Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journalInformation?journalCode=rina20

Download by: [ James Madison University] Date: 23 October 2016, At: 14:24

International Journal of AdvertisingThe Review of Marketing Communications

ISSN: 0265-0487 (Print) 1759-3948 (Online) Journal homepage: http://www.tandfonline.com/loi/rina20

Celebrity endorsements, self-brand connectionand relationship quality

Abhishek Dwivedi, Lester W. Johnson & Robert McDonald

To cite this article: Abhishek Dwivedi, Lester W. Johnson & Robert McDonald (2016) Celebrityendorsements, self-brand connection and relationship quality, International Journal ofAdvertising, 35:3, 486-503

To link to this article: http://dx.doi.org/10.1080/02650487.2015.1041632

Published online: 13 May 2015.

Submit your article to this journal

Article views: 621

View related articles

View Crossmark data

Celebrity endorsements, self-brand connection and relationship

quality

Abhishek Dwivedia*, Lester W. Johnsonb and Robert McDonaldc

aSchool of Management & Marketing, Charles Sturt University, P.O. Box 789, Albury, NSW 2640,Australia; bSwinburne Business School, Department of Marketing, Tourism and Social Impact, 200Leicester Street, Carlton, VIC 3053, Australia; cRawls College of Business, Texas Tech University,

Area of Marketing, MS 42101, Lubbock, TX 79409-2101, USA

(Received 29 October 2012; accepted 1 April 2014)

The extant literature on celebrity endorsement effects largely focuses on theendorsement effects on consumer evaluations of the endorsed brand. The currentstudy extends the literature by assessing the impact of endorser credibility on twoconsumer�brand relationship-oriented outcomes � brand relationship quality andconsumer self-brand connections. Additionally, the self-brand connection ispositioned as a partial mediator of the effect of endorser credibility on relationshipquality. A conceptual model is developed and estimated on a sample of 535Generation Y (youth) consumers from India. The hypotheses are supported and themodel demonstrates acceptable fit to the data. Overall, the present study introduces arelationship-building perspective to the celebrity endorsement literature. The resultssuggest that celebrity endorsers possess the ability to provide meaningful self-definitional benefits to consumers as well as cultivate enhanced relationship qualitywith the endorsed brands, thus contributing novel insights into celebrity endorsementdynamics.

Keywords: celebrity endorser credibility; self-brand connection; relationship quality

Introduction

Celebrity endorsements represent a multi-billion dollar industry globally, and continue to

be used regularly as part of marketing campaigns (Eisend and Langner 2010; Kelting and

Rice 2013). A reason for their continued usage is that celebrity endorsers possess sym-

bolic and aspirational associations (Escalas 2004) that are readily transferred to brands

that they endorse (Elliott and Wattanasuwan 1998). Brands with such added associations

are able to generate high visibility and persuasive impact in the market, generally culmi-

nating in greater levels of advertisement and brand recall (Stafford et al. 2002). Academic

research reveals that a major determinant of celebrity endorsement effects is consumer-

perceived credibility associated with an endorser (Amos et al. 2008; Ohanian 1990), rep-

resenting a dominant paradigm in the endorsement literature (Stafford et al. 2002).

Endorser credibility, as typically comprising consumer perceptions of a celebrity’s attrac-

tiveness, expertise with the endorsed product and trustworthiness (Ohanian 1990), is gen-

erally positively associated with greater persuasiveness of the delivered message

(Erdogan 1999), enhanced brand information recall (Speck et al. 1988), endorsed brand

evaluations (Stafford et al. 2002), as well as consumer-based brand equity (Spry et al.

*Corresponding author. Email: adwivedi@csu.edu.au

� 2015 Advertising Association

International Journal of Advertising, 2016

Vol. 35, No. 3, 486�503, http://dx.doi.org/10.1080/02650487.2015.1041632

2011). Despite advancements in examining the effects of endorser credibility on con-

sumer behavior, the celebrity endorsement literature provides relatively limited insights

into the influence of endorser credibility on certain critical aspects of consumer behavior

that are relevant to brand marketing. Two such aspects that are the focus of the current

investigation are consumer�brand relationships and the consumer self-brand connection.

The marketing discipline has adopted a relationship-based paradigm whereby build-

ing long-term consumer�brand relationship with consumers is considered vital to long-

term business success (Gr€onroos 2006; Hennig-Thurau et al. 2002). However to date,

explicit empirical investigations into celebrity endorsements effects on consumer�brand

relationships has been limited to indirect support in the trade literature. For instance,

celebrity endorsements have shown to improve consumers’ emotional engagement with

an endorsed brand, culminating in stronger consumer relationships with the endorsed

brand (Yelin and Kinnear 2012). Similarly, practitioners also stress the need for the

endorsement process to be able to resonate with the intended target audience (Cahal

2012), thereby implying the need for investigating a relationship-building role of the

endorsement process.

From a theoretical standpoint, development of a consumer relationship with brands

serves the purpose of providing functional and emotional benefits such as enabling con-

sumers to look and feel better, as well as providing an opportunity to behave consistently

with valued ideology (MacInnis et al. 2009). Brands by virtue of fulfilling such emotional

and hedonic needs are increasingly being viewed as relational partners by consumers

(Fournier 1998). For firms, development of consumer�brand relationships favorably

impact valued consumer outcomes, such as enhanced positive word-of-mouth, brand loy-

alty, brand forgiveness and involvement in brand communities (MacInnis et al. 2009).

Given the importance of consumer�brand relationships to consumers and firms alike,

and the extensive use of celebrity endorsements, it is worthwhile to investigate the influ-

ence of endorser credibility on the strength of consumer�brand relationships, i.e., rela-

tionship quality (Gr�egoire et al. 2009; Hennig-Thurau et al. 2002). We define relationship

quality as perceived by consumers as their trust in and commitment to a brand, along

with the perceived social benefits gained from the relationship. Theoretically, the current

study contributes by extending the scope of endorsement effects from brand-related out-

comes to relationship-orientated outcomes. From a managerial perspective, the current

investigation has implications for using celebrity endorsers as facilitators of long-term

relationships with consumers.

The second aspect investigated in the current study is the potential impact of celeb-

rity endorsements on the consumer self-brand connection. The construct of self-brand

connection was first introduced by Escalas and Bettman (2003) who define it as the

extent to which a consumer integrates a brand into his or her self-concept. The construct

is relevant to understanding celebrity endorsement effects since consumers use celebri-

ties as reference groups for constructing and redefining their desired self-identities

(Escalas and Bettman 2005). Literature on self-concept theory (Belk 1988; Sirgy 1982)

and meaning-transfer models (McCracken 1989) supports this logic. Consumers actively

engage in the process of self-identity creation, and purchase of products and brands

helps to address such self-definitional needs (Belk 1988). Furthermore, brands endorsed

by celebrities can facilitate the transfer of cultural meaning as the meaning of a brand

for consumers ultimately draws from the culturally constituted world (McCracken 1989)

and the celebrity acts a conduit of meaning transfer to the brand. Consumers in turn use

this shared meaning as a resource (Holt 2002) to construct a more individualized mean-

ing (McCracken 1989), which is vital to fulfillment of self-definitional needs (Eisend

International Journal of Advertising 487

and Langner 2010; Escalas and Bettman 2013), resulting in formation of self-brand

connections.

Additionally, investigation of self-brand connection as an outcome of the endorse-

ment process helps address a vital issue in the advertising literature. A perplexing concern

typically raised in the context of celebrity endorsements is that despite consumers being

able to recall a particular celebrity-based advertisement, consumers are unable to recall

or relate to the advertised brand (Keller 2013; Rossiter and Percy 1987). This situation

raises questions about the effectiveness of an ad campaign. Therefore, within an endorse-

ment-type advertising scenario, experts have highlighted the need for the creation of

stronger links between consumers and an endorsed brand per se (Van Kuilenburg et al.

