Matthew Thomson

Human Brands: InvestigatingAntecedents to Consumers' Strong

Attachments to CelebritiesThis article explores recent advances in self-determination research to address why consumers develop strongattachments to "human brands," a term that refers to any well-known persona who is the subject of marketingcommunications efforts. Study 1 uses a survey that is analyzed with structural equation modeling. Study 2 isqualitative and offers corroborating evidence for the proposed theoretical model. Study 3 extends the model with amore naturalistic sample and tests several alternative hypotheses using hierarchical regression. The resultssuggest that when a human brand enhances a person's feelings of autonomy and relatedness and does notsuppress feelings of competence, the person is likely to become more strongly attached to it. This article documentsthat strong attachments are predictive of satisfied, trusting, and committed relationships and proposes thatattachment strength may be a parsimonious proxy for consumer-brand relationship strength. The results imply thatbenefits would accrue to organizations such as entertainment firms and political parties that establish direct androutine interaction between human brands and consumers, that human brands to which consumers are attachedoffer significant potential as endorsers, and that organizations should address how to make the human brands theymanage more authentic.

"When I go to the grocery store, people hug me. But that'sOK. People do know me. They've seen me grow up ontelevision."

—Michelle Kwan'

I n the past decade, academic and practitioner interest inconsumer relationships has flourished (e.g., Sheth andParvatiyar 1995). One approach to studying these rela-

tionships focuses on feelings of attachment and love (e.g.,Carroll and Ahuvia 2006). which lay "at the core of allstrong brand relationships" (Fournier 1998. p. 363).

This article focuses on consumers' attachments tohuman brands, a tetrn that refers to any well-known personawho is the subject of marketing communications efforts(e.g., Rindova, Pollock, and Hayward 2006). The centralpurpose of this article is to address why consumers formstrong attachments to human brands. From a practicalstandpoint, advancing the understanding of the relationshiphetween consumers and human brands is important becausemuch of the success of the $190 billion U.S. media andentertainment industry hinges on the successful positioningof one of its key assets, the celebrity (Plunkett Research

'Michelle Kwan, world champion figure skater, quoted inHorovitz (2tX)3).

Matthew Thomson is an assistant professor. Queen's School of Business(e-mail: mthomson@business.queensu.ca). The author thanks DebbieMaclnnis, C.W. Park, Brian Lickel, Allen Weiss, Jay Handelman, PeterDacin. Allison Johnson, and the anonymous JM reviewers for feedback onprevious drafts of this article.

To read or contribute to reader and author dialogue on this article, visithttp://www.marketingpower.comfimblog.

2004). Organizations spend vast sums annually in an effortto establish psychological connections between consutnersand human brands, such as Michael Jordan, Tiger Woods,and Oprah (Serwer 2(X)l). Firms even specialize in brandmanagement with respect to dead celebrities, such as SteveMcQueen and John Wayne (Halpern 2005). Similarly, in the2004 U.S. presidential election, each of the two main politi-cal parties spent in excess of $1 billion dollars in an effortto position their candidates favorably with voters (Edsalland Grinialdi 2004). That is, organizations are already dedi-cated to managing human brands and building emotionalbonds with consumers. This article is an effort to explorethis phenomenon methodically; it does so by advancinghypotheses and testing them in three studies. Study 1 uses asurvey that is analyzed with structural equation modeling.Study 2 is qualitative and offers corroborating evidence forthe proposed theoretical model. Study 3 extends the modelwith a more naturalistic sample and tests several alternativehypotheses using hierarchical regression.

Conceptual FoundationHuman Brands and the Branding Literature

"Your client, whether they are an athlete or an actor or anactress, has intangible assets: a name, a reputation, a cred-ibility and an image. All of those aitributes may be com-bined into something that could he made into a brand."

—Brian Dubin (quoted in Towle 20()3)

This article's underlying premise is that human brands maybe viewed as one of several operationalizations of thebroader concept of a brand. In marketing, the term "brand"is typically applied to firms, products, and services, and ingeneral, marketers accept that brands may be described in

© 2006, American Marketing AssociationISSN: 0022-2429 (print), 1547-7185 (eiectronic) 104

Journal of MarketingVoi. 70 (July 2006). 104-119

terms of perceived quality, image, and so forth. Celebritiescan also be considered brands because they can be profes-sionally managed and because they have additional associa-tions and features of a brand. Evidence for this idea can per-haps be most readily found in the context of politicalcampaigns, in which the candidate's message, publicappearance, endorsements, and so on. are all controlled byconsultants and political parties in the hopes of managingperceived quality and brand image to increase "marketshare" at the polls (Simon, Gilgoff, and Samuel 2004).Likewise, the National Basketball Association recentlyestablished an off-the-court dress code for its players in aneffort to improve the league's appeal by managing theimage of its individual human brands (Lee 2005). However,many of the players (e.g., Allen Iverson) had previouslyestablished and lucrative images that were misaligned withthe new dress code; this underscores the importance ofunderstanding how to better manage the brand equity ofindividual human brands.

Why Attachments MatterConsiderable research has examined features of the varioustypes of relationships that people form with human brands.For example, idolatry (Houran, Navik, and Zerrusen 200.^),fandoin (Leets, De Becker, and Giles 1995), and celebrityworship (Dietz et al. 1991) have been explored. Mucb ofthis research has been advanced under the mhric of attach-ment theory and its associated literature, which is both welliuticulated and relevant to marketing (Kleine and Baker2004). Attachments are a type of strong relationship thatpeople usually first experience as children with their par-ents; later in life, these attachments routinely develop withother "targets," such as human brands (Leets, De Becker,and Giles 1995). A person immersed in such an emotionallysignificant relationship normally perceives the relationshippartner as differentiated and irreplaceable (i.e.. target spe-cific; see Hazan and Shaver 1994). When these types ofrelationships are experienced in reference to human brands,they are typically referred to as "secondary object" attach-ments and have been descrihed as "intimacy at a distance"(Horton and Wohl 1956).

This article focuses on the strength of attachment,defined as the intensity of a person's target-specific emo-tional bond with a human brand. Prior research reveals thatattachments arc differentiated from other constructs. Forexample, the strength of an attachment is orthogonal toinvolvement, satisfaction, loyalty, and attitude favorability(Ambler et al. 2002; Thomson, Maclnnis, and Park 2005).

This literature suggests that attachment theory can makea contribution to marketing because of the distinctive quali-ties of an attachment. For example, much as marketers mayattempt to create relationships that are trusting, committed,and -satisfied (Fournier, Dobscha, and Mick 1998), peopleroutinely report elevated levels of each in their attachmentsto a variety of objects (Rempel, Ross, and Holmes 2001;Spake et al. 2004). Prior literature points to the indepen-dence of attachment strength from trust, satisfaction, andcommitment, but the correlation among these constructssuggests that understanding how to create or intensifyattachments could offer both an effective and an economical

means of achieving stronger marketing relationships.Indeed, Foumier's (1998) article on consumer-brand rela-tionships suggests that feelings linked to attachments arefundamental to strong brand relationships. Therefore,attachment strength may provide a parsimonious and unidi-mensional indicator of "relationship quality" or strength.

Furthermore, more intense attachments are linked toseveral consequences that are desirable to marketers. Forexample, attachments in a marketing context may preventconsumer defections (Liljander and Strandvik 1995),increase consumers' forgiveness in tbe face of negativeinformation (Ahluwalia. Unnava. and Bumkrant 2001), andpredict brand ioyalty and willingness to pay (Thomson,Maclnnis, and Park 2005). If marketers understand whatdetermines the strength of attachments, they may be in abetter position to foster more durable relationships withconsumers.

Finally, research implies that attachments may be piv-otal to understanding customer-based brand equity, inwhich certain brands "resonate" with consumers and indi-cate financial value for the firm (Ambler et al. 2002; Keller2001). For various reasons, therefore, contemplating whatstrengthens attachments seems to be an endeavor that isworthwhile from both a practical and a theoreticalstandpoint,

Relationship similarities. In some respects, relation-ships involving human brands are a hybrid of other relation-ships. For example, they are pertinent to marketing becauseof the central role of the consumer, but they are also a rele-vant extension of interpersonal research because they impli-cate a bona fide human being (e.g., David Beckham. HilarySwank), not a comparatively inanimate object (e.g.. Tidelaundry detergent). Prior research has considered celehri-ties, well-known figures, and group entities, such as sportsteams or musical acts, and has documented how they aresimilar to interpersonal relationships in many ways. Forexample, people often experience "seeming face-to-face"relationships with human brands who are "met as if theywere in the circle of one's peers" (Horton and Wohl 1956, p.215). These relationships evince many of the same expecta-tions, cognitions, emotions, and behaviors that operate innormal interpersonal relationships to the point that a con-sumer might view a human brand as a pleasant companion,good friend, or romantic mate (Cole and Leets 1999; Rubinand McHugh 1987). Many of the same variables that pro-vide the basis for interpersonal attachments operate withre.spect to human brands, such as increasing perceptions ofa shared background and elevated social appeal (Cole andLeets 1999; Perse and Rubin 1989).

Relationship differences. Altbougb researcb in market-ing has embraced the idea that consumer relationships maybe analogous to interpersonal bonds (Fournier 1998), thereare differences between the two; however, in general, theseimportant difference are underresearched. For example,Rubin and McHugh (1987) note that unlike interpersonalrelationships, people's relationships with celebrities are lesslikely to be truly interactive (e.g., mutual self-disclosureand interrogation are not possible). However, the effect ofthese and other differences in relationship functioning is notwell understood (Kleine and Baker 2004). Until their

Human Brands/105

impact is addressed, comprehension of how consumers'human-brand attachments form may be advanced by attend-ing to theory development that is dependent on similaritiesbetween these contexts. That is, many of the same causes ofstrong attachments operate when the focal object is a per-son, a brand, or some hybrid of the two (i.e., humanbrands).

