Rokeah

American Association for Public Opinion Research
Attitude Change and Behavioral Change

Author(s): Milton Rokeach
Source: The Public Opinion Quarterly, Vol. 30, No. 4 (Winter, 1966-1967), pp. 529-550
Published by: Oxford University Press on behalf of the American Association for Public Opinion Research
Stable URL: http://www.jstor.org/stable/2746961 .
Accessed: 01/09/2013 11:44
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 Association for Public Opinion Research and Oxford University Press are collaborating with JSTOR
to digitize, preserve and extend access to The Public Opinion Quarterly.
http://www.jstor.org
This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

All use subject to JSTOR Terms and Conditions
ATTITUDE CHANGE AND
BEHAVIORAL CHANGE*
BY MILTON ROKEACH
The literatureon opinion and attitude change continues to probe deeper.

The author of this paper arguesthat attitude change is a function not merely
of attitude toward an object but also of attitude toward a situation. This
proposition further complicatesthe study of expressedopinion change and
behavior change in relation to attitude, and leads the author to consider
severalnew methodsfor assessingattitudechange.
Dr. Milton Rokeach is Professorof Psychologyat Michigan State Univer-
sity.
SNINCE
World War II many experimental studies of opinion
change, carried out within a variety of conceptual frameworks,
have been designed to increase our theoretical understanding
of the conditions under which men's minds and men's behavior
may change. While the main empirical focus of these studies is on
behavioral changes in the expression of opinion, their main theoreti-
cal concern is with the conditions facilitating and inhibiting change
in underlying beliefs and attitudes. To what extent have these experi-
mental studies actually advanced our theoretical understanding of
processes leading to attitude and behavior change? And to what ex-
tent have they improved our understanding of the fundamental struc-
ture of underlying attitudes, the way attitudes are organized with re-
spect to one another, and the way attitude and attitude change may
affect behavior?
To discuss these questions I would like to begin with certain con-
siderations, not about attitude change, but about the nature of atti-
tude, and about the relationship between attitude and behavior. In
contemporary approaches to "attitude change" the accent seems to be
on the understanding of "change" rather than on the understanding
of "attitude"; that is, one may note an interest in attitude theory as
such only insofar as that interest is necessary to formulate testable
hypotheses about attitude change. The point of view to be developed
* This is one of a series of papers on attitude organization and modification sup-

ported by a grant from the National Science Foundation. Earlier versions of this pa-
per were presented at a Symposium on Attitude Change and Behavior Change held
at the State University of New York at Buffalo in March 1965, and at the meetings
of the World Association for Public Opinion Research in Dublin in September 1965.

530 MILTON ROKEACH
here will differ somewhat from that expressed by the late Arthur R.
Cohen who, in the preface to his recent book, Attitude Change and
Social Influence, wrote: "This book does not take up the definition
and conceptualization of attitude, but instead assumes that there is a
commonly accepted core of meaning for the term 'attitude change'."'
I will try to show that the concept of "attitude change" can have no
"commonly accepted core of meaning" apart from the concept of
attitude-that, indeed, theory and research on the nature, determi-
nants, and consequents of attitude formation and maintenance are
prerequisite to and inseparable from theory and research on attitude
change.
ATTITUDE AND BEHAVIOR
For purposes of this discussion let me offer the following coordi-

nated definitions of attitude and attitude change. An attitude is a rela-
tively enduring organization of beliefs about an object or situation
predisposing one to respond in some preferential manner.2 A ttitude
change would then be a change in predisposition,3 the change being
either a change in the organization or structure of beliefs or a change
in the content of one or more of the beliefs entering into the attitude
organization.
Especially important for the major thesis of this paper is that an
attitude may be focussed either on an object or on a situation. In the
first instance we have in mind an attitude-object, which may be con-
crete or abstract, involving a person, a group, an institution, or an
issue. In the second instance the attitude is focussed on a specific situa-
tion, an event, or an activity. To say that a person has an enduring
attitude toward a given object is to say that this attitude will, when
activated, somehow determine his behavior toward the attitude-object,
across situations; conversely, to say that a person has an enduring atti-
tude toward a given situation is to say that this attitude will, when
activated, determine his behavior toward the situation, across atti-
tude-objects.
Social scientists have generally been far more interested in the the-
ory and measurement of attitudes toward objects, across situations,
than in the theory and measurement of attitudes toward situations,
'Arthur R. Cohen, Attitude Change and Social Influence, New York, Basic Books,
1964, p. xi.
2 A detailed elaboration of the nature of attitudes is presented in M. Rokeach,
"The Nature of Attitudes," in International Encyclopedia of Social Sciences, New
York, Macmillan, 1966.
3 A predisposition would be defined as a hypothetical state of the organism which,
when activated by a stimulus, causes a person to respond selectively, affectively, or
preferentially to the stimulus,

ATTITUDE CHANGE AND BEHAVIORAL CHANGE 531
across objects. Thus, we have scales that measure attitudes toward the
Negro, the Jew, liberalism-conservatism, and religion. But we do not
have scales that measure attitudes toward such situations as managing
or being a guest in a Southern hotel, being a passenger or driver of a
bus, buying or selling real estate. As a result, the study of attitudes-
toward-situations has become more or less divorced from the study of
attitudes-toward-objects.
This separation of attitude-toward-object from attitude-toward-
situation has, in my opinion, severely retarded the growth not only of
attitude theory but also of attitude-change theory. It has resulted in a
failure to appreciate that an attitude-object is always encountered
within some situation about which we also have an organized attitude.
It has resulted in unsophisticated attempts to predict behavior or be-
havioral change on the basis of a single attitude-toward-object, ignor-
ing the equally relevant attitude-toward-situation. And it has resulted
in unjustified interpretations and conclusions that there are often
inconsistencies between attitudes and behavior, or between attitude
changes and behavioral change.
A preferential response toward an attitude-object cannot occur in a
vacuum; it must necessarily be elicited within the context of some
social situation about which, as already noted, we also have an atti-
tude. If you like, conceive of the attitude-object you encounter as the
figure and the situation in which you encounter it as the ground.
How a person will behave toward an object-within-a-situation will
depend, on the one hand, on the particular beliefs or predispositions
activated by the attitude-object and, on the other hand, on the particu-
lar beliefs or predispositions activated by the situation. Thus it fol-
lows that a person's social behavior must always be a function of at
least two attitudes-one activated by the attitude-object, the other
activated by the situation.
If one focusses only on attitude-toward-object, he is bound to ob-
serve some inconsistency between attitudes and behavior, or, at least,
a lack of dependence of behavior on attitude. Most frequently men-
tioned as evidence of this are such studies as those by LaPiere4 and
Kutner et al.,5 in which restaurant owners and innkeepers showed
marked discrepancies between their verbal expressions of discrimina-
tion toward Chinese and Negroes by letter or phone and their non-
discriminatory, face-to-face behavior. The present analysis suggests
one possible explanation for such apparent inconsistency: the in-
4 R. T. LaPiere, "Attitudes vs. Actions," Social Forces, Vol. 13, 1934, pp. 230-237.
5 B. Kutner, Carol Wilkins, and Penny R. Yarrow, "Verbal Attitudes and Overt
Behavior Involving Racial Prejudice," Journal of Abnormal and Social Psychology,
Vol. 27, 1952, pp. 649-652.

