CONFIDENCE, DO'GWFISTM, DIS$ONANCE,
AND RESPONSE SET

Thesis {or Hm Degree of M. A.
MICHIMN STATE UNIVERSHY
Miles E. Simpson

1965

THESIS

 

 
 
  

LIBRAR 1/
Michigan State

University

“

 

ABSTRACT

CONFIDENCE, DOGMATISM, DISSONANCE,
AND RESPONSE SET

By Miles Ea Simpson

This study undertakes to explore possible person-
ality correlates of generalized over—confidence in judgment
in an ego—involving task (‘Defensive Confidence“). DOgma-
tism or closed—mindedness was hypothesized to be related to
‘Defensive Confidence'. Directional response set, eig.,
positive setting and negative setting, were hypothesized to
be unrelated to Dogmatism as conceptualized by Rokeach; in—
stead directional response set was seen to be an artifact
of the Likert Scale and not related to CDefensive Confidence”o
Confidence in judgment scores were taken after each
item of the Pettigrew Category Width Scale which was given
under ego-involving instructionso High confidence in judg-
ment, due to the difficulty level of the Pettigrew Category
Width Scale, was considered evidence of 'Defensive Confi-
dence'. DOgmatism was measured by the Rokeach Dogmatism
Scale and directional response set by the Bass Social Acquies-
cence Scale. Measures of confidence in belief were taken for

both belief scales”

Miles E0 Simpson

A U-shaped distribution was found between confidence
in judgment and the Dogmatism Sealer This was contrary to
the main hypothesis°

To explore the U-distribution, positive and nega—
tive setters were hypothesized to be closed-minded) Both
groups will seem to be acting in terms of the irrelevant
aspects of the testing situation by responding to the
structure or ‘field forceB of the items rather than their
contento The data were reanalyzeda

Negative setters, positive setters, and non-response
setting dogmatics were expected to exhibit ‘Defensive Confi-
dence' while only the non-response setting non—dogmatics
were expected not to exhibit defensive confidenceo The two
belief scales were combined to form a more accurate measure
of directional response set and the types were isolated by
arbitrarily quartering the sampler

Results strongly support the revised hypothesiso
Positive response—setters, negative response—setters, and
non—setting dogmatics were significantly more confident in
their judgments than non-setting nonndogmatics.

As positive setters and negative setters, as well
as, non—response setting dogmatics manifest "Defensive
Confidence" and both directional response setting and
"Defensive Confidence" were hypothesized to manifest closed

mindedness, closed mindedness as defined by Rokeach (1960)

Miles E. Simpson

was interpreted to be a generic concept which encompasses
the more particular concepts of dogmatism, positive setting,

and negative setting.

APPROVED:__€) E ”Lg/5% __

DATE : CZhw/rq ] é /7 65
0 i f7

CONFIDENCE, DOGMATISM, DISSONANCE,
AND RESPONSE SET

BY

Miles E. Simpson

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Department of Psychology

1965

ACKNOWLEDGMENT

The author wishes to express his gratitude and
appreciation to Dr. C. F. Wrigley, chairman of his com-
mittee, for his guidance in this research and for his

assistance in the development of this manuscript.

ii

TABLE OF CONTENTS

IN‘I‘RODUCT ION O 0 0 O 0 C O C O O O O O 0 0 O 0 0

Confidence in Judgment . . . . o a . o .
Confidence in Belief . . . . a o o . o .
Ego Involvement and COgnitive Dissonance
Defensive Confidence . . . . . . o . . .
Open and Closed Mindedness . . . o . . .
Response Set . . . . . . . o . o . . . 0

DESIGN OF THE STUDY . o . . . . o . . . o o o .

SUbjeCtS I; o O O o O 0 O O O O O O O 0 O O 0
Instruments . . o . . . . . . . . o o . . .

The Pettigrew Category Width Scale . . .
Confidence in Judgment . . . . o . . a .
The Rokeach Dogmatism Scale . . . . . .
The Bass Social Acquiescence Scale . . .
Confidence in Belief . . . . . . o . . .
Extremeness Set . . . . o . . . . . . .
Summary of Variables . . . o . o . o . .

AnalYSiS O 0 0 CI 0 0 0 0 0 O 0 O 0 0 0 0 0 O
RESIJI‘TS 0 O O 0 O O O O O 0 O O O 0 0 O O 0 O 9
Means, Standard Deviations, and

Reliabilities . . . o . . . . . o . .
Factor Analysis . . . . . . . . . o o .
Non Linear AnaIySis . . . . . . . . . .
Partial Correlation . . o . o o . . . .
Reanalysis . . . . . . . . . . . . . . .

DISCUSSION 0 O O O O O O 0 O C O O 0 O O O O O 0

Theoretical . . . . .-. . . . . . . . .
Me thOdOlogy 0 O O O O 0 O 0 O O O O 0 6

SIJMMARY O 0 0 0 0 0 0 0 O 9 9 O O O O 0 0 O O .

BIBLIOGRAPHY . . o o . o . o . . o o . . . o o 9

iii

Page

i—J

I-‘KOCDWOJI-J

22
22

22
22
23
24
24
25
25

26
27
27
3O
33
34
35
41

41
45

47

51

LIST OF TABLES

Table Page
1. Means, Standard Deviations and Odd—Even
Reliabilities . . . . . . . . . . . . . . . 28
2. Correlations of the Test Battery . . . . . . . 29
3. Quartimax Factor Loadings . . . . . . . . . . 31
4. Dogmatism and Confidence in Judgment—-Mean
category Score 0 O O O O O O 0 O 0 O O O O O 34
5. Types of Respondents and Confidence in

Judgment . . . o . . . . . . . . . . . . . . . 40

iv

INTRODUCTION

The purpose of this study is to explore the relation-
ship between confidence in judgment, dogmatism, and response
set.

Confidence is defined operationally to be the extent
to which a person reports his action, decision, or response
to be appropriate. It is possible to use confidence in
terms of 'self—worth' or the individual‘s sense of adequacy
in dealing with the world. While this may be related to
'confidence in judgment' as defined here, we will only con—
cern ourselves with the former as there has been little
evidence relating the two. Therefore, decisiveness of action,
vacillation, etc., will not be considered measures of confi-
dence in judgment unless found to relate to the present
measure.

The subject's confidence will be elicited by the
method of 'subjective probabilities' where the subject
selects a number ranging from zero to one hundred which he
judges to approximate the probability of his response being

'right' or 'correct.'

Confidence in Judgment

Confidence in judgment generated a great deal of

work in the past; yet there appears to have been but one

major study relating it to any type of personality variable.

Block and Peterson (1955), in a study of confidence
in judgment and personality, found over-confidence to be
related to their criteria of dogmatism. Using a psycho-
physical apparatus and design developed by Festinger (1943),
they tested 53 'normal' army officers and then compared the
results with a Q-sort made by clinical psychologists who had
been observing the subjects in a 'living in' situation. It
was found those subjects overestimating their performance
(i.e., were over-confident) tended to be dogmatic, and those
cautious tended to be introspective and self-abasing. Indi-
viduals with realistic confidence in their decisions appeared
to be more self-reliant. Even though the extent of over- or
under-confidence was not considered, the preceding results
were significant.

Confidence in judgment arose out of the psychophysi-
cal tradition, with the first studies made by Fullerton and
Cattell (1892), Griffing (1895), Henmon (1911), Hollingsworth
(1913) and Metcalf (1917). Here the emphasis was on the
subject‘s estimate of his performance in a situation which
he thought to have an objectively right and wrong answer.

In general, these early studies found a strong relationship
between accuracy and confidence.

Johnson (1939) found a generality between confidence
in three psychophysical tasks and a vocabulary test; also,

he found a correlation between latency and lack of confidence.

In changing the instructions (i.e., emphasizing speed or
accuracy), changes were brought about in latency but the
changed instructions had little effect on confidence.

Adams and Adams (1958) linked confidence in judgment
and learning by training subjects to improve their confidence
estimates. The confidence judgments were made on psychophysi-
cal tasks. Subjects were told of their performance after
completing a series of tasks. Their performance was com-
pared with a control group who simply repeated the tasks with
no feed—back on their performance. The experimental group
improved while the control group's performance showed no

improvement.

Confidence in Belief

Another line of research beginning with Williamson
(1892), Sumner (1898), and Okabee (1910) studied confidence
in belief. The first major study (F. Allport and Hartmann,
1925) showed extreme opinions to be related to extreme con-
fidence. Both radicals and reactionaries had higher confi-
dence in controversial issues than did conservatives and
liberals. Also, radicals and reactionaries had a higher
and inappropriate estimate of their I.Q.'s.

Allport and Hartmann concluded that confidence in
belief was related to the extremeness of the person's opin-
ion with extreme confidence compensating for a feeling of

insecurity.

To test this conclusion, Johnson (1940) took two
uncorrelated Thurstone attitude scales (war) and censor—
ship) and compared them in terms of confidence ratings. The
correlation between confidence on the two scales was 0.72 i
0.04 (raw) or 0.84_: 0.02 (corrected), supporting Johnson's
hypothesis of a general trait of confidence. It is impor-
tant to note that Johnson never compared confidence in
belief with confidence in judgment.

Thouless (1935 p. 30) points to what he calls 'the
tendency to certainty' in religious belief which he finds
no less strong in nonbelievers than in believers. He
concludes:

. . . there is a real tendency amongst people to
certainty. Doubt and skepticism are for most
peOple unusual, and I think an unstable state of
mind . . . the ability to adopt the attitude of
partial belief or to hold prOpositions with less
than full certainty is rare, and its acquirement
should be one of the aims of a liberal education.

