ASAC 2006

Document Sample
ASAC 2006 Powered By Docstoc
					ASAC 2006                                                                 Frank Pons
Banff, Alberta                                                            University of San Diego

                                                                          Mehdi Mourali
                                                                          University of New Hamoshire


        This paper studies the relationship between consumers’ density and satisfaction in service
        settings. In particular, it focuses on the role of affective reactions as a mediator in this
        relationship and on the role of density expectations and shopping situations as moderator
        of this relationship. The results support the mediating effect and show that expectations
        and shopping situation interact to affect consumers’affective reactions, turning an
        unpleasant shopping situation in a pleasant one. These findings offer managerial insights
        to deal with various consumers’ density in shopping situations.


     Crowding is an important environmental factor in consumers’ evaluation of the retail experience (e.g.
Eroglu and Harrell, 1986; Eroglu and Machleit, 1990; Harrell et al., 1980; Machleit et al, 2000).
However, it is a paradoxical phenomemon. On one hand, retailers who successfully attract crowd of
shoppers in their stores are often penalized by their very success since, as most studies show, high
crowding is associated with low shopping satisfaction. On the other hand, hedonic services or events
(amusement parks, concerts…) often acquire value only if they are crowded (Holt, 1995; Brown et al.,
2000) and this pleasant feature of crowded settings has been totally ignored so far in the marketing
literature on crowding.
     Therefore, the purpose of the present study is to examine the relationship between consumer density
and consumer’s satisfaction in shopping situations and to study the role of affective reactions in this
relationship as well as the moderating effects of two variables, namely density expectations and shopping
situation (Eroglu et al., 2004; Machleit et al, 2000; Holt, 1995; Hui and Bateson, 1991). The proposed
framework aims at clarifying the ambiguous influence (positive/negative) that crowds can have in a retail
setting. Using an experimental design and manipulating consumer expectations in particular, this paper
should offer a potential solution to solve the paradox that retailers face when dealing with successful and
crowded stores.

                                      Literature Review and Hypotheses

    1. Influence of other customers in the retail setting
         In their seven Ps’ framework, Booms and Bitner (1981) first identified co-consumers as important
actors in the service encounter. Several researchers (Grove and Fisk, 1997; Iacobucci, 1998; Stauss and
Mang, 1999) also suggest that customer-customer interactions in the service experience are often drivers
of the overall customer satisfaction/dissatisfaction (Jones, 1995, Martin, 1996).
        Only few articles, mainly dealing with experiential products such as river rafting, baseball or
sports spectatorship, mention that other customers contribute to create an enjoyable experience (Eastman
and Land, 1997; Holt, 1995; Price et al., 1995).
        In fact, the majority of the studies describe the interaction between customer as a noise or a
disturbance occuring during the service delivery. These studies typically tackle waiting line issues
(Schmidt et al., 1992; Hui et al., 1997, 1998) or critical incidents in services delivery (Edvardsson, 1992;

Grove and Fisk, 1997). In addition, the negative impact of other customers is studied at the individual
level, where rude or unexpected behaviors from others spoil the nature of the service, but rarely at a more
aggregate level (crowd) (Grove and Fisk, 1997).

    2. The critical distinction between density and crowding
        The crowd’s impact in a retail environment has been conceptualized in the retail crowding model
(Harrell et al.; 1980; Eroglu and Harrell, 1986). In this model, the researchers highlight the difference
between density and crowding and the role of density as a driver of crowding:

                   “Density alone does not produce adaptation behaviors. Only when it produces
                   perceived crowding do shoppers act. Perhaps, then, environmental designs
                   can be created which provide for increased density but lessen the feeling of
                   being crowded.” (Harrell et al., 1980: p.48).

                                             Insert Figure 1 about here.

         In environmental psychology, Rapoport (1975) also emphasizes the importance of density but
insists on the existence of two different kinds of “density”. In his research, perceived density is the driver
of a second concept called “affective density”:

         “The difference is the following. Density is the perception and estimate of the number of people
present in a given area, the space available, and its organization, whereas affective density is the
affective evaluation or judgment of that perceived density” (Rapoport, 1976).

         In both of the previous conceptualizations, a two step approach is used to describe how
consumers react to crowded settings (Rapoport, 1976; Harrell et al. ; 1980; Eroglu and Harrell, 1986).
First, consumers roughly assess how dense the shopping environment (perceived density) is and then
affectively react to this density level. If they enjoy the density level, the density is described as functional
whereas if the dislike the density level, it is described as dysfunctional and called crowding. Indeed, in
line with processes also described in sociology and environmental psychology literature, crowding is
clearly defined as a negative affective evaluation.

               •      “Crowding is the negative subjective evaluation of excessively high
                      densities… Crowding is experienced when the environment is judged as
                      being dysfunctionnaly dense” (Eroglu and Harrell, 1986).
               •       “Crowding is a negative experiential state associated with spatial aspects
                      of the environment” (Rustemli, 1993).

