You are on page 1of 18

Store Environment

and Consumer Purchase


Behavior: Mediating Role
of Consumer Emotions
Elaine Sherman and Anil Mathur
Hofstra University
Ruth Belk Smith
University of Baltimore

ABSTRACT

This article presents a large-scale cross-sectional field study of


the effect of store environment on consumer emotions and the
resulting influence on aspects of consumer behavior with
actual shopping behavior used as an example. Cast into a
stimulus – organism – response framework, the results suggest that
a consumer’s emotions can be a mediating factor in the purchase
process. In this study, we identify and explore how store
environment and emotional states may influence various dimensions
of purchase behavior. This research confirms that although cognitive
factors may largely account for store selection and for most planned
purchases within the store, the environment in the store and the
emotional state of consumers may be important determinants of
purchase behavior. This research has many pragmatic applications,
because pleasure was associated with the amount of money spent
and affinity for the store, whereas arousal was associated with
money spent in the store, time spent in the store, and the number of
items purchased in the store. ©1997 John Wiley & Sons, Inc.

Recently, there has been a research focus on the role of affective or


emotional states on decision making and on consumer behavior. An

Psychology & Marketing Vol. 14(4):361 – 378 (July 1997)


© 1997 John Wiley & Sons, Inc. CCC 0742-6046/97/040361-18
361
individual’s emotional state can be important and can influence such
areas as decision making, Isen, Means, Patrick, and Nowicki (1982) ex-
plored decision making and found that people in more positive emo-
tional states will tend to reduce decision complexity and have shorter
decision times. For marketers, knowledge about how shopping behav-
ior of potential customers can be influenced by their emotional states
can be of considerable importance. With the current emphasis on cre-
ating more lasting relationships with customers, the role of store envi-
ronments and what can foster or sustain pleasant emotional reactions
becomes strategically more important. Assessing what factors con-
tribute to creating pleasant or unpleasant shopping experiences can
affect future strategic planning.
Our research is a field study of a large sample of consumers. The
study attempts to overcome the problems of laboratory studies, where
subjects’ various emotional states are artificially induced and con-
sumers’ moods at the point of purchase are measured. In contrast to
experimentally induced moods, naturally occurring emotional states
are more subtle. They do not have the demand artifacts associated
with affect manipulation (e.g., success feedback). Thus, investigation of
naturally occurring emotional states may overcome some of the
threats to validity engendered by manipulation (Simon, 1982). The
present purpose is to investigate the potential influence of the store’s
environment on the emotional states of the consumer and how these
emotions may affect various aspects of purchase behavior and feelings
toward the store.

BACKGROUND LITERATURE

The emotional nature of shopping has been characterized by re-


searchers for some time. Oxenfeldt (1974) suggests customers will
have both opinions and feelings toward certain stores that will influ-
ence their perceptions. Consumer-behavior research has heavily re-
flected the influence of cognitive psychology, focusing and relying upon
the traditional information-processing paradigm to explain or predict
consumer decision making processes and their outcomes (Bettman,
1979). Belk (1975) asserted that the situation must influence the con-
sumer, and offered a topology of situations. One of the antecedent
states, suggested by Belk, included the consumer’s mood. Until then,
mood was not legitimized as a variable of concern in consumer re-
search. It was, however, being studied by psychologists, and much of
what we know about the concept has come from those efforts (e.g.,
Bower, 1981; Clark, 1982; Clark & Isen, 1982; Isen, 1984; Isen et al.,
1982; Mandler, 1979). In recent years, affect has been recognized as
more important in decision making. An individual’s mood can appar-
ently influence behavior without interfering with other cognitive

362 SHERMAN ET AL.


processes (Clark, 1982; Clark & Isen, 1982). Thus, moods may be
thought of as feeling states that gain ascendancy and “describe a phe-
nomenological property of an individual’s subjectively perceived affec-
tive state which will refer to the general, pervasive, affective states
that are transient and particularized to specific times and situations”
(Gardner, 1985, p. 284).
Most recent explorations on the influence of mood have been con-
ducted within the confines of the laboratory, with the use of carefully
delineated tasks and techniques. Specifically, researchers “have at-
tempted to study mood effects by using a mood manipulation to induce
a ‘positive’ or ‘negative’ affective state in experimental subjects” (Hill
& Ward, 1989, p. 97). However, as in any laboratory research, the ques-
tion of external validity arises. As Hill and Ward (1989) point out,
“commonly used mood manipulations may influence more than sub-
jects’ moods and thus confound studies” (p. 97). With regard to con-
sumer behavior, laboratory studies indicate that mood interacts with
involvement and shopping experiences (Swinyard, 1993).
Although it has been more recently suggested that emotion can be a
synonym for affect and mood, some authors suggest that “emotion ap-
pears to be the more encompassing terms, with affect and mood partic-
ular types or examples of emotion” (Babin, Darden, & Griffen, 1992).
Specific references to emotion in the marketing literature are less fre-
quent than references to affect or mood. Often emotion and mood have
been treated interchangeably. However, when Holbrook and Hirschman
(1982) recognized the role of consumers’ feelings and fantasies, they
resurrected interest in the area after the era of motivational research
led by Dichter (1964) lost research momentum. Accumulated research
(Isen, 1984) suggests that there are consistent interrelationships be-
tween emotional states and decision processes. Other studies, such as
Holbrook and Westbrook (1989) explored the emotional responses of
television commercials, and Holbrook and Gardner (1993) explored the
consumption experience.
Earlier research by Mehrabian and Russell (1974) offered a multidi-
mensional perspective in environmental psychology. This framework
consists of antecedents (the attributes of the environment), the inter-
vening emotional state, and a taxonomy of outcomes based on the ap-
proach – avoidance concept suggested by Wundt (1905). Russell (1978)
proposed a three-dimensional schema of pleasure, arousal, and domi-
nance. Donovan and Rossiter (1982) cast these variables onto the stim-
ulus – organism – response (SOR) framework and tested the linkage
between the O and the R variables, with promising results. They sug-
gested that environmental stimuli affect the emotional states of con-
sumers in ways of which they may not be fully aware, but which can
affect approach or avoidance behavior. Such behavior may be observed
in retail patronage, store search, interactions with store personnel,
and in-store behavior. However, the Donovan and Rossiter early study

