USF Libraries
USF Digital Collections

The effect of celebrity endorsements on gift-giving purchases

MISSING IMAGE

Material Information

Title:
The effect of celebrity endorsements on gift-giving purchases an application of the elaboration likelihood model
Physical Description:
Book
Language:
English
Creator:
Anghel, Christine
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Product involvement
Oprah Winfrey
Advertising
Marketing
Attitudes
Dissertations, Academic -- Mass Communications -- Masters -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: The purpose of this study is to determine how effective celebrity endorsements are in regards to the type of gift purchase one decides to make (i.e., buying for someone who has a high significant meaning to the buyer, such as a best friend, versus buying for someone who has a low significant meaning to the buyer, such as a casual friend). The study seeks to extend upon the anthropology research exploring gift-giving and marketing research exploring celebrity endorsements by applying the tenants of the Elaboration Likelihood Model (ELM). This study uses an experimental procedure in order to determine the effect of using celebrity endorsements on buyers' attitudes and purchase intentions for gift-giving purchases in low and high involving categories. Results indicate that celebrity endorsements have no influence on attitudes and purchase intention in different product involvement and gift giver-receiver conditions.
Thesis:
Thesis (M.A.)--University of South Florida, 2009.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Christine Anghel.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 92 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002045516
oclc - 495351595
usfldc doi - E14-SFE0002979
usfldc handle - e14.2979
System ID:
SFS0027296:00001


This item is only available as the following downloads:


Full Text

PAGE 1

The Effect of Celebrity Endorsements on Gift -Giving Purchases: An Application of the Elaboration Likelihood Model by Christine Anghel A thesis submitted in partial fulfillment of the requirement for the degree of Master’s of Arts College of Mass Communications University of South Florida Major Professor: Scott Liu, Ph.D. Kelli Burns, Ph.D. Kenneth Killebrew, Ph.D. Date of Approval: July 7, 2009 Keywords: product involvement, Oprah Winfre y, advertising, marketing, attitudes, purchase intent, behavioral intent Copyright 2009, Christine Anghel

PAGE 2

DEDICATION This thesis is dedicated to anyone who has ever thought that anything is impossible to achieve.

PAGE 3

ACKNOWLEDGEMENTS I am always reminding myself to be th ankful to the many people who love and support me in everything that I do. For this, I owe my thanks. To my thesis committee members, Dr. Liu, Dr. Burns and Dr. Killebrew, thank you for your assistance and support th at aided in making this thesis the best that it can be. To my boyfriend, Jordan, for helping me understand the statistical information that made me feel overwhelmed at times; I thank you for your patience. Most importantly, thank you for your encouragemen t as I traveled along in the program. Finally, to my parents, Dan and Lou Ann, who have always been shoulders to lean on. Thank you for constantly giving me the positive advice that I needed, especially when I became discouraged in my final m onths of graduate school. Without you, I would not have had the opportunities to achieve my dreams thus far. From deep down, thank you.

PAGE 4

TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES v ABSTRACT vi CHAPTER 1: INTRODUCTION 1 CHAPTER 2: LITERATURE REVIEW 3 Endorsements Defined 3 Celebrity Defined 4 Source Characteristics 6 Source Credibility 6 Source Attractiveness 7 Matching Products with Endorsers 8 The Impact of Using Celebrity versus A Non-Celebrity 8 Elaboration Likelihood Model and Celebrity Endorsements 10 Gift-Giving 15 Elaboration Likelihood Model and Gift-Giving Involvement 16 Determinants of Processing Stra tegy: Motivation and Ability 16 Gift-Giving and Involvement 17 CHAPTER 3: RESEARCH HYPOTHESES 23 i

PAGE 5

CHAPTER 4: METHODOLOGY 26 Experimental Design 26 Stimulus Material 29 Pretest 29 Spokesperson 30 Product 30 Product Recipient 30 Procedure 31 Manipulation Checks 31 Reliability 34 CHAPTER 5: RESULTS 36 Descriptive Statistics 36 Sample Profile 40 Hypothesis 1 42 Hypothesis 2 48 Hypothesis 3 54 CHAPTER 6: DISCUSSION 62 Other Findings 62 Study Limitations 63 The Use of Oprah Winfrey 65 CHAPTER 7: CONCLUSION 67 REFERENCES 71 ii

PAGE 6

APPENDICES 76 Appendix A. 1. Participant Background 77 Appendix B. 1. Participant Instructions 78 Appendix C. 1. Non-Celebrity, MP3 Treatment 79 Appendix C. 2. Non-Celebrity, Photo Album Treatment 80 Appendix C. 3. Celebrity, MP3 Treatment 81 Appendix C. 4. Celebrity, Photo Album Treatment 82 Appendix D. 1. Advertisement Questions 83 Appendix E. 1. Celebrity Manipulation Check 87 Appendix E. 2. Non-Celebrity Manipulation Check 88 Appendix E. 3. High Involvement Product Manipulation Check 89 Appendix E. 4. Low Involvement Product Manipulation Check 90 Appendix E. 5. High Involvement Friend Manipulation Check 91 Appendix E. 6. Low Involvement Friend Manipulation Check 92 iii

PAGE 7

LIST OF TABLES Table 1. Distribution of Partic ipants to Treatment Groups 26 Table 2. 2X2X2 Factorial Design 29 Table 3. Spokesperson Manipulation Ch eck (2X2 Chi-Square Test) 32 Table 4. Manipulation Check for Pr oduct Involvement (t-test) 33 Table 5. Manipulation Check for Friend Importance (t-test) 34 Table 6. Reliability Statistics for ATTA 34 Table 7. ATTA Item Statistics 35 Table 8. ATTB Reliability Statistics 35 Table 9. ATTB Item Statistics 35 Table 10. Dependent Variable Desc riptive Statistics 37 Table 11. Sample Profile 41 Table 12. ATTA Means for Spokesperson+Product Category 42 Table 13. ATTB Means for Spokesperson+Product Category 43 Table14. Purchase Probabil ity Means for Spokespers on+Product Category 46 Table 15. ATTA Means for Spokesperson+Friend Relationship 48 Table 16. ATTB Means for Spokesperson+Friend Relationship 49 Table17. Purchase Probab ility Means for Spokesper son+Friend Relationship 52 Table 18. ATTA Means for Spokesperson+Product Involvement+Friend 54 Relationship Table 19. ATTB Means for Spokesperson+Product Involvement+Friend 55 Relationship Table 20. Purchase Probabil ity Means for Spokesperson+ Product Involvement+ 58 Friend Relationship Table 21. ANOVA Results 60 iv

PAGE 8

LIST OF FIGURES Figure 1.0. The Elaboration Likelihood Model 12 Figure 2.1. ATTA Means 44 Figure 3.1. ATTB Means 45 Figure 4.1. Purchase Probability Means 47 Figure 5.2. ATTA Means 50 Figure 6.2. ATTB Means 51 Figure 7.2. Purchase Probability Means 53 Figure 8.3. ATTA Means 56 Figure 9.3. ATTB Means 57 Figure 10.3. Purchase Probability Means 59 v

PAGE 9

The Effect of Celebrity Endorsements on Gift-Giving Purchases: An Application of the Elaboration Likelihood Model Christine Anghel ABSTRACT The purpose of this study is to determ ine how effective celebrity endorsements are in regards to the type of gift purchase one decides to make (i.e., buying for someone who has a high significant meaning to the buye r, such as a best friend, versus buying for someone who has a low significant meaning to the buyer, such as a casual friend). The study seeks to extend upon the anthropology research exploring gift-giving and marketing research exploring celebrity e ndorsements by applying the tenants of the Elaboration Likelihood Model (ELM). This st udy uses an experime ntal procedure in order to determine the effect of using cel ebrity endorsements on buyers’ attitudes and purchase intentions for gift-giving purcha ses in low and high involving categories. Results indicate that celebrity endorsement s have no influence on attitudes and purchase intention in different product involv ement and gift giver-receiver conditions. vi

PAGE 10

CHAPTER 1: INTRODUCTION With Hollywood being a reflection of Ameri can culture, it is not a surprise that approximately 25 percent of American comme rcials use celebrity endorsers (Shimp, 2000). Celebrity icons are found in roughly 20 pe rcent of all televi sion advertisements (Boyd & Shank, 2004). But why do advertisers spend so much and have such confidence in celebrities? According to Till (1998) when celebrity endorsements are used appropriately, “[they] can serve a valuable ro le in developing brand equity and enhancing a brand’s competitive position” (p. 401). Understanding the effectiveness of e ndorsers is a central issue for both practitioners and academics (Till & Busler, 1998). With American culture’s obsession with celebrities, it is import ant to understand the use of cel ebrities in advertising. More importantly, understanding the effect that ce lebrity endorsements have on how consumers spend their money will better he lp advertisers market their products when using celebrity endorsers. Since approximately 95 percent of the gifts given in th e United States are purchased products rather than services or products (Belk 1982), it can be assumed that gift-giving is an important issue to marketer s. For instance, Christmas holiday sales can make up 30-50 percent of a reta iler’s total yearly sales and earnings alone (Smith & Beatty, 1985). 1

PAGE 11

A great deal research has been done on th e effect of celebrity endorsements based on the following characteristics: Source credib ility (e.g. trustworthiness and expertise), source attractiveness (e.g. lik eability) and celebrity/product match. Past research has consistently shown that source credibility, especially expertise, is the most significant factor in determining the effectiveness of celebrity endorsements, which influences the consumers’ intention to purchase (Silvera & Austad, 2004). On the other hand, little to no research has been done in order to determ ine the influence of these advertisements on attitudes and intention to purchase for some one other than the purchaser. For instance: Does the use of a celebrity aid in the consum er’s decision to purchase a product as a gift? The purpose of this study is to determ ine how effective celebrity endorsements are in regards to the type of gift purchase one decides to make (i.e., high involving product versus low involving product) and the type of gift giver-receiv er relationship that exists (i.e., buying for someone who holds a special meaning to the buyer versus buying for someone who does not hold a special mean ing to the buyer). Th is study seeks to extend upon the anthropology research explor ing gift-giving and marketing research exploring celebrity endorsements by applying the tenants of the Elaboration Likelihood Model (ELM). This effort will be able to further discuss how i nvolvement levels for celebrity endorsements and gift purchasing are used in changing attitudes and purchase intention. Since research shows that celebrity endorsements are often successfully used in advertisements and gift purchases make up a la rge percentage of sales, this subject should be of interest to marketers. 2

PAGE 12

CHAPTER 2: LITERATURE REVIEW Endorsement Defined According to the Federal Trade Commission (1980), an endorsement is defined as: Any advertising message (including ve rbal statements, demonstrations, or depictions of the name, signature, li keness or other identifying personal characteristics of an individual or th e name or seal of an organization) which message consumers are likely to believe reflects the opinions, beliefs, findings, or expertise of a party other than the sponsoring advertiser. The party whose opinions, beliefs, findings, or expertise the message appears to reflect will be called the endorser and may be an individual, group or institution. Endorsements have shown to be successf ul in advertisements. For instance, a study by Hastak & Mazis (2003) factoring te stimonials and disclosures in dietary supplement booklets, found that numerous testimonials about a product positively and effectively communicates that the product is successful in the us es described in the testimonials and that the product will work for at least half of pe ople who use it (Hastak & Mazis, 2003). Although this may be true, endorsers us ed in advertisements have certain restrictions and guidelines that must be considered, according to the FTC (1980). “Endorsements must always reflect the honest opinions, findings, beliefs, or experience of the endorser” (FTC, 1980). 3

