University of Southern California Dissertations and Theses

Powerful brand influentials: conceptualization, measurement, and distinctiveness of a brand’s influential consumers

(USC Thesis Other)

Powerful brand influentials: conceptualization, measurement, and distinctiveness of a brand’s influential consumers

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Running head: POWERFUL BRAND INFLUENTIALS 1

Powerful Brand Influentials: Conceptualization, Measurement, and Distinctiveness of a Brand’s
Influential Consumers

by

Zhong Wan

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BUSINESS ADMINISTRATION)

December 2012

Copyright 2012 Zhong Wan
POWERFUL BRAND INFLUENTIALS 2

Dedication
To my beloved daughter Evelyn Wan Kim, who is the most spirited and beautiful child in the
whole wide world.
Please be independent, passionate, and headstrong in whatever you decide to pursue in your life.

POWERFUL BRAND INFLUENTIALS 3

Acknowledgments
I am eternally grateful for several people who have made it possible for the completion of
this dissertation and who have supported me through the journey of my graduate years.
To my advisor, Deborah MacInnis: you have not only been an incomparable mentor but
also an irreplaceable part of my life over all these years. You have shown me the essence of a
true researcher, a brilliant teacher, a loving person, and an intellectual friend. You have given
endless support and love to me and my family through the ups and downs. You have always
believed in me and encouraged me to conquer the difficult times and embrace the wonderfulness
of life. I cherish every bit of your guidance on research, every moment of our interactions, and
every piece of our friendship. You have been and will always be on my mind and in my heart.
To my husband, Sang In: you are my best friend and a witty one. I thank you for loving
me, loving our daughter, and standing by us regardless of what happens in life. I thank you for
always believing in me, enduring me, and enabling me to focus on my endeavor throughout the
eight years we have been together. I thank you for being the one who I could share my world
with and the one who I could always depend on learning the truth about the world. Most of all, I
am grateful that you are the father of my daughter and that you scarified your own dreams to
give undivided attention, care, and love to our daughter.
My baby girl, Evy, you are my heart. You were born during the journey of this
dissertation. Thank you for making me see new meanings of life and for making me feel
complete. Thank you for all the joys and tears that you have brought to my life: for reading with
me all the boring papers when you were growing in my belly, for tearing my papers into pieces
and leaving me with the rest to read, and for tolerating long hours of my work away and still
greeting me with hugs and kisses.
POWERFUL BRAND INFLUENTIALS 4

To my mom and dad: you taught me to be an independent thinker and woman. Thank you
for your unconditional love, support, comfort, and protection. Because of you, I have always felt
safe, confident, and loved in this world. To my sister: I thank you for being my company and
buddy. I thank you for being proud of me in fulfilling my dreams. I thank you for being an
individual with heart pure as snow.
Last but not least, to my committee members, C.W. Park, Gulden Ulkumen, and Michael
Cody. Thank C.W. for all the wonderful time in the seminar, all the charisma, all the spirited and
passionate conversions about research topics, and all the philosophical, caring and loving words
that never fail to put a big smile on my face. Thank Gulden for giving me the most valuable
experience in teaching assistance and for being so sweat in helping me through the dissertation
process. Thank Michael for being tremendously supportive and directive in shaping and
broadening the horizon of my dissertation and for always being kind and caring in the entire
process.

POWERFUL BRAND INFLUENTIALS 5

Table of Contents
Dedication 2

Acknowledgments 3

List of Tables 9

List of Figures 12

Abstract 13

Chapter 1: Introduction 15

Chapter 2: Evidence for the Existence of a Brand’s Influential Consumers 18

Chapter 3: Conceptualization of the Powerful Brand Influential 20
3.1. Conceptual Properties of the Influential Consumer Construct 20
3.2. Conceptual Properties of the Powerful Brand Influential Construct 23
3.3. Conceptual Definition of the Powerful Brand Influential 29

Chapter 4: Differentiating the Powerful Brand Influential from
Market Maven, Opinion Leader, Brand Attachment,
and Brand Ambassador 30
4.1. The Powerful Brand Influential is Related to Market Maven,
Opinion Leader, Brand Attachment, and Brand Ambassador 30
4.2. The Powerful Brand Influential is Distinct from Market Maven,
Opinion Leader, Brand Attachment, and Brand Ambassador 33
4.3. The Powerful Brand Influential is Distinct on the Engagement
Level of Brand Word-of-Mouth and Brand Advocacy Behavior 34
4.4. The Powerful Brand Influential is Distinct on the Level of
Influence on others’ Brand Decisions 42
POWERFUL BRAND INFLUENTIALS 6

4.5. The Powerful Brand Influential is Distinct on the Level of
Intrinsic Pleasure Derived from Brand Word-of-Mouth 47
4.6. Further Differentiating the Powerful Brand Influential from
Brand Attachment 49

Chapter 5: Differentiating the Powerful Brand Influential from
Brand Attitude Strength and Brand Satisfaction 66
5.1. The Powerful Brand Influential is Related to Brand
Attitude Strength and Brand Satisfaction 66
5.2. Differentiating the Powerful Brand Influential from Brand
Attitude Strength and Brand Satisfaction on Brand
Behavioral Outcomes 67

Chapter 6: Overview of Studies 1 – 4 74

Chapter 7: Study 1: Development and Purification of the
Powerful Brand Influential Scale (PBIS) 75
7.1. Item Generation 75
7.2. Objectives and Method 76
7.3. Measurement Results of the PBIS 76

Chapter 8: Study 2: The Powerful Brand Influential is Distinct from
Market Maven, Opinion Leader, Brand Attachment,
and Brand Ambassador 82
8.1. Objective and Method 82
8.2. Measurement Results of the PBIS 82
8.3. Testing Convergent and Discriminant Validity 84
8.4. Testing Predictive Validity and Propositions 97

Chapter 9: Study 3: The Powerful Brand Influential is Distinct from
Market Maven, Opinion Leader, Brand Attachment,
Brand Attitude Strength, and Brand Satisfaction 115
9.1. Objective and Method 115
9.2. Measurement Results of the PBIS 116
POWERFUL BRAND INFLUENTIALS 7

9.3. Testing Convergent Validity and Discriminant Validity of
the PBIS in Relation to the Market Maven, Opinion
Leadership, and Brand Attachment Constructs (P1 – P2) 117
9.4. Testing Convergent Validity and Discriminant Validity of
the PBIS in Relation to Brand Attitude Strength and
Brand Satisfaction 123
9.5. Testing Predictive Validity of the PBIS in Relation to Market
Maven, Opinion Leader and Brand Attachment (P4a – 4d) 129
9.6. Comparing Predictive Validity of the PBIS and Brand Attachment
for Behavioral and Psychological Outcomes (P4d, P7 and P8) 131
9.7. Comparing the PBIS and Brand Attachment on Emotional
Expressiveness 138
9.8. Testing Predictive Validity of the PBIS in Relation to
Brand-Level Constructs for Behavioral Outcomes
(P8, P9, P13, P14, and P15) 139

Chapter 10: Study 4: Differentiating the Powerful Brand Influential from
Brand Attachment in the Context of a Positive Brand Event 144
10.1. Objective and Methodology 144
10.2. Measurement Results of the PBIS 155
10.3. Testing Known Group Validity 156
10.4. Testing Convergent Validity and Discriminant Validity of
the PBIS in Relation to Brand Attachment (P1d and P2d) 156
10.5. Testing Predictive Validity of the PBIS vs. Brand Attachment
for Positive Self-Conscious Emotions and Challenge
Emotions (P10a – 10b) 158
10.6. Testing Predictive Validity of the PBIS vs. Brand Attachment for
the Level of Actual Word-of-Mouth Engagement and Types of
Word-of-Mouth Following a Positive Brand Event (P11a – 11c) 161

Chapter 11: General Discussion 168
11.1. Contribution and Implications 168
11.2. Summary of Findings 169
11.3. Limitation and Further Research 174

References 178
POWERFUL BRAND INFLUENTIALS 8

Appendices 197
Appendix A. (Chapter 4) Brand Attachment and Behaviors from
Park, MacInnis, and Priester (2006), “Beyond Attitudes:
Attachment and Consumer Behavior,” Seoul Journal of
Business, 12 (December), p. 19 197
Appendix B. (Chapter 8) Measures of Constructs Related to
the Powerful Brand Influential (Study 2 – Study 3) 198
Appendix C. (Chapter 9) Dependent Variable Measures (Study 3) 199
Appendix D. (Chapter 10) Dependent Variable Measures (Study 4) 200
POWERFUL BRAND INFLUENTIALS 9

List of Tables
TABLE 7-1. The Initial 13-Item Pool of the PBIS and Corresponding
Defining Components 77

TABLE 7-2. The Final PBIS and Corresponding Defining Components 79

TABLE 7-3. Correlations of the PBIS with Criterion Variables in Pilot Study 80

TABLE 7-4. Exploratory Factor Analysis Factor Loadings of
the PBIS (Study 1 – 4) 81

TABLE 8-1. Confirmatory Factor Analysis Results of the PBIS (Study 2 – 4) 84

TABLE 8-2. Exploratory Factor Analysis Results (Oblique Rotation) with
the PBIS and Measures of Market Maven, Opinion Leader, Brand
Attachment, and Brand Ambassador 86

TABLE 8-3. Correlations based on TABLE 8-2. Exploratory Factor
Analysis Results 87

TABLE 8-4. Confirmatory Factor Analysis Results of Model 1
– Five Correlated Factors on the PBIS and Measures of 4
Related Constructs 89

TABLE 8-5. Interfactor (Φ) Correlations of Model 1 – Correlations between
5 Factors based on TABLE 8-4. Confirmatory Factor Analysis Results 90

TABLE 8-6. Results for Comparing Models of Confirmatory Factor
Analysis on the PBIS and Measures of 4 Related Constructs 92

POWERFUL BRAND INFLUENTIALS 10

TABLE 8-7a. Hierarchical Regression Results Predicting the Engagement
Level of Brand Word-of-Mouth (indicator 1 – the Number of
Receivers): Comparing Predictive Ability of the PBIS, Market
Maven, Opinion Leader, Brand Attachment, and Brand Ambassador 102

TABLE 8-7b. Hierarchical Regression Results Predicting the Engagement
Level of Brand Word-of-Mouth (indicator 2 – the Selectiveness of
Brand Word-of-Mouth Receivers): Comparing Predictive Ability of
the PBIS, Market Maven, Opinion Leader, Brand Attachment, and
Brand Ambassador 104

TABLE 8-8. Hierarchical Regression Results Predicting Brand Advocacy
Behaviors: Comparing Predictive Ability of the PBIS, Market
Maven, Opinion Leader, Brand Attachment, and Brand Ambassador 107

TABLE 8-9. Hierarchical Regression Results Predicting the Tendency to
Influence Others’ Brand Decisions: Comparing Predictive Ability
of the PBIS, Market Maven, Opinion Leader, Brand Attachment,
and Brand Ambassador 110

TABLE 8-10. Hierarchical Regression Results Predicting the Level of
Intrinsic Pleasure Derived from Brand Word-of-Mouth: Comparing
Predictive Ability of the PBIS, Market Maven, Opinion Leader,
Brand Attachment, and Brand Ambassador 111

TABLE 9-1. Confirmatory Factor Analysis Results of Model 1 – Four
Correlated Factors on the PBIS and Measures of 3 Related Constructs 119

TABLE 9-2. Interfactor (Φ) Correlations of Model 1 – Correlations
between 4 Factors based on Table 9-1. Confirmatory Factor
Analysis Results 120

TABLE 9-3. Results for Comparing Models of Confirmatory Factor
Analysis on the PBIS and Measures of 3 Related Constructs 122

TABLE 9-4. Confirmatory Factor Analysis Results of Model 1 – Four
Correlated Factors including the PBIS and Brand Attachment,
Satisfaction, and Attitude Strength 127
POWERFUL BRAND INFLUENTIALS 11

TABLE 9-5. Interfactor (Φ) Correlations of Brand-Level Construct CFA
– Correlation between the PBIS, Brand Attachment, Brand
Satisfaction, and Brand Attitude Strength 128

TABLE 10-1. Pre-Election: Confirmatory Factor Analysis Results of Model 1
– the PBIS and Brand Attachment as Two Correlated Factors 157

TABLE 10-2. Post-Election: Regression Results Comparing Predictive
Ability of the PBIS and Brand Attachment for Behavioral Responses
to a Positive Brand Event 162

TABLE 10-3. Post-Election: Regression Results Comparing Predictive
Ability of the PBIS and Brand Attachment for Types of
Word-of-Mouth after a Positive Brand Event 164

POWERFUL BRAND INFLUENTIALS 12

List of Figures
FIGURE 9-1. Simultaneous Comparison of Predictive Ability of the
PBIS and Brand Attachment, Market Maven, and Opinion
Leader for Brand Advocacy Behaviors 131

FIGURE 9-2. Predictive Ability of the PBIS and Brand Attachment for
Behavior Outcomes
Panel A: Predictive ability of the PBIS for behavior outcomes 134

FIGURE 9-2. Predictive Ability of the PBIS and Brand Attachment for
Behavior Outcomes
Panel B: Predictive ability of Brand Attachment for
behavior outcomes 135

FIGURE 9-2. Predictive Ability of the PBIS and Brand Attachment for
Behavior Outcomes
Panel C: Simultaneous comparison of predictive ability the
PBIS and Brand Attachment for behavior outcomes 137

FIGURE 9-3. Simultaneous Comparison of Predictive Ability of the PBIS,
Brand Attachment, Brand Satisfaction, and Brand Attitude Strength
for Behavior Outcomes 141

FIGURE 10-1. Comparison of Predictive Ability of the PBIS and Brand
Attachment for Positive Self-Conscious Emotions and Challenge
Emotions
Panel A: Predictive Ability for Pre-Election Emotional Responses 159

FIGURE 10-1. Comparison of Predictive Ability of the PBIS and Brand
Attachment for Positive Self-Conscious Emotions and Challenge
Emotions
Panel B: Predictive Ability for Post-Election Emotional Responses 161

POWERFUL BRAND INFLUENTIALS 13

Abstract
This article focuses on a unique and important consumer construct, a brand’s influential
consumers, namely “the powerful brand influential”. In the marketplace, influential consumers
designate those who exert exceptional influence on other consumers’ decision making. Extant
literature has established opinion leaders as influential consumers of a specific product category
and market mavens as influential consumers of the general marketplace. Although the existence
and power of a brand’s influential consumers are well-documented and the designations of these
consumers are proliferated, we have limited understanding about these consumers. The
conceptualization of a brand’s influential consumers, the distinctiveness of the construct from
existing concepts, and the measurement for the construct are largely missing.
Drawing from various domains of literature, this article provides a systematic
conceptualization of a brand’s influential consumers that reflects four defining components.
Corresponding to these conceptual properties, the powerful brand influential is defined as
consumers (1) who are strongly attached to a given brand and (2) who have high tendency to
influence other consumers’ brand decisions through (3) a high engagement of (4) passionate and
emotionally evocative word-of-mouth about the brand. This article then details the differentiation
of the new construct from six conceptually related constructs and posits a set of corresponding
propositions. Study 1 develops a 6-item measurement, the Powerful Brand Influential Scale
(PBIS), to represent each of the four conceptual properties. Study 2, 3, and 4 provide converging
support that the powerful brand influential construct is theoretically and empirically
distinguishable from the six constructs including market maven, opinion leader, brand
attachment, brand ambassador, brand attitude strength, and brand satisfaction. Our proposition
testing substantiates the distinctiveness and added value of the powerful brand influential
POWERFUL BRAND INFLUENTIALS 14

construct – demonstrating that this new construct better predicts many key brand outcomes
including brand word-of-mouth, intrinsic pleasure derived from brand word-of-mouth, brand
advocacy behavior, tendency to influence others’ brand decision, brand loyal behavior, actual
brand purchase as well as desirable emotional and word-of-mouth responses toward a positive
brand event than the related constructs examined. Moreover, using different samples and brands
from diverse product categories, the four studies establish the reliability, robustness, known
group, convergent, discriminant, and predictive validity of the PBIS. The article concludes by
discussing theoretical and managerial contribution, findings, and implications of the powerful
brand influential construct and the PBIS along with limitations and areas of future research.

POWERFUL BRAND INFLUENTIALS 15

Powerful Brand Influentials: Conceptualization, Measurement, and Distinctiveness of a Brand’s
Influential Consumers
Chapter 1: Introduction
Marketers and academics have been chasing the most valuable consumer groups for
decades. In the marketplace, all consumers are not created equal as some exert disproportionate
influence on other consumers’ decision making. They are influential consumers (Clark and
Goldsmith 2005; Feick and Price 1987; Keller and Berry 2003; Price, Feick, and Higie 1987;
Weimann 1994). They influence other consumers’ purchases through the medium of word-of-
mouth (Feick and Price 1987; King and Summers 1970; Summers 1971). Although the power of
word-of-mouth has long been established, in the new era of social media, it is amplified to a
stunning level. A consumer may tell ten people about his or her product experience in the older
times while today, he or she has the tools to tell 10 million (Gillin 2007). Influential consumers
as the powerhouses of word-of-mouth, now harness a greater sphere of influence than they ever
did. Consequently, it is extremely critical for a company to identify and reach the influential
consumers of its brand(s). Moreover, it is enormously profitable for a company to engage
influential consumers as persuasive sources of its brand(s).
In spite of their importance and value to marketers, a brand’s influential consumers have
not been systematically explored. On one hand, the existence of consumers who influence others’
brand purchases through word-of-mouth has been documented by academic research in various
domains, such as research on fandom (Kozinets 2001; O’Guinn 1991), consumer loyalty,
commitment and devotion (Pimentel and Reynolds 2004; Reichheld 2003), brand community
(Muniz and O’Guinn 2001), consumer referral and word-of-mouth of brands (Godes and
Mayzlin 2004; Ryu and Feick 2007), and customer relationship management (Bhattacharya and
POWERFUL BRAND INFLUENTIALS 16

Sen 2003; Winer 2001). The evidence and impact of these consumers have been highlighted by
popular press authors (McConnell and Huba 2003; Rusticus 2005) and marketers (Rozanski,
Baum, and Wolfsen 1999; Yahoo! and comScore Networks 2006). The designations of these
consumers have been proliferating (e.g., brand advocate, brand evangelist, brand enthusiast,
brand zealot, brand ambassador).
On the other hand, both academic researchers and marketers have limited understanding
of what these consumers are like, whether they are distinct from previously identified influential
consumers (e.g., opinion leaders and market mavens), whether a brand’s influential consumer
constitutes a new marketing construct that is differentiated from existing brand constructs (e.g.,
brand attitude, brand satisfaction, brand attachment), and if so, how to effectively identify these
consumers and to measure the represented new construct.
To address these issues, we first review evidence of a brand’s influential consumers
named as the powerful brand influential
1
in this paper. We then provide a theoretical
conceptualization of the powerful brand influential construct. We do so by defining the key
components of being a powerful brand influential. We then conceptually differentiate the
powerful brand influential from two previously identified types of influential consumers (i.e.,
opinion leaders and market mavens) and from several related brand constructs (i.e., brand
attachment, brand ambassador, brand attitude strength, and brand satisfaction). Following the
conceptualization, we develop a scale (the Powerful Brand Influential Scale – the PBIS) to
measure the represented concept and to help identify a brand’s influential consumers. In four
subsequent studies, we assess the reliability, validity, and stability of PBIS across different

1
We designate a brand’s influential consumers as the powerful brand influential since the term explicitly denotes the nature of
these consumers. The term serves as an umbrella term for previous designations of a similar concept (i.e., brand advocate, brand
evangelist, brand enthusiast, brand zealot). However, the designation of brand ambassador denotes a somewhat different
concept from ours (as discussed later in the paper).
POWERFUL BRAND INFLUENTIALS 17

product categories. We also demonstrate that the PBIS is empirically discriminable from
measures of related constructs in the literature and hence documents the distinctness of the
powerful brand influential from the same.

POWERFUL BRAND INFLUENTIALS 18

Chapter 2: Evidence for the Existence of a Brand’s Influential Consumers
There exist consumers who are extraordinary in terms of the power they possess on
influencing other consumers’ decisions regarding a specific brand. They are interchangeably
called brand advocates, brand enthusiasts, brand zealots, and brand evangelists
2
in popular
press, marketers’ and academic research. All designate consumers who voluntarily and actively
promote a brand or company that they feel passionate about via word-of-mouth (McConnell and
Huba 2007; Rozanski et al. 1999; Rusticus 2005; also see Wikipedia for “Evangelism Marketing”
http://en.wikipedia.org/wiki/Evangelism_marketing). They are a brand’s influential consumers.
In this paper, they are called the powerful brand influential, a new designation to distinguish our
conceptualization and highlight the phenomenon.
The power and value of these consumers in the marketplace has been well documented.
In their book Creating Customer Evangelists: How Loyal Customers Become a Volunteer Sales
Force, McConnell and Huba (2003) conducted year-long case studies on companies that have
successfully harnessed the force of the company’s influential consumers. Their qualitative
account demonstrated that real-life brands’ influential consumers drive and sustain the success of
a company in good and bad times. A study conducted by Yahoo! and comScore Networks (2006)
on 2297 respondents has provided both qualitative and quantitative evidence on the influence of
these powerful consumers. The study shows that a brand’s influential consumers stand apart
from the rest of consumer population in affecting others’ purchases through word-of-mouth (both
online and offline). It also documents the increased sphere of influence possessed by today’s
influential consumers of a brand through the combined forces of word-of-mouth and social
media. Some compelling data on the value of these consumers to a company or brand include: (1)

2
The marketing term of “brand evangelist” is drawn from the religious term of “evangelist”: literally means “a
bringer of the glad tidings” and denotes the religious believers who roam the back ways of the world to spread the
word of their faith (McConnell and Huba 2007).
POWERFUL BRAND INFLUENTIALS 19

they are significantly more likely (at least twice as likely) to convert another consumer to
purchase the same brand than are non-brand’s influential; (2) their reach through online and
offline word-of-mouth across product categories is significantly greater than that of non-brand’s
influential (they reach 25% to 30% more consumers than do non-brand’s influential); and (3)
they are more likely to talk about positive than negative experiences (about 90% of them shared
something positive about a purchase they made).
A brand’s influential consumers may also drive brand growth. Previous studies have
identified a positive correlation between consumer advocacy and brand growth (Marsden,
Samson, and Upton 2005; Reichheld 2003). Reichheld (2003) conducted the seminal study on
the relationship between consumers who would recommend a company or brand (termed as
“promoters”) and the growth rate of the company or brand. The study shows a strong correlation
between the percentage of a company’s promoters and the company’s annual revenue growth. As
Reichheld (2003) suggests, one simple statistic, the percentage of a company’s promoters,
“seemed to explain the relative growth rates across the entire industry” and “the only path to
profitable growth may lie in a company’s ability to get its loyal customers to become, in effect,
its marketing department” (p. 51).

POWERFUL BRAND INFLUENTIALS 20

Chapter 3: Conceptualization of the Powerful Brand Influential
The powerful brand influential, is first of all an influential consumer (Clark and
Goldsmith 2005; Feick and Price 1987; Keller and Berry 2003; Price et al. 1987; Weimann 1994).
Such consumers possess a set of fundamental characteristics that make a consumer influential in
the marketplace. At the same time, they are not simply just influential consumers – they are
influential at a specific brand-level. They have features that make them influential regarding a
brand, features that are not captured by the concept of influential consumers. Correspondingly,
we take a two-step approach in conceptualizing the powerful brand influential. First, we
synthesize the key components of the influential consumer concept by understanding what
defining features several influential consumer groups previously identified in the literature share.
Second, we draw insights from popular press, academic and practitioner research to
conceptualize the defining features (in addition to those already captured by the influential
consumer concept) that constitute a brand-specific influential consumer construct.
3.1. Conceptual Properties of the Influential Consumer Construct
The concept of influential consumer has been explored in various research domains, such
as interpersonal communication
3
or word-of-mouth (Engel, Kegerreis, and Blackwell 1969;
Walsh, Gwinner, and Swanson 2004), interpersonal influence (Feick and Price 1987; Katz and
Lazarsfeld 1955; Weimann 1994), and the diffusion of innovations (Baumgarten 1975;
Hirschman and Adcock 1978; Midgley 1976; Turnbull and Meenaghan 1980). Approaching the
phenomenon with different perspectives, extant research has concentrated on identifying various
types of influential consumers. A unified conceptualization of what defines a consumer as an

3
In this paper, the term “interpersonal communication” specifically means “word-of-mouth”. By definition,
interpersonal communication incorporates all forms of “exchange of information between individuals” (King and
Summers 1970, p44). It consists of word-of-mouth as a type of verbal communications and other non-verbal
communications (e.g., facial expression, body languages, and actions). However, it is particularly the word-of-
mouth communication that has been investigated in the literature on the influential consumers.
POWERFUL BRAND INFLUENTIALS 21

influential consumer is largely missing in the literature. Therefore, we provide a formal
conceptualization of the influential consumer construct by synthesizing the fundamental features
shared by extant influential consumer types. The fundamental feature that defines consumers as
influential consumers of some sort (i.e., of the general marketplace, of a specific product
category, of innovation within a specific product category, or of a specific brand) sets these
consumers apart from the general population.
Previously identified types of influential consumers. Three distinctive influential
consumer types have been previously identified. Market mavens (i.e., “individuals who have
information about many kinds of products, places to shop, and other facets of markets, and
initiate discussions with consumers and respond to requests from consumers for market
information” Feick and Price 1987, p. 85) are influential consumers of the general marketplace.
Opinion leaders (i.e., individuals who exert disproportionate influence on the decision of others
in a specific product category through interpersonal communication) are influential consumers of
a specific product category (Childers 1986; Flynn, Goldsmith, and Eastman 1996; King and
Summers 1970; Rogers and Cartano 1962; Summers 1971). Product innovators (i.e., early
adopters who learn about and adopt innovations within a specific product category earlier than
others and who initiate the diffusion process) are also influential consumers of a specific product
category with an orientation towards innovation adoption and diffusion of the product category
(Goldsmith and Hofacker 1991; Rogers 1962; Summers 1971).

POWERFUL BRAND INFLUENTIALS 22

A review of these existing types of influential consumers reveals that influential
consumers are characterized by 1) a high tendency to exert personal influence on other
consumers’ decisions 2) through a high engagement of word-of-mouth (i.e., informal, consumer-
to-consumer communication between a receiver and a communicator whom the receiver
perceives as non-commercial, regarding a brand, a product, an organization, or a service,
Anderson 1998; Arndt 1967).
Influential consumers have a high tendency to influence other consumers’ decisions.
Influential consumers are thus those who tend to initiate great personal influence on others’
decisions. Previous research on opinion leaders (Engel and Blackwell 1982; Flynn, Goldsmith,
and Eastman 1994, 1996; Rogers 1962; Weimann 1994), product innovators (Goldsmith and
Flynn 1992; Kotler and Zaltman 1976; Rogers 1962), and market mavens (Feick and Price 1987;
Higie and Feick 1987) all emphasize that influential consumers are essentially defined by their
disproportionate exertion of personal influence on others. It is noteworthy that influential
consumers can exert either passive or active influence on other consumers (Feick and Price
1987). They could influence others in a passive or unintentional way through mechanisms of
“imitation” and “contagion” (Flynn et al. 1994; Hamilton 1971). Nevertheless, they are
predominantly characterized as those who are attempting and motivated to influence others’
decisions in an active manner (Bloch and Richins 1983; Dichter 1966; Engel et al. 1969;
Midgley and Dowling 1978; Walsh et al. 2004). Additionally, most influential consumer scales
could only capture personal influence of an intentional nature (Hamilton 1971). Correspondingly,
influential consumers are consumers who intend to exert a high influence on other consumers’
decisions.
POWERFUL BRAND INFLUENTIALS 23

Influential consumers exert influence through a high engagement of word-of-mouth.
Opinion leaders (Childers 1986; King and Summers 1970; Meyers and Robertson 1972; Richins
and Root-Shaffer 1988), product innovators (Engel et al. 1969; Midgley and Dowling 1978;
Rogers 1962), and market mavens (Feick and Price 1987; Higie, Feick, and Price 1987;
Wiedmann, Walsh, and Mitchell 2001) are all characterized by a high propensity toward word-
of-mouth. They act as information diffusers who initiate information exchange in the
marketplace. Researchers have highlighted word-of-mouth as the medium by which influential
consumers exercise their influence on others (Feick and Price 1987; King and Summers 1970;
Summers 1971). Word-of-mouth has been documented as the most influential form of
communications that affects consumer decision making (Katz and Lazarsfeld 1955). Therefore, a
high level of word-of-mouth engagement empowers influential consumers in their attempt to
influence others’ decisions.
3.2. Conceptual Properties of the Powerful Brand Influential Construct
The powerful brand influential construct shares the two conceptual components of the
influential consumer construct. Incorporating both defining components into a brand-level
conceptualization: powerful brand influential consumers are 1) consumers who have a high
tendency to influence other consumers’ brand decisions about a given brand and 2) they do so
through a high engagement of word-of-mouth about the brand. We further propose two
additional defining components to capture the key features of them being influential consumers
of a given brand: 3) they are consumers who have strong attachment to a given brand and 4) they
conduct passionate and emotionally evocative word-of-mouth about the brand. Theoretically,
powerful brand influentials could also have strong aversion toward a given brand and they
conduct strong negative word-of-mouth about the brand. Nevertheless, we focus on capturing
POWERFUL BRAND INFLUENTIALS 24

consumers for whom attachment is high (and aversion is low) in that they are most desirable for
a brand from a marketing perspective.
The powerful brand influential has strong brand attachment. Brand attachment
refers to "the strength of the bond connecting the brand with the self”; (Park et al. 2009). Brand
attachment is indicated by two fundamental components: brand-self connectedness and
prominence of brand-relevant thoughts and feelings (Park et al. 2009; Park et al. 2010).
Specifically, brand-self connectedness refers to the strength of connectedness between a given
brand and the self. Prominence of brand-relevant thoughts and feelings refers to the ease with
which brand-related thoughts and feelings are retrieved and the retrieval frequency of such
thoughts and feelings. Brand attachment constitutes “the core of all strong brand relationships”
(Fournier 1998, p. 363).
Several previous documentations suggest that the powerful brand influential has strong
attachment to a given brand. McConnell and Huba (2003) profile powerful brand influential
(also called “brand advocates”) as influential consumers who have built a strong connection with
the brand or company at an intrinsic, emotional level. They emphasized that these consumers are
not simply repeat purchasers of a brand. They are profiled as those who are passionate about a
brand, those who feel that a brand or company’s product is a must-have not simply a nice-to-
have, and those who wouldn’t stop talking to other consumers about how the brand changed their
life (McConnell and Huba 2003; Yaverbaum et al. 2006). All these profiles suggest the presence
of strong brand-self connectedness and prominent brand-related feelings and thoughts, both of
which indicate a strong brand attachment.

POWERFUL BRAND INFLUENTIALS 25

We further propose that a strong brand attachment acts as a key motivator for the
powerful brand influential’s high engagement of brand word-of-mouth. A strong (brand)
attachment has distinct behavioral implications. When people are strongly attached to an object,
they are highly motivated to maintain closeness to it (i.e., proximity seeking) and they suffer
from emotional and physical distress when they face potential or real separation from it (i.e.,
separation distress) (Bowlby 1973; Hazan and Zeifman 1999). When people are strongly
attached to a brand, they proactively create opportunities to devote personal discretionary
resources in sustaining brand-self relationship (Park, MacInnis, and Priester 2006, Park et al.
2010, Thomson, MacInnis, and Park 2005). They engage in acts that reveal brand commitment
(Fournier 1998; Thomson et al. 2005; Pimentel and Reynolds 2004).
Brand word-of-mouth and brand advocacy have been well-documented as brand
commitment activities performed by strongly brand-attached consumers to foster their needs for
brand relationship maintenance (Park et al. 2006, 2007; Park et al. 2010). Ample empirical
evidence shows that brand passion (one of the key emotional aspects of brand attachment) leads
to high engagement of brand word-of-mouth. Passionate consumers of a brand are highly
engaged in brand word-of-mouth (Matzler, Pichler, and Hemetsberger 2007; Pimentel and
Reynolds 2004; Rozanski et al. 1999). Matzler et al. (2007) investigated a construct they term
“brand evangelism” as a behavioral outcome of brand passion
4
. Brand evangelism is defined as
an “active and committed way of spreading positive opinions and trying fervently to convince or
persuade others to get engaged with the same brand” (p. 27). They note, “consumers who
evangelize are passionate about their brand and feel the need to share their emotions with others”

4
Matzler et al. (2007) have conceptualized brand passion (a key emotional aspect of brand attachment) and brand
evangelism (a behavioral outcome) as two separate constructs and intended to examine the causal relationship at the
construct level. We are interested in a type of consumers, the powerful brand influential (sometime named as brand
evangelists) who possess features that are integral to their essence. Hence our approach is different in that brand
attachment is a fundamental characteristic of the powerful brand influential (or brand evangelists).
POWERFUL BRAND INFLUENTIALS 26

(p. 27). We therefore propose that a third defining component of the powerful brand influential is
that they have strong attachment to a given brand.
The powerful brand influential conduct passionate and emotionally evocative brand
word-of-mouth. We propose that the powerful brand influential’s brand word-of-mouth is
featured by a passionate and emotionally evocative theme. This final defining component gets at
specific facets of word-of-mouth that characterize the powerful brand influential’s word-of-
mouth. Previous literature has predominantly concentrated on investigating the level (e.g.,
frequency, number of people told) and valence (i.e., positive, neutral, or negative) of word-of-
mouth. Limited work on word-of-mouth goes beyond a favorability or valence categorizations of
word-of-mouth. Nevertheless, we posit that the theme of word-of-mouth is a crucial aspect that
differentiates the powerful brand influential’s brand word-of-mouth from all other’s, and that
this feature essentially defines the powerful brand influential.
In the domain of communication, the theme of word-of-mouth relates to the message (M)
component of Berlo’s S-M-C-R communication model (Berlo 1960). The message (M) is the
information that is transmitted from the source (S) over the word-of-mouth channel (C) to the
receiver (R). Two essential dimensions of the message are its content and treatment (Berlo 1960).
In the marketing literature, Mazzarol, Sweeney, and Soutar (2007) have empirically identified
two word-of-mouth themes specific to the marketplace. These two themes significantly influence
how much impact word-of-mouth has on the receiver. The first word-of-mouth theme is
“richness of the message”. Richness relates to the content dimension of message and denotes the
message’s depth, intensity, and vividness. Word-of-mouth messages that include highly
descriptive and evocative terms and involve a high degree of storytelling are perceived as vivid
(Mazzarol et al. 2007; Herr, Kardes, and Kim 1991). Vividness refers to the extent to which
POWERFUL BRAND INFLUENTIALS 27

message is “(a) emotionally interesting, (b) concrete and imagery-provoking, and (c) proximate
in a sensory, temporal, or spatial way” (Nisbett and Ross, 1980, p. 45). Vivid messages are more
inherently interesting, thought provoking, and attention holding. Moreover, vivid messages are
more accessible from memory and have more influence on consumer judgment (Herr et al. 1991;
Kisielius s and Sternthal 1986; Nisbett and Ross 1980).
The second word-of-mouth theme is “strength or power of advocacy”. Strength, which
relates to the treatment dimension of the message (i.e., the manner in which a message is
conveyed), denotes the potency of the recommendation. A recommendation is not equivalent to
word-of-mouth since a recommendation can be strong or weak and explicit or implicit (Mazzarol
et al. 2007). The advocacy of word-of-mouth is strong when the message is conveyed in a firm,
assertive, and passionate manner and through the usage of body language (e.g., gestures, eye
contact) (Mazzarol et al. 2007). Considering both word-of-mouth themes, word-of-mouth is a
more significant event for both the source and receiver when the message involves rich
descriptive, evocative language and storytelling and when the message is conveyed in a strong
and passionate manner.
We propose that brand word-of-mouth conducted by the powerful brand influential is
featured by rich messages with strong advocacy themes – that is, their brand word-of-mouth is
featured as passionate (in the manner of message convey) and emotionally evocative (in the
richness of message). A passionate and emotionally evocative theme that is specific to a brand
differentiates the powerful brand influential’s brand word-of-mouth from others’ at the same
time making their brand communication disproportionately influential.

POWERFUL BRAND INFLUENTIALS 28

We further provide theoretical reasoning for such proposition. As previously discussed,
the powerful brand influential has strong brand attachment, which is reflected by two properties
the brand-self connection property and the prominence of brand-relevant thoughts and feelings
property (Park et al. 2009; Park et al. 2010). The brand-self connection property of brand
attachment implicates “hot” affect originated from the brand’s linkage with the self (Ball and
Tasaki 1992; Mikulincer et al. 2001; Thomson et al. 2005). “Hot” affect towards the brand will
channel the manner in which the powerful brand influential conveys the brand to others – “hot”
affect is conveyed in a passionate manner through word-of-mouth and often involving the usage
of body language. As evidenced by McConnell and Huba’s (2007) study, a key feature of the
powerful brand influential is that they passionately recommend the brand or company to others,
and “passionate is the keyword” (p. 4). As they spread the word about the brand, their love and
passion for the brand is “highly visible” and their enthusiasm about the brand is “contagious”
(Yaverbaum et al. 2006). Consumers who are passionate about a brand not only spread positive
word-of-mouth but engage in a more active and committed way of brand advocacy: they
fervently convince or persuade others to engage with their beloved brand (Matzler et al. 2007;
Pimentel and Reynolds 2004; Rozanski et al. 1999).
Furthermore, the prominence of brand-relevant thoughts and the feelings property of
brand attachment implicate highly accessible personal thoughts and feelings associated with the
brand. Whenever a conversation about the brand is initiated, the personal brand-associated
thoughts and feelings will be automatically retrieved. Correspondingly, the powerful brand
influential’s brand word-of-mouth is filled with rich and vivid content involving personal
thoughts and feelings. Such content engenders descriptive, emotionally evocative language and
storytelling. Matzler et al.’s (2007) account of brand evangelism echoes this facet of the powerful
POWERFUL BRAND INFLUENTIALS 29

brand influential’s brand word-of-mouth – “preaching the brand’s most loved aspects and all
positive associations that come with it to people who have so far not acknowledged ‘the wonder
of it’” (p. 27).
3.3. Conceptual Definition of the Powerful Brand Influential
Integrating the four abovementioned defining characteristics, we define the powerful
brand influential as consumers who are strongly attached to a given brand and who have high
tendency to influence other consumers’ brand decisions through a high engagement of passionate
and emotionally evocative word-of-mouth about the brand.

