb & Hupp, 1968; Sirgy, 1982). The significant findings also support the Theory of Symbolic Consumption, Model of Consuming Behavior, and the adapted model by explaining how consumers and celebrity endorsers use the symbolic goods (endorsed product and endorsed charitable cause) to explain how self-concept and behavior interact. The original Model of Consuming Behavior (Grubb & Grathwohl, 1967) is significant because it outlines clearly the role of the consumer in a consumption opportunity. The significant results for consumer fit variables extended the concept of symbolic good to charitable causes. Prior studies had not examined charitable causes as a good to be consumed and thus had not evaluated charitable causes as a consumable symbolic good that could enhance self-concept. The adapted Model of Consuming Behavior proposed in this study posits that celebrity endorsers enter into endorsement opportunities with charitable causes to enhance their self-concept. Similarly, this proposed adaptation is significant because prior studies had not addressed celebrity endorser actions, communications, and outcomes within the context of a charitable cause. The significant finding that celebrity fit variables influence consumer purchase intention for a CECCP is supported in the adapted model. This study presents managerial implications for marketing a CECCP. First, this study found purchase intention toward CECCP was influenced most by consumer fit factors (CONSR- EP and CONSR-ECC). As charitable causes seek to capture consumer attention, this finding implies that a CRM campaign is likely to garner the support of consumers when the charitable cause and associated product fit the consumers? self-concept. Even if a CRM employs a celebrity 89 endorser, the potential consumer is more likely to consider their fit with the charitable cause and associated product than their fit with the celebrity endorser in determining their purchase intent. However, it is possible that having a celebrity endorser raises the consumers? awareness of the charitable causes? mission and associated products available for purchase. Future studies would need to consider the influence of celebrity endorsers in earlier phases such the ?consumer awareness? phase of the CRM campaign and compare it with their influence in the ?consumer decision-making? phase. Next, consumer fit?s influence on purchase intention is significant because it supports prior Cone Inc. survey findings. Specifically, the 2010 Cone Nonprofit Marketing Trend Tracker, found 59 percent of all Americans are more likely to purchase a product if it is associated with a charitable cause (Cone, Inc., 2010b). Consistent messages, as exemplified by Feed Projects, indicate that charitable causes that market their own products have an opportunity to increase revenues when there is CONSR-EP and CONSR-ECC. CRM campaigns are also likely to find success with the female demographic group when the charitable cause and associated product match the consumers? personality. In the 2010 Cone Cause Evaluation survey, 18-24 year old consumers were the most likely group to purchase a good that benefits a cause (Cone Inc., 2010a). This study used a sample composed of female millennials, consistent with the Cone study. The female millennial response in this study may also suggest that this powerful consumer group is aware of marketing techniques that use celebrity endorsers to attract their attention and purchasing power. The consideration that celebrity fit variables were not perceived as having the strongest influence on purchase intents suggests that this consumer group acts cautiously before supporting products and charitable causes by considering whether the charitable cause product reinforce their self-concept. 