PAGE 13

Endorsements should not be presented that is in any way out of context or reworded in any way that would alter the e ndorser’s view or occurr ence with the product (FTC, 1980). Additionally, a celebrity can be used as an endorser only when the advertiser has good reason to believe that the endorser continues to promise to the opinions presented. The endorser must have been a true user of the product at the time the endorsement was given and the advertisem ent can only be run for as long as the advertiser believes that the endorse r still remains a user (FTC, 1980). According to Friedman, Termini, and Wa shington (1976), there are four major different endorsers: The typical consumer, professional expert, company president and celebrity. The typical consumer is a real pe rson, not an actor, and a true user of the product. In fact, the only knowledge of the produc t is the result of the typical consumer’s use of the product. The company president is le ader of the company’s product in which is being promoted and the profe ssional expert is recognized ba sed on their expertise within the product class that is being endorsed. This person’s special understanding or training of the product is more advanced than that gained by average people. The celebrity is a recognized individual who is known for their accomplishments in areas that are not associated to the product class that is be ing endorsed (Fredman, Termini & Washington, 1976). Celebrity Defined According to Marshall (1997), celebrities are celebrated i ndividuals that are unique, in some way, from the average citize n. Boorstin (1961) defines celebrity as someone “who is well-known for their well-known-ness” (p. 58). 4

PAGE 14

According to Fredman, Termini and Washington, a celebrity is sometimes a sports figure, actor, comedian or other type of entertainer (1976). Howeve r, unlike heroic figures, the modern celebrity may not have achieved anything exceptional except, merely, public attention and a product of me dia representation (Tuner, 2004). This is evident in contestants from reality shows, such as Big Brother and Survivor According to Tuner (2004) fame is developed, not by the achievement of great things, but, by differentiating one’s persona lity from those of their competitors. Consequently entertainers lead the ranks of celebrity “because they are skilled in the marginal differentiation of their personalities” (Boorstin, 1961, p. 65). One of the “know all” resources that keep our society up-to-date about the latest celebrity news is the tabloid press. As not ed by Marshall (1997), th e tabloid press gives us an outrageous twist on the connotation of th e celebrity because it presents the general public a possibility that these “unique talents, ” that makes one a celebrity, are vulnerable. As a result, these public individuals are subj ect to ups and downs in their career, and ultimately their life. As a result, these ups and downs can influence the brand or product that has been endorsed by a particular celebrity. According to McCracken (1989), the success of a celebrity endorsement has to do with the cultural meaning of the celebrity endorser. These meanings vary across status, class, gender and age. In addition, unique celebrity pers onalities and lifestyles can influence the success of an endorser depending on cultura l norms. (McCracken, 1989). These and other source characteristics are si gnificant to advertising research and, more specifically, research that focuses on th e effect of celebrity endorsements. 5

PAGE 15

Source Characteristics The source characteristics of a spokesp erson, more specifi cally a celebrity spokesperson, have been studied by research ers for years since each characteristic influences audiences in their own unique wa ys (Marshall, 1997). “Who is shown in an advertisement can say much to the consumer about the intended users of a product and about the benefits resulting from pr oduct use” (Lynch & Schuler, 1994, p. 418). Therefore, these characteristic s are important to briefly to uch upon in order to entirely comprehend the impact of a celebrity endorser. Source Credibility Silvera & Austad (2004) note that sour ce credibility is the primary factor determining how influential the celebrity endor ser is perceived. According to Goldsmith, Lafferty and Newell (2000), the credibility of the endorser, celebrity or not, is influential, especially on the attitudes towards the adve rtisement. Source credibility is classically seen as a function of trustworthines s and expertise (Silvera & Austad, 2004). “Trustworthiness refers to the general believab ility of the endorser, and is thus broader but conceptually similar to correspondent inferences about the endorser” (p. 1511). For example, this was apparent in regards to George Foreman and the George Foreman Grill. “The key to the success of the grill was marrying a grea t product with a credible personality to endorse it” (H oyer & MacInnis, 2007, p. 124). The other function of source credibility is expertise. This refers to the product knowledge that the endorser show s, which ultimately leads to the validity of his or her arguments in regards to the pr oduct (Silvera & Austad, 2004). 6

PAGE 16

In addition, “[expertise] is believed to be a factor that increases persuasiveness above and beyond the effects of trustwor thiness” (Silvera & Austa d, 2004 p. 1512). For instance, a study conducted by Ohanian (1991), examining the impact of celeb rity spokespersons’ perceived image on consumers’ intention to purchase, found that expertise was the only significant characteristic that impacted purchase intention. Source Attractiveness Another source characteristic that has been of interest, in regards to celebrity endorsement research is source attractivene ss. Source attractiveness has been argued to increase the likeability of the source and the a dvertisement. Most studies have shown that a physically attractive source as sists in changing consumers’ attitudes (Baker & Gilbert, 1997; Caballero & Prince, 1984; Chaiken, 1979; Horai et al., 1974; Joseph, 1982; Kulka & Kessler, 1978; Mills & Ar onson, 1965; Mills & Harvey, 1972; Petty & Cacioppo, 1980 as cited in Kahle & Homer, 1985; Silvera & Austad, 2004). However, researchers such as Benoy (1982) found source expertise to be more important than source attractiveness. When the source was expert, [their] physical attractiveness made little difference in terms of subjects’ preferences; however, when she was inexpert, subjects agreed more with the high attractive source than with the medium or low attractive source. Apparently, when objective or taskrelated source characteristics (e.g., expert ise) are weak, subjects resort to “irrelevant” cues (such as physical at tractiveness) to form opinions (p. 19). 7

PAGE 17

Matching Products with Endorsers Other research has focused on the “fit” between the celebrity and the endorsed product (Till & Busler, 1998; Till & Shimp, 1998). Till & Busler (1998) note that certain celebrity/product matches work better than others due to a match-up between the celebrity and product. For example, profe ssional sports players match up better with sporting good products rather than cosmetic s. In general, understanding the source characteristics that have been researched ar e key in order to fully grasp an understanding of how celebrity endorsements are used to pers uade consumers. However, this particular research is concerned with if and when a spokesperson (famous or not famous) is influential. The Impact of Using a Celebrit y versus Using a Non-Celebrity Overall, research has shown that c onsumers’ views and opinions about an advertisement that involve celeb rities are positive. Research has shown that a significant amount of money is invested by corporations in order to align itself and its products with celebrities (O’Mahony & Meenaghan, 1997/1998). This is done in an effort to “draw attention to endorsed products/services and transfer image values to these products/services by virtue of their celebrity profile and engaging attributes” (O’Mahony & Meenaghan, 1997/1998, p. 15). According to Sherman (1985), approximately 20 percent of all television a dvertisements include famous people and approximately 10 percent of the money spent on television advertisements are used on celebrity endorsements. 8

PAGE 18

In addition, high profile bra nds side themselves with high named celebrities. For example, Coca Cola Company spent $25 million dollars for an advertising campaign in order to presen t a certain celebrity as an endorser for Coke (Advertising Age, 1986 as cited in Jagdish, Kamakura, & Wagner, 1995). However, according to Fredman, Termin i and Washington (1976), endorsements, no matter who the endorser is, have been f ound to be worthwhile (Fredman, Termini & Washington, 1976). Endorsers effect consumer s’ expectations, inte nt-to-purchase and believability than advertisements show n with no endorser (Fredman, Termini & Washington, 1976). Although Till (1998) suggests th at there are certain risks involved in using a celebrity endorser, cele brity ads can serve as effectiv e ways to market a brand. Ultimately celebrity endorsements are assu med to produce a greater probability of customers’ choosing the endorsed bra nd (Heath, McCarthy, & Mothersbaugh, 1994; Kahle & Homer, 1985; Kamins et al., 1989; Ohanian, 1991 as cited in Agrawal, Jagdish, Kamakura & Wagner, 1995). A study was conducted in an effort to better understand if celebrity endorsers we re economically worthwhile. By vi ewing the impact of celebrity contract announcements on stock returns, resu lts indicate a positive correlation (Agrawal, Jagdish, Kamakura & Wagner, 1995). Therefore, celebrity endorseme nt contracts are a worthwhile investment in advertising as opposed to using non-celebrities. (Agrawal, Jagdish, Kamakura & Wagner, 1995). However, do consumers use a celebrity as a cue when purchasing a gift for someone else? This question will be discussed later. But first, it is important to grasp an understanding of the importance of the Elaboration Likelihood Model and how it interacts w ith celebrity endorsements. 9

PAGE 19

Elaboration Likelihood Model and Celebrity Endorsements According to Petty, Cacioppo and Schum ann (1983), understanding attitudes has become of key interest within consumer beha vior research. Resear chers and advertisers have devoted a great deal of time and effort in order to determine how to change the buyer’s attitude to sell their brand. As a result, Petty and Cacioppo’s Elaboration Likelihood Model (ELM) is an influential to ol in research studying the attitudes and persuasion in consumer behavior research. In order to understand the ELM, it is first important to define the terms attitude and influence and persuasion in regards to attitude, since these words are imperative in understanding the model. Attitudes are define d as “general evaluations people hold in regards to themselves, other people, object s, and issues” (Petty & Cacioppo, 1986, p. 4). Influence is defined as any change in thes e general attitudinal evaluations. Finally, persuasion refers to “any change in attitudes that results from exposure to a communication” (Petty & Cacioppo, 1986, p. 5). The ELM is a two-process model of re sponse to advertisi ng stimuli (Petty, Cacioppo, & Schumann, 1983). In regards to th e “elaboration” portion of the ELM, “on one end…there is virtually no thinking a bout the issues discussed in the message, whereas at the other end there is an enormous amount of mental activity, as the individual mulls over and cognitively elaborates on message arguments” (Perloff, 2003, p. 118119). Under conditions of high involvement the attitude change is processed through the central route 10

PAGE 20

This is when “consumers are more likely to devote a lot of effort toward and invest considerable personal involvement in forming or changing attitudes and making decisions” (Hoyer & MacInnis, 2007, p. 127). Th e central route is used to process information that the viewer finds to be vital to the true qualities of a particular attitudinal position (Petty, Cacioppo, & Schumann, 1983). Under conditions of low involvement attitude change is processed through the peripheral route (Petty, Cacioppo, & Schumann, 1983). This is when “consumers’ attitudes are based on a more ta ngential or superficial analysis of the message, not on an effortful analysis of its true merits” (Hoyer & MacInnis, 2007, p. 127). For example, rather than considering the pros and cons of an argument, an individual may decide to accept an argument simply because a celebri ty was used in the endorsement. Figure 1.0. offers a diagram of the model. 11

PAGE 21

Figure 1.0. The Elaboration Likelihood Model of Persuasion (Petty & Wegener, 1999). 12

PAGE 22

As noted by Clark and Horstmann (2005) products that are often endorsed by celebrities are running shoes, beauty products, soft drinks and other beverages, and the like. These types of products are often in the low involving product categories, as noted by the ELM. This is because these types of products are well estab lished, have little apparent quality variation, and on which th eir manufacturers constantly spend large amounts for advertising. In addition, the author s note that an item in the low involvement product category risks being forgotten or being passed up for a similar product. Essentially, the explanation is that producers of these types of products advertise because of the negative inferences that consumers not seeing an ad for a particular product draw ab out the value of that product. A consumer who does not see an ad fo r firm 2’s product, say, but does for firm 1’s, attaches greater probability to firm 1’s product, the advertised product, having large sales, and so be ing of higher value, than firm 2’s product, the non-advertised product. As a result, this consumer is less likely to purchase firm 2’s product. Had this consumer failed to observe an ad for either product, he would value firm 1’s product less than in the previous situation and so would be less likely to purchase 1’s product than previously. These negative inferences, and consequent purchase decisions, provide the profit incentive for a firm to advertise (Clark & Horstmann, 2005, p. 380). Therefore, effective advertising, such as using celebrity endorsements, is imperative to the lower involving product ca tegories in order to make a product memorable. Other researchers have concluded that celebrities pers uaded consumers when products were less costly, low involving a nd few differences apparent among existing brands (Callcoat & Phillips, 1996). 13