532 MILTON ROKEACH
vestigators did not obtain all the relevant attitudinal information
needed to make accurate predictions. The subjects not only had atti-
tudes toward Chinese and Negroes, but, as managers of businesses, also
had attitudes about the proper conduct of such businesses. The in-
vestigator's methods, however, are typically focussed on attitude-
toward-object and are generally insensitive to attitude-toward-
situation.
The proposition that behavior is always a function of at least two
attitudes (to be called Ao and As) has at least two implications worth
noting. First, a given attitude-toward-object, whenever activated, need
not always be behaviorally manifested or expressed in the same way or
to the same degree. Its expression will vary adaptively as the attitude
activated by the situation varies, with attitude-toward-situation facili-
tating or inhibiting the expression of attitude-toward-object, and vice
versa. Any attitude-toward-object has the inherent property of being
differentially manifested along a range of values rather than as a single
value, depending on the situation within which the attitude-object is
encountered.6 This same property is inherent in any attitude-toward-
situation. Consequently, a significant change of opinion toward an
object may indicate nothing more than that a given attitude-toward-
object was activated, and thus behaviorally expressed, in two different
situations, S1 and S2, activating, respectively, two different attitudes-
toward-situation, A.1 and As,.
Second, in principle there is no difference between the verbal and
the nonverbal expression of a given attitude. Every expression in be-
havior, verbal or nonverbal, must be a confounding and a compound-
ing function of at least two underlying attitudes. Thus, any verbal
expression of opinion, like any nonverbal behavior, is also a function
of at least two attitudes-attitude-toward-object and attitude-toward-
situation-and ascertaining the extent to which the opinion is a
manifestation of one attitude or the other, or both, requires careful
inference rather than careless assumption.
COGNITIVE INTERACTION BETWEEN TWO ATTITUDES
It is not enough merely to assert that social behavior is a function

of two attitudes. To predict behavioral outcome requires a model
6 H. C. Kelman expresses a similar view: "The attitudes expressed by an individual

may vary from situation to situation, depending on the requirements of the situa-
tion in which he finds himself and the motivations which he brings into this situa-
tion. What the individual says will be determined at least in part by what he con-
siders to be proper in this situation and consonant with group norms, and also by
what he considers to be most conducive to the achievement of his personal goals ....
The amount of discrepancy depends on the situational requirements, on the per-
son's goals, on his relation to the group, and on some of his personal characteristics."

ATTITUDE CHANGE AND BEHAVIORALCHANGE 533
about the manner in which the two attitudes will cognitively interact
with one another. In a recent paper, Rokeach and Rothman7 have
presented such a model of cognitive interaction, called the belief
congruence model, which has been shown to be approximately three
times as accurate as Osgood and Tannenbaum's congruity model8 in
quantitatively predicting the outcome of cognitive interaction.
Applying this model to the present context, we can then conjecture
that whenever a person encounters an attitude-object within some
situation, two attitudes, Ao and A, are activated; further, a compari-
son process regarding the relative importance of these two attitudes
with respect to one another is also activated. The two attitudes are
assumed to affect behavior in direct proportion to their perceived im-
portance with respect to one another. The more important Ao is per-
ceived to be with respect to A., the more will the behavioral outcome
be a function of Ao; conversely, the more important A8 is perceived to
be with respect to AO, the more will the behavioral outcome be a
function of A8.
How can the relative importance of two attitudes be determined?
One way is by strictly empirical means-for example, by the method
of paired comparison. In this instance we would not be able to pre-
dict the behavioral outcome of cognitive interaction between two
attitudes of varying importance because we would have no way of
knowing in advance their importance relative to one another or the
absolute degree of that importance. Fortunately, however, the com-
parison of relative importance of the two attitudes does not occur in a
vacuum; it takes place within the general framework of one's total
belief system, wherein all beliefs and attitudes are arranged along a
central-peripheral dimension of importance. Thus, the two attitudes,
Ao and A., can be compared as if to determine their relative position
along this central-peripheral dimension. I have already presented else-
where some tentative and admittedly incomplete conceptualizations of
the types of beliefs and attitudes that may be found along this central-
peripheral dimension;9 such conceptualizations enable us to make at
least some educated guesses in advance about which of two attitudes
is likely to be the more important. I will have more to say about the
("Social Influence and Personal Belief: A Theoretical and Experimental Approach to
the Study of Behavior Change," unpublished manuscript, 1958, pp. 25-26.)
7 M. Rokeach and G. Rothman, "The Principle of Belief Congruence and the
Congruity Principle as Models of Cognitive Interaction," Psychological Review, Vol.
72, 1965, pp. 128-142.
8 C. E. Osgood and P. H. Tannenbaum, "The Principle of Congruity in the Predic-
tion of Attitude Change," Psychological Review, Vol. 62, 1955, pp. 42-55.
9 M. Rokeach, The Open and Closed AMind:Investigations into the Nature of
Belief Systems and Personality Systems, New York, Basic Books, ig6o; The Three
Christs of Ypsilanti: A Psychological Study, New York, Knopf, 1964.