2011; Zdravkovic and Till 2012; Wang and Muehling 2012). More recently, in a large

practitioner-initiated multi-national study of brand consumption behavior, it was revealed

that self-identification with a brand is a vital factor leading to brand preference among

consumers (Sbarbaro et al. 2011; Van den Bergh and Behrer 2011). Despite possessing

theoretical appeal and practitioner relevance, empirical investigations into the impact of

celebrity endorsements on consumer self-brand connection are only beginning to emerge

in the literature. For instance, Escalas and Bettman (2008) as cited in Escalas and

Bettman (2013) observe that well-liked celebrity endorsers lead to formation of self-brand

connections among consumers. The present study extends this work, addressing a vital

gap in the literature by empirically examining the impact of endorser credibility on the

self-brand connection, representing the second contribution of the current study. Further-

more, we argue that the effect of endorser credibility on brand relationship quality may

materialize via two pathways. Endorser credibility is likely to have direct influence on

relationship quality, as well as indirectly via impacting a self-brand connection. Thus, a

self-brand connection acts as a partial mediator of the influence of endorser credibility on

relationship quality. The current study represents an attempt to examine such effects

under a holistic framework of celebrity endorsement effects, empirically examining the

proposed effects using structural equation modeling. The conceptual model of the current

study is presented next.

The conceptual model

The conceptual model (Figure 1) is developed to address the core issue of the present

research � how do celebrity endorsements impact consumers’ brand relationship quality?

A starting point in model development is to understand and conceptualize endorser credi-

bility � the prime endorser characteristic driving subsequent consumer behavioral

effects. With origins in source credibility theory (Hovland et al. 1953), endorser credibil-

ity receives substantial investigation in the literature as a major determinant of consum-

ers’ endorser-initiated behavior in the marketplace (Amos et al. 2008). Initially

conceptualized as being composed of perceived trustworthiness and expertise of a source

(Hovland et al. 1953), the construct was later expanded in the celebrity endorsement

domain by Ohanian (1990) who investigated it as a three-dimensional construct represent-

ing consumer evaluations of a celebrity endorser’s level of trustworthiness, expertise and

attractiveness. Ohanian’s (1990) conceptualization of endorser credibility is adopted in

the present research. Trustworthiness refers to the honesty and integrity and the degree of

confidence associated with an endorser (Erdogan 1999; Ohanian 1990). Some experts

claim that trustworthiness of a source is a critical component of credibility in the persua-

sion process (Perloff 2010). Expertise is a consumer perception of knowledge, experience

or ability associated with an endorser with the product being endorsed (Erdogan 1999).

488 A. Dwivedi et al.

The expertise associated with a source is found to be positively associated with attitude

change (Ohanian 1990) as well as enhanced persuasive impact (Erdogan 1999). Finally,

the attractiveness dimension captures consumer perceptions of physical appeal of an

endorser (Ohanian 1990). Physical appeal of a communicator (endorser) facilitates

enhanced receiver (consumer) attention to message arguments, leading to favorable atti-

tude change (Erdogan 1999). The focus of the current research is on understanding the

effect of endorser credibility on relationship quality, which is hypothesized next.

Endorser credibility and relationship quality

Consumers tend to be generally attracted to celebrities (Thompson 2006) and to the

objects associated with them (Radford and Bloch 2013) as celebrities represent reposito-

ries of cultural meaning that is valued by consumers (McCracken 1989). This added

meaning is attached to the brands that are endorsed by admired celebrities, thereby imbu-

ing the endorsed brands with added value as perceived by consumers (Richins 1994). For

instance, a wrist watch endorsed by Leonardo DiCaprio is likely to be valued more by

consumers as a result of the added meaning (say, elegant and trustworthy) being associ-

ated with the product. Specifically, if an endorser is perceived as credible, that is, per-

ceived as attractive, trustworthy and an expert in the use of the endorsed product, such

perceptions are readily transferred to the endorsed brand (McCracken 1989). Moreover,

as a result of such endorser-infused value, consumers may begin to form emotional bonds

or relationships with an endorsed brand (Fournier 1998). Such relational bonding may

begin to manifest over time as various relationship-oriented behaviors, such as develop-

ment of enhanced trust in the brand (Arantola 2002), a greater commitment towards main-

taining the relationship (Sheth and Parvatiyar 1995) and enhanced perception of deriving

social benefits (Hennig-Thurau et al. 2002). Additional support for the expected relation-

ship is derived from reference-group theory (Bearden and Etzel 1989) that states people

compare themselves to reference groups to whom they look for guidance for their own

Endorser credibility

Trustworthiness

Expertise

Endorser-brand fit

Self-brand connection

Relationship quality

Trust CommitmentSocial

benefits

H1+

H2+

Consumer satisfaction

H3+

Attractiveness

H4+

H6+

H5+

Figure 1. The conceptual model.Note: Endorser-brand fit and consumer satisfaction are specified in the model as control variables.

International Journal of Advertising 489

behavior. For consumers, celebrities represent aspirational reference groups (Escalas

2004), and consumers refer to the normative practices of celebrities and develop values

and standards for their own behavior (Sheth and Parvatiyar 1995; Bearden and Etzel

1989). Shaping one’s behavior in line with that of a reference group entails a dual benefit

for consumers. On one hand, there is a fulfillment of an aspirational motivation, and on

the other hand, a reduction of psychological risks (Sheth and Parvatiyar 1995). Therefore,

we argue that consumers will engage in favorable relationship behaviors with celebrities

that are perceived as credible (Sheth and Parvatiyar 1995; Schiffman et al. 2011). Thus:

H1: Endorser credibility has a direct positive impact on relationship quality.

Endorser credibility and self-brand connection

The self-brand connection is an emergent concept in the marketing literature (Moore and

Homer 2008), and is defined as the extent to which a consumer integrates a brand into his

or her self-concept (Escalas and Bettman 2005). The concept of self-brand connection is

related to the self-identity theory (Belk 1988), the central premise of which is that con-

sumers use brands as tools for creation and maintenance of self-identity (Belk 1988).

Despite being related to the concept of a consumer’s sense of self, the current study con-

siders the self-brand connection construct as conceptually distinct from ‘brand-engage-

ment in self-concept’ (BESC) (Sprott et al. 2009). The distinction is important as the two

constructs pertain to consumers’ sense of self and the role of brands in relation to consum-

ers’ sense of self. The difference lies primarily in the levels of abstraction of the two con-

structs. Sprott et al. (2009) define BESC as ‘consumers’ propensity to include important

brands as part of how they view themselves’ (92). Where BESC is a generalized con-

sumer tendency to incorporate brands as a part of the self-concept, self-brand connection

is brand-specific in its level of abstraction and conceptualization. This distinction is also

acknowledged by Sprott et al. (2009). Additionally, a self-brand connection is individual-

specific in nature. Celebrity endorsers imbue brands with symbolic associations and con-

sumers use the brand symbolism to shape their self-concept and behavior (Cutright et al.

2013). The strength of this highly individualized self-definitional meaning is captured by

the construct of self-brand connection (Escalas and Bettman 2013).

Given that celebrities represent symbolic and aspirational reference groups for consum-

ers (Escalas 2004), a positive relationship is expected between endorser credibility and

self-brand connection. Support for this logic is derived from meaning transfer dynamics

(McCracken 1989) and the value-expressive function of reference groups (Bearden

et al. 1989). As stated previously, celebrity endorsements are a mechanism by which mean-

ing from the culturally constituted world transfers to brands (McCracken 1989). However,

brands not only possess a shared (common) culturally derived meaning, but also a highly

individualized meaning that is created by and is unique to a consumer (Alan et al. 2008).