POWERFUL BRAND INFLUENTIALS 30

Chapter 4: Differentiating the Powerful Brand Influential from Market Maven, Opinion
Leader, Brand Attachment, and Brand Ambassador
The powerful brand influential construct should be distinguished from two previously
identified influential consumer constructs, the market maven and the opinion leader
5
. As
influential consumers, the powerful brand influential shares similar features with existing types
of influential consumers. The powerful brand influential construct should also be distinguished
from brand attachment and the notion of a brand ambassador. Both brand attachment and brand
ambassador are brand-level constructs. Both share certain conceptual similarities to the powerful
brand influential. It is therefore crucial to demonstrate that the two extant brand-level concepts
do not capture a brand’s influential consumers at the same time justifying that the powerful
brand influential construct does. In the next session, we discuss how the powerful brand
influential construct relates to market maven, opinion leader, brand attachment and brand
ambassador.
4.1. The Powerful Brand Influential is Related to Market Maven, Opinion Leader, Brand
Attachment, and Brand Ambassador
The relatedness of the powerful brand influential and market maven, opinion leader.
The market maven and opinion leader constructs share conceptual similarities with the powerful
brand influential construct. Conceptually, as influential consumers, they share the two
fundamental components that make them influential and differentiate them from the general
consumer population. Previous studies have shown that different groups of influential consumers
overlap to certain extent: market mavens also can be opinion leaders or product innovators

5
Product innovator as another identified influential consumer construct is not further discussed in this paper since
product innovators tend to be the influential consumers of innovation adoption and diffusion. This research
investigates the influential consumers of established brands and does not tap into the domain of new products and
innovations.
POWERFUL BRAND INFLUENTIALS 31

(Feick and Price 1987) and product innovators also can be opinion leaders (Baumgarten 1975;
Summers 1971). Likewise, we expect that the powerful brand influential can also be market
mavens and opinion leaders and wise versa.
A subpopulation of opinion leaders and market mavens can be brand influentials. By
definition, opinion leaders and market mavens are consumers who have a high tendency to
influence other consumers’ decisions through high engagement of word-of-mouth. When these
influential consumers are strongly attached to a given brand, they acquire the defining features of
the powerful brand influential: a strong brand attachment will motivate them to be highly
engaged in passionate and emotionally evocative word-of-mouth about the brand. Comparing
market mavens to opinion leaders, we expect the latter will overlap with the powerful brand
influential to a greater extent than the former in general. Because they are passionate about and
involved in the product category, they may also develop a strong attachment to a given brand
within the category. Consumers whose involvement is in the general marketplace (i.e., market
mavens) may be less inclined to do the same. Based on the differential overlapping ratio, we
expect that the brand influential construct is more closely related to the opinion leader than the
market maven construct.
The relatedness of the powerful brand influential and brand attachment. The
powerful brand influential construct has a deep-seated relationship to the brand attachment
construct. As conceptualized previously, a strong brand attachment is a defining component of
the powerful brand influential construct. We propose that a high brand attachment is a necessary
but not sufficient condition for the powerful brand influential. A simplified way to summarize
the relationship between the two constructs is: All powerful brand influential consumers are
strongly attached to a given brand, whereas consumers who are strongly attached to a given
POWERFUL BRAND INFLUENTIALS 32

brand are not necessarily powerful brand influential consumers. Powerful brand influential
consumers constitute a subgroup of strongly brand-attached consumers. We propose that the
tendency to become a powerful brand influential corresponds closely to the level of brand
attachment as indicated by brand-self connection and brand prominence together. In other words,
as the brand attachment level is high, the tendency to be a powerful brand influential is high; and
as the brand attachment level is low, the tendency to be a powerful brand influential is low.
The relatedness of the powerful brand influential and brand ambassador. Brand
ambassadors are often described in the popular press as those who are “enlisted” by a company
via incentives and rewards (e.g., prizes, cash, store credits, product samples, gifts, discounts) to
advertise and promote a brand to other consumers (Collier 2007; Voight 2007; also see
Wikipedia for “Evangelism Ambassador” http://en.wikipedia.org/wiki/Brand_ambassador).
Brand ambassadors may be celebrity figures. Yet they need not be celebrities. With the advent of
social media, several companies such as Sony, Unilever, Microsoft, McDonald’s and JetBlue
have launched brand ambassador programs recruiting everyday consumers as brand ambassadors.
Brand ambassadors regardless of whether they are celebrities or everyday consumers are
influential consumers who are motivated to promote a given brand through high engagement of
word-of-mouth and advocacy behaviors (online or offline)
(http://en.wikipedia.org/wiki/Brand_ambassador; Sweeney 2002; Voight 2007). Therefore, the
brand ambassador construct shares similarities with the powerful brand influential in that it
denotes a high motivation to influence others’ brand decision and a high engagement in brand
word-of-mouth and advocacy as means to do so. Correspondingly, they share two defining
components with the powerful brand influential: high motivation to influence others’ brand
decision and high engagement in brand word-of-mouth and advocacy. We also expect that the
POWERFUL BRAND INFLUENTIALS 33

powerful brand influential and brand ambassadors overlap in the market to some extent. The
powerful brand influential is likely to be one who has already voluntarily acted as a brand. Their
brand advocacy is highly visible and motivated, which makes them identifiable and desirable to a
company that intends to recruit brand ambassadors. When reached, they are more likely to take
on the role of brand ambassador as additional opportunities to sustain, expand, and restore the
brand-self relationship.
In sum, we have discussed how the powerful brand influential construct relates to the
construct of market maven, opinion leader, brand attachment, and brand ambassador. Though the
similarities reside in different groundings, we expect that
P1: The powerful brand influential construct is significantly related to the construct of
market maven (P1a), opinion leader (P1b), brand attachment (P1c), and brand
ambassador (P1d).
4.2. The Powerful Brand Influential is Distinct from Market Maven, Opinion Leader,
Brand Attachment, and Brand Ambassador
Though closely related, the powerful brand influential construct is distinctly different
from market maven, opinion leader, brand attachment, and brand ambassador. The powerful
brand influential construct can be differentiated from the rest in three key aspects that are crucial
to capture influential consumers of a given brand. These aspects include: 1) the level of
engagement in word-of-mouth and advocacy behaviors of a given brand, 2) the level of influence
on other consumers’ brand decisions, 3) the level of intrinsic pleasure derived from word-of-
mouth about a given brand. With such differentiation, we intend to show that the powerful brand
influential adds to the literature as a unique construct that captures a brand’s influential
consumers in that none of other related constructs adequately captures a brand’s influential
POWERFUL BRAND INFLUENTIALS 34

consumers. We develop a set of propositions that pertain to the distinction between the powerful
brand influential construct and other related constructs on the three dimensions. Tests for these
propositions will demonstrate the validity of the powerful brand influential construct and our
scale to measure this construct. Brand attachment, among all the related constructs, shares the
most fundamental conceptual overlapping with the powerful brand influential construct. We
thereby devote a separate section to further differentiate the powerful brand influential from
brand attachment and develop a set of propositions for additional distinctions between the two
constructs.
Based on all the subsequent differentiation, we first propose that:
P2: The powerful brand influential construct is distinctly different from the construct of
market maven (P2a), opinion leader (P2b), brand attachment (P2c), and brand
ambassador (P2d).
4.3. The Powerful Brand Influential is Distinct on the Engagement Level of Brand Word-
of-Mouth and Brand Advocacy Behavior
The powerful brand influential vs. market maven and opinion leader. Powerful
brand influentials are differentiated from market mavens and opinion leaders on the level of
motivation to influence others’ brand decisions and correspondingly the level of engagement in
word-of-mouth and advocacy behavior about a given brand.
First, we expect that powerful brand influentials are more motivated to influence others’
decision about a brand than market mavens and opinion leaders. A strong brand attachment
constitutes the motivational basis of the powerful brand influential for influencing others’ brand
decisions. As previously mentioned, strong brand attachment acts as an innate motivator for
active investment in and devotion of personal resources to maintain and expand the brand-self
POWERFUL BRAND INFLUENTIALS 35

relationship (Park et al. 2006, 2007; Park et al. 2010; Thomson et al 2005). The motivational
power of brand attachment resides in the fundamental psychology of self-expansion theory (Aron
et al. 2005). Self-expansion theory contends that people possess the innate desire to expand the
“self” by incorporating others (applies to brands as well) into their sense of self. Studies have
shown that strongly attached consumers are more motivated to devote their personal resources to
a brand as ways of self-expansion (Johnson and Rusbult 1989; Park et al. 2010; Thomson et al
2005). For example, they are more motivated to allocate their social resources to defend the
brand to others (Johnson and Rusbult 1989). They are more motivated to allocate their time
resources to promote the brand and brand communities (Muñiz and O’Guinn 2001; Schouten and
McAlexander 1995). Park et al. (2010) proposed that brand attachment has exceptional
predictive power for difficult behaviors including those oriented towards influencing others’
brand decisions such as promoting the brand and defending the brand.
Market mavens and opinion leaders (other than those who overlap with powerful brand
influential consumers) do not possess a strong motivational basis for influencing other’s brand
decisions. Market mavens are characterized by high marketplace and purchasing involvement
(Feick and Price 1987), which motivates them to influence others’ decisions about the general
marketplace and purchase decisions. Opinion leaders are characterized by high involvement in
product category (Bloch and Richins 1983; Richins and Root-Shaffer 1988), which motivates
them to influence others’ decisions about a given product category. Consumers who possess
continuing interest or enthusiasm regarding marketplace activities or a given product category
are not necessarily interested in any given brand. Market mavens and opinion leaders might not
have strong motivational basis for influencing others’ at a specific brand level. By the same
token, powerful brand influential are not necessarily highly involved in the marketplace or a
POWERFUL BRAND INFLUENTIALS 36

product category and hence might not possess a strong motivation to influence others in those
domains. Therefore, the powerful brand influential should be more motivated to influence others’
brand decisions than is true for market mavens and opinion leaders.
Second, powerful brand influentials differ from market mavens and opinion leaders in the
engagement level of word-of-mouth and advocacy behavior about a given brand. Word-of-mouth
is the medium by which influential consumers exert their personal influence on others’ decisions
(Feick and Price 1987; King and Summers 1970; Summers 1971). As influential consumers,
powerful brand influentials, market mavens, and opinion leaders are all highly engaged in word-
of-mouth to exert personal influence on other consumers. However, the primary domain in which
they conduct word-of-mouth is different. Since they are motivated to influence others in different
domains, they conduct word-of-mouth primarily in the corresponding domain of their intended
influence.
Specifically, market mavens are highly engaged in word-of-mouth about the general
marketplace to influence others’ decisions about purchase decisions in the general marketplace.
They conduct word-of-mouth about the general marketplace and purchase activities: they talk
about changes in marketing mix (price, availability, new products and stores) spanning multiple
product categories (Feick and Price 1987; Slama and Williams 1990), attributes of retailers
(grocery stores, department stores, and discount stores), and mostly about special sales, prices,
product quality and product variety (Higie et al. 1987). Opinion leaders conduct word-of-mouth
about the product category that they are highly involved in (Bloch and Richins 1983; Richins and
Root-Shaffer 1988) to influence others’ decisions about the category. A given brand is neither
the primarily word-of-mouth domain of market mavens nor that of opinion leaders. As we will
discuss in more detail in a subsequent section, market mavens and opinion leaders might talk to
POWERFUL BRAND INFLUENTIALS 37

like-minded others and respond to others about a brand when they are requested for information
(Feick and Price 1987; Richins and Root-Shaffer 1988; Weimann 1994). However, they lack of
motivation to actively reach out to other consumers to talk about a specific brand. If they do
engage in word-of-mouth about a specific brand they are selective regarding to whom they talk
to and when. Additionally, they are not likely to engage in brand advocacy behaviors that are
more resource-demanding than word-of-mouth (e.g. promoting a brand committee, defending a
brand).
In contrast, powerful brand influentials are highly engaged in word-of-mouth and all
other forms of advocacy behaviors about their attached brand (Park et al. 2006, 2007; Park et al.
2010; Pimentel and Reynolds 2004) as means to influence others’ decisions about the brand. We
expect that they proactively create opportunities (to anyone and on any occasion) to talk about
the brand. Comparing to market mavens and opinion leaders, powerful brand influentials should
be less selective in choosing the audience and the occasions to talk about the brand. They preach
about the brand, actively recruit anyone who has so far not acknowledged “the wonder of it”, and
continuously talk about the brand (Matzler et al. 2007; McConnell and Huba 2007; Pimentel and
Reynolds 2004). In sum, the powerful brand influential construct should be distinguishable from
market maven and opinion leader by a higher engagement level of brand word-of-mouth and
advocacy.
The powerful brand influential vs. brand attachment. We propose that the powerful
brand influential is distinguished from brand attachment by a higher level of motivation to
influence others’ brand decisions and correspondingly a higher engagement level of brand word-
of-mouth and advocacy. Though abundant existing studies (Johnson and Rusbult 1989; Muñiz
and O’Guinn 2001; Park et al. 2010; Schouten and McAlexander 1995; Thomson et al 2005)
POWERFUL BRAND INFLUENTIALS 38

have shown that brand attachment largely predicts brand word-of-mouth and advocacy behaviors,
these studies have not explored the powerful brand influential concept. All powerful brand
influential consumers are highly motivated to influence others’ brand decisions and they do so
through a high engagement of brand word-of-mouth and advocacy behaviors. By contrast, not all
strongly brand-attached consumers possess such features. Hence we distinguish powerful brand
influentials (as a subgroup of all strongly brand-attached consumers) from the rest of strongly
brand-attached consumers.
All strong brand-attached consumers are highly motivated to devote personal resources to
sustain, maintain and expand the brand-self relationship (Park et al. 2006, 2007; Park et al. 2008;
Park et al. 2010; Thomson et al 2005). Due to fundamental needs for self-expansion, all strongly
brand-attached consumers are highly engaged in brand commitment behaviors of some sort
(Fournier 1998; Pimentel and Reynolds 2004; Thomson et al. 2005). Previous research has
revealed that commitment behaviors take various forms (Bhattacharya and Sen 2003; Johnson
and Rusbult 1989; Madrigal 1995; Park et al. 2006; Park et al. 2010; Pimentel and Reynolds
2004; Thomson et al. 2005). We examined the forms of commitment behaviors that have been
identified in the brand attachment literature. We propose that brand word-of-mouth and
advocacy behaviors constitute a unique category of brand commitment behaviors. To facilitate
the discussion, we divide brand commitment behaviors into two general categories.
The first category of brand commitment behaviors involves an inward orientation
towards one’s own relationship with a brand. This category includes brand loyalty behaviors
(e.g., purchasing, repurchasing, waiting to purchase, switching from competitor to the brand,
paying a price premium) as well as other commitment behaviors that are inwardly oriented such
as participating in personal rituals, making personal sacrifices, collecting brand-related objects,
POWERFUL BRAND INFLUENTIALS 39

and personally creating brand-related items. These behaviors reflect a focus on fostering one’s
own relationship with the attached brand with an inward orientation that only involves oneself
(and not others). They require devotion of one’s financial resources and time resources into a
brand. In this paper, we focus on brand loyal behaviors only since other inwardly oriented
behaviors are rather difficult to capture with quantitative scales (they are often captured by
qualitative methods such as interview, observation, ethnography). Correspondingly, we refer to
the first category as brand loyal behavior and differentiate it from the second category, which
includes brand word-of-mouth and advocacy behavior.
The second category of brand commitment behaviors involves an outward orientation on
influencing other consumers’ decisions about a brand. This category includes brand word-of-
mouth and all forms of brand advocacy behaviors such as defending the brand to others,
publically derogating alternatives, promoting the brand, participating in brand community
activities, and voluntarily recruiting new brand users. All these behaviors reflect a focus to
influence others’ decisions regarding one’s own attached brand with an outward orientation that
involves others. In addition to financial and time resources, they require devotion of one’s social
resources into a brand. Both categories serve as a means for strongly brand-attached consumers
to satisfy their needs of sustaining, maintaining and expanding brand-self relationship. However,
the category of brand word-of-mouth and advocacy behavior demands unique requirements that
are not present in the category of brand loyal behavior (and the other inwardly-oriented
commitment behaviors). Brand word-of-mouth and advocacy behavior involve the motivation to
influence others’ brand decisions, the opportunity and ability (self-perceived or real) to fulfill
such motivation with enactment. Inwardly-focused brand loyalty behavior does not have these
requirements. Not all strongly brand-attached consumers are motivated to influence others’ brand
POWERFUL BRAND INFLUENTIALS 40

decisions and/or have the opportunity and ability to do so. We propose that there exist strongly
brand-attached consumers who are highly engaged in brand loyalty behavior but are not highly
engaged in brand word-of-mouth and advocacy behavior. By contrast, all powerful brand
influentials are highly engaged in the latter behavior category.
These propositions are in line with the hierarchy of brand attachment and behaviors
discussed by Park et al. (2006). Park et al. (2006) categorize the hierarchy of brand attachment
strength (i.e., low, moderate, and high), the dimensions of personal resource sacrifice (i.e., self-
image resources and personal discretionary resources), and the type of brand behavioral
engagement (behaviors including brand loyalty acts and brand advocacy acts). Their hierarchy
(see Appendix A) posits three scenarios representing the high level of brand attachment. In the
first scenario, the high level of attachment involves low/moderate allocation of self-image
resources (or social resources) but high allocation of personal discretionary resources (or
financial and time resources). The corresponding type of brand behavioral engagement resembles
our proposed brand loyal behavior category. This scenario corresponds to our proposition that
strongly brand-attached consumers could have a high level of engagement in brand loyal
behavior but low or moderate level of engagement in brand advocacy behavior.
In Park et al.’s (2006) second and third scenarios, the high level of attachment involves
high allocation of self-image resources. The corresponding type of brand behavioral engagement
is word-of-mouth and brand advocacy behavior while the engagement level of brand loyal
behavior could also be high. These two scenarios correspond to our proposition that there exist
strongly brand-attached consumers (powerful brand influential consumers) who have a high
engagement level in both categories of brand-commitment behavior. Comparing to the first
POWERFUL BRAND INFLUENTIALS 41

scenario of strongly brand-attached consumers, powerful brand influentials are differentiated by
a higher engagement level of brand word-of-mouth and advocacy behavior.
Based on the reasoning, we expect that the powerful brand influential construct predicts
brand word-of-mouth and advocacy behavior better than does brand attachment. We anticipate
that the powerful brand influential construct largely accounts for predictive ability of brand
attachment for brand word-of-mouth and advocacy behavior that has been suggested by previous
studies (Fedorikhin, Park, and Thomson 2008; Park et al. 2006, 2007; Park et al. 2010).
The powerful brand influential vs. brand ambassador. We propose that the powerful
brand influential consumer is distinguished from the brand ambassador by a higher level of
motivation to influence others’ brand decisions and correspondingly a higher engagement level
in brand word-of-mouth and advocacy. Brand attachment is not a defining feature of brand
ambassadors. Brand ambassadors are not necessarily attached to the brand that they advocate.
They might be highly engaged in brand word-of-mouth and advocacy for the sake of extrinsic
incentives and rewards. Motivated by extrinsic values, their engagement in brand advocacy is
proportional to and falls within the limit of the rewards they receive from their engagement.
Though brand ambassadors are likely to seek a broader audience and search for occasions to talk
about the brand to acquire incentives, they are unlikely to seek for additional audiences and
opportunities that go beyond the incentives they would receive from doing so.
Powerful brand influentials are intrinsically driven by strong brand attachment, which
motivates them to constantly expand their brand-self relationship to the next level. There is no
limit to the motivational and engagement level of their brand word-of-mouth and advocacy
behaviors. They are more engaged in seeking broader audiences and occasions for their brand
word-of-mouth and advocacy than are brand ambassadors. Therefore, we expect that the
POWERFUL BRAND INFLUENTIALS 42

powerful brand influential construct is distinguishable from brand ambassador concept by a
higher engagement level of brand word-of-mouth and advocacy.
To summarize, we have discussed why the powerful brand influential is differentiated
from other constructs: market maven, opinion leader, brand attachment, and brand ambassador
by a higher engagement level of brand word-of-mouth and advocacy behavior. Based on these
reasoning, we anticipate that
P3: The powerful brand influential construct predicts the engagement level of word-of-
mouth about a brand better than does the construct of market maven. (P3a), opinion
leader (P3b), brand attachment (P3c), and brand ambassador (P3d).
P4: The powerful brand influential construct predicts brand advocacy behavior better
than does the construct of market maven (P4a), opinion leader (P4b), brand attachment
(P4c), and brand ambassador (P4d).
4.4. The Powerful Brand Influential is Distinct on the Level of Influence on others’ Brand
Decisions
We propose that powerful brand influentials have more impact on other consumers’
brand decisions than do market mavens, opinion leaders, strongly brand-attached consumers, and
brand ambassadors. The previous section indicated that they are more motivated to influence
others’ brand decisions and therefore more engaged in brand word-of-mouth and advocacy
behaviors than are the other three consumer groups. They talk to more people, seek a broader
audience and take more opportunities to influence others’ brand decisions than the rest. All these
make them exceptionally influential regarding a given brand.

POWERFUL BRAND INFLUENTIALS 43

We acknowledge that market mavens and opinion leaders as influential consumers in the
market might also be highly engaged in word-of-mouth about a brand. Drawing from previous
research, we expect that they could conduct word-of-mouth about a given brand for several
motivational reasons (Arndt 1967; Dichter 1966; Walsh et al. 2004). They could conduct brand
word-of-mouth for self-involvement motivations wherein word-of-mouth is conducted to meet
the need for self-confirmation in front of others, more for the sake of the self than for the
product/brand. Specifically, they could talk about a brand to show their own connoisseurship and
to assert their own superiority in front of others (Dichter 1966). They could also conduct brand
word-of-mouth for altruistic motivations wherein word-of-mouth is conducted to help the
receivers make a better decision. Since market mavens and opinion leaders are those who posses
superior expertise and reputation, they could talk about a brand to help others and to respond to
request from others (Arndt 1967; Walsh et al. 2004). As designated information transmitter and
adviser-giver in the market, they are likely to be highly engaged in providing information and
responding to requests from others about a given brand.
For different motivational reasons from powerful brand influentials, market mavens and
opinion leaders might also engage in word-of-mouth about a given brand. Therefore, we further
distinguish the powerful brand influential from opinion leaders and market mavens regarding
their level of influence on others’ brand decisions. Since strongly brand-attached consumers and
brand ambassadors are not well-known influential consumers, they are unlikely to be highly
engaged in brand word-of-mouth for those motivations that are mostly relevant to designated
influential consumers in the marketplace. Strongly brand-attached consumers and brand
ambassadors are less influential than powerful brand influentials because they are less engaged
in word-of-mouth and brand advocacy behaviors.
POWERFUL BRAND INFLUENTIALS 44

Powerful brand influentials’ brand word-of-mouth vs. market mavens’ and opinion
leaders’ brand word-of-mouth. We expect that powerful brand influentials conduct more
persuasive of word-of-mouth about a brand than do market mavens and opinion leaders. Their
brand word-of-mouth is different from market mavens’ and opinion leaders’ in several
fundamental facets: the content of brand word-of mouth (i.e., what they talk about a brand), the
treatment of brand word-of-mouth (i.e., how they talk about the brand), and the level of intrinsic
pleasure derived from brand word-of-mouth (i.e., how much they enjoy talking about the brand).
Specifically, powerful brand influentials differ from mavens and opinion leaders in what
they say about a brand (the content) and how they talk about a brand (the treatment). Market
mavens and opinion leaders exchange knowledge and expertise-based information (what they
know and advice about a brand) more than personal experience. They possess superior
knowledge and expertise about the marketplace and product, respectively (Assael 1984; Feick
and Price 1987; Jacoby and Hoyer 1981). They exchange more advice giving, product news, and
marketing mix information (Feick and Price 1987; Higie et al. 1987; Richins and Root-Shaffer
1988; Slama and Williams 1990) than personal experience information (Richins and Root-
Shaffer 1988). Although they may possess great knowledge about a brand, market mavens and
opinion leaders are not necessarily users of the brand that they talk about. They might lack of
personal experience with the brand and hence may not be likely to exchange user-based or
experiential brand information. Additionally, they are respected and sought by others because of
their superior credibility to provide correct information (Bristor 1990; Dichter 1966; Feick and
Price 1987; Roger 1962). Therefore, they are likely to conduct factual and objective word-of-
mouth about a brand. The valence of their brand word-of-mouth content could be positive,
POWERFUL BRAND INFLUENTIALS 45

neutral, or negative depending on what they know is factual and true about the brand. Factual
and objective word-of-mouth is primarily evaluative and “cold” in nature.
The features of their word-of-mouth content should influence how they talk about a brand.
Since the content of their brand word-of-mouth is knowledge-based, objective, evaluative and
“cold”, market mavens and opinion leaders are likely to talk about a brand in an objective and
factual manner. They provide the receivers with knowledge and advice about a brand rather than
passionate and strong advocacy about the brand.
By contrast, powerful brand influentials as strongly brand-attached consumers do not
simply have great brand knowledge and expertise (Park et al. 2009). They have developed a
strong self-brand connection over time (Mikulincer and Shaver 2003). They possess prominent,
autobiographical and episodic memories, thoughts, and feelings about the brand (Collins 1996;
Mikulincer 1998; Park et al. 2010). The brand-self connection and brand prominence
components of brand attachment imply that whenever the brand is inferred (e.g., when talking
about the brand), the self is inferred and rich brand-experiential thoughts and feelings are
brought to mind (Park et al. 2009; Park et al. 2010). Therefore, powerful brand influentials tend
to conduct experiential and self-related brand word-of-mouth. Moreover, brand attachment
implicates “hot” affect and strong emotions built from a strong linkage connecting the brand to
the self. Therefore, powerful brand influentials tend to engage in “hot” and subjective brand
word-of-mouth. Since brand attachment denotes a positive valence of the brand-self connection
and brand-related thoughts and feelings (Park et al. 2010), powerful brand influentials’ word-of-
mouth can only be positive in valence. Furthermore, their word-of-mouth is featured by rich
content including personal thoughts and feelings and involves emotionally evocative descriptions,
frequent usage of body language, and storytelling. When they talk about the brand to other
POWERFUL BRAND INFLUENTIALS 46

consumers their word-of-mouth goes beyond “what they know about the brand”. They talk about
“how much they love the brand”, “what the brand means to them”, and “how the brand changed
their life” (McConnell and Huba 2003; Yaverbaum et al. 2006).
Regarding how they talk about the brand, powerful brand influentials convey strong
advocacy in their brand word-of-mouth. Their brand word-of-mouth is featured by a passionate
and emotionally evocative theme. They do not just spread positively word-of-mouth but
advocate the brand by fervently convincing or persuading others to get engaged with their
beloved brand (Matzler et al. 2007; Pimentel and Reynolds 2004; Rozanski et al. 1999).
As previously discussed, the richness of content and the strength of advocacy are two
crucial themes of word-of-mouth that significantly influence how influential word-of-mouth is
on the receiver (Herr et al. 1991; Mazzarol et al. 2007; Nisbett and Ross 1980). Word-of-mouth
is more influential when the content involves rich descriptive, emotionally evocative, and
storytelling descriptions (Herr et al. 1991; Mazzarol et al. 2007). Furthermore, word-of-mouth is
more influential when the content is conveyed in a strong and passionate advocacy manner. The
brand word-of-mouth conducted by powerful brand influentials involves rich content and is
conveyed with a strong advocacy theme. In contrast, the brand word-of-mouth conducted by
market mavens and opinion leaders lacks of features of rich content and strong advocacy. Hence,
powerful brand influentials conduct brand word-of-mouth is more influential than that of market
mavens and opinion leaders.

POWERFUL BRAND INFLUENTIALS 47

To summarize the distinction regarding the level of influence on other consumers’ brand
decisions, we propose that:
P5: The powerful brand influential construct predicts the tendency to influence others’
brand decisions better than does the construct of market maven (P5a), opinion leader
(P5b), brand attachment (P5c), and brand ambassador (P5d).
4.5. The Powerful Brand Influential is Distinct on the Level of Intrinsic Pleasure Derived
from Brand Word-of-Mouth
Powerful brand influentials also differ from market mavens, opinion leaders, strongly
brand-attachment consumers, and brand ambassadors in the level of intrinsic pleasure derived
from word-of-mouth about a given brand. That is they differ in how much sheer pleasure and
intrinsic satisfaction they derive from the action of word-of-mouth about a brand.
First, we differentiate powerful brand influentials from market mavens and opinion
leaders on this dimension. As those who have a high propensity to engage in the act of word-of-
mouth, market mavens and opinion leaders find word-of-mouth activities generally intrinsically
satisfying (Feick and Price 1987; Price et al. 1987; Walsh et al. 2004). When compared to
powerful brand influentials, however, we expect that market mavens and opinion leaders derive a
lower level of intrinsic pleasure from word-of-mouth about a brand. Previous literature has
shown that how people conduct word-of-mouth influences and reflects how much intrinsic
pleasure they derive from word-of-mouth (Bloch 1986; Feick and Price 1987; Price et al. 1987;
Walsh et al. 2004). When people conduct emotionally evocative and storytelling word-of-mouth,
they tend to derive great joy from it (Mazzarol et al. 2007). Word-of-mouth is a more significant
event not only to the receivers but also to the source of word-of-mouth when it is rich in content
and when it is conveyed in a strong and passionate manner (Herr et al. 1991; Mazzarol et al.
POWERFUL BRAND INFLUENTIALS 48

2007; Nisbett and Ross 1980). Furthermore, people driven by sheer passion who talk about a
brand should find the event more intrinsically enjoyable than those who talk about a brand driven
by responsibilities to provide information and advice to others. Correspondingly, powerful brand
influentials should derive a higher level of intrinsic pleasure from word-of-mouth about their
strongly-attached brand than do market mavens and opinion leaders.
Second, we expect that powerful brand influentials derive greater intrinsic pleasure than
do other strongly brand-attached consumers from their brand word-of-mouth. We propose that
there exist strongly brand-attached consumers who derive low intrinsic pleasure from word-of-
mouth about the brand. When strongly brand-attached consumers talk about their attached brand,
their word-of-mouth is filled with rich brand-self related feelings and emotionally evocative.
Talking about their attached brand therefore entails socially expressing their emotions related to
the brand. Individuals differ in their dispositional emotional expressiveness, the tendency of
which individuals outwardly display their emotions (Kring, Smith, and Neale 1994). Some
strongly brand-attached consumers might have low emotional expressiveness. These consumers
should find brand word-of-mouth taxing and derive low intrinsic pleasure from the act. We
expect that these consumers have low propensity to engage in brand word-of-mouth. By
comparison, powerful brand influentials are those who chose brand word-of-mouth as the means
to fulfill their needs for brand-self relationship. We expect that they have higher emotional
expressiveness than other strongly brand-attached consumers. We also expect that they drive a
higher intrinsic value from the act of brand word-of-mouth than do other strongly brand-attached
consumers.

POWERFUL BRAND INFLUENTIALS 49

Last, we expect that powerful brand influentials derive greater intrinsic pleasure from
word-of-mouth than do brand ambassadors. Brand ambassadors are largely driven by extrinsic
values to conduct brand word-of-mouth. Since motivation is more extrinsic than intrinsic
motivation, they might not find word-of-mouth about a given brand intrinsically satisfying. They
are engaged in brand word-of-mouth more for the sake of incentives than enjoyment.
In sum, we proposed that:
P6: The powerful brand influential construct predicts the level of intrinsic pleasure
derived from brand word-of-mouth better than does the construct of market maven (P6a),
opinion leader (P6b), brand attachment (P6c), and brand ambassador (P6d).
4.6. Further Differentiating the Powerful Brand Influential from Brand Attachment
We regard the powerful brand influential construct as distinctly different from brand
attachment because it captures key defining components that are not captured by brand
attachment. We emphasize that the powerful brand influential construct is not equivalent to a
high level of the brand attachment construct. We posit that the two constructs are distinguishable
regarding key psychological, emotional, and behavioral outcomes.
The powerful brand influential vs. brand attachment on separation distress and
brand commitment. Separation distress and brand commitment are psychological outcomes that
indicate the existence of attachment to an entity (Bowlby 1979; Hazan and Zeifman, 1999; Park
et al. 2009; Park et al. 2010; Thomson et al. 2005). Separation distress (i.e., the experience of
emotional and physical distress at the prospect of potential or real separation from an attached
entity) is a natural product of attachment (Bowlby 1980). Brand commitment is defined as “a
decision or pledge to maintain a long-term relationship with a brand into the future” (Park et al.
2009, p. 331). Commitment is a psychological outcome of attachment regarding future behaviors
POWERFUL BRAND INFLUENTIALS 50

(e.g., Levinger 1980; Park et al. 2009; Rosenblatt 1977; Rusbult et al. 1991). We do not consider
brand commitment as a behavior construct. It is a psychological pledge regarding future
behavior, a pre-cursor to brand-commitment behavior. Brand-commitment behavior is the
intentional or actual enactment of the decision or pledge to sustain brand-self relationship into
the future. As we discussed earlier, brand-commitment behavior takes various forms and could
be classified into two categories: brand loyal behavior and brand advocacy behavior. The
distinction between brand commitment and brand-commitment behavior is important in that the
relative predictive ability of the powerful brand influential and brand attachment for the two
outcome constructs are different. We will discuss the outcome of brand-commitment behavior in
the subsequent section.
Separation distress and brand commitment are theoretical indictors of brand attachment.
As strongly brand-attached consumers, powerful brand influentials should experience distress
when they face separation from the brand. They also should have high brand commitment.
Nevertheless, brand attachment is the driver of separation distress and brand commitment, not
the powerful brand influential construct.
As such, we anticipate that
P7: Brand attachment predicts separation distress (P7a) and brand commitment (P7b)
better than does the powerful brand influential construct.
Impact on brand loyal behaviors and actual purchase. Earlier, we proposed that the
powerful brand influential should be discriminable from brand attachment by a higher
engagement level of the first brand-commitment behavior category (i.e., brand word-of-mouth
and advocacy behavior). In this section, we discuss the second category of brand-commitment
behavior (i.e., intentions to conduct brand loyal behavior and actual purchase behavior).
POWERFUL BRAND INFLUENTIALS 51

Our prediction of whether the powerful brand influential and brand attachment have
differential impact on brand loyal behavior is exploratory. Both constructs significantly predict
brand loyal behavior. The adult attachment literature shows that attachment leads to loyal
behaviors toward an attached figure. People remain loyal to the partner, resist competing
alternatives, and forgive mishaps of the partner (Drigotas and Rusbult 1992; Johnson and
Rusbult 1989; McCullough 1998). The marketing literature reveals brand attachment as a
powerful predictor of brand loyal behaviors. Strong brand attachment leads to high investment of
personal financial and time resources and high engagement of various brand loyal behaviors
(e.g., purchasing, repurchasing, waiting to purchase, switching from competitor to the brand,
paying a price premium) (Esch et al. 2006; Park et al. 2010; Thomson et al. 2005). In particular,
Park et al. (2010) empirically show that brand attachment not only significantly predicts brand
loyalty intention but also actual purchase. Actual purchase behavior is more difficult- to-enact
than loyalty intention. The powerful brand influential construct also significantly predicts brand
loyal behavior. Popular press and academic work provide prevailing evidence that powerful
brand influentials are first of all loyal consumers and repeated purchasers of a brand (Carr 1996;
McConnell and Huba 2003; Muñiz and O’Guinn 2001; Pimentel and Reynolds 2004; Reichheld
2003; Yaverbaum et al. 2006).
We have little literature to draw from regarding whether the powerful brand influential
and brand attachment predict brand loyal behavior to a significantly different extent. In Park et
al.’s (2006) hierarchy of brand attachment and behaviors (see Appendix A), powerful brand
influential could be placed in the second (i.e., high allocation of self-image resources but
low/moderate allocation of personal discretionary resources) or third scenario (i.e., high
allocation of self-image resources and personal discretionary resources). This hierarchy would
POWERFUL BRAND INFLUENTIALS 52

suggest that powerful brand influentials do not necessarily have a high engagement level of
brand loyalty behaviors. By contrast, other strongly brand-attached consumers could only be
positioned in the first scenario (i.e., low allocation of self-image resources but high allocation of
personal discretionary resources). Correspondingly, they should have a high engagement level of
brand loyal behavior. If this is true, when simultaneously estimated, we would see that brand
attachment better predicts brand loyal behavior than does the powerful brand influential
construct. Furthermore, a higher predictive powerful of brand attachment should be more
pronounced for actual purchase behavior than intention to conduct brand loyal behavior. This is
because the former involves a higher level of actual resource allocation than does the latter.
Nevertheless, we speculate that powerful brand influentials are highly engaged in both
brand advocacy behavior and brand loyal behavior. They correspond primarily to the third
scenario of high brand attachment in Park et al.’s (2006) hierarchy. As mentioned earlier, brand
advocacy behavior demands requirements that are unique from brand loyal behavior. Powerful
brand influentials fulfill these requirements while other strongly brand-attached consumers do
not necessarily do so. We believe that powerful brand influentials differentiate from other
strongly attached consumers on a motivational, opportunity and ability basis.
From the motivational perspective, brand word-of-mouth and advocacy behavior involves
the motivation to influence others’ brand decisions. Some strongly brand-attached consumers
might lack of the motivation to influence other’s brand decisions. These consumers focus on
fostering their own brand-self relationship. They fully satisfied their brand-self relational needs
by engaging in brand loyal behavior. They do not feel the need to influence others’ brand
decisions. These consumers should engage in high level of brand loyal behavior and actual
purchase since they rely heavily on this behavior category to fulfill their brand-self relational
POWERFUL BRAND INFLUENTIALS 53

needs. By contrast, powerful brand influentials are those who are motivated to go the extra mile
to influence others as additional ways to satisfy their needs for expanding brand-self relationship
(McConnell and Huba 2003; Muñiz and O’Guinn 2001; Pimentel and Reynolds 2004). They are
not fully satisfied by a high engagement of brand loyal behavior alone. They have a similarly
high level of intention to conduct brand loyal behavior and actual purchase behavior as other
strongly brand-attached consumers.
From the opportunity perspective, some strongly brand-attached consumers might lack of
the opportunity to engage in brand word-of-mouth and advocacy behaviors due to constraints of
financial, time and social resources. These consumers have limited personal resources to share
with the brand (e.g., limited financial resources to promote a brand community on their own
expense, limited time or social resources to advocate the brand). Prior studies suggest that
influential consumers tend to be relatively affluent (Weimann 1994, Yahoo! and comScore
Networks 2006) and more socially-connected (Goldsmith, Flynn, and Clark 2012; Weimann
1994; Yahoo! and comScore Networks 2006) than non-influential consumers. Powerful brand
influentials might be more resourceful than other strongly brand-attached consumers. Therefore,
powerful brand influentials should have a similarly high level of intention to conduct brand loyal
behavior as strongly attached consumers who are limited in personal resources. However, their
actual purchase behavior should be higher.
From the ability perspective, some strongly brand-attached consumers might lack of the
ability to influence others’ brand decisions due to personality, self-concept, knowledge, and skill
reasons. As noted earlier, they might have low emotional expressiveness, which hinders their
propensity to publically exert influence and actual ability to be influential. Factors such as
introversion, low self-esteem, and lack of skills might also inhibit one’s ability (self-perceived or
POWERFUL BRAND INFLUENTIALS 54

real) to influence others. Previous literature suggests that influential consumers tend to be
extroversive, confident, more educated and skillful at technologies and means of information
searching (Goldsmith, Flynn, and Clark 2012; Matzler et al. 2007; Weimann 1994; Yahoo! and
comScore Networks 2006). Comparing to strongly attached consumers who are limited in ability
and skills, powerful brand influentials should have a similarly high level of intention to conduct
brand loyal behavior. However, their actual purchase behavior might be higher. They might be
more skillful at accessing brand’s information (e.g., sales, promotions, investment opportunities)
and therefore purchase and invest in the brand more extensively.
Considering these reasoning, we speculate that powerful brand influentials largely
correspond to the third scenario of high attachment proposed by Park et al. (2006). They
represent the ultimate level of a brand’s desirable consumers. If this is true, the powerful brand
influential and brand attachment, when simultaneously estimated, will contribute independently
significant predictions of intention to perform brand loyal behavior and actual purchase. By our
reasoning, it seems plausible that powerful brand influential predicts actual purchase behavior
better than does brand attachment. This prediction however, is highly speculative. We refrain
from making the prediction about whether the two constructs will differ significantly in
predictive ability for actual purchase behavior.
Hence, we propose that
P8: The powerful brand influential construct and brand attachment both contribute
independently significant prediction of intention to conduct brand loyal behavior.
P 9: The powerful brand influential construct and brand attachment both contribute
independently significant prediction of actual purchase behavior.
POWERFUL BRAND INFLUENTIALS 55

