You are on page 1of 11

Journal of Personality and Social Psychology Copyright 1986 by the Am rican Psychological Association, Inc.

1986, Vol. 51, No. 1,55-65 0022-3514/86/S00.7J

Physiological Arousal, Dissonance, and Attitude Change: Evidence for a


Dissonance-Arousal Link and a "Don't Remind Me" Effect

Roger A. Elkin and Michael R. Leippe


Adelphi University

Two experiments replicated and extended research by Croyle and Cooper (1983) indicating that
cognitive dissonance involves physiological arousal. In Experiment 1, subjects wrote counterattitudi-
nal essays under conditions of high or low choice, and, to assess arousal effects owing to effort, with
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

or without a list of arguments provided by the experimenter. In high-choice conditions only, and
This document is copyrighted by the American Psychological Association or one of its allied publishers.

regardless of effort, subjects showed both arousal (heightened galvanic skin response) and attitude
change. Arousal, however, did not decline following attitude change. The more effortful task (no
arguments provided) produced increased arousal but not greater attitude change. In Experiment
2, the opportunity to change one's attitude following a freely chosen counterattitudinal essay was
manipulated. As in Experiment 1, arousal increased following the essay but did not decline following
a postessay attitude change opportunity. When subjects were not given an attitude change opportu-
nity, however, arousal did decline. Thus, dissonance seems to create arousal, but attitude change
sustains rather than reduces the arousal. It is suggested that if dissonance is a drive state, drive
reduction typically may be accomplished through gradual cognitive change or forgetting.

Cognitive dissonance exists when two or more cognitive ele- involves undifFerentiated physiological arousal that, in turn, is
ments (e.g., an attitude and knowledge of one's behavior) are labeled and experienced as a negative (unpleasant) state if the
experienced as logically inconsistent (Festinger, 1957). This in- individual attributes the arousal to his or her attitude-discrep-
consistency is presumed to produce discomfort or tension that ant behavior (Cooper & Fazio, 1984; see, also, Kiesler & Pallak,
the individual is motivated to reduce. Dissonance occurs, for 1976). It has been demonstrated, for example, that subjects ex-
example, when someone engages in attitude-discrepant behav- perimentally induced into a high-dissonance state make more
ior (e.g., writing a counterattitudinal essay) while under the im- dominant responses than control subjects on word recognition
pression that he or she freely chose such behavior. When this (Cottrell, Rajecki, & Smith, 1974) and color-word interference
happens, and the individual also is committed to the attitude- (Pallak & Pittman, 1972) tasks, just as do people who are anx-
discrepant behavior (Brehm & Cohen, 1962) and perceives its ious. The most suggestive indirect evidence that dissonance is
consequences as negative (Cooper & Goethals, 1974), the per- arousing, though, has come from the work of Cooper and his
son reliably shows evidence of dissonance reduction, often in colleagues. Zanna and Cooper (1974) found that subjects did
the form of attitude change in the direction of the behavior. not change their attitude following counterattitudinal behavior
Festinger (1957) originally posited that the attitude change if they believed that a drug (actually a placebo) they had been
is motivated by the psychological discomfort that accompanies given caused arousal. Presumably, these subjects misattributed
cognitive inconsistency. This idea of "motivated attitude dissonance-produced arousal to the drug, thereby negating the
change," however, has been difficult to assess. Indeed, among need to reduce tension through attitude change. Subsequently,
the most serious challenges to dissonance theory have been in- Cooper, Zanna, and Taves (1978) used actual drugs to mimic
terpretations based on dispassionate, self-attributional pro-
cesses (e.g., Bern, 1967). To strengthen support for the classic 1
Evidence of discomfort and alleviation of discomfort will not be
motivational explanation of dissonance effects, it must be dem- construed in this article as supporting dissonance theory over alterna-
onstrated that "psychological discomfort" is present under tive explanations of the effects of counterattitudinal behavior. Recent
high-dissonance conditions and that this discomfort is allevi- self-presentation analyses such as Schlenker's (1982) identity-analytic
ated by attitude change or some other means of dissonance re- model propose that heightened arousal following counterattitudinal be-
duction.1 havior may reflect social anxiety or embarrassment about public viola-
tion of one's identity-relevant standards (Schlenker, 1982, pp. 222-223).
Arousal as a Concomitant of Dissonance It seems reasonable that reduced discomfort following attitude change
There is some evidence for the first of these required condi- could similarly be construed by a self-presentation model as a conse-
tions: Numerous studies suggest that the high-dissonance state quence of the psychic relief that accompanies "setting things straight."
Thus, no "crucial test" seems possible at this time (Tetlock & Manstead,
1985), and the present contribution is offered in the spirit of paradig-
Special thanks to Bob Bedell for his help in conducting Experiment matic progress within the current, "revised" dissonance framework (cf.
1, and to four anonymous reviewers for their helpful comments. Greenwald & Ronis, 1978). To use Schlenker's (1982) description, dis-
Correspondence concerning this article should be addressed to Mi- sonance is used here to define a recognition by the individual that his
chael R. Leippe, Department of Psychology, Adelphi University, Gar- or her attitude-discrepant behavior carries "responsibility-for-negative-
den City, New York 11530. consequences-that-threaten-the-self" (p. 238).

55
56 ROGER A. ELKIN AND MICHAEL R. LEIPPE

the arousal states proposed to exist under high- and low-disso- not previously considered supporting evidence for the opposing
nance conditions. They found that a sedative suppressed atti- point of view. Also, more effort may be exerted if the counterat-
tude change and an amphetamine enhanced attitude change titudinal essay is written under high- relative to low-choice con-
among high-dissonance subjects who believed they had taken a ditions. Involvement may be less when behavior is seen as
placebo. Cooper et al. reasoned that in the amphetamine group, "manded" (low choice), resulting in little effort investment. In
subjects attributed the otherwise inexplicable drug-induced Experiment I , we tested this alternative explanation by manip-
arousal to their inconsistent cognitions thereby increasing the ulating the effortfulness of an essay-writing task independently
drive to restore consistency. In the sedative group, on the other of choice and direction of essay.
hand, arousal associated with counterattitudinal behavior was
suppressed, and subjects felt no "drive" to change their atti-
tudes.
Arousal Reduction as a Consequence of Attitude
Cooper et al.'s results constitute impressive, albeit indirect, Change (Dissonance Reduction)
evidence for a dissonance-arousal link. Of course, the drugs uti-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Dissonance, as noted, has been viewed traditionally as a ten-


lized in their study can only be assumed to have effects similar
This document is copyrighted by the American Psychological Association or one of its allied publishers.