McGuire (1960) studied the effects of logical reason-
ing and personal desires in attitude expression by having
subjects rate the desirability and probability of truth of
the premises and conclusions in a set of syllogisms.

The correlation between the judged truth and desir—
ability of the items was 0.40. Further experimental manipu-

lation made either desire or belief more salient and re—

sulted in a shift away or toward logical conSistency.

Ego—Involvement and Cognitive
Dissonance

 

Klein and Schoenfeld hypothesized that the general—
ity in confidence found by Johnson (1939) was due to the
subject's ego-involvement in the test which produced defen—
siveness. To test this hypothesis, Klein and Schoenfeld
first took a group of subjects aside in the 'low involvement
condition' and told them they were observers in the experi-
ment and they were to watch the other students' reactions
while pretending to take the group of psychological tests.
In the 'high involvement' condition, the group was told they
were taking important I.Q. tests that would be recorded. No
generality was found with the low involvement condition while
a generality across the measures appeared with the high in-
volvement condition. The problem then becomes: why does
high ego—involvement result in a generality in confidence in
judgment across uncorrelated psychOphysical measures?

According to Klein and Schoenfeld, subjects entering
a psychological testing situation are on their guard for
evaluations of themselves in terms of good or bad, intelli-
gent or not, balanced or unbalanced, healthy or unhealthy.
Hence, they become quite apprehensive and ego-involved lest
they have a bad label put on them. This leads to the raising
of confidence to reduce this apprehension and enxiety.

Using the notion of post—decisional cognitive dis-
sonance as formulated by Festinger (1957), a more complete

hypothesis can be develOped. According to Festinger,

cognitive dissonance occurs when a person confronts a situa-
tion where one set of cognitive elements follows from the
obverse of another set. 0n the other hand, two clusters of
cognitive elements are consonant when one follows logically
from the other. Cognitive dissonance motivates the indi-
vidual to reduce it and avoid its recurrence. Thus, presence
of cognitive dissonance brings about pressure for change
which can be reduced by changing either cognitive cluster

by:

1. . . . changing one or more of the elements
involved in dissonant relations.

2. . . . adding new cognitive elements that are
consonant with already existing cognition.

3. . . . decreasing the importance of the ele-
ment involved in the dissonant relations
(p. 264).

Festinger describes post—decisional dissonance where
. . . the cognitive elements corresponding to
positive characteristics of the rejected alter-
natives, and those corresponding to the negative
characteristics of the accepted alternative are
dissonant with the knowledge of the action that
has been taken (p. 262).

Post—decisional dissonance may be reduced by increas-
ing the attractiveness of the chosen alternative, decreasing
the attractiveness of the unchosen alternative or both (p.
264).

Further, the amount of dissonance created by two

conflicting sets of cognitions will be a function of the

resistance to change of the weakest element.

A slightly different position is taken by Brehm and
Cohen (1962) who introduce commitment as a condition under
which ". . . the specification of dissonance and the manner
in which it is likely to be reduced are relatively unequivo-
cal. . . ." Evidence for this hypothesis was gathered by
Deutsch, Krauss, and Rosenau (1962). They hypothesized
that:

The objects involved in the choice have relevance
to the post-decisional dissonance only in so far
as they determine what the act of choices signifies
concerning the chooser (e.g., to what extent and
how reliably it signifies that the chooser is
"intelligent," "prudent," "moral," "tasteful,"
"nice").
In essence, the production of post-decisional dissonance
requires the existence of ego—involvement in the decision.
The findings supported the hypothesis.

Both Deutsch, Krauss, and Rosenau (1962) and Klein
and Schoenfeld (1941) emphasize the importance of ego—
involvement in a post—decisional situation which leads us
to hypothesize that when a subject has sufficient ego—
involvement in a task he will experience dissonance and
attempt to reduce it. He has three choices: devalue the
situation or reduce ego—involvement (i.e., 'I.Q. tests don't
measure anything' or 'what is one item'), reorder his self
concept (i.e., I'm not as smart as I thought), or positively
reinterpret his performance. The latter leads to over con-
fidence. Festinger (1957) found a final confidence decision

increased from that reported on a prior tentative confidence

estimate regardless of whether the last piece of evidence

presented supported or contradicted evidence given prior

to the tentative confidence estimate. In this study, some

but not all subjects increased their confidence. This over-
confidence was interpreted to be the result of post—decisional

dissonance.

Defensive Confidence

 

It seems that certain subjects will be over-confident
in the face of public contrary evidence, e.g., Festinger
(1957), and this is general across uncorrelated judgmental
measures under ego-involvement, e.g., Klein and Schoenfeld
(1941). Festinger saw post-decisional dissonance as a
sufficient condition for increased confidence and post—
decisional dissonance arising after any decision; but, the
Deutsch, Krauss and Rosenau (1962) study demonstrated the
importance of ego—involvement for certain post decisional
dissonance 'effects' and Klein and Schoenfeld demonstrated
the same for confidence in judgment.

The employment of over-confidence in ego—involving
judgments as a defense against 'dissonance' will be referred
to as 'defensive confidence' in this study. Defensive con-
fidence will be manifested by certain subjects when they are

sufficiently ego—involved.l

 

1It would be useful to know whether the attempt to
induce ego—involvement or evaluation apprehension occurs in
almost all subjects only some of whom react by evaluating
their confidence or whether only a certain portion of the
subjects became ego-involved but almost all of these diSplay
over-confidence.

Defensive confidence seems quite general over a
variety of situations provided ego-involvement exists and
therefore, is best characterized as a trait. G. Allport
(1939) defines a trait as

. . . a generalized and focalized neuropsychic
system (peculiar to the individual) with the
capacity to render many stimuli functionally
equivalent, and to initiate and guide consistent
(equivalent) forms of adaptive and expressive
behavior (p. 295).
In a confidence in judgment situation, the items judged are
functionally equivalent in that they arouse 'dissonance'

which is handled by a particular adaptive function, over-

confidence.

Open and Closed Mindedness

One of the most provocative and productive areas in
social psychology has been that of personality structure.
Beginning with the seminal work of Adorno (1950) "The
Authoritarian Personality," the area has accumulated a mass
of research. In "The Authoritarian Personality," as in psy-
choanalytic theory, Adorno emphasized the social environment
in which a person grows. What develops is

. . . a structure within the individual, something
which is capable of self-initiated action upon the
social environment and of selection with respect
to varied impinging stimuli, something which al-
ways modifiable is frequently very resistant to
fundamental change. This conceptualization ac-
counts for consistencies of behavior in a wide
variety of situations and the persistence of
ideological trends in the face of contradicting
facts and radically altered social conditions . . .
(1950, p. 6).

10

The focus of "The Authoritarian Personality" was on
the relationship of ideology and personality structure and
in particular the ideology of the extreme right; hence,
ethnocentrism and Fascist tendencies were its first areas
of concern. As Rokeach (1958) points out, the F or Fascist
Scale was generalized to the "Authoritarian Scale." This
resulted in a certain amount of conceptual confusion, be-
cause in the shift from fascism and authoritarianism, there
is an unwitting leap from the particular to the general
(1958, p. 13). Since the high F—scorer was generally more
ethnocentric, anti-Semitic, and politically conservative,
the existence of people who were none of these but intolerant
and authoritarian posed a problem. To rectify this anomaly,
Rokeach sharply distinguishes the content and the structure
of belief systems. He next develOped the concept of open
and closed mindedness which was independent of ideoloqy and
devised the Rokeach Dogmatism Scale to tape this dimension.

In his conceptualization of the structure of belief
systems, Rokeach (1960) delineates a continuum of Open- and
closed—mindedness.

We assume that, in any situation in which a person
must act, there are certain characteristics of the
situation that point to the appropriate action to be
taken. If a person reacts in terms of such character-
istics, his response should be correct or appropriate.
The same situation also contains irrelevant factors,
not related to the inner structure or requirements of
the situation. To the extent that response depends
on such irrelevant factors, it should be unintelligent
or inappropriate . . . this leads us to suggest a

basic characteristic that defines the extent to which
a person's system is Open or closed; namely the extent

11

to which the person can receive, evaluate, and act
on relevant information received from the outside on
its own intrinsic merits unencumbered by irrelevant
factors in the situation arising from within the
person or from the outside. Examples of irrelevant
internal pressures that interfere with the realistic
reception of information are unrelated habits, be-

liefs, and perceptual cues, irrational ego-motives,
power needs, the need for self aggrandisement, the
need to allay anxiety, and so forth. By irrelevant
external pressures we have in mind most particular-
ly the pressures of reward and punishment arising
from external authority . . . (1960, p. 157).

Using this definition, I am defining the relevant
aspects of the confidence in judgment situation as the
performance cues, while the irrelevant aspects of the
situation involve the need to reduce dissonance produced
by the conflict of the perceptual cues and ego needs. Of
the two cognitive clusters, the more open-minded subjects
ego needs will tend to be less resistant to change, but for

the more closed minded individual, the performance cues will

be more modifiable.

Response Set

 

Studies on judgmental tasks have revealed certain
clustering of responses at particular points on a scale
seemingly independent of the content, e.g., Cronbach (1946,
1950). Mathews (1927), for example, reported position
biases that produce responses clustering near the tOp of a
vertical scale and near the left end of a horizontal scale.
Similarly the directional response bias has been noted in

attitudinal studies, e.g., Bass (1955). Also, there is a

12

tendency to use polar extremes, e.g., Osgood and Tannenbaum
(1955) and a tendency to concentrate responses at the mid~
point on the scale, e.g., Lorge (1937).

