Title: Perceptual reactance as predictive of self-image discrepancies
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00098386/00001
 Material Information
Title: Perceptual reactance as predictive of self-image discrepancies
Physical Description: xii, 106 leaves. : ; 28 cm.
Language: English
Creator: Ostrowski, Francis Joseph, 1930-
Publication Date: 1972
Copyright Date: 1972
 Subjects
Subject: Self-perception   ( lcsh )
Psychology thesis Ph. D   ( lcsh )
Dissertations, Academic -- Psychology -- UF   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Thesis: Thesis -- University of FLorida.
Bibliography: Bibliography: leaves 100-104.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00098386
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000577491
oclc - 13983126
notis - ADA5186

Downloads

This item has the following downloads:

PDF ( 3 MBs ) ( PDF )


Full Text













PERCEPTUAL REACTANCE AS PREDICTIVE OF
SELF-IMAGE DISCREPANCIES


Francis Joseph Ostrowski


DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY





UNIVERSITY OF FLORIDA
1972




















In Memory of My Dad and in Honor of My Mother,

Without Whose Assistance and Encouragement

This W'ork Would Never Have Been Possible











ACKNOWLEDGEMENTS


I should like to express deep gratitude to the many

persons who so generously contributed their time, money,

advice, and encouragement to this project.

Thanks go to the advisory committee, Chairman

Audrey S. Schumacher, Jacqucline R. Goldman, Madeleine M.

Ramey, Everette E. Hall, and Joseph S. Vandiver, for their

contributions, suggestions, and helpfulness.

Thanks also go to the Florida Correctional Institutions

staff members who were so cooperativ'.e in granting me

permission to come into the prison at Lowell and conduct my

study. Thanks are due likewise to the ..omen prisoners at

Lowell who had to endure a very hot room on a very hot day

in August to participate in the experiment.

I wish to express gratitude to the staff in the College

of Education at the University of Florida for allowing me

to come into their classrooms to recruit subjects. -And I

wish to thank those same subjects who served without any

inducement.

I am indebted to Miss Jean Crago, w.ho was of inestimable

help in scoring, recording, and card punching the information.

Above all, I am indebted to my wife, Sarita, who

suffered long and patiently through the months of preparation,


iii










constant help and encouragement the whole venture would

ave come to nought.













CONTENTS


Page

ACKNOWLEDGEMENTS . . . . . . . . . iii

LIST OF TABLES . . . . . . . . . .. viii

ABSTRACT . . . . . . . . . . . x

CHAPTER I. INTRODUCTION . . . . .. ... 1

Theoretical Positions . . . . . . . 1

Self-Ideal Discrepancy and Maladjustment 1
Self-Ideal Discrepancy and Repression-
Sensitization . . . . . . . 3
Self-Ideal Discrepancy and Maturity . . 4
Self-Ideal Discrepancy and Perceptual
Reactance . . . . . . . . 6
The personality of the reducer . 9
The personality of the augmenter . 9
The implications of reduction-
augmentation for self-image discrep-
ancy . . . . . . . . 10

Interrelationships Between the Four Theoretical
Positions . . . . . . . . . . 11

Self-Image Disparity, Maladjustment, and
Perceptual Reactance . . . . .. 11
Self-Image Disparity, Defensive Style, and
Perceptual Reactance . . . . . 14
Self-Image Disparity, Developmental Theory,
and Perceptual Reactance . . . .. 16

The Problem of Testing Perceptual Reactance . 16

Hypotheses . . . . . . . . . 18

CHAPTER II. METHODS . . . . . . . . 20

Subjects . . . . . . . . . . 20

Measures . . . . . . . . . . 20









Page

Self-Concept Measure . . . . . 20
Perceptual Reactance Measure . . . .. 22
Affect Adjective Check List . . . .. 24
Repression-Sensitization Scale . . .. 25
Social Comoetence . . . . . ... .26
Biographical Data Form . . . . ... 26

Procedures . . . . . . . . . . 27

CHAPTER III. RESULTS . . . . . . . ... .31

Reliability and Validity of the Squares Estimation
Task . . . . . . . . . . . 31

Rationale for Subdividing the Analysis of the Data 34

The Results Relative to the Hypotheses . . .. 36

Hypothesis 1 . . . . . . ... 36
Hypothesis 2 . . . . . . ... 41
Hypothesis 3 . . . . . . ... 42
Hypothesis 4 . . . . . . . . 42
Relative predictive power of %OE, anxiety,
RS, and status with regard to Hypotheses
1-4 . . . . . . . ... 43
Hypotheses 5 and 6 . . . . . .. 50

Additional Analysis Relative to the Prisoners'
Stimulus-Governed Tendency . . . . .. 57

Summary of Results . . . . . . . .. 58

CHAPTER IV. DISCUSSION . . . . . . . . 61

Perceptual Reactance and Set . . . . .. 61

Perceptual Style and Repression-Sensitization . 63

Overestimation and Self-Other Discrepancy . . 64

The Anxiety of the Prisoners . . . . ... 66

The Personality of the Stimulus-Governed . .. 68

Situational Factors and Perceptual Style ..... 70

Lines of Suggested Future Research . . . .. 73

CHAPTER V. SUMMARY . . . . . . . . . 75












APPENDIX A. Self-Concept Measures . . .

APPENDIX B. Squares Estimation Test . .

APPENDIX C. Order of Square Sizes in the Set

APPENDIX D. Affect Adjective Check List .

APPENDIX E. Biographical Data Sheet . .

APPENDIX F. Raw Data . . . . .. .

BIBLIOGRAPHY . . . . . . . .

BIOGRAPHICAL SKETCH . . . . . . .


Page

* . 78

* . 86

* . 89

* . 91

. . 93

. . 95

. . 100

. . 105


vii












LIST OF TABLES


able Page

1. %OE in the A and D Series . . . . ... 32

2. Intercorrelations between A, D, and Total %OE 32

3. Means and standard deviations for the sample
divided on order of presentation of stimuli . 34

4. Means and standard deviations of the sample
divided on status . . . . . . ... 35

5. Means and standard deviations for the four sub-
groups . . . . . . . . . . 37

6. Summary of correlations between predictor and
criterion variables for the various groups 39

7. C6rrelations between %OE and RS . . . .. 41

8. Intercorrelations of predictor variables . 43

9. Regression equations for predicting self-image
measures for subjects given AD order of presen-
tation of stimuli (N = 48) . . . . . 45

LO. Regression equations for predicting self-image
measures for subjects given the DA order of
presentation of stimuli (N = 48) . . .. 46
Ll. Intercorrelations of predictor variables for
subgroups (I = 24 for each subgroup) . .. 47

12. Regression equations for predicting self-image
measures for education students given the AD
order of presentation of stimuli (Group EAD,
N = 24) . .. . . . . . . . . 48

L3. Regression equations for predicting self-image
measures for prisoners given the AD order of
presentation of stimuli (Group PAD, N = 24) 48


viii











14. Regression equations for predicting self-image
measures for education students given the DA
order of presentation of stimuli (Group EDA,
N = 24) . . . . . . . .... .. 49