sion state with drive properties. Drive theories of behavior (e.g.,


to those of dissonant behavior. It might be argued instead that
Hull, 1943) postulate that internal needs (e.g., hunger) and
the drugs merely influenced information processing ability.
goals (e.g., solution of a complex cognitive problem) energize
Drugs may affect the extent to which people expend energy on
behavior (increase drive) directed toward the satisfaction of
sorting out inconsistencies and on seeking information about
those needs or goals and that once the need or goal is satisfied,
the environment.
behavior returns to "less energetic" levels. If autonomic arousal
Perhaps a more fundamental shortcoming of the misattribu-
is a correlate of drive (e.g., Berlyne, 1965), behavior that re-
tion studies is that they did not directly observe the arousal ac-
duces underlying need should be followed by arousal reduction.
tually produced by attitude-discrepant behavior. Croyle and
This view is supported by research on goal-directed cognitive
Cooper (1983) recently reported two experiments designed for
activity. Blatt (1961), for example, found increases in physiolog-
this purpose. In the first experiment, the traditional attitude
ical arousal among individuals while they pondered a complex
change following counterattitudinal essay writing occurred for
problem and abrupt decreases once they discovered the solu-
high- but not low-choice subjects. In a second experiment,
tion. Similarly, tension that occurs among people who are inter-
Croyle and Cooper used exactly the same procedure, but mea-
rupted in their quest to complete a task (the "Zeigarnik effect")
sured subjects' arousal (skin conductance response; SCR) be-
decreases following the opportunity to complete the task (Man-
fore and following the essay-writing task. Greater increases in
dler, 1975).
arousal were observed among subjects in the high-choice condi-
If dissonance creates drivelike tension, a similar reduction in
tion, leading the authors to conclude that dissonance caused
arousal should follow behavior that restores consonance. Ac-
heightened autonomic activity. Interestingly, however, no atti-
cordingly, some reduction in any dissonance-produced arousal
tude change differences were found in the second study, an out-
should be evident after dissonance is reduced through attitude
come Croyle and Cooper argued was because of high-choice
change or some other means. To our knowledge, no direct dem-
subjects' misattribution of their arousal to the tension produc-
onstration of this derivation of dissonance theory exists. There
ing effects of the physiological recording apparatus. Thus, al-
is indirect evidence, however. Pallak and Pittman (1972) found
though it was shown that a reliable dissonance manipulation
that like other drive states, dissonance produced by an attitude-
produces arousal, it was not demonstrated that attitude change
discrepant, freely chosen decision was associated with a high
and arousal are common, co-occurring effects of dissonance.
frequency of dominant responses that inhibited performance
One purpose of the present research was to look for the co-
on a difficult verbal task and improved it on an easy verbal task.
occurrence of arousal and attitude change among subjects expe-
However, when the experimenter provided high-dissonance
riencing cognitive dissonance. To do so, we used a procedure
subjects with a strong justification for their attitude-discrepant
designed to minimize the possibility of misattribution of any
decision, these effects disappeared. This implies that once disso-
dissonance-produced arousal to our physiological recording
nance is reduced, drive and its associated tension is reduced.
equipment.
We sought more direct evidence for this relation by tracking the
The lack of within-experiments correspondence of attitude
course physiological arousal takes following dissonance-reduc-
change and arousal points up the complexity of the measure-
ing attitude change.
ment and meaning of arousal effects. Electrodermal activity as
measured by Croyle and Cooper may result from many specific
neural systems and therefore from many types of activities (So- Experiment 1
kolov, 1963). Cacioppo and Petty (1979), for example, found
that heightened arousal evidenced by five different physiological In Experiment 1, we modeled Croyle and Cooper's strategy
measures accompanied subjects' anticipatory counterarguing of comparing arousal changes of subjects in the low- and high-
of an expected attitude-discrepant message. This suggests that dissonance conditions of a well-established forced compliance
the effort involved in mental activity may itself be arousing. manipulation. Several changes in the Croyle and Cooper proce-
Conceivably, the effort-arousal link rather than a dissonance- dure were made, however. Galvanic skin response (GSR) was
arousal association could account for Croyle and Cooper's data. our measure of arousal. GSR is a measure of skin resistance,
Writing a counterattitudinal essay normally may require more the reciprocal of the skin conductance response measured by
effort than writing a proattitudinal essay because subjects have Croyle and Cooper. Research (e.g., Sourek, 1965; Wang, 1957,
AROUSAL AND DISSONANCE 57
1958) suggests that GSR reliably correlates with nervous sys- tion and seldom found significant attitude change (relative to a
tem activity known to be indicative of tension and anxiety.2 control group) among low-choice subjects.
To acclimate subjects to the physiological recording proce- In sum, the following predictions were made:
dure, four filler tasks were given subjects prior to baseline re- 1. Relative to low-dissonance subjects, high-dissonance sub-
cording of GSR. This contrasts with two filler tasks used by jects will show (a) greater increases in arousal and (b) more atti-
Croyle and Cooper. In addition, unlike in Croyle and Cooper's tude change after they write a counterattitudinal essay.
procedure, no statements implying the possibility of apparatus- 2. Relative to those in the low-effortfulness condition, sub-
related tension were made by the experimenter. These proce- jects in the high-effortfulness condition will show (a) greater in-
dures were followed to reduce the likelihood of misattribution creases in arousal but (b) equivalent attitude change after they
of dissonance arousal to the recording procedures. We also took write a counterattitudinal essay.
a postattitude change measure of arousal in addition to pre- 3. Following an opportunity to reduce their dissonance
and postdissonance manipulation measures to assess whether through attitude change, high-dissonance subjects will experi-
attitude change serves to reduce arousal. ence a decline in arousal.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Finally, to test the possibility that the cognitive effort in-


This document is copyrighted by the American Psychological Association or one of its allied publishers.