The powerful brand influential vs. brand attachment on positive emotional
responses related to a positive brand event. We explore potential differentiation between the
powerful brand influential and brand attachment regarding emotional outcomes. Brand
attachment is primarily characterized by positive emotions (Thomson et al. 2005). Hence, we are
particularly interested in positive emotions. We discuss whether the powerful brand influential
construct and brand attachment predict different emotional responses toward a positive brand
event. Drawing from literature on appraisal theory of emotions and brand attachment, we explore
two general categories of positive emotions: positive self-conscious emotions and challenge
emotions.
According to appraisal theory, self-conscious emotions (positive or negative) arise from
attributing an action or outcome to the self (Lazarus 1991; Roseman 1984, 1991, 2001; Scherer
2001; Tracy and Robines 2004; Weiner 1985). Positive self-conscious emotions (e.g., pride and
positive self-esteem feelings) are emotional responses when a situation is congruent with
personal goals (the primary appraisal dimension of goal congruence) and the causal agency is the
self (the secondary appraisal dimension of causal agency) (Roseman 1984, 1991). In a goal
congruent situation such as a positive event, an individual could attribute the situation to
him/herself (Self-Agency), to others (Other-Person-Agency), or to luck and circumstances
(Circumstance-Agency) (Roseman 1984, 1991). The more an individual attributes oneself as
responsible for a positive event or outcome, the greater he/she experiences pride and positive
self-esteem feelings (Tangney 1999; Tracy and Robines 2004, 2007; Weiner 1985).
We expect that powerful brand influentials experience a higher level of positive self-
conscious emotions than do other strongly brand-attached consumers. First, powerful brand
influentials are highly engaged in influencing others’ brand decisions, advocating the brand, and
POWERFUL BRAND INFLUENTIALS 56

participate in brand events. Through these frequent activities, they have more opportunities to
evaluate self-performance and attribute achievements to the self (e.g., a positive brand outcome
occurs because of “my” abilities and efforts). Correspondingly, they have more opportunities to
experience positive self-conscious emotions than other strongly brand-attached consumers.
Second, powerful brand influential tend to be influential on others. They are likely to receive
frequent social rewards from others and develop a high sense of self-confirmation (e.g., I am
capable, persuasive and influential on others; others acknowledge and appreciate my authority,
efforts, and contribution to the brand). They are more likely to experience positive self-conscious
emotions than other attached consumers since they receive more social rewards and have higher
sense of self-confirmation. Third, self-conscious emotions are social emotions. They are more
likely to develop from social settings (Keltner and Buswell 1997; Tracy and Robins 2004).
Moreover, they reinforce social behaviors and motivate further attainment of social goals
(Tangney and Dearing 2002). Powerful brand influentials not only are more likely to experience
these social emotions but also more motivated by the same. Fourth, more experiences of positive
self-conscious emotions should enhance one’s awareness and expressiveness of the same. Hence,
powerful brand influentials are more likely to recognize, label, and express positive self-
conscious emotions than other strongly brand-attached consumers.
Last, powerful brand influentials and other strongly attached consumers might experience
positive self-emotional responses toward the same brand event through different mechanisms.
According to the “working model” of attachment, the attachment entity develops into a core
component of one’s self-representation (Bowlby 1973; Collins 1996; Collins and Read 1994).
The self-representation of brand-attached consumers contains rich brand-self memories,
thoughts, and feelings linking the self with the brand (e.g., Berman and Sperling 1994; Collins
POWERFUL BRAND INFLUENTIALS 57

1996; Collins and Read 1994; Mikulincer and Shaver 2007; Park et al. 2010). Whenever a brand
is inferred, the contents of self-representation are automatically retrieved and become highly
salient (Park et al. 2009; Park et al. 2010). The self-representation of powerful brand influentials
contains rich histories of self-achievement memories, thoughts, and positive self-conscious
emotions linked with the brand. When a brand event occurs, these contents are immediately
retrieved leading to immediately experiences of intense self-conscious emotions. Therefore,
powerful brand influentials might experience more spontaneous positive self-conscious emotions
in response to a brand event via a direct mechanism (i.e., “schema-triggered affect”, Fiske and
Pavelchak 1986). By contrast, other strongly brand-attached consumers might experience these
emotions less spontaneously via the indirect mechanism of cognitive appraisals (see Collins and
Read 1994; Collins and Allard 2001 for discussion on the two pathways of how attachment
working models shape emotional responses).
To sum up, we anticipate that the powerful brand influential construct is differentiated
from brand attachment by a better predictive power for positive self-conscious emotions in
response to a positive brand event.
The second type of positive emotions of our interest is challenge emotions. According to
appraisal theory, challenge emotions (e.g., eagerness, hopefulness, and excitement) are emotional
consequences of challenge appraisals (Folkman and Lazarus 1985; Lazarus 1991; Lazarus and
Folkman 1984). Challenge appraisals are evaluative consequences of primary appraisal of a
situation. Through primary appraisal, a person evaluates whether a situation affects his or her
well-being by judging what is happening against the person’s enduring beliefs, valued goals, or
what the person desires (Lazarus and Folkman 1984). Primary appraisal yields three types of
evaluative consequences: (1) irrelevant, (2) benign-positive, and (3) stressful. In the current
POWERFUL BRAND INFLUENTIALS 58

context, a brand event might be appraised as (1) irrelevant when the event carries no
implications for a person’s goals and the person has no stake in the possible outcomes of the
event, (2) benign-positive when the event implies only good outcomes and the person does not
perceive any uncertainty that good outcomes will happen, and (3) stressful when the event
demands taxing resources of a person and the person has significant investment in the possible
outcomes (possibly good or bad) (Folkman and Lazarus 1985; Lazarus 1999; Lazarus and
Folkman 1984). Challenge appraisals occurs when a situation is construed as stressful but
containing also opportunities for success, the potential for gain, growth, mastery, or learning
(Lazarus 1991; Lazarus and Folkman 1984; Skinner and Brewer 2002).
Challenge appraisals will not happen when a brand event does not demand one’s
allocation of resources to handle or cope with the demand. In other words, when a brand event is
construed as irrelevant or benign-positive, a person will not conduct challenge appraisals or
experience subsequent challenge emotions. Specifically, when the brand event is appraised as
irrelevant, the person will experience a lack of emotions (Folkman and Lazarus 1985). When the
brand event is appraised as benign-positive, the person will experience pleasurable emotions
(e.g., joy, happiness, peacefulness, and love) (Folkman and Lazarus 1985; Lazarus 1999; Lazarus
and Folkman 1984). Moreover, challenge appraisals only occur when a person judges a stressful
situation as providing opportunities for success or the potential for gain. When a person
evaluates a stressful situation as signaling the potential for harm or loss more than the potential
for gain and success, threat appraisals are more likely to happen than challenge appraisals.
While challenge appraisals elicit positive emotions, threat appraisals evoke negative emotions
(e.g., fear, anxiety, and worry). Therefore, positive interpretations of a stressful event are crucial
for challenge appraisals and subsequent challenge emotions.
POWERFUL BRAND INFLUENTIALS 59

In sum, challenge appraisals and emotions are more likely when an event and the
outcomes (anticipated or actual outcomes) are more relevant to a person’s goals and when the
event is resource taxing but also signaling opportunities for success, the potential for gain,
growth, mastery, or learning. We expect that consumers are more likely to construe a brand event
as a challenge and experience a higher level of challenge emotions when brand attachment is
strong (versus weak). Strongly brand-attached consumers (including powerful brand influentials
and other strongly brand-attached consumers) should experience greater challenge emotions than
do low-attachment consumers.
First, a brand event and the possible outcomes carry greater relevance to the personal
goals of consumers when brand attachment is stronger. Applying self-expansion theory, strongly
brand-attached consumers have developed a sense of “oneness” with the attached brand and the
attached brand is experienced as part of the self (Aron, Aron, and Smollan 1992; Park et al.
2010). Goals, needs, and achievement outcomes of the brand are incorporated into the goal
structures of strongly attached-consumers. What a brand is motivated to achieve is experienced
as one’s own motivation and a brand’s outcomes are viewed as one’s own performance
outcomes. Inclusion of a brand’s goals and outcomes in one’s self-concept is a process that is
mutual to strongly brand-attached consumers and powerful brand influentials. They all have a
big stake in the possible outcomes before the event happens. They all experience a sense of
personal achievement after the brand succeeds. Therefore, strong brand attachment precludes
irrelevant appraisals and enhances challenge appraisals leading to intense challenge emotions in
response to a brand event.

POWERFUL BRAND INFLUENTIALS 60

Second, a brand event and the possible outcomes are more stressful or resource taxing to
consumers when brand attachment is stronger. As discussed earlier, brand attachment motivates
consumers to devote personal resources into the brand (Park et al. 2006, 2007; Park et al. 2010;
Thomson et al 2005). During the course of a brand event, strongly brand-attached consumers
should be more engaged in expending resources of their own to facilitate the brand’s goals.
Although powerful brand influentials and other strongly brand-attached consumers differ in the
types of personal resources allocated to a brand, they share a higher level of resource allocation
than low-attachment consumers. Therefore, a brand event is more resource taxing to strongly
brand-attached consumers since they are more engaged in allocating personal resources to the
event. Moreover, strong attachment implicates heightened level of vigilance and stress towards
potential threats related to the attached entity (Bowlby 1973; Mikulincer and Shaver 2005).
Previous attachment literature suggests that attached people adopt hyperactivating strategies to
alleviate separation distress, which involve constant monitoring of potential threats related to the
attached entity and lowered threshold for detecting cues of threats (Bowlby 1973; Mikulincer
and Shaver 2005). In a brand event context, consumers are more likely to detect and monitor
cues of threats to a brand success when brand attachment is stronger. These hyperactivating
strategies tend to be resource taxing and stressful. Since a brand event is more likely to be
appraised as stressful when brand attachment is stronger, the likelihood of benign-positive
appraisals is lower. We expect that powerful brand influentials and other strongly brand-attached
consumers perceive a brand event as more stressful than do low-attachment consumers.

POWERFUL BRAND INFLUENTIALS 61

Third, consumers are more likely to appraise a brand event as providing opportunities for
success and the potential for gain, growth, mastery, or learning when brand attachment is
stronger. Strongly brand-attached consumers have more to gain from a brand event and
successful outcomes than other consumers. A brand event provides strongly-attached consumers
with additional opportunities (e.g., campaign for the brand and donate to the brand) to fulfill their
needs for expanding brand-self relationship. A brand event also provides them with means of
supporting the brand and the future of it in a significant way. Furthermore, brand attachment
shapes how consumers interpret a brand event and the possible outcomes. A brand’s goals and
achievement outcomes are included as part of the personal goals and performance of strongly
brand-attached consumers. During a brand event, opportunities and success of a brand are likely
to be experienced as their own. Last, attachment induces selective attention and biases
information processing related to the attached entity (Collins and Allard 2001; Jain and
Maheswaran 2000). Strongly attached-consumers are more likely to attend to positive
information, resist negative information, and process information about the brand in a positive
way to be consistent with their beliefs about the brand (Jain and Maheswaran 2000). They are
more likely to focus on positive aspects (e.g., opportunities for success, the potential for gain,
mastery, or learning) associated with a brand event and more optimistic about the possible
outcomes. Although a brand event is more stressful and resource taxing to strongly brand-
attached consumers, they are more likely to conduct challenge appraisals than threat appraisals.
Correspondingly, they should experience greater challenge emotions than threat emotions.

POWERFUL BRAND INFLUENTIALS 62

Based on the discussion, strongly brand-attached consumers (including powerful brand
influentials and other strongly brand-attached consumers) should be distinguishable from other
consumers by greater challenge emotions in response to a positive brand event. To sum up our
prediction for positive self-conscious emotions and challenge emotions, we propose that
P10: The powerful brand influential construct predicts positive self-conscious emotions
related to a positive brand event better than does brand attachment (P10a) but both
constructs contribute independently significant prediction of challenge emotions (P10b).
The powerful brand influential vs. brand attachment on the level of actual word-of-
mouth engagement and types of word-of-mouth following a positive brand event. In the last
section of comparison between the powerful brand influential and brand attachment, we explore
behavioral responses following a positive brand event. In particular, we focus on comparing
powerful brand influentials and other strongly brand-attached consumers on the level of actual
word-of-mouth engagement (how much they are engaged in word-of-mouth activities) and types
of word-of-mouth (what they would talk about) in response to a positive brand event.
By our conceptualization, powerful brand influentials have an overall higher engagement
level of word-of-mouth than other strongly brand-attached consumers. Word-of-mouth
engagement constitutes an important component of their brand-self relationship – it is the means
by which they fulfill their needs for brand-self relationship, their habitual ways of sustaining and
expanding their brand-self relationship. By contrast, other strongly brand-attached consumers
tend to adopt inwardly-oriented means as their habitual ways of engaging with the brand.
Following a brand success, we expect word-of-mouth as a predominant type of responses
employed by powerful brand influentials. Moreover, they find the general act of word-of-mouth
intrinsically more enjoyable than do other strongly brand-attached consumers. They are more
POWERFUL BRAND INFLUENTIALS 63

likely to adopt the same as means of gratification following a brand success. Last, the validity of
the powerful brand influential construct will be evidenced by showing that it predicts actual
word-of-mouth engagement better than does brand attachment in a specific brand context (versus
in general). Correspondingly, we propose that the powerful brand influential construct predicts
the level of actual word-of-mouth engagement following a positive brand event better than does
brand attachment.
Powerful brand influentials should be distinguishable from other strongly brand-attached
consumers not only by a higher engagement level of word-of-mouth but also by types of word-
of-mouth. We posit that powerful brand influentials are more likely to engage in action-focused
brand advocacy voice and emotion-focused voice. We use action-focused brand advocacy voice
to capture word-of-mouth responses to a positive brand context that involve actively and
constructively talking about a brand with the intent to advocate the brand. Powerful brand
influentials are defined as those who are highly engaged in influencing others’ brand decisions
through word-of-mouth about the brand. Advocating the brand is a key purpose of their word-of-
mouth engagement. Correspondingly, their word-of-mouth should reflect a brand advocacy
nature. Their word-of-mouth should also reflect an active (versus passive) nature as well as
constructive (versus destructive) impact on relationship with the brand. We expect that they
actively attempt to influence others and reach out to others with word-of-mouth aiming at doing
something that is constructive about the brand and their relationship with the brand.
Active and constructive word-of-mouth resembles the concept of voice in the literature on
close relationship. Previous research has established a typology of people’s responses to
dissatisfaction incidents in close relationship. The typology is differentiated along two
dimensions: activity/passivity and constructiveness/destructiveness (Ping 1993; Rusbult and
POWERFUL BRAND INFLUENTIALS 64

Zembrodt 1983). Voice captures behavioral responses of actively and constructively expressing
dissatisfaction with the intent to maintain or revive the relationship conditions. Although it
pertains to events of dissatisfaction with a close relationship, voice response captures an active
and constructive nature of word-of-mouth. On the passivity end of the dimension is the concept
of loyalty. Loyalty captures behavioral responses of remaining passively loyal to the relationship
by optimistically waiting for relationship conditions to improve without actually doing
something about the relationship. Both voice and loyalty responses are constructive (i.e.,
constructive regarding the impact of the response on the relationship). Loyalty response,
however, does not involve talking and thus is less relevant to word-of-mouth. Although the
typology is originally developed to understand dissatisfaction, it has relevance to our research
context. The typology pertains to how people react to attachment relationship scenarios and has
been applied to explain how consumers with different attachment styles respond to transgression
in consumer-brand relationships. Additionally, voice response and word-of-mouth both involve
talking as the means of responding to relationship incidents. To make the distinction from voice
response, we apply a new concept (i.e., action-focused brand advocacy voice) to denote a similar
response type in a positive relationship context. We propose that powerful brand influentials are
more likely to engage in action-focused brand advocacy voice than other strongly brand-attached
consumers in response to a positive brand event.
We use emotion-focused voice to capture word-of-mouth responses to a positive brand
context that involve talking about a brand with the intent to regulate emotional responses. We
draw from previous literature on coping and particularly emotion-focused coping (Folkman and
Lazarus 1985; Lazarus 1991, 1999; Lazarus and Folkman 1984). Emotion-focused coping
strategies (e.g., emotional support and emotional venting strategies) include talking or expressing
POWERFUL BRAND INFLUENTIALS 65

to others about personal emotions and experiences as ways of reacting to stressful situations
(Duhachek 2005; Duhachek and Iacobucci 2005; Lyubomirsky, Sousa, and Dickerhoof 2006).
Emotion-focused coping apply not only to a stressful and negatively appraised situation (i.e., a
threat) but also to a stressful but positively appraised situation (i.e., a challenge). As discussed
earlier, a dispositional differentiation between powerful brand influentials and other strongly
brand-attached consumers is that the former has a higher emotional expressiveness than the
latter. Powerful brand influentials are more likely to engage in word-of-mouth activities as
means of socially expressing their emotions in addition to advocating the brand. Their natural
propensity to express emotions socially should be evident following a positive brand event.
Therefore, we propose that powerful brand influentials are more likely to engage in emotion-
focused voice than other strongly brand-attached consumers in response to a positive brand event.
To sum up our prediction for word-of-mouth responses to a positive brand event, we
propose that
P11: The powerful brand influential construct predicts the level of actual word-of-mouth
engagement following a positive brand event better than does brand attachment (P11a).
The powerful brand influential construct predicts action-focused brand advocacy voice
(P11b) and emotion-focused voice (P11c) following a positive brand event better than
does brand attachment.

POWERFUL BRAND INFLUENTIALS 66

Chapter 5: Differentiating the Powerful Brand Influential from Brand Attitude Strength
and Brand Satisfaction
Brand attitude strength and brand satisfaction are two other brand-level constructs that
relate to the powerful brand influential construct. We define brand attitude strength in the same
way as recently conceptualized by Park et al. (2010). Brand attitude strength incorporates both
attitude valence (i.e., the degree of positivity and negativity of the evaluation towards an attitude
objective) and the confidence or certainty of the valence (i.e., the extent to which the attitude
valence is considered valid). It is operationalized as brand attitude valence weighted by the
confidence/certainty with which it is held. Park et al. (2010) posit that brand attitude strength is
conceptually more similar to brand attachment than brand attitude valence alone. Hence, it is
more rigorous to differentiate brand attitude strength than brand attitude valence from brand
attachment. By the same token, we intend to differentiate the powerful brand influential
construct from brand attitude strength instead of attitude valence.
5.1. The Powerful Brand Influential is Related to Brand Attitude Strength and Brand
Satisfaction
Researchers have shown that brand attitude (attitude valence or attitude strength) and
brand satisfaction are significantly related yet distinctly different from brand attachment (Park et
al. 2010; Thomson et al. 2005). As brand attachment defines the powerful brand influential, the
powerful brand influential should also be related to but distinct from brand attitude strength and
brand satisfaction. The relatedness and differentiation between brand attachment and the two
brand constructs (Park et al. 2010; Thomson et al. 2005) also largely explain how the powerful
brand influential construct is related to yet distinct from the two. Powerful brand influentials are
likely to have a strong favorable attitude towards the brand and be satisfied with the brand they
POWERFUL BRAND INFLUENTIALS 67

are highly engaged with. However, they are not simply those who favor and are satisfied with the
brand.
As such, we expect that,
P12: The powerful brand influential construct is significantly related to but is distinctly
different from the construct of brand attitude strength (P12a) and brand satisfaction
(P12b).
5.2. Differentiating the Powerful Brand Influential from Brand Attitude Strength and
Brand Satisfaction on Brand Behavioral Outcomes
Previous literature has highlighted that brand attitude strength and brand satisfaction
predict various behavioral outcomes that are desirable for a brand (Fazio, Powell, and Williams
1989; Holmes and Lett 1977; LaBarbera and Mazursky 1983; Oliver 1980; Oliver, Rust, and
Varki 1997, Park et al. 2010; Petty, Haugtvedt, and Smith 1995; Priester et al. 2004; Swan and
Oliver 1989; Thomson et al. 2005). Some key behavioral outcomes of brand attitude strength and
satisfaction are brand loyal behaviors including both intentions and actual purchase (e.g.,
purchase, repurchase, brand consideration, brand choice, willingness to pay a price premium),
and advocacy behavior (e.g., recommending the brand, defending the brand, promoting the brand,
using products with the brand logo). Brand advocacy behavior, brand loyal behavior, and actual
purchase are relevant outcomes of the powerful brand influential construct as well. Therefore, we
compare predictive ability of the powerful brand influential to that of brand attitude strength and
brand satisfaction for these shared behavioral outcomes.

POWERFUL BRAND INFLUENTIALS 68

Impact on brand advocacy behaviors. Many studies show that brand attitude strength
and brand satisfaction significantly predict brand word-of-mouth and advocacy behavior
(Holmes and Lett 1997; Park et al. 2009; Park et al. 2010; Swan and Oliver 1989; Thomson et al.
2005). Brand advocacy is a core behavioral manifestation of the powerful brand influential
construct. The ability to predict brand advocacy behavior beyond extant constructs is crucial for
demonstrating the value and validity of the powerful brand influential construct. The powerful
brand influential construct should be distinguishable from brand attitude strength and brand
satisfaction by a better predictive power for brand advocacy behavior.
In great details, Park et al. (2010) show that brand attachment predicts difficult behaviors
better than does brand attitude strength. Behaviors differ in the extent to which the enactment
expends personal resources including economic, social, psychological, temporal, or physical
resources (Park et al. 2010). Since the powerful brand influential is defined by strong brand
attachment, we expect that this construct also predicts difficult behaviors better than does brand
attitude strength. Park et al. (2010) show that brand advocacy behaviors such as promoting the
brand and defending the brand rank high on the hierarchy of behavior difficulty. These behaviors
require great allocation of personal resources spanning across not only financial and time
resources but social, psychological, and physical resources. Correspondingly, we expect that the
powerful brand influential construct predicts brand advocacy behavior better than does brand
attitude strength. As previously proposed, the powerful brand influential exceeds brand
attachment in predicting brand advocacy behaviors. Empirically, we expect that the powerful
brand influential construct predicts brand advocacy behavior better than does brand attitude
strength even when taking brand attachment into account.
POWERFUL BRAND INFLUENTIALS 69

Brand satisfaction also produces brand word-of-mouth (Holmes and Lett 1997; Swan and
Oliver 1989). It is noteworthy that there is insufficient evidence showing that brand satisfaction
influences other forms of brand advocacy behaviors such as promoting a brand community and
defending a brand to others. Brand satisfaction is primarily considered to be an evaluative
concept that is differentiated from brand attachment, which is an emotionally laden construct
(Mano and Oliver 1993; Thomson et al. 2005). Although empirical evidence shows that
satisfaction produces a tendency toward word-of-mouth, “it is the affect that stimulates the ‘web
of word-of-mouth’ rather than satisfaction per se” (Westbook 1987, p. 267). Moreover, brand
attachment predicts positive word-of-mouth propensity beyond brand satisfaction. Carroll and
Ahuvia (2006) found that among consumers who were satisfied with a brand, those who feel
stronger emotional attachment to the brand engaged in more positive brand word-of-mouth.
Moreover, brand satisfaction is a brand-outcome concept and often transaction-specific
and situational. Satisfactory outcomes from transactions could stimulate positive brand word-of-
mouth that is situational and short-term in nature. Like brand attachment, the powerful brand
influential construct is a brand relationship concept that is long-term and sustained. The powerful
brand influential construct predicts enduring engagement in brand advocacy. Last, it is less likely
that transaction-driven satisfaction would lead to more resource-demanding advocacy behaviors
(e.g., promoting and defending a brand to others). Therefore, we propose that the powerful brand
influential construct better predicts brand advocacy behavior than does brand satisfaction.
Empirically, we also expect that the powerful brand influential construct predicts brand advocacy
behavior better than does brand satisfaction even when taking brand attachment into account the
prediction of brand attachment.

POWERFUL BRAND INFLUENTIALS 70

To summarize the discussion regarding brand advocacy behaviors, we propose that,
P13: The powerful brand influential construct predicts brand advocacy behavior better
than does brand attitude strength (P13a) and brand satisfaction (P13b).
Impact on brand loyal behaviors and actual purchase. Both brand attitude strength
and brand satisfaction significantly predict intention to conduct brand loyal behavior and actual
purchase behavior (Fazio et al. 1989; Oliver 1980; Oliver et al. 1997; Park et al. 2010; Petty et al.
1995; Priester et al. 2004; Thomson et al. 2005). Park et al. (2010) suggest that actual purchase is
a behavior that is more difficult- to-enact than are loyalty intentions. Actual purchase, that
involving repeated purchase behaviors, requires that consumers overcome potential challenges
for actualizing intentions into enactment (Park et al. 2010).
As theorized by Park et al. (2010), brand attachment better predicts actual purchase
behavior than does brand attitude strength, whereas brand attachment predicts intention to
perform less difficult behaviors similarly as brand attitude strength. They empirically found that,
when simultaneously estimated with brand attachment, brand attitude strength remained as a
significant predictor of intentions to perform brand loyal behaviors (i.e., intention to purchase, to
switch to the brand, to pay more) but became an insignificant predictor of actual purchase. Brand
attachment remained significant in predicting actual purchase. In light of these results, we expect
that the powerful brand influential and brand attitude strength constructs both bring
independently significant contributions to intention to conduct brand loyal behavior, whereas the
powerful brand influential construct better predicts actual purchase behavior than does brand
attitude strength.

POWERFUL BRAND INFLUENTIALS 71

Marketers have commonly understood brand satisfaction as a crucial driver of brand
loyalty, which is often reflected by intention to perform loyal behaviors and/or repeated purchase
(Anderson, Fornell, and Lehmann 1994; Jacoby and Kyner 1973; Jones and Sasser 1995;
Szymanski and Henard 2001). Increasingly research has shown that brand satisfaction has a
differential impact on loyal behavior depending on the nature of the consumer’s relationship with
the brand (Garbarino and Johnson 1999; McAlexander, Kim, and Roberts 2003). The
relationship marketing literature emphasizes that consumers want and form different
relationships with a brand and they differ in their relationships with a brand on a continuum of
translational to relationship orientations (Anderson and Narus 1991; Dwyer, Schurr, and Oh
1987; Jackson 1985). This line of theory is in agreement with our proposition that there exist
strongly brand-attached consumers and the powerful brand influential who have extreme levels
of relational bonds with a brand and who are differentiated from other consumers of a brand
(their relationships with a brand are less relational). Drawing from relationship marketing
theories, researchers have shown that satisfaction significantly produces intention to perform
loyal behaviors for consumers who hold more transactional-oriented relationships with a brand,
whereas satisfaction yields to commitment-related constructs (e.g., attachment, brand community
integration) and becomes an insignificant driver of loyal intentions for consumers who hold more
relational-oriented affiliation with the brand (Garbarino and Johnson 1999; McAlexander et al.
2003).
We derive two implications from the body of relationship marketing research to
differentiate the powerful brand influential from brand satisfaction regarding predictive power
for intentions to perform loyal behavior and actual purchase, respectively. First, brand
satisfaction should be a significant driver of intention to perform loyal behavior when estimated
POWERFUL BRAND INFLUENTIALS 72

simultaneously with the powerful brand influential construct. This is because brand satisfaction
will be driven by a considerable population of consumers who are less experienced with a brand
and/or less relationally connected to a brand to significantly predict intention to perform loyal
behavior. At the same time, the powerful brand influential construct will also significantly
predict intention to perform loyal behavior driven by a population of consumers who have
formed relational bonds with the brand. Correspondingly, we expect that the powerful brand
influential and brand satisfaction constructs both bring independently significant contributions to
intention to conduct brand loyal behavior. Second, the powerful brand influential better predicts
actual purchase behavior than does brand satisfaction. Prior research exploring the differential
impact of brand satisfaction on loyal behaviors has been limited to behavioral intentions or self-
reported patronage instead of actual purchase. Nevertheless, since actual purchase accounts for
repeated purchase behaviors during a period of time, it should be significantly less driven by
consumers who are less experienced and/or less relationally connected to the brand than by
consumers who have developed strong relational bonds with the brand over the time. Hence, we
anticipate that the powerful brand influential construct predicts actual purchase better than does
brand satisfaction.
Though for different reasons, our predictions regarding brand attitude strength and brand
satisfaction in comparison to the powerful brand influential are in the same pattern. To
summarize, we propose that,
P14: The powerful brand influential construct and brand attitude strength both contribute
independently significant prediction of intention to conduct brand loyal behavior (P14a).
The powerful brand influential construct and brand satisfaction both contribute
independently significant prediction of intention to conduct brand loyal behavior (P14b).
POWERFUL BRAND INFLUENTIALS 73

P15: The powerful brand influential construct predicts actual purchase behavior better
than does brand attitude strength (P15a) and brand satisfaction (P15b).

POWERFUL BRAND INFLUENTIALS 74

Chapter 6: Overview of Studies 1 – 4
We conducted four studies to test the above noted propositions. Study 1 develops a scale
that maps the conceptual properties of the powerful brand influential construct. The scale aims to
capture consumers’ tendency to be influential consumers of a given brand. We also provide
preliminary evidence for the reliability of the scale. Studies 2, 3, and 4 demonstrate the stability
and robustness of the scale and demonstrate its convergent, discriminant and predictive validity
using different samples and brands across different product categories. Specifically, Study 2 tests
differentiations of the powerful brand influential construct from the market maven, opinion
leader, brand attachment and brand ambassador constructs. Study 3 repeats testing for key
differentiations of the powerful brand influential construct from all related constructs examined
in Study 2 excluding brand ambassador (see justification for exclusion later in Study1).
Additionally, it tests the differentiation of the powerful brand influential construct from brand
attitude strength and brand satisfaction. Finally, Study 4 advances further differentiations of the
powerful brand influential from the most overlapped construct, brand attachment and tests for
additional propositions regarding the two constructs.

POWERFUL BRAND INFLUENTIALS 75

Chapter 7: Study 1: Development and Purification of the Powerful Brand Influential Scale
(PBIS)
7.1. Item Generation

An initial set of 18 items was constructed to represent the conceptualization of the
powerful brand influential construct. The Powerful Brand Influential Scale (PBIS) taps each of
the four defining components of the construct. We constructed approximately equal numbers of
items (four to five items with reverse-coded items) to reflect each component. Item generation is
grounded in extant research and scales of influential consumers, brand attachment, and
interpersonal communication. For the component of high influence tendency on other consumers’
brand decisions, we included items to measure the level of motivation to exert personal influence
and the level of self-perceived effectiveness at influencing others’ brand decisions. For the
component of high engagement of brand word-of-mouth, we included items measuring the
tendency to seek various time, occasion, and audience opportunities to talk about the brand. We
also included an item measuring overall tendencies to seek opportunities to talk about the brand.
For the component of strong brand attachment, we measured the ability to appreciate and
maximize the emotional experiences in brand usage, the motivation to share brand emotional
experiences with others, and the degree to which brand attachment motivates brand
communication. For the component of passionate and emotionally evocative brand word-of-
mouth, we measured the degree of passionate and emotional communication style as well as
usage of passion-driven verbal messages and body language in brand communication.
All items were administered on a 9-point scale that ranged from 1 = strongly disagree to 9
= strongly agree. Marketing scholars with great expertise in the areas of branding, brand
attachment, interpersonal communication and influence reviewed these original items. The initial
POWERFUL BRAND INFLUENTIALS 76

set of 18 items of PBIS was then edited to eliminate ambiguous, double-barreled, and repetitive
statements and to ease comprehension. A revised 16-item scale resulted.
7.2. Objectives and Method
Study 1 was conducted to refine the conceptually developed 16-item PBIS using one
hundred sixty seven undergraduate students. The students were recruited from various business
majors to participate for class credit. We intended to choose a brand in a product category that is
highly relevant to the respondent population. The brand of Apple iPod was chosen to assess the
scale in a brand-specific context. The corresponding product category is MP3 players. In
addition to the 16-item PBIS, a set of criterion variables was included to select items that are
reliable at the same time keeping the scale parsimonious. To keep the number of questions
manageable, these criterion variables were administered to three independent sample groups (N
= 58, N = 55, N = 54) and the PBIS as administered to the whole sample (N=167). Items of the
PBIS as well as the presentation order of criterion variable measures were randomized.
7.3. Measurement Results of the PBIS
Three steps were conduced to purify the scale. First, coefficient alpha and item-to-total
correlations of the 16 items were acquired for reliability analyses. Items with substantially lower
item-to-total correlations were deleted if the deletion increased alpha (Churchill 1979). The
iterative process of assessing item-to-total correlations and alpha after deleting items yielded 13
remaining items (coefficient alpha = .92, item-to-total correlations ranged from .63 to .86). All
items are listed in Table 7-1. Then, Principle Component Analysis was conducted to test the
dimensionality of these remaining items. Similar to previous scales for existing types of
influential consumers (e.g., market mavens, opinion leaders), the PBIS was shown as a
unidimensional scale (with factor loadings ranging from .69 to .89).
POWERFUL BRAND INFLUENTIALS 77

TABLE 7-1. The Initial 13-Item Pool of the PBIS and Corresponding Defining Components

Item
Number
Items Represented Component
PBIS 1.
I want to change how other people think about or understand
this brand.
High influence tendency
on others’ brand
decisions
PBIS 2.
I want others to know about and appreciate the benefits of this
brand.
PBIS 3.
I am quite persuasive at convincing others how great this brand
is.
PBIS 4. I spend very little time talking about this brand to others.
R

High engagement of
brand word-of-mouth
PBIS 5.
I tend to talk about this brand to anyone who is near me or
willing to listen to me
PBIS 6. I talk about this brand whenever I find an opportunity to do so
PBIS 7.
I deeply appreciate the emotional experiences I’ve had from
using this brand
Strong brand attachment
PBIS 8.
I want people to have the same emotional experiences with this
brand as I do.
PBIS 9. I tell others about this brand because I am passionate about it.
PBIS 10.
When I talk to others about this brand, I tend to be excited and
emotional.
Passionate and
emotionally evocative
brand word-of-mouth
PBIS 11 I talk passionately about this brand to others.
PBIS 12
I often use colorful adjectives to describe this brand and my
experiences with it.
PBIS 13
One can tell how much I love this brand just by observing the
body language that I use when I talk about it.
Note:
Scale ranges from 1 (strongly disagree) to 9 (strongly agree).
R
Reverse coded items.

The final step involved further reducing the number of items to achieve a parsimonious
scale. We included four criterion variables and examined the correlations between each item and
the criterion variables to assess the contribution of an item to the total scale’s criterion validity
(Spector 1992). The first criterion variable was brand usage (How much do you use your MP3
player? 1 = Hardly at all to 7 = Extremely Much) and was expected to positively associate with
the powerful brand influential the same way as product category usage and opinion leaders
(Flynn et al. 1996). Only those respondents who own an iPod (N = 114) were used for the
correlation analysis. All the other criterion variables pertained to those who own and do not own
an iPod.
POWERFUL BRAND INFLUENTIALS 78

The second criterion variable was respondents’ self-ratings of the extent to which they
possess the characteristics of the powerful brand influential, a measure suggested by Bagozzi
(1993) for validity test. An overall description of the powerful brand influential was presented as
“Some people are very attached to a particular brand (e.g., the Apple iPod). But more than being
just emotionally bonded to the brand, they are also motivated to influence other people's beliefs
and decisions about the brand, and enjoy taking every opportunity they can to talk passionately
about it”. Respondents were asked two following questions: 1) How well does this description
characterize you and your relationship with the iPod? (1 = not well at all to 7 = 3xtremely well)
and 2) Does this description accurately characterize how you think, feel and act toward the iPod?
(1 = absolutely not to 7 = absolutely yes). The self-rating items were administered to all groups
of respondents and were averaged to form a composite score, which should strongly correlate to
the PBIS.
The third criterion variable was brand involvement and was assessed using Thomson et
al.’s (2005) 6-item short version of Personal Involvement Inventory (Zaichkowsky 1985). The
last criterion variable is brand advocacy behavior including the advocacy behaviors of
recommending, defending, and promoting the brand (Park et al. 2010). Each specific advocacy
behavior was measured by two questions: respondents’ self-report of whether they have
performed the specific advocacy behavior in the past (0 = no and 1 = yes) and their intentions to
perform the behavior in the future (1 = not at all to 9 = definitely). The brand involvement scale
(N = 54) and the brand advocacy behavior scale (N = 55) were administered to two separate
groups of respondents. Brand involvement and brand advocacy behavior are expected to
positively associate with the PBIS.
POWERFUL BRAND INFLUENTIALS 79

Following procedures for item selection by previous research (Flynn et al. 1996;
Goldsmith and Hofacker 1991; Messick 1981), we examined the average correlations between
each PBIS item and the four criterion variables. We selected the items with strong average
correlations. At the same time, we intended to balance the number of items measuring each of
the four defining components of the powerful brand influential. The final PBIS scale consists of
six items with one to two items measuring each defining component (see Table 7-2). These six
items were strongly correlated with all criterion variables (see Table 7-3).