15. Regression equations for predicting self-image
measures for prisoners given the DA order of
presentation of stimuli (Group PDA, N = 24) 49

16. Means of low (less than 50%OE) and high (more
than 50%OE) scoring Ss on the SET, divided
according to low (less than the median, 5.5) and
high (more than 5.5) ACL in the AD condition on
self-image measures . .. . . . . 51

17. Analyses of variance for the AD order ..... 53

18. Means of low (less than 50%OE) and high (more
than 50%OE) scoring Ss on the SET, divided
according to low (less than the median, 5.5) and
high (more than 5.5) ACL in the DA condition on
self-image measures . . . .. . . 54

19. Analyses of variance for the DA order ..... 56

20. Scores of subgroups by series . . . ... 57


Table


Page








b\hstract of Discertation Pr.esenLcd to thr: Craduate Council
:f the Universit-y of Florida in ParLial lulfill.ment of tlhe
Requirenimnts for the Deyrei.c of Dou'JLc of I'll i .o-'iO. ly

PERCEPTUAL REACTAlJdCr AS PRI;DICTIV'.: OF
SELF'-111AGE DISCRl:PANC1LSE;

3<-

Francis Joseph Ostro*.'s-ki

March, 1972

chairman: Audrey S. Schumacher
4ajor Departiient: Psychology

This study set out to test the hypothesis that discrep-

ancies between perceived and ideal self, and between perceived

and thought-perceived-by -others self,are related not only

to anxiety, defensive style, and social competence, but also

to such a conflict-free ego dimension as perceptual reac-

tance.

Ninety-six white women, half of them graduate and

undergraduate students at the University of Florida, and

ialf of them prisoners in a Florida institution of maximum

security, served as subjects for the study.

Subjects were administered Achenbach and Zigler's

30-item questionnaire under three sets of instructions:

self as seen by self, self as ideally desired, and self as

thought perceived by others. The predictor variables were

a Squares Estimation Task, Zuckerman's Affect Adjective

Check List, Byrne's Revised Repression-Sensitization Scale,

and a dichotomization of the sample by student vs. prisoner.

The Squares Estimation Task consisted of 150 black squares



sC








on a white background, projected on a screen. Half of the

presentations progressed in ascending, half in descending,

order, and order of presentation of the two series was

counterbalanced for the sample. Since the two orders of

presentation were found to have widely divergent reliabil-

ities, and since subject status was found to reflect two

very different sets of personality variables, the data were

analyzed within these variables rather than as independent

main effects.

The results gave some support to the predictive power

of perceptual reactance for self-image discrepancies. Those

given the Descending Series of perceptual stimuli first

produced less reliable results, and hence straightforward

interpretation of that part of the study was made difficult.

Those given the Ascending Series first, on the other hand,

were consistent on the total series and showed some

interesting results.

The students in Ascending Series first who underestimated

were significantly more prone than their overestimating

counterparts to use extreme response categories in respond-

ing, and to have a high self-other discrepancy. Among this

subgroup perceptual reactance constituted a better predictor

of these criteria than did anxiety or defensive style. On

the other hand, perceptual reactance did not predict self-

ideal discrepancy with any significance, although the

underestimator was prone to have a high self-ideal discrepancy.








Among the prisoners given the Ascending Series first,

n the other hand, it was the overestimator who tended to

se extreme response categories. No strong relationship

n the prisoner group, however, was found to exist between

erceptual reactance and the self-image discrepancy measures.

The influence that order of presentation had on per-

ormance was discussed in terms of a need to modify perccp-

ual reactance theory to include the concept of set. The

influence that subject status had on performance was

discussed in terms of environmental and social history.


xii













CHAPTER I

INTRODUCTION


This study investigated the relative power of certain

predictors of self-image discrepancies. The predictors

chosen were perceptual reactance, a variable defined by

Petrie (19E7), referring to a tendency to over- or under-

estimate the size of physical stimuli; anxiety; defensive

style, i.e., repression or sensitization; and social com-

petence, defined here as student vs. prison populations.

The chief discrepancies of interest were perceived self vs.

ideal self (self-ideal discrepancy), and perceived self

vs. self-as-thought-perceived-by-others (self-other discrep-

ancy).

Theoretical Positions

Self-Ideal Discrepancy and Maladjustment

The discrepancy between the way an individual sees

himself and the way he would either like to be or the way

he feels others see him has been the object of increasing

study in recent years. The view of Rogers and his colleagues

(Rogers and Dymond, 1954) was that high self-ideal discrep-

ancy was related to maladjustment. Rogers' prediction of

a positive relationship between self-acceptance and adjust-

ment has been supported by numerous studies (Rogers









and Dymond, 1954; Calvin and Holtzman, 1953; Bills, 1954;

Chase, 1957; Cowen, Heiliver, Axelrod, and Alexander, 1957;

Hanlon, Hofstaetter and O'Conner, 1954; Shlien, Mosak and

Dreikurs, 1962; Turner and Vanderlippe, 1958; and Williams,

1962). Nonetheless, some others found essentially no

relationship between self-acceptance and adjustment (Borislow,

1962; Kamano, 1961; Zuckerman and Monashkin, 1957; Achenbach

and Zigler, 1963; Katz and Zigler, 1967; Feder, 1968), and

still others observed a curvilinear relationship between

the two (Chodorkoff, 1954; Hillson and Worchel, 1957).

Rogers and Dymond (1954) mention the characteristics

of the "maladjusted" individual, i.e., the student who

sought therapy at their counseling center and served as

subject for their study:

He experiences frequent anxiety, a sense of
failure, and guilt. He feels that he is socially
inadequate, that he lacks positive goals for his
life, and that he has difficulty making decisions.
In addition, he is concerned about one or more
specific problems such as an unhappy marital life,
impending failure on the job or in school, homo-
sexual urges (usually not being acted upon), or
aggressive impulses (p. 39).

They go on to say that this description fits the middle range

of their Ss, and that another "sizable number" were "more

seriously disturbed even to the point of being borderline

psychotic." Another small group, they said, "evidenced

more healthy personalities," who "came to counseling be-

cause of uncertainties and emotional tension associated

largely with some immediate environmental problem."








It would seem that the outstanding characteristic of

the majority of the "maladjusted" group was anxiety.

Therefore, in the present study anxiety was used as the

measure of maladjustment.

Rogers and those following him have investigated the

relationship between maladjustment and one form of self-

image discrepancy especially, namely, self-ideal discrepancy.

However, self-ideal discrepancy is only one form of self-

image discrepancy, since an individual sees himself at

variance not only with his ideal, but also with what he

thinks others see him to be. The latter form of discrepancy,

self-other, could have a great relationship to the ease a

person feels in his dealings interpersonally. Consequently,

it seems that a study of prediction of self-image discrep-

ancies and their relation to maladjustment (or anxiety)

should include self-other discrepancy as well.