The self-brand connection concept allows consumers to ascribe their own meanings to a

brand (Escalas and Bettman 2013), which is vital to fulfillment of their self-definitional

needs (Belk 1988; Escalas and Bettman 2013). For instance, consider a consumer who

aspires to be as successful and confident as Roger Federer, who endorses Gillette razors. In

order to appropriate Federer’s meanings to construct his own identity, this consumer may

engage in tennis and use Gillette razors. As a result, he may form a self-brand connection

with the endorsed brand. Furthermore, celebrities by virtue of their reference group status

provide a value-expressive function to consumers, meaning that consumers are motivated

to enhance individual self-concept by identifying themselves with the celebrity reference

490 A. Dwivedi et al.

group (Bearden et al. 1989). Therefore, it is expected that when a brand is endorsed by a

celebrity who is perceived as credible, it is likely perceived by consumers as salient to ful-

fillment of self-definitional needs. Consumers are thus likely to emulate a celebrity

endorser’s behavior and use the endorsed brand as a resource to craft an individual self-

meaning (Holt 2002), increasing the likelihood of integrating the brand as a part of their

self-concept. Emergent empirical evidence supports this theorization, whereby well-liked

celebrity endorsers were observed to facilitate self-brand connections among consumers

(Escalas and Bettman (2008) as cited in Escalas and Bettman (2013)).

Endorser trustworthiness, i.e., dependability, honesty, reliability and sincerity

(Ohanian 1990), may help reduce the psychological risk associated with incorporating the

brand into the self-concept. Endorser expertise may boost the consumer’s belief that the

brand will perform well, possibly increasing the consumer’s confidence enough to incor-

porate the brand into her/his self-concept. Endorser attractiveness, i.e., likeability, famil-

iarity and similarity (Amos et al. 2008), may affectively draw consumers closer to the

brand. Erdogan’s (1999) findings that endorser credibility leads to greater persuasiveness

may create positive associations with the brand. Enhanced brand information recall due

to endorser credibility (Speck et al. 1988) may indicate that the brand is more salient to

the consumer. Finally, endorsed brand evaluations (Stafford et al. 2002) may lead the

consumer to be more open to incorporating the brand into his or her self-concept. Thus,

the following hypothesis is formulated:

H2: Endorser credibility has a direct positive impact on self-brand connection.

Self-brand connection and relationship quality

It is expected that the self-brand connection is positively associated with relationship

quality. As outlined in the preceding section, development of a self-brand connection

directly fulfills self-definitional needs of consumers (Belk 1988; Sirgy 1982). Once

such vital self-concept motivations are addressed, consumers may begin to develop a

favorable relationship with the brand. Consumers may begin to develop greater trust in

the particular brand as it has served a vital purpose of self-identity creation (Wang and

Bloemer 2008). In addition, consumers may commit to using the brand in the future in

order to maintain the desired sense of self (Albert et al. 2013). Furthermore, a percep-

tion of social benefits may materialize as a result of possible enhancement in endorser-

inspired social self-image (Schiffman et al. 2011). Therefore, when consumers per-

ceive a brand as fulfilling self-definitional needs, they are likely to engage in long-

term relational behavior with such brands (Sheth and Parvatiyar 1995). Additional the-

oretical support for this logic is drawn from the theory of reciprocal action (Li and

Dant 1997), a central tenet of which is that consumers often develop a strong sense of

commitment towards a firm in response to its perceived relationship-building efforts.

The theory has been validated across diverse consumer�brand relationship contexts

(Kim et al. 2008; Yoon et al. 2008), and can be extended to the present context. Thus,

consumers who perceive that a brand provides them with valuable self-identity-defin-

ing benefits are likely to reciprocate by engaging in long-term relational behavior with

the brand, enabling them to continue to derive such self-definitional benefits (Dolich

1969). This dynamic likely enhances the importance and quality of the relationship for

the consumer (Brodie et al. 2013). Hence:

H3: A self-brand connection has a direct positive impact on relationship quality.

Thus, we posit that endorser credibility influences consumer�brand relationship qual-

ity through two routes. The first is a direct effect of endorser credibility on relationship

International Journal of Advertising 491

quality as per theoretical mechanisms outlined for Hypothesis H1. The second route is

mediated through the self-brand connection.

Impacts of endorser-brand fit and consumer satisfaction

Consumer�brand relationship quality may be influenced by a number of factors other

than endorser credibility and a self-brand connection. Two factors that likely play an

important role in developing the consumer�brand relationship are endorser-brand fit

and consumer satisfaction. Therefore, these are used as covariates in the current study.

Consumer-perceived endorser-brand fit is expected to positively influence perceived

relationship quality. Endorser-brand fit is defined as a degree of consumer-perceived

similarity between the image of the celebrity endorser and the brand being endorsed.

Our conceptualization of endorser-brand fit draws upon cognitive categorization mech-

anism that consumers use to make sense of the consumption environment (Cohen and

Basu 1987). Categorization theory explains how people group various stimuli in their

environment into discrete categories so as to make sense of the consumption environ-

ment, thereby enhancing information processing efficiency (Cohen and Basu 1987). In

a celebrity endorsement scenario, consumers would likely judge fit (or similarity)

between an endorser and a brand using na€ıve ‘mental explanations’ that are not based

on much scientific foundation (Murphy and Medin 1985, 290). Ultimately, endorser-

brand fit is judged based on the degree to which the pairing ‘makes sense to the

perceiver’ (291). Such mechanism of consumer assessment of fit between entities

receives widespread support in the branding and advertising literatures (e.g., Bridges

et al. 2000; Zdravkovic and Till 2012). Given that celebrities personify consumers’

diverse aspirational associations, enhanced endorser and brand fit potentially leads to

favorable relational behaviors on the part of consumers such as the development of a

greater trust in and relationship commitment towards the endorsed brand. The

endorser-brand match-up framework observes a positive relationship between

perceived endorser-brand fit and favorable outcomes such as enhanced brand attitude

(Misra and Beatty 1990) and purchase intentions (Kahle and Homer 1985). A similar

positive effect on relationship quality is expected. Therefore, the following hypotheses

are formulated:

H4: Endorser-brand fit has a direct positive impact on relationship quality.

Consumer satisfaction with a product (service) is also likely to play an explanatory

role in the hypothesized model. Hence, we include satisfaction as a covariate in the cur-

rent study. Satisfaction is defined as a customer’s overall evaluation of the performance

of an offering to date (Gustafsson et al. 2005). Conceptualized as relative (dis)confirma-

tion of expectations (positive or negative), the literature outlines a prominent role of over-

all satisfaction in affecting consumer relationships (Anderson and Mittal 2000; Hennig-

Thurau et al. 2002). Thus, a positive influence of satisfaction on relationship quality is

suspected. Additionally, satisfaction likely exerts an impact on consumer self-brand con-

nection. Brands are tools for consumers to differentiate themselves and to express their

personality and individuality (Escalas 2004). However, such creation and maintenance of

the desired self-image necessitates that consumers be satisfied with a brand’s performance

(i.e., a brand either meets or exceeds expectations), because non-performing brands carry

negative associations of incompetence, resulting in a possible dissonance. Thus, consum-

ers would not like to associate their self-identities with poorly performing brands (Markus

1977). Hence:

492 A. Dwivedi et al.

H5: Consumer satisfaction has a direct positive impact on relationship quality.

H6: Consumer satisfaction has a direct positive impact on self-brand connection.