TABLE 7-2. The Final PBIS and Corresponding Defining Components

Item Number PBIS Items Represented Component
PBIS 1.
I am quite persuasive at convincing others how
great this brand is. High influence tendency on others’
brand decisions
PBIS 2.
I want to change how other people think about
or understand this brand.
PBIS 3.
I talk about this brand whenever I find an
opportunity to do so.
High engagement of brand word-of-
mouth
PBIS 4.
I tell others about this brand because I am
passionate about it.
Strong brand attachment
PBIS 5.
I deeply appreciate the emotional experiences
I’ve had from using this brand.
PBIS 6.
When I talk to others about this brand, I tend to
be excited and emotional.
Passionate and emotionally evocative
brand word-of-mouth
Note:
Scale ranges from 1 (strongly disagree) to 9 (strongly agree).

POWERFUL BRAND INFLUENTIALS 80

TABLE 7-3. Correlations of the PBIS with Criterion Variables in Pilot Study

Item
Number
Brand
Usage
(N=114)
Self-Ratings as the Powerful
Brand Influential
(N=167)
Brand
Involvement
(N=54)
Brand
Advocacy
Behaviors
(N=55)
Average
a

PBIS 1. .30** .62** .41** .62** .49
PBIS 2. .31** .58** .33* .51** .43
PBIS 3. .33** .52** .39** .48** .43
PBIS 4. .35** .63** .57** .61** .54
PBIS 5. .34** .49** .57** .39** .45
PBIS 6. .35** .61** .49** .49** .49
Note:
a
is the average score of the correlations between each PBIS item and the four criterion variables.
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.

The mean score of the 6-item PBIS was 4.08 (range = 1 - 9, standard deviation = 1.67).
Mean scores of individual PBIS items ranged from 3.0 to 4.92 with standard deviations ranged
from 1.94 to 2.18. Kolmogorov-Smirnov one-sample test was conducted to test the normality of
the scale and yielded z score of .76 (two-tailed p = .605), which supported the null hypothesis
that the PBIS has a normal distribution. Similar to opinion leadership scale (Flynn et al. 1996),
the powerful brand influential construct is normally distributed. Principle Component Analysis
was conducted on the selected six items and confirmed the unidimensional nature of the PBIS.
The scale explained 65.86% of the total variance (factor loadings ranged from .74to .90) (see
Table 7-4). The scale yielded a coefficient alpha of .90 with item-to-total correlations ranging
from .63 to .84. These results showed that the 6-item PBIS is a satisfactory measure for the
powerful brand influential construct.

POWERFUL BRAND INFLUENTIALS 81

TABLE 7-4. Exploratory Factor Analysis Factor Loadings of the PBIS (Study 1 – 4)

Item
Number
PBIS Items
Factor Loading (N=167)
Study 1
(N=167)
Study 2
(N=121)
Study 3
(N=697)
Study 4
Pre-
election
(N=182)
PBIS 1.
I am quite persuasive at convincing
others how great this brand is.
.75 .83 .99 .90
PBIS 2.
I want to change how other people think
about or understand this brand.
.78 .71 .99 .71
PBIS 3.
I talk about this brand whenever I find
an opportunity to do so.
.82 .87 .98 .89
PBIS 4.
I tell others about this brand because I
am passionate about it.
.90 .91 .99 .83
PBIS 5.
I deeply appreciate the emotional
experiences I’ve had from using this
brand.
.74 .81 .95 .92
PBIS 6.
When I talk to others about this brand, I
tend to be excited and emotional.
.88 .83 .99 .93

POWERFUL BRAND INFLUENTIALS 82

Chapter 8: Study 2: The Powerful Brand Influential is Distinct from Market Maven,
Opinion Leader, Brand Attachment, and Brand Ambassador
8.1. Objective and Method

Study 2 has three objectives. First, we examine the stability and reliability of the PBIS
using a different sample of student respondents. Second, we test the convergent and discriminant
validity of the PBIS, providing evidence that the powerful brand influential is related to but
empirically differentiable from other influential consumers (i.e., opinion leaders and market
mavens) and from two brand constructs (i.e., brand attachment and brand ambassador). Last, we
test predictive validity of the PBIS and a set of propositions regarding to differentiations between
the powerful brand influential and the four existing constructs.
We again chose the Apple iPod brand but collected new data after more updated versions
of the iPod (i.e., iPod Shuffle and iPod Touch) had penetrated the MP3 market. One hundred
twenty one undergraduate students answered a survey containing the 6-item PBIS, measures of
market maven, opinion leader, brand attachment and brand ambassadors, and a set of scales to
assess validity. Items of the PBIS as well as the presentation order of additional scales (of related
consumers groups) were randomized.
8.2. Measurement Results of the PBIS
Scale norm. The distribution of the PBIS of Study 2 falls within similar range as Study 1.
The mean score of the PBIS was 4.1 (range = 1 - 9, standard deviation = 1.95). Mean scores of
individual PBIS items ranged from 3.45 to 4.55 with standard deviations ranged from 2.17 to
2.59. Kolmogorov-Smirnov one-sample test for normality yielded a z score of .77 (two-tailed p
= .596) and supported normal distribution. The coefficient alpha was .91 with item-to-total
correlations ranging from .60 to .86.
POWERFUL BRAND INFLUENTIALS 83

Scale dimensionality. Two steps were conducted to confirm the dimensionality of the
PBIS. In the first step, Exploratory Factor Analysis (hereafter EFA) was conducted on the PBIS.
Principle Component Analysis produced a single-factor solution with 68.81% explained total
variance (factor loadings ranged from .71 to .91) (see Table 7-4, results for Study 2).
In the second step, Confirmative Factor Analysis (hereafter CFA) was conducted on the
variance-covariance matrix using LISREL 8.80 (Jöreskog and Sörbom 1996). Table 8-1 provides
the standardized factor loadings, goodness-of-fit statistics, construct reliability and average
variance extracted. All standardized factor loadings were significant (p < .001). The average
standardized factor loading was .80. Although the chi-square test was significant, this test value
tends to be sensitive to sample size (Jöreskog and Sörbom 1993). All the other fit statistics
provided converging evidence for a good fit of the PBIS: NFI, CFI, and GFI all exceeded the .90
cutoff value (Bentler 1992); the SRMR was less than the .05 recommendation (Byrne 1998); and
the construct reliability well exceeded the criterion of .70 with the average variance extracted
greater than the criterion of .50 (Fornell and Larcker 1981). Overall, the CFA results confirmed
the unidimensionality of the PBIS and reflected a good measurement model for the PBIS.

POWERFUL BRAND INFLUENTIALS 84

TABLE 8-1. Confirmatory Factor Analysis Results of the PBIS (Study 2 – 4)

Item
Number
PBIS Items
Standardized Factor Loadings
Study 2
(N=121)
Study 3
(N=697)
Study 4 Pre-
election
(N=182)
PBIS 1.
I am quite persuasive at convincing others
how great this brand is.
.79 .99 .83
PBIS 2.
I want to change how other people think
about or understand this brand.
.66 .99 .64
PBIS 3.
I talk about this brand whenever I find an
opportunity to do so.
.83 .98 .89
PBIS 4.
I tell others about this brand because I am
passionate about it.
.91 .99 .79
PBIS 5.
I deeply appreciate the emotional experiences
I’ve had from using this brand.
.78 .92 .91
PBIS 6.
When I talk to others about this brand, I tend
to be excited and emotional.
.83 .97 .93

χ² ( df=9) 31.19 347.17 43.36
NFI .95 .95 .96
CFI .96 .95 .97
GFI .92 .94 .92
SRMR .043 .005 .038
CR .91 .98 .93
AVE .64 .87 .70
Note:
Factor loadings are standardized factor loadings. All factor loadings are significant at p < .001. The chi-
square statistics are significant at p < .001.
Abbreviations: NFI = Normed Fit Index. CFI = Comparative Fit Index. GFI = Goodness-of-Fit Index.
SRMR = Standardized Root Mean Square Residual. CR = Construct Reliability. AVE = Average Variance
Extracted (Fornell and Larcker 1981).

8.3. Testing Convergent and Discriminant Validity
Measurement results of the PBIS and related constructs. A list of well-established
scales was adopted to measure the market maven, opinion leader, and brand attachment
constructs (see Appendix B). Specifically, the extent to which one is a market maven was
measured using a 3-item short version of the 6-item market maven scale by Feick and Price
(1987) as previous studies adopted
6
. Opinion leadership was measured using the 6-item opinion

6
Previous studies found that the 3-item short version market maven scale yielded better reliability than did the 6-
item scale. In Study 2, we used the 6-item market maven scale and our data confirmed that the 3-item scale had
better reliability than the 6-item scale. Therefore, results of the 3-item scale were reported in this research.
POWERFUL BRAND INFLUENTIALS 85

leadership scale by Flynn et al. (1996). Brand attachment was measured using the 4-item
measure developed by Park et al. (2009). These adopted measures yielded good internal
consistency: The Cronbach’s alpha for market maven is .88, opinion leadership is .86, and brand
attachment is .89. A single-item was constructed for the brand ambassador construct that
measures consumers’ interests in becoming brand ambassadors. A description of the brand
ambassador phenomenon was constructed based on references from the popular press (see
Appendix B). Respondents were asked to rate the extent to which they would be interested in
becoming a brand ambassador for the target brand (1 = not at all to 9 = completely).
Convergent and discriminant validity (P1 – P2). Two sets of propositions pertain to
the convergent and discriminant validity of the powerful brand influential construct and the PBIS.
We proposed that the powerful brand influential construct is significantly related to but distinctly
different from the construct of market maven (P1a-P2a), opinion leader (P1b-P2b), brand
attachment (P1c-P2c), and brand ambassador (P1d-P2d). We simultaneously test these two sets
of propositions. We first conducted an EFA on all measurement items of all constructs using
oblique rotation. Table 8-2 reports the EFA results. The EFA yielded a four-factor solution
(eigenvalues > 1) and explained a total variance of 68.54%. One factor included all the PBIS
items plus the single-item measure of brand ambassador. Though the brand ambassador measure
loaded on the PBIS factor, the factor loading of brand ambassador item is the lowest. A single
measure might lack of the power to constitute a unique factor by its own. Additionally, since the
item measures consumers’ tendency to become brand ambassadors, this result could simply
indicate that the powerful brand influential have higher tendency to become brand ambassadors
than do market mavens, opinion leaders, and strongly brand-attached consumers. The other three
factors represented opinion leader, brand attachment, and market maven, respectively.
POWERFUL BRAND INFLUENTIALS 86

TABLE 8-2. Exploratory Factor Analysis Results (Oblique Rotation) with the PBIS and
Measures of Market Maven, Opinion Leader, Brand Attachment, and Brand Ambassador

Item Number
Factor Loadings
PBIS and Brand
Ambassador
Opinion Leader
(OL)
Brand
Attachment
(BA)
Market Maven
(MM)
PBIS 4. .88 .29 .68 .17
PBIS 3. .87 .28 .47 .13
PBIS 6. .85 .28 .54 .35
PBIS 2. .84 .46 .49 .22
PBIS 5. .80 .11 .64 .23
PBIS 1 .70 .27 .40 .20
Brand Ambassador 1. .60 .22 .47 .28
OL 3. .22 .86 .05 .17
OL 6. .38 .85 .14 .12
OL 2. .25 .80 .11 .23
OL 5. .40 .80 .24 .21
OL 4. .24 .73 .23 .28
OL 1. .19 .60 .31 .08
BA 3. (PRO 1) .57 .24 .86 .19
BA 4. (PRO 2) .58 .24 .86 .17
BA 2. (BSC 2) .61 .14 .85 .26
BA 1. (BSC 1) .60 .14 .85 .19
MM 3. .21 .23 .16 .91
MM 1. .32 .24 .26 .90
MM 2. .20 .17 .19 .88
Eigenvalue 7.61 2.97 2.09 1.03
Percentage of Variance 38.05 14.87 10.46 5.17
Note:
Abbreviations: PBIS (the Powerful Brand Influential Scale). OL (Opinion Leader). BA (Brand
Attachment). PRO (Prominence of Thoughts and Feelings indicator of brand attachment). BSC (Brand-
Self Connection indicator of brand attachment). MM (Market Maven).

To establish initial evidence for the convergent validity of the PBIS, we examined the
correlations between the PBIS factor and the other construct factors extracted from the EFA. In
this step of analysis, the single measure of brand ambassador was analyzed as its own individual
factor (see Table 8-3). The correlations between the PBIS and each of the other factors were all
positive and significant, which supports our proposition (P1a-P1d) that the powerful brand
influential construct is related to the other four constructs. Additionally, the PBIS factor has the
POWERFUL BRAND INFLUENTIALS 87

highest correlation with brand attachment, followed by brand ambassador, and next with opinion
leadership. It was least correlated with market maven. Such pattern of the correlations is
consistent with our conceptualized overlapping pattern between the powerful brand influential
and other constructs. The correlations among opinion leader, market maven and brand
ambassador were all positive and significant, which suggests that all these consumer constructs
are interrelated to some extent. Moreover, brand ambassador factor is highly related to the PBIS
and brand attachment while moderately related to opinion leader and market maven. This result
shows that the powerful brand influential and strongly brand-attached consumers of a given
brand are more likely to become brand ambassadors of the brand than are opinion leaders and
market mavens.

TABLE 8-3. Correlations based on TABLE 8-2. Exploratory Factor Analysis Results

Factor
Interfactor Correlations (Pearson's r)
PBIS
Opinion
Leader
Brand
Attachment
Market
Maven
Brand
Ambassador
PBIS 1.00
Opinion Leader .32** 1.00
Brand Attachment .61** .18* 1.00
Market Maven .26** .22* .23* 1.00
Brand Ambassador .51** .22* .46** .25** 1.00
Note:
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.

The EFA results provided initial empirical evidence that the PBIS is related to, but
distinct from market maven, opinion leader, brand attachment and brand ambassador measures.
Next, a series of CFA models were constructed that assume different conceptual relationships
between the powerful brand influential and the related constructs. Model 1 represents our
proposed theoretical structure of how all constructs are related. It assumes that the PBIS, opinion
leader, market maven, brand attachment and brand ambassador are five distinct factors that are
POWERFUL BRAND INFLUENTIALS 88

correlated with each other. The hypothesized Model 1 corresponds similarly to the previous EFA
oblique factor solution. At the same time, Model 1 advances the EFA by representing the
theoretical conceptualization more precisely in two aspects. First, the concept of brand
attachment is structured in CFA as a second-order factor structure. Brand attachment is best
represented as a second-order factor structure indicated by both Brand-Self Connection (BSC)
and Prominence of Thoughts and Feelings (PRO) (Park et al. 2009). Such structure of brand
attachment cannot be modeled using the EFA. Second, brand ambassador is modeled as a unique
factor instead of loading together with the PBIS factor.
Table 8-4 reports the results of Model 1. All items loaded significantly on their predicted
factors (p < .001) (the factor loading of the single brand ambassador item was constrained to 1).
The construct reliability (CR) of all measures was well above the criterion of .70 and the average
variance (AVE) extracted for all factors was greater than the criterion of .50. Thus, the model has
a satisfactory fit with the data.

POWERFUL BRAND INFLUENTIALS 89

TABLE 8-4. Confirmatory Factor Analysis Results of Model 1 – Five Correlated Factors
on the PBIS and Measures of 4 Related Constructs

Factor Factor Indicators
Standardized Factor
Loadings
AVE CR
The Powerful Brand Influential
(PBIS)
PBIS 1. .81
.63 .91
PBIS 2. .65
PBIS 3. .82
PBIS 4. .89
PBIS 5. .78
PBIS 6. .80
Opinion Leader (OL)
OL 1. .52
.56 .88
OL 2. .72
OL 3. .77
OL 4. .68
OL 5. .81
OL 6. .86
Market Maven (MM)
MM 1. .86
.71 .88 MM 2. .77
MM 3. .89
Brand Attachment (BA)
BA 1. (BSC 1) .93
.84 .95
BA 2. (BSC 2) .95
BA 3. (PRO 1) .89
BA 4. (PRO 2) .90
Brand Ambassador Brand Ambassador 1. 1.00
Note:
Factor loadings are standardized factor loadings. All factor loadings are significant at p < .001.
Abbreviations: CR = Construct Reliability. AVE = Average Variance Extracted. PBIS (the Powerful Brand
Influential Scale). OL (Opinion Leader). BA (Brand Attachment). PRO (Prominence of Thoughts and
Feelings indicator of brand attachment). BSC (Brand-Self Connection indicator of brand attachment). MM
(Market Maven).

To test for convergent validity, interfactor Φ coefficients were acquired. Interfactor Φ
coefficients are the correlations between all exogenous variables in a structural equation model.
Table 8-5 reports all pair-wise correlation coefficients (Interfactor Φ coefficients) among the
PBIS, opinion leader, market maven, brand attachment and brand ambassador factors. As shown,
Model 1 CFA yielded significant positive correlations among the five factors. This result
supports our P1a-1d that the powerful brand influential is significantly related to other constructs.
Φ coefficients have also been used as criteria for assessing discriminant validity in the literature
POWERFUL BRAND INFLUENTIALS 90

(Anderson and Gerbing, 1988; Bagozzi, Yi, and Phillips 1991; Fornell and Larcker, 1981).
Although significant, none of the Φ coefficients between the set of latent variables is higher
than .90, a correlation cutoff value indicating that the latent factors might measure the same
construct (Bagozzi et al. 1991). Additionally, Fornell and Larcker (1981) suggest that evidence
for the discriminant validity between constructs is established when the squared Φ coefficients
between each construct and all the other constructs in the model are less than the average
variance extracted for each construct itself (AVE reported in the previous Table 8-4). All 10
pair-wise Φ coefficients squared in Table 8-5 are less than the AVE extracted for the
corresponding construct in Table 8-4. Overall, the Φ coefficient results supported that all
constructs in Model 1 are related but distinct constructs.

TABLE 8-5. Interfactor (Φ) Correlations of Model 1 – Correlations between 5 Factors
based on TABLE 8-4. Confirmatory Factor Analysis Results

Factor
Interfactor Correlations (Φ)
(t-Value; Std. Error)
The Powerful
Brand
Influential
Opinion
Leader
Market
Maven
Brand
Attachment
Brand
Ambassador
The Powerful
Brand Influential
―

Opinion Leader
.13**
(2.85; 0.05)
―
Market Maven
.20**
(2.71; 0.04)
.09*
(2.19; 0.04)
―
Brand Attachment
.53***
(5.51; 0.10)
.09*
(2.18; 0.04)
.19*
(2.25; 0.08)
―
Brand Ambassador
.44***
(4.87; 0.09)
.10*
(2.20; 0.05)
.22*
(2.54; 0.09)
.41***
(4.27; 0.10)
―
Note:
* Φ coefficients are significant at p < .05 for two-tailed tests. ** Φ coefficients are significant at p < .01
for two-tailed tests. *** Φ coefficients are significant at p < .001 for two-tailed tests.

POWERFUL BRAND INFLUENTIALS 91

To further establish discriminant validity of the PBIS, the hypothesized Model 1 was
compared to a set of alternative conceptual models, Model 2 – Model 6. Support for convergent
and discriminant validity would be achieved by showing that each and all alternative models are
inferior to our hypothesized Model 1. Table 8-6 summarizes the CFA results including the
goodness-of-fit statistics, specification of model comparisons, and chi-square difference test
results of all the model comparisons.

POWERFUL BRAND INFLUENTIALS 92

TABLE 8-6. Results for Comparing Models of Confirmatory Factor Analysis on the PBIS
and Measures of 4 Related Constructs

Model
χ²
(df)
NFI CFI SRMR
Model
Comparison
∆χ²
(∆df)
∆p
M 1: Five Correlated Factors
304.74
(160)
.94 .95 .048 NA NA NA
M 2: Five Uncorrelated Factors
456.46
(170)
.86 .91 .26 M2-M1
151.7
(10)
.001
M 3: Four Correlated Factors –
PBIS & OL belong to the same
factor
331.61
(161)
.89 .93 .30 M3-M1
26.87
(1)
.001
M 4: Four Correlated Factors –
PBIS & MM belong to the same
factor
310.70
(161)
.89 .94 .25 M4-M1
18.55
(1)
.001
M 5: Four Correlated Factors –
PBIS & BA belong to the same
factor
310.70
(161)
.90 .94 .15 M5-M1
5.96
(1)
.025
M 6: Four Correlated Factors –
PBIS & Ambassador belong to the
same factor
310.84
(161)
.90 .94 .16 M6-M1
6.1
(1)
.025
Note:
The chi-square statistics are significant at p < .001.
Abbreviations: NFI = Normed Fit Index. CFI = Comparative Fit Index. SRMR = Standardized Root Mean
Square Residual. ∆χ² (∆df) = Chi-Square Difference Tests. ∆p = Significance of Chi-Square Difference
Tests. Brand Advocates (PBIS). Market Mavens (MM). Opinion Leaders (OL). Brand Attachment (BA).
Model Specification:
M 1: Five Correlated Factors: The same model as Table 8-5, assumes that PBIS, OL, MM, BA,
and Brand Ambassador are 5 factors, allowing factors to correlate freely.
M 2: Five Uncorrelated Factors: The same 5 factors as M 1 except that correlation between 5
factors is set to zero.
M 3: Four Correlated Factors – PBIS & OL belong to the same factor: assumes that PBIS and
OL are loaded on 1 factor while MM, BA, and Brand Ambassador are separate factors, allowing factors to
correlate freely.
M 4: Four Correlated Factors – PBIS & MM belong to the same factor: assumes that PBIS
and MM are loaded on 1 factor while OL, BA, and Brand Ambassador are separate factors, allowing
factors to for correlate freely.
M 5: Four Correlated Factors – PBIS & BA belong to the same factor: assumes that PBIS and
BA are loaded on 1 factor while OL, MM, and Brand Ambassador are separate factors, allowing factors to
for correlate freely.
M 6: Four Correlated Factors – PBIS & Brand Ambassador belong to the same factor:
assumes that PBIS and Brand Ambassador are loaded on 1 factor while OL, MM, and Brand Attachment
are separate factors, allowing factors to for correlate freely.

POWERFUL BRAND INFLUENTIALS 93

Specifically, Model 2 assumes a unique factor for each construct as with Model 1 except
that all factors are forced to be uncorrelated. This model deviates the most from the hypothesized
Model 1 wherein all five constructs are conceptualized as fully unrelated. We expect that Model
2 represents the most implausible conceptual structure and worst model-fitting scenario. Models
3 – Model 6 take a step-wise approach to assess the discriminant validity of the PBIS from a
specified individual construct. Each of these models has the same structure as Model 1 with only
one difference in model specification – the PBIS factor is modeled as representing the same
construct as one specified construct (the Φ coefficient between the PBIS construct and a
specified construct was fixed to unity, forced as perfectly correlated)
7
. For example, In Model 3,
the PBIS and opinion leader constructs are modeled as perfectly correlated with each other,
while the market maven, brand attachment, and brand ambassador scales are modeled as three
separate constructs that are allowed to correlate freely with all other constructs including the
PBIS and opinion leader. Hence, Model 3 assumes that the powerful brand influential construct
is not differentiated from opinion leadership but is differentiated from the market maven, brand
attachment and brand ambassador constructs.
The goodness-of-fit statistics and chi-square difference tests provide converging evidence
that the proposed Model 1 is the best fitting model. First, Model 1 has a good overall model fit:
NFI and CFI both exceed the .90 cutoff values (Bentler 1992) and the SRMR is less than the .05
recommendation (Byrne 1998). None of the alternative models meets the cutoff criteria for all fit
statistics. Second, compared to alternative models, Model 1 has the lowest χ² statistic and SRMR
as well as the highest NFI, and CFI among all models. Third, the chi-square difference tests

7
The objective of the model comparison tests here was to assess the discriminant validity of the PBIS from all the
other related consumer group scales, not to test the discriminant validity among the other related consumer group
scales. Therefore, only pairwise comparisons between the PBIS and each related consumer group scale were
modeled. No pairwise comparison was conducted to discriminate related consumer group scales from each other.
POWERFUL BRAND INFLUENTIALS 94

showed that the chi-square of Model 1 is significantly better [∆χ² ( ∆df) is at p < .05 or less] than
each and all alternative models. As expected Model 2 was the worst fitting model. Following
Model 1, the next best fitting model is Model 5, which supports our conceptualization that the
powerful brand influential construct is most related to brand attachment (hence when
constraining the PBIS and brand attachment into one construct, the decrease in model fit is the
least among all alternative models). Also as proposed, the brand ambassador and the powerful
brand influential constructs are closely related. This is shown by the comparison between Model
6 and Model 1. The Opinion Leader and market maven constructs are relatively more
differentiated from the powerful brand influential construct (hence when constraining the PBIS
and opinion leader, market maven into one construct, the decrease in model fit is the highest
among all). This is shown by the comparison between Model 2 and Model 1 as well as Model 3
to Model 1.
In sum, both EFA and CFA results support the two sets of our propositions
simultaneously – the powerful brand influential construct is significantly related to but distinctly
different from the construct of market maven (P1a-P2a), opinion leader (P1b-P2b), brand
attachment (P1c-P2c), and brand ambassador (P1d-P2d). Moreover, the interrelatedness of the
PBIS and other constructs is in line with our previous conceptual discussions. As the results,
these results simultaneously support convergent and discriminant validity of the PBIS.
Indictor variable results of the PBIS and related constructs. The next step to test the
validity of the PBIS is to show that the PBIS is significantly related to a set of indictors of the
powerful brand influential construct (convergent validity) and that the PBIS is more strongly
related to these indicators than are the market maven, opinion leadership, brand attachment, and
brand ambassador measures (discriminant validity). We compared the correlations between the
POWERFUL BRAND INFLUENTIALS 95

PBIS and an indicator to the correlation between each of the other four constructs and the same
indicator. Hotelling's t-test was conducted to examine the difference in two dependent
correlations (i.e., correlations drawn from the same sample).
The first indictor variable is the self-rating measure of the powerful brand influential
employed in Study 1. The self-rating measure is a recommended indicator by Bagozzi (1993) for
validity test of a construct. The self-rating measure (α = .93) and the PBIS were significantly
correlated (r = 0.78, p < .001) supporting convergent validity. Discriminant validity was tested
by comparing the correlation between the PBIS and the self-rating scale to the correlations
between other constructs and the self-rating scale. Hotelling's t-test results support discriminant
validity: the PBIS correlates more strongly with the self-rating scale than does the market maven
scale (r = 0.15, p > .10; t (118) = 9.17, p < .001), the opinion leadership scale (r = 0.39, p < .001;
t (118) = 6.11, p < .001), the brand attachment scale (r = 0.56, p < .001; t (118) = 5.24, p < .001),
and the brand ambassador measure (r = 0.49, p < .001; t (118) = 5.40, p < .001). Furthermore, the
pattern of differences in these correlations was as expected, showing that the powerful brand
influential construct is more strongly related to brand attachment and the brand ambassador
construct and is relatively less related to the opinion leadership and market maven constructs.
Another indictor variable concerns consumers’ ability to identify others as the powerful
brand influential. This variable was adapted from Feick and Price (1987). Respondents read the
same description of the powerful brand influential used in the self-rating measure. It was
highlighted that a powerful brand influential consumer could be an influential of the given brand
(Apple iPod) as well as any other brands. They then answered an item measuring their ability to
identify others as the powerful brand influential: “How many people do you know, other than
POWERFUL BRAND INFLUENTIALS 96

yourself, who fits this description?” The PBIS is expected to correlate with identification of like-
minded others in the marketplace more so than other four constructs.
Convergent validity was supported by a significant correlation between the PBIS and
identification of other powerful brand influential consumers (r = 0.51, p < .001). At the same
time, the PBIS is significantly more related to this indictor than are the market maven (r = 0.21,
p < .05; t (118) = 3.15, p < .01), opinion leadership (r = 0.24, p < .01; t (118) = 3.07, p < .01),
brand attachment (r = 0.26, p < .01; t (118) = 4.28, p < .001), and brand ambassador constructs (r
= 0.21, p < .05; t (118) = 3.93, p < .001). These results support discriminate validity of the PBIS.
The last indicator is brand expertise. As previously mentioned, the powerful brand
influential should possess great brand expertise. Opinion leaders as the influential consumers of
the product class and brand ambassadors as those who intend to influence others at the brand-
level might also have great brand expertise. Additionally, strongly brand-attached consumers
should have high brand expertise in that they possess rich experiential knowledge. In comparison,
market mavens might have less expertise about a given brand since their knowledge primarily
involves facets about the general marketplace. Therefore, the convergent validity would be
supported by showing that the PBIS is strongly related to the brand expertise measure and so are
opinion leadership, brand ambassador and brand attachment. Support for discriminant validity
would be shown by a stronger relationship between the PBIS and brand expertise than the
relationship between the market maven measure and brand expertise.
The subjective expertise scale by Mitchell and Dacin (1996) was adapted to a brand-
specific measure (α = .88). The convergent validity was confirmed by a significant correlation
between the PBIS and subjective brand expertise (r = 0.52, p < .001). As predicted, the PBIS
correlated more strongly with brand expertise than did the market maven scale (r = 0.24, p < .01;
POWERFUL BRAND INFLUENTIALS 97

t (118) = 3.0, p < .01). Also as expected, the PBIS correlated with brand expertise as strongly as
the opinion leadership scale (r = 0.44, p < .001; t (118) = .96, p > .10), the brand attachment
scale (r = 0.48, p < .001; t (118) = .64, p > .10), and the brand ambassador measure (r = 0.42, p <
.001; t (118) = 1.42, p > .10).
8.4. Testing Predictive Validity and Propositions
Three propositions (P3 – P6) were construct to test predictive validity of the PBIS by
showing that the PBIS predicts core conceptual outcomes better than does all other four
constructs. Two propositions (P7 – P8) compare predictive validity of the powerful brand
influential and that of brand attachment on psychological and behavioral outcomes. All
propositions were tested using hierarchical regression analyses on the list of dependent variables.
The dependent variable list includes the engagement level of brand word-of-mouth, brand
advocacy behavior, the tendency to influence others’ brand decisions, the level of intrinsic
pleasure derived from brand word-of-mouth, separation distress, brand commitment, and
intentions to conduct brand loyal behaviors.
Using hierarchical analysis, we constructed predictive models in a sequential manner: we
input the other constructs as antecedents of each outcome variable before entering the PBIS as an
additional predictor to those in the previous step. Since the PBIS is entered as a last-step
predictor, this analysis approach allows for previously entered predictors to take precedence in
contributing to the outcome before the PBIS. It is a rigorous way to test the independent
predictive power of the PBIS relative to that of the other constructs including market maven,
opinion leadership, brand attachment and brand ambassador.

POWERFUL BRAND INFLUENTIALS 98

We used mean scores for all the predictor constructs. The concept of brand attachment is
represented by a composite score of all items including both BSC and PRO items. Though the
second-order factor structure of brand attachment is not an option in regression analysis, the
composite score takes into account the contribution of both indicators of brand attachment
8
.
Dependent measures for assessing predictive validity. The first dependent variable is
the engagement level of brand word-of-mouth. We adopted two indicators to assess the degree of
which a consumer is engaged in brand word-of-mouth. The first indicator is the number of
people that respondents would tell about the Apple iPod brand. It measures the audience size of
their word-of-mouth about the brand. The more people they would tell about the brand, the
higher their level of engagement in brand word-of-mouth. The distribution of the measure
showed that the average number of people that the respondents would tell about the brand is 10.
The range and variation on this measure is relatively large (range = 0 - 100, standard deviation =
19.35).
The second indicator of brand word-of-mouth engagement level is the selectiveness of
brand word-of-mouth receivers and to whom consumers would like to talk about the brand. As
previously discussed, the higher the level of brand word-of-mouth engagement, the more a
consumer seeks an audience and finds occasions to talk about the brand, and the less selective
they are in who they would talk to. Correspondingly, the selectiveness of brand word-of-mouth is
a negative indicator of the level of engagement. Respondents rated the item, “I tend to talk about
this brand to like-minded others as opposed to just anybody” on a 9-point scale that ranged from
1 = strongly disagree to 9 = strongly agree. The higher the score, the more selective they are in

8
In order to fully represent the second-order factor structure of brand attachment, structural models are needed.
Since each predictive validity test involves many variables, the data requirements for structural models cannot be
satisfied by the current study.
POWERFUL BRAND INFLUENTIALS 99

choosing their audience and the lower their engagement level in brand word-of-mouth. The
scaled yielded a mean of 5.74 (range = 1 - 9, standard deviation = 2.31).
The second dependent variable is brand advocacy behavior. We assessed the degree of
which a consumer intends to perform various brand advocacy behaviors. A list of brand
advocacy behaviors was adopted from previous studies (Park et al. 2006; Park et al. 2010). The
list includes 1) recommending the brand to others, 2) defending the brand when others speak
poorly about it, 3) spending money, time, and energy to participate in activities to promote the
brand, 4) actively resisting negative information about the brand, and 5) frequently using
products with the brand logo. Each item is measured on a 9-point scale ranging from 1= not at all
to 9 = extremely. These brand advocacy items loaded on a single factor with explained variance
of 65.51% (factor loadings ranging from .74 to .90) and good internal consistency (α = .87).
Hence, we averaged all five items and from a composite score as the brand advocacy behavior
measure.
The third dependent variable is the tendency to influence others’ brand decision. We used
respondents’ self-perceived influence on others’ brand decisions to measure their tendency to
influence others’ brand decisions. Respondents rated two items including the perceived
importance of influencing other consumers’ brand decisions (1 = not important at all to 7 =
extremely important) and perceived persuasiveness at convincing other consumers’ beliefs and
decisions about the brand (1 = not persuasive at all to 7 = extremely persuasive). The two items
yielded good internal consistency (α = .93) and are averaged as the influence tendency measure.

POWERFUL BRAND INFLUENTIALS 100

The fourth dependent variable is the level of intrinsic pleasure derived from brand word-
of-mouth. We measured the level at which a consumer would derive sheer pleasure and intrinsic
satisfaction from the action of brand word-of-mouth. A 3-item scale was adapted from Walsh et
al.’s (2004) scale of intrinsic pleasure in the general act of information sharing to brand-level
information sharing: “I get pleasure out of telling others about the brand”, “I can’t say that I
particularly enjoy telling others about the brand” (a reverse- coded item), and “Telling others
about the brand gives me a good feeling even if no one paid me for doing it” on a 9-point scale
anchored by 1 = strongly disagree and 9 = strongly agree (α = .77). We acquired the average
score for this indicator.
Three additional dependent variables pertain only to the comparison between the PBIS
and brand attachment, which includes separation distress, brand commitment, and brand loyal
behaviors. First, we measured separation distress using a 2-item scale used by Park et al. (2010).
These items are “To what extent would you be distressed if the brand were discontinued?” and
“To what extent is it difficult for you to imagine life without the brand?” (α = .89) on a 9-point
scale ranging from 0 = “not at all” to 9 = “completely”. Second, we measured brand commitment
using the 3-item scales of Park et al. (2009) including “To what extent have you made a pledge
to yourself to use the brand in the future?”, “To what extent are you committed to the brand” and
“To what extent are you loyal to the brand (i.e., not buying competitors’ brands)?” All items
used 9-point scales anchored by 0 = “not at all” to 9= “completely”. They all loaded on a single
factor that explained 83.57% of variance (factor loadings ranged from .90 to .94) with high
internal consistency (α = .90). Hence, we averaged the items.

POWERFUL BRAND INFLUENTIALS 101

Last, we included measures for intention to conduct brand loyal behaviors, the behavioral
manifestations of brand commitment (as a pre-cursor). We incorporated a comprehensive list of
brand-commitment behaviors (Park et al. 2010; Thomson et al. 2005) with the exclusion of brand
word-of-mouth and advocacy type of brand-commitment behaviors. The brand loyal behavior
scale measure behavioral intentions to 1) buy the brand for oneself, 2) buy the brand for others, 3)
always buy the new model of the brand, 4) pay more for the brand than for a comparable brand
made by a competitor, 5) switch from the brand to brands of competitors, and 6) wait for several
months to buy the brand instead of buying an alternative brand more immediately. All items used
9-point scales ranging from 0 = “not at all” to 9 = “extremely”. These brand loyal behaviors (α
= .85) loaded on a single factor with explained variance of 60.10% (factor loadings ranging
from .73 to .85). The six items were averaged to form the brand loyal behavior measure.
Testing predictive validity for engagement in brand word-of-mouth (P3a – 3d). We
proposed that the powerful brand influential construct predicts the engagement level of brand
word-of-mouth better than does the market maven (P3a), opinion leadership (P3b), brand
attachment (P3c), and brand ambassador (P3d) constructs. Two hierarchical regressions were
conducted on the two indicators of brand word-of-mouth engagement separately. For both
regressions, the first set of predictor variables entered into the model (Model 1) consists of only
non-brand level constructs (i.e., influential consumers: market maven and opinion leadership).
The second entry into the model (Model 2) added in extant brand-level predictors of interests:
brand attachment and brand ambassador. The last entry into the model (Model 3) added the PBIS.
Table 8-7a and Table 8-7b reported the regression of each indicator of the engagement level of
brand word-of-mouth, respectively.

POWERFUL BRAND INFLUENTIALS 102

TABLE 8-7a. Hierarchical Regression Results Predicting the Engagement Level of Brand
Word-of-Mouth (indicator 1 – the Number of Receivers): Comparing Predictive Ability of
the PBIS, Market Maven, Opinion Leader, Brand Attachment, and Brand Ambassador

Model
β
(Standardized
Coefficients)
R R
2
Adjusted
R
2
R
2
Change

F Change df
Model 1 .24 .06 .04 .06 3.42* 2, 116
Predictors
Market Maven .10
Opinion Leader .19*
Model 2 .39 .15 .12 .10 6.49** 2, 114
Predictors
Market Maven .03
Opinion Leader .11
Brand Attachment .22*
Brand Ambassador .17
Model 3 .48 .23 .20 .08 11.30*** 1, 113
Predictors
Market Maven -.01
Opinion Leader .05
Brand Attachment -.03
Brand Ambassador .07
PBIS .44***
Note:
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.

