BUYER-SELLER RELATIONSHIP QUALITY AND BRAND EQUITY IN THE
THOROUGHBRED CONSIGNMENT INDUSTRY
ADAM JEFFERSON MARQUARDT
A DISSERTATION
Presented to the Department of Marketing
and the Graduate School of the University of Oregon
-• in partial fulfillment of the requirements
for the degree of
Doctor of Philosophy
September 2007
"Buyer-Seller Relationship Quality and Brand Equity in the Thoroughbred Consignment
Industry,"  a dissertation prepared by Adam Jefferson Marquardt in partial fulfdlment of
the requirements for the Doctor of Philosophy degree in the Department of Marketing at
the University of Oregon. This dissertation has been approved and accepted by;
Dr. Lynn R. Kahle, Chair offhe Examining Committee
A.,, ic
Committee in Charge: Dr. Lynn R. Kahle, Chair
Dr. David M. Boush
Dr. Susan L. Golieic
Dr. Diane Dunlap
Accepted by:
Dean of the Graduate School
Copyright © 2007 by Adam Jefferson Marquardt
All rights reserved.
An Abstract of the Dissertation of
Adam Jefferson Marquardt for the degree of Doctor of Philosophy
in the Department of Marketing to he taken September 2007
Title: BUYER-SELLER RELATIONSHIP QUALITY AND BRAND EQUITY IN
THE THOROUGHBRED CONSIGNMENT INDUSTRY
Approved:
Jr. Lynn R. Kahle
In the relationship marketing literature, our knowledge of husiness-to-husiness
buyer-seller relationship quality and its role in influencing seller brand equity remains
incomplete. This study develops a theory-grounded conceptual framework regarding
the mediating role of buyer-seller relationship quality between buyer attitudes toward
the seller's corporate and product brands, and seUer brand equity in a business-to-
business context. A mail survey was administered to buyers of Thoroughbred horses
regarding the past purchase of a racing prospect (product brand) from a Thoroughbred
consignor (seller brand), generating 249 buyer responses. Structural equation
modeling was used to test the hypotheses. Results reveal that buyer-seller relationship
quality fully mediates the positive direct path between buyer attitude toward the seller
and seller brand equity, and partially mediates the negative direct path between buyer
attitude toward the product and seller brand equity. The finding that buyer attitude
toward the product is inversely predictive of seller brand equity in this context is
particularly interesting, because it imphes that in spite of the expense and uncertainty
attached to the purchase of a Thoroughbred racing prospect (product brand), the value
the buyer ascribes to the consignor (seller) is marginalized when the buyer has a more
favorable attitude toward the horse. The finding that buyer-seller relationship quality
partially mediates this path is also very interesting, because it implies that as the buyer
exhibits a less favorable attitude toward the racing prospect (product brand), the value
attributed to the consignor (seller) increases. The results of this study have significant
implications for sellers within speculative and competitive business climates.
CURRICULUM VITAE
NAME OF AUTHOR: Adam Jefferson Marquardt
GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED:
University of Oregon
University of Tennessee
Salisbury University
DEGREES AWARDED:
Doctor of Philosophy in Marketing, 2007, University of Oregon
Master of Business Administration, 2002, University of Tennessee
Bachelor of Science in Business Administration, 1993, Salisbury University
AREAS OF SPECIAL INTEREST:
Brand Development and Management
Entrepreneurship
Sports and Entertainment Marketing
PROFESSIONAL EXPERIENCE:
Graduate Teaching Fellow, Department of Marketing, University of Oregon, Eugene,
OR, 2003-2007
Graduate Assistant, Department of Marketing, University of Tennessee, Knoxville,
TN, 2001-2002
Owner and President, Due Respect, Howes Cave, NY, 1995-2001.
Director of Sales and Marketing, Story House Corp., Charlotteville, NY,2 000-2001
Owner and President, Calendar Club, Bellevue, WA, 1999-2000
General Manager, Urschel Ltd., Cooperstown, NY, 1997-1999
Owner and President, Chatter's Ice Cream and Coffee Shop, Lake George, NY, 1996-
1997
Marketing Manager, Support Services Alliance, Inc., Schoharie, NY, 1987-1996
GRANTS, AWARDS AND HONORS:
Lundquist College of Business GSPHDP Research Award (2006)
Nominated by the small business track chair for the best student paper award at the
2006 USASBE/SBl Entrepreneurship Conference
Kauffman Foundation Ph.D. Seminar in Entrepreneurship and Seminar Scholarship
Award, Case Western University (2005)
Merle King Smith Marketing Scholars Award, University of Oregon (2003-2007)
Calvin Reed Smith Research Grant (2003-2007)
Merle King Smith Travel Grant (2003-2007)
Graduate Teaching Fellow, University of Oregon (2003-2007)
Richland Ventures Fellowship Recipient, University of Tennessee (2001-2002)
Graduate Assistant, University of Tennessee (2001-2002)
Graduated with Honors, Salisbury University (1993)
Empire State Regents Scholarship Recipient (1988-1989)
PUBLICATIONS:
Davis, Donna F., Susan L. Golicic, and Adam J. Marquardt (2007)," Branding a B2B
Service: Does a Brand Differentiate a Logistics Service Provider?" Industrial
Marketing Management,( Forthcoming).
Kim, Woo Sung, David M. Boush, Adam J. Marquardt, and Lynn R. Kahle (2005),
"Values, Brands, and Image," in Lynn R. Kahle and Chung-Hyun Kim, Eds.
Creating Images and the Psychology of Marketing Communication, Mahwah, NJ:
Lawrence Erlbaum.
ACKNOWLEDGEMENTS
There are a number of people I would like to thank for helping me achieve this
milestone. I am especially grateful to my family and friends. Thank you for aU your
support through the ups and downs of life. You have always encouraged me to push
beyond my fears and limitations, inspiring me through your thoughts, words and actions;
and for that I am eternally indebted. In particular, I would like to thank my parents
Robert and JiU, my sister Valerie, my brother Lucas, Trish Lingo, Helder Sebastiao, and
Jim and Doris Wright for all your encouragement and support.
I am also extremely appreciative of the mentorship efforts provided by my
committee members. Dr. Lynn R. Kahle, Dr. David M. Boush, Dr. Susan L. Golicic, and
Dr. Diane Dunlap. Your tireless guidance and encouragement throughout not only my
dissertation, but my entire Ph.D. program have been invaluable. I am proud to have had
the chance to befriend and learn from each of you. I am especially indebted to Lynn and
Susan for your guidance and mentoring through all of my most challenging decisions.
Lastly, I would like to thank the faculty and Ph.D. students in the University of
Oregon's Lundquist College of Business for all your contributions in helping me grow as
a scholar, a teacher and a person. You have helped make my experience at the University
of Oregon one that I will cherish forever.
DEDICATION
This dissertation is dedicated to my parents Robert and Jill, and my siblings Valerie and
Lucas. Thank you so much for your love, patience and support. You are my inspiration!
TABLE OF CONTENTS
Chapter
I. INTRODUCTION
Research Opportunity 2.
Contributions 4
Document Framework 7
II. BUILDING THE THEORY 10
Theoretical Model 11
Resource-Advantage Theory 12
Brand Attitudes 1^
Corporate Brand 15
Product Brand 18
Brand Equity 20
Relationship Quality 21
Satisfaction 24
Trust 24
Commitment 26
Relational Resources 27
Research Hypotheses 29
Research Context, Description and Application 32
Thoroughbred Consignment 36
Consignor's (Seller's) Corporate Brand 37
Racing Prospect (Product) Brand 38
Converting Brand Attitudes into Brand Equity 40
III. METHODOLOGY 42
Structural Model 43
Research Hypotheses 43
Research Method 45
Unit of Analysis 47
Chapter Page
Sample
Pilot Study: Developing and Screening Scale Measures 49
Scale Development & Measurement 50
Buyer Attitude toward the Horse (Product Brand) 54
Buyer Attitude toward the Consignor (Seller Brand) 55
Buyer-Consignor (Seller) Relationship QuaUty 56
Satisfaction 57
Trust 58
Commitment 59
Consignor (Seller) Brand Equity 60
Main Study 63
Survey Response Rate 65
Data Analysis Approach: Structural Equation Modeling (SEM) 67
Summary 68
IV. DATA ANALYSIS AND RESULTS 69
Developing and Refining the Measurement Model 70
Descriptive Statistics 70
Response Bias Checks 71
Preliminary Factor Analyses 72
Refining the Construct Measures 73
Building the Measurement Model 75
Base and Comparison Measurement Models 76
Measurement Model Revision 1 78
Measurement Model Revision 2 80
Measurement Model Revision 3 81
Construct Reliability and Yahdity 83
Construct Reliability 84
Discriminant Validity 84
Convergent Validity 85
Arriving at the Final Measurement Model 86
Confirming the Measurement Model 87
Chapter
Buyer Attitude toward the Horse - Three Item Measures for
ATH 87
Buyer Attitude toward the Consignor (Seller) - Two Item
Measures for ATC 87
Satisfaction - Three Item Measures for SATIS 88
Trust - Two Item Measures for TRST 88
Commitment - Four Item Measures for COM 88
Relationship Quality - Nine Item Measures for RQual 88
Consignor (Seller) Brand Equity - Three Item Measures for
BrEq 89
Hypothesis Testing 89
Hypothesis 1: Supported 93
Hypothesis 2: Supported 93
Hypothesis 3: Not Supported 93
Hypothesis 4: Not Supported 93
Hypothesis 5: Supported 94
Hypothesis 6: Supported 94
Hypothesis 7: Supported 94
Summary 94
V. CONCLUSIONS AND CONTRIBUTIONS 97
Conclusions 98
Findings and Implications 100
Hypothesis 1 100
Hypothesis 2 101
Hypothesis 3 103
Hypothesis 4 104
Hypothesis 5 106
Hypothesis 6 107
Hypothesis 7 108
Findings and Implications Summary 109
Theoretical Contributions 111
Managerial Contributions 113
Chapter
Study Limitations
Future Research
APPENDICES 120
A. SURVEY INSTRUMENT 121
B. COYER LETTER FOR SURVEY MAILING 130
C. REMINDER AND THANK YOU POSTCARD 132
D. FOLLOW UP LETTER FOR SURVEY MAILING 134
E. RESPONSE RATE CALCULATIONS 136
F. DESCRIPTIVE STATISTICS 138
G. INITIAL MEASUREMENT MODEL MODIHCATION INDICES 147
H. HNAL MEASUREMENT MODEL AND RESULTS 152
I. RELIABILITY TESTS 165
J. DISCRIMINANT AND CONVERGENT VALIDITY TESTS 168
K. FACTOR CORRELATIONS 177
L. MEDIATION TEST RESULTS 179
REFERENCES 1^^
LIST OF HGURES
Figure
2.1 Theoretical Model and Postulated Hypotheses ..
2.2 Hierarchy of Brands
2.3 Brand Tangibility Continuum
2.4 Customer Relationship Continuum
2.5 Thoroughbred Consignment Conceptual Model
3.1 Theory Grounded Structural Model
4.1 Full Structural Model
4.2 Final Model
LIST OF TABLES
2.1 Relationship Quality Dimensions 23
2.2 Glossary of Thoroughbred Terminology 33
3.1 Items Measuring Buyer Attitude toward the Horse (Product Brand) 54
3.2 Items Measuring Buyer Attitude toward the Consignor (Seller Brand) .... 56
3.3 Items Measuring Satisfaction 57
3.4 Items Measuring Trust 59
3.5 Items Measuring Commitment 60
3.6 Items Measuring Consignor (Seller) Brand Equity 62
3.7 Non-Applicability Rationale Provided by Potential Respondents 66
4.1 Measurement Item Coding Summary 77
4.2 Construct Factor Loadings and Measurement Properties 82
4.3 Comparison of Key Measurement Model Fit Statistics 86
4.4 Comparison of Mediation Results 92
CHAPTER I
INTRODUCTION
Firms that leverage resourees in ways that create superior customer value are
more likely to develop advantages relative to competitors (Barney 1996, 1991; Hunt
1997, Srivastava, Shervani and Fahey 1999; Wemerfelt 1984). Although firm resourees
have traditionally been thought of in a tangible sense, some of a firm's most important
resourees are intangible in nature (Barney 1996; Berry 2000; Hunt 1997; Keller 2003).
Within this study the author explores the relationships among three firm-enabling
resources that demonstrate significant levels of intangibility within business-to-business
(B2B) contexts where products demonstrate aspects of both goods and services: 1)
sellers' corporate brands, 2) sellers' product brands, and 3) sellers' customer
relationships.
Existing literature supports the position that sellers' corporate and product brands
are valuable, enabling resources that help to build and reinforce firm-level competitive
advantages (Aaker 2004, 1996; Keller 2003). Corporate and product brands act as
important signaling mechanisms that convey critical and differentiating information to
current and prospective customers concerning a firm's good and service offerings (Keller
2003; Kirmani and Rao 2000; Park, Jun and Shocker 1996). Such differentiation serves
to reinforce brand positions, rewarding firms with perfonnance, growth and profit (Aaker
2004, 1996; Keller 2003; Drucker 1954; Penrose 1959).
Existing research also supports the notion that firms' critical resources and
sources of competitive advantage are not necessarily proprietary, but rather may span
firm boundaries (Davis, Golicic and Marquardt 2007; Dyer and Singh 1998; Hunt 1997;
Johnson and Seines 2004). Consequently, huyer-seller relationships are recognized as
valuable, enabling resources that facilitate positive customer response and superior
performance outcomes (Dyer and Singh 1998; Hunt 1997; Subramani and Venkatraman
2003). These attributions of customer value refer to the concept of brand equity, which
occurs when customers react more favorably to sellers' product offerings and the way
they are marketed when the brands are identified than when they are not (Aaker 2004,
1996; Keller 2003, 1993).
Research Opportunity
Relationship marketing and customer relationship management( CRM)a re
proximal areas that, in spite of the increased attention they have received in recent years,
remain compelling areas of research (Agrawal 2003; Eng 2004; Johnson and Seines
2004; Zablah, Bellenger and Johnston 2004). In spite of widespread embracement
regarding the role of different types of customer relationships (Garbarino and Johnson
1999; Hunt 1997; Stringefellow, Nie and Bowen 2004), there is no consensus on how
best to organize and leverage a firm's relationships (Agrawal 2003; Ang and Taylor
2005; Hunt 1997; Johnson and Seines 2004; Ryals 2003; Skaates and Seppanen 2005).
One of the key coneepts that has drawn attention out of the relationship marketing
and customer relationship management literatures is buyer-seller relationship quality
(Garbarino and Johnson 1999, Huntley 2006; Johnson and Seines 2004; Kim and Chan
Olmsted 2005; Ulaga and Eggert 2006). Buyer-seller relationship quality describes the
degree to which buyers are satisfied with the overall relationship with a seller, as
manifested through the buyer's overall satisfaction, trust and commitment toward the
seller (Johnson and Seines 2004; Ulaga and Eggert 2006). Consequently, as buyer-seller
relationship quality improves, firm-level competitive advantages are gained and
strengthened over time, thereby improving finn perfonnance outcomes and seller brand
equity accrual (Huntley 2006; Johnson and Seines 2004; Kim and Chan-Olmsted 2005;
Ulaga and Eggert 2006).
In spite of the "preponderance" of fimis selling in business-to-business contexts,
"virtually all discussions of branding" have been framed in consumer goods contexts
(Webster and Keller 2004, pg. 388). Consequently, although there are far more dollars
spent in business-to-business( B2B) markets than in business-to-consumer( B2C) markets
(Armstrong and Kotler 2005), industrial branding is far less developed and understood
than is consumer branding (Webster and Keller 2004). By extension and not
surprisingly, branding in business-to-business contexts where products demonstrate
aspects of both goods and services (i.e., hybrid products) is also far less developed and
understood than is branding in hybrid business-to-consumer contexts (Berry 2000; Keller
2003; McDonald, de Chernatony and Harris 2001; Webster and Keller 2004).
As a result, in spite of the growing attention that business-to-business branding
and buyer-seller relationship quality have generated within the popular press and
academic circles, their complementary roles in contributing to the building of seller brand
equity remain a promising research area. This study seeks to address the prominent gap
in the current relationship marketing literature concerning the concept of buyer-seller
relationship quality. This study examines the concept of business-to-business buyer-
seller relationship quality and its role in influencing seller brand equity. This study
thereby serves to address the current need for conceptual and empirical studies of this
important topic. Consequently, the research question that is asked within this study is:
"How does buyer-seller relationship quality mediate the relationships between buyer
attitude toward sellers' corporate and product brands, and seller brand equity?" 
Contributions
This research contributes to the existing body of knowledge in multiple ways.
First, this study will help provide insight into how buyer-seller relationship quality
influences the relationship between buyer brand attitudes and seller brand equity.
Second, tbis research will belp extend our knowledge of business-to-business branding,
particularly as it relates to branding speculative products that exhibit aspects of both
goods and services. This research thus helps to illuminate the opportunity to build brand
equity in contexts exhibiting this dual nature, especially when the product offerings are
highly uncertainty (for example, graduating Ph.D. candidates, college and professional
A
athletes, prospective university students, venture capital, real estate, high-end wine and
artistic offerings).
A significant portion of the evaluation of these product offering examples is based
on an envisioned future outcome associated with the product offering. For example, the
evaluation and hiring of newly-minted Marketing Ph.D.s typically occurs following the
candidates third or forth years, and usually six to twelve months ahead of when they will
be granted their degrees. Additionally, most Ph.D. candidates (product offerings) have
had only a limited chance to make their marks on the marketing field; consequently, the
hiring of ABDs is speculative. The hiring institution is projecting aspects of the
Marketing Ph.D. candidate into the future. These include issues related to current status
(i.e., what does the candidate's academic and professional pedigree look like, has he/ she
completed his/ her comprehensive/ qualifying exams, defended his/ her dissertation
proposal, collected dissertation data) and long-range potential (i.e., how good is his/ her
dissertation idea, who are his/ her committee members and chair, what else is he/ she
working on, who else is he/ she working with, what type of teacher and/ or researcher
will this candidate be). These evaluation criteria are subjective in nature, and hence the
evaluation of candidates will vary based on what the prospective hiring institutions think
of the Ph.D. candidate's potential to be successful at that institution. The same principles
apply to the evaluation of numerous other product offerings, including the ones listed
above, which also demonstrate the current evaluation of the product offering based on
uncertain and speculative factors.
Third, this study is designed with the intent of theory elaboration - i.e., to apply
and extend the existing theory within new settings or contexts (Lee 1999). The
Thoroughbred industry is a new contextual area for business research, as the limited
numbers of studies that have been published related to Thoroughbreds have almost
universally involved gambling rationale and decision-making process research. The use
of theory elaboration is thus an important contribution on three fronts. First, it
encourages the application of existing branding and relationship quality theory in order to
further our knowledge of a developing, but currently incomplete, concept. Second,
theory elaboration opens the door for research in a novel and exciting contextual area,
i.e., the Thoroughbred industry. Finally, the use of theory elaboration is an important
contribution because it provides support for a valuable but relatively overlooked and
underutilized research technique (Lee 1999). The approach was valuable within this
study because it encouraged the application and extension of existing brand attitude,
brand equity and relationship marketing concepts, in order to develop a theory-grounded
conceptual framework that will enhance our future understanding of the buyer-seller
relationship quality construct and how it influences the pathways between buyer attitudes
toward the seller's corporate and product brands, and seller brand equity.
Finally, this research helps to provide both the impetus and the roadmap for B2B
firms to understand better the mediating role of buyer-seller relationship quality between
buyer attitudes toward sellers' corporate and product brands, and seller brand equity.
This practical insight will help B2B firms, and particularly sellers within the
Thoroughbred consignment industry, to understand better how the dimensions of buyer
seller relationship quality (satisfaction, trust and commitment) collectively contribute to
enhancing seller brand equity. This research thus affords the opportunity to provide
several theoretical and applied contributions.
Document Framework
The framework for the remainder of the document is as follows. Chapter II is
entitled Building the Theory. Chapter II develops and presents a literature review that is
designed to build the theory underlying the theory-grounded conceptual framework that
is developed and presented within the chapter. The focus of Chapter II is on the four
model constnacts, buyer attitude toward the seller's corporate brand, buyer attitude
toward the seller's product brand, buyer-seller relationship quality and seller brand
equity, and how these constructs have been discussed in the existing relationship
marketing and branding literatures. The relationships among the four constructs are laid
out, paving the way to be able to test these relationships.
Chapter III is entitled Methodology and describes the methodological approach
that was used in this study. The intent of the design was theory elaboration - i.e., to
apply and extend existing theory into new settings or contexts (Lee 1999). This study
utilized expert review as a pilot study pretest effort designed to develop an appropriate
survey instrument by establishing survey question face validity and wording through
expert opinion (Dillman 2000; Elsbach 1994). Following the development of the survey
instrument, a mail survey approach was used to collect data from business-to-business
buyers of Thoroughbred racing prospects (product brands). A review of the data-
collection steps and processes and a brief summary of the data collection effort are also
presented.
Chapter IV is entitled Data Analysis and Results. Collected data are analyzed
using conventional exploratory data analysis tools in SPSS (Version 13.0) and structural
equation modeling tools in AMOS (Version 5.0). Results indicate significant direct
positive relationships between buyer attitude toward the seller's corporate brand and
buyer-seller relationship quality, buyer attitude toward the seller's product brand and
buyer-seller relationship quality, and buyer-seller relationship quality and seller brand
equity, as well as a direct negative relationship between buyer attitude toward the seller's
product brand and seller brand equity. The results also indicate that buyer-seller
relationship quality fiilly mediates the positive direct path between buyer attitude toward
the seller and seller brand equity, and partially mediates the negative direct path between
buyer attitude toward the product and seller brand equity. The finding that buyer attitude
toward the product is inversely predictive of seller brand equity in this context is
particularly interesting. This finding suggests that in spite of the expense and uncertainty
attached to the purchase of a Thoroughbred racing prospect (product brand) the value the
buyer ascribes to the consignor (seller) is marginalized when the buyer has a more
favorable attitude toward the horse. The finding that buyer-seller relationship quality
partially mediates this path is also very interesting. This finding implies that as the buyer
demonstrates a less favorable attitude toward the racing prospect (product brand) the
value attributed to the consignor (seller) increases.
Chapter V is entitled Conclusions and Contributions. This chapter explores the
findings presented in Chapter IV in more detail. Specifically, this chapter serves to
explore the theoretical and managerial implications of the findings of this research,
discussing the role of buyer-seller relationship quality in business-to-business contexts
where products demonstrate aspects of both goods and services. The findings from this
research have significant implications for sellers within speculative and competitive
business climates. Chapter V also presents and discusses the limitations associated with
this research, as well presents recommendations for future research.
CHAPTER II
BUILDING THE THEORY
Chapter II begins with a literature review that is designed to build the theory-
grounded framework that is developed and presented within this chapter. The presented
literature review provides a discussion of the study's theoretical framework, and draws a
variety of important concepts from both the marketing and management strategy
literatures. The foundation for this research is resource-advantage theory. Resource-
advantage theory focuses on the resources available to a finn (Hunt 1997; Hunt and
Morgan 1995), and provides a theoretical grounding platform for three important firm
resources: 1) sellers' corporate brands, 2) sellers' product brands, and 3) sellers' customer
relationships.
The approach that is adopted within the literature review begins with the
introduction of the theoretical model used within this study. Next, this study's theoretical
grounding platform, resource-advantage theory, is outlined and discussed. Following the
discussion of resource-advantage theory, the model's three branding constructs: 1) buyer
attitude toward the seller's corporate brand, 2) buyer attitude toward the seller's product
brand, and 3) seller brand equity, are presented. The theoretical basis for each is
provided, as well as the relationships that exist between and among these constructs. The
review of the model's constructs is completed with a discussion of the model's fourth and
final construct, buyer-seller relationship quality. As was done with the three branding
constructs, the theoretical basis of buyer-seller relationship quality is outlined and
discussed, specifically focusing on the three dimensions that comprise the construct,
satisfaction, trust and commitment. The literature review concludes with a discussion of
the theoretical justification for the relationships that exist between the model's four
constructs, thereby providing the rationale for the hypotheses that were posed for this
research.
Theoretical Model
Existing literature affirms that buyer attitude toward sellers' corporate and
product brands varies between and across customer relationships (Garbarino and
Johnson; Nowak, Thach and Olsen 2006) and that the quality of buyer-seller relationships
is an important determinant of a firm's brand perfonnance outcomes (Garbarino and
Johnson; Johnson and Seines; Nowak, Thach and Olsen 2006; Ulaga and Eggert 2006).
Consequently, buyer-seller relationship quality influences the relationship between buyer
attitudes toward sellers' corporate and product brands, and seller brand equity. These
sentiments are articulated within the postulated construct relationships depicted in the
theory-grounded framework presented in Figure 2.1, and they are justified within the
literature review that follows.
Figure 2.1: Theoretical Model and Postulated Hypotheses
Buyer
Attitude
toward the
Seller's
Corporate
Brand
Buyer-Seller
Relationship Seller's Brand
Quality Equity
Buyer
Attitude
toward the
Seller's
Direct Relationships:
Product
Brand Mediating Relationships:
Adapted from Garbarino and Johnson 1999; Keller 1993
Resource-Advantage Theory
Resource-advantage theory focuses on the resourees available to the firm, thereby
helping to explain from a resouree perspective how brands are built and why one
corporate or product brand is able to outperform another (Hunt 1997; Hunt and Morgan
1995). Brand performance is eontingent upon and determined by the firm's use of
available resources (Hunt and Morgan 1995; Penrose 1959; Wemerfelt 1984). Resourees
that are difficult for competitors to imitate provide the firm with advantages relative to
competitors (Barney 1996, 1991; Hunt 1997; Hunt and Amett 2003; Johnson and Seines
2004; Teece, Pisano and Shuen 1997). Sellers' corporate brands, product brands and
customer relationships are three such enabling resources.
The cornerstone of resource-advantage theory is that "social structures and trust-
based governance can be competition enhancing"( Hunt and Arnett 2003, p. 1). This
tenet underlies the argument that "firms should develop a relationship portfolio that is
comprised of relationships that constitute relational resources" in order to build and
leverage their brand offerings (Hunt 1997, p. 431). Hunt suggests resource-advantage
theory is a logical grounding platform because "it expands the view of resources to
include all entities that have an enabling capacity"( 1997, p. 441). Such a view
encourages thinking beyond tangible resources, to include intangible entities such as
corporate brands, product brands, and customer relationships. Such resources provide
firms with competitive advantages, which foster superior performance levels (Hunt 1997;
Hunt and Amett 2003; Hunt and Morgan 1995).
Resource-advantage theory is in itself a relatively nascent theoretical platform
(formally proposed in the mid-nineties), and its focus on resource heterogeneity and
imperfect mobility across the financial, physical, legal, human, organizational,
infonnational and relational areas of business has helped it gain support in both
marketing and non-marketing circles (Hodgson 2000; Hunt and Amett 2003; Savitt 2000;
Schlegelmileh 2002). In utilizing resource-advantage theory as a grounding mechanism,
sellers' corporate brands, product brands, and customer relationships are recognized as
important signaling resources that affect buyer assessment and behavior, and as a result
firm perfomiance (Garbarino and Johnson 1999; Hunt and Morgan 1995; Johnson and
Seines 2004; Park, Jun and Shocker 1996).
Resource-advantage theory is closely aligned with the idea that the goal of any
firm should be to create superior value for its customers (Drucker 1954; Levitt 1960).
Firms strive to create superior value for their customers by developing and employing the
right combination of resources (Barney 1996, 1991; Kogut and Zander 1992; Penrose
1959). Sellers' corporate brands, product brands and customer relationships thus serve as
valuable firm resources which help to create favorable positions in the minds of
customers and superior differentiated positions relative to competitors.
Brand Attitudes
A brand attitude conveys the extent to which the seller has been able to create a
level of connection with an actual or prospective buyer( Kim and Chan-Olmsted 2005;
Keller 2003). A brand attitude essentially reflects the actual or prospective buyer's
feelings toward a particular brand, and it is formally defined as a predisposition to
respond in a consistently favorable or unfavorable manner towards a particular brand
(Ajzen and Fishbein 1980; Kim and Chan-Olmsted 2005). A brand, in turn, is defined as
"a name, term, sign, symbol, or design, or combination of them, which is intended to
identify the goods and services of one seller or group of sellers and to differentiate them
from those of competitors"( Armstrong and Kotler 2005, p. 234).
Buyer attitudes toward brands are generally arrived at as a function of their
attitudes toward the brand's multiple attributes and benefits (Ajzen and Fishbein 1980;
Keller 2003, 1993). These attributes and benefits differ in terms of their respective
importance; however, in aggregate they help to form buyers' attitudes toward speeifie
brands (Ajzen and Fishbein 1980; Myers 2003). Brand attitudes thus serve to reflect
buyer evaluations of brand attributes and benefits, as well as their overall evaluations of
the brand, thereby forming the basis for buyer appraisal and choice (Ajzen and Fishbein
1980; Keller 2003; Kim and Chan-Olmsted 2005; Wilkie 1980).
Communicating the brand's appeal to the right group of prospective customers
(i.e., the firm's target market) is critical and is the essence of branding (Armstrong and
Kotler 2005; Keller and Kotler 2005). Branding has received considerable attention in
mature business-to-consumer contexts (Berry 2000; Keller 2003), and is now generating
increased attention in business-to-business contexts (Marquardt, Golicic and Davis 2007;
Webster and Keller 2004). In spite of this attention and the widespread acceptance of
branding as a valuable source of competitive advantage (Aaker 1996; Berry 2000; Keller
2005), business-to-business corporate and product brand elaboration remains relatively
underdeveloped in the literature (Aaker 2004; Gordon, Calantone and di Benedetto 1993;
Webster and Keller 2004).
Corporate Brand
The seller's corporate brand is the highest level of brand within an organization's
hierarchy of brands and serves to define the organization that will deliver and stand
behind the product (good and/ or service) offering (Aaker 2004, 1996; Keller 2003). In
this sense the corporate brand serves in an umbrella capacity (Erdem 1998), conveying a
brand promise for the more specifically-defined and targeted brands that occur in
subsequent (lower) levels of the hierarchy. The brand hierarchy concept is visually
depicted in the General Motors' hierarchy of brands illustration that follows in Figure 2.2.
Figure 2.2: General Motors' Hierarchy of Brands in 2001
Corporate General
Brand Motors
Family
Oldsmobile
Brand
'  Trucks ^ Carfl
Silverado/C1500/ Cavalier
(Individual) Grand PrixK1500/C/K/ Camaro Alero
Grand Am Century
S-10 Pickups Prizm Cutlass
Product Bonneville RegalTahoe Malibu Intrigue
Trans Sport LeSabre
Venture Monte Carlo Eighty-eight
Brand Firebird/Trans Am Park AvenueSuburban Lumina Aurora
Sunfire
Astro Metro
Brand Modifiers for Each Individual
Brand
Modifier Product Brand (ex. Z06 Model of the
Chevrolet Corvette)
Adapted from Keller 2003
Based on the illustration presented within Figure 2.2, the following General
Motors' hierarchy of brands can be witnessed: 1) Corporate Brand - e.g., General Motors;
2) Family Brand - e.g., Chevrolet; 3)( Individual) Product Brand - e.g.. Corvette; 4)
Brand Modifier - e.g., Z06. This study focuses on two seller brands, corporate and
product. As a result, the focus is on the first and third levels of the brand hierarchy - or in
the case of the hierarchy of brands illustration described immediately above, the
corporate brand is General Motors, while the product brand is the Corvette.
The definitive goal of the corporate brand is to signal important infonnation and
added value to current and prospective customers such that they look to purchase/
repurchase product offerings from the firm at future points of time (Aaker 2004; de
Chernatony 2001; Harris and de Chematony 2001). The corporate brand thus symbolizes
the relational promise of what the firm will deliver and also by extension what the firm's
offerings will deliver (Aaker 2004; de Chernatony 2001; Marquardt, Golicic and Davis
2007). Because the corporate brand communicates a commitment to provide a certain
kind of experience, it is critical that the fimr convey the right message to customers
(Gombeski Jr., Kantor and Klein 2002).