PAGE 23

Also consistent with the ELM, Petty Cacioppo, & Schumann (1983) conducted a study, specific to celebrity endorsements, which found that under conditions of high involvement (central route), ar guments but not celebrities in fluenced attitudes. On the other hand, under conditions of low involvement (peripheral route), celebrities but not arguments influenced attitudes. This experi ment showed that there are two relatively distinct routes to persuasion (Petty, Cacioppo, & Schumann, 1983). According to Byrne et al. (2003), incor porating the use of a celebrity to a product’s image, such as a grocery store item can positively affect the overall image of a corporation. For instance, a l eading European grocery store, J. Sainsbury, incorporated the use of a famous chef from th e popular Food Network Channel’s show The Naked Chef. The store had the famous chef endorse cert ain grocery store items, such as low fat food products, in an attempt to rebuild their br and. As a result, J. Sainsbury succeeded in their brand renovation because that they used the well-known chef for their low-involving grocery items (Byrne et al., 2003). Thus, per the ELM, celebrity endorsements have shown to be successful when the product is low-involving by activ ating the consumer’s peripheral cue. However, it is important to note that the eff ectiveness of the spokesperson, such as a celebrity, depends on the receiver. “Much depends on the biases and beliefs that audience members bring to the persuasion situation and the extent to which receivers are mo tivated and able to process the message” (Perloff, 2003, p. 153). 14

PAGE 24

So, is the receiver motivated because the me ssage is personally relevant to them? Personal relevance is a significant component of the ELM that is especially relevant to the gift-giving portion of this st udy and discussed in more detail at a later time. But first, let’s explain the gift and the fact ors that influence gift-giving. Gift-Giving Gift-giving is a subject chiefly studie d by anthropologists for the reason that different cultures exchange gifts in various ways. Anthropologists such as Malinowski (1922), Levi-Strauss (1969), & Sahlins (1972) have contributed to gift-giving research (as cited in Komter & Vollebergh, 1997). These researchers “emphasized that gift exchange fulfills important functions in the development and continuity of society and culture” (Komter & Vollebergh, 1997, p. 747). Certain gift-giving research incorporates the emotions that go along with the process. “We all understand that we are expect ed to give gifts to certain people on certain occasions, and that the value of the gift de pends on the occasion and our relationship with the recipient” (Laroche et al., 2000). Anth ropologists who have re searched the act of gift-giving have pointed to feelings of oblig ation and patterns of reciprocity that are involved in the gift exchange process (K omter & Vollebergh, 1997). However, “pure” gift-giving is described by th e lack of these feelings. The reciprocity within the pure gift-giving act may not be in the form of tangible items. Instead, reciprocal gift exchan ge creates moral ties between people…” (Malinowski, 1923 as cited in Komter & Vollebergh, 1997, p. 748). In other words, giftgiving is a way of maintain ing social relationships. 15

PAGE 25

Another type of obligation, defined by Goodwin et al. (1990), is ritual. Certain occasions involve certain traditions, whic h are dictated by the culture of one’s society. For instance, one of the most popular occasions is Christma s where the gift exchange is a key ritual. Elaboration Likelihood Model and Gift-Giving Involvement Determinants of processing stra tegy: Motivation and ability According to the ELM, people must be motivated, influenced by the level of involvement and the need for cognition or t hought, to elaborate on a message and must have the ability, influenced by knowledge or situ ational factors, to do so in order to bring out central processing (Petty & Cacioppo, 1981). A motivational factor that can affect a person’s ability to analyze “issue-relevant arguments in a relatively objective manner” (Petty & Cacioppo, 1981, p. 81). There are three va riables that have an important effect on the motivation to process: personal relevance of an issu e, personal responsibility for message evaluation, and the number of message sources. However, pers onal relevance is the variable that is most applicable to this study. Personal relevance, defined by Petty & Cacioppo, 1981, have also be labeled “ego-involvement,” “issue involvement,” “perso nal involvement,” “vested interest,” etc. Personal relevance occurs when people expect the issue to have c onsequences that can affect their own lives. These consequences can exist over long periods of time (e.g. changing certain laws), exist for a more conf ined period and/or audi ence (raising college tuition), or exist only under temporary condi tions (e.g. computer advertisements have a higher relevance when a person is search ing for a new computer) (Petty & Cacioppo, 1981). 16

PAGE 26

Petty and Cacioppo (1986) proposed that “as personal relevance increases, people become more motivated to process the issue -relevant arguments presented” (p. 82). In other words, as the personal consequences increase, it is more important for people to evaluate the true merits of the proposal by processing info rmation centrally in order to form an opinion. Gift-giving and involvement Although purchasing something as a gift fo r someone else is not initially thought of as personally relevant and involving because it is not being purchased for one’s self, the opposite is true. According to Belk (1982) the gift-giving pro cess is more involving activity than self purch asing activity. There ar e at least two types of involvement with which are concerned with gift-giving. The first is item-specific and the second is purchase situation-specific The item-specific form of involve ment has been called "issue involvement" (Lastovika, 1976 as cited in Belk, 1982), "importance of purchase" (Howard & Sheth, 1969), "enduring involve ment" (Rothchild, 1977) and "product involvement" (Clarke & Belk, 1979). Essentially the consumer who is high in purchase item-specific involvement cares more about th at item and is more interested in the purchase outcome. The situation-specific invol vement of the purchase centers on the consumer’s particular objective or task to be complete d in the shopping situati on. For instance, "the task may be highly involving either because it entails important immediate goals (e.g. find a coat which is the least expensive wool co at in town), or because the intended usage situation involves important goals (e.g. find a dress to wear to th e prom)" (Belk, 1979). 17

PAGE 27

Laroche et al. (2000) additionally found that involvement level varies depending on the item or budget the buyer has predeter mined. Consistent w ith the ELM, the study by Laroche et al. found that in-store informa tion sources (information that buyers used to go out of their way and seek), such as store di splays and sales clerks, were generally used for items that were costly (e.g. high invol ving). Additionally, an individual with a predetermined gift selection was more likely to specifically search for information about that predetermined gift instead of a genera l information search (Laroche et al, 2000). According to Gronhaug (1972), consumer s who bought tableware as a gift reported spending a significant amount of time seeking out information about the product (e.g. considering more alternativ e choices, shopping at more d ealers, seeking more advice from others, and reading dealers’ brochures more thoroughly). Additionally, previous research has also found higher levels of percei ved risk (Hart, 1974 as cited in Belk, 1982) and prices less of an issue (Shapiro, 1975) in gift buying ra ther than self purchase. However, this paper focuses on the different gift giver-recipient re lationships that can ultimately impact how the buyer chooses a gift. Belk’s (1982) study was composed of less involving gift items that were characterized by low cost, ease of purchase and low quality. The low involvement gift situations were the following: a thank-you gift to repay for a favor and a birthday gift for a casual friend. High involvement gift situations were as follows: a birthday gift for a close friend and a wedding gift for a close relative. As stated earlier, gift-giving involvement levels are accompanied by other f actors in regards to pur chasing strategies. 18

PAGE 28

In this study, Belk (1982) found th at the expectations that are attached to the specific giftgiving occasion and the recipient relationshi p influenced the results. For instance, wedding gifts were judged as having more e xpensive selections, and therefore higher expectations, than birthday gifts. Therefore, the buyer was highly involved for this giftgiving occasion. In addition, these high involving occasions were both to be given as a gift to someone with greater salient meaning (e.g. close friend or relative versus casual friend). All in all, the gift giver-recipien t relationship showed to influence purchase strategy involvement. According to Laroche, et. al. (2000), “In terms of search effort, more demanding selection strategies are likely to be used when the relationship is more salient to the giver” (p. 4). Gift recipien ts are often described as “difficult” or “easy,” which is influenced by the giver-receiver relationship or commonality that they both share, noted by Otnes et al. (1993). For instance, difficult re cipients in Otnes’ (1993) study were older or more distant relatives, wh ile easy recipients tended to be young (children) and samegendered friends. Givers also noted that they perceive difficult recipients as misinterpreting gifts that are intended to express a specific social role. The following is a list of top (six or more; the rest was four or less) difficult re cipients as noted by respondents: in-laws, fathers, grandparents and elderly relatives, a nd step relatives. The list of top (11 or more; the rest was four or less) easy recipients we re children and samegendered friends. Mothers, sisters, husbands or boyfriends, and opposite-gendered friends were spread among the two lists of recipients (Otnes et al, 1993). 19

PAGE 29

Komter and Vollebergh (1997) studied di fferent Western soci al relationships between the giver and recipien t by analyzing the feelings th at go along with gift-giving. The study contrasted between feelings of affection, as indicated by love and companionship, and obligation, as indicated by indebtedness, personal obligation, and customary obligation. Results point out that extended kin and friends each are given more than a quarter of all gifts. However, showing to have lesser salient meaning to the giver were acquaintances and neighbors who receive the least amount of gifts. Results also found that friends and family are as emotionally as close. Overall, family and friends are the most salient individuals in a person’s social network. Different from what the older anth ropological sociological theories suggest in this respect, gift-giving to these family members is not only based on feelings of affection in our own society, but it also springs from feelings of moral obligati on. Gift-giving to friends; however, is more often accompanied by feelings of affection. ...The explanation for this may be that family ties are given, and most people traditionally feel a certain moral obligation to sustain these ties. Ties to friends are chose, not born out of obligation or tradition, but out of mutual affection (Komter & Vollebergh, 1997, p. 756). As shown in the research previously stat ed, different gifts ar e given to different people. For instance, one may gi ve a gift to someone that they consider particularly special in their life, such as a family memb er. However, the consumer may feel obligated to give a gift to someone even if they do not consider this recipient to be as special. Therefore, personal relevance and the level of involvement in the gift choosing process differ. 20

PAGE 30

As stated earlier, Petty and Cacioppo (1976) found that as the personal consequences increase, it is more important for people to evaluate the true merits of the proposal by processing information centrally in order to form an opinion. Therefore, as shown in the research, gift-giving is a personally relevant and involving process (especially if it is high ly priced), which makes it important for buyers to centrally process information about the item. More specifical ly, the more significant one feels that the recipient is to the buyer, the more involved th e buyer is in the gift-giving process. But how does the use of a spokesperson, such as a celebrity, influence gift purchasing? Celsi and Olson (1988) note that personal relevance is a subjective feeling, which they label “felt involvement.” Even objects or events, such as gift-giving, that are extremely important to an individual are no t felt as personally relevant at all times. Therefore, just because gift-giving has been noted as being high involving does not mean that it is high involving at all times. Instead, outside factors, such as the giver-receiver relationship, influences how involving the gi ft-giving process may be. This entails that the situational context, such as the use of a celebrity endorsement or giver-receiver relationship, is vital in determining the degr ee and type of personal relevance experienced by a consumer (Celsi & Olsen, 1988). 21

PAGE 31

As argued earlier, per the ELM, there ha s been a significant amount of research that has shown that the use of a celebrity act ivates the consumer’s peripheral route of persuasion. Celebrity endorsements are used for low involving situ ations while giftgiving is a high involving, cogni tive-based process, especially when purchasing for a recipient that has salient meaning to the buye r. Therefore, even though gift-giving is a high involving process, an out side factor, such as the use of a celebrity in an advertisement, may influence the consumer. However, one specific study, conducted by Ohanian (1991), which focused on the impact of celebrity spokespersons on consumers’ intention to purchase, only briefly mentioned anything about gift purchasi ng. Again, compared to other source characteristics such as physical attractiven ess and trustworthiness, this study found that expertise was the only characteri stic that significantly relate d to respondents’ intention to purchase for themselves and as a gift not for themselves (Ohanian, 1991). The research done for this study seeks a more specific and better understanding of the impact, if any, of celebrity endorsements on a consumer’s atti tudes and intention to purchase an item as a gift. The information gathered from this study will be able to aid marketers’ understanding of what in fluences consumers. 22

