Sei sulla pagina 1di 8

Does customer sex influence the relationship between

perceived quality and share of wallet?


Emin Babakus
a,

, Ugur Yavas
b
a
Department of Marketing and Supply Chain Management, Fogelman College of Business and Economics,
The University of Memphis, Memphis, TN 38152, United States
b
College of Business and Technology, East Tennessee State University, United States
Received 21 August 2006; accepted 30 October 2007
Abstract
This study examines the within and between sex differences on the relative importances of merchandise quality and interaction quality as
drivers of behavioral loyalty (share of wallet). A sample of customers of a national retail chain serves as the study setting. The results indicate that
for female customers, interaction quality and merchandise quality have equally strong total influence on share of wallet. For male customers, the
total effect of merchandise quality on share of wallet is significantly stronger than the total effect of interaction quality. Across sex comparisons
reveal that the total effect of interaction quality on share of wallet is stronger for female customers. However, merchandise quality exerts a stronger
total influence on share of wallet for male customers. Implications of the results are discussed and avenues for future research are offered.
2007 Elsevier Inc. All rights reserved.
Keywords: Store loyalty; Share of wallet; Customer sex; Interaction quality; Merchandise quality
1. Introduction
Facing stiffening competition, increasing operational costs
and declining profit margins these days, retailers constantly
strive to have loyal customers and capture a larger portion of
tightly held consumer dollars. This is because loyal customers
buy and spend more; generate more predictable sales; and are
less costly and less price sensitive (Harris and Goode, 2004;
Kumar and Shah, 2004). Yet, research on store loyalty is limited
(Bloemer and de Ruyter, 1998; Macintosh and Lockshin, 1997;
Odekerken-Schroder et al., 2001) and several gaps remain.
First, few studies examine the differential effects of the
relational (interaction with customers) and core (tangible goods/
merchandise) components of perceived quality on loyalty (Brady
and Cronin, 2001; Iacobucci and Ostrom, 1993). Second, the
evidence on the influence of customer characteristics on store
loyalty is scanty (Mittall and Kamakura, 2001; Noble et al., 2006).
In particular, little is known about the potential effects of sex on
the relationship between perceived quality and store loyalty
(Odekerken-Schroder et al., 2001) as well as the relative
importance each sex places on relational and core components
of quality. Third, prior research primarily uses behavioral inten-
tions as a measure of loyalty (Brady et al., 2005). While customer
loyalty encompasses attitudinal components including behavioral
intentions (Dick and Basu, 1994; Oliver, 1999), intentions do not
necessarily translate into behavior and managers are interested in
actual behavior (Mittal and Kamakura, 2001).
This study aims to partially fill in these gaps by developing a
model that examines the relative effects of perceived merchan-
dise quality and interaction quality on store loyalty. This model,
based on well-accepted precepts in the literature, serves as a
theoretical platform for testing four specific sex-related research
hypotheses to address the following overall research question:
Are there within and between sex differences on the relative
importances of merchandise quality and interaction quality as
drivers of behavioral loyalty to a store? Share of wallet or
expenditure at a specific store as a fraction of total category
expenditures is used as a loyalty measure. Share of wallet is the
Available online at www.sciencedirect.com
Journal of Business Research 61 (2008) 974981

Corresponding author. Tel.: +1 901 678 3857.


E-mail address: ebabakus@memphis.edu (E. Babakus).
0148-2963/$ - see front matter 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jbusres.2007.10.005
most critical behavior-related marketing performance metric
(Zeithaml, 2000) and when several stores offer similar products/
services, share of wallet is the key outcome indicating loyalty
from a customer's perspective (Babin and Attaway, 2000).
The model and hypotheses are tested using automotive parts
and accessories retail customers. Although still dominated by
men, the composition of customers in this sector has been
changing rapidly. For instance, close to 85% of American
women are responsible for the maintenance of their vehicles,
and more than a third (34%) of the automotive do-it-yourself
customers are female (Guyette, 2004, p. 63). Considering most
prior research focuses on female dominated retail shopping
contexts (e.g., Noble et al., 2006), the current setting, where
male customers are dominant, provides a fresh platform for
testing important theoretical propositions regarding sex effects.
The study also carries implications regarding the proper focus
on merchandise (core service) and service (interaction with
customers) to attract a larger share of wallet. Such insights are
particularly important for managers in automotive parts and
accessories sector, who strive for better ways to serve the fast
growing female segment (Alexander, 2001).
2. Conceptual framework and hypotheses
2.1. Research model
The model in Fig. 1 posits that two critical variables under
management control, interaction quality (quality of the frontline
employees' interaction with customers) and merchandise
quality (the core tangible aspects of the service including
availability, variety and quality of merchandise) play significant
roles in the formation of store loyalty. They influence share of
wallet directly and indirectly, mediated by perceived value and
customer satisfaction.
Bagozzi's (1992) attitude framework suggests that an
individual's appraisal of his/her environment (cognitions)
produces particular affective responses that subsequently lead
to various behaviors. This is consonant with Oliver (1999), who
expanded the definition of loyalty by arguing that understanding
of loyalty requires careful assessments of consumers' cognitive
evaluations of a brand relative to other brands, affective
preferences for the particular brand, conations or behavioral
intentions to buy the brand, and action loyalty (actual purchase).