The corporate brand "has access to organizational as well as product associations"
(Aaker 2004 p. 6), thus it both drives and is driven by buyer attitudes toward the firm's
product brands. Consequently, the corporate brand "explicitly and unambiguously
represents an organization as well as a product"( Aaker 2004 p. 10). As a result, the
corporate brand serves to facilitate and enhance the product brand-building process
(Aaker 1996; Erdem 1998; Schreiber 2002; Wemerfelt 1989).
Product Brand
The product brand is a single good and/ or service, assigned its own individual
name and accorded its own individual positioning (Kapferer 1992; Keller 2003). The
product brand refers to "anything that can be offered to a market for attention,
acquisition, use, or consumption that might satisfy a want or need"( Annstrong and
Kotler 2005, p. 223). Product brands are therefore not constrained to being purely
tangible, but can also exhibit high degi-ees of intangibility. In fact, while either extreme
is possible, most product brands exhibit both tangible and intangible aspects, and as a
result are really hybrid offers that may fall anywhere along the Brand Tangibility
Continuum illustrated in Figure 2.3.
Figure 2.3: Brand Tangibility Continuum
Purely Tangible Hybrid Purely Intangible
Goods Offers Services
Adapted from Armstrong and Kotler 2005
An additional factor that detennines where a product falls along the brand
tangibility continuum and that distinguishes how the produet needs to be treated, relates
to the product's brand attributes (Annstrong and Kotler 2005; Da Silva and Syed Alwi
2006; Hoffe, Lane and Nam 2003). Product brands possess both tangible attributes
which are funetional - for example, the fuel efficiency of a car, and intangible attributes,
which are more perceptual in nature - for example, the perceived safety or reliability of a
car( Da Silva and Syed Alwi 2006; Hoffe, Lane and Nam 2003). Produet brands that
exhibit high degrees of tangible attributes are usually perceived as less complex, risky
and uncertain than are brands that exhibit high degrees of intangible attributes.
Existing literature supports the notion that product brands that are less tangible in
nature, such as those possessing a significant service component, need to be handled
differently than do brands that are highly tangible (Annstrong and Kotler 2005; Berry
2000; Gombeski, Kantor and Klein 2002; Harris 2002; Krishnan and Hartline 2001). Of
central importance is the idea that the strength and the perfonnance of these brands are
largely driven by the people associated with the brands and by the quality of the buyer-
seller relationships (Berry 2000; Gordon, Calantone and di Benedetto 1993; Marquardt,
Golicic and Davis 2007; McDonald, de Chematony and Harris 2001). This idea is
particularly evident within B2B contexts (Davis, Golicic and Marquardt 2007; Subramani
and Venkatraman 2003), where purchase expenditures tend to be larger, and where buyer
involvement tends to be greater (Annstrong and Kotler 2005; Webster and Keller 2004).
In these settings, customer relationships become an even more important resource.
developing critical competitive advantages for the finn (Aaker 2004; Hunt 1997; Keller
2003). These valuable competitive advantages, along with the competitive advantages
derived via the seller's corporate brand, serve as the foundation for building brand equity
(Keller 2005, 2003).
Brand Equity
When a seller's competitive advantages are effectively leveraged, brand equity is
more likely to emerge (Beverland 2005; Dawar 2004; DelVecchio and Smith 2005;
Keller 2005; Rao and Monroe 1996). Brand equity occurs when customers have a high
level of awareness, familiarity and associations with a brand. Brand equity is fonnally
defined as the differential effect that brand knowledge has on customer response to the
marketing of that brand (Keller 2003). Brand equity is thus the differential effect arising
from everything a customer takes into account when looking to purchase a good or
service, including all the tangible and intangible attributes of the product offering.
Aaker (1996, 1991) suggests that brand equity is a multidimensional construct
consisting of brand loyalty, brand awareness, perceived quality, brand associations, and
other assets derived from the brand. Brand equity is generated over time( Nowak, Thach
and Olsen 2006), and it provides the firm with competitive advantages based on both
non-price factors (Aaker 1996, 1991; Keller 2003) and on the incremental monetary
value that the brand provides the product (Farquhar 1989).
Brand equity is typically manifested through a buyer's willingness to pay a
premium for the seller's product and/ or through the buyer's future intentions (Aaker
1996, 1991; Keller 2003, 1993). In a practical sense, brand equity can thus be thought of
as the additional value endowed to a product above and beyond what the customer would
pay for the same product from a neutral entity (Aaker 1996; Farquhar 1989; Hunt and
Morgan 1995; Park, Jun and Shocker 1996), or that results in a more favorable customer
response to that brand (Aaker 1996; Keller 2003).
Relationship Qualitv
A wide assortment of customer relationship terminology has been broached
within the relationship marketing literature, including relationship strength, closeness and
quality. Bove and Johnson (2001) do an excellent job discussing and dissecting the
subtleties of each. They conclude that strength is most appropriate when "describing the
magnitude of a relationship between two individuals" (i.e., a customer and a salesperson
or service employee), closeness is most appropriate when "describing romantic, friendly
or family relationships," and quality is most appropriate when discussing "buyer-seller"
relational dyads( Bove and Johnson, p. 195). Consequently, the use of buyer-seller
relationship quality within this research both follows and supports Bove and Johnson's
(2001) call for researchers to use the appropriate terminology when conducting studies
relating to the classifications and characterizations of different types of customer
relationships.
The management of different types of customer relationships continues to be a
popular topic in the marketing and management strategy literatures. It has garnered
considerable support as a means of creating superior competitive resource positions
(Boulding, Staelin, Ehret and Johnson 2005; Hunt 1997; Morgan and Hunt 1995). In
spite of this increased attention, our knowledge of huyer-seller relationship quality is still
far from complete (Huntley 2006; Johnson and Seines 2004; Ulaga and Eggert 2006).
Within this study, relationship quality is discussed from the buyer's perspective and is
defined as the degree to which buyers are content with the relationship they have with the
seller at a particular point of time, as manifested through the buyer's satisfaction, trust
and commitment toward the seller (Johnson and Seines 2004; Ulaga and Eggert 2006).
As the mix of satisfaction, trust and commitment changes, the quality of the buyer-seller
relationship changes, thereby influencing brand perfomiance outcomes (Bae and
Gargiulo 2004; Dyer and Singh 1998; Garharino and Johnson 1999; Gimeno 2004;
Leonidou, Barnes and Talias 2006). Consequently, firm-level competitive advantages are
gained and strengthened over time by fostering higher quality customer relationships
(Huntley 2006; Kim and Chan-Olmsted 2005).
The existing literature supports relationship quality as a second-order construct,
with each dimension of relationship quality contributing to the rationale behind customer
buying behavior( Bove and Johnson 2001; Dwyer and Oh 1987; Johnson and Seines
2004; Kim and Chan-Olmsted 2005; Ulaga and Eggert 2006). This translates to the
notion that satisfaction, trust and commitment individually and collectively contribute to
informing buyer product evaluation and purchasing behavior. As a result, the integrity of
relational resources is fortified as relationship quality improves (Hunt 1997; Johnson and
Seines 2004). This fortification process involves moving from highly transactional
activities in low quality relationships to highly collaborative activities in high quality
relationships (Garbarino and Johnson 1999).
It is important to note that while the constituent components of relationship
quality are closely correlated, satisfaction, trust and commitment are discreet dimensions.
This observation means that it is not prerequisite to have high buyer evaluations on one
of the dimensions (for example, trust), in order to have high buyer evaluations on another
dimension (for example, commitment), and that relationship quality is the byproduct of
an additive mix of the dimensions. These concepts are illustrated in Table 2.1 and in the
subsections that follow the table.
Table 2.1: Relationship Quality Dimensions
Relationship Quality Rationale for the Buying Activitj
Dimension
Satisfaction Satisfaction facilitates and reinforces the buying aetivity,
reducing the need to search for market infonnation
Trust Trust facilitates the buying activity in the absence of perfect
infonnation
Commitment Commitment in the form of information sharing improves
relationship perfonnance outcomes
Adapted IVom Johnson and Seines (2004)
Satisfaction
Satisfaction is defined as the buying firm's level of contentment based on its
personal and tangential purchase and consumption experiences with a selling fmn and the
selling firm's products over time (Anderson, Fomell and Lehmann 1994; Garbarino and
Johnson 1999). The rationale behind the satisfaction construct is to go beyond
transaction-specific encounters to capture the buyer's overall level of satisfaction with the
seller. This definition of satisfaction provides a more comprehensive view of the buyer-
seller relationship than does buyer satisfaction toward a transaction-specific experience,
because it accounts for buyer satisfaction across product offerings and/ or time.
The broad and encompassing nature of the satisfaction construct pemiits great
latitude in determining what influences a particular buyer's evaluation. Satisfaction can
extend beyond idiosyncratic buyer-seller relationship boundaries to be derived from not
only the customer's personal experiences with the seller, but also his or her tangential
experiences. As the buyer's satisfaction increases, the affective state he/ she feels toward
the seller and expectations for the relationship also increase( De Wulf, Odekerken-
Schroder and lacobucci 2001; Hon and Grunig 1999). Consequently, as the level of the
buying firm's satisfaction increases, the pereeived risk and need to search for infonnation
both decrease (Johnson and Seines 2004).
Trust
Trust is the second of the three constructs that constitute the higher order
relationship quality construct( Dwyer and Oh 1987; Johnson and Seines 2004; Kim and
Chan-Olmsted 2005; Ulaga and Eggert 2006). Two proximate, but slightly differing
definitions of trust fonn the basis for a large number of marketing studies involving trust.
The first of these was put forth by Moorman, Deshpande and Zaltman, who define trust
as "a willingness to rely on an exchange partner in whom one has confidence"( 1993 p.
82). They add that seller trustworthiness results from the seller's reliability, intentionality
and ability to perform (i.e., competence/expertise). The second commonly used
definition of trust was put forth by Morgan and Hunt( 1994 p. 23), who define trust as
one party's "confidence in the exchange partner's reliability and integrity."
The basic separation between these two conceptualizations of trust deals with the
behavioral intention of willingness utilized in Moorman, Deshpande and Zaltman's
definition (Morgan and Hunt 1994). Moorman, Deshpande and Zaltman (1993) argue
willingness is a critical facet of the trust conceptualization, while Morgan and Hunt
(1994) argue willingness is implicit and therefore redundant when included in the
definition. Having confidence in and expectations for the exchange partner is the key
facet of the trust conceptualization (Dwyer and Oh 1987), and a willingness to rely on the
exchange partner is a by-product of this confidence. This paper therefore utilizes Morgan
and Hunt's (1994) definition of" confidence in the exchange partner's reliability and
integrity" for the conceptualization of trust.
Trust thus refers to the level of expectation that the buyer feels towards the seller
and their products( Dwyer and Oh 1987), helping to minimize the complexities inherent
within the buyer's decision-making process by reducing their fear, uncertainty and doubt
(Schreiber 2002). Trust signals added value such that customers look to continue their
buying activities without perfect infonnation (Berry 2000; Johnson and Seines 2004;
Schreiber 2002) and is thus a valuable ingredient for building successful long-term
relationships (Dwyer and Oh 1987; Moorman, Deshpande and Zaltman's 1993; Morgan
and Hunt 1994; Ulaga and Eggert 2006).
Commitment
Commitment is the third construct that constitutes the higher order relationship
quality construct (Johnson and Seines 2004; Ulaga and Eggert 2006). Following
Moonnan, Zaltman and Deshpande (1992, pg. 316), commitment is defined as "an
enduring desire to maintain a valued relationship" and describes the degree to which the
buyer believes the relationship with a particular seller is worth committing resources to in
order to maintain and promote the relationship (Hon and Grunig 1999; Morgan and Hunt
1994).
Commitment is made significantly easier when firms are satisfied with and trust
each other (Bae and Gargiulo 2004; Dyer and Singh 1998; Garbarino and Johnson 1999;
Johnson and Seines 2004; Morgan and Hunt 1994); however, neither is prerequisite for
commitment to exist. Commitment to the relationship elicits better informational
exchanges between buyer and seller (Hunt 1997; Johnson and Seines 2004; Seines and
Sallis 2003), thereby serving to facilitate better relationship perfonnance outcomes. This
consequence is particularly true as it relates to creating economies of scale and reducing
buyer search and due diligence costs.
Commitment from the buyer firm's perspective is thus the implicit or explicit
pledge of relationship continuity( Dwyer, Schurr and Oh 1987), and is most frequently
demonstrated by committing resources to the relationship (Monczka, Petersen, Handheld,
and Ragatz 1998). In order for commitment to occur, the buyer must perceive some level
of intrinsic value related to their perceptions of involvement with the seller and/ or the
seller's brands. Levels of commitment vary markedly across relationships, with better
quality relationships typically reflecting higher levels of commitment (Johnson and
Seines 2004; Kim and Chan-Olmsted 2005; Ulaga and Eggert 2006).
Relational Resources
Firms that recognize the critical opportunities afforded by relational resources are
more likely to benefit (Davis, Golieic and Marquardt 2007; Hunt 1997; Hunt and Amett
2003; Marquardt, Golicie and Davis 2007). While the notion of relationships as
resources is critical, not all customer relationships are equal. As a result buyer-seller
relationships are thought to fall along a continuum (Garbarino and Johnson 1999;
Webster 1992), with locations along the continuum representing differing levels of
relationship quality (Dwyer, Sehurr and Oh 1987; Garbarino and Johnson 1999; Johnson
and Seines 2004; Ulaga and Eggert). Higher points on the continuum represent higher
quality relationships, while lower levels on the continuum represent lower quality
relationships. This concept is illustrated immediately below in Figure 2.4.
Figure 2.4: Customer Relationship Continuum
Discreet Buyer-Seller
Transactions Partnerships
Lower Quality Higher Quality
Relationships Relationships
Adapted from Garbarino and Johnson 1999; Jotinson and Seines 2004; Webster 1992
The quality of buyer-seller relationships is not static (Huntley 2006; Kim and
Chan-Olmsted). This idea suggests that some huyer-seller relationships reflect discreet
transactional convenience, while others reflect high-involvement relational bonds, and
that buyers may move up or down the relational continuum (Garbarino and Johnson
1999; Johnson and Seines 2004). As huyer-seller relationship quality improves, so does
the opportunity for partner infonnational exchanges and relational learning (Anand and
Khanna 2000; Johnson and Seines 2004; Seines and Sallis 2003; Stuart 2000). As a
result, customer relationships differ in terms of their enabling capacity (Hunt 1997),
serving as variable hut valuable resources within the hrand-huilding process (Davis,
Golieie and Marquardt 2007; Seines and Sallis 2003; Marquardt, Golieie and Davis
2007).
Building high quality, mutually beneficial relationships helps stimulate customer
advocacy (Prahalad and Ramaswamy 2000) and is an effective branding strategy within
competitive and complex industries (Keller 2005). High quality relationships are
manifested through higher combinations of satisfaction, trust and commitment, which
have been shown to be individually and collectively predictive of superior performance
outcomes (Garbarino and Johnson 1999; Johnson and Seines 2004; Kim and Chan-
Olmsted 2005). Park and Russo (1996) posit that higher quality relationships expand the
asset positions of each relational partner, promoting reciprocity and benefiting each.
Anand and Khanna (2000) and Stuart (2000) reinforce this view through their respective
discussions of the benefits associated with the transfer of tacit knowledge that occurs as
relationship quality improves. However, in spite of the growing coverage that buyer-
seller relationship quality has received in the relationship marketing and customer
relationship management literatures, our knowledge of its role in influencing
perfonnance outcomes remains profoundly underdeveloped (Garbarino and Johnson
1999; Huntley 2006; Johnson and Seines 2004; Kim and Chan-Olmsted 2005).
Research Hypotheses
Brands play an important signaling role (Aaker 1996; Berry 2000; Dawar 1998;
Kirmani and Rao 2000), thereby influencing buyer attitudes toward sellers' corporate and
product brands, buyer-seller relationship quality, and seller brand equity (Davis, Golicic
and Marquardt 2007; Hunt and Morgan 1995; Huntley 2006; Keller 2003; Kim and
Chan-Olmsted 2005). Existing research suggests that buyer-seller relationship quality
increases as buyer attitudes toward sellers' corporate and product brands increase
(Johnson and Seines 2004), and empirical research affimrs the direct effects buyer
attitudes have on the buyer-seller relationship quality dimensions of satisfaction, trust and
commitment within the business-to-consumer Broadway theater context (Garbarino and
Johnson 1999) and within business-to-business manufacturing contexts (Ulaga and
Eggert 2006). Hypotheses HI and H2 are offered as an extension of these previous
studies, testing buyer-seller relationship quality as an aggregate construct within a new
business-to-business context, i.e., the thoroughbred consignment industry.
HI) Buyer attitude toward the seller's corporate brand has a positive direct
effect on buyer-seller relationship quality.
H2) Buyer attitude toward the seller's product brand has a positive direct
effect on buyer-seller relationship quality.
Buyers' decision-making and choice processes are guided by their attitudes
toward sellers' corporate and product brands (Aaker 2004, 1996; Keller 2003). Empirical
research supports the notion that buyer attitudes toward sellers' corporate and product
brands directly effect brand equity (Chaudhuri 1999; Netemeyer, Krishnan, Pullig, Wang,
Yagci, Dean, Richs, Wirth 2004; Subrahmanyan 2004). This line of thinking provides
the foundation for offering hypotheses H3 and H4 which test the concept of seller brand
equity within a new business-to-business context, i.e., the thoroughbred consignment
industry.
Hi) Buyer attitude toward the seller's corporate brand has a positive direct
effect on seller brand equity.
H4) Buyer attitude toward the seller's product brand has a positive direct
effect on seller brand equity.
Buyer-seller relationship quality has a significant effect on finu performance
(Johnson and Seines 2004; Ulaga and Eggert 2006). The three dimensions of buyer-seller
relationship quality, satisfaction, tnast and commitment have been empirically
demonstrated as important predictors of relational outcomes such as buyer behavior,
buyer willingness to pay a premium for a product, future purchasing intentions,
propensity to leave a relationship, and seller market share (Chaudhuri and Holbrook
2001; Garbarino and Johnson 1999; Johnson and Seines 2004; Ulaga and Eggert 2006).
Hypothesis H5 is offered as an extension of these previous studies, testing buyer-seller
relationship quality and seller brand equity as aggregate constructs within a new
business-to-business context, i.e., the thoroughbred consignment industry.
H5) Buyer-seller relationship quality has a positive direct effect on seller
brand equity.
The three dimensions of buyer-seller relationship quality, satisfaction, trust and
commitment have also been demonstrated to serve in a mediating capacity between
predictor and criterion variables (Garbarino and Johnson 1999; Johnson and Seines 2004;
Morgan and Hunt 1994; Ulaga and Eggert 2006). Predictor variables can include
antecedents such as buyer attitudes toward the seller's corporate and product brands, and
buyer perceptions of relationship benefits and costs, while criterion variables can include
outcomes such as buyer behavior, buyer willingness to pay a premium for a product,
buyer attributions of seller values, and future purchasing intentions (Garbarino and
Johnson 1999; Johnson and Seines 2004; Morgan and Hunt 1994; Ulaga and Eggert
2006). Hypotheses H6 and H7 are offered as extensions of these previous studies, testing
the mediating role of buyer-seller relationship quality between buyer brand attitudes and
seller brand equity in a new business-to-business context, i.e., the thoroughbred
consigmnent industry.
H6) Buyer-seller relationship quality mediates the effect ofb uyer attitude
toward the seller's corporate brand on seller brand equity.
H7) Buyer-seller relationship quality mediates the effect ofb uyer attitude
toward the seller's product brand on seller brand equity.
Research Context, Description and Application
This research draws on the Thoroughbred industry as its business-to-business
context. The Thoroughbred industry possesses its own jargon; consequently, a number of
the terms used within this paper might be confusing to those readers from outside the
industry. Table 2.2 introduces and describes some of these terms. Also, note that the
word Thoroughbred refers to a specific breed of horse and is thus a proper noun.
Table 2.2: Glossary of Thoroughbred Terminology
Thoroughbred Tenn Conceptual Description
2-year-old-in-training 2-year-old horse that is in training to race. Typically sold
between February and May of their freshman (2-year old) season
Breeding cross The genetic lineage on both the sire and dam sides
Claiming Races Classification of race where entered horses can be purchased by
another party
Colt Male offspring (up to 4-years old)
Conformation The physical structure and alignment of a horse
Consignor Business entity representing a for-sale horse
Dam Mother horse
Family line Historical lineage on the sire or dam sides
Filly Female offspring (up to 4-years old)
Lineage Prior generations of a horse's family on both the sire and dam
sides
Mare-and-foal Mother and a baby horse
Mare-in-foal Pregnant female horse
Older horse Horse that is 5-years old or older
Pinhooker Specialty consignor that purchases racing prospects at one point
of time with the intent of reselling them at another point of time
Progeny A horse's offspring
Racing prospect Horse that is targeted for or owned with the intent of racing
Sire Father horse
Split Time it takes a racing prospect to run a predetemiined distance
Stud Male horse that is in demand to produce future generations
Suckling Baby horse that is too young to be separated from its mother
Weanling Baby horse that has been separated from its mother
Yearling 1 -year-old colt or fdly that is typically sold between September
and November of the year following the year it was bom
* It should be noted that regardless oft he horse's actual date ofb irth, all Thoroughbreds
share a January f birthdayf or sales and race classification purposes
The Thoroughbred industry provides a compelling research context because it is
an academically novel and an informational-rich B2B context in which to study the
interplay of buyer brand attitudes, buyer-seller relationship quality, and seller brand
equity. Within the industry, there are a number of ways horses are bought and sold.
These ways include sale through a wide array of targeted and inclusive auctions
(Weanling, Yearling 2-Year-Old-in-Training, Racing Horses, Broodmare, Broodmare-in-
Foal, Broodmare-and-Foal), private contract, and claiming races. This study focuses on
Thoroughbred racing prospects (product brands) sold through sales auctions by a selling
entity that has consigned them to the sale. These selling entities are in business to sell
their own racing prospects (product brands), the racing prospects (product brands) of
others, or some combination or the two, and are referred to as consignors (sellers).
The Thoroughbred consignment industry is an excellent environment in which to
study husiness-to-husiness relationships involving the sale of products exhibiting aspects
of both goods and services. While it may he intuitive to think of Thoroughbred racing
prospects (product brands) as being highly tangible, Thoroughbred racing prospects
(product brands) also demonstrate a high degree of intangibility, due in large part to the
speculative, future-based nature of the racing business. Consequently, there are
numerous specialists within the Thoroughbred industry, who offer their own toolkit of
distinctive capabilities. Consignors (sellers) are one such group of industry specialists.
The level of distinctive capabilities varies widely among consignors (sellers),
ranging from trivial involvement where the consignor (seller) is simply a listing agent or
storefront with no-to-minimal knowledge of the consigned racing prospect (product
brand), to full involvement where the consignor (seller) has extensive familiarity with,
and tacit knowledge of, the racing prospect (product brand). As a result, there is a widely
varying service component inherent within the sale of Thoroughbred racing prospects
(product brands), and these racing prospects (product brands) exhibit aspects of both
goods and services.
The second reason the Thoroughbred consignment industry is such an intriguing
area is that outside of a trivial amount of research dealing with pari-mutuel gambling
rationale and decision-making processes, there have been no known marketing studies
done in the Thoroughbred context. The Thoroughbred industry as a whole is a
microcosm of business-to-business service activity. It has a rich history not only within
the United States where it was once the country's most popular sport, but also around the
world. This pervasiveness makes it a compelling context to explore a wide range of topic
areas in an international business-to-business context possessing some of the world's
greatest business minds.
The third reason this context is such an appealing one deals with the inherently
complex, risky and speculative nature of the Thoroughbred industry. The gestational
horizon for a racing prospect (product brand) is often substantially distant, and the
temporal and capital investments substantially high. As a result, racing prospect (product
brand) evaluation is a critical antecedent to success in the Thoroughbred racing industry.
Racing prospects (product brands) are also complex in that they possess a number of
attributes, which may or may not be important to each individual evaluator. As a result.
the dynamics between buyers and sellers in the foiTn of the relationship quality that exists
between them is partieularly compelling, especially as relates to the building of consignor
(seller) brand equity.
Thoroughbred Consignment
Thoroughbred consignment is the process of offering horses (product brands) for
sale at auction with the intent of generating the best price possible. Consignors (sellers)
can offer for sale horses they own. that others own, or some combination of the two. The
vast majority of horses that are offered for sale at auction are racing prospects (product
brands), meaning that they are being sold with the intent that they will at some point in
time race for the new owner. Thus, for the purposes of this research, consignors (sellers)
are service-based business entities who perfonn a variety of selling activities, such as
preparing a racing prospect (product brand) for an auction and marketing the racing
prospect to prospective buyers.
Although it is possible for the sales of racing prospects (product brands) to occur
through private contract and claiming races, this study focuses on racing prospects that
were purchased through Thoroughbred sales auctions. Consequently, this study focuses
on the business-to-business buyer-seller relationships involving Thoroughbred racing
prospects (product brands) that were purchased via a Thoroughbred sales auction
distribution channel.
The United States Thoroughbred consignment industry provides the benefit of
having a finite number of consignors (sellers ̂  roughly 80-120 national and 250-300
predominantly regional), which represent racing prospects (product brands) at auction.
This characteristic facilitates the opportunity to delve into the theoretically-important
topic area concerning the mediating role of buyer-seller relationship quality between
buyer attitude toward sellers' corporate and product brands, and sellers' brand equity. In
spite of the relatively limited number of consignors (sellers), Thoroughbred consignment
is an extremely competitive, highly uncertain activity; therefore, brand development at
both of these levels is critieal. Thoroughbred buyers help to shape, evolve and reinforee
eonsignors'( sellers') corporate and product brands through the attitudes they hold toward
these brands and the equity that accrues to consignors' (sellers') brands within buyers'
minds; consequently these brands will not be assessed and evaluated equally by buyers.
Consignor's (Seller's) Corporate Brand
The definitive goal of the consignor's (seller's) corporate brand is to signal
important infonnation and added value such that prospective buyers value the
consignor's involvement and look to purchase the consignor's racing prospect (product
brand) offerings. The researcher's experience in the industry paired with qualitative
interviews with the industry experts who helped evaluate and shape the survey
instrument, suggest that consignors' (sellers') corporate brands play a valuable signaling
role to potential Thoroughbred racing prospect (product brand) buyers. As is the case
with many corporate brands (Aaker 2004, 1996; Armstrong and Kotler 2005; Keller
2003), the Thoroughbred consignor (seller) brand demonstrates a high degree of
intangibility, and similar to most service brands (Blankson and Kalafatis 1999; Krishnan
and Hartline 2001), the attributes that make up the consignors'( sellers') corporate brands
are largely experiential and/ or credential in nature.
An important attribute of the consignor's (seller's) corporate brand is the
consignor's (seller's) reputation, which is analogous to status. The number of actors/
participants in the Thoroughbred industry is relatively small, and word travels quickly.
One of the key aspects of consignor (seller) reputation relates to historical aspects of
buyer/ consignor (seller) relationship outcomes. These potentially relationship-specific
outcomes, however, can also extend beyond the idiosyncratic relationship boundaries to
include perfomiance outcomes from external relational dyads, such as consignor (seller)
success with racing prospects (product brands) sold to other buyers. These personal and
tangential performance outcomes serve as signals of the consignor's (seller's) dynamic
capabilities (i.e., skill set and abilities).
Racing Prospect (Product) Brand
The racing prospects (product brands) that Thoroughbred consignors (sellers) sell
are complex offerings that are significantly intangible in nature. Thoroughbred racing
prospects (product brands) are significantly intangible because of their long gestational
horizon (e.g., the wait until a weanling or yearling makes the races can be years) and
extremely speculative nature because one can never be certain what one will get out of an
individual racing prospect (product brand). Consequently, although it may be natural to
think of a Thoroughbred racing prospect (product brand) as a highly tangible entity, the
racing prospect is actually substantially intangible in nature.
Several attributes of the racing prospect (product brand) contribute to how buyers
receive the horse (product hrand). The first of these is the racing prospect's (product
brand's) lineage or pedigree (e.g., family lines on both the sire and dam sides, and how
these breeding lines cross). This attribute is slightly different from the other four
subsequently presented attributes of the racing prospect (product brand), because the
lineage information on each racing prospect (product hrand) is directly measurable,
relates highly to both family lines' past performance, and is immediately available to all
prospective buyers.
The remaining four attributes of the racing prospect (product brand) are much
more subjective in nature and carry a future orientation (i.e., tend to be projected
attributes). The second attribute of the racing prospect (product hrand) is the horse's
(product brand's) physical size and muscularity. Depending in large part on what the
buyer's goal for the purchased racing prospect (product hrand) is, the physical aspects of
the horse convey different merit in different circumstances. For example, in general,
smaller and lighter horses (product brands) tend to be preferred to very large, heavily
muscled horses at shorter race distances, on the grass, and/ or on racecourses where it
rains a great deal. Conversely, in general, large heavily muscled horses (product brands)
tend to be preferred for races run on the dirt and at longer distances.
The racing prospect's (product brand's) conformation is the third attribute and
refers to how well formed the racing prospect is. Conformation deals with the fonnation
and alignment of the racing prospect's (product brand's) body. When a Thoroughbred
horse runs, his/ her entire body will leave the ground and all of his/ her weight will come
down on a single front leg before the other three touch the ground. Consequently, the
confonnation of the horse (product brand) is very important, because a poorly conformed
horse is at substantially higher risk of injury than is a well conformed horse.
The racing prospect's (product brand's) disposition is the fourth attribute and
refers to the racing prospect's attitude, temperament and competitiveness. Thoroughbred
racing prospects (product brands) display the same competitive spirit and temperament
characteristics demonstrated by other world-class athletes. The horse's (racing prospect's)
disposition reveals characteristics of the horse's attitude and competitiveness, which can
foreshadow future performance. Consequently, the disposition of the racing prospect
(product brand) can be a very useful evaluation attribute.
The final attribute relates to the racing prospect's (product brand's) athletic
potential. Racing prospect athletic potential refers to how the buyer envisions the racing
prospect will be able to perform in the future, and is both highly subjective and buyer-
specifie in nature. It is also critical because this attribute underlies the rationale for most
Thoroughbred purchases. For most Thoroughbred buyers, their perception of the horse's
athletic potential is the most important racing prospect (product brand) attribute.
Converting Brand Attitudes into Brand Equity
The evaluation of racing prospect (product brand) talent is an extremely
challenging task, because no one knows or is able to evaluate the racing prospect with
any degree of certainty. The consignors (sellers) that consign these racing prospects
(product brands) to the sales auctions may have intimate or tacit knowledge related to the
racing prospect that the prospective buyer desires; consequently, the involvement of, or
recommendation from certain consignors (sellers) could influence the buyer's
evaluations. The value that buyers ascribe to the involvement of, or recommendation
from, a consignor (seller) reflects seller brand equity. Each of these ideas is articulated in
the scale measures (see Chapter III) for the latent constructs captured in the theory-
grounded eontextual model depicted in Figure 2.5.
Figure 2.5: Thoroughbred Consignment Conceptual Model
Buyer
Attitude
toward the
Con.signor
Brand
(ATC)
Buyer-Consignor
Relationship Consignor Brand
Quality Equity
^ ( RQual) (BrF.q)
Buyer
Attitude
toward the
Horse Brand Satisfaction Commitment
(ATI!) (SATISFY: (COM)
CHAPTER III
METHODOLOGY
This study is designed to fill gaps in the relationship marketing and branding
literatures by looking at a valuable strategic branding tool, the seller's customer
relationships (Aaker 1996; Aaker 1997; Foumier 1998; Keller 2003; McAlexander,
Schouten and Koenig 2002). Chapter III outlines and discusses the research
methodology that is used to test the theory conceptualized and developed within this
study. The goal of this research is to develop a more complete understanding of the
buyer-seller relationship quality construct and to explore its role as a mediator between
buyer attitudes toward the seller's corporate and product brands, and seller brand equity
in a business-to-business context.