STORE ENVIRONMENT AND PURCHASE BEHAVIOR 363


(1982) used student subjects and studied attitudes and intention
rather than actual shopping behavior. Sherman and Smith (1986), in a
study using this framework, found support for the findings of Donovan
and Rossiter on the relationship between consumer emotional states
(O) and buying behavior (R). Both studies, however, termed store im-
age as an organismic variable, and both used small sample sizes.
Anderson (1986) also used this framework in a retail environment.
Recently, Baker, Grewal, and Parasurman (1994) explored components
of store environment, how they affect the image of the store, and how
consumers evaluate the quality of the merchandise.
Although much of this past research has helped our understanding
of store environment and how its components may affect consumer de-
cision making, the process of how all these affect the consumers in-
store moods and purchase behavior needs further research. In a more
recent study, Donovan, Rossiter, Marcoolyn, and Nesdale (1994) at-
tempted to replicate and extend Donovan and Rossiter’s (1982) re-
search. These researchers suggest that emotional responses in stores
can be affected by the store’s environment. Although they extended
their research to measure emotions during the shopping experience,
their results had some limitations. They were still based on a small
sample limited to discount department stores. Some of their results
were inconsistent and failed to replicate their prior finding that
arousal was significant in pleasant environments. Although merchan-
dising perceptions and emotional responses independently contributed
to extra time and unplanned shopping, they suggest the need for
studying other types of stores to see if these factors would interact in
other environments. Their study, while contributing to our current
knowledge in this area, indicates the necessity for additional research
to understand the relationship between a shopper’s emotional states
within the store and actual purchase behavior.

Extension of Prior Research to Purchase Behavior


The current study attempts to reinterpret the conceptualization of
store atmosphere, consumer emotions, and buying behavior as stimu-
lus – organism – response in a manner consistent with the environmen-
tal psychology approach (Mehrabian & Russell, 1974) and to add to our
knowledge of the role of emotions in consumption experiences. This cur-
rent research, although it is an extension of related thought, uses a
causal modeling approach in modeling and analysis, and has a large
sample size to ensure external validity. It explores the relationship of
such factors as store image and social dimensions, on pleasure and
arousal and how they can be linked to time and money spent in stores.
The study further extends research by measuring emotions at the time
of purchase to eliminate any bias between actual behavior and inten-
tions. Respondents’ cognition in the store, rather than laboratory set-

364 SHERMAN ET AL.


ting, and their perceptions of various environmental factors, such as
lighting, color design, merchandise, and salespeople are also important
factors explored in terms of impact on their emotional states at the
time of purchase.

THE MODEL

Stimulus
The conceptualization of stimulus as something that rouses or incites
to action or increased action has been used and accepted in the litera-
ture (e.g., Bagozzi, 1980, 1986; Belk, 1975; Kelly, 1955). In a consumer
decision-making context, the stimulus can be conceptualized as those
external factors associated with a pending decision. According to
Bagozzi (1986), when consumer behavior is depicted as a stimulus –
organism – response system, the stimuli are “external to the person”
and consist of both marketing mix variables and other environmental
inputs (p. 46). The decision may be about whether to purchase or save,
what categories of goods or services to purchase, what brands to pur-
chase, how much money to spend, how many different purchases to
make, and/or how products will be used and discarded (Robertson,
Zielinski, & Ward, 1984). In the classical SOR model, the stimulus is
that which affects internal states of the individual. In this model, the
stimulus is the store atmosphere as it affects the mood of the con-
sumer. Baker (1986) presented a typology categorizing the elements of
store environment into three categories: social factors, design factors,
and ambient factors. Social factors relate to other people present in
the store (Baker et al., 1994). Of all such people, salespeople are the
most important because, as components of the marketing mix, a mar-
keter has a significant control over their number, type, and behavior.
Ambient factors relate to nonvisual elements of a store’s environment
(e.g., smell, lighting, etc.). Design factors, on the other hand, are visual
in nature (e.g., layout, color, cleanliness, clutter, space, etc.) (Baker et
al., 1994). Finally, overall store image has also been shown to have an
impact on consumer’s behavior (e.g., store choice) (Nevin & Houston,
1980). Thus, the stimulus is represented by various elements of store
atmosphere and is expected to influence consumer’s mood while shop-
ping.