534 MILTON ROKEACH
several types of beliefs varying along the central-peripheral dimension
later on in this paper.
The question may now be raised whether it is ever possible to ob-
tain a behavioral measure of a given attitude-toward-object that is
uncontaminated by interaction with attitude-toward-situation. The
extent to which this is possible is a function of the extent to which
the situation is a "neutral" one-that is, a situation carefully struc-
tured by the experimenter to activate a relatively unimportant atti-
tude-toward-situation that is, hence, of relatively little influence in the
context of its interaction with attitude-toward-object. Learning how
to structure the test or interview situation so that it is a neutral one is,
of course, a major objective of attitude and survey research methodol-
ogy. But this is only a methodological ideal to strive for and is prob-
ably rarely achieved in practice.
ATTITUDE CHANGE AND BEHAVIORAL CHANGE
The proposition that behavior is always a function of two inter-

acting attitudes has, I believe, important and disturbing implications
for theory and research on attitude change and behavioral change. If
expressing an opinion is a form of behavior, then expressing a
changed opinion is also a form of behavior; thus, a changed opinion
must also be a function of the two attitudes previously discussed-
attitude-toward-object and attitude-toward-situation. Similarly, any
change in nonverbal behavior is also a form of behavior, and hence
must also be a function of the same two attitudes. The question there-
fore arises: When there is a change in opinion or behavior, how can
we tell whether or not there has also been a change in attitude and, if
so, which attitude?
Although a reasonably clear distinction can be made between an
underlying attitude and an expression of opinion (or, if you will, be-
tween a covert and overt attitude, or between a private and public
attitude), and between an underlying attitude change and an ex-
pressed opinion change, one may nevertheless observe in the experi-
mental and theoretical literature a general tendency to use these
concepts interchangeably and thereby to shift the discussion back and
forth between "attitude" and "opinion," and between "opinion
change" and "attitude change." It thereby becomes difficult to tell
whether one is dealing in any instance with phenomena involving
attitude change, expressed opinion change, or both. Many writers
have ridden roughshod over the distinction between attitude and
expressed opinion by using the phrases "attitude change," "opinion
change," "attitude and opinion change" and "attitude or opinion
change" more or less arbitrarily in the context of a single discussion.

In this way the impression is created that a significant change in the
expression of an opinion also represents a change in underlying atti-
tude. For example, Hovland opens his paper, "Reconciling Conflict-
ing Results Derived from Experimental and Survey Studies of Atti-
tude [sic] Change":
Two quite different types of research design are characteristicallyused to
study the modificationof attitudes [sic] through communication.In the first
type, the experiment, individuals are given a controlled exposure to a com-
munication and the effects evaluated in terms of the amount of change in
attitude or opinion [sic] produced.10
Festinger opens his article entitled, "Behavioral Support for Opinion
[sic] Change":
The last three decades have seen a steady and impressive growth in our
knowledge concerning attitudes and opinions [sic].ll
Both of these writers, like many others, then employ the concepts of
"attitude" and "opinion" indiscriminately in carrying forward their
discussions, which are usually discussions about how some empirical
data involving a change in expressed opinion bear on some hypothesis
or theory regarding a change in attitude.
As one tries to assimilate the growing experimental literature on
opinion change, he becomes increasingly aware that this literature
concerns primarily the conditions affecting change in the expression
of opinion. But this literature, considered as a whole, does not seem
to have much to say about the conditions leading to a change in the
content or structure of underlying predispositions (or, as Doob would
have it,12 of implicit responses) toward objects or toward situations.
Theories of attitude change, with certain exceptions (e.g. Kelman's
work on processes of social influence, related work on the public-
private variable, work on the "sleeper" effect), seem to be generally
unconcerned with whether an expressed opinion change does or does
not represent an underlying attitude change. Indeed, the classical
paradigm employed in experimental studies of opinion change-
pre-test, treatment, post-test-is not capable of telling us whether an
expressed opinion change indicates an attitude change; it can tell us
only whether an expression of opinion has or has not changed as a
result of a particular experimental treatment. If the main theoretical
concern of experimental studies on expressed opinion change is with
10 C. I. Hovland, "Reconciling Conflicting Results Derived from Experimental and
Survey Studies of Attitude Change," American Psychologist, Vol. 14, 1959, p. 8.
11 L. Festinger, "Behavioral Support for Opinion Change," Public Opinion Quar-
terly, Vol. 28, 1964, p. 404.
12 L. W. Doob, "The Behavior of Attitudes," Psychological Review, Vol. 54, 1947,
pp. 135-156.

536 MILTON ROKEACH
the conditions leading to attitude change, then the classical paradigm
is basically a faulty one and should be replaced with other or modified
experimental designs (to be discussed later) more suited to deal with
this issue.
A closely related point concerns the relationship among attitude
change, expressed opinion change, and behavioral change. In his
presidential address to Division 8 of the American Psychological As-
sociation, Leon Festinger expresses astonishment over the "absence
of research, and of theoretical thinking, about the effect of attitude
change on subsequent behavior."13 He says that he could find only
three empirical studies relevant to this problem and that they all
show "the absence of a relationship between opinion change . . . and
resulting behavior.114 Festinger stresses in his closing remarks that we
ought not to ignore this problem or simply assume "a relationship
between attitude change and subsequent behavior....." He concludes
that "The problem needs concerted investigation."1'5And Cohen, in a
similar vein, writes: "Until experimental research demonstrates that
attitude change has consequences for subsequent behavior, we cannot
be certain that our procedures . . . do anything more than cause cogni-
tive realignments."'6 It should be noted, first, that we cannot even be
certain whether the experimental procedures employed "cause cognitive
realignments" and, second, that the absence of relationship noted by
Festinger and Cohen is not between attitude change and subsequent
behavior but between two forms of behavior-expressed opinion
change, and subsequent nonverbal behavioral change. My main point,
then, is that there would seem to be not one but two problems re-
quiring "theoretical thinking" and "concerted investigation." First:
Why is it so difficult to demonstrate a relationship between attitude
change and behavioral change? And, second: Why is it so difficult to
demonstrate a relationship between one form of behavioral change
and another form?
I have proposed that expressed opinion or behavioral change is al-
ways a function of at least two attitudes. This proposition only com-
plicates our attempts to determine whether or not a particular change
in expressed opinion or behavior represents a change in attitude. Be-
cause we have to contend with two types of underlying attitudes, we
now have four possible determinants of a change in expressed opinion
or behavior: (i) interaction between attitude-toward-object and at-
titude-toward-situation, neither of which has changed, or (2) a
13 Festinger, op. cit., p. 405.