These non—content factors of the judgment process
may have a profound influence on the reporting of attitudes
and may be responsible for many inconsistencies in results
(Peabody, 1962).

The Likert Scale produces at least two types of
response set: directional (i.e., positive set or acquies-
cence and negative set) and extremeness set. This means
that a person may consistently agree or disagree to items
regardless of their content, e.g., Cronbach (1946); Bass
(1956); Chapman and Campbell (1957). In this study persons
tending toward positive set will be referred to as positive
setters, and persons employing negative set are called nega-
tive setters.

As noted by Lorge (1937) and Osgood and Tannenbaum
(1955), there is a tendency for some subjects to pile up
their responses in the middle of the scale and for some
others to use only the extreme ends of the scale. Peabody
(1962) calls these tendencies extremeness sets.

Since Cronbach's (1946) article on the effects of
response set, considerable attention has been given the
topic. Of positive response set, the most thoroughly ex—
plored form, Christie and Lindauer (1963) state:

One thing can be said about the tendency to agree
to items on personality inventories: there is a

l3

consensus that some people will say yes to some
items without paying much attention to the content
of the item. How many and what sorts of peOple,
and what kinds of items are moot points (1963,

p. 202).

They also point out that as yet there is no evidence which
contradicts Cronbach's contention
. . . the greater the ambiguity of an item the
more likely it is to elicit response set. The
problem is that one man's certainties are often )

another man's ambiguities (1963, p. 202).

Bass (1956) developed a Social Acquiescence Scale

 

made up of aphorisms, chosen for their heterogeneous content,
which had a reliability of 0.91 and correlated moderately
with the Authoritarian Scale. Others followed with reversed
items, measures of response in consistence, e.g., Chapman
and Campbell (1957); Christie, Havel, and Seidenberg (1958);
Jackson, Messick, and Solley (1957); Zuckerman, Norton, and
Sprague (1958). All found evidence of positive response set
and Christie and Lindauer (1963) noted that "positive corre—
lations were reported between every paper and pencil test of
agreement when they were compared" (1963, p. 205).

In their theories, Gage, Leavitt, and Stone (1957),
Zuckerman, Norton, and Sprague (1958), and Leavitt, Hax, and
Roche (1958), all postulated positive response set to be
compatible with authoritarian submission; yet Messick and
Frederickson (1958) have found that the content and acquies-
cence scores from a reversed authoritarian scale did not
correlate, suggesting that they measured independent

dimensions.

14

The notion of negative response set appears later in
the literature. Christie, Havel, and Seidenberg (1958) seem
to have given the first report of negative response set. A
nation-wide sample of college—trained adults displayed more
negative than positive response set (5% showed a strong
positive set, 28% a slight positive set, 23% were completely
consistent (i.e., not set), 33% a slight negative set, and
11% a strong negative set). Yet college students showed
almost Opposite results with strong positive set (23%) and
strong negative set (3%). Washington lobbyists and graduate
students showed little response set. The authors inter-
preted this to mean that they were ". . . least confused
ideologically."

Broen and Wirt (1958) factored results from 15
response set scores. They found a factor which seemed
associated with positive set and one associated with negative
set. Messick (1962), using Edward's social desirability
scale and an acquiescence scale, found three factors:
impulse-control or negative set, social desirability, and
acquiescence.

Couch and Keniston (1960) attempted to identify
personality correlates of the direction response sets. They
found, using the MMPI with Harvard students, that positive
set was related to ". . . impulsivity, dependency, anxiety,

mania, anal preoccupation, and anal resentment" (p. 173).

 

15

On the other hand, negative setters were character-
ized by ego strength, stability, responsibility, tolerance,
and impulse control.

Asch (1958) began a study of negative set by citing
Karen Horney (1937) who had described the neurotic of our
time as "basically hostile and negative in orientation. . . .
Three hypotheses were made: (a) the non-setter is better
adjusted than persons demonstrating negative response bias;
(b) neurotic tendency and negative set are positively re-
lated; and (c) obsessive-compulsive persons demonstrate more
negative response bias than those with hysterical tendencies.
Using the MMPI, Goodenough 'Draw a Person," and a blind
Rorschach with a sample of normal veterans, Asch found all
relationships to be significant at the .001 level.

The syndromes presented by Asch and by Couch and
Keniston are quite different. Asch found negative setting
veterans to be neurotic, while Couch and Keniston indicate
negative setting Harvard students to have greater ego
strength, impulse control and other desirable qualities. Two
hypotheses could account for this difference in the Harvard
sample like Christie, Havel, and Seidenberg's (1958) college
sample may contain only a small number of extreme negative
setters while Asch's veterans like Christie, Havel, and
Seidenberg's college trained adults may have had a higher
proportion of extreme negative setters; or negative setting

Harvard students and veterans differ in personality due to

 

16

some other factor. The former hypothesis seems more plausible
at this point, although no conclusion is possible until more
research is done.

Jackson (1958) finds that a positive set represents
a generalized tendency to yield to the "field forces of a
situation" (i.e., reversal of a Nector cube). Disagreement,
on the other hand, may be considered as more or less actively
resisting these "field forces." Also James and Stone (1960)
find that negative set is associated with retroactive facili—
tation and positive set with retroactive inhibition. In a
Trinidad study, Mischel (1961) found an inverse relationship
between the capacity to delay gratification and acquiescence.
A study by Zucker (1962) revealed significant correlations
between the Bass Social Acquiescence Scale and measures of
succurance-autonomy. On the other hand, Foster (1961) found
no behavioral correlates of Acquiescence.

Two recent studies of response set and personality
by Adams (1962) and Weitman (1962) have delineated more
carefully between the positive setters, negative setters and
the content responders or non-setters. Adams (1962), using
a reversed F-scale, divided his sample of 48 college stu-
dents into negative setters, positive setters, authoritarian
non—setters (i.e., those selecting authoritarian responses
regardless of the items direction) and equalitarian non-
setters. He then compared the groups on a measure of con-

ceptual rigidity, viz., Berg's 'Perceptual Reaction Test'.

 

17

The negative setters were found to be the most rigid; the
positive setters, the authoritarians; and the equalitarians
followed in order. Thus non-setters were lower in rigidity
than both types of response setters.

In a larger study, Weitman (1962) attempted to
reconcile response set and authoritarism. To do this, he
hypothesized four orientations to authority:

a. Allo-authoritarian individuals (e.g., hermits,
recluses, and perhaps some catatonics) whose
predominant orientation toward authority is

avoidance.

b. Pro-authoritarian individuals ('classical')
who are overly concerned with authority.

c. Anti-authoritarian individuals (e.g., chronic
opportunists, perennial rebels, 'trouble makers',
and many criminals) who are over concerned with
authority and resistant to said authority.

d. Non—authoritarians—-those individuals hOpe—
fully constituting a majority of the population
who have no special authority problems (p. 194).
From these definitions, pro—authoritarians were

hypothesized to tend to agree, anti-authoritarians to dis—
agree, and non—authoritarians not to exhibit any directional
response set. Allo-authoritarians were not expected in the
sample.

In the first study, 581 students were screened with
an acquiescence test. The author selected 151 out of which
121 showed up [high scorers (50), middle scorers (35), low
scorers (36)]. A highly significant relationship (p .001)

was found between the acquiescence classification and both

18

of the authoritarian classifications (the F—scale and a
special sentence completion test).

As the theory characterized the anti-authoritarian
personality as rebellious and generally anti-social, Weitman
examined male and female prisoners with the hypothesis that
there would be a greater percentage of anti—authoritarians
and pro—authoritarians than non-authoritarians. Out of 103
prisoners only two could be classified as non-authoritarians.

Here again like Adams' (1962) data we find differences be-

 

tween setting (regardless of direction) and non—setting
subjects. The extremes in both cases look more like those
in the other extreme groups than they do the middle; non-
setting does not represent a half way mark between positive

and negative response setting.

 

THE DESIGN OF THE STUDY

As previously stated, the main purpose of this
research is to determine the relationship between an indi—
vidual's location on the Open- and closed—mindedness con— )
tinuum and his confidence in judgments. The general hypothe-

sis, derived from Rokeach's second definition of the Open

 

and closed mindedness continuum, is that the closed minded
subject will be more influenced by the irrelevant aspects of
the confidence decision, that of his own ego involvement and
dissonance reduction, than the more open individual. Under
ego involving conditions the closed minded subject will
employ 'defensive confidence' in an attempt to reduce the
ego threat or 'dissonance' resulting from the disparity be-
tween his performance and his expectations.

In relating directional response set to the present
study, Adams' (1962) study of response set, authoritarianism,
and rigidity poses another problem. Here the subjects making
content responses and scoring high on the authoritarian
scale (i.e., Rokeach's persons classified to be dogmatic) are
less rigid than either negative or positive setters. While
this demonstrates further the importance of directional re-
sponse set as a personality variable, it furnishes evidence

for the independence of dogmatism and directional response set.

19

20

Fruchter, Rokeach, and Novak (1958) likewise found the Gough
Rigidity Scale and the Rokeach Dogmatism Scale to measure
virtually independent psychological dimensions. In de-
lineating between dogmatism and rigidity, Rokeach (1960)
states:

the referent of dogmatic thinking seems to be a

total cognitive configuration of ideas and be—

liefs organized into a relatively closed system;

rigidity, on the other hand, points to diffi-

culties in overcoming single sets or beliefs en-

countered in attacking, solving, or learning

specific tasks or problems (1960, p. 183).