The Organismic Variable


In a stimulus – organismic – response model of consumer behavior, or-
ganism refers to “internal processes and structures intervening between
stimuli external to the person and the final actions, reactions, or re-
sponses emitted. Notice that the intervening processes and structures
consist of perceptual, physiological, feeling, and thinking activities”

STORE ENVIRONMENT AND PURCHASE BEHAVIOR 365


(Bagozzi, 1986, p. 46). Consistent with this definition, emotional state is
conceptualized as the organismic (intervening) variable. This is to imply
that the effect of store atmosphere (the stimulus) on consumer behavior
is mediated by the consumer’s emotional state. Emotional state is con-
ceptualized as consisting of three domains: pleasure – displeasure,
arousal – nonarousal, and dominance – submissiveness (Mehrabian &
Russell, 1974). However, previous studies have found the dominance fac-
tor to be of little predictive value in similar situations (e.g., Donovan &
Rossiter, 1982; Donovan et al., 1994; Russell & Pratt, 1980), and was
therefore not included in the model. Thus, elements of store atmosphere
are expected to positively influence the components of consumer emo-
tional state (pleasure and arousal).

The Response
Bagozzi (1986) defines response as the outcome or final action toward
or reaction of consumers, including psychological reactions such as at-
titudes and/or behavioral reactions. Wundt (1905) argued that behav-
iors due to mood and environment evaluation can be classified as
approach or avoidance. Donovan and Rossiter (1982) suggested several
responses in a retail environment, also used by Sherman and Smith
(1986), that represent approach or avoidance. They include (1) number
of items purchased, (2) amount of time spent in the store, (3) actual
amount of money spent in the store, and (4) whether the shopper liked
the store environment. Thus, consumer’s emotional state (pleasure and
arousal) will positively influence these four outcome variables. Figure
1 shows the SOR conceptualization of the model. It is consistent with
the SOR models proposed by Lewin (1936), Kelly (1955), and Rotter
(1954) in that the stimulus affects the intervening mental event,
which, in turn, affects the response. Implicit in this conceptualization
is that there is a positive effect of the stimulus upon the response
which is mitigated by the effect of the intervening variable.

METHOD

Sample
The data were collected by intercepting shoppers as they emerged from
fashion stores in shopping malls and requesting them to complete a self-
administered questionnaire. The questionnaire contained items to mea-
sure consumers’ emotional state, their perceptions of store’s atmosphere,
outcome variables, and simple demographic questions. Although it may
be argued that emotional state should be measured in the store, during
the shopping experience, this would require an experimental design, per-
mission of the retailer, and would intrude upon and interrupt the con-
sumer’s emotional state, causing demand artifacts and bias, even
irritation or anger. Possibly for this reason, studies attempting to mea-

366 SHERMAN ET AL.


STIMULUS ORGANISM RESPONSE

Note: Correlations among structural error terms are not shown for clarity.
Figure 1 SOR model of retail shopping behavior.

sure in-store emotional state have had inconsistent results (e.g.,


Donovan et al., 1994). In addition, measurement just after the shopping
experience and associated emotional feelings occur adheres to the notion
that, although feelings can be transient, behaviors can be retroactively
ascribed to one’s emotional state, if the measurement occurs before it has
changed or is forgotten (Donovan & Rossiter, 1982).
In order to ensure that emotional state and store environment
would be relevant to the shopper’s buying experience, and for the sake
of consistency, only fashion store shoppers were sampled. These stores
are considered to carry the type of merchandise most amenable to this
study in that they carry high-involvement goods (Flicker & Speer,
1990). Prior to being used in the study, stores were screened. Only
those stores that matched the characteristics of a fashion store or bou-
tique were selected. The concept of involvement is consumer related,
“defined in terms of the consumer’s evaluation of the importance of
and identity with the product” (Assael, 1987, p. 101). One’s emotional
state or perception of store image is far less likely to be a salient factor
in a quick trip to a drug store for razor blades, for example. Respon-
dents were selected randomly, on different days of the week and at dif-
ferent times of day, immediately as they exited the store. A total of
1480 such shoppers were approached, of which 61.4% agreed to partic-
ipate, reflecting a non-participation rate of 38.6%. The sample is com-
prised of 508 responses gathered in a large mid-Atlantic metropolitan
area and 401 responses from a major northeastern metropolitan area.

STORE ENVIRONMENT AND PURCHASE BEHAVIOR 367


T tests showed no significant differences between the geographic
groups in demographic characteristics and no significant differences in
the stimulus, organismic, or response variables; therefore, the entire
sample was analyzed as a whole.