While a consensus has yet to emerge on the definition of loyalty
(East et al., 2005; Odin et al., 2001), recent emphasis is on
behavioral metrics with a focus on capturing a larger share of the
customer's wallet (Babin and Attaway, 2000).
Consumers' cognitive appraisals of stores may involve a
variety of factors (Ghosh, 1994; Grewal et al., 1998). The focus
here is on three cognitive constructs (merchandise quality,
interaction quality, and perceived value). Merchandise quality
and interaction quality being two components of perceived
quality are both critical for retail success (Ghosh, 1994; Grewal
et al., 1998; Odekerken-Schroder et al., 2001). They directly
influence share of wallet, and perceived value cognitions follow
from evaluations of merchandise quality and interaction quality
Fig. 1. Research model.
975 E. Babakus, U. Yavas / Journal of Business Research 61 (2008) 974981
(Brady et al., 2005). Perceived quality (merchandise and inter-
action quality) and perceived value lead to customer satisfaction
and consequently to share of wallet.
Research suggests that perceived quality and perceived value
may also have direct influences on loyalty in addition to indirect
effects through customer satisfaction (Brady et al., 2005; Cronin
et al., 2000; Odekerken-Schroder et al., 2001). For instance,
Brady et al. (2005) found that perceived quality and value had
direct influences on repeat purchase intentions, and quality had
direct effects on value and satisfaction while value had a
direct impact on satisfaction. Also, consistent with the current
conceptualization, Sirohi et al. (1998) found that merchandise
quality affects store loyalty through perceived value.
The preceding discussion suggests that the model in Fig. 1
provides a viable representation of retail store loyalty and offers
a framework to guide this study. Given the strong theoretical
and empirical support for the linkages in the model, a specific
hypothesis for each linkage is not advanced. Instead, a test of
the model as an overall default hypothesis is provided and the
model is used as a platform for testing sex-related hypotheses.
2.2. Role of customer sex
The importance of relational and core components of service
quality differs across customers. Some customers value
relational aspects of the retail experience, while others care
only about the quality of the core offering (Fruchter and Sigue,
2005 ; Iacobucci and Ostrom, 1993). Customer sex may be such
a differentiator concerning the effects of interaction quality and
merchandise quality on share of wallet.
The literature suggests that the differences between the two
sexes partially stem from biological make-up and partially from
socialization experiences (Putrevu, 2001). Biological explana-
tions focus on hormonal and brain lateralization differences.
Hormones are responsible for the differences in mood and
personality, and subsequent differences in activities and in-
terests of the two sexes (Berenbaum, 1999). Brain lateralization
researchers contend that females' hemispheres may be more
symmetrically organized while men's hemispheres are more
specialized (Saucier and Elias, 2001). Such differences lead to
distinct information processing capabilities where women
process information in a more integrative and detail-oriented
fashion while men focus on few details.
The socialization perspective holds that societal norms and
practices create traditional sex roles that lead to differences in
motivation, behavior and personality (Eagly, 1987). Males are
primarily guided by societal norms that require control, mastery
and self-efficacy to pursue self-centered goals. Females, on the
other hand, are guided by concerns for self and others. They
emphasize affiliation and harmonious relationships with others.
Evidence supports these premises and shows that men are more
task or goal-oriented and women are more relationship-oriented
(Meyers-Levy, 1988).
Regardless of the sources of the differences, biological or
social/psychological, the unique characteristics of each sex and
distinctions between them have implications regarding how
each sex observes the environment, processes, evaluates and
retrieves information, and makes judgments. Women tend to
approach information processing in a more integrative fashion
by assimilating details whereas men tend to focus less on details
and employ a more selective, heuristics-based information
processing approach (Phillip and Suri, 2004; Meyers-Levy,
1988). Women are guided more by emotion-laden communal
goals, which seek harmony, affiliation, and betterment of self
and others. They are also more likely to value and use infor-
mation provided by others in decision making whereas men
emphasize independence and achievement (Meyers-Levy,
1988). These findings carry implications regarding the relative
importances of merchandise and interaction quality as drivers of
share of wallet within and between sexes.
2.2.1. Within sex hypotheses
Female customers elaborate and integrate both interaction
and core aspects of service quality. Hence, they are expected to
place more emphasis on interaction quality. This is because
women have strong desire for affiliation and place high value on
interacting with the store employees who provide them with
informational and interpersonal cues (Noble et al., 2006).
On the other hand, male customers are expected to attach
more importance to merchandise quality relative to interaction
quality. Research shows that male customers emphasize a few
key attributes in their decision process (Putrevu, 2001) and their
primary focus is on the core/tangible merchandise attributes
(Iacobucci and Ostrom, 1993). This does not mean that
interaction quality is not a significant predictor of store loyalty
for males. Rather, because of their tendency to be independent,
self-focused and mainly motivated by the purchase transaction,
men favor merchandise quality more than interaction quality.