The directional response set which arises in the
'attacking' of the problem of placing items on a Likert
Scale, appears to be a form of rigid behavior. Considering
that rigidity is associated both with positive and negative
set, we hypothesize that response set and the Open- and
closed-mindedness continuum as conceptualized are independent
but that as the Rokeach Dogmatism Scale is a Likert scale,
the ambiguity of certain items for certain persons results
in response setting. Confidence in judgment and response
set will be therefore uncorrelated, while the Rokeach
Dogmatism Scale and any measure of directional response set
will be moderately correlated.

In any factor analysis involving measures of dog—
matism, confidence, and response set, we hypothesize the
emergence of two factors: one, on which the 'Dogmatism
Scale' and confidence in judgment load highly; and a second

factor of directional response set on which the 'Dogmatism

Scale' and measures of directional response set load.

21

As stated before, the generality of confidence in
belief presents a problem for the ego-involvement theory.

Why did Johnson (1940) find a high correlation between
confidence scores on two uncorrelated belief scales? An
answer other than ego—involvement is provided in terms of
Peabody's (1962) extremeness set. Peabody demonstrated that
extremeness set or the tendency to use answers at the ex-
treme ends of the scale or in the middle range had a high
reliability over a number of Likert Scales, including the
Semantic Differential, and the Dogmatism Scale. If all sub—
jects evidence a high degree of confidence in belief (i.e.,
most subjects averaging above the mid-point), and assuming
that the more ambiguous the task, the more response set to
be expected, it will be hypothesized that extremeness set as
measured by Peabody and confidence in belief will be linearly
related.

Therefore, the specific hypotheses are: (l) The
closed minded subjects will have higher confidence in their
judgments than Open minded subjects under ego-involvement;
(2) Confidence in judgment will not be related to directional
response set; (3) With directional response set removed from
the Dogmatism Scale Scores, the relationship between dogmatism

and confidence in judgment will improve.

22

Subjects

The sample for this study consisted of 55 female
and 37 male college students taking sophomore—level educa-

tional psychology.

Instruments

The Pettigrew Category Width Scale

This is a paper and pencil test developed by
Pettigrew (1958) to measure response style of judgment. The
subject is given a category (e.g., width of windows, speed
of birds, etc.) and its mean. He then estimates the largest
member 'on record' of the category from four alternatives,
and the smallest member 'on record' from another set of four
alternatives. Pettigrew (1958) did not attempt to include
correct answers among the alternatives.

The Pettigrew Scale reduces the factor of accuracy
nearly to zero. Subjects might have special knowledge which
would allow them to give the best estimate of the correct
answer for one or two items. However, the range and highly
specialized nature of the items minimizes the probability of
the subject having any notion of what a correct answer would

be for a majority of the items.

Confidence in Judgment

 

Confidence in judgment will be measured by having

the subject respond to each item of the Pettigrew Category

23

Width Scale (CW scale) after which he indicates his confi-
dence in that particular item on a six—point confidence
scale (0, 20, 40, 60, 80, 100).

To generate ego-involvement without raising the sub—
ject's suspicions, the usual title used with the Pettigrew
CW Scale (Estimation Scale) was left on and the following

instructions were given:

This is a 'standard test' of the ability to make
'qualitative estimations'. On each item care—
fully study the mean of the category, then from
the first set of alternatives, select what you
consider to be the largest member of the cate—
gory, and from the second set of alternatives,
select what you consider to be the smallest
member of the category.

The inclusion of the terms 'ability', 'standard test' and
'quantitative estimation' should result in ego—involvement.
The confidence instructions were:

Next, I want you to select the number which to you
best represents your confidence in your last judg-
ment. The alternatives range from chance (0) to
certainty (100). In each case make your confidence
judgment before going on to the next alternatives.
Any questions?

The Rokeach Dogmatism Scale

The DOgmatism scale is Rokeach's instrument for
measuring the Open and Closed mindedness continuum. Rokeach

describes the instrument as follows:

The primary purpose of this scale is to measure
individual differences in Openness or closedness

of belief systems. Because of the way we have
defined open and closed, the scale should also
serve to measure general authoritarianism and
general intolerance. Our procedure in constructing

24

the Dogmatism scale was essentially deductive. We
scrutinized the various defining characteristics
of open and closed systems. We then tried to con—
struct statements designed to tap these character-
istics (1960, pp. 71-72).

Items were constructed to tap the major traits of
the closed minded person as described by Rokeach: isolated
belief systems, intolerance, feelings of helplessness and
anxiety, etc.

The DOgmatism Scale form D (66 items) was used for
this study. The items are scored unidirectionally and the
subjects respond on a six-point Likert Scale (+3 = I Agree
Very Much; +2 = I Agree on the Whole; +1 = I Agree a Little;
—1 = I Disagree a Little; -2 = I Disagree on the Whole; —3 =
I Disagree Very Much). The Dogmatism Scale items were ad-

ministered with and embedded in the Bass Social Acquiescence

Scale.

The Bass Social Acquiescence Scale

To construct the Social Acquiescence Scale, Bass
(1956) collected a heterogeneous set of ambiguous aphorms.
The Bass Social Acquiescence Scale was administered

under the same instructions as the Dogmatism Scale.

Confidence in Belief

Six point confidence intervals (0, 20, 40, 60, 80,
100, were administered along with each item on the Bass

Social Acquiescence Scale and the Rokeach DOgmatism Scale.

25

Scores were taken separately for a confidence response to

each belief scale.

Extremeness Response Set

Measures of set toward extremeness were taken on all
five scales: Rokeach Dogmatism Scale, Bass Social Acquies-
cence Scale, Confidence in Pettigrew Judgments, Confidence
in Rokeach Beliefs, and Confidence in Bass Beliefs. On the
belief scales, set toward extremeness was scored by summing
absolute values for the items. The confidence intervals
were scored by weighting responses of 40 and 60 as 1, 80 and

20 as 2, and 100 and 0 as 3, and then summing over the scale.

Summary of Variables

Thus eleven variables were included in the statisti-
cal analysis:

1. Score on the Pettigrew Estimation Scale

2. Score on the Rokeach Dogmatism Scale

3. Score on the Bass Social Acquiescence Scale

4. Confidence in Pettigrew Judgments

5. Confidence in Rokeach Beliefs

6. Confidence in Bass Beliefs

7. Extremeness of Response, Rokeach Scale

8. Extremeness of Response, Bass Scale

9. Extremeness of Confidence in Pettigrew Judgment

10. Extremeness of Confidence in Rokeach Beliefs

ll. Extremeness of Confidence in Bass Beliefs.

26

Analysis

Non—linear correlations (eta's) were calculated to
test the first two specific hypotheses. Product movement
correlations were then calculated for all measures and
factored. The factor analysis was a Quartimax rotation
analysis with a Kiel-Wrigley criterion of two requiring each
factor to have at least two tests with highest loading upon
that factor.

The DOgmatism Scales scores and Social Acquiescence
scores were normalized and the Social Acquiescence was
partialled out from Dogmatism and the adjusted dogmatism
score 8 was correlated with confidence in judgment. The
partialling of directional response set variance from the
Dogmatism scores was accomplished by multiplying the corre-
lation (0.57) between the two scales and the normalized Bass
Social Acquiescence Scale scores. The product was then sub-
tracted from the normalized DOgmatism scale scores yielding

o
D.

RESULTS

Scores for all 92 subjects were obtained on all 11
variables.

Means, Standard DeviationsL and
Reliabilities

All 11 measures compare favorably with the means,
standard deviations, and reliabilities reported in previous
studies (see Table l).

The confidence measures showed high internal con-
sistency; e.g., Confidence in Pettigrew Judgments (0.98),
Confidence in Rokeach Beliefs (0.96), Confidence in Bass
Beliefs (0.95). The internal consistency of the extremeness
measures were also high; e.g., Extremeness of Confidence in
Pettigrew Judgments (0.95), Extremeness Of Response, Rokeach
Scale (0.89), Extremeness of Response, Bass Scale (0.93),
Extremeness of Confidence in Rokeach Beliefs (0.96), and
Extremeness of Confidence in Bass Beliefs (0.95).

A summary of intercorrelations between the 11 vari-
ables is located in Table 2.

The hypothesis that confidence in belief and confi-
dence in judgment measures would not be related did not hold

(r = 0.366, p (.0. and r = 0.329, p < .01).

27

28

Table 1. Means, Standard Deviations and Odd—Even

Reliabilities

 

 

 

Mean S.D. r
xx
1. Pettigrew Estimation Scale 63.05 17.70 .86
2. Rokeach Dogmatism Scale:
Form E 188.95 32.98 .87
3. Bass Social Acquiescence
Scale 217.85 36.96 .90
4. Confidence in Pettigrew
Judgments 47.55 20.48 .98
5. Confidence in Rokeach
Beliefs 70.49 14.56 .96
6. Confidence in Bass
Beliefs 66.14 15.35 .95
7. Extremeness of Confidence
in Pettigrew Judgments 65.83 18.02 .95
8. Extremeness of Response,
Rokeach Scale 122.57 22.23 .89
9. Extremeness of Confidence
in Rokeach Beliefs 125.50 29.29 .96
10. Extremeness of Response,
Bass Scale 100.18 20.42 .93
11. Extremeness of Confidence
in Bass Beliefs 104.98 23.14 .95

 

29

Table 2. Correlations of the Test Battery

 

 

 

l. Pettigrew
Estimation
Scale --

2. Confidence in
Pettigrew
Judgments .08 --

3. Rokeach Dog—
matism Scale -.03 .07 —-

4. Confidence in

Rokeach
Beliefs .11 .37 .04 ~—
5. Bass Social
Acquiescence
Scale -.ll-.15 .57 103 --

6. Confidence in
Bass Beliefs .02 .32 .15 .91 .10 --

7. Extremeness of
Confidence in
Pettigrew
Judgments -.01 131 .07-.00 .03 .03 —-

8. Extremeness of
Response,
Rokeach
Scale .07 .05 .05 .54 .02 .49 .26 —-

9. Extremeness of
Confidence in
Rokeach
Beliefs .16 .24-306 .74-u07 .66 .21 .61 --

10. Extremeness of
Response, Bass
Scale -.02 .06 .27 .60 .19 .67 .21 .76 .67 --

ll. Extremeness of
Confidence in
Bass Beliefs .13 .12 .03 .65 .04 .68 .30 .65 .84 .74

(r 3.21, r< -.21, pj>.05; r)>.28, r<f.-.28, p (.01)

 

30

There was a (0.907) correlation between the two
confidence in belief scales. Johnson (1941) found a similar
relationship using uncorrelated belief scales. On the other
hand, the correlation between confidence in judgment and the
Dogmatism Scale was (0.07).

