A Longitudinal Assessment of Consumer Satisfaction/Dissatisfaction

advertisement
A Longitudinal Assessment of Consumer Satisfaction/Dissatisfaction: The Dynamic Aspect of
the Cognitive Process
Author(s): Priscilla A. LaBarbera and David Mazursky
Source: Journal of Marketing Research, Vol. 20, No. 4 (Nov., 1983), pp. 393-404
Published by: American Marketing Association
Stable URL: http://www.jstor.org/stable/3151443 .
Accessed: 12/10/2013 07:33
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected]
.
American Marketing Association is collaborating with JSTOR to digitize, preserve and extend access to
Journal of Marketing Research.
http://www.jstor.org
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
I
A. LaBARBERA
and DAVIDMAZURSKY*
PRISCILLA
A simplifiedcognitivemodelis proposedto assess the dynamicaspect of consumer
in consecutivepurchasebehavior. Satisfactionis found
satisfaction/dissatisfaction
to have a significantrole in mediatingintentionsand actual behaviorfor five product
classes that were analyzed in the context of a three-stage longitudinalfield study.
The asymmetriceffect found demonstratesthat repurchaseof a given brand is affected by lagged intentionwhereas switchingbehavioris more sensitiveto dissatisfactionwithbrandconsumption.An attemptto predictrepurchasebehavioron the
basis of the investigatedcognitivevariablesyielded weak results. However,repurchase predictionswere improvedwhen the model was extended to a multipurchase
setting in whichprior experiencewith the brandwas taken into account.
A
of
Consumer
Longitudinal Assessment
The
Satisfaction/Dissatisfacti
Dynamic
of
the
Aspect
Cognitive Process
III
A simple model of satisfaction as a function of expectations and disconfirmation (i.e., the extent to which
the expectationsare met) was suggested by Oliver (1980a)
to summarize the early research on the antecedents of
satisfaction (see Anderson 1973, Cardozo 1965, Oliver
1977, Olshavsky and Miller 1972, and Olson and Dover
1979 for representative reports). Oliver (1980b) also
proposed a cognitive model to account for the consequences as well as the antecedents of satisfaction in the
formation of purchase intentions. This model postulates
that satisfaction acts as a mediator between preexposure
and postexposure attitudes. In contrast to treating satisfaction as a static dependentvariable, the cognitive model
recognizes that satisfaction is part of the dynamic purchase process and influences repurchase intentions.
Longitudinal research is needed to test the feedback
effects of postpurchase behavior hypothesized by an ex-
tended cognitive model (Oliver 1980a). Although a few
longitudinal studies of consumer behavior have been reported (see Oliver 1977, automobile purchases; Oliver
1980b, flu vaccination program; Swan 1977, department
store shopping; Swan and Trawick 1981, restaurants),
the analysis of the purchase process in the consumer satisfaction literature has not progressed much beyond the
prepurchaseand postpurchase conditions pertaining to a
single purchase activity. Only in other areas of research,
such as job satisfaction, have multiple-exposure studies
analyzed the relationship between satisfaction and two
or more trials (see Ilgen 1971; Ilgen and Hamstra 1972;
Weaver and Brickman 1974). The purpose of our article
is to extend the understanding of consumer satisfaction
by examining multiple consecutive product purchases in
the context of a field study.
Oliver (1980a) identified adaptation level theory (Helson 1964) as a promising theoretical framework to support several findings of consumer satisfaction studies.
This theory seems to be appropriate for explaining how
past cognitions, which serve as the adaptation levels, are
mediated by present satisfaction in influencing repeat
purchase behavior. The theory, which has its roots in
physiology, stipulates that previous experience with the
phenomenon under study generates an adaptation level
or anchor for subsequent judgments. According to Helson (1964, p. 128), "Stimuli above [the] adaptation level
*Priscilla A. LaBarberais Associate Professor of Marketing, Graduate School of Business Administration, New York University. David
Mazursky is Lecturer, School of Business Administration, Hebrew
University, Jerusalem.
The authors acknowledge financial support of this research from
the New York University Research Challenge Fund. Thanks are expressed to Jacob Jacoby, Srinivas Reddy, and Carol Surprenant for
their valuable comments on drafts of this article.
393
Journal of Marketing Research
Vol. XX (November 1983), 393-404
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
NOVEMBER
1983
JOURNALOF MARKETING
RESEARCH,
394
give rise to positive gradients of excitation and responses
of one kind, while stimuli below [the] adaptation level
give rise to negative gradients of excitation and responses of an opposite kind." In our study, prior intentions were hypothesized to act as the adaptation level for
future intentions (Oliver 1980b, p. 462) and the satisfaction/dissatisfaction from the brand consumption represented the stimulus that mediates the changes.
Thibout and Kelley's (1959) comparison level theory
provides another useful framework to explain the role of
satisfaction in mediating preconsumption and postconsumption cognitions. The theory predicts that satisfaction develops from an interactionamong individuals, and
results from the discrepancy between outcomes and a
certain comparison level. LaTour and Peat (1979) applied the theory in the context of a product purchase setting. The theory is particularly appealing because it adds
a dynamic aspect whereby outcomes from a present experience modify the comparison level for future consideration. According to comparison level theory, a high
level of satisfaction results in a greater difference between outcome and the comparison level which yields a
higher intention to repurchase the brand.
A COGNITIVEMODEL OF REPEAT
PURCHASE BEHAVIOR
Several researchers have examined the dynamic aspect
of the purchase-repurchaseprocess in which satisfaction
is treated as the feedback of purchase behavior. In the
revised Howard and Sheth model (Howard 1974) the
consumptionof the branddeterminesthe satisfactionlevel
which, in turn, affects the revised attitude as well as the
intention to repurchase the brand. In Oliver's (1980b)
cognitive model of the purchase process, the revised intention to repurchase the brand is formulated in a way
similar to that suggested by Howard (1974):
I, = f (l-1, SAT,ATT,).
The intention (I,) in the present period is determined
by three components, (1) the previous intention (I,-_),
which serves as the adaptation level for the current intention, (2) the satisfaction level (SAT) from the brand
consumption, which mediates the changes between the
prepurchase and postpurchase intentions, and (3) the
current attitudinal level (ATT,). In reviewing the consumer satisfaction literature, LaTour and Peat (1979, p.
434) conclude, "Given that attitude and satisfaction are
both evaluative responses to products, it is not clear
whether there are any substantive differences between
the two." However, Oliver (1981) suggested a conceptual difference between attitude and satisfaction. He defined satisfaction as "an evaluation of the surprise inherent in a product acquisition and/or consumption
experience." The surprise or excitement is of finite duration, so that satisfaction soon decays into attitude toward purchase. Because the purpose of our study was to
extend the investigation to behavior in a multiple-purchase setting, with short time spans between purchases
and cognitive evaluation, we used a measure of satisfaction/dissatisfaction to estimate postpurchase evaluation. In addition to the theoretical problem of differentiating between attitude and satisfaction for an extended
experience (i.e., repurchase behavior), Oliver (1980b)
confronted practical difficulties in testing his cognitive
model due to multicollinearity among both preexposure
and postexposure cognitions.