Indeed, men consider shopping as a needs-driven task that has
to be completed efficiently (Noble et al., 2006). In addition,
men's grab and go and get it over with as soon as possible
tendencies and less interest in interaction (Otnes and McGrath,
2001) suggest that they place higher importance on merchandise
quality relative to interaction quality. Thus:
H1a. For female customers, interaction quality will have a
stronger total influence on share of wallet than merchandise
quality.
H1b. For male customers, merchandise quality will have a
stronger total influence on share of wallet than interaction
quality.
2.2.2. Between sex hypotheses
Due to women's desires for more affiliation and interperso-
nal connection, greater openness to persuasion and tendencies
to value opinions of others relative to men (Putrevu, 2001), they
are expected to place more emphasis on interaction quality
relative to men. Men, on the other hand, are less emotional, less
relationship-oriented and more task-focused as they attempt to
simplify the decision process (Phillip and Suri, 2004). Such
tendencies, combined with men's greater emphasis on self-
efficacy should lead them to focus on merchandise quality more
than women. This reasoning reflects Iacobucci and Ostrom's
(1993) proposition that, in the short run, females attach more
976 E. Babakus, U. Yavas / Journal of Business Research 61 (2008) 974981
importance to the interaction (relationship) quality of the service
relative to the core service quality. However, contrary to
Iacobucci and Ostrom (1993), sex differences are not expected
to disappear in the long run (e.g., ongoing service relationships
as in this study setting) and female customers will continue to
emphasize interaction quality more than males, while males will
continue to emphasize merchandise quality more than females.
Chiu (2002) supports this contention and suggests a stronger
relationship between interaction quality and females' behavioral
intentions and a stronger relationship between merchandise
quality and males' behavioral intentions. Also, empirically
Mittal and Kamakura (2001) found that service quality
(including interaction quality) was more important to females
whereas males placed more emphasis on the tangible product
quality. Thus:
H2a. The total influence of interaction quality on share of wallet
will be stronger for female customers than male customers.
H2b. The total influence of merchandise quality on share of
wallet will be stronger for male customers than female
customers.
3. Methodology
The research model and hypotheses were tested using data
from the customer database of a national retailer of auto-
mobile accessories and replacement parts which operates
more than 3000 stores across 30 states. Females constitute
30% of the customers of this retailer. Random samples of 750
female and 750 male customer responses were selected from
the database.
Interaction quality captured customers' perceptions of their
interaction with store employees relative to competition. The
scale had eight items, each scored on a five-point scale ranging
from 1=XYZ is much worse to 5=XYZ is much better than the
competitors. The same scale was used to measure merchandise
quality with five items reflecting various aspects of the products
carried by the retailer, and perceived value via a single item that
captures relative value as a function of quality and price (Fornell
et al., 1996; Sirohi et al., 1998). Relative measures of perceived
quality and value are considered superior to measures based
on noncomparative approaches since they explicitly require
respondents to engage in cognitive evaluations of the focal
retailer relative to competition (Gale, 1994). Customer
satisfaction was measured via three items scored on five-point
scales (1=very dissatisfied to 5=very satisfied).
Share of wallet was measured via two questions. The first
question asked the respondents to indicate where they met the
majority of their shopping needs for this product category.
Responses were recoded as 0 (competition) and 1 (focal
retailer). The second question asked them to indicate the
percentage of total category purchases they made from the focal
retailer on a 10-point scale with 10% increments (1=less than
10% to 10=90% to 100%). Both items were standardized to
create a common metric. The items used to operationalize the
constructs are consonant with the established measures on
service quality (e.g., Cronin et al., 2000; Parasuraman et al.,
1991), perceived value (e.g., Sirohi et al., 1998), customer
satisfaction (e.g., Fornell et al., 1996), and share of wallet (e.g.,
Babin and Attaway, 2000; de Wulf et al., 2001; Macintosh and
Lockshin, 1997).
Table 1
Psychometric properties of scales and two-group CFA results
Scales Common metric
Standardized
loadings
t-values
Relative Interaction Quality (IQ)
(Female: =.93, AVE=.70,
2
=.14.67)
(Male: =.92, AVE=.68,
2
=.21.58)
Please rate XYZ employees relative to competitors (much bettermuch worse)
IQ1. Putting your needs first .84
a
IQ2. Willing to go the extra mile to help you .83 38.4
IQ3. Ability to diagnose your problem .78 34.9
IQ4. Honesty/trustworthiness .83 38.3
IQ5. Listening carefully to you .86 41.1
IQ6. Knowledge about types of products .84 38.8
Relative Merchandise Quality (MQ)
(Female: =.92, AVE=.70,
2
=.14.67)
(Male: =.90, AVE=.65,
2
=.23.58)
Please rate XYZ relative to its competitors (much bettermuch worse)
MQ1. Variety of products to choose from .83
a
MQ2. Being able to get the products you need .86 39.4
MQ3. Having products in stock .86 39.7
MQ4. Quality of the product warranties .76 33.4
MQ5. Carrying name brands .78 34.6
Relative Perceived Value (VAL): (single item)
b
(Female: AVE=.94,
2
=.14.40)
(Male: AVE=.94,
2
=.17.34)
Please rate XYZ relative to its competitors (much bettermuch worse)
VAL. Overall quality for the prices you pay .97
a
Customer Satisfaction (SAT)
(Female: =.88, AVE=.73,
2
=.14.58)
(Male: =.87, AVE=.72,
2
=.24.58)
How satisfied have you been with XYZ? (very satisfiedvery dissatisfied)
CS1. Your overall satisfaction .91 40.8
CS2. Products you have purchased .74 (.86)
c
26.1 27.0)
c
CS3. Services you have received .83
a
Share of Wallet (SOW)
(Female: =.67, AVE=.52,
2
=.13.14)
(Male: =.76, AVE=.63,
2
=.17.24)
SOW1. Where do you buy the majority of your
(product category)?