Measurement of Variables
As the stimulus variable, store environment was measured with the
use of items suggested by Dickson and Albaum (1977). In addition,
several items were included to measure in-store environment. Thirty-
one items were measured on an 8-point semantic-differential scale.
These items were fashioned to measure traditional aspects of store en-
vironment, such as spaciousness, price range, and assortment of mer-
chandise (e.g., Fisk, 1961) as well as the store’s “atmospherics,” such as
lighting, odor, and music (Kotler, 1974). Exploratory factor analysis re-
vealed seven factors; however, many items had low loadings or double
loadings. Thus, following a procedure suggested by Churchill (1979) for
scale refinement, items with low loadings or double loadings were
dropped. Further scale refinement was done by examining item-to-
total correlations and reliability, and dropping items with low item-to-
total correlation to improve the reliability. This led to the retention of
20 items, which represented four dimensions: social factor (4 items, al-
pha 5 0.70), overall image (2 items, alpha 5 0.67), design factor (10
items, alpha 5 0.79), and ambience factor (4 items, alpha 5 0.64).
The consumer’s emotional state served as the organismic, or inter-
vening variable. It was measured with the use of the Mehrabian –
Russell (1974) scale of 18 items on an 8-point semantic-differential
scale. Factor analysis of this scale showed the three distinct factors
postulated by Mehrabian and Russell (1974) to mediate approach –
avoidance behavior in environmental situation: pleasure, arousal, and
dominance (or their opposites). The scale revealed these factors quite
clearly, except the dominance factor. In this case, the dimension would
be better described as alertness. As discussed earlier, based on theoret-
ical and empirical considerations, the items representing dominance
(or alertness, as we found them) were dropped from further analysis.
When the same procedure was followed, to improve the scale reliabil-
ity, 11 items were retained, representing two dimensions: pleasure
(6 items, alpha 5 0.87) and arousal (5 items, alpha 5 0.74).
The response variables were single items about shopping behavior
in the store, suggested by Mehrabian and Russell (1974), Donovan and
Rossiter (1982), and Sherman and Smith (1986), in line with the ap-
proach – avoidance concept of Wundt (1905) in response to environ-
ment. The respondents indicated the number of items purchased,
amount of money spent in the store, amount of time spent in the store,
and whether they liked the store environment. Items used in all the
scales are given in Table 1.

368 SHERMAN ET AL.


Table 1. Items Used to Measure Various Constructs
Reliability
Stimulus variables
Social factor (4 items) 0.70
Lively – unlively
Cheerful – depressing
Boring – stimulating (R)
Courteous salespeople – discourteous salespeople
Overall image (2 items) 0.67
Good – bad
Negative – positive (R)
Design factor (10 items) 0.79
Large – small
Roomy – cramped
Colorful – drab
Unattractive – attractive (R)
Dirty – clean (R)
Comfortable – uncomfortable
Cluttered aisles – uncluttered aisles (R)
Crammed merchandise – well-spaced merchandise (R)
Impressive interior – unimpressed interior
Well-organized layout – unorganized layout
Ambience factor (4 items) 0.64
Pleasant – unpleasant
Relaxed – tense
Dull – bright (R)
Pleasant smelling – unpleasant smelling

Organism Variables 0.87


Pleasure (6 items)
Happy – unhappy
Bored – relaxed (R)
Unsatisfied – satisfied (R)
Pleased – annoyed
Contented – melancholic
Despairing – hopeful (R)
Arousal (5 items) 0.74
Frenzied – sluggish
Stimulated – relaxed
Calm – excited (R)
Dull – jittery (R)
Unaroused – aroused (R)
Note: Items identified with R were reverse coded.

ANALYSIS AND RESULTS

The data were analyzed with the use of LISREL 8 (Jöreskog & Sörbom,
1993a). A preprocessor to LISREL, PRELIS 2 (Jöreskog & Sörbom,
1993b) was used for screening and for the computation of the correla-
tion matrix used in the analysis. Although the initial sample contained

STORE ENVIRONMENT AND PURCHASE BEHAVIOR 369


909 responses, the listwise deletion method used for computation of the
correlation matrix reduced the sample size to 688. According to gener-
ally accepted guidelines, this sample size is more than adequate for
testing the model (cf. Bagozzi, 1980; Gerbing & Anderson, 1985).
Multiple items used to measure stimulus (social, image, design, and
ambience) and organism (pleasure and arousal) variables were ran-
domly split into two groups (separately for each variable) to represent
two indicators of each construct, in line with the procedure suggested
by Bagozzi and Heatherton (1994). The scores for all the items for each
indicator were summed, and that sum was used as respective indica-
tors. Further, the structural error terms (zetas) for the two organism
variables (pleasure and arousal) were allowed to correlate to each
other, but not to any other structural error term, because these two
represented the two dimensions of consumer’s emotional state. Be-
cause the four response variables cannot be assumed to represent a
single unidimensional construct, they represented four single-item
constructs. For the purpose of identification, factor loadings of these
single indicators were fixed to one and the respective error terms to
zero. However, it was expected that these response variables will be
correlated to each other. Therefore, structural error terms for these re-
sponse variables (zetas) were allowed to correlate with each other but
not to any other structural error term.
Maximum-likelihood estimates for the various parameters and indi-
cators of the overall fit of the model are given in Table 2. Although the
x2 statistics suggests that the data did not fit the model (x2 5 452.80,
p , .01), the overall evaluation of the fit was based on multiple indica-
tors (cf. Bagozzi & Yi, 1988). These multiple indicators suggest that
the data fit the model quite well, justifying further interpretation. The
goodness-of-fit index was 0.92; the Bentler and Bonett (1980) normed-
fit index and nonnormed-fit index were 0.91 and 0.89, respectively.
Furthermore, the standardized root-mean-square residual was 0.046,
RMSEA was 0.083, and CFI was 0.93.