Because of the theoretical and practical issues related
to the distinction between satisfaction and attitude, a
simplified model is suggested for the investigation of
successive purchase-repurchaseevents. In this model the
repurchaseis tested as a categorical variable (repeat purchasers versus brand switchers), satisfaction/dissatisfaction estimates postpurchase evaluations, and intention is
believed to act as the adaptation variable. This model
permits investigation of the adaptation level formation
in which satisfaction mediates the adjustments. A high
satisfaction level will have a positive impact on the intention level for repeat purchasers, whereas dissatisfaction could have a negative influence on the intention level
for brand switchers.
The cognitive process studied is postulated to proceed
as follows.
It- -
Pt-, SAT-> It-- Pt+
This process yields the following general model.
(1)
P,t+ = f (I, SAT, P,, I,t-)
Let RP denote the repeat brand purchase behavior and
SW be the switching behavior. Because P,+l is a categorical variable denoting overt behavior (P,+i = RP for
repeat purchasers and P,+l = SW for switchers) and is a
function of 1, (Howard 1974), a simple path analytical
model is suggested:
(2a)
I,= gl (I,_-,SAT)
and
(2b)
SAT = g2 (I,-).
The model should be analyzed separately for repurchasers and switchers because of possible asymmetry in the
relationship between the cognitive variables and consumer decision implications. Several hypotheses about
the average intensitiesof satisfactionand intentions across
consumers for a given brandare postulated by the model:
l,IRP> I,fSW,
(3)
(4)
> SAT,ISW,
SAT,IRP
and
(5)
i,-,\RP > i,- ISW.
That is, equations 3 and 4 stipulate that the average
revised intention and satisfaction levels are higher for
consumers who subsequently repeat the purchase of the
same brand than for those who switch. Likewise, hy-
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
CONSUMER
SATISFACTION/DISSATISFACTION
395
pothesis 5 implies that the decision to repurchase is preceded by differences in the intention level during the
lagged period. This hypothesis is attributed to the fact
that a repeat purchase decision (or switch) is generated
gradually, and therefore is reflected in higher (lower) expectation and attitudinal favorability which form intentions in the preceding period.
The hypothesized rdle of satisfaction in mediating the
intention level implies that within each group (P,+l =
RP or P,+1 = SW) the postpurchase cognitions will score
higher than prepurchase cognitions for repurchasers. In
contrast, postpurchase cognitions will be lower than prepurchase cognitions for switchers. Thus, for the repurchasers
I,IRP>
(6)
,t-
IRP
and for the switchers
(7)
!,ISW
< !,-,ISW.
Finally, the satisfaction/dissatisfaction from brand consumption is hypothesized to amplify the differences in
the intention levels between the groups over time. This
relationship is summarized by
(8)
(IIRP)- (IISW)> (I,- IRP) - (I,- SW).
Figure 1 illustrates the hypothesized cognitive process
for the repurchaserand switcher groups for two consecutive brand purchases.
In addition to the bivariate differences, a simultaneous
relationshipamong the three cognitive variables (It_l, SAT,
It) is hypothesized to yield the sequence It-_ -> SAT ->
It. Satisfaction is predicted to mediate the difference between the intention levels to produce a relationship that
is stronger than the direct effect of the lagged intention
on revised intention.
Figure 1
A PATHMODELOF SATISFACTIONDECISIONSAND INTENTIONLEVELREVISIONSUNDERLYINGOVERTBEHAVIOR
PERIOD
t-l
INTENTION AND
SATISFACTION
J
4%
PERIOD
1
I
PERIODt
Ir-
I
z
0
H
H
:4
Ci2
Repurchase
Brand X
H
z
0
H
Ez
zH
Pe
rX,
En
F4
to
:r
H
H
Switch
to
Brand Y
H
0
H
(a)
(b)
TIME
denotes
the direct
of prior
effect
intention
on the revised
intention
of prior
denotes
the indirect
intention
on the revised
effect
intention
on revised
intention
as mediated
with brand consumption).
by satisfaction
(i.e.,
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
the
effect
396
JOURNAL
OF MARKETING
NOVEMBER
1983
RESEARCH,
STUDY OBJECTIVES
To date consumer satisfaction research has concentrated on a single purchase period. We investigated the
role of satisfaction in determining repurchase intentions
and overt behavior within the context of a longitudinal
study. The dual purpose of our study was to (1) investigate the direct and indirect effects of satisfaction/dissatisfaction on the formation of revised intentions and
overt behavior and test a simplified cognitive model over
two consecutive purchase periods for five different product classes and (2) test the extent to which reported satisfaction/dissatisfaction and the appropriate intention
levels directly predict repeat purchase behavior in the
context of multiple consecutive purchasesin a field study.
METHOD
To obtain a consumer panel, a sample of 500 individuals was selected at random from five randomly selected
telephone directories of residential communities in Connecticut. The potential participants received information
aboutthe intendedstudy, theirresponsibilitiesif they chose
to and were qualified to participate, and the reward for
their cooperation.1 One hundred and eighty consumers
who indicated a willingness to participate were asked to
complete several items to determine how often they purchased each of 24 grocery products and to pretest the
questionnaire designed for the study. One hundred and
twenty-five consumers who purchased and used at least
12 of the 24 products on a biweekly basis were invited
to join the panel.2
The research instrument was revised on the basis of
the pretest findings. The final questionnaire consisted of
three items pertainingto each of 24 grocery productssuch
as tissues, coffee, and dish detergent. Respondents were
asked to report the name of the brand currently in use,
satisfaction with the current brand, and intention to repeat the purchase of the current brand. In addition, respondentswere asked whether they had purchasedor used
another brand between the current and last questionnaires. Panel members completed questionnaires biweekly over a five-month period.
Although 125 consumers responded to the first questionnaire, 38 panel members dropped out during various
stages. The final analysis includes 87 consumers who
completed each of the 10 questionnaires. This group represents 70% of the 125 consumers who were eligible to
participateand 48% of the 180 consumers who indicated
their willingness to participate.
After collection of the data, the product classes were
screened to identify those that most closely approximated a two-week interpurchase interval. Consumer recall of repurchase frequency served as the basis for the
screening. The purpose was to eliminate bias in the ag'The reward was five books of S & H Green Stamps.
2Consumerswho did not qualify for panel participation received one
book of S & H stamps.
gregate analysis in which the disconfirmation period is
assumed to be equal for all consumers. The five product
classes that qualified for the analysis were margarine,
coffee, toilet tissue, paper towels, and macaroni.
Purchasebehavior was tracked indirectly by pretesting
interpurchasefrequency and by inserting an additional
question about purchases of other brands between periods. This indirect method was used to simplify the respondents' task, which would otherwise have entailed
the completion of a purchase diary in addition to the 10
questionnaires.Though pretest results indicated that most
consumers purchased a single item of a certain brand
from a specific product category every two weeks (for
the five product classes investigated), three exceptions
to this pattern might have occurred during the study period. First, an "inventory" of a brand purchased previously could have been in use rather than the purchased
brand. Second, two or more purchases could have been
made during the two-week period. Third, two or more
brands could have been purchased or used concurrently
during the period. On the basis of the pretest findings
we assumed that such patterns would not seriously affect
the results. In addition, in several instances, consumers
indicated that other brands were purchased or used between periods. Consecutive purchases that were interrupted by the purchase or consumption of other brands
were deleted from the analyzed data files.
Added to the last questionnaire was a section pertaining to respondent demographic characteristics. In many
consumer behavior studies a comparison between respondents and nonrespondents indicates whether a response bias is present, but this procedure was determined not to be appropriate in our study because past
researchdemonstratesthat purchase satisfaction does not
correlate significantly with demographic variables
(Westbrook and Newman 1978).
Measures
In measuring satisfaction/dissatisfaction, several researchers prefer an overall summary measure whereas
others argue that satisfaction should be measured by a
combination of attributes. Day (1977a) claims that there
is no difficulty with measuring overall consumer satisfaction. Czepiel and Rosenberg (1976) justify the use of
a single-item overall measure because it represents a
summaryof subjective responses to several different facets. Other researchers use summative scales to measure
satisfaction/dissatisfaction (Oliver 1980b; Swan and
Trawick 1981; Westbrook and Oliver 1980).
The pretest demonstrated that multiple-item measures
designed to obtain informationabout many productclasses
substantiallylimited the number of consumers willing to
participate. A poor response due to questionnaire length
could have caused a significant nonresponse bias. Moreover, because the questionnaire did not investigate many
different issues, the use of several items to measure certain variables (i.e., satisfaction, intention)repeatedly over
time might have resulted in artificial answers by re-
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
397
CONSUMER
SATISFACTION/DISSATISFACTION
spondents. Therefore, the use of summative scales in our
research would probably have decreased rather than increased reliability. As suggested by psychometricians
(Deutcher 1966; Hirsci and Selvin 1967; Kimberly 1976),
a modification of the scales in a longitudinal study is
appropriate;in our study this entailed the use of overall
satisfaction and intention measures. Accordingly, the intention scale used ranged from "extremelylikely" to "not
at all likely" and the satisfaction scale ranged from
"completely satisfied" to "not at all satisfied." Both
variables were measured on 5-point scales.
ANALYSIS
To investigate repurchase behavior, lagged variables
for brand consumption within each product class were
computed (Hull and Nie 1979). After deletion of missing
values for consumers who did not purchase any brand
within the product class during any of the 10 periods,
and deletion of consecutive purchases which were interrupted by the purchase of another brand between periods, 2568 cases of two consecutive purchases of the
same brand within the product category were detected
and consisted of 1969 repeat purchasesand 599 switches.
A computed categorical variable called CH, to denote
change for repurchase (CH = 0) versus switching behavior (CH = 1), was matched with the intention and
satisfaction level of the appropriateperiods. Each of the
two consecutive purchase periods constituted the unit of
analysis (because each consumer had more than one such
"unit").
The Role of Satisfaction as a Mediating Factor
To test the role of satisfaction in mediating intention,
the analysis was performed separately for repeat purchasers and switchers who demonstrated the opposite
overt behavior in the next successive period. A just identified general path analytical model (Duncan 1966) was
applied to the two groups. The analysis first concentratedon the total population. For purposes of testing the
generalizability of the findings, the total population was
subsequently split into the five product categories in an
individual analysis. The variables were standardized to
represent the standardized regression path coefficients.
Path analysis observations were analyzed for three points
in time. For the first observation, the intention to purchase the previously consumed brand was measured (I,_in Figure 1). In the next time period the brand purchase
observation (Pt) was considered as well as the satisfaction (SAT) from the brand consumption and the intention (It) to repurchase the same brand. For the third
observation, the actual repurchase (or switching) observation (P,t+) was obtained. In this process, which was
applied on the aggregate level for three successive periods, the individual cases were used as input.
Correlational analysis revealed similar patterns of relationships investigated for repeat purchasers and brand
switchers among the three variables (I,_,, SAT, I,). For
the total population, the strongest relationship is between
the two postpurchase variables (r = .73 for repeat purchasers and r = .77 for switchers). Simple correlations
exceed the r = .34 level. The mean values of lagged
intention, satisfaction from the previous purchase stage,
and revised intentionare higher for repeat purchasersthan
for brand switchers (as postulated in equations 2, 3, and
4). Table 1 demonstrates that the differences are significant across four of five product classes.
Two alternative explanations may account for this difference. First, the intuitive explanation is that, in comparison with brand switchers, consumers intending to repurchase the same brand score higher on the revised
intention as well as the satisfaction from the previous
consumption. Even the previous intention shows significant differences between the two groups. Alternatively,
the previous purchase may not necessarily be the first
trial of the brand and, by taking consumption experience
into account, a longer history might yield a higher level
of previous intention. Different purchase histories of the
same brand between the two groups could affect the initial investigated intention level. In a subsequent section
we demonstratethat such an effect accounted for part of
this difference.
Our study concentrated on the deviations of the revised cognitions from the lagged intention (within each
of the two groups) with the lagged intention serving as
a reference point for the satisfaction mediation. Thus, it
emphasized the relative changes of the revised intention
from the original level. However, the absolute value of
the previous intention (which is different for the two
Table 1
IN MEAN SCORES OF LAGGEDINTENTION,SATISFACTION,AND INTENTIONTO REPURCHASEBETWEEN
DIFFERENCES
AND SWITCHING
GROUPS
THEREPURCHASE
,-i
SAT
I,
Aggregate
Repur- Switchchase
ing
4.41
4.07
4.48
4.07
4.46
3.91
Diff.
0.34a
4.41a
0.55a
Margarine
Repur- Switchchase
ing
Diff.
4.23 0.11
4.34
4.48
4.12 0.36a
3.93 0.49a
4.42
Coffee
Repur- Switchchase
ing
4.24
4.44
4.16
4.55
4.03
4.51
Toilet tissues
Diff.
0.20b
0.39a
0.48a
Repur- Switchchase
ing
Diff.
4.41
3.95 0.46a
4.44
3.97 0.47a
4.45
3.81 0.64a
Paper towels
Repur- Switchchase
ing
Diff.
3.75 0.62a
4.17
4.42
3.91 0.51a
3.65 0.77a
4.42
'p < 0.01.
bp < 0.05.
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
Macaroni
Repur- Switchchase
Diff.
ing
4.16
0.32a
4.48
4.53
4.20
0.32a
4.14
4.49
0.35a
JOURNAL
OF MARKETING
NOVEMBER
1983
RESEARCH,
398
groups; see Figure 1) might account for some of the differential changes in the revised intention as affected by
the satisfaction intervention. To control for both the absolute level and the relative intention adjustments, a ttest3 was performed on the difference over time in intention levels between the two groups (see equation 8).
The test shows that for the two groups the difference in
the intention levels increases significantly (p < .01) between the two periods.
Table 2 shows the results of the paired t-test for the
differencebetween the means of the intentionlevels within
the two groups (see equations 6 and 7). The differences
between the previous intentions and the revised adaptation levels support the expected direction in all cases
although not all are statistically significant.
Table 3 summarizes the path analysis findings for the
effects of prepurchase intention and postpurchase satisfaction on the revised intention to repurchase the brand.
In both groups (and across all product classes) the effect
of the lagged intention through the intervention of satisfaction is larger than the direct effect of the initial adaptation level on revised intentions. Likewise, the residual path coefficients of the model, in predictingthe change
in the revised intention, indicate that a relatively large
part of the variance is explained by both the prepurchase
and postpurchase predictors (E3 in Table 3). Potential
multicollinearity among the independent variables (I,-,,
SAT) might influence those effects in an arbitrarydirection. The average correlation among the two independent variables is r = 0.37. However, in 10 of the 12
simple correlation matrices the correlations among the
independent variables are the lowest. This finding might
indicate that the chance occurrence of multicollinearity
was minimized.
A systematicdifference between the two groups is found
in the path analysis presented in Table 3. The direct effect of the previous intentionon the revised level is higher
and more significant for the repeat purchasers than for
brand switchers. The opposite dominance appears in the
case of brand switchers, for whom the effect of dissatisfaction on the revised intention yields larger path coefficients in all cases, with the exception of one product
class. Such an observation was expected because the
variability in satisfaction/dissatisfaction scores is larger
3Equation 8 was computed as follows:
AI,_, = I,_- I RP - I,_, I SW = 0.338
AI, = I, I RP - I, I SW = 0.541
A(AX) = A(I,) - A(l,_,) = 0.204
Let S denote the pooled standard error; then
S
rnl,, s12+ n2s22
n, + n2
= 0.03,
n
\
i
=
+
L
n
n
2
nl + n2 - 2
ni ?n2
J
t
=
6.8.
Table 2
PURCHASE
INTENTIONS
RELATED
MEANSOF PREVIOUS
TO REPURCHASE
VS. SWITCHING
TO MEANINTENTION
BEHAVIOR
It-I
Repeat purchase group
Aggregate
Margarine
Coffee
Toilet tissue
Paper towels
Macaroni
Switching group
Aggregate
Margarine
Coffee
Toilet tissue
Paper towels
Macaroni
->
,
I,-1
Expected
change
Actual
(n =)
4.41
4.34
4.43
4.41
4.37
4.48
+
+
+
+
+
+
+c
+
+
+
+
+
1969
326
385
428
416
414
-
_b
c
-
-
599
121
130
116
118
114
4.07
4.23
4.23
3.97
3.75
4.15
aThesignificance of differences was computed by a correlated (pairs)
t-test. The minus (-) sign denotes decrease whereas plus (+) denotes
increase between the two variables.
p < 0.01.
cp < 0.05.
for switchers than for repeat purchasers.
To investigate the direct and indirect effects of lagged
intentions on intention formation, consumers' evaluations of the cognitive variables were subsequently crossclassified. Whereas the path analysis focused on the
changes in mean intensities of the appropriate cognitions, a three-way frequency analysis was conducted to
examine changes in the frequencies of consumers represented in each of the cells. The three lowest values of
both the intention and satisfaction scales were grouped
to obtain these matrices. The first reason for this grouping was the upward skewness in the scales' frequency
(which is a typical phenomenon in most satisfaction research). The second reason was to reduce the matrix of
scales to low, medium, and high satisfaction and intention.
Table 4 summarizes the results of this analysis. Entries of equality of SAT and I, intensities represent adjustment of the revised intention to the satisfaction/dissatisfaction level as denoted by footnote b. These figures
may indicate the mediating effects of satisfaction/dissatisfaction on revised intention. In addition, equal intensities of lagged intention (I,1-) and revised intention
(I,), denoted by footnote c, represent those respondents
whose intention level was not affected by satisfaction/
dissatisfaction levels. The cells denoted by both footnotes b and c comprise all cases in which intensities of
cognitions do not vary throughoutthe cognitive process.
If the entries denoted by footnote c are larger than the
other entries, we might expect that lagged intentions and
satisfaction levels do, in fact, influence the level of the
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
CONSUMER
SATISFACTION/DISSATISFACTION
399
Table 3
DIRECT
ARECT
AND INDIRECT
EFFECTS
OF LAGGEDINTENTION
TO
(-) AND SATISFACTION
(SAT)ON INTENTION
REPURCHASE
(4)
Direct effecta
SW
Variable
Product class
1,-,
Aggregate
0.19b
0. 10b
Margarine
Coffee
0.27b
0.14b
0.07
0.01
Toilet tissues
Paper towels
Macaroni
Aggregate
Margarine
Coffee
0.16b
0.16b
0.20b
0.65"
0.56"
0.79b
0.07
0.10
Toilet tissues
0.62b
Paper towels
Macaroni
Aggregate
Margarine
Coffee
Toilet tissues
Paper towels
Macaroni
0.65b
0.68b
SAT
RP
Indirect Effect'
RP
Total Effecta
SW
RP
SW
0.27b
0.25b
0.46b
0.35b
0.16b
0.14c
0.37b
0.24b
0.43b
0.46b
0.21c
0.38b
0.39b
0.47b
0.53b
0.65b
0.56b
0.79b
0.31"
0.31b
0.53b
0.74b
0.73b
0.83b
0.79"
0.62b
0.79b
0.71b
0.58
0.65b
0.68"
0.7 lb
0.68b
0.33b
0.74b
0.73
0.83b
0.32b
0.23b
0.31b
0.33b
-
-
0.21
0.20b
-
-
Coefficients' residual path
RP
SW
E2 =
E2 =
E2 =
E2 =
E2 =
E2 =
E2 = 0.94 E3 = 0.62
E3 = 0.66
E3 = 0.73
Ez = 0.98 E3 = 0.67
E2 = 0.89 E3 = 0.55
E3 = 0.51
E2 = 0.95 E3 = 0.57
E3 = 0.72
E2 = 0.95 E3 = 0.66
E3 = 0.66
E2 = 0.94 E3 = 0.64
E3 = 0.60
aStandardizedregression coefficients (P,) were obtained by applying a regression analysis on equations 2a and 2b. Let us denote I,_- by g, I,
by r, and SAT by s. Accordingly, the indirect effect of g on r is equal to P ,s x P,sr and the direct effect of g on r is equal to Pg,r where Pi are
obtained from equation 2a. The direct effect of s on r is equal to Ps,r obtained from equation 2b.
bp 0.01.