What Strengthens Attachments?In general, it is construed that the chief function of attach-ments is to confer emotional security to the attached partyby being responsive to a person's needs (Hazan and Shaver1994). This theory is expanded on by recent work on funda-mental human needs that suggests that if an object isresponsive to a person's needs for autonomy, relatedness,and competence (i.e., A-R-C), intense attachments mayresult (La Guardia et al. 2000; Ryan and Deci 2000).

Defining "autonomy." Autonomy refers to a person'sneed to feel that his or her activities are self-chosen, self-governed, and self-endorsed (Deci and Ryan 2000; Ryanand Deci 2000). Associated with the fulfillment of this needis a person's perception that he or she is free from pressureto behave in certain ways and is able to express bim- or her-self as he or she wishes. A person whose need for autonomyis satiated Is likely to report feelings of "volition, agency,and initiative" (La Guardia et al. 2000, p. 368). Autonomyis similar to what Austin and Vancouver (1996, p. 357) label"self-determination." that is, a person's sense of freedom tomake his or her own choices and to avoid feeling con-strained or coerced.

Defining "relateduess." Relatedness refers to a person'sneed to feel a sense of closeness with others (Deci and Ryan2000). The need for relatedness is a homonomous tendency;it is a desire to belong to a social sphere (Ryan and Deci2000) and to avoid feeling isolated (Austin and Vancouver1996). A person whose need for relatedness is satisfied i.slikely to report feeling "connected with and cared for byanother" (La Guardia et al. 2000, p. 368).

Defining "competence." Competence refers to a per-son's innate, life-span tendency to seek feelings of effec-tiveness, achievement, and challenge in his or her activities(Deci and Ryan 2000). This idea is similar to what previousresearchers have termed "mastery," which refers to a per-son's avoidance of mediocrity and decrements in perfor-mance (Austin and Vancouver 1996). A person whose needfor competence is satiated will report feeling curious andskilled (La Guardia et al. 20(K)).

The identification of these three variables emerges fromrecent studies on human motivation, and "although motiva-tion is often treated as a singular construct, even superficialreflection suggests that people are moved to act by very dif-ferent types of factors" (Ryan and Deci 2000, p. 69). Ratherthan considering motivation as rising from general humanneeds, the current research attempts to parcel out the spe-cific motivations and needs that are active in particularsocial spheres by "asking what kind of motivation is beingexhibited at any given time" (Ryan and Deci 2000, p. 69).As such, this article proposes that the consumer-human-

brand dyad may describe a relationship context in whichthese three needs are potentially important because theirfulfillment in that context may lead to carefully targetedfeelings of attachment.

A-R-C as responsiveness. Autonomy, relatedness. andcompetence are fundamental human needs (Deci and Ryiin2000). They are different from most needs because they areuniversal, innate, und enduring. However, this article is notinterested in the existence of these needs but rather in thenotion that specific partners, such as human brands, mightprovide an individual with relational inputs to satisfy theseneeds; this concept is referred to as "responsiveness."

The idea of responsiveness is foundational to attach-ment research, and self-determination theorists propose thatthe fulfillment of the A-R-C needs qualifies as responsive-ness (e.g., Deci and Ryan 2000). Responsive relational part-ners "are ones who respond in ways that promote a person'sexperienced satisfaction of these basic lA-R-Cl psychologi-cal needs" (La Guardia et al. 2000, p. 368). People gravitatetoward relationships that serve their A-R-C needs; socialexperiences that make people fee! autonomous, related, andcompetent also promote stronger attachments.

HypothesesThere is conceptual and empirical support for the view thata responsive object contributes to the creation of intenseattachments, but much of this literature provides only indi-rect support, has been overlooked by marketing researchers,or has not been applied to the realm of human brands. Con-ceptually. Ryan and Deci (20(X)) argue that the satisfactionof A-R-C needs leads to the "development of secure attach-ments within specific relationships" (p. 326) because "prox-imal supports for basic psychological needs in any relation-ship ... play a crucial role in predicting feelings ofattachment" (p. 262). Although Ryan and Deci focus on aninterpersonal context, they propose that the A-R-C needsmight be fulfilled in other ways, inviting the pt)ssibility thathuman brands qualify as responsive relationship partners.

Furthermore, La Guardia and colleagues (2{K)0) askedparticipants to rate their levels of attachment security withrespect to difierent interpersonal relationships. They alsoassessed the extent to which relationship partners satisfyparticipants' A-R-C needs. The results show that greaterA-R-C feelings positively predict attachment security. How-ever, because the study uses a composite measure of theA-R-C needs, it is not possible to determine ihe distinctcontribution of eacb need in predicting attachment .strength.

Autonomy. Empirical research suggests that there is apositive relationship between felt autonomy and attachmentsecurity in teenagers (Ryan and Lynch 1989). By providingan atmosphere in which autonomy is encouraged, parentsfacilitate the strengthening of attachment bonds. Prior workalso documents a positive correlation between respondents"attachment and the extent to which their relationship part-ners make them feel autonomous in both interpersonal andhuman-brand contexts (Giles and Maltby 2(X)4; Leak andCooney 2001). Likewise, Schultz, Kleine, and Kernan(1989) argue that strong attachments are linked to thetheme of individuation, reflecting a person's desire to feel

106 / Journal of Marketing, July 2006

autonomous and in control of the self. They report thatmany objects to which respondents are strongly attachedshow evidence of this theme. Taken jointly, these resultssuggest that the degree to which a relationship partnerserves the need for autonomy should positively affect theintensity of a person's attachment to that partner.

Hj: The more a person perceives a human brand as fulfillinghis or her autonomy needs, the more intensely attached heor she will be to the human brand.

Relatedness. Similar to Deci and Ryan's (2000) previ-ously discussed arguments, Baumeister and Leary (1995)argue that the fulfillment of the need for relatedness by anobject results in increased positive affect, such as feelingsof attachment and love. In addition, prior literature hasdemonstrated empirically that many consumers who arestrongly attached to particular objects show evidence of"integration," which, in piut, refers to their preference to beconnected and joined with others (Schultz, Kleine, and Ker-nan 1989, p. 360).

H2: The more a person perceives a human brand a.s fulfillinghis or her relatedness, the more intensely attached he orshe will be to the human brand.

Competence. As discussed previously, it is not possibleto distinguish the independent effects of each A-R-C needon attachment strength in all prior empirical work (e.g.. LaGuardia et al. 2000). This is important when consideringthe dearth of empirical work that addresses the independentcontribution of competence. Conceptually, research sug-gests that attachments can develop for reasons tied to self-cultivation and self-development (Kleine and Baker 2004),themes that are at least consistent with the idea of a personfeeling competent as a result of consuming a particularobject. For example, elderly people in nursing homes sur-round themselves with special possessions that serve asreminders of a lost sense of proficiency (Rubinstein andParmelee 1992). Similarly, consumers wbo partake in cer-tain high-risk consumption activities can develop a sense ofboth mastery and emotional bonds in those contexts(Arnould and Price 1993; Ceisi, Rose, and Leigh 1993),though it is unclear whether these bonds qualify as attach-ments because they may not be appropriately target spe-cific. This research has not been tested in the context ofhuman brands and provides only conceptual or indirectempirical support.

Evidently, the effect of the satiation of the need forcompetence on attachment is "a bit less straightforward"(La Guardia et al. 2000, p. 368). Although the satisfactionof both autonomy and relatedness needs is likely to occurprimarily in emotionally significant relationships, a per-son's competence needs may be satiated in contexts that areunrelated to attachment, for example, at leisure, work, orelsewhere (CeIsi, Rose, and Leigh 1993). Thus, if a husbandfails to make his wife feel effective and capable (but doesnot actively make her feel ineffective or incapable), she mayyet remain intensely attached to him because he satisfies herneeds for relatedness and autonomy. Although theoryimplies that the fulfillment of the need for competence mayqualify as responsiveness and thus produce intense attach-

ments, prior empirical research is less conclusive. Thus,there is reason to predict that the linkage between compe-tence and stronger attachments may be either positive ornonnegative.

Hi : The more a person pea'eives a human brand as fulfillinghis or her competence needs, the more intensely aUachedhe or she will be to the human brand.

H31,: The relationship between how a person perceive.s ahuman brand to be fulfilling his or her competence needsand intense attachments is nonnegative.

Study 1Study 1 is designed to test H|-Hi directly. In addition,because prior literature has described how strong attach-ments are associated with satisfied, trusting, and committedrelationships (Feeney and Noller 1996; Simpson 1990) andbecause these variables are also pivotal to marketing rela-tionsbips (Foumier, Dobscha, and Mick 1998), Study Iexamines whether strong attachments positively predict thethree variables. This endeavor extends prior work to a novelcontext, permits the testing of the independence of theseconstructs from attachment strength, and highlights themanagerial implications of consumer brand attachments.

MethodOne hundred sixty-four undergraduate volunteers, solicitedin signs posted around a university campus, were paid $5 inexchange for filling out a survey that took most respondentsapproximately 15 minutes to complete. Young-adult respon-dents may be particularly well suited to studying thedomain of human brands because they constitute a covetedtarget market for human-brand marketers and because it iscommon for them to develop attachments to human brands(Boon and Lomore 2001; Perse and Rubin 1989). Volun-teers received one of two versions of an instrument wbosepreamble read as follows: "Many people tbink of them-selves as being attached to certain celebrities or well-knownfigures (e.g., actors, athletes, singers ...). Think about acelebrity or well-known figure to whom you are very Inot]attached." The two conditions (i.e., "very attached" and "notattached") were designed to create variance in responses onthe dimension of attachment strength. Next, respondentsanswered a series of questions (in order) that mapped ontothe dependent (i.e., attachment strength), independent (e.g.,autonomy), and outcome (e.g., satisfaction) variables.