As shown in Table 8-7a, between market maven and opinion leadership, the former is not
a significant predictor of the number of people to talk to while the latter is (see Model 1; (β = .19,
p < .05). This result is consistent with our discussion that opinion leaders are more likely to
develop attachment to a given brand within the product category about which they have an
opinion. Hence, they are more likely to engage in brand word-of-mouth than are market mavens.
Model 2 shows that the addition of brand attachment and brand ambassador to the model
explains a significant increment of variance in the dependent variable (R
2
change is significant as
assessed by F change = 6.49, p < .01 from Model 1 to Model 2). Moreover, only brand
attachment significantly predicts the number of people to who consumers talk (β = .22, p < .05),
POWERFUL BRAND INFLUENTIALS 103

whereas market maven, opinion leadership, and brand ambassador are insignificant predictors.
The previously found predictive power of opinion leadership yielded to brand attachment and
became insignificant when brand attachment was included into the equation. This result again
indicates that brand attachment largely explains predictive power of opinion leadership found in
Model 1. Moreover, when simultaneously estimated, brand ambassador is not a significant
predictor while brand attachment is. This result supports our prediction that brand ambassadors
who do not have strong attachment to the brand are less engaged in brand word-of-mouth than
are strongly brand-attached consumers.
The last model (Model 3) directly supports our propositions that the PBIS is a better
predictor of engagement in brand word-of-mouth than the market maven (P3a), opinion
leadership (P3b), brand attachment (P3c), and brand ambassador constructs (P3d). The results
showed that the PBIS remained a highly significant predictor (β = .44, p < .001), whereas all the
other predictor variables became insignificant. Furthermore, the PBIS explained a considerable
amount of additional variance (a significant increment of R
2
change assessed by F change =
11.30, p < .001 from Model 2 to Model 3). This result confirmed that the powerful brand
influential construct predicts engagement in brand word-of-mouth to a substantial extent and to
an extent that exceeds predictive power of other existing constructs. Last, the fact that brand
attachment becomes insignificant when simultaneously estimated with the PBIS shows that the
powerful brand influential construct largely accounts for predictive ability of brand attachment
on brand word-of-mouth that has been previously found in the literature. More importantly, the
powerful brand influential construct exceeds brand attachment in predicting brand word-of-
mouth engagement.

POWERFUL BRAND INFLUENTIALS 104

TABLE 8-7b. Hierarchical Regression Results Predicting the Engagement Level of Brand
Word-of-Mouth (indicator 2 – the Selectiveness of Brand Word-of-Mouth Receivers):
Comparing Predictive Ability of the PBIS, Market Maven, Opinion Leader, Brand
Attachment, and Brand Ambassador

Model
β
(Standardized
Coefficients)
R R
2
Adjusted
R
2
R
2
Change

F Change df
Model 1 .30 .09 .07 .09 5.63** 2, 118
Predictors
Market Maven -.24**
Opinion Leader -.12
Model 2 .57 .32 .30 .24 20.65*** 2, 116
Predictors
Market Maven -.13
Opinion Leader .01
Brand Attachment -.34***
Brand Ambassador -.27**
Model 3 .64 .41 .38 .08 15.89*** 1, 115
Predictors
Market Maven -.10
Opinion Leader .07
Brand Attachment -.09
Brand Ambassador -.18*
PBIS -.45***
Note:
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.

Table 8-7b reports the hierarchical regression results for the selectiveness of brand word-
of-mouth receivers, a negative indicator of the engagement level of brand word-of-mouth. Model
1 revealed that market maven is a significant negative predictor (β = -.24, p < .01) while opinion
leadership has no significant effect. This pattern is consistent with previous literature suggesting
that market mavens are sought by a broader audience (Feick and Price 1987) and hence less
selective in word-of-mouth receivers than are opinion leaders. Model 2 shows that the brand
attachment and brand ambassador measures contribute to additional variance explained in the
dependent variable (R
2
change is significant as assessed by the F change = 20.65, p < .001 from
Model 1 to Model 2). Both constructs negatively influence the selectiveness of the brand word-
POWERFUL BRAND INFLUENTIALS 105

of-mouth audience. Comparatively, brand attachment has more predictive power (β = -.34, p
< .001) than does brand ambassadorship (β = -.27, p < .01). This result provides more evidence
that brand ambassadors who do not always have strong attachment to a brand might have lower
brand word-of-mouth engagement than strongly brand-attached consumers. Additionally, when
these brand-level predictors were included, market maven and opinion leader became
insignificant.
Model 3 fully supports our propositions that the PBIS is a better predictor of the
engagement level of brand word-of-mouth than market maven (P3a), opinion leadership (P3b),
brand attachment (P3c), and brand ambassadorship (P3d). After accounting for the contribution
of all other predictors, the PBIS remains significant (β = -.45, p < .001). Moreover, the PBIS
adds in a high amount of explained variance in the dependent variable (a significant increment of
R
2
change assessed by F change = 15.89, p < .001 from Model 2 to Model 3). Consistently with
the results for the first indicator of brand word-of-mouth engagement, when simultaneously
estimated, brand attachment yields to the PBIS in predicting this outcome variable and becomes
insignificant. Interestingly, though to a lower degree, brand ambassador remains significant (β =
-.18, p < .05) in the entry together with the PBIS. This result suggests that the powerful brand
influential construct does not fully account for the ability of the brand ambassadorship construct
to predict the tendency to seek for broad audience. Brand ambassadors and the powerful brand
influential might both be highly engaged in seeking a broad audience, even if they do so for
different reasons. The former is driven by extrinsic incentives while the latter is driven by brand
attachment. Since different motivational antecedents drive them, brand ambassadorship and the
PBIS both independently drive this outcome. In sum, regression results on both indicators of the
brand word-of-mouth engagement level consistently supported P3a-3d.
POWERFUL BRAND INFLUENTIALS 106

Testing predictive validity for brand advocacy behavior (P4a – 4d). We proposed that
the powerful brand influential construct better predicts brand advocacy behavior than does the
market maven (P4a), opinion leader (P4b), brand attachment (P4c), and brand ambassador (P4d)
construct. A hierarchical regression analysis was conducted on the composite score measuring
consumers’ intentions to perform various brand advocacy behaviors. The independent variables
included the same set of predictors and were entered into the model in the same sequential
manner as the previous hierarchical regressions. The results are shown in Table 8-8. In Model 1,
opinion leadership significantly predicts brand advocacy behavior (β = .37, p < .001) while
market maven does not. This pattern again confirmed that opinion leaders are more likely to
develop brand attachment and hence more likely than does market mavens to engage in brand
advocacy behaviors.

POWERFUL BRAND INFLUENTIALS 107

TABLE 8-8. Hierarchical Regression Results Predicting Brand Advocacy Behaviors:
Comparing Predictive Ability of the PBIS, Market Maven, Opinion Leader, Brand
Attachment, and Brand Ambassador

Model
β
(Standardized
Coefficients)
R R
2
Adjusted
R
2
R
2
Change

F Change df
Model 1 .37 .14 .12 .14 9.36*** 2, 118
Predictors
Market Maven .02
Opinion Leader .37***
Model 2 .69 .47 .45 .34 36.76*** 2, 116
Predictors
Market Maven -.12
Opinion Leader .22**
Brand Attachment .33***
Brand Ambassador .39***
Model 3 .77 .60 .58 .13 35.78*** 1, 115
Predictors
Market Maven -.15
Opinion Leader .14*
Brand Attachment .02
Brand Ambassador .27***
PBIS .55***
Note:
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.

In Model 2, brand attachment and brand ambassadorship were added to the model and
considerably increased the explained variance in the dependent variable (R
2
change is significant
as assessed by F change = 36.76, p < .001 from Model 1 to Model 2). Both brand attachment (β
= .33, p < .001) and brand ambassadorship are significant predictors (β = .39, p < .001) of brand
advocacy behavior and both have similarly high predictive power. The fact that both constructs
are significant supports our previous discussion that predictive power of brand ambassadorship is
driven by extrinsic motivation (external incentives and rewards) while that of brand attachment
is driven by intrinsic motivation (the need to expand brand-self relationship). By similar logic,
opinion leadership also remained as a significant predictor of brand advocacy behavior in Model
POWERFUL BRAND INFLUENTIALS 108

2 (β = .22, p < .01) though to a less degree of significance. As previously discussed, we expect
that opinion leaders could engage in brand advocacy behaviors to a certain degree for different
motivations such as asserting superiority in front of others and securing their status as influential
consumers of the brand’s category.
In Model 3, the PBIS is highly significant (β = .44, p < .001) even after accounting for
the variance explained by all other predictor variables. The PBIS explained a substantial amount
of additional variance (a significant increment of R
2
change assessed by F change = 35.78, p
< .001 from Model 2 to Model 3). As proposed, predictive power of the PBIS exceeds that of the
opinion leadership (β = .14, p < .05), brand ambassadorship (β = .27, p < .001), and market
maven (insignificant) scales. From the motivational perspective, the powerful brand influential’s
engagement in brand advocacy behaviors is driven by strong brand attachment. Also as we
anticipated, brand attachment yielded to the PBIS as a key driver of brand advocacy behavior
and became insignificant. This result again confirmed that the powerful brand influential
construct largely accounts for predictive ability of brand attachment on brand advocacy behavior,
which has been suggested by previous studies. Moreover, the powerful brand influential
construct also exceeds brand attachment in predicting brand advocacy behavior (in addition to
the engagement level of brand word-of-mouth).
Overall, the hierarchical regression results fully support our propositions that the
powerful brand influential construct predicts brand advocacy behavior better than does the
construct of market maven (P4a), opinion leader (P4b), brand attachment (P4c), and brand
ambassador (P4d).

POWERFUL BRAND INFLUENTIALS 109

Testing predictive validity for the tendency to influence others’ brand decision (P5a
– 5d). It is proposed that the powerful brand influential construct predicts the tendency to
influence others’ brand decision better than does the market maven (P5a), opinion leadership
(P5b), brand attachment (P5c), and brand ambassadorship (P5d) constructs. Table 8-9 reports the
hierarchical regression results that included all other predictive variables before allowing the
PBIS to predict the outcome. The results provide clear-cut support that only the PBIS
significantly captures a consumer’s tendency to influence others’ brand decisions (β = .39, p
< .01). As shown in Model 1 and 2, none of the other previous constructs has a significant effect
on this outcome. This result highlights that extant constructs the ability to designate a brand’s
influential consumers, those who are extraordinarily influential on others’ decisions about a
given brand. The PBIS contributes to the literature by filling this gap.

POWERFUL BRAND INFLUENTIALS 110

TABLE 8-9. Hierarchical Regression Results Predicting the Tendency to Influence Others’
Brand Decisions: Comparing Predictive Ability of the PBIS, Market Maven, Opinion
Leader, Brand Attachment, and Brand Ambassador

Model
β
(Standardized
Coefficients)
R R
2
Adjusted
R
2
R
2
Change

F
Change
df
Model 1 .16 .03 .01 .03 1.61 2, 118
Predictors
Market Maven .11
Opinion Leader .10
Model 2 .31 .09 .06 .07 4.26* 2, 116
Predictors
Market Maven .05
Opinion Leader .03
Brand Attachment .19
Brand Ambassador .13
Model 3 .40 .16 .12 .06 8.76** 1, 115
Predictors
Market Maven .03
Opinion Leader -.03
Brand Attachment -.03
Brand Ambassador .05
PBIS .39**
Note:
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.

Testing predictive validity for the level of intrinsic pleasure derived from brand
word-of-mouth (P6a – P6d). We proposed that the powerful brand influential construct better
predicts the level of intrinsic pleasure derived from brand word-of-mouth than does the market
maven (P6a), opinion leadership (P6b), brand attachment (P6c), and brand ambassadorship (P6d)
constructs. Table 8-10 reports the regression results on the level of intrinsic pleasures derived
from brand word-of-mouth as the dependent variable. Model 1 shows that opinion leadership
significantly predicts the level of intrinsic pleasure derived from brand word-of-mouth (β = .24,
p < .05) while market maven does not. As influential consumers, market mavens and opinion
leaders tend to find word-of-mouth activities intrinsically pleasant in general (Feick and Price
POWERFUL BRAND INFLUENTIALS 111

1987; Price et al. 1987; Walsh et al. 2004). Our results suggest that opinion leaders derive certain
level of pleasure from talking about a brand in the category they are highly involved in while
market mavens do not find talking about a give brand particularly enjoyable.

TABLE 8-10. Hierarchical Regression Results Predicting the Level of Intrinsic Pleasure
Derived from Brand Word-of-Mouth: Comparing Predictive Ability of the PBIS, Market
Maven, Opinion Leader, Brand Attachment, and Brand Ambassador

Model
β
(Standardized
Coefficients)
R R
2
Adjusted
R
2
R
2
Change

F Change df
Model 1 .26 .07 .05 .07 4.24* 2, 118
Predictors
Market Maven .07
Opinion Leader .24*
Model 2 .29 .08 .05 .02 1.02* 2, 116
Predictors
Market Maven .06
Opinion Leader .21*
Brand Attachment -.06
Brand Ambassador .15
Model 3 .51 .26 .23 .18 28.15*** 1, 115
Predictors
Market Maven .02
Opinion Leader .12
Brand Attachment -.43***
Brand Ambassador .01
PBIS .66***
Note:
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.

In Model 2, it is interesting to see that brand attachment and brand ambassadorship do not
significantly predict intrinsic pleasure derived from brand word-of-mouth. This result is line with
our discussions that there is a considerable division among strongly brand-attached consumers:
some do not find word-of-mouth activities intrinsically pleasant while others (i.e., the powerful
brand influential) do. Also as expected, brand ambassadors tend to conduct brand word-of-mouth
for extrinsic reasons rather than for intrinsic pleasure. Opinion leadership remained a significant
POWERFUL BRAND INFLUENTIALS 112

predictor (β = .21, p < .05), suggesting that opinion leaders derive a generally higher level of
pleasure from the act of word-of-mouth than do strongly brand- attached consumers and those
who act as brand ambassadors.
As shown in Model 3, the PBIS is a highly significant predictor of intrinsic pleasure
derived from brand word-of-mouth (β = .66, p < .001). The inclusion of the PBIS into the model
produced a substantial variance (a significant increment of R
2
change assessed by F change =
28.15, p < .001 from Model 2 to Model 3). We proposed that strongly brand-attached consumers
who are not the powerful brand influential are those who derive low intrinsic pleasure from the
act of brand word-of-mouth. This proposition is directly supported by the fact that brand
attachment becomes a significant negative predictor (β = -.43, p < .001) while the PBIS remains
a significant positive predictor in Model 3. Moreover, opinion leadership is no long significant
when the PBIS is entered into the equation, suggesting that predictive power of opinion
leadership for intrinsic pleasure was largely driven by powerful brand influentials who overlap as
opinion leaders in the brand category. Overall, these results support our propositions that the
powerful brand influential construct predicts the level of intrinsic pleasure derived from brand
word-of-mouth better than do the market maven (P6a), opinion leadership (P6b), brand
attachment (P6c), and brand ambassadorship (P6d) constructs.
Comparing predictive validity of the PBIS and brand attachment for separation
distress and brand commitment (P7a – 7b). Brand attachment should better predict separation
distress (P7a) and brand commitment (P7b) than should the powerful brand influential construct.
To test separation distress, brand attachment was first regressed on separation distress by itself
and yielded a significant effect (β = .73, p < .001). As the second step, brand attachment and the
PBIS were simultaneously regressed on separation distress. The result showed that, when adding
POWERFUL BRAND INFLUENTIALS 113

the PBIS into the model, brand attachment kept the same level of predictive power (β = .73, p
< .001) while the PBIS is insignificant (β = .01, p > .05). Moreover, the inclusion of the PBIS
into the model did not significantly contribute to additional variance explained in separation
distress (an insignificant increment of R
2
change assessed by F change = .01, p > .05 from the
first model to the second model). These results support P7a, showing that brand attachment fully
explains separation distress.
Brand commitment was tested in the same sequential way as separation distress. In the
first step, brand attachment by itself significantly influence brand commitment (β = .76, p < .001).
In the second step, when estimated simultaneously, brand attachment is a stronger predictor (β
= .59, p < .001) than the PBIS (β = .24, p < .01) of brand commitment. While we find that the
PBIS does not add any additional prediction for separation distress beyond brand attachment, the
results on brand commitment suggest that the powerful brand influential construct provides
additional contribution to predict brand commitment (a significant increment of R
2
change
assessed by F change = 8.05, p < .01 from the first model to the second model). Nevertheless, as
proposed, brand attachment exceeds the powerful brand influential in predictive powerful for
separation distress as well as brand commitment. These results support P7a-P7b. More
importantly, these results demonstrate that the powerful brand influential construct and brand
attachment are two distinct constructs – the powerful brand influential is not the representation of
a high level of brand attachment.
Comparing predictive validity of the PBIS and brand attachment for intention to
conduct brand loyal behavior (P8). We proposed that the powerful brand influence construct
and brand attachment both contribute independently significant prediction of intention to conduct
brand loyal behavior (P8). To test the similarity between the two constructs, both the PBIS and
POWERFUL BRAND INFLUENTIALS 114

brand attachment were simultaneously regressed on the measure of brand loyal behaviors. The
results confirmed that, when simultaneously estimated, both the PBIS (β = .43, p < .001) and
brand attachment (β = .27, p < .01) significantly predict this behavioral outcome. As shown by
our results, powerful brand influential consumers are not only highly engaged in brand word-of-
mouth and advocacy behaviors but also in brand loyal behaviors. The rest of strongly brand-
attached consumers are highly engaged in the latter forms of brand commitment behaviors but
not the former types of brand commitment behaviors. From a marketing standpoint these results
show the marketplace value of the powerful brand influential consumers.

POWERFUL BRAND INFLUENTIALS 115

Chapter 9: Study 3: The Powerful Brand Influential is Distinct from Market Maven,
Opinion Leader, Brand Attachment, Brand Attitude Strength, and Brand Satisfaction
9.1. Objective and Method

The first objective of Study 3 is to establish the robustness and further test the validity of
the PBIS scale by using a non-student consumer sample from a different product category
(financial services) and in a different country. The second objective is to replicate some key
propositions that have been tested in Study 2 but with a much larger sample size. The third
objective is to more precisely represent the higher-order structure of the brand attachment scale
and to retest some key propositions as well as testing additional propositions. The final objective
is to further establish the convergent, discriminant and predictive validity of the PBIS by
introducing additional brand-level constructs (i.e., brand attitude strength and brand satisfaction)
and testing related propositions.
One of the largest banks in Austria served as the tested brand. A random sample of two
thousand customers of a bank branch was invited to participate in a mail survey as part of a
research project. Respondents were informed of the voluntary nature of the survey and were
assured complete confidentiality of their responses. The survey was translated from English to
German and the German version was then blind back-translated into English. The two versions
were compared by two native speakers who were blind to the purpose of the study.
Final sample. Prior to the data collection, a subsample of forty-one customers was
randomly selected for telephone interviews to ensure comprehension of all measures. The survey
was then pretested on thirty-five bank customers to further refine our measures. Our final
surveys were mailed to bank customers in two waves with stamped prepaid envelopes. We
obtained a total of six hundred ninety seven usable questionnaires, with a response rate of 35%.
POWERFUL BRAND INFLUENTIALS 116

There were no significant differences in responses to measures across the two waves (p > .10)
and hence data were pooled for further analyses.
The final sample acquired is heterogeneous, ranging in age from 18 – 69 with an
approximate mean age of thirty eight. The sample is 34% female, 66% male; 10% have a middle
school degree, 70% have a high school degree, and 20% have a degree from an institution of
higher education; annual gross income ranges from 12,000 Euro to 255,000 Euro, with the
median income of 34,000 Euro; 62% are married, 31% are single; 31% have no children, 48%
have one child, and 21% have more than one child.
Measures. The survey included the PBIS and measures of market mavens, opinion
leadership, and brand attachment in addition to a set of scales to further evaluate the validity of
the PBIS. Moreover, respondents’ purchase data was acquired directly from the bank. The data
of total portfolio value of a customer (as described later) is used to assess and compare the
objective value of the Powerful Brand Influential with related constructs. The survey also
included demographic variables (e.g., age, gender, education, income, marital status, and number
of children).
9.2. Measurement Results of the PBIS
Scale norm. The distribution of the PBIS across Study 1, Study 2 and Study 3 was
relatively stable. The mean score of the PBIS in Study 3 sample was 4.06 (range = 1 - 9,
standard deviation = 1.94). Mean scores of individual PBIS items ranged from 3.92 to 4.42 with
standard deviations ranged from 1.78 to 2.43. The coefficient alpha was .99 with item-to-total
correlations ranging from .93 to .99.

POWERFUL BRAND INFLUENTIALS 117

Scale dimensionality. Similar to Study 2, two steps of analysis (EFA followed by CFA
modeling) were conducted to confirm the dimensionality and to establish the stability of the
PBIS with a different and much larger sample. In the first step, Principle Component Analysis on
the PBIS consistently supports that the scale is unidimensional. With a much larger sample size,
the total variance explained, the factor loadings and the reliability coefficient generally increased
in values. The scale explained 97% of the total variance (factor loadings ranging from .95 to .99)
(see Table 7-4, results for Study 3).
In the second step, CFA results on the PBIS (see Table 8-1, results for Study 3) showed
that all standardized factor loadings were significant (p < .001) with an average standardized
factor loading of .97. The fit statistics of NFI, CFI, and GFI all exceeded the .90 cutoff value
(Bentler 1992) at the same time the SRMR is satisfactorily small, meeting the less than the .05
recommendation (Byrne 1998). The construct reliability surpasses the criterion of .70 and the
average variance extracted is greater than the criterion of .50 (Fornell and Larcker 1981). With a
large sample size and its’ ability to provide robust model estimates, these CFA results advance
good confidence in the PBIS measurement model.
9.3. Testing Convergent Validity and Discriminant Validity of the PBIS in Relation to the
Market Maven, Opinion Leadership, and Brand Attachment Constructs (P1 – P2)
One key objective of Study 3 is to replicate the results showing that the powerful brand
influential is significantly related to but distinctly different from the construct of market maven
(P1a-P2a), opinion leader (P1b-P2b) and brand attachment (P1d-P2d). Brand ambassador was
excluded from the current study since the phenomenon is less relevant to the examined brand (a
foreign bank). The same scales as Study 2 were used to measure opinion leader (α = .98), market
maven (α =.97), and the brand attachment (α = .89).
POWERFUL BRAND INFLUENTIALS 118

Measurement model of the PBIS, market maven, opinion leadership, and brand
attachment. EFA was conducted using oblique rotation on the four measurement scales. The
results consistently supported the convergent and discriminant validity of the PBIS in relation to
the other three constructs. A four-factor structure (eigenvalues > 1) resulted from the EFA with
all items loading on the corresponding concept and a total explained variance of 91.13%. The
first factor was the PBIS factor (factor loadings ranged from .95 to .99; explained variance =
62.41%). The second factor was the opinion leadership factor (factor loadings ranged from .96
to .97; explained variance = 15.15%). The third factor was the market maven factor (factor
loadings ranged from .94 to .98; explained variance = 9.26%), followed by the fourth factor of
brand attachment (factor loadings ranged from .81 to .92; explained variance = 4.31%).
The correlations among consumer group factors in Study 3 generally increase from Study
2, which is most likely due to the much larger sample size. Nevertheless, the structure of the
correlation between the PBIS and the related constructs resembled that in Study 2. As
hypothesized, the PBIS factor is significantly related to all three constructs (p < .001, in all
cases). It was most strongly correlated with the brand attachment factor (r = 0.58), followed by
the opinion leadership factor (r = 0. 49), and least with market maven factor (r = 0.47). All three
factors are positively correlated with each other as predicted (Opinion Leadership - Market
Maven, r = 0.66; Opinion Leadership - Brand Attachment, r = 0.53; Brand Attachment - Market
Maven, r = 0.50; p < .001, in all cases).

POWERFUL BRAND INFLUENTIALS 119

Five CFA models were constructed to further substantiate the validity of the PBIS. Model
1 was the proposed theoretical structure of the PBIS, which constructs a correlated yet distinct
four-factor structure for the PBIS, opinion leadership, market maven, and brand attachment
constructs. As shown in Table 9-1, all items loaded significantly on their predicted factors (p
< .001), the construct reliability (CR) of all measures well exceeded the criterion of .70 and the
average variance (AVE) extracted for all factors was well above the criterion of .50.

TABLE 9-1. Confirmatory Factor Analysis Results of Model 1 – Four Correlated Factors
on the PBIS and Measures of 3 Related Constructs

Factor Factor Indicators
Standardized Factor
Loadings
AVE CR
The Powerful Brand Influential
PBIS 1. .99 .87 .98
PBIS 2. .99
PBIS 3. .98
PBIS 4. .99
PBIS 5. .92
PBIS 6. .97
Opinion Leader
OL 1. .98 .86 .97
OL 2. .89
OL 3. .98
OL 4. .98
OL 5. .87
OL 6. .86
Market Maven
MM 1. .95 .87 .95
MM 2. .89
MM 3. .95
Brand Attachment
BA 1. (BSC 1) 0.96 .84 .95
BA 2. (BSC 2) 0.88
BA 3. (PRO 1) 0.87
BA 4. (PRO 2) 0.95
Note:
Factor loadings are standardized factor loadings. All factor loadings are significant at p < .001.
Abbreviations: CR = Construct Reliability. AVE = Average Variance Extracted. PBIS (the Powerful Brand
Influential Scale). OL (Opinion Leader). BA (Brand Attachment). PRO (Prominence of Thoughts and
Feelings indicator of brand attachment). BSC (Brand-Self Connection indicator of brand attachment). MM
(Market Maven).

POWERFUL BRAND INFLUENTIALS 120

Evidence for convergent validity of the PBIS was further demonstrated by significantly
positive interfactor correlation (Φ) coefficients between the PBIS and all three factors in Model 1
(see Table 9-2). The pattern of Φ coefficients also confirmed that the powerful brand influential
is most related to the concept of brand attachment, next to opinion leadership, and least to market
maven. Although the Φ coefficients are significant, none of the Φ coefficient is higher than the
correlation cutoff value of .90 (Bagozzi et al. 1991) and all 6 pair-wise coefficients squared are
less than the AVE (reported in Table 9-1) extracted for the corresponding factors (Fornell and
Larcker 1981). Therefore, there is no concern about the discriminant validity of the factors in
Model 1.

TABLE 9-2. Interfactor (Φ) Correlations of Model 1 – Correlations between 4 Factors
based on Table 9-1. Confirmatory Factor Analysis Results

Factor
Interfactor Correlations (Φ)
(t-Value; Std. Error)
The Powerful
Brand Influential
Opinion
Leader
Market
Maven
Brand
Attachment
The Powerful
Brand Influential
―
Opinion Leader

.47***
(11.56; 0.04)
―
Market Maven

.42***
(10.90; 0.04)
.43***
(11.40; 0.04)
―
Brand Attachment

.57***
(13.35; 0.04)
.55***
(12.96; 0.04)
.53***
(13.18; 0.04)
―
Note:
*** Φ coefficients are significant at p < .001 for two-tailed tests.

To further establish discriminant validity of the PBIS, the proposed Model 1 was
compared to a set of alternative conceptual models. It is particularly important to take a model
comparison approach to confirm the discriminant validity since all four factors have rather high
correlations. At the same time, the sample size is ample to allow for estimating CFA in a robust
way. Alternative models, Model 2 – Model 5, were constructed following the same conceptual
POWERFUL BRAND INFLUENTIALS 121

structure to those in Study 2 (except that Study 2 had the additional factor of brand
ambassadorship). Table 9-3 reports the results of each model and comparison results of each
alternative model to the proposed Model 1 on model fitness and chi-square difference test. The
results of Study 3 replicated the conclusions found in Study 2. Relatively, Study 3 results showed
more robust, clear cut model estimates and comparisons than did Study 2 results due to the
advantage of a larger sample size for CFA modeling. First, the hypothesized Model 1 is a good
fitting model by itself (NFI and CFI both exceed the .90 cutoff value and the SRMR is less than
the .05 recommendation). In Study 3, all alternative models failed the cutoff value of .90 for NFI
and CFI (Bentler 1992) and the recommended SRMR of less than .05 (Byrne 1998).

POWERFUL BRAND INFLUENTIALS 122

TABLE 9-3. Results for Comparing Models of Confirmatory Factor Analysis on the PBIS
and Measures of 3 Related Constructs

Model
χ²
(df)
NFI CFI SRMR
Model
Comparison
∆χ²
(∆df)
∆p
M 1: Four Correlated Factors
333.64
(145)
.95 .96 .042 NA NA NA
M 2: Four Uncorrelated Factors
667.57
(152)
.74 .78 .44 M2-M1
333.93
(7)
.001
M 3: Three Correlated Factors –
PBIS & OL belong to the same
factor
355.60
(146)
.83 .89 .33 M3-M1
21.96
(1)
.001
M 4: Three Correlated Factors –
PBIS & MM belong to the same
factor
350.29
(146)
.83 .90 .25 M4-M1
16.65
(1)
.001
M 5: Three Correlated Factors –
PBIS & BA belong to the same
factor
346.27
(146)
.83 .91 .24 M5-M1
12.63
(1)
.001
Note:
The chi-square statistics are significant at p < .001.
Abbreviations: NFI = Normed Fit Index. CFI = Comparative Fit Index. SRMR = Standardized Root Mean
Square Residual. ∆χ² (∆df) = Chi-Square Difference Tests. ∆p = Significance of Chi-Square Difference
Tests. Brand Advocates (PBIS). Market Mavens (MM). Opinion Leaders (OL). Brand Attachment (BA).
Model Specification:
M 1: Four Correlated Factors: The same model as Table 9-2, assumes that PBIS, OL, MM, and
BA are 4 factors, allowing factors to correlate freely.
M 2: Four Uncorrelated Factors: The same 4 factors as M 1 except that correlation between 4
factors is set to zero.
M 3: Three Correlated Factors – PBIS & OL belong to the same factor: assumes that PBIS
and OL are loaded on 1 factor while MM and BA are separate factors, allowing factors to correlate freely.
M 4: Three Correlated Factors – PBIS & MM belong to the same factor: assumes that PBIS
and MM are loaded on 1 factor while OL and BA are separate factors, allowing factors to for correlate
freely.
M 5: Three Correlated Factors – PBIS & BA belong to the same factor: assumes that PBIS
and BA are loaded on 1 factor while OL and MM are separate factors, allowing factors to for correlate
freely.
POWERFUL BRAND INFLUENTIALS 123

Second, Model 1 is the best fitting model among all models with the lowest χ² statistic
and SRMR at the same time highest NFI, and CFI among all models. Third, the chi-square
difference tests showed that the chi-square of Model 1 is significantly better [∆χ²(∆df) is at p
< .001] than each and all alternative models. Additionally, the model comparison results showed
the exact pattern as Study 2 by showing that (1) the worst fitting model is Model 2, which
assumes that all four factors are uncorrelated concepts; (2) the next best model to Model is
Model 5, which supported that the PBIS is most correlated to brand attachment; and (3) opinion
leadership and market maven status are relatively the most differentiated from the PBIS,
indicated by the comparison results of Model 2 and Model 3 to Model 1.
With a much larger sample size than Study 2 generated from a different category, Study 3
replicated the support for our proposition that the concept of powerful brand influential is
significantly related to but distinctly different from the market maven, opinion leadership, and
brand attachment constructs (P1 – P2). Moreover, Study 3 establishes additional confidence in
the convergent and discriminant validity of the PBIS as a measure to capture the construct of the
powerful brand influential.
9.4. Testing Convergent Validity and Discriminant Validity of the PBIS in Relation to
Brand Attitude Strength and Brand Satisfaction
Another objective of Study 3 is to establish the validity of the PBIS in relation to a set of
brand-level constructs including brand satisfaction, brand attitude strength, and brand attachment.
We proposed that the powerful brand influential construct is significantly related to but distinctly
different from the construct of brand attitude strength (P12a) and brand satisfaction (P12b). This
proposition pertains to the convergent and discriminant validity of the powerful brand influential
construct and the PBIS in relation to the two brand-level constructs. Furthermore, we concentrate
POWERFUL BRAND INFLUENTIALS 124

on assessing predictive validity of the PBIS relative to brand satisfaction, brand attitude strength,
and brand attachment for several key behavioral outcomes related to a given brand. The previous
section has provided evidence of the convergent and discriminant validity of the powerful brand
influential construct in relation to the brand attachment construct. Nevertheless, brand
attachment was included again here with brand satisfaction and brand attitude strength and
compared with the PBIS to predict brand-specific behavior outcomes. Since brand attachment is
significantly related to brand satisfaction and brand attitude strength, inclusions of brand
attachment is crucial for demonstrating that the additional predictive power of the PBIS after
accounting for that of brand attachment.
Measures of brand satisfaction and brand attitude strength. Brand satisfaction was
measured by 3 items including “I am satisfied with the brand”, “I am content with the brand”,
and “I am happy with the brand” on a 9-point scale ranging from “strongly disagree” to “strongly
agree” (α = .97). This scale intends to capture a consumer’s satisfaction level with a brand in
general instead of satisfaction driven by situational or specific transaction outcomes. The items
were worded by taking key items from Oliver’s (1997) consumption satisfaction scale and
adaptations of the scale (Zboja and Voorhees 2006).
As previously discussed, brand attitude strength is conceptualized in this paper as brand
attitude valence weighted by the confidence/certainty with which this attitude is held. Following
the measurement approach by Park et al. (2010), our measure included both items measuring
brand valence and items measuring confidence/certainty. Specifically, brand valence was
measured by 3 items (α = .97) on a 9-pointsemantic differential scale that indicates the extent to
which the brand is “unfavorable” (1) to “favorable” (9), “bad” (1) to “good” (9), and “negative”
(1) to “positive” (9). Confidence/certainty with which the attitude is held was also measured by 3
POWERFUL BRAND INFLUENTIALS 125

items (α = .90) on a 9-pointsemantic differential scale: the first item measured the extent to
which the brand is “not important” (1) to “extremely important” (9); the second item measured
the extent to which respondents are certain about their brand evaluation ranging from “not
certain” (1) to “extremely certain” (9); and the last item measured the extent to which
respondents are confident about their brand evaluation ranging from and “not confident” (1) to
“extremely confident” (9). Following the recommendation by Park et al. (2010), the brand
attitude strength measure was constructed by the average score of all brand valence items
multiplied by the average score of all confidence/certainty items.
We conducted EFA on all brand attitude items including both brand valence and
confidence/certainty items. The EFA results showed that all items loaded on a single factor
(factor loadings ranged from .80 to .91 with a total explained variance of 70%), which suggests
that a single composite score that takes into account all items adequately reflects the single
dimensionality of the brand attitude strength scale.
Measurement model of the PBIS, brand attitude, brand satisfaction and brand
attitude strength (P12a – P12b). EFA was conducted on all brand-level construct measures. A
five-factor solution was expected for the four brand-level constructs since brand attachment
reflects BSC and PRO as separate second-order factors. The EFA results confirmed that all items
of the PBIS loaded on the first factor (factor loadings ranged from .90 to .94; explained variance
= 39.25%). The second factor included all items of brand satisfaction (factor loadings ranged
from .78 to .85; explained variance = 21.95%). The third and fourth factor included the PRO
component (factor loadings ranged from .70 to .85; explained variance = 12.28%) and the BSC
component of brand attachment (factor loadings ranged from .70 to .81; explained variance =
POWERFUL BRAND INFLUENTIALS 126

11.51%). The fifth factor was brand attitude strength (factor loadings ranged from .69 to .82;
explained variance = 8.48%).
We also conducted a CFA to represent the five-factor measurement model with all items
of the PBIS, brand satisfaction, brand attachment and brand attitude strength loading on their
intended constructs. Brand attachment was constructed as a second-order factor structure that
included both Brand-Self Connection (BSC) and Prominence of Thoughts and Feelings (PRO).
All factors were allowed to correlate to reflect the conceptually related nature of these constructs.
As reported in Table 9-4, the CFA results are favorable since all items loaded significantly on
their intended factors (p < .001) with high construct reliability (CR) exceeding the criterion
of .70 and average variance (AVE) exceeding the criterion of .50.

POWERFUL BRAND INFLUENTIALS 127

TABLE 9-4. Confirmatory Factor Analysis Results of Model 1 – Four Correlated Factors
including the PBIS and Brand Attachment, Satisfaction, and Attitude Strength

Factor Factor Indicators
Standardized
Factor
Loadings
AVE CR
The Powerful Brand Influential
PBIS 1. .99 .87 .98
PBIS 2. .99
PBIS 3. .98
PBIS 4. .99
PBIS 5. .92
PBIS 6. .97
Brand Attachment
BA 1. (BSC 1) .97 .84 .95
BA 2 .(BSC 2) .87
BA 3. (PRO 1) .85
BA 4. (PRO 2) .96
Brand Satisfaction
BS 1. .96 .87 .95
BS 2. .89
BS 3. .96
Brand Attitude Strength
BAS 1. (Brand Valence 1) .92 .82 .93
BAS 2. (Brand Valence 2) .96
BAS 3. (Brand Valence 3) .96
BAS 4. (Confidence/Certainty 1) .68
BAS 5. (Confidence/Certainty 2) .70
BAS 6. (Confidence/Certainty 3) .69
Note:
Factor loadings are standardized factor loadings. All factor loadings are significant at p < .001.
Abbreviations: CR = Construct Reliability. AVE = Average Variance Extracted. PBIS (the Powerful Brand
Influential Scale). BA (Brand Attachment). PRO (Prominence of Thoughts and Feelings indicator of brand
attachment). BSC (Brand-Self Connection indicator of brand attachment). BS (Brand Satisfaction). BAS
(Brand Attitude Strength).

Table 9-5 reports the interfactor correlation (Φ coefficients) between all constructs. The
PBIS, brand satisfaction, brand attachment and brand attitude strength are all significantly related
to each other. Convergent validity of the PBIS is supported by the fact that the PBIS is
significantly related to these brand-level constructs. Although the Φ coefficients are relatively
high, none of the Φ coefficients is higher than the correlation cutoff value of .90 (Bagozzi et al.
1991) and all 6 pair-wise coefficients squared are less than the AVE (reported in Table 9-4)
POWERFUL BRAND INFLUENTIALS 128

extracted for the corresponding factors (Fornell and Larcker 1981). Therefore, there is no
concern about the discriminant validity of the factors in the model.
Moreover, the PBIS has relatively lower Φ coefficients with the other constructs than the
Φ coefficients between those previously established constructs (i.e., Φ coefficients between
brand attitude strength and brand satisfaction, between brand satisfaction and brand attachment,
and between brand attitude strength and brand attachment). This pattern provides additional
confidence in the discriminant validity of the PBIS. Overall, the measurement results support the
convergent and discriminant validity of the PBIS in relation to brand attachment, brand
satisfaction, and brand attitude strength. Thus, these results support P12, showing that the
powerful brand influential construct is significantly related to but distinctly different from the
construct of brand attitude strength (P12a) and brand satisfaction (P12b).