Table 2. Parameter Estimates for the Structural Model


Parameter MLE Standard Error T Value
lx1 0.71 0.04 20.25
lx2 0.80 0.03 23.50
lx3 0.68 0.04 18.97
lx4 0.76 0.04 21.56
lx5 0.77 0.04 21.62
lx6 0.79 0.04 22.19
lx7 0.69 0.04 18.47
lx8 0.64 0.04 17.29
ly1 0.81 0.10 8.17
ly2 0.80 0.10 8.16

370 SHERMAN ET AL.


Table 2. (Continued from previous page.)
Parameter MLE Standard Error T Value
ly3 0.63 0.06 10.64
ly4 0.96 0.08 11.73
ly5 1.00*
ly6 1.00*
ly7 1.00*
ly8 1.00*
e11 0.35 0.02 14.81
e22 0.36 0.02 14.96
e33 0.60 0.04 13.58
e44 0.08 0.07 1.20
d11 0.50 0.03 15.73
d22 0.36 0.03 12.42
d33 0.54 0.03 15.95
d44 0.42 0.03 13.53
d55 0.41 0.03 12.62
d66 0.38 0.03 11.84
d77 0.53 0.04 14.63
d88 0.59 0.04 13.74
g11 1.04 0.42 2.48
g12 20.61 0.57 21.07
g13 0.32 0.16 2.05
g14 0.20 0.31 0.64
g21 20.40 0.58 20.69
g22 0.33 0.70 10.48
g23 20.92 0.25 23.72
g24 1.21 0.61 2.00
b31 0.14 0.05 2.92
b32 0.15 0.05 3.32
b41 0.87 0.11 8.16
b42 20.02 0.03 20.55
b51 0.08 0.05 1.60
b52 0.11 0.05 2.44
b61 0.02 0.05 0.32
b62 0.10 0.05 2.06
c21 0.07 0.06 1.08
c43 0.05 0.02 2.12
c53 0.29 0.04 7.55
c54 0.08 0.02 3.47
c63 0.21 0.04 5.51
c64 0.13 0.02 5.50
c65 0.30 0.04 7.72
f21 1.04 0.03 38.91
f31 0.72 0.03 21.16
f32 0.79 0.03 23.60
f41 1.02 0.03 30.73
f42 1.04 0.04 29.07
f43 0.95 0.03 28.45
Overall indicators: x2 5 452.80 (79 df, p < .01), GFI 5 0.92, AGFI 5 0.86.
Standardized RMSR 5 0.046, normed-fit index5 0.91, nonnormed-fit index 5 0.89, CFI 5 0.93,
RMSEA 5 0.083.
*Fixed parameter.

STORE ENVIRONMENT AND PURCHASE BEHAVIOR 371


An examination of the parameter estimates reveals that the four stim-
ulus variables were highly correlated, suggesting a possible lack of dis-
criminant validity of the four variables. However, if the four variables
indeed lack such discriminant validity (i.e., js were redundant) we would
expect them to have the same impact on the two organism variables (i.e.,
four gammas on h1 would be equal and the four gammas on h2 would be
equal. As seen in the table, this is not true. Therefore, one would con-
clude that even though the four exogenous variables are highly corre-
lated, they still have different impacts on the endogenous variables.
The model suggested that elements of store environment (stimulus)
will have a positive influence on the organismic variables (pleasure and
arousal). Although social factors had a positive impact on pleasure
(maximum-likelihood estimate 5 1.04, t value 5 2.48), it did not have a
significant impact on arousal (maximum-likelihood estimate 5 2 0.40,
ns). Image did not have a significant impact on either pleasure or
arousal. Design factor had a positive influence on pleasure, as expected
(maximum-likelihood estimate 5 0.32, t value 5 2.05). However, it had
a negative impact on arousal (maximum-likelihood estimate 5 20.92,
t value 2 3.72), contrary to expectation. Ambience had a positive im-
pact on arousal (maximum-likelihood estimate 5 1.21, t value 5 2.00);
however, it did not have a significant influence on pleasure (maximum-
likelihood estimate 5 0.20, ns).
The second part of the model suggested that organism variables
(pleasure and arousal) will have a positive influence on response vari-
ables. Both pleasure and arousal had a positive influence on the money
spent in store (maximum-likelihood estimates 5 0.14 and 0.15, respec-
tively, t values 5 2.92 and 3.32, respectively). Although pleasure had a
positive influence on store liking (maximum-likelihood estimate 5 0.87,
t value 5 8.16), arousal did not have a significant impact (maximum-
likelihood estimate 5 2 0.02, ns). Pleasure did not have a significant in-
fluence on the number of items purchased (maximum-likelihood
estimate 5 0.08, ns) or the time spent in store (maximum-likelihood es-
timate 5 0.02, ns). Arousal, on the other hand, had a positive impact on
both variables, that is, number of items purchased (maximum-likeli-
hood estimate 5 0.11, t value 5 2.44) and the time spent in store (maxi-
mum-likelihood estimate 5 0.10, t value 5 2.06).
As encouraging as these findings are, there are limitations to the
study that must be considered. First, as Isen et al. (1982) indicate, peo-
ple in good moods are more likely to engage in helping behavior, such
as filling out questionnaires, so that the influence of positive mood
may be somewhat exaggerated, because it may be assumed that dis-
proportionately more people in negative moods were nonrespondents.
More research is needed to explore those negative-mood respondents.
Second, consumers may deliberately choose to shop in stores that in-
duce positive moods, and the relative effects of the store’s image may
be difficult to determine without a pretest. This, of course, would intro-