.67 14.3
SOW2. What percentage of all your
(product category) purchases made at XYZ?
.83
a
Fit statistics (model df =231)

2
=1053.82
RMSEA=.075, NFI =.94, NNFI =.94, CFI =.95
a
Loadings were initially fixed to 1.0.
b
Error variance for this item was fixed as (1reliability) s
2
, where s
2
is the
observed item variance. The communality estimate from exploratory factor
analysis was used as the reliability estimate.
c
CFA results are based on a two-group simultaneous estimation using male
and female sample covariance matrices by imposing equality constraints on, all
but one, factor loadings only (partial metric invariance). Only the factor loading
of the second satisfaction item was free across gender.
977 E. Babakus, U. Yavas / Journal of Business Research 61 (2008) 974981
4. Results
4.1. Measurement results
The measures of interaction quality (8 items), merchandise
quality (5 items), customer satisfaction (3 items), share of wallet
(2 items), and perceived value (single item) were first subjected to
principal components analyses, with oblique rotation, for male
and female data separately. These analyses identified each item
strongly with its respective a priori dimension for each group with
the exception of the same two items in the interaction quality
scale. These items were discarded and a reanalysis produced
highly parsimonious five-factor solutions that accounted for
78.7%and 77.7%of the variance for the female and male samples
respectively. All factor loadings were in excess of .60 and none of
the items cross-loaded heavily (all cross-loadings were below.30
cut-off value suggested by Hair et al. (2006)).
Measurement invariance was examined via a series of con-
firmatory factor analyses using the covariance matrices of
female and male samples simultaneously as input to LISREL
8.51 (Joreskog and Sorbom, 1993). Communality estimates
from exploratory factor analyses were used as reliability esti-
mates for the single-item perceived value measure (Wanous and
Hudy, 2001). The error variance for the perceived value indi-
cator was set to (1reliability)

s
2
where s
2
is the observed item
variance.
First, a five-factor measurement model was tested as the base
model with no equality constraints across sexes. The base model fit
statistics (
2
0
=1034.06, RMSEA=.076, NFI =.94, NNFI =.94,
CFI =.95) and the t-values suggested that the measures exhibited
factorial invariance and that each itemwas a significant indicator of
its underlying construct (Joreskog and Sorbom, 1993).
Second, factor loadings were constrained to be equal across
sexes. The result (
2
232
=1064.83) indicated a significant
deterioration in overall fit over the base model (difference

2
12
=30.77, pb.05). Thus, full metric invariance was unten-
able. However, when the loading equality restriction was
removed for one of the customer satisfaction items (satisfaction
with products purchased), the resulting chi-square (
2
231
=
1053.82) did not indicate a significant deterioration relative to
the base model (difference
2
13
=19.76, pN.05), indicating
partial invariance across sexes (Hair et al., 2006). Third, in
addition to the metric invariance constraints, the factor covari-
ance matrices were set to be equal across the two groups. The
resulting chi-square (
2
246
=1080.53) indicated a significant
deterioration of model fit over the metric invariance model
(difference
2
15
=26.71, pb.05), suggesting that sex may have
a moderating effect (Hair et al., 2006). Table 1 presents scale
items and metric invariance results.
Table 1 also provides coefficient alphas and average
variance extracted (AVE) by each underlying latent variable,
and the shared variances (
2
) among study constructs across
groups. The AVE of each construct was above .50 and con-
sistently higher than the shared variance (
2
) with other con-
structs. These results indicate that the measures are
psychometrically sound (Fornell and Larcker, 1981) and show
partial equivalence across the two samples. Table 2 presents the
correlations, means and standard deviations of the scale items for
each sample.
4.2. Overall model test
The overall viability of the model in Fig. 1 was tested using the
combined female and male data. The sample covariance matrix of
Table 2
Correlations, means and standard deviations of the observed variables
Females (n=693)
IQ1 IQ2 IQ3 IQ4 IQ5 IQ6 MQ1 MQ2 MQ3 MQ4 MQ5 VAL CS1 CS2 CS3 SOW1 SOW2 Mean S.D.