volved in high-choice counterattitudinal essay writing, and not


Method
dissonance per se, is responsible for heightened arousal, Experi-
ment 1 included a manipulation of essay task effortfulness that Subjects, Design, and Apparatus
was crossed with the dissonance variable of high versus low
choice. This involved varying the availability to subjects of in- Twelve male and 28 female undergraduates, distributed by sex in rela-
tively equal proportions across conditions, participated to partially ful-
formation to be used in their essays. Half the subjects were given
fill the requirements of 100-level psychology courses. At the time of
the arguments that they were to write into an essay (low effort-
recruitment, they were told that physiological measurements would be
fulness), whereas the other half were required to generate their taken during several tasks. The subjects were run individually. The de-
own arguments (high effortfulness). We predicted that subjects sign was a 2 X 2 (High vs. Low Choice X Essay Arguments Available vs.
in the high-effortfulness condition would become more aroused Not Available) between-subjects factorial (« = 10 subjects/cell) in
than those in the low-effortfulness condition, but would not en- which, as described later, repeated measures were taken on attitude (2
gage in greater attitude change. When there is no dissonance occasions) and GSR (3 occasions).
A Biofeedback Instruments G.S.R. Model P-976 preamplifier was
associated with writing the essay, subjects should correctly attri-
used to measure skin resistance. A current of 10 microamps was passed
bute effort-produced arousal to task difficulty. Only when there
through copper tipped electrodes. All measures were taken from a digi-
is dissonance should arousal be tied to dissonance-reducing
tal readout. The P-976 model requires no skin preparation and no elec-
drive. Although it is possible that effort-produced arousal will trolyte. Measurement drift is no greater than 1 ohm per 15-min mea-
combine with dissonance-produced arousal among high-disso- surement period.3
nance subjects to yield a particularly strong dissonance-reduc-
tion motive, the fact remains that following dissonance theory,
1
arousal and attitude change should only co-occur when disso- Of course, it is also the case that electrodermal activity is sensitive
nance is experienced. to factors (e.g., attentional changes) that other physiological measures
It should be noted that the effortfulness manipulation is not (e.g., heart rate) may not be measurably influenced by. We relied on the
logic of experimentation and random assignment to control for addi-
expected to itself influence dissonance. Although some studies
tional influences on electrodermal activity across conditions. We should
have found that attitudes toward the outcome of a task are more
also note that physiological researchers believe skin conductance is a
positive the more effortful the task (e.g., Aronson & Mills, 1959; more preferable measure of electrodermal activity than our measure
Zimbardo, 1965), other studies have failed to observe this effort of skin resistance (see Venables & Christie, 1980). Conductance more
justification effect (cf. Insko, 1967). More importantly, effort precisely reflects what is occurring on the surface of the skin. Our pri-
justification studies differ from the present research in that in mary purpose, however, was not a precise measure of the sweat re-
the former, variation in effortfulness is the dissonance manipu- sponse. Rather, we looked for evidence of underlying physiological
change and whether such change is systematically related to manipu-
lation. In the present case, dissonance is initially created by
lated psychological variables. Resistance serves this purpose (see
freely choosing to perform attitude-discrepant behavior. Stud-
Sourek, 1965; Wang, 1957, 1958). A related, and more relevant short-
ies that used manipulations resembling our effortfulness vari-
coming of resistance relative to conductance is that the size of a resis-
able (e.g., reading versus improvising a counterattitudinal mes- tance response to a stimulus (phasic activity) is more highly correlated
sage) have observed no effect of these manipulations on attitude with existing (tonic) resistance activity. Known as the "law of initial
change among subjects in high-choice, high-commitment con- values," this means that the meaning (in terms of change in physiologi-
ditions (Rabbie, Brehm, & Cohen, 1959; Zimbardo, 1965). As cal activity) of a resistance response varies as a function of prestimulus
Carlsmith and Freedman (1968) have suggested, task details in- activity levels. The law of initial values mandates that premanipulation
levels of tonic activity should be within the same range in all experimen-
troduced following a dissonance-arousing decision should not
tal conditions in order for phasic changes to be comparable. Checks on
affect the magnitude of dissonance unless they could have been
this requirement in both experiments, as reported in the Results sec-
foreseen before the decision (see also Cooper & Goethals, 1974).
tions, found no evidence of a confounding because of the law of initial
Finally, in low-choice conditions, we would not expect differen- values. Accordingly, changes in skin resistance are a valid index of
tial dissonance or attitude change across effortfulness levels, be- changes in arousal in the present context.
3
cause most of the great many counterattitudinal advocacy stud- Personal communication with Philip Brotman of Biofeedback In-
ies have used the high effortfulness instance of our manipula- struments Co., Westchester, NY, July 1985.
58 ROGER A. ELKJN AND MICHAEL R. LEIPPE

Procedure and Dependent Measures forceful arguments in support of parking fees for students. Your
arguments will be sent directly to the committee for evaluation.
Subjects were seated at a desk in a temperature and humidity con-
trolled room and given a consent form to sign that gave a general de- Subjects in the high-choice conditions were further given a release
scription of the tasks to be performed and explained that "harmless and form to sign that stated:
painless recordings" would be taken from the surface of their skin. After
I realize what is involved in this task and that I am performing it
this form was signed, the experimenter stressed the innocuous nature of my own free will. The essay I write will be sent to a committee
of the recording apparatus, stating that absolutely no electric shock on campus that is intending to make decisions on this issue based
would be run through the wires. The electrodes were attached to the on the arguments it receives from me and other students. I am
distal phalanx of the index and second fingers of the subject's nonpre- aware that I may stop participating in this survey now without loss
ferred hand. It was explained that the study was concerned with what of participation credit. . .
naturally occurs on the surface of the skin as people engage in everyday
tasks. All of five subsequently administered tasks were then performed All subjects signed and dated the form and checked a space indicating
while the experimenter was out of the room. release of their essay.
Low-choice condition subjects were not given the release form but
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

The first task was a "Student Attitudes Questionnaire" containing 10


This document is copyrighted by the American Psychological Association or one of its allied publishers.

campus issues for which subjects expressed their attitudes on 31-point were given slightly different information about the task:
scales ranging from disagree to agree. Issue 5 was the focal issue and
In the past it has been shown that a good way of doing this is simply
read as follows: "All students should pay a parking fee of $15 per
to ask people, no matter how they feel personally, to list arguments
semester."4 This served as the attitude premeasure. When subjects had
from a particular side of the issue. Therefore, I have assigned you
completed the entire questionnaire, the experimenter returned, intro- to write strong, forceful arguments in support of parking fees for
duced the first filler task (an anagrams task), and left the room for 3 students. \bur arguments will be sent directly to the committee for
min. The second filler task (a math problem) was introduced next, and evaluation.
subjects were given 3 min to grapple with it.
Subjects were then asked whether they felt relaxed. It was hoped that All subjects were given 5 min to write the essay. Subjects then placed
a response to such a question would have the force of a commitment their essays in an envelope addressed to the "Bursar's Committee on
and would help avoid misattribution. All subjects reported that they
were indeed relaxed. The experimenter then called to his assistant in Availability manipulation. Half of the subjects in the experiment
the next room to inquire as to whether this corresponded to the readings were provided with five arguments (arguments-available condition).
being recorded. The assistant came to the room and said, "Yes, accord- They were told they could elaborate on these but that they should use
ing to the machine you seem relaxed." This statement was made to re- them exclusively. The other half of the subjects were not provided with
duce nervousness about the experimental setting by suggesting that these arguments (arguments-unavailable condition), but, instead, gen-
none existed. No mention was made during any phase of this study erated their own.
about anxiety or tension produced by the machine. After the essay-writing task another 3-min rest period ensued during
The subjects next spent 3 min each on a third (memory) and fourth which GSR was recorded. This postessay measurement procedure was
(brightness magnitude estimation) filler task. the same as during the baseline measurement. Afterward, subjects were
After this, there was a 3-min rest period during which baseline GSR presented with a single 31-point item identical to the premeasure on
data were taken. During the 3 min, 18 recordings (one every 10s) were parking fees. It was explained that the committee was interested in their
taken from a digital readout. opinion on this issue and that this item was to accompany their essay in
At the end of this 3-min period, the experimenter entered the room the envelope. When subjects had placed this measure in the envelope, a
and read the following: final 3-min rest period followed during which postattitude change op-
portunity GSR was recorded.
In exchange for these research facilities, the psychology depart- Manipulation checks then were administered to determine subjects'
ment has agreed to do a survey for the administration. The Univer- perceptions of how much freedom they had to decline to write the essay
sity in conjunction with the Board of Trustees has set up a commit- and of the effortfulness of the essay-writing task. Both these measures
tee on campus to investigate the possibilities of a parking fee for were on 31-point scales in which higher numbers indicate greater free-
those students who either live on campus or drive to school every-
dom and effort. When these were completed subjects placed these mea-
day. The fee would be about $ 15 per semester. After reviewing what
sures with the other material in the envelope. The subjects were then
they find, the committee will make a recommendation to the ad-
ministration regarding a change in parking policy. asked what they thought the study was concerned with. No subjects
expressed suspicion of deception regarding the campus committee. The
subjects were fully debriefed and thanked.
Choice manipulation. In the high-choice condition, subjects were fur-
ther told that Results