14 Ibid., p. 416.
15 Ibid., p. 417-
16 Cohen, op. cit., p. 138.

change only in the attitude-toward-object, or (3) a change only in
the attitude-toward-situation, or (4) a change in both attitude-toward-
object and attitude-toward-situation.
Changes in expressed opinion or behavior as a result of (2), (3),
and (4) are more or less self-evident. But the first determinant of
expressed opinion or behavioral change-the interaction between
AOand A8 neither of which has changed-merits further consideration
because (a) it goes against the widely held assumption that behavioral
and expressed opinion changes cannot take place without a preceding
change in attitude; (b) it has implications for experimentally ori-
ented and personality-oriented studies of attitude and behavioral
change; and (c) it may open up some fresh possibilities for bringing
about changes in expressed opinion and behavior that do not de-
pend on antecedent attitude change.
Let us consider a variety of instances in which a change of ex-
pressed opinion or behavior may be observed and understood without
positing a change in underlying attitude. First, there are those actions
which represent public conformity or compliance without private ac-
ceptance. Kelman has shown that a subject exposed to an authority
who is in a position to reward and punish will display a change of
opinion in the direction of authority's opinion,17 but this change of
opinion is manifested only under conditions of surveillance by author-
ity and not under conditions of nonsurveillance. The surveillance
condition represents a situation, S1, activating the attitude, A*1. The
nonsurveillance condition represents another situation, S2, activating
another attitude, A8.. A change in expression of opinion from condi-
tions of nonsurveillance to surveillance can readily be accounted for
without assuming a change in underlying attitude-toward-object. One
measure of opinion toward a specified object is the behavioral result
of the interaction between Ao and A,,; the other measure is a result of
the interaction of the same Ao but another attitude, A .,. The change of
expressed opinion toward the specified attitude-object can be best
understood as a reflection of the two different situations, each acti-
vating a different attitude-toward-situation. And there is no need to
assume that any one of the activated attitudes-AO or A., or A,.has
undergone any change.
But not all instances of expressed opinion change unaccompanied
by attitude change necessarily represent acts of public compliance or
conformity. Consider, for example, expressed opinion changes
brought about as a result of what Orne has called the "demand char-
17 H. C. Kelman, "Compliance, Identification, and Internalization: Three Processes

of Attitude Change," Journal of Conflict Resolution, Vol. 2, 1958, pp. 51-60.

538 MILTON ROKEACH
acteristics" of the experimental situation,18 or what Rosenberg has
called "evaluation apprehension."19 Both terms refer to methodologi-
cally unwanted situational variables that may or may not motivate
compliant behavior, variables which exist during the post-test period
and not during the pre-test period, and which activate some attitude-
toward-situation persisting beyond that activated by the experimental
treatment as such. Changes in expressed opinion toward an object
from pre. to post-test thus result because two different situations ac-
tivated two different attitudes, A81 and A,8; and we can therefore ac-
count for such changes without adding further assumptions regard-
ing underlying attitude change.
Incidentally, changes in expressed opinion thus obtained are difficult
to interpret because they violate a basic principle of measurement
theory, namely, that repeated measurements designed to assess
the effects of some experimental variable should be obtained un-
der constant test conditions. Unlike survey research methodology, ex-
perimental studies of opinion change employing the pre-test, treat-
ment, and post-test paradigm cannot by their very nature guarantee
the required constancy of testing conditions. The post-test situation is
bound to be psychologically different from the pre-test situation, the
former activating different attitudes from the latter. Moreover, a
post-test situation following one experimental treatment is not neces-
sarily comparable with another post-test situation following a different
experimental treatment. Orne expresses a similar view when he
writes: "It should be clear that demand characteristics cannot be
eliminated from experiments; all experiments will have demand
characteristics and these will always have some effect."20 Neverthe-
less, the proposition that behavior is a function of A, and A, would, if
valid, require us to assess the relative effects of A, and A, in the pre-test
and in the post-test situations separately, in order to determine the
meaning of a given change in expressed opinion.
Let me turn now to another illustration, this time not of an opinion
change but of a change in behavior, real-life behavior which is differ-
ent from what we would ordinarily expect, which does not necessarily
involve an attitude change, and which does not necessarily represent
an act of public compliance or conformity.
This experiment, which was carried out in collaboration with
18 M. T. Orne, "On the Social Psychology of the Psychological Experiment: With

Particular Reference to Demand Characteristics and Their Implications," American
Psychologist, Vol. 17, 1962, pp. 776-783.
19 M. J. Rosenberg, "When Dissonance Fails: On Eliminating Evaluation Appre-
hension from Attitude Measurement," Journal of Personal and Social Psychology,
Vol. i, 1965, pp. 28-42.
20 Orne, op. cit., p. 779.

Louis Mezei, takes place in the natural setting of the personnel offices
of two state mental hospitals near Detroit, Michigan. The subjects are
twenty-six Negro and twenty-four white males applying for jobs as
janitor, attendant, laundry worker, and the like. Each subject, after
he has filled out his application forms, is made to wait for his inter-
view in a room with four other people who are apparently also wait-
ing to be interviewed. These four are really our confederates. Two of
them are white and two are Negro. These four, along with the real
applicant, are told that while they wait they might want to look at a
mimeographed list of issues that are used in a training program for
hired personnel. These issues, concerning problem situations involv-
ing difficult mental patients, can be resolved in one of two ways: by
being permissive or harsh. For example, one issue concerns a mental
patient who asks to be fed after dining hours. How should one deal
with such a request, the subjects are asked. The confederates "spon-
taneously" launch into a discussion of the issues, with one white and
one Negro confederate advocating the permissive solution and with
one white and one Negro confederate advocating the harsh solution.
The naive subject is gradually drawn into the discussion by the four
confederates so that he, too, states whether he favors the permissive or
harsh solution. He is then asked which two of the four confederates
he would most prefer to work with. Which two will he choose and on
what basis?
One might reasonably expect that of fifty persons applying for low-
status jobs a substantial number would, under the conditions de-
scribed, choose two partners of their own race, given the salience of
racial attitudes in our culture. But the results shown in Table 1
(which are the results from only one of three experiments reported
more fully elsewhere by Rokeach and Mezei)21 do not confirm the
expectation that attitude-toward-race is at all important. There are
six possible ways of choosing the two partners under the specified con-
ditions. These are indicated in the column heads of Table i: (i)
S+O+, a person of the same race and a person of the other race, both
agreeing with the subject; (2) S-O-, a person of the same race and
a person of the other race, both disagreeing with the subject, etc.
Only two subjects of the fifty chose attitude-objects of the same race
(pattern 3, S+S-), considerably less than would be expected even
on a purely chance basis. That similarity of race is not an important
criterion of choice of work partners, either for white or for Negro
subjects, is indicated further in that three more subjects chose two
work partners of the other race (pattern 4, O+O-). It is clear that
21 M. Rokeach and L. Mezei, "Race and Shared Belief as Factors in Social Choice,"
Science, Vol. 151, 1966, pp. 167-172.