IQ1 1.000 .781 .621 .712 .727 .676 .563 .558 .565 .523 .558 .409 .526 .435 .587 .220 .185 4.04 0.99
IQ2 .777 1.000 .658 .711 .724 .664 .545 .536 .554 .518 .511 .421 .556 .447 .615 .234 .222 4.18 1.03
IQ3 .601 .613 1.000 .693 .673 .700 .555 .541 .558 .571 .553 .459 .510 .440 .531 .248 .243 3.90 1.04
IQ4 .672 .637 .667 1.000 .725 .675 .609 .582 .576 .588 .589 .455 .559 .471 .589 .249 .190 4.11 0.96
IQ5 .718 .679 .652 .674 1.000 .747 .571 .640 .615 .563 .576 .451 .603 .516 .652 .227 .230 4.08 1.02
IQ6 .665 .634 .668 .673 .754 1.000 .598 .636 .639 .605 .635 .443 .547 .516 .652 .249 .243 4.09 1.02
MQ1 .465 .461 .423 .523 .441 .500 1.000 .737 .735 .691 .719 .527 .502 .501 .503 .235 .202 4.18 0.88
MQ2 .496 .520 .503 .538 .517 .557 .737 1.000 .803 .629 .644 .492 .500 .466 .501 .261 .220 4.14 0.97
MQ3 .502 .533 .504 .535 .543 .580 .667 .768 1.000 .681 .667 .531 .527 .503 .558 .240 .254 4.12 1.01
MQ4 .436 .411 .405 .498 .435 .495 .564 .544 .574 1.000 .740 .515 .5.32 .520 .497 .238 .229 4.22 0.93
MQ5 .457 .434 .474 .486 .489 .549 .583 .623 .591 .641 1.000 .515 .488 .509 .468 .212 .207 4.07 0.96
VAL .428 .403 .384 .419 .401 .432 .438 .431 .456 .441 .433 1.000 .437 .409 .418 .282 .231 4.25 0.86
CS1 .572 .550 .456 .563 .543 .535 .434 .480 .472 .480 .446 .423 1.000 .741 .778 .254 .240 4.57 0.75
CS2 .486 .452 .408 .479 .510 .492 .435 .470 .456 .496 .475 .381 .752 1.000 .634 .216 .208 4.57 0.76
CS3 .609 .644 .508 .540 .588 .542 .371 .426 .423 .354 .389 .349 .725 .629 1.000 .228 .221 4.48 0.88
SOW1 .351 .368 .346 .360 .307 .363 .368 .363 .363 .378 .331 .366 .429 .355 .328 1.000 .509 0.00 1.00
SOW2 .290 .318 .296 .275 .242 .278 .237 .237 .240 .300 .219 .275 .311 .242 .285 .608 1.000 0.00 1.00
Males (n=691)
Mean 3.87 3.99 3.65 4.00 3.86 3.77 3.97 3.91 3.94 4.06 3.69 4.12 4.44 4.32 4.30 0.00 0.00
S.D. 0.96 1.02 1.05 0.89 0.98 1.03 0.96 1.00 0.96 0.95 0.97 0.86 0.81 0.91 0.97 1.00 1.00
Perceptual measures were recorded on 5-point scales, where a higher number indicates a more favorable rating. SOW1 and SOW2 were standardized as indicators of
SOW, and higher values indicate higher share of purchases from the focal retailer. Sample sizes were reduced due to listwise deletion of missing cases.
978 E. Babakus, U. Yavas / Journal of Business Research 61 (2008) 974981
the observed variables was used as input to LISREL. Table 3
shows that the model fits the data relatively well (
2
0
=899.93,
RMSEA=.075, NFI =.95, NNFI =.94, CFI =.95). It accounts for
37% of the variance in perceived value, 59% in customer
satisfaction, and 26% in share of wallet. With the exception of the
direct path from interaction quality to share of wallet (
5
=.07,
t =1.16), all linkages are significant (t N2.00). Interaction quality
has no direct effect but an indirect effect on share of wallet,
mediated by perceived value and customer satisfaction. Results
also show that merchandise quality exerts significant direct
and indirect effects on share of wallet. These results validate
the research model as a viable platform for testing sex-related
hypotheses.
4.3. Tests of research hypotheses
The model in Fig. 1 was reestimated using female and male
sample covariance matrices simultaneously as input to LISREL.
Initially, all structural model parameters were freely estimated for
each group without any constraints (except for the constraints
imposed for measurement invariance). Results in Table 3 suggest a
good overall model fit (
2
231
=1058.23, RMSEA=.075, NFI =.94,
NNFI =.94, CFI =.95). This initial set of two-group results served
as the basis for testing the sex hypotheses.
4.3.1. Within sex differences
The total effects reported under the two-group results in
Table 3 lead to two observations regarding H1a and H1b. First,
for female customers, the total effect of interaction quality on
share of wallet (.20) is almost the same as the total effect of
merchandise quality (.21). Second, for male customers, mer-
chandise quality has a stronger total effect on share of wallet
(.30) than interaction quality (.17). While these estimates may
be considered stable due to relatively large sample sizes, H1a
and H1b were formally subjected to
22
234
values were 1064.01
for females and 1082.30 for males. The
222
231
=1058.23),
revealed that the total effects of interaction quality and mer-
chandise quality on share of wallet were not significantly dif-
ferent for females (pN.05), but significantly different for males
(pb.05). These results indicate that H1a is not tenable, but H1b
is supported.