Research design

A self-administered cross-sectional survey was used to collect data from a convenience

sample of 620 users of telecom services. Data were collected from a management institute

in New Delhi, India, in 2012. The respondents, enrolled in business courses, were contacted

in classrooms at the beginning of lecture time for the purpose of the survey. The academics

were given explicit instructions by the research team regarding how to brief the respondents

prior to the start of the survey. The specific research objectives were not revealed to the

respondents so as to limit social desirability responding. The respondents were instructed

to attempt the survey on their own without consulting other participants. Data collection

was completed within three weeks of first contact. The respondents all belong to the Gener-

ation Y (youth, 18�25 years old) cohort. The young consumer segment (under 25 years

old) constitutes roughly half of the Indian population (Euromonitor 2007), and is consid-

ered as a major consumer segment for telecoms services (Evalueserve 2009). The segment

is speculated to be a driver of future growth in the telecom services sector (Chaturvedi

2011). The survey was targeted at respondents belonging to middle-income households,

which represent the biggest consumer group in the country (Ablett et al. 2007).

Stimulus design

The study was conducted in the telecom services industry. We chose this context for pri-

marily two reasons. First, the telecom service sector in India represents one of the domi-

nant adopters of celebrity endorsers (Indiantelevision.com 2009), thereby being suitable

for the current study. Second, India is one of the fastest growing mobile phone markets in

the world (Euromonitor 2007), with approximately 890 million subscribers nationally

(Telecomcircle.com 2012). Furthermore, in New Delhi, where the data for the current

study were collected, there are approximately 42 million active subscribers (Telecom-

circle.com 2012).

Around three-quarters of the market share of the Indian telecom services is controlled

by five brands, viz., Airtel (20% share), Reliance (17% share), Vodafone (16% share), Idea

(11% share) and BSNL (11% share) (Telecom Regulatory Authority of India 2011). Three

of the top five brands (i.e., Reliance, Idea and BSNL) were offered as choices to respond-

ents. Airtel was not offered as a choice in the current study since it is endorsed by multiple

celebrity endorsers, thereby potentially introducing confounding effects of different endors-

ers. Vodafone was excluded since its current advertising campaign in India uses spokes-

characters (called Zoo Zoos), study of which is beyond the scope of the current research

(see Garretson et al. 2012 for a discussion). The three brands that are offered as choices to

respondents are endorsed by Indian movie celebrities, namely, Hritik Roshan (for

Reliance), Abhishek Bachchan (for Idea) and Deepika Padukone (for BSNL), respectively.

The respondents were asked to base their responses on the endorser-brand combination for

a service brand that they use most of the time (i.e., their primary service brand).

Questionnaire, measurement and analysis

Several steps were taken to enhance the validity of the study. The survey questionnaire

was structured and formatted to minimize the potential of common method bias

International Journal of Advertising 493

(Podsakoff et al. 2003). Construct items are operationalized clearly (consistent with their

usage in the literature) in order to reduce the risks of ambiguity and item-demand charac-

teristics. Item intermixing was not adopted, thus eliminating a potential source of com-

mon method bias. Finally, respondent anonymity was ensured, and the questionnaire was

broken up into sections, design aspects that help to minimize common method bias

(Podsakoff et al. 2003). A post hoc test of common method bias was conducted to test its

severity in the data. Also, the context of the survey pertains to consumer assessment of

their telecom service brands (a general topic), which does not seem to be susceptible to a

social desirability bias.

We use previously validated scales to operationalize the constructs, measured on

multi-item five-point Likert scale type format. Relationship quality is the dependent vari-

able in the present study, and is investigated for the first time in the context of celebrity

endorsements. In the relationship marketing literature, relationship quality is operational-

ized variously in the literature, ranging from unidimensional measures (Goodman and

Dion 2001) to multi-dimensional ones (Gr�egoire et al. 2009), with no consensus to date

regarding its operationalization. In the present study, relationship quality is conceptual-

ized as a fully reflective second-order construct, defined as consumer perceptions of trust

in a brand, commitment towards maintaining a relationship and perceptions of social ben-

efits received from a brand. The first dimension, trust, is defined as consumer perception

that a brand is dependable and can be relied on to do the right thing (Ganesan 1994). The

second dimension, relationship commitment, is defined as consumers’ tendency to main-

tain the relationship (Aurier and N’Gaola 2010). Perception of social benefits received

(the third dimension) is defined as consumer perception of personal recognition by a

brand (Hennig-Thurau et al. 2002). This conceptualization of relationship quality is con-

sistent with prominent conceptualizations adopted in the literature (Gr�egoire et al. 2009;

Hennig-Thurau et al. 2002). Moreover, relationship quality is conceptualized as a fully

reflective construct, that is, reflective at the first- and second-order levels (referred to as a

Type-I construct by Jarvis et al. (2003). Endorser credibility is operationalized as a sec-

ond-order fully reflective construct (Ohanian 1990), that is, measured using five items

each for perceived attractiveness, expertise and trustworthiness as dimensions. Self-brand

connection is measured using four items derived from Escalas and Bettman (2003).

Finally, three items from Till and Busler (2000) operationalize endorser-brand fit, and

overall satisfaction is measured using three items derived from Aurier and N’Goala

(2010).

Structural equation modeling is used as the analytic tool. A measurement model anal-

ysis is first conducted for the purpose of assessing the validity of the proposed factor

structure, followed by an analysis of the hypothesized structural model. Model fit is

assessed using four indices as recommended by Bagozzi and Yi (2012). These are Com-

parative Fit Index (CFI), Tucker and Lewis Index (TLI), Root Mean Square Error of

Approximation (RMSEA) and Standardized Root Mean Square Residual (SRMR).

Data analysis and results

Preliminary analyses

The raw data were first cleaned. Records with substantial extreme and mid-point

responses, and substantially large missing values, were eliminated. A total of 535 usable

responses were available as the usable dataset. Missing values were less than 3% of the

data, and were substituted using a full-information maximum-likelihood estimation

494 A. Dwivedi et al.

procedure (Enders and Bandalos 2001). The usable sample adequately meets the 10:1

respondent-to-item ratio (Hair et al. 2009). We first examine the extent of common

method bias in the data using the single-factor test suggested by Podsakoff et al. (2003).

We compared the five factor measurement model of the current study with a common-

methods factor. The single-factor model elicits a significant Chi-square (x2 D 3336.42,

degrees of freedom D 350, p < 0.01) with unacceptable fit to data (CFI D 0.58, TLI D0.55, SRMR D 0.120 and RMSEA D 0.126). The research measurement model, on the

other hand, is significantly different from and better fitting than the single-factor model

(i.e., x2 (334) D 583.10; p < 0.01; CFI D 0.96, TLI D 0.96, SRMR D 0.039 and

RMSEA D 0.037), meeting the model fit cut-offs suggested by Bagozzi and Yi (2012).

Thus, no serious threat of common-method bias is suggested in the data. Furthermore,

these results indicate that the model fits the data well.

Descriptives

The means, standard deviations, inter-construct correlations and square root of average

variance extracted (AVE) scores are reported in Table 1. The construct means range from

3.09 to 3.67 (out of 5.0), and corresponding standard deviations range from 0.84 to 0.98.

Inter-construct correlations are positive and significant (p < 0.05) for all pairs of con-

structs. The only exception is the correlation between perceived endorser attractiveness

and consumer satisfaction that is significant at the 10% level.