MeasuresThe measures for the A-R-C needs were taken from thework of La Guardia and colleagues (2000), who developedmetrics appropriate for an interpersonal context (see theAppendix). Using two pretests, these measures wereadapted to suit the current human-brand context, thoughminimal changes were actually required. Respondents wereasked to indicate the extent of their agreement (anchored by1 = "strongly disagree" and 7 = "strongly agree") with aseries of statements about the human brand (denoted as"XYZ"). The items for relatedness (a = .81) were as fol-lows: "XYZ makes me feel cared about" and "I feel a lot of

Human Brands/107

closeness with XYZ." Items for autonomy (a = .80)included *'XYZ makes me feel controlled and pres.sured tobe certain ways" (reversed) and "XYZ makes me feel freeto be who I am." Items for competence (a = .70) included"Generally. XYZ makes me feel very capable and effective"and "XYZ makes me feel inadequate or incompetent"(reversed).

Previous research has argued that the degree to whichpeople demonstrate separation distress is a good indicatorof the strength of their attachment bunds (Berman and Sper-ling 1994). That is, with the real, imagined, or threatenedseparation from the object (e.g., death, loss), a stronglyattached person will experience a negative emotional reac-tion that researchers can use as a measure of the intensity ofthe attachment bond. Prior research has advanced multi-item measures (Hazan and Shaver 1994; Hazan and Zeif-man 1994) that were adapted and extended to the currentcontext. The four items (anchored by 1 = "strongly dis-agree" and 7 = "strongly agree") show good reliability (a =.89): (1) *i feel better if I am not away from or withoutXYZ for long periods of time"; (2) ' i miss XYZ when XYZis not iu-ound"; (3) "If XYZ were permanently gone frommy life. I'd be upset"; and (4) "Losing XYZ forever wouldbe distressing to me."

Finally, the instrument included several metrics togauge some of the potential positive qualities associatedwith people's stronger attachments and to provide furtherevidence that attachment strength is differentiate fromprior constructs. Specifically, respondents were adminis-tered three metrics that were adapted from the perceivedrelation.ship quality scale (Fletcher, Simpson, and Thomas2000). which taps the degree to which respondents assesstheir relationships to he satisfying, trusting, and committed.Although this scale has not been widely used in marketing,it shows excellent internal consistency and validity. Theitems (all anchored by 1 = "not at all" and 7 = "very much")for satisfaction (a = .97) were as follows: "How satisfiedare you with the relationship?" "How content are you withthe relationship?" and "How happy are you with the rela-tionship?" Similarly, items for trust (a = .92) were as fol-lows: "How much can you count on XYZ?" "How much doyou trust XYZ?" and "How dependable is XYZ?" Finally,items for commitment (a = .98) were as follows: "Howdedicated are you to the relationship?" "How committed areyou to the relationship?" and "How devoted are you to therelationship?"

Results

There was an enormous range in the types of human brandsthat respondents listed, including actors, musicians, politi-cians, and media personalities. These figures represent atarget of investment by, for example, movie studios, recordlabels, and political parties in terms of promoting thehuman brands and managing their brand associations. Amanipulation check composed of an altemative measure ofattachment strength (Thomson, Maclnnis. and Park 2005)showed a significant difference between respondents in theweak- (M = 1.86) and those in the strong- (M = 4.05. p <.05) attachment conditions.

Results of a muUivariate analysis of variance. Theresults of a muUivariate analysis of variance provide initialsupport for the idea that A-R-C may be important in theproduction of stronger attachments. The mean of autonomyin the strong-attachment condition (M = 3.66, SE = .18)was higher than that in the weak-attachment condition (M =1.98, SE = .17, p < .05). The mean of relatedness in thestrong-attachment condition (M = 2.94, SE = . 14) was sig-nificantly different from that in the weak-attachment condi-tion (M = 1.30, SE = .\3,p< .05), as was the mean of com-petence in the weak- (M = 2.71. SE = .18) versus thestrong- (M = 3.67, SE = ,19, p < .05) attachment condition.

Correlational analysis. A consideration of the correla-tion between the A-R-C needs and strong attachments mar-shals evidence of a significant linkage. Autonomy (r = .62,p < .05), relatedness (r = .70, p < .05), and competence (r -.44, p < .05) are each significantly correlated with attach-ment strength. In addition, the correlations among theA-R-C needs are positive (r's = .59-.75; see Table I).

FuU structural models. The remaining analysis is con-ducted with structural equation modeling (AMOS). It iscustomary to evaluate structural models in reference toalternative conceptualizations of the data and by comparingfit statistics. As such, the analyses included three models.

Model I is the hypothesized model, with the A-R-Cindicators loading onto their re.spective factors, which areallowed to correlate and, in turn, predict attachmentstrength. Model 2 is the same as Model I, except that thelatent A-R-C needs are modeled as uncorrelated. Finally,Model 3 depicts all the indicators for the A-R-C needs load-ing onto a single latent factor.

In each model, the relationships between the indicatorsand the latent variables are positive and significant. Model 1

TABLE 1Correlation Matrix of Study 1 Measures

1. Autonomy2. Relatedness3. Competence4. Attachment strength5. Satisfaction6. Commitment7. Trust

.75

M

.11

.30

.12*

. ^^—

.58

.72

.48

. «

.46

.55

.69M

m.33

Notes: All ps < .05, except 'p = .14.

108 / Journal of Marketing, July 2006

shows excellent minimum sample discrepancy divided bydegrees of freedom (CMIN/DF = 1.92). comparative fitindex (CFI = .96), and nornied fit index {NFI = .92) statis-tics. Mode! 2. in which the A-R-C needs are uncorrelated.has universally unacceptable fit statistics and therefore isrejected. Specifically, the CMIN/DF statistic is inflated(6.04). and both the CFI (.77) and the NFI (.74) statisticsare inferior. The fit statistics for Model 3 are mixed, with

acceptable CMrN/DF (2.98) and CFI (.91) values but a mar-ginal NFI value (.87). Because of its better fit and because ithas a significantly different chi-square {%- = 92.27, d.f. =48) value from Model 3 (p < .05), Model I is superior (seeFigure 1).

H[ proposes that when a human brand fulfills autonomyneeds, a person is more likely to become strongly attachedto it. The path between autonomy and attachment strength

FIGURE 1Summary of Study 1 Structural Models

A: Model 1

o

Autonomy 1

Autonomy 2

Autonomy 3

Relatedness 1

Reiatedness 2

Competence 1

Competence 2

Competence 3

6 9 ^

83*- >

90*

85'

Separation

Separation

Separation

Separation

1

2

3

4

B: Model 2

O

O

o

Autonomy 1

Autonomy 2

Autonomy 3

1.05*

^.76* ^ ^

ooo

Relatedness 1

Reiatedness 2

Competence 1

Competence 2

Competence 3

Human Brands/109

FIGURE 1Continued

C: Model 3

o

Autonomy 1

Autonomy 2

Autonomy 3

.88*

.85*

.54* "^

O " * Relatedness 1

o

Relatedness 2

Competence 1

Competence 2

Competence 3

*'p< .05.Notes: From Study 1. Model 1 shows all items for the independent variables loading onto their respective latent factors, which predict attach-

ment strength (x^ = 92.26. d.f. = 48; CMIN/DF = 1.92; CFI = .96: and NFI = .92). Model 2 Is identical to Model 1. except that the A-R-Clatents are uncorrelated (x^ = 307.95, d,f. = 51; CMIN/DF = 6.04; CFI = .77; and NFI = .74). Model 3 shows all items for the independentvariables loading onto one latent factor, which predicts attachment strength (x^ = 158.07, d.f. = 53; CMIN/DF = 2.98; CFI = .91; andNFI = .87). All path estimates are standardized and shown above the respective path.

in Model 1 is positive (y = .41, /? < .05), in support of H|.The path between relatedness and attachment strength isalso positive (y = .49, p < .05), in support of H^. Finally, H3ais rejected, but H3(, is supported; there is an insignificantrelationship between competence and attachment strength.This finding allows for the possibility that a human brandmust not negatively affect competence for a strong attach-ment to exist.

Outcome variables. Although increasingly intenseattachments may have positive qualities, these are not dueto a lack of construct orthogonality. An exploratory factoranalysis that included the attachment .strength items and thesatisfaction, trust, and commitment items was conducted(see Table 2). The results show a clean four-factor solution;no inappropriate loadings exceed .3, suggesting that theseconstruct.s are distinguishable.

Furthermore, stronger attachments are characterized byconsumers* perceptions of their relationships with thehuman brand as satisfied, trusting, and committed. The cor-relation matrix (Table I) shows that attachment strength ispositively correlated with satisfaction (r = .25, p < .05),trust (r = .66. p < .05), and commitment (r = .66, p < .05). Inaddition, a multivariate analysis of variance of the differ-ences in satisfaction, commitment, and trust between weak-and strong-attachment conditions suggests that there arebenefits associated with strong attachments. In the strong-versus the weak-attachment conditions, respondents

reported having greater satisfaction (M = 4.46, SE = .18versus M = 3.14, SE = .19; F = 25.17.p < .05). commitment(M = 3.59. SE = .16 versus M = 1.66, SE = .17; F = 71.56,p < .05), and trust (M = 3.68, SE = .16 versus M = L90,SE = .17; F = 60.16, p<.05).