540 MILTON ROKEACH
TABLE 1
FREQUENCIES OF CHOOSING EACH INTERPERSONAL PATTERN
(1) (2) (3) (4) (5) (6)

S+O+ S-0- S+S- 0+0- S+O- S-0+ Total
Negro 15 3 1 2 3 2 26
White 15 0 1 1 1 6 24
Total 30 3 2 3 4 8 50
S = same race; 0= other race; + = agree;- = disagree.
the most frequent basis of choice is not similarity of race but similar-
ity of belief-which is fifteen times more frequent. Thirty of the fifty
subjects chose two work partners, one white and one Negro, both of
whom agreed with the subject, as compared with only two subjects
who chose on the basis of similarity of race.
Even though we have no direct pre-test and post-test data showing
that there has been an actual change of behavior in the particular
individuals studied here, I would nevertheless regard these data as
relevant to the issue of behavioral change not preceded by attitude
,change. For the choice of work partners shown in Table 1 is not
what we would ordinarily expect from fifty lower-class persons looking
for a job, given the harsh facts of social discrimination in contem-
porary American culture. The data presented in Table 1 represent
only a small portion of a larger body of data suggesting that the ob-
served absence of racial discrimination is a function of the subject's
knowing the stand taken by a Negro or white on an important issue.
Assuming that there has been a change in behavior from discrimina-
tion on the basis of race to discrimination on the basis of belief, we
are again not required to posit any changes in attitudes underlying
that behavior (although such changes can come about subsequently,
as dissonance theory suggests). We can more simply understand such
behavior as arising from an interaction between two attitudes, acti-
vated by an object encountered within a situation in which the acti-
vated attitude-toward-situation far outweighs in importance the ac-
tivated attitude-toward-object.
As a final example of behavioral change occurring without under-
lying attitude change, let me discuss an as yet unpublished study by
Jamias and Troldahl.22 These investigators were studying differences
in willingness to adopt new agricultural practices recommended by
agricultural extension agents as a function of personality and social
22 J. F. Jamias, and V. C. Troldahl, "Dogmatism, Tradition and General Innova-
tiveness," 1965, unpublished manuscript.

system. The frequency of adoption of recommended agricultural prac-
tices-the dependent variable-was measured by a series of questions,
each designed to determine which of several alternative procedures
the dairy farmer followed in the day-to-day management of his farm.
On each question, one of the alternatives was the one recommended
by agricultural extension agents. Personality differences in receptivity
to new information were measured by the Dogmatism Scale, and social
system differences in receptivity were determined by identifying two
types of rural townships in Michigan, one type identified as high and
the other as low in their "value for innovativeness." The two types of
social systems were readily identified by agricultural extension agents
on the basis of the generally positive or negative attitude of the people
in the townships toward extension activities, size of attendance at ex-
tension meetings, etc. The results are shown in Table 2.
TABLE 2
MEAN ADOPTION RATE BY HIGH AND Low DOGMATIC Groups LIVING IN SOCIAL
SYSTEMS HIGH AND Low IN "VALUE FOR INNOVATIVENESS"
Social Systemin Which Valuefor Innovativenessis

Low High
Low dogmatism group 7.3 6.2
High dogmatism group 4.9 6.8
Correlationbetween dogmatism
and adoption rate -.40 -.09
Statistical analyses show a highly significant interaction between re-

ceptivity in the personality system and receptivity in the social sys-
tem. In the social system having a low value for innovativeness, the
correlation between scores on the Dogmatism Scale and adoption rate
is -.40; in the social system having a high value for innovativeness,
the correlation is -.og. Highly dogmatic persons (scoring above the
median on a modified form of the Dogmatism Scale constructed by
Troldahl and Powell)23 living in social systems having a high value
for innovativeness more frequently adopt recommendations of agri-
cultural extension agents than highly dogmatic persons living in so-
cial systems having a low value for innovativeness. But low dogmatic
subjects, regardless of the social system in which they live, have a rela-
tively high adoption rate for new practices recommended by agricul-
tural experts.
Table 2 shows that behavioral changes in highly dogmatic persons
are the result of compliance or identification with social norms, and
23 V. C. Troldahl and F. A. Powell, "A Short-form Dogmatism Scale for Use in
Field Studies," Social Forces, Vol. 44, 1965, pp. 211-214.

542 MILTON ROKEACH
behavioral changes in low dogmatic persons are the result of a gen-
eralized receptivity to new information, which is routinely internalized
according to its intrinsic correctness and usefulness. Again, no change
of underlying attitude need be postulated to account for behavioral
change, either in the unreceptive highly dogmatic subjects or in the
more receptive low dogmatic subjects. Thus, the results suggest that
we can produce changes in the behavior of different individuals
through knowledge of personality organization, that is, a knowledge
of how a particular situation will activate different beliefs and atti-
tudes in persons who vary in the structure of their belief systems.
In summary, I have tried to suggest that a behavioral change (and
this includes an expressed opinion change) may be determined by a
change in attitude-toward-object, or in attitude-toward-situation, or
both, or neither. I have concentrated mainly on behavioral changes
that do not involve any kind of underlying attitude change, and have
cited various instances of behavioral change in real life and in the
laboratory which are attributable to compliance, demand character-
istics, evaluation apprehension, the activation of salient beliefs and
attitudes within the context of on-going activity, or the activation
of different beliefs and attitudes in persons with differing personality
structures. All the illustrations cited have, I believe, a common
thread. They all involve expressed opinion or behavioral changes that
can be analyzed and reduced to two component attitudes, A, and A.,
interacting within a figure-ground relationship, carrying differential
weights, and affecting a behavioral outcome in proportion to their
relative importance with respect to one another. And all these ex-
amples suggest that many kinds of behavioral change can be brought
about by learning which attitude-object to combine with which situ-
ation, which attitudes are activated by attitude-object and situation in
different personalities, and which outcome to expect from such inter-
actions in different personalities.
Before terminating this portion of the discussion, however, let me
mention one other relevant consideration that has thus far been al-
together overlooked in contemporary theory and research on attitude
change. If expressed opinion changes may be observed when there has
been no underlying attitude change, then the converse is also true: an
absence of expressed opinion change may be observed even after a
change in underlying attitude has already taken place. For example,
a dutiful son may continue to express pro-religious sentiments even
after he has changed his underlying attitudes toward religion, in
order not to hurt his parents; a disillusioned Communist may con-
tinue to engage in Party activities because he is afraid of social ostra-
cism; a person may continue to say "I love you" even after he has