PAGE 32

CHAPTER 3: RESEARCH HYPOTHESES The main purpose of this study is to explore the influe nce that celebrity endorsements have on the gift-giving pro cess and attitudes toward the brand and advertisement they endorse. The experiment s eeks to find if there is any effect that the independent variables— product involvement (high versus low), spokesperson (celebrity versus non-celebrity), and gift recipient re lationship (weak or lo w significant meaning versus strong or high meaning), have on th e dependent variable — attitudes (attitude toward the ad and attitude toward the brand) and buying intention. The Elaboration Likelihood Model states that a message or advertisement can change attitudes or create resi stance to change either centra lly or peripherally (Petty & Cacioppo, 1986). A message can influence at titudes “…by getting the person to do a great deal of thinking about the message, or by inducing the i ndividual to focus on simple, but compelling, cues that are periphe ral to the message content” (Perloff, 2003, p. 119). Research has consiste ntly found through empirical analysis that more thought and processing of message arguments were f ound in individuals in high involvement situations. However, the peripheral cues, such as the use of a cel ebrity, were found in individuals under low involvement situations Under high involvement, argument quality was important. Based on the tenets of th e ELM, the following hypotheses are suggested: 23

PAGE 33

H1a: Celebrity endorsements will be more effective in generating favorable attitudes than non-celebrity endorsem ents only in low involvement product situations, while celebrity endorsements and non-celebrity endorsements will have no effect in high involvement product situations in regards to the buyer’s attitudes. H1b: Celebrity endorsements will be more effective in increasing buying intention than non-celebrity endorsem ents only in low involvement product situations, while celebrity endorsements and non-celebrity endorsements will have no effect in high involvement product situations in regards to the buyer’s buying intentions. Furthermore, research has found that the giver-receiver relationship impacts the level of involvement consumer’s use when choosing a gift for someone. “In terms of search effort, more demanding selection stra tegies are likely to be used when the relationship is more salient to the giver” (L aroche, et al., 2000, p. 4). In addition to the celebrity and involvement re lationship stated earlier, the following hypotheses were constructed: H2a: Celebrity endorsements will be more effective in regards to the buyer’s attitudes than non-celebrity endorsements only when the gift recipient has low significant relationship with the buyer. 24

PAGE 34

H2b: Celebrity endorsements will be more effective in regards to the buyer’s buying intentions than noncelebrity endorsements only when the gift recipient has a low significant relationship with the buyer. H3a: The effect of the celebrity endorsem ent on the buyer’s attitudes will be the strongest in low involvement product s ituations where the recipient has low significant meaning to the buyer. H3b: The effect of the celebrity endo rsement on the buyer’s buying intentions will be the strongest in low involvement product situations where the recipient has low significant meaning to the buyer. 25

PAGE 35

CHAPTER 4: METHODOLOGY Experimental Design A 2X2X2 factorial design was chosen in order to determine if the independent variables influence the depende nt variables. Each particip ant was randomly assigned to one of eight treatment cells, varying the spokesperson, product involvement level, and recipient. The goal of this design is to de termine if any difference in attitudes and the likelihood of purchasing exists across (1) the affective relati onship between receiver and giver; (2) the level of involvement (high or low); (3) celebrity endorsements. Table 1 illustrates the distribution of partic ipants to six treatment groups. Table 1. Distribution of Participants to Treatment Groups Friend Product MP3 Photo Album Best Friend Spokesperson Oprah 31 26 Regular 27 25 Casual Friend Spokesperson Oprah 27 27 Regular 25 31 26

PAGE 36

In order to empirically assess the hypotheses mentioned previously, a convenience sample of 219 stude nts enrolled in summer cour ses at the University of South Florida were involved in the expe riment. Although a total of 226 experiment packets were collected, the 219 packets that were tabulated were those that found Oprah a celebrity. The measurement instrument was a 26-item experiment that measured consumers’ attitudes toward th e advertisement and brand, as well as purchase intention. These items serve as the primary dependent variables. The packets experimentally and successfully manipulated three conditi ons: The spokesperson endorsing a common consumer product, the product featured in an advertisement and the recipient/friend for whom the product is intended. These meas ures serve as the primary independent variables. Altogether, there were a to tal of four advertisements that were given to eight experimental groups. The first advertisemen t was of the celebrity endorsing a low involving product where groups one and two r eceived this advertisement. This low involving product was to be bought for either a best friend or a casual friend. The second advertisement was of a celebrity endorsi ng a high involving produc t where groups three and four received this advertisement. Th is high involving product was to be bought for either a best friend or a ca sual friend. The third advert isement was of a non-celebrity endorsing a low involving product where groups five and six rece ived this advertisement. This low involving product was to be purchased for either a best fr iend or a casual friend. 27

PAGE 37

The fourth advertisement was of a non-cel ebrity endorsing a high involving product where groups seven and eight received this advertisemen t. This high involving product was to be purchased for either a best friend or a casual friend. These groups mentioned above are more specifically note d, in regards to placement, as follows: The first group view ed a low involvement product advertisement where a celebrity endorses the product and the recipient has a less significant relationship with the buyer ( n =27). The second group viewed a low involvement product advertisement where a celebrity endorses the product and the recipient has a significant relationship with the buyer ( n =26). The third group viewed a low involvement product where a celebrity does not endorse the produc t and the recipient has a less significant relationship with the buyer ( n =31). The fourth group viewed a low involvement product where a celebrity does not endorse the pr oduct and the recipien t has a significant relationship with the buyer ( n =25). The fifth group viewed a high involvement product where a celebrity endorses the product and the recipient has a less significant relationship with the buyer ( n =27). The sixth group viewed a high involvement product where a celebrity endorses the product and the recipient has a si gnificant relati onship with the buyer ( n =31). The seventh group viewed a high involvement product where a celebrity does not endorse the product and the recipient has a less significant relationship with the buyer ( n =25). The eighth group viewed a high involvement product where a celebrity does not endorse the product and the recipien t has a significant re lationship with the buyer ( n =27). 28

PAGE 38

The dependent variables of advertising e ffectiveness were measured using Likert scales. Three general categorie s were examined: attitudes towa rd the ad, attitudes toward the brand and the likelihood of purchasing the product as a gift. The list of questions that tested these variables can be found in Appendix D. The experiment used a 2X2X2 factorial design to test the hypotheses. A more visual explanation of the factorial de sign hypotheses testing is shown in Table 2. Table 2. 2X2X2 Factorial Design High Involvement (MP3) Low Involvement (Photo Album) Significant Relationship (Best Friend) Insignificant Relationship (Casual Friend) Significant Relationship (Best Friend) Insignificant Relationship (Casual Friend) Celebrity (Oprah) Celeb-MP3Best Celeb-MP3Casual Celeb-PhotoBest Celeb-PhotoCasual Non-Celebrity Non-CelebMP3-Best Non-CelebMP3-Casual Non-CelebPhoto-Best Non-Celeb-PhotoCasual Stimulus Material Each treatment was given one of four adve rtisements, which are available to view in Appendix C. Each advertisement featured a female spokesperson, product, brand and written text. The brand, “Companion,” and wr itten text were consciously chosen to posses the same attributes in order to co mpliment both products. Every advertisement remained the same with the exception of the spokesperson and product. Pretest Before the experimental packets were designed, a pretest was conducted in order to identify the appropriate inde pendent variables. A list of 12 celebrities was presented in the pretest. 29

PAGE 39

Based on respondents’ opinions of the celeb rities in regards to familiarity, popularity, likeability, and celebrity, the s pokesperson was chosen. Additi onally, a list of different product categories were listed, while responde nts were to rank how important the decision of buying the product would be when it came to buying for themselves and for someone other than themselves. Lastly, a list of gift recipients were presented and ranked in terms of importance, significance and saliency. Based on the pretest results, the questions for the experiment were designed. Spokesperson Two spokespeople were featured in the a dvertisements that were included in the experiment: Oprah Winfrey and an ordinary spokesperson. Product Both low and high involvement products were featured in the experiment in order to stay true to the ELM. It was necessary that both products were consciously chosen to be purchased by a college-aged student and as a gift. Therefore, an MP3 player and photo album were the products chosen. Product Recipient The two recipients in the experiment were labeled “best friend” and “casual friend.” The recipients were noted as being separate from each other. A best friend was defined in the experiment as someone the res pondent speaks to or se es on a regular basis and who is very close with and trusts. A casual friend is someone the respondent may socialize with in a group or even one-on-one. 30

PAGE 40

Procedure The respondents were informed at the out set of the experiment that the questions were to be for a Master’s thesis. Participan ts were told to simply read the packets carefully and follow the instructions that were given on the packets. Participant instructions are available in Appendix B. After the verbal instructions were gi ven, the packets were passed out to the participants. Since the experiment took place in a classroom, the experiment was done in a group setting of approximately 20-30 st udents per classroom. On average, it took participants a total of 10-15 minutes to complete the experiment. After each person was done with the experiment, they handed the p acket in to the front of the room to be collected by the experimental conductor. Manipulation Checks Manipulation checks were performed to test the internal validity of the experimental conditions. In other words, respond ents were asked a seri es of questions at the end of the experiment designed to conf irm (1) that respondents who received the Oprah Winfrey advertisement believed she had traditional celebrity attributes compared to the non-celebrity advertisement; (2) that respondents who were in the MP3 player condition considered it a hi gh involvement product, as opposed to the photo album, which was designed to be a low involvement product; 31

PAGE 41

(3) that respondents whose purch ase intention was for a recipi ent defined as a best friend considered them to be a high involving reci pient, as opposed to a casual friend, which was designed to be a low involving recipient. For a complete list of the manipulation questions and answers, pleas e refer to Appendix E. Respondents were asked whether they believed that the spokesperson was a celebrity by simply checking “Yes” or “No.” Since this question contained only two response choices a 2X2 chi-square table wa s created. Table 3 reveals that of the 219 respondents, 111 of the respondents in the Opra h treatment recognized her as a celebrity and 108 of the respondents in the non-celebrity treatme nt did not recognize the spokesperson as a celebrity. The trends in th e data exhibit a perfect relationship. For each variable, the results confirm that the measures are valid as the observed frequencies are clearly non-random and in the intended direction. Table 3. Spokesperson Manipulatio n Check (2X2 Chi-Square Table) SP-Celeb/Non-Celeb Total No Yes Spokesperson Oprah 0 111 111 Regular 108 0 108 Total 108 111 219 32

PAGE 42

Participants were then asked a series of questions pertai ning to whether the product category they were assigned was hi gh or low involving. The questions asked whether they thought the product was somethi ng they care about when buying as a gift, whether it was something that was important to them in regards to making the right choice of the product, and whether it was some thing that was important to the them, in regards to the outcome of their choice. To assess the effec tiveness of the product involvement and perceptions of tw o different product categories, a t test was performed, t =7.122, p <.001. These results, shown in Table 4, confirm that the measures are significantly valid. Table 4. Manipulation Check fo r Product Involvement (t-test) t-test for Equality of Means t df Sig. Mean Difference Std. Error Difference Prod Involvement 7.122 217 .000 1.51991 .21341 The final manipulation check was designed to test whether the gift recipient was low or high involving; high bei ng a best friend and low being a casual friend. Participants were asked about the friend they were assigne d to in terms of importance, significance and saliency. To evaluate the effectiveness of the product involvement and perceptions of two different friend categories, a t test was performed, t =10.07, p <.001. As expected, the results shown in Table 5 confirm th at the manipulation was successful. 33

PAGE 43

Table 5. Manipulation Check for Friend Importance (t-test) t-test for Equality of Means t df Sig. Mean Difference Std. Error Difference Friend Importance 10.066 202 .000 .99115 .09847 Reliability A reliability check was performed for th e seven attitudinal measures to ensure they measure a single construct. The measur es for ATTA (a combined index for the all attitudinal questions about the ad) were found to have “accep table” reliability; Cronbach’s Alpha was estimated to be .729 (“What does Cronbach’s alpha mean?,” 2009). However, ATTB (a combined index fo r all attitudinal questi ons about the brand) had a Cronbach’s alpha was .676, wh ich is rather low. Reliabili ty statistics for ATTA and ATTB are illustrated in Tables 6 and 8. Tables 7 and 9 are pr esented as reference for each attitudinal measure. Table 6. Reliability Statistics for ATTA Cronbach's Alpha Cronbach's Alpha Based on Standardized Items N of Items .729 .744 4 34