Self-Ideal Discrepancy and Repression-Sensitization

Rogers' position was challenged (Altrocchi et al., 1960;

Byrne, 1961; Hillson and Worchel, 1957; Feder, 1969). It

was suspected that self-ideal discrepancy was related not

to every form of maladjustment, but to that form of malad-

justment whose defensive style is repression. Byrne (1964)

characterizes responses indicating repression as

. those responses which involve avoidance of the
anxiety arousing stimulus and its consequences.
Included here are repression, denial, and many types
of rationalization. (Sensitization includes) .
behaviors which involve an attempt to reduce anxiety









by approaching or controlling the stimulus and its
consequents. (These) mechanisms include intellec-
tualization, obsessive-compulsive behaviors, and
ruminative worrying (p. 169).

Those who appeal to defensive style as an explanation

for self-image disparity qualify the Rogerian position.

Hillson and Worchel (1957) predicted that a) maladjusted

individuals characterized by anxiety would present a depre-

ciative self-picture, report high ideals, and show a high

discrepancy between self and ideal concepts; and b) malad-

justed subjects with defensive patterns (denial) would show

little discrepancy between self and ideal, and would present

a picture similar to that of normals. Their results confirmed

their predictions.

Altrccchi et al. (1960), Byrne (1961), and Feder (1968)

all found that sensitizers had larger self-ideal discrepancy

scores than did repressors.

Thus, these investigators relate self-image discrep-

ancy to one form of maladjustment, namely, a defensive

style characterized by anxiety and sensitization. The

opposite pole of this defensive (and maladjusted) style

yields low self-image discrepancies, presumably by means

of denial.

Self-Ideal Discrepancy and Maturity

A third approach to self-ideal disparity is the develop-

mental view proposed by Achenbach and Zigler (1963). This

position holds that an individual's self-ideal discrepancy

is not directly a function of either maladjustment or








defensive style, but rather of his level of maturity.

Derived from their theory is a Social Competence Scale

(Phillips, 1968), used to assess social competence in terms

of intellectual, educational, occupational, employment

history, and marital status variables. The view of Achenbach

and Zigler (1963) is that

If it is true that psychological development
as reflected in social competence constitutes a
general dimension of behavior which cuts across the
usual distinctions of "normality" and "pathology,"
then a correlate of personality as important as
self-image disparity should be at least partially
a function of such development (p. 198).

They base their contention on two premises:

a) The factor of guilt, they hold, should be greater in

individuals with a high level of maturity. Such a person

has a greater capacity for irtrojecting societal demands,

mores and values than an individual at a lower developmental

level. Zigler and Phillips (1960) found that persons who

could be classified as having high social competence showed

just such a high degree of introjection of societal demands.

Thus, it could be argued that such individuals would be more

self-critical.

b) Greater cognitive differentiation, they hold,

following Piaget (1951) and Werner (1948),accompanies higher

levels of development. Thus, more highly developed individ-

uals were expected to utilize more categories and make finer

distinctions than less mature individuals. In general,

Achenbach and Zigler's findings supported their hypotheses.









Katz and Zigler (1967) showed that among children

self-image discrepancy scores increased along with chrono-

logical age and/or IQ. They suggest that "self-image

disparity might be better conceptualized as an index of

development rather than as a measure of maladjustment."

Plotnik (1969) also tested out the relationship between

the Social Competence Scale and various measures of self-

ideal discrepancy. She too found significant positive

correlations between social competence and self-ideal

discrepancy.

Self-Ideal Discrepancy and Perceptual Reactance

Another approach to the question of self-image discrep-

ancy which may help to shed light on the three theories

discussed above is through perceptual reactance styles.

"Perceptual reactance" as used by Petrie (1967) refers to

the differences between individuals in their ways of pro-

cessing experience of the sensory environment. "Reducers"

are those who tend to decrease subjectively what is perceived,

"augmenters" are those who tend to increase what is perceived,

and "moderates" are those who neither reduce nor augment

what is perceived.

The assumption could be made that reducers would tend

to lessen not only perceived physical stimuli, but social

stimuli as well. A lower self-ideal discrepancy would

result from not having received societal norms at full









intensity. A lower self-other discrepancy would result

from not receiving cues from peers at full intensity, and

hence not viewing oneself as different from what one feels

his peers see him to be. Augmenters would, on the other

hand, have high self-image discrepancies.

Perceptual reactance is not identical to perceptual

defense. Perceptual reactance is the reaction of an individ-

ual to a stimulus perceived to contain no apparent ego-

threat. Perceptual defense is conceived to be a reaction

to a stimulus generally thought to be taboo in some way.

Perceptual defense has been the subject of numerous studies

since the late 40s (Bruner and Fostman, 1947; Postman,

Bruner, and McGinnies, 1948; McGinnies, 1949; Lazarus and

McCleary, 1951; Eriksen, 1951 and 1954; Postman and Schneider,

1951; McGinnies and Sherman, 1952; Postman, Bronson, and

Gropper, 1953; Eriksen and Browne, 1956; Cofer and Shepp,

1957; Walters, Banks,and Ryder, 1959; Walters and Pilipec,

1964; Forest, Gordon,and Taylor, 1965; Cable, 1969; Lapidus,

1969; Millimet, 1969; and Porzemsky, 1969). Certain of

these studies supported the hypothesis of perceptual defense,

others did not. Still others (Forest et al., 1965; Cable,

1969) showed that there was a set effect, and that recognition

of task words following presentation not only of taboo words,

but also of nonsense words took longer than recognition

following neutral words.









Perceptual reactance theory concerns itself not with

the ego-threatening qualities of the stimulus, but merely

with its physical size. This study attempted to relate

differential perception of a non-threatening stimulus

(physical size of visually perceived squares) to differ-

ential perception of self-ideal and self-other discrepancies.

What is proposed, therefore, is an investigation into whether

the "conflict-free ego sphere"(Hartmann, 1958) is important

in helping to determine certain self-image variables in the

personality.

One measure of perceptual reactance which has been

employed in personality research is the Petrie Block

Comparison Test (Petrie, Collins, and Solomon, 1960; Petrie,

1967), also referred to as tile Kinesthetic Figural After-

effect by Silverman (1967). The Block Comparison Test is

one in which a blindfolded subject gives subjective estimates

of a block of wood. He does this after having rubbed

(following an initial quiet period of 45 minutes) a block

of wood either larger or smaller than the one he estimates,

for 90, 180, and 300 seconds. Petrie found that charac-

teristically some Ss are reducerss," i.e., tend to under-

estimate the size of the block, and some are "augmenters,"

i.e., tend to overestimate it. Petrie and her co-workers

have related the reduction-augmentation dimension to several

personality variables which may have some bearing on self-

ideal discrepancy.








The personality of the reducer

The reducer is tolerant of pain, but intolerant of

sensory deprivation (Petrie, Collins, and Solomon, 1960).

In other words, he seems to feel input less. He tends to

be extroverted ibidd.), to have a lower Hs score on the

IMPI and to be somewhat accident prone (cf. Petrie, 1967,

pp. 24-25). Reducers are overrepresented among schizophre-

nics, and reduction may be related to their typical flatness

of affect (Petrie, 1967, pp. 60-66). Reducers are likewise

overrepresented among male delinquents (although more delin-

quents than normals are "stimulus-governed," i.e., reducing

and augmenting both, depending on which way the stimulus is

presented (Petrie, 1967, Ch. 5).