         Using overload theory and behavioral constraint theory, most of the studies in marketing focus on
crowding rather than density (Eroglu and Machleit, 1990; Eroglu et al., 1994; Eroglu et al., 2004;
Machleit et al, 2000). They explain how consumers use adaptation strategies such as shorter shopping
trips, fewer interactions with others (employees or consumers) or faster choices (familiar brands) to cope
with the stress triggered by the high density encountered in the store. Surprisingly, the potential positive
influence of density on the retail experience is not studied and the empirical studies only deal with
perceived crowding (a negative affective state). Therefore, the scope of research on crowds in services
remains limited to the negative impact of density on consumers’ satisfaction (Eroglu and Machleit, 1990;
Machleit et al., 2000) with some marginal findings suggesting potential positive influences (Eroglu et al.,
2004). The concept of density is overlooked (only the negative influence) in spite of the theoretical
support for its major role in the relationship with retail shopping satisfaction. Building on the model of

retail crowding and the clear conceptual distinction between density and the subsequent affective
reactions (positive or negative), we propose to empirically demonstrate that these affective reactions such
as crowding only represent a mediator in the density-satisfaction relationship. Therefore, we posit that:

        Hypothesis 1: Affective reactions to the density encountered in a retail setting mediate the
density-satisfcation relationship.

         Consequently, this paper breaks away from the traditional stream of research on crowding by
repositioning density as the initial driver of consumers’ shopping experience and crowding as a mediator
of the density-satisfaction relationship. It also examines conditions (moderators) under which consumers’
affective reactions to similar density levels may switch from negative to positive. Some of these
conditions and related hypotheses are discussed in the next section.

    3. Moderating effects on the density-affective evaluation relationship.

         Most of the previous studies on crowding also underline the central role played by
antecedents/conditions that may affect the intensity of the negative affective reactions to density
(crowding) in the retailing environment (Eroglu and Harrell, 1986; Eroglu and Machleit, 1990; Hui and
Bateson, 1991; Eroglu et al., 2004; Machleit et al., 2000) but, as they focus solely on the moderating
effects of these variables on crowding and not density, these effects mainly stress out a potential reduction
of the dissatisfaction experienced by the consumers. However, these studies provide fruitful insights on
these moderators’ role and their influence. Two of these moderators are detailed hereafter along with
formal hypotheses of their influence on the density-satisfaction relationship.

        3.1. Expectations
         Consumers often form expectations about what they will find in a store or a service setting. These
expectations about products, atmospherics, people, etc also deal with crowd and density levels (Rapoport,
1976; Sinha and Nayyar, 2000; Machleit et al., 2000).
         Expectations are described as anticipations of future consequences (Taylor, 1994) or comparison
standards (Oliver, 1989). They form the cornerstone of the satisfaction literature based on the expectancy-
disconfirmation model (Tse and Wilton, 1988; Fournier and Mick, 1999; Oliver, 1997, 1993, 1981;
Miller, 1977; Zeithaml et al, 1993). In this model, consumers use comparative standards to assess their
satisfaction level. The value of the expectancy-disconfirmation approach is supported by numerous
studies (Erevelles and Leavitt, 1992; Oliver, 1981; Tse et al., 1990; Yi, 1990).
         Rapoport (1976) stresses the central role played by expectations in the evaluation of dense
conditions. For instance, in his research, perceived density is always assessed through standards of
comparison with previous experience or desired density levels. In addition, the central role of
expectations has been well documented in several other sociology and psychology studies (Ford, 2001;
Klein and Harris, 1979; Martin, 1996; Sinha and Nayyar, 2000; Webb and Worchel, 1993). These studies
demonstrate that expectations about crowded or highly dense environments clearly influence outcomes
for the individual of subsequent high-density situations.
         Only recently Machleit et al. (2000) empirically tested the role of density expectations in a retail
setting. They posit that when perceived crowding are met or fall short of expectations, consumers should
be less dissatisfied than when perceived crowding exceed expectations. Unfortunately, they offer
inconclusive results regarding this hypothesis, as they did not manipulate this variable an experiment and
could not obtain an optimum measure of crowding expectations.