Construct validity

We examine construct reliability and validity using traditional tests. Standardized factor

loadings, reliability estimates and AVE scores are reported in Table 2. All Cronbach

alphas are 0.75 or greater for all the constructs, exceeding the traditionally accepted level

of 0.70. Similarly, composite reliability estimates are above 0.75 for all constructs. Con-

vergent validity is demonstrated as factor-loadings are highly significant and load

strongly on respective constructs. The AVE score for each construct is greater than 0.50,

further supporting convergent validity (Fornell and Larcker 1981). Furthermore, discrimi-

nant validity is evidenced as the square-root of AVE (refer to Table 1) for any given

Table 1. Descriptives, bivariate correlations and square root of AVE estimates.

Construct Mean S.D. TWR ATT EXP FIT SBC BT COM SOC SAT

Trustworthiness (TWR) 3.39 0.84 0.76

Attractiveness (ATT) 3.67 0.98 0.36�� 0.79

Expertise (EXP) 3.13 0.84 0.49�� 0.23�� 0.71

Endorser-brand fit (FIT) 3.28 0.86 0.36�� 0.19�� 0.48�� 0.73

Self-brand connection(SBC)

3.09 0.93 0.20�� 0.09�� 0.25�� 0.25�� 0.78

Brand trust (BT) 3.33 0.93 0.27�� 0.14�� 0.30�� 0.34�� 0.49�� 0.79

Commitment (COM) 3.12 0.98 0.23�� 0.17�� 0.23�� 0.32�� 0.57�� 0.65�� 0.77

Social benefits (SOC) 3.31 0.88 0.23�� 0.15�� 0.23�� 0.32�� 0.59�� 0.59�� 0.52�� 0.77

Satisfaction (SAT) 3.33 0.96 0.17�� 0.11� 0.14�� 0.24�� 0.62�� 0.60�� 0.60�� 0.60�� 0.79

Note: ��Significance at 0.01 level; �significance at 0.10 level. The square root of the average variance extracted isgiven in italics along the diagonal.

International Journal of Advertising 495

Table 2. Reliability and validity estimates.

Construct and itemsStandardizedloading (sig.) Alpha CR AVE

Trustworthiness 0.81 0.81 0.58

[Endorser] is trustworthy 0.80�� � � �[Endorser] is honest 0.77�� � � �[Endorser] is reliable 0.72�� � � �Attractiveness 0.83 0.83 0.63

[Endorser] is good looking 0.85�� � � �[Endorser] is attractive 0.77�� � � �[Endorser] is sexy 0.75�� � � �Expertise 0.75 0.75 0.50

[Endorser] is an expert 0.71�� � � �[Endorser] is knowledgeable 0.70�� � � �[Endorser] possesses experience 0.70�� � � �Endorser-brand fit 0.78 0.78 0.54

There is match-up 0.76�� � � �There is similarity 0.73�� � � �Combination is appropriate 0.71�� � � �Self-brand connection 0.86 0.86 0.61

I consider my telecom brand as a part of myself 0.85��

I have a special bond with my telecom brand 0.78�� � � �I feel a personal connection with my brand 0.77�� � � �My brand is an important indication of who I am 0.71�� � � �Brand trust 0.86 0.84 0.63

The [brand] is honest in addressing my concerns 0.83�� � � �Cares about its customers 0.81�� � � �Can count on it to do what is right 0.74�� � � �Relationship commitment 0.84 0.82 0.60

Proud to be a customer 0.80�� � � �Intend to maintain an indefinite relationship 0.77�� � � �The brand has a lot of personal meaning 0.75�� � � �Social benefits 0.82 0.81 0.59

Relation based on brand’s ability to make me feel important 0.78�� � � �Relation based on brand’s ability to know my needs 0.77�� � � �Relation based on brand’s ability to build a one on one

connection0.75�� � � �

Performance satisfaction 0.83 0.83 0.62

I did the right thing when I signed up 0.84�� � � �I am satisfied with my service 0.80�� � � �My choice is a wise one 0.72�� � � �Note: Alpha refers to Cronbach’s alpha reliability estimate; CR refers to composite reliability estimate; AVErefers to average variance extracted. ��Significance at 0.01 level.

496 A. Dwivedi et al.

construct is greater than the standardized correlation coefficient of that construct with all

other constructs (Fornell and Larcker 1981). Moreover, the measurement model analysis

using a confirmatory factor analysis demonstrates acceptable fit to data, as noted previ-

ously. These results suggest that adequate construct validity can be assumed in the study.

We would like to clarify that we collected data on the original 15-item Ohanian

(1990) scale (i.e., five items for each of the three dimensions). However, preliminary

measurement analysis revealed that two items of each of the dimensions load weakly

(i.e., standardized factor loading < 0.70) on the respective construct. For instance, items

‘endorser is classy’ and ‘endorser is elegant’ load inadequately on the attractiveness

dimension. Similarly, items ‘endorser is dependable’ and ‘endorser is sincere’ load

weakly on the expertise dimension. These items were eliminated from further analyses,

resulting in the use of a shortened version of the Ohanian scale.

Structural model analysis and hypotheses tests

The structural model analysis elicits a significant Chi-square statistic (i.e., x2 (335) D583.12, p < 0.01). Other fit indices reveal an acceptable fit to data (CFI D 0.96, TLI D0.96, SRMR D 0.039 and RMSEA D 0.037). The individual hypotheses are examined

next. Table 3 reports the parameters of the structural model. As expected, endorser credi-

bility exerts a direct significant impact on relationship quality (standardized beta coeffi-

cient, b D 0.16, C.R. D 2.48, p < 0.05). Thus, Hypothesis H1 is supported. Endorser

credibility exerts a direct significant impact on the self-brand connection (b D 0.16,

C.R. D 3.40, p < 0.01), eliciting support for Hypothesis H2. Relationship quality is sig-

nificantly impacted by self-brand connection (b D 0.19, C.R. D 3.10, p < 0.01), thus sup-

porting Hypothesis H3.

Endorser-brand fit exerts a direct significant influence on relationship quality (b D0.15, C.R. D 2.92, p < 0.05), supporting Hypothesis H4. Overall satisfaction exerts a

strong significant impacts on relationship quality (b D 0.64, C.R. D 9.33, p < 0.01), sup-

porting Hypothesis H5. Finally, Hypothesis H6 was supported as satisfaction exerted a

significant impact on the self-brand connection (b D 0.70, C.R. D 13.15, p < 0.01).

Overall, the estimated model explains 56% of variation in self-brand connection and 82%

variation in relationship quality.

Table 3. Structural parameter estimates.

b CR Sig. 95% C.I. Bootstrap p

Hypothesized path

H1: Endorser credibility! Relationship quality 0.16 2.48 0.013 0.01�0.32 0.040

H2: Endorser credibility! Self-brand connection 0.16 3.40 ��� 0.05�0.26 0.006

H3: Self-brand connection! Relationship quality 0.19 3.10 0.002 0.05�0.31 0.009

Controlled influences

H4: Endorser-brand fit! Relationship quality 0.15 2.92 0.012 0.01�0.29 0.041

H5: Satisfaction! Relationship quality 0.64 9.33 ��� 0.52�0.75 ���

H6: Satisfaction! Self-brand connection 0.70 13.15 ��� 0.59�0.78 ���

Note: b refers to the standardized beta coefficient; CR refers to critical ratio; n.s. refers to ‘not significant’. ���p< 0.001. Bootstrap p refers to the bootstrapped significance level (using 5000 bootstrap samples).