The personality of the augmenter

At the other end of the continuum is the augmenter.

Tending to overestimate the size of the comparison block,

he is intolerant of pain but very tolerant of sensory

deprivation (Petrie, Collins, and Solomon, 1960). In other

words, he has too much perceptual input to begin with. He

is more introverted ibidd.), has higher Hs scores on the

MHIPI than the reducer, and is not as accident prone (Solon,

1967, in Petrie, 1967, pp. 24-25). Under alcohol, aspirin,

or audioanalgesia the augmenter changes into a reducer, i.e.,

he tends to perceive incoming stimuli less intensely ibidd.,

pp. 62-67). Augmenters are overrepresented among alcoholics

ibidd., pp. 91-92).









The implications of reduction-auamentation for self-image
discrepancy

The reducer seems to be one whose need of stimulation

leads him to disregard harmful or unpleasant characteristics

of things in the environment. Such a lowered sensitivity

may be, at least for male delinquents, the reason for their

popularity, but it may also be a more general style of life

which attracts people to those who complain less of what

ails them.

What is of interest here is that unpleasantness seems

to ail reducers less not because they decide to let them-

selves be bothered less (which might be argued were we con-

cerned only with their tolerance for pain), but because it

may be theorized that the very perceptual mechanism which

presents incoming stimuli to them is of such a nature that

they have less intense experience. Such individuals,

whether they be delinquents, schizophrenics, alcoholics or

normalss," would perceive their world as less intense and

less differentiated than would augmenters.

Augmenters, on the other hand, have overintense percep-

tion of their environment. The sense-world is always keenly

present to them. Sensations that come to them are strong,

so that they tend to see differences rather than similarities.

Such a heightened sense of differences in the sense-world

should also cause them to see sharper discrepancies both

between what they see themselves to be and what they perceive









others see them to be. Experience both in sensory and

social perception would seem to be acutely perceived and

even magnified.

Therefore, reducers should tend to have a concept of

themselves which is not, on the one hand, too heavily over-

burdened by their own shortcomings, nor on the other, too

keenly aware that they could be more than they are nor that

others perceived them differently from what they perceive

themselves to be. Augmenters, however, should tend in the

opposite direction. All sensation, all external stimulation,

should come in at full intensity, and they should see them-

selves as not only filled with defects, but as needing to

achieve high goals and as under constant surveillance from,

and judgment by, others.

Interrelationships Between the Four Theoretical
Positions

The four explanations for self-image discrepancy

presented above, maladjustment, defensive style, developmental

level, and perceptual reactance, are obviously not necessarily

mutually exclusive. Of the four, only perceptual reactance

has not bcen investigated in its contribution to self-image

discrepancies. We shall examine the relation perceptual

reactance should have with each of the other three singly.

Self-Imace Disparity, Maladjustment, and
Perceptual Reactance

An expectation of this study was that maladjustment and

perceptual reactance are orthogonally related. Therefore,









the reducer was expected either to be or not be maladjusted

and the augmenter either to be or not be so, likewise.

What this study investigated was which of the two dimensions,

perceptual reactance or maladjustment, is a better predictor

of self-image discrepancy.

However, the question of independence between the two

variables must be addressed. Rogers and his colleagues have

focused on maladjustment as underlying a high self-image

discrepancy. This relationship was important theoretically

for them, since in client-centered therapy improvement means

precisely a lowering of the discrepancy within the subject

of the felt difference between real and ideal self. Malad-

justment was equated with stress and tension.

The perceptual augmenter, i.e., the one for whom the

sensations from the environment are always experienced as

stronger than they actually are, is an individual in whom

one might expect to find more stress than in the perceptual

reducer. It is, therefore, noteworthy that Petrie has con-

sistently failed to find a relationship between reduction-aug-

mentation and the Neuroticism scale of the Maudsley Personality

Inventory (Petrie, Collins, and Solomon, 1960; Petrie, 1967).

This could lead us to suspect either a) that perceptual

augmentation is unrelated to tension and therefore unrelated

to self-image discrepancy, or b) that "tension" does not

mean maladjustment, and a person can have high tension

(consequent, let us hypothesize, on perceptual augmentation)








and possibly also high self-image discrepancy without being

maladjusted. The Rogerians would hold the former and reject

the latter.

Plotnik (1969) notes some reasons why Rogers' findings

must be looked at more critically. First, she believes

that the equating of tension and stress with maladjustment

is unjustified. She notes that a certain level of tension

is easily admissible with proper development, and even a

high degree of it does not necessarily mean that a person is

maladjusted.

It should be remarked, however, that stress or tension

would be impossible were the person to have had the intensity

of his perception of the environment cut off at its very

intake. Conversely, should he have this intensity increased

at intake, one could expect that he will be subjected to

greater stress. Hence reducers should experience less, and

augmenters more, tension.

Secondly, Plotnik notes that the academic community which

served as Rogers' subject pool is probably best viewed as

high competence, at least as judged by variables of intel-

ligence, education, and occupation. Consequently, she

argues that the finding by

. Rogers and his colleagues that their clients'
decrease in self-image discrepancy was not great or
perfect after successful therapy might support the
idea that a) these investigators were working with
a selectively high competence population, and
b) that high social competence individuals tend to
manifest self-ideal discrepancies (p. 17).









While Plotnik's suggestions are in accord '..ith the social

competence theory which she espouses, one might also ask a

different question. If perceptual augmentation is a style

previous to either maladjustment or social competence, and

one that is minimally influenced by non-physiological means,

and if it is perceptual reactance which forms the basis for

self-image discrepancy, then two things should be expected:

a) A great change could be wrought in objective measures

of adjustment with a minimal change in self-image discrep-

ancy, whether this '..:as a student, or any other population.

An individual's perceptual reactance style would determine

his self-image discrepancy, no matter who he was or how

maladjusted he was.

b) Augmenters, having (supposedly) a high self-image

discrepancy, might he expected to want to reduce the tension

cuased by such a discrepancy. Thus, we might expect aug-

menters to be overrepresented among those exhibiting a high

need for achievement. If college students might be viewed

as displaying just such a high need for achievement, then

we could expect augmenters to be overrepresented among

students.

Self-Image Disparity, Defensive Style,
and Perceptual Reactance

While maladjustment and perceptual reactance were

expected in this study to be orthogonal, an opposite hypothe-

sized relationship, one of congruence, was conceived to








underlie the relationship between perceptual reactance and

the defensive styles of repression and sensitization.

Byrne et al. (1963) conceive that those on the repressive

extreme of the repression-sensitization continuum,which

they hypothesize,employ avoidance defenses, such as denial.

Those on the sensitizing extreme they conceive of as using

approach defenses, such as intellectualization. Now, by

definition the reducer is one who does not feel sensory

stimulation as intensely as it is, while the augmenter feels

it more intensely. Therefore, it might be expected that

the reducer, like the repressor, will use avoidance defenses,

while the augmenter, like the sensitizer, will use approach

defenses. Hence, we might expect positive correlations

between augmentation and sensitization scores.