        3.2. Shopping situation
         The shopping situation in which the dense situation occurs appears to be also an essential
moderator of the density-affective evaluation relationship. An extensive amount of marketing research
describes how a retail experience unfolds and how others (consumers and employees), during the
encounter, may enhance/hinder the experience of the individual shopper (Eastman and Land, 1997;
Eroglu and Machleit, 1990; Holt, 1995; Machleit et al., 2000; Price and Arnould, 1993; Sherry, 1998;
Wann et al., 2000).
         In the majority of these studies, a high density of shoppers consistently triggers a negative
influence on affect (crowding) and satisfaction (Hui and Bateson, 1991; Eroglu and Machleit; 1990).
These results support qualitative studies suggesting that highly dense situations in a service setting often
lead to dissatisfaction and stressful situations (Grove and Fisk, 1997; Holt, 1995).
         However, some research suggests that the nature of the shopping trip or even the setting itself
may reduce the negative influence of density on affect and satisfaction. For instance, Hui and Bateson
(1991) even suggest in their conclusion that “it seems that density produces positive emotional and
behavioral effects in some settings and negative effects in other settings (Freedman, 1975)”.
    The studies on crowding deal with utilitarian shopping situations such as shopping malls, stores, or
banks. Only a handful of research describes leisure situations such as a concert, a ball game or a show.
(Holt, 1995). This contextual variable (leisure shopping situation versus utilitarian shopping situation;
Babin et al., 1994) has yet to be tested in a research that would allow both positive and negative
consequences to dense situations. For instance, in some leisure situations, individuals expect the crowd to
be there as much as they expect a great performance (sports team, rock artist).

        3.3. Hypotheses
         The literature review clearly highlights the contribution made by previous studies on retail
crowding (Eroglu and Harrell, 1986; Eroglu and Machleit, 1990; Holt, 1995; Machleit et al., 2000).
However, it also suggests several gaps that this paper investigates.
         Unlike most of the previous studies, this research suggests that high density itself is not a
negative feature of a shopping situation. In fact, specific conditions will interact to create either a positive
or a negative experience for the consumer. Two of these conditions relate to the nature of the shopping
situation (utilitarian or leisure) and the level of density expected by the consumer. It is not hypothesized
that they individually have an effect on satisfaction but rather that there is an interaction effect. For
instance, the density level (expected and encountered) only takes a particular meaning (benefit or
disbenefit) when interacting with the shopping situation. Using the expectancy-disconfirmation approach,
in which a positive disconfirmation means that you perceive more people than you expected whereas a
negative disconfirmation means that you perceive less people than previously expected and a
confirmation means that you perceive as many people as expected, (Fournier and Mick, 1999), the
following hypotheses are presented:

        Hypothesis 2a: In a leisure situation, a positive disconfirmation of density trigger a more positive
        affective reaction than a confirmation of density, which in turn trigger a more positive affective
        reaction than a negative disconfirmation of density.
        Hypothesis 2b: In a utilitarian situation, negative disconfirmation of density trigger a more
        positive affective reaction than a confirmation of density, which in turn trigger a more positive
        affective reaction than a positive disconfirmation of density.
        Hypothesis 3a: When there is a positive disconfirmation of density, a leisure situation triggers a
        more positive affective reaction than a utilitarian one.
         Hypothesis 3b: When there is a negative disconfirmation of density, a leisure situation triggers a
        more negative affective reaction than a utilitarian one.

        Using the previous hypotheses, our research aims at showing the role played by density
expectations in a retail environment. This is the first attempt made at manipulating consumers’
expectations in a research on crowd issues. The methodology used is described in the next section.


    1. Research Design and Sample
         In this study, the nature of the shopping situation (leisure versus utilitarian) and the
confirmation/disconfirmation of the human density level are manipulated independently and their effects
are studied. In short, a 2 (leisure shopping situation and utilitarian shopping situation) X 2 (high density
versus low density) X 2 (high density expected versus low density expected) factorial design is used in
this experiment to test the hypotheses. Written scenarios and video stimuli are used to operationalize the
manipulated variables. The scenarios were written by the researcher, reviewed by experts and pretested.
The video stimuli were shot and edited by a professional. The two situations chosen for the final
questionnaire, as either leisure or utilitarian, were identified through pretests among 153 students who
were asked to rate 6 situations (bar, restaurant, hockey game, mall, bank and bookstore) using the hedonic
side of the hedonic and utilitarian shopping values scale (Babin et al, 1994). The two situations with the
lowest (bookstore) and the highest (bar) average scores are significantly different on a 7-point Likert scale
(Mbar = 6.2 vs Mbookstore = 3.1; F(1,151) = 95.52, p<0.001). Therefore, a bar and a bookstore situation are
adopted in this study.
         The confirmation/disconfirmation variable is manipulated in two steps. First, expectations of
density are manipulated through written scenarios, in which clear statements about what density to expect
in the retail setting are made. Then, after consumers start filling the questionnaire, a short video is
presented. The density in the store is manipulated. In one condition, there are a lot of consumers whereas
in the other one there are only a few. Everything else is kept identical. All the pretests indicated adequate
manipulations. Significant differences are found for expectations on a 7-point Likert scale (Mhighexp = 6.1
vs Mlowexp = 3.1; F(1,62) = 164.3, p<0.001) and for the estimation of the number of persons (1question) in
the video (Mmany = 118.9 vs Mfew = 6.9; F(1,62) = 53.8, p<0.001). Eight groups were necessary to implement
the experimental design. Therefore, a sample of about 480 subjects (i.e around 60 per cell) was targeted
for the final data collection. 25 introductory business courses with at least 30 students registered were
randomly selected. The final sample is composed of 572 respondents (groups ranging from 44 to 33
respondents), of which 54.6% are female and 45.4% are male. 97.4 % of our sample is under the age of