PAGE 44

Table 7. ATTA Item Statistics Mean Std. Deviation N A dCon 2.62 .933 225 A dInfo 2.52 1.065 225 A dFee l 2.57 .837 225 A dLike 2.75 .781 225 Table 8. ATTB Reliability Statistics Cronbach's Alpha Cronbach's Alpha Based on Standardized Items N of Items .676 .680 3 Table 9. ATTB Item Statistics Mean Std. Deviation N BrandFav 2.73 .824 225 BrandFeel 3.07 .630 225 BrandLike 2.98 .664 225 35

PAGE 45

CHAPTER 5: RESULTS Descriptive Statistics Along with demographic information of par ticipants, this stu dy also sought to measure participants’ attitudes about the advertisement, specifically asking how convincing, informative, and likeable the ad wa s, as well as the f eelings the ad brought upon the respondent. Additionally, respondents were asked how favorable and likeable they found the brand, as well as the feelings the participant felt about the brand. The research instrument included items to meas ure purchase intent. Table 10 contains the descriptive statistics for the eight dependent variables used in this study. The results reveal that each variable is approximately normali in distribution and has sufficient variation for quantitative analysis. 36

PAGE 46

Table 10. Dependent Variable Descriptive Statistics Spokesp erson Product Friend Mean Std. Deviation N ATTA Oprah MP3 Best Friend 2.7083 .79080 30 Casual Friend 2.6759 .59167 27 Total 2.6930 .69760 57 Photo Album Best Friend 2.5577 .60954 26 Casual Friend 2.3889 .60579 27 Total 2.4717 .60776 53 Total Best Friend 2.6384 .71019 56 Casual Friend 2.5324 .61053 54 Total 2.5864 .66225 110 Regular MP3 Best Friend 2.8269 .65867 26 Casual Friend 2.7000 .60381 25 Total 2.7647 .62931 51 Photo Album Best Friend 2.4200 .67593 25 Casual Friend 2.6210 .76341 31 Total 2.5312 .72623 56 Total Best Friend 2.6275 .69169 51 Casual Friend 2.6563 .69177 56 Total 2.6425 .68861 107 Total MP3 Best Friend 2.7634 .72835 56 Casual Friend 2.6875 .59176 52 Total 2.7269 .66411 108 Photo Album Best Friend 2.4902 .64024 51 Casual Friend 2.5129 .69840 58 Total 2.5023 .66883 109 Total Best Friend 2.6332 .69814 107 Casual Friend 2.5955 .65313 110 Total 2.6141 .67440 217 37

PAGE 47

Table 10. Dependent Variable Des criptive Statistics (Continued) ATTB Oprah MP3 Best Friend 2.7222 .49583 30 Casual Friend 2.9630 .54954 27 Total 2.8363 .53126 57 Photo Album Best Friend 2.9231 .53589 26 Casual Friend 2.9506 .61813 27 Total 2.9371 .57385 53 Total Best Friend 2.8155 .51998 56 Casual Friend 2.9568 .57933 54 Total 2.8848 .55196 110 Regular MP3 Best Friend 3.0513 .58646 26 Casual Friend 2.9067 .48610 25 Total 2.9804 .53918 51 Photo Album Best Friend 2.8933 .63625 25 Casual Friend 3.0323 .52603 31 Total 2.9702 .57657 56 Total Best Friend 2.9739 .61044 51 Casual Friend 2.9762 .50795 56 Total 2.9751 .55646 107 Total MP3 Best Friend 2.8750 .55981 56 Casual Friend 2.9359 .51572 52 Total 2.9043 .53739 108 Photo Album Best Friend 2.9085 .58149 51 Casual Friend 2.9943 .56710 58 Total 2.9541 .57282 109 Total Best Friend 2.8910 .56779 107 Casual Friend 2.9667 .54173 110 Total 2.9293 .55474 217 38

PAGE 48

Table 10. Dependent Variable Des criptive Statistics (Continued) Purchase Intent Oprah MP3 Best Friend .3000 .19564 30 Casual Friend .1852 .16860 27 Total .2456 .19069 57 Photo Album Best Friend .2600 .22900 26 Casual Friend .2796 .24511 27 Total .2700 .23527 53 Total Best Friend .2814 .21077 56 Casual Friend .2324 .21375 54 Total .2574 .21269 110 Regular MP3 Best Friend .3923 .20369 26 Casual Friend .1988 .22002 25 Total .2975 .23136 51 Photo Album Best Friend .2728 .25633 25 Casual Friend .3097 .24847 31 Total .2932 .25038 56 Total Best Friend .3337 .23648 51 Casual Friend .2602 .24061 56 Total .2952 .24037 107 Total MP3 Best Friend .3429 .20298 56 Casual Friend .1917 .19319 52 Total .2701 .21147 108 Photo Album Best Friend .2663 .24042 51 Casual Friend .2957 .24521 58 Total .2819 .24230 109 Total Best Friend .3064 .22387 107 Casual Friend .2465 .22721 110 Total .2760 .22703 217 39

PAGE 49

Sample Profile The study sampled 75 male and 151 female respondents. With a total of 24,591 students enrolled for summer 2009 courses at the USF Tampa campus, 14,724 were female and 9,856 were male, which may explai n the reason for such a higher number of female respondents. Additionally, the major ity of respondents were white (n= 147), of senior class rank (n = 136), single (n = 213), and had a mean age of 22.2. The ethnicity of respondents also follows the suite of stude nts enrolled for summer 2009 courses, since the majority of students were white (n = 15,255). For a more specific summary of the sample profile, refer to Table 11 for a summary of the sample. All student profile data for USF summer 2009 students is available on the University of South Florida Info Center Web page.ii Sample profile information was collected for this study using questions that are availabl e to view in Appendix A. 40

PAGE 50

Table 11. Sample Profile n % Gender Male 75 33.0 Female 151 66.5 Missing 1 0.4 Ethnicity White 147 64.8 Black 34 15.0 Hispanic 25 11.0 Asian or Pacific Islander 10 4.4 American Indian 1 0.4 Other 8 3.5 Missing 2 0.9 College Rank Freshman 4 1.8 Sophomore 17 7.5 Junior 63 27.8 Senior 136 59.9 Graduate 3 1.3 Other 2 0.9 Missing 2 0.9 Marital Status Single 213 93.8 Married 11 4.8 Widowed 0 0.0 Separated/Divorced 1 0.4 Missing 2 0.9 Age Mean 22.2 18 3 1.3 19 20 8.8 20 32 14.1 21 56 24.7 22 46 20.3 23 24 10.6 24 15 6.6 25 7 3.1 26 8 3.5 27 4 1.8 28 5 2.2 29 1 0.4 30 1 0.4 31 1 0.4 32 1 0.4 44 1 0.4 55 1 0.4 Missing 1 0.4 41

PAGE 51

Hypothesis 1 H1a. The first hypothesis (H1a) predicted that celebrity endorsement will be more effective in generating favorable attitudes than non-celebrity e ndorsements only in low involvement product situations, while cel ebrity endorsements and non-celebrity endorsements will have no effect in high invol vement product situations in regards to the buyer’s attitudes. To test this hypothesis, the mean of the attitudes were compared across the product categories and spokespersons, as show in Tables 12 and 13. In regards to ATTA, the mean for the celebrity and low involveme nt product condition (M = 2.48) is higher than the mean for the non-celebrity and low involvement product condition (M= 2.54). Table 12. ATTA Means for Spokesperson+Product Category Spokesperson Product Friend Mean Std. Deviation N Oprah MP3 Best Friend 2.7083 .79080 30 Casual Friend 2.6759 .59167 27 Total 2.6930 .69760 57 Photo Album Best Friend 2.5577 .60954 26 Casual Friend 2.3889 .60579 27 Total 2.4717 .60776 53 Regular MP3 Best Friend 2.8269 .65867 26 Casual Friend 2.7000 .60381 25 Total 2.7647 .62931 51 Photo Album Best Friend 2.4200 .67593 25 Casual Friend 2.6210 .76341 31 Total 2.5312 .72623 56 42

PAGE 52

For ATTB, the mean for the celebrity and low involvement product condition (M= 2.94) is slightly lower than the mean for the non-celebrity and low involvement product condition (M= 2.97). In both cases ANOVA indicated no support for the hypothesis, (ATTA) F(1, 209)= .017, p =.895 and (ATTB) F(1, 209)= .534, p =.466 as illustrated in Table 21. Table 13. ATTB Means for Spokesperson+Product Category Spokesperson Product Friend Mean Std. Deviation N Oprah MP3 Best Friend 2.7222 .49583 30 Casual Friend 2.9630 .54954 27 Total 2.8363 .53126 57 Photo Album Best Friend 2.9231 .53589 26 Casual Friend 2.9506 .61813 27 Total 2.9371 .57385 53 Regular MP3 Best Friend 3.0513 .58646 26 Casual Friend 2.9067 .48610 25 Total 2.9804 .53918 51 Photo Album Best Friend 2.8933 .63625 25 Casual Friend 3.0323 .52603 31 Total 2.9702 .57657 56 Figure 2.1. displays the means of A TTA, comparing the spokespersons and product categories. Results show that att itudes were most favorable in the high involvement product category, while attit udes were least favorable in the low involvement product category for both spokespersons. 43

PAGE 53

In addition, the fact that the lines are near ly parallel suggests that the mean for each spokesperson treatments is similar. In other wo rds, there is no effect of the spokesperson. Figure 2.1. ATTA Means Figure 3.1. demonstrates the means of ATTB, comparing spokespersons and product categories. Results show that brand attitudes are most fa vorable in both product categories when a non-celebrity is used. This is in contrast to the hypothesis. 44

PAGE 54

Figure 3.1. ATTB Means H1b. The second hypothesis states that ce lebrity endorsements will be more effective in increasing buying intention th an non-celebrity endorsements only in low involvement product situations, while cel ebrity endorsements and non-celebrity endorsements will have no effect in high involvement product situations in regards to the buyer’s buying intentions. 45

PAGE 55

To test this hypothesis, the mean of the purchase intent was compared across the product categories and spokespersons, show n in Table 14. In the celebrity and low involvement product condition, the mean (M= .27) was lower than the mean for the non-celebrity and low involvement product condition (M= .29). ANOVA indicated no support for the hypothesis, F(1, 209)= .017, p =.603 as demonstrated in Table 21. Table 14. Purchase Probability Means for Spokesperson+Product Category Spokesperson Product Friend Mean Std. Deviation N Oprah MP3 Best Friend .3000 .19564 30 Casual Friend .1852 .16860 27 Total .2456 .19069 57 Photo Album Best Friend .2600 .22900 26 Casual Friend .2796 .24511 27 Total .2700 .23527 53 Regular MP3 Best Friend .3923 .20369 26 Casual Friend .1988 .22002 25 Total .2975 .23136 51 Photo Album Best Friend .2728 .25633 25 Casual Friend .3097 .24847 31 Total .2932 .25038 56 46

PAGE 56

Figure 4.1. shows the means of purchase in tent, comparing product categories and spokespersons. Purchase intent shows to be highest for both product categories when a non-celebrity is used. Figure 4.1. Purchase Probability Means 47

PAGE 57

Hypothesis 2 H2a. The third hypothesis (H2a) states that celebrity endorsements will be more effective in regards to the buyer’s attitude s than non-celebrity endorsements only when the gift recipient has low significan t relationship with the buyer. In order to test this hypothesis, the m ean of the attitudes were compared across the friend categories and spokespersons, as s hown in Tables 15 and 16. In regards to ATTA, the mean for the celeb rity and low significant rela tionship condition (M= 2.54) is lower than the mean for the non-celebrity and low significant relationship condition (M= 2.66). Table 15. ATTA Means for Spokesperson+Friend Relationship Spokesperson Product Friend Mean Std. Deviation N Oprah Regular Total Best Friend 2.6384 .71019 56 Casual Friend 2.5324 .61053 54 Total 2.5864 .66225 110 Total Best Friend 2.6275 .69169 51 Casual Friend 2.6563 .69177 56 Total 2.6425 .68861 107 48