Moreover, since it has already been shown (Byrne et al.,

1963) that a relationship exists between repression-

sensitization and self-ideal discrepancies, it was hypothe-

sized here that a similar relationship would exist between

perceptual reactance and self-ideal discrepancies. Thus,

since sensitizers have been shown to have a high self-ideal

discrepancy, it might be presumed that augmenters will also

have a high self-ideal discrepancy, and of approximately the

same order of magnitude as their sensitization scores. It

is also proposed that similar relationships hold in the area

of self-other discrepancies.









Self-Imaoe Disparity, Developmental
Theory, and Perceptual Reactance

As with repression-sensitization, there was expected

also to be a positive relationship between perceptual

reactance and social competence.

Plotnik (1969) showed social competence to be related

to repression-sensitization and that both help to predict

self-image discrepancy about equally well, depending on the

measure of self-image discrepancy used. This, then, both

supports the developmental theory on the origin of self-

image discrepancy, and at the same time lends support to the

view that defensive style and developmental level may not be

totally independent. In other words, it does raise a ques-

tion as to whether maturity might not also be consistent

with at least one form of maladjustment, i.e., a tendency

to defend against unwanted impulses by approach, or what

those investigating the repression-sensitization dimension

label the tendency to sensitize.

Therefore, since this paper postulates that reducers

should be repressors, and that sensitizers should be aug-

mentors, it also postulates that reducers will be found to

be low social competence, and augmenters high social

competence, individuals.

The Problem of Testing Perceptual Reactance

To date, the Petrie Block Comparison Test has been very

little used in research. The nature of the test is this:








a subject is asked to give subjective estimates of a physical

stimulus with respect to a standard. Some estimate relatively

accurately, some consistently overestimate, and some con-

sistently underestimate.

Since Petrie has shown that there is some generalizability

between modalities of width- and weight-measurement (Petrie,

1967), it seems logical that other modalities as '..'ell should

yield comparable results. In this experiment a visual

modality was selected, with the subject writing down his

estimate.

The stimulus materials .:ere 150 slides, showing squares

of 14 different sizes. The score of interest for percep-

tual reactance is the percentage of square-presentations

which the subject overestimates over the presentation of the

150 slides.

Petrie's Block Comparison Test, from .'.hich the concept

of perceptual reactance is derived, is in two parts, a "small-

block stimulation" and a "large-block stimulation" session.

In both sessions the subject is made to sit with thumbs and

forefingers not touching for a period of 45 minutes before

testing begins. In the large-block stimulation the subject

(blindfolded, as he is for the other session as well) is

asked first to feel a 2 1/2-inch block and then 1 1/2-inch

block with the thumb and index finger of his dominant hand.

Still touching the smaller block, he is then asked to give

subjective estimates with the other hand on a tapered block,

first two estimates for practice, and then four more estimates









to establish a baseline. Then the process is repeated with

the subject rubbing the larger block for first 90 seconds,

after which he makes four estimates of the small block on the

tapered block; then 90 seconds more of rubbing, and again

four estimates; then 120 seconds more of rubbing, and four

more estimates. This is followed by a 15-minute rest period,

and then four more estimates without rubbing. The small-

block stimulation period uses a 1-inch block for stimulation

and a 2-inch block for estimation. Reducers are defined

as those who have had an average reduction over the three

rubbings of either period of more than -1.18 inch; augmen-

ters are those who have had a similar figure of +1.18 inch.

One who has had a reduction in the large-block stimulation

of more than -1.18 inch and an augmentation in the small-

block stimulation of more than +1.18 inch falls in the

"stimulus-governed" category.

An approximation of the ascending and descending orders

of stimuli presentation was incorporated into the method of

using the stimuli material. The 150 squares were presented

in two series, Series A (Ascending), and Series D (Descending).

This will be described more fully below in the Methods

section.

Hypotheses

The questions which this study hopes to shed some light

on are the following, stated as hypotheses:








1. Perceptual reactance has predictive power with
regard to self-ideal and self-other discrepancy.

2. Reducers w..ill have low, and augmenters high,
Repression-Sensitization scores.

3. Perceptual reactance will have greater predictive
power than will anxiety for self-ideal and self-
other discrepancies.

4. Perceptual reactance will have greater predictive
power than will social competence for self-ideal
and self-other discrepancies.

And additionally:

5. Reducers who are highly anxious will have lower
self-ideal and self-other discrepancy than reducers
who are less anxious.

6. Augmenters w.ho are highly anxious will have a
higher self-ideal and self-other discrepancy than
those who are less anxious.












CHAPTER II

METHODS


Subjects

Ninety-six white female subjects participated in the

experiment. Forty-eight of these were students (undergrad-

uate and graduate) enrolled for the summer quarter of 1970

at the University of Florida. The other 48 were prisoners

in the Florida Correctional Institution at Lowell, Florida.

The prisoners had been convicted of a wide variety of crimes

ranging from narcotics violations through forgery, passing

bad checks, breaking and entering, grand larceny, armed

robbery, and murder in the third, second, and first degrees.

The mean age of the students was 24.19 (s.d., 6.10), and

that for the prisoners 25.29 (s.d., 5.23).

Measures

Self-Concept Measure (cf. Appendix A)

The self-concept measure used was a 30-item question-

naire after Achenbach and Zigler (1963) and employed by

Plotnik (1969). This was administered under three sets of

instructions:

a. Myself as I see me (Questionnaire R = Real)

b. Myself as I would like to be (Questionnaire I = Ideal)

c. Myself as I think others see me (Questionnaire =
Other)








Achenbach and Zigler do not give any reliability, norma-

tive, or validational data for their self-image discrepancy

measure. However, employing as a measure of disparity

between real and ideal questionnaires the number of times

the response on the ideal differed from that on the real

(the maximum thus being 30), and similarly for the self-

other discrepancy, they found High Social Competence related

to high real-ideal disparity as well as to real-social-self

disparity, although the latter was less strong. Moreover,

although their N (40) is too small for normative purposes,

they found that their High Competence Ss had a mean score

for what is labelled herein (cf. below under Perceptual

Reactance Measure) "TOTE" (total number of extreme responses

on the three questionnaires combined) of 21.7, while their Low

Competence Ss had a mean of 49.1. Hence, their results pose

the paradox of a group which used many extremes in responding

having a low self-image discrepancy. It could be that their

measure of self-image discrepancy needs refinement.

It was decided in the present study to employ total

magnitude of differences rather than number of changes

between two questionnaires, since this total magnitude of

differences score was also used by Plotnik, and seems to

be more sensitive to the actual felt discrepancy between

two questionnaires. Moreover, it was decided to isolate

the number of extremes on the "Real" (self as perceived by

self) questionnaire to determine whether individuals who









vary on perceptual reactance, repression-sensitization, and

social status vary also in the definiteness of the view they

take of themselves.

Scores of interest, thus, are:

1. The total magnitude of differences between Real

and Ideal questionnaires (Self-Ideal Discrepancy, S-I D).