stopped loving. All these examples suggest a possible constancy in ex-
pressing an opinion despite a change in attitude. The conditions un-
der which such phenomena will occur deserve more study than they
have received so far.
METHODS FOR ASSESSING ATTITUDE CHANGE
Thus far, I have tried merely to suggest that a change in behavior

or expressed opinion may arise in different ways, and may or may not
involve a change in underlying attitude-toward-object or attitude-
toward-situation. I have also tried to suggest that the classical para-
digm employed in experimental studies of opinion change cannot
yield information about attitude change as such. The question is,
then, how should we proceed if we wish to ascertain whether a given
attitude has undergone change, or, at least, if we wish to increase the
probability of correctly inferring that a given expressed opinion
change represents an attitude change.
I would like to discuss three methods and I will try to illustrate
each with some relevant research studies.
1. Test for opinion change across different situations. If verbal or
nonverbal behavior toward an object is observed in only one situation
following an experimental treatment, we hardly have a basis for in-
ferring a change of attitude. Orne has pointed out: "If a test is given
twice with some intervening treatment, even the dullest college stu-
dent is aware that some change is expected, particularly if the test is
in some obvious way related to the treatment."24 But the more post-
test situations in which a changed opinion is manifested, the more
confident we may be that a change in attitude has actually taken
place. Any experimental study of expressed opinion change, if it is
to qualify as a study in attitude change, should demonstrate the ex-
istence of change in at least two reasonably different situations.
One research design in which there are several post-tests of opinion
change is Kelman's study of three processes of social influence. Opin-
ion change was assessed in three different post-test situations, under
conditions of surveillance and salience, nonsurveillance and salience,
and nonsurveillance and nonsalience. In the last condition, the post-
test was administered "one or two weeks after the communication ses-
sion, in a different place, under different auspices, and by a different
experimenter."25 Kelman has shown that subjects who were given the
experimental treatment designed to favor compliance manifested an
opinion change only under conditions of surveillance and salience,
thereby suggesting that there was no change in underlying attitude.
24 Orne, oP. cit., P. a779.
25 Kelman,"Compliance,Identification,and Internalization,"p. 56.

544 MILTON ROKEACH
Subjects who were given the internalization treatment, however,
manifested opinion changes in all three post-test situations.
Another illustration of repeated post-tests is my study on The Three
Christs of Ypsilanti, which is concerned with underlying changes in
delusional attitudes and beliefs among three paranoid schizophrenic
patients who believed they were Jesus Christ. After several months of
confrontation with each other over who was the real Christ, the
youngest of the three, Leon Gabor, announced one day that he was no
longer married to the Virgin Mary. Our problem was to determine
whether or not this change in expressed opinion represented an un-
derlying change in delusional belief. Our confidence that it did in-
deed represent such a delusional change increased as Leon Gabor re-
peatedly told us during the next few weeks and months in various
contexts that he was about to get divorced, then that he was divorced,
then that he had a new brother, then that his brother had married
the Virgin Mary, then that he himself was about to re-marry, then
that he had re-married another woman, and so forth. Had we relied
only on one "post-test" expression of a changed opinion about whom
Leon was married to, our claim of a change in delusion would have
been extremely weak.
In the experimental literature on opinion change, one may ifind an
occasional study in which two or more post-test situations are em-
ployed, but, unless I am mistaken, such studies are the exception
rather than the rule. In the typical experiment the post-test is given
only once, usually within a short time after the experimental treat-
ment; thus the meaning of the expressed opinion change in relation
to attitude change is highly equivocal.
The preceding remarks concern the assessment of change in under-
lying attitudes-toward-object in several post-test situations. I have not
said anything about assessing change in underlying attitude-toward-
situation because this type of attitude is typically not employed in
experimental studies of attitude change. But the principle would
seem to be the same: instead of testing for a change in opinion to-
ward an attitude-object across situations, a change in attitude-toward-
situation would be tested by substituting various attitude-objects
which might be encountered within that situation.
2. Test for changes of several opinions in one situation. In the
classical paradigm only one opinion is pre-tested, experimentally
treated, and then post-tested. Evaluation of the nature of the opin-
ion change is difficult because the expressed opinion is compared only
with itself. But suppose several opinions that are thought to be system-
atically related to one another in some way were pre-tested, experi-
mentally treated, and then assessed for change all in one post-test