'p c 0.05.
revised intention. The results of Table 4 support the contention that lagged intention and the mediating effect of
satisfaction contribute to the formation of intention levels. The magnitudes of the frequencies in each of the
three investigated classes-the cells denoted by footnote
b, the cells denoted by footnote c, and the rest of the
entriesin the matrices-reveal a clear pattern.Both lagged
intentions and satisfaction show high frequencies for all
the values that coincide with similar values of revised
intention. Moreover, between the two cognitions that are
0.91
0.96
0.91
0.93
0.88
0.87
postulated to affect the revised intention, the mediating
effect of satisfaction appears to be more substantial than
the lagged intention.
The total population subsequently was split into repurchaserand switcher categories. One observation from
the tables is noteworthy. Although an upward skewness
in the satisfaction and intention scales is expected for
repeat purchasers, a substantial skewness is also observed for switchers. For example, 261 respondents
(43.5% of the switcher group) reported high satisfaction
Table 4
A CROSS-CLASSIFICATION
OF THECOGNITIVE
VARIABLES
I,
Medium
Low
RP
I,-1
SAT
Count
Low
SW
%
Count
Low
118
60
79
Med.
23
59
16
70
19
8
High
Med.
Low
39
61
25
Med.
20
71
8
9
4
69
High
Low
38
45
46
High
Med
22
51
21
14
6
70
High
values
are
at
c
0.01.
X\2
significant p
bEqualSAT -+ I, intensities.
RP
High
SW
RP
SW
%
Total
Count
%
Count
%
Totala
Count
%
Count
40
41
30
39
29
31
55
49
30
197bc
10
59
20
14
151
10
5
84
19
62
72
67
70
74
62
71
82
73
6
23
10
6
54
6
2
18
7
38
28
33
30
26
38
29
18
27
16
82b
30
20C
12
9
77
5
13
100
20
61
1001
80
75
72
100
81
79
83
74
86
3
3
30
0
3
27
4
21
163
39c
27c
64b
28
13
84b
43
20
205bc
16c
7
102b
26
CEqual I,- -> I, intensities.
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
%
20
25
18
0
19
21
17
26
14
Total
15
12
107b
5
16
127b
24c
82c
1164b.c
400
JOURNALOF MARKETING
NOVEMBER
RESEARCH,
1983
with the brand although they switched brands in the subsequent period.
The Effects of Satisfaction/Dissatisfaction on Future
Behavior
Consumersatisfaction/dissatisfactionresearchmay have
substantial marketing implications if postpurchase evaluations largely predict future purchase behavior. The assumption that satisfaction/dissatisfaction meaningfully
influences repurchase behavior underlies most of the research in this area of inquiry. However, the extent to
which satisfaction/dissatisfaction influences future purchase behavior has not been investigated rigorously. One
possible explanation for the lack of research on the linkage between satisfaction and repeat purchase may be a
past concentration on the antecedents of satisfaction and
research utilizing social psychological experimentation,
which inhibited such investigations (Russo 1979). The
purposeof this section is to investigatethe degree to which
satisfaction/dissatisfaction and intentions (lagged and
revised) predict future purchases. Equation 1 is a model
demonstrating these relations.
In contrast to the dearth of empirical research on the
satisfaction/repeat purchase behavior link, the relationship between intention and repurchase behavior is well
established (see Warshaw 1980 for a detailed review).
For the most part these relationships have been found to
be poor, particularlyin product-relatedsettings. For example, Bonfield (1974) calculated a correlation of 0.44
between intentions and fruit drink choices and Harrell
and Bennett (1974) obtained r = 0.37 when intentions
and physician prescribingbehavior were correlated. Thus,
in our study a weak relationship between intentions and
repurchase behavior was expected. Moreover, because
satisfaction was posited in our proposed cognitive process to precede intentions in the sequence of causality,
satisfaction/dissatisfaction was expected to yield correlations with purchase that are weaker than the intention/
future behavior relationship.
A multiple discriminant analysis was used in which
repeat purchase behavior (P,+l) was treated as the dependent variable. The dependent variable comprised repurchasers (RP) and switchers (SW). The independent
variables were the lagged intention, satisfaction, and revised intention. Similar to the path analysis procedure,
the analysis first was performed for the total population
and subsequently employed for each of the five product
classes.
The results of the discriminant analysis are summarized in Table 5. Column 7 of this table shows the simple
correlationsof the various cognitive variables and repeat
purchase behavior. These correlations reveal that in all
Table 5
THE EFFECTSOF LAGGEDINTENTION(It-,), SATISFACTION/DISSATISFACTION
(SAT), AND REVISEDINTENTION(I,) ON
BRANDREPURCHASE
BEHAVIOR
BETWEEN
AND SWITCHING
REPURCHASING
BEHAVIOR
(P,+1):A COMPARISON
(5)
11 1
(3)
Standardized
function
coefficients
0.758
0.176
0.199
Margarine
I,
SAT
I,_
0.728
0.421
-0.175
0.953
0.961
0.998
21.8
18.2
0.96
P < 0.001
P < 0.001
N.S.
Coffee
I,
SAT
It-1
0.864
0.177
-0.043
0.953
0.962
0.991
25.13
20.01
4.26
P < 0.001
P < 0.001
P < 0.05
Toilet
tissues
I,
SAT
0.717
0.153
0.319
0.932
0.952
0.968
39.76
27.31
18.06
P < 0.001
P < 0.001
P < 0.001
Paper
towels
I,
SAT
ItI1
0.874
-0.164
0.414
0.906
0.951
0.942
54.79
27.26
32.53
P < 0.001
P < 0.001
P < 0.001
Macaroni
I,
SAT
0.251
0.500
0.430
0.967
0.966
0.972
17.45
18.36
14.69
P < 0.001
P < 0.001
P < 0.001
(1)
Product
Aggregate
(2)
Variable
I,
SAT
Univariate F-ratios
(4)
Wilks'
lambda
Level
0.943
0.959
0.979
153.6
108.5
55.9
Significance
P < 0.001
P < 0.001
P < 0.001
(6)
Percent of
cases correctly
classified
(7)
Correlations
with P,+1
0.24a
0.20a
0.15a
64.5
0.22a
0.20a
0.04
63.8
0.22a
0.19
0.09"
63.5
0.26a
0.22a
0.18a
68.8
0.30a
0.22a
0.24a
67.4
I,-_
0.18a
0. 18a
0.16a
64.4
aSignificant at p < 0.01 level.
bSignificant at p < 0.05 level.