This chapter discusses the methodology that is used to investigate the integrity of
the theory-grounded framework developed and presented in Chapter II, through the
testing of the postulated hypotheses and model subsequently performed in Chapter IV.
Due to the high degree of intercorrelation between the model's six latent constructs, as
well as its prominent use and acceptance within other relationship marketing studies
(Garbarino and Johnson 1999; Morgan and Hunt 1994), structural equation modeling is
deemed the appropriate analysis technique to test the main study survey responses. The
following sections discuss aspects of the research methodology in more detail.
Structural Model
This section reintroduces the postulated hypotheses and discusses the conceptual
framework that was developed in Chapter II in the form of a structural equations model.
The structural model consists of two exogenous constructs, buyer attitude toward the
seller's eorporate brand (ATC)a nd buyer attitude toward the seller's product brand
(ATH), as well as four endogenous constructs, satisfaction (SATISF), trust (TRST),
commitment( COM)a nd seller brand equity (BrEq). Buyer-seller relationship quality
(RQual) is a second-order construct composed of three endogenous constructs,
satisfaction (SATISF), trust (TRST), and commitment( COM). The research hypotheses
are presented immediately below, followed by the theory-grounded structural model in
Figure 3.1.
Research Hypotheses:
HI) Buyer attitude toward the seller's corporate brand( ATC) has a positive
direct effect on buyer-seller relationship quality (RQual).
H2) Buyer attitude toward the seller's product brand( ATH) has a positive
direct effect on buyer-seller relationship quality (RQual).
H3) Buyer attitude toward the seller's corporate brand (ATC) has a positive
direct effect on seller brand equity (BrEq).
H4) Buyer attitude toward the seller's product brand (A TH) has a positive
direct effect on seller brand equity (BrEq).
H5) Buyer-seller relationship quality (RQual) has a positive direct effect on
seller brand equity (BrEq).
H6) Buyer-seller relationship quality mediates the effect of buyer attitude
toward the seller's corporate brand on seller brand equity.
H7) Buyer-seller relationship quality mediates the effect ofb uyer attitude
toward the seller's corporate brand on seller brand equity.
Figure 3.1: Theory-Grounded Structural Model
(ATC)
Buyer-Seller Seller's Brand
Relationship Equity
Quality (BrEq)
(RQual)
^ 1 4 ^
Satisfaction Trust Commitment
(SATISF) (TRST) (COM)
ll ^ 12 13 ̂
Research Method
This study utilized a survey methodology designed with the intent of theory
elaboration, to apply and extend existing theory into new settings or contexts (Lee 1999).
Theory elaboration was the appropriate choice for this research because, although the
basis for relationship quality has been discussed within the relationship marketing
literature, our knowledge of the relationship quality construct is still far from complete
(Huntley 2006; Ulaga and Eggert 2006). The methodological technique used in
approaching the application and extension of the existing theory is a mail survey.
Following Dillman (2000) and Malhotra (1993), the mail survey research design
provides several benefits that make it the appropriate choice for this research. First, mail
questionnaires are relatively quick and easy to administer. The method is an effective
way to reach a large number of potential respondents in a relatively short period of time,
while also yielding economic advantages in terms of the cost efficiencies associated with
reaching the targeted respondents. Second, the collected data tend to be reliable because
of the availability of valid and reliable existing scale measures, as well as the limited
number of stated alternatives, which serve to reduce researcher-induced response
variability. Third, data coding, analysis and interpretation are relatively simple and
straightforward, in addition to aligning well with the use of structural equation modeling.
Finally, survey research is a particularly effective technique to utilize when collecting
perceptual data from respondents.
While the mail survey methodology provides benefits, there are also limitations
(Dillman 2000; Malhotra 1993). One of the primary limitations concerns measure
validity and reliability, which is why scale measurement purification is a highly
recommended pursuit (Churchill 1979; Garver and Mentzer 1999; Peter and Churchill
1986; Netemeyer, Bearden and Shanna 2003). A second limitation of the mail survey
method relates to potential response bias. In survey research we try to address potential
bias issues in order to make sure our data are truly representative of the sample
population. Two common forms of response bias are non-response and false reporting
bias (Dillman 2000; Kahle and Page 1976; Page and Kahle 1976).
Non-response bias is the idea that informants who respond at different intervals,
or potential informants who did not respond at all, would respond differently than those
responding at time 1. There are common methods utilized to address the issue of non-
response bias (Armstrong and Overton 1977; Mentzer and Flint 1997). First, data
collected from different survey waves can he compared against each, in order test
whether significant differences exist. A second check that can he perfonned to assure the
consistency of the data involves comparing randomly drawn subsets of survey response
data against other randomly drawn subsets of the response data. A third response bias
cheek involves asking non-respondents to complete a subset of the survey items and then
comparing the responses on these items hack to the original data.
False reporting bias, which is also known as social desirability or demand
characteristics, is the idea that respondents are not telling the truth in their responses.
whether intentionally or unintentionally (Dillman 2000; Kahle and Page 1976; Page and
Kahle 1976). Although test questions can be integrated into the survey instrument to
explore the existence of false reporting bias, there is no failsafe method with which to use
to evaluate its existence, particularly as it relates to perceptual issues. A discussion of
this issue with committee members and industry experts suggested that the incorporation
of extra questions would serve as a deterrent to potential respondents, while there would
be only a minimal likelihood of false reporting bias occurring in this industry with this
relatively benign topic; hence embedded false reporting bias control questions were not
integrated into the survey. Response bias issues in the context of this study are discussed
in more detail in Chapter IV.
Unit ofAnalysis
Because this study investigated the mediating role of buyer-seller relationship
quality between buyer attitudes toward a consignor's (seller's) corporate and product
brands, and consignor (seller) brand equity, the unit of analysis for this research was the
buyer-consignor (seller) relationship at the time of a specific racing prospect (product
brand) purchase from the buyer's perspective.
Sample
Potential respondents were chosen through the random selection of roughly one-
third of the 2700 names contained within the 2006-07 Thoroughbred Owners and
Breeders Association( TOBA) Membership Directory (n = 942). This total was chosen
for several reasons. First, budgetary constraints prevented the acquisition of an electronic
mailing list; consequently the mailing infonnation for the potential respondents needed to
be manually entered from a hardcopy of the TOBA Membership Directory. The time
available for this process allowed for the entry of roughly one-third of the names and
addresses. Second, this study is designed to develop and test the theory-grounded
framework presented in Chapter II. In order to do so, a minimum of one hundred
returned and usable surveys were necessary. It was estimated that a response rate of
I0%-15% would be realistic for this study, thereby providing sufficient data with which
to test the hypothesized model. Third, this study is part of an ongoing stream of research.
The remaining two-thirds of the initial population of TOBA members will be targeted in
follow-up studies.
TOBA's mission is to improve the economics, integrity and pleasure of the sport
on behalf of Thoroughbred owners and breeders. The association manages a wide range
of programs that support its mission, including the Sales Integrity Program, The Greatest
Game Campaign, as well as a variety of other ownership support initiatives (Message
from the President - TOBA Membership Directory 2006-07). Per TOBA management,
and as a broad categorization, roughly half of the membership is involved in the purchase
and ownership of Thoroughbreds (i.e., individual ownership, partnerships/ syndications,
pedigree research/ consultants, bloodstock agents representing the buyer, etc.), and
roughly half is involved in the breeding and/ or sale of Thoroughbreds (i.e..
Thoroughbred breeders and breeding fanns, bloodstock sales agents, sales preparation
agents, stallion syndicates, etc.). One challenge associated with using TOBA's
membership directory is that it does not discern between Thoroughbred owners and
breeders. Due to temporal and fiseal eonstraints, there was not ample opportunity to
prescreen TOBA members to ensure their relevance for this study; consequently, it was
expected that approximately half of the 942 members the initial survey packets were
mailed to would not be appropriate respondents for the discussed topic.
Pilot Study: Developing and Screening Scale Measures
A pilot study was conducted in order to arrive at and refine the scale measures
utilized in this research. Following the work of Elsbach (1994), who used a similar
approach in pilot testing instnaments with California cattle industry experts, this design
utilized review by knowledgeable colleagues, and review and cognitive interviews with
Thoroughbred (product brand) buyers, as a pilot study effort designed to establish survey
question face validity and wording through expert opinion (Dillman 2000; Elsbach 1994).
Pretesting in this manner is effective at establishing the legitimacy and real-to-life nature
of the survey instrument (Elsbach 1994). Thus, utilizing cognitive interviews with
industry experts in order to develop and screen the survey instrument in cases where
traditional survey development is not possible serves as a useful supplemental method to
addressing the issue of measurement scale validity and reliability.
The sampling frame for the pilot study was arrived at through theoretical and
opportunistic sampling (Eisenhardt 1989; Yin 1994) and consisted of industry and
academic experts. The industry expert group consisted of eiglit purposefully selected
Thoroughbred (product brand) buyers, while the academic expert group consisted of four
puiposefully selected scholars who were all well-versed in survey methodology.
Purposeful selection is the recommended sampling technique to utilize when facing
sampling constraints (Dillman 2000; Eisenhardt 1989; Yin 1994) and an in-depth
interview sample of eight is considered sufficiently large to provide the required expertise
(Eisenhardt 1989; McCracken 1988).
This sample was used to screen the face validity and readability of the survey
instrument, making sure the survey instrument was comprised of the right questions,
which in turn were worded using the proper terminology (Churchill 1979). The findings
from this pilot study were then used to reword confusing items and/ or restructure the
survey as required. Following the rewording of the confusing survey items, the survey
was finalized for use in the principal survey mailing. The subsequent subsections provide
an overview and discussion of the list of measurement items developed in the pilot study
and included in the main study survey instrument.
Scale Development & Measurement
Existing scales and item measures were used as the foundation for variable
measurement whenever possible. All of the utilized scales and their item measures are
relevant within this study, as the respective intent of each is proximal to the purposes
discussed in this research design; however, because the intent is not identical, each
needed to be modified, in order to be relevant within the Thoroughbred consignment
context. Each of the dimensions is measured using multiple scale measures, in order to
maximize the validity and reliability of the results (Churchill 1979).
The scale measures used within this study were arrived at through the interplay of
two processes. The first process involved scanning the existing literature for
measurement items appropriate to this study. The context-specific terminology
associated with the Thoroughbred consignment industry and the Thoroughbred racing
prospect (product brand) made it impossible to directly adopt these measures;
consequently, all the item measures utilized within this study were adapted from their
original form to fit this context.
This adaptation serves an important purpose in ensuring that the survey
instrument is not only readable (i.e., possesses face validity), but also that it comes across
as "legitimate" and seems "real-life"( Elsbach 1994). Existing measures that were easily
adaptable to the context (i.e., required only minor wording changes such as changing the
seller term to consignor) were available for items utilized in three of the six latent
constructs; consequently, existing measures are adapted with relatively minor
modification for satisfaction, trust and commitment. However, although the wording in
these items are consistent with the wording of the original scale item measures, the
experts indicated that questions coming from several different scales were a better
reflection of the industry, and, as a result, items measuring satisfaction and trust came
from multiple scales.
The remaining three constructs, buyer attitude toward the horse (product brand),
buyer attitude toward the consignor (seller), and consignor (seller) brand equity, require
more extensive levels of modification. The existing brand attitude and brand equity
literatures provided frameworks with which to approach these constructs; however, the
idiosyncrasies of the Thoroughbred consignment industry make qualitative interviews
essential for developing and refining the items that will aptly tap these constructs and
make them real-to-life (Elsbach 1994). Scale measure development for these constructs
follows Churchill (1979), Peter and Churchill (1986) and Netemeyer, Bearden and
Sharma (2003), supplementing the existing literature with qualitative interviews in order
to develop the scales measuring buyer attitude toward the horse (product brand), buyer
attitude toward the consignor (seller), and consignor (seller) brand equity.
All of the scale items measuring buyer attitude toward the horse (product brand)
are measured using a seven-point scale anchored at 1 (Favorable) and 7( Very Favorable).
The rationale behind the anchors chosen for the brand attitude toward the horse construct
is that Thoroughbred racing prospect (product brand) buyers are making purchases
ranging from thousands of dollars, to in some cases millions of dollars. Discussions with
industry and academic experts suggested that inelusion of unfavorable scale items on this
construct would not provide any additional infonnation, as respondents would only
purchase horses toward which they hold favorable attitudes. It was concluded that the
inclusion of these items would generate "ceiling effects" where variance would be
restricted to a range of 5 (Slightly Favorable) to 7( Very Favorable), with the most
pronounced restricted range clustering occurring at 6 (Moderately Favorable) and 7( Very
Favorable). This logic follows the justification and protocol used in other perceptual
measurement research, such as values research, where scales are developed to capture
respondent sentiment at the point variance exists (Kahle, Beatty and Homer 1983; Kahle
and Homer 1984).
All the scale items for buyer attitude toward the consignor (seller), satisfaction,
trust, commitment, and for three of the five consignor (seller) brand equity measures are
measured using a seven-point scale anchored at 1 (Strongly Disagree) and 7 (Strongly
Agree). In addition to utilizing seven-point scales, two perceptual questions that relate to
the effects associated with a consignor's (seller's) recommendation, also contained a "0"
(Non Applicable) option. The first of these perceptual questions deals with the strength
of a consignor's (seller's) recommendation and is anchored by 1 (Very Weakly) and 7
(Very Strongly). The second of these perceptual questions is the fourth consignor (seller)
brand equity measure and deals with the impact this recommendation had on the buyer's
decision to purchase the racing prospect (product brand). In addition to containing a "0"
(Non Applicable) option, the scale is anchored by 1 (No Affect) and 7 (Strong Affect).
The final consignor (seller) brand equity measure asks about the buyer's intention to
consider this consignor's (seller's) racing prospect (product brand) offerings in the future,
and originally utilized a seven-point scale anchored by 1 (Strongly Disagree) and 7
(Strongly Agree). This item was however subsequently modified to also include a "0"
option, after one respondent indicated that the consignor (seller) that he was responding
about had passed away and another respondent indicated that the consignor (seller) that
he was responding about had left the industry. The comprehensive list of scale items
used to tap and measure each of the six conceptual latent constructs follows in the
descriptions and summaries appearing below.
Buyer Attitude toward the Horse (Product Brand)
The first independent variable utilized in this research design draws on the
attitude toward the product brand scales of Kim and Chan-Olmsted (2005) in order to
frame the item structure, and then industry expert opinion in order to arrive at the items
and wording utilized within the multidimensional buyer attitude toward the horse
(product brand) scale. Buyer attitude toward the horse (product brand) reflects the
buyer's attitude toward the consignor's (seller's) product brand offering. The buyer's
attitude is based on the characteristics and traits the buyer sees within the offered racing
prospect (product brand) that project into future performance outcomes. The author's
experience in the industry, coupled with feedback from industry experts, suggests that the
following five measurement items aptly tap the construct. The original survey item
measures follow immediately below in Table 3.1.
Table 3.1: Items Measuring Buyer Attitude toward the Horse (Product Brand)
Five Scale Items:
Adapted Based On: Seven-Point Scale Ranging from 1 (Favorable) to 7( Very
Favorable)
Kim and Chan- At the time of purchase, how favorable were your pereeptions
Olmsted (2005)and of the racing prospect's breeding cross/ pedigree?
Qualitative Interviews
Kim and Chan- At the time of purchase, how favorable were your perceptions
Olmsted (2005)and of the racing prospect's physical appearance (size and
Qualitative Interviews muscularity)?
Kim and Chan- At the time of purchase, how favorable were your perceptions
Olmsted (2005)and of the racing prospect's confonnation?
Qualitative Interviews
Kim and Chan- At the time of purchase, how favorable were your perceptions
Olmsted (2005) and of the racing prospect's disposition (attitude, temperament
Qualitative Interviews and competitiveness)?
Kim and Chan- At the time of purchase, how favorable were your perceptions
Olmsted (2005) and of the racing prospect's athletic potential?
Qualitative Interviews
Buyer Attitude toward the Consignor (Seller Brand)
The second independent variable utilized in this research design focuses on the
multidimensional nature of buyer attitude toward the consignor (seller) and draws on the
predictive variable attitudinal measurement items utilized by Garbarino and Johnson
(1999). Buyer attitude toward the consignor (seller) reflects the buyer's attitude toward
the consignor's (seller's) corporate brand. It is derived from the buyer's attitude toward
various attributes of the seller's corporate brand. In speculative environments such as the
Thoroughbred consignment industry, where product variation is high and no two products
are ever the same, the seller's corporate brand can be viewed as being significantly more
stable and predictable than are his/ her product brand offerings. The initial item measures
follow immediately below in Table 3.2.
i . ..
Table 3.2: Items Measuring Buyer Attitude toward the Consignor (Seller Brand)
Adapted Three Scale Items: Seven-Point Scale
Based On: Ranging from 1 (Strongly Disagree) to 7 (Strongly Agree)
Garbarino and At the time of purchase this consignor had a good reputation
Johnson(1999) and within the industry.
Qualitative Interviews
Garbarino and At the time of purchase, I felt good about this consignor's
Johnson(1999)and abilities.
Qualitative Interviews
Garbarino and At the time of purchase, I felt good about the involvement of
Johnson(1999)and this consignor with this racing prospect.
Qualitative Interviews
Buyer-Consignor (Seller) Relationship Quality
The mediating variable utilized in this research design focuses on the
multidimensional nature of buyer-seller relationship quality and draws on the item
measures of several different researchers. Relationships are not equal, and relationship
quality varies markedly across buyer-consignor (seller) relationships. Buyer-consignor
(seller) relationship quality is a higher-order constmct driven by different combinations
of satisfaction, trust and commitment (Garbarino and Johnson 1999; Johnson and Seines
2004; Ulaga and Eggert 2006). As a result, we must draw upon existing theory to
provide us with scales to measure the variables that constitute relationship quality - i.e..
satisfaction, trust and commitment (Garbarino and Johnson 1999; Johnson and Seines
2004; Ulaga and Eggert 2006).
Satisfaction
Satisfaetion is an endogenous eonstruet that refers to the buyer's level of
eontentment based on his/ her personal and tangential purchase and consumption
experiences with a consignor's (seller's) racing prospect (product brand) offerings over
time (Anderson, Fomell and Lehmann 1994; Garbarino and Johnson 1999). Satisfaction
can thus extend beyond idiosyncratie buyer-seller relationship boundaries also to include
buyer satisfaction with how the consignor (seller) conduets business on an ongoing basis.
Based on the feedbaek from the cognitive interviews with industry experts, individual
scale items were seleeted from three different measurement scales. The initial item
measures follow immediately below in Table 3.3.
Table 3.3: Items Measuring Satisfaction
Adapted Five Scale Items: Seven-Point Seale
Based On: Ranging from 1 (Strongly Disagree) to 7 (Strongly Agree)
Garbarino and At the time 1 purehased this racing prospect, 1 was satisfied
Johnson(1999)and with the quality of racing prospects this consignor offered.
Qualitative Interviews
Garbarino and At the time I purehased this raeing prospeet, I was satisfied
Johnson(1999)and with the quality of service this consignor provided.
Qualitative Interviews
Maddox (1982) and At the time I purchased this raeing prospect, I was satisfied
Qualitative Interviews that this consignor's offerings were a good value for the price
paid.
Kim and Chan- At the time I purchased this racing prospect, I was satisfied
Olmsted (2005) and with my working relationship with this consignor.
Qualitative Interviews
Kim and Chan- At the time I purchased this racing prospect, I was satisfied
Olmsted (2005) and that this consignor was a good seller with which to do
Qualitative Interviews business.
Trust
Following the work of Dwyer and Oh (1987) and Morgan and Hunt (1994), trust
is an endogenous construct that occurs when the buyer has confidence in the consignor's
(seller's) reliability and integrity. Trust is a critical concept in most business contexts,
but it is heightened in high-risk, speculative climates such as the one that exists within the
Thoroughbred industry. Issues of trust within the Thoroughbred industry have come to
light over the past couple of years, drawing both trade and national attention with a
handful of lawsuits alleging consignor (seller) impropriety by taking kickbacks on racing
prospects (product brands) they have consigned to auction. Following Johnson and
Seines (2004), trust in the consignor (seller) is thus made even more poignant as it refers
to the level of confidence that the buyer feels towards the consignor (seller) and his/ her
racing prospect (product brand) offerings in the absence of perfect information, which
will always be the case within uncertain and speculative contexts. The initial item
measures follow immediately below in Table 3.4.
Table 3.4: Items Measuring Trust
Adapted Five Scale Items: Seven-Point Scale
Based On: Ranging from 1 (Strongly Disagree) to 7( Strongly Agree)
Dwyer and Oh (1987) At the time I purchased this racing prospect, 1 could count on
and Qualitative this consignor to be sincere.
Interviews
Dwyer and Oh (1987) At the time I purchased this racing prospect, I was suspicious
and Qualitative of this consignor's recommendations,( reverse scored)
Inteiwiews
Dwyer and Oh (1987) At the time I purchased this racing prospect, I questioned the
and Qualitative integrity of this consignor,( reverse scored)
Interviews
Dwyer and Oh (1987) At the time I purchased this racing prospect, this consignor
and Qualitative and I worked together to ensure the best outcomes for both of
Interviews us.
Morgan and Hunt At the time 1 purchased this racing prospect, this consignor
(1994) and Qualitative was a relational partner I trusted completely.
Interviews
Commitment
Following Moorman, Zaltman and Deshpande (1992) and Morgan and Hunt
(1994), commitment is conceptualized as an enduring desire to maintain a valued
relationship, and describes the degree to which the buyer believes the relationship with
the consignor (seller) is worth investing resources. Commitment to the relationship is
envisioned to elicit better informational exchanges between buyer and seller, thereby
reducing buyer search and due diligence costs (Hunt 1997; Johnson and Seines 2004;
Seines and Sallis 2003). In order for commitment to occur, the buyer must perceive a
higher level of intrinsic value and motivation because of the consignor's (seller's)
involvement; otherwise, there would he no reason for the buyer to feel committed to the
consignor (seller). The initial item measures follow immediately below in Table 3.5.
Table 3.5: Items Measuring Commitment
Adapted Four Scale Items: Seven-Point Scale
Based On: Ranging from 1 (Strongly Disagree) to 7( Strongly Agree)
Morgan and Hunt At the time I purchased this racing prospect, the relationship I
(1994) and Qualitative had with this consignor was something to which I was very
Interviews committed.
Morgan and Hunt At the time 1 purchased this racing prospect, the relationship I
(1994) and Qualitative had with this consignor was something I planned to maintain
Interviews indefinitely.
Morgan and Hunt At the time 1 purchased this racing prospect, the relationship I
(1994) and Qualitative had with this consignor deserved my maximum effort to
Interviews maintain.
Morgan and Hunt At the time 1 purchased this racing prospect, the relationship I
(1994) and Qualitative had with this consignor was something I cared a great deal
Interviews about long-term.
Consignor (Seller) Brand Equity
The criterion variable utilized in this research design focuses on the
multidimensional nature of consignor (seller) brand equity. Consignor (seller) brand
equity is the differential effect that the consignor's (seller's) corporate brand has on buyer
evaluations of his or her racing prospect offerings (product brands) verses those
evaluations of the average consignor (Aaker 1996; Keller 2005). Consignor (seller)
brand equity is derived from everything the buyer takes into account when considering
the consignor's (seller's) racing prospect (product brand) offerings and can be positive or
negative (Keller 2003, 1993).
Positive consignor (seller) brand equity occurs when buyers have a high level of
positive associations with the consignor's (seller's) corporate brand, thereby resulting in a
more favorable buyer response to that consignor's (seller's) racing prospect (product
brand) offerings (Aaker 1996, 1991; Agarwal and Rao 1996; Mackay 2001). Consignor
(seller) brand equity therefore captures the differential effect that the consignor's
(seller's) corporate brand has on buyer response to the marketing of that consignor's
racing prospect (product brand) offerings, and can be thought of as the additional value
endowed to the racing prospect (product brand) above and beyond what the buyer would
pay for the same horse from the average consignor (seller)( Aaker 1996; Agarwal and
Rao 1996; Farquhar 1989; Keller 2003).
Scale measures draw on Agarwal and Rao (1996) and Mackay (2001) in order to
frame the item structure and then industry expert opinion in order to arrive at the items
and wording utilized within the multidimensional consignor (seller) brand equity scale.
The author's experience in the industry, coupled with feedback from industry experts,
suggest that the following five measurement items aptly tap the consignor (seller) brand
equity construct. The initial survey item measures follow immediately below in Table
Table 3.6: Items Measuring Consignor (Seller) Brand Equity
Adapted Five Scale Items: Seven-Point Scale
Based On: Ranging from 1 to 7( Multiple Scales)
Agarwal and Rao (1996), Had this prospect been offered by the average consignor
Mackay (2001) and instead of the consignor 1 purchased from, 1 would have
Qualitative Interviews been willing to pay the same amount for this horse.
Agarwal and Rao (1996), At the time of purchase, 1 would have been willing to pay
Mackay (2001) and more for a racing prospect sold by this consignor than for
Qualitative Interviews the same horse sold by the average consignor, because of
what it meant to have this consignor's involvement
(endorsement).
Agarwal and Rao (1996), 1 intend to consider racing prospects offered by this
Mackay (2001) and consignor again in the future.
Qualitative Interviews
Agarwal and Rao (1996), How much did this (consignor's) recommendation affect
Mackay (2001) and your decision to purchase this prospect?
Qualitative Interviews
Agarwal and Rao (1996), At the time of purchase, the involvement of this
Mackay (2001) and consignor helped in my decision to purchase this
Qualitative Interviews prospect.
Following Elsbach (1994), who had pilot tested her instruments with California
cattle industry experts, it was detemiined that using Thoroughbred industry experts to
detemrine which measurement items to include and discard within the final survey
instrument provided the best opportunity to ensure the legitimacy and real-to-life nature
of the survey instrument to be used in the main study. The final survey instrument is
presented in Appendix A.
Main Study
The research design for the main study consisted of mailing the survey instrument
that was developed in the pilot study to 942 Thoroughbred racing prospect (product
brand) buyers who had purchased racing prospects (product brands) from consignors
(sellers) at auction in the past. There are approximately two dozen national auctions
where more than 7,500 Thoroughbred racing prospects (product brands) are sold annually
by approximately 80-120 national consignors (sellers) and 250-300 predominantly
regional consignors (sellers).
Potential respondents were mailed a cover letter (refer to Appendix B)a sking for
their participation in the study provided they had previously been involved in the
purchase of Thoroughbred racing prospects (product brands), two categorical versions of
the survey, and a postage-guaranteed return envelope. The first version of the survey
asked the potential respondent (buyer) to think about a strong (stronger) relationship he/
she had with a consignor (seller) at the time of a specific Thoroughbred (product brand)
purchase. The second version asked the potential respondent to think about a weak
(weaker) relationship he/ she had with a consignor (seller) at the time of a specific
Thoroughbred (product brand) purchase. Potential respondents were offered an incentive
to complete the survey in the form of a $25 charitable donation that would be made to the
Thoroughbred Retirement Foundation (TRF)f or each completed questionnaire, up to a
maximum donation of $1,000 for all responses. The mail survey was administered in a
manner consistent with Dillman (2000), except Dillman's first two steps were combined
into a single step within this study. The utilized protocol is as follows.
1. Initial contact and distribution of the survey packets
2. Follow up postcard reminder of the survey due date (refer to Appendix C)
3. Second distribution of the survey packets to non-respondents (refer to
Appendix D)
4. Follow-up telephone calls
In the survey, respondents were asked a variety of questions related to study's six
latent constructs. Respondents were asked questions regarding their attitudes toward a
specific consignor's (seller's) corporate and racing prospect (product) brands and the
attributes that constitute each, questions specific to the buyer-seller relationship quality
dimensions of satisfaction, trust and commitment, and questions that reflected the buyer's
perceptions of various aspects reflective of the seller's brand equity. Additional
questions were also asked in regard to several other areas. Respondents were asked how
many years they had been involved in the Thoroughbred industry, what the age category
of the purchased horse (product brand) was at the time of purchase, and how strongly the
consignor (seller) had recommended the racing prospect (product brand) to the buyer, in
order to ascertain if there were any differences and to use in future studies. Respondents
were also asked to provide any comments they thought were important, as well as the
lineage of the purchased horse, for use in follow-up studies. Finally, respondents were
asked if they would like a summary of the findings sent to them at the completion of the
study. In all, 138 usable strong relationship surveys and 111 usable weak relationship
surveys were completed and returned, for a total of 249 usable surveys.
The initial contact and distribution of the survey packets occurred April 5th,
2007. This timeframe was chosen because the early season yearling sales had all been
completed. Roughly 85% of the racing prospects (product brands) that are sold at auction
in a given year are yearlings; hence, this buyer group was considered the most likely to
respond to this research request. Additionally, although the 2-year-old-in-training sales
season was underway at time the survey was mailed, there was a nearly two week lag that
coincided with the mailing of the initial survey packets. It was believed that by timing
the survey mailing around this lag, that there would be an increased opportunity to
generate responses from buyers considering 2-year-old-in-training racing prospects
(product brands) at the spring 2007 sales. Finally, choosing the first week of April as the
time period for the initial mailing allowed for the second distribution of the survey
packets to occur at the beginning of May. Similar to the rationale used with the initial
mailing in April, the time period around the second mailing wave was also a slower
period of the 2-year-old-in-training auction season, thereby again increasing the
likelihood of generating responses from this group of buyers.
Survey Response Rate
Of the 1884 surveys that were initially mailed, 68 were returned as undeliverable
due to bad addresses, 10 more were returned as undeliverable due to the passing of the
targeted individual, 36 were returned after the data analysis cutoff point, and 63 were
returned with only one of the two surveys completed. In support of TOBA's belief that a
significant portion of their membership roster would not be relevant for this study.
another 256 surveys were returned because respondents stated the content of the surveys
were not relevant or applicable for them. Table 3.7 highlights the rational provided by
these respondents, followed by a descriptive summary of the study's response rate. The
full response rate calculations, which were based on the most conservative estimates,
appear in Appendix E.
Table 3.7: Non-Applicability Rationale Provided by Potential Respondents
Provided Rationale . Number
Thoroughbred Breeder/ Race Only Homebreds 210
Questions about Consignors do not Reflect how
Thoroughbred Purchases are Actually Made
Part of a Partnership/ Syndicate where Someone Else
Evaluates
Conflicts with the Pinhooking Sales/ No Time for Surveys
Total
A total of 255 completed surveys were returned. In addition to the adjustments
discussed above, adjustment were also conducted related to invalid survey responses and
tbe respondent's indication of survey non-applicability, thereby reducing the sample size
estimates to 1445, and the number of responses to 249, for a final response rate of 17.2%.
It is important to note that while this response rate is respectable within B2B contexts
where response rates are typically lower than they are in B2C contexts, that this response
rate was calculated based on the most conservative estimates. For example, while the
actual population in this study is unknown because the initial mailing list contained
numerous names that were not eligible for this study( TOBA estimates suggest
approximately 470 mailing recipients and 940 mailed surveys were not eligihle), all
potential respondents were retained in the in the response rate calculations unless an
unsolicited admission of respondent inapplicability for this study was received.
Consequently, the actual response rate may actually be substantially greater than 17.2%;
however, there is no way to he sure to what degree.
Data Analysis Approach: Structural Equation Modeling( SEM)
Structural equation modeling( SEM) was deemed the appropriate technique to
analyze the data collected through the 249 usable surveys collected within this study.