TABLE 9-5. Interfactor (Φ) Correlations of Brand-Level Construct CFA – Correlation
between the PBIS, Brand Attachment, Brand Satisfaction, and Brand Attitude Strength

Factor
Interfactor Correlations (Φ)
(t-Value; Std. Error)
The Powerful
Brand Influential
Brand
Attachment
Brand
Satisfaction
Brand
Attitude Strength
The Powerful
Brand Influential
―
Brand Attachment

.57***
(13.37; 0.04)
―
Brand Satisfaction

.58***
(13.21; 0.04)
.70***
(15.23; 0.05)
―
Brand Attitude
Strength
.66***
(14.56; 0.05)
.67***
(14.90; 0.04)
.74***
(15.75; 0.05)
―
Note:
*** Φ coefficients are significant at p < .001 for two-tailed tests.

POWERFUL BRAND INFLUENTIALS 129

9.5. Testing Predictive Validity of the PBIS in Relation to Market Maven, Opinion Leader
and Brand Attachment (P4a – 4d)
The ability to predict brand advocacy behavior lies at the core of the powerful brand
influential construct. Hence, we replicate predictive validity test of the PBIS for brand advocacy
behavior and compare the PBIS to market maven, opinion leadership and brand attachment
measures regarding this key outcome. We measured a consumer’s willingness to conduct three
key brand advocacy behaviors in the future (α = .75), which was a subset of the brand advocacy
behavioral lists used in Study2. The adopted items include recommending the brand to others,
defending the brand to others who speak negatively of the brand, and donating time to promote
the brand community. These items represent the type advocacy behaviors that are most relevant
considering that the brand here is a bank. Each item is measured on a 9-point scale ranging from
“not at all” to “extremely”.
To replicate predictive validity of the PBIS in comparison to other predictors, we adopted
a different type of analysis from Study 2. While Study 2 used hierarchical regression analysis to
examine predictive power of the PBIS sequentially to other predictors, Study 3 uses structural
modeling, which simultaneously evaluates predictive ability of the PBIS and all other predictors.
One of the key advantages of structural modeling is allowing us to represent brand attachment as
a higher-order construct indicated by both BSC and PRO indicators. Regression analysis in
Study 2 operationalized brand attachment using the average of BSC and PRO items, which
assumes a single-order structure of brand attachment and potentially underestimates predictive
ability of brand attachment. Study 3 advances Study 2 by representing brand attachment as a
second-order factor in the structural modeling and replicates our test of the prediction of this
construct on the key outcome variables.
POWERFUL BRAND INFLUENTIALS 130

We constructed a structural model (see Figure 9-1) that included the PBIS, brand
attachment, market maven, and opinion leader measures as the predictor variables and brand
advocacy behavior as the outcome variable. The results showed that the PBIS significantly
predicts brand advocacy behavior (γ = 0.17, p < .01), whereas none of the brand attachment,
market maven, and opinion leadership measures measures significantly predicts brand advocacy
behavior. Consistent with Study 2, Study 3 results confirmed predictive validity of the PBIS for
brand advocacy behavior and our proposition that the powerful brand influential construct
predicts brand advocacy behavior better than does the construct of market maven (P4a), opinion
leader (P4b), and brand attachment construct (P4d).

POWERFUL BRAND INFLUENTIALS 131

FIGURE 9-1. Simultaneous Comparison of Predictive Ability of the PBIS and Brand
Attachment, Market Maven, and Opinion Leader for Brand Advocacy Behaviors

χ² = 25.43, NFI = 0.99, CFI = 0.99, SRMR = 0.017

Note:
The three values reported above the associated path are in the sequence of γ estimation, standard error,
and t value. Significant paths are shown by solid lines. Insignificant paths are shown by dotted lines.
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.
Abbreviations: NFI = Normed Fit Index. CFI = Comparative Fit Index. SRMR = Standardized Root Mean
Square Residual. BSC (Brand-Self Connection). PRO (Prominence of Thoughts and Feelings). PBIS (the
Powerful Brand Influential Scale). CPAM (Connection-Prominence Attachment Model).

9.6. Comparing Predictive Validity of the PBIS and Brand Attachment for Behavioral and
Psychological Outcomes (P4d, P7 and P8)
We concentrate on comparing predictive ability of the PBIS and brand attachment for
two behavioral outcomes (i.e., brand advocacy behavior and brand loyal behavior) and two
psychological outcomes (i.e., separation distress and brand commitment). We intended to
replicate the results of Study 2 in supporting three key propositions: First, the powerful brand
BSC
PRO

The Powerful
Brand
Influential
(PBIS)
Brand
Attachment
(CPAM)

Market
Maven
0.17
(0.04)
4.79***

Brand
Advocacy
Behavior
0.05
(0.04)
1.13
0.02
(0.05)
0.45
0.02
(0.06)
0.47

Opinion
Leader
POWERFUL BRAND INFLUENTIALS 132

influential construct predicts brand advocacy behavior better than does brand attachment (P4d).
This proposition is partially supported by the previous structural model that included the PBIS
and brand attachment as a subset of predictive variables of brand advocacy behavior. We repeat
our test of this proposition by excluding market mavens and opinion leaders from the model and
providing a more direct comparison between the PBIS and brand attachment. Second, the brand
attachment construct better predicts separation distress (P7a) and brand commitment (P7b) than
does the powerful brand influential construct. Third, the powerful brand influential and brand
attachment constructs both contribute independently significant predictions for intention to
conduct brand loyal behavior (P8).
We measured intention to conduct brand loyal behavior (α = .88) and separation distress
(α = .74) using the same scales as in Study 2. We measured brand commitment using two items
that were adopted from the scale of De Wulf, Odekerken-Schröder and Iacobucci (2001). The
wording of the items was tailored toward the current context. The two items (α = .70) measured a
consumer’s willingness and intent to sustain a long-term relationship with the brand into the
future. Specifically, the items were “I am willing ‘to go the extra mile’ to keep banking with the
brand” and “Even if the brand would not be located in a convenient place, I would still keep
doing business with it” on a 9-point scale (0 = strongly disagree to 9 = strongly agree).
We tested all three propositions with a set of structural models. We intend to first show
the independent ability of the PBIS and brand attachment to separately predict the four outcomes
of interest. Then, we include both the PBIS and brand attachment as predictors to directly
compare the relatively predictive ability of both constructs for the four outcomes. Specifically,
the first model included only the PBIS as the predictor of the four outcome variables (see Figure
9-2, Panel A). The second model included only brand attachment as the predictor of the same
POWERFUL BRAND INFLUENTIALS 133

outcome variables (see Figure 9-2, Panel B). The results of these two models showed that the
PBIS and brand attachment, when independently estimated, both significantly predict brand
advocacy behavior (the PBIS as the predictor: γ = .23, p < 0.01; brand attachment as the
predictor: γ = .13, p < 0.01), intention to conduct brand loyal behavior (the PBIS as the predictor:
γ = .55, p < 0.001; brand attachment as the predictor: γ = .53, p < 0.001), separation distress (the
PBIS as the predictor: γ = .33, p < 0.01; brand attachment as the predictor: γ = .41, p < 0.001)
and brand commitment (the PBIS as the predictor: γ = .57, p < 0.001; brand attachment as the
predictor: γ = .67, p < 0.001).

POWERFUL BRAND INFLUENTIALS 134

FIGURE 9-2. Predictive Ability of the PBIS and Brand Attachment for Behavior Outcomes

Panel A: Predictive ability of the PBIS for behavior outcomes

χ² = 158.39, NFI = 0.86, CFI = 0.86, SRMR = 0.092

Note:
The three values reported above the associated path are in the sequence of γ estimation, standard error,
and t value. Significant paths are shown by solid lines. Insignificant paths are shown by dotted lines.
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.
Abbreviations: NFI = Normed Fit Index. CFI = Comparative Fit Index. SRMR = Standardized Root Mean
Square Residual. PBIS (the Powerful Brand Influential Scale).

The Powerful
Brand
Influential
(PBIS)
0.55
(0.03)
17.18***
0.23
(0.04)
6.34**

Separation
Distress
Brand
Loyal
Behavior
Brand
Advocacy
Behavior
0.33
(0.04)
9.21**

Brand
Commitment
0.57
(0.03)
18.33***

POWERFUL BRAND INFLUENTIALS 135

FIGURE 9-2. Predictive Ability of the PBIS and Brand Attachment for Behavior Outcomes

Panel B: Predictive ability of Brand Attachment for behavior outcomes

χ² = 160.74, NFI = 0.93, CFI = 0.93, SRMR = 0.079

Note:
The three values reported above the associated path are in the sequence of γ estimation, standard error,
and t value. Significant paths are shown by solid lines. Insignificant paths are shown by dotted lines.
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.
Abbreviations: NFI = Normed Fit Index. CFI = Comparative Fit Index. SRMR = Standardized Root Mean
Square Residual. BSC (Brand-Self Connection). PRO (Prominence of Thoughts and Feelings). PBIS (the
Powerful Brand Influential Scale). CPAM (Connection-Prominence Attachment Model).

The last model (see Figure 9-2, Panel C) included both the PBIS and brand attachment as
predictors of the four outcomes. The results showed that, when simultaneously estimated, (1) the
PBIS significantly predicts brand advocacy behavior (γ = .23, p < 0.001) while brand attachment
becomes insignificant (γ = .01, p > 0.05). This result supports P4d. (2) Both the PBIS (γ = .36, p
< 0.001) and brand attachment (γ = .33, p < 0.001) remain significant predictors of intention to
perform brand loyal behavior and at similarly influential levels judging by the difference
0.53
(0.03)
17.38***
0.13
(0.04)
3.63**

Separation
Distress

Brand
Loyal
Behavior

Brand
Advocacy
Behavior
0.41
(0.03)
12.39***

BSC
PRO

Brand
Attachment
(CPAM)
0.67
(0.03)
24.20***

Brand
Commitment
POWERFUL BRAND INFLUENTIALS 136

between the two γ coefficients (z = .6, p > .05). This result supports P8. (3) Brand attachment
predicts separation distress (γ = .32, p < 0.001) better than does the powerful brand influential (γ
= .15, p < 0.01). The difference between the γ coefficients is significant (z = 3.01, p < .01). This
result supports P7a. (4) Brand attachment predicts brand commitment (γ = .51, p < 0.001) better
than does the powerful brand influential (γ = .28, p < 0.001). Again, the difference between the
two coefficients is significant (z = 5.42, p < .001). This result supports P7b.

POWERFUL BRAND INFLUENTIALS 137

FIGURE 9-2. Predictive Ability of the PBIS and Brand Attachment for Behavior Outcomes

Panel C: Simultaneous comparison of predictive ability the PBIS and Brand Attachment for
behavior outcomes

χ² = 104.46, NFI = 0.97, CFI = 0.97, SRMR = 0.047

Note:
The three values reported above the associated path are in the sequence of γ estimation, standard error,
and t value. Significant paths are shown by solid lines. Insignificant paths are shown by dotted lines.
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.
Abbreviations: NFI = Normed Fit Index. CFI = Comparative Fit Index. SRMR = Standardized Root Mean
Square Residual. BSC (Brand-Self Connection). PRO (Prominence of Thoughts and Feelings). PBIS (the
Powerful Brand Influential Scale). CPAM (Connection-Prominence Attachment Model).

The Powerful
Brand
Influential
(PBIS)
0.36
(0.04)
10.05***
0.23
(0.04)
5.32***

Separation
Distress

Brand
Loyal
Behavior

Brand
Advocacy
Behavior
BSC
PRO

Brand
Attachment
(CPAM)
0.33
(0.03)
10.05***
0.01
(0.04)
0.19
0.15
(0.04)
3.66**
0.32
(0.04)
8.46***

Brand
Commitment
0.28
(0.03)
8.83***
***
0.51
(0.03)
16.35***
POWERFUL BRAND INFLUENTIALS 138

9.7. Comparing the PBIS and Brand Attachment on Emotional Expressiveness
We discussed that emotional expressiveness might be an individual trait that
differentiates powerful brand influential consumers from other strongly brand-attached
consumers. As previously noted, emotional expressiveness captures individuals’ dispositional
tendency to outwardly express emotions. We expect that strongly brand-attached consumers who
are emotional expressive find the act of brand word-of-mouth intrinsically pleasant and
correspondingly tend to engage in brand word-of-mouth. The other way around, those who have
low emotional expressiveness are inhibited from engaging in the act of brand word-of-mouth and
correspondingly are unlikely to be powerful brand influential consumers. Therefore, the validity
of the PBIS would be further established by showing that when estimated simultaneously, the
PBIS leads to emotional expressiveness while brand attachment does not.
We measured emotional expressiveness using a short-version of the Emotional
Expressivity Scale (EES) by Kring et al. (1994). The scale consists of six items (see Appendix C
for the list of emotional expressiveness items) and yielded high internal consistency (α = .99). A
structural model was constructed to compare the PBIS and brand attachment as two predictors of
emotional expressiveness. The model result showed that the PBIS is a significant predictor (γ =
.78, p < 0.001) while brand attachment is not a significant predictor (γ = 0.02, p > 0.05) of
emotional expressiveness. This result indicates that powerful brand influential consumers tend to
be emotional expressive while the rest of strongly brand-attached consumers do not.

POWERFUL BRAND INFLUENTIALS 139

9.8. Testing Predictive Validity of the PBIS in Relation to Brand-Level Constructs for
Behavioral Outcomes (P8, P9, P13, P14, and P15)
In this section of the paper, we compare the PBIS and related brand-level constructs
regarding predictive validity for key brand behavioral outcomes. The brand-level constructs
included brand attachment, brand attitude strength, and brand satisfaction. We examined brand
attachment again in this set of analyses since it is important to show predictive power of the
PBIS after accounting for brand attachment. Using a structural model, we simultaneously test the
ability of the PBIS to predict four brand behavioral outcomes. The first brand behavioral
outcome is brand advocacy behavior. We proposed that the powerful brand influential construct
predicts brand advocacy behavior better than does the construct of brand attachment (P4a), brand
attitude strength (P13a), and brand satisfaction (P13b). The second behavioral outcome is
intention to perform brand loyal behavior. A set of propositions pertain to this behavioral
outcome: The powerful brand influential construct and each of (1) brand attachment (P8), (2)
brand attitude strength (P14a), and (3) brand satisfaction (P14b) contribute independently
significant predictions of intention to conduct brand loyal behavior.
The third behavioral outcome is actual purchase behavior. We proposed that the powerful
brand influential construct and brand attachment both bring independently significant
contributions to actual purchase behavior (P9) while the powerful brand influential construct
predicts actual purchase behavior better than does brand attitude strength (P15a) and brand
satisfaction (P15b). To measure actual purchase behavior, we acquired respondents’ purchase
data directly from the bank. The bank is interested in the ‘total worth’ of certain customer groups
and is willing to share with us the “total portfolio value” data on customers. The “total portfolio
value” data sums the value of the saving, investment, and all other types of accounts a customer
POWERFUL BRAND INFLUENTIALS 140

holds at the bank. This objective measure accounts for both the breadth (spreads across various
types of bank accounts) and depth (the absolute amount) of a customer’s actual brand purchase.
This measure also captures a customer’s all forms of repeated purchase behaviors towards the
brand.
Figure 9-3 depicts the structural model that simultaneously compares predictive power of
the PBIS and that of brand attachment, brand attitude strength, and brand satisfaction for all four
behavioral outcomes. The results of brand advocacy behavior reveal that, when simultaneously
estimated, the PBIS is the only significant predictor of brand advocacy behavior (γ = .21, p <
0.01). None of the other brand constructs is significant (brand attachment: γ = .02, brand attitude
strength: γ = .01, and brand satisfaction: γ = .04, all at p > 0.05). This result replicates
substantiation for P4d and supports additional propositions (P13a – 13b). We expected and
empirically find that the powerful brand influential construct predicts brand advocacy behavior
better than does brand attitude strength and brand satisfaction even accounting for brand
attachment.

POWERFUL BRAND INFLUENTIALS 141

FIGURE 9-3. Simultaneous Comparison of Predictive Ability of the PBIS, Brand
Attachment, Brand Satisfaction, and Brand Attitude Strength for Behavior Outcomes

χ² = 22.15, NFI = 0.99, CFI = 0.99, SRMR = 0.016

Note:
The three values reported above the associated path are in the sequence of γ estimation, standard error,
and t value. Significant paths are shown by solid lines. Insignificant paths are shown by dotted lines.
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.
Abbreviations: NFI = Normed Fit Index. CFI = Comparative Fit Index. SRMR = Standardized Root Mean
Square Residual. BSC (Brand-Self Connection). PRO (Prominence of Thoughts and Feelings). PBIS (the
Powerful Brand Influential Scale). CPAM (Connection-Prominence Attachment Model).

The Powerful
Brand
Influential
(PBIS)

Brand
Satisfaction
0.21
(0.05)
4.39**
0.40
(0.05)
8.51***
0.02
(0.04)
0.47
0.17
(0.03)
4.86**
0.21
(0.05)
4.39**
0.04
(0.06)
0.60

Brand
Loyal
Behavior

Actual
Purchase
Behavior

Brand
Advocacy
Behavior
0.10
(0.04)
2.47*
0.07
(0.06)
1.23
0.21
(0.04)
4.67**
0.01
(0.06)
0.24
0.01
(0.06)
0.12
BSC
PRO
Brand
Attachment
(CPAM)

0.18
(0.04)
4.68**
Brand Attitude
Strength
(Attitude Valence x
Confidence/
Certainty)
POWERFUL BRAND INFLUENTIALS 142

The results of intention to perform brand loyal behavior showed that, when
simultaneously estimated, the PBIS and all three brand constructs remain significant predictors
of intention to conduct brand loyal behavior. The PBIS (γ = .18, p < 0.01) and brand attachment
(γ = .17, p < 0.01) both contribute independently significant prediction of intention to conduct
brand loyal behavior and at similarly influential levels judging by the difference between the two
coefficients (z = .2, p > .05). This result replicates our previous support for P8. Additionally, the
PBIS (γ = .18, p < 0.01) and brand attitude strength (γ = .21, p < 0.01) both bring independently
significant contributions to intention to conduct brand loyal behavior and also at similarly
influential levels (z = -.53, p > .05). This result supports P14a. Last, the PBIS (γ = .18, p < 0.01)
and brand satisfaction (γ = .21, p < 0.01) both independently contribute to intention to conduct
brand loyal behavior and at similarly influential levels (z = -.47, p > .05). This result supports
P14b.
The results of actual purchase behavior showed that, when simultaneously estimated, the
PBIS (γ = .40, p < 0.001) and brand attachment (γ = .10, p < 0.05) both bring independently
significant contributions to actual purchase behavior. This result confirms P9. We noted earlier
that the powerful brand influential might predict actual purchase behavior better than does brand
attachment. We empirically find support for this speculation. The PBIS better predicts actual
purchase behavior than does brand attachment (z = 5.30, p < .001). Moreover, the PBIS predicts
actual purchase behavior (γ = .40, p < 0.001) better than does brand attitude strength (γ = .01, p >
0.05) and brand satisfaction (γ = .07, p > 0.05) even after accounting for brand attachment. These
results support P15a – 15b, respectively. Neither brand attitude strength nor brand satisfaction,
when simultaneously estimated with the PBIS, significantly predicts actual purchase of the
brand.
POWERFUL BRAND INFLUENTIALS 143

To summarize these results of behavioral outcomes, Study 3 provides converging
evidence that the powerful brand influential exceeds brand attachment in predicting brand
advocacy behavior and that both constructs contribute independently significant predictions for
intention to conduct brand loyal behaviors. Furthermore, Study 3 compares the ability of the
powerful brand influential to that of brand attitude strength and brand satisfaction in producing
desirable behavioral outcomes. From a marketing perspective, the powerful brand influential
construct and its scale capture a brand’s most valuable consumer – They have as strong intention
as other favorable consumers of a brand (i.e., other strongly brand-attached consumers,
consumers who hold strongly favorable attitude toward the brand, and consumers who are
satisfied with the brand) to conduct brand loyal behaviors, are more engaged in brand advocacy
and actual purchase behaviors than other favorable consumers of the brand.

POWERFUL BRAND INFLUENTIALS 144

Chapter 10: Study 4: Differentiating the Powerful Brand Influential from Brand
Attachment in the Context of a Positive Brand Event
10.1. Objective and Methodology

Study 4 has two purposes: 1) to confirm the stability of the PBIS using brands in another
product category and 2) to differentiate brand advocates from the most overlapped consumer
group, strongly brand-attached consumers, in a specific brand context (i.e., a positive brand
event). We conducted a field study in the arena of politics during the 2008 Presidential Election.
The two presidential candidates (Barack Obama and John McCain) served as the chosen brands.
We administered two online questionnaires; a pre-election questionnaire (collected within the
week before the Election Day) and a post-election questionnaire (collected within the week after
the Election Day). Respondents answered a list of questions regarding their thoughts and feelings
towards the 2008 Presidential Election and the presidential candidates. They were informed that
the survey was part of a doctoral dissertation and were promised confidentiality of their
responses.
Pre-election questionnaire sample. One hundred eighty two respondents from the state
of California (64.3%), Arizona (10.4%), and Connecticut (25.3%) completed the pre-election
survey. The majority (75.8%) were students recruited from two large universities in California
and Arizona. The rest (24.2%) were recruited from the three states via email invitation. The
sample ranged in age from 18 - 66, with an approximate mean age of 24. Fifty three percent were
female and 45.6% were male. Among the respondents were two special groups of consumers
recruited from the student organizations “Students for Obama” (N = 20, 11%) and “Students for
McCain” (N = 18, 10%). These individuals were the real life powerful brand influentials. We
POWERFUL BRAND INFLUENTIALS 145

intend to compare them with the rest of student respondents and test known group validity of the
PBIS.
Pre-election questionnaire measures. Respondents were informed that the purpose of
the survey was to understand people’s thoughts and feelings toward 2008 presidential candidates
and the upcoming election. They were also informed that there were two parts of the study: the
present questionnaire constitutes Part I; and Part II questionnaire will be emailed to them after
the presidential Election Day. The pre-election questionnaire included the PBIS, brand
attachment scale, and emotion measures. Brand attachment is measured using the same 4-item
scale as our previous studies (α = .80). The main purpose of the pre-election survey was to
acquire respondents’ scores of the PBIS and brand attachment scale. There were two versions of
the pre-election survey: a version with Obama as the targeted brand and a version with McCain
as the targeted brand. The two versions were identical except for the evaluated brand. We
distributed an even number of the two versions and obtained a relatively balanced final sample
for each brand (the Obama version: N = 87, 48%; the McCain version: N = 95, 52%). To
facilitate the test of known group validity, we administered the Obama version to Students for
Obama and the McCain version to Students for McCain. We expected these respondents would
score significantly higher on the PBIS than would the remaining respondents. We randomly
distributed the two versions to the remaining sample. There were no significant differences
between the two brand versions in responses measures (p > .10). Hence we pooled data from the
two versions for further testing.
All emotion measures were executed on a 9-point scale ranging from 1 = “not at all” to 9
= “very much”. Respondents were asked to indicate the extent to which they feel a list of
emotions about the upcoming election. The list included both positive and negative emotions
POWERFUL BRAND INFLUENTIALS 146

though we are only interested in the former and particularly positive self-conscious emotions and
challenge emotions. We adopted the 3-item measure of positive self-conscious emotions from
Ramanathan and Williams (2007) including “proud”, “self-respect”, and “confident” (α = .77).
The scale yielded a mean of 5.46 (range = 1 - 9, standard deviation = 2.05). We adopted two
items from Folkman and Lazarus (1985) to measure challenge emotions including “hopeful” and
“eager” (α = .82). The scale yielded a mean of 6.44 (range = 1 - 9, standard deviation = 2.24).
Post-election questionnaire sample. Following Election Day, we emailed all
respondents the post-election survey. One hundred thirty one (72%) of them completed the post-
election survey within one week after Election Day. We are only interested in those who
perceive the election as a positive event. We used two items to identify these respondents. The
first item asked respondents to indicate their intended candidate (“Obama”, “McCain”, or “not
sure”) and was administered in the pre-election survey. We dropped the data of those whose
intended candidate was McCain (N=30). The low responses incident from the lost party is not
surprising. The majority of those who faced the undesirable election outcome refrained from
attending the post-election survey. The second item measured goal congruence of the election
outcome, which was acquired from the post-election survey. Respondents were asked the extent
to which the election was consistent with what they wanted on a 9-point scale (1 = not at all to 9
= extremely). We dropped the data of those who responded lower than the neutral point of the
scale (the score of 5). We again checked the remaining respondents’ intended candidate. Six
respondents were further dropped from the final data, all of whom were undecided in their
intended candidate before the election. As such, we ensure that our final data (N = 95) only
includes respondents who indeed perceive the election as a positive event. The post-election
POWERFUL BRAND INFLUENTIALS 147

sample ranged in age from 18 – 66, with an approximate mean age of 25. Fifty eight percent
were female and 42% were male.
Post-election questionnaire measures. Respondents were informed that the purpose of
the survey was to understand people’s thoughts, feelings, and actions with respect to the election.
Then they answered measures of emotions, behaviors, and different types of word-of-mouth in
response to the election. The main objective of the post-election questionnaire was to examine
and compare how powerful brand influentials and other strongly brand-attached consumers react
to a positive brand event by testing a set of propositions.
Emotional response measures were executed on the same 9-point scale as the emotion
measures before the election (1 = not at all to 9 = very much). Respondents were asked to
indicate the extent to which they experienced a list of feelings following the election. The list of
emotions included both positive and negative emotions while we focused on testing only positive
self-conscious emotions and challenge emotions. We repeated the same 3-item measure of
positive self-conscious emotions from the pre-election questionnaire (α = .78). The scale yielded
a mean of 6.55 (range = 1 - 9, standard deviation = 1.93). We also repeated the same 2-item
measure of challenge emotions from the pre-election survey (α = .74). The scale yielded a mean
of 7.27 (range = 1 - 9, standard deviation = 1.70). The mean scores of both emotion types are
relatively high suggesting that respondents experienced a relatively high level of these emotions
following a positive brand event.
To measure behavioral responses toward the election, respondents were asked to indicate
the extent to which they act toward the election in a way described by a list of behaviors. We
specifically instructed them to indicate what they actually did rather than what “most people”
would do. All behavior items were executed on a 9-point scale ranging from 1 = “I didn’t do this
POWERFUL BRAND INFLUENTIALS 148

at all” to 9 = “I did this a lot”. EFA with oblique solution was conducted on the entire list of
behavior items and yielded four factors (eigenvalues > 1) that explained a total of 70.15% of the
variance. All behavior items, the corresponding factors, and internal consistency of factor
measures are listed in Appendix D.
The first factor included four items labeled as action-focused behavior (factor loadings
ranged from .74 to .87, explained variance = 37.51%). These items pertain to taking actions that
aim at improving the situation. They reflect a factor that is very similar to the essence of what
previous coping literature terms as problem-focused coping (Folkman and Lazarus 1985;
Lazarus 1991, 1999; Lazarus and Folkman 1984). In the current context, they are behavioral
responses toward a positive event instead of a stressful event. Specifically, the action-focused
behavior factor contained two items adopted from active coping measures (Duhachek 2005;
Duhachek and Iacobucci 2005) including “I tried to make a plan of action” and “I followed a
plan to make things better”. These items do not involve word-of-mouth engagement. The action-
focused behavior factor also contained two items adopted from measures of seeking social
support for instrumental reasons (Carver and Scheier 1989) including “I asked someone with
similar experiences what they did” and “I tried to get advice from someone about what to do”.
Though these two items involve word-of-mouth components, they are oriented toward seeking
for information from others as the word-of-mouth receiver instead of the word-of-mouth source.
This distinction is important as we will discuss in the subsequent section a second factor that
captures word-of-mouth engagement as the word-of-mouth source. Carver and Scheier (1989)
proposed that both active coping and seeking social support for instrumental reasons fall under
problem-focused coping. Our study results support their proposition by showing that the two
behavioral types loaded on a single factor. The final scale of action-focused behavior (α = .84)
POWERFUL BRAND INFLUENTIALS 149

yielded a relatively low mean of 2.96 (range = 1 - 9, standard deviation = 1.88). A low mean
indicates that respondents, on average, have a low engagement in action-focused behavioral
responses toward a positive brand event. This behavioral response type should be more relevant
to a stressful brand context than a positive brand context.
The second factor included four items labeled as emotion-expressive communication
(factor loadings ranged from .79 to .85, explained variance = 23.28%). These items involve
communicating and expressing emotional responses and experiences related to the event. They
reflect a factor that previous coping literature terms as emotion-focused coping (Folkman and
Lazarus 1985; Lazarus 1991, 1999; Lazarus and Folkman 1984), which pertains to behaviors that
focus on regulating one’s emotional responses. The majority of the items were adopted from
extant measures of emotional/social support and emotional venting measures (Carver and
Scheier 1989; Duhachek 2005; Duhachek and Iacobucci 2005). These items apply not only to a
stressful event but also to a positive situation. Specially, the emotion-expressive communication
factor included items of wanting to talk to others, talking and expressing emotions to others, and
celebrating the event. All the measures are oriented toward initiating (emotional) word-of-mouth
as the source of word-of-mouth to others as the receiver. They loaded on a distinct factor from
measures of seeking social support for instrumental reasons in that the latter involves word-of-
mouth from the receiver perspective. The emotion-expressive communication factor directly
captures word-of-mouth engagement, whereas the first factor of behavioral responses does not.
The final scale of emotion-expressive communication (α = .78) yielded a mean of 5.97(range = 1
- 9, standard deviation = 2.08).

POWERFUL BRAND INFLUENTIALS 150

The third factor included three items labeled as positive interpretation and growth (factor
loadings ranged from .82 to .85, explained variance = 9.35%). Positive interpretation and growth
coping is initially proposed by Carver and Scheier (1989). Others term this coping type as
positive thinking (Duhachek 2005; Duhachek and Iacobucci 2005) and positive reappraisal
(Lazarus and Folkman 1984). This factor captures behaviors that aim at construing a situation in
positive terms and personal growth opportunities. We adopted measures from these previous
studies, which apply to a stressful situation and a positive event alike. This factor does not
involve word-of-mouth engagement. The final scale of positive interpretation and growth (α
= .79) yielded a mean of 6.17 (range = 1 - 9, standard deviation = 2.10).
To measure types of word-of-mouth responses to the election, we constructed a battery of
items that taps various contents of word-of-mouth (i.e., what one would talk about) following the
election. With a broad list of items, we aim at covering action-focused brand advocacy voice,
emotion-focused voice and other potential word-of-mouth contents given the current context. A
total of twelve word-of-mouth items were administered and repeated for two different word-of-
mouth scenarios: a scenario of in-group word-of-mouth and a scenario of out-group word-of-
mouth. For the scenario of in-group word-of-mouth, respondents were asked to take time to
picture a scenario of talking to someone who supported the candidate they supported. They then
rated how well a list of statements characterizes what they would say to this person on a 9-point
scale (1 = strongly disagree to 9 = strongly agree). For the scenario of out-group word-of-mouth,
respondents were asked to take time to picture a scenario of talking to someone who supported
the candidate they opposed. They then rated how well the same list of statements characterizes
what they would say to this person on the same 9-point scale.
POWERFUL BRAND INFLUENTIALS 151

The list of measures first included an item measuring overall voice tendency: “I would
really want to talk about the election”. This item had a mean score of 6.14 (range = 1 - 9,
standard deviation = 2.55) for the in-group word-of-mouth scenario and a mean score of 3.85
(range = 1 - 9, standard deviation = 2.66) for the out-group word-of-mouth scenario. On average,
respondents showed a relatively low tendency to talk to out-group members. The variation
among their responses is fairly large for both the in-group and out-group word-of-mouth
scenarios.
Next, items measuring specific types of word-of-mouth were included. The items of
action-focused brand advocacy voice were adapted from measures of voice response and
complaint responses in the literature of satisfaction and dissatisfaction (Bougie, Pieters, and
Zeelenberg 2003; Ping 1993; Rusbult and Zembrodt 1983). We constructed additional items
related to recommending, sharing information, and promoting a brand. The items of emotion-
focused voice were drawn from extant measures of emotional support and emotional venting
strategies in the coping literature (Duhachek 2005; Duhachek and Iacobucci 2005; Lyubomirsky,
Sousa, and Dickerhoof 2006). To explore additional potential types of word-of-mouth, our
measurement list goes beyond the above items. We conducted EFA on all word-of-mouth items
for the in-group and out-group scenario separately. All word-of-mouth items, the corresponding
factors, and internal consistency of the scale (for both the in-group and out-group scenarios) are
listed in Appendix D.
For the scenario of in-group word-of-mouth, EFA with oblique solution was conducted
on all measures and yielded three factors (eigenvalues > 1). The three factors explained a total of
68.10% of the variance. The first factor included measures of action-focused brand advocacy
voice and measures of emotion-focused voice. However, measures of emotion-focused voice had
POWERFUL BRAND INFLUENTIALS 152

high cross loadings on the other two factors (cross loadings ranged from .41 to .68) suggesting
that these measures do not clearly loaded on the first factor. Conceptually, we expect that
measures of action-focused brand advocacy voice and emotion-focused voice should be
represented by two individual factors.
The second factor included two items (factor loadings ranged from .87 to .94, explained
variance = 12.37 %). Both items were originally adopted from complaint measures in the
dissatisfaction literature. Since the current context is positive and satisfactory in nature, these
items do not capture complaint behaviors in response to brand dissatisfaction. In the current
context, they reflect the concept of derogation of alternatives, a behavioral outcome of
attachment as previously noted (Johnson and Rusbult 1989; Park et al. 2006; Thompson,
Rindfleisch and Arsel 2006). Derogation of alternatives involves disparaging others who use
competing brands and rejecting what they stand for (Park et al. 2006). In the current context, this
concept relates to engagement of negative word-of-mouth to derogate members of the opposition
party (brand) and to reject what they stand for. Though derogation of alternative items loaded on
a factor of its own, we consider derogation of alternatives as a type of action-focused brand
advocacy voice. The derogation of alternative voice factor pertains to actively influencing others
regarding the attached brand and advocating the brand through disparaging the opposition brand.
It is a unique type of action-focused brand advocacy voice since it involves a destructive
intention toward the word-of-mouth receiver, whereas other measures of action-focused brand
advocacy voice involve a constructive intention toward the word-of-mouth receiver. The final
measure for derogation of alternative voice (α = .85) yielded a mean of 4.67 (range = 1 - 8.50,
standard deviation = 2.29).
POWERFUL BRAND INFLUENTIALS 153

The third factor included a single item: “I would focus more on my experiences of the
election than on the election per se” (factor loading = .80, explained variance = 9.60 %). This
factor reflects an inward orientation toward the self and lacks of the intention to influence others
regarding the brand or brand event. We label this factor as self-experience focused voice. The
item had a mean of 4.83 (range = 1 - 9, standard deviation = 2.11).
We attempted to address the high cross loading issues of the first factor. We conducted
EFA with oblique solution on items of action-focused brand advocacy voice and emotion-
focused voice again but fixed the number of factor to extract as 2 factors. The two-factor
structure explained a total of 65.07% of the variance. All items of action-focused brand advocacy
voice loaded on the first factor with eigenvalue greater than 1 (factor loadings ranged from .77
to .86, explained variance = 53.51%). All items of emotion-focused voice loaded on the second
factor with eigenvalue of 0.93 (factor loadings ranged from .76 to .84, explained variance =
11.56%). The final scale of action-focused brand advocacy voice (α = .83) yielded a mean of
5.43 (range = 1 - 9, standard deviation = 1.99). The final scale of emotion-focused voice (α = .79)
yielded a mean of 5.0 (range = 1 - 9, standard deviation = 2.01). Overall, the four factors
adequately represent all word-of-mouth items. A four-factor structure is also more theoretically
meaningful than the initial three-factor solution. Hence, we kept the four factors for proposition
testing.
For the scenario of out-group word-of-mouth, an initial EFA with oblique solution
yielded two factors (eigenvalues > 1). The two factors explained a total of 60.55% of the
variance. Different from the scenario of in-group word-of-mouth, items of action-focused brand
advocacy voice and those of derogation of alternative voice loaded together on the first factor
(factor loadings ranged from .77 to .86, explained variance = 44.17%). This result supports that
POWERFUL BRAND INFLUENTIALS 154

derogation of alternative voice represents a type of action-focused brand advocacy voice to a
certain extent. The second factor included items measuring emotion-focused voice and the single
item measuring self-experience focused voice (factor loadings ranged from .58 to .78, explained
variance = 16.38%). However, the second factor had fairly low factor loadings and cross
loadings were high (ranged from .13 to .62). This result suggests that measures were not loaded
clearly on the current two-factor structure. It is desirable to apply the same four-factor solution of
the in-group scenario to the out-group scenario. In this way, we could directly compare responses
between the two scenarios.
Therefore, we conducted a second EFA with oblique solution and fixed the number of
factor to extract as 4 factors. The four-factor structure explained a total of 74.77% of the variance.
The first factor included items of derogation of alternative voice with eigenvalue greater than 1
(factor loadings ranged from .84 to .88, explained variance = 44.17%). The second factor
included items of emotion-focused voice with eigenvalue greater than 1 (factor loadings ranged
from .74 to .87, explained variance = 16.38%). The third factor included measures of action-
focused brand advocacy voice with eigenvalue of .89 (factor loadings ranged from .72 to .80,
explained variance = 8.06%). The single item of self-experience focused voice loaded alone on
the fourth factor with eigenvalue of .68 (factor loadings was .91, explained variance = 6.16%).
We checked internal consistency and descriptive statistics of the resulted four factors: (1)
Derogation of alternative voice (α = .76) had a mean of 3.92 (range = 1 - 8.50, standard deviation
= 2.21). (2) Action-focused brand advocacy voice (α = .77) had a mean of 4.85 (range = 1 - 9,
standard deviation = 1.97). (3) Emotion-focused voice (α = .75) had a mean of 4.32 (range = 1 -
8.5, standard deviation = 1.89). (4) The item of self-experience focused voice had a mean of 5.45
(range = 1 - 9, standard deviation = 2.42). The mean scores for the out-group scenario were
POWERFUL BRAND INFLUENTIALS 155

relatively lower than those for the in-group scenario with the exception of self-experience
focused voice. It is not surprising that people of the winning party are less likely to derogate the
opposition party, share their emotions, and take joint actions with the losing party than with their
own party.
10.2. Measurement Results of the PBIS
Scale norms and reliability. Pre-election data (N=180) was used to evaluate the stability
and reliability of the PBIS. The distribution of the PBIS is similar to that found in Study 2 and
Study 3. The mean score of the PBIS for Study 4 pre-election sample (N = 182) was 3.85 (range
= 1 - 9, standard deviation = 2.32). Mean scores of individual PBIS items ranged from 3.02 to
5.02 with standard deviations ranging from 2.44 to 3.05. The internal consistency reliability of
the PBIS was .93 with item-to-total correlations ranging from .62 to .89.
Scale dimensionality. We conducted similar steps to confirm the dimensionality and
stability of the PBIS using brands in a different product category. EFA results converge with
Study 2 and Study 3 by showing a single factor, which confirmed that the PBIS is a
unidimensional scale. The scale explained 75.12% of the total variance with factor loadings
ranging from .71 to .93 (see Table 7-4, results for Study 4 Pre-election). Furthermore, CFA
results on the PBIS (see Table 8-1, results for Study 4 Pre-election) showed that all standardized
factor loadings were significant (p < .001). The average standardized factor loading was .83. The
measurement model of the PBIS has a good fit to the data. The fit statistics of NFI, CFI, and GFI
all exceeded the .90 cutoff value (Bentler 1992) and the SRMR was less than .05 as
recommended (Byrne 1998). The construct reliability was greater than the criterion of .70 with
the average variance extracted exceeding the criterion of .50 (Fornell and Larcker 1981). Using a
POWERFUL BRAND INFLUENTIALS 156

different product category, the results again demonstrate that the PBIS is a robust and good
measure.
10.3. Testing Known Group Validity
Using Study 4 pre-election data, we tested the known group validity of the PBIS. We
examined whether the PBIS could differentiate real-life powerful brand influentials (Students for
Obama and Students for McCain) from the larger heterogeneous population. We combined
Students for Obama and Students for McCain samples to form a single group of powerful brand
influentials (N = 38) and compared them against the rest of student population (N = 144) by the
PBIS mean score. The t-test result showed that the unique group of powerful brand influentials
(M = 6.22) scored significantly higher on the PBIS than the comparison group (M = 3.32, t (136)
= 9.68, p < .001). The known group validity of the PBIS is supported.
10.4. Testing Convergent Validity and Discriminant Validity of the PBIS in Relation to
Brand Attachment (P1d and P2d)
Using brands from a different category, we intend to replicate the results that the
powerful brand influential is significantly related to (P1d) but distinctly different from brand
attachment (P2d). Similar to Study 3, brand attachment was constructed as a second-order factor
structure that includes Brand-Self Connection (BSC) and Prominence of Thoughts and Feelings
(PRO) as two indicators. We constructed three CFA models to confirm the convergent and
discriminant validity of the PBIS in relation to the brand attachment.
Model 1 was our proposed model: the PBIS and brand attachment were modeled as two
correlated yet distinct constructs. Table 10-1 reports Model 1 results. The PBIS items and brand
attachment items loaded on the powerful brand influential factor and brand attachment factor,
respectively. The model fits the data well. All factor loadings were significant (p < .001). The
POWERFUL BRAND INFLUENTIALS 157

construct reliability (CR) of both factors exceeded the criterion of .70 and the average variance
(AVE) extracted for both factors exceeded the criterion of .50. Model 1 also had a good fit [χ
2

(32) = 151.93, NFI = .95, CFI = .96, SRMR =. 049]. The interfactor correlation coefficient
between the PBIS and brand attachment factor was significant (Φ = .36, t = 5.01, p < .001)
confirming that the measured two constructs are significantly related. Hence, convergent validity
of the PBIS is supported. The Φ coefficient squared was less than the AVE extracted for both
factors, which supports discriminant validity of the PBIS from brand attachment.