90 Lastly, consumer donation behavior was found to be a small but significant as a predictor of purchase intention. This implies that consumers who have previously contributed to a charitable cause at least once are more likely to purchase a CECCP than consumers who have never donated to a charitable cause. Consistent with the 2010 Cone Cause evaluation survey, 41 percent of consumers said they had previously purchased a charitable cause product before (Cone Inc., 2010a). This indicates that CRM campaigns should target consumers who have previously donated to charitable causes. Although the finding suggest consumer fit variables were more influential in predicting purchase intent, the finding for celebrity fit factors (CE-EP and CE-ECC) is consistent with the reviewed literature that attractive celebrities influence consumer purchase intentions (Friedman and Friedman, 1979). It can be inferred that attractive celebrity endorsers with good fit between product and cause will have some influence on consumer purchase intentions for a CECCP. Traditional marketing uses celebrity endorsers to market products across a wide range of categories from luxury to every day products and services. CE-CONSR was significant in influencing consumer purchase intention for a CECCP. Although not as strong as consumer fit factors that influence purchase intention, the link between celebrity endorsers and consumers may be explained by a shared postmaterialism world view. Consumers may be sensitive to the socially responsible actions outlined in the stimuli scenario (Giacalone et al., 2008). Celebrity endorsers? postmaterialism values may lead consumers to adopt and purchase CECCP, which has implications for marketing strategies. Successful CRM campaigns should consider the alignment of celebrities? and consumers? desires to make socially ethical decisions and enhance their own self-concepts (Giacalone et al., 2008). 91 Suggestions for Future Research This research considered how celebrity endorser ? cause related marketing influenced consumer response to CECCP. The results of this study provide several suggestions for future research to augment the literature on celebrity endorsers, CRM, Theory of Planned Behavior, and the Theory of Consuming Behavior. Figure 6 illustrates a modified version of this study?s conceptual model. Regression analyses employed to evaluate the original conceptual model found significant predictors of purchase intention but did not establish causation of purchase intention. The modified model presents a causal model to examine the effects of the three groups of fit constructs on purchase intention. The three groups in the modified model are based on the hypotheses developed in this paper. Consumer fit, Celebrity-Consumer fit, and Celebrity fit are suggested constructs that influence PI. As noted in the limitations, the measures used did not fully capture the extent to which these variables influenced consumer attitude or purchase intention. Future studies could benefit through the use of other measures to evaluate these fit constructs. Structural Equation Modeling could be used to evaluate if the new model is valid and has a better fit with data. This study examined consumer fit with an endorsed product and an endorsed charitable cause independently to determine if there was a dominant variable predictor of purchase intention for a CECCP. Product personality measures were adapted for this study to measure the influence of CONSR-EP fit and CONSR-ECC fit on consumer attitudes. Noted in the limitations, this measure was not implemented in pretesting. Therefore, an expanded use of this measure may provide greater insight into influence on purchase intention. Additional variables of fit may also be considered within this construct such as consumer-audience fit. This suggested variable is illustrated in the original Grubb and Grathwhol (1967) model of consuming behavior and can be 92 examined within the framework of the TPB?s subjective norms. Greater attention to the influence a consumers? peer group has on the consumers? purchase intention may provide valuable theoretical and managerial implications. Parasocial relationships dominated the study of celebrity-consumer fit in this study. However, the findings were unclear due to poor manipulation checks. An expansion of this construct could include previous exposure to the celebrity endorser. Celebrities as brands were recalled in this study to support the concept that the celebrity embodies meanings that may transfer to the endorsed object. Measures of brand familiarity, brand personality, brand knowledge, and brand recall may be adapted in context to celebrities to provide greater insight into how celebrity-consumer fit influences purchase intention. With additional research and analysis, the