Structural equation modeling provides valuable capabilities when analyzing latent
variable relationships demonstrating significant degrees of shared variance (Bollen and
Long 1992; Fomell and Larcker 1981, 1981) and affords several henefits over other
statistical options. First, SEM provides the benefit of allowing for the simultaneous
estimation of multiple latent variable paths (Hair, Anderson, Tatham and Black 1998;
Garbarino and Johnson 1999; Morgan and Hunt 1994). Second SEM allows for the
accounting of measurement error in latent variables (Hair, Anderson, Tatham and Black
1998; Fomell and Larcker 1981, 1981). Third, structural equation modeling provides the
ability to compare and contrast the hypothesized model and construct paths verses rival
models, thereby allowing the researcher to observe the effects of specific variables and to
adjust his/ her model accordingly (Bollen and Long 1992; Garbarino and Johnson 1999;
Carver and Mentzer 1999; Morgan and Hunt 1994).
Summai
The goal of this research is to enhance our knowledge and understanding of
buyer-seller relationship quality and to explore its role as a mediator between buyer
attitudes toward the seller's corporate and product brands, and seller brand equity in a
business-to-business context. Chapter III outlines and discusses the research
methodology adopted in order to test the theory-grounded framework that is
conceptualized and developed in Chapter II. Chapter III presents a comprehensive list of
scale measures that, following the work of Elsbach (1994), was developed via the Pilot
Study through the interplay of existing scale measures and cognitive interviews with
industry experts. This technique is used in order to establish the "legitimacy" and real-to-
life nature of the survey instrument (Elsbach 1994). The byproduct is a list of multiple
seale items that can be used in measuring each of the survey's six latent constructs.
Structural equation modeling was deemed the appropriate analysis technique to test the
data collected from 249 survey responses. The data analysis process and results are
presented in Chapter IV, followed by the conclusions and contributions of these findings
in Chapter V.
CHAPTER IV
DATA ANALYSIS AND RESULTS
Chapter IV discusses the process utilized to test the measurement model, as well
as presents the survey results. Data analysis was conducted using both SPSS (Version
13.0) and AMOS (Version 5.0). The result was a measurement model that provided good
fit statistics, as well as construct validity and reliability.
The structural equations model shows that 1) buyer attitude toward the seller (i.e.,
the consignor's corporate brand), 2) buyer attitude toward the product (i.e., the
consignor's racing prospect offering), and 3) buyer-seller relationship quality (i.e., the
buyer-consignor relationship, all have a significant direct effect on seller brand equity
(i.e., the consignor's brand equity). The model also reveals that buyer-seller relationship
quality fully mediates the positive direct effect of buyer attitude toward the seller on
seller brand equity, and it partially mediates the negative direct path between buyer
attitude toward the product and seller brand equity.
The balance of Chapter IV delves into the process and data used to investigate the
research hypotheses and to arrive at the study findings. Chapter V then discusses the
implications of these finding and provides recommendations for future research.
Developing and Refininti the Measurement Model
Running exploratory data analyses prior to running the final measurement model
helps to produce a model with acceptable fit statistics. AMOS and SPSS are used in
conjunction to perform a number of exploratory analyses designed with the intent of
developing the best possible test of the model. These analyses include calculating
descriptive statistics, conducting factor analyses, and conducting construct validity and
reliability assessments.
Descriptive Statistics
The first step in the process of developing an acceptable measurement model
involves analyzing the collected data for nonnality. The results suggest the data can be
considered nonnally distributed. The survey contains three negatively-worded items
(suspicious of the consignor's recommendations, questioned the consignor's integrity, and
would have paid the same regardless of the consignor), which are analyzed in their raw
form for most of the analyses. These items are, however, reverse-scored for use within
construct indices, prior to analyzing the data. Missing values in the data set (8 out of
more than 4,250 responses) are replaced using the average of the remaining construct
item measures. One respondent left multiple items blank and that individual's responses
are omitted from the final analysis.
The means of the scale items range from 2.63 to 5.65, with standard deviations
ranging from 1.401 to 2.279. All items received the full range of responses. Six items
have kurtosis values above one, with buyer attitude toward the racing prospect's (horse's
—> product brand's) athletic potential exhibiting a positive kurtosis value of 1.805, and
impact of the consignor's (seller's) involvement (-1.352), cared about the relationship
long-tenu (-1.096), trusted the consignor (seller) completely (-1.093), planned to
maintain the relationship indefinitely (-1.050), and worked together with the consignor
(seller)( -1.009) exhibiting slightly negative kurtosis values. While none of these values
are thought to be problematic, additional tests of noraiality showed statistics for both the
Shapiro-Wilk test of nonuality (typically recommended for small and medium samples)
and the Kolmogorov-Smimov test of nomiality (typically used for larger samples) both
affirmed the normal distribution of the data for all measurement items. Complete
descriptive statistics for the survey data are provided in Appendix F.
Response Bias Checks
The utilized protocol for this study involved two waves of survey mailings. In
order to address non-response bias issues, data collected from the two survey waves are
compared against each other, in order to test for significant differences in responses
between the waves (Anustrong and Overtoil 1977). An independent samples t-test
comparing the means of the two survey waves reveals no significant differences between
the response waves. A second check is then conducted in which randomly drawn subsets
of the survey response data are eompared against other randomly drawn subsets of the
response data. Again no significant differences in response data are found. As a final
response bias check, thirty non-respondents were telephoned and asked to complete a
subset of the original survey items (Mentzer and Flint 1997). The contacted individuals
were asked to answer six questions, with each representing one of the six first-order
model constructs. Their responses are then compared to sample data, and again no
significant response differences were found. The independent sample t-test results for
these response bias checks appear in Appendix F.
There is no failsafe method with which to evaluate false reporting bias (also
known as social desirability or demand characteristics), particularly as it relates to
perceptual issues. One strategy to avoid false reporting bias is to target appropriate
respondents. False reporting bias is not thought to be an issue in this study as the
perceptions of the respondents seems to be appropriate for this study's conceptual
framework. However, the data were screened for outlier responses as a step to address
any additional bias. Additionally, many respondents took the time to provide extensive
comments related to their purchase of racing prospects (product brands) and of the
Thoroughbred industry as a whole. As a result, false-reporting bias was not thought to be
a major concern in this study.
Preliminary Factor Analyses
The next step involves conducting exploratory and confirmatory factor analyses
using SPSS. Factor analyses reveal when measurement items load at acceptable or
unacceptable levels on the factors in the model. Based on the existing theory, we would
expect the twenty-seven original scale items to load onto six separate first-order
constnacts: 1) buyer attitude toward the product (racing prospect/ horse) —> ATH, 2)
attitude toward the seller (consignor) —> ATC,3 ) satisfaction —> SATIS, 4) trust —>
TRST,5 ) commitment —^ COM,a nd 6) seller (consignor) brand equity ̂  BrEq.
An exploratory factor analysis allows the measurement items to freely load on an
undetennined and unconstrained number of constructs. The number of constructs is
reached by retaining the principal components that demonstrate an eigenvalue over a
certain threshold, which is typically set at 1.0. In running an exploratory factor analysis
for the data in question, several of the scale items fail to load cleanly on the expected
construct, meaning that these items correlate highly with constructs beyond the construct
that theory supports. Further, although theory suggests that six first-order constmcts
should emerge when conducting an exploratory factor analysis (four if satisfaction, trust
and commitment are to collapse into the second-order relationship quality construct),
only three constructs achieved eigenvalues greater than one. Confirmatory factor
analyses affirmed factor cross-loading issues; consequently, it became necessary to
eliminate bad scale items (i.e., scale items demonstrating high degrees of factor cross-
loading) in order to refine the construct measures.
Refining the Construet Measures
In order to refine the construct measures, it is necessary to adopt criteria to guide
the process. The decision criteria used to evaluate the retention or rejection of the
measurement items involves an iterative process using both the modification indices( MI)
feature in AMOS and the confirmatory factor analysis feature in SPSS. The decision
criteria involved first looking at the AMOS modification indices report to determine
which items are most problematic. The AMOS modification indices feature produces a
report that highlights measurement issues such as scale measures cross-loading on
multiple constructs and correlating highly with other non-related scale items. The
modification indices feature is therefore an effective tool in arriving at the best fitting and
most parsimonious construct measures.
Measurement items appearing in the modification indices multiple times and/ or
displaying high Ml scores were considered for removal. The initial modification indices
which served as the starting point for measurement item removal to arrive at the final
model are depicted in Appendix G. These items were then subjected to three additional
criteria tests. First, considered items were investigated using the SPSS confirmatory
factor analysis feature. Using this criterion, considered items were expected to display
factor loadings of greater than 0.60 on the theoretically-expected construct, and display
factor loadings of less than 0.50 on all the other constructs. Second, the standardized
regression weights for the considered items were also considered. Item standardized
regression weights were expected to be equal to or greater than 0.60. Finally, and most
importantly, the theoretical domain of the item was considered. Constituent items that
failed to add substantial explanatory capacity to the parent constructs were also dropped,
while items that are theoretically important but fail to meet the other two criteria could be
retained. The following section discusses the building of the measurement model
through the refinement of the measurement model constructs.
Building the Measurement Model
As previously mentioned, AMOS and SPSS are used in conjunction to refine the
scale measures used to build a measurement model to test the relationships between 1)
buyer attitude toward the product (racing prospect/ horse) —^ ATH, 2) buyer attitude
toward the seller (consignor) ̂  ATC, buyer-seller (consignor) relationship quality —>
RQual, and 4) seller (consignor) brand equity —>■ BrEq. In order to develop the best
fi tting measurement model, several model refinement iterations are perfonned, where
poorly fitting scale measures and/ or excess items related to the constructs of interest are
removed. The removal of these items followed the decision criteria outlined above.
Following Anderson and Gerbing (1988) and Hair, Anderson, Tatham and Black
(1998), the process uses the maximum likelihood estimation procedure for model testing.
The first step involves examining the results for theoretical inconsistencies - i.e.,
offending estimates (Hair, Anderson, Tatham and Black 1998). Next, several key
components of structural equation modeling (SEM) output are utilized to evaluate overall
model fi t. In addition to considering standardized regression weights, error teiTns, and
outlier effects. Hair, Anderson, Tatham and Black (1998) suggest the use of key
indicators, including the comparative fi t index (CFl), goodness of fi t index (GFl), root
mean square error of approximation (RMSEA), and nonned chi-square (CMIN/DF).
Although there are no steadfast requirements for these values, there are general
guidelines (Hair, Anderson, Tatham and Black 1998). The general acceptance level for
the comparative fi t index (CFl), which is both a baseline and a parsimonious fi t measure,
is 0.90. Similarly, the general acceptance level for the goodness of fi t index (GFl), which
is an absolute fit measure, is 0.90. The general acceptance levels for the root mean error
of approximation (RMSEA), another measure of absolute fit, is between 0.05 and 0.08.
Finally, the general guideline for the normed chi-square (CMIN/DF), which is a
calculated measure of overall model goodness of fit, is a value of under 3.0.
The initial base model contained all twenty-seven original endogenous scale
measures, and all six latent constructs. In the base model, all latent variables are allowed
to freely correlate. In addition to the full base model, a comparison model is also run.
This model is a specified modification of the base model, in which all the latent variable
covariances are set equal to one. The goal of this model is to provide a basis of
comparison. We would expect that our predicted model(s) should outperform the
constrained comparison model.
Base and Comparison Measurement Models
The comparison model demonstrates significantly weaker key fit indicator values
than does the base model. The comparison model CFl (0.737) and GFI (0.619) are
significantly below the targeted thresholds, while the RMSEA (.143) and CMIN/DF
(6.069) are significantly above targeted thresholds. The unconstrained base model
displays key fit indicator values that are better than the comparison model, but still well
outside the targets for each. The key fit measures for the base model are: CFI (0.772),
GFI (0.685), RMSEA (.136), and CMIN/DF (5.616). The results of a Mahalanobis d-
squared test, which is an analysis tool used to observe cases farthest from the centroid.
show there to be a single outlier. This outlier has a Mahalanobis distance of 129.673,
which is significantly higher than the next highest value of 69.755. The measurement
model was rerun without the outlier data; however, deletion of the outlier had only trivial
impact on the model results, so the outlier data is retained.
The coding for the measurement items utilized within the statistical analyses that
follow appears immediately below in Table 4.1.
Table 4.1: Measurement Item Coding Summary
Coding Term Conceptual Description
lORSEl *** Buyer attitude toward the horse's breeding cross/ pedigree
fORSE2 Buyer attitude toward the horse's physical appearance (size and
muscularity)
10RSE3 Buyer attitude toward the horse's conformation
10RSE4* Buyer attitude toward the horse's disposition (attitude, temperament
and competitiveness)
10RSE5 Buyer attitude toward the horse's athletic potential
!10N1 Buyer felt the consignor had a good reputation in the industry
rON2 Buyer felt confident in this consignor's abilities
30N3** Buyer felt good about the involvement of the consignor with this horse
SATl Buyer was satisfied with the quality of the consignor's horses
iAT2 Buyer was satisfied with the quality of the consignor's service
)AT3 Buyer was satisfied this consignor's horses were a good value for the
money
iAT4* Buyer was satisfied with his/ her working relationship with the
consignor
1AT5*** Buyer was satisfied this was a good seller with which to do business
^RUSTl* Buyer could count on the consignor to he sincere
"RUST2 Buyer was suspicious of this consignor's recommendations (reverse
coded)
'RUST3 Buyer questioned this consignor's integrity (reverse coded)
'RUST4* Buyer worked together with the consignor to ensure the best outcomes
for both
'RUST5** Buyer trusted the consignor completely
/OMMITl Buyer was very committed to the relationship with this consignor
COMMIT2 Buyer planned to maintain the relationship with this consignor
indefinitely
COMMITS Buyer felt the relationship deserved his/ her maximum effort to
maintain
C0MMIT4 The relationship was something the buyer cared a great deal about
long-tenn
PAYSAME* Buyer would have been willing to pay the same price had this horse
instead been offered by the average consignor
PAYMORE Buyer would have been willing to pay more for a horse sold by this
consignor than for the same horse sold by the average consignor
Buyer intends to consider horses offered by this consignor again in the
future
AFFECT Effect of the consignor's recommendation on the buyer's decision to
purchase this horse
CONIMPAC The involvement of this consignor helped in the buyer's decision to
purchase this horse
* D ropped in Measurement Model Revision 1
** Dropped in Measurement Model Revision 2
*** Dropped in Measurement Model Revision 3
Measurement Model Revision 1
The base model modification indices reveal that Fl, PAYSAME,T RUSTl,
TRUST2, TRUST3, TRUST4, SAT4 and HORSE4 are all loading with other constmcts
and measurement items. A review of the data under the decision criteria guidelines
suggests that two brand equity measures (Fl and PAYSAME),t wo trust measures
(TRUSTl and TRUST4), and one satisfaction measure( SAT4) are also demonstrating
high factor cross-loadings in SPSS, while a review of the attitude toward the horse
construct with industry experts suggests that the H0RSE4 item, which deals with the
horse's disposition, is not important to the integrity of the construct and should be
removed; hence HORSE4 is dropped. Additionally, the theoretical domain and the
integrity of the other constructs are also considered, should their respective constituent
measurement items be deleted.
Brand equity is typically manifested through a buyer's willingness to pay a
premium for the seller's product and/ or through the buyer's future intentions (Aaker
1996, 1991; Keller 2003, 1993). While typically an indicator of brand equity accrual, FI
fails to load with any of the other brand equity measures. The failure of FI to load on the
brand equity construct likely occurs because Thoroughbred racing prospects (product
brands) are usually purchased in an auction environment. The nature of this practice
creates an environment in which the focus is predominantly on the racing prospect
(product brand), and the consignor (seller) brand simply serves to reinforce this product
brand. PAYSAMF is a negatively worded item that is constructed to delve into buyers'
willingness to pay a premium for the involvement of a particular consignor (seller). It
becomes clear in reviewing survey responses that the PAYSAMF question is not clear to
a number of respondents, while PAYMORF is a much better worded and understood item
that asks for the same infonnation in a more direct manner. Consequently, the
PAYMORF item is retained while both the FI and PAYSAMF items are removed.
The modification indices and SPSS tests of convergent and discriminant validity
show that TRUSTl cross-loads with the items measuring satisfaction, while TRUST4
cross-loads with the items measuring commitment, and that both cross-load with the
attitude toward the consignor (seller) items. Similarly, SAT4 also cross-loads with the
items measuring trust, commitment and attitude toward the consignor (seller). A review
of the remaining items suggests that deletion of these items does not compromise the
integrity of the relevant latent construct, and these items are also deleted.
The model is then rerun, and the test of this modified model signifieantly
improves the CFI (0.934), GFI (0.838), RMSEA (0.090) and CMIN/DF (2.996) model fit
statistics. While these statistics are either acceptable (CFI or CMIN/DF), or approaching
general acceptance thresholds (GFI and RMSEA), the modification indices indicate that
TRUSTS and C0N3 are now disrupting the model.
Measurement Model Revision 2
The Revision 1 modification indices reveal that TRUSTS and CON3 are tending
to load with other constructs and measurement items. The modification indices and SPSS
tests of eonvergent and discriminant validity show that TRUSTS cross-loads with attitude
toward the consignor (seller), satisfaction, commitment and consignor (seller) brand
equity, while C0N3 cross-loads with consignor (seller) brand equity. The trust and
consignor (seller) brand equity constructs are reexamined to detennine if the respective
deletions of TRUSTS and CONS will compromise the integrity of either, and it is
detennined that they will not. Both items are deleted and the model rerun.
The test of this modified model once again significantly improves the CFI
(0.968), GFI (0.896), RMSEA (0.061) and CMIN/DF (1.92S) model fit statistics.
Although these statistics are well in line with general acceptance thresholds, two
measurement items, HORSE 1 and SATS, are still demonstrating problematic tendencies.
. f ;  m. , i i"' "■
Measurement Model Revision 3
Although the second revision of the measurement model results in a model with
good fit statistics, two items continue to be problematic. The first item, HORSE 1, asks
about the lineage of the horse. In spite of being an important aspect in a prospective
buyer's initial evaluation of a racing prospect (product), the scale item continuously
exhibits substantially lower factor loadings than the other buyer attitude toward the horse
measures (Revision 2 loadings: HORSEl = 0.571, HORSE2 = 0.905. HORSE3 = 0.866,
H0RSE5 = 0.895). The large discrepancy between these regression weights is at first
perplexing; however, after speaking with several buyers, it becomes apparent that the
horse's lineage only serves as an initial evaluation criterion for many buyers, and once the
horse makes their consideration list, the buyers' evaluation is predicated on the horse's
appearance, confomiation and athletic potential (H0RSE2, HORSE3 and HORSE5). As
a result of this finding, HORSEl is dropped.
SPSS factor analyses show that SAT5 cross-loads with buyer attitude toward the
consignor (seller), in fact, demonstrating a significantly higher loading on the buyer
attitude toward the consignor (seller) construct, than on the satisfaction construct. The
wording of this item aligns well with the other buyer attitude toward the consignor
(seller) measures; consequently, several tests of convergent and discriminant validity are
conducted to detemiine whether to 1) leave the SAT5 scale item with satisfaction, 2)
reassign the item to buyer attitude toward the consignor (seller), or 3) to remove the item
from the analysis. Ultimately these tests suggest that the best course of action is to
remove the item, so SAT5 is also dropped.
The third iteration results in the final measurement model. This revision once
again significantly improves the CFl (0.979), GFl (0.920), RMSEA (0.053) and
CMIN/DF (1.706) model fit statistics. These key indicators reflect a parsimonious model
with good fit statistics. The construct factor loadings and measurement properties for the
final measurement model appear immediately below in Table 4.2. The complete AMOS
output related to the final model appears in Appendix H.
Table 4.2: Construct Factor Loadings and Measurement Properties
Factor Coefficient Calculated Variance Average
t-Value
Loading Alpha Reliability Extracted
ATH 0.919 0.919 0.791 0.027
HORSE2 0.913 20.221
HORSE3 0.861 18.543
H0RSE5 0.893 20.223
0.916 0.846 0.363
CONl 0.885 21.010
CON2 0.953 21.029
SATISFACTION 0.880 0.878 0.706 0.354
SATl 0.823 14.363
SAT2 0.893 16.498
SAT3 0.802 14.372
TRUST 0.930 0.869 0.225
TRUST2 0.947 19.853
TRUST3 0.917 28.467
COMMITMENT 0.845 0.295
COMMITl
COMMIT2 0.958 30.317
COMMITS 0.857 28.467
COMMIT4 0.942 28.472
BRAND 0.759 0.759 0.515 0.265
EQUITY
0.663 10.084
PAYMORE
0.655 8.657
AFFECT
0.822 10.081
CONIMPAC
Construct Reliability and Validity
In order to ensure that the tested eonstruets are separate, distinct and measuring
what they are supposed to be measuring, additional tests of construct reliability and
validity reliability are conducted. Construct reliabilities can be calculated in several
different manners to ensure the measure is free from random error (Hoyle, Harris and
Judd 2002). Proving measures reliable is a necessary precursor to establishing construct
validity which is the extent to which the constmct measures only the construct of interest
(Hoyle, Harris and Judd 2002). Two common approaches that are recommended in this
pursuit are establishing discriminant and convergent validity (Fomell and Larcker 1981;
Hoyle, Harris and Judd 2002; Peter and Churchill 1986). Discriminant validity refers to
the degree to which scale measures are distinct from, and fail to correlate with, the scale
items measuring other constructs. Convergent validity refers to the degree to which
different scale items measuring the same construct converge on that construct.
Construct Reliability
Construct reliability is assessed in three ways. The first way constmct reliability
is assessed is by calculating coefficient alpha values for all constructs with three or more
items (buyer attitude toward the horse, satisfaction, commitment, and seller brand
equity). The coefficient alpha values for all these constructs are greater than the
reeommended 0.70 threshold, ranging from 0.759 to 0.955. Reliability values were also
calculated using the formulae recommended by Carver and Mentzer (1999). Their
formulae permits testing two-item constructs, as well as provide a more accurate adjusted
measure of reliability than does Cronbach's Alpha. Calculated reliabilities range from
0.759 to 0.956, once again exceeding the 0.70 threshold for all constructs. Finally,
reliabilities are assessed through the variance extracted for each construct. The variance
extracted ranges from 0.515 to 0.869, all exceeding the recommended threshold value of
0.50 (Carver and Mentzer 1999). A summary of these values appears above in Table 4.2.
The complete summary of reliability measures appears in Appendix I.
Discriminant Validity
The first step in testing for discriminant validity is to run a six-factor confinnatory
factor analysis with the revised set of measures. The appropriate items load onto the six
latent constructs, with all the items exhibiting factor loadings greater than 0.60. These
lambda values appear in the reliability summary presented in Appendix I. The second
step involves running factor analyses on paired tests of measurement items representing
all possible combinations of the six latent constructs. In all instances the appropriate
items load on the appropriate constructs (refer to Appendix 1). The final step in testing
for discriminant validity is to compare and contrast each constmcfs variance extracted
verses its average shared variance with each of the other constructs. In all cases, the
variance extracted is significantly higher than is the average shared variance. A summary
of the variance extracted and average variance shared statistics appears above in Table
4.2. The complete discriminant validity statistical output appears in Appendix J.
Convergent Validity
The first step in testing for convergent validity is to look at the regression weights
of the measurement items on the six latent variables. All the regression weights are
greater than 0.60, in the appropriate direction, and demonstrate statistical significance.
The next step in testing for convergent validity involves comparing and contrasting the
average inter-factor correlations (shared variance) with the variance extracted for each of
the constructs (Fomell and Larcker 1981). The average shared variance ranges from
0.027 to 0.354, while the variance extracted ranged from 0.515 to 0.869, thereby again
supporting the unidimensionality of the constructs. The results from the convergent
validity tests appear in both Table 4.2 and in Appendix J.
Arriving at the Final Measurement Model
Although analysis results show higher-than-desired degrees of inter-factor
correlation between buyer-seller relationship quality and buyer attitude toward the
consignor, r = 0.551, and between buyer-seller relationship quality and seller brand
equity, r = 0.361,( refer to Appendix K for the factor correlations), these values are not
high enough to be problematic; consequently, valuable insight and explanatory capacity
can still be derived. A summary of the key fit statistics for each of the model iterations
appears immediately below in Table 4.3, while the final measurement model and the
complete test results for the model appear in Appendix H.
Table 4.3: Comparison of Key Measurement Model Fit Statistics
. .  v.:.: ' CFI RMSEA CMIN/DF.
Comparison Model 0.726 0.602 0.142 5.986
Base (Full) Model 0.760 0.663 0.136 5.578
Revision 1 (Dropped FI, PAYSAME, 0.924 0.838 0.090 2.996
TRUST 1, TRUST4, SAT4, H0RSE4)
Revision 2( Also Dropped TRUST5, 0.968 0.896 0.061 1.925
C0N3)
Revision 3 (Also Dropped SAT5, 0.979 0.920 0.053 1.706
HORSEl)
Confirming the Measurement Model
Once the information elicited from the struetural equation measurement model is
defined and utilized, a confirmatory factor analysis is rerun in SPSS, specifying six
factors and utilizing a Principal Component Analysis - Varimax with Kaiser
Normalization Rotation Method. The previously discussed deeision criteria are utilized,
and the remaining seventeen endogenous item measures meet these criteria, and therefore
are retained.
The six expected constructs of buyer attitude toward the horse (produet brand),
buyer attitude toward the consignor (seller), satisfaetion, trust, commitment, and
eonsignor (seller) brand equity appear as expected, with eaeh of the item measures
loading on the appropriate factor and demonstrating low factor loadings on the other
constructs. The six latent constructs and their respective item measures follow
immediately below.
Buyer Attitude toward the Horse - three item measuresf or ATH:
• H ORSE2 - Buyer Attitude toward the Physical Appearance of the Horse
• H 0RSE3 - Buyer Attitude toward the Conformation of the Horse
• H ORSES - Buyer Attitude toward the Athletic Potential of the Horse
Buyer Attitude toward the Consignor (Seller)-  two item measuresf or ATC:
• C ON 1 - Buyer Attitude toward the Consignor's (Seller's) Reputation
C0N2 - Buyer Attitude toward the Consignor's (Seller's) Abilities
Satisfaction - three item measures for SATIS:
• S ATl - Satisfaction with the Quality of the Consignor's (Seller's) Prospect
(Product) Offerings
• S AT2 - Satisfaction with the Quality of the Consignor's (Seller's) Provided
Service
• S AT3 - Satisfaction that this Consignor's Prospect (Product) Offerings were a
Good Value for the Price
Trust - two item measures for TRST:
• T RUST2 - Suspicious of this Consignor's Recommendations (reverse scored)
• T RUST3 - Questioned this Consignor's Integrity (reverse scored)
Commitment -f our item measures for COM:
• C OMMITl - Committed to the Relationship
• C 0MM1T2-  Planned to Maintain the Relationship Indefinitely
• C 0MM1T3 - Relationship Deserved the Maximum Effort to Maintain
• C 0MMIT4 - Cared about the Relationship Long-Term
Relationship Quality - nine item measures for RQual:
• S ATI - Satisfaetion with the Quality of the Consignor's (Seller's) Prospect
(Product) Offerings
SAT2 - Satisfaction with the Quality of the Consignor's (Seller's) Provided
Service
SATS - Satisfaction that this Consignor's Prospect (Product) Offerings were a
Good Value for the Price
TRUST2 - Suspicious of this Consignor's Recommendations (reverse scored)
TRUSTS - Questioned this Consignor's Integrity (reverse scored)
COMMIT 1 - Committed to the Relationship
C0MM1T2 - Planned to Maintain the Relationship Indefinitely
COMMITS - Relationship Deserved the Maximum Effort to Maintain
C0MMIT4 - Cared about the Relationship Long-Term
Consignor (Seller) Brand Equity - three item measures BrEq:
• P AYMORE - Willing to Pay More for this Consignor's (Seller's) Prospect
(Product) Offerings
•  AFFECT - Impact of this Consignor's Recommendation
• C ONIMPAC - Impact of this Consignor's (Seller's) Involvement
Hypothesis Testimi
Once the process to arrive at the final measurement model (i.e., the measurement
model which demonstrates the best fit, validity and reliability statistics) is complete, the
next step is to test the hypotheses using the full structural model. Figure 4.1 follows
immediately below and presents the full structural model.
Figure 4.1: Full Structural Model
m
ATH ATC
t. ^tSiS
SATI SF 
COM RQuali
m TRST
m
BrEq
m
Utilized variations of the structural model include a direct effects (unmediated)
model, a full mediation model, and a partial (base) mediation model. In the direct effects
model, the pathways between ATH and RQual (G17), between ATC and RQual (G27),
and between RQual and BrEq (B76) are set to zero. Setting these pathways equal to zero
allows us to see how significant the direct effects of ATH and ATC on BrEq are. In the
full mediation variation, the pathways between ATH and BrEq (G16) and between ATC
and BrEq (G26) are set to zero. Setting these pathways equal to zero allows us to see
how significant the indirect effects of ATH and ATC on BrEq through RQual are. In the
partial mediation model, none of the latent construct pathways are constrained.
The unmediated model provides poor fit statistics, while the partially and fully
mediated models provide acceptable and comparable fit statistics. Within the fully
mediated model, the path from ATH to RQual is insignificant (p = 0.233), while the paths
from ATC to RQual (p < 0.001) and from RQual to BrEq (p < 0.001) are highly
significant. Within the partially mediated model, the path from ATC to BrEq is
insignificant (p = 0.465), while the paths from ATC to RQual (p < 0.001), ATH to BrEq
(p = 0.003), RQual to BrEq (p = 0.004) are all significant at the 0.01 level, and ATH to
BrEq (p = 0.05) was significant at the 0.05 level. Hair, Anderson, Tatham and Black
(1998) have recommend using a more conservative significance level than 0.05 when
using the maximum likelihood estimate (for example, 0.025 or 0.01); however, the
critical ratio for the ATH to BrEq pathway meets the acceptable threshold of 1.96,
thereby suggesting the significance of the ATH to BrEq path.
The theory-grounded framework developed in Chapter 11 and supported through
higher statistical values in regard to model fit, coefficient values and variance explained
warrants the assessment of the posed hypotheses in regard to the partial mediation model.
A review and discussion of the hypotheses findings derived from these assessments
follows the summary of the results from the direct effects, full mediation and partial
JSTOMER # illlM I SPINE LETTERING SET OF DATE SENT
COVER COLOR PRINT DATE
PRINT COLOR ENTRY DATE
HEIGHT UPDATE
SN. / ISBN. BIND. REQ. PUB. FREQ.
F^LTF'.P-SPJ.TrFP
WIDTH SPINE
RELATIONSHIP
□ SAVE TITLE (NEW) □ PERMANENT CHANGE osAjjxx'x and
□ 00 NOT SAVE TITLE □ SAMPLE / RUB SENT BINDERY USE ONLY
BRAND EQUITY. ,
PID LEAF □ HAND TRIM □ SP. BK CUT
TITLE PAGE FRONT COVER □ STF (MACH) □ STUBBING
TABLE CONT. BACK COVER □ STF (HAND) □ SEPARATORS
INDEX AOS □ STF (SINGLE) □ MENDING
JSTRUCTIONS □ RECASE □ REM. STAPLES
^  I' j S V"£ kT X i.JAxj □ PKT / PAPER □ TRIM XRX. PGS.
'  ■. ■ □ MOUNT COVER □ FOLD XRX. PGS
□ AU AOS OUT □ ENTERING
□ IMPRINT □ PANEL LINES □ IMPRINT □ PANEL UNES
□ RECASE / NEW COVER □ HAND TRIM
□ SEW THRU FOLD □ BIND AS IS □ EXTRA TIME _ MIN.
□ ADHESIVE BIND □ POCKET / CLOTH REASON
□ MOUNT COVER □ POCKET / PAPER
TYPE OF MATERIAL - PLEASE DESIGNATE B I BAR CODE DATA
•I* □ STANDARDIZED BDOK (NO COLOR / FOIL CHOICE) ■ □ MUSIC BOOK / SCORE (CHOICE OF COLOR / FOIL)
□ CUSTOM BOOK (CHOICE OF COLOR / FOIL) □ FLEXBINO-S PAPERBACK (NO COLOR CHOICE)
□ THESIS (CHOICE OF COLOR AND / OR FOIL) □ FLEXBINO-M PAPERBACK (MYLAR) LIBRARY COPY
□ OTHER
mediation model tests that appear in Table 4.4. The complete statistical output related to
the hypotheses testing appears in Appendix L.
Table 4.4: Comparison of Mediation Results
Hypothesized Model
Direct Effects Full Mediation Partial Mediation
0.892 0.957 0.960
0.845 0.883 0.886
RMSEA 0.117 0.074 0.072
CMIN/DF 4.401 2.354 2.299
Est. C.R. Sig. Est. C.R. Sig. Est. C.R. Sig.
ATC^RQual .895 12.407 .000 .894 12.40 .000
ATH^RQual .055 1.192 .233 .093 1.960 .050
ATC^BrEq .602 6.732 .000 -.212 -.730 .465
ATH^BrEq -.137 -2.20 .028 -.224 -3.01 .003
RQual^BrEq .671 6.911 .000 .942 2.868 .004
Note: Critical Ratio (should be > ±1.96)
Hypothesis 1: Supported
Hypothesis 1 states that there is a positive direet effeet between buyer attitude
toward the seller( ATC)a nd buyer-seller relationship quality (RQual). The results
support this hypothesis at the 0.01 level.
Hypothesis 2: Supported
Hypothesis 2 states that there is a positive direct effect between buyer attitude
toward the product( ATH)a nd buyer-seller relationship quality (RQual). The results
support this hypothesis at the 0.05 level.
Hypothesis 3: Not Supported
Hypothesis 3 states that there is a positive direct effect between buyer attitude
toward the seller( ATC)a nd seller brand equity (BrEq). The results do not support this
hypothesis (p = 0.465).
Hypothesis 4: Not Supported
Hypothesis 4 states that there is a positive direet effeet between buyer attitude
toward the product( ATH)a nd seller brand equity (BrEq). The results do not support the
positive direct effect between buyer attitude toward the product( ATH)a nd seller brand
equity (BrEq), but rather a significant negative direct effect between buyer attitude
toward the product( ATH)a nd seller brand equity (BrEq) at the 0.01 level.
Hypothesis 5: Supported
Hypothesis 5 states that there is a positive direet effeet between buyer-seller
relationship quality (RQual) and seller brand equity (BrEq). The results strongly support
this hypothesis at the 0.01 level.
Hypothesis 6: Supported
Hypothesis 6 states that buyer-seller relationship quality (RQual) mediates the
effect of buyer attitude toward the seller( ATC)o n seller brand equity (BrEq). The
results strongly support this hypothesis at the 0.01 level; with buyer-seller relationship
quality (RQual) fully mediating the path between buyer attitude toward the seller( ATC)
and seller brand equity (BrEq).
Hypothesis 7: Supported
Hypothesis 7 states that buyer-seller relationship quality (RQual) mediates the
effect of buyer attitude toward the product( ATH)o n seller brand equity (BrEq). The
results support this hypothesis at the 0.05 level; with buyer-seller relationship quality
(RQual) partially mediating the negative direct path between buyer attitude toward the
product( ATH)a nd seller brand equity (BrEq).
Summary
Chapter IV discusses the process utilized to arrive at the best possible
measurement model, as well as presents the data analysis and results. An iterative
process is utilized, resulting in the removal of ten of the original twenty-seven scale items
(HORSEl, H0RSE4, C0N3, SAT4, SATS, TRUST 1, TRUST4, TRUSTS, FI and
PAYSAME). The result is a measurement model that provides good fit statistics, as well
as construct validity and reliability. The next step is to test three competing model
alternatives, a direct effect (unmediated) model, a fully mediated model, and a partial
mediation model. The hypothesized partial mediation model provided the best model fit
statistics (CFl = 0.979. GFl = 0.920, RMSEA = 0.0S3 and CMIN/DF = 1.706),
coefficient values and variance explained.
Structural analysis results fail to support the positive direct effect relationships of
buyer attitudes toward the seller's corporate and product brands on seller brand equity.
Hypothesis 3 states that buyer attitude toward the seller( ATC) would have a positive
direct effect on seller brand equity (BrEq). Structural analysis results failed to support
this hypothesis. The analyses reveal that the reason the path is not significant, is that
buyer-seller relationship quality fully mediates the path between buyer attitude toward
the seller( ATC) and seller brand equity (BrEq) in support of Hypothesis 6. Hypothesis 4
states that buyer attitude toward the product( ATH) would have a positive direct effect on
seller brand equity (BrEq). Structural analysis results also fail to support this hypothesis.
The analysis reveals that the reason this hypothesis is not supported is because this path
demonstrates a significant negative rather than positive relationship. This finding
suggests that as the buyer feels better about the product offering, as reflected through his/
her attitude toward the attributes of product brand, buyer perceptions concerning the
importance of the seller's involvement become less important (i.e., the seller brand carries
correspondingly less value to the buyer - thereby lessening the effect of seller brand
equity within this dyad). The results do however show that buyer-seller relationship
quality (RQual) partially mediates the path between buyer attitude toward the product
(ATH)a nd seller brand equity (BrEq) in support of Hypothesis 7. The final model
depicting the theoretical constructs and path weights appears immediately below in
Figure 4.2.
Figure: 4.2: Final Model
Buyer
Attitude
toward the
Seller's
Corporate
Brand
, ( ATC) ,
Buyer-Seller
Seller's Brand
Relationship Equity
Quality (BrEq)
^(RQual)_^
Buyer
Attitude
toward the
Seller's
Product
Brand
, ( ATH) . : I nsignificant path estimates
: Significant path estimates
Chapter V next discusses the conclusions and contributions of these findings and
provides recommendations for future research.
CHAPTER V
CONCLUSIONS AND CONTRIBUTIONS
The intangible nature of certain speculative offerings (for example, graduating
Ph.D. candidates, college and professional athletes, prospective university students,
venture capital, real estate, high-end wine and artistic offerings) makes it difficult to
determine outcome quality until some future point of time - if at all. This research
examines one such industry from the business-to-business context.
Previous research dealing with building brand equity implies that brand equity is
directly affected in a positive manner as buyers feel better about a particular seller's
corporate brand. This research presents a different lens with which to view the building
of seller brand equity, suggesting that instead buyer attitude toward the seller serves to
improve buyer-seller relationship quality, and that it is buyer-seller relationship quality
that facilitates the building of seller brand equity. These research findings also indicate
that seller brand equity is directly affected in a negative manner as buyer attitude toward
the seller's product offering increases. This is an interesting concept that has not received
much in the way of attention in the literature. This concept translates to the notion that as
the seller does a more effective job offering the products that prospective buyers desire,
that his or her buyer-attributed value within the exchange is marginalized, thereby
diminishing seller brand equity. It also translates to the notion that as the buyer feels less
confident about the product offering, that the value attributed to the role of the seller is
enhanced, thereby increasing seller brand equity. This concept could have great
signifieance across a wide range of B2B and B2C business situations.
These research findings also indicate that buyer-seller relationship quality not
only directly influences seller brand equity, but that it also serves a facilitating function.
Buyer-seller relationship quality fully mediates the positive direct path between buyer
attitude toward the seller and seller brand equity. Buyer-seller relationship quality also
partially mediates the negative relationship between buyer attitude toward the product
and seller brand equity.
Conclusions
Existing relationship marketing research suggests that buyer-seller relationships
constitute an important enabling resource which improves performance outcomes.
Although relationship marketing theories diseuss the importance of developing quality
customer relationships, our knowledge of buyer-seller relationship quality is still far from
complete (Huntley 2006; Johnson and Seines 2004; Ulaga and Eggert 2006). This is
particularly evident when discussing buyer-seller relationship quality in business-to-
business settings. The motivation behind this study was to fill gaps in the existing
relationship marketing and branding literatures regarding buyer-seller relationship quality
and its role in influencing seller brand equity in a business-to-business context. This
study contributes to the existing relationship marketing and brand management
discussions by developing a theory-grounded conceptual framework concerning the
mediating role of buyer-seller relationship quality between buyer attitudes toward sellers'
corporate and product brands, and seller brand equity in a business-to-business setting.
The Thoroughbred consignment industry was chosen as the business-to-business
context for this study. Buyers of Thoroughbred horses were surveyed regarding the
previous purchase of a racing prospect (product brand) from a Thoroughbred consignor
(seller brand). Respondents were asked to rate scale items related to this encounter that
measured buyer attitude toward the seller( ATC), buyer attitude toward the product
(ATH), buyer-seller relationship quality (RQual), and seller brand equity (BrEq).
Hypotheses of the relationships between these constructs were tested using structural
equation modeling( SEM)t o analyze data collected from 249 usable surveys.
Stmctural equation modeling results reveal that buyer-seller relationship quality
fully mediates the positive direct path between buyer attitude toward the seller and seller
brand equity, and partially mediates the negative direct path between buyer attitude
toward the product and seller brand equity. The results of this study therefore suggest
that within speculative and competitive business climates, the role of buyer-seller
relationship quality in mediating the relationship between buyer attitudes toward tbe
seller's corporate and product brands, and seller brand equity is important, but that it
varies significantly depending upon which brand attitude is being studied.
Study results imply that buyer attitude toward the seller's corporate brand is
important in generating seller brand equity, but that it is the quality of the buyer-seller
relationship that facilitates that occurrence. The results also support the notion that buyer
attention is largely on the product within this uncertain context. This concept is critical
and conveys two important ideas. First, in uncertain contexts as exemplified by the
Thoroughbred industry, that the seller's corporate brand is highly valued by buyers
whose attitudes toward the product brand are less certain in nature and/ or that are shaped
in large part because of the involvement of the particular seller. Second, this suggests
that as the buyer's attitude toward the product improves (i.e., the buyer feels more
eonfident about the product), the value attributed to the role of the seller diminishes.
These findings are discussed in more detail in the following section, along with a
discussion of the other study findings and the implications of each. A discussion of the
contributions this study provides follows in the next section, followed by this study's
limitations and a discussion of future research opportunities.
Findings and Implications
This study's findings were derived from the analysis of seven hypotheses that
were elicited from a theory-grounded conceptual framework concerning the mediating
role of buyer-seller relationship quality between buyer attitudes toward the seller's
corporate and product brands, and seller brand equity in a business-to-business eontext.
The following subsections present a brief discussion of the postulated hypotheses,
findings and implications.
Hypothesis 1
HI) Buyer attitude toward the seller's corporate brand has a positive direct effect on
buyer-seller relationship quality. Supported
Hypothesis 1 stated that there would be a positive direct effect between buyer
attitude toward the seller( ATC)a nd buyer-seller relationship quality (RQual). The
results strongly support this hypothesis at the 0.01 level.
Buyer attitude toward the seller consisted of two measurement items, 1) buyer
attitude toward the consignor's (seller's) reputation( CON 1), and 2) buyer attitude toward
the consignor's (seller's) abilities (CON2). These two items deal with buyer evaluation of
the seller's coiporate brand. The author's experience in the industry and feedback from
industry experts suggests that buyer attitude toward the seller is an important factor for
many buyers. Empirical results supported this position, as buyer attitude toward the
seller was highly predictive of buyer-seller relationship quality. Even when buyers
indicated that relationships were not important to them, many indicated that their attitude
toward certain consignors (sellers) would preclude them from considering those
consignors' racing prospect (product brand) offerings. These findings suggest that within
speculative and competitive business-to-business environments that sellers should focus
on presenting favorable corporate brand messages to current and prospective buyers (i.e.,
convey a favorable corporate brand image) if they want to develop higher quality
relationships with these entities.
Hypothesis 2
H2) Buyer attitude toward the seller's product brand has a positive direct effect on
buyer-seller relationship quality. Supported
Hypothesis 2 stated that there would be a positive direct effect between buyer
attitude toward the product( ATH)a nd buyer-seller relationship quality (RQual). The
results supported this hypothesis at the 0.05 level.
Hypothesis testing provided support for the direct effect relationship; however the
support for this hypothesis was only achieved at the 0.05 level. The author's experience
in the industry and feedback from industry experts suggests that a reason that the strength
of the support for this hypothesis was not stronger is that the Thoroughbred consignment
industry represents a situation where the seller's image is more constant and predictable
than are the products he or she offers. In essence, Thoroughbred consignors (sellers)
offer a heterogeneous array of products, where no two products are ever the same.
The Thoroughbred racing prospect product category is markedly different from
the product brands that have been studied within the consumer goods context (for
example, soft drinks) as well as the industrial goods context (for example, lighting
fixtures) where product variability is trivial to non-existent. This product category is
even substantially more variant and uneertain than are the intangible, experiential and
credential brands that have been studied within B2C and B2B services contexts (for
example banking, medical and logistics services) where consistency is both desired and
rewarded. However, in spite of this being a variable product class, the context is very
representative of other speculative contexts where product brands demonstrate a great
deal of heterogeneity, for example, the Ph.D. student market, college and professional
athletics, university admissions, venture capital markets, real estate markets, the wine,
art, music and entertainment industries.
The results support this hypothesis through the positive direct effect between
buyer attitude toward the product( ATH)a nd buyer-seller relationship quality (RQual) at
the 0.05 level. This suggests that as the buyer feels better about the product offering that
they are considering, they also feel better about the relationship they have with the seller.
With that said the 0.05 support for this hypothesis indicates that while buyers may seek to
develop higher quality relationships with some of the consignors they purchase from;
they are sparing in doing so.
Hypothesis 3
H3) Buyer attitude toward the seller's corporate brand has a positive direct ejfect on
brand equity. Not Supported
Hypothesis 3 stated that there would be a positive direct effect between buyer
attitude toward the seller( ATC)a nd seller brand equity (BrEq). The results do not
support this hypothesis (p = 0.465).
Existing brand attitude and equity literature supports the direct effect of buyer
brand attitudes on the generation of seller brand equity. It was hypothesized in this study
that there would be both a positive direct effect of buyer attitude toward the seller's
corporate brand on seller brand equity (H3), and a positive indirect effect of buyer
attitude toward the seller's corporate brand on seller brand equity through buyer-seller
relationship quality (H6). The mediating effect that will be discussed in detail in the H6
subsection over-road the direct effect that was hypothesized in H3. This means that
buyer attitude toward the seller is not sufficient on its own to explain the generation of
seller brand equity.
Branding theory suggests that the attitudes buyers hold toward the seller help to
predict a variety of criterion variables, including the accrual of brand equity. This
research suggests that the direct relationship between buyer attitude toward the seller and
seller brand equity is not optimal, and that buyer-seller relationship quality enhances this
process. This means that within uncertain business-to-business contexts such as the
Thoroughbred consignment industry, that buyer-seller relationship quality (and perhaps
other mediating variables) can enhance the building of seller brand equity.
Hypothesis 4
H4) Buyer attitude toward the seller's product brand has a positive direct effect on
brand equity. Not Supported
Hypothesis 4 stated that there would be a positive direct effect between buyer
attitude toward the product( ATH)a nd seller brand equity (BrEq). Data analysis results
do not support a positive direct effect between buyer attitude toward the product( ATH)
and seller brand equity (BrEq), but rather strongly support the negative direct effect
between buyer attitude toward the product( ATH)a nd seller brand equity (BrEq) at the
0.01 level.
Although the existing branding literature provides support for the positive direct
effect relationship between buyer attitude toward the product( ATH)a nd brand equity
(BrEq), the results from this research did not support this hypothesis. In fact, and quite to
the contrary, the results provide support for the negative direct effect relationship
between buyer attitude toward the product( ATH)a nd seller brand equity (BrEq). The
directional discrepancy between the hypothesis and actual finding is at first confusing,
however, a review of the branding literature suggests that the author's use of the specific
seller brand equity construct was the underlying cause for this inconsistency. The
common brand equity language would justify the positive direct relationship hypothesis;
however, this study investigates brand equity accrual at the corporate brand level.
Consequently, the study findings make sense. This translates to the notion that the seller
adds more to the perceived value of product offerings the buyer is uncertain about, than
they do to product offerings about which the buyer is confident.
Feedback from industry experts also suggests that the reason a negative direct
effect was found was because if the buyer feels confident about the product he or she is
purchasing, which tends to be the case more often than not when making a purchase
commitment of this magnitude, that the buyer will rely predominantly on their own
judgment when making the purchase decision. What this means in the Thoroughbred
consignment industry is that buyers that have more positive attitudes towards the racing
prospects (product brands) they are considering, don't look for reinforcement from fhe
consignor (seller). It further means that they don't attribute successful racing prospect
(product brand) purchases to anything the consignor (seller) has done, but rather their
own "horsemanship" skills. This once again is logical and intuitive. If the buyer feels
very confident about the racing prospect (product offering), it doesn't matter who is
selling the horse. This further suggests that consignors (sellers) not interested in
committing resources to building better quality customer relationships should instead
focus on offering racing prospects (product brands) that possess the attributes most
desired by their target customers.
Hypothesis 5
H5) Buyer-seller relationship quality has a positive direct effect on brand equity.
Supported
Hypothesis 5 stated that there would be a positive direct effect between buyer-
seller relationship quality (RQual) and seller brand equity (BrEq). The results strongly
support this hypothesis at the 0.01 level.
Buyer-seller relationship quality consisted of nine measurement items that were
pulled from the three dimensions that constitute relationship quality, satisfaction, trust
and commitment. These nine measurement items are: 1) satisfaction with the quality of
the consignor's (seller's) prospect (product brand) offerings - SATl, satisfaction with the
quality of the consignor's (seller's) provided service - SAT2, satisfaction that this
consignor's prospect (product brand) offerings were a good value for the price - SAT3,
suspicious of this consignor's recommendations (reverse scored) - TRUST2, questioned
this consignor's integrity (reverse scored) - TRUSTS, committed to the relationship -
COMMIT!, planned to maintain the relationship indefinitely - COMM1T2, relationship
deserved the maximum effort to maintain - COMMITS, and cared about the relationship
long-term - C0MM1T4. These nine items deal with buyer perceptions of the quality of
the relationship the buyer has with the consignor (seller) as reflected by his or her overall
satisfaction, trust and commitment.
The role of these buyer-seller relationship quality dimensions are well supported
in the literature as facilitators of buyer value perceptions. The author's experience and
feedback with industry experts suggest the same holds true in the Thoroughbred industry.
In spite of the large dollar volume spent on the purchase of racing prospects (product
brands) every year, the Thoroughbred industry is not a large industry in terms of the
number of buyers and sellers. Consequently, the potential for buyers and sellers to
perform more than an isolated transaction always exists. Further, the limited number of
actors in the industry also facilitates the dissemination of infonnation beyond a specific
relational dyad. Consequently, the individual and collective attainment of buyer
satisfaction, trust and commitment in the seller can be important relational resources for
the consignor (seller) to build brand equity at the corporate brand level. This suggests
that in speculative and competitive business-to-business contexts, that buyer-seller
relationship quality is an important factor in building seller brand equity.
Hypothesis 6
H6) Buyer-seller relationship quality mediates the effect ofb uyer attitude toward the
seller's corporate brand on seller brand equity. Supported
Hypothesis 6 stated that buyer-seller relationship quality would mediate the
relationship between buyer attitude toward the seller( ATC) and seller brand equity
(BrEq). The results strongly support this hypothesis at the 0.001 level.
This suggests that buyer attitude toward the seller not only helps to shape the
buyer-seller relationship, as was discussed in HI, but that the relationship actually
facilitates the generation of seller brand equity that occurs as the result of the buyer's
attitude toward the seller. Within the Thoroughbred consignment industry, this means
that while the buyer's attitude toward the consignor's (seller's) capabilities is important,
that it actually serves to build higher quality buyer-seller relationships, which in turn
serve to build seller brand equity. This is an important concept as it suggests a definitive
brand development strategy that business-to-business sellers can use to build brand equity
at their corporate brand level. This further suggests that within speculative and
competitive B2B contexts, sellers should look to focus on developing high quality
relationships with current and prospective buyers if they want to generate brand equity at
the corporate brand level.
Hypothesis 7
H7) Buyer-seller relationship quality mediates the effect ofb uyer attitude toward the
seller's product brand on seller brand equity. Supported
Hypothesis 7 stated that buyer-seller relationship quality would mediate the
relationship between buyer attitude toward the product( ATH)a nd seller brand equity
(BrEq). The results support this hypothesis at the 0.05 level.
This suggests that buyer attitude toward the seller's product brand not only helps
to shape the buyer-seller relationship, as was discussed in H2, but that the relationship
can actually change the negative direct effect that buyer attitude toward the product has
on generating seller brand equity. This is an interesting finding as it suggests that buyer-
seller relationship quality can add value to buyers, even when they feel confident about
the product offering they are evaluating. It is further interesting because it implies that as
the buyer exhibits a less favorable attitude toward the racing prospect (product brand),
that the value attributed to the consignor (seller) increases. This suggests that while
buyers in this context exhibiting positive attitudes toward the product offerings they are
considering are hesitant to place value on the role of the seller, they do tend to attribute
value to a seller with which they have a good relationship. This means that while the
buyer's attitude toward the product is important in generating sales, that building high
quality buyer-seller relationships is important for generating seller brand equity.
Findings and Implications Summary
The final structural equations model shows that buyer attitude toward the seller
(ATC)a nd buyer attitude toward the product( ATH)b oth have positive direct effects on
buyer-seller relationship quality (RQual), while neither has positive direct effects on
seller brand equity (BrEq) - although the rationale for the absence of positive direct
effects is considerably different for the two. The reason a positive direct path between
buyer attitude toward the seller( ATC) and seller brand equity (BrEq) does not exist is
likely because buyer attitude toward the seller is not sufficient on its own to explain the
generation of seller brand equity. The reason a positive direct path between buyer
attitude toward the product( ATH)a nd seller brand equity (BrEq) does not exist is
because an inverse relationship actually exists.
Hypothesis testing supports that buyer-seller relationship quality (RQual) has a
positive direct effect on seller brand equity (BrEq). The structural model also reveals that
buyer-seller relationship quality (RQual) fiilly mediates the positive direct effect of buyer
attitude toward the seller( ATC)o n seller brand equity (BrEq), and partially mediates the
negative direct path between buyer attitude toward the product( ATH)a nd seller brand
equity (BrEq). As a result, three of the five paths that were hypothesized were supported,
and five of the seven hypotheses were supported.
The results of this study have significant implications for sellers within
speculative and uncertain business environments where product brands demonstrate
variability. The results suggest that buyer attitude toward the seller significantly
contributes to generating seller brand equity tlirough buyer-seller relationship quality,
while buyer attitude toward the product also contributes to the building of seller brand
equity through the quality of the buyer-seller relationship, though to a significantly lesser
extent when buyer attitude toward the product is high.
The finding that buyer attitude toward the product is inversely predictive of seller
brand equity in this context is particularly interesting, because it suggests that in spite of
the expense and uncertainty attached to the purchase of Thoroughbred racing prospects
(product brands), the role of the seller is of little consequence within buyer purchasing
decisions when the seller's attitude toward the racing prospect (product brand) is good. It
further suggests that the role of the seller is of consequence within buyer purchasing
decisions when the seller's attitude toward the racing prospect (product brand) is less
favorable. This is discussed in more depth within the sections that follow.
Theoretical Contributions
The objective within this research agenda is to further our understanding of buyer-
seller relationship quality and to fill gaps within the existing theory by contributing to our
overall understanding of the role buyer-seller relationship quality plays in influencing the
paths between buyer attitudes toward the seller's coiporate and product brands, and seller
brand equity in a business-to-business context. The theoretical and practical implications
of this research are significant, as this study serves to fill gaps within the relationship
marketing and branding literatures, while also contributing to our overall body of
knowledge concerning buyer-seller relationships.
First, this study provides insight into how buyer-seller relationship quality
influences the relationship between buyer brand attitudes and seller brand equity, by
empirically examining the interactions between and among these important determinants
of customer buying behavior. Structural equation modeling results show that buyer-seller
relationship quality fully mediates the positive direct path between buyer attitude toward
the seller and seller brand equity, and partially mediates the negative direct path between
buyer attitude toward the product and seller brand equity. This research thus serves to
validate the role of husiness-to-husiness buyer-seller relationship quality in influencing
the direct paths between buyer attitudes toward sellers' corporate and product brands, and
seller brand equity. Such contributions are essential to establishing a theory's external
validity and generalizability (Aaker 1997; Calder, Phillips and Tybout 1982; McGrath
and Brinberg 1983).
Second, this research extends our knowledge of business-to-business branding,
particularly as it relates to the corporate branding efforts of sellers that offer products
exhibiting aspects of both goods and services, i.e., hybrid products (Anustrong and
Kotler 2005). Such a pursuit is a valuable contribution because virtually all discussions
of branding have been framed in consumer goods contexts, and consequently our
knowledge of B2B branding is far less developed and understood than is our knowledge
of B2C branding (Webster and Keller 2004). This research thus helps illuminate the
process of building seller brand equity in B2B contexts, as well as in contexts exhibiting
this dual nature. This is intriguing, especially when product offerings demonstrate high
degrees of speculation and uncertainty (for example, the Ph.D. student market, college
and professional athletics, university admissions, venture capital markets, real estate
markets, the wine, art, music and entertainment industries, etc.).
Finally, this study is designed with the intent of theory elaboration - i.e., to apply
and extend existing theory into new settings or contexts (Lee 1999). This technique
provides the dual benefits of supporting the development of a conceptual framework that
furthers our knowledge of buyer-seller relationship quality, while also introducing a new
research context. The Thoroughbred industry is a new contextual area for business
research, as the limited studies that have been published related to Thoroughbreds have
almost universally involved gambling rationale and decision-making processes. The use
of theory elaboration is thus an important contribution as it supports the development of a
theory-grounded buyer-seller relationship framework, as well as research in a novel and
exciting contextual area, i.e., the Thoroughbred industry. A positive byproduct is that
this study will also serve to provide support for a valuable but relatively overlooked and
underutilized research technique, i.e., theory elaboration (Lee 1999).
Managerial Contributions
In addition to the theoretical contributions this research provides, this research
also affords several applied contributions. In terms of practical contributions, this
research will help business-to-business firms, and more specifically Thoroughbred
consignment firms, by empirically establishing the role of buyer-seller relationship
quality as a mediator of the relationship between buyer attitudes toward sellers' corporate
and product brands, and seller brand equity. This research thus serves to infonn B2B
fmns by affirming the role that business-to-business buyer-seller relationship quality has
as a facilitator in converting buyer attitudes toward sellers' corporate and product brands
into seller brand equity. What this means within the Thoroughbred consignment industry
is that building high quality buyer-consignor (seller) relationships can help consignors
(sellers) to generate higher levels of brand equity, while also helping buyers in their
decision making processes.
Another valuable managerial contribution that comes out of this study is that this
research suggests that if sellers (consignors) elect not to build high quality customer
relationships, they should focus on offering the products that demonstrate the attributes
and qualities that their target customers will value the most. The rationale behind this
logic is for the buyers who have the most positive attitudes toward the product brand
offering (this would be the racing prospect in the Thoroughbred consignment industry),
that the seller (consignor) themselves does not add enough additional value to develop a
differential effect.
Finally, this research provides both the impetus and the roadmap for business-to-
business firnis to understand the role buyer-seller relationship quality plays or can play,
within their marketing and branding strategies. This practical insight will help these
finns, and particularly sellers within the Thoroughbred consignment industry, to
understand how to best structure their customer relationship portfolios.
Study Limitations
There are several limitations concerning the design of this study. First, there are
study design concerns as this study relied on review by knowledgeable colleagues and
cognitive interviews with industry experts in order to inforni the survey instrument. This
technique essentially used expert opinion in place of conventional pretesting methods.
While this technique improves the perceived "legitimacy" of the questionnaire and makes
it seem more "real-life" (Elsbach 1994), it does lose the benefits associated with utilizing
whole scales and grouped item measures that have previously proven valid and reliable.
As the result of these choices, a number of measurement issues and limitations
exist. First, the choice of measurement items is a concern. The final survey questions
were assembled in a patchwork fashion. While this strategy helped to enhance the
perceived legitimacy and real-to-life nature of the survey instrument, it also introduced
significant item and factor correlation issues. These correlation issues reduced the
number of survey measures from an initial count of twenty-seven, to a final count of
seventeen. Additionally, high degrees of inter-factor correlation still existed between
buyer attitude toward the consignor (seller), satisfaction and trust, as well as between
commitment and seller brand equity. There are also directionality concerns. This study
developed a theory-grounded framework utilizing buyer-seller relationship quality as a
mediating construct between buyer attitudes toward the seller's corporate and product
brands, and seller brand equity. While this framework fit the data well, so do did the first
order model where satisfaction, trust and commitment were independent mediating
variables, as well as the alternate model that utilized buyer-seller relationship quality as
the predictor variable, and buyer attitude toward the seller and buyer attitude toward the
product as mediating variables. A final measurement limitation that exists is that this
study investigated four constructs; consequently, omitted variable concerns also exist.
Limitations also exist related to the chosen sampling frame and methodological
approach. The utilized sample drew on the comprehensive Thoroughbred Owners and
Breeders Association( TOBA) Membership Directory. Per TOBA management, roughly
half of the members listed in the directory would not be appropriate for this study.
Consequently, sampling concerns do exist. Survey limitations also exist, particularly in
regard to survey response biases, such as nonresponse bias, false reporting bias and
potential dependency issues which could arise as a result of asking respondents to
complete surveys on both a strong and weak relationship. Potential response bias issues
were addressed as best as possible within the study, however it is impossible to be sure
they do not exist.
Another potential limitation of this research design is that this study relies on
buyer perceptions to investigate buyer-seller relationship dyads. This single source
approach introduces limitations associated with having one-way perceptual measures for
two-way relational exchanges. Finally, this study focuses solely on the thoroughbred
consignment industry. While it is suggested that the findings of this study are
generalizable beyond this industry, particularly within speculative and competitive
business environments, the unique aspects of the Thoroughbred consignment industry
may cast doubt regarding the extendibility of the findings.
Future Research
There is a wide array of future research opportunities that are derived by
addressing the limitations listed immediately above. First refining and extending the
theory-grounded framework developed in this paper through the incorporation and testing
of other measurement items offers significant opportunity. There are significant
opportunities to address the measurement issues articulated in the previous section, and to
refine the buyer-seller relationship quality measurement scale. Additionally, the testing
of the framework in other contexts would also be very interesting. Interesting areas with
which to test the framework include both other eontexts where the product offerings
exhibit aspects of goods and services, as well as traditional business-to-business and
business-to-eonsumer contexts. Proximate expansion of the sample would also be very
interesting. Consignors are not restrieted to acting as sellers. It would be very interesting
to investigate the developed framework in regard to their supply-side relationships.
This research utilizes an interdisciplinary lens and draws upon several marketing
and management strategy literatures to suggest that buyer-seller relationship quality is an
important faetor when looking to translate buyer attitudes toward sellers' corporate and
product brands into seller brand equity. In exploring the role of buyer-seller relationship
quality, this study utilized three theory-derived buyer-seller relationship quality
dimensions - satisfaction, trust and commitment (Garbarino and Johnson 1999; Johnson
and Seines 2004; Ulaga and Eggert 2006). However, it is possible that the independent
dimensions of buyer-seller relationship quality, satisfaction, trust and commitment are
better observed as separate variables than they are as an aggregate eonstruet. On the flip
side, it is also plausible that buyer-seller relationship quality as a second-order construct
is the more effeetive mediating variable choice and that other variables might also be
relevant constituent dimensions of buyer-seller relationship quality. As a result, future
researeh could draw upon measures used in other contexts to see if they would add
explanatory or predictive capacity. For example, social exchange theory suggests that
reciproeity can serve as an important variable under certain conditions. From the buyer's
standpoint, overall satisfaction, trust and commitment are outwardly focused. It is
conceivable that in order for buyers' to feel higher levels of relationship quality, that they
must also perceive reeiprocity in the form of certain seller behaviors (Sawhney and Zabin
2002). Consequently, additional or alternative sources of buyer-seller relationship quality
miglit be an interesting direction for future research.
Additionally, this study utilized a unique perspective to investigate the effects of
buyer-seller relationship quality on brand equity. This study used the concept of seller
brand equity. It would also be very interesting to explore other aspects and variable
combinations of brand equity. The investigation of brand equity effects on the actual
sales, race and breeding performance of the purchased horse (product brand) would be
particularly interesting. Further, the finding that buyer attitude toward the product is
inversely predictive of seller brand equity in this context is a compelling finding. Future
research could further explore this concept by investigating the inverse relationship that
was demonstrated in this context between buyer attitude toward the seller's product brand
and seller brand equity. This finding could have enormous implications for the allocation
of marketing and branding resources.
Thoroughbred consignment is but one business-to-business context, and at that is
speculative and uncertain in nature. Consequently, studying the role buyer-seller
relationship quality plays in other industries and contexts would be highly informative. It
would also be interesting to explore the role buyer-seller relationship quality plays in
influencing the paths between buyer attitude toward the seller's corporate and product
brands, and seller brand equity accrual within traditional tangible good contexts, where
product variability is low and intangible factors are largely external to the actual product
offering. Consequently, testing the generalizability of the theory and dissertation model
in other contexts, both B2B and B2C, and tangible and intangible, would be interesting.
Lastly, thoroughbred consignors can find themselves in the interesting position of
acting as both buyers and sellers (i.e.. Thoroughbred pinhooking). While both sides of
the equation are important, this study focused on just the downstream side of the value
chain, with consignors behaving as sellers. Consequently, a valuable extension of this
study would be to explore specialty consignors' (pinhookers') supply-side relationships to
detennine the impact of buyer-seller relationship quality when the consignor (pinhooker)
behaves as the buyer, rather than as the seller.
.1 ii V «
§*''■'i'.^V
APPENDICES
APPENDIX A
SURVEY INSTRUMENT
LUNDQUIST
COLLEGE OF
Survey of Thoroughbred Racing
BUSINESS
Prospect Buyers
University of Oregon
The two parts of this survey have been designed to take approximately 10 minutes to
complete. Please check, circle or provide the responses that best reflect your opinion and
level of agreement with the following questions and statements.
1. Approximately how many years have you been involved in the Thoroughbred
industry (please check only one)?
Years
For questions 2-15, think about a strong relationship you had with a consignor (selling
agent/ pinhooker) at the time of a specific Thoroughbred purchase. The term consignor
will be used throughout the survey to represent this individual. Base your responses for
the following questions on this purchase (to the best of your recollection). ALL
RESPONSES WILL BE KEPT COMPLETELY CONFIDENTIAL AND REPORTED
ONLY IN AGGREGATE.
2. What was the age category of this horse at the time of purchase (please check only
one).
Weanling
Yearling
2-Year-Old
3-Year-Old
Older Horse
3. What is the lineage of the purchased Thoroughbred racing prospect?
Sire: _
Dam:
Please answer each of the items in Question 4 using the following scale:
Very
Favorable
Favorable
7
4. Most people perceive attributes of their purchases favorably, but not always to the
same degree. At the time of purchase, how favorable were your perceptions of the
following attributes of this racing prospect?
a) Breeding cross/ pedigree
b) Physical appearance (size and muscularity)
c) Conformation
d) Disposition (attitude, temperament and competitiveness)
e) Athletic potential
Please answer each of the items in Questions 5-12 using the following scale:
Strongly Strong
Disagree Agre
1 I 2 I 3 I 4 I 5 I 6 I 7
5. At the time of purchase, ...
a) this consignor had a good reputation within the industry.
b) 1 felt good about this consignor's abilities.
c) 1 felt good about the involvement of this consignor with this racing prospect.
At the time 1 purchased this racing prospect, 1 was satisfied ...
a) with the quality of the racing prospects this consignor offered.
b) with the quality of the service this consignor offered.
c) that this consignor's offerings were a good value for the price paid.
d) with my working relationship with this consignor.
e) that this consignor was a good seller with which to do business.
7. At the time 1 purchased this racing prospect, ...
a) 1 could count on this consignor to be sincere.
b) 1 was suspicious of this consignor's recommendations,( reverse coded)
c) 1 questioned the integrity of this consignor, (reverse coded)
d) this consignor and I worked together to ensure the best outcomes for both of us.
e) this consignor was a relationship partner 1 trusted completely.
8. At the time I purchased this Thoroughbred racing prospect, the relationship I had with
this consignor, ...
a) was something to which I was very committed.
b) was something I planned to maintain indefinitely.
c) deserved my maximum effort to maintain.
d) was something I cared a great deal about long-term.
9. Had this Thoroughbred racing prospect been offered by the average consignor instead
of the consignor 1 purchased from, 1 would have been willing to pay the same amount
for this horse,( reverse coded)
10. At the time of purchase, 1 would have been willing to pay more for a racing prospect
sold by this consignor than for the same horse sold by the average consignor because
of what it meant to have this consignor's involvement (endorsement).
11. At the time of purchase, the involvement of this consignor helped in my decision to
purchase this racing prospect.
12.1 intend to consider Thoroughbred racing prospects offered by this consignor again in
the future.
Please answer Question 13 using the following scale:
Not Very Very
Applicable Weakly Strongly
0 1  2 6 7 
13. How strongly did the consignor recommend this racing prospect to you?
Please answer Question 14 using the following scale:
Not No Strong
Applicable Affect Affect
0 1 6 7 
14. How much did this recommendation affeet your decision to purchase this racing
prospect?

For questions 15-27, think about a weak relationship you had with a consignor (selling
agent/ pinhooker) at the time of a specific Thoroughbred purchase. The term consignor
will be used throughout the survey to represent this individual. Base your responses for
the following questions on this purchase (to the best of your recollection). ALL
RESPONSES WILL BE KEPT COMPLETELY CONFIDENTIAL AND REPORTED
ONLY IN AGGREGATE.
15. What was the age category of this horse at the time of purchase (please check only
one).
Weanling
Yearling
2-Year-Old
3-Year-Old
Older Florse
16. What is the lineage of the purchased Thoroughbred racing prospect?
Sire: _
Dam:
Please answer each of the items in Question 4 using the following scale:
Very
Favorable
Favorable
7
17. Most people perceive attributes of their purchases favorably, but not always to the
same degree. At the time of purchase, how favorable were your perceptions of the
following attributes of this racing prospect?
f) Breeding cross/ pedigree
g) Physical appearance (size and muscularity)
h) Confomiation
i) Disposition (attitude, temperament and competitiveness)
j) Athletic potential
Please answer each of the items in Questions 18-25 using the following scale:
Strongly Strongly
Disagree Agree
1 I 2 I 3 I 4 I 5 I 6 I 7
18. At the time of purchase, ...
a) this consignor had a good reputation within the industry.
b) I felt good about this consignor's abilities.
c) 1 felt good about the involvement of this consignor with this racing prospect.
19. At the time 1 purchased this racing prospect, 1 was satisfied ...
a) with the quality of the racing prospects this consignor offered.
b) with the quality of the service this consignor offered.
c) that this consignor's offerings were a good value for the price paid.
d) with my working relationship with this consignor.
e) that this consignor was a good seller with which to do business.
20. At the time 1 purchased this racing prospect, ...
a) I could count on this consignor to be sincere.
b) I was suspicious of this consignor's recommendations,( reverse coded)
c) 1 questioned the integrity of this consignor, (reverse coded)
d) this consignor and 1 worked together to ensure the best outcomes for both of us.
e) this consignor was a relationship partner 1 trusted completely.
21. At the time I purchased this Thoroughbred racing prospect, the relationship 1 had with
this consignor, ...
a) was something to which 1 was very committed.
b) was something 1 planned to maintain indefinitely.
c) deseiwed my maximum effort to maintain.
d) was something 1 cared a great deal about long-term.
22. Had this Thoroughbred racing prospect been offered by the average consignor instead
of the consignor 1 purchased from, 1 would have been willing to pay the same amount
for this horse,( reverse coded)
23. At the time of purchase, 1 would have been willing to pay more for a racing prospect
sold by this consignor than for the same horse sold by the average consignor because
of what it meant to have this consignor's involvement (endorsement).
24. At the time of purchase, the involvement of this consignor helped in my decision to
purchase this racing prospect.
25.1 intend to consider Thoroughbred racing prospects offered by this consignor again in
the future.
Please answer Question 26 using the following scale:
Not Very Very
Applicable Weakly Strongly
0 1 6 7 
26. How strongly did the consignor recommend this racing prospect to you?
Please answer Question 27 using the following scale:
Not No Strong
Applicable Affect Affect
0 1 6 7 
27. How much did this recommendation affect your decision to purchase this racing
prospect?
Please provide any comments you think are important.
Please complete Question 28 if you'd like a summary of the findings sent to you at the
completion of this study.
28. Please send me a summary of the findings at the completion of this study ...
a) by e-mail at the following address:
b) by U.S. Mail at the following address:
THANK YOU FOR YOUR VALUABLE INPUT!
P
'Tm
tf
•» * 1
APPENDIX B
COVER LETTER FOR SURVEY MAILING
«Date»
«First_Name» «Middle_Initial» «Last_Name»
«Stable_Name»
«Address»
«City», «State» «Zip»
Dear «Greeting»,
I am a Ph.D. Candidate in Marketing at the University of Oregon and am writing to ask
your help with my dissertation research. You were selected as a possible participant
because of your previous purchase of thoroughbred racing prospects. This survey will
take approximately 10 minutes to complete and is part of an effort to learn more about
buyer-consignor relationships and the factors that affect purchasing decisions in the
thoroughbred industry.
Your answers are completely confidential and will be released only in aggi'egate, where
your answers cannot be identified with you or the consignors with which you do
business. Additionally, in order to address any potential risks, all individuals' names will
be replaced with ID numbers and the list of names will be destroyed once the data
collection is complete. Your completion of this survey will serve as your consent, so
please retain a copy of this letter for your records. Please return your completed survey
in the accompanying postage-paid envelope at your earliest convenience. The target
return date is «two weeks hence»; however if you are not able to meet this date, please
do still complete and return the survey as soon as you're able. This survey is voluntary
and there is no penalty for refusing to complete it, though your assistance is greatly
appreciated. As a token of appreciation for your participation, a $25 donation will be
made to the Thoroughbred Retirement Foundation for each completed questionnaire (up
to $1,000), and you will have the option of receiving a summary of the research findings.
To receive a summary, simply indicate your desire to receive a copy in the final question
of your completed survey.
If you have any questions about this study, I would be happy to talk with you. Please
contact me by phone at (541) 515-0169 or by e-mail at amarquar@,uoregon.edu. or my
dissertation advisor Lynn Kahle by phone at (541) 346-3373 or by e-mail at
lkahle@,uoregon.edu. If you have any questions regarding your rights as a research
participant, please contact the Office of Human Subjects Compliance at the University of
Oregon at (541) 346-2510. Thank you for your help!
Sincerely,
Adam Marquardt
Ph.D. Candidate in Marketing
APPENDIX C
REMINDER AND THANK YOU POSTCARD
^ " -• •■v.
* £-^S r ^
-1- =
KS'
•*
REMINDER & THANK YOU POSTCARD
Last week, you were sent a questionnaire asking your opinions about buyer/ seller
relationships in the thoroughbred industry. If you have already returned the
questionnaire, please accept my sincere thanks. If not, please complete and return it at
your earliest convenience. I am especially grateful for your help because it not only
assists in my dissertation research, but it also benefits the Thoroughbred Retirement
Foundation, a very important and worthwhile cause.
If you did not receive the questionnaire, or if it was misplaced, please contact me at (541)
515-0169 or at amarquar@uoregon.edu. and 1 will send a replacement copy immediately.
Thank you very much for your help and best wishes for a fantastic 2007 season!
Adam Marquardt
Ph.D. Candidate in Marketing
APPENDIX D
FOLLOW UP LETTER FOR SURVEY MAILING
is* - ■
■ f i0:f . V 
«Date»
«First Name» «Middle Initial» «Last Name»
«Stable Name»
«Address»
«City», «State» «Zip»
Dear «Greeting»,
A few weeks ago I sent you a brief questionnaire asking your opinions concerning
aspects of buyer/ consignor (selling agent) relationships in the thoroughbred industry. If
you've returned the completed questionnaire, please accept my deepest thanks. However,
to the best of my knowledge it's not yet been received, and 1 am following up because of
the importance that your answers have in helping me to arrive at accurate and
representative results within this study.
Your answers are completely confidential and will he released only in aggregate, where
your answers cannot he identified with you or your consignor (selling agent) counterpart.
Protecting the confidentiality of your answers is very important to me, as well as to the
University of Oregon. Your completion of this survey will serve as your consent, so
please retain a copy of this letter for your records.
I would he most appreciative if you would fill out and return the questionnaire in the
provided envelope at your earliest convenience. As a token of appreciation, a $25
donation will he made to the Thoroughbred Retirement Foundation (up to a total
contribution of $ 1,000). Thank you very much for helping me with my dissertation
research, and best wishes for a fantastic 2007 season!
Kind regards.
Adam Marquardt
Ph.D. Candidate in Marketing
1208 University of Oregon
Lundquist College of Business
Eugene, OR 97403
(541) 346-1452
amarquar(@uoregon.edu
APPENDIX E
IP * % ? A,
p- ^' 1# ■.#; ■ «%
S.'^- <>
t
'?•
/s^i- «fe,""S.iw, ,,
Survey Response Rates
Surveys Mailed Initially - (942 Packets * 2 Surveys)
Undeliverable Due to Flawed Addresses (34 * 2)
Undeliverable Due to the Passing of the Individual (5 * 2)
Not Applicable or Relevant (128 * 2)
- Thoroughbred Breeder/ Race Only Hoinebreds (105)
- Questions about Consignors do not Reflect how Thoroughbreds
Purchases are Actually Made (10)
- Part of a Partnership/ Syndicate where Someone Else Evaluates (8)
- Conflicts with the Pinhooking Sales/ No Time for Surveys (5)
Completed Only One of the Two Surveys (63 * 1)
Returned Survey Instrument Contained too many Missing Values( 1*2)
Returned Survey Instrument Contained Invariant Responses( 2 * 2)
Returned Survey/ Response After the Data Analysis Cutoff Point( 18*2)
Adjusted Total Sample Size
Adjusted Total Sample Size
Respondent Surveys Complying with the Utilized Protocol
Effective Response Rate
w
Independent Samples T-Test for Survey Waves
Independent Samples Test
assumed
Equal variances
not assumed
I Satisfaction with
I Consignor's Prospoct
I Ofteiings
Satisfaction with
Consignor s Serwce
Quality
Satisfaction with Working
Relationship with
Consignor
Comiiiitted to the
Relationship with
Wotikl have Paid the
Same Regaidless of tt>e
Consignor (reverse
Future liileiilion to
Consider Prospect
Olleiings from this
Consignor
Independent Samples T-Test lor Randomly Drawn Sample Group 1
a
1
Independent Samples T-Test for Phone Survey Participants
independent Samples Test
Levene's Test for
iquality of Variances t-test for Equality of Means
95% Confidence
Interval of the
Mean Std. Error Difference
df Siq. (2-tailed) Difference Difference Lower I Upper
Attitude toward Prospe Equal variance
.047 277 .963 .013 .282 -.543 .569
Athletic Potential assumed
Equal variance
.060 43.220 .953 .013 .222 -.434 .461
not assumed
Attitude toward Equal variance
-.567 277 .571 -.170 -.760 .420
Consignor's Abilities assumed
Equal variance
-.628 38.689 .533 -.170 .270 -.717 .377
not assumed
Satisfaction that Equal variance
-.768 277 .443 -.220 .287 -.785 .344
Prospects were a Goo( assumed
Value for the Price Equal variance
-.877 39.528 .386 -.220 .251 -.727 .287
not assumed
Suspicious of Equal variance
-.697 277 .486 -.247 .355 -.946 .451
Consignor's assumed
Recommendations Equal variance
(reverse coded) -.697 36.347 .490 -.247 .355
-.967 .472
not assumed
Committed to the Equal variance
-1.443 277 .150 -.543 .376 -1.284 .198
Relationship with this assumed
Consignor Equal variance
-1.614 38.933 .115 -.543 .336 -1.223 .137
not assumed
Would have been Willii Equal variance
-.828 277 .408 -.301
to Pay a Premium for tl assumed
Involvement of this Equal variance
Consignor not assumed -.782 35.305 .440 -.301 -1.082 .480
4^(1
vS
* I 
CJTm
t ' 
!t=5m
mm
'mmn
Means, Standard Deviations, Skewness and Kurtosis - All Items
Descriptive Statistics
Statistic I Statistic | Statistic I Statistic | Std. Error | Statistic | Std. Error
Attitude toward Prospect's
Lineage
Attitude toward Prospect's
Appearance
Attitude toward Prospect's
Conformation
Attitude toward Prospect's
Disposition
Attitude toward Prospect's
1.499 -1.460
Athletic Potential
Attitude toward
Consignor's Reputation
Attitude toward
Consignor's Abilities
Attitude toward
Consignor's Involvement
Satisfaction with
Consignor's Prospect
Offerings
Satisfaction with
Consignor's Service
Quality
Satisfaction that
Prospects were a Good
Value for the Price
Satisfaction with Working
Relationship with
Consignor
Belief this Consignor was
a Good Seller to do
Business with
Count on the Consignor
to be Sincere
Suspicious of
Consignor's
Recommendations
(reverse coded)
Questioned Consignor's
Integrity (reverse coded)
Worked Together with the
Consignor
Trusted the Consignor
Completely
Committed to the
Relationship with this
Consignor
Planned to Maintain the
Relationship Indefinitely
Relationship Deserved
Maximum Effort to
Maintain
Cared about the
Relationship Long-Term
Would have Paid the
Same Regardless of the
Consignor (reverse
coded)
Would have been Willing
to Pay a Premium for the
Involvement of this
Consignor
Future Intention to
Consider Prospect
249 4.59 1.907 -.524 .154 -.493 .307
Offerings from this
Consignor
Impact of Consignor
249 2.63 2.279 .508 .154 -.989 .307
Recommendation
Impact of Consignor's
249 3.71 2.134 .156 .154 -1.352 .307
Involvement
Valid N (listwise)
Means, Standard Deviations, Skewness and Kurtosis - Final Items
Descriptive Statistics
Mean Std. Skewness Kurtosis
Statistic Statistic Statistic Std. Error Statistic Std. Error
Altitude toward Prospect's
5.37 1.537 -.960 .154 .301 .307
Appearance
Attitude toward Prospect's
5.33 1.486 -.974 .154 .348 .307
Conformation
Attitude toward Prospect's
5.65 1.499 -1.460 .154 1.805 .307
Athletic Potential
Attitude toward
5.36 1.507 -.839 .273 .307
Consignor's Reputation
Attitude toward
5.20 1.570 -.619 .154 -.324
Consignor's Abilities
Satisfaction with
Consignor's Prospect 5.11 1.401 -.512 .150 .307
Offerings
Satisfaction with
Consignor's Service 4.96 1.548 -.471 .154 -.331 .307
Quality
Satisfaction that
Prospects were a Good 4.98 1.507 -.415 .154 -.348
Value for the Price
Suspicious of
Consignor's
2.82 1.837 .729 .154 -.578 .307
Recommendations
(reverse coded)
Questioned Consignor's
Integrity (reverse coded)
Committed to the
Relationship with this
Consignor
Planned to Maintain the
Relationship Indefinitely
Relationship Deserved
Maximum Effort to
Maintain
Cared about the
Relationship Long-Term
Would have been Willing
to Pay a Premium for the
2.80 1.864 .747 .154 -.562 .307
Involvement of this
Consignor
Impact of Consignor
2.63 2.279 .508 .154 -.989 .307
Recommendation
Impact of Consignor's
3.71 2.134 .156 .154 -1.352 .307
Involvement
Valid N (listwise)
Range - All Items
Descriptive Statistics
Attitude toward Prospect's
Lineage
Attitude toward Prospect's
Appearance
Attitude toward Prospect's
Conformation
Attitude toward Prospect's
Disposition
Attitude toward Prospect's
Athletic Potential
Attitude toward
Consignor's Reputation
Attitude toward
Consignor's Abilities
Attitude toward
Consignor's Involvement
Satisfaction with
Consignor's Prospect
Offerings
Satisfaction with
Consignor's Service
Quality
Satisfaction that
Prospects were a Good
Value for the Price
Satisfaction with Working
Relationship with
Consignor
Belief this Consignor was
a Good Seller to do
Business with
Count on the Consignor
to be Sincere
Suspicious of
Consignor's
Recommendations
(reverse coded)
Questioned Consignor's
Integrity (reverse coded)
Worked Together with the
Consignor
Trusted the Consignor
Completely
Committed to the
Relationship with this
Consignor
Planned to Maintain the
Relationship Indefinitely
Relationship Deserved
Maximum Effort to
Maintain
Cared about the
Relationship Long-Term
Would have Paid the
Same Regardless of the
Consignor (reverse
coded)
Would have been Willing
to Pay a Premium for the
Involvement of this
Consignor
Future Intention to
Consider Prospect
Offerings from this
Consignor
Impact of Consignor
Recommendation
Impact of Consignor's
Involvement
Valid N (listwise)
Range - Final Items
Descriptive Statistics
N  Minimum Maximum
Attitude toward Prospect's
Appearance
Attitude toward Prospect's
Conformation
Attitude toward Prospect's
Athletic Potential
Attitude toward
Consignor's Reputation
Attitude toward
Consignor's Abilities
Satisfaction with
Consignor's Prospect
Offerings
Satisfaction with
Consignor's Service
Quality
Satisfaction that
Prospects were a Good
Value for the Price
Suspicious of
Consignor's
Recommendations
(reverse coded)
Questioned Consignor's
Integrity (reverse coded)
Committed to the
Relationship with this
Consignor
Planned to Maintain the
Relationship Indefinitely
Relationship Deserved
Maximum Effort to
Maintain
Cared about the
Relationship Long-Term
Would have been Willing
to Pay a Premium for the
Involvement of this
Consignor
Impact of Consignor
Recommendation
Impact of Consignor's
Involvement
Valid N (llstwise)
Tests of Normality
Tests of Normality
Kolmogorov-Smirnov ^ Shapiro-Wilk
df
Attitude toward Prospect's
.196 249 .000 .895 249 .000
Lineage
Attitude toward Prospect's
.214 249 .000 .869 249 .000
Appearance
Attitude toward Prospect's
.239 249 .000 .871 249 .000
Conformation
Attitude toward Prospect's
.208 249 .000 .896 249 .000
Disposition
Attitude toward Prospect's
.268 : 249 .000 .799 249 .000
Athletic Potential
Attitude toward
.198 249 .000 .877 249 .000
Consignor's Reputation
Attitude toward
.177 249 .000 .897 249 .000
Consignor's Abilities
Attitude toward
.193 249 .000 .904 249 .000
Consignor's Involvement
Satisfaction with
Consignor's Prospect .167 249 .000 .903 249 .000
Offerings
Satisfaction with
Consignor's Service .152 249 .000 .922 249 .000
Quality
Satisfaction that
Prospects were a Good .168 249 .000 .916 249 .000
Value for the Price
Satisfaction with Working
Relationship with .156 249 .000 .919 249 ,000
Consignor
Belief this Consignor was
a Good Seller to do .168 249 .000 .910 249 .000
Business with
Count on the Consignor
.155 249 .000 .915 249 .000
to be Sincere
Suspicious of
Consignor's
.210 249 .000 ,858 249 .000
Recommendations
{reverse coded)
Questioned Consignor's
.217 249 .000 .833 249 .000
Integrity (reverse coded)
Worked Together with the
.144 249 .000 .917 249 .000
Consignor
Trusted the Consignor
.138 249 .000 .915 249 .000
Completely
Committed to the
Relationship with this .168 249 .000 .890 249 .000
Consignor
Planned to Maintain the
.152 249 .000 .893 249 .000
Relationship Indefinitely
Relationship Deserved
Maximum Effort to .175 249 .000 .880 249 .000
Maintain
Cared about the
.160 249 .000 .885 249 .000
Relationship Long-Term
Would have Paid the
Same Regardless of the
.198 249 .000 .865 249 .000
Consignor (reverse
coded)
Would have been Willing
to Pay a Premium for the
.204 249 .000 .849 249 .000
Involvement of this
Consignor
Impact of Consignor
.185 249 .000 .892 249 .000
Recommendation
Impact of Consignor's
.146 249 .000 .891 249 .000
Involvement
a  Lilliefors Significance Correction
APPENDIX G
INITIAL MEASUREMENT MODEL MODIFICATION INDICES
Initial Modification Indices( MI)-  Covariances:( Group number 1 - Default model)
Regression Weights:( Group number 1 - Default model)
M.I. Par Change
FI <— ATH 14.966 .48S
FT <— SATISF 38.6SS .7IS
FI <— TRST 18.440 .437
FI <— ATC 30.661 .498
FI <— PAYSAME 26.430 .32S
FI SATS 37.609 .408
FI <— SAT4 30.17S .3S6
FI <— SAT3 24.8S6 .379
FI <— SAT2 30.0S2 .406
FI <— SATl 26.428 .421
FI <— TRUSTS 11.030 .192
FI <— TRUST4 13.398 .246
FI <— TRUSTS 12.S04 -.161
FI <— TRUST2 IS.348 -.1S8
FI <— TRUSTl 26.476 .380
FI <— CON3 20.393 .30S
FI <— C0N2 2S.010 .36S
FI <— CON I 34.308 .446
FI <— HORSES 11.742 .262
FI <— H0RSE4 14.300 .277
FI <— H0RSE2 II.4S0 .2S2
CONIMPAC<— CON3 20.2S7 .2S0
AFFECT <— FI 1I.29S -.206
PAYSAME <— ATH 11.27S .394
PAYSAME <— SATISF 21.664 .S02
PAYSAME <— ATC 24.S81 .417
PAYSAME <— FI 27.SS3 .296
PAYSAME <— SATS 18.947 .271
PAYSAME <— SAT3 20.24S .321
PAYSAME <— SAT2 20.600 .3IS
PAYSAME <— SATl 12.982 .276
PAYSAME <— TRUSTl 18.8IS .300
PAYSAME <— C0N3 16.273 .2SS
PAYSAME <— C0N2 24.S63 .339
M.l. Par Change
PAYSAME <- CONl 23.808 .348
PAYSAME <- HORSEl 11.0S3 .217
SAT4 <- COM 10.61S .111
SAT4 <- C0MMIT3 13.82S .139
SAT4 <— C0MMIT2 10.396 .107
SAT4 <— COMMIT 1 11.692 .107
SAT3 <- SATl 17.946 .202
SAT2 <- SATl 10.466 .126
SATl <- SAT3 17.S23 .174
TRUST4 <- BrEq 2S.010 .334
TRUST4 <— SATISF 14.369 .300
TRUST4 <- TRST 22.S69 .332
TRUST4 <— COM 23.813 .204
TRUST4 <— AFFECT 18.404 .149
TRUST4 <— PAYMORE 18.81S .184
TRUST4 <— SATS 14.86S .176
TRUST4 <— SAT4 11.380 .ISl
TRUST4 <— SAT2 11.462 .173
TRUST4 <— TRUSTS 28.446 .213
TRUST4 <— TRUST 1 17.870 .21S
TRUST4 <— COMM1T4 18.906 .163
TRUST4 <— COMMITS 2S.380 .231
TRUST4 <— C0MM1T2 21.283 .189
TRUST4 <— COMMITI 20.446 .173
TRUST4 <— CON3 12.186 .162
TRUST3 <— BrEq 88.070 -1.207
TRUST3 <— SATISF 162.309 -1.943
TRUST3 <— TRST 1SS.102 -1.681
TRUST3 <— COM 89.811 -.763
TRUST3 <— ATC 148.092 -1.4S0
TRUST3 <— F1 61.S87 -.626
TRUST3 <—CONIMPAC 49.903 -.S03
TRUST3 <— AFFECT 1S.931 -.266
TRUST3 <— PAYMORE 26.913 -.423
TRUST3 <— SATS 13S.994 -1.027
TRUST3 <— SAT4 98.67S -.8S4
TRUST3 <— SAT3 97.768 -.998
TRUST3 <— SAT2 124.934 -1.097
M.l. Par Change
TRUSTS <-- SATl 8S.209 ^990
TRUSTS <-- TRUSTS 127.8S0 -.868
TRUSTS <- TRUST4 116.8S6 -.962
TRUSTS <- TRUST2 77.4S4 .472
TRUSTS <-- TRUSTl 192.042 -1.SS7
TRUSTS <- COMMIT4 82.089 -.6SS
TRUSTS <- COMMITS 61.102 -.691
TRUSTS <-- C0MM1T2 78.249 -.697
TRUSTS <-- COMMITl 64.69S -.S9S
TRUSTS <-- CONS 119.280 -.978
TRUSTS <-- CON2 125.182 -1.08S
TRUSTS <-- CONl 116.S87 -1.089
TRUSTS <-- HORSE4 14.S67 -.S70
TRUST2 <- BrEq 101.7S9 -1.422
TRUST2 <-- SATISF 177.668 -2.229
TRUST2 <— TRST 17S.980 -1.9S2
TRUST2 <— COM 102.S44 -.89S
TRUST2 <— ATC 1S6.944 -1.6S7
TRUST2 <— FI 78.SS8 -.774
TRUST2 <— CONIMPAC SS.98S -.S8S
TRUST2 <— AFFECT 20.S82 -.SS2
TRUST2 <— PAYMORF S2.826 -.512
TRUST2 <— SATS 1SS.648 -1.197
TRUST2 <— SAT4 11S.S27 -I.OIS
TRUST2 <— SATS 97.897 -1.09S
TRUST2 <— SAT2 128.US -1.219
TRUST2 <— SATl 88.709 -1.121
TRUST2 <— TRUSTS 1S8.201 -.989
TRUST2 <— TRUST4 142.149 -1.164
TRUST2 <— TRUSTS 9S.714 .6S9
TRUST2 <— TRUSTl 207.719 -1.S48
TRUST2 <— COMM1T4 9S.S18 -.764
TRUST2 <— COMMITS 66.167 -.789
TRUST2 <— C0MM1T2 89.SS2 -.816
TRUST2 <— COMMITl 74.218 -.696
TRUST2 <— CONS 121.466 -1.08S r.
TRUST2 <— C0N2 1S1.46S -1.217
TRUST2 <— CONl 12S.SS0 -1.2S8 <- '2'
TRUST2 <— H0RSE4 1S.702 -.S94
TRUSTl <— SATISF 44.S9S .4S4
TRUST 1 <— TRST 2S.040 .289
TRUSTl <— ATC 4S.904 .SSS
TRUSTl <— FI 1S.06S .129
TRUSTl <— SATS 46.042 .267
TRUSTl <— SAT4 26.8S7 .199
TRUSTl <— SATS S2.171 .2S6
TRUSTl <— SAT2 40.410 .279
e-f-
TRUSTl <— SATl 22.4S1 .2S0
W
TRUSTl <— TRUSTS 26.207 .176
TRUSTl <— TRUST4 17.092 .16S
TRUSTl <— TRUSTS S0.64S -.149
TRUSTl <— TRUST2 2S.889 -.122
TRUSTl <— CONS 28.81S .21S
TRUSTl <— C0N2 40.S77 .27S
TRUSTl <— CONl 40.728 .288
C0MM1T4 <— AFFECT 16.010 -.090
COMMITS <— COM 11.087 .120
COMMITS <— AFFECT 10.008 .09S
COMMITS <— C0MMIT2 1S.S17 .ISO
COMMITS <— COMMITI 10.078 .lOS
CONS <— BrEq 12.491 .17S
CONS <—CONIMPAC 2S.701 .ISS
CONS <— TRUST4 10.0S4 .108
H0RSE4 <— SATISF 17.S72 .292
H0RSE4 <— TRST 1S.S96 .227
HORSE4 <— ATC 20.2S0 .246
H0RSE4 <— SATS ll.SSl .1S6
HORSE4 <— SAT4 14.728 .1S2
H0RSE4 <— SATS 1S.021 .180
HORSE4 <— SAT2 17.821 .190
H0RSE4 <— TRUSTl 1S.69S .167
H0RSE4 <— CONS 11.426 .1S9
H0RSE4 <— C0N2 18.221 .190
H0RSE4 <— CONl 2S.864 .226

H0RSE2 IHORSE
ATC
p|i14 ph56
TRS;
Variable Summary (Group number 1)
Your model contains the following variables (Group number 1)
Observed, endogenous variables
H0RSE2
H0RSE3
HORSES
CONl
C0N2
COMMIT]
C0MMIT2
COMMITS
C0MMIT4
TRUST2
TRUSTS
SATl
SAT2
SATS
PAYMORE
AFFECT
CONIMPAC
Unobserved, exogenous variables
EH2
EHS
EH5
ATC
EC I
EC2
COM
EMI
EM2
EMS
EM4
TRST
ET2
ETS
SATISF
ESl
ES2
ESS
BrEq
EP2
ATH
Variable counts (Group number 1)
Number of variables in your model;
Number of observed variables:
Number of unobserved variables:
Number of exogenous variables:
Number of endogenous variables:
Parameter summary (Group number 1)
Weights Covariances Variances Means Intercepts Total
Fixed
Labeled
Unlabeled
Total
Assessment of normality (Group number 1)
Variable min max skew c.r. kurtosis c.r.
CONIMPAC 1.000 7.000 .155 1.001 -1.349 -4.345
AFFECT .000 7.000 .505 3.255 -.993 -3.199
PAYMORE 1.000 7.000 .743 4.785 -.575 -1.852
SAT3 1.000 7.000 -.412 -2.657 -.365 -1.174
SAT2 1.000 7.000 -.468 -3.017 -.348 -1.122
SATl 1.000 7.000 -.509 -3.280 .123 .397
TRUST3 1.000 7.000 .892 5.746 -.244 -.785
TRUST2 1.000 7.000 .725 4.670 -.591 -1.902
C0MMIT4 1.000 7.000 .377 2.430 -1.099 -3.538
C0MM1T3 1.000 7.000 .559 3.600 -.695 -2.238
C0MM1T2 1.000 7.000 .404 2.600 -1.053 -3.390
COMMITl 1.000 7.000 .454 2.925 -.962 -3.099
C0N2 1.000 7.000 -.616 -3.966 -.341 -1.099
CONl 1.000 7.000 -.834 -5.371 .244 .785
HORSES 1.000 7.000 -1.451 -9.348 1.745 5.621
Variable skew c.r. kurtosis c.r.
H0RSE3 1.000 7.000 -.968 -6.236 .317 1.020
H0RSE2 1.000 7.000 -.954 -6.148 .271 .871
Multivariate 75.365 23.395
Observations farthest from the centroid (Mahalanobis distance)( Group number 1)
Observation number Mahalanobis d-squared
235 84.521
67 65.981
28 56.045
162 43.171
193 42.979
114 41.249
33 40.775
98 39.624
223 39.000
207 37.746
116 37.644
50 37.516
188 37.242
36.535
36.530
36.452
35.756
34.621
33.719
32.373
31.637
30.945
30.443
30.194
29.670
29.520
28.926
28.484
28.151
27.125
Observation number Mahalanobis d-squared pi p2
171 27.117 .056 .000
132 27.025 .058 .000
217 26.869 .060 .000
62 26.812 .061 .000
70 26.633 .064 .000
192 26.025 .074 .000
65 25.993 .075 .000
91 25.556 .083 .000
198 25.549 .083 .000
154 25.058 .093 .001
61 25.055 .093 .000
15 24.917 .097 .000
78 24.915 .097 .000
147 24.414 .109 .001
111 24.391 .109 .001
13 23.830 .124 .004
31 23.793 .125 .003
205 23.660 .129 .003
179 23.566 .132 .003
187 23.538 .133 .002
245 23.399 .137 .002
39 23.014 .149 .007
170 22.736 .158 .013
131 22.305 .173 .044
56 21.865 .190 .124
32 21.709 .196 .145
191 21.596 .201 .152
23 20.996 .226 .428
208 20.954 .228 .398
242 20.708 .240 .505
156 20.318 .258 .707
74 20.258 .261 .692
231 20.176 .265 .693
22 19.844 .282 .830
89 19.809 .284 .809
166 19.808 .284 .768
19 19.767 .286 .747
241 19.672 .291 .758
Observation number Mahalanobis d-squared pi p2
239 19.626 .294 .740
236 19.424 .305 .810
134 19.253 .314 .855
216 19.171 .319 .858
9 19.062 .325 .873
150 18.909 .334 .903
224 18.856 .337 .897
155 18.734 .344 .913
127 18.698 .346 .902
69 18.663 .348 .890
202 18.515 .357 .917
76 18.489 .359 .903
141 18.432 .362 .899
18 18.253 .373 .933
136 18.104 .382 .952
161 18.033 .387 .953
212 18.009 .388 .944
14 17.878 .397 .958
145 17.876 .397 .945
80 17.418 .426 .992
221 17.413 .427 .989
17 17.348 .431 .989
119 17.145 .445 .995
215 17.141 .445 .993
87 17.096 .448 .993
30 17.008 .454 .994
86 16.883 .462 .996
90 16.876 .463 .994
35 16.836 .466 .993
120 16.741 .472 .995
103 16.583 .483 .997
240 16.533 .486 .997
Estimates( Group number 1 - Default model)
Sealar Estimates (Group number 1 - Default model)
Maximum Likelihood Estimates
Regression Weights:( Group miiiiber 1 - Default model)
Estimate S.E. C.R. P Label
CONl <-- ATC 1.000
C0N2 <-- ATC 1.121 .OSS 21.029 ***
COMMIT 1 <-- COM 1.000
COMMIT2 <-- COM 1.078 .0S6S0.S17 ***
COMMITS <-- COM .9S7 .OSS 28.467 ***
C0MMIT4 <-- COM 1.087 .0S8 28.472 ***
TRUST2 <-- TRST 1.000
TRUSTS <-- TRST .9S7 .OSS 28.467 ***
SATl <-- SATISF 1.000
SAT2 <-- SATISF 1.198 .07S 16.498 ***
SATS <-- SATISF 1.047 .07S 14.S72 ***
PAYMORE <-- BrEq 1.000
HORSES <-- ATH 1.000
HORSES <-- ATH .956 .052 18.S4S ***
H0RSE2 <-- ATH 1.048 .0S2 20.22S ***
AFFECT <-- BrEq 1.208 .140 8.657 ***
CONIMPAC<-- BrEq 1.420 .141 10.081 ***
Standardized Regression Weights:( Group number 1 - Default model)
Estimate
CONl <— ATC .885
C0N2 <— ATC .95S
COMMITI <— COM .917
C0MM1T2 <— COM .958
COMMITS <— COM .857
COMM1T4 <— COM .942
TRUST2 <— TRST .947
TRUSTS <— TRST .917
SATl <— SATISF .82S
SAT2 <— SATISF .89S
SATS <— SATISF .802
PAYMORE <— BrEq .663
HORSES <— ATH .89S
HORSES <— ATH .861
H0RSE2 <— ATH .91S
AFFECT <— BrEq .655
Estimate
CONIMPAC<— BrEq M2
Covariances:( Group number 1 - Default mode!)
Estimate S.E. C.R. P L abel
ATC <-->COM 1.441 .192 7.491 *** ph25
ATC <-->TRST -1.537 .196 -7.860 *** ph24
ATC <-->SATISF 1.215 .147 8.242 *** ph23
ATC <--> BrEq .973 .158 6.154 *** ph26
COM <-->TRST -1.236 .226 -5.465 *** ph45
COM <-->SATISF 1.189 .169 7.045 *** ph35
COM <--> BrEq 1.678 .231 7.253 *** ph56
TRST <-->SATISF -1.328 .174 -7.622 *** ph34
TRST <--> BrEq -.669 .174 -3.850 *** ph46
SATISF<--> BrEq .762 .135 5.654 *** ph36
ATC <-->ATH .316 .124 2.544 .011 phl2
COM <-->ATH .277 .161 1.727 .084 phi 5
TRST <-->ATH -.220 .161 -1.363 .173 phl4
BrEq <-->ATH -.013 .123 -.108 .914 phl6
SATISF<-->ATH .442 .114 3.883 *** phi 3
Correlations:( Group number 1 - Default model)
Estimate
ATC <--> COM
ATC <--> TRST -.657
ATC <-> SATISF
ATC <-> BrEq
COM <-> TRST
COM <-> SATISF
COM <--> BrEq
TRST <-> SATISF
TRST <-> BrEq
SATlSF<-> BrEq
ATC <-> ATH
COM <-> ATH
TRST <--> ATH
BrEq <--> ATH
4
SATlSF<-> ATH
Variances:( Group number 1 - Default model)
Estimate S.E. C.R.
1.774 .205 8.649
3.139 .326 9.638
3.088 .314 9.843
1.325 .172 7.683
1.520 .280 5.428
1.784 .203 8.779
.393 .066 5.938
.569 .069 8.211
.453 .065 6.939
.489 .068 7.198
.225 .068 3.289 ,
.597 .067 8.964
.328 .051 6.465
.997 .099 10.030
.474 .061 7.782
.358 .107 3.353
.514 .101 5.114
.630 .073 8.633
.485 .075 6.494
.807 .090 9.012
1.941 .206 9.405
2.955 .312 9.481
1.470 .231 6.367
Modification Indices (Group number 1 - Default model)
Covariances:( Group number 1 - Default model)
M.l. Par Change
Variances:( Group number 1 - Default model)
M.l. Par Change
Regression Weights:( Group number 1 - Default model)
M.I. Par Change
COMMITK—AFFECtI 13.112 .086
Minimization History (Default model)
Negative Smallest
Iterati Conditi Diamet NTri
eigenvalu eigenval F Ratio
on on # er es
es ue
9999.0 3761.2 9999.0
0 e 18 -.809 0
00 92 00
2282.3
1 e 23 -.628 2.483 19 .530
46
1597.6
2 e 17 -.892 .766 5 .989
08
e 1188.8
3 11 -.518 .497 4 .877
* 35
e 656.38
4 3 -.332 .771 5 .904
* 3
341.64
5 e 0 300.693 .612 5 .907
3
261.79
6 e 0 395.659 .973 2 .000
3
190.12
7 e 0 284.046 .313 1 1.153
2
179.75
8 e 0 289.292 .126 1 1.135
9
179.14
9 e 0 292.761 .032 1 1.049
0
179.13
10 e 0 304.300 .003 1 1.005
6
179.13
11 e 0 306.643 .000 1 1.000
6
Model Fit Summary
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 48 179.136 105 .000 1
Meas Test Model 33 595.032 120 .000 ^
Model NPAR CMIN CMIN/DF
Saturated model 153 .000
Independence model 17 3680.981 27.066
RMR,G FI
Model RMR GFI AGFI PGFl
Default model .144 .920 .884 .632
Meas Test Model 1.176 .805 .751 .631
Saturated model .000 1.000
Independence model 1.340 .245
Baseline Comparisons
NFI IFl TLl
Model
Delta 1 Delta2 rho2
Default model ^79 .979
Meas Test Model .838 .867 .848 .866
Saturated model 1.000 1.000 1.000
Independence model .000 .000 .000 .000
Parsimony-Adjusted Measures
Model PRATIO PNFI PCFI
Default model .772 .734 .756
Meas Test Model .882 .740 .764
Saturated model .000 .000 .000
Independence model 1.000 .000 .000
Model NCP LO 90 HI 90
Default model 74.136 40.998 115.150
Meas Test Model 475.032 402.662 554.926
Saturated model .000 .000 .000
Independence model 3544.981 3350.815 3746.440
FMIN
FMIN LO 90 HI 90
.722 .165 .464
Model FMIN FQ LO 90 HI 90
Meas Test Model 2.399 1.915 1.624 2.238
Saturated model .000 .000 .000 .000
Independence model 14.843 14.294 13.511 15.107
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model .053 .040 .066 .327
Meas Test Model .126 .116 .137 .000
Independence model .324 .315 .333 .000
Model AlC BCC BIC CMC
Default model 275.136 282.649 443.974 491.974
Meas Test Model 661.032 666.197 777.108 810.108
Saturated model 306.000 329.948 844.170 997.170
Independence model 3714.981 3717.642 3774.778 3791.778
ECVI
Model ECVl LO 90 HI 90 MECVl
Default model 1.109 .976 1.275 1.140
Meas Test Model 2.665 2.374 2.988 2.686
Saturated model 1.234 1.234 1.234 1.330
Independence model 14.980 14.197 15.792 14.990
nOELTER
HOELTER HOELTER
Model
.05 .01
Default model 180 197
Meas Test Model 62 67
Independence model 12 12
Nested Model Comparisons: Assuming model Defanlt model to be correct:
A. DDFF  CMIN
D ^Fl IFl RFl TLl
CMIN P Delta-2 rho-1 rho2
Meas Test Model 15 415.895 .000 .113 .116 .120 .125
APPENDIX I
RELIABILITY TESTS
v^Al.i. '
Jh v,i
*• fc
: 2&.V > Jk^- 'i^
.f /it
Sii!|^
-6*
M
vkijr. 1^'
m
mi
W'
I
k5
V
BW>-.'v- -'iv?:;"* ̂  I- •
;. 'iv."
H#.''
&&S-"7
Tinniiir'TW'tf"''1^'' ^'' 
Calculated Reliabilities Using Cronbach's Alpha
Buyer Attitude toward the Horse
Reliability Statistics
Cronbach's
Alpha N of Items
.919 3
Buyer Attitude toward the Consignor (N/A - 2 Items)
Satisfaction
Reliability Statistics
Cronbach's
Alpha N of Items
.880 3
Trust (N/A - 2 Items)
Commitment
Reliability Statistics
Cronbach's
Alpha N of Items
.955 4
Relationship Quality
Reliability Statistics
Cronbach's
Alpha N of Items
.909 9
Brand Equity
Reliability Statistics
Cronbach's
Alpha N of Items
.759 3
Calculated Reliabilities Using Carver and Mentzer Formulae
Item Lambda Lambda Meas Construct Variance
(ABS) squared Error Reliability Extracted
H0RSE2 0.91 S 0.8S4 0.166
HORSES 0.861 0.741 0.259
HORSES 0.89S 0.797 0.20S
ATH 2.667 2.S72 0.628 0.919 0.791
C0N1 0.885 0.78S 0.217
C0N2 0.95S 0.908 0.092
ATC 1.8S8 1.691 0.S09 0.916 0.846
SAT1 0.82S 0.677 0.S2S
SAT2 0.89S 0.797 0.20S
SATS 0.802 0.64S 0.S57
SATIS 2.518 2.118 0.882 0.878 0.706
TRUST2 0.947 0.897 0.10S
TRUSTS 0.917 0.841 0.159
TRST 1.864 1.7S8 0.262 0.9S0 0.869
C0MMIT1 0.917 0.841 0.159
C0MMIT2 0.958 0.918 0.082
COMMITS 0.857 0.7S4 0.266
G0MMIT4 0.942 0.887 0.11S
COM S.674 S.S80 0.620 0.956 0.845
PAYMORE 0.66S 0.440 0.560
AFFECT 0.655 0.429 0.571
GONIMPAC 0.822 0.676 0.S24
BREQ 2.140 1.544 1.456 0.759 0.515
APPENDIX J
DISCRIMINANT AND CONVERGENT VALIDITY TESTS
Total Variance Explained
Initial Eigenvalues Rotation Sums of Squared Loadings
Component otal % of Variance Cumulative % Total % of Variance Cumulative %
1 7.437 43.748 43.748 3.682 21.657 21.657
2.697 15.863 59.611 2.641 15.537 37.194
2.075 12.206 71.817 2.458 14.461 51.655
.855 5.027 76.844 2.113 12.429 64.084
.816 4.803 81.646 1.905 11.205 75.288
.632 3.719 85.366 1.713 10.077 85.366
.511 3.008 88.374
.337 1.980 90.354
.299 1.757 92.111
.256 1.507 93.618
.239 1.406 95.024
.201 1.183 96.207
.173 1.017 97.224
.146 .859 98.083
.130 .765 98.848
.112 .660 99.509
.084 .491 100.000
Extraction Method: Principal Component Analysis.
Rotated Component Matrij?
Component
Attitude toward Prospect's
Appearance
Attitude toward Prospect's g^^
Conformation
Attitude toward Prospect's
Athletic Potential
Attitude toward
Consignor's Reputation
Attitude toward
Consignor's Abilities
Satisfaction with
Consignor's Prospect
Offerings
Satisfaction with
Consignor's Service
Quality
Satisfaction that
Prospects were a Good
Value for the Price
Suspicious of
Consignor's
Recommendations
(reverse coded)
Questioned Consignor's
Integrity (reverse coded)
Committed to the
Relationship with this .836
Consignor
Planned to Maintain the
Relationship Indefinitely
Relationship Deserved
Maximum Effort to .854
Maintain
Cared about the
Relationship Long-Term
Would have been Willing
to Pay a Premium for the
Involvement of this
Consignor
Impact of Consignor
Recommendation
Impact of Consignor's
Involvement
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization,
a- Rotation converged in 7 iterations.
Paired Construct Validity Check
Rotated Component Matrix
Component
1 2 
Attitude toward Prospect's ^^2
Appearance
Attitude toward Prospect's
Conformation
Attitude toward Prospect's
Attiletic Potential
Attitude toward ggg
Consignor's Reputation
Attitude toward
Consignor's Abilities
Extraction Mettrod: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization,
a. Rotation converged in 3 iterations.
Rotated Component Matr'n?
Component
Attitude toward Prospect's g^^
Appearance
Attitude toward Prospect's g^2
Conformation
Attitude toward Prospect's
Athletic Potential
Satisfaction with
Consignor's Prospect
Offerings
Satisfaction with
Consignor's Service
Quality
Satisfaction that
Prospects were a Good .891
Value for the Price
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization,
a. Rotation converged in 3 iterations.
Rotated Component Matrij 
Component
Attitude toward Prospect s
Appearance
Attitude toward Prospect's
Conformation
Attitude toward Prospect's
Athletic Potential 3' iPi
Suspicious of
Consignor's ggg
Recommendations
(reverse coded)
Questioned Consignor's ggg
integrity (reverse coded)
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization,
a. Rotation converged in 3 iterations.
Rotated Component Matri)?
Attitude toward Prospect's
Appearance
Attitude toward Prospect's .
Conformation
Attitude toward Prospect's
Athletic Potential
Committed to the
Relationship with this .942
Consignor
Planned to Maintain the
, Relationship Indefinitely
Relationship Deserved
Maximum Effort to .901
Maintain
Cared about the g
Relationship Long-Term
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
3- Rotation converged in 3 iterations.
Rotated Component Matrl)?
: Attitude toward Prospect's
Appearance
Attitude toward Prospect's
Conformation
Attitude toward Prospect's
Athletic Potential
Would have been Willing
to Pay a Premium for the
Involvement of this
Consignor
Impact of Consignor
Recommendation
Impact of Consignor's
Involvement
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization,
a. Rotation converged in 3 iterations.
Rotated Component Matri)?
Attitude toward
Consignor's Reputation
Attitude toward gg^
Consignor's Abilities
Satisfaction with
Consignor's Prospect .858
Offerings
Satisfaction with
Consignor's Service .734
Quality
Satisfaction that
Prospects were a Good .864
Value for the Price
Extraction Method; Principal Component Analysis.
Rotation Method; Varimax with Kaiser Normalization,
a. Rotation converged in 3 iterations.
Vi
.r *
Rotated Component Matrix
Attitude toward
Consignor's Reputation
Attitude toward
Consignor's Abilities
Suspicious of
Consignor's
Recommendations
(reverse coded)
Questioned Consignor's
Integrity (reverse coded)
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization,
a. Rotation converged in 3 iterations.
Rotated Component Math)!
Component
1 2 
Attitude toward
Consignor's Reputation
Attitude toward
Consignor's Abilities
Committed to the
Relationship with this .893
Consignor
Planned to Maintain the
Relationship Indefinitely
Relationship Deserved
Maximum Effort to .900
Maintain
Cared about the
Relationship Long-Term
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization,
a- Rotation converged in 3 iterations.
Rotated Component Matrii?
Attitude toward
Consignor's Reputation
Attitude toward
Consignor's Abilities
Would have been Willing
to Pay a Premium for the
Involvement of this
Consignor
Impact of Consignor ^^2
Recommendation
Impact of Consignor's
Involvement
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization,
a. Rotation converged in 3 iterations.
Rotated Component Matrix
Satisfaction with
Consignor's Prospect .893
Offerings
Satisfaction with
Consignor's Service .819
Quality
Satisfaction that
Prospects were a Good .854
Value for the Price
Suspicious of
Consignor's
Recommendations
(reverse coded)
Questioned Consignor's
Integrity (reverse coded)
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization,
a. Rotation converged in 3 iterations.
Rotated Component MatrlJt
Satisfaction with
Consignor's Prospect .897
Offerings
Satisfaction with
Consignor's Service .833
Quality
Satisfaction that
Prospects were a Good .862
Value for the Price
Committed to the
Relationship with this ,907
Consignor
Planned to Maintain the
Relationship Indefinitely
Relationship Deserved
Maximum Effort to .881
Maintain
Cared about the
Relationship Long-Term
Extraction Method; Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization,
a- Rotation converged in 3 iterations.
Rotated Component Matrix
Satisfaction with
Consignor's Prospect .902
Offerings
Satisfaction with
Consignor's Service .849
Quality
Satisfaction that
Prospects were a Good .891
Value for the Price
Would have been Willing
to Pay a Premium for the
Involvement of this
Consignor
Impact of Consignor
Recommendation
Impact of Consignor's
Involvement
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization,
a- Rotation converged in 3 iterations.
Rotated Component Matrix
Suspicious of
Consignor's ^
Recommendations
(reverse coded)
Questioned Consignor's ^
Integrity (reverse coded)
Committed to the
Relationship with this .926
Consignor
Planned to Maintain the
Relationship Indefinitely
Relationship Deserved
Maximum Effort to .903
Maintain
Cared about the
919
Relationship Long-Term
Extraction Method: Principal Component Analysis.
Rotation Method: Varimaxwith Kaiser Normalization,
a. Rotation converged in 3 iterations.
Rotated Component Matri)!
Suspicious of
Consignor's
Recommendations
(reverse coded)
Questioned Consignor's
Integrity (reverse coded)
Would have been Willing
to Pay a Premium for the
Involvement of this
Consignor
Impact of Consignor g^^
Recommendation
Impact of Consignor's g^^
Involvement
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization,
a* Rotation converged in 3 iterations.
^. 1>J.V
'-t. T,
Shared Variance vs. Explained Variance
•'"^v *
Variance
Construct Covariate Correlation Extracted (Highest) Average R
ATH 0.79 0.027
ATC 0.178 0.032
8 0 .287 0.082
T - 0.094 0.009
C 0 .117 0.014
BE -0.008 0.000
0.846
0.178 0.032
0.792 0.627
-0.657 0.432
0.61 0.372
0.592 0.350
0.706 0.354
0.287 0.082
0.792 0.627
-0.657 0.432
0.583 0.340
0.537 0.288
0.869
-0.094 0.009
-0.657 0.432
-0.657 0.432
-0.397 0.158
-0.309 0.095
0.845
0.117 0.014
0.61 0.372
0.583 0.340
-0.397 0.158
0.768 0.590
0.515
-0.008 0.000
0.592 0.350
0.537 0.288
-0.309 0.095
0.768 0.590
APPENDIX K
FACTOR CORRELATIONS
), ty
"■'v. '-•1?
\j-'
4" *-3r^ r tf* ■.'■[I^^ir"
k  ''^
• ^
'V 145
It
*i ^?i
*■*
rj .
»  w'4^f>ll<"i' -»• mlV I -T.! <| v J^vv
1.^ i>"ij
"S < ^
tr«r r^r dK
v-jTi
-pjrft ■  st**^ !ij5"^liK ^ 3?^
Factor Correlations
Correlations
ATHORSE ATCONSIG RQUALITY |B RANDEQU
ATHORSE Pearson Correlation
Sig. (2-tailed)
N
ATCONSIG Pearson Correlation
Sig. (2-tailed) IiII
N
RQUALITY Pearson Correlation
Sig. (2-talled)
N
BRANDEQU Pearson Correlation
Sig. (2-talled)
N
Correlation Is significant at the 0.05 level (2-tailed).
Correlation is significant at the 0.01 level (2-talled).