372 SHERMAN ET AL.


duce its own type of bias, both statistical and nonstatistical, with a
field study (e.g., possible testing effect, difficulty of obtaining such data
from shoppers). Third, because the survey was conducted at the point
of purchase, there is no way to assess the effects of other variables,
such as the atmosphere of stores previously visited. Finally, the act of
purchasing may have contributed to the better emotional feelings of
the respondents upon exiting. In addition, feeling states may reflect
emotions brought to the environment. As Donovan and Rossiter indi-
cate, more research is needed in this area.

DISCUSSION AND IMPLICATIONS

In this study, we identified and explored how store environment and


emotional states may influence various dimensions of purchase behav-
ior. This research confirms that although cognitive factors may largely
account for store selection and for most planned purchases within the
store, the environment in the store and the emotional state of con-
sumers may be important determinants of purchase behavior. Al-
though it differs on some factors with the research of Donovan et al.
(1994), the present study reaffirms that retailers should pay attention
to consumer’s in-store emotional state (pleasure and arousal), because
the emotions of consumers are important factors in buyer behavior.
In the present study, social factors and the design of the store had a
positive impact on pleasure, and ambience positively affected arousal.
This research also found that pleasure had a positive influence on
money spent and liking the store, and arousal had a positive impact on
money spent in the store, time spent in the store, and the number of
items purchased in the store. These findings differ from those of
Donovan et al. (1994). Although their research indicates that pleasure
is a significant predictor of both extra time and unplanned spending,
our study suggests that pleasure is more closely associated with store
liking and money spent in the store.
Together with the traditional utilitarian explanation of shopping be-
havior, these more recently identified reasons for behavior in the store
present a new challenge to the marketer. Efforts have shifted from a
sole concern with merchandise breadth, depth, and quality to include
an emphasis on creating a pleasant, entertaining experience for the
consumer who is interested in more than just the product. Unless a
store has a distinct product offering or pricing strategy, retailers must
distinguish their store by building on the relationship between store
atmosphere and consumers’ emotional states. Arousing the consumer
by bright colors, or upbeat music, as suggested by Holbrook and
Garner (1993), may have a positive effect on profitability. As Donovan
et al. (1994) indicate, care needs to be taken that in unpleasant envi-
ronments, arousal might need to be dampened. Furthermore, as they

STORE ENVIRONMENT AND PURCHASE BEHAVIOR 373


suggest, more research should be undertaken to measure the pleasure
and arousal dimensions prior to entering the store, because some of
the motives for spending may have been to relieve negative emotions.
First, the utilitarian shopper who goes to the store to buy a specific
product may purchase additional items if his or her attention is cap-
tured by creative displays or demonstrations. Second, the shopper who
goes to the store because she or he likes the environment may unex-
pectedly spend more money as a result of a positive-mood – inducing
atmosphere in the store. Even if a consumer is in a negative emotional
state upon entering, she or he may become more emotionally uplifted
while in a store where the atmosphere creates positive feelings, and
may spend more than intended. In fact, more research is needed on
this aspect.
It should be noted, of course, that just as a positive in-store experi-
ence may induce a positive emotional state and subsequent positive af-
firmative behavior, a negative in-store experience could nullify the
influence of preexisting good feelings, enhance a negative state, or
even create a mood state. Thus, the retailer must not only ensure that
positive good feelings remain positive or become even more so, but
must also take steps to help change negative feelings to positive ones.
For some customers, this may amount to simply not being offended by
discourteous salespeople, unpleasant odors or noise, or temperatures
in store.
This study indicates that a consumer’s emotional state may affect
his or her shopping behavior after the decision to shop has been made.
At the point of purchase there are many ways to make a customer feel
better: suitable layout, cleanliness, colors, and salesperson training.
These findings seem important for retailers to note, especially if all
that is necessary is to properly train salespeople and to make even
small capital investments such as adjusting the light level or store lay-
out in ways that would appeal to the store’s clientele. Such atmos-
pheric changes or additions would be well worth the effort if they
positively influence consumers’ emotions and stimulate positive pur-
chasing behavior, such as buying and spending more. Having a pleas-
ant experience can also encourage a shopper to build a more lasting
relationship with the store and seek to return.