theoretical and conceptual models presented in this paper will aid marketers in identifying the proper fit among the celebrity endorser, the charitable cause, the endorsed product, and the target market. Future research should consider consumer motivations for purchase decisions in the specific context of celebrity endorsers and charitable causes. Likewise, future studies should evaluate within the context of CRM if celebrity endorsers are useful in increasing a charitable causes? traction with a target market. The theoretical models of Grubb and Grathwohl (1969) and Ajzen (1991) provide a foundation for understanding why consumers enter into a purchase decision. Further research will lead to the development of celebrity endorser CRM techniques that connect the charity?s mission to achieve its philanthropic goals, the celebrity?s desire to enhance her self-image, and the consumer?s need to fulfill her behavioral intentions to purchase a celebrity endorsed charitable cause product. 93 References Aaker, J.L., (1997). Dimensions of brand personality. Journal of Marketing Research, 34(3), 347-356. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179-211. Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In J. Kuhl & J. Beckman (Eds.), Action-control: From cognition to behavior (pp. 11- 39). Heidelberg, Germany: Springer. Allison, N. K., Golden, L. L., Mullet, G. M., & Coogan, D. (1980). Sex-typed product images: The effects of sex, sex role self-concept and measurement implications. Advances in Consumer Research, 7, 604-609. Ashe, D. D., & McCutcheon, L. E. (2001). Shyness, loneliness, and attitude toward celebrities. Current Research in Social Psychology, 6, 124-133. Baker, M. J. & Churchill, G. A., Jr. (1977). The impact of physically attractive models on advertising evaluations. Journal of Marketing Research, 14, 538-555. Barton, N., & Hall, H. (2010, October 17). Donations dropped 11% at nation's biggest charities last year. Retrieved from: http://philanthropy.com/article/A-Sharp-Donation-Drop-at- Big/125004 Becker-Olsen, K. L., Cudmore, A., & Hill, R. P. (2006). The impact of perceived corporate social responsibility on consumer behavior. Journal of Business Research, 59, 46-53. Beer, R. (2010, March 31). Grazia world exclusive: Thandie Newton signs as new face of Olay total effects. Retrieved from: http://www.graziadaily.co.uk/beauty/archive/2010/03/31/grazia-world-exclusive-- thandie-newton-signs-as-new-face-of-olay-total-effects.htm 94 Bhatti,V. (2009, July 7). Kiehl?s and Brad Pitt make it right with JPF Eco Systems. Retrieved from: http://optimistworld.com/Kiehls-Brad-Pitt-Make-It-Right-JPF-Eco-Systems/ Birdwell, A. E. (1968). A study of the influence of image congruence on consumer choice. Journal of Business, 41(1), 76-88. Bower, A. & Landreth, S. (2001), Is beauty best? Highly vs. normally attractive models in advertising. Journal of Advertising, 30 (1), 1-14. Bremner, R. H. (1994). Giving: Charity and philanthropy in history. New Brunswick, NJ: Transaction Publishers. Cailleba, P. & Casteran, H. (2010). Do ethical values work? A quantitative study of the impact of fair trade coffee on consumer behavior. Journal of Business Ethics, 97(4), 613-624. Chiou, J. S., Huang, C. Y., & Chuang, M. C. (2005). Antecedents of Taiwanese adolescents? purchase intention towards the merchandise of a celebrity: The moderating effect of celebrity adoration. Journal of Social Psychology, 145(3), 317?332. Cohen, J., Cohen, P., West, S.G., & Aiken, L.S. (2003). Applied multiple regression/correlation: Analysis for the behavioral sciences, 3rd ed. Mahwah, NJ: Lawrence Erlbaum Associates. Cone, Inc. (2010a). 2010 Cone Cause Evolution study, Boston, MA. Cone, Inc. (2010b). 2010 Cone Nonprofit Marketing Trend Tracker, Boston, MA. Coolspotters.com (2011a). Feed bag and celebrities. Retrieved from http://coolspotters.com/handbags/feed-bag/and/celebrities Coolspotters.com. (2011b). Tom?s shoes and celebrities. Retrieved from http://coolspotters.com/shoes/toms-shoes-canvas-slip-ons/and/celebrities Crane, A. & Matten, D. (2004). Business ethics: A European perspective. Oxford University Press, Oxford, UK. De Backer, C. J. S., Nelissen, M., Vyncke, P., Braeckman, J., & McAndrew, F. T. (2007). Celebrities: From teachers to friends?A test of two hypotheses on the adaptiveness of celebrity gossip. Human Nature, 18, 334?354. de Waal, A. (2008). The humanitarian carnival. World Affairs, 43-55. Dolich, I.J. (1969). Congruence relationship between self image and product brands. Journal of Marketing Research, 6 (1), 80-4. 