International Journal of Advertising 497

Assessing parameter stability

In order to assess stability of the parameter estimates in the preceding analysis, a boot-

strapping procedure in structural equation modeling was carried out. Bootstrapping is a

form of resampling in which the original data are repeatedly sampled with replacement

for model estimation. Multiple sub-samples of the same size are randomly drawn with

replacement from the original sample and data for the parameter distributions for each of

the spawned sub-samples are generated (Byrne 2010). Parameter stability is assessed

through an examination of bootstrap standard errors, bias estimates (i.e., difference

between the bootstrap estimates and the original estimates) and an examination of the

bias-corrected confidence intervals (Byrne 2010). In the current study, bootstrapping is

conducted using 5000 samples with 95% confidence intervals. The analysis reveals bias

(i.e., the difference between original and bootstrapped estimate) values are all below

0.01. More importantly, the 95% confidence intervals for the parameter estimates exclude

a zero point (refer to the last two columns of Table 3). These results suggest that the

parameters of the estimated model are stable, implying plausibility in a much larger sam-

ple than the one used in the current study (Byrne 2010).

Discussion and implications

The present study adds a consumer�brand relational perspective to the celebrity endorse-

ment literature. Specifically, the impact of endorser credibility was modeled on consumer-

perceived relationship quality with the endorsed brand. The results suggest that endorser

credibility exerts a direct significant impact on consumer-perceived relationship quality

with an endorsed brand. A theoretical implication of our study pertains to extending the

conceptual scope of celebrity endorsement effects. Traditionally, such effects are modeled

on brand evaluations such as brand attitudes (Silvera and Austad 2004) and brand equity

(Till 1998). We add relationship quality as an outcome of celebrity endorsement effects,

thereby complementing the literature on consumer�brand relationships (Fournier 1998). A

managerial implication of the finding is that celebrity endorsers can be used as relationship

building and enhancement tools. In an era of media fragmentation, a credible celebrity may

be able to break through the clutter and directly facilitate long-term relational ties with a

brand’s consumers.

The self-brand connection is introduced as another outcome in the nomological net-

work of celebrity endorsement effects. A desire to maintain and enhance one’s self-con-

cept is considered as a core consumer motivation (Sirgy 1982). The present study

supports the notion that celebrities are repositories of cultural meaning, aspects of which

are used by consumers to construct a desired sense of self (Holt 2002). The results of the

present study reveal that endorser credibility directly impacts consumers’ self-brand con-

nection. A theoretical implication of this finding is that endorsement effects have more of

a multi-faceted impact than previously envisaged. Credible celebrity endorsers are thus

likely to engage the brand as part of long-term self-image development. A practical impli-

cation of the finding is that brand managers can consider using celebrities not only to

enhance or expand brand awareness, but to engage with consumers on a much deeper

self-concept level. A practitioner need for facilitating deeper connections with consumers

has been recently highlighted in the sponsorship/endorsement literature (Wang and

Muehling 2012; Zdravkovic and Till 2012). Consumer beliefs and perceptions that reso-

nate in relation to the self-concept are thus likely to be more enduring than simpler aware-

ness effects. Additionally, the results support the logic that celebrities can potentially play

498 A. Dwivedi et al.

a more meaningful role in brand (re)positioning that is based on consumers’ aspirational

values. A case in point is that of the US automobile brand Lincoln that is attempting a

comeback in the domestic luxury auto sector (Black 2008). Lincoln’s use of the celebrity

rap artist Common for the brand’s Navigator model is designed to attain deep connections

(engagement) with the target market. A clear implication for brand managers is that

advertisements featuring celebrity endorsers must communicate deeper self-concept-

related benefits, possibly through the use of a transformational ad executional style and

symbolic visuals. However, choice of a celebrity endorser per se is also of critical impor-

tance. Practitioners suggest that celebrity-brand fit is a vital ingredient to attaining favor-

able outcomes and are likely to materialize only if credible celebrities endorse a brand

(Pringle 2012; Thompson 2010). In the present study, although celebrity-brand fit was

specified as a control variable, its positive impact on relationship quality is consistent

with the literature on the effects of endorser-brand fit (e.g., Misra and Beatty 1990).

Furthermore, self-brand connection was hypothesized as a partial mediator of the

effect of endorser credibility on relationship quality, which is empirically supported in

the analysis. Theoretically, the partial mediation via self-brand connection supports

McCracken’s (1989) cultural meaning transfer mechanisms that underline endorsement

effects. Specifically, the study supports McCracken’s logic of transference (through a

celebrity with mass appeal) of culturally constituted shared meaning into a more individu-

alized meaning for consumers in the form of self-brand connection. Thus, culturally

shared meaning created through credible endorsers not only directly leads to improve-

ment in consumer perception of brand relationship quality, but also indirectly via self-

brand connection � that is the shared meaning first translates into individualized meaning

for consumers, which subsequently influences relationship quality. From a managerial

perspective, the partial mediation result implies that investments into recruitment of cred-

ible celebrity endorsers entail a multiple return on investment. In the first instance, con-

sumer experience enhanced relationship quality with the endorsed brand because of the

credible reference group associations. Second, consumers ascribe a highly individualized

meaning to the brand which further impacts relationship quality.

Limitations and future research

The findings of the current study should be accepted in light of its limitations. The current

study is conducted using Indian consumer responses to local celebrities, thereby limiting

the scope of our study’s external validity. However, given the global similarity of Genera-

tion Y (youth) segment’s consumption of the celebrity culture, this limitation is not likely

to seriously impact the external validity of the findings. Future investigations should,

nonetheless, consider replicating the research model internationally in order to achieve

greater generalizability. The use of convenience sampling represents potential limitation

of the study though the demographic profile of the respondents closely resembled those

of the target population. A limitation pertains to the potential influence of self-generated

validity, which arises when a respondent’s answer to a survey question is influenced by

the preceding questions (Feldman and Lynch 1988), likely influencing the correlation

estimates. However, self-generated validity is not likely to be a serious threat as critical

questions were separated by sub-headings and page-breaks in the questionnaire. More

importantly, items measuring relationship quality and self-brand connection were listed

on separate pages. Future research may consider using multiple methods for measuring

antecedent and outcome variables to further mitigate self-generated validity concerns.

Furthermore, the current study investigates the hypothesized relationships at a given point

International Journal of Advertising 499

in time, thus limited with regard to capturing a change in perceptions. Consumer percep-

tions of endorser credibility may change over time as a result of a celebrity’s behavior in

personal (but well documented) life. Future research should consider examining the

research model using a longitudinal design to model such changes in consumer percep-

tions of relationship quality over the course of a celebrity’s involvement as an endorser of

a brand. Finally, the present study investigates direct effects, some of which are likely

moderated by consumer behavioral variables. For instance, consumer perception of the

level of a celebrity’s social connectedness (i.e., high versus low) may moderate the

impact of endorser credibility in relationship quality. Celebrities that are perceived as

being highly involved in the social media circuit may facilitate a greater transfer of aspi-

rational values from the celebrity to the endorsed brand, and vice versa.

Conclusion

The current study shows that the credibility of a celebrity endorser may play a crucial role

in strengthening the quality of the relationship between the consumer and the brand.

We show that this link occurs directly and through the consumers’ incorporation of the

brand’s meaning into their self-concept. These results enhance our understanding of the

role of celebrity endorsers and the mechanisms by which they effectively promote a

brand.

Acknowledgement

The authors would like to thank Dr. D. N. Dwivedi, Prof. N. K. Kakkkar and other academic staffmembers at the Maharaja Institute of Management Sciences, New Delhi, India for their assistanceduring the data collection phase in India.

Disclosure statement

No potential conflict of interest was reported by the authors.

References

Ablett, J., A. Baijal, E. Beinhocker, A. Bose, D. Farrell, U. Gersch, E. Greenberg, S. Gupta, and S.Gupta. 2007. The ‘Bird of Gold’: The rise of India’s consumer market. Report by McKinseyGlobal Institute. http://www.mckinsey.com/Insights/MGI/Research/Asia/The_bird_of_gold.