4.3.2. Between sex differences
The total effects reported in Table 3 provide a comparative
perspective between sexes that are relevant to H2a and H2b.
First, consistent with H2a, the total effect of interaction quality
on share of wallet is .20 and .17 for the female and male samples
respectively. Second, the total effect of merchandise quality on
share of wallet for male customers (.30) is larger than for female
customers (.17), which is consistent with H2b. In going beyond
these observations and formally testing H2a and H2b, two
additional
2
difference tests were conducted. First, to compare
the total effect of interaction quality on share of wallet across
females and males, the direct and indirect paths from interaction
quality to share of wallet were constrained to be equal across the
two groups. The resulting
2
237
(1074.34), compared to the
base value (
2
231
=1058.23), indicated that the
2
difference
test was significant (pb.05). Therefore, H2a is supported.
Second, to compare the total effect of merchandise quality on
share of wallet between females and males, the direct and
indirect paths from merchandise quality to share of wallet were
constrained to be equal across the groups. The resulting
2
237
(1071.53), compared to the base value (
2
231
=1058.23), indi-
cated that the
2
difference test was significant (pb.05). Hence,
H2b receives support from data.
5. Discussion
Findings suggest that while both interaction quality and
merchandise quality exert significant influences on share of
wallet, the strengths of these effects vary within and between
sexes. For male customers, the total effect of merchandise
quality on share of wallet is stronger than the total effect of
interaction quality. This is consistent with Iacobucci and
Ostrom's (1993) finding that male customers place a higher
degree of importance on core service (merchandise) quality
relative to relational (interaction) quality. However, contrary to
Iacobucci and Ostrom's (1993) and Mittal and Kamakura's
(2001) findings, female customers view interaction quality and
merchandise quality equally important. This finding may be
attributable to women's elaborate information processing styles
(Putrevu, 2001) in that they integrate detailed cues of both
interaction quality and merchandise quality in their share of
wallet decisions. The finding that male customers place higher
importance on merchandise quality relative to interaction
Table 3
Tests of structural model and research hypotheses
a
Hypothesized
effects
Combined sample results
(n=1384)
Two-group results
(n
f
=693, n
m
=691)
Standardized
estimate
t-value R
2
Standardized
estimate
t-value R
2
Direct effects
IQ VAL (
1
) .16 3.70 .08 (.12) 1.23 (2.20)
MQ VAL (
2
) .47 10.80 .37 .57 (.49) 8.74 (6.81) .40
(.34)
IQ SAT (
3
) .59 14.68 .57 (.60) 9.78 (11.5)
MQ SAT (
4
) .16 3.83 .17 (.15) 2.70 (2.84)
VAL SAT (
1
) .09 3.21 .59 .09 (.08) 2.34 (2.16) .59
(.59)
VAL SOW (
2
) .18 4.72 .20 (.16) 3.35 (3.45)
SAT SOW (
3
) .22 4.21 .14 (.17) 1.87 (2.96)
IQ SOW (
5
) .07 1.16 .10 (.05) 1.04 (0.68)
MQ SOW (
6
) .13 2.28 .26 .07 (.19) 0.72 (3.41) .16
(.31)
Total effects of
IQ on SOW .23 4.29 .20 (.17) 3.29 (2.72)
MQ on SOW .26 4.74 .21 (.30) 3.44 (4.85)
Model fit
statistics

2
=899.93, df =110,
RMSEA=.075

2
=1058.23, df =231,
RMSEA=.075
NFI =.95, NNFI =.94,
CFI =.95
NFI =.94, NNFI =.94,
CFI =.95
Measurement estimates were identical to those from confirmatory factor
analyses up to the second decimal. Therefore, they are not presented. Under two-
group results, female sample results are presented first followed by male sample
results (in parentheses).
979 E. Babakus, U. Yavas / Journal of Business Research 61 (2008) 974981
quality than females is consistent with Mittal and Kamakura
(2001).
Customer sex did influence the total effects of interaction quality
and merchandise quality on share of wallet. Specifically, when
compared with males, females expect a higher quality experience in
their ongoing interactions with store employees. This may be
attributable to female customers' needs to interact with store
employees and receive expert advice from them because of the
nature of the retailer's core product. However, when allocating their
purchases among competing retailers, male customers' decisions
are more heavily influenced by merchandise (core service) quality
compared with females.
The indirect influences of interaction quality and merchandise
quality (mediated by perceived value and customer satisfaction) on
share of wallet were significant regardless of sex. However,
contrary to expectations, interaction quality did not have a
significant direct impact on share of wallet while merchandise
quality had a significant direct influence only for male customers.
Perhaps females' tendency to elaborate information may have
generated heightened levels of perceived value and satisfaction.
Hence, perceived value and satisfaction completely mediated the
effects of both interaction quality and merchandise quality for
female customers. Auto parts and accessories are primarily
utilitarian products. Hence, the nature of merchandise and men's
focus on shopping efficiency and heuristics-based decision making
may have elevated merchandise quality to a prominent role in their
minds, exerting a significant direct effect on share of wallet in
addition to an indirect effect.