DiscussionThe main contribution of Study I is that the fulfillment ofautonomy and relatedness needs explains attachmentstrength, but the need for competence does not. Althoughtheory and prior research suggests that an object that makesa person feel cotnpetent may bolster attachments (LaGuardia et al. 2000; Rubinstein and Parmelee 1992), theresults suggest that competence does not play a role instrengthening attachments to human brands. Although Table1 reveals that competence is correlated with stronger attach-ments, its contribution is limited when the variance ofautonomy and relatedness is taken into account. However,these same results suggest that the relationship betweencompetence and attachment strength must be nonnegative.That is, an object that suppresses a person's competenceneeds may also diminish the strength of a person's attach-ment. This issue is considered again in the subsequentstudies.

Although the results of Study 1 are encouraging, thechoice of survey design dictates that some of the affectiverichness evident in consumers" attachments to human

110 / Journal of Marketing, July 2006

TABLE 2Study 1 Exploratory Factor Analysis with Reiationsiiip Constructs

Factor

Item SatisfactionAttachment

Strength Trust Commitment

Satisfaction 1Satisfaction 2Satisfaction 3Attachment Strength 1Attachment Strength 2Attachment Strength 3Attachment Strength 4Trust 1Trust 2Trust 3Commitment 1Commitment 2Commitment 3

.99

.95

.94.98.87.75.67

.96

.88

.72.97.93.91

Notes: Principal axis factoring with Promax rotation; ail ioadings >.3 are reported.

brands i.s lost (e.g.. Rubin and McHugh 1987). It is clearfrom ihe literature that respondents use the same words todescribe their attachments to buman brands as tbey do todescribe their strong interpersonal relationsbips (a resultfound by Foumier [1998] in reference to product brands).Study 1 builds primarily on tbe reliability and validity of themeasures, whereas Study 2 employs a qualitative approachto corroborate tbe results of Study I and to Oesh out someof the affective ricbness of consumers" attacbments tobuman brands. In tbe process. Study 2 is intended to gamerevidence that relationsbips cbaracterized by strong attach-ments lie between and are comparable with botb person-to-person (Hazan and Zeifman 1994) and product-brand(Fournier 1998) attachments.

Study 2Study 2 is a qualitative study tbat evaluates evidence of theA-R-C needs in tbe context of human brands to comple-ment the empirical results of Study 1. Tbe use of qualitativeprocedures (Study 2) in conjunction witb quantitativeapproaches (Studies 1 and 3) may be particularly wellsuited to researcb on attachments (Belk 1992) because,together, they ricbly describe tbe focal variable while statis-tically establishing tbe linkages among variables.

Method

Twenty-five undergraduate respondents were recruited on auniversity campus; each was paid $5 in exchange foranswering questions about buman brands in writing. Nolimitations were provided about whicb or wbat type ofhuman brands to discuss: respondents were instructed todiscuss any buman brand ("celebrity or well-known per-son") to whom tbey felt '^attached." In addition to docu-menting the emotional quality of these stronger attach-ments, the aim was to determine wbetber tbe responsivenessof human brands, revealed by tbe A-R-C needs, was evidentin the context of tbese attachments because prior research isnot conclusive on this issue. A series of questions wereposed; for example, respondents were asked to tell the story

of tbeir first experience witb the human brand, to describetbeir best memory involving the buman brand, and to con-vey what the human brand meant to them. Thus, attach-ments were examined from tbe consumer's first-hand per-spective on the basis of relatively unstructured descriptionsof salient experiences with tbe brand (Foumier and Mick1999). Two independent raters, wbo were blind to tbe pur-pose of tbe study, coded respondent feedback for evidenceof strong attachments and the A-R-C needs. Disagreementsbetween coders were resolved in discussions with theauthor.

Results

All responses showed evidence of the types of feelings typi-cally associated with attachments of moderate to intensestrength. In addition. 48% of responses indicate evidence ofautonomy, and 60% show evidence of relatedness. Only oneresponse shows evidence of the respondent experiencingfeelings of competence. Incorporated into tbe subsequentsections are excerpts from tbe responses that tended to beextensive and thoughtful. In each, the respondent self-selected a human brand as the object of feelings of attacb-ment. For example, one woman described a musical act intbe following terms:

I can't really remember when I siarted loving them. Atleast 15 years ago. I think probably I fell in love wilh thesong "Rose Coloured Glasses." and then they played arink in a town close to me and a friend and I went to seethem. It was a great place to see them play because youcould dance and get really close to the stage. They had allthese table and floor lamps on the staye. A really intimatesetting, and there weren't that many people there. I thinkwe paid $7 a ticket. They absolutely rocked. I've beenaddicted to them since then.

In this excerpt, the respondent describes a "loving" rela-tionship that bas endured for years. This pattem was com-mon; many respondents referred to powerful emotionswhile describing facets of tbe relationships that were inter-preted as evidence of specific A-R-C needs. For example, arespondent wrote the following about a band:

Human Brands/111

I bought all of their albums without hearing them (1 hadonly had the live one previous) and became obse.s.sed withiheir talent. 1 went to see them for the first time ever andwas so incredibly moved, I often found myself starringIsic] in awe at them during the show. 1 was even moved totears on a few songs. After. I met Jim [lead singer| and gola picture taken with him, and I jusi couldn't believe athow nice he was, I had become from that experience, asother fans like to say, a Rodeohead. This was less than ayear ago.

It is evident that this person's response to the band isaffect laden. Furthermore, she now considers herself a"Rotleohead." a group of enthusiastic followers of the band(Blue Rodeoj. Although this group appears to be fairlyunstructured, many of its members gather in online forums(e.g., www.discussions.bluerodeo.com) to share .stories,news, and opinions about the band. This excerpt demon-strates thai a Rodeohead (and this respondent) enjoys socialrewards that are linked to the band and that are interpretedas serving a need for relatedness. The following providesadditional evidence of feelings of relatedness:

Ah, now this is a better story. Another live show. This wasmy 3rd of 3 concerts;... it was in April of 2000 I think (1haven't gotten to a show .since then—starving student andall) at the Grand Theatre.... It was really a fajita.stic night,my dad and I went together. He's a bit of a fan himself,though not quite as die-hard as myself. It's something thathas sort of brought us together. We didn't have the great-est of seats because we bought our tickets late, but itdidn't matter too much because the venue was fairlysmall, it was really a pretty magical evening. I foundmyself getting lost in the music now and then, and I justcouldn't keep my eyes from glossing over.... Anyway,I've rarely felt so emotionally connected to anything.

Here, the respondent's relationship with the band isexplained by the notion that it brings her closer to herfather, a fairly clear representation of relatedness, evenwhile her feelings of attachment are also directed at theband itself. A similar result with respect to feelings of relat-edness is evident from the following excerpt in reference toa sports team, the Detroit Pistons;

I used to attend games with my family as a child. 1 grewup watching Joe Dumars and Grant Hill all throughouthigh school. Basketball was a family event—a sanctu-ary—and I did my first internship with the company. Thefirst game we went to was against the Portland Trailblaz-ers—I was about 11 years old. Dennis Rodman grabbed18 rebounds, and we cheered in the mid tier seats as theybeat the Blazers by about 10 points. We had so much funthat game that going to the contests became a regular fam-ily event.

Other excerpts described people's adopting the manner-isms of various human brands, such as a male respondentwho wrote, "I'm even a fan in their club. I own all of theirCDs plus bootlegs. I even write and sign my name to reflectmy love for them." This excerpt reveals evidence of related-ness ("fiin in their club") and the type of self-endorsedbehavior lying at the core of autonomy, reflected in thedecision to model his signature after the style of print asso-ciated with the band's (Metallica) album covers.

Other passages showed similar modeling behavior thatsignifies a freedom to express the self. For example, a male

respondent wrote about a band, "I strive to be like thembecause they haven't sold out and they are nice people whotreat others well," and another respondent indicated that aparticular performer (DMX) "inspires me to be a true manwho will love my family and give back to my community.He isn't just a rapper. He's a legend and a great role model.He's one crazy hard-knocks guy, but he's also normal cuz[sici he has kids and a wife and I respect that." Perhaps noexample better captures the intersection of self-expression,self-endorsed behavior, and the feeling of freedom associ-ated with autonomy as the follov /ing excerpt provided by amale respondent in relation to a musical band:

As a musician myself (guitar, banjo, accordion, bass) Ienjoy their folky-jazzy-rock vibe mixed with lyrics thathit home to me. Their songs are intelligent, well-crafted,and presented with impeccable taste. I think the fact thatthe band is accessible before and after many shows addsto their appeal. They sound like rock stars on stage—butact like normal guys off-stage. They provide songs thatspeak to me. They provide a middle-agreed tsic] familyguy with the responsibilities of a family and mortgage andstressful career, an outlet to cui loose. They also inspiredme to pick up my guitar and write songs again. My wifeadores Jim Cuddy. Back in the mid-80's/early 9O's theband was unique—now there is a whole industry(alt.country) dedicated to this blend of music. I amunique—as 1 follow a band that most everyone 1 meet ha.snever heard of.

Finally, many respondents discussed objects they per-ceived as possessing a high degree of skill. For example,one respondent wrote about tennis player Martina Hingi.shaving "a pretty solid game." Another respondent discussedTori Amos as having "a spectacular singing voice, plays thepiano masterfully, and is an amazing lyricist." One thoughtthat Matthew McConaughey was "really good. We're talk-ing Oscar here," whereas another described the WorldWrestling Federation as follows: "The acting is stiperb andnot just movie actors. These athlete/actors have to get eachdetail right because they are in front of a life [sic] audience5-6 days a week." However, in all but one case, respondentsattributed skill to the human brand. With one exception,none of the human brands appeared to make the respon-dents themselves feel competent. The single exception wasa respondent who wrote the following about a band:

I like to emulate what they have raised me on (so I feel).They were a positive role model for me in my teen years,providing a degree of sanity and a beacon of relief in aturbulent world I dealt witb during thai lime. I admire allthese traits in them and try to make them a part of my lifein my own, different ways. Even when I meet a goal. Ithen set out for a new one to challenge me.