2. The total magnitude of differences between Real and

Other questionnaires (Self-Other Discrepancy, S-O D).

3. The total number of responses answered in an

extreme direction (a or f in a six-point scale--cf. Appendix

A for description) on the R questionnaire (Extreme Real, ER).

4. The total number of responses answered in an

extreme direction on the three questionnaires together

(Total Extreme Responses, TOTE).

Perceptual Reactance Measure: Squares Estimation
Task (SET)--cf. Appendices B and C

The stimulus materials for the Squares Estimation Task

were borrowed from a test used by Gardner et al. (1959), and

adapted for the present study. They consist of 150 slides

showing black squares of varying sizes on a white background.

There are 14 different-sized squares used, in addition

to the two anchor sizes of 1 inch and 18 inches. The sizes

are 1.2, 1.6, 2.0, 2.4, 2.8, 3.2, 3.8, 4.6, 5.5, 6.6, 7.9,

9.6, 11.4, and 13.7 inches, sizes which Gardner et al. had

determined to be discriminably different. The order of








presentation is shown graphically in Appendix C. In the

present experiment the first five sizes occupied the first

set, and were presented once in ascending order, followed

by two random orders. The second set was composed of the

third through the seventh sizes, presented in the same

fashion, and so on through the ninth to the thirteenth sizes

in the fifth set. Sets 1-5, each presented three times,

constitute Series A (slides 1-75).

Series D (slides 76-150) begins with set 6, i.e., sizes

14 (13.7 inches) through 10 (6.6 inches), presented in

descending order first, then twice in random order. This

is followed by set 7, sizes 12-8 in descending order first,

then twice randomly, and so on through set 10, which stops

at size 2 (1.6 inches). It will be noted that, despite

parallel procedures in Series A and D, the two series are

not mirror images, since Series A contains the smallest

size but not the largest, while Series D contains the largest

but not the smallest. Something of a parallel to Petrie's

procedures can be seen, since it was noted above in

Chapter I that she used both a larger stimulus and a larger

comparison block for the large-block stimulation (which

parallels the Descending Series in the present experiment)

than for the small-block stimulation (which parallels the

Ascending Series in the present experiment).

Deriving from the Squares Estimation Task is one major

score of interest. Percent Overestimation (%OE) is the total









percent of squares overestimated, and is arrived at by the

formula

%E Number of overestimations
%OE =
150 minus number of correct responses

Using this formula, a 40% overestimation score implies a

60% underestimation and so on, since the "correct," i.e.,

exact, responses have been eliminated.

Hence, overestimators should have %OE scores of more

than 50, underestimators less, and those with neither should

be around 50. Overestimation is equated here with "augmen-

tation," underestimation with "reduction."

Two additional scores of interest are Ascending Series

Percent Overestimation (A%OE) and Descending Series Percent

Overestimation (DOE).

E Number of overestimations in Series A (slides 1-75)
AOE =
75 minus number of correct responses in Series A

NE Number of overestimations in Series D (slides 76-150)
75 minus number of correct responses in Series D

Affect Adjective Check List (cf. Appendix D)

Zuckerman's (1960) Affect Adjective Check List was used

as the measure of anxiety. It is a 60-item inventory of

which 21 items are scorable for anxiety, 11 in the positive,

and 10 in the negative direction. High scores are taken as

indicating high anxiety.

This inventory may be given in either a "General" or

"Today" form, i.e., as asking how the adjectives listed apply

to oneself in general or how they apply to one on the day








on which the person i; responding. In the present study the

"Today" form was used, since how an:-:iety related to the indivi-

dual's performance on otiihr tasks ai. the time was the variable

of interest.

Zuckerman (1960) found a mean of 7.38 (s.d. = 3.98) for

the "Today" form (N = 8.-). li also obtained a respectable

K-R 20 reliability of .85 (p -: .001), but only a .31 (p < .05)

retest reliability (1N = 50). The lo'..' retest reliability was

anticipated, since it is a transient type of anxiety which

this instrument purports to measure.

Zuckerman (1960) also showed a fairly high relationship

of the scores to external criteria in two different samples.

One criterion was exam vs. non-exam day anxiety for students.

The other was a correlation with the Taylor Mlanifest

Anxiety Scale on a group of pregnant women. He also (1962)

has shown that the AACL in the "Today" form correlates

highly with the .-1;4PI scales D, Pt, THAS, Welch AI, and

Rosen ARSI.

Repression-Sensitization Scale (RS)

The defensive styles of repression and sensitization

were measured by the Byrne Revised Repression-Sensitization

Scale (Byrne, et al., 1963). This is a 182-item True-False

instrument employing I.1MPI items. It contains 127 scorable

items and 55 buffer items.

A high score indicates a sensitizing type of defense;

a low score, a repressing one. For a population of 134









students.(58 male, 76 female) enrolled in a sophomore course

in adjustment, the authors found a Brown-Spearman split-half

reliability of .94. For the 78 individuals (32 males, 46

females) in the same group who were available for a second

test three months later, a test-retest correlation of .82

was obtained. Norms on 1,304 students (933 males, 571

females) at the University of Texas over a two-year period

yielded a mean of 42.25 (s.d., 20.10) for males, and 42.68

(s.d., 18.66) for females. The authors had hypothesized a

relationship between RS and self-ideal discrepancy, as well

as between RS and negative self-description. They found

that RS correlated with self-ideal discrepancy .63 (p < .01),

and with negative self-description .68 (p < .01).

Social Competence (Status: St)

The variable of social competence in this study was

measured by considering students to be high and prisoners

low in social competence. In the analysis of the data

following, students were given a rating of 1, prisoners 0,

so that positive point-biserial correlations with the

Status variable indicate that students scored higher on the

criterion, while negative correlations indicate that

prisoners scored higher on the criterion.

Biographical Data Form (cf. Appendix E)

Subjects were asked name, age, grade point average,

sex, marital status, and some questions about smoking habits








and sleeping habits, which w,,ere not analyzed in the present

study.

Procedures

Filling out the Biographical Data Form was the first

procedure for all Ss. The SET was given next, followed by

the personality measures in randomized order. In a few

cases Ss took the personality tests before the SET, however.

Student Ss participated in groups no larger than 6.

The prisoners took part in groups of 12, 11, 12, and 13.

In all cases the Ss were between 6 and 9 feet from the

screen for the SET.

The slides were presented for 3 seconds on the screen,

followed by 1 second of darkness. In a pilot study it was

found that this was the optimum interval to help Ss avoid

boredom and yet ensure that they could make a reasonable

judgment. The 8 seconds on, 3 seconds off,of Gardner et al.

(1959),was reported by the participants in the pilot study

to be too slow to allow them to sustain interest.

The slides were illuminated by a Leica projector and

the intervals set by a Hunter Timer. The projector held

cartridges of 50 slides each, so that natural breaks occurred

after slides 50, 100, and 150. However, two orders of

presentation were employed. In the Ascending-Descending

(AD) order the Ss were shown slides 1-75 first (Series A),

with a break of approximately 2 minutes occurring after

slide 75. This was followed by slides 76-150 (Series D).