. . . the committee has asked to have arguments gathered from Choice Manipulation Check
both sides of the issue in order to make a more educated decision.
In the past it has been shown that a good way of doing this is simply Analysis of variance (ANOVA) revealed that subjects in the
to ask people to list arguments from a particular side of the issue. high-choice conditions reported greater perceived decision free-
dom (M = 24.3) than subjects in the low-choice conditions
The experimenter added that he had just been informed that morning (M= 5.3), F(\, 36) = 196.14, p < .001. Effortfulness had no
that the committee had finished gathering antiparking fee essays and
effects on perceived freedom, either through a main effect or
was now ready to gather proparking fee essays. The experimenter con-
interaction, Fs < 1.
tinued:

So while / would like to stress the voluntary nature of which side


of the issue you decide to write on, the committee needs strong, * Parking is free for students at the subjects' campus.
AROUSAL AND DISSONANCE 59

dence or "law of initial values" does not invalidate the between-


16 conditions results described later. First, a two-way ANOVA of the
Available
mean GSR estimates revealed no premanipulation differences
among the four conditions, all Fs < 1. Next, for each subject,
«o> 12 Not Available absolute GSR change for each 10-s measurement interval was
determined and the mean of these 17 change scores was com-
I
o puted. This mean change score constitutes an estimate of base-
line GSR fluctuation. One-way ANOVA revealed that baseline
1 • fluctuation did not differ across conditions, F < 1. Overall,
•*•«
mean within-subjects fluctuation was 9.474 kohms. In addi-
i tion, a Levene test for equal variances (Levene, 1960) revealed

I ' that within-groups variability in level of fluctuation did not


differ across conditions, F(3,36) = 1.75, p > . 17,
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Mean GSR estimates served as the unit of analysis for all sub-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

sequent analyses. To examine manipulation and attitude mea-


surement effects on GSR, a repeated measures ANOVA was
performed that included the choice and availability between-
Low High subjects factors and timing of GSR assessment—baseline, post-
essay, and postattitude change opportunity—as the repeated
CHOICE measure. This analysis revealed a main effect for timing, F(2,
Figure 1. Experiment 1: Mean pre- to postessay attitude change 72) = 50.82, p < .001, characterized by an overall decrease in
as a function of choice and argument availability. GSR (increase in arousal) between baseline and postessay as-
sessment. The main effect, however, was qualified by significant
interactions between timing and choice, F(2, 72) = 34.94, p <
.001, and timing and availability, F(2, 72) = 2.98, p = .057.
Effort Manipulation Check Figure 2 indicates that both high choice and high effort (argu-
The only significant effect on perceived effortfulness of the ments-unavailable) associated with the essay task served to
essay-writing task was a main effect for availability of essay ar- heighten arousal at the postessay measurement period, in a
guments, F(\, 36) = 142.01, p < .001. (Fs < 1 for choice main fashion that was primarily additive [the three-way interaction
effect and Choice X Availability interaction.) Subjects provided of choice, availability, and timing was nonsignificant, F(2,72) =
with arguments saw the essay task as less effortful (M = 7.4) 1,02, p > .36]. Using the Bonferroni procedure (Myers, 1972), a
than those not provided with arguments (M = 25.5). series of planned comparisons (contrasts-wide alpha = .05, df=
72), revealed an increase in arousal (decrease in GSR) from
baseline to postessay measurement in the high-choice/argu-
Attitudes ments-unavailable condition, M = 344, M - 249 (in kohms),
There were no differences among the four experimental / = 9.65, and the high-choice/arguments-available condition,
groups on the premeasures of attitude toward a student parking M = 310, M = 250, t = 6.10, but not in the low-choice/argu-
fee, F < 1 (overall M - 1 .58). Attitude change was determined ments-unavailable condition, M = 2\3,M=202,t = 1.15, and
by subtracting each subject's attitude premeasure score from low-choice/arguments-available condition, M = 266, M = 263,
his or her postessay score. Mean attitude change in each condi- t = .26.
tion is depicted in Figure L. A main effect was found for choice, No postessay to postattitude change opportunity trend in
F(l, 36) = 60.39, p < .001, with high-choice subjects (M= 13.7) GSR was apparent within any of the experimental conditions,
showing greater attitude change than low-choice subjects (M = all » < .2 (Ms - 249. 245, 201, 262, in the high/unavailable,
0.85). Neither the main effect of argument availability nor the high/available, low/unavailable, and low/available conditions,
interaction of choice and availability approached significance, respectively).5

Attitude-Arousal Relation
GSR
The overall correlation between attitude change and arousal
During each of the three 3-min recording periods, a GSR was shift (from baseline to postessay) was .73 (df= 38, p < .001).
taken every 10 s. A subject's score for a measurement period The correlation between attitude change and arousal shift was
was the mean of the 18 GSR estimates for that subject within a nonsignificant .36, df= 18,p> .10, when calculated forhigh-
a recording period. The overall correlation (n = 40) between
individuals' mean baseline GSR and their baseline-mean-to-
postessay-mean GSR change was .53 (p < .001 ), suggesting that s
In the two high-choice conditions in which the baseline to postessay
as expected (see Footnote 2), stimulus-produced (phasic) GSR increase in arousal was significant, the increase continued to be signifi-
responses depend in part on prestimulus baseline (tonic) GSR cant in the two high-choice conditions as evidenced by Bonferroni t
activity. Accordingly, several preliminary analyses of baseline contrasts between baseline and postattitude change opportunity levels
responses were made, which collectively suggest that this depen- of arousal, /s> 6.5.
60 ROGER A. ELK1N AND MICHAEL R. LE1PPE

100 explained on the basis of differential arousal-producing effort


Not Available
across their high- and low-dissonance conditions. This possibil-
ity is further ruled out by the fact that subjects did not perceive
80 the task as more eifortful when given high as opposed to low
choice.
The results raise interesting questions about the combined
effects of two sources of arousal: dissonance tension and task
60 difficulty. Among high-choice subjects, essay requirements did
not influence attitude change. If all arousal was misattributed
Available to the recognizable task difficulty in the arguments-unavailable
40 condition, we would expect no attitude change among high-
cc choice/high-effort subjects, because they would "feel" no disso-
nance. This form of misattribution seems untenable given the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

large number of studies showing significant attitude change fol-


This document is copyrighted by the American Psychological Association or one of its allied publishers.