   2. Procedure
        Subjects receive a questionnaire and are told to wait for directions from the researcher. When
turning the first page, they answer a first set of questions pertaining to general personality traits. Then on
page 3, a scenario describing a service situation is presented. The researcher read it aloud while students
also read it. They are then asked to picture themselves in the situation. This point is emphasized. Then,
they turn the page and answer the first part of the questionnaire where they are asked about their density
expectations for the situation previously described in the scenario. On the next page, the scenario is
repeated and read again by the interviewer. In addition, they are told that they are about to enter the
service setting previously described. Then, a short video (1 minute), supposed to depict the situation is
shown. They are told to watch carefully the video. After viewing the video, they answer the last part of
the questionnaire pertaining to perceived density, affective reactions and satisfaction.

        3.      Measures
        A self-administered questionnaire is used to gather the data. All the items are measured on a 7-
point Likert scale.

  Density expectations and perceptions are both measured by using similar questions but with a different
wording underlining the pre or post encounter situation. In the case of expectations, items such as “I
expect this place to be crowded”, “I expect a lot of people to be in this place” or “I expect this place to
be virtually empty (reversed)” are used whereas in the case of perceptions items are “There are a lot
people in this place” or “This place is crowded”. Most of these items are borrowed and adapted from
existing scales and studies on crowding (Machleit et al., 1994; Webb and Worchel, 1993; Hui and
Bateson, 1991) but some were developed and added to reflect environmental psychology findings on
density (Rapoport, 1976, Altman, 1975).
  The affective evaluation is assessed by asking respondents how they feel about the situation they just
encountered in the video. Items such as “I feel happy” or “I feel good” are borrowed from scale
measuring affective evaluations (Derbaix, 1995; Murry and Dacin, 1996).
  Satisfaction is measured for the overall service experience as suggested by Oliver (1997). Items are
borrowed from his consumption satisfaction scale and Machleit et al’s scale (2000). They are adapted to
the situation at stake.
  As these measures are borrowed from different studies and are developed in different contexts, a series
of analyses were performed on each of the latent variables used in the model to determine their
psychometric properties and particularly assess their reliability and validity. Results from exploratory
factor analyses using the principal component extraction method suggested adequate dimensionalities and
satisfying reliability indicators for each factor present in the model (all Cronbach alphas above 0.86).
Moreover, confirmatory factor analyses performed on these latent factors indicated a reasonably good fit
to the data and allowed us to accept the measurement model.

                                           Results and Discussion

        1. Manipulation Checks
        The confirmation/disconfirmation variable manipulation was checked. A paired sample T-tests
was used in order to ensure that the 4 confirmation/disconfirmation groups reflected adequately the
confirmation/disconfirmation variable manipulation. These results supported adequate manipulations and
the conformity with the experimental design of the study.
        The two situations chosen for the final questionnaire used a video shot in bar and in a bookstore.
Respondents did not report any difference in the perceived similarity between them and the people in the
video (F(15,558) = .957, p=.413) for each condition. These results indicate that the respondents felt similar
enough to the subjects used in the respective service situations.

    2. Analyses of Variance

     2.1 Main Effects
          The emphasis in this study is on the fact that density alone does not necessarily lead to crowding
(negative affective evaluation) and dissatisfaction. It is suggested here that the triggering mix is more
complicated and interactions between several variables may be necessary.
When only looking at the impact of the density confirmation/disconfirmation variable, across the service
situations, the mean comparisons of affective evaluation do not reveal any main effect of the density
confirmation/disconfirmation. Indeed, there is no significant mean difference in terms of affective
evaluation (F(3,570) = 1.52, p=0.206) between positive disconfirmation (Mpositive disconfirmation = 4.03), high
density confirmation (Mhigh confirmation = 3.73), low density confirmation (Mlow confirmation = 3.84) and negative
disconfirmation (Mnegative disconfirmation = 3.99). This result suggests that perceiving more people than
expected in a service situation do not forcely lead to more dissatisfaction. Machleit et al. (2000), in the
first introduction of expectations in an empirical study, suggested that shopper affective reactions and
satisfaction would be higher in negative disconfirmation and lower in positive disconfirmation cases.
They found mixed results in their three studies. In our framework, we find inconclusive results for the

lone effect of density expectations and perceptions on satisfaction. This result suggests that density alone
does not impact affective reactions.
        Also, as expected, there is no main effect of the shopping situation (leisure versus utilitarian) on
the affective evaluation of the density. For instance, one should not be more satisfied with being in a bar
or in a bookstore. It should all depend on the shopping motives. The situation has no effect on the
affective evaluation of the density encountered (Mbar = 5.08 vs Mbookstore = 4.96; F(1,572) = 1.825, p=0.177)

         2.2 Interaction Effects
         Interaction effects of the manipulated variables on affective evaluations are presented hereafter.
Results are then discussed. The ANOVA results show a significant two-way interaction. The individual
effects are pinpointed in table 1.

                                              Insert Table 1 about here.

Results in figure 2 suggest that the confirmation/disconfirmation of a density level in a shopping situation
interact with the nature of the shopping situation itself to alter consumers’ affective evaluation of the
situation. These results are discussed hereafter.

                                             Insert Figure 2 about here.

First, significant differences in affective reactions exist between the leisure setting (bar) and the utilitarian
setting (bookstore) for each level of confirmation/disconfirmation. The mean comparisons between the
two settings for a negatively disconfirmed situation (Mbar = 2.8 vs Mbookstore= 5.1; F(1,145) = 180.6,
p<0.001), for the confirmation of a high density (Mbar = 4.3 vs Mbookstore= 3.1; F(1,135) = 71.6, p<0.001), for
the confirmation of a low density situation (Mbar = 3.2 vs Mbookstore= 4.4; F(1,141) = 41.8, p<0.001) and for a
positively disconfirmed situation (Mbar = 5 vs Mbookstore= 2.8; F(1,145) = 270.1, p<0.001) are all significant.
These results mainly show that in a positive disconfirmation or in a confirmation of a high density,
affective reactions to the setting are significantly more positive in the leisure situation than in the
utilitarian one. On the contrary, in the case of a negative disconfirmation or a confirmation of a low
density, the relationship is reversed and affective evaluations are significantly more positive in the
utilitarian situation than in the leisure one. This supports hypotheses 3a-3b. Dense shopping settings
clearly lead to different outcomes for the consumer depending on the shopping situation. The utilitarian
setting results in our research support previous studies where high density is shown to have a negative
impact on satisfaction (Eroglu and Machleit, 1990); however, this impact is reversed in the case of leisure
service situations with high density triggering strong positive affective reactions from the consumer. This
is the first time an empirical study in a controlled environment demonstrates the potential positive
influence of crowds in a shopping situation.
Second, significant differences in the affective reactions also exist between the four
confirmation/disconfirmation levels for each setting The mean comparisons between the
confirmation/disconfirmation levels for the bar (Mposdisc = 5 vs Mhighconf= 4.3 vs Mlowconf = 3.2 vs Mnegdisc=
2.8) (F(1,287) = 60.5, p<0.001) and the bookstore (Mposdisc = 2.8 vs Mhighconf= 3.1 vs Mlowconf = 4.4 vs
Mnegdisc= 5.1) (F(1,279) = 157.6, p<0.001) are all significant. These results show that in a leisure situation
(Bar), the affective reaction is significantly more positive when positive density disconfirmation occurs
than when confirmation (high or low density) does, which in turn is significantly higher than in a negative
density disconfirmation situation. In the case of a utilitarian (bookstore) setting, these relationships are
reversed. These findings support hypotheses 2a-2b. This last result on the role of expectations in crowded
situations is critical to studies on crowding and represents the key finding in this research. In fact, in any
individual situation (bar and bookstore), respondents in the positive disconfirmation situation or the
confirmation of high density are always presented with the same video stimulus (a huge crowd in the
retail setting). However, they present significant differences in the way they affectively react to the
situation. In the bar situation, they are not as happy about the high density level when they expect it

(confirmation) (MPos Disc = 5 vs MHigh Conf = 4.3; F(1,149) = 18.4, p<0.001) and in the bookstore situation,
they are not as unhappy of this dense setting when they expect it (MPos Disc = 2.8 vs MHigh Conf = 3.1; F(1,153)
= 21.5, p<0.001). This finding emphasizes that xpectations always make a difference in consumers’
affective reactions to dense situations. This supports the appeal made for considering relative perceptions
of density in a given situation. It is critical to consider these perceptions relatively to expectations formed
prior to the encounter as they can increase positive feelings and/or decrease negative feelings due to the
crowd in retail settings. This finding strengthens the important role played by expectations in crowd
assessment and may offer retailer options and crowd management tools.

    3. The density-satisfaction relationship: the mediating effect of the affective reaction to density
         In this section, the relationship that ties density and satisfaction is analyzed. The model
hypothesized in figure 3 is first tested on the overall sample. The goal of this first evaluation is to check if
an underlying constant pattern of relationship exists between density and satisfaction. In a second round
of analyses, the moderating effect of the situational shopping situation variable (leisure versus utilitarian)
interacting with different levels of density is analyzed through multigroup studies.

                                             Insert Figure 3 about here.