The Rokeach Dogmatism scale and the Bass Social

AcquiescenCe scale correlated (0.57).

Factor Analysis

The Quartimax factor analysis of the 11 variables
yielded three factors which were labeled 'Extremeness Set',
'Directional Response Set' and 'Defensive Confidence' (see
Table 3).

The_Extremeness set factor, which accounted for 41%
of the variance had six high loadings: Confidence in
Rokeach's Dogmatism Beliefs (0.856), Confidence in Bass's
Social Acquiescence Beliefs (0.842), and Extremeness of Re-
sponse for the two belief scales (0.781 and 0.858), and
Extremeness of Response for both confidence in belief
measures (0.888 and 0.888). The loading of Extremeness of
Response for both belief scales and confidence in belief
measures on the same factor indicates the importance of
Extremeness set for these measures. On the other hand, set
effects these measures differently. When all-most—all sub-
jects have mean scores above or all-most—all below the mid—

point of the scale as in the case of the confidence in belief

31

.uouomm mom OCHUMOH amen oBu mo cofluwufluo hmamfluzlaflmxe

 

wosmfihm>

mmmm CH wocwcflmsou mo

mamom mmmm .mmcommmm wo

commxom CH mocmpflmsoo mo

sommxom .mmcomwmm mo

mpcwempsb
CH oocmpflmsou mo

mo “:00 you

wmmflaom
mmmcmeuuxm .HH

WWOCTEOHHXW .OH

mwmflawm
mmmd®EmHuxm .m

mamom
mmwcmEmupxm .m

3mumflpumm
wmwcmeupxm .n

mwmflamm mmmm OH OOCOOHMGOU .o

mamow woswommwsqum

Hmfluom mmmm .m

mmmflamm nommxom CH mosmpflmcoo .w

mamom Emflumemon nommxom .m

mucwEmUsb Boumfluumm OH OOCOUHMCOU .m

mamom soaumaflpmm Boumfluumm .H

 

 

Xm.ma fim.ma XO.HV
HVH.I moo.| mmw.
mma.| Hmm. mmm.
moo. mma.| mmm.
oom.l nHo.| awn.
mmh.| Hmo. new.
mom. moo. New.
mmo.| mom. Nmo.
mmm. Ono. 0mm.
hmo. mmm. vHH.
won. ooa.l vmm.
HBO. onm.l ONH.

ucmEmpsb CH umm Oncommmm wwcommwm mo
wocmpflmcoo Hmsofluomuflm mmmcmEmuuxm
muouomm

nonemmwz

 

 

smmcflpmoq Monomm meHuumsO .m OHQmB

32

scales, then the scale scores will linearly correlate with
the extremeness scores. Both 40 and 60 are scored as l in
extremeness which means that if all subjects had mean scores
above 40 and used the lower end of the scale (20 and 0) in-
frequently then the linear effect Of extremeness set would
be maximized. In the case of the belief scores the lack of
linear relationship between scale scores and extremeness set
means that the subjects are using both ends of the scale
extensively, e.g., Peabody (1962).

The second factor has as its highest loadings the
Dogmatism Scale (0.859) and the Social Acquiescence Scale
(0.868). As both the Bass' Social Acquiescence Scale and
the Dogmatism Scale have been shown to be related to other
measures of directional response set, the second factor will
be designated as the 'Directional Response Set Factor'.

The third factor has as high loading measures confi-
dence in Pettigrew Judgments (0.794) and Extremeness of
Confidence in Judgment, Pettigrew Scale (0.758). Neither of
these measures loaded highly on the extremeness set factor.

There would theoretically be two factors operating
in Confidence in Judgment: Accuracy and 'Defensive Confi—
dence'. The accuracy factor should be minimized by the
Pettigrew Category-Widths Scales' difficulty level. 'Defen-
sive Confidence', on the other hand, would be expected to
be present due to the Pettigrew Category-Width Scale

instructions.

33

Klein and Schoenfeld (1941) demonstrated that by in-
creasing the subjects ego-involvement in a series of unre-
lated tasks, you produce a generality in confidence across
tasks. In this case, the negative correlation between
Confidence in Judgment and Extremeness of Confidence in
Judgment, is an artifact. Subjects with high confidence in
Judgment use more 40's and 60's which are scored as l on the
Extremeness in Response measures; whereas, subjects who are
more accurate tend to use 0, 20, and 40 which are scored 3,
2, and 1 respectively. As Extremeness of Confidence in
Judgment does not highly correlate with other Extremeness of
Response measures, Extremeness set cannot be used to account
for its homogeneity or its relationship to confidence in
Judgment. 'Defensive Confidence' of some subjects and the
'accuracy' of others (low confidence) would account for

'Confidence in Judgments' homogeneity of extremeness.
Non Linear Analysis

The hypothesis that closed minded individuals would
be over—confident did not hold when the Rokeach Dogmatism
Scale and Confidence in Judgment were linearly correlated
(.07).

A non-linear correlation between Confidence in
Judgment, Pettigrew Scale (y) and the Dogmatism Scale (x)
was significant (/q'yx = .52, F = 4.39, p <.0.010Mx y, = .37,
F = 1.90, not significant) but the distribution was not hypo-

thesized (see Table 4).

34

Table 4. DOgmatism and Confidence in Judgment——Mean Category

 

 

Score

Dogmatism

Scores 152 165 178 191 204 217

239+
-151 —164 -l77 —l90 -203 -216 -229

Mean
Confidence

in Judg- .

ment 66.1 37.7 40.2 46.5 44.8 47.5 51.5 55.6

in

N 9 13 ll 17 16 8 9 9

 

 

1‘.-

‘fdl

The Confidence in Judgment means were high for high
dogmatism, low for middle to low dogmatism, and high for
extremely low dogmatism. The lowest dogmatic group (9
cases) had the highest mean confidence of any group on the
continuum. This group due to its limited size creates a
serious interpretational problem. In comparing the ”4 yx =
0.52 and r = 0.07 for linearity, the distribution is non—
linear (f = 4.38 p <‘.01). The Bass Social Acquiescence
scale was found not to be related to confidence in judgment

(/4 xy = .37 N.S., yx = .33 N.S.).

Partial Correlation

 

When directional response set as measured by the
Bass Social Acquiescence Scale is partialled from the Rokeach
DOgmatism Scale, assuming a 0.574 correlating, the resulting

o
dogmatism scores D linearly correlates (r 0.215,

CD/A =

p<< .025) with confidence in judgment.

35

Reanalysis

 

The emergence of a U—type distribution between dog-
matism and confidence in judgment requires a rethinking of
this study's basic hypothesis. Rokeach (1960) defines
'dogmatism' or 'authoritarism' (regardless of political
affiliation) to be synonymous with closed mindedness. This
study will consider closed mindedness as a more generic term
of which dogmatism is but one particular variety.

DOgmatism will operationally be defined to mean
agreement with dogmatic items without the influence of
directional response sets, e.g., positive set, and negative
set.

The open— and closed-mindedness continuum will re—
tain Rokeach's and Restle's definition. The closed minded
individual is hypothesized to act more in terms of irrelevant
aspects of a situation. These irrelevant aspects may include
any of the external and internal pressures cited by Rokeach
(i.e., "examples of internal pressures . . . are unrelated
habits, beliefs, perceptual cues, irrational ego motives,
power needs, the need for self—aggrandizement, the need to
allay anxiety, and so forth . . . by irrelevant external
pressures . . . of reward and punishment arising from external
authority. . . .") (Rokeach, 1960, p. 57).

Directional response set, positive set and negative
set, will be defined in terms of the basic definition of

closed mindedness. Any response to an item on a personality

36

inventory can be based on two basic aspects of the item
(Jackson and Messick, l958)--the relevant or content and the
irrelevant or non-content (stylistic). To the extent a sub—
ject's responses to questionnaire items are influenced by
directional response set, he is responding closed mindedly.

As stated previously, the conceptualization of dog—
matism advanced by Rokeach implies that dogmatism is synony-
mous with closed mindedness; yet, directional response set
as defined fits the second definition of closed mindedness.

There appears to be a confusion between the generic
concept closed-mindedness as defined in the second definition
and the more particular concept dogmatism similar to Adorno's
(1950) confusion of the particular concept Facism and the
more generic concept authoritarism as described by Rokeach
(1960).