20 lowing arguments-unavailable essay writing. If all arousal was


misattributed to dissonance tension, high-choice/high-effort
subjects should experience the most dissonance and possibly
change more. That this did not occur suggests two possibilities.
Subjects may indeed correctly differentiate the two sources of
Low High arousal. Misattribution may require special environmental cir-
cumstances, such as an ingested drug (a la Zanna & Cooper,
CHOICE
1974), camouflaged influences on bodily cues such as facial ex-
Figure 2. Experiment 1: Mean arousal shifts (skin resistance on a kohms pressions (Zuckerman, Klorman, Larrance, & Spiegel, 1981),
scale) from baseline to postessay as a function of choice and argument or temporal lag that removes attention from the "true" source
availability. (Shifts are in the direction of increased arousal [decreased of arousal (Zillmann, 1971). Alternatively, combining of the
GSR].) arousal from dissonance and the mental exertion required by
the task may indeed have occurred. However, the increase in
arousal over that produced by dissonance alone may need to
choice conditions only. The only significant within-cells correla- reach a certain threshold before it compels more extreme disso-
tion was that for the high/available condition, r = .63, p = .05. nance-reducing cognitive change. This threshold may not have
been reached in the high-choice/high-effort condition.
Discussion In contrast to the presumed tension reduction properties of
dissonance-driven attitude change, high-dissonance subjects'
The results are consistent with the findings of Croyle and arousal did not decline after they indicated their postessay atti-
Cooper (1983). As predicted, subjects who performed counter- tude. This finding may be because of any number of factors.
attitudinal behavior under dissonance-producing conditions of Perhaps attitude does not change in order to reduce tension, but
high choice showed heightened postbehavior arousal, whereas rather as a result of cognitive reorganization prompted by the
subjects who performed counterattitudinal behavior without awareness of dissonance. That is, arousal may signal an incon-
perceived free choice, and therefore should not have experi- sistency rather than drive its resolution (cf. Schlenkei; 1982). If
enced dissonance, did not. Moreover, as also predicted, these so, dissonance is not a drive state in the classic sense. Alterna-
same groups of subjects showed patterns of attitude change that tively, the sustained measured arousal may reflect residual pal-
were strikingly isomorphic with the arousal patterns: High- mar activity that obscured observation of subjects' relief. Fi-
choice subjects changed their attitude following their attitude- nally, perhaps the attitude change opportunity makes the indi-
discrepant behavior, low-choice subjects did not. These parallel vidual's dissonant state more acutely salient and therefore
patterns of arousal and attitude change concur with research sustains arousal or recreates dissonance that the individual re-
that suggests arousal is necessary for dissonance-produced atti- duced in his or her own way before the attitude change opportu-
tude change (cf. Cooper et al., 1978; Kiesler & Pallak, 1976). nity. Experiment 2 examined these possibilities.
The conclusion that dissonance produces a tension which
"drives" attitude change is bolstered further by the effects of Experiment 2
the argument availability manipulation. Judging from subjects'
perceptions of the effort required by the essay tasks, generating Our second experiment (a) tracked the course of arousal for
counterattitudinal arguments was, as expected, more effortful a longer postessay time period and, (b) for some subjects, did
than embellishing arguments made available. And, indeed, not provide an opportunity for explicit (i.e., written) attitude
arousal increased more in the former case. Importantly, how- change. The first of these procedures should allow for assess-
ever, no difference in attitude change occurred between the ar- ment of arousal reduction after any obscuring residual physio-
guments-unavailable and arguments-available groups, suggest- logical activity has ceased. The second procedure, in combina-
ing it is not arousal per se that is related to attitude change. tion with the first, should permit a test of whether explicit atti-
Rather, it is the arousal uniquely associated with dissonance. tude change sustains or reduces arousal or, in fact, is irrelevant
These results suggest that Cooper and Croyle's results cannot be to the postdissonance induction course of arousal, by yielding
AROUSAL AND DISSONANCE 61

comparisons of the temporal course of arousal among subjects changes. One filler task was eliminated. The three GSR recording peri-
given and not given an explicit attitude change forum. Although ods occurred at the same points in the procedure as in the previous
we acknowledge the possibility of private attitude change study, but were followed by a fourth 3-min recording period, the post-
among subjects whose postessay attitude is left unmeasured, it postattitude change opportunity measure. This measure was taken be-
seems reasonable that attitude change will have greater impact ginning 1 min after completion of the postattitude change opportunity
period.
on subjects' phenomenology if it is symbolized on an attitude
There were four high-choice conditions with 10 subjects in each. In
scale.
the replication group, subjects performed counterattitudinal behavior
In addition to high-dissonance conditions in which subjects and were given an attitude change opportunity as in the high-choice/
either did or did not receive an attitude posttest, Experiment 2 arguments-unavailable condition of Experiment I . In the no-attitude-
included a third theoretically relevant high-dissonance group change opportunity condition, the procedures were identical to the rep-
that not only received the posttest on the focal dissonance atti- lication group except that instead of the attitude change opportunity,
tude issue but also posttests on all the issues comprising a pre- subjects were provided with a magazine to browse through during that
manipulation questionnaire. This condition was included to ex- same time period (60 s). A third group, the dissonance/multi-item con-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

amine the possibility that dissonance arousal involves an un- dition, performed the counterattitudinal behavior, received the attitude
This document is copyrighted by the American Psychological Association or one of its allied publishers.

differentiated state of cognitive unrest that results in erratic change opportunity, and then was given the remaining nine items of the
initial student attitudes questionnaire received at the start of the study.
attitudinal responses. That is, we wondered if attitude change
Finally, in the consonance/multi-item condition, subjects wrote a proat-
on the focal issue might occur, not for the purpose of reducing titudinal essay and received the multi-item attitude change opportunity.
dissonance, but as a byproduct of diffuse, arousal-related cogni- This group served as an arousal control because heightened arousal
tive activity. If attitude change is dependent on arousal in this should not occur when performing a proattitudinal behavior. These lat-
undifferentiated way, shifts should be found also on issues for ter two groups were told that "we thought that the committee would
which counterattitudinal behaviors are not performed. also be interested in how the students felt about the remaining issues.
So, I'd like you to fill out this questionnaire again, except this time I'd
Method like you to include it in the envelope."

Subjects Results
Seven male and 33 female undergraduates participated to partially
fulfill a requirement for 100-level psychology courses. Males and fe-
Decision Freedom
males appeared in each condition in similar proportions. No differences in perceived decision freedom were obtained
between groups, F < 1. All of the group means were on the
Procedure and Design high choice side of the 31-point scale (replication = 17.20; no-
The same GSR equipment as employed in Experiment 1 was used. attitude-change opportunity = 18.80; dissonance/multi-item =
The basic procedure was identical to that of the first study with a few 17.90; consonance/mum-item = 21.70).