Alan, C.T., S. Fournier, and F. Miller. 2008. Brands and their meaning makers. In Handbook of con-sumer psychology. 1st ed., ed., C.P. Haugtvedt, P.M. Herr, and F.R. Kardes, 781�822. NewYork: Taylor and Francis.

Albert, N., D. Merunka, and P. Valette-Florence. 2013. Brand passion: Antecedents and consequen-ces. Journal of Business Research 66, no. 7: 904�9.

Amos, C., G. Holmes, and D. Strutton. 2008. Exploring the relationship between celebrity endorsereffects and advertising effectiveness. International Journal of Advertising 27, no. 2: 209�34.

Anderson, E., and V. Mittal. 2000. Strengthening the satisfaction�profit chain. Journal of ServiceResearch 3, no. 2: 107�20.

Arantola, H. 2002. Consumer bonding: A conceptual exploration. Journal of Relationship Market-ing 1, no. 2: 93�107.

Aurier, P. and G. N’Goala. 2010. The differing and mediating roles of trust and relationship com-mitment in service relationship maintenance and development. Journal of the Academy of Mar-keting Science 38, no. 3: 303�25.

Bagozzi, R.P., and Y. Yi. 2012. Specification, evaluation, and interpretation of structural equationmodels. Journal of the Academy of Marketing Science 40, no. 1: 8�34.

500 A. Dwivedi et al.

Bearden, W.O., and M.J. Etzel. 1982. Reference group influence on product and brand purchasedecision. Journal of Consumer Research 9, no. 2: 183�94.

Bearden, W.O., R.G. Netemeyer, and J.E. Teel. 1989. Measurement of consumer susceptibility tointerpersonal influence. Journal of Consumer Research 15, no. 4: 473�81.

Belk, R.W. 1988. Possessions and the extended self. Journal of Consumer Research 15, no. 2:139�68.

Black, I. 2008. Adjusting the consumer crosshairs: The use of celebrities in automotive marketing.Automotive marketing reports. World Advertising Research Centre Database. www.warc.com.

Bridges, S., K.L. Keller, and S. Sood. 2000. Communication strategies for brand extensions:Enhancing perceived fit by establishing explanatory links. Journal of Advertising 29, no. 4:1�11.

Brodie, R.J., A. Ilic, B. Juric, and L. Hollebeek. 2013. Consumer engagement in a virtual brandcommunity: An exploratory analysis. Journal of Business Research 66, no. 1: 105�14.

Byrne, B.M. 2010. Structural equation modeling: Basic concepts, applications and programming.New York: Routledge.

Chahal, M. 2012. Does your celebrity have the $-factor? Marketing Week, 16�18. http://search.proquest.com/docview/1082293565?accountidD10344.

Chaturvedi, P. 2011. India’s ‘big three’ mobile brands and life beyond the price war (WARC exclu-sive), 1�11. World Advertising Research Centre Online Database. www.warc.com.

Cohen, J.B., and K. Basu. 1987. Alternative models of categorization: Toward a contingent process-ing framework. Journal of Consumer Research 13, no. 4: 455�72.

Cutright, K.M., A. Samper, and G.J. Fitzsimons. 2013. We are what we buy. In The Routledge com-panion to identity and consumption (Chapter 9). 1st ed., ed. A.R. Ayalla and R.W. Belk, 91�8.New York: Routledge.

Dolich, I.J. 1969. Congruence relationships between self-images and product brands. Journal ofMarketing Research 6, no. 1: 80�4.

Eisend, M., and T. Langner. 2010. Immediate and delayed advertising effects of celebrity endorsers’attractiveness and expertise. International Journal of Advertising 29, no. 4: 527�46.

Elliott, R., and K. Wattanasuwan. 1998. Brands as symbolic resources for the construction of iden-tity. International Journal of Advertising 19, no. 2: 131�44.

Enders, C.K., and D.L. Bandalos. 2001. The relative performance of full information maximumlikelihood estimation for missing data in structural equation models. Structural EquationModeling 8, no. 3: 430�57.

Erdogan, B.Z. 1999. Celebrity endorsement: A literature review. Journal of Marketing Management15, no. 4: 291�314.

Escalas, J.E. 2004. Narrative processing: Building consumer connections to brands. Journal of Con-sumer Psychology 14, no. 1/2: 168�80.

Escalas, J.E., and J.R. Bettman. 2003. You are what they eat: The influence of reference groups onconsumers’ connections to brands. Journal of Consumer Psychology 13, no. 3: 339�48.

Escalas, J.E., and J.R. Bettman. 2005. Self-construal, reference groups, and brand meaning. Journalof Consumer Research 32, no. 3: 378�89.

Escalas, J.E., and J.R. Bettman. 2013. The brand is “Me”: Exploring the effect of self-brand connec-tions on processing brand information as self-information. In The Routledge companion to iden-tity and consumption (Chapter 37). 1st ed., ed. A.R. Ayalla and R.W. Belk, 366�74. New York:Routledge.

Euromonitor. 2007. Top 10 consumer trends in India, 1�7. Business Source Complete, EbscoHostOnline Database.

Evalueserve. 2009. Uptake of 3G services in India, 1�8. Business Source Complete, EbscoHostOnline Database.

Feldman, J.M., and J.G. Lynch Jr. 1988. Self-generated validity and other effects of measurementon belief, attitude, intention and behaviour. Journal of Applied Psychology 73, no. 3: 421�35.

Fornell, C., and D. Larcker. 1981. Evaluating structural equation models with unobservable varia-bles and measurement error. Journal of Marketing Research 18, no. 1: 39�50.

Fournier, S. 1998. Consumers and their brands: Developing relationship theory in consumerresearch. Journal of Consumer Research 24, no. 4: 343�73.

Ganesan, S. 1994. Determinants of long-term orientation in buyer�seller relationships Journal ofMarketing 58, no. 2: 1�19.

Garretson, J.A., R.G. Folse, and S. Burton. 2012. How the personality traits of sincerity, excitement,and competence help to build equity. Journal of Advertising 41, no. 1: 17�32.

International Journal of Advertising 501

Goodman, L.L., and P.A. Dion. 2001. The determinants of commitment in the distributor � manu-facturer relationship. Industrial Marketing Management 30, no. 3: 287�300.

Gr�egoire, Y., T.M. Tripp, and R. Legoux. 2009. When customer love turns into lasting hate: Theeffects of relationship strength and time on customer revenge and avoidance. Journal of Mar-keting 73, no. 6: 18�32.

Gr€onroos, C. 2006. Adopting service logic for marketing.Marketing Theory 6, no. 3: 317�33.Gustafsson, A., M.D. Johnson, and I. Roos. 2005. The effects of customer satisfaction, relationship

commitment dimensions, and triggers on customer retention. Journal of Marketing 69, no. 4:210�8.

Hair, J.F., W.C. Black, B.J. Babin, and R.E. Anderson. 2009. Multivariate data analysis. NewJersey: Prentice Hall.

Hennig-Thurau, T., K.P. Gwinner, and D.D. Gremler. 2002. Understanding relationship marketingoutcomes: An integration of relational benefits and relationship quality. Journal of ServiceResearch 4, no. 3: 230�47.

Holt, D. 2002. Why do brands cause trouble? A dialectical theory of consumer culture and branding.Journal of Consumer Research 29, no. 1: 70�90.

Hovland, C.I., I.L. Janis, and H.H. Kelley. 1953. Communication and persuasion: Psychologicalstudies of opinion change. New Haven, CT: Yale University Press.