Whereas this excerpt shows evidence of self-expression("make them part of my life"), the reference to being chal-lenged is interpreted as fulfilling his need for competence.In contrast to the needs for autonomy and relatedness, therewas only limited evidence for respondents' feeling attachedto a human brand because of feelings of competence.

DiscussionThe juxtaposition of feelings indicative of the A-R-C needsand feelings of attachment provides correlational support

1121 Journal of Marketing, July 2006

for the first and second hypotheses. The relative absence offeelings of competence experienced by the respondentsdoes not contradict tbe results of Study 1. Both suggest thatcompetence may be present but does not explain strongattachments in the context of human brands. However, thereis also another possible explanation for (his result, namely,that the nature of the human brands selected by respondents(primarily musical perfonners, athletes, and actors) or theage (18-25 years) of most respondents ruled out discussionof competence, which, in relation to other human brands(e.g., politicians, religious figures), migbt actually be linkedto stronger attacbments. However, it is clear that youngeradults develop human-brand attachments (Boon andLomore 2001; Cole and Leets 1999), suggesting that thechoice of sample was appropriate (this issue is reexaminedin Study 3). The instructions lo respondents allowed themto discuss any human brand tbey desired. The predomi-nance of actors, musicians, and sports figures in the samplecorresponds to prior research that suggest that these are themost common categories of buman brands thai people typi-cally describe (North, Bland, and Ellis 2005). Furtbermore,respondents were clearly familiar witb and actively contem-plating tbe idea of skill and competence (e.g., MartinaHingis, Tori Amos), except tbat the attachment object wasalmost universally the target of sucb attributions. Study 3further addresses tbe role of competence.

Study 2 presents evidence that respondents use tbe samewords (e.g., love, adoration, emotional connection) tode.scribe their attachments to human brands as tbey use inthe context of otber attachments (e.g., romantic). Thus, itwould be premature to consider consumers" buman-brandattachments trivial or unimportant; the meaningfulness andintense emotions ascribed to these relationsbips argueagainst this conclusion. Furthermore, although much of theliterature cited in laying out the theoretical model originatesin the interpersonal domain, the results suggest tbat thehuman-brand context is sufficiently similar to apply tbeA-R-C framework constructively.

In conjunction with Study 1, the results of Study 2 areencouraging, but additional support for tbe hypotheses andfor tbe role of competence can be obtained by assessingattachment strengtb in a more natural environment. This ispursued in Study 3. In addition, consistent with the effortsof Study 2 to describe the phenomena more fully, Study 3uses expanded measures of the A-R-C needs to ensure thatthe domain of these constructs is adequately captured.

Study 3The hypotheses tested in Study 1 pertain to tbe role of theA-R-C needs in reference to human brands. Each is speci-fied by self-determination theorists as potentially importantto strengthening attachments. However, to address severalalternative hypotheses, tbe analysis is expanded beyond tbeA-R-C needs. Specifically, there are other important humanneeds that are at the same high-order level of analysis as theA-R-C needs. For example, a recent publication thatreviews and consolidates prior research on important humanneeds (see Sheldon et al. 2001) advances measures of vari-

ous needs discussed in the overlapping research of scholarssucb as Maslow, Epstein, and Atkinson; these needs arereferred to as (1) "self-actualization," (2) "pleasure-stimulation," (3) "money-luxury," (4) "self-esteem," and (5)"popularity-intluence." This background is important in tbecurrent context because the extant literature seems toadvance indirectly the more expansive view tbat servicingneeds unrelated to tbe A-R-C needs may also bolster attach-ments, Although the articles cited subsequently support orare consistent with the view that tbe A-R-C needs areimportant, they also propose that tbrougb the course of anobject's serving additional needs, a consumer may developfeelings that are emblematic of stronger attacbments.

For example, Keller (2001) indicates tbat when a con-sumer gains feelings of fun, excitement, and self-respectfrom a brand, he or she may be more likely to becomeattached it. These feelings seem to mirror the needs forpleasure-stimulation and self-esteem (Sheldon et al. 2001).Therefore, a positive relationship between fulfillment ofthese two needs and increasingly strong attacbments tohuman brands might be expected.

Similarly, Escalas and Bettman (2005) propose tbat aconsumer wili form self-brand connections when the brandhelps him or her achieve goals, sucb as providing self-esteem, piulicularly wben tbe brand is symbolic and when aperson's in-group has already embraced tbe brand. This rea-soning indicates that a person's needs for self-esteem,money-luxury (i.e., the brand's symbolic stature), andpopularity-influence (i.e., tbe brand as a tool for a person tomaintain position within a group) may be active in tbeprocess of attachment formation, in turn, this suggests thatas an object services a person's needs for self-esteem,money-luxury, and popularity-influence, attachmentsbecome more intense.

Finally, Pimentel and Reynolds (2004) describe con-sumers who become more affectively committed to sportsteams after undergoing a transcendent process of "sacraliza-tion," in which the team takes on meaningful status in theconsumer's mind. Tbis process seems consistent with Shel-don and colleagues* (2001) description of self-actualization,which refers to a person's identification of importantaspects of life's meaning and purpose. Accordingly, it migbtbe expected tbat the relationship between attacbmentstrengtb and self-actualization, as well as the other basicneeds, will be positive and significant.

H4: The more a person perceives a human brand as fulfillinghis or her needs for (a) self-actualization, (b) pleasure-stimulation, (c) money-luxury, (d) self-esteem, and (e)popularity-intluence, the more he or she will be intenselyattached to that human brand.

Tbis hypothesis does not undermine the role of theA-R-C needs in strengtbening attachments. It is possibletbat tbese additional needs are complementary. Study 3retests the hypotheses with a nonstudent sample in a morenaturalistic setting and considers these five alternative needsin an effort to address tbe rival hypothesis that the satisfac-tion of any important need explains stronger attacbments.

Human Brands/113

MethodAdult respondents were recruited by posting a letter ofsolicitation on Web sites organized as online communitiesfor or by fans of various celebrities. These sites provideforums in which people gather electronically to share theirthoughts and feelings about a celebrity with like-mindedpeople. These sites were targeted because they are accessi-ble and populated by people who are communicative andfamiliar with (though not necessarily attached to) at leastone celebrity. Along other dimensions, such as age. geo-graphic location, and gender, there is enormous heterogene-ity in membership within and among sites.

A soliciting message was posted at each forum, whichdirected volunteers to send an e-mail to a university-affiliated address thai was set up to for this study. To pro-vide an incentive, the soliciting letter informed people thatrespondents who completed the survey would be enteredinto a random drawing for one of five prizes of $50 andwould be eligible to receive a summary of the research.Those who sent an e-mail were directed to the location (i.e.,a URL) of the survey, which was constructed for the Webusing the nTreePoinl software package. The study descrip-tion, instructions, and order of presentation for the mea-sures were nearly identical to those of Study 1, and respon-dents were randomly assigned to list a human brand towhom they either were or were not strongly attached.

Measures

The metric used to assess attachment strength was com-posed of the same four items used in Study 1 and showedgood reliability (a = .93; see the Appendix). The A-R-Cneeds were each measured with the items employed inStudy I; additional items were adapted from the work of LaGuardia and colleagues {2000) and Sheldon and colleagues(2001). Compared with Study I, these showed improvedinternal consistency (as = .83-.89).

The preamble for each metric that assessed needs read,"When I see. hear, think about or interact with thiscelebrity,..." The need for relatedness was measured withthe two items from Study I plus four additional items: "Ifeel loved and cared about," "I often feel a lot of distancefrom this celebrity." "I feel like I am in contact with some-one who cares for me and who I care for," and "'I feel closeand connected to a person who is important to me." Theneed for autonomy was measured with three items in addi-tion to those used in Study I: "I feel that my choices arebased on my true interests and values," "'I feel free to dothings my own way," and "I feel that my choices expressmy 'true' self," The need for competence included threeadditional items: "I feel that I can successfully completedifficult tasks anti projects." "1 feel that 1 can take on andmaster hard challenges," and "I feel very capable in what Ido."

Measures of other needs were included to assess therival scenario that, in general, need fulfillment leads tostrong attachments; these measures were adapted from thework of Sheldon and colleagues (2001). Respondents wereasked in relation to each need the extent of their disagree-ment ("strongly disagree") or agreement ("'strongly agree")

with certain statements. Items for self-actualization (a =.95) were as follows: "I feel like 1 am 'becoming who Ireally am,"' "I feel a sense of deeper purpose in my life,"and "I feel a deeper understanding of myself and my placein the universe." Pleasure-stimulation (a = .88) was mea-sured with the following item.s: "I feel I am experiencingnew sensations and activities," "I feel intense physical plea-sure and enjoyment." and "1 fee! 1 have found new sourcesand types of stimulation for myself." The three items tap-ping money-luxury (a = .91) were as follows: "I feel I amable to buy most of the things I want,"' "1 feel that I havenice things and possessions," and "I feel that I have plentyof money." The measure of self-esteem (a = .84) was com-posed of the following: "I feel I have many positive quali-ties," '1 feel quite satisfied with who I am," and "I feel Ihave a strong sense of self-respect." Finally, people's needfor popularity-influence (a = .85) was assessed by the fol-lowing: "1 feel that I am a person whose advice others seekout and follow." "I feel that I strongly influence otherpeople's beliefs and behavior," and "I feel that I have astrong impact on what other people do."

Results

There were 107 respondents (26 males and 81 females), andthey ranged in age from 18 to 60 years, with an average ageof 33 years. Although gender and age were initially consid-ered covariates in the subsequent analysis, neither predictedattachment strength and thus were dropped from the analy-sis. Although the instrument did not explicitly collect dataon nationality, it was apparent from respondents' e-mailaddresses and signatures that the sample was geographi-cally diverse. For example, respondents appeared to belocated in Australia, Canada. Hong Kong, New Zealand,South Africa, Spain, Sweden, Uruguay, and the UnitedStates.