Future Research
This study has attempted to present a broad-based, theoretically
guided field study of the effects of store image on consumers’ emo-
tional states and resulting influences on shopping behavior. The re-
sults are promising and indicate the possibility of fruitful future
research. It would be helpful, for example, to ascertain the effects of
store environment on preexisting positive or negative consumer emo-
tional states. Further research regarding consumers who are in nega-

374 SHERMAN ET AL.


tive states after shopping may be helpful. It would also be useful to at-
tempt nonintrusive in-store measurement. Investigation into different
types of shoppers and their motivations while entering the store could
be made in order to study how various shopper segments react to the
store image and atmosphere. Finally, a variety of products and stores
could be used to broaden these findings. An interesting extension of
this research could follow suggestions made by Keaveney and Hunt
(1992). They suggest that different categories of stores have different
cues, and if retailers want to change their image, they need to be care-
ful about these cues. Category processing can be extended to explore
how it affects consumer’s emotions in the store and how consumer
evaluations of store images and atmosphere can affect emotional
states. This research found that the four dimensions of store environ-
ment are correlated. Perhaps a second-order latent variable could rep-
resent overall store atmosphere. Future researchers might model
these exogenous variables as first-order indicators of a second-order
latent variable and examine its impact on consumer mood and behav-
ior.
There has been renewed interest in current competitive environ-
ments and in exploring how managers can effectively use store images
and consumer mood to stimulate purchase. It is hoped that future re-
search will continue to explore new directions of how components of
the shopping experiences can be improved to create more positive rela-
tionships between sellers and store patrons. What determines a pleas-
ant environment for consumers and how this affects buyer behavior
are exciting regions for continued exploration. Different types of in-
store experiences can also be examined, along with the notions of con-
sumer expectations and how they affect their emotional states. For
example, if a consumer goes to a discount retailer, would he or she be
disconcerted if the store was more lavish than his or her expectations?
Would this modify his or her emotional state and possible purchase be-
havior? These are just a few of the exciting avenues for future re-
search.

REFERENCES

Anderson, P. (1986). Personality, perceptions and emotion-state factors in ap-


proach-avoidance behaviors in the store environment. In Terence A. Shimp
et al. (Eds.), 1986 AMA educators’ proceedings (pp. 35 – 39). Chicago: Ameri-
can Marketing Association.
Assael, H. (1987). Consumer behavior and marketing action (3rd ed.). New
York: Kent.
Babin, B. J., Darden, W. R., & Griffin, M. (1992). Some comments on the role of
emotions in consumer behavior. In R. P. Leone and V. Kumar (Eds.), 1992
AMA educators’ proceedings: Enhancing knowledge development in market-
ing (pp. 130 – 135). Chicago: American Marketing Association.

STORE ENVIRONMENT AND PURCHASE BEHAVIOR 375


Bagozzi, R. P. (1980). Causal models in marketing. New York: John Wiley &
Sons, Inc.
Bagozzi, R. P. (1986). Principles of marketing management. Chicago: Science
Research Associates, Inc.
Bagozzi, R. P., & Heatherton, T. F. (1994). A general approach to representing
multifaceted personality constructs: Application to state self-esteem. Struc-
tural Equation Modelling, 1, 35 – 67.
Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation mod-
els. Journal of the Academy of Marketing Science, 16, 74 – 94.
Baker, J. (1986). The role of environment in marketing services: The consumer
perspective. In J. A. Cepeil et al. (Eds.) The services challenge: Integrating
for competitive advantage (pp. 79 – 84). Chicago, IL: American Marketing
Association.
Baker, J., Grewal, D., & Parasuraman, A. (1994). The influence of store envi-
ronment on quality inferences and store image. Journal of the Academy of
Marketing Science, 22, 328 – 339.
Belk, R. (1975). Situational variables and consumer behavior. Journal of Con-
sumer Research, 2, 157 – 164.
Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of
fit in the analysis of covariance structures. Psychological Bulletin, 88,
588 – 606.
Bettman, J. R. (1979). An information processing theory of consumer choice.
Reading, MA: Addison-Wesley.
Bower, G. H. (1981). Mood and memory. American Psychologist, 31, 129 – 148.
Clark, M. (1982). A role for arousal in the link between feeling states, judge-
ments and behavior. In M. Clark and S. Fiske (Eds.) Affect and cognition
(pp. 263 – 269). Hillsdale, NJ: Lawrence Erlbaum Associations.
Clark, M., & Isen, A. M. (1982). Toward understanding the relationship be-
tween feeling states and social behavior. In A. Hastorf and A. M. Isen (Eds.),
Cognitive social psychology. New York: Elsevier.
Churchill, G. A. (1979). A paradigm for developing better measures of market-
ing constructs. Journal of Marketing Research, 16, 64 – 73.
Dichter, E. (1964). Handbook of consumer motivations: The psychology of the
world of objects. New York: McGraw-Hill.
Dickson, J., & Albaum, G. (1977). A method of developing tailor made seman-
tic differentials for specific marketing content areas. Journal of Marketing
Research, 14, 87 – 91.
Donovan, R., & Rossiter, J. (1982). Store atmosphere: An environmental psy-
chology approach. Journal of Retailing, 58, 34 – 57.
Donovan, R., Rossiter, J., Marcoolyn, G., & Nesdale, A. (1994). Store atmos-
phere and purchasing behavior. Journal of Retailing, 70, 283 – 294.
Fisk, G. (1961). A conceptual model for studying consumer image. Journal of
Retailing, 37, 1 – 8, 54.
Flicker, M. H., & Speer, W. C. (1990). Emotional responses to store layout and
design: An experimental approach. In A. Parasuraman et al. (Eds.) 1990
AMA educators’ proceedings: Enhancing knowledge development in market-
ing (pp. 1 – 5). Chicago, IL: American Marketing Association.
Gardner, M. P. (1985). Mood states and consumer behavior: A critical review.
Journal of Consumer Research, 12, 281 – 300.