In the Descending-Ascending (DA) order the Ss were shown

slides 76-150 first (Series D), followed by a 2-minute

break, then slides 1-75 (Series A). Forty-eight subjects

were given the AD order, and 48 subjects were given the

DA order.

The slides were numbered as indicated in Appendix B.

Subjects were instructed where to begin on the response

sheet (e.g., at number 76 for the DA order, at number 1

for the AD order), and the experimenter called out the

number of approximately every fifth slide to ensure that

Ss were not giving their responses under the wrong number.

The instructions in the SET were flashed on the screen

at the beginning of the test and simultaneously read by E.

The instructions were:

We wish to see how well you can judge the size
of squares. We're going to show you a number of
squares on the screen and want you to tell us how
big they are. The squares may range anywhere between
1 inch and 18 inches. This doesn't necessarily mean
you will get a square which is 1 inch or 18 inches,
though you may. The squares will always be somewhere
within this range.

To help you judge the size of the squares, we
will show you what a 1-inch square looks like--the
smaller end of the range--and what an 18-inch square
looks like--the larger end of the range.

Squares of 1 inch and 18 inches were then flashed on

the screen for approximately 6 seconds apiece. These (and

all the squares) were solid black on a white background.

Following display of the two reference-squares, the instruc-

tions continued:








SWe will sho'. them to you again. You will see
150 squares during the course of the hour and you
have 150 numbered spaces on your sheet. Write your
estimation of the size of each square in its own
numbered space. Thus, for square number 1 record
its size in inches next to number one, etc.

Don't go back over your judgments to change
them. In changing them they are more likely to be
inaccurate. Please don't compare your estimate with
anyone or make any comment during the hour. Make
your judgments independently.

Now. to remind you once again of the range in
which the squares will fall, we will show you again
the smaller and larger end of the range.

Again, the 1-inch and 18-inch squares were flashed on

the screen for approximately 6 seconds apiece. Then,

depending on whether the Ss were in the AD or DA condition,

the final instructions were:

Now we a"e readv to begin. You will see each
of the following squares for only a few seconds.
Look at it all the time it is on the screen and make
your estimation of it when it disappears. The next
square you will see will be number 1. (order AD).

Or:

Now we are ready to begin. You will see each of
the following squares for only a few seconds. Look
at it all the time it is on the screen and make your
estimation when it disappears. The next square you
will see will be number 76. Please begin there on
your answer sheet. The numbers will proceed from 76
through 150, then from 1 through 75. (order DA).

The Ss were then asked if there were any qeustions before

proceeding further. The only questions that arose were

concerning whether the squares were all whole numbers of

inches on a side. This was raised only a few times, and

was always answered by saying that it was up to each









individual to judge that for herself. It may have developed

among the members of those groups where this question arose

a set for greater precision. However, since "overestimation"

in the SET refers to any overestimation, however small, it

is unlikely that such a set would truly confound the results,

since accurate responses were in tenths of an inch, while

the most any S (except one) ever became precise was to

estimate in half rather than whole inches.

Order of presentation was counterbalanced, so that

half of the students and half of the prisoners received the

AD order, while the other half of both groups received the

DA order.

Order of presentation of the various personality

measures was randomized. The three successive administra-

tions of the self-concept measure, however, were always

presented in the order R, I, and O.













CHAPTER III

RESULTS


Reliability and Validity of the
Squares Estimation Task

Before proceeding to results relative to the proposed

hypotheses, the meaning of the instrument used here to

measure "perceptual reactance" was investigated. This was

done relative to Petrie's Block Comparison Test, since the

present instrument was devised to parallel hers. Hence,

use of words such as "reduction," "augmentation," and

"perceptual reactance" depends on the comparability of the

two measures.

First of all, perceptual reactance as used by Petrie

et al. (1960, 1967) refers to a general tendency to augment

or reduce incoming stimulation subjectively. In the Petrie

test the measure of augmentation or reduction is whether a

subject augmented subjectively more than 1.18 inch or

reduced more than -1.18 inch as evidenced by their location

of experienced width on the tapered block. This augmentation

or reduction is estimated irrespective of whether it is in

the large- or small-block stimulation, and irrespective of

which of these had come first. In the present study it

becomes apparent on inspection of the data that not only









was there. an effect of increased overestimation in the

A Series (comparable to the small-block stimulation in the

Petrie) over the D Series (comparable to the large-block

stimulation in the Petrie), but also that it made a great

difference which series was presented first (i.e., whether

we have an AD or DA order of presentation). Tables 1 and

2 make this clear. Table 1 shows that in the AD condition


Table 1.


%OE in the A and D Series


Series


Order of presentation


AD DA

(I = 48) (IJ = 48)
Mean S.D. Mean S.D.

A%OE 65.3 30.4 48.9 31.8

D%OE 55.7 31.0 38.5 27.4

%OE 60.7 31.0 42.8 25.6



Table 2. Intercorrelations between A, D, and Total %OE


A%OE x D%OE

A%OE x%OE

D%OE x %OE


Order of presentation

AD DA

(N = 48) (II = 48)

.819 .440

.950 .896

.952 .739








Ss tend to overestimate both when the stimuli are increasing

and when they are decreasing in size, while those in the DA

condition tend to underestimate in both cases.

Those in the AD condition began with small sizes, moved

to larger ones, then back to smaller ones. They tended to

overestimate throughout. Those in the DA condition began

with large sizes, moved to smaller ones, then back to

larger ones. They tended to underestimate throughout.

Consequently, it appears that there was a "set" effect at

work here. This may be labelled a "contra-anchor" set, since

it appears that those who began with the smaller and moved to

the larger sizes tended to overestimation, while the reverse

is true for those who began with the larger sizes and moved

to the smaller ones.

Table 2, on the other hand, reveals that internal

consistency was not at all similar for the two conditions.

Thus, while the two series correlated in the AD condition

at .819, in the DA condition the correlation sank to only

.440. Moreover, in the AD condition either series correlated

better with overall %OE than in the DA condition.

Thus, two things may be noted. A) The effect of "set"

on estimation of squares confounds the interpretation of

overestimation as simple "augmentation" and underestimation

as simple "reduction." If those in the AD order tend to

overestimate on both A and D Series, and those in the DA

order tend to underestimate on both, it would seem amiss

to treat the underestimators in both groups as reducerss,"




.34



since the "set" effect is diametrically opposed in the two

groups. B) Internal consistency was greater in the AD

condition than it was in the DA.

Rationale for Subdividing the Analysis
of the Data

Besides the difference in internal consistency between

AD and DA orders, there was also found a great difference

between the two in overall %OE. Table 3 shows the Ss in


Table 3. Means and standard deviations for the sample divided
on order of presentation of stimuli

Order of presentation
Variable AD DA p*
(N = 48) (; = 48)

Mean S.D. Mean S.D.

ACL 6.81 4.68 6.56 5.05 n.s.

RS 44.92 23.47 44.77 25.38 n.s.