There was an enormous range in the types of humanbrands selected by respondents, including actors, comedi-ans, models, directors, radio personalities, writers, singers,athletes, musicians, politicians, and royalty. For example,respondents listed Ben Affleck, King Albert II. LucindaWilliams. Mario Lemieux, Viggo Mortensen. MichaelBiehn. and Joan Crawford. Respondents also varied signifi-cantly in their popularity, activity level (e.g., retirement sta-tus), and profession. . ^

Correlation. Each of the three A-R-C needs was posi-tively correlated with the measure of attachment strength(see Table 3). The correlations between attachment andautonomy (r = .37, /; < .05), relatedness (r = .31. p < .05),and competence (r = .20, p < .05) were all positive and sig-nificant. In generai, the correlation between attachmentstrength and the other needs was weak or nonsignificant:self-actualization (r = .16, p = not signiflcant |n.s.]),pleasure-stimulation (r = .19, p < .05). money-luxury (r =.15,/j = n.s.), self-esteem (r= .09,p = n.a.), and popularity-influence (r = .19,/^ < .05).

Hierarchical regression. Using hierarchical regressionallows variables entered in prior stages to account for vari-ance in the dependent variable (i.e., attachment strength)

114 / Journal of Marketing, July 2006

TABLE 3Correlation Matrix of Study 3 Measures

Variable

1. Attachment strength2. Autonomy3. Relatedness4. Competence5. Self-actuali2ation6. Pleasure-stimulation7. Money-luxury8. Self-esteem9. Popularity-infiuence

.38 3137

.20

.60

.52

.16"

.47

.67

.56—

.19

.20

.52

.35

.76—

.15'

.35

.45

.46

.56

.33—

.09'

.33

.42

.69

.48

.38

.50

.19

.33

.36

.45

.37

.27

.44

.66

Notes: All ps < .05, except 'p = n.s.

befot^ the independent contribution of variables entered insubsequent steps is assessed. Tbus, it provides a relativelyconservative test of tbe contributions of the variablesentered last.

Considering the weak relationships between attachmentstrength and the non-A-R-C needs that were revealed bythe previous correlation analysis, it might be expected thatthere is little or no relationship between an object's fulfill-ing tbese alternative needs and attachment strength. A con-sideration of tbe hierarchical regression with tbe alternative(i.e.. non-A-R-C) needs entered in tbe first step and tbeA-R-C needs entered second shows tbat this is true.

In Regression 1 (see Table 4), in which the A-R-C needsare not entered, none of the alternative needs significantlypredicts attachment strength. Tbe coefficients for self-actualization (P = .01, p = n.s.), pleasure-stimulation (p =.20, p = n.s.), money-luxury (P = .13. /? = n.s.), self-esteem(P = -.06, p = n.s.), and popularity-influence (p = .10, p =n.s.) are uniformly insignificant. In light of (I) the weak orinsignificant correlations between attachment and the inde-pendent alternative needs and (2) the weak to moderate cor-relations among the alternative needs, it is unlikely that thisresult is due to multicollinearity.

Likewise, in Regression 2, wbicb includes the A-R-Cneeds in tbe second step, none of tbe alternative needs is asignificant predictor of attachment strength. Combined with

the results from Regression I, this demonstrates a lack ofsupport for H4. However, tbe links between two of the tbreeA-R-C needs and the dependent measure are significant.Specifically, autonomy (P = .31, p < .05) and relatedness(p = .29, p < .05) are significant predictors, but competence(P = -.03, p = n.s.) is not. Tbus, tbe results of Study 3 sup-port H|, Hi, and H^^,.

DiscussionStudy 3 mirrors Study 1, insofar as botb reveal tbat the sat-isfaction of autonomy and relatedness predicts strongattachments. In interacting with buman brands, people wboare made to feel capable of self-expression and self-governance and those wbo are made to feel a sense of close-ness and intimacy are likely to develop strong attachments.In Studies 1 and 3 and in support of H-^^,, the relationshipbetween competence and attachment strength was not sig-nificant, suggesting that at least in a celebrity context, feet-ings of efficacy are not fundamental to promoting strongerattachments. Furtbennore, competence and five additionalneeds were considered but were found to he insignificant inStudy 3; tbis suggests tbat general need fulfillment does notexplain tbe creation of strong attachments. Although otherpredictors of attachment strength may be identified in fur-ther research, tbe results suggest tbat autonomy and related-ness play an important role.

TABLE 4Study 3 Hierarchicai Regression Results

HierarchicalRegression

Step 1

Step 2

Predictor

(Constant)Self-actualization

Pleasure-stimulationMoney-luxurySelf-esteem

Popularity-influence

AutonomyRelatednessCompetence

Regression 1

StandardizedBeta

.01

.20M

-.06.10

;———

p Value

.00

.94

.24

.43

.72

.52

———

Regression

StandardizedBeta

-J31.26.14

-.13.10 • -

.31

.29-.03

2

p Value

.01

.15

.12

.37

.50

.52

.03

.05

.89

Notes: Dependent variable = attachment strength.

Human Brands/115

General DiscussionStudy I revealed that consumer attachments to humanbrands resuU in elevated levels of satisfaction, trust, andcommitment. Study 2 revealed the affective richness of theexperience for consumers attached to human hrands. Study3 offered additional support for the theoretical model usinga separate sample. On the basis of this and other research(e.g.. Ambler et al. 2002; Thomson Maclnnis, and Park2005), it is reasonable to conclude that attachment strengthis separable from trust, satisfaction, commitment, and avariety of other constructs, such as favorable attitudes andbrand loyalty, that have been linked to strong relationships.The promise of attachment strength to marketers is that theconstnict is predictive of these and other relationship fea-tures (e.g.. Abluwalia, Unnava, and Bumkrant 2001; Liljan-der and Strandvik 1995). is theoretically rigorous and wellarticulated, and is achievable through two means (i.e., creat-ing feelings of autonomy and/or relatedness). Few, if any,relationship constructs appear to offer the same comprehen-sive benefits as attachment strength, which may provide auseful means of conceptualizing and an efficient metric forassessing the strength of consumer-brand relationships.

Manageriai implicationsIf consumers can be made to feel autonomous and related,organizations may be able to foster strong attachments andthus enjoy the benefits of superior relationships with con-sumers. In general, human brands that make consumers feelappreciated, empowered, and understood succeed in creat-ing feelings of autonomy. Similarly, feelings of relatednessare likely to be produced when a human brand promotesacceptance, openness, and belonging.

General conditions. For an attachment to form, the con-sumer must initially be minimally attracted in some way tothe human brand, whether socially or in some other manner(Boon and Lomore 2001). In other words, an attachment isunlikely to develop if the starting point of the relationship ischaracterized by intense negative feelings or thoughts. Thissuggests that managers must introduce a human brand tothe world carefully and deliberately and choose a position-ing that is appealing and sustainable over time.

Need for interaction. The importance of human brandsinteracting with consumers was underscored by the qualita-tive results from Study 2, in which one respondent spoke ofhaving his picture taken with the human brand. Repeatedinteraction between a consumer and the human brandreduces uncertainty and provides the basis for an attach-ment to grow (Berman and Sperling 1994), but in the con-text of human brands, it is not just the quantity but also thequality of interaction that matters.

With direct interaction, consumers are more likely toview the human brand as accessible, increasing the opportu-nity for feelings of autonomy and relatedness to be pro-moted. Although "being seen" (e.g., public appearances)may be acceptable as a form of indirect interaction, moredirect means offer a richer, more rewarding choice. Firmsshould focus on determining how to make their humanbrands approachable. For example, technological alterna-

tives for direct interaction show potential not only becausethey expand the human brand's reach and exposure. If thehuman brand had a meaningful presence online through theuse of blogs or chat rooms, attached consumers wouldgather to ask questions or to learn more about the humanbrand in his or her own words. An example of this directapproach is Mark Cuban, owner of the Dallas Mavericksbasketball team, whose immense popularity is linked to hisspending several hours daily answering e-mails from fans(Fielding 2005). The direct approach not only improves thequality of interaction but also may encourage feelings ofautonomy and relatedness by reducing hierarchy, makingconsumers feel appreciated and understood, and reducingthe prominence of economic motives by the human brand(Silvera and Austad 2004),

Interaction should also occur on a consistent basis(Beech, Chadwick, and Tapp 2000). Consumers mayrespond better to a human brand who has a routine that con-sumers may view as a guarantee of interaction. For exam-ple, Ellen DeGeneres is known for a daily segment on hershow in which she invites the audience (both in the studioand watching on television) to dance with her. This is anespecially astute move because of its regularity and becauseit simultaneously promotes feelings of autonomy (whatDeGeneres calls "an expression of freedom") and closeness(dancing "with" a celebrity; see Elber 2005). Ensuringregular interaction between a consumer and a human brandhelps create the conditions under which an attachment islikely to grow.

Role of authenticity. Human brands that are perceived asauthentic are more readily embraced (Cole and Leets 1999;Perse and Rubin 1989). Research has elaborated on authen-ticity (Rose and Wood 2005), but what the term means inreference to a human brand remains an underresearchedtopic. However, several ideas are promising (see Tolson2001). For example, authenticity is best developed slowlyand deliberately, thus signaling a long-term view, and byavoiding perceived opportunism by the human brand, whichcan signal that Ihe human brand has "sold out." Likewise,when human brands are seen publicly in a nonpertbrmingrole, they should not be "perceived as acting" (Tolson 2(X)1,p. 445), because this may lead consumers to feel manipu-lated and to view the human brand as lacking credibility. InStudy 2, one consumer spoke appreciatively of the band towhich he was attached acting "like normal guys off-stage."Away from their trade, human brands should not be viewedas trying too hard to convey an image, or they will riskbeing viewed as "pretenders."