95 Elliott, R. & Wattanasuwan, K. (1998). Brands as symbolic resources for the construction of identity. International Journal of Advertising, 17(2), 131-144. Edrogan, B. Z. (1999). Celebrity endorsement: A literature review. Journal of Marketing Management, 15(4), 93-105. Erdogan, E.Z., Baker, M.J. & Tagg, S. (2001). Selecting celebrity endorsers: The practitioner?s perspective, Journal of Advertising Research, 41(3), 39-48. Feed Foundation, The. (2011). About us. Retrieved from: http://www.thefeedfoundation.org/Who-We-Are Fenn, D. (2010, July 19). Inc.com's 2010 30 under 30 list: Lauren Bush and Ellen Gustafson, founders of FEED Projects. Retrieved from: http://www.inc.com/30under30/2010/profile-lauren-bush-ellen-gustafson-feed- projects.html Finch, C. (2010, October 3). Changing the world...one brand at a time: Feed. [blog]. Retrieved from http://webpub.allegheny.edu/employee/I/iroy/weblogcmrt260fa2010/2010 Fishbein, M., & Ajzen, I. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice Hall. Fishbein, M., & Ajzen, I. (1975). Beliefs, attitude, intention and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley. Fofana, F. (2009, September 24). LTTS exclusive: Celebrities bowl for sickle cell disease awareness. Retrieved from: http://www.looktothestars.org/news/3147-ltts-exclusive- celebrities-bowl-for-sickle-cell-disease-awareness Francis, J.J., Eccles, M.P., Johnston, M., Walker, A., Grimshaw, J., Foy, R., Kaner, E.F.S., Smith, L. & Bonetti, D. (2004). Constructing questionnaires based on the Theory of Planned Behaviour. University of Newcastle upon Tyne: Centre of Health Services Research, Newcastle upon Tyne, UK. Friedman, H. & Friedman, L. (1979). Endorser effectiveness by product type, Journal of Advertising Research, 19, 63-71. Fromherz, K. (2006). Cause-related marketing. American Nurseryman, 203(6), 46-50. Gabor, A., Thorton, J., & Wienner, D. P. (1987, December 7). Star turns that can turn star- crossed. U.S. News and World Report. pp. 57. Giacalone R.A., Jurkiewicz, C.L., & Deckop, J.R. (2008). On ethics and social responsibility: The impact of materialism, postmaterialism, and hope. Human Relations, 61(4), 483-515. 96 Gotschi, E., Vogel, S., Lindenthal, T., & Larcher, M. (2010). The role of knowledge, social norms, and attitudes toward organic products and shopping behavior: Survey results from high school students in Vienna. The Journal of Environmental Education, 41(2), 88?100. Govers, P. C. M., & Mugge, R. (2004). ?I love my Jeep, because its tough like me?: The effect of product-personality congruence on product attachment. In A. Kurtg?z? (Ed.), Proceedings of the Fourth International Conference on Design and Emotion. Ankara, Turkey. Govers, P.C.M & Schoormans J.P.L. (2005). Product personality and its influence on consumer preference. Journal of Consumer Marketing, 22(4), 189-197. Grounds, J. (2005). Special issue on charity branding. International Journal of Nonprofit & Voluntary Sector Marketing, 10, 65-67. Grubb, E. L. & Grathwohl, H. L. (1967). Consumer self-concept, symbolism, and market behavior: A theoretical approach. Journal of Marketing, 31, 22?27. Grubb, E.L. & Hupp, G. (1968). Perception of self, generalized stereotypes and brand selection. Journal of Marketing Research, 5(1), 58-63. Hall, C. S., & Lindsay, G. (1957). Theories of personality. New York: Wiley. Herrera, M. (2009, July 3). Michael Jackson, Pepsi made marketing history. Retrieved from http://www.reuters.com/assets/print?aid=USTRE56252Z20090703 Hoovers.com. (2011). Apparel and accessories retail industry description. Retrieved from: http://www.hoovers.com/industry/apparel-accessories-retail/1519-1.html Horton, D., & Wohl, R. R. (1956). Mass communication and para-social interaction: Observations on intimacy at a distance. Psychiatry, 19, 215-229. Hovland C. I., Irving, J. L. & Harold K. H. (1953). Communication and Persuasion. New Haven, CT: Yale University Press. Howard, A. (1979, July 30). More than just a passing fancy. Advertising Age, S-2. Howe, N., & Strauss, W., (2001). Millennials rising: The next great generation. New York: Vintage. Hung, K., Chan, K.W., & Tse, C.H. (2011). Assessing celebrity endorsement effects in China: A consumer-celebrity relational approach. Journal of Advertising Research, 51(4), 608-623. Inglehart, (1977). The silent revolution: Changing value and political styles among Western publics. Princeton, NJ: Princeton University Press. 