PAGE 58

In regards to ATTB, the mean for the celebrity and low significant relationship condition (M= 2.96) is slightly lower than the mean for the non-celebrity and low significant relationship condition (M = 2.98). ANOVA shows no support for the hypothesis, (ATTA) F(1, 209)=.567, p =.452 and (ATTB) F(1, 209)=.822, p =.366, as shown in Table 21. Table 16. ATTB Means for Spokesperson+Friend Relationship Spokesperson Product Friend Mean Std. Deviation N Oprah Regular Total Best Friend 2.8155 .51998 56 Casual Friend 2.9568 .57933 54 Total 2.8848 .55196 110 Total Best Friend 2.9739 .61044 51 Casual Friend 2.9762 .50795 56 Total 2.9751 .55646 107 Figure 5.2. illustrates the means of ATTA, comparing spokesperson and friend categories. Results show that ad attitudes are least favorable when a celebrity is used in the casual friend category and most favorable wh en a non-celebrity is used in the casual friend category. This is in c ontrast to the hypothesis. 49

PAGE 59

Figure 5.2. ATTA Means Figure 6.2. shows the means of ATTB comparing spokesperson and friend categories. Results show that brand attitude s are most favorable, and almost equal, in both friend categories when a non-celebrity is used. 50

PAGE 60

Figure 6.2. ATTB Means H2b. The fourth hypothesis states that cel ebrity endorsements will be more effective in regards to the buyer’s buying intentions than non-celebrity endorsements only when the gift recipient has a low significant relati onship with the buyer. To test this hypothesis, the mean of the purchase intent was compared across the friend categories and spokespersons. 51

PAGE 61

For purchase intent, the mean of the celebrity and low significant relationship condition (M= .24) is lower than the non-ce lebrity and low significant relationship condition (M= .26). This is show in Ta ble 17. ANOVA indicated no support for the hypothesis, F(1, 209)=.257, p =.613, as demonstrated in Table 21. Table 17. Purchase Probability Means for Spokesperson+Friend Relationship Spokesperson Product Friend Mean Std. Deviation N Oprah Regular Total Best Friend .2814 .21077 56 Casual Friend .2324 .21375 54 Total .2574 .21269 110 Total Best Friend .3337 .23648 51 Casual Friend .2602 .24061 56 Total .2952 .24037 107 Figure 7.2. shows the means of purchase intent, comparing spokespersons and friend categories. Results show that purchase intent is hi ghest when a non-celebrity is used for both friend categories, which is opposit e of the hypothesis. Nearly parallel lines suggest that there is no di fference in purchase intent for both product categories and spokespersons. 52

PAGE 62

Figure 7.2. Purchase Probability Means 53

PAGE 63

Hypothesis 3 H3a. The fifth hypothesis states that the eff ect of the celebrity endorsement on the buyer’s attitudes will be the strongest in lo w involvement product situations where the recipient has low significant meaning to the buye r. In order to test this hypothesis, the mean of the attitudes were compared acro ss the product categories friend categories and spokespersons. This is displayed in Tables 18 and 19. In regards to ATTA, the mean for the celebrity, low product involvement and low significant relationship conditi on (M= 2.39) is lower than th e mean for the non-celebrity, low product involvement and low signif icant relationship condition (M= 2.63). Table 18. ATTA Means for Spokesperson+Product Involvement+Friend Relationship Spokesperson Product Friend Mean Std. Deviation N Oprah Regular Photo Album Best Friend 2.5577 .60954 26 Casual Friend 2.3889 .60579 27 Total 2.4717 .60776 53 Photo Album Best Friend 2.4200 .67593 25 Casual Friend 2.6210 .76341 31 Total 2.5312 .72623 56 54

PAGE 64

In regards to the ATTB, the mean for the celebrity, low product involvement and low significant relationship condition (M= 2.96) is lower than the mean for the noncelebrity, low product involvement and low significant condition (M= 3.04). ANOVA designated no support for the hypothesis, (ATTA) F(1, 209)=1.164, p =.205 and (ATTB) F(1, 209)=2.701, p =.102, as reported in Table 21. Table 19. ATTB Means for Spokesperson+Product Involvement+Friend Relationship Spokesperson Product Friend Mean Std. Deviation N Oprah Regular Photo Album Best Friend 2.9231 .53589 26 Casual Friend 2.9506 .61813 27 Total 2.9371 .57385 53 Photo Album Best Friend 2.8933 .63625 25 Casual Friend 3.0323 .52603 31 Total 2.9702 .57657 56 Figure 8.3. shows the means of the ATTA when product categories and friend categories are compared. Both friend categor ies were least favorable in the low product category and most favorable in the high pr oduct category, which is conflicting with the hypothesis. Since the lines are n early parallel, this proposes th at there is no difference in ad attitudes when both product a nd friend categories are used. 55

PAGE 65

Figure 8.3. ATTB Means Figure 9.3. shows the means of ATTB, co mparing product and friend categories. Results show that ATTB is most favorab le in both low involving categories. 56

PAGE 66

Figure 9.3. ATTB Means H3b. The final hypothesis states that the e ffect of the celebrity endorsement on the buyer’s buying intentions will be the str ongest in low involvement product situations where the recipient has low signi ficant meaning to the buyer. To test this hypothesis, the mean of the purchase intent was compar ed across the produc t categories, friend categories and spokespersons, shown in Table 20. 57

PAGE 67

The mean of the celebrity, low product invo lvement and low significant relationship condition (M= .28) is lower than the mean for the non-celebrity, lo w product involvement and low significant relations hip condition (M= .31). ANOVA i ndicated no support for the hypothesis, F(1, 209)=.627, p =.429, as illustrates in Table 21. Table 20. Purchase Probability Means for Spokesperson+Product Involvement+Friend Relationship Spokesperson Product Friend Mean Std. Deviation N Oprah Regular Photo Album Best Friend .2600 .22900 26 Casual Friend .2796 .24511 27 Total .2700 .23527 53 Photo Album Best Friend .2728 .25633 25 Casual Friend .3097 .24847 31 Total .2932 .25038 56 Figure 10.3. demonstrates the means of purchase intent when both spokesperson means for each product and friend treatments we re averaged. This figure shows that the purchase intent for a casual friend and MP3 player is lower than the purchase intent for a casual friend and photo album. The opposite is true for a best friend. 58

PAGE 68

Figure 10.3. Purchase Probability Means 59

PAGE 69

Table 21. ANOVA Results Indep Variable Dep Variable df F Sig. Corrected Model ATTA 7 1.309 .247 ATTB 7 .970 .454 Purchase Intent 7 2.259 .031 Intercept ATTA 1 3270.279 .000 ATTB 1 6015.033 .000 Purchase Intent 1 329.043 .000 Spokesperson ATTA 1 .421 .517 ATTB 1 1.154 .284 Purchase Intent 1 1.507 .221 Product ATTA 1 6.388 .012 ATTB 1 .267 .606 Purchase Intent 1 .143 .706 Friend ATTA 1 .121 .728 ATTB 1 .755 .386 Purchase Intent 1 4.317 .039 Spokesperson Product ATTA 1 .017 .895 ATTB 1 .534 .466 Purchase Intent 1 .271 .603 Spokesperson Friend ATTA 1 .567 .452 ATTB 1 .822 .366 Purchase Intent 1 .257 .613 Product Friend ATTA 1 .275 .601 ATTB 1 .054 .816 Purchase Intent 1 9.062 .003 Spokesperson Product Friend ATTA 1 1.614 .205 ATTB 1 2.701 .102 Purchase Intent 1 .627 .429 60

PAGE 70

Table 21. ANOVA Results (Continued) Error ATTA 209 ATTB 209 Purchase Intent 209 Total ATTA 217 ATTB 217 Purchase Intent 217 61

PAGE 71

CHAPTER 6: DISCUSSION Other Findings As far as the non-significant statistical re sults that were measured based on the hypotheses, significance was indeed found in ot her areas. The ANO VA results indicate that respondents in the MP3 category had sign ificantly higher attitudes towards the ad than respondents in the phot o album category, F(1, 209)= 6.39, p = .01. This may be because photo albums do not hold as much salie ncy as MP3 players, as indicated in the manipulation checks. In other words, a photo al bum is not something that one finds to be noteworthy or striking. Sin ce the hypothesis, according to the ELM, was stated in relation to the low involvement product categor y (the photo album) it is possible that the null results are due to an aversion to th e product chosen. The significant difference reported above suggests this is the case. While a photo album is certainly a low involving product, it may not be something pe ople are willing to purchase regardless of the celebrity endorsements. Basically, ther e are limits to the effects of celebrity endorsements. 62

PAGE 72

The results also indicate that responden ts in the best friend category were significantly more likely to pur chase the product than respondents in the casual friend category, F(1,209)=4.31, p=.04. This results does not tell us much about the ELM, rather it confirms the findings in gift-giving rese arch and our general intuition about human relationships—that people are more likely to buy a gift for someone who holds a significant meaning to the buyer than for so meone who holds a less significant meaning to the buyer. Finally, the results indicate that re spondents in the high involvement product category who were buying the MP3 player for a best friend were more likely to purchase the product than respondents who were buying the MP3 for a casual friend, F(1, 209)= 9.062, p <.01. These results are in line with the ELM’s explanation of motivation and the gift-giving research presented earlier. In ot her words, people are mo tivated to consider the true merits of a high involving product, such as an MP3 player, especially if it is for someone that holds special meaning to th e gift-giver, such as a best friend. Study Limitations As with any other study, it cannot be said that this study ha s reached perfection. Based on an overall review, there are certa in limitations that can be summarized First, it is important to note that when choosing the products for this study, it was essential to keep in mind that these products were to be purchase d by a college student and for someone as a gift. It was imperative to choose products that were affordable for college students. However, the products chosen may have influenced the results of this study. 63

PAGE 73

For instance, an MP3 player has recently become so commonplace and may not be viewed as high involving. Furthermore, even though manipulation ch ecks confirmed that both products were viewed as high and low involving, it may be that conceptually an MP3 player and photo album are more sim ilar than suspected. High and low involving products can be somewhat subjective, de pending on the consumer’s opinions. The experiment should have pointed out that th e MP3 player was more expensive than the photo album in order to clarify how high and low involving the products were. Since both products may not be far enough apart, in rega rds to involvement, this may have created an unwanted bias to the study. The same limita tion explanation can be said for the friend relationships chosen. Second, the celebrity chosen, Oprah Winfrey is unlike any other celebrity. She is unlike any ordinary celebrity, especially known for successfully endorsing various products, and even people. The unique nature of the celebrity chosen for this study may introduce unrecognized biases (t his possibility is addresse d more fully in the next section). Third, the adve rtisements that were used in the experiment were not true ads. These ads were not professionally designe d based on marketing research, unlike the various ads seen in magazines. Therefore, there may have been a lack of realism or practicality that respondents ar e used to seeing these types of ads, especially those that Oprah has been a part of in the past. Fourth, as noted in the reliability portion, the items designed to measure brand attit udes are not internally consis tent, as reflected by in the reliability portion of this manuscript. 64