The role of competence. In all three studies, competencewas not effective in bolstering attachments in the context ofhuman brands. If the popular press is any indication, con-sumers can form attachments to humiin brands who lackany appreciable skills (e.g., Paris Hilton). However, humanbrands who make consumers feel incompetent are likely toweaken attacbments or prevent them from forming in thefirst place. For example, when comedian Dennis Millerjoined the Monday Night Football broadcast team, his addi-tion was intended to halt a slide in ratings. However, on theshow. Miller used indecipherable analogies, complex dic-tion, and obscure references that alienated his audience and

116 / Journal of Marketmg, July 2006

made them feel inadequate and unintelligent. In otherwords, many viewers were made to feel incompetent; thisnot only caused a weakening of any possible preexistingattachments but also prevented any new attachments fromforming. Although consumers do not become attached to ahuman brand through feelings of competence, the resultssuggest that managers must nonetheless be careful not topermit the degradation of competence to be linked to theirhuman brand.

However, caution must be taken in extending theseresults to other domains in which attachment and self-determination theories may provide a useful road map tobuilding attachment bonds. For example, with respect toproduct functional brands that are intended to solveconsumption-related problems (Park. Jaworski, and Mac-lnnis 1986), consumers may experience feelings of compe-tence as a result of brand use. Similarly, consumer relation-ships that implicate particular service providers, such ascoaches or psychologists, may be explicitly designed todevelop particular skills, and it is common for strongattachments to develop in such contexts (Colin 1996). How-ever, only further empirical study of the issue can ascertainwhether attachments in these contexts become strongbecause of the servicing of competence needs or whethercompetence is merely a correlate without any meaningfuleffect when the variance of autonomy and relatedness aretaken into account.

Potential as endorsers. Marketing research that has con-sidered human brands has tended primarily to address thechoosing and measuring of the effectiveness of celebrityendorsers to enhance the appeal of a brand (Agrawal andKamakura 1995). Although the current article does notdirectly address the use of human brands as endorsers, priorresearch and this article imply that human brands to whomconsumers are attached may be unusually effective. Oneway to assess the appeal of a particular celebrity is the Qratio, a commonly used but sometimes "misleading" mea-sure of popularity indicated by the celebrity's familiarityand likeability (Knott and St. .rames 2002, p. 1428). Com-pared with the likeability dimension, an attachment is moresignificant and profound to an individual (Thomson,Maclnnis, and Park 2005). Although consumers may likemany celebrities, they will be strongly attached to only afew and be willing to sacrifice for and invest in these humanbrands to a greater extent.

In a related vein, research suggests that consumers whohave positive feelings toward a particular brand are morelikely to embrace a brand extension because of a more auto-matic and direct transfer of positive affect to the affiliatedproduct (Yeung and Wyer 2005). This result occurs evenwhen the extended brand has comparatively poor fit withthe core brand, which implies that attachment may facilitatea highly effective transfer of feelings from the human brandto the sponsored object. In the context of celebrity

endorsers, this suggests that there does not necessarily needto be a high degree of fit between an endorser and a particu-lar brand if there is a strongly attached target market. A firmemploying a human brand to whom consumers are stronglyattached as an endorser may anticipate greater returns andinsulation from transgressions (e.g.. celebrity involvementin scandals; see Knott and St. James 2002) than a firm usinga human brand who is merely well liked.

Limitations and Further ResearchA result of these studies that has not yet been addressed isthe overall weakness of attachments to human brands. InStudy 2, respondents used language tbat suggested that theirattachments were strong and emotionally significant. Tbisresult mirrors prior qualitative work on different types ofconsumer-brand relationships (Foumier 1998; Schultz,Kleine, and Kernan 1989). However, respondents in Studyl's strong-attachment condition reported attachment scoresthat, on average, hovered around the scale midpoint. Theattachment scores in Study 3 were stronger than those inStudy 1. Both studies used the same measure of attachmentstrength, but respondents in Study 3 may have been evan-gelical about the human brand they selected. After all.Study 3 respondents were recruited from online communi-ties revolving around human brands. Although many of theparticipants in the low-attachment condition reported weak-attachment scores, (e.g., Orlando Bloom = 1.8/7, MattCrane = 1.0/7), most respondents in the high-attachmentcondition reported scores greater than 6.5/7 (e.g.. BenAffleck. Julie Andrews, Bobby Valentine). These humanbrands seemed to play a more central role in the Study 3respondents' lives than might otherwise be normallyexpected. Furthermore, although the same language mightbe used to describe interpersonal and consumption attach-ments (e.g.. Study 2), relationships implicating the lattermay be weaker in general. Human-brand attachments areprobably best viewed as supplementary and subsidiary(Colin 1996). The implication for qualitative research onmarketing relationships is that quantitative methods providea useful perspective that may not be evident from qualita-tive approaches alone (Belk 1992).

Systematic research into what tactical choices are likelyto promote autonomy and relatedness is needed. For exam-ple, do self-service models of service delivery or productcustomization promote autonomy by making consumersfeel more self-expressive or self-governing? Do autonomyand relatedness need to be created from scratch, or canfirms draw on existing feelings targeted elsewhere? Doespossessing one or more strong consumption-related attach-ments mean that a person experiences elevated life satisfac-tion, or as alluded to by prior research (Kleine and Baker2004), are there other, negative effects? Given the power ofthe strong attachments, answering these questions would bea worthwhile undertaking.

Human Brands/117

APPENDIXSummary of Measures in Studies 1 and 3

Measure

Attachment strengthAutonomyRelatednessCompetenceSelf-actualizationPleasure-stimulationMoney-luxurySelf-esteemPopularity-influence

Items

4323

Study 1

a

.89

.80

.81

.70

Notes: All measures use seven-point Likert scales anchored by

M

2.33

%m1.903.10

"strongly disagree" and

items

4%6633333

"strongly agree."

Study 3

a

.93M.89.83.95.88.91.84.85

M

5.725.684.235.194.424.493,465.184.45

REFERENCESAgrawai. Jagdish and Wagner A. Kamakura (1995). "The Eco-

nomic Worth of Celebrity Endorsers: An Event Study Analy-sis." Journal of Marketing, 59 (July), 56-63.

AhluwaJia, Rohini, H. Rao Unnava, and Robert E. Bumkrant(2001), "The Moderating Role of Commitment on theSpillover EtTect of Marketing Communications." Journal ofMarketing Research. 38 (November), 458-70.

Ambler. Tim. C.B. Bhattacharya, Julie Edell. Kevin Lane Keller.Katherine N. Lemon, and Vikas Mittal {2(X)2). "Relating Brandand Customer Perspectives on Marketing Management," Jour-nal ofSen'ice Research, 5 (I) . 13-25.

Amoutd. Eric J. and Linda L. Price (1993). "River Magic: Extra-ordinary Experience and the Extended Service Encounter,"Journal of Consumer Research. 20 (June). 24-45.

Austin, James T and Jeffrey B. Vancouver (1996). "Goal Con-structs in Psychology: Structure. Process and Content," Psv-chologkul Bulletin. 120 (3), 338-75.

Baumeister. Roy F, and Mark R. Leary (1995) "The Need toBelong: Desire for Interpersonal Attachments as a Fundamen-tal Human Motivation," Psychological Bulletin. 117 (3).497-529.

Beech. John. Simon Chadwick. and Alan Tapp (2000), "Towards aSchema for Football Clubs Seeking an Effective Presence onthe Internet." European Journal for Sport Management. (Spe-ciai Issue). 30-50.

Befk. Russell W. (1992), "Attachment to Pos.sessions." in PlaceAttachmeni: Human Behavior and Environment, Setha M. Lowand Irwin Altman, eds. New York: Plenum Press, 37-62.

Berman. William H. and Michael B. Sperling (1994). "The Struc-ture and Function of Adult Atiachmenl," in Attachment inAdults: Clinical and Developmental Perspectives, Michael B.Sperling and William H. Berman, eds. New York: GuilfordPress, 3-28.

Boon. Susan D. and Christine D. Lomore (2(X)I). "Admirer-Celebrity Relaiionships Among Young Adults: E.nplaining Per-ceptions of Celebrity Influence on Identity." Human Communi-cation Research, 27 (July), 432-65.

Carroll. Barbara A. and Aaron C. Ahuvia (2006). "SomeAntecedents and Outcomes of Brand Love," Marketing Letters,forthcoming.

Ceisi, Richard L., Randall L. Rose, and Thomas W. Leigh (1993)"An Exploration of High-Risk Leisure Consumplion ThroughSkydiving," Journal of Con.sumer Research, 20 (June) 1-23.

Cole, Tim and Laura Leets (1999). "Altachmeni Styles and Inti-mate Television Viewing: Insecurely Forming Relationships ina Parasocial Way," Journal of Social and Personal Relation-ships, 16(4), 495-511.

Colin. Virginia L. (1996), Human Attachment. Philadelphia: Tem-ple University Press.

Deci, Edward L. and Richard M. Ryan (2000), 'The 'What' and'Why' of Goal Pursuits: Human Needs and the Self-Deterniination of Behavior." Psychological htquiry, 11 (4),227-68.

Dietz. Park E.. Daryl B. Matthews, Cindy Van Duyne. DanielAllen Martell, Charles D.H. Parry. Tracy Stewan. Janet War-ren, and J. Douglas Crowder (1991), "Threatening and Other-wise Inappropriate Letters to Hollywood Celebrities." Journalof Forensic Sciences, 36 (1). 185-209.