97 Jekielek, S. & Brown, B. (2005). The transition to adulthood: Characteristics of young adults ages 18-24 in America, www.aecf.org. Jones, P. (2010a, November 30). Target's gift card cause marketing for St. Jude. Retrieved from http://www.causemarketing.biz/ Jones, P. (2010b, November 26). Salvation Army kettle drive kickoff during the Dallas Cowboys Thanksgiving day game. Retrieved from http://www.causemarketing.biz/ Jones, P. (2010c, November 23). Give your cause marketing a little extra punch with a coupon hangtag. Retrieved from http://www.causemarketing.biz/ Jones, P. (2010d, November 5). Cause marketing the happy meal. Retrieved from http://www.causemarketing.biz/ Josephson, N. (1984). American Express raises corporate giving to marketing art. Advertising Age, 55(4), 10-11. Kahle, L.R. & Homer, P.M. (1985) Physical attractiveness of the celebrity endorser: A social adaptation perspective. Journal of Consumer Research, 11, 954-961. Kamins, M.A. (1990.) An investigation into the match-up hypothesis in celebrity advertising: When beauty may be only skin deep. Journal of Advertising, 18(2), 4-13. Kamins, M.A. & Gupta, K. (1994). Congruence between spokesperson and product type: A matchup hypothesis perspective. Psychology & Marketing, 11(6), 569-586. Keller, K.L. (1993). Conceptualizing, measuring and managing customer-based brand equity. Journal of Marketing, 57(1), 1-22. Kotler, P. (2003). Marking insights from A-Z: 80 concepts every manager needs to know. Hoboken, NJ: John Wiley & Sons. LaVallee. A. (2009, November 20). AT&T counters Verizon ads with Luke Wilson?s help. Retrieved from: http://blogs.wsj.com/digits/2009/11/20/att-counters-verizon-ads-with- luke-wilsons-help/ Lee, J.G., & Thorson, E. (2008). The impact of celebrity product incongruence on the effectiveness of product endorsement. Journal of Advertising Research, 48(3), 433-449. Levene, H. (1960). Robust tests for the equality of variances. In I. Olkin (Ed.), Contribution to probability and statistics(pp. 278?292). Palo Alto, CA: Stanford University Press. Macy?s Inc. (2010, June 29). Macy?s continues tradition of fashion and compassion with passport presents glamorama. Retrieved from: http://www.macysinc.com/pressroom/macys/press.aspx?cid=3&scid=32&mkid=567 98 Maltby, J., Houran, J., Lange, R., Ashe, D.D., & McCutcheon, L.E. (2002). Thou shalt worship no other gods?unless they are celebrities. Personality and Individual Differences, 32, 1157-1172. McCarley, N.J. & Escoto, C.A., (2003). Celebrity worship and psychological type. North American Journal of Psychology, 5(1), 117-120. McCracken, G. (1989). Who is the celebrity endorser? Cultural foundations of the endorsement process. Journal of Consumer Research, 16, 310-321. McCracken, G. (1986). Culture and consumption: A theoretical account of the structure and movement of cultural meaning of consumer goods. Journal of Consumer Research, 13(1), 71-84. McCutcheon, L. E., Ashe, D. D., Houran, J., & Maltby, J. (2003). A cognitive profile of individuals who tend to worship celebrities. The Journal of Psychology, 137( 4), 309-322. McCutcheon, L. E. & Maltby, J. (2002). Personality attributions about individuals high and low in tendency to worship celebrities. Current Research in Social Psychology, 7, 325-338. McCutcheon, L. E., Lange, R., & Houran, J. (2002). Conceptualization and measurement of celebrity worship. British Journal of Psychology, 93, 67-87. Miller, G.P. & Basehart, J. (1969). Source trustworthiness, opinionated statements and response to persuasive communication. Speech Monographs, 36(1), 1-7. Misra, S. & Beatty, S. E. (1990). Celebrity spokesperson and brand congruence: An assessment of recall and affect. Journal of Business Research, 21, 159-173. Mittelstaedt, J. D., Riesz, P.C., & Burns, W.J. (2000). Why are endorsements effective? Sorting among theories of product and endorser effects. Journal