Not only may a person closed mindedly accept a dog—
matic item because of a dogmatic belief system but he may
close mindedly accept the prevailing authority or asserted
position when faced with a certain amount of ambiguity, or
he may, as Weitman (1962) suggests, react negatively to the
prevailing authority when faced with an ambiguous item and
reject it without considering its content. Most likely with
ambiguous items the subjects respond without attempting to
arrive at the intended meaning of the item. Again to
clarify our concepts, we hypothesize dogmatism and direction-

al response sets to be subordinates of closed mindedness.

37

Closed-mindedness and mental impairments as Rokeach (1960)
noted are subsumed under the generic concept of 'rigidity'.

A reanalysis of the data must discriminate between
the two types of directional response set and also the two
types of content responders. A subject may agree or dis—
agree with a majority of the items on either scale on the
basis of either content or response set. Despite the assumed
ambiguity of the Bass Social Acquiescence Scale items, many
of the items may have meaning for the subject which is
constant with the subjects' value system. For instance,
"Love is the greatest of all arts," may very well be re—
sponded to by the subject on the basis Of his definition of
'art' and his views on 'love'. Both the Social Acquiescence
Scale and Dogmatism Scales means are significantly different
from the midpoint of 200 which would be gotten if the sub-
jects had accepted as many items as they rejected; yet, on
the Dogmatism Scale, the group rejected more items than they
accepted while on the Social Acquiescence Scale they accepted
more items than they rejected.

This together with the fact that both scales have
appreciable correlations with other measures of directional
response set such as reversed F Scale items and ambiguous
true and false tests (Peabody, 1961; McGee, 1962 b) leads us
to conclude that the difference between the means and the
mid—point of the two scales is due to the group's reaction

to the content and not to directional response set. If, for

38

instance, the Social Acquiescence Scale elicited more
positive set from the group because it was a better measure
of response set it would contribute more to the directional
response set factor; the same holds true for the DOgmatism
Scale. This does not occur.

To reduce the effect of content on directional re-
sponse set, Couch and Keniston (1960) selected items from a
wide variety of standard scales to reduce the effects of
content in forming the All-over-agreement Scale. Also the
same approach was used by Bass (1955) in forming the Social
Acquiescence Scale. Instead of borrowing items from other
standard scales Bass used aphorism of hetergeneous content.

To reduce the effects of scale content and increase
the effect of directional response set, the Dogmatism Scale
scores and Social Acquiescence scores were combined to
provide a more stable measure of directional response set.
This produces a "Directional Response Set Scale" with a mean
of 406.31 and a standard deviation of 62.02. The odd-even
reliability is (0.91) corrected. This compares favorably
with both the DOgmatism Scale (0.87) and the Social Acquies-
cence Scale (0.90). A scale designed to measure directional
response set should have a moderately high reliability in
that the non—response-setting subjects are making their re-
sponses on the basis of content and their scores should add
error variance which the new all-over-directional response

set measure includes.

39

To divide the sample into types, the top and bottom

 

quartiles of the 'Directional Response Set Scale' were arbi—
trarily designated positive response setters and negative
response setters. Subjects in the second and third quartiles
were designated as non-response setters and divided into

high and low dogmatism g. The division between the high and
low dogmatism groups was arbitrary with 50 per cent appear- '
ing in each group.

We now expect to find groups in the following order

 

in terms of confidence in judgment; (a) the negative setting
subjects, (b) positive setting subjects, (c) nonsetting dog-
matics, and lowest (d) nonsetting low dogmatic subjects.
These expectations correspond with Adams' (1962) results,
and were confirmed in part (see Table 5).

The results of this analysis are clear; nonsetting
non-dogmatics have less confidence in their judgment than
negative setters, positive setters, and nonsetting dogmatics.
Unlike Adams' (1962) study which used the F scale and
rigidity, the negative setters, positive setters, and dog-
matic or authoritarian subjects were not significantly
different in terms of their degree of confidence. It is im-
portant to note that Adams (1962) used the F scale which
allowed for the possibility that left wing authoritarians

would score low.

40

Table 5. Types of Respondents and Confidence in Judgment

 

 

 

 

Mean Confi— Differences
dence in

Type of Respondent N Judgment 2 3 4
1. Negative Setting 23 54.04 1 .91 3.56 21.52**
2. Positive Setting 23 53.13 2 -- 2.65 20.61**
3. Dogmatic Non-

Response Setting 23 50.48 3 —— ---- 17.96**
4. Non-Dogmatic

Non—Setting 23 32.52 4 _— ____ ______

 

1i:

p <1.001

Hrp < .0005.

This data strongly supports the hypothesis that
closed mindedness is a generic concept under which dogmatism,
negative setting and positive setting fall while Open minded-
ness is a concept referring only to non-response-setting non-
dogmatics (equalitarians).

'Defense Confidence' relates to the basic definition
of closed mindedness and appears to be a 'Trait' of at least
three personality typologies. An important problem yet to
be answered is: does the four way classification system
described here represent fundamental differences or are
there other important types and personality dimensions
hidden within this system? The negative setters or positive
setters may subdivide into typologies differing in terms of

some important unidentified dimension.

DISCUSSION

The typologies isolated in this study present both

theoretical and methodological questions.

Theoretical

 

Adorno (1950) became well aware of the existence of
different 'types' of high and low F scale responders; how-
ever, he did not consider response set as a means of differ-
entiating between types. High F scale scores were broken

down into six different types: "Authoritarian," "Conven—

tional,‘ "Surface resentment," Tough guy," "Crank," and the

"Manipulative" types (Adorno, 1951, p. 753). While no
percentages were cited, Adorno pointed out that, in his
sample, the conventional and the authoritarian types seemed
to be by far the most frequent (Adorno, 1950, p. 753).

The "Conventional" syndrome is marked by

. stereotype which come from the outside, but
which has been integrated within the personality
as part and parcel of a general conformity. In
women there is special emphasis on neatness and
femininity, in men upon being a "regular" he-man.
. . . Thinking in terms of in group and out group
prevails. Prejudice apparently does not fulfill a
decisive function within the psychological house-
hold of the individuals, but is only a means of
facile identification with the group to which
they belong. They are prejudice in the specific
sense of the term: taking over current judg—
ments of others without having looked into

41

42

matters themselves. Their prejudice is a "matter
of course," possibly "preconscious," and not even
known to the subjects themselves. It may become
articulate only under certain conditions (1950,
p- 756).

In the psycho—dynamics of the conventional type
". . . the superego was never firmly established and the
individual is under sway of its external representatives"
(1951, p. 751).

It is not difficult to see the conventional type as
a candidate for the positive response setter. Jackson
(1958) found that even with holding the up position on
Nector Cube reversals the positive setter yields to the
field force of a situation. Both Asch (1958) and Couch and
Keniston (1960) found positive set related to 'hysteria‘
which is marked by 'conventionality' and hyper-femininity
and masculinity.

More important, the conventional types lack of a
completely integrated value system (super-ego) and the ten-
dency to take over the current judgments of others seems to
be prime characteristics of positive setters. Further study
is needed to determine if these categories are the same.

The true 'Authoritarian' type ". . . follows the
'classic' psycho—analytic pattern involving a sadomasochis-
tic resolution of the Oedipus complex" (Adorno, 1950, p.
759), and Erich Fromm labeled it the 'Sadomasochistic'

character (1950, p. 759). 'Internalization' of a value

system does take place in this sort of individual resulting

43

in a strong irrational super—ego and the individual achieving
". . . his own social adjustment only by taking pleasure in
obedience and subordination" (Adorno, 1950, p. 759).

The presence of a strong irrational super-ego and
internalized value system indicates a personality that while
authoritarian or dogmatic would not be a positive setter.
Most likely, the true 'Authoritarian' type is also the non-
response setting dogmatic.

Among the low scorers there existed five types:
"Protesting," "Easy Going," "Rigid," "Implusive" and the
"Genuine Liberal." The "Protesting" and "Easy Going" were
the most frequent; however, the low scorers were on the
whole "less typed." The "Protesting" low scorer was psycho-
dynamically similar to the "Authoritarian" high scorer with
the main difference being that

. . . the further sublimation of the father idea,
concomitant with an undercurrent of hostility
against the father, leads to the conscientious
rejection of heteronomous authority instead of its
acceptance. The decisive feature is opposition to
whatever appears to be tyranny (1950, p. 771).
Adorno also noted that most of the neurotic low scorers were
Protesting low scorers and sometimes show compulsive symptoms.
This typology fits the description of Asch's (1958) "nega-
tive setter" and Weitman's (1962) "anti-authoritarian."

The "Easy going" low scorer is the Opposite of the

"Manipulative" high scorer being that they have a "let

things go" attitude and ". . . a profound unwillingness to

do violence to any object" (1950, p. 778). The Genuine

44

Liberal has a strong sense of autonomy and independence and

has a well develOped ego. Adorno sees both the "Easy going"
and ”Genuine Liberal" low scorers as having well established
rational value systems or super-ego's. The "Easy going" and
"Genuine Liberal" types most likely make up the non-setting

low dogmatic group.

As the Adorno types which appear in personality most
like the descriptions of the non—setting types (e.g., Author-
itarian," "Easy going," and "Genuine Liberals") are described
as having well internalized value systems and the possible
response setting types are characterized as not having a well
internalized value system, the degree of the internalization
of the person's value system may be the key difference be-
tween non-setting and directional setting subjects. This
brings in a possible new dimension 'Internalization'.

At present no independent measure of the degree of
internalization of a value system or lack of it is known to
the author. Before further research on determining the
existence of such a dimension and its possible relationship
to directional response set can begin, some independent
measure of the degree of internalization of a subject's
value system must be devised.