Edsall. Thoma.s B. and James V. Grimaldi (2004), "On Nov. 2,GOP Got More Bang for Its Billion, Analysis Shows." Wash-ington Post, (December 30), AOL

Eiber, Lynn (2005). "Ellen DeGeneres, Returning to Daytime TV,Still Revels in Freedom of Being Herself," Sun Sentinel, (Sep-tember 6). (accessed December 12. 2005). [available atwww.sun-sent inel .com ].

Escalas, Jennifer Edson and James R, Bettman (2005). "Self-Construal, Reference Groups, and Brand Meaning." Journal ofConsumer Research. 32 (3), 378-90.

Feeney. Judith A. and Patricia Noller (1996). Adult Attachment.London: Sage Publications.

Fielding. Michael (2005). "Mark, His Words: Marketing's Maver-ick on Blogs, Brands and Why Sales Is King." MarketingNews, (November 1), 13-14.

Fletcher. Garth J.O., Jeffrey A. Simpson, and Geoff Thoma.s(2000), "The Measurement of Perceived Relationship QualityComponents: A Confirmatory Factor Analytic Approach," Per-sonality- and Social Psychology Bulletin, 26 (3). 340-54.

Foumier. Susan (1998), "Consumers and Their Brands: Develop-ing Relationship Theory in Consumer Research," Journal ofCon.sumer Research, 24 (4), 343-73.

. Susan Dobscha, and David Glen Mick (1998), "Prevent-ing the Premature Death of Relationship Marketing," Har\>ardBusiness Review, 76 (January), 42-51.

and David Glen Mick (1999), "Rediscovering Satisfac-tion," Journal of Marketing, 63 (October). 5-23.

Giles. David C. and John Maltby (2004), "The Role of Media Fig-ures in Adolescent Development: Relations Between Auton-omy, Attachment and Interest in Celebrities." Personality andIndividual Differences, 36 (4). 813-22.

Halpem, Jake (2005), "An Agent for Dead Celebrities," broadcaston All Things Considered. National Public Radio, (August 9),(accessed December 12. 2005), [available at www.npr.orgj.

118 / Journal of Marketing, July 2006

Hazan. Cindy and Phillip R. Shaver (1994). "Attachment as anOrganizational Framework for Research on Close Relation-ships," Psychological Inquiry, 5 (1 ) . 1-22.

and Debra Zeifman (1994). "Sex and the PsychologicalTether." in Advances in Personal Relationships: AttachmentProcesses in Adulthood, Vol. 5, Kim Bartholomew and DanielPerlman. eds. London: Jessica Kingsley. 151-77.

Horovii/. Bruce (2003). "The Good. Bad and Ugly of America'sCeleb Obsession." USA Today, (December 19), (accessedDecember 12. 2005). [available at www.usatoday.com].

Horton. Donald and Richard R. Wohl (1956). "Mass Communica-tion and Para-Social Interaction: Observations on Intimacy at aDistance." Psychiatry: Journal for the Study of InterpersonalProcesses, 19.215-29.

Houran, James. Samir Navik, and Keeli Zerrusen (2(K)5). "Bound-ary Functioning in Celebrity Worshippers," Personality andindividual Differences, 38 (1). 237-48.

Keller. Kevin Lane (2001). "Building Customer-Based BrandEquity: A Blueprint for Creating Strong Brands." MarketingManagement, (July-August). 15-19.

Kleine. Susan Schultz and Stacey Menzel Baker (2004), "An Inte-grative Review of Material Possession Attachment," Academyof Marketing Science Review, 1. 1-39.

Knott. Cynthia L. and Melissa St. James (2(X)2). "An AnalyticHierarchy Process Melhodology Solution for CelebrityEndorser Decisions in Marketing." in Decision Sciences Insti-tute 2002 Annual Meeting Proceedings. San Diego: DecisionSciences Institute. 1427-32.

La Guardia. Jennifer G.. Richard M. Ryan. Charles E. Couchman,and Edward L. Deci (2000). "Within-Person Variation in Secu-rity of Attachment: A Self-Determination Theory Perspectiveon Attachment. Need-Fulfillment and Well-Being." Journal ofPersonality ami Social Psychology, 79 (September). 367-84.

Leak. Gary K. and Ramie R. Cooney (2(X)1), "Self-Determination,Attachment Styles and Well-Being in Adult Romantic Rela-tionships." Representative Research in Social Psychology, 25.55-62.

Lee. Michael (2005), "Iverson. 76ers Not Suited for Success."Washington Post, (November 2). EOl.

Leets, Laura. Gavin de Becker, and Howard Giles (1995), "FANS:Exploring Expressed Motivations of Contacting Celebrities,"Journal of Language and Social Psychology, 14 (March).102-123.

Liljander. Veronica and Tore Strandvik (1995), "The Nature ofCustomer Relationships in Services" in Advances in ServicesMarketing and Management, Teresa A. Swartz, David E.Bowen. and Stephen W. Brown, eds. London: JAI Press,141-67.

North. Adrian C.. Victoria Bland, and Nicky Ellis (2005). "Distin-guishing Her )es from Celebrities."' British Journal of Psychol-ogy, 9(y{\), 39-52.

Park, C. Whan. Bernard J. Jaworski, and Deborah J. Maclnnis(1986), "Strategic Brand Concept/Image Management." Jour-nal of Marketing, 50 (October). 135^5.

Perse. Elizabeth M. and Rebecca B. Rubin (1989), "Attribution inSocial and Parasocial Relationships." CommunicationResearch. 16 (February), 59-77.

Pimentel. Robert W. and Kristy E. Reynolds (2004). "A Model forConsumer Devotion: Affective Commitment witb ProactiveSustaining Behaviors." Academy of Marketing Science Review,(accessed March 10. 2006). [available at http://www.amsrev iew.org/artic I e.s/pimenteI05-2004. pdtl

Plunkett Research (2004), "Entertainment & Media IndustryOverview," Entertainment and Media Industry Statistics,(accessed December 12. 2005). [available at www.plunkettresearch.comj.

Rempel, John K.. Michael Ross, and John G. Holmes (2001)."Trust and Communicated Attributions in Close Relation-ships," Journal of Personality and Social Psychology, 8!(July). 57-64.

Rindova. Violina P.. Timothy Pollock, and Mathew Hayward(2(K)6), "Celebrity Firms: The Social Construction of MarketPopularity." Academy of Management Review, 31 (I), 279-92.

Rose. Randall L. and Stacy L. Wood (2(K)5), "Paradox and theConsumption of Authenticity Through Reality Television."Journal of Consumer Research, 32 (1), 284—96.

Rubin, R.B. and M.P McHugh (1987), "Development of Paraso-cial Interaction Relationships." Journal of Broadcasting andElectronic Mediu,^\ (3), 279-92.

Rubinstein, Robert L. and Patricia A. Parmelee (1992), "Attach-ment to Place and the Representation of the Life Course by theElderly." in Place Attachment, Irwin Altman and Setha M.Low. eds. New York: Plenum Press, 139-63.

Ryan. Richard M. and Edward L. Deci (2000), "Self-Determination Theory and the Facilitation of Intrinsic Motiva-tion. Social-Development, and Well-Being." American Psy-chologist, 55 {\),6^-lS.

and John H. Lynch (1989). "Emotional Autonomy VersusDetachment: Revisiting the Vicissitudes of Adolescence andYoung Adulthood." Child Development, 60 (2). 340-56.

Schultz, Susan E., Robert E. Kleine, and Jerome B. Kernan(1989), "These Are a Few of My Favorite Things: Toward anExplication of Attachment as a Consumer Behavior Construct."in Advances in Consumer Research, Vol. 16, Thomas K. Scrull,ed. Provo, UT: Association for Consumer Research, 359-66.

Sheldon, Kennon M.. Andrew J. Elliot. Youngmee Kim, and TimKasser (2001), "What Is Satisfying About Satisfying Events?Testing 10 Candidate Psychological Needs." Journal of Person-ality and Social P.sychology. 80 (2). 325-39.

Sheth. Jagdish N. and Atul Parvatiyar (1995), "Relationship Mar-keting in Consumer Markets: Antecedents and Consequences."Journal of the Academy of Marketing Science, 23 (4), 255-71.

Silvera. David H. and Benedikte Austad (2004). "Factors Predict-ing the Effectiveness of Celebrity Endorsement Advertise-ments," European Journal of Marketing, 38 (11-12).1509-1526.

Simon, Roger. Dan Gilgoff. and Terence Samuel (2004). "In theHouse of the Believers." US News and World Report. (August9). (accessed March 10. 2006). [available at www.usnews,com].

Simpson, Jeffrey A. (1990), "Influence of Attachment Styles onRomantic Relationships." Journal of Personality and SocialPsychology, 59 (5), 91 \-m.

Spake, Deborah F.. Sharon E. Beatty. Beverly K. Brockman, andTammy Neal Crutchfield (2004), "Consumer Comfort in Ser-vice Relationships," Journal of Service Research. 5 (4),316-32.

Thomson, Matthew, Deborah J. Maclnnis. and C. Whan Park(2005). "The Ties that Bind: Measuring the Strength of Con-sumers* Emotional Attachments to Brands," Journal of Con-sumer Psychology, 15 (1). 77-91.

Tolson, Andrew (2001). "'Being Yourself: The Pursuit of Authen-tic Celebrity," Discourse Studies, 3 (4), 443-57.

Towle. Angela Phipps (2003). "Celebrity Branding." HollywoodReporter, (November 18), (accessed December 12, 2005),[available at www.hollywoodreporter.com].

Yeung, Catherine W.M. and Robert S. Wyer Jr. (2005). "Does Lov-ing a Brand Mean Loving Its Products? The Role of Brand-Elicited Affect in Brand Extension Evaluations," Journal ofMarketing Research, 42 (November), 495-506.

Human Brands/119