of Current Issues and Research in Advertising, 22(1), 55-66. Mortimore, R. (2010, January 7). Ethical goods extend beyond costly coffee. Retrieved from: http://www.marketingweek.co.uk/ethical-goods-extend-beyond-costly- coffee/3008269.article Motavalli, J. (1988, January 11). Advertising blunders of the rich and famous, Adweek's Marketing Week, 18-19. Munson, J.M. & Spivey, W. A. (1980). Assessing self-concept. In J. C. Olson (Eds), Advances in Consumer Research: Vol. 7. Association for Consumer Research, (pp.598-603). Naughton J. &Born P. (August 18, 2008). A twist on celebrity: Kiehl's goes green with Brad Pitt deal. WWD: Women's Wear Daily, 196(36). 99 North, A. & Sheridan, L.P. (2009). Death, attractiveness, moral conduct, and attitudes to public figures. Omega, 60(4), 351-363. Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers? perceived expertise, trustworthiness, and attractiveness. Journal of Advertising, 19(3), 39- 52. Osgood, C.E., Suci, G. J., & Tannebaum, P. H., (1957).The measurement of meaning. Urbana, IL: University of Illinois Press. Panepento, P. (2006, November 9). Young people buy products tied to social causes, but doubt their value. Retrieved from: http://find.galegroup.com/gtx/infomark.do?&contentSet=IAC- Documents&type=retrieve&tabID=T002&prodId=AONE&docId=A153998222&source= gale&srcprod=AONE&userGroupName=naal_aub&version=1.0 Petty, R. E., Cacioppo, J. T., & Schumann, D. (1983). Central and peripheral routes to advertising effectiveness: The moderating role of involvement. Journal of Consumer Research, 10, 135-146. Pornpitakpan, C. (2003). Validation of the celebrity endorsers? credibility scale: Evidence from Asians. Journal of Marketing Management, 19, 179-195. Rogers, C.R. (1951). Client-centered therapy. Boston, MA: Houghton Mifflin. Rosenthal, J. (2010, January 27). Products with purpose will change the world. Retrieved from http://www.huffingtonpost.com/jeff-rosenthal/products-with-purpose-wil_b_437917.html Shimp, T. E. (1997). Advertising, promotion and supplemental aspects of integrated marketing communication, 4th Edition. Fort Worth, TX: The Dryden Press. Sirgy, M.J. (1982). Self-concept in consumer behavior: A critical review. Journal of Consumer Research, 9(3), 287-300. Sirgy, M.J., Grewal, D., Mangleburg, T.F., Park, J., Chon, K.S., Claiborn, C.B., Johar, J.S. & Berkman, H. (1997). Assessing the predictive validity of two methods of measuring self- image congruence. Journal of the Academy of Marketing Science, 25(3), 229-41. Spielman, H. M. (1981, November/December). The celebrity sell: Making it work. Marketing Times, 28, 13-14. Strahilevitz, M, & Myers, J.G. (1998). Donations to charity as purchase incentives: How well they work may depend on what you are trying to sell. Journal of Consumer Research, 24, 434-446. 100 Target Corporation. (2009, May 20). Target announces new partnership with the Black Eyed Peas. Retrieved from: http://pressroom.target.com/pr/news/target-announces-new- partnership-153105.aspx T-Mobile USA. (2005, June 2). T-Mobile and Catherine Zeta-Jones Agree to 'Get More Advertising?. Retrieved from: http://www.tmobile.com/company/PressReleases_Article.aspx?assetName=Prs_Prs_2005 0602&title=TMobile%20and%20Catherine%20ZetaJones%20Agree%20to%20%27Get %20More%20Advertising%27 Urban Institute, National Center for Charitable Statistics. (2008). The Nonprofit sector in brief: Facts and figures from the nonprofit almanac 2008: Public charities, giving, and volunteering. Washington, DC: Blackwood, J., Wing, K.T., & Pollak, T.H. Vasquez, A. (2008, December 17). Quannum records and Reebok team up to aid Darfur. Retrieved from: http://www.hiphopdx.com/index/news/id.8271/title.quannum-records- and-reebok-team-up-to-aid-darfur Yurchisin, J., Kwon, Y. J., & Marcketti, S.B. (2009). Consumers of charity bracelets: Cause- supporters or fashion-followers? Journal of Fashion Marketing and Management, 13(3), 448-457. 