As this study provides no way in which Adorno's
types can be related to the types isolated in this study,
further study should be made to determine if there are

relationships between them.

45

Methodology

 

The literature has shown that the directional re-
sponse setters have personality characteristics of their
own. The lumping of content responding and directional re-
sponse setting most likely has resulted in the failure of
many studies to uncover existing relationships. Loevinger
(1959) in the concluding remarks of her article in the
Annual Review of Psychology states, ". . . the proliferation
of tests of high sounding psychological constructs in dis—
regard of response bias is conspicuous waste of research"
(1959, p. 306).

A major problem of future research in the area of
attitude and personality measurement is to develop methods
that either account for response set or to eliminate it.
Peabody (1962) has devised formulas which give the percent-
age of variance accounted for by extremeness set on Likert
Scales and a correction factor could possibly be devised for
it.

Several researchers (Jackson and Messick, 1960;
Messick and Jackson, 1961; Helmstadter, 1958; Webster, 1957)
have constructed methods for separating directional response
set from content; yet, directional response set creates
special problems. If a person has high directional response
set, you have little or no idea of how the contents of your
items fit into his belief system, and if he has answered on

the basis of directional response set to a larger number of

46

items, it would be next to impossible to determine his con-
tent score by partialling out directional response set.

To eliminate the effects of response set new methods
of assaying beliefs and attitudes must be devised. Forced
choice methods seem to be promising. Each alternative would
represent a true difference in content. Heineman (1957)
developed a tripartite multiple choice Manifest Anxiety
Scale using items from the Taylor Manifest Anxiety Scale
plus additional alternatives. Silverman (1957) in comparing
the Heineman Manifest Scale and the Taylor Manifest Anxiety
Scale found that the Taylor Scale correlates with the K
scale of the MMPI, whereas the Heineman Scale does not.
Furthermore, the Heineman Scale correlates with changes in
skin conductivity under the threat of shock, while the
Taylor Scale does not. Jackson and Messick (1960) have
introduced forced choice in the MMPI to reduce the effects
of directional response set and social desirability set.
While response set itself seems indicative of certain
personality types and important in itself, to get at the
actual beliefs held by the person at a given moment measures

other than the Likert Scale must be used.

SUMMARY

This study undertakes to explore possible person—
ality correlates of generalized over-confidence in judgment
in an ego—involving task ('Defensive Confidence'). DOgma-
tism or closed-mindedness was hypothesized to be related to
'Defensive Confidence'. Directional response set, e.g.,
positive setting and negative setting, were hypothesized to
be unrelated to Dogmatism as conceptualized by Rokeach; in-
stead directional response set was seen to be an artifact of
the Likert Scale and not related to 'Defensive Confidence'.

Confidence in Belief and Confidence in Judgment were
not expected to correlate; instead confidence in belief was
hypothesized to be related to extremeness set.

Confidence in judgment scores were taken after each
item of the Pettigrew Category Width Scale which was given
under ego—involving instructions. High confidence in judg-
ment, due to the difficulty level of the Pettigrew Category
Width Scale, was considered evidence of 'Defensive Confi—
dence'. Dogmatism was measured by the Rokeach Dogmatism
Scale and directional response set by the Bass Social Acquies-
cence Scale. Measures of confidence in belief were taken for

both belief scales.

47

48

Extremeness set measures were obtained for both be-
lief scales, the confidence in judgment scale, and both
confidence in belief scales. The eleven measures were
factored.

The factor analysis yielded three factors: (1)
Extremeness set, (2) Directional response set, and (3) De-
fensive Confidence.

A U—shaped distribution was found between confidence
in judgment and the Dogmatism Scale. This was contrary to
the main hypothesis.

To explore the U-distribution, positive and negative
setters were hypothesized to be closed-minded. Both groups
will seem to be acting in terms of the irrelevant aspects of
the testing situation by responding to the structure or
'field force' of the items rather than their content. The
data were reanalyzed.

Negative setters, positive setters, and non—response
setting dogmatics were expected to exhibit 'Defensive Confi-
dence' while only the non—response setting non-dogmatics
were expected not to exhibit defensive confidence. The two
belief scales were combined to form a more accurate measure
of directional response set and the types were isolated by
arbitrarily quartering the sample.

Results strongly support the revised hypothesis.

Positive response-setters, negative response-setters, and

49

non-setting dogmatics were significantly more confident in
their judgments than non-setting non-dogmatics.

As positive setters and negative setters, as well
as, non-response setting dogmatics manifest "Defensive
Confidence" and both directional response setting and "De-
fensive Confidence" were hypothesized to manifest closed
mindedness, closed‘mindedness as defined by Rokeach (1960)
was interpreted to be a generic concept which encompasses
the more particular concepts of dogmatism, positive setting,
and negative setting.

The results were found to compare favorably with
Adorno's (1950) descriptions of types of F scale responders:
the positive setter with the "Conventional" type; the nega-
tive setter with the "Protesting" low scorer; the non-
response setting dogmatic with the 'true' "Authoritarian;"
and the non-response setting non-dogmatic with the "Easy
going" and "Genuine Liberal."

From Adorno's (1950) typologies it was suggested
that directional response setters and non-setting subjects
differed in terms of the presence or absence of a well
developed super—ego or an 'Internalized' value system.
Content responders (e.g., non-response setting non-dogmatics
or Adorno's "Easy going" and "Genuine Liberal" and non—
response setting dogmatics or Adorno's true "Authoritarian")
were marked by internalized value system or strong super—

ego's; whereas the response setters seem to have weaker

50

super—ego's. Further research into the compatibility of
this study's typology and those described by Adorno (1950)

was suggested.

BIBLIOGRAPHY

Adams, Henry E. (1962) "Authoritarism, statistical rigidity,
and a sampling control of response set," J. Psychol.,

 

 

53:47-53.

Adams, Joe K. (1957) "A confidence scale defined in terms
of expected percentages," Amer. J. Psychol., 70:432-
436.

Adams, Pauline A., and Adams, J. K. (1958) "Training in
confidence-judgments," Amer. J. Psychol., 71:747-751.

 

Adams, Pauline A., and Adams, J. K. (1960) "Confidence in
the recognition and reproduction of words difficult
to sell," Amer. J. Psychol., 73:544-552.

Adorno, T. W., Frenkel-Brunswick, Else, Levinson, Daniel J.,
and Sanford, R. Nevitt. (1950) The authoritarian
personality. New York: Harper.

 

Allport, F. H., and Hartmen, D. A. (1925) "The measurement
and motivation of a certain group," Amer. Polit. Sci.
Rev., 19:753-763.

 

Allport, G. W. (1937) Personality: a psychological
interpretation. New York: Henry Holt.

 

Asch, M. J. (1958) "Negative response bias and personality
adjustment," J. Counsel. Psychol., 5:206-210.

Bass, B. M. (1955) "Authoritarianism or acquiescence?" .J.
abnorm. soc. Psychol., 51:616—623.

Bass, B. M. (1956) "Development and evaluation of a scale
for measuring social acquiescence," J. abnorm. soc.
Psychol., 53:266-299.

Block, J., and Peterson, P. (1955) "Some Personality
correlates of confidence, caution and speed in a
decision situation," J. abnormal. soc. Psychol., 51:
34-41.

51

52

Brehm, J., and Cohen, A. C. (1962) Explorations in
ngpitive Dissonance. New York: Wiley.

 

Broen, W. E., Jr., and Wirt, R. D. (1958) "Varieties of
Response Sets," J. consult. Psychol., 22:237-241.

Carey, Gloria, Rogow, A. A., and Farrell, C. (1957) "The
relationship between the F scale and aphorm usage
and agreement," J. Psychol., 43:163-167.

Chapman, L. J., and Campbell. (1957) ”Response set in the
F scale," J. abnorm. soc. Psychol., 54:129-132.

Christie, R., Havel, Joan, and Seidenberg, B. (1958) "Is
the F scale irreversible?" J. abnorm. soc° Psychol.
56:143-195.

 

Christie, R., and Lindauer, Florence. (1963) "Personality
structure," In R. P. Farnsworth and Q. McNemar
(eds.), Annu. Rev. Psychol., 14:201-230.

Couch, A., and Keniston, K. (1960) "Yea sayers and nay
sayers: Agreeing response set as a personality
variable," J. abnorm. soc. Psychol., 60:151-174.

Cronbach, L. J. (1946) "Response sets and test validity,"
Educ. psychol. Measmt., 6:475-494.

Cronbach, L. J. (1950) "Further evidence on response sets
and test design," Educ. psychol. Measmt., 10:3-31.

 

Deutsch, M., Krauss, R. M., and Rossenau, Norah. (1962)
"Dissonance or defensiveness," J. Pers., 30:16-28.

Edwards, A. L. (1957) The social desirability variable in
personality assessment and research. New York:
Dryden.

Edwards, A. L., and Walker, J. N. (1961) "Social
desirability and agreement response set," J. abnorm.
soc. Psychol., 62:180-183.

 

Feather, N. T. (1959) "Subjective probability and decision
under uncertainty," Psychol. Rev., 66:150-164.

Festinger, L. (1943) "Studies in decision: Decision-time
relative frequency of judgment and subjective confi-
dence as related to physical stimulus difference,"
J. exp. Psychol., 32:291-306.

 

.Festinger, L. (1957) A theory of cognitive dissonance.
Stanford, California: Stanford University Press.

53

Foster, J. R. (1961) "Acquiescent response set as a
measure of Acquiescence," J. abnorm. soc. Psychol.,
63:296-299.

Frederiksen, N., and Messick, S. (1959) "Response set as
a measure of personality,“ Educ. psychol. Measmt.,
19:137-157.