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
CONSUMER
SATISFACTION/DISSATISFACTION
cases satisfaction/dissatisfaction is significantly correlated with repeat purchase behavior (p < 0.01). Although these correlations are generally poor in magnitude they are stable across all the studied product classes
(r is between 0.18 and 0.22).
In addition, the simultaneous analysis of all three variables (column 3) shows that the relative contribution of
satisfaction/dissatisfaction to the function is low in comparison with that of the revised intention level. This pattern is consistent across all product classes except macaroni, for which the relative importance of satisfaction
is the highest among the three cognitions.
Although significant differences appear for most comparisons between the groups (column 5), indicating that
switchers and repurchasers differ in their cognitive process, the overall strength of the three investigated cognitions in predicting future purchase behavior appears to
be low. This conclusion is based on the high scores of
Wilks' lambda(column 4) and the low percentageof cases
correctly classified (column 6).
Extention to a Multiple Repurchase Framework
We have focused on a model of purchase-repurchase
behavior pertaining to two consecutive periods. An interesting issue is the change in satisfaction levels as a
function of consumers' experience with the consumed
brand. In this section we extend the model to encompass
additional consecutive purchase periods. The question
addressed is: "To what extent are brand loyal consumers
likely to switch brands?" That is, by knowing that the
repurchaser group comprises consumers who repurchased the same brand several times successively, can
we improve the predictions about repurchase behavior
provided by the three-stage model?
Equation 9 specifies the extended model in which PE
(denoting prior experience) is believed to affect the intensities of the investigated cognitions, namely, lagged
intentions, satisfaction, and revised intention to repurchase. The general extended model is
(9)
P,+ = f(It, SAT,,It-i, PE).
More specifically, in the extended model we add two
consecutive periods4 to examine an experience record of
at least three repeat purchase occurrences. Equation 10
is the model tested.
(10)
Pt+l = f(I,
SAT, P,, Itl,
P,_-, P,-2).
Consumer cognitions, mediated by additional experience derived from consecutive repurchasebehavior, were
expected to result in greater discriminating power between repurchase and switching in comparison with the
first analysis. In addition, because the repurchasergroup
4The decision to add two periods is a compromise between including more previous purchase experience and a sufficient representation
of respondents in the multivariate analysis.
401
(in the previous discriminant analysis) was replaced by
brand loyal consumers, the relative importance of satisfaction in predicting repurchase behavior was expected
to decrease and the opposite observation was expected
for intention levels. The assumption that adaptation levels stabilize over time underlies this proposition.
A computed variable called LOY to denote the number
of successive repurchases of the same brand was computed on the basis of the lagged values of the previously
defined CH variable (i.e., switching versus repurchase
behavior). Consumers who repurchased the same brand
at least two consecutive times constituted the first discriminantgroup whereas consumers' overt switching behavior was the criterion for generating the second group.
Table 6 summarizes the results of a discriminant analysis conducted to test the effects of experience and loyalty on consumer satisfaction/dissatisfaction and overt
repurchasebehavior. A comparison of this analysis (Table 6) with the first discriminant analysis (Table 5) reveals certain patterns. First, predictions of repurchase
behavior are improved when experience with consecutive purchase of the same brand is taken into consideration. This finding is derived by comparing columns 4
and 6 of both tables. Both the decrease in Wilks' lamba
and the increase in the correctly classified cases demonstrate that such a pattern is consistent across the studied products. The findings indicate that the low probability of brand loyal consumers switching brands is
influencedby higher satisfactionand, consequently, higher
intentions to repurchasethe brand. However, despite this
improvement the overall predictions of these cognitions
in repurchase decisions are relatively weak.
A second pattern is indicated by changes in the relative importance of the three investigated cognitions. On
the aggregate level, the relative contribution of satisfaction and revised intention to the overall function decreases (as shown in column 2 in both tables) and the
relative contribution of lagged intentions increases for
brand loyal behavior. Whereas the relative importance
of revised intentions to repurchase is the highest, as posited in the assumption that revised intention is the last
cognitive stage prior to purchase decisions, it is not enhanced when brand loyal behavior is investigated. However, a comparison of lagged intention contributions between the two analyses reveals that this cognition's relative
importance is increasing across four of the five investigated product classes. This finding reinforces the conclusion that intentions to repurchase increase and then
gradually stabilize as consumers repeat their purchase of
the same brand.
DISCUSSION
Our findings strongly support the role of satisfaction
in mediating revised intentions and overt behavior. Our
study differs from much of the past research in that it is
modified for repetitive brand purchases. However, the
patterns found support previous theoretical approaches
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
402
JOURNAL
OF MARKETING
NOVEMBER
1983
RESEARCH,
Table 6
THEEFFECTS
WITHBRANDON BRANDREPURCHASE
OF EXPERIENCE
BEHAVIOR
THROUGH
COGNITIONFORMATION:
A COMPARISON
BETWEEN
BRANDLOYAL
AND SWITCHING
BEHAVIOR
Aggregate
1,
SAT
,I-i
(3)
Standardized
function
coefficients
0.727
0.096
0.336
Margarine
/,
SAT
,_-i
0.638
0.351
0.194
0.892
0.909
0.970
35.1
29.2
9.0
P < 0.001
P < 0.001
P < 0.01
Coffee
/,
SAT
1,_,
0.611
0.230
0.306
0.896
0.908
0.937
38.4
33.4
22.2
P < 0.001
P < 0.001
P < 0.001
Toilet
tissues
1,
SAT
1,,-
0.784
-0.065
0.444
0.870
0.919
0.912
55.6
32.6
35.9
P < 0.001
P < 0.001
P < 0.001
Paper
towels
/,
SAT
1/,1
0.801
-0.109
0.444
0.802
0.898
0.860
89.7
41.3
59.0
P < 0.001
P < 0.001
P < 0.001
Macaroni
/,
SAT
1,-1
0.422
0.477
0.232
0.923
0.925
0.950
30.2
29.4
19.3
P < 0.001
P < 0.001
P < 0.001
(1)
Product
(2)
(1)
Variable
(5)
(4)
Wilks'
lambda
0.889
0.915
0.929
Uvate
Univariate
F-ratio
F-cases
Level
Significance
238.7
P < 0.001
160.3
P < 0.001
131.1
P < 0.001
(6)
Percent of
correctly
classified
(7)
Correlation
with P,+1
0.35'
0.29a
0.26'
69.2
0.35'
0.30a
0.26a
67.0
0.32a
0.30a
0.25a
65.6
0.36a
0.28a
0.30"
71.9
0.44a
0.32a
0.37"
73.1
0.28'
0.27a
0.22'
66.0
'Significant at p < 0.01 level.
addressing the cognitive process underlying purchase behavior. Our study, which examines the cognitive process
for more than a single purchase activity, supports an extension of the role of satisfaction in purchase behavior
from a static to a longitudinal perspective. The upward
and downward shifts of satisfaction, which result in
changes of the adaptation level in parallel directions and
the dominance of the I,-i -- SAT-> I, relationship, demonstrate the importance of satisfaction/dissatisfaction in
explaining the behavior of repeat purchasers and brand
switchers.
Although the changes in the adaptation level confirm
the expected shifts, those changes in most cases are not
statistically significant. This finding is consistent with
Oliver's (1980b) observation that "the adaptation level
effect is remarkably resistant to extinction." The statement can be generalized to the three-stage analysis. Although satisfaction affects the intention level, the shifts
in the adaptationlevel for each group are not substantial.
It is important to note, however, that the difference between the two groups' intention levels increased significantly between the two periods. This finding indicates
that overall satisfaction has a meaningful role in determining the revised intention to repurchase.
As one might expect, the mean revised intentions are
significantly different for consumers who eventually repeated the purchase of the same brand and those who
switched. Likewise, the satisfaction level is a significant
explanation of intention formation. The mean lagged intention to purchase the brand in the previous period is
also different for the two groups. This finding supports
the proposition that the underlying beliefs which give
rise to the formation of postpurchase cognitions are internalized to the extent that attitude, or perhaps intention, persists over a long period of time (Oliver 1980b).
Although lower in magnitude than the difference in the
revised intention, a significant difference in lagged intentions is found in four of five product classes between
consumers who made a decision to repurchase and those
who switched in the following period. This finding indicates that the repeat purchase group comprises mainly
brandloyal consumers. It is congruentwith Day's (1969)
statement that there is a difference between intentional
loyalty and spurious loyalty associated with consistent
repurchaseof one brand.
When the population was split on the basis of whether
or not the brand was repurchased, an asymmetrical relationship among the three variables was revealed for the
two groups. The direct relationships between the preand postpurchase intention levels are stronger for brand
loyal consumers than for brand switchers. This observation was expected because the beliefs about the brand
are more stable and internalized for brand loyal consumers. However, the findings suggest that switcher be-
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
403
CONSUMER
SATISFACTION/DISSATISFACTION
liefs are more sensitive to the dissatisfaction from consuming the brand.
Despite the empirical support for the cognitive process
underlying repurchasebehavior, the strength of the studied cognitions in directly predicting repurchase or
switching behavior is not substantial. Only when consumers' prior experience with the brand was incorporated into the model were predictions somewhat improved. This finding suggests that other situationalfactors,
such as coupons and sales, may influence purchase decisions despite a high satisfaction with the currently consumed brand.
Satisfaction and intention are found to increase as the
loyalty to the brand increases (when brand loyalty is
measuredin a numberof successive purchasesof the same
brand). However, the relative importance of satisfaction
in predicting repurchase appears to decrease as loyalty
increases. Thus, it is likely that a certain threshold of
satisfaction must be met to lead to a repeat purchase of
the brand. Moreover, the longer the sequence of repeat
purchases, the more the experience with the brand accounts for repurchase behavior. One managerial implication may be that marketing efforts appealing to consumer satisfaction with the brand should be expended in
the early stages of consumer brand experience. In later
stages of consumers' experience with the brand a reasonable level of satisfaction should be maintained,so that
long-run intention levels do not decrease.
A second managerial implication is that reports indicating consumers are "extremelysatisfied" with the brand
followed by indications of high intention to repurchase
do not necessarily indicate that switching behavior will
be insignificant. Although the cognitive process is important in predicting repurchase behavior, situational
factors must be incorporated into future models to improve predictions;
Finally, an implication for marketing research is that
because adaptation levels tend to stabilize as brand loyalty increases, both satisfaction and intention should be
measured for those segments (or products) primarily
characterized by low brand loyalty. For segments (or
products) characterized primarily by brand loyal consumers, intention levels may be sufficient to represent
the cognitive influence on repurchase behavior.
LIMITATIONSAND FUTURE DIRECTIONS
Our study tested an abbreviated cognitive model. The
goal was to test in a field study situation the dynamic
aspect of the role of satisfaction in revising repurchase
intentions for several consecutive purchase activities. A
limitation of the study is that it did not simultaneously
test all the variables involved in the dynamic cognitive
model proposed by Oliver (1980b) because of theoretical
and practical issues.
The study was based on panel data. Most consumer
satisfaction studies are conducted in experimental settings or are field studies investigating a single purchase
event. Although a panel study using reported purchase
data might be considered more reliable, it could be biased
if panel members report a higher satisfaction or intention
level to justify their repeat purchases. However, because
the study data were not based on recall of previous periods and the time interval between questionnaires was
two weeks, we believe such potential bias was minimized.
In addition, frequently purchased goods were the focus of study. Although a relatively large sample of frequently purchased products was included, further research is needed to examine other products, such as high
involvementgoods. In relationto low involvement goods,
Day (1977b) argued that the consumer may not make
any evaluation for many simple products. Swan and Trawick (1979) found support for the contention that the
extent of product evaluation may depend on the product
class, each product class representing a different level of
involvement.
Despite this criticism, we believe that measuring satisfaction may be appropriate in the context of longitudinal studies even for low involvement products. The
first reason for this contention relates to the timing of
measurement. Many of the past investigations reported
in the consumer satisfaction literature are based on definitions that do not consider the importance of measurement timing. Only recently has Oliver (1981) suggested
a definition of satisfaction which stipulates that satisfaction rapidly decays into attitude. According to this view,
a measure of immediate past usage satisfaction would
yield the highest construct validity of satisfaction measurement. Because of the frequency of reporting used in
our study, the measurement timing limitation was probably minimized.
The second reason for believing that measuring satisfaction may be appropriatefor low involvement goods
relates to the importance of satisfaction in repurchase decisions. Although some consumers may not make any
evaluation when satisfaction is measured as a dependent
variable, we believe they do think of past satisfaction
prior to a repurchase decision. Thus, though consumers
might often respond, "I don't think about it," when they
do not have to confront a repurchase decision, the question, "How satisfied are you?" may be appropriate in a
repurchase decision situation.
The order of successive purchases of a brand within
the product class was used as the measure of brand loyalty in our study. In reality, however, consumers might
be loyal to several brands concurrently. Different measures of brandloyalty might lead to results different from
ours.