TABLE 10-1. Pre-Election: Confirmatory Factor Analysis Results of Model 1 – the PBIS
and Brand Attachment as Two Correlated Factors

Factor Factor Indicators
Standardized Factor
Loadings
AVE CR
The Powerful Brand Influential
PBIS 1. .85 .95 .99
PBIS 2. .63
PBIS 3. .88
PBIS 4. .81
PBIS 5. .91
PBIS 6. .91
Brand Attachment
BA 1. (BSC 1) .86 .91 .98
BA 2. (BSC 2) .88
BA 3. (PRO 1) .91
BA 4. (PRO 2) .93
Note:
Factor loadings are standardized factor loadings. All factor loadings are significant at p < .001.
Abbreviations: CR = Construct Reliability. AVE = Average Variance Extracted. PBIS (the Powerful Brand
Influential Scale). BA (Brand Attachment). PRO (Prominence of Thoughts and Feelings indicator of brand
attachment). BSC (Brand-Self Connection indicator of brand attachment).

Furthermore, we compared Model 1 to two alternative models. In Model 2, the two
factors (i.e., the PBIS and brand attachment factors) were forced to be uncorrelated [χ
2
(34) =
216.65, NFI = .90, CFI = .91, SRMR = .33]. This model assumes that there is no significant
relationship between the two constructs. The change in chi-square [∆χ² (2) = 64.72, p < .001]
suggests that Model 1 was significantly better than Model 2. In Model 3, the two factors were
POWERFUL BRAND INFLUENTIALS 158

forced to be perfectly correlated [χ
2
(33) = 174.67, NFI = .94, CFI = .95, SRMR = .32]. This
model assumes that the two constructs are not distinct from each other. The chi-square difference
test suggests that Model 1 is also significantly better than Model 3 [∆χ² (1) = 22.74, p < .001].
Overall, with brands in a different category, Study 4 replicates Study 2 and Study 3 in supporting
that the powerful brand influential construct is significantly related to (P1d) but distinctly
different from (P2d) the brand attachment construct.
10.5. Testing Predictive Validity of the PBIS vs. Brand Attachment for Positive Self-
Conscious Emotions and Challenge Emotions (P10a – 10b)
We proposed that the powerful brand influential construct better predicts positive self-
conscious emotions related to a positive brand event than does brand attachment (P10a) but both
constructs contribute independently significant predictions of challenge emotions (P10b). We
tested the proposition with both pre-election data and post-election data. People experience
anticipatory emotions upon an emotion-eliciting event and experience outcome emotions after
the event actually happens (Folkman and Lazarus 1985). We expect that people experience
positive self-conscious emotions and challenge emotions that are anticipatory before the election.
They should also experience these emotions after the election driven by the positive outcome.
Using pre-election data, we conducted a structural model with the PBIS and brand
attachment as the predictors of positive self-conscious and challenge emotions (see Figure 10-1,
Panel A). The pre-election model results showed that the PBIS significantly predicts positive
self-conscious emotions (γ = .45, p < 0.01). When simultaneously estimated with the PBIS,
brand attachment is an insignificant predictor of positive self-conscious emotions (γ = .01, p >
0.05). This result supports P10a by showing that the powerful brand influential better predicts
positive self-conscious emotions than does brand attachment. The model also showed that the
POWERFUL BRAND INFLUENTIALS 159

PBIS (γ = .24, p < 0.05) and brand attachment (γ = .19, p < 0.05) contribute independently
significant prediction of challenge emotions. The difference between the two coefficients is not
significant (z = .47, p > .05), which indicates that the two constructs predict challenge emotions
experienced before election at similarly influential levels. This result confirms P10b.

FIGURE 10-1. Comparison of Predictive Ability of the PBIS and Brand Attachment for
Positive Self-Conscious Emotions and Challenge Emotions

Panel A: Predictive Ability for Pre-Election Emotional Responses

χ² = 24.75, NFI = 0.93, CFI = 0.94, SRMR = 0.07

Note:
The three values reported above the associated path are in the sequence of γ estimation, standard error,
and t value. Significant paths are shown by solid lines. Insignificant paths are shown by dotted lines.
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.
Abbreviations: NFI = Normed Fit Index. CFI = Comparative Fit Index. SRMR = Standardized Root Mean
Square Residual. BSC (Brand-Self Connection). PRO (Prominence of Thoughts and Feelings). PBIS (the
Powerful Brand Influential Scale). CPAM (Connection-Prominence Attachment Model).

The Powerful
Brand
Influential
(PBIS)
0.45
(0.07)
6.09**

Challenge
Emotions
Self-
Conscious
Emotions
BSC
PRO
Brand
Attachment
(CPAM)
0.01
(0.04)
0.19
0.24
(0.08)
3.11*
0.19
(0.07)
2.83*
POWERFUL BRAND INFLUENTIALS 160

Next, the same structural model was applied to post-election data (see Figure 10-1, Panel
B). The post-election model replicates the results that the PBIS significantly predicts positive
self-conscious emotions (γ = .40, p < 0.05) while brand attachment remains an insignificant
predictor (γ = .07, p > 0.05). The post-election data provides additional support for P10a. On the
other hand, tests of P10b yielded mixed results. We expected that, when simultaneously
estimated, both constructs should independently predict challenge emotions. As expected, the
PBIS significant predicts challenge emotions (γ = .32, p < 0.05). However, when simultaneously
estimated, brand attachment becomes an insignificant predictor of challenge emotions (γ = .03, p >
0.05) after the election. This result needs further validation. The post-election data had relatively
small sample size. Our structural modeling on the data is limited in robustness of the estimation.
Overall, P10b is fully supported by pre-election data but partially supported by the post-election
data.

POWERFUL BRAND INFLUENTIALS 161

FIGURE 10-1. Comparison of Predictive Ability of the PBIS and Brand Attachment for
Positive Self-Conscious Emotions and Challenge Emotions

Panel B: Predictive Ability for Post-Election Emotional Responses

χ² = 18.81, NFI = 0.98, CFI = 0.99, SRMR = 0.05

Note:
The three values reported above the associated path are in the sequence of γ estimation, standard error,
and t value. Significant paths are shown by solid lines. Insignificant paths are shown by dotted lines.
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.
Abbreviations: NFI = Normed Fit Index. CFI = Comparative Fit Index. SRMR = Standardized Root Mean
Square Residual. BSC (Brand-Self Connection). PRO (Prominence of Thoughts and Feelings). PBIS (the
Powerful Brand Influential Scale). CPAM (Connection-Prominence Attachment Model).

10.6. Testing Predictive Validity of the PBIS vs. Brand Attachment for the Level of Actual
Word-of-Mouth Engagement and Types of Word-of-Mouth Following a Positive Brand
Event (P11a – 11c)
A set of propositions compare the powerful brand influential and brand attachment
constructs regarding behavioral responses toward a positive brand event (P11a – 11c). This set of
propositions was tested using post-election data. Given the sample size constraint of using
structural modeling, we adopted regression analysis to test these propositions. Specifically, we

The Powerful
Brand
Influential
(PBIS)
0.40
(0.14)
2.84*

Challenge
Emotions

Self-
Conscious
Emotions
BSC
PRO

Brand
Attachment
(CPAM)
0.07
(0.14)
0.49
0.32
(0.14)
2.32*
0.03
(0.14)
0.24
POWERFUL BRAND INFLUENTIALS 162

proposed that the powerful brand influential construct better predicts the level of actual word-of-
mouth engagement following a positive brand event than does brand attachment (P11a). To test
this proposition, the PBIS and brand attachment were regressed on each of the three behavioral
response factors including emotion-expressive communication, action-focused behavior, and
positive interpretation and growth. Table 10-2 reports the results of all three regression analyses.

TABLE 10-2. Post-Election: Regression Results Comparing Predictive Ability of the PBIS
and Brand Attachment for Behavioral Responses to a Positive Brand Event

Predictor
Dependent Variable
β
(Standardized
Coefficients)
R
2
Emotion-Expressive Communication .24
PBIS .42**
Brand Attachment .10
Action-Focused Behavior .01
PBIS -.03
Brand Attachment -.06
Positive Interpretation & Growth .15
PBIS .27
Brand Attachment .14
Note:
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.

As discussed earlier, emotion-expressive communication is the only behavioral response
factor that captures word-of-mouth engagement from the perspective of initiating word-of-
mouth. Therefore, we expect that the PBIS should predict emotion-expressive communication
better than does brand attachment. As shown in Table 10-2, the PBIS significantly predicts
emotion-expressive communication (β = .42, p < .01), whereas brand attachment is an
insignificant predictor (β = .10, p > .05). This result supports P11a. Furthermore, both the PBIS
and brand attachment are insignificant predictors of the other two behavioral response factors. As
mentioned before, respondents reported an overall low engagement of action-focused behaviors
POWERFUL BRAND INFLUENTIALS 163

following the brand success. This behavioral type might have low relevance to a positive brand
event. It is therefore not surprising that powerful brand influentials and strongly brand-attached
consumers are no different from the rest in the engagement level of this behavioral type. Positive
interpretation and growth has relatively high relevance to a positive brand event. Our result
suggests that powerful brand influentials and strongly brand-attached consumers do not differ
from others in the engagement level of this behavioral response type toward a positive brand
event.
P11b - P11c posit that the powerful brand influential construct predicts action-focused
and emotional-focused types of word-of-mouth following a positive brand event better than does
brand attachment, respectively. We explored the types of word-of-mouth for two different
scenarios: an in-group and an out-group word-of-mouth scenario. The PBIS and brand
attachment were regressed on each word-of-mouth variable for the scenario of in-group and out-
group word-of-mouth separately. The results were presented in Table 10-3. We first checked the
measure of overall voice tendency. For the in-group word-of-mouth scenario, the PBIS
significantly predicts overall voice tendency (β = .49, p < .001) while brand attachment does not
(β = -.002, p > .05). For the out-group word-of-mouth scenario, neither the PBIS (β = .28, p > .05)
nor brand attachment (β = .02, p > .05) significantly predicts overall voice tendency. These
results suggest that powerful brand influentials are differentiated from strongly brand-attached
consumers by a higher overall tendency to initiate word-of-mouth to in-group but not out-group
members. Overall, respondents showed a relatively low tendency to talk to out-group members
in the current context of a brand success. It is important to note that the measure of overall voice
tendency does not capture specific types of word-of-mouth. In the next section, we consider
POWERFUL BRAND INFLUENTIALS 164

specific voice types and examine whether powerful brand influentials tend to initiate certain
types of word-of-mouth to out-group members.

TABLE 10-3. Post-Election: Regression Results Comparing Predictive Ability of the PBIS
and Brand Attachment for Types of Word-of-Mouth after a Positive Brand Event

Predictor
Dependent Variable
In Group Out Group
β
(Standardized
Coefficients)
R
2

β
(Standardized
Coefficients)
R
2
Overall Talking
Tendency
.24 .09
PBIS .49*** .28
Brand Attachment -.002 .02

Action-Focused Brand
Advocacy Voice
.25 .15
PBIS .49*** .38**
Brand Attachment .02 .002

Derogation of
Alternative Voice
.16 .10
PBIS .42** .32*
Brand Attachment -.04 -.01
Emotion-Focused Voice .12 .02
PBIS .35* .11
Brand Attachment .001 .03

Self-Experience Focused
Voice
.09 .05
PBIS -.22 .08
Brand Attachment .41** .16
Note:
* Significant at p < .05 for two-tailed tests. ** Significant at p < .01 for two-tailed tests.
*** Significant at p < .001 for two-tailed tests.

Our data identified four different types of word-of-mouth including action-focused brand
advocacy voice, derogation of alternative voice, emotion-focused voice, and self-experience
focused voice. For the in-group word-of-mouth scenario, the PBIS significantly predicts action-
focused brand advocacy voice (β = .49, p < .001), derogation of alternative voice (β = .42, p
< .01), and emotion-focused voice (β = .35, p < .05). Brand attachment does not significantly
influence action-focused brand advocacy voice (β = .02, p > .05), derogation of alternative voice
POWERFUL BRAND INFLUENTIALS 165

(β = -.04, p > .05), or emotion-focused voice (β = .001, p > .05). Noted earlier, derogation of
alternative voice is a unique type of action-focused brand advocacy voice. Therefore, these
results confirm that the PBIS predicts action-focused brand advocacy voice (P11b) and emotion-
focused voice (P11c) better than does brand attachment. Though not hypothesized, brand
attachment significantly predicts self-experience focused voice (β = .41, p < .01) while the PBIS
is not a significant predictor (β = -.22, p > .05). We found that strongly brand-attached
consumers who are not powerful brand influentials talk more about their personal experiences
than the brand or the positive brand event when they talk to in-group members.
For the out-group word-of-mouth scenario, the PBIS remains significant for both action-
focused brand advocacy voice (β = .38, p < .01) and derogation of alternative voice (β = .32, p
< .05). Brand attachment remains insignificant for action-focused brand advocacy voice (β
= .002, p > .05) and derogation of alternative voice (β = -.01, p > .05). These results replicate our
support for P11b found in the scenario of in-group word-of-mouth. We further compared the
coefficients between the out-group and in-group word-of-mouth scenarios. The comparison
showed that the PBIS was as strong a predictor of action-focused brand advocacy voice (z = .71,
p > .05) and derogation of alternative voice (z = .54, p > .05) for the out-group scenario as for the
in-group scenario. These results suggest that powerful brand influentials are highly engaged in
the type of action-focused brand advocacy voice regardless of whether they talk to in-group or
out-group members.
For the out-group scenario, however, the PBIS became an insignificant predictor of
emotion-focused voice (β = .11, p > .05). This result indicates that powerful brand influentials
have a higher tendency to express their emotions than other consumers only when they talk to in-
group members. Brand attachment remains an insignificant predictor of emotion-focused voice
POWERFUL BRAND INFLUENTIALS 166

(β = .03, p > .05). Hence, P11c only holds true for the in-group word-of-mouth scenario.
Regarding self-experience focused voice, the PBIS remains insignificant (β = .08, p > .05) while
brand attachment also became insignificant (β = .16, p > .05) for the out-group scenario. Our
results reveal that strongly brand-attached consumers who are not powerful brand influentials do
not differ from other consumers in any type of word-of-mouth when they talk to out-group
members.
To summarize these results for behavioral responses, Study 4 provides insights in the
nature of word-of-mouth engaged by powerful brand influentials following a positive brand
event. They are not simply those who tend to talk about their attached brand. Following a
positive brand event, they actively and constructively advocate the brand and derogate brand
alternatives regardless of whom they talk to. As those who are highly emotionally expressive,
they also share their emotional experiences with supporters of the brand. When talking to
members of the opposition brand, their word-of-mouth is more centered on advocating the brand
and discussing brand-related issues than sharing personal feelings about the brand event. Our
results also provide initial understanding on how other strongly brand-attached consumers would
talk following a positive brand event. Overall, they have a lower tendency to talk about the brand
or brand related issues than powerful brand influentials. When talking to others, they would
focus more on their personal experiences and lack of the focus on brand-related substances.
Comparing to powerful brand influentials, they have a significantly lower tendency to advocate
the brand and derogate brand alternatives regardless of whom they talk to. They also have a
lower tendency than powerful brand influentials to initiate emotional sharing when talking to
supporters of the brand. Using the context of positive brand event, we demonstrate that powerful
POWERFUL BRAND INFLUENTIALS 167

brand influentials are unique from other strongly brand-attached consumers in that their word-of-
mouth features a brand advocacy and emotionally expressive nature.

POWERFUL BRAND INFLUENTIALS 168

Chapter 11: General Discussion
11.1. Contribution and Implications

Despite the prevailing evidence for the existence of a brand’s influential consumers and
their prominent value to marketers, our understanding about these consumers is limited.
Abundant research has touched the phenomenon with perspectives from different domains
including brand community, fandom, word-of-mouth, relationship marketing, consumer loyalty,
brand commitment, and consumer devotion. However, a conceptualization of a brand’s
influential consumers and a scale to identify these consumers is largely missing in the literature.
This article fills the gap by focusing on a brand’s influential consumers termed “the powerful
brand influential” as a unique construct. We systematically conceptualize the defining properties
of the construct and theoretically differentiate it from key related constructs with a set of
propositions. Moreover, this paper contributes to the literature by developing a scale to measure
the new construct, empirically proving the proposed differentiations, and simultaneously
establishing the scale validity.
By synthesizing various streams of research related to the phenomenon, we conceptualize
four key properties of the powerful brand influential construct. We differentiate the construct
from six existing constructs including two influential consumer constructs (i.e., opinion leader
and market maven) and four brand-level constructs (i.e., brand attachment, brand ambassador,
brand satisfaction, and brand strength). We demonstrate the uniqueness of the powerful brand
influential construct by supporting a total of fifteen propositions on its related but distinctive
nature from those existing concepts. More importantly, we substantiate the added value of the
powerful brand influential construct – demonstrating that this construct predicts many outcomes
of key interests to a brand beyond predictive power of those existing constructs. These key brand
POWERFUL BRAND INFLUENTIALS 169

outcomes include brand word-of-mouth, intrinsic pleasure derived from brand word-of-mouth,
brand advocacy behavior, tendency to influence others’ brand decision, brand loyal behavior,
actual brand purchase along with desirable emotional and word-of-mouth responses toward a
positive brand event. An added predictive power offered by this new construct highlights the
managerial significance and importance of being able to identify these desirable consumers of a
brand. The newly developed 6-item scale, the PBIS, enables marketers to reach the powerful
brand influential with a viable scale. Across four studies using different samples and brands from
diverse product categories, we have established the reliability, robustness, known group,
convergent, discriminant, and predictive validity of the PBIS. The scale is short, easily-
administered, and adaptable to various contexts. Hence, this article also makes significant
practical contribution by providing brand managers with the tool to identify their brand’s
influential consumers and manage these consumers more efficiently and effectively.
11.2. Summary of Findings
In this article, we develop and test a total of fifteen propositions using four studies. We
summarize these propositions and findings in the current section. We first propose that the
powerful brand influential construct is significantly related to but distinctly different from all six
marketing constructs (P1, P2, and P12). We then articulate in great details how our construct is
differentiated from opinion leader, market maven, brand attachment, and brand ambassador by a
higher level of three outcomes theoretically linked with the defining properties of our construct.
A core set of propositions (P3 – P6) corresponds to the four theoretical outcomes of the powerful
brand influential construct. Study 2 provides initial support that the powerful brand influential
construct is significantly related to but distinctly different from market maven, opinion leader,
brand attachment, and brand ambassador. Additionally, Study 2 indicates the reliability,
POWERFUL BRAND INFLUENTIALS 170

convergent, and discriminant validity of the PBIS by showing that it has good internal
consistency and measures a distinct latent construct from the related four constructs. Moreover,
we hypothesize and support with Study 2 that the powerful brand influential construct predicts (1)
the engagement level of brand word-of-mouth and (2) brand advocacy behavior, (3) the level of
influence tendency on other consumers’ brand decisions, and (4) the level of intrinsic pleasure
derived from brand word-of-mouth better than the four related constructs. These findings prove
that none of those related constructs capture the defining properties of the powerful brand
influential construct, and simultaneously, justify the proposed construct as a new construct.
Using a much larger sample size and a drastically different brand, Study 3 replicates the
support that the powerful brand influential construct is distinct from the related constructs
investigated in Study 2 (brand ambassador was deleted from Study 3). Study 3 repeats the testing
for a core conceptual outcome of the powerful brand influential construct, which is brand
advocacy behavior. Predictive validity of the new construct and the PBIS is provided by showing
that the construct and its measure better predict brand advocacy behavior than does all three
related constructs and their measures.
Study 2 and Study 3, as taken together, conclude the uniqueness of the powerful brand
influential construct from two existing types of influential consumers (i.e., market maven and
opinion leader) as well as two related brand-level constructs (i.e., brand attachment and brand
ambassador). Additionally, brand attachment is conceptualized and empirically shown in Study 2
and Study 3 as the most overlapping construct to the powerful brand influential construct.
Correspondingly, we detail further differentiation between the two constructs by comparing their
predictive ability for a wide range of key outcomes. The outcomes compared consist of
psychological outcomes including separation distress and brand commitment (P7) and brand
POWERFUL BRAND INFLUENTIALS 171

behavioral outcomes including intention to conduct brand loyal behavior and actual purchase (P8
– P9). Study 2 and Study 3 converge in proving that the powerful brand influential construct is a
significantly weaker predictor of the conceptual outcomes of brand attachment (i.e., separation
distress and brand commitment). Moreover, both studies reveal that the two constructs bring
independently significantly prediction of intention to conduct brand loyal behavior. These results
provide crucial substantiation that the powerful brand influential construct does not represent a
high level of the brand attachment construct in that it does not replace the significant prediction
of brand attachment for its key conceptual outcomes. It is important for marketers to understand
the independent significance of the powerful brand influential construct and brand attachment for
key brand outcomes so that they could target customer-brand relationship management according
to the respective value of each construct.
It is also crucial to recognize fundamental differences between powerful brand influential
consumers and other strongly brand-attached consumers. Though powerful brand influential
consumers have strong brand attachment, they are not simply strongly brand-attached
consumers. Study 3 supports that they tend to be emotional expressive while other strongly
brand-attached consumers do not have such dispositional tendency. As a result of emotional
expressiveness, powerful brand influentials find the act of brand word-of-mouth intrinsically
pleasant. By contrast, strongly brand-attached consumers who are not powerful brand influentials
derive low intrinsic pleasure from the act of brand word-of-mouth. Study 1, 2 and 3 all point to a
key differentiating feature of powerful brand influentials from other attached consumers – being
the highly engaged, brand-attached consumers. Engagement is the key word. Across three
studies, we prove that powerful brand influentials compared with other strongly brand-attached
consumers (1) are more engaged in influencing others’ brand decisions; and they do so via (2) a
POWERFUL BRAND INFLUENTIALS 172

higher engagement level of brand word-of-mouth and (3) advocacy behavior; and (4) are not
only as highly motivated to conduct brand loyal behavior but also more engaged in actual brand
purchase.
Study 4 further differentiates powerful brand influentials from other strongly brand-
attached consumers regarding their responses toward a specific brand event. We propose that the
two constructs should be distinguishable in predictive ability for emotional and behavioral
responses toward a positive brand event. We investigate desirable emotional responses include
positive self-conscious emotions and challenge emotions (P10). We examine behavioral
responses focusing on the engagement level of word-of-mouth about the brand and types of
word-of-mouth following a positive brand event (P11). Using the 2008 Presidential Election as
the study context, Study 4 supports that the powerful brand influential construct better predicts
anticipatory positive self-conscious emotions prior to the positive event as well as actual positive
self-conscious emotions after the positive event than does brand attachment. As expected, Study
4 also shows that both constructs contribute independently significant prediction of challenge
emotion before the positive event. Unexpectedly, our post-election result suggests that the
powerful brand influential construct predicts actual challenge emotions after a positive event
better than does brand attachment. We call for further investigation of such empirical finding due
to the sample size constraint of the post-election data for robust modeling.
Study 4 also advances findings in Study 1, 2, and 3 regarding a highly brand-engaged and
emotionally-expressive nature of powerful brand influentials – showing that they actually
respond to a positive brand event in a more engaged and emotionally expressive manner than do
other strongly attached consumers. Specifically, Study 4 reveals that they are significantly
differentiated from other consumers (including other strongly attached consumers) by how they
POWERFUL BRAND INFLUENTIALS 173

respond to a positive brand event and how they would talk about it. Overall, their actual
engagement level of word-of-mouth following a brand success is higher than other consumers.
Regardless of whom they talk to (in-group or out-group members), they engage in higher level of
brand advocacy and derogation of alternative types of voice than other consumers. Comparing to
other consumers, they are also more engaged in emotion sharing when talking to in-group
members. By contrast, other strongly brand-attached consumers are primarily undifferentiated
from low-attachment consumers in how they actually respond to a positive event and how they
would talk about it. Comparing to powerful brand influentials, they have an overall lower
tendency to talk about the brand and express brand related emotions following a positive brand
event. When talking to in-group members, they would focus more on their personal experiences
than on the brand or brand event. When talking to out-group members, they do not show any
differences from low-attachment consumers in how they would talk about it. More importantly,
regardless of whom they talk to, their word-of-mouth lacks of substances pertaining to the
attached brand and does not involve a brand advocacy or emotionally expressive nature.
The last part of our conceptual discussion compares the powerful brand influential
construct from two additional brand constructs, brand attitude strength and brand satisfaction.
We concentrate on comparing predictive ability of our construct from brand attitude strength and
brand satisfaction for three key brand outcomes: brand advocacy (P13), intention to conduct
brand loyal behavior (P14), and actual purchase (P15). Study 3 confirms that the powerful brand
influential construct (1) is significantly related to but distinctly different from both existing
constructs (P12); (2) better predicts brand advocacy and actual purchase behavior than both
existing constructs; (3) it predicts brand loyalty intention at a similarly significant level as both
constructs.
POWERFUL BRAND INFLUENTIALS 174

11.3. Limitation and Further Research
We consider this article as a first step to address several unexplored issues – to
systematically conceptualize the powerful brand influential construct, to develop a scale, and to
establish the distinctiveness and validity of the new construct and its scale. Although we address
some critical issues, several areas are worthy of further investigation. First, at the very core of
the powerful brand influential is their high tendency to influence others’ brand decisions through
a high engagement of passionate and emotionally evocative word-of-mouth and advocacy
behavior. This article shows that the powerful brand influential construct better predicts the
tendency to influence others’ brand decisions through a high level of self-reported brand word-
of-mouth and advocacy behavior than six related constructs. Although previous literature on
influential consumers has used self-reports as the predominant method to measure criterion
outcomes (Feick and Price 1987; Flynn et al. 1996), objective measures of these key conceptual
outcomes will advance the managerial values of the powerful brand influential. Future research
might use objective measures to capture actual influence of the powerful brand influential and
test whether they actually exert disproportionate influence on others’ real-life decisions about a
brand. Similarly, future research might measure brand word-of-mouth and advocacy behavior
using objective measures and confirm that the PBIS indeed captures those who are more engaged
in these activities. Moreover, additional research is needed to analyze the content of word-of-
mouth and support that the powerful brand influential not only are more engaged in brand word-
of-mouth but also conduct more passionate and emotionally evocative natured word-of-mouth
than other consumers.

POWERFUL BRAND INFLUENTIALS 175

Second, future studies should investigate the generalizability of our findings. In
particular, we find that the powerful brand influential construct is as strong a predictor as brand
attachment, brand attitude strength and brand satisfaction of intention to conduct brand loyal
behavior, whereas it is a stronger predictor than the three constructs of actual purchase behavior.
These findings should be further evaluated for different contexts considering the types of brands
(e.g., functional brands, symbolic brands, hedonic brands), the nature of product categories (e.g.,
high involvement products, low involvement products), and the characteristics of samples (e.g.,
socioeconomic status, demographics, psychographics). Additional research should systematically
explore variables that might moderate the relative predictive ability of the powerful brand
influential and existing marketing constructs for key brand outcomes. Marketers need to
understand the boundary and applicable contexts of our findings to target toward their most
desirable group(s) of consumers. A valuable research approach is to construct mathematic
models that directly link the PBIS and other parameters such as existing consumer-brand
relationship measures (e.g., brand attachment, brand attitude strength, brand satisfaction),
consumer previous purchase data, consumer characteristics (e.g., demographics, psychographics,
socioeconomic status, social connectedness), marketing variables (e.g., price, promotion, store
location), brand variables (e.g., brand equity, the type of brand, country of origin) into brand
sales. Such models should be able to help marketers assess the prediction of the PBIS compared
with that of other variables. More importantly, such models should enable prediction of future
sales.

POWERFUL BRAND INFLUENTIALS 176

Third, although we demonstrate the significance of the powerful brand influential, we
have not explored how to acquire, foster, manage and sustain these valuable consumers of a
brand. In order to address these questions, future research should investigate the antecedents,
responses, and dynamics of the powerful brand influential construct. This article suggests a
strong brand attachment and a high emotional expressiveness as potential antecedents of the
powerful brand influential. Researchers should explore what additional antecedents lead to the
formation of the powerful brand influential, which will guide ways of fostering these consumers.
Regarding responses of the powerful brand influential, this article taps positive self-
conscious emotions, challenge emotions, engagement level of word-of-mouth, and types of
word-of-mouth given a positive brand event. Researchers should validate our findings related to
these examined responses with larger sample sizes than our current data to allow for more robust
estimation. Future studies might expand the research contexts from positive to negative brand
events and examine the responses of powerful brand influentials compared with that of other
types of consumers (e.g., other strongly brand-attached consumers, satisfied consumers,
consumers who hold strong attitude toward the brand). We believe that the active, constructive,
and emotionally expressive nature of powerful brand influential should prevail not only in
positive but also negative brand contexts. They should conduct constructively action-focused and
expressively emotion-focused behaviors when they face a negative brand circumstances such as
brand transgression. Nevertheless, additional works are needed for testing these speculations.
More importantly, researchers should systematically explore various types of responses (e.g.,
cognitive responses, additional types of emotional and behavioral responses) of the powerful
brand influential and understand the mechanism by which they experience these responses in
different forms of positive and negative brand situations (e.g., situations that attributable to
POWERFUL BRAND INFLUENTIALS 177

different causes, involve differential magnitude of impact, involve social meanings of different
natures).
Finally, the temporal relationship of the powerful brand influential to other types of
consumers and the dynamics of these consumers are also important areas of further exploration.
The question as whether or not the formation of a strong brand attachment necessarily precedes
the formation of the powerful brand influential has not been answered. It is uncertain whether
(1) powerful brand influentials have longer attachment-relationship with the brand than other
strongly attached consumers or (2) they do not differ in the length of attachment-relationship
from other attached consumers but become powerful brand influentials due to certain factors
(e.g., psychographic, demographics, social connectedness). Additionally questions as whether,
when, and how the powerful brand influential die away are crucial for managing this advanced
form of customer-brand relationship. Future research might answer these questions by
monitoring the powerful brand influential over time using the PBIS and probing factors that lead
to potential termination.