101 Appendix A Defined Abbreviations 102 Table 22 Defined Abbreviations CRM Cause Related Marketing CE Celebrity Endorser CONSR Consumer EP Endorsed Product ECC Endorsed Charitable Cause CE-CONSR Celebrity Endorser ? Consumer CE-EP Celebrity Endorser ? Endorsed Product CE-ECC Celebrity Endorser ? Endorsed Charitable Cause CONSR-EP Consumer ? Endorsed Product CONSR-ECC Consumer ? Endorsed Charitable Cause CECCP Celebrity Endorsed Charitable Cause Product MUH Match-up Hypothesis TPB Theory of Planned Behavior TRA Theory of Reasoned Action 103 Appendix B Pretest Information Letter 104 Auburn University College of Human Sciences Department of Consumer Affairs 308 Spidle Hall Auburn, AL 36849-5601 (334)844-4084 INFORMATION LETTER for a Research Study Pretest entitled ?Cause Related Marketing: Understanding Celebrity, Consumer, and Product Fit Relationships? You are invited to participate in a research study pretest to examine fit relationships between consumers, causes, products, and celebrity endorsers of charitable cause products. The study is being conducted by Whitney Skinner Nunnelley, Ph.D. student, under the direction of Dr. Carol Warfield, Department Head, and Dr. Veena Chattaraman, Associate Professor, in the Auburn University Department of Consumer Affairs. You were selected as a possible participant because you are female and are between the ages of 19-24. If you decide to participate in this research study, you will be asked to click on the ?Next? link at the bottom of this page. This link will lead you to an online questionnaire. Your total time commitment for completing the questionnaire will be approximately 15 minutes. There is no direct benefit to you for participating in the study. There are no foreseeable risks associated with participating in this study. Research will be conducted confidentially. To thank you for your time, after you have completed the survey you will be offered the chance to submit your email address to enter in a drawing to win one of ten cash prizes of $25. It is hoped that findings from this study will increase understanding of students? interaction with female celebrities and with various charitable causes. Your participation is completely voluntary. If you change your mind about participating while you are completing the questionnaire, you can stop filling out the questionnaire and close your browser without submitting the data. Once you have submitted anonymous data, it cannot be withdrawn due to it being unidentifiable. Your decision about whether or not to participate or to stop participating will not jeopardize your future relations with Auburn University or the Department of Consumer Affairs. Your privacy will be protected. Any information obtained in connection with this study will remain anonymous. To protect your privacy, if you choose to submit your email address to enter the prize drawing, your email address will be collected in a separate database from your survey responses. Information collected through your participation may be presented at professional meetings and published in professional journals with no personal identification of the participants. No IP address or any other identifiable data 105 will be recorded as a result of your online survey participation. If you have questions about this study, please contact Whitney Skinner Nunnelley by email, skinnwa@auburn.edu or telephone, 334-844-4084 or my faculty advisors, Dr. Chattaraman, by email vzc0001@auburn.edu or by telephone, 334-844- 3258, or Dr. Warfield by email, warficl@auburn.edu, or telephone 334-844-1329. If you have questions about your rights as a research participant, you may contact the Auburn University Office of Human Subjects Research or the Institutional Review Board by phone (334)-844-5966 or e-mail at hsubjec@auburn.edu or IRBChair@auburn.edu. HAVING READ THE INFORMATION PROVIDED, YOU MUST DECIDE IF YOU WANT TO PARTICIPATE IN THIS RESEARCH PROJECT. IF YOU DECIDE TO PARTICIPATE, PLEASE CLICK ?CONSENT TO PARTICIPATE? TO ACCESS THE QUESTIONNAIRE. YOU MAY PRINT A COPY OF THIS LETTER TO KEEP. The Auburn University Institutional Review Board has approved this document from July 1, 2011 to June 30, 2012. Protocol #11-216 EX 1107. NEXT 106 Appendix C Pretest 1 Online Survey Questions 107 108 109 110 111 112 113 114 Appendix D Pretest 2 Online Survey Questions (Showing one of eight celebrity-cause match-ups and one of eight celebrity-product match-ups) 115 116 117 118 119 Appendix E Main Study Stimuli 120 Table 23 Main Study Stimuli by Condition Condition Celebrity Charitable Cause Product 1) High CE-CONSR Fit High CE-ECC Fit High CE-EP Fit 2) High CE-CONSR Fit High CE-ECC Fit Low