Beruflich Dokumente
Kultur Dokumente
net/publication/312808968
CITATIONS READS
0 878
1 author:
Susan Brudvig
Indiana University East
9 PUBLICATIONS 247 CITATIONS
SEE PROFILE
All content following this page was uploaded by Susan Brudvig on 20 December 2017.
Susan Brudvig
ABSTRACT
Trust is generally recognized as antecedent to long-term, profitable marketing relationships. Since the
publication of Morgan and Hunt’s theory of relationship marketing (1994), there has been a plethora of studies
investigating the role of trust in developing relationships. But only more recently has trust been explicitly examined
in a B2C context, particularly in the consumer-brand domain (Delgado-Ballester 2004). Although attempts to
conceptualize and measure brand trust are numerous, there is no generally accepted brand trust scale. This research
presents a psychometric evaluation of recently developed brand trust scales. The results suggest key distinctions
between scales measuring overall brand trust. It is argued that researchers should consider these distinctions in their
trust is “central to successful relationship marketing, not power and its ability to condition others” (22). Trust is
central to the maintenance of long-term relationships and can be important to a firm’s success. As such, trust has
proliferated as a key construct in the B2B domain (Geyskens, Steenkamp, and Kumar 1998). More recently trust has
emerged as a defining construct in the B2C domain, particularly with regard to trust-building in C2B relationships
(Sirdeshmukh, Singh and Sabol 2002) and consumer trust in brands (Delgado-Ballester 2004). Despite the
significance of trust in building and maintaining customer relationships, few studies have specified and validated of
consumer-based measures of the trust construct (Delgado-Ballester and Munuera-Aleman 2001; Hess 1995).
Brand trust is a particularly relevant construct in the relationship marketing literature, which considers trust
and commitment as key mediating variables in relational exchanges that are highly valued (Morgan and Hunt 1994).
There have been few attempts to develop and validate brand trust scales which are specific to the B2C domain, and
researchers have yet to develop a consensus measure of brand trust. Most typically, studies utilizing consumer-based
measure of brand trust have adapted items from the B2B domain. As such, studies of consumer-based trust may be
characterized by a lack of definitional and methodological standardization. This lack of standardization creates a key
problem for brand trust research: Measures used cannot be compared across empirical studies, perhaps hindering
theory testing and development. As such, the purpose of this paper is to provide a psychometric evaluation and
assessment of consumer-based brand trust scales currently used in the marketing literature.
The outline of the article is as follows. First, a review of the concept of trust is presented. Second, the
methods used to identify overall measures of brand trust are discussed. Then, a psychometric assessment is
undertaken using consumer-based measures of overall brand trust and dimensions of brand trust. The paper
concludes with a discussion of limitations, directions for future research, and managerial relevance.
1. BRAND TRUST
Several definitions of trust are in the literature. Although researchers have invoked somewhat different
conceptualizations, a common theme among them is a willingness to rely on a partner. For example, in the B2B
marketing literature, Moorman, Zaltman, and Deshpande (1992) define trust as “willingness to rely on an exchange
partner in whom one has confidence” (315). Similarly, Ganesan (1994) defines trust as a willingness to depend upon
another. The focus of trust can be inter-personal or inter-firm, such as trust in a sales representative or trust in
supplier. In contrast, in B2C contexts, the role of a trademark or brand, in addition to the role of employee
-2-
behaviors, may be salient. In particular, brand trust is thought to have a role in developing positive, favorable
attitudes toward a brand, resulting in brand commitment or loyalty (Delgado-Ballester and Munuera-Aleman 2001).
Overall brand trust has been defined as the “willingness of the average consumer to rely on the ability of
the brand to perform its stated function” (Chaudhuri and Holbrook 2001). This brand trust conceptualization taps
into behavioral intent, which is relevant in situations of uncertainty. In particular, trust reduced uncertainty,
especially when consumers feel vulnerable when they can rely on a trusted brand (Chaudhuri and Holbrook 2001).
Brand trust has also been defined as the “a feeling of security held by the consumer that the brand will meet his or
her consumption expectations” (Delgado-Ballester and Munuera-Aleman 2001). This conceptualization taps into
consumer inferences of promised or expected performance (Delgado-Ballester and Munuera-Aleman 2001). Like
Chaudhuri and Holbrook’s (2001) conceptualization, brand trust is particularly relevant in the context of inferring
how a brand will behave under uncertainty – in situations not previously experienced (Delgado-Ballester and
Munuera-Aleman 2001). While not specific to branded products, Sirdeshmukh, Singh and Sabol’s (2002) definition
is relevant to branded services in that they define consumer trust as the “expectations held by the consumer that the
service provider 1 is dependable and can be relied on to deliver on its promises.” As highlighted by the above
discussion, consumer’s beliefs about reliability, integrity, and competence are importance facets of brand trust
2. LITERATURE SEARCH
Overall brand trust items and dimensions of brand trust were generated through an extensive literature
search conducted in the spring of 2004. 2 The goal was to identify published consumer-based measures relating to
trust in branded products or branded services, rather than to identify scales that may be adaptable for this purpose.
As a result, the search specifically excluded marketing management research on consumers’ trust in highly
researched targets such as management, contact employees, and salespersons. Searches were conducted of electronic
databases (Proquest ABI/Inform), scale books (Bearden and Netemeyer 1999; Bruner and Hensel 1994), the
contents of top marketing journals (e.g., Journal of Marketing, Journal of the Academy of Marketing Science,
Journal of Business Research, etc.) reverse citation searches of key articles using Social Science Citation Index, and
Overall brand trust is a construct at a higher level of abstraction than specific dimensions of brand trust. To
evaluate a multi-item measure of overall brand trust, the literature search goal was to identify scales that were
-3-
developed for consumers, rather than scales that were adapted to a B2C domain without validation. This would
ensure that an overall brand trust scale was assessed in an exploratory context on a minimum of one independent
sample prior to use in this confirmatory context. Several additional criteria were used for selecting brand trust scales
to evaluate. First, the investigation was restricted to overall, global assessment rather than specific dimensions of
trust (e.g., Delgado-Ballester and Munuera-Aleman 2001; Gurviez and Korchia 2002; Hess 1995). Second, the
measures must be consumer-based, investigating specific branded versions of products or services. Third, consistent
with recommendations for scale construction (Spector 1992), the search was restricted to scales containing four or
more items.
As highlighted in Table 1, the search identified three consumer-based overall brand trust scales (Chaudhuri
and Holbrook 2001; Lau and Lee 1999; and Sirdeshmukh, Singh and Sabol 2002). Two of the studies in which the
scales were measured on respondent-selected brands and one on researcher-selected brands. Two of the studies were
based on goods and products, and one of the studies was based on services. Likert-style agreement and bipolar
semantic differential scales have been used. All scales exceeded alpha reliability of 0.70. As indicated in Table 1,
one common element among the scales is their lack of established validity. The authors have measured overall brand
trust by adapting trust measures from the B2B literature. So, these research studies appear to have proceeded without
3. METHODOLOGY
The purpose of this study was to evaluate and validate published overall measures of consumer-based brand
trust. Participants in this study included 231 student consumers at a large southeastern university in marketing and
management classes. Students received course credit in exchange for study participation. Subjects were recruited to
participate in “market research regarding a UK company considering entering the US market” by the author posing
as an executive representing the company. Student subjects were debriefed following data collection.
Prior to administration of the brand trust items, the subjects were exposed to realistic manipulation
materials consisting of Wall Street Journal (WSJ) articles describing the potential market entry of a branded, but
fictitious, cellular service provider. The WSJ article contained information about the company’s network
performance and customer satisfaction ratings. Subjects were instructed to read one of two versions of the WSJ
manipulation materials – one version primed a high level of trust in the company and the second version primed a
-4-
lower level of trust. Cellular service was chosen as the product due the relevance of the category to student
consumers. However, the use of a fictitious company may eliminate potential confounds with brand experience and
familiarity.
A questionnaire was used to collect data for the scale validation. In addition to the three overall measures of
brand trust, the questionnaire included several multi-item constructs to be used in the validation. Two dimensions of
trust – a multi-item reliability measure (Delgado-Ballester, Munuera-Aleman, and Yague-Guillen 2003) and a multi-
item benevolence measure (Garbarino and Lee 2003) – were included in order to determine specific dimensions that
the overall measures of brand trust tap. Company expertise (Newell and Goldsmith 2001) and modified personal
involvement (Mittal 1995) were included for discriminant and convergent validity tests. Questionnaire items used in
the final questionnaires are presented in Table 2. The questionnaire was pre-tested on 38 student respondents in the
spring of 2004. The validation sample was collected in the summer of 2004 and spring of 2005
It is generally agreed that some method must be used to evaluate dimensionality, reliability, and validity
before measures may be used an indicators of a concept, regardless of the statistical technique to be used to analyze
data (Spector 1999). Because reliability presumes unidimensionality and validity presumes reliability, a three-step
sequential approach was followed to analyze the scales’ psychometric properties. First, the overall scales were tested
for unidimensionality and they scales were pruned if necessary. Then, construct reliability was evaluated. Finally,
several dimensionality checks were performed. Given that this study used published scales and a hypothesized
structure for overall brand trust scales, the measurement testing approach is confirmatory.
4.1. Unidimensionality
Some constructs can be specific and narrow while others can be broad with each item representing some
underlying dimension (i.e., multi-dimensional representation in a single construct). Whether specific and narrow or
broad and encompassing, a construct must be homogeneous, and homogeneity is tested with unidimensionality
analysis (Spector 1992). The importance of unidimensionality cannot be understated. It is a necessary condition for
both reliability and validity (Gerbing and Anderson 1988; Hattie 1985). In addition, demonstrating
unidimensionality in structural analyses avoids interpretational confounding (Bagozzi 1980) and is a requisite for
-5-
It is generally agreed that some method must be used to assess unidimensionality. However, procedures for
assessing unidimensionality appear to be not well-known and apparently infrequently used (Ping 2004). In a review
of recent research, Ping (2004) found that most marketing studies assumed unidimensionality in the specification of
a measurement model. Recently published recommendations for structural equation studies do not provide
information on assessing unidimensionality (Shook, Ketchen, Hult and Kacmar 2004). Assessments of internal
consistency (i.e., Cronbach’s reliability or the alpha coefficient) are not a substitute for assessing unidimensionality,
and the popular approach of omitting items to improve reliability or fit may actually undermine unidimensionality,
average variance extracted, content validity, and/or face validity (Ping 2004).
Unfortunately, the published overall brand trust scales provide little guidance on this issue. The overall
brand trust scales (Chaudhuri & Holbrook 2001; Lau & Lee 1999; Singh and Sirdeshmukh 2000) do not explicitly
address unidimensionality of their respective overall brand trust scales. While the authors assess reliability with
coefficient alpha, a measure of internal consistency is not an adequate substitute as a measure of unidimensionality
(Ping 2004). The scales also contain three or four items, and assessing unidimensionality with scales that contain
three or four items can be difficult to demonstrate because these scales can be under- or just-identified.
In a confirmatory context, unidimensionality can be suggested by a measurement model that fits the data
when its constructs are specified as unidimensional (i.e., each observed variable or indicator is connected to one
latent construct). In addition to overall measures of fit, one should examine the loadings, significance and error.
Unidimensional constructs should have standardized loadings greater than 0.50 (Shook et al. 2004), significant
loadings (Gerbing and Anderson 1988), and a ratio of standard error to the loading of greater than 10. Finally, the
shared residual variance of individual items, with values less than the absolute value of 2.58, are considered to be
To assess unidimensionality of the overall brand items, the approach adopted was to evaluate the each
scale’s items in a single factor measurement model (Ping 2004). The process was iterative, with items pruned, if
necessary, to meet the criteria of unidimensionality noted above. As indicated in Table 3, only one scale
(Sirdeshmukh, Singh and Sabol 2002) passed the unidimensionality criteria without eliminating items. For two
scales (Chaudhuri and Holbrook 2001; Lau and Lee 1999), one item was eliminated in each scale to meet
unidimensionality criteria. However, no scale met model fit criteria, which would be suggested by CFI of 0.95 or
greater and RMR less than 0.09 (Hu and Bentler 1999).
-6-
--- Table 3 about here ---
4.2. Reliability
To test reliability and validity, a measurement model was prepared and assessed. Using covariance
structure modeling, the measurement model tests all constructs where each item is restricted to load on its
corresponding construct and all constructs are allowed to correlate (Bryne 1998). Therefore, three measurement
models were prepared, each containing one of the overall brand trust scales and the remaining measures: reliability
(RELY), benevolence (BENEV), involvement (INVOLVE), and corporate expertise (CORPEXP). The fits
Reliability is frequently characterized as the repeatability of a measure (Strube 2000), but may be more
appropriately referred to as the degree to which as set of latent construct indicators are consistent with their
measurement (Hair, Anderson, Tatham and Black 1998). For computing reliability of items measured with error in a
confirmatory context, a frequently used measure is construct reliability (Fornell & Larcker 1981). While there is no
firm rule of thumb, a level of 0.70 is generally considered the minimum level of reliability, and many authors
suggest an acceptable level of reliability at 0.80 or more (Nunnally and Bernstein 1994). Each of the overall brand
trust scales exhibited strong construct reliability ranging from 0.80 to 0.93. (See Table 4.)
In addition to construct reliability, average variance extracted (AVE) can be used as a reliability criterion.
AVE is the percent amount of shared or common variance among the indicators for a construct, with higher values
representing a greater degree of shared representation of the indicators with the construct (Hair et al. 1998). As a test
of reliability, it assesses the amount of variance captured by the construct in relation to the amount of variance due
to random error. Fornell and Larcker (1981) suggest that a construct's measure should extract at least fifty percent of
the variance (i.e., AVE should be 0.50 or above). AVE of each of the overall brand trust scales exceeded 0.50,
4.3. Validity
Validity is the ability of a construct’s indicators to measure the construct accurately (Hair et al. 1998).
However, interpreting what the scales represent is difficult because validation “can only occur within a system of
hypothesized relations between the construct of interest and other constructs. Tests for validity involve simultaneous
tests of the hypotheses about constructs and hypotheses about scales” (Spector 1992, 46). Validity reflects how well
-7-
a measure reflects its unobservable construct. Validity is established using relationships between observed variables
and their unobserved variable, and observed variables' relationships with other sets of observed variables. While
there are several different types of validity, the following will be tested here: convergent, discriminant, concurrent,
and know-groups.
Convergent validity – the degree to which the items converges on the theoretical construct – can be
assessed in an SEM context. Adequate convergent validity is frequently hinted at by reliabilities of .8 or higher
and/or demonstrated by an AVE above 0.50 (Ping 2004). A measure’s overall validity is then qualitatively assessed
considering this minimal set of criteria. The three overall brand trust scales pass the test of convergent validity.
Discriminant validity is the degree to which measures of different constructs are distinct (Bryant 2000).
Measures of different constructs should relate only modestly to each other if they are valid measures of different
constructs. In an SEM framework, the Fornell and Larcker (1981) test of discriminant validity is used. It is
established by comparing shared variance between pairs of constructs to average variance extracted. (See Table 5.)
All pairs of constructs passed the test of discriminant validity with the exception of the reliability dimension for two
overall brand trust scales (i.e., the shared variance between overall brand trust and reliability was larger than the
average variance extracted of overall brand trust). This result suggests that two overall brand trust scales –
Chaudhuri and Holbrook (2001) and Lau and Lee (1999) – can not be discriminated from reliability as a sub-
dimension of brand trust. This may suggest that neither scale is a broad, global assessment of brand trust, but rather
Concurrent validity is tested by simultaneously collecting data from a sample using the scales of interest
and a criterion hypothesized to relate to the scale of interest (Spector, 1992; Bryant 2000). The hypotheses are that
the will correlate with one or more criteria, and statistically significant relationships are taken as support for validity.
Concurrent validity was established by embedding the corporate expertise scale (Newell and Goldsmith 2001) in the
questionnaire. It is hypothesized that corporate expertise will be positively and strongly related to brand trust.
Corporate credibility was found to be strongly and significantly related to the three brand trust scales. (See Table 6.)
Known-groups validity can be established if it can be hypothesized that different groups of respondents will
score differently on the scale of interest (Spector 1992). Known-groups validity of the overall brand trust scales was
-8-
assessed by investigating whether the scales’ mean score could distinguish between the groups of subjects in the
high and low trust conditions. ANOVA was performed to test for difference in the means of the known-groups. All
The goal of this study was to identify and evaluate consumer-based overall brand trust scales. (See Table
8.) One overall brand trust – Sirdeshmukh, Singh and Sabol (2001) – measure satisfied the criteria of
unidimensionality, reliability, and validity through a series of statistical tests. The overall scales of Chaudhuri and
Holbrook (2001) and Lau and Lee (1999) passed unidimensionality after an item was eliminated from each scale.
On inspection of the items for these two scales, it can be seen that beliefs about reliability, safety and honesty are
important facets of trust operationalized by these two scales. As such, it is understandable that the scales would not
pass discriminant validity with an underlying dimension of brand trust (e.g., reliability).
5. CONCLUSIONS
While definitional considerations have received some attention in the brand trust literature, this study is the
first known attempt to assess the psychometric properties of currently used brand trust scales. While reliabilities of
the three scales were high, the results suggest that two of the scales tested were not unidimensional and did not
discriminant from a sub-dimension of trust. Because most empirical studies have been found to not examine
unidimensionality or validity, this finding calls into question the use of these scales as measures of overall brand
trust.
One measure satisfied the criteria of unidimensionality, reliability, and validity through a series of
statistical tests. However, these results should be interpreted with caution. First, the study used student consumers, a
fictitious brand, and an experimental context. While prior experience and brand familiarity was eliminated as a
potentially confounding explanation, further validation may shed light on the degree to which these factors might
account for consumers’ perception of brand trust. In addition, the two scales that did not pass unidimensionality --
Chaudhuri and Holbrook (2001) and Lau and Lee (1999) – contained reverse items, and reverse items have
previously been demonstrated to hinder unidimensionality (Herche and Engelland 1996). Finally, method effects
-9-
may account for the overall strong performance of Sirdeshmukh, Singh and Sabol (2002) because it was the only
scale utilizing ten points and the only semantic differential brand trust scale.
It was found that two scales – Chaudhuri and Holbrook (2001) and Lau and Lee (1999) – tapped into a
specific dimension of brand trust. This has both managerial and research implications. If the scales are utilized as a
managerial assessment tool, practitioners should be aware of potentially different conclusions that could result for
including one brand trust scale over another. For researchers who test theories, there are also several implications.
The conceptualization and measurement of overall brand trust differs by authors. If overall brand trust scales tap into
different dimensions of trust, theory-testing and theory-building could be impacted depending on which scale is
used. This suggests a need for a review of the literature, development of theoretical meaning, and further empirical
testing.
It may be that overall brand trust is factorially complex, with many dimensions being tapped. Alternatively,
it may be that a second-order factorial model better represents the brand trust construct. Nevertheless, the findings
suggest that researchers should carefully examine scales for content and face validity, in addition to examining
scales for unidimensionality. It is certainly possible that different (and perhaps better) constructs could exist for
Future research should be initiated. First, future research could address the limitations of the current study.
For instance, adult consumers might be used rather than a convenience sample of student respondents. In addition,
other dimensions of brand trust that have been proposed (e.g., altruism, honesty, credibility) might be selected to
further enhance our understanding of consumer-based measures of brand-trust. Also, a real product, rather than a
fictitious product could be used. Finally, because scale re-specification was undertaken, cross-validation of the
scales is warranted.
- 10 -
REFERENCES
Bearden, William O. and Richard G. Netemeyer. 1999. Handbook of Marketing Scales: Multi-Item Measures for
Marketing and Consumer Behavior Research, 2nd Ed. Palo Alto, CA: Sage.
Bryant, Fred B. 2000. “Assessing the Validity of Measurement,” In Reading and Understanding More Multivariate
Statistics. Eds. Laurence G. Grimm and Paul R. Yarnold. Washington, DC: American Psychological Association,
99-146.
Hess, Jeffery S. 1995. “Construction and Assessment of a Scale to Measure Consumer Trust. In 1995 AMA
Educator’s Proceedings, Vol 6. Eds. Barbara B. Stern and George M. Zinkhan. Chicago: American Marketing
Association, 20-26.
Bruner II, Gordon C. and Paul J. Hensel. 1994. Marketing Scales Handbook, Volume II. Chicago: American
Marketing Association.
Bryne, Barbara M., 1998. Structural Equation Modeling with LISREL, PRELIS, and SIMPLIS. Lawrence Erlbaum,
New Jersey.
Chaudhuri, Arjun and Morris B Holbrook. 2001. “The Chain of Effects from Brand Trust and Brand Affect to Brand
Performance: The Role of Brand Loyalty.” Journal of Marketing 65 (April): 81-93.
Delgado-Ballester, Elena. 2004. “Applicability of a Brand Trust Scale across Product Categories: A Multigroup
Invariance Analysis.” European Journal of Marketing 38 (5/6): 573-592.
---------- and Jose Luis Munuera-Aleman. 2001. “Brand Trust in the Context of Consumer Loyalty,” European
Journal of Marketing, 35 (11/12), 1238-1259.
---------- and ----------. 2005. “Does Brand Trust Matter to Brand Equity?” Journal of Product and Brand
Management 14 (2/3): 187-196.
----------, ----------, and Maria Jesus Yague-Guillen. 2003. “Development and Validation of a Brand Trust Scale,”
International Journal of Market Research 45 (1), 35-55.
Fornell, Claus and David F. Larcker. 1981. “Evaluating Structural Equation Models with Unobserved Variables and
Measurement Error.” Journal of Marketing Research 18 (Feb): 39-50.
Garbarino, Ellen and Olivia F. Lee. 2003. “Dynamic Pricing in Internet Retail: Effects on Consumer Trust.”
Psychology & Marketing 20 (June): 495-513.
Gefen, David and Detmar W. Straub. 2004. “Consumer Trust in B2C e-Commerce ad the Importance of Social
Presence: Experiments in e-Products and e-Services.” Omega 32: 407-424.
Gerbing, David W. and James C. Anderson. 1988. “An Updated Paradigm for Scale Development Incorporating
Unidimensionality and Its Assessment.” Journal of Marketing Research 25 (2): 186-192.
Geyskens, Inge, Jan-Benedict E M Steenkamp, and Nirmalya Kumar. 1998. “Generalizations about Trust in
Marketing Channel Relationships using Meta-Analysis.” International Journal of Research in Marketing 15 (July):
223-247.
- 11 -
Gurviez, Patricia and Michaël Korchia. 2002. “Proposition d'une échelle de Mesure Multidimensionnelle de la
Confiance dans la Marque.” Recherche et Applications en Marketing 17 (3): 41-61.
Ha, Hong-Youl. 2004. “Factors Influencing Consumer Perceptions of Brand Trust Online.” Journal of Product and
Brand Management 13 (4/5): 329-342.
Hair, Joseph F. Jr., Rolph E. Anderson, Ronald L. Tatham, and William C. Black. 1998. Multivariate Data Analysis,
5th Ed. Upper Saddle River, NJ: Prentice Hall.
Hattie, John R. 1985. “Methodological Review: Assessing Unidimensionality of Tests and Items,” Applied
Psychological Measurement 9 (June): 139-164.
Hess, Jeffery S. 1995. “Construction and Assessment of a Scale to Measure Consumer Trust. In 1995 AMA
Educator’s Proceedings, Vol 6. Eds. Barbara B. Stern and George M. Zinkhan. Chicago: American Marketing
Association, 20-26.
Herche, J. and B. Engelland. 1996. “Reversed-Polarity Items and Scale Unidimensionality.” Journal of the Academy
of Marketing Science 24 (Fall): 366-374.
Hu, Li-tze and Peter M. Bentler. 1999. “Cutoff Criteria for Fit Indexes in Covariance Structure Analysis:
Conventional Criteria Versus New Alternatives.” Structural Equation Modeling 6 (1): 1-55.
Lau, Geok Theng and Sook Han Lee. 1999. “Consumers’ Trust in a Brand and the Link to Brand Loyalty.” Journal
of Market Focused Management 4: 341-370.
Mittal, Banwari. 1995. “A Comparative Analysis of Four Scales of Consumer Involvement.” Psychology &
Marketing 12 (7): 663-682.
Moorman, Christine, G. Zaltman, and R. Deshpande. 1992. “Relationships between Providers and Users of Market
Research: The Dynamics of Trust within and between Organizations.” Journal of Marketing Research 29: 314-328.
Morgan, Robert M. and Shelby D. Hunt. 1994. “The Commitment-Trust Theory of Relationship Marketing.” Journal
of Marketing 58 (July): 20-38.
Newell, Stephen J. and Ronald E. Goldsmith. 2001. “The Development of a Scale to Measure Perceived Corporate
Credibility,” Journal of Business Research, 52, 235-247.
Nunnally, Jum C. and Ira H. Bernstein. 1994. Psychometric Theory, 3rd, New York: McGraw-Hill.
Ping Jr., Robert A. 2004. “On Assuring Valid Measures for Theoretical Models using Survey Data.” Journal of
Business Research 57 (February): 125-141.
Reast, Jon D. 2005. “Brand Trust and Brand Extension Acceptance: The Relationship.” Journal of Product and
Brand Management 14 (1): 10-19.
Shook, Christopher L., David J. Ketchen, Jr., G. Tomas M. Hult, and K. Michele Kacmar. 2004. “An Assessment of
the Use of Structural Equation Modeling in Strategic Management Research.” Strategic Management Journal 25:
397-404.