POWERFUL BRAND INFLUENTIALS 178

References
Anderson, Eugene W. (1998), “Customer Satisfaction and Word of Mouth,” Journal of Service
Research, 1 (1), 5-17.
Anderson, Eugene W., Claes Fornell, and Donald R. Lehmann (1994), “Customer Satisfaction,
Market Share, and Profitability,” Journal of Marketing, 58 (July), 53-66.
Anderson, James C. and David W. Gerbing (1988), “Structural Equation Modeling in Practice: A
Review and Recommended Two-Step Approach,” Psychological Bulletin, 103(3), 411-
423.
Anderson, James C. and James A. Narus (1991), “Partnering as a Focused Market Strategy,”
California Management Review, 33 (3), 95-113.
Arndt, Johan (1967), Word of Mouth Advertising: A Review of the Literature, New York:
Advertising Research Foundation Inc.
Aron, Arthur, Elaine N. Aron, and Danny Smollan (1992), “Inclusion of Other in the Self-Scale
and the Structure of Interpersonal Closeness,” Journal of Personality and Social
Psychology, 63 (4), 596-612.
Aron, Arthur, Debra Mashek, Tracy McLaughlin-Volpe, Stephen Wright, Gary Lewandowski,
and Elaine N. Aron (2005), “Including Close Others in the Cognitive Structure of the
Self,” in Interpersonal Cognition, ed. Mark W. Baldwin, New York: Guilford Press, 206-
232.
Assael, Henry (1981), Consumer Behavior and Marketing Action, Boston, MA: Kent Publishing
Company.
Bagozzi, Richard P. (1993), “Assessing Construct Validity in Personality Research: Applications
to Measures of Self-Esteem,” Journal of Research in Personality, 27 (1), 49-87.
POWERFUL BRAND INFLUENTIALS 179

Bagozzi, Richard P., Youjae Yi, and Lynn W. Phillips (1991), “Assessing Construct Validity in
Organizational Research,” Administrative Science Quarterly, 36 (3), 421-458.
Ball, Dwayne A. and Lori H. Tasaki (1992), “The Role and Measurement of Attachment in
Consumer Behavior,” Journal of Consumer Psychology, 1 (2), 155-172.
Baumgarten, Steven A. (1975), “The Innovative Communicator in the Diffusion Process,”
Journal of Marketing Research, 12, 12-18.
Bentler, Peter M. (1992), “On the Fit of Models to Covariances and Methodology to the
Bulletin,” Psychological Bulletin, 112 (3), 400-404.
Berlo, David K. (1960). The Process of Communication: An Introduction to Theory and Practice.
New York: Holt, Rinehart and Winston.
Berman, William H. and Michael B. Sperling (1994), “The Structure and Function of Adult
Attachment,” in Attachment in Adults: Clinical and Developmental Perspectives, ed.
Michael B. Sperling and William H. Berman, New York: Guilford Press, 3-28.
Bhattacharya, C.B. and Sankar Sen (2003), “Consumer–Company Identification: A Framework
for Understanding Consumers’ Relationships with Companies,” Journal of Marketing, 67
(April), 76-88.
Bloch, Peter H. (1986) "The Product Enthusiast: Implications for Marketing Strategy,” Journal
of Consumer Marketing, 3 (3), 51-62.
Bloch, Peter H. and Marsha L. Richins (1983), “A Theoretical Model for the Study of Product
Importance Perceptions,” Journal of Marketing, 47 (Summer), 69-81.
Bloch, Peter H., Daniel L. Sherrell, and Nancy M. Ridgway (1986), “Consumer Search: An
Extended Framework.” Journal of Consumer Research, 13 (1), 119-126.
POWERFUL BRAND INFLUENTIALS 180

Bougie, Roger, Pieters Rik, and Zeelenberg Marcel (2003), “Angry Customers Don’t Come Back,
They Get Back: the Experience and Behavioral Implications of Anger and Dissatisfaction
in Services,” Journal of the Academy of Marketing Science, 31(4), 377-393.
Bowlby, John (1973), Attachment and Loss: Vol 2. Separation: Anxiety and Anger, New York:
Basic Books.
——— (1979), The Making and Breaking of Affectional Bonds, London: Tavistock.
——— (1980), Loss: Sadness and Depression, New York: Basic Books.
Bristor, Julia (1990), “Enhanced Explanations of Word-Of-Mouth Communications: The Power
of Relationships,” Research in Consumer Behavior, 4, 51-83.
Byrne, Barbara M. (1998), Structural Equation Modeling With LISREL, PRELIS, and SIMPLIS:
Basic Concepts, Applications, and Programming, Mahwah, NJ: Lawrence Erlbaum
Associates.
Carr, Steven D. (1996), “The Cult of Brand Personality,” Marketing News, 30 (10), 4-9.
Carroll, Barbara A. and Aaron Ahuvia (2006), “Some Antecedents and Outcomes of Brand
Love,” Marketing Letters, 17 (2), 79-89.
Carver, Charles S. and Michael F. Scheier (1989), “Assessing Coping Strategies: A Theoretically
Based Approach,” Journal of Personality and Social Psychology, 56 (2), 267-283.
Childers, Terry L. (1986), “Assessment of the Psychometric Properties of an Opinion Leadership
Scale,” Journal of Marketing Research, 23 (5), 184-188.
Churchill, Gilbert A., Jr. (1979), “A Paradigm for Developing Better Measures of Marketing
Constructs,” Journal of Marketing Research, 16, (February), 64-73.
Clark, Ronald A. and Ronald E. Goldsmith (2005), “Interpersonal Influence and Consumer
Innovativeness,” International Journal of Consumer Studies, 30 (1), 34-43.
POWERFUL BRAND INFLUENTIALS 181

Collier, Mack (2007), “Ten Steps to Creating a Brand Ambassador Program,” MarketingProfs,
Retrieved from: http://www.marketingprofs.com/7/collier5.asp.
Collins, Nancy L. (1996), “Working Models of Attachment: Implications for Explanation,
Emotion, and Behavior,” Journal of Personality and Social Psychology, 71 (4), 810-832.
Collins, Nancy L. and Stephen J. Read (1994), “Cognitive Representation of Adult Attachment:
The Structure and Function of Working Models,” in Advances in Personal Relationships:
Attachment Processes in Adulthood, Vol. 5, ed. Kim Bartholomew and Daniel Perlman,
London: Jessica-Kingsley, 53-90.
Collins, Nancy L. and Lisa M. Allard (2001), “Cognitive Representations of Attachment: The
Content and Function of Working Models,” in Blackwell Handbook of Social
Psychology: Interpersonal Processes, Vol. 2, ed. Garth J. O. Fletcher and Margaret S.
Clark, United Kingdom: Blackwell Publishers, 60-85.
De Wulf, Kristof, Gaby Odekerken-Schröder, and Dawn Iacobucci (2001), “Investments in
Consumer Relationships: A Cross-Country and Cross-Industry Exploration,” Journal of
Marketing, 65 (October), 33-50.
Dichter, Ernest (1966), "How Word-of-Mouth Advertising Works," Harvard Business Review,
44 (November-December), 147-157.
Drigotas, Stephen M. and Caryl Rusbult (1992), “Should I Stay or Should I Go: A Dependence
Model of Break-Ups,” Journal of Personality and Social Psychology, 62 (1), 62-87.
Duhachek, Adam (2005), “Coping: A Multidimensional, Hierarchical Framework of Responses
to Stressful Consumption Episodes,” Journal Consumer Research, 32(June), 41-53.
Duhachek, Adam and Dawn Iacobucci (2005), “Consumer Personality and Coping: Testing
Rival Theories of Process,” Journal of Consumer Psychology, 15(1), 52-63.
POWERFUL BRAND INFLUENTIALS 182

Dwyer, Robert F., Paul H. Schurr, and Sejo Oh (1987), “Developing Buyer–Seller
Relationships,” Journal of Marketing, 51(April), 11-27.
Engel, James F., Robert J. Kegerreis, and Roger D. Blackwell (1969), “Word-Of-Mouth
Communications by the Innovator,” Journal of Marketing, 33 (7), 15-19.
Engel, James F. and Roger D. Blackwell (1982), Consumer Behavior, 4th ed., New York: The
Dryden Press.
Esch, Franz-Rudolf, Tobias Langner, Bernd H. Schmitt, and Patrick Geus (2006), "Are Brands
Forever? How Brand Knowledge and Relationships Affect Current and Future
Purchases," Journal of Product & Brand Management, 15 (2), 98-105.
Fazio, Russell H., Mary C. Powell, and Carol J. Williams (1989), “The Role of Attitude
Accessibility in the Attitude-to-Behavior Process,” Journal of Consumer Research, 16
(December), 280-288.
Fedorikhin, Alexander, C. Whan Park, and Matthew Thomson (2008), “Beyond Fit and Attitude:
The Effect of Emotional Attachment on Consumer Responses to Brand Extensions,”
Journal of Consumer Psychology, 18 (4), 281-291.
Feick, Lawrence F. and Linda L. Price (1987), “The Market Maven: A Diffuser of Marketplace
Information,” Journal of Marketing, 51 (1), 83-97.
Fiske, Susan T. and Pavelchak, Mark A. (1986), “Category-based versus Piecemeal-based
Affective Responses: Developments in Schema-Triggered Affect,” in Handbook of
Motivation and Cognition: Foundations of Social Behavior, ed. Richard M. Sorrentino
and Edward Tory Higgins, New York: Guilford Press, 167-203.
POWERFUL BRAND INFLUENTIALS 183

Flynn, Leisa R., Ronald E. Goldsmith, and Jacqueline K. Eastman (1994), “The King and
Summers Opinion Leadership Scale: Revision and Refinement,” Journal of Business
Research, 31, 51-64.
——— (1996), “Opinion Leaders and Opinion Seekers: Two New Measurement Scales,”
Journal of the Academy of Marketing Science, 24 (2), 137-147.
Folkman, Susan and Richard S. Lazarus (1985), “If It Changes It Must Be a Process: Study of
Emotion and Coping During Three Stages of a College Examination,” Journal of
Personality and Social Psychology, 48(1), 150-170.
Fornell, Claes and David F. Larcker (1981), “Evaluating Structural Equation Models with
Unobservable Variables and Measurement Error,” Journal of Marketing Research, 18
(February), 39-50.
Fournier, Susan (1998), “Consumers and Their Brands: Developing Relationship Theory in
Consumer Research,” Journal of Consumer Research, 24 (March), 343-373.
Garbarino, Ellen and Mark S. Johnson (1999),”The Different Roles of Satisfaction, Trust, and
Commitment in Customer Relationships,” Journal of Marketing, 63 (April), 70-87.
Gillin, Paul (2007), The New Influencers: A Marketer’s Guide to the New Social Media, Sanger,
CA: Quill Driver Books.
Godes, David and Dina Mayzlin (2004), “Using Online Conversations to Study Word-of-Mouth
Communication,” Marketing Science, 23, 545-560.
Goldsmith, Ronald E. and Charles F. Hofacker (1991), “Measuring Consumer Innovativeness,”
Journal of the Academy, 19 (3), 209-221.
Goldsmith, Ronald E. and Leisa Flynn (1992), “Identifying Innovators in Consumer Product
Markets,” European Journal of Marketing, 26 (12), 42-55.
POWERFUL BRAND INFLUENTIALS 184

Gounaris, Spiros and Vlasis Stathakopoulos (2004), “Antecedents and Consequences of Brand
Loyalty: An Empirical Study,” Journal of Brand Management, 11 (4), 283-306.
Hallberg, Garth (2004), “Is Your Loyalty programme Really Building Loyalty? Why Increasing
Emotional Attachment, Not Just Repeat Buying, is Key to Maximising Programme
Success,” Journal of Targeting, Measurement and Analysis for Marketing, 12 (3), 231-
241.
Hamilton, Herbert (1971), “Dimensions of Self-Designated Opinion Leadership and Their
Correlates,” The Public Opinion Quarterly, 35 (2), 266-274.
Hazan, Cindy and D. Zeifman (1999), “Pair Bonds as Attachment: Evaluating the Evidence,” in
Handbook of Attachment: Theory, Research, and Clinical Applications, ed. Jude Cassidy
and Phillip R. Shaver, New York: Guilford Press, 336-354.
Herr, Paul M., Frank R. Kardes, and John Kim (1991), "Effects of Word-of-Mouth and Product-
Attribute Information on Persuasion: An Accessibility-Diagnosticity Perspective,"
Journal of Consumer Research, 17 (March), 454-462.
Higie, Robin A., Lawrence Feick F., and Linda L. Price (1987), "Types and Amount of Word-of-
Mouth Communications About Retailers," Journal of Retailing, 63 (Fall), 260-278.
Hirschman, Elizabeth C. and William O. Adcock (1978),"An Examination of Innovative
Communicators, Opinion Leaders and Innovators for Men's Fashion Apparel,” in
Advances in Consumer Research, Vol. 5, ed. Kent Hunt, Provo, UT: Association for
Consumer Research, 308-314.
Holmes, John H. and John D. Lett (1977), “Product Sampling and Word of Mouth,” Journal of
Advertising Research, 17(5), 35-40.
POWERFUL BRAND INFLUENTIALS 185

Jackson, Barbara B. (1985), Winning and Keeping Industrial Customers: The Dynamics of
Customer Relationships, Lex-ington, MA: Lexington Books.
Jacoby, Jacob and David B. Kyner (1973), “Brand Loyalty vs. Repeat Purchasing Behavior,”
Journal of Marketing Research, 10 (February), 1-9.
Jacoby, Jacob and Wayne D. Hoyer (1981), "What if Opinion Leaders Didn't Know More? A
Question of Nomological Validity," in Advances in Consumer Research, Vol. 8, ed. Kent B.
Monroe, Provo, UT: Association for Consumer Research, 299-303.
Jain, Shailendra and Durairaj Maheswaran (2000), “Motivated Reasoning: A Depth of
Processing Perspective”, Journal of Consumer Research, Vol. 26 (March), 358-372.
Johnson, Dennis J. and Caryl E. Rusbult (1989), “Resisting Temptation: Devaluation of Alternative
Partners as a Means of Maintaining Commitment in Close Relationships,” Journal of
Personality and Social Psychology, 57 (6), 967- 980.
Jones, Thomas and W. Earl Sasser, Jr. (1995), “Why Satisfied Customers Defect,” Harvard Business
Review, 73 (November/December), 88-89.
Jöreskog, Karl G. and Dag Sörbom (1993), LISREL 8: Structural Equation Modeling with the
SIMPLIS Command Language, Mooresville, IL: Scientific Software.
——— (1996), LISREL 8: User’s Reference Guide, Chicago, IL: Scientific Software International.
Katz, Elihu and Paul F. Lazarsfeld (1955), Personal Influence: The Part Played by People in the
Flow of Mass Communications, Glencoe, IL: The Free Press.
Keller, Ed and Jon Berry (2003), The Influentials: One American in Ten Tells the Other Nine
How to Vote, Where to Eat, and What to Buy, New York: Free Press.
Keltner, Dacher and Brenda N. Buswell (1997), “Embarrassment: Its Distinct Form and
Appeasement Functions,” Psychological Bulletin, 122 (3), 250-270.
POWERFUL BRAND INFLUENTIALS 186

King, Charles W. and John O. Summers (1970), “Overlap of Opinion Leadership Across
Consumer Product Categories,” Journal of Marketing Research, 7 (2), 43-50.
Kisielius, Jolita and Brian Sternthal (1986), "Examining the Vividness Controversy: An
Availability-Valence Interpretation," Journal of Consumer Research, 12 (March), 418-
431.
Kotler, Philip and Gerald Zaltman (1976), “Targeting Prospects for a New Product,” Journal of
Advertising Research, 16 (1), 7-18.
Kozinets, Robert V. (2001), “Utopian Enterprise: Articulating the Meanings of Star Trek’s Culture of
Consumption,” Journal of Consumer Research, 28 (June), 67-88.
Kring, Ann M., David A. Smith, and John M. Neale (1994), “Individual Differences in
Dispositional Expressiveness: Development and Validation of the Emotional Expressivity
Scale,” Journal of Personality and Social Psychology, 66(5), 934-949.
LaBarbera, Priscilla and David Mazursky (1983). “A Longitudinal Assessment of Consumer
Satisfaction/Dissatisfaction: The Dynamic Aspect of the Cognitive Process.” Journal of
Marketing Research, 20 (November): 393-404.
Lazarus, Richard S. (1991), Emotion and Adaptation, New York: Oxford University Press.
——— (1999), Stress and Emotion, New York: Springer.
Lazarus, Richard S. and Susan Folkman (1984), Stress, Appraisal, and Coping, New York:
Springer.
Levinger, George (1980), “Toward the Analysis of Close Relationships,” Journal of
Experimental Social Psychology, 16 (6), 510-544.
POWERFUL BRAND INFLUENTIALS 187

Lyubomirsky, Sonja, Lorie Sousa, and Rene Dickerhoof (2006), “The Costs and Benefits of
Writing, Talking, and Thinking about Life’s Triumphs and Defeats,” Journal of
Personality and Social Psychology, 90 (4), 692-708.
Madrigal, Robert (1995), “Cognitive and Affective Determinants of Fan Satisfaction with Sporting
Event Attendance,” Journal of Leisure Research, 27 (3), 205-227.
Mano, Haim and Richard L. Oliver (1993), “Assessing the Dimensionality and Structure of
Consumption Experience: Evaluation, Feeling, and Satisfaction,” Journal of Consumer
Research, 20 (December), 451-466.
Marsden, Paul, Alain Samson, and Neville Upton (2005), “Advocacy Drives Growth,” Brand
Strategy, 198 (December), 45-48.
Matzler, Kurt, Elisabeth A. Pichler, and Andrea Hemetsberger (2007), “Who is Spreading the Word?
The Positive Influence of Extraversion on Consumer Passion and Evangelism,” in
Proceedings of the American Marketing Association, Volume 18, ed. Andrea L. Dixon and
Karen A. Machleit , San Diego: American Marketing Association, 25-32.
Mazzarol, Tim, Jillian C. Sweeney, and Geoffrey N. Soutar (2007), “Conceptualizing Word-of-
Mouth Activity, Triggers and Conditions: An Exploratory Study,” European Journal of
Marketing, Volume 41, Issue 11-12, 1475-1494.
McAlexander, James H., Stephen K. Kim, and Scott D. Roberts (2003), “Loyalty: The Influences of
Satisfaction and Brand Community Integration,” Journal of Marketing Theory and Practice,
11(Fall), 1-11.
McConnell, Ben and Jackie Huba (2007), Creating Customer Evangelists: How Loyal Customers
Become a Volunteer Sales Force, Chicago: Kaplan Business.
POWERFUL BRAND INFLUENTIALS 188

McCullough, Michael E., K. Chris Rachal, Steven J. Sandage, Everett L. Worthington Jr., Susan
Wade Brown, Terry L. Hight (1998), “Interpersonal Forgiving in Close Relationships: II.
Theoretical Elaboration and Measurement,” Journal of Personality and Social
Psychology, 75(6), 1586-1603.
Messick, Samuel (1981), “Constructs and Their Vicissitudes in Educational and Psychological
Measurement,” Psychological Bulletin, 89 (3), 575-588.
Midgeley, David F. (1976), “A Simple Mathematical Theory of Innovative Behavior.” Journal of
Consumer Research, 3 (June), 31-41.
Midgley, David F. and Grahame R. Dowling (1978), “Innovativeness: The Concept and its
Measurement,” Journal of Consumer Research, 4 (March), 229-242.
Mikulincer, Mario (1998), “Attachment Working Models and the Sense of Trust: An Exploration of
Interaction Goals and Affect Regulation,” Journal of Personality and Social Psychology,
74(5), 1209-1224.
Mikulincer, Mario and Phillip R. Shaver (2003), “The Attachment Behavioral System in Adulthood:
Activation, Psychodynamics, and Interpersonal Processes,” in Advances in Experimental
Social Psychology, ed. Mark P. Zanna, New York: Academic Press, 53-152.
——— (2005), “Mental Representations of Attachment Security: Theoretical Foundation for a
Positive Social Psychology,” In Interpersonal Cognition, ed. Mark W. Baldwin, New
York: The Guilford Press, 233-266.
——— (2007), Attachment in Adulthood: Structure, Dynamics and Change. New York: Guilford
Press.
POWERFUL BRAND INFLUENTIALS 189

Mikulincer, Mario; Gilad Hirschberger, Orit Nachmias, and Omri Gillath (2001), “The Affective
Component of the Secure Base Schema: Affective Priming with Representations of
Attachment Security,” Journal of Personality and Social Psychology, 81(2), 305-321.
Mitchell, Andrew A. and Peter A. Dacin (1996), “The Assessment of Alternative Measures of
Consumer Expertise,” Journal of Consumer Research, 23 (December), 219-239.
Muniz, Albert M., Jr. and Thomas C. O’Guinn (2001), “Brand Community,” Journal of
Consumer Research, 27 (March), 412-432.
Myers, James H. and Thomas S. Robertson (1972), “Dimensions of Opinion Leadership,” Journal of
Marketing Research, 9 (February), 41-46.
Nisbett, Richard E. and Ross, Lee (1980), Human Inference: Strategies and Shortcomings of
Social Judgment, Englewood Cliffs, NJ: Prentice-Hall.
O’Guinn, Thomas C. (1991), “Touching Greatness: The Central Midwest Barry Manilow Fan
Club,” in Highways and Buyways, ed. Russell W. Belk, Provo, UT: Association for
Consumer Research, 102-111.
Oliver, Richard L. (1980), “A Cognitive Model of the Antecedents and Consequences of
Satisfaction Decisions.” Journal of Marketing Research, 17 (November): 460-469.
——— (1997), Satisfaction: A Behavioral Perspective on the Customer, New York: McGraw-
Hill.
Oliver, Richard L., Roland T. Rust, and Sajeev Varki (1997), “Customer Delight: Foundations,
Findings, and Managerial Insight,” Journal of Retailing, 73 (Fall), 311-336.
Park, C. Whan, Deborah J. MacInnis, and Joseph R. Priester (2006), “Beyond Attitudes:
Attachment and Consumer Behavior,” Seoul National Journal, 12 (2), 3-36.
POWERFUL BRAND INFLUENTIALS 190

______ (2007), “Brand Attachment as a Strategic Brand Exemplar,” in Handbook of Brand and
Experience Management, ed. Bernd H. Schmitt and David Rogers, Nrthampton, MA:
Edward Elgar.
Park, C. Whan, Joseph Priester, Deborah J. MacInnis, and Zhong Wan (2009), “The Connection-
Prominence Attachment Model (CPAM): A Conceptual and Methodological Exploration
of Brand Attachment,” in Handbook of Brand Relationships, Chap. 17, ed. Deborah J.
MacInnis, C. Whan Park, and Joseph Priester, Armonk, NY: M.E. Sharpe, 327-341.
Park, C. Whan, Deborah J. MacInnis, Joseph Priester, Andreas B. Eisengerich, and Dawn
Iacabucci (2010), “Brand Attachment and Brand Attitude Strength: Conceptual and
Empirical Differentiation of Two Critical Brand Equity Drivers,” Journal of Marketing,
74(6), 1-17.
Paulssen, Marcel and Richard B. Bagozzi (2009), "Customer Coping in Response to Relationship
Transgressions: An Attachment Theoretic Approach," in Handbook of Brand
Relationships, Chap. 19, ed. Deborah J. MacInnis, C. Whan Park, and Joseph Priester,
Armonk, NY: M.E. Sharpe, 358-376.
Petty, Richard E., Curtis P. Haugtvedt, and Stephen M. Smith (1995), “Elaboration as a
Determinant of Attitude Strength: Creating Attitudes That are Persistent, Resistant, and
Predictive of Behavior,” in Attitude Strength: Antecedents and Consequences, ed.
Richard E. Petty and Jon A. Krosnick, Mahwah, NJ: Lawrence Erlbaum Associates, 93-
130.
Pimentel, Ronald W. and Kristy E. Reynolds (2004), “A Model for Consumer Devotion:
Affective Commitment with Proactive Sustaining Behaviors,” Academy of Marketing
Science Review, 2004 (5), 1-45.
POWERFUL BRAND INFLUENTIALS 191

Ping, Robert A., Jr. (1993), “The Effects of Satisfaction and Structural Constraints on Retailer
Exiting Voice, Loyalty, Opportunism, and Neglect,” Journal of Retailing 69 (3), 320-352.
Price, Linda L., Lawrence F. Feick, and Robin A. Higie (1987), “Information Sensitive
Consumers and Market Information,” The Journal of Consumer Affairs, 21 (2), 328-341.
Priester, Joseph R., Dhananjay Nayakankuppam, Monique A. Fleming, and John Godek (2004),
“The A2SC2 Model: The Influence of Attitudes and Attitude Strength on Consideration
and Choice,” Journal of Consumer Research, 30 (March), 574-587.
Ramanathan, Suresh and Patti Williams (2007), “Immediate and Delayed Emotional
Consequences of Indulgence: The Moderating Influence of Personality Type on Mixed
Emotions,” Journal of Consumer Research, 34 (August), 212-223.
Reichheld, Frederick F. (2003), “The One Number You Need to Grow,” Harvard Business
Review, 81 (12), 46-54.
Richins, Marsha L. and Teri Root-Shaffer (1988), “The Role of Involvement and Opinion
Leadership in Consumer Word-Of-Mouth: An Implicit Model made Explicit,” in
Advances in Consumer Research, Vol. 15, ed. Michael J. Houston, Provo, UT:
Association for Consumer Research, 32-36.
Rogers, Everett M. (1962), Diffusion of Innovations, Glencoe, IL: The Free Press.
Rogers, Everett M. and David G. Cartano (1962), "Methods of Measuring Opinion Leadership,"
Public Opinion Quarterly, 26 (Fall), 435-441.
Roseman, Ira J. (1984), “Cognitive Determinants of Emotion: A Structural Theory,” in
Review of Personality and Social Psychology, 5, 11-36.
______ (1991), “Appraisal Determinants of Discrete Emotions,” Cognition and Emotion, 5, 161-
200.
POWERFUL BRAND INFLUENTIALS 192

______ (2001), “A Model of Appraisal in the Emotion System: Integrating Theory, Research,
and Applications, in Appraisal Processes in Emotion: Theory, Methods, Research, ed.
Klaus R. Scherer, Angela Schorr, and Tom Johnstone, New York: Oxford University
Press, 68-91.
Rosenblatt, Paul C. (1977), “Needed Research on Commitment in Marriage,” in Close
Relationships, ed. George Levinger and Harold L. Raush, Amherst, MA: University of
Massachusetts Press.
Rozanski, Horacio D., Allen G. Baum, and Bradley T. Wolfsen (1999), “Brand Zealots:
Realizing the Full Value of Emotional Brand Loyalty,” Strategy and Business, Issue 17,
51-62.
Rusbult, Caryl E. and Isabella Zembrodt (1983), “Responses to Dissatisfaction in Romantic
Involvements: A Multidimensional Scaling Analysis,” Journal of Experimental Social
Psychology, 19 (3), 274-293.
Rusbult, Carly E., Julie Verette, Gregory A. Whitney, Linda F. Slovik, and Issac Lipkus (1991),
“Accommodation Processes in Close Relationships: Theory and Preliminary Empirical
Evidence,” Journal of Personality and Social Psychology, 60 (1), 53-78.
Rusticus, Sven (2005), "Creating Brand Advocates." In Connected Marketing: The Viral, Buzz
and Word of Mouth Revolution, ed. Justin Kirby and Paul Marsden, London:
Butterworth-Heinemann, 47-58.
Ryu, Gangseog and Lawrence Feick (2007), “A Penny for Your thoughts: Referral Reward
Programs and Referral Likelihood,” Journal of Marketing, 71 (1), 84-94.
POWERFUL BRAND INFLUENTIALS 193

Scherer, Klaus R. (1988), “Criterion for Emotion-Antecedent Appraisal: A Review,” in
Cognitive Perspective on Emotion and Motivation, ed. Vernon Hamilton, Gordon H.
Bower, and Nico H. Frijda, New York: Kluwer Academic/Plenum Publishers, 89-126.
______ (2001), “Appraisal Considered as a Process of Multilevel Sequential Checking,” in
Appraisal Processes in Emotion: Theory, Methods, Research, ed. Klaus R. Scherer,
Angela Schorr, and Tom Johnstone, New York: Oxford University Press, 92-120.
Schouten, John W. and James H. McAlexander (1995), “Subcultures of Consumption: An
Ethnography of the New Bikers,” Journal of Consumer Research, 22 (June), 43- 61.
Skinner, Natalie and Neil Brewer (2002), “The Dynamics of Threat and Challenge Appraisals
Prior to Stressful Achievement Events,” Journal of Personality and Social Psychology,
83 (3), 678-692.
Slama, Mark E. and Terrell G. Williams (1990), “Generalization of the Market Maven's
Information Provision Tendency Across Product Categories,” in Advances in Consumer
Research, Vol. 17, ed. Marvin E. Goldberg, Gerald Gorn, and Richard W. Pollay, Provo,
UT: Association for Consumer Research, 48-52.
Smith, Craig A. and Phoebe C. Ellsworth (1985), “Patterns of Cognitive Appraisal in Emotion,”
Journal of Personality and Social Psychology, 48 (4), 813-838.
Spector, Paul E. (1992), Summated Rating Scale Construction: An Introduction, Newbury Park,
CA: Sage.
Summers, John O. (1971), “Generalized Change Agents and Innovativeness,” Journal of
Marketing Research, 8 (8), 313-316.

POWERFUL BRAND INFLUENTIALS 194

Sundaram, D. S., Kaushik Mitra, and Cynthia Webster (1998), “Word-of-Mouth
Communications: A Motivational Analysis,” in Advances in Consumer Research, Vol. 25,
ed. Joseph W. Alba and Wesley Hutchinson, Provo, UT: Association for Consumer
Research, 527-531.
Swan, John E. and Richard L. Oliver (1989), “Postpurchase Communications by Consumers,”
Journal of Retailing, 65 (4), 516-533.
Sweeney, Ric (2002), “‘Brand ambassadors’ Give Your Business A Boost”, Business Courier,
Retrieved from:
http://www.bizjournals.com/cincinnati/stories/2002/04/15/smallb3.html?page=all.
Szymanski, David M. and David H. Henard (2001), “Customer Satisfaction: A Meta-Analysis of
the Empirical Evidence,” Journal of the Academy of Marketing Science, 29 (Winter), 16-
35.
Tangney, June Price (1999), “The Self-Conscious Emotions: Shame, Guilt, Embarrassment and
Pride,” in Handbook of Cognition and Emotion, ed. Tim Dalgleish and Mick J. Power,
New York: Wiley and Sons, 541-568.
Tangney, June Price and Ronda L. Dearing (2002), Shame and Guilt, New York: Guilford.
Thompson, Craig J., Aric Rindfleisch, and Zeynep Arsel (2006), “Emotional Branding and the
Strategic Branding of the Doppelganger Brand Image,” Journal of Marketing, 70
(January), 50-65.
Thomson, Matthew, Deborah J. MacInnis, and C. Whan Park (2005), “The Ties That Bind:
Measuring the Strength of Consumers’ Emotional Attachments to Brands,” Journal of
Consumer Psychology, 15 (1), 77-91.
POWERFUL BRAND INFLUENTIALS 195

Tracy, Jessica L. and Richard W. Robins (2004), “Putting the Self into Self-Conscious Emotions:
A Theoretical Model,” Psychological Inquiry, 15(2), 103-125.
______ (2007), “The Psychological Structure of Pride: A Tale of Two Facets,” Journal of
Personality and Social Psychology, 92(3), 506−525.
Turnbull, Peter W. and Meenaghan A. (1980), "Diffusion of Innovation and Opinion
Leadership," European Journal of Marketing, 14 (1), 3-33.
Voight, Joan (2007), “The New Brand Ambassadors”, Adweek, Retrieved from:
http://www.adweek.com/news/advertising/new-brand-ambassadors-91501.
Walsh, Gianfranco, Kevin P. Gwinner, and Scott R. Swanson (2004), “What makes Mavens
Tick? Exploring the Motives of Market Mavens’ Initiation of Information Diffusion,”
Journal of Consumer Marketing, 21 (2), 109-122.
Weimann, Gabriel (1994), The Influentials: People Who Influence People, Albany: State
University of New York.
Weiner, Bernard (1985), “An Attributional Theory of Achievement Motivation and Emotion,”
Psychological Review, 92(4), 548-573.
Westbrook, Robert A. (1987), “Product / Consumption-Based Affective Responses and
Postpurchase Processes,” Journal of Marketing Research, 24 (3), 258-270.
Wiedmann, Klaus-Peter, Gianfranco Walsh, and Vincent-Wayne Mitchell (2001), “The
Mannmaven: An Agent for Diffusing Market Information,” Journal of Marketing
Communications, 7 (4), 195-212.
Wikipedia: “Brand Ambassador”, Retrieved from:
http://en.wikipedia.org/wiki/Brand_ambassador.
POWERFUL BRAND INFLUENTIALS 196

Wikipedia: “Evangelism Marketing”, Retrieved from:
http://en.wikipedia.org/wiki/Evangelism_marketing.
Winer, Russell S. (2001), “A Framework for Customer Relationship Management,” California
Management Review, 43 (Summer), 89-105.
Yahoo! and comScore Networks (2006), “Engaging Advocates through Search and Social Media”,
Retrieved from: http://createwithcontext.com/media/yahoo-summit-series.pdf.
Yaverbaum, Eric, Ilise Benun, Richard Kirshenbaum, and Robert W. Bly (2006), Public Relations
for Dummies, Indiana: Wiley Publishing.
Zaichkowsky, Judith Lynne (1985), “Measuring the Involvement Construct,” Journal of Consumer
Research, 12 (December), 341-352.
Zboja, James J. and Clay M. Voorhees (2006), "The Impact of Brand Trust and Satisfaction on
Retailer Repurchase Intentions," Journal of Services Marketing, 20 (5), 381-390.

POWERFUL BRAND INFLUENTIALS 197

Appendices

Appendix A.

TABLE A-1. (Chapter 4) Brand Attachment and Behaviors from Park, MacInnis, and
Priester (2006), “Beyond Attitudes: Attachment and Consumer Behavior,” Seoul Journal of
Business, 12 (December), p. 19

Attachment
Types of Resource Sacrifice
Brand Supporting Behaviors
Self-Image
Resources
Personal
Discretionary
Resources
Low Low Low None
Moderate
Low

Moderate

Repeat purchase behavior accompanied with paying a
price premium, postponement of purchase, or
prolonged brand search
Moderate

Low

Repeat purchase behavior accompanied with public
display, public defending of a brand, or
recommendation to others
Moderate Moderate
Repeat purchase behavior accompanied with
participation in the brand community
High
Low/Moderate High
Stronger repeat purchase behavior accompanied with
more willingness to pay a price premium, postpone
purchase, or prolong brand search
Additional brand supporting behaviors: investing
in a firm, applying for a job to work, refusal to
exchange the attached product for financial gains
High

Low/Moderate

Stronger repeat purchase behavior accompanied with
more willingness to display, defend, or recommend a
brand to others
Additional brand supporting behaviors: investing
in a firm, applying for a job to work, refusal to
exchange the attached product for financial gains
High High
Stronger repeat purchase behavior accompanied with
more willingness to participate in the brand
community
Additional brand supporting behaviors: investing
in a firm, applying for a job to work, refusal to
exchange the attached product for financial gains

POWERFUL BRAND INFLUENTIALS 198

Appendix B.

TABLE B-1. (Chapter 8) Measures of Constructs Related to the Powerful Brand
Influential (Study 2 – Study 3)

Construct Item Number Measures
Market
Maven
MM 1. People ask me for information about products, places to shop, or sales.
MM 2.
If someone asked where to get the best buy on several types of products,
I could tell him
MM 3.
My friends think of me as a good source of information when it comes to
new products
Opinion
Leader
OL 1.
My opinion on (PRODUCT CATEGORY) seems not to count with other
people.
R

OL 2.
When they choose a (PRODUCT CATEGORY), other people do not
turn to me for advice.
R

OL 3.
Other people rarely come to me for advice about choosing (PRODUCT
CATEGORY).
OL 4.
People that I know pick (PRODUCT CATEGORY) based on what I
have told them.
OL 5.
I often persuade other people to buy the (PRODUCT CATEGORY) that
I like.
OL 6. I often influence people’s opinions about (PRODUCT CATEGORY).
Brand
Attachment
BA 1.
(BSC 1)
To what extent do you feel that you are personally connected to
(BRAND NAME)?
BA 2.
(BSC 2)

To what extent is (BRAND NAME) part of you and who you are?
BA 3.
(PRO 1)
To what extent are your thoughts and feelings toward (BRAND NAME)
often automatic, coming to mind seemingly on their own?
BA 4.
(PRO 2)
To what extent do your thoughts and feelings toward (BRAND NAME)
come to you naturally and instantly?
Brand
Ambassador
Brand
Ambassador
1.
Nowadays, several companies such as Unilever, Sony, McDonald’s,
JetBlue have launched brand ambassador programs. They “hire”
consumers, via incentives and rewards (e.g., prizes, cash, store credits,
product samples, gifts, discounts) to advertise and promote their brands
to other consumers. Brand ambassadors are advised by marketers to
openly reveal that they’re the brand representatives when promoting the
brand. To what extent would you be interested in becoming a brand
ambassador of (BRAND NAME)?
Note:
R
Reverse coded items.
Abbreviations: MM (Market Maven). OL (Opinion Leader). BA (Brand Attachment). BSC (Brand-Self
Connection indicator of brand attachment). PRO (Prominence of Thoughts and Feelings indicator of brand
attachment).

POWERFUL BRAND INFLUENTIALS 199

Appendix C.

TABLE C-1. (Chapter 9) Dependent Variable Measures (Study 3)

Construct
Internal
Consistency
(α)
Item
Number
Measures
Emotion
Expressiveness
0.99
EE 1. I keep my feelings to myself.
R

EE 2. I display my emotions to other people.
EE 3. People can read my emotions.
EE 4.
Even if I am feeling very emotional, I don’t let
others see my feelings.
R

EE 5. I think of myself as emotionally expressive.
EE 6. I hold my feelings in.
R

Note:
R
Reverse coded items.
Abbreviations: EE (Emotion Expressiveness).

POWERFUL BRAND INFLUENTIALS 200

Appendix D.

TABLE D-1. (Chapter 10) Dependent Variable Measures (Study 4)

Construct
Internal
Consistency (α)
Item
Number
Measures
Behavioral Responses to a Positive Brand Event
Emotion-
Expressive
Communication
0.78
1. I told others how I feel.
2. I took time to express my emotions.
3. I want to talk to someone about the election.
4. I celebrated the event.
Action-Focused
Behavior
0.84
1. I tried to get advice from someone about what to do.
2. I tried to make a plan of action.
3. I followed a plan to make things better.
4.
I asked someone with similar experiences what they
did.
Positive
Interpretation &
Growth
0.79
1. I realized that my feelings are valid and justified.
2. I grew as a person in a good way.
3. I looked for the good in what happened.
Types of Word-of-Mouth In-Group / Out Group in Responses to a Positive Brand Event
Overall Talking
Tendency
NA 1. I would really want to talk about the election.
Derogation of
Alternative Voice
in-group
= .85;
out-group
= .76
1.
I would express my dissatisfaction with the
opposition party.
2.
I would complain about the opposition party and
point out that their party has some issues to address.
Action-Focused
Brand Advocacy
Voice
In-group
= .83;
Out-group
= .77
1. I would praise our party.
2.
I would try to discuss what our party could do to help
the future of our country.
3. I would give this person information and suggestions.
4.
I would try to work jointly with this person to help
improve the situation of our country.
Emotion-Focused
Voice
In group
= .79;
Out group
= .75
1. I would try to talk to this person to feel better.
2.
I would try to talk to this person to understand and
make sense of what happened in the election.
3. I would give this person my emotional support.
4.
I would replay my feelings about the election
outcome with this person.
Self-Experience
Focused Voice
NA 1.
I would focus more on my experiences of the election
than on the election per se.

Abstract (if available)

Abstract This article focuses on a unique and important consumer construct, a brand’s influential consumers, namely “the powerful brand influential”. In the marketplace, influential consumers designate those who exert exceptional influence on other consumers’ decision making. Extant literature has established opinion leaders as influential consumers of a specific product category and market mavens as influential consumers of the general marketplace. Although the existence and power of a brand’s influential consumers are well-documented and the designations of these consumers are proliferated, we have limited understanding about these consumers. The conceptualization of a brand’s influential consumers, the distinctiveness of the construct from existing concepts, and the measurement for the construct are largely missing. ❧ Drawing from various domains of literature, this article provides a systematic conceptualization of a brand’s influential consumers that reflects four defining components. Corresponding to these conceptual properties, the powerful brand influential is defined as consumers (1) who are strongly attached to a given brand and (2) who have high tendency to influence other consumers’ brand decisions through (3) a high engagement of (4) passionate and emotionally evocative word-of-mouth about the brand. This article then details the differentiation of the new construct from six conceptually related constructs and posits a set of corresponding propositions. Study 1 develops a 6-item measurement, the Powerful Brand Influential Scale (PBIS), to represent each of the four conceptual properties. Study 2, 3, and 4 provide converging support that the powerful brand influential construct is theoretically and empirically distinguishable from the six constructs including market maven, opinion leader, brand attachment, brand ambassador, brand attitude strength, and brand satisfaction. Our proposition testing substantiates the distinctiveness and added value of the powerful brand influential construct – demonstrating that this new construct better predicts many key brand outcomes including brand word-of-mouth, intrinsic pleasure derived from brand word-of-mouth, brand advocacy behavior, tendency to influence others’ brand decision, brand loyal behavior, actual brand purchase as well as desirable emotional and word-of-mouth responses toward a positive brand event than the related constructs examined. Moreover, using different samples and brands from diverse product categories, the four studies establish the reliability, robustness, known group, convergent, discriminant, and predictive validity of the PBIS. The article concludes by discussing theoretical and managerial contribution, findings, and implications of the powerful brand influential construct and the PBIS along with limitations and areas of future research.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

How sequels seduce: consumers' affective expectations for entertainment experiences

PDF

Creation and influence of visual verbal communication: antecedents and consequences of photo-text similarity in consumer-generated communication

PDF

The motivational power of beauty: how aesthetically appealing products drive purchase effort in consumers

PDF

Forgiveness: elucidating the underlying psychological processes that foster brand forgiveness and interpersonal forgiveness

PDF

The effects of a customer's comparative processing with positive and negative information on product choice

Asset Metadata

Creator Wan, Zhong (Heather) (author)

Core Title Powerful brand influentials: conceptualization, measurement, and distinctiveness of a brand’s influential consumers

School Marshall School of Business

Degree Doctor of Philosophy

Degree Program Business Administration

Publication Date 12/10/2012

Defense Date 08/17/2012

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag brand advocacy,brand ambassador,brand attachment,brand attitude strength,brand influentials,brand satisfaction,influential consumer measurement scale,influential consumers,interpersonal communication,market mavens,OAI-PMH Harvest,opinion leaders,word-of-mouth

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor MacInnis, Deborah J. (committee chair), Cody, Michael J. (committee member), Park, C. Whan (committee member), Ulkumen, Gulden (committee member)

Creator Email heather.wan@unilever.com,heer330@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-125653

Unique identifier UC11292253

Identifier usctheses-c3-125653 (legacy record id)

Legacy Identifier etd-WanZhongHe-1383.pdf

Dmrecord 125653

Document Type Dissertation

Rights Wan, Zhong (Heather)

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA