"AFTER I HAD MADE THE DECISION, I...:"
TOWARD A SCALE TO MEASURE COGNITIVE DISSONANCE
Douglas R. Hausknecht, University of Akron
“ilan C. Sweeney, University of Western Australia
Geoffrey N. Soutar, ith Cowan University
‘Lester W. Johnson, Monash University
ABSTRACT
The objective of the preset research study is
to begin the development ofa sale to measure
cognitive dissonance that arses from fee cesce in
consumer purchase decisions. This represents an
Advance in de measurement of cognitive
fissonance. Cognitive, emotional and behavioral
aspects ofthe dissonance consruc are slated and
desenbed, as are relaionshins with other post
purchase variables, In particlar, i is noted that
the reduction of copetive dissonance is 2
recesiry condition for the cccurrence of
fausietion, Multiple tems, developed from focus
groups, were evaluate by a series of independent
judges (consumer behavior resarchers). A oul of
Si measurement items are offered for further
refinement,
INTRODUCTION
One hopes tht the contrition, vliaion
‘and dissemination of comprehensive. dissonance
‘cals will be forthcoming." (Over 1997, p- 261)
‘Wiens plea, Oliver conciadd hie chap length
feview of ssonance research as part of
landmark treatise on consumer satisfaction, Since
Leon Festinger coined. the term “cognitive
dissonance" in 1957, the concept has been
fnterprete, debate, ad re-iterpreted with some
frequency and ferocity. Many ofthe ealy ential
Iswoes, however, romain at the center of
disagreements. Underying these conned debates
fe attempts to ft the se concept into 2 act of
‘volving theories and paradigns
‘Vaous concepaualization: have been used in
sempts 10 identify the relationships. between
ognitve dissonance and consumer satisfaction!
Aissasstction [CS/D]. Some authors ive
discussed bow dissonance genealy fits theories of
consumer behavior (e-g., Cummings and
Venkatesan 1976; Schewe 1973), while others
hve made explicit disinsions among. the
consis and’ their relationships (8.
Montgomery and Barnes 1993; Olver 1997), To
de, however, no one has sted convincingly the
‘conflict tht have been rated.
Whereas the siisfaton construct has been
widely discussed, and measures and models
developed ound it (@g., Churhil and
‘Supremant 1982; Johnson and Fornell 1991;
Oliver 1980), fewer measurement studies have
examined the concep of dissonance. Some early
Aisonance studies have been ertized a Lapping
related consracis, such as anxiety, rater than
Gissonance itself” (Cummings and) Venkatesan
1976). Other earlier studies did not measure
dissonance but, rather, inferred the occurence of
tlissonance from evidence of dissonance reducing
behaviors. For example, Engl (1963) asessed the
tention paid to advertising of a brand after
purchasing that brand (Bagel 1963), while
Lsuito and Perloff (1967) measured atitude
changes towards seleced and nonselected
products. Both are examples of how people strive
Toachineecrmerence song atinides, knowledge
and behavior Schewe 1973)
Despite four decades of discussion, an agreed
‘upon and measurable construct continues to clude
fal selmi. Tt bas been suggested recently
that, "disonance, a least as presealy measured,
may noc have discriminant validity when compared
to ofier post purchase consructs™ (Sweeney.
Soutar and Johnson 1996, p. 138), Over (1997)
discusses the relationships among these constructs
‘with an acknowledgment ofthe nee for improved
‘measures consistent wih theoeetcal modes
"THE DISSONANCE CONCEPT
Festinger’ early explanation of the dissonance
construct doesnot ientify clearly whether
“alssonance" is cognitive or emotional. The
cognitive view is supported by his definition tht
“he obverse of one element follows from the
other” (Festinger 1957, p, 261). Yet, he seems
flo to have intended’ an emotional
Cconcepmalizaton, suggesting tat “for” some20. Journal of Consumer Satisfaction, Dissatisfaction and Complaining Behave
people, dissonance is an enremely painful and
Intolerable thing" (Festinger 1957, p. 266). In a
recent social psychology treatment, Gilovih,
Medvec and Chen (1995, appest (© teat
dissonance ia a tration cognitive. sense.
Conversely, Harmon-Jones el, (1996) sem t0
‘be more concerned with te emotions aroused by
‘he “aversive consequences” of an individuals
action. Infact, they subseibe to the theory that it
is these consequences that are necessary and
sufficient to create the emotion, rather than mere
cognitions. However, they vere working within
the" induced-compitance” paradigm and thet
conclusions may not have diet relevance wo the
“fee-choice” paradigm thal is most often of
intrest to consumer reseacbes
(Oliver (1997) revived thr fre-choice version
‘of dissonance, which he characterized as having
Ini dormint for some time. in Oliver's model of
satisfaction, the dissonance soncept is stretched
‘across wo-thids of the satisfaction process
‘Ocginting in a presprchase phat, the construct
is laeled “apprehension,” These seme copaitions
and felings mutate into er dissonance ater the
‘decison is made. With ue and experience,
lissonance dissipates and yes to dissatisfaction
(as can be seen in Oliver's igure 1-3, 1997, p
2)
‘wile Ouver (1997) argues tat dissonance
‘eeurs at various stages of the consumption
roots, it is generally reegnized a a post
ecisional, but pre-use phenomenon (e.
Festinger, 1957: Insko and Schopler, 1972)
Indeed, Otiver (1997, p. 24), in 9 subsequent
‘esi, views dissonance a resulting “from 2
personal decision or acon” He termed. this
hartower window the "Ganma™ phase (Ohver
1997, p. 242). The concept of dissonance
aeessed in the present research best fis tis
period that immediately fellows the purchase
‘Secision but precedes use or experience with the
result of the purchase decision.
‘This rstionsip is made explicit in Figure 1,
in which the horizontal axe represents changes
‘over a purchase and consumption proces, but docs
ot presume causality. Dissmance constructs aise
‘only after the decison ie made and in response to
‘a mimber of faeors,
Tes important to note tit the presentation of
‘atisfacton models and dissonance models isnot
 