160

183
§
O 206
o
o
= 229
ill
O
252

co 275
K
? 298
CO
321

344

BL PE PACOPPACO BL PE PACOPPACO BL PE PACO PPACO BL PE PACO PPACO


REPLICATION NO ATTITUDE CHANGE DISSONANCE/ CONSONANCE/
OPPORTUNITY MULTI-ITEM MULTI-ITEM
Figure 3. Experiment 2: Mean arousal levels within and between conditions. (BL = Baseline; PE = Postes-
say; PACO = Postattitude Change Opportunity; PPACO = Post-Postattitude Change Opportunity. The
numbers on the ordinate axis have been placed in inverse order to make the numbers placed higher on the
axis correspond to heightened levels of arousal.)
62 ROGER A. ELK1N AND MICHAEL R. LE1PPE

Attitude Change however, a significant effect of time of measurement was ob-


tained, F(2, 54) = 5.51, p < .01, characterized by a significant
There were no differences among the groups on the premea- decreasing linear trend component, jFU, 54) = 5.48, p < .05,
sures of attitude toward a student parking fee, F < I . ANOVA on and a nonsignificant quadratic component, F < 1. Postdisso-
pre- to postessay attitude change scores on the focal parking fees nance arousal then tended to decrease over time for subjects
issue revealed that the three groups that received an attitude whose postessay attitudes were not assessed.
change opportunity differed in amount of attitude change, F(2,
27) = 8.96, p< .001. A Duncan range test showed that the con-
Attitude-Arousal Correlations
sonance/multi-item group (M = 0.50) changed significantly less
than the replication and dissonance/multi-item groups (Ms = The overall, across-conditions correlation between attitude
9.50 and 13.10, respectively) at the .01 level. These latter two change and baseline to postessay changes in arousal was signifi-
groups did not differ from each other. cant, r = .34 (dj'= 28, p = .05). In the dissonance conditions
To assess whether erratic responding to attitude scales of any only (i.e., the replication and dissonance/multi-item groups),
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

sort was occurring, absolute change scores (postessay score — the correlation between attitude change and arousal change was
This document is copyrighted by the American Psychological Association or one of its allied publishers.

preessay score!) were computed for all the remaining attitude nonsignificant, r = -.01. None of the within-cells correlations
items completed by subjects in the dissonance/multi-item and were significant.
consonance/multi-item conditions. Neither group showed sig-
nificant absolute change on any of the nine items, all ps > . 12. Discussion

Dissonance and Arousal


GSR
As in Experiment 1, subjects who performed counterattitudi-
Mean baseline GSR did not differ across conditions, F(3, nal behavior under high-choice conditions showed evidence of
36) = 1.29, p > .29.' Average fluctuation, as measured in Exper- heightened arousal following that behavior. This arousal oc-
iment 1, was 10.487 kohms overall, and did not differ across curred in all three high-dissonance conditions and was accom-
conditions, F < 1. Within-conditions variances in fluctuation panied by attitude change in the two conditions in which post-
were equivalent, Levene test F < 1. behavior attitude was assessed. Thus, arousal appears to be a
A Group X Timing of Measurement repeated measures AN- reliable consequence of dissonance. Moreover, attitude change
OVA revealed a main effect for time of measurement, f\3, only occurred on the issue that was the focus of counterattitudi-
108) = 6.93, p < .001, and an interaction between time of mea- nal behavior. Dissonance-produced arousal, then, does not sim-
surement and group, F(9, 108) = 2.75, p < .006. No main effect ply energize cognitive changes of any sort.
for group was obtained, ^3, 36) = 2.15,;» .1. These relations, Relevant to the evidence that both attitude change and
depicted in Figure 3, were examined further in a series of analy- arousal are coresults of cognitive dissonance are the attitude
ses that first assessed pre- to postessay changes in arousal and change-arousal correlations observed in Experiments 1 and 2.
then the course of postessay arousal among dissonance groups. Interestingly, across the high-dissonance conditions of both
As seen in Figure 3, arousal means are higher at postessay experiments, only one of six within-cells and within-high-
measurement than at baseline in all three dissonance condi- dissonances conditions correlations was significant. This overall
tions, whereas baseline and postessay arousal are the same in absence of a relation may have both methodological and sub-
the consonance condition. Because the dissonance conditions stantive roots. Methodologically, the correlations are based on
were identical in procedure through postessay measurement, a small sample sizes and apt to be unstable estimates of a relation.
statistical contrast between baseline and postessay arousal in In addition, any actual relation may be obscured by restriction
the three dissonance conditions combined was performed using of range of both arousal and attitude changes within high levels.
the overall repeated measures error term (df= 108) from the Theoretically, arousal and attitude change may be genuinely un-
ANOVA just described. This test revealed that arousal increased correlated among individuals experiencing dissonance. Physio-
significantly following dissonant behavior,; = 4.43, p < .001. A logical responses to inconsistent behavior, as well as the amount
contrast of baseline and postessay arousal in the consonance of attitude change prompted by a given level of arousal, proba-
condition revealed no such increase, t <l. bly vary widely from person to person, depending on individual
To investigate the course of postessay arousal over the three
measurement periods in the dissonance conditions, a repeated
6
measures ANOVA was performed excluding the consonance/ Though not nearly significant, the presence of an F value greater
multi-item group and the baseline GSR measure. This analysis than 1.00 creates a concern that small baseline differences between con-
revealed a significant main effect for condition, F(2,27) = 3.64, ditions might "grow" into large differences on subsequent measurement
of galvanic skin response (GSR) in accord with the law of initial values.
p < .04, and near-significant Time of Measurement X Condi-
Such a phenomenon does not seem to pose a rival explanation of the
tion interaction, F(4, 54) = 2.43, p = .058, and no overall effect
experimental results of Experiment 2. Figure 3 depicts mean arousal in
for time of measurement, F < 1. the four conditions at each GSR measurement period. It can be seen
Breakdown of the interaction into time of measurement sim- that baseline GSR is not agood predictor of GSR shift between baseline
ple main effects and trend tests for each condition revealed, first, and postessay measurement [r(3) between condition mean baseline val-
that there were no overall postessay changes in GSR among sub- ues and condition mean shifts = - .09]. In addition, the correlation (« =
jects in either the replication or dissonance/multi-item condi- 40) between individual mean baseline GSR and baseline-to-postessay
tions, Fs < 1. In the no-attitude-change opportunity group, change was a modest .32 (p - .05).
AROUSAL AND DISSONANCE 63

tolerances for dissonance and characteristic ways of handling an irrelevant GSR spike as discussed earlier). Moreover, after
dissonance (cf. Zanna & Aziza, 1976).7 resolving the inconsistency via attitude change, a new sense of
dissonance may be experienced—between original and postes-
say attitude espousal—to the extent that the original response
Dissonance Reduction and Arousal
is recalled. This new dissonance may be readily resolved (e.g.,
As in Experiment 1, arousal did not decline once high-disso- "My essay convinced me to change") but be arousing nonethe-
nance subjects indicated change on the postessay scale, even 4- less. Though not inconsistent with dissonance theory, the idea
7 min following the attitude response. In fact, in one condition that rearousal occurs when an explicit and public attitude
(dissonance/multi-item), arousal increased following attitude change opportunity is offered perhaps can be viewed as more
change. Arousal, however, did decrease for some high-disso- indicative of self-presentational concerns (Baumeister, 1982)
nance subjects, specifically those who were not asked to indicate and concomitant anxiety (Schlenker, 1982) rather than a disso-
their postessay attitude. These results cast some doubt on the nance state of heightened salience. Newly cognizant of his or
view that dissonance arousal reflects a tension state that moti- her previous attitude-discrepant behavior, the individual may
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

vates cognitive change in and of itself. But, clearly they do not now be confronted with the task of how to "account" for the
This document is copyrighted by the American Psychological Association or one of its allied publishers.