Singh, Jagdip and Deepak Sirdeshmukh. 2000. “Agency and Trust Mechanisms in Consumer Satisfaction and
Loyalty Judgements.” Journal of the Academy of Marketing Science 28 (Winter): 150-167.
Sirdeshmukh, Deepak, Jagdip Singh and Barry Sabol. 2002. “Consumer Trust, Value, and Loyalty in Relational
Exchanges.” Journal of Marketing 66, (January): 15-37.
- 12 -
Smith, JB and WB Barclay. 1997. “The Effects Organizational Difference and Trust on the Effectiveness of Selling
Partner Relationships,” Journal of Marketing 61: 3-21.
Spector, Paul E., 1992. Summated Rating Scale Construction. Sage, Newbury Park, CA.
Strube, Michael J. 2000. “Reliability and Generalizability Theory,” In Reading and Understanding More
Multivariate Statistics. Eds. Laurence G. Grimm and Paul R. Yarnold. Washington, DC: American Psychological
Association, 23-66.
- 13 -
TABLE 1 – OVERALL BRAND TRUST SCALES USED IN PREVIOUS RESEARCH
- 14 -
TABLE 3 – UNIDIMENSIONALITY INDICATORS
# Items 4 5 4
Significant Loadings 4 5 4
Standardized > 0.5 4 5 4
Range of Loadings 0.54 - 0.82 0.52 - 0.78 0.86 - 0.89
Ratio > 10 3 3 4
Error Variance > |2.58| N BT6 - BT8 N
Contains reversed items? N Y N
Respecified Single
Factor Models OBT - CH OBT - LL OBT - SSS
No. Items 3 4 --
Significant Loading 3 4 --
Standardized > 0.5 3 4 --
Range of Loadings 0.68 - 0.87 0.52 - 0.83 --
Ratio > 10 3 3 --
Error Variance > |2.58| * N --
Contains reversed items? N Y --
CFI * -- --
RMR * 0.02 --
Chi-Sq * 0.1 --
df 0 2 --
p * 0.64 --
Chi-Sq/df * 0.0 --
- 15 -
TABLE 4 – MEASUREMENT MODELS
Fit
CFI 0.94 0.94 0.95
RMR 0.09 0.09 0.11
RMSEA 0.07 0.07 0.07
Chi-Sq/df 2.1 2.1 2.0
- 16 -
TABLE 5 – DISCRIMINANT VALIDITY TEST
OBT - CH 0.51 *
OBT - LL 0.47 *
* p<0.01
- 17 -
TABLE 7 – KNOWN-GROUPS VALIDITY TEST
ANOVA Table SS df MS F
* p<0.01
Unidimensionality 2 2 1
Reliability 3 2 1
Variance Extracted 3 2 1
Validity
Discriminant 2 2 1
Known-Groups 1 1 1
Convergent 2 1 3
Concurrent 1 1 1
- 18 -
ENDNOTES
1 In the context of the services literature, it is relevant to consider the “service provider” (e.g., American Express, Cingular,
Ballester and Munuera-Alemán 2005; Reast 2005; Ha 2004). However, the scales used would not qualify for testing: Two brand
trust scales are dimensional (Delgado-Ballester and Munuera-Alemán 2005; Reast 2005) and one overall brand trust scale has
only two items (Ha 2004).
3 In addition to overall definitions of brand trust, the multi-dimensional view of trust has had growing acceptance (Garbarino &
Lee 2003; Gefen and Straub 2004). A number of different dimensions of the trust concept have been proposed, such as honesty,
dependability, credibility, motives and altruism (cf. Ganesan 1994; Smith & Barclay 1997). However, in the consumer brand
domain, trust has been conceptualized as a function of perceived benevolence (Guivez & Korchia 2002; Singh & Sirdeshmukh
2000; Garbarino & Lee 2003) perceived competence (Sirdeshmukh, Singh, & Sabol 2002; Singh & Sirdeshmukh 2000;
Garbarino & Lee 2003), perceived problem solving orientation (Delgado-Ballester, Munuera-Aleman, and Yague-Guillen 2003),
and perceived reliability (Ganesan 1994). While many researchers have posited a multi-dimensional conceptualization of
consumer brand trust, few have tested it (for an exception, see Gefen and Straub 2004).
4 Although not discussed in the body of the paper, the four measures also in the measurement models – reliability, benevolence,
involvement, and corporate expertise – were subjected to, and passed, unidimensionality, reliability, and discriminant validity
tests.
- 19 -