Indiantelevision.com. 2009. Snapshot on celebrity endorsement on TV during H1’09. http://www.indiantelevision.com/tamadex/y2k9/aug/tam32.php.

Jarvis, C.B., S.B. MacKenzie, and P.M. Podsakoff. 2003. A critical review of construct indicatorsand measurement model misspecification in marketing and consumer research. Journal of Con-sumer Research 30, no. 2: 199�218.

Kahle, L.R., and P.M. Homer. 1985. Physical attractiveness of the celebrity endorser: A social adap-tation perspective. Journal of Consumer Research 11, no. 4: 954�61.

Keller, K.L. 2013. Strategic brand management: Building, measuring and managing brand equity(Chapter 7). 4th ed. Essex, UK: Pearson Education.

Kelting, K., and D.H. Rice. 2013. Should we hire David Beckham to endorse our brand? Contextualinterference and consumer memory for brands in a celebrity’s endorsement portfolio. Psychol-ogy and Marketing 30, no. 7: 602�13.

Kim, J.W., J.C. Chonnam, W. Qualls, and K. Han. 2008. It takes a marketplace community to raisebrand commitment: The role of online communities. Journal of Marketing Management 24, no.3/4: 409�31.

Li, Z.G., and R.P. Dant. 1997. An exploratory study of exclusive dealing in channel relationships.Journal of the Academy of Marketing Science 25, no. 3: 201�13.

MacInnis, D.J., C.W. Park, and J.R. Priester. 2009. Introduction: Why brand relationships? In Hand-book of brand relationships. 1st ed., ed. D.J., MacInnis, C.W. Park, and J.R. Priester, ix�xxi.New York: ME Sharpe.

Markus, H. 1977. Self-schemata and the processing of information about the self. Journal of Per-sonality and Social Psychology 35, no. 2: 63�78.

McCracken, G. 1989. Who is the celebrity endorser? Cultural foundations of the endorsement pro-cess. Journal of Consumer Research 16, no. 3: 310�21.

Misra, S., and S.E. Beatty. 1990. Celebrity spokesperson and brand congruence. Journal of BusinessResearch 21, no. 2: 159�73.

Moore, D.J., and P.M. Homer. 2008. Self-brand connections: The role of attitude strength and auto-biographical memory primes. Journal of Business Research 61, no. 7: 707�14.

Murphy, G.L., and D.L. Medin. 1985. The role of theories in conceptual coherence. PsychologicalReview 92, no. 3: 289�316.

Ohanian, R. 1990. Construction and validation of a scale to measure celebrity endorsers’ perceivedexpertise, trustworthiness, and attractiveness. Journal of Advertising 19, no. 3: 39�52.

Perloff, R.M. 2010. The dynamics of persuasion: Communication and attitudes in the 21st century.New York: Taylor & Francis.

Podsakoff, P.M., S.B. MacKenzie, J.Y. Lee, and N.P. Podsakoff. 2003. Common method biases inbehavioral research: A critical review of the literature and recommended remedies. Journal ofApplied Psychology 88, no. 5: 879�903.

Pringle, H. 2012. How to use celebrity in brand building. WARC best practice. World AdvertisingResearch Centre Database. www.warc.com

Radford, S.K., and P.H. Bloch. 2013. Consumers’ online responses to the death of a celebrity.Marketing Letters 24, no. 1: 43�55.

502 A. Dwivedi et al.

Richins, M.L. 1994. Valuing things: The public and private meanings of possessions. Journal ofConsumer Research 21, no. 3: 504�21.

Rossiter, J.R., and L. Percy. 1987. Advertising and promotion management (Chapter 11). NewYork: McGraw Hill International.

Sbarbaro, S., J. Van der Bergh, E. Veris, and T. De Ruyck. 2011. We got a crush on you(th)! Involv-ing influential Gen Y’ers from 15 global cities to learn why something is cool. In ESOMARresearch paper, collection: qualitative. World Advertising Research Centre Database. www.warc.com (accessed November 13, 2011).

Schiffman, L., A. O’Cass, A. Paladino, S. D’Alessandro, and D. Bednall. 2011. Consumer behavior.5th ed. Frenchs Forest, NSW: Pearson Australia.

Sheth, J.N., and Parvatiyar, A. 1995. Relationship marketing in consumer markets: Antecedents andconsequences. Journal of the Academy of Marketing Science 23, no. 4: 225�71.

Silvera, D.H., and B. Austad. 2004. Factors predicting the effectiveness of celebrity endorsementadvertisements. European Journal of Marketing 38, no. 11/12: 1509�26.

Sirgy, M.J. 1982. Self-concept in consumer behavior: A critical review. Journal of ConsumerResearch 9, no. 3: 287�300.

Speck, P.S., D.W. Schumann, and C. Thompson. 1988. Celebrity endorsements � scripts, schemaand roles: Theoretical framework and preliminary tests. Advances in Consumer Research 15,no 1: 69�76.

Sprott, D., S. Czellar, and E. Spangenberg. 2009. The importance of a general measure of brandengagement on market behavior: Development and validation of a scale. Journal of Marketing46, no. 1: 92�104.

Spry, A., R. Pappu, and T.B. Cornwell. 2011. Credibility endorsement, brand credibility and brandequity. European Journal of Marketing 45, no. 6: 882�909.

Stafford, M.R., T.F. Stafford, and E. Day. 2002. A contingency approach: The effects of spokesper-son type and service type on service advertising perceptions. Journal of Advertising 31, no. 2:17�35.

Telecom Regulatory Authority of India. 2011. Information note to the press (press releaseNo. 41 /2011), July 7, New Delhi. www.trai.gov.in

Telecomcircle.com. 2012. Snapshot: India telecom market. http://www.telecomcircle.com/2009/06/india-telecom/.

Thompson, D. 2010. Celebrities must be relevant to increase brand engagement. Admap report.World Advertising Research Centre Database. www.warc.com.

Thompson, M. 2006. Human brands: Investigating antecedents to consumers’ strong attachments tocelebrities. Journal of Marketing 70, no. 3: 104�19.

Till, B.D. 1998. Using celebrity endorsers effectively: Lessons from associative learning. Journal ofProduct and Brand Management 7, no. 5: 400�9.

Till, B.D., and M. Busler. 2000. The match-up hypothesis: Physical attractiveness, expertise, andthe role of fit on brand attitude, purchase intent and brand beliefs. Journal of Advertising 29,no. 3: 1�13.

Van den Bergh, J., and M. Behrer. 2011. How cool brand stay hot: Branding to generation Y. 2nded. London: Kogan Page.

Van Kuilenburg, P., M.D.T. De Jong, and T.J.L. Van Rompay. 2011. That was funny, but what wasthe brand again? International Journal of Advertising 30, no. 5: 795�814.

Wang, J., and J.M.M. Bloemer. 2008. The impact of value congruence on consumer-service brandrelationships. Journal of Service Research 11, no. 2: 161�78.

Wang, A., and D. Muehling. 2012. The moderating influence of brand status and source confirma-tion on third-party endorsement effects in advertising. International Journal of Advertising 31,no. 3: 605�22.

Yelin, H., and K. Kinnear. 2012, September. Celebrity endorsement: Socialise the endorsement.Admap report. World Advertising Research Centre Database. www.warc.com.

Yoon, D., S.M. Choi, and D. Sohn. 2008. Building customer relationships in an electronic age: Therole of interactivity of e-commerce web sites. Psychology and Marketing 25, no. 7: 602�18.

Zdravkovic, S., and B.D. Till. 2012. Enhancing brand image via sponsorship: Strength of associa-tion effects. International Journal of Advertising 31, no. 1: 113�32.

International Journal of Advertising 503