         As the construct validity and reliability of the measurement model was assessed in the previous
analysis (cf. measure section), the hypothesized model is directly estimated (figure 2). The fit indicators
further suggest an adequate performance of the specified model. The indicators support the fact that the
model fits the data well (Chi-square = 261.7 with 75 degrees of freedom (χ2/ df = 3.32). The CFI was
0.938 and the RMSEA was 0.065).
         The standardized estimates of the parameters and their respective t-values are presented in table
2. As shown in the table, all of the structural relationships but one were significant at p < 0.001 (tvalue
>1.96; Anderson and Gerbing, 1988).

                                             Insert Table 2 about here.

         These results support the existence of a complex model linking density in a service setting and
satisfaction. In this model, as suggested by Machleit et al (2000), density has a major indirect impact on
satisfaction through affective reaction. As hypothesized, perceived density only influences satisfaction
through affective reactions, which in turn positively influence satisfaction. Therefore, H1 is supported.
The path analysis also seems to confirm the over-described negative relationship between crowding and
satisfaction (Eroglu and Machleit, 1990, Machleit et al, 2000). However, as suggested through our
previous analysis, the shopping situation should moderate this relationship. Therefore, a multigroup
analysis is performed to assess structural differences between the bar and bookstore situation. After
checking for configural and metric invariance across the two situations, the structural invariance of the
model is evaluated.
         All the parameters in the causal structure are constrained to be equal across the two situations.
The fit indicators of the resulting model are a χ² of 871.2 with 161 degrees of freedom, χ²/df = 5.21, CFI
= .856 and RMSEA = 0.090. This suggests a relatively poor fit to the data. Furthermore, the Lagrange
Multiplier (LM) test for releasing constraints indicated that one constraint (out of 3) should be released,
supporting the poor similarity of structural coefficients between the two groups. The identical paths
across the two groups are the non-significant effect of density on satisfaction and the impact of affective
evaluation on satisfaction. Once the constraint on the density-affective reaction path is released, the model
is reevaluated. The final indicators are a χ² of 426.7 with 160 degrees of freedom, χ²/df = 3.71, CFI =
.938 and RMSEA = 0.069. These results suggest a fairly good fit to the data. The parameter estimates are
presented in table 3.

                                             Insert Table 3 about here.

         A closer analysis of the results yields to major findings. The poor quality of the overall fit
indicators of the constrained model supports the non-equivalence of one parameter estimate in the two
situations. The effect of shopping situation seems to simply switch the nature of the influence of
perceived density on the affective evaluation of the situation. The traditional negative influence of density
is only working in the utilitarian situation, on the contrary, in the leisure situation, the density enhances
the retail experience of the consumer. This results support the findings presented in the ANOVA analysis.


         The presence of other customers in the service factory has turned the typical service consumption
experience in encounters where, in many situations, our evaluation of the service is partly or totally based
on the interactions we have with non-service providers. The group (crowd) influences have surprisingly
been studied in a limited manner and mainly to show the negative impact of having too many people in a
retail setting (Eroglu and Harrell, 1986; Eroglu and Machleit, 1990). Through the integration of sociology
or psychology related research in the services marketing literature on crowding, this paper tests the
existing model of retail crowding and offer alternative interpretations to the potential impact of perceived
density in service situations.
         The main contribution of this study lies in the role played by expectations in the way consumers
deal with density in a retail setting. As a result, significant interactions between the service situation at
stake and the level of density confirmation/disconfirmation suggest that expectations can increase
pleasure in a leisure situation with high unexpected levels of density whereas they can decrease negative
reactions in utilitarian situation with high expected levels of density. This result suggests the important
role played by the surprise effect in building value for the consumer in a leisure setting and increasing
stress in the utilitarian solution. Overall, it seems that to better control consumers’ reactions to crowd, one
should manage their expectations.
         From a managerial standpoint, this research contributes to a better understanding of the way a
crowd should be managed or even how a crowd can be useful and under what conditions. For instance,
managers may want to create and use feelings of high density if they find themselves in a situation where
the crowd contributes to the experience. They also may want to warn customers and advertise how many
fellow customers may be expected in the retail setting. The identification of other moderating variables
such as similarity with the crowd or scarcity of the service offered represents the next step in research
crowd issues. As presented in this study, it is critical to understand circumstances and contexts in order to
better manage crowd issues. In conclusion, it is important to understand that too many is not always bad
and that one can control specific key aspects to use overcapacity its advantage.