5.1. Implications
The foremost implication emerging from this study is that to
increase customers' share of wallet, the retailer should focus on
both interaction quality and merchandise quality. Stronger
emphases on merchandise quality for males and interaction quality
for females are in order. Also, the retailer should continue to
maintain and enhance the interaction skills of its employees through
proper recruitment, selection, training and empowerment. To
maintain and enhance merchandise quality, the retailer should
monitor the quality of the physical products it purchases from the
manufacturers. Additionally, as the average scores in Table 2
suggest, the retailer can capitalize on its favorable perception in
name brand products and also benefit from introducing store
brands.
The emergence of perceived value, another cognitive factor, as a
driver of share of wallet is insightful. Perceived value is a mental
estimate of consumers' cognitive tradeoffs between the value of the
experiences they gain and the sacrifice they make to receive the
gain (cf. Grewal et al., 2004). Thus, to enhance customers'
perceived value perceptions, management should focus on factors
that would reduce the costs and increase the gains. Given the study
results, management can reduce search, time and travel-related
costs to customers by having a wide variety of products available in
stock at all times. Likewise, identifying customers' problems
through free diagnostic services could pay dividends. Similarly, by
training employees to be knowledgeable about the products they
sell and in interpersonal skills, the management can enhance
customer perceptions of the gain side of the perceived value
equation.
6. Limitations and future research directions
This study was conducted among customers of one retailer
offering a specific category of merchandise. Replications among
customers of different retail businesses would be illuminating. For
instance, in fashion clothing, merchandise quality might be of more
interest to women than men.
The model explains a healthy portion of the variance in share of
wallet. Yet, the portion that remains unaccounted for presents a
formidable challenge. Customer share of wallet is a complex
function of store attributes and competitive structure as well as other
factors including enduring relationships with store employees and
preference for particular brands of merchandise (Macintosh and
Lockshin, 1997). Inclusion of other variables besides interaction
quality and merchandise quality that are under direct management
control (e.g., convenience) could improve the predictive power of
the model. Likewise, the affective component of the model could be
expanded by incorporating affective commitment to a retail store
(cf. Pritchard et al., 1999).
Additionally, there is roomfor improvement in the measures. In
particular, perceived value should be measuredvia multiple items to
capture its broader domain (Sweeney and Soutar, 2001) and
achieve sound psychometric properties. Similarly, loyalty measure
comprised of two questions can be complemented by additional
items.
Inclusion of control variables in future studies might lead to
better delineation of sex effects. For instance, demographics (e.g.,
income, age) may affect various relationships posited in the model
(Mittal and Kamakura, 2001). More importantly, although the
female customer segment is growing in the automotive aftermarket,
these products may not have as much personal relevance to females
compared with males. Therefore, consumer involvement (Gabbott
and Hogg, 1999) should be incorporated as a control variable in
future studies. Additionally, here potential differences due to mix of
products sold in various stores/regions and purchased by female
and male customers were not considered. Inclusion of such factors
as control variables is in order. Finally, future studies should
incorporate the role of sex from both retail employee and customer
perspectives (Gutek et al., 1999) and determine if the findings hold
true across the same-sex versus mixed-sex settings.
Acknowledgement
This research was partially supported by a summer research
grant awarded to the first author by the Fogelman College of
Business and Economics at The University of Memphis.
References
Alexander SK. Go the extra mile with your women customers. Aftermark Bus
2001;111(7):567.
Babin BJ, Attaway JS. Atmospheric affect as a tool for creating value and
gaining share of customer. J Bus Res 2000;49(2):919.
Bagozzi RP. The self regulation of attitudes, intentions, and behavior. Soc
Psychol Q 1992;55(2):178204.
980 E. Babakus, U. Yavas / Journal of Business Research 61 (2008) 974981
Berenbaum SA. Effects of early androgens on sex-typed activities and interests
in adolescents with congenital adrenal hyperplasia. Horm Behav 1999;35:
10210.
Bloemer J, de Ruyter K. On the relationship between store image, store
satisfaction and store loyalty. Eur J Mark 1998;32(5/6):499513.
Brady MK, Knight GA, Cronin JJ, Hult GTM, Keillor BD. Removing the
contextual lens: a multi-national, multi-setting comparison of service
evaluation models. J Retail 2005;81(3):21530.
Brady MK, Cronin Jr JJ. Some new thoughts on conceptualizing perceived
service quality: a hierarchical approach. J. Mark. 2001;65(3):3449.
Chiu HC. A study on the cognitive and affective components of service quality.
Total Qual Manag 2002;13(2):26574.
Cronin JJ, Brady MK, Hult GTM. Assessing the effects of quality, value, and
customer satisfaction on behavioral intentions in service environments.
J Retail 2000;76(2):193218.
de Wulf K, Odekerken-Schroder G, Iacobucci D. Investments in consumer
relationships: a cross-country and cross-industry exploration. J Mark
2001;65(4): 3350.
Dick AS, Basu K. Customer loyalty: toward an integrated conceptual
framework. J Acad Mark Sci 1994;22(2):99113.