situation? Suppose, further, that we find differential opinion changes
and that these differential changes are systematically related to one
another in the same way as the original opinions are related to one
another?
In an as yet unpublished study by Rokeach, Reyher, and Wiseman,
over fifty beliefs ranging from highly central beliefs about self-iden-
tity to highly inconsequential ones were subjected to change through
hypnotic induction. Immediately prior to the hypnotic trance, the
subject is pre-tested on these beliefs and attitudes, indicating his
agreement or disagreement on an ii-point scale ranging from -5 to
+5-
The subject is then hypnotized. By suggestion, the experimenter
attempts to alter one of five types of beliefs varying in centrality. It is
suggested to the subject that he will experience an impulse to respond
to such statements in a manner opposite to the way he had responded
previously. He is then asked to respond while still under hypnosis to
one of the five types of statements to which he was subjected in the
experimental treatment. He then responds again to all the statements,
first under hypnosis, and again immediately after hypnosis.
On several later occasions, about a week apart, the same subject is
brought back for additional testing. He is re-hypnotized and, once
again through suggestion, attempts are made to alter, in turn, each of
four other kinds of beliefs varying in centrality-peripherality. For
example, in the first experimental session an attempt is made to alter
through suggestion a subject's primitive beliefs about self-identity; in
the second session, his beliefs about authority; in the third session, be-
liefs which he has derived from external authority, which we call pe-
ripheral beliefs, and so on, for the remaining sessions, until each of
five kinds of beliefs has, in turn, been subjected to the hypnotic ma-
nipulation.
It may be helpful to give a few of the belief statements to which the
subject is asked to respond before, during, and after hypnosis. Two
belief statements are: "My name is [real name]," and "My name is
Ivan Petrov." In the pre-hypnotic test the subject would, of course, be
expected to agree strongly with the first statement and disagree
strongly with the second statement. Another pair of statements reads:
"My mother's name is [real name]," and "My mother's name is Olga
Petrov." Again, the subject can normally be expected to accept
strongly the first statement, and to reject strongly the second state-
ment. In the same way, we presented other biographical statements,
paired with fictitious biographical statements-that the subject's
father is Boris Petrov, that the subject was born and lives in Moscow,
that he is an atheist and member of the Communist Party. Our sub-

546 MILTON ROKEACH
jects, who were neither atheist nor Communist, strongly reject each
of these latter statements in the pre-hypnotic session and strongly ac-
cept the parallel statements based on their real biographical data.
The belief statements just described are samples of one kind of
primitive belief, ones we have identified as taken-for-granted beliefs
supported by a unanimous social consensus. Other statements are
presented to sample a second kind of primitive belief, beliefs that, we
assume, do not depend on social support. For example: "I believe my
mother loves me." "I can remember having been very curious and
sexually excited about the thought of seeing my mother without
clothes on." Statements are then presented to test a third kind of be-
lief ranging along the central-peripheral dimension we call author-
ity beliefs, for example, "The philosophy of Nikita Khrushchev is
basically a sound one and I am all for it." Similarly worded state-
ments are presented about such positive and negative authority fig-
ures as Hitler, Faubus, Eisenhower, Kennedy, Lincoln, and Castro.
Some examples of the fourth kind of belief, which we call pe-
ripheral beliefs, beliefs derivable from authority, are: "The Russians
were justified in putting down the Hungarian revolt in 1956," "The
Gettysburg Address does not really say anything important," and "I
think this country would have been better off if the South had won
the Civil War."
Finally, there are inconsequential beliefs, beliefs which, if changed,
are not expected to produce any significant changes in other beliefs:
"There is no doubt in my mind that Elizabeth Taylor is more beauti-
ful than Dinah Shore," and "I think summer is a much more enjoy-
able time of the year than winter."
This experimental situation was designed to determine whether dif-
ferent kinds of beliefs ranging systematically along a theoretically
postulated central-peripheral dimension can be changed through hyp-
nosis and, if so, which kinds of beliefs will be the easiest to change
and which the most difficult. In other words, we were interested in
determining whether differential changes in several beliefs will oc-
cur as a result of a single experimental treatment.
This is not the place to present the full results, which will be con-
tained in a technical report presently in preparation. But Table 3
shows that the hypothesized differential changes in many expressed
opinions were obtained as a result of one experimental treatment. As
expected, the amount of change varies in inverse proportion to cen-
trality of belief, the primitive beliefs changing least as a result of
hypnotic suggestion, authority beliefs changing more, peripheral be-
liefs changing still more, and inconsequential beliefs changing the
most. These differential results were obtained on three occasions, I

might add: immediately after the hypnotic suggestion, a short time
later while the subject was still under hypnosis, and immediately after
he was awakened. It would be difficult to attribute such differential
changes to compliance or to the operation of any other post-test situ-
ational variable, because the post-test situation, and the attitude ac-
tivated by that situation, is a constant. Yet opinion changes vary sys-
tematically in the same way the original opinions vary, namely, as a
function of centrality-peripherality. To account for such differential
changes we would have to infer that they are manifestations of differ-
ential changes in underlying attitudes.
TABLE 3
MEAN CHANGES UNDER HYPNOSIS OF FIvE TYPES OF BELIEFS
VARYING IN CENTRALITY
As an Immediate Post-test1:
Result of Hypnotic While Still Post-test2:
Type of Belief Suggestion underHypnosis Post-hypnosis
Primitive, Type 1
(unanimous social
support) 2.48 2.66 2.56
Primitive, Type 2
(no social support) 3.50 3.52 2.82
Authority 3.54 3.47 2.87
Peripheral 3.65 3.62 3.55
Inconsequential 4.25 4.08 3.48
Another illustration from our own research program where several

opinion changes are obtained in one post-test situation may again be
taken from The Three Christs study. As a result of experimental con-
frontation with others over the issue of who was the real Christ, Leon
Gabor reported that he had changed his name from Dr. Domino Dom-
inorum et Rex Rexarum, Simplis Christianus Pueris Mentalis Doktor,
the reincarnation of Jesus Christ of Nazareth to Dr. R. I. Dung. Again
the problem was whether this expressed opinion change regarding a
new identity represented a true underlying change. And, again, our
confidence increased that it did indeed represent a true change when
he expressed a network of additional differential changes in opinion
that were wholly consistent with the expressed change of name.
3. Test for other behavioral changes accompanying a given opinion
change. If a single expressed opinion change truly represents a change
in underlying attitude, it is reasonable to expect that such a change
will be accompanied by other changes-cognitive, affective, or be-
havioral-which theoretically should be related to the change in atti-
tude. It is difficult to believe that a change in expressed opinion repre-
senting a true change in attitude would have no other behavioral