Fruchter, B., Rokeach, M., and Novak, E. G. (1958) "A
factorial study of dogmatism, opinionation, and re-
lated scales," Psychol. Rep., 4:19-22.

Frenkle—Brunswik, Else. (1951) "Patterns of social and
cognitive outlook in children and parents," Amer. J.
Orthopsychiat., 21:543-558.

 

Fullerton, G. S., and Cattell, J. M. (1892) "On the
perception of small differences," Univ. Penn. Pub.,
2:159.

Gage, N. L., and Chatterjei, B. B. (1960) "The psychologi-
cal meaning of acquiescence set: Further evidence,“
J. abnorm. soc. Psyghol., 60:280-283.

Gage, N. L., Leavitt, G. S., and Stone, G. C. (1957) "The
psycholigical meaning of acquiescence set for
authoritarianism," J. abnorm. soc. Psychol., 55:98-
103.

Green, B. F. (1954) "Attitude measurement," In G. Lindzey
(ed.), Handbook of social psychology, Vol. 1.
Cambridge, Massachusetts: Addison-Wesley.

Griffing, H. (1895) "On sensation from pressure and impact,"
Psychol. Monog., l.

Heineman, C. E. (1953) "A forced choice form of the Taylor
Anxiety Scale," J. consult. Psychol., 17:447-454.

Helmstadter, G. C. (1957) "Procedures for obtaining
separate set and content components of a test score,"
Psychometrika., 22:381-394.

Henmon, V. A. C. (1911) "The relation of the time of a
judgment to its accuracy," Psychol. Rev., 18:186—201.

Hollingsworth, H. L. (1913) "Experimental Studies in
Judgment," Arch. of Psychol., No. 29.

 

Horney, Karen. (1937) The neurotic personality of our time.
New York: Norton.

54

Jackson, D. N. (1958) "Independence and resistence to
perceptual field forces," J. abnorm. soc. Psychol.,
56:279-281.

 

 

Jackson, D. N. (1959) "Cognitive energy level, response
acquiescence, and authoritarianism," J. soc. Psychol.,
49:65-69.

Jackson, D. N. (1960) "Stylistic response determinants in

the California psychological inventory," Educ.
psychol. Measmt., 20:339-346.

 

Jackson, D. N., and Messick, S. J. (1958) ”Content and
style in personality assessment," Psychol. Bull.,
55:243-252.

 

Jackson, D. N., and Messick, S. J. (1960) "Acquiescence
and desirability as response determinants on the
MMPI," Princeton, N. J., Educational Testing Service,
Report RB-60-10.

Jackson, D. N., Messick, S. J., and Solley, G. M. (1957)
"How 'rigid' is the 'authoritarian'," J. abnorm. soc.
Psychol., 54:137-140.

 

James, R. L., and Stone, L. A. (1960) "Individual differ-
ences in the production of retroactive facilitation
and inhibition," Percept. mot. Skills, 11:223-227.

 

Jersild, A. (1929) "Determinants of Confidence," Amer. J.
Psychol., 41:640-642.

Johnson, D. M. (1939) "Confidence and speed in the two-
category judgment," Arch. of Psychol., No. 241:52.

Johnson, D. M. (1940) "Confidence and the expression Of
Opinion," J. soc. Psychol., 12:213-220.

Johnson, D. M. (1955) The psychology of thought and
judgment. New York: Harper.

Klein, G. S., and W. Schoenfeld. (1941) "The influence of
ego-involvement on confidence," J. abnorm. soc.
Psychol., 36:249-258.

Kogan, N., and Wallach, M. A. (1960) "Certainty of judg-
ment and the evaluation of risk," Psychol. Rep., 6:
207-213.

55

Leavitt, H. J., Hax, H., and Roche, J. H. (1955)
"'Authoritariansim' and agreement with things author-
itarian," J. Psyghol., 40:215-221.

Lichtenstein, E., Quinn, R., and Hover, G. L. (1961)
"Dogmatism and acquiescent response set," J. abnorm.
soc. Psychol., 63:636-638.

 

 

Lorge, I. (1937) “Gen-like: halo or reality," Psychol.
Bull., 34:545-546.

Loevinger, Jane. (1959) "Theory and technique of
assessment," In P. R. Farnsworth and Q. McNemar
(eds.), Annu. Rev. Psychol., 10:287-317.

Lund, F. H. (1926) "Criteria of Confidence," Amer. J.
Psychol., 37:372-381.

Mathews, C. O. (1927) The effects of position of printed
response words upon children's answers to questions
in two-response types of tests," J. Exp. Psychol.,
18:445-457.

McGee, R. K. (1962a) "The relationship between response
style and personality variables: The measurement of
response acquiescence," J. abnorm. soc. Psychol., 54:
229-233.

McGee, R. K. (1962b) "The response style as a personality
variable: By what criterion?" Psychol. Bull., 59:
284-295.

McGuire, William J. (1960) "A syllogistic analysis of
cognitive relationships," In M. J. Rosenberg, C. I.
Hovland, W. J. McGuire, R. P. Abelson, J. W. Brehm,
Attitude organization and change: An analysis of
consistency among attitude comppnents. New Haven:
Yale University Press, 65-112.

Messick, S. (1961) "Personality structure," In P. R.
Farnsworth and Q. McNemar (eds.), Annu. Rev.
Psychol., 10:287-317.

Messick, S. (1962) "Response style and content measure
from personality inventories," Educ. psychol.
Measmt., 22(1):41-56.

Messick, S., and Jackson, D. N. (1961) "Desirability scale
values and dispersion for MMPI items," Psychol. Rep.,
8:409-414.

56

Metcalf, J. T. (1917) "An experimental study of the
conscious attitudes Of certainty and uncertainty,"
Psychol. Monog., 23:181-240.

Mischel, W. (1961) "Preference for delayed reinforcement
and social responsibility," J. abnorm. soc. Psychol.,

 

62:1-7.

Neuhaus, J. G., and Wrigley, C. F. (1954) "The quartimax
method: an analytical approach to orthogonal simple
structure," Brit. J. statist. Psychol., 7:81-91.

Okabee, T. (1910) ”An experimental study of belief,"
Amer. J. Psychol., 21:563-596.

Osgood, C. E., and Tannenbaum, P. H. (1955) "The principle
of congruity in the prediction of attitude change,"
Psychol. Rev., 62:42-85.

Peabody, Dean. (1961) “Attitude content and agreement set
in scales of authoritarianism, dogmatism, anti-
semitism, and economic conservatism," J. abnorm.
soc. Psychol., 63:155-160.

Peabody, Dean. (1962) "Two components in bipolar scales:
directional and extremeness," Psychol. Rev., 69:65—73.

Pettigrew, T. F. (1958) "The measurement and correlates
of category width as a cognitive variable," J. Pers.,
26:532-544.

Rokeach, M. (1956) "Political and religious dogmatism:
An alternative to the authoritarian personality."
Psychol. Monogr., 70, No. 18.

Rokeach, M. (1960) The open and closed mind. New York:
Basic Books.

Schultz, R. E., and Foster, R. J. (1963) A factor anaiytic
study of acquiescent and extreme response set,“ Educ.
Psychol. Measmt., 23:435-446.

Seward, G. H. (1928) "Recognition time as a measure of
confidence," Arch. Psychol., No. 99.

Silverman, R. E. (1957) "The manifest anxiety scale as a
measure of drive," J. abnorm. soc. Psychol., 55:94-
97.

57

Small, D. 0., and Campbell, D. T. (1960) "The effect of
acquiescence response-set upon the relationship of
the F scale and conformity," Sociometry, 23:69-70.

 

Stone, L., and James, R. L. (1961) "Stability of agreeing
response set over a period of time," Psychol. Rep.,
8:350—351.

Sumner, F. B. (1898) "A statistical study of belief,"
Psychol. Rev., 5:616-631.

Thouless, R. H. (1935) "The tendency to certainty in
religious belief," Brit. J. Psychol., 27:16-31.

Trow, W. C.. (1923) '"The psychology of confidence," Arch.
Psychol., No. 67, 33:47.

Volkmann, John. (1934) "The relation of time of judgment
to the certainty of judgment," Psychol. Bull., 31:
672-673.

Walker, H. M., and Lev, J. (1953) Statistical inference.
New York: Holt, Rinehart and Winston.

 

Webster, H. (1958) "Correcting personality scales for re-
sponse sets or suppression effects," Psychol. Bull.,
55:62-64.

Weitman, M. (1962) "More than one kind of authoritarian,"

J. Pers., 33:193-208.

Williamson, G. F. (1915) "Individual differences in belief
measured and expressed by degree Of confidence," J.
Phil. Psychol., etc., 12:127-137.

Zajonc, R. B., and Moriessette, J. (1960) "Cognitive be—
havior under uncertainty and ambiguity," Psychol.
R p., 6:31-36.

Zucker, M., and Eisen, B. (1962) "Relationship Of acquies—
cence response set to authoritarianism and de-
pendency," Psychol. Rep., 10:95—102.

Zuckerman, M., Norton, J., and Sprague, D. S. (1958)
"Acquiescence and extremeness set and their role in
tests of authoritarianism and parental attitudes,"
In Social aspects of psychiatry, Psychiat. res. Rep.,
10:28-45.

'LY

’
I

Ci

‘5 1-
UP.

Hat-
U. .3

—

-‘~--- k..—

 

 

. O

‘l r.“-

at

 

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

12930013174 489

 

 

.1 ~ .3.___.
w—AJ-h
A b..-

--....—-