meant to pre-suppose tat every purchase results in
the arousal of elter or oth of these processes
For example, ic has been argued that satisfaction
andor dlsatisfacion may not aise in. low
Involvement situations (Hauskneche 1988; Oliver
1997). Using Oliver's (1997, p. 13) conepe of
satisfaction a 2 flillment respons, its apparent
‘hat ether te cognitive or emotional components
‘may not be aroused in given situations
Bell (1967) suggested long ago that some
individuals simply may not experince dissonance
‘The literature has established well the necessary
conditions for dissonance arousal (Oliver 1997),
although there is sme confusion caused by mixing
paradigms (Le, forced compliance versus free
choice). A physiological state, arousal, hus been
‘suggested as another necessary condition (Eot
and Devine 1994), but no one has demonstrated
‘ficient conditions to force the process, The
forced compliance paradigm has been used mest
fofien in diasonance studies 10 ensure having
something to measure, but has been assailed as not
Tikely, or even rare, in consumption situations
‘Cummings sad Venkatesan, 1976; Olver 1997).
“The previously mentioned controversy’ ab 1
te tetment of dissonance as cognitive or
‘motional is similar to that which exists ia the
atiudelerature between proponents of the
(parte ate mogel and those wao favor the
attde--affect version. To clarity the ature of
‘he constructs, the present” model “induces
Aiscintions overtime and separates cognitive from
emotional concepis. The frst of these, decision
config, i not usually presented as part of the
consumer decision process. Davidson and Kiesler
(1968). cied Festinger's distinction between
eciion conflict, pre-dcision concept, and
Gissonance, a post-dcision concep. The” same
authors also presenta contrary view, "Janis (1959)
fon the other hand, thinking in terms of ‘confit
‘resolution’, implied that there is litle or 0
Aiscintion between pre-and pst decision behavior
fand that systematic reevaluation occurs both
‘before and afte the decision” (Davidson and
‘Keder 1964, p10). Even if the two concepts are
similar structurally, they ae divided inte by the
act (behavior) of having made 2 decision. In any
as, they describe decision conic asa cognitive
Jmbolance that is resolved by the decison
‘Whatever the relationship of dissonance withVolume 11, 1998
nt
Figure 1
‘Temporal Relations Among Dissonance Concepts
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  
 
pre-decsion variables may be, there is consensus
hac etic dat are tacondnent with © decision
may persist after a purchase. For clarity of
txponton, Figure bels this “decison
fissonance’. If decision conflict exist up 10 the
‘moment of decision, its decision dissonance that
follows immediatly. Insko and Schopler (1972p.
105) reasoned smiley, suggesting that
“dissonance is thus  postecisional conflict.”
Further, they suggested. tat there may be a
“spreading” of the evalstion of the decision
ltematives 25 part of the conflict or the
Gissonnnce. Oshawa (1972, p. 65) agreed with
the temporal positioning of dissonance but labeled
it's only, "an intervening variable: afer &
decison is made and before stempts at dissonance
redveton.”
‘More substantially, tht period in a decision
process hat boen characionzed as represeating 2
ange in confidence (Knoe and Inkster 1968),
‘motivation to solve a puzzle regarding one's ov,
betavir (Greenwald and Ronis 1978) ora wonder
about the wisdom of the decision (Lowe and
 
 
 
Steiner 1968). All ofthese conceps appear 1 be
‘Sevoid of feling, evaluation, emotion. Thur
‘ecision dissonance appeat o be the same kindof
imental enity a8 cognitions apd éecision confit.
‘Cognitive disonance is most -commoaly
‘efi a psychological discomfort ( Carismih
‘and Aronson. 1963, Eliot and Devine 1996;
Festinger 1957), a peychologically uncomfortable
sine (Menasco and Hawkins 1978), linked with
tmxity and uncerainy or doubt ( Monigomery
tint Barnes 1993; Mowen 1995; Oshikawa 1972)
‘or synonymous with the repre or remorse eponed
in falespeople's aneedotes (Insko and Schopler
1972; Mowen 1995). Thus, the fr-year history
of the Iieratue appears 10 have created &
theoretical oxymoron in which an essentially
finotionsl construct tears the burden of
Cognitive" in ie mame, Nevertheless, "to date
there have been no systematic atempts to dietly
mpinially validate the prycholgieal dscomfor:
Component of dissonance” (Elli and. Devine
10994, p. 383) The emotional aspect of dissonance,
represeting the pryeboloical discomfortBa Journal of Consumer Saifction, Dissatisfaction and Complain
eseribed above, stemmed “dissonance ~
‘emotional component” in Figure 1. Subsequent to
‘he arousal of cognitive disonance, dissonance
‘eduction behaviors may arise and these have often
‘been used as surrogate measres for dissonance in
prior work. Finally, satisfacion is assessed pos
Purchase. and postuse, when performance is
Compared 10 expectations. Satisfaction has been