548 MILTON ROKEACH
manifestations. That Festinger and Cohen find virtually no studies in
the experimental literature showing behavioral change following an
attitude change only serves to reinforce my suspicion that most current
experimental studies on opinion change do not deal with true attitude
change, but with superficial opinion changes.
But I would like to draw attention to some data, again from our
own research program, which illustrate that behavioral changes fol-
lowing or accompanying opinion changes can indeed be obtained.
In The Three Christs study, one may note many changes in Leon
Gabor's behavior following changes in expressed opinion. After he
verbally claimed to have a new wife, Madame Yeti Woman, his be-
havior with respect to money changed. He accepted, handled, and
spent money when it allegedly came from Madame Yeti; these were
actions we had never before observed. Furthermore, Leon Gabor, fol-
lowing suggestions allegedly coming from his new wife, changed the
song with which he always opened the meetings on days he was chair-
man from America to Onward Christian Soldiers, a permanent be-
havioral change. Again, these behavioral changes serve to increase our
confidence in the inference that Leon's expressed opinion change
represented a true change in underlying delusion.
Perhaps more impressive in this respect is a study of changes in ex-
pressed values by Kemp.26 His subjects were all religiously minded
persons enrolled in a special training curriculum designed to pre-
pare them for positions as Boy Scout executives or YMCA or YWCA
secretaries. Kemp was interested in determining whether changes in
values and in behavior were a function of personality. The subjects
were given the Allport-Vernon Scale of Values while still in college
in 1950. Six years later they were contacted again, were given the
Dogmatism Scale, and were retested with the Allport-Vernon Scale. As
shown in Table 4, closed, middle, and open-minded subjects all ex-
pressed identical value patterns in 1950. But in 1956, the rank order
of values remained the same only for the middle group, and had
changed for the closed and open groups:
Although religious values were still predominant in all groups, the closed
group increased in political and economic values and decreased markedly in
social values. The open group remained unchanged in its religious and social
values but increased in theoretical values and decreased in economic and
political values... . The vocational choices .. . follow closely these changes or
non-changes in value patterns. Roughly 70 per cent of the middle group be-
came Boy Scout executives as planned, or entered closely related professions.
But most of the open and closed subjects changed their vocational choice
after leaving college; the open subjects more frequently entered vocations
26 G. C. Kemp, "Change in Values in Relation to Open-Closed Systems," in
M. Rokeach, The Open and Closed Mind.

requiring more advanced professional training in careers involving social
welfare, and the closed subjects more frequently entered military and com-
mercial careers of an administrative nature.27
TABLE 4
RANK ORDER OF IMPORTANCE OF SIX VALUES FOR THE TOTAL GROUP AND FOR
OPEN, MIDDLE, AND CLOSED SUBGROUPS IN 1950 AND IN 1956
(N) Religious Social Political Economic Theoretical Esthetic

Test 1950:
Open (25) 1 2 3 4 5 6
Middle (54) 1 2 3 4 5 6
Closed (25) 1 2 3 4 5 6
Total
group (104) 1 2 3 4 5 6
Retest 1956:
Open (25) 1 2 4 6 3 5
Middle (54) 1 2 3 4 5 6
Closed (25) 1 5 2 3 4 6
Total
group (104) 1 2.5 2.5 5 4 6
SOURCE: M. Rokeach, The Open and Closed Mind, New York, Basic Books,
1960, p. 339.
That vocational changes accompanied changes in scores on the All-
port-Vernon Scale of Values strengthens the likelihood that the
changes in expressed values represent real changes in underlying
values.
CONCLUSION
In closing, I would like to concede that the point of view I have

presented will probably not appeal to those who, disliking to think in
terms of genotypes and phenotypes, would insist on equating an atti-
tude with its operational measurement by some opinion questionnaire.
But, starting with a conception of attitude as a hypothetical construct,
I have proposed that the literature on opinion change does indeed
tell us a good deal about the social influence variables and cognitive
processes affecting changes in expressed opinion. But this is a litera-
ture concerning changes which, in the main, seem to be localized in
the region of the lips and do not seem to affect the mind and heart,
nor the hands and feet.
The view developed here on the relations existing among attitude,
attitude change, and behavioral change is incomplete, however. It
has neglected other kinds of change which must sooner or later be
27 Ibid., pp. 345-346-

550 MILTON ROKEACH
considered if there is to be a truly systematic consideration of ante-
cedents and consequents of attitude and behavioral change, namely,
the problem of changes in values, in ideology, in total belief systems,
in therapy, and in personality. It seems to me that contemporary
theory and research on opinion change, dealing as they typically do
with changes in single and isolated expressions of opinion, and se-
lecting as they typically do opinions that are, as Hovland points out,
"relatively uninvolving"28 and thus easily capable of manipulation
within the context of a relatively brief experimental session, have
somehow lost touch with such broader issues. I hope that the present
paper will serve as a contribution toward our regaining contact with
these broader and more significant kinds of change, which may affect
and be affected by our everyday life in local, national, and inter-
national affairs.
28 Hovland, op. cit.


Rokeah

Uploaded by

Copyright:

Available Formats

Rokeah

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Rokeah

Uploaded by

Copyright:

Available Formats

American Association for Public Opinion Research

Attitude Change and Behavioral Change

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

The literatureon opinion and attitude change continues to probe deeper.

* This is one of a series of papers on attitude organization and modification sup-

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

ATTITUDE AND BEHAVIOR

For purposes of this discussion let me offer the following coordi-

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

COGNITIVE INTERACTION BETWEEN TWO ATTITUDES

It is not enough merely to assert that social behavior is a function

6 H. C. Kelman expresses a similar view: "The attitudes expressed by an individual

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

ATTITUDE CHANGE AND BEHAVIORAL CHANGE

The proposition that behavior is always a function of two inter-

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

13 Festinger, op. cit., p. 405.

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

17 H. C. Kelman, "Compliance, Identification, and Internalization: Three Processes

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

18 M. T. Orne, "On the Social Psychology of the Psychological Experiment: With

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

Science, Vol. 151, 1966, pp. 167-172.

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

(1) (2) (3) (4) (5) (6)

22 J. F. Jamias, and V. C. Troldahl, "Dogmatism, Tradition and General Innova-

tiveness," 1965, unpublished manuscript.

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

Social Systemin Which Valuefor Innovativenessis

Statistical analyses show a highly significant interaction between re-

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

METHODS FOR ASSESSING ATTITUDE CHANGE

Thus far, I have tried merely to suggest that a change in behavior

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

Another illustration from our own research program where several

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

(N) Religious Social Political Economic Theoretical Esthetic

In closing, I would like to concede that the point of view I have

27 Ibid., pp. 345-346-

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

This content downloaded from 150.108.161.71 on Sun, 1 Sep 2013 11:44:12 AM

You might also like