rule out this thesis. First, although subjects showed no evidence discrepancy in a way that presents the self in a positive light.
of "relief" after indicating attitude change on a scale, it is still For the impression-conscious person, this task, like the essay
possible that the dissonance reduction accomplished by their task itself, may prove anxiety provoking.
attitude change would eventually serve to quell their arousal. An alternative (but not incompatible with the heightened sa-
Indeed, dissonance theorists typically construe dissonance re- lience account) interpretation of the arousal results is that in
duction as a cognitive process that requires time (Festinger, the natural course of dissonance, people typically use a rather
1964; Pallak & Pittman, 1972; see, also, McGuire, 1960). Sec- undramatic (and unarousing) method of dealing with their dis-
ond, there is no reason to believe that those subjects not pro- sonance, namely, forgetting. As several researchers (e.g., Har-
vided with an explicit attitude change opportunity—whose dyck & Kardush, 1968; Rosenberg, 1960; Zanna & Aziza,
postessay arousal did decrease—did not also resolve their in- 1976) have suggested, dissonance may create tension, but, in a
consistency. They may have changed their attitudes privately, Freudian-like fashion, people might deal with the inner conflict
for example. by repressing the inconsistency, or, in a way that suggests recent
There are other possible accounts of the postattitude change research on "mindlessness" (e.g., Langer & Imber, 1980), they
arousal pattern. Among those offered earlier, the notion of re- might do so by ceasing to think about it. Accordingly, our no-
sidual palmar activity seems rather untenable given the ex- attitude-change opportunity subjects may have achieved reduc-
tended GSR recording period and the evidence of arousal de- tion in arousal because we allowed them to forget their disso-
cline in one condition. It is possible, though, that our extended nance. Indeed, we may have made it easy for them to forget by
recording period was still not long enough to observe a decline providing a distraction in the form of a magazine to browse
in arousal in those conditions involving a postessay attitude through (cf. Zanna & Aziza, 1976). Our remaining high-disso-
question. Any form of mental stimulation, including being nance subjects enjoyed no luxury of forgetting because we inter-
asked a question, may create a "spike" in electrodermal activity rupted the forgetting process with an attitude question, thereby
which, in the present study, would add to existing dissonance sustaining the conflict and its concomitant arousal. Perhaps the
arousal. In turn, even if attitude change mitigated arousal, a dissonance-plagued person would prefer that you "don't re-
lengthier period might be required for the combined arousal to mind me."
dissipate than for dissonance-produced arousal alone (i.e., the Two observations enhance the plausibility of forgetting as a
no-attitude-change opportunity condition) to be reduced. common dissonance reduction strategy. First, research suggests
We also suggested that arousal may be a built in response to many people are unable to recall their previous attitudes follow-
inconsistency, but one that functions to signal inconsistency ing counterattitudinal behavior (e.g., Bern & McConnell, 1970;
rather than drive its reduction. This information processing ac- Ross & Schulman, 1973). Second, cognitive inconsistency is
count is plausible if it is assumed that the postessay attitude surely something people deal with every day (cf. Aronson,
change opportunity brings the inconsistency back into focus, 1969). Forgetting may be a more efficient, less time-consuming
creating a second arousal signal. Thus, arousal may have begun strategy than cognitive reorganization, attitude shifts, and so
to decline for all subjects once the signal was registered, but this on, particularly because many attitude changes are reverber-
decline may have been interrupted for those given an attitude ating—changing one attitude may require other attitudes and
beliefs to change.
The notion of the attitude change opportunity as a rearouser
or sustainer of arousal may also make sense within a dissonance General Discussion and Conclusions
framework. As noted, when individuals are not confronted with A major implication of the foregoing discussion is that disso-
a request for attitude espousal in the immediate aftermath of nance may not be the forceful attitude-changing process it has
their attitude-discrepant behavior, dissonance- and arousal-re-
ducing cognitive activity may occur in a gradual fashion that ' The arousal-attitude changerelationmay be much like the message
does not require the direct attention of the individual. Asking comprehension-attitude change relation in persuasion. Variables that
people to state their attitudes just after their inconsistent act influence comprehension also influence persuasion (Eagly, 1974), but
may heighten conscious awareness of the inconsistency and, within-conditions comprehension-attitude change correlations are sel-
thus, "rearouse" dissonance tension (rather than simply adding dom reliable (Greenwald, 1968).
64 ROGER A. ELK1N AND MICHAEL R. LEIPPE

long been considered. Dissonance and dissonance reduction Brehm, J. W., & Cohen, A. R. (1962). Explorations in cognitive disso-
certainly may be everyday phenomena, but the sharp and rapid nance. New York: Wiley.
cognitive and behavioral changes attributed to dissonance re- Cacioppo, J. T, & Petty, R. E. (1979). Attitudes and cognitive responses:
An electrophysiological approach. Journal of Personality and Social
duction may occur mainly when, as in laboratory studies, indi-
Psychology, 37, 2181-2199.
viduals are questioned (by others or by themselves) about their
Carlsmith, J. M., & Freedman, J. L. (1968). Bad decisions and disso-
attitudes shortly following their attitude-discrepant behavior.
nance. In R. Abelson et al. (Eds.), Theories of cognitive consistency:
Otherwise, people may do less, less effortfully, to deal with in- A sourcebook. Chicago: Rand McNally.
consistencies than we think. Perhaps future research should be- Cooper, J., & Fazio, R. H. (1984). A new look at dissonance theory. In
gin studying more subtle effects of dissonance in situations that L. Berkowitz (Ed.), Advances in experimental social psychology (Vol.
approximate the more steady flow of everyday interplay be- 17, pp. 229-266). New York: Academic Press.
tween attitudes, behavior, and physiology. Cooper, J., & Goethals, G. R. (1974). Unforeseen events and the elimi-
The major issue addressed in these two studies is the motiva- nation of cognitive dissonance. Journal of Personality and Social Psy-
tional conceptualization of both arousal and attitude change in chology. 29, 441-445.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

cognitive dissonance. Cognitive dissonance theory argues that Cooper, J., Zanna, M. P., & Taves, P. A. (1978). Arousal as a necessary
This document is copyrighted by the American Psychological Association or one of its allied publishers.