CE-EP Fit 3) High CE-CONSR Fit Low CE-ECC Fit High CE-EP Fit 4) High CE-CONSR Fit Low CE-ECC Fit Low CE-EP Fit 5) Low CE-CONSR Fit Low CE-ECC Fit High CE-EP Fit 6) Low CE-CONSR Fit Low CE-ECC Fit Low CE-EP Fit 7) Low CE-CONSR Fit High CE-ECC Fit High CE-EP Fit 8) Low CE-CONSR Fit High CE-ECC Fit Low CE-EP Fit 121 Appendix F Main Study Information Letter 122 Auburn University College of Human Sciences Department of Consumer Affairs 308 Spidle Hall Auburn, AL 36849-5601 (334)844-4084 INFORMATION LETTER for a Research Study entitled ?Cause Related Marketing: Understanding Celebrity, Consumer, and Product Fit Relationships? You are invited to participate in a research to study examine fit relationships between consumers, causes, products, and celebrity endorsers of charitable cause products. The study is being conducted by Whitney Skinner Nunnelley, Ph.D. student, under the direction of Dr. Carol Warfield, Department Head, and Dr. Veena Chattaraman, Associate Professor, in the Auburn University Department of Consumer Affairs. You were selected as a possible participant because you are female and are between the ages of 19-24. If you decide to participate in this research study, you will be asked to click on the ?Next? link at the bottom of this page. This link will lead you to an online questionnaire. Your total time commitment for completing the questionnaire will be approximately 15 minutes. There is no direct benefit to you for participating in the study. There are no foreseeable risks associated with participating in this study. Research will be conducted confidentially. To thank you for your time, after you have completed the survey you will be offered the chance to submit your email address to enter in a drawing to win one of ten cash prizes of $25. It is hoped that findings from this study will increase understanding of students? interaction with female celebrities and with various charitable causes. Your participation is completely voluntary. If you change your mind about participating while you are completing the questionnaire, you can stop filling out the questionnaire and close your browser without submitting the data. Once you have submitted anonymous data, it cannot be withdrawn due to it being unidentifiable. Your decision about whether or not to participate or to stop participating will not jeopardize your future relations with Auburn University or the Department of Consumer Affairs. Your privacy will be protected. Any information obtained in connection with this study will remain anonymous. To protect your privacy, if you choose to submit your email address to enter the prize drawing, your email address will be collected in a separate database from your survey responses. Information collected through your participation may be presented at professional meetings and published in professional journals with no personal identification of the participants. No IP address or any other identifiable data 123 will be recorded as a result of your online survey participation. If you have questions about this study, please contact Whitney Skinner Nunnelley by email, skinnwa@auburn.edu or telephone, 334-844-4084 or my faculty advisors, Dr. Chattaraman, by email vzc0001@auburn.edu or by telephone, 334-844- 3258, or Dr. Warfield by email, warficl@auburn.edu, or telephone 334-844-1329. If you have questions about your rights as a research participant, you may contact the Auburn University Office of Human Subjects Research or the Institutional Review Board by phone (334)-844-5966 or e-mail at hsubjec@auburn.edu or IRBChair@auburn.edu. HAVING READ THE INFORMATION PROVIDED, YOU MUST DECIDE IF YOU WANT TO PARTICIPATE IN THIS RESEARCH PROJECT. IF YOU DECIDE TO PARTICIPATE, PLEASE CLICK ?CONSENT TO PARTICIPATE? TO ACCESS THE QUESTIONNAIRE. YOU MAY PRINT A COPY OF THIS LETTER TO KEEP. The Auburn University Institutional Review Board has approved this document from July 1, 2011 to June 30, 2012. Protocol #11-216 EX 1107. NEXT 124 Appendix G Main Study Scenario Example (Showing one of eight scenarios) 125 126 Appendix H Main Study Online Survey Questions (Showing one of eight conditions) 127 128 129 130 131 132 133 134 135 Appendix I Demographic Survey Questions 136 137 138 Appendix J Main Study Debriefing Page and Optional Email Submission 139