376 SHERMAN ET AL.


Gerbing, D. W., & Anderson, J. C. (1985). The effects of sampling error and
model characteristics on parameter estimation for maximum likelihood
confirmatory analysis. Multivariate Behavioral Research, 20, 255 – 271.
Hill, R. P., & Ward, J. (1989). Mood manipulation in marketing research: An
examination of potential confounding effects. Journal of Marketing Re-
search, 26, 97 – 104.
Holbrook, M., & Gardner, M. (1993). An approach to investigating the emo-
tional determinants of consumption durations: Why do people consume
what they consume for as long as they consume it? Journal of Consumer
Psychology, 2, 123 – 142.
Holbrook, M., & Hirschman, E. (1982). The experimental aspects of consump-
tion: Consumer fantasies, feelings, and fun. Journal of Consumer Research,
9, 132 – 140.
Holbrook, M., & Westbrook, R. (1989). The role of emotion in advertising re-
visited: Testing a typology of emotional responses. In P. Cafferata and
A. Tybout (Eds.), Cognitive and affective responses to advertising. Lexing-
ton, MA: Lexington Books.
Isen, A. M. (1984). Toward understanding the role of affect in cognition. In R.
Wyer and T. Srull (Eds.), Handbook of social cognition (pp. 179 – 236). Hills-
dale, NJ: Lawrence Erlbaum Associates.
Isen, A. M., Means, B., Patrick, R., & Nowicki, G. (1982). Some factors influ-
encing decision-making strategy and risk taking. In M. Clark and S. Fiske
(Eds.), Affect and cognition (pp. 243 – 261). Hillsdale, NJ: Lawrence Erl-
baum Associates.
Jöreskog, K. G., & Sörbom, D. (1993a). LISREL 8 user’s reference guide.
Chicago, IL: Scientific Software International, Inc.
Jöreskog, K. G., & Sörbom, D. (1993b). PRELIS 2 user’s reference guide.
Chicago, IL: Scientific Software International, Inc.
Keaveney, S. M., & Hunt, K. A. (1992). Conceptualization and operationaliza-
tion of retail store image: A case of rival middle-level theories. Journal of
the Academy of Marketing Science, 20, 165 – 175.
Kelly, G. A. (1955). The psychology of personal constructs. New York: Norton.
Kotler, P. (1974). Atmosphere as a marketing tool. Journal of Retailing, 49,
48 – 64.
Lewin, K. (1936). Principles of topological psychology. New York: McGraw-
Hill.
Mandler, G. (1979). Emotion. In E. Hearst (Ed.), The first century of experi-
mental psychology (pp. 275 – 321). Hillsdale, NJ: Lawrence Erlbaum Associ-
ates.
Mehrabian, A., & Russell, J. A. (1974). An approach to environmental psychol-
ogy. Cambridge, MA: MIT Press.
Nevin, J. R., & Houston, M. (1980). Images as a component of attractiveness
to intra-urban shopping areas. Journal of Retailing, 56, 77 – 93.
Oxenfeldt, A. (1974). Developing a favorable price-quality image. Journal of
Retailing, 50, 8 – 14.
Robertson, T. S., Zielinski, J., & Ward, S. (1984). Consumer behavior. Glenview,
IL: Scott, Foresman, and Company.
Rotter, J. B. (1954). Social learning and clinical psychology. Englewood Cliffs,
NJ: Prentice-Hall.

STORE ENVIRONMENT AND PURCHASE BEHAVIOR 377


Russell, J. A. (1978). Evidence of convergent validity on the dimensions of af-
fect. Journal of Personality and Social Change, 36, 1152 – 1168.
Russell, J. A., & Pratt, G. (1980). A description of the affective quality attrib-
uted to environments. Journal of Personality and Social Psychology, 38,
311 – 322.
Sherman, E., & Smith, R. B. (1986). Mood states of shoppers and store image:
Promising interactions and possible behavioral effects. In P. Anderson (Ed.),
Advances on consumer research (Vol. 14). Provo, UT: Association for Con-
sumer Research.
Simon, H. A. (1982). Comments. In M. Clark and S. Fiske (Ed.), Affect and
cognition (pp. 333 – 342). Hillsdale, NJ: Lawrence Erlbaum Associates.
Swinyard, W. R. (1993). The effects of mood, involvement and quality of store
experience on shopping intentions. Journal of Advertising Research, 20,
271 – 280.
Wundt, W. (1905). Grundriss der psychologic. Leipzig: Wilhelm Engelmann.
Correspondence regarding this article should be sent to: Anil Mathur, Depart-
ment of Marketing and International Business, Hofstra University, 222
Weller Hall, Hempstead, NY 11550-1090.

378 SHERMAN ET AL.

You might also like