PAGE 74

Instead, the Cronbach’s alpha for ATTB is .676, which is rather low. Lastly, the experiment was conducted only with college st udents, which means that the results can only be generalized of similar populations. A ll of these limitations are important to be considered when evaluating the results. The Use of Oprah Winfrey The results indicate that the use of a celeb rity does not generate more favorable attitudes or increase purchas e intention than non-celebrit ies in low involving product categories and low significant friend relations hips. A reason for these findings can be reflected on the celebrity used in the experiment. A more recent example of Oprah’s endorsement influence is the result of th e 2008 presidential nomi nation campaign. After examining data that studied the influence of her endorsement of Barack Obama, results showed that the endorsement did not direc tly influence people’s opinions toward Obama or the level of Obama’s likeability (Pease & Brewer, 2008). Instead, Oprah’s endorsement had an indirect effect, causing respondents to perceive Obama as likely to win the nomination and thereby helping hi s campaign (Pease & Brewer, 2008). This shows that the effect of Oprah, and celebrity endorsers in general, can have complex effects of attitudes. In the case of the 2008 election and, potentially the present study, the effect of Oprah may be different than we expect. This may explain the null results reported. 65

PAGE 75

It is apparent that Oprah is someone who is known to endorse an array of products. As mentioned in the review of litera ture, celebrity endorsers are most effective when they possess credibility. In other word s, celebrity endorser s can be successful because consumers trust that celebrities are genuine in their endorsement of the product rather than motivated simply by endorsement fees (Atkin & Block, 1983). However, this may not be the case in all similar situa tions. Silvera & Austad (2004) found that participants were cynical towa rd the endorsers’ motives (i.e., receiving standard fees), which may have influenced this study’s resu lts. Research has shown that multiple product endorsements by celebrities affect consumers’ attitudes and intent ions (Tripp, Jenson & Carlson, 1994). In other words, the more e ndorsements a celebrity is involved in, the lower their credibility and likeability becomes in the eyes of the consumer. In short, it is possible that while respondents clearly perceive d Oprah as a credible celebrity, they were nonetheless fatigued by Oprah’s overexposure and thus unaffected by the celebrity stimulus. In the simplest of terms, Oprah ma y be too much of a celebrity and, in certain situations, her endorsement may have no effect. The intended contribution of this research was to introduce a ne w variable to the ELM literature: gift giver-receiver rela tionship. Though the study found no signifying results in this domain, additional research is needed to understand how celebrity endorsements influence attitudes and purchase intention. If the effect of celebrity endorsements on gift-giving for low involveme nt produces is truly non-existent this would suggest s limitation of the ELM. In al l likelihood, however, the null effect is due to the problems outlined in the limitations section. 66

PAGE 76

CHAPTER 7: CONCLUSION This experimental study sought a better unde rstanding of consumer’s attitudes and purchase intentions in different product i nvolvement and gift gi ver-receiver conditions. While past research has shown that the us e of a celebrity as a spokesperson impacts buyers, the subject has not been studied when consumers’ intentions are to purchase a product solely as a gift. Over the years, re search has shown that celebrity endorsements are effective (Agrawal et al., 1995; Byrn e et al., 2003; Callcoat & Phillips, 1996; Ohanian, 1991; Silera & Austad, 2004). More importantly, celebrity endorsements are useful in changing people’s at titudes and buying intentions, es pecially when it comes to a low involving product, according to the ELM (Clark & Horstmann, 2005; Petty & Cacioppo, 1981; Petty et al., 1983). Contrary to previous res earch, the results of this study do not support the hypotheses and show th at the use of a celebrity does not influence buyers’ attitudes or purchase inten tions in low involving product situations and giver-receiver relationships, compar ed to a non-celebrity spokesperson. Although non-significant results were measured based on the hypotheses, significance was found in the pr oduct category, friend category, and product and friend interaction. Unlike the hypothesi zed results, these additional results are consistent with ELM and gift-giving research. 67

PAGE 77

As previously mentioned, ELM has two di fferent routes of persuasion: Central and peripheral. Research shows that the use of a celebrity is benefi cial in low involving product situations. However, one needs to be motivated and possess ability to process a message before anyone can be persuaded. In terms of ELM, “Much depends on the biases and beliefs that audience members bri ng to the persuasion situ ation and the extent to which receivers are motivated and able to process the message” (Perloff, 2003, p. 153). Motivation is influenced by one’s need for cognition. Possessing some sort of motivation is necessary in order to ela borate on a message and bring out central processing (Petty & Cacioppo, 1981). In other word s, one must be motivated or driven to think or learn about a product before attit udes are changed. However, one person’s need for cognition may be different from anot her’s need. While some people may be influenced by certain superficial or periphe ral factors, such as celebrities, to buy a product, others have a constant need for cognition and will not be influenced by peripheral factors. Instead, those who have a constant need for cognition will seek out the true merits of a product, even if it is a low involving. Another central aspect of ELM that effect s one’s motivation to process is personal relevance (Petty & Cacioppo, 1981). Personal re levance is something that determines whether the consumer is willing to pay attenti on to an advertisement. For instance, if one is currently in the market to buy a car, a car advertisement may catch the consumer’s attention and motivate them to cognitively pay attention to the message and its information. However, if the person is not in the market for such a product, the message may be ignored. 68

PAGE 78

Although past research has found that gift-giving is more i nvolving than se lf-purchasing, it may not be the case for all consumers (Belk, 1982). Buying something for someone else may not be as personally relevant as buying something for one’s self. Therefore, the viewer of a message is not motivated to seek information about the product. Although non-significant results were measured based on the hypotheses, significance was found in the pr oduct category, friend category, and product and friend interaction. These other resu lts are consistent with ELM and gift-giving research. With the lack of research combining cel ebrity endorsements and gift-giving, and the findings of this study imply that there is more research to be done. In order for marketers to better appreciate buyers’ attitudes a nd purchase intentions in the area of giftgiving, suggestions are offered fo r future studies. Future resear ch may want to merely use participants who find gift-giving high i nvolving, because although gift-giving research has found that gift-giving is a high involving process, it may not be for all givers. This may lead to significant results that would fu rther tap into consumer s’ attitudes and buying intentions and better aid mark eters in various popular gift -giving and gift exchanging occasions, such as Christmas. Previous research on celebrity endorsement s has not been done over long periods of time. Future research ma y benefit from tracking celeb rity endorsements over longer periods of time in order to find greater sign ificance in regards to people’s attitudes and purchase intentions, depe nding on the celebrity. 69

PAGE 79

For the reasons mentioned above, it can be assumed that the use of an atypical celebrity may have created factors that hamper ed the results of this study. Future studies are encouraged to use one or more “ordinary” celebrities in order to determine if gift purchases are influenced by th ese endorsers. In addition, a mo re diverse selection of high and low product categories for futu re studies may be beneficial. Finally, it is important to note that the number and type of respondents used in this study may have caused certain limitations. Additional studies may want to incorporate the use of both college and non-co llege samples in order to achieve a more diverse sample. Since the number of responde nts is quite low, future studies are encouraged to obtain a larger sample. 70

PAGE 80

REFERENCES Agrawal, Jagdish, Kamakura & Wagner ( 1995). The economic worth of celebrity endorsers: An event study analysis. Journal of Marketing, 59 (3). Atkin, C. & Block, M. (1983), “Effec tiveness of celebrity endorsers.” Journal of Advertising Research 23 (1). Belk, R. W. (1982). Effects of Gift-Giving Involvement on Gi ft Selection Strategies. In Advances in Consumer Research, 9. (A. Mitchell, Ed.) Ann Arbor, MI: Association for Consumer Research. Benoy, J.W. (1982). The credibility of physi cally attractive communicators; A review Journal of Advertising 11 (3), 19. Boorstin, D. J. (1961). The Image: A Guide to Pseudo-Events in America New York, NY: Macmillian Publishing Co. Boyd, T.C. & Shank (2004). Athletes as pr oduct endorsers: The effect of gender and product relatedness. Sport Marketing Quarterly, 13 (2). Byrne, A., Whitehead, A., & Breen, S. (2003). The naked truth of celebrity endorsement. British Food Journal 105 (4). Callcoat, M. F. & Phillips, B. J. (1996). Observations: Elves make good cookies: creating likable spokes-chara cter advertising. Journal of Advertising Research 36 (5), 739. Celsi, R. L. & Olson, J. C. (1988). The role of involvement in attention and comprehension processes. The Journal of Consumer Research, 15 (2). Clarke, K. & Belk, R. W. (1979). The E ffects of Product Involvement and Task Definition on Anticipated Consumer Effort. In Wilkie, W. L. (Ed.), Advances in consumer research (Vol. 6, pp. 313-318). Ann Arbor, MI: Association for Consumer Research. Clark, C. R. & Horstmann, I. J. (2005). A dvertising and coordination in markets with consumption scale effects. Journal of Economics & Management Strategy, 14 (2), 380. 71

PAGE 81

Federal Trade Commission. (1980). FTC Guides Concerning Use of Endorsements and Testimonials in Advertising. Friedman, H.H., Termini, S. & Washi ngton, R. (1976). The effectiveness of advertisements utilizing four types of endorsers. Journal of Advertising, 5 (3). Goldsmith, R. E., Lafferty, B. A., & Newell, S. J. (2000). The impact of corporate credibility and celebrity credibility on consumer reaction to advertisements and brands. Journal of Advertising, 29 (3). Goodwin, C., Smithe, K.L., & Spiggle, S. ( 1990). Gift-giving: Consumer motivation and the gift purchase process. Advances in Consumer Research, 17 (1). Gronhaug, K. (1972). Buying situation and buyer’s information behavior. European Marketing Research Review, 7. Howard, J. A. & Sheth, J. N. (1969). The Theory of Buyer-Behavior New York, NY: John Wiley & Sons, Inc. Hoyer, W. & MacInnis, D. (2007). Consumer Behavior. Boston, MA: Houghton Mifflin Co. Kahle, L.R. & Homer, P.M. (1985). Physical attractiveness of the celebrity endorser: A social adaptation perspective. Journal of Consumer Research 11 959. Komter, A. & Vollebergh, W. (1997). Gift-g iving and the emotional significance of family and friends. Journal of Marriage and the Family, 59 747, 756. Laroche, M., Gad, S., Browne, E., Cleveland, M., & Kim, C. (2000). Determinants of instore information search strategies pe rtaining to a Christmas gift purchase. Canadian Journal of Admini strative Sciences, 17 (1), 4. Lynch, J. & Schuler, D. (1994). The matc h up effect of spokesperson and product congruency: A schema theory interpretation. Psychology & Marketing, 11 (4), 418. McCracken, G. (1989). Who is the celebrity endorser? Cultural foundations of the endorsement process. The Journal of Consumer Research, 16 (3). Ohanian, R. (1991). The impact of celeb rity spokespersons’ perceived image on consumers’ intention to purchase. Journal of Advertising Research 31 (1). 72

PAGE 82

O’Mahony, S. & Meenaghan, T. (1997/1998). The impact of celebrity endorsements on consumers. Irish Marketing Review, 10 (2), 15. Otnes, C., Lowrey, T.M., & Kim, Y.C. ( 1993). Gift selection for easy and difficult recipients: A social roles interpretation. Journal of Consumer Research, 20 Pease, A. & Brewer, P. R. (2008). The Op rah factor: The effects of a celebrity endorsement in a presidential primary campaign. The International Journal of Press/Politics, 13 (386). Perloff, R.M. (2003). The dynamics of persuasion: Co mmunication and Attitudes in the 21st Century (2nd Ed.). Mahwah, NJ.: LEA. Petty R. E. & Cacioppo, J. T. (1981). Attitudes and Persuasion: Classic and Contemporary Approaches. Dubuque, IA: WM. C. Brown. Petty R. E. & Cacioppo, J. T. (1986). Communication and Persuasion: Central and Peripheral Routes to Attitude Change. New York, N.Y.: Springer-Verlag New York Inc. Petty, R.E., Cacioppo, J.T., & Schumann, D. ( 1983). Central and peri pheral routes to advertising effectiveness: The moderating role of involvement. Journal of Consumer Research, 10. Petty, R. E. & Wegener, D. T. (1999). The Elaboration Likelihood M odel: Current Status and Controversies. In Chaiken, S. & Trope, Y. (Eds.), Dual Process Theories in Social Psychology (p. 43). New York: Guilford Press. Rothchild, M. L. (1977). Advertising St rategies for High and Low Involvement Situations. In Maloney, J. C. & Silverman, B. (Ed.), Attitude research plays for high stakes Chicago, IL: American Marketing Association. Silvera, D.H. & Austad, B. (2004). Factors predicting the effectiv eness of celebrity endorsement advertisements. European Journal of Marketing, 38 (11), 1511-2, 1520. Shimp, T.A. (2000). Advertising Promotion: Supplemental Aspects of Integrated Marketing Communications. Fort Worth: TX: Dryden Press. Shapiro, B. A. (1970). The Effects of Price on Purchase Behavior. In Sparks, D. (Ed.), Broadening the Concept of Marketing Chicago, IL: American Marketing Association. 73