       Shopping Motives        Expectations             Constraints                Environmental Cues


                                   Affective Evaluation
                                    Functional   Crowding

                                     Service experience

                                                        Adapted from Eroglu and Harrell, 1986, Journal of Retailing.
                     Figure 1.      An extended Model of Retail Crowding
                  (Adapted from Eroglu and Harrell, 1986, Journal of Retailing)

                          Source                                 Dependent Variable                                    F      P
Situation (A)

A within B (1)                                                Affective Evaluation                           357.1***      0.000

A within B (2)                                                Affective Evaluation                           236.9***      0.000

A within B (3)                                                Affective Evaluation                           209.1***      0.000

A within B (4)                                                Affective Evaluation                           804.6***      0.000

Confirmation/Disconfirmation (B)

B within A (1)                                                Affective Evaluation                           104.3***      0.000

B within A (2)                                                Affective Evaluation                           302.6***      0.000

       **p<0.05; ***p<0.001.

   Numbers in () represents respective levels of variables (Situation: Bar =1, Bookstore =2;
   Confirmation/Disconfirmation: Neg Disc = 1, High conf = 2, Low Conf = 3 and Pos Disc = 4)

                 Table 1. Simple Effects Analysis for significant Two-way interaction

 Affective       5.5

 Evaluation              5.1
                 5.0                                                  5.0

                                          4.3               4.4



                                          3.1                                   Bar
                         2.8                                          2.8
                 2.5                                                            Bookstore
                Neg. Disc              High Conf         Low Conf    Pos.Disc

Figure 2. Situation by Confirmation/Disconfirmation Interaction Effect on Affective Evaluation

                                                                     General Model
                        FIT INDICATORS
                               CFI                                        0.938
                              χ2/Df                                   261/ 75= 3.32
                             RMSEA                                        0.065
                         PATH TESTED                                STANDARDIZED
                Density        Affective Evaluation                   -.194 (-2.05)
                       Density     Satisfaction                              NS
              Affective Evaluation        Satisfaction                 .867 (20.72)

        Table 2. Standardized estimates for the density-satisfaction general model

                                                                  Service experience


                                         Affective Evaluation

                     Figure 3. The density-satisfaction general model

                                                            Bar              Bookstore
           PATH TESTED                                STANDARDIZED       STANDARDIZED
                                                        ESTIMATE           ESTIMATE
                                                         (T-Value)           (T-Value)
  Perceptions   Affective Evaluation                     .760 (8.1)         -.661 (-8.9)
       Perceptions   Satisfaction                           NS                  NS
  Affective Evaluation    Satisfaction                   .455 (4.2)          .497 (4.7)

Table 3. Standardized estimates for the density-satisfaction model in the bar and bookstore


Altman, I (1975). The Environment and Social Behavior. Montererey, Calif.: Brooks/Cole.
Babin, Barry J., William R. Darden, and Mitch Griffin. (1994). “Work and/or Fun: Measuring Hedonic
  and Utilitarian Shopping Value”. Journal of Consumer Research, 20: 644-656.
Baum, A and Epstein, Y.M (1978). Human Response to Crowding. Gillsdale, NJ : Wiley & Sons.
Baum, A. & Valins, S. (1977). Architecture and Social Behaviour. Hillsdale, NJ : Erlbaum.
Bitner, Mary Jo, William T, Faranda, Amy R. Hubbert and Valarie A. Zeithaml (1996). “Quality, and
  Productivity in Service Experiences: Customers’ Roles”. In Advancing Service Quality: A Global
  Perspective, Bo Edyardsson, Stephen W. Brown, Robert Johnston and Eberhard E. Scheuing (eds).
  Jamaica, NY: International Service Quality Association: 289-298.
Bitner, M.J. (1992). Servicescapes: The Impact of Physical Surroundings on Customers and Employees.
  Journal of Marketing 56 (2), 57 – 71.
Brown, T.D, Van Raalte, J and Andersen, M.B (2000). World Cup soccer Home Advantage. Journal of
  Sport Behavior, 25 (2), 134-144.
Chebat, J.-C. (2002).The Interplay of Cognitions and Emotions in Building Customers’ Loyalty Toward
  Services. In: Woodside, A.G., Moore, E. (Eds.), Advances in Marketing Theory and Strategy, Chapter
  2, JAI Press Advances Series, JAI Press: New York.
Choi, S. C., Mirjafari. A. and Weaver. H. B (1976). The Concept of Crowding : A Critical Review and
  Proposal of an Alternative Approach. Environment and Behavior, 8, 345-362.
Czepiel, J. A., Solomon M. R., Surprenant, C. F. and Gutman, E. G. (1985). Service Encounters: An
  overview. In J. A. Czepiel M. R. Solomon and C. F. Surprenant (Eds.), The Service Encounter (pp. 1-
  15) Toronto: Lexington Books.
Donovan, R.J, Rossiter, J.R, Marcoolyb, G and Nesdale A (1994), Store Athmosphere and Purchasing
  Behavior. Journal of Retailing, 70(3), 283-294.
Eastman, S.T and Land, A.M (1997). The Best of Both Worlds : Sports Fans Find Good Seats at the Bar.
  Journal of Sport and Social Issues, 21(2), 156-178.
Eroglu, S.A and Harrell, G.D. (1986). Retail Crowding : Theoretical and Strategic Implications. Journal
  of Retailing, 62 (4), 347-363.
Eroglu, S.A and Machleit, K.A. (1990). An Empirical Study of retail Crowding : Antecedents and
  Consequences. Journal of Retailing, 66 (2), 201-220.
Evans, G and Lepore, S.J (1993). Household Crowding and Social Support : A Quasiexperimental
  Analysis. Journal of Personality and Social Psychology, 65(2), 308-316.
Felmlee, D and Sprecher, S (2000). Close Relationships and social Psychology : Intersections and Future
  Paths. Social Psychology Quarterly, 63(4), 365-376.
Grove, S.J. and R.P. Fisk. (1997). The Impact of Other Customers on Service Experiences: A Critical
  Incident Examination of “Getting Along”. Journal of Retailing 73 (1), 63 – 88.
Harrell, G. D., M. D. Hutt, and J. C. Anderson (1980). Path Analysis of Buyer Behavior under Conditions
  of Crowding. Journal of Marketing Research, 17 (February), 45-51.
Holbrook, M. B. (2000). The Millenial Consumer in the Text of Our times : Experience and
  Entertainment. Journal of Macromarketing, vol 20 (2), 178-192.
Holt Douglas (1995), How consumers Consume: a Typology of Consumption Practices, Journal of
  Consumer Research, 22(2), 1-21.
Hui, M.K. and J.E. Bateson. (1991). Perceived Control and the Effects of Crowding and Consumer
  Choice on the Service Experience. Journal of Consumer Research, 18 (2), 174-184.
Iacobucci, D (1998). Services : What Do We Know and Where Shall We Go? A View for Marketing. in
  Advances in Services Marketing and Management, 7, 1-96.
Insel, P.M and Lindgren, H.C (1978). Too Close for Comfort : The Psychology of Crowding. Englewood
  Cliffs, NJ : Prentice-Hall.