r
l-Jt-
sC
r h ̂
-■
l::,
-  =- LiSajSwi
APPENDIX L
MEDIATION TEST RESULTS
Direct( Unmedlated) Model Fit Summari
CMIN
Model NPAR CMIN CMIN/DF
Direct Effects on RQual Model 40 497.308 4.401
Saturated model 153 .000
Independence model 17 3680.981 27.066
RMR,G FI
Model RMR GFI AGFl PGFl
Direct Effects on RQual Model .857 .845 .791 .624
Saturated model .000 1.000
Independence model 1.340 .245
Baseline Comparisons
NEl IFl TI I
Model CFl
Delta 1 Delta2 rho2
Direct Effects on RQual Model .870 .892
Saturated model 1.000 1.000 1.000
Independence model .000 .000 .000 .000
Parsimony-Adjusted Measures
Model PRATIO PNFl PCFl
Direct Effects on RQual Model .831 .719 .741
Saturated model .000 .000 .000
Independence model 1.000 .000 .000
Model NCP LO 90 HI 90
Direct Effects on RQual Model 384.308 318.970 457.190
Saturated model .000 .000 .000
Independence model 3544.981 3350.815 3746.440
FMIN
Model LO 90 HI 90
Model FMIN FO LO 90 HI 90
Direct Effects on RQual Model 2.005 1.550 1.286 1.844
Saturated model .000 .000 .000 .000
Independence model 14.843 14.294 13.511 15.107
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Direct Effects on RQual Model .117 .107 .128 .000
Independence model .324 .315 .333 .000
Model AlC BCC BIC
Direct Effects on RQual Model 577.308 583.568 718.006
Saturated model 306.000 329.948 844.170
Independence model 3714.981 3717.642 3774.778
Model ECVI LO 90 HI 90 MECVI
Direct Effects on RQual Model 2.328 2.064 2.622 2.353
Saturated model 1.234 1.234 1.234 1.330
Independence model 14.980 14.197 15.792 14.990
HOELTER
HOELTER HOELTER
Model
.05 .01
Direct Effects on RQual Model 70 76
Independence model 12 12
Estimates (Group number 1 - Direct Effects on RQual Model)
Scalar Estimates (Group number 1 - Direct Effects on RQual Model)
Maximum Likelihood Estimates
Regression Weights:( Group number 1 - Direct Effects on RQual Model)
Estimate S.E. C.R. P L abel
RQual <— ATH .000
RQual <— ATC .000
BrEq <— RQual .000
SATISF <— RQual 1.000
COM <— RQual .945 142 6.657 *** B75
TRST <— RQual -1.003 146 -6.884 *** 374
BrEq <— ATH -.110 050 -2.198 028 G16
BrEq <— ATC .526 078 6.732 *** 026
CONl <— ATC 1.000
C0N2 <— ATC 1.312 097 3.552 *** LC2
COMMITl <—COM 1.000
C0MMIT2 <—COM 1.056 037 28.613 *** LM2
C0MMIT3 <—COM .877 044 20.068 *** LM3
C0MMIT4 <—COM 1.070 039 27.482 *** LM4
SATl <— SATISF 1.000
SAT2 <— SATISF 1.165 074 15.778 *** LS2
SAT3 <— SATISF 1.066 072 14.707 *** LS3
affect <— BrEq 1.290 171 7.562 lp3
PAYMORE<— BrEq 1.000
HORSES <— ATH 1.000
H0RSE3 <— ATH .957 051 18.606 *** LH3
H0RSE2 <— ATH 1.044 052 20.138 *** LH2
conimpac <— BrEq 1.805 217 8.322 ^ ̂
TRUST3 <— TRST 1.024 057 17.954 ** *  3X3
TRUST2 <— TRST 1.000
Standardized Regression Weights:( Group number 1 - Direct Effects on RQual Model)
Estimate
RQual <— ATH .000
RQual <— ATC .000
BrEq <— RQual .000
SATISF <— RQual .978
COM <— RQual .593
TRST <— RQual -.674
BrEq <— ATH -.137
BrEq <— ATC .602
CONl <— ATC .819
C0N2 <— ATC 1.031
Estimate
COM
COM .957
COM .845
COM .945
SATISF .829
SATISF .874
SATISF .822
BrEq .610
BrEq .578
ATM .894
ATH .863
ATH .910
BrEq .912
TRST .945
TRST .921
Covariaiices:( Group number 1 - Direct Effects on RQual Model)
Estimate S.E. C.R. P Label
ATC<->AfH .296 .110 2.688 .007 Phl2
Correlations:( Group number 1 - Direct Effects on RQual Model)
Estimate
ATC<--> ATH 4^
Variances:( Group number 1 - Direct Effects on RQual Model)
Estimate S.E. C.R. P Label
1.517 .212 7.146 ***
1.788 .203 8.789 ***
1.287 .225 5.706 ***
.750 .166 4.516 ***
.058 .149 .388 .698
2.117 .263 8.044 ***
1.555 .234 6.648 ***
.402 .067 6.029 ***
.563 .069 8.135 ***
.449 .065 6.866 ***
Estimate S.E. C.R. P Label
7m .113 6.615 ***
-.154 .157 -.980 .327
.612 .069 8.851 ***
.334 .052 6.382 ***
1.004 .099 10.099 ***
.445 .060 7.402 ***
.611 .075 8.107 ***
.562 .084 6.697 ***
.733 .089 8.278 ***
3.245 .336 9.662 ***
2.304 .232 9.950 ***
.764 .303 2.522 .012
.357 .137 2.610 .009
.512 .135 3.795 ***
Fullv Mediated Model Fit Summary
CMIN
Model NPAR CMIN DP P CMIN/DF
Full Mediation Model 41 263.613 112 .000 2.354
Saturated model 153
Independence model 17 3680.981 136 .000 27.066
RMR,G FI
Model RMR GFl AGFl PGFl
Full Mediation Model .312 .883 .840 .646
Saturated model .000 1.000
Independence model 1.340 .245 .151 .218
Baseline Comparisons
NFl IFl TLI
Model
Delta 1 Delta2 rho2
Full Mediation Model ^58 ^4^
NFl RFI IFI TLl
Model
Delta 1 rhol Delta2 rho2
Saturated model 1.000 Tooo
Independence model .000 .000 .000 .000
Parsimony-Adjusted Measures
Model PRATIO PNFl PCFI
Full Mediation Model .824 .765 .788
Saturated model .000 .000 .000
Independence model 1.000 .000 .000
Model NCP LO 90 HI 90
Full Mediation Model 151.613 108.018 202.919
Saturated model .000 .000 .000
Independence model 3544.981 3350.815 3746.440
FMIN
Model FMIN FO LO 90 HI 90
Full Mediation Model 1.063 .611 .436 .818
Saturated model .000 .000 .000 .000
Independence model 14.843 14.294 13.511 15.107
RMSEA
Model RMSEA LO 90 HI 90 PCLOS^
Full Mediation Model .074 .062 .085 .001
Independence model .324 .315 .333 .000
Model AlC BCC BIC CAIC
Full Mediation Model 345.613 352.030 489.828 530.828
Saturated model 306.000 329.948 844.170 997.170
Independence model 3714.981 3717.642 3774.778 3791.778
ECVI
Model ECVI LO 90 HI 90 MECVl
Full Mediation Model 1.394 1.218 1.600 1.419
Saturated model 1.234 1.234 1.234 1.330
Independence model 14.980 14.197 15.792 14.990
HOELTER
HOELTER HOELTER
Model
.05 .01
Full Mediation Model 130 141
Independence model 12 12
Estimates (Group number 1 - Full Mediation Model)
Scalar Estimates (Group number 1 - Full Mediation Model)
Maximum Likelihood Estimates
Regression Weights:( Group numher 1 - Full Mediation Model)
Estimate S.E. C.R. P  Label
■ATH .033 1.192 .233 G17
■ATC .682 .055 12.407 *** G27
■ RQual .757 .110 6.911 *** B76
■ RQual 1.000
■ RQual 1.246 .126 9.930 *** B75
- RQual -1.161 .121 -9.579 **=1= B74
-ATH .000
■ATC .000
■ATC 1.000
■ATC 1.134 .054 20.903 *** LC2
■COM 1.000
■COM 1.057 .037 28.716 *** LM2
■COM .875 .044 20.004 *** LM3
■COM 1.069 .039 27.543 *** LM4
-SATISF 1.000
■SATISF 1.200 .073 16.395 *** LS2
-SATISF 1.047 .073 14.300 *** LS3
Estimate S.E. C.R. P Label few
affect <--BrEq 1.273 .161 7.916 *** lp3
PAYMORE<---BrEq 1.000
H0RSE5 <--ATH .955 .047 20.115 *** LH5
H0RSE3 <---ATH .914 .048 19.020 *** LH3
H0RSE2 <-"ATH 1.000
conimpac <-"BrEq 1.631 .186 8.757 *** lp4
TRUST3 <-"TRST 1.012 .054 18.597 *** LT3
TRUST2 <-"TRST 1.000
Standardized Regression Weights:( Group number 1 - Full Mediation Model)
Estimate
RQual <— ATH
RQual <— ATC .895
BrEq <— RQual .671
SATISF <— RQual .877
COM <— RQual .696
TRST <— RQual -.690
BrEq <— ATH .000
BrEq <— ATC .000
CONl <— ATC .880
C0N2 <— ATC .959
COMMITI <— COM .918
COMM1T2 <— COM .957
C0MM1T3 <— COM .844
COMM1T4 <— COM .945
SATl <— SATISF .823
SAT2 <— SATISF .894
SAT3 <— SATISF .801
affect <— BrEq .637 ,f. ilt-
PAYMORE<— BrEq .612
HORSE5 <— ATH .893
H0RSE3 <— ATH .862
H0RSE2 <— ATH .912
conimpac <— BrEq .872
TRUST3 <— TRST .939
TRUST2 <— TRST .926
^ IJI
Covariances:( Group number 1 - Full Mediation Model)
Estimate S.E. C.R. P Label
ATC<-->ATH .332 .129 2.568 .010 Phl2
Correlations:( Group number 1 - Full Mediation Model)
Estimate
ATC<--> ATH .179
Variances:( Group number 1 - Full Mediation Model)
Estimate S.E. C.R. P Label
ATC 1.754 .205 8.571 ***
ATH 1.955 .216 9.063 ***
Ps7 .183 .052 3.529 ***
Ps6 .712 .157 4.546 ***
Ps3 .304 .067 4.531 ***
Ps5 1.685 .203 8.318 ***
Ps4 1.510 .192 7.866 ***
EH2 .397 .067 5.948 ***
EH3 .563 .069 8.133 ***
EH5 .453 .066 6.905 ***
ECl .509 .069 7.387 ***
EC2 .199 .069 2.897.004
EMI .610 .069 8.880 ***
EM2 .331 .052 6.422 ***
EM3
EM4
ESl
ES2
ES3
EP3
EP2
EP4
ET3
ET2
Partiallv Mediated Model Fit Summai
Model Fit Summai y
CMIN
Model NPAR CMIN DF P CMIN/DF
Partial Mediation Model 43 252.871 110 .000 2.299
Saturated model 153 .000 0
Independence model 17 3680.981 136 .000 27.066
RMR,G FI
Model RMR GFI AGFl PGFl
Partial Mediation Model .299 .886 .841 .637
Saturated model .000 1.000
Independence model 1.340 .245
Baseline Comparisons
NFl IFl
Model
Delta 1 Delta2
Partial Mediation Model .950 .960
Saturated model 1.000 1.000 1.000
Independence model .000 .000 .000 .000
Parsimony-Adjusted Measures
Model PRATIO PNFl PCFl"
Partial Mediation Model .809 .753 .776
Saturated model .000 .000 .000
Independence model 1.000 .000 .000
Model NCP LO 90 HI 90
Partial Mediation Model 142.871 100.407 193.057
Saturated model .000 .000 .000
Independence model 3544.981 3350.815 3746.440
FMIN
Model FMIN ^ LO 90 HI 90
Partial Mediation Model 1.020 .576 .405 .778
Saturated model .000 .000 .000 .000
Independence model 14.843 14.294 13.511 15.107
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Partial Mediation Model .072 .061 .084 .001
Independence model .324 .315 .333 .000
Model AlC BCC BIC CMC
Partial Mediation Model 338.871 345.602 490.122 533.122
Saturated model 306.000 329.948 844.170 997.170
Independence model 3714.981 3717.642 3774.778 3791.778
ECVI
Model ECVI LO 90 HI 90 MECVl
Partial Mediation Model 1.366 1.195 1.569 1.394
Saturated model 1.234 1.234 1.234 1.330
Independence model 14.980 14.197 15.792 14.990
HOELTER
HOELTER HOELTER
Model ^ ^
Partial Mediation Model 133 145
Independence model 12
Estimates (Group iiiiniber 1 - Partial Mediation Model)
Scalar Estimates (Group number 1 - Partial Mediation Model)
Maximum Likelihood Estimates
19 1 
Regression Weights: (Group number 1 - Partial Mediation I lod el) 
Estimate S.E. C. R. p Label 
RQual <---ATH .067 .034 1.960 .050 Gl7 
RQual <---ATC .676 .055 12.395 *** G27 
BrEq <---RQual 1.054 .368 2.868 .004 B76 
SATISF <--- RQual 1.000 
COM <--- RQual 1.268 .126 10.062 *** B75 
TRST <--- RQual -1.141 .122 -9.376 *** B74 
Br q <---ATH -.180 .060 -3.011 .003 Gl6 
BrEq <---ATC -.1 79 .245 -.730 .465 G26 
co I <---ATC 1.000 
CON2 <--- ATC 1.1 34 .054 20.908 *** LC2 
COMMIT] <--- COM 1.000 
COMMIT2 <--- COM 1.057 .037 28.744 *** LM2 
COMMIT3 <--- COM .875 .044 20.003 *** LM3 
COMMIT4 <--- COM 1.069 .039 27 .529 *** LM4 
SAT1 <---SATISF 1.000 
SAT2 <--- SATISF 1.20 l .073 16.434 *** LS2 
SAT3 <--- SATISF 1.045 .073 14.283 *** LS3 
affect <---BrEq 1.264 .162 7. 802 .. j! .. ~ .. j< lp3 
PAYMORE <---BrEq 1.000 
HORSES <--- ATH .960 .047 20.2 13 *** LH5 
HORSE3 <---ATH .914 .048 18.955 *** LH3 
HORSE2 <---ATH 1.000 
commpac <---BrEq 1.684 .19 1 8.8 12 ·'<i<* lp4 
TRUST3 <---TRST 1.012 .055 18.262 *** LT3 
TRUST2 <--- TRST 1.000 
Sta ndardized Regression Weights: (Group number 1 - Partial Mediation Model) 
Estimate 
RQual <--- ATH .093 
RQual <--- ATC .894 
BrEq <--- RQual . .942 
SATISF <--- RQual .. 871 
COM <--- RQual .703 
TRST <--- RQual -.673 
BrEq <--- ATH -.224 
BrEq <--- ATC -.212 
Estimate
CONl <— ATC .880
C0N2 <— ATC .959
COMMITI <— COM .918
C0MMIT2 <— COM .958
C0MMIT3 <— COM .844
C0MMIT4 <— COM .945
SATl <— SATISF .823
SAT2 <— SATISF .894
SAT3 <— SATISF .800
affect <— BrEq .623
PAYMORE<— BrEq .603
H0RSE5 <— ATH .896
H0RSE3 <— ATH .861
H0RSE2 <— ATH .910
conimpac <— BrEq .887
TRUST3 <— TRST .940
TRUST2 <— TRST .926
Covariances:( Group number 1 - Partial Mediation Model)
Estimate S.E. C.R. P Label
ATC<->ATH .330 .129 2.561 .010 Phl2
Correlations; (Croup number 1 - Partial Mediation Model)
Estimate
ATC<--> ATH .179
Variances:( Group number 1 - Partial Mediation Model)
Estimate S.E. C.R. P Label
ATC 1.754 .205 8.571 ***
ATH 1.948 .215 9.042 ***
Ps7 .163 .054 3.003 .003
Ps6 .592 .149 3.971 ***
Ps3 .321 .070 4.607 ***
Ps5 1.653 .199 8.301 ***
Ps4 1.575 .198 7.969 ***
EH2 .404 .067 6.066 ***
Estimate S.E. C.R. P
.069 8.198 ***
.442 .065 6.795 ***
.509 .069 7.390 ***
.199 .069 2.898 .004
.610 .069 8.884 ***
.330 .051 6.412 ***
1.013 .100 10.130 ***
.446 .060 7.485 ***
.631 .074 8.576 ***
.479 .076 6.329 ***
.815 .091 8.997 ***
3.163 .330 9 579 ***
2.204 .226 9 767 ***
.968 .276 3.507 ***
.391 .133 2.944 .003
.479 .132 3.623 ***
J- ̂
f r-x
(4 i \  r
^ A
^4
V"* - 'v ":
^4
v-'t..
-^TisuCiau."
-j .7^•-
- ' f-■-VA.C !■
PPirMiS"
•iS % iNaBga*^»r> ■
m.
^ "
W:ih
f "1 if <' i
■^..1 IfclA.T.-'f Ji V M
'hm-iMtfi' i \ ■>«'tT» -Vl
'fesa m
REFERENCES
Aaker, David A.( 2004), "Leveraging the Corporate Brand," California Management
Review, 46, 3, 6-18.
Aaker, David A.( 1996), "Measuring Brand Equity across Products and Markets,"
California Management Review, 38, 3, 102-120.
Aaker, Jennifer L.( 1997)," Dimensions of Brand Personality," Jo//rna/ ofMarketing
Research, 34, 347-356.
Agrawal. M.L.( 2003). "Customer Relationship Management( CRM)&  Corporate
Renaissance," ofS endees Research, Vol. 3, 2, 149-171.
Agrawal, Manoj K. and Vithala R. Rao (1996)," An Empirical Comparison of Consumer-
Based Measures of Brand Equity," Marketing Letters, 7, 3, 237-247.
Ajzen, leek and Martin Eishbein (1980), Understanding Attitudes and Predicting Social
Behavior, Englewood Cliffs, NJ: Prentice-Hall.
Anand, Bharat N. and Klianna, Tarun (2000)," Do Finns Learn to Create Value? The
Case of Alliances," Strategic Management Journal, 21, 295-315.
Anderson, Eugene W., Claes Fomell, and Donald R. Lehmann (1994), "Customer
Satisfaction, Market Share, and Profitability: Findings from Sweden," Journal of
Marketing, 58, 3, 53.
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.
Ang, Lawrence and Ben Taylor (2005)," Managing Customer Profitability Using
Portfolio Matrices," Journal ofD atabase Marketing & Customer Strategy
Management, 12, 4, 298-304.
Amistrong, Gary and Philip Kotler (2005), Marketing: An Introduction. Upper Saddle,
NJ: Prentice-Hall, Inc.
Annstrong, J. Scott and Terry S. Overton (1977), "Estimating Nonresponse Bias in Mail
Surveys,'" Journal ofM arketing Research, 14 (August), 396-402.
Bae, Jonghoon and Gargiulo, Martin (2004)," Partner Substitutability, Alliance Network
Structure, and Finn Profitability in the Telecommunications Industry," ylcaJemy of
Management Journal, 47, 843-859.
Barney, Jay B.,( 1996), "The Resource-based Theory of the Firm," Organization Science,
7, 5, 469-469.
Barney Jay B.( 1991)," Firm Resources and Sustained Competitive Advantage," Journal
ofManagement 17, 99-120.
Berry, Leonard L.( 2000)," Cultivating Service Brand Equity," Journal oft he Academy of
Marketing Science, 28 (Winter), 128-137.
Beverland, Michael B.( 2005), "Crafting Brand Authenticity; The Case of Luxury
^ines," Journal ofManagement Studies 42, 5, 1003-1029.
Blankson, Charles and Stavros P. Kalafatis (1999), "Issues and Challenges in the
Positioning of Service Brands: A Review," Journal ofP roduct & Brand
Management, 8, 2, 106-116.
Bollen, Kenneth and J. Seott Long (1992), "Tests for Structural Equation Models:
Introduction," Sociological Methods and Research, 21, 123-131.
Boulding, William; Staelin, Richard; Ehret, Michael; Johnston, Wesley J. (2005)," A
Customer Relationship Management Roadmap: What Is Known, Potential Pitfalls,
and Where to Go," Journal of Marketing, 69, 4, 155-166.
Bove, Liliana L. and Lester W. Johnson (2001)," Customer Relationships with Service
Personnel: Do We Measure Closeness, Quality or Strength?" Journal ofB usiness
Research, 54, No. 2, pp. 189-197.
Calder, Bobby J., Lynn W. Phillips, and Alice M. Tybout (1982)," Beyond External
Validity," Journal of Consumer Research, 9, 240-244.
Chaudhuri, Arjun and Morris B. Holbrook (2001), "The Chain of Effects from Brand
Trust and Brand Affect to Brand PerfoiTnance: The Role of Brand Loyalty," Journal
ofM arketing, 65, 2, 81-93.
Chaudhuri, Arjun (1999), "The Relationship of Brand Attitudes and Brand Perfonnance:
The Role of Brand Loyalty," Journal ofM arketing Management, 9, 3, 1-9.
Churchill, Jr., Gilbert A.( 1979)," A Paradigm for Developing Better Measures of
Marketing Constructs," Jo«/"«a/ of Marketing Research, 16 (February), 64-73.
Da Silva, Rui Vinhas and Sharifac Fariday Syed Alwi (2006), "Cognitive, Affective
Attributes and Conative, Behavioural Responses in Retail Corporate Branding,"
Journal of Product and Brand Management, 15, 4/5, 293-305.
Davis, Donna P., Susan L. Golicic and Adam J. Marquardt (2007)," Branding a B2B
Service: Does a Brand Differentiate a Logistics Service Provider," Industrial
Marketing Management, Forthcoming.
Dawar, Niraj (2004), "What Are Brands Good For," MIT Sloan Management Review, 46,
1,31-37.
Dawar, Niraj (1998)," Product-Hami Crises and the Signaling Ability of Brands,"
International Studies ofM anagement & Organization. 28. 3. 109-119.
De Chernatony, Leslie (2001), "A model for strategically building brands," Journal of
Brand Management, 9, 1, 32-44.
De Wulf, Kristof, Gaby Odekerken-Schroder and Dawn lacobucci (2001)" Investments in
Consumer Relationships: A Cross-Country and Cross-Industry Exploration," Journal
ofMarketing, 65, 4, 33-50.
DelVecchio, Devon and Danile C. Smith (2005), "Brand-Extension Equity: The Effects
of Perceived Fit and Extension Product Category Risk," Journal oft he Academy of
Marketing Science, 33,2, 184-196.
Dillman, Donald (2000), Mail and Internet Surveys: The Tailored Design Method, New
York, NY: John Wiley & Sons, Inc.
Drucker, Peter F.( 1954), The Practice ofManagement, New York: Harper & Bros.
Dwyer, F. Robert, Paul H. SchuiT, and Sejo Oh (1987), "Developing Buyer-Seller
Relationships," Journal ofM arketing, 51,2, 11-27.
Dwyer, F. Robert, and Sejo Oh,( 1987)" Output Sector Munificence Effects on the
Internal Political Economy of Marketing Channels," Journal of Marketing Research,
24, 4, 347-358.
Dyer, Jeffrey H. and Singh, H.( 1998)," The Relational View: Cooperative Strategy and
Sources of Interorganizational Competitive Advantage," Hcnr/cmy of Management
Review, 23, 660-679.
Eisenhardt, Kathleen M.( 1989), "Building Theories from Case Study Research,"
Academy of Management Review, 14, 4, 532-550.
Elsbach, Kimberly D.( 1994), "Managing Organizational Legitimacy in the California
Cattle Industry: The Construction and Effectiveness of Verbal Accounts,"
Administrative Science Quarterly, 39, 1, 57-88.
Eng, Teck-Yong (2004)," Does Customer Portfolio Analysis Relate to Customer
Performance? An Empirical Analysis of Alternative Strategic Perspective," Journal
ofB usiness & B2B Marketing, 19, 1, 49-67.
Erdem, Tulin (1998), "An Empirical Analysis of Umbrella Branding," Journal of
Marketing Research, 35, 3, 339-351.
Farquhar, Peter H.( 1994), "Managing Brand Equity," Marketing Research, 1, 3, 24-33.
Farquhar, Peter H.( 1989), "Managing Brand Equity," Marketing Research, 1, 3, 24-33.
Fomell, Claes and David F. Larcker (1981), "Structural Equation Models with
Unobservable Variables and Measurement Error: Algebra and Statistics," Journal of
Marketing Research, 18, 3, 382-388.
Fomell, Claes and David F. Larcker (1981), "Evaluation Structural Equation Models with
Unobservable Variables and Measurement Error," Journal ofM arketing Research,
18, 1, 39-50.
Foumier, 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,
Tmst, and Commitment in Customer Relationships," Journal of Marketing, 63, 2, 70-
87.
Carver, Michael S. and John T. Mentzer (1999), "Logistics Research Methods:
Employing Structural Equation Modeling to Test for Constmct Validity,"
Journal ofB usiness Logistics, 20, 1, 33-57.
Gimeno, Javier (2004)," Competition within and between Networks: The Contingent
Effect of Competitive Embeddedness on Alliance Foimation," Academy of
Management Journal, 47, 820-842.
Gombeski Jr., William, David Kantor, and Rob Klein (2002)," Physician-Based Branding
Strategies," Marketing Health Services, 22, 4, 28-34.
Gordon, Geoffrey L., Roger J. Calantone, and C. Anthony di Benedetto (1993)," Brand
Equity in the Business-to-Business Sector: An Exploratory Study," Journal of
Product & Brand Management, 2, 3, 4-16.
Hair, Joseph C., Rolph E. Anderson, Ronald L. Tatham, and William C. Black (1998),
Multivariate Data Analysis: 5''' edition. Upper Saddle River, NJ; Prentice Hall.
Harris, Fiona and Leslie de Chematony (2001), "Corporate Branding and Corporate
Brand Perfonnance," European Journal ofMarketing, 35, 3/4, 441-456.
Harris, Greg (2002), "Brand Strategy in the Retail Banking Sector; Adapting to the
Financial Services Revolution," Brand Management, 9, 6, 430-436.
Hodgson, Geoffrey M.( 2000), "The Marketing of Wisdom: Resource-Advantage
IhQOxyf Journal ofM acromarketing, 20, 1, 68-72.
Hoffe, Jason, Kevin Lane and Victoria M. Nam (2003), McKinsey Quarterly (Special
Edition), 4, 14-17.
Hon, Linda Childers and Larissa A. Grunig (1999), Guidelines for Measuring
Relationships in Public Relations, (Gainesville, PL: The Institute for Public
Relations).
Hoyle, Rick H., Monica J. Harris and Charles M. Judd (2002), Research Methods in
Socicd Relations, 7"' Edition, Stamford, CT: Thompson Learning.
Hunt, Shelby D.( 1997)" Competing Through Relationships: Grounding Relationship
Marketing in Resource-Advantage Theory," Journal ofMarketing Management, 13,
5,431-445.
Hunt, Shelby D. and Dennis B. Amett (2003) "Resource-Advantage Theory and
Embeddedness: Explaining R-A Theory's Explanatory Success," Journal of
Marketing Theory & Practice, 11, 1, 1-17.
Hunt, Shelby D. and Robert M. Morgan (1995)" The Comparative Advantage Theory of
Competition," Jowr«a/ ofMarketing, 59 (April), 1-15.
Huntley, Julie K.( 2006), "Conceptualization and Measurement of Relationship Quality:
Linking Relationship Quality to Actual Sales and Recommendation Intention,
Industrial Marketing Management, 35, 6, 703-714.
Johnson, Michael D. and Fred Seines (2004)," Customer Portfolio Management: Toward
a Dynamic Theory of Exchange Relationships," Journal of Marketing, 68 (April), 1-
17.
Kahle, Lynn R. and Monte M. Page (1976), "The Deprivation-Satisfaction Effect in
Attitude Conditioning without Deprivation but with Demand Characteristics,"
Personality and Social Psychology Bulletin, 2, 4, 470-473.
Kapferer, Jean-Noel (1992), Strategic Brand Management, London: Kogan-Page.
Keller, Kevin L.( 2005), "Branding Shortcuts"( Cover story). Marketing Management,
14, 5, 18-23.
Keller, Kevin L.( 2003), Strategic Brand Management: Building, Measuring, and
Managing Brand Equity, Upper Saddle River, NJ: Prentice-Hall, Inc.
Keller. Kevin L.( 1993). "Conceptualizing. Measuring. Managing Customer-Based Brand
Equity," Journal ofM arketing, 57, 1, 1-22.
Philip Kotler and Keller, Kevin L.( 2006), Marketing Management, Upper Saddle River,
NJ: Prentice-Hall, Inc.
Kim, Jeesun, and Sylvia M. Chan-Olmsted,( 2005)," Comparative Effects of
Organization-Public Relationships and Product-related Attributes on Brand Attitude,"
Journcd ofM arketing Communications, 11, 3, 145-170.
Kinnani, Amna and Akshay R. Rao (2000)," No Pain, No Gain: A Critical Review of the
Literature on Signaling Unobservable Product Quality," Journal ofMarketing, 64, 2,
66-79.
Kogut Bruce and Udo Zander (1992)," Knowledge of the Firm, Combinative
Capabilities, and the Replication of Technology," Organization Science 3: 383-397.
Krishnan, Balaji C. and Michael D. Hartline (2001)," Brand Equity: Is It More Important
in Services," Jonrna/ ofS ervices Marketing, 15, 5, 328-342.
Lee, Thomas W.( 1999), Using Qualitative Methods in Organizational Research,
Thousand Oaks, CA: Sage Publications, Inc.
Leonidou, Leonidas C., Bradley R. Barnes, and Michael A. Talias (2006)," Exporter-
Importer Relationship Quality: The Inhibiting Role of Uncertainty, Distance, and
CowiUcQ'lndustrial Marketing Management, 35, 5, 576-588.
Levitt, Theodore (1960)," Marketing Myopia," Harvard Business Review, Jul-Aug, 45-
56.
Mackay, Marisa M.( 2001), "Evaluation of Brand Equity Measures: Further Empirical
Results," Journal ofP roduct and Brand Management, 10, 1, 38-51.
Maddox, Neil R.( 1982), "The Structure of Consumers' Satisfaction: Cross Product
Comparisons," Journal oft he Academy ofMarketing Science, 10, 1, 37-53.
Malhotra, Naresh K.( 1993), Marketing Research: An Applied Orientation, Englewood
Cliffs, NJ: Prentice-Hall, Inc.
Marquardt, Adam J., Susan L. Golicie, and Donna F. Davis (Under Review)," Exploring
the Branding of B2B Services," Submitted to the Journal ofP roduct & Brand
Management.
McAlexander, James H., John W. Schouten and Harold Koenig (2002), "Building Brand
Community." Journal of Marketing. 66 (January). 38-54.
McCracken, Grant (1988), The Long Interview, Beverly Hills, CA: Sage Publications,
Inc.
McDonald, Malcolm H.B., Leslie de Chernatony, and Fiona Harris (2001), "Corporate
Marketing and Service Brands: Moving Beyond the Fast-Moving Consumer Goods
Model," European Journal ofM arketing, 35, 3 /4, 335-352.
McGrath, Joseph E. and David Brinberg (1983), "External Validity and the Research
Process, A Comment on the Calder/Lynch Dialogue," Journal of Consumer
Research, 10, 115-124.
Mentzer, John T. and Daniel J. Flint (1997)," Validity in Logistics Research," Journal of
Business Logistics, 18, 1, 199-216.
Monczka, Robert M., Kenneth J. Petersen, Robert B. Handfield and Gary L. Ragatz
(1998), "Success Factors in Strategic Supplier Alliances: The Buying Company
Perspective," Decision Sciences, 29, 3, 553-577.
Moorman, Christine, Rohit Deshpande, and Gerald Zaltman (1993), "Factors Affecting
Trust in Market Research Relationships," Jo/;rt7u/ ofM arketing, 57, 1, 81.
Moorman, Christine, Gerald Zaltman, and Rohit Deshpande (1992), "Relationships
between Providers and Users of Market Research: The Dynaiuics of Trust within and
Between Organizations," Journal ofMarketing Research, 29, 3, 314-328.
Morgan, Robeif M. and Shelby D. Hunt (1994)," The Commitment-Trust Theory of
Relationship Marketing," Jottrtto/ ofMarketing, 58 (July), 20-38.
Myers, Chris A.( 2003), "Managing Brand Equity: A Look at the Impact of Attributes,"
Journal ofP roduct & Brand Management, 12, 1,39.
Netemeyer, Richard G., Balaji Krishnan, Chris Pullig, Guangping Wang, Mehmet Yagci,
Dwane Dean, Joe Ricks and Ferdinand Wirth (2004)," Developing and Validating
Measures of Facets of Customer-Based Brand Equity," Journal ofB usiness Research,
57, 2, 209-224.
Netemeyer, Richard G., William O. Bearden and Subhash Shanna (2003), Scaling
Procedures: Issues and Applications. Thousand Oaks, CA: Sage Publications,
Inc.
Nowak. Linda. Liz Thach. and Janeen E. Olsen (2006)." Wowing the Millennials:
Creating Brand Equity in the Wine Industry," Journal ofP roduct & Brand
Management, 15,4/5,316-323.
Monte M. Page and Lynn R. Kahle (1976), "Demand Characteristics in the Satisfaction-
Deprivation Effect on Attitude Conditioning," Journal ofPersonality and Social
Psychology, 33, 5, 553-562.
Park, C. Whan, Bernard J. Jaworski and Deborah J. Maclnnis (1986), "Strategic Brand
Concept-Image Management," Journal of Marketing, 50, 4, 135-145.
Park, C. Whan, S. Y. Jun and Alan D. Shocker (1996)," Composite Branding Alliances:
An Investigation of Extension and Feedback Effects," Jourur// ofM arketing
Research, 33, 4, 453-466.
Park, S. H. and Michael V. Russo (1996)," When Competition Eclipses Cooperation: An
Event History Analysis of Joint Venture Failure," Management Science, 42, 875-890.
Penrose Edith T.( 1959), The Theory oft he Growth oft he Firm, Wiley: New York.
Peter, J. Paul and Gilbert A. Churchill, Jr. (1986), "Relationships Among Research
Design Choices and Psychometric Properties of Rating Scales:
A Meta-Analysis," Journal ofMarketing Research, 23 (February), 1-10.
Prahalad, C.K. and Ramaswamy, Venkatram (2000)," Co-opting Customer Competence,"
Harvard Business Review, 78, 79-87.
Rao, Akshay R. and Kent B. Monroe (1996)" Causes and Consequences of Price
Premiums," Journal ofBusiness, 69, 4, 511.
Riebe, Erica, and John Dawes. 2006. Recall of radio advertising in low and high
advertising clutter formats. International Journal ofAdvertising 25 (l):71-86.
Ryals, Lynette (2003)," Making Customers Pay: Measuring and Managing Customer
Risk and Returns," ofS trategic Marketing, 11,3, 165-175.
Sawhney and Zabin (2002), "Managing and Measuring Relational Equity in the Network
Economy," Journal oft he Academy ofMarketing Sciences, 30, 4, 313-332.
Savitt, Ronald (2000)," A Philosophical Essay about a General Theory of Competition:
Resources, Competences, Productivity, Economic Growth," Journal of
Macromarketing, 20, 1, 73.
Schlegelmilch, Bodo B.( 2002), "Special Symposium on Shelby D. Hunt's 'A General
Theory of Competition: Resources. Competences, Productivity Economic Growth'
Part \ ," Journal ofM arketing Management, 18, 1/2, 221-227.
Schreiber, Elliot S.( 2002)," Brand Strategy Frameworks for Diversified Companies and
Brand Management. 10,2, 122-138.
Seines, Fred, and James Sallis (2003)," Promoting Relationship Learning," Jourwn/ of
Marketing, 67 (July), 80-95.
Skaates, Maria Anne and Veikko Seppanen (2005), "Market-Oriented Resource
Management in Customer Relationships," Qualitative Market Research: An
International Journal, 8, 1, 77-96.
Srivastava, Rajendra, Tasadduq A. Shervani and Liam Fahey (1999)," Marketing,
Business Processes, and Shareholder Value: An Organizationally Embedded View of
Marketing Activities and The Discipline of Marketing," Journal of Marketing, 63
(Special Issue), 168-179.
Stringefellow, Anne, Winter Nie and David E. Bowen (2004)," CRM: Profiting from
Understanding Customer Needs," Business Horizons, 47, 5, 45-52.
Stuart, Toby E.( 2000), "Interorganizational Alliances and the Performance of Firms: A
Study of Growth and Innovation Rates in a High-Technology Industry," Strategic
Management Journal, 21, 791 -811.
Subrahmanyan, Saroja (2004)," Effects of Equity and Product Type on the Choice of
Cause-Related Brands: A Singapore Perspective," Journal ofP roduct & Brand
Management, 13,2, 116-124.
Subramani, Mani R. and N. Venkatraman (2003)," Safeguarding Investments in
Asymmetric Interorganizational Relationships: Theory and Evidence," Academy of
Management Journal, 46, 1, 46-62.
Teece David J., Gary Pisano and Amy Shuen (1997)," Dynamic Capabilities and
Strategic Management," Strategic Management Journal 18, 7, 509-533.
Thoroughbred Owners and Breeders Membership Directory (2006-07), Lexington, KY.
Ulaga, Wolfgang and Andreas Eggert (2006), "Relationship Value and Relationship
Quality," European Journal ofMarketing, 40, 3/4, 311-327.
Webster, Frederick E. Jr. (1992)," The Changing Role of Marketing in Corporation",
Journal ofMarketing, 56 (October), 1-17.
Webster. Frederick E. Jr.. and Kevin L. Keller (2004)." A Roadmap for Branding in B2B
Markets," Brand Management, 11,5, 388-402.
Wemerfelt, Birger (1989), "From Critical Resource to Corporate Strategy," Journal of
General Management, 14, 3, 4-12.
Wemerfelt, Birger (1984)," A Resource Based View of the Finn," Strategic Management
Journal, 5, 171-80.
Wilkie, William L.( 1990), Consumer Behavior, Second Edition, New York: John Wiley
& Sons.
Yin, Robert K.( 1994), Case Study Research: Design and Methods, Applied Social
Research Methods Series, Volume 5, Thousand Oaks, California: Sage Publications,
Inc.
Zablah, Alex R., Danny N. Bellenger and Wesley J. Johnson (2004)," An Evaluation of
Divergent Perspectives on Customer Relationship Management: Towards a Common
Understanding of an Emerging Phenomenon," Industrial Marketing Management, 33,
6, 475-489.