PAGE 83

Smith, S.M. Beatty, S.E. (1985). An Examin ation of Gift-Giving Purchasing Behavior: Do Shoppers Differ in Task Involvement Search Activity or Perceptions of Product Selection Risk? In Lusch (Eds.), AMA Educator’s Proceedings Chicago, IL: American Marketing Association. Till, B.D. (1998). Using celebrity endorsers effectively: Lessons from associative learning. The Journal of Product and Brand Management, 7 (5), 401, 403-5. Till, B.D. & Busler, M. (1998). Matching produc ts with endorsers: Attractiveness versus expertise. The Journal of Consumer Marketing, 15 (6). Tripp, C., Jensen, T.D., & Carlson, L. (1994). The effects of multiple product endorsements by celebrities on consum ers’ attitudes and intentions. Journal of Consumer Research, 20 UCLA Academic Technology Services. (2009). SPSS FAQ: What does Cronbach’s alpha mean ? Retrieved 6/14/2009, from http://www.ats.ucla.edu/st at/Spss/faq/alpha.html i As indicated by histogram gr aphs (not reported here). ii http://usfweb3.usf.edu/infocenter/?silverh eader=2&report_category=ADM&report_type= SMAJF&reportid=1145 74

PAGE 84

APPENDICES 76

PAGE 85

Appendix A. 1. Participant Background Section 1—Background Instructions: Please check the box that best answers the question. 1. Gender Male Female 2. Ethnic Origin White (non-Hispanic) 1. Asian or Pacific Islander 4. Black (non-Hispanic) 2. American Indian/Alaskan Native 5. Hispanic 3. Other: ___________________ 6. 3. Age _______ (write age here) 4. College Rank Freshman 1. Senior 4. Sophomore 2. Graduate Student 5. Junior 3. Other: ___________________ 6. 5. Are You Currently (check only one) Single 1. Widowed 3. Married 2. Separated/Divorced 4. 77

PAGE 86

Appendix B. 1. Participant Instructions Section 2— Advertisem ent Instructions Instructions: In a moment you will flip to page 2 of the questionnaire. On page 2 is an advertisement for a product. Please take a moment to examine the ad. After examining the advertisement flip to p age 3 and answer a few short questions pertaining to the ad. Please DO NOT return to the advertisement on page 2 while answering the questions. Remember to do your own work and do not refer to anyone else’s packet. 78

PAGE 87

Appendix C. 1. Non-Celebrity, MP3 Treatment 79

PAGE 88

Appendix C. 2. Non-Celebr ity, Photo Album Treatment 80

PAGE 89

Appendix C. 3. Celebrity, MP3 Treatment 81

PAGE 90

Appendix C. 4. Celebrity, Photo Album Treatment 82

PAGE 91

Appendix D. 1. Advertisement Questions Section 3—Advertisement Questions Instructions: In this section you are an swering a series of questions about the advertisement you just examin ed. Remember not to flip back to the advertisement After carefully reading the questions and considering what the question is asking, circle the response that best describes what you thought about the ad. Please take your time. 6. Regarding the advertisement as a whole how convincing is it? Would you say it is very convincing, somewhat convincing, so mewhat unconvincing, very unconvincing? Very Convincin g Somewhat Convincin g Neutral Somewhat Unconvincin g Very Unconvincin g 5 4 3 2 1 7. Regarding the advertisement as a whole how informative is it? Very informative, somewhat informative, somewhat uninformative, very uninformative? Very Informative Somewhat Informative Neutral Somewhat Uninformative Very Uninformativ 5 4 3 2 1 8. Regarding the advertisement as a whole do you feel that it was a very good ad, a good ad, a bad ad, or a very bad ad? Very Unfavorable Somewhat Unfavorable Neutral Somewhat Favorable Very Favorable 5 4 3 2 1 83

PAGE 92

9. Regarding the advertisement as a whole how likeable is the ad? Do you like very much, like, dislike, dislike very much? Like Very Much Like Neutral Dislike Dislike Very Much 5 4 3 2 1 10. Regarding the brand in this advertisement, how favorable do you feel toward this brand? Very unfavorable, somewhat unfavorable, somewhat favorable, very favorable. Very Unfavorable Somewhat Unfavorable Neutral Somewhat Favorable Very Favorable 1 2 3 4 5 11. Regarding the brand in this advertisement, what kind of feelings do you have about the brand? Do you have very negative feelings, somewhat negative feelings, some what positive feelings, very positive feelings? Very Negative Feelin g s Somewhat Ne g ative Neutral Somewhat Positive Very Positive Feelin g s 1 2 3 4 5 12. Regarding the brand in this advertisement, how likeable is the brand? Do you like very much, like, dislike, dislike very much? Like Very Much Like Neutral Dislike Dislike Very Much 5 4 3 2 1 84

PAGE 93

13. Regarding the spokesperson in this advertisement, how effective do you find her? Is she very effective, somewhat effective, somewhat ineffective, very ineffective? Very Effective Somewhat Effective Neutral Somewhat Ineffective Very Ineffective 5 4 3 2 1 14. Regarding the spokesperson in this advertisement, what kind of feelings do you have towards her? Do you have very positive, somewhat positive, somewhat negative, or very negative feeli ngs toward her? Very Positive Feelings Somewhat Positive Feelings Neutral Somewhat Negative Feelings Very Negative Feelings 5 4 3 2 1 15. Regarding the spokesperson in this advertisement, how likeable is the spokesperson? Very likeable, lik eable, unlikable, very unlikable? Very Likeable Somewhat Likeable Neutral Somewhat Unlikable Very Unlikable 5 4 3 2 1 85

PAGE 94

Instructions: Please use the following sc ale to write in your answer (0%-100%). Zero percent means you would definitely not purchase the gift and 100% means you would definitely purchase the gift. 16. What is the likelihood you will purchase this product as a gift for a [best friend/casual friend inserted here] ? Please note that a best friend is not a casual friend. Instead, a best friend is someone you speak to or see on a regular basis and is very close to you because you trust them / Instead, casual friend is someone you might socialize with in a group or even one on one, but they are not your best friend Zero percent (0%) means you would definitely not purchase the gift and 100% means you would definitely purchase the gift. What is your likelihood you would purch ase this product as a gift for a [best friend/casual friend] ? _________ (write percentage here from 0% to 100%) 0% 50% 1 00% Definitely Not Purchase Maybe Purc hase Definitely Purchase 86

PAGE 95

Appendix E. 1. Celebrit y Manipulation Check Instructions: In regards to the spok esperson, check each that apply: 17. The spokesperson is: _____ a celebrity or _____ not a celebrity 18. The spokesperson is: _____ popular or _____ unpopular 19. The spokesperson is: _____ familiar or _____ unfamiliar 20. The spokesperson is: _____ likable or _____ unlikable 87

PAGE 96

Appendix E. 2. Non-Celebrity Manipulation Check Instructions: In regards to the spok esperson, check each that apply: 17. The spokesperson is: _____ a celebrity or _____ not a celebrity 18. The spokesperson is: _____ popular or _____ unpopular 19. The spokesperson is: _____ familiar or _____ unfamiliar 20. The spokesperson is: _____ likable or _____ unlikable 88

PAGE 97

Appendix E. 3. High Involvement Product Manipulation Check Instructions: Answer each question for the product category listed as if you were buying the product for a [best friend/casual friend inserted here] Please circle the number that best represents your response to the product MP3 Player 21. In selecting from many types and brands of this product available in the market, would you say that: I would not care a 1 2 3 4 5 6 7 I would care a great great deal as to which one deal as to which one I buy as I buy as a gift a gift 22. How important would it be to you to make a right choice of this product? Not at all important 1 2 3 4 5 6 7 Extremely important 23. In making your selection of this produc t, how concerned would you be about the outcome of your choice? Not at all concerned 1 2 3 4 5 6 7 Very much concerned 89

PAGE 98

Appendix E. 4. Low Involvement Product Manipulation Check Instructions: Answer each question for the product category listed as if you were buying the product for a best friend Please circle the number that best represents your response to the product. Photo album 21. In selecting from many types and brands of this product available in the market, would you say that: I would not care a 1 2 3 4 5 6 7 I would care a great great deal as to which one deal as to which one I buy as I buy as a gift a gift 22. How important would it be to you to make a right choice of this product? Not at all important 1 2 3 4 5 6 7 Extremely important 23. In making your selection of this produc t, how concerned would you be about the outcome of your choice? Not at all concerned 1 2 3 4 5 6 7 Very much concerned 90

PAGE 99

Appendix E. 5. High Involving Friend Manipulation Check Instructions: Please rank a best friend in terms of importance, significance and saliency Circle the number 1 to 5 for each question. Best Friend 24. Very important Somewhat important Neutral Somewhat unimportant Very unimportant 5 4 3 2 1 25. Very significant Somewhat significant Neutral Somewhat insignificant Very insignificant 5 4 3 2 1 26. Very salient Somewhat salient Neutral Somewhat non-salient Very non-salient 5 4 3 2 1 91

PAGE 100

Appendix E. 6. Low Involving Friend Manipulation Check Instructions: Please rank a casual friend in terms of importance, significance and saliency Circle the number 1 to 5 for each question. Casual Friend 24. Very im p ortant Somewhat im p ortant Neutral Somewhat unim p ortant Very unim p ortant 5 4 3 2 1 25. Very si g nificant Somewhat si g nificant Neutral Somewhat insi g nificant Very insi g nificant 5 4 3 2 1 26. Very salient Somewhat salient Neutral Somewhat non-salient Very non-salient 5 4 3 2 1 92


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200385Ka 4500
controlfield tag 001 002045516
005 20091229123329.0
007 cr mnu|||uuuuu
008 091229s2009 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002979
035
(OCoLC)495351595
040
FHM
c FHM
049
FHMM
090
P90 (Online)
1 100
Anghel, Christine.
4 245
The effect of celebrity endorsements on gift-giving purchases :
b an application of the elaboration likelihood model
h [electronic resource] /
by Christine Anghel.
260
[Tampa, Fla] :
University of South Florida,
2009.
500
Title from PDF of title page.
Document formatted into pages; contains 92 pages.
502
Thesis (M.A.)--University of South Florida, 2009.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
3 520
ABSTRACT: The purpose of this study is to determine how effective celebrity endorsements are in regards to the type of gift purchase one decides to make (i.e., buying for someone who has a high significant meaning to the buyer, such as a best friend, versus buying for someone who has a low significant meaning to the buyer, such as a casual friend). The study seeks to extend upon the anthropology research exploring gift-giving and marketing research exploring celebrity endorsements by applying the tenants of the Elaboration Likelihood Model (ELM). This study uses an experimental procedure in order to determine the effect of using celebrity endorsements on buyers' attitudes and purchase intentions for gift-giving purchases in low and high involving categories. Results indicate that celebrity endorsements have no influence on attitudes and purchase intention in different product involvement and gift giver-receiver conditions.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
590
Advisor: Scott Liu, Ph.D.
653
Product involvement
Oprah Winfrey
Advertising
Marketing
Attitudes
0 690
Dissertations, Academic
z USF
x Mass Communications
Masters.
773
t USF Electronic Theses and Dissertations.
856
u http://digital.lib.usf.edu/?e14.2979