Laurent, Gilles, and Jean-Noel Kapferer, (1985). Measuring Consumer Involvement Profiles. Journal of
  Marketing Research, 22, 41-53.
Lepore. S. J., Evans, G. W., and Schneider, M. L. (1992). Role of Control and Social Support in
  Explaining the Stress and Hassles of Crowding. Environment and Behavior. 24, 795-811.
Machleit, K.A., Eroglu, S.A and Powell Mantel, S. (2000). Perceived Retail Crowding and Shopping
  Satisfaction: What Modifies the Relationship. Journal of Consumer Psychology, 9 (1), 29-42.
Machleit, K.A., Kellaris, J.J and Eroglu, S.A (1994). Human Versus Spatial Dimensions of Crowding
  Perceptions in Retail Environments: A Note on Their Measurement and Effect on Shopper Satisfaction.
  Marketing Letters, 5 (2), 183-194.
Martin, C.L. (1996). Consumer-to-Consumer relationships : Satisfaction with Other Consumers’ Public
  Behavior. The Journal of Consumer Affairs, 30(1), 146-158.
Oliver, Richard L. (2000). Customer Satisfaction with Service. In Handbook of Services Marketing and
  Management. Thousand Oaks, California: Sage.
Oliver, Richard L. (1997), Satisfaction: A Behavioral Perspective on the Consumer. New York: McGaw-
Rapoport, A (1976). Toward a Redefinition of Density. In S. Saegert (Ed.), Crowding in Real
  Environments. Beverly Hills, California : Sage.
Richins, M (1997). Measuring Emotions in the Consumption Experience. Journal of Consumer Research,
  24(2), 127-147.
Rustemli, A (1993). Crowding Effects of Density and Interpersonal Distance. Journal of Social
  Psychology, 132(1), 51-58.
Sherry, J.F. (1998). “The Soul of the Company Store” in Managing the Servicescape, ed. J.F. Sherry, 109
  – 146.
Shostack, L. (1985). Planning the service encounter. In Czepiel, J, A, Solomon, M.R. and Surprenant,
  C.F. (Eds), The Service Encounter Managing Employee/Customer Interaction in Service businesses,
  Lexington Books, Lexington, MA, 243-54.
Sinha, S.P and Nayyar, P (2000). Crowding Effects of density and Personal Space requirements Among
  Older People : The Impact of Self-Control and Social Support. Journal of Social Psychology, 140 (6),
Stokols, D (1972). On the Distinction between Density and Crowding : Some Implications for Future
  Research. Psychological Review, 79(3), 275-278.
Stott, C and Drury, J (2000). Crowds, Context and Identity : Dynamic Categorization Processes in the
  ‘Poll Tax riot’. Human Relations, 53(2), 247-273.
Wakefield, K.L and Blodgett, J.G (1994). The Importance of Servicescapes in Leisure Service Settings.
  Journal of Services Marketing, 8(3), 66-76.
Webb, W.M and Worchel, S (1993). Prior Experience and Expectation in the Context of Crowding.
Journal of Personality and Social Psychology, 65(3), 512-521.
Yi, Y. (1990). A Critical Review of Consumer Satisfaction. Review of Marketing 1990, AMA, Chicago,