condition for attitude change following induced compliance. Journal


dissonance is unpleasantly arousing and that attitude change
of Personality and Social Psychology, 36, 1101-1106.
reflects motivation to reduce the unpleasantness. We indeed
Cottrell, N. B., Rajecki, D. W, & Smith, D. K. (1974). The energizing
found that attitude-discrepant behavior under the illusion of
effects of postdecision dissonance upon performance of an irrelevant
high choice reliably produces increases in arousal as indexed
tusk. Journal of Social Psychology, 93, 81-92.
by GSR, and that this arousal is followed reliably by attitude Croyle, R. T., & Cooper, J. (1983). Dissonance arousal: Physiological
change when people are asked to make an attitude judgment. evidence. Journal of Personality and Social Psychology, 45, 782-791.
However, the presumed unpleasantness of the arousal cannot Eagly, A. H. (1974). Comprehensibility of persuasive arguments as a
be determined from GSR data because GSR does not reliably determinant of opinion change. Journal of Personality and Social
indicate evaluative reactions (Lacey, 1959). Therefore, it is only Psychology, 29, 758-773.
through the arousal's subsequent reduction that motivation can Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA:
be implied, and we found no evidence that explicit attitude Stanford University Press.
change reduced arousal. On the contrary, arousal only de- Festinger, L. (1964). Conflict, decision, and dissonance. Stanford, CA:
Stanford University Press.
creased in the absence of an attitude change opportunity.
Greenwald, A. G. (1968). Cognitive learning, cognitive response to per-
Cognitive dissonance, then, may or may not be a motivational
suasion, and attitude change. In A. G. Greenwald, T. C. Brock, &
state. Reminiscent of the dissonance versus self-perception de-
T. M. Ostrom (Eds.), Psychological foundations of attitudes. New
bate of more than a decade ago (see Greenwald, 1975) and in
York: Academic Press.
line with the current intrapsychic-impression management Greenwald, A. G. (1975). On the inconclusiveness of "crucial" tests of
controversy (Tetlock & Manstead, 1985), our research leaves a dissonance versus self-perception theories. Journal of Experimental
major question unanswered: Does dissonance arousal just sig- Social Psychology, 11,490-499.
nal inconsistency to the individual or, as dissonance theorists Greenwald, A. G., & Ronis, D. L. (1978). Twenty years of cognitive
would have it, is arousal a consequence of inconsistency that dissonance: Case study of the evolution of a theory. Psychological Re-
fuels a drive to resolve it in some way? Though an answer to this view, 85, 53-57.
question would have theoretical value, an equally important is- Hardyck, J. A., & Kardush, M. (1968). A modest modish model for
sue, as we have implied, may be how people typically deal with dissonance reduction. In R. Abelson et al. (Eds.), Theories of cogni-
tive consistency: A sourcebook. Chicago: Rand McNally.
cognitive inconsistency, driven or not.
Hull, C. L. (1943). Principles of behavior. New York: Appleton-Century-
Crofts.
Insko, C. A. (1967). Theories of attitude change. New York: Appleton-
References
Century-Crofts.
Aronson, E. (1969). The theory of cognitive dissonance: A current per- Kiesler, C. A., & Pallak, M. S. (1976). Arousal properties of dissonance
spective. In L. Berkowitz (Ed.), Advances in experimental social psy- manipulations. Psychological Bulletin, 83.1014-1025.
chology (Vol. 4, pp. 2-34). New York: Academic Press. Lacey, J. I. (1959). Psychophysiological approaches to the evaluation of
Aronson, E., & Mills, J. (1959). The effect of severity of initiation on psychotherapeutic process and outcome. In E. A. Rubinstein &
liking for a group. Journal of Abnormal and Social Psychology, 59, M. B. Parloff (Eds.), Research in psychotherapy. Washington, DC:
177-181. American Psychological Association.
Baumeister, R. F. (1982). A self-presentational view of social phenom- Langer, E. J., & Imber, L. (1980). Role of mindlessness in the perception
ena. Psychological Bulletin, 91, 3-26. of deviance. Journal of Personality and Social Psychology, 39, 360-
Bern, D. J. (1967). Self-perception: An alternative interpretation of cog- 367.
nitive dissonance phenomena. Psychological Review, 74, 183-200. Levene, H. (1960). Robust tests for equality of variance. In I. Olkin
Bern, D. J., & McConnell, H. K. (1970). Testing the self-perception ex- (Ed.), Contributions to probability and statistics. Palo Alto, CA: Stan-
planation of dissonance phenomena: On the salience of premanipula- ford University Press.
tion attitudes. Journal of Personality and Social Psychology, 14, 23- Mandler, G. (1975). Mind and emotion. New York: Wiley.
31. McGuire, W. J. (1960). Cognitive consistency and attitude change. Jour-
Berlyne, D. E. (1965). Structure and direction in thinking. New York: nal of Abnormal and Social Psychology, 60, 345-353.
Wiley. Myers, J. L. (1972). Fundamentals of experimental design (2nd ed.).
Blatt, S. J. (1961). Patterns of cardiac arousal during complex neural Boston: Allyn and Bacon, Inc.
activity. Journal of Abnormal and Social Psychology, 63, 272-282. Pallak, M. S., & Pittman, T. S. (1972). General motivational effects of
AROUSAL AND DISSONANCE 65

dissonance arousal. Journal of Personality and Social Psychology, 21 works from a physiologic point of view. Part 1. American Journal of
349-358. Physical Medicine, 36, 295-320.
Rabbie, J. M., Brehm, I. W., & Cohen, A. R. (1959). Verbalization and Wang, G. H. (1958). The galvanic skin reflex. A review of old and recent
reactions to cognitive dissonance. Journal of Personality, 27, 407- works from a physiologic point of view. Part 2. American Journal of
417. Physical Medicine, 37, 35-57.
Rosenberg, M. J. (I960). An analysis of affective-cognitive consistency. Zanna, M. P., & Aziza, C. (1976). On the interaction of repression-
In C. I. Hovland & M. J. Rosenberg (Eds.), Attitude organization and sensitization and attention in resolving cognitive dissonance. Journal
change. New Haven, CT: Yale University Press. of Personality, 44, 577-593.
Ross, M., & Schulman, R. F. (1973). Increasing the salience of initial Zanna, M. P., & Cooper, J. (1974). Dissonance and the pill: An attribu-
attitudes: Dissonance versus self-perception theory. Journal of Per- tion approach to studying the arousal properties of dissonance. Jour-
sonality and Social Psychology, 28, 138-144. nal of Personality and Social Psychology, 29, 703-709.
Schlenker, B. R. (1982). Translating actions into attitudes: An identity- Zillmann, D. (1971). Excitation transfer in communication-mediated
analytic approach to the explanation of social conduct. In L. Berkow- aggressive behavior. Journal of Experimental Social Psychology, 7,
itz (Ed.), Advances in experimental social psychology (Vol. 15, pp. 419-434.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

151-181). New York: Academic Press. Zimbardo, P. G. (1965). The effect of effort and improvisation on self-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Sokolov, A. N. (1963). Perception and the conditioned reflex. Oxford:


persuasion produced by role playing. Journal of Experimented Social
Pergamon Press.
Psychology, 1, 103-120.
Sourek, K. (1965). The nervous control of skin potentials in man.
Zuckerman, M., Klorman, R., Larrance, D. X, & Spiegel, N. H. (1981).
Prague: Czechoslovak Academy of Sciences.
Facial, autonomic, and subjective components of emotion: The facial
Tetlock, P. E., & Manstead, A. S. R. (1985). Impression management
feedback hypothesis versus the externalizer-internalizer distinction.
versus intrapsychic explanations in social psychology: A useful di-
Journal of Personality and Social Psychology, 41,929-944.
chotomy? Psychological Review, 92, 59-77.
Venables, P. H., & Christie, M. J. (1980). Electrodermal activity. In I.
Martin & P. H. Venables (Eds.), Techniques in psychophysiology. New
York: Wiley. Received March 11, 1985
Wang, G. H. (1957). The galvanic skin reflex. A review of old and recent Revision received January 22, 1986

You might also like