Eagly AH. Sex differences in social behavior: a social-role interpretation.
Hillsdale, NJ: Erlbaum; 1987.
East R, Gendall P, Hammond K, Lomax W. Consumer loyalty: singular, additive
or interactive. Australasian Mark J 2005;13(2):1026.
Fornell C, Larcker DF. Evaluating structural equation models with unobservable
variables and measurement error. J Mark Res 1981;18(2):3950.
Fornell C, Johnson MD, Anderson EW, Cha J, Bryant BE. The American customer
satisfaction index: nature, purpose, and findings. J Mark 1996;60(4):718.
Fruchter GE, Sigue SP. Transactions vs. relationships: what should the company
emphasize? J. Serv. Res 2005;8(1):1836.
Gabbott M, Hogg G. Consumer involvement in services: replication and
extension. J Bus Res 1999;46(2):15966.
Gale BT. Managing customer value. New York, NY: The Free Press; 1994.
Ghosh A. Retail management. Fort Worth, TX: The Dryden Press; 1994.
Grewal D, Levy M, Lehmann DR. Retail branding and customer loyalty: an
overview. J Retail 2004;80(4):ixxii.
Grewal D, Krishnan R, Baker J, Borin N. The effect of store name, brand name,
and price discounts on consumers' evaluations and purchase intentions.
J Retail 1998;74(3):33152.
Gutek BA, Cherry B, Groth M. Gender and service delivery. In: Powell GN,
editor. Handbook of gender and work. Thousand Oaks, CA: Sage; 1999.
p. 4768.
Guyette JE. Don't think pink. Aftermar Bus 2004;114(1):1, 6364, 6768, 70.
Hair JF, Black B, Babin B, Anderson RE, Tatham RL. Multivariate data
analysis. 6th edition. Orlando, FL: Prentice Hall; 2006.
Harris LC, Goode MH. The four levels of loyalty and the pivotal role of trust: a
study of online service dynamics. J Retail 2004;80(2):13958.
Iacobucci D, Ostrom A. Gender differences in the impact of core and relational
aspects of services on the evaluation of service encounters. J Consum
Psychol 1993;2(3):25786.
Joreskog KG, Sorbom D. LISREL 8: structural equation modeling with
SIMPLIS command language. Hillsdale, NJ: Scientific Software Interna-
tional; 1993.
Kumar V, Shah D. Building and sustaining profitable customer loyalty for the
21st century. J Retail 2004;80(4):31730.
Macintosh G, Lockshin LS. Retail relationships and store loyalty: a multi-level
perspective. Int J Res Mark 1997;14(3):48797.
Meyers-Levy J. The influence of sex roles on judgment. J Consum Res
1988;14(3):52230.
Mittal V, Kamakura WA. Satisfaction, repurchase intent, and repurchase
behavior: investigating the moderating effect of customer characteristics.
J Market Res 2001;38(2):13142.
Noble SM, Griffith DA, Adjei MT. Drivers of local merchant loyalty:
understanding the influence of gender and shopping motives. J Retail
2006;82(3):17788.
Odekerken-Schroder G, de Wulf K, Kasper H, Kleinjnen M, Hoekstra J,
Commandeur H. The impact of quality on store loyalty: a contingency
approach. Total Qual Manag 2001;12(3):30722.
Odin Y, Odin N, Valette-Florence P. Conceptual and operational aspects of brand
loyalty: an empirical investigation. J Bus Res 2001;53(2):7584.
Oliver RL. Whence customer loyalty. J Mark 1999;63(special issue):3344.
Otnes C, McGrath MA. Perceptions and realities of male shopping behavior.
J Retail 2001;77(1):11137.
Parasuraman A, Berry LL, Zeithaml VA. Refinement and reassessment of the
SERVQUAL scale. J Retai 1991;67(4):42050.
Phillip MV, Suri R. Impact of gender differences on the evaluation of
promotional emails. J Advert Res 2004;44(4):3609.
Pritchard MP, Havitz ME, Howard DR. Analyzing the commitment-loyalty link
in service contexts. J Acad Mark Sci 1999;27(3):33348.
Putrevu S. Exploring the origins and information processing differences
between men and women: implications for advertisers. Acad Mark Sci Rev
2001(10) [Online]www.amsreview.org/articles/putrevu10-2001.
Saucier DM, Elias LJ. Lateral and sex differences in manual gesture in
conversations. Laterality 2001;6(3):23945.
Sirohi N, McLaughlin EW, Wittink DR. A model of consumer perceptions and
store loyalty intentions for a supermarket retailer. J Retail 1998;74(2):22345.
Sweeney JC, Soutar GN. Consumer perceived value: the development of a
multiple item scale. J Retail 2001;7(2):20320.
Wanous JP, Hudy MJ. Single-item reliability: a replication and extension. Organ
Res Methods 2001;4(4):36175.
Zeithaml VA. Service quality, profitability, and the economic worth of customers:
what we know and what we need to learn. J Acad Mark Sci 2000;28(1):
6785.
981 E. Babakus, U. Yavas / Journal of Business Research 61 (2008) 974981

Potrebbero piacerti anche