Finally, a broadening of the consumer satisfaction literatureto longitudinal designs should be continued. An
extension of the design to encompass a longer sequence
of successive purchase behavior would permit further
brand experience to be taken into account. In addition,
including situational factors, such as company reputation
and sales, would probably improve predictions of brand
repurchase behavior.
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
404
JOURNALOF MARKETINGRESEARCH,NOVEMBER1983
REFERENCES
Anderson, R. (1973), "Consumer Dissatisfaction: The Effect
of Disconfirmed Expectancy on Perceived Product Performance," Journal of Marketing Research, 10 (February), 3844.
Bonfield, E. (1974), "Attitude, Social Influence, Personal
Norms, and Intention Interaction as Related to Brand Purchase Behavior," Journal of Marketing Research, 11 (November), 379-89.
Cardozo, R. (1965), "An Experimental Study of Consumer
Effort, Expectations and Satisfaction," Journal of Marketing Research, 2 (August), 244-9.
Czepiel, J. A. and L. T. Rosenberg (1976), "Consumer Satisfaction: Toward an Integrative Framework," Proceedings
of the Southern Marketing Association, 169-71.
Day, G. S. (1969), "A Two-Dimensional Concept of Brand
Loyalty," Journal of Advertising Research, 9 (Number 3),
29-35.
Day, R. L. (1977a), "Alternative Definitions and Designs for
Measuring Consumer Satisfaction," in The Conceptualization and Measurement of Consumer Satisfaction and Dissatisfaction, H. K. Hunt, ed. Cambridge, MA: Marketing
Science Institute.
(1977b), "Toward a Process Model of Consumer Satisfaction," in The Conceptualization and Measurement of
Consumer Satisfaction and Dissatisfaction, H. K. Hunt, ed.
Cambridge, MA: Marketing Science Institute, Report No.
77-103, 153-83.
Deutcher, I. (1966), "Words and Deeds: Social Science and
Social Policy," Social Problems, 13, 237-47.
Duncan, O. D. (1966), "Path Analysis: Sociological Examples," American Journal of Sociology, 72, 1-16.
Harrell, G. and P. Bennett (1974), "An Evaluation of the Expectancy/Value Model of Attitude Measurement for Physician Prescribing Behavior," Journal of Marketing Research, 11 (August), 269-78.
Helson, H. (1964), Adaptation Level Theory: An Experimental
and Systematic Approach to Behavior. New York: Harper
and Row.
Hirsci, T. and H. C. Selvin (1967), Delinquency Research:
An Appraisal of Analytical Methods. New York: Free Press.
Howard, J. A. (1974), "The Structure of Buyer Behavior," in
Consumer Behavior: Theory and Application, John V. Farley, John A. Howard, and L. Winston Ring, eds. Boston:
Allyn & Bacon, 9-32.
Hull, H. and Nie, N. H. (1979), SPSS Update. New York:
McGraw-Hill Book Company, 111-14.
Ilgen, D. R. (1971), "Satisfaction with Performance as a
Function of the Initial Level of Expected Performance and
the Deviation from Expectations," Organizational Behavior
and Human Performance, 6, 345-61.
and B. W. Hamstra (1972), "PerformanceSatisfaction
as a Function of the Difference Between Expected and Reported Performance at Five Levels of Reported Performance," Organizational Behavior and Human Performance, 7, 359-70.
Kimberly, J. R. (1976), "Issues in the Design of Longitudinal
Organizational Research," Sociological Methods and Research, 4, 321-47.
LaTour, S. and N. Peat, (1979), "Conceptual and Methodological Issues in Satisfaction Research," in Advances in
ConsumerResearch, Vol. 6, W. L. Wilkie, ed. Miami: Association for Consumer Research. 431-37.
Oliver, R. C. (1977), "Effect of Expectation and Disconfirmation on Postexposure Product Evaluations: An Alternative Interpretation,"Journal of Applied Psychology, 62 (August), 480-6.
(1980a), "Conceptualization and Measurement of Disconfirmation Perceptions in the Prediction of Consumer Satisfaction," in Proceedings of Fourth Annual Conference on
Consumer Satisfaction, Dissatisfaction, and Complaining
Behavior, H. K. Hunt and R. L. Day, eds. Bloomington:
School of Business, Indiana University.
(1980b), "A Cognitive Model of the Antecedents and
Consequences of Satisfaction Decisions," Journal of Marketing Research, 17 (November), 460-9.
(1981), "Measurement and Evaluation of Satisfaction
Process in Retail Settings," Journal of Retailing, 57 (Fall),
29-31.
Olshavsky, R. W. and J. A. Miller (1972), "Consumer Expectations, Product Performance and Perceived Product
Quality," Journal of Marketing Research, 9 (February), 1921.
Olson, J. C. and P. A. Dover (1979), "Disconfirmation of
Consumer Expectations Through Product Trial," Journal of
Applied Psychology, 64 (April), 179-89.
Russo, E. (1979), "ConsumerSatisfaction/Dissatisfaction: An
Outsider's View," in Advances in Consumer Research, Vol.
6, W. Wilkie, ed. Miami: Association for Consumer Research.
Swan, J. E. (1977), "ConsumerSatisfaction with a Retail Store
Related to the Fulfillment of Expectations on an Initial
Shopping Trip," in Consumer Satisfaction, Dissatisfaction
and Complaining Behavior, R. L. Day, ed. Bloomington:
School of Business, Indiana University, 10-17.
and I. F. Trawick (1979), "Triggering Cues and the
Evaluation of Products as Satisfactory or Dissatisfactory,"
in Educators' Conference Proceedings, N. Beckwith et al.,
eds. Chicago: American Marketing Association, 231-4.
and
(1981), "Disconfirmation of Expectations
and Satisfaction with a Retail Service," Journal of Retailing, 57 (Fall), 49-67.
Thibout, J. W. and H. H. Kelley (1959), The Social Psychology of Groups. New York: John Wiley & Sons, Inc.
Warshaw, P. R. (1980), "Predicting Purchase and Other Behaviors from General and Contextually Specific Intentions,"
Journal of Marketing Research, 17 (February), 26-33.
Weaver, D. and P. Brickman (1974), "Expectancy, Feedback
and Disconfirmation as Independent Factors in Outcome
Satisfaction," Journal of Personality and Social Psychology, 30 (3), 420-8.
Westbrook, R. A. and J. W. Newman (1978), "An Analysis
of Shopper Dissatisfaction for Major Household Appliances," Journal of Marketing Research, 15 (August), 45666.
and R. L. Oliver (1980), "Developing Better Measures of Consumer Satisfaction: Some Preliminary Results,"
in Advances in Consumer Research, K. B. Monroe, ed. Ann
Arbor: Association for Consumer Research, 94-9.
This content downloaded from 132.64.30.218 on Sat, 12 Oct 2013 07:33:21 AM
All use subject to JSTOR Terms and Conditions
Download