%OE 60.66 30.96 42.83 25.60 .01

S-I D 37.98 16.31 38.63 20.41 n.s

S-O D 23.65 10.74 24.52 11.98 n.s

ER 9.73 7.61 10.10 7.05 n.s

TOTE 34.71 22.36 34.96 21.45 n.s

Age 24.71 5.56 24.76 5.86 n.s

Education** 1.69 0.75 1.81 0.79 n.s

*p values were considered significant at the .05 level
of confidence.
**"Education" is a variable which had four categories:
grade school or less = 0, high school = 1, college under-
graduate work = 2, and graduate school = 3.








the two orders to be dissimilar on %OE, while they were

alike on age, education, anxiety and sensitization. There-

fore it seemed warranted to do separate analyses of the data

by order of presentation.

However, Table 4 shows that there was a great discrep-

ancy between students and prisoners on anxiety and sensiti-

zation, even though their overall %OE's did not differ

significantly.


Table 4. Means and standard deviations of the
sample divided on status

Status
Variable Students Prisoners p*
(N = 48) (N1 = 48)

Mean S.D. Mean S.D.

ACL 5.17 4.12 8.21 5.07 .002

RS 33.31 17.47 56.38 24.87 .002

%OE 48.71 29.07 54.78 30.21 n.s.

S-I D 32.40 13.65 44.21 20.62 .002

S-O D 19.98 6.18 28.19 13.67 .002

ER 6.02 4.92 13.81 7.24 .002

TOTE 23.44 16.63 46.23 20.45 .002

Age 24.19 6.10 25.29 5.23 ---

Education** 2.34 0.48 1.15 0.46 .002

*p values were considered significant at the .05 level
of confidence.
**cf. Table 3 for explanation of "Education" variable.









Therefore, since order of presentation was a prominent

influence on %OE, and subject status was a prominent influ-

ence on anxiety and sensitization, the data were analyzed

within order and status rather than as independent main

effects. However, where the relative effects of status were

concerned, the data were analyzed only within order, with

status treated as a main effect.

The total sample was subdivided into four subgroups,

each having 24 Ss, education students who had been given

the AD order (group EAD), prisoners given the AD order (PAD),

education students given the DA order (EDA), and prisoners

given the DA order (PDA). Table 5 lists means and standard

deviations for each subgroup on the relevant variables.

"Education" was included in the tables since it is quite

relevant and informative, but was not treated as a predictor

variable, and mostly paralleled closely the St variable.

The Results Relative to the Hypotheses

Hypothesis 1

Hypothesis 1 proposed that %OE has predictive power

for self-image discrepancies. Table 6 shows that in the

EAD subgroup %OE showed a strong negative relationship to

self-other discrepancies and to use of extreme response

categories in self-description. In the same subgroup the

predictive power of %OE was evident for self-ideal discrep-

ancy, but did not reach statistical significance.














SCO r 00 I T O CO


II (N (N ( r(

Z

C m 0o r- r- Ln ,- Co
u - (T r- r- s 'o r- Ln

O m m -i r- Ni
S0oo0 n -,T -

Un



II -- r- -, m (N



II C -"
< On (N mn Ln 0 0c r- m






0 Q W * .
0 2


0





IIi
. ( N N m * *%

U- (N N H ( a N














-.-.LA,- ~ (N .-
r-















II
--







. 4J






(U< (N *
m r c > rl m--I ID
> r-4 a% n u o C un I












T1 M 0 (N 00 0 N LA 0
II C4 C14 -4 -4-4












z
S (Nu (M N m L -4 rm i 0 CO




-r) C) o r- r q a 00
0) 00 'n LA 0 LA IT
IIO



ci-4
.0 %






m Ln m CN i N






dU U) 0 1 0 I 0
> < U 0. U) C E-
o~ U "' n o n N ri o ii t




























0






cn














cn
u





r-
Q)

rin
.4




0)







u
U

-i









.4
U-l
-I


l-4








(U


U) 0




0 0 0





c) C) C
U0 0 I






Jn
Q ) I









:3






4E 4
4- 0


l--
C














S0
m IOI .
Q) 0









Ul Q
Q'0 0


Fn C


'0 C:


C) L 4 UD CA D. a H
0 0 0 0 O o 0





















in H r-4 ui LA ,'-4
000 0 0 0 0 0




















o o o o0 0 0 0
CI) U r- U) I) I) -- l



























U .
mn in Un En ul En 1



















o o 0 E-4o
* .















o 0 a 1 1 0



< 0 dP w i C) (i () E4







U V) O I I O
< o; dp y Ln ui E-


CJl
'-
.0









Table 6. Sunulary of corrciatio,',s bct\.'cOn predictor and
criter~ron vari''ables for the various groups


Vari- AD DA ELAD PAD EDA PDA
able (N = 43) (] 4 ,) (N = 24) (! 24) (11 = 241) (N = 24)

First criterion: S-I D

ACL .527** .44** .591** .458* .687** .196

RS .423** .607** .432* .380 .698^' .386

%OE .004 -.054 -.338 .255 -.035 -.093

St -.172 -.448;:**- --- -----

Second criterion: S-O D

ACL .337* .156 .392 .221 .263 -.053

RS .095 .438** .338 -.139 .433' .243

LOE -.097 -.042 -.658** -.079 -.110 -.043

St -.327* -.399** -- --

Third criterion: ER

ACL .254 .151 .272 .023 -.098 .021

RS .308* .327* .300 .069 -.029 .079

%OE .104 -.022 -.456* .375 .181 -.166

St -.556** -.516** ---- -- ---

Fourth criterion: TOTE

ACL .350* .155 .300 .193 -.035 .004

RS .270 .299* .209 .034 .015 .009

%OE .076 .012 -.489* .400* .213 -.150

St -.550** -.500** --- --- ---

*p < .05.
**p < .01.
The divergence in correlational strength between AD and DA
in personality variables was due to sampling error.









For the prisoner subsample of the AD order (PAD), the

predictive power of %OE was rather high for the use of

extreme response categories. For self-ideal discrepancy in

this subgroup there was a tendency for %OE to show some

predictive power, while for self-other discrepancy the pre-

dictive power of %OE was low.

The strong moderating effect of subject status was

evident in the direction of the correlations in the EAD and

PAD subsamples. The negative correlations in the EAD sub-

sample between %OE and the criteria indicate that it was

student underestimators who had the higher self-other

discrepancies, used extreme response categories more, and

tended to have high self-ideal discrepancies. The positive

correlations in the PAD subsample between %OE and the

criteria indicate that it was prisoner overestimators who

used extreme response categories and tended to have a high

self-ideal discrepancy.

In the less reliable (cf. sup.) DA order of presentation

the correlations between %OE and the criteria were generally

low. The effect of order of presentation was evident also

in the tendency of correlations in the EDA and PDA subsamples

of %OE with the use of extreme response categories (ER and

TOTE) to be opposite in sign from those in the AD order.

Thus, underestimating students in the AD order, but over-

estimating students in the DA order, tended to use extreme

response categories. Similarly, overestimating prisoners in










the AD order, but underestimating prisoners in the DA order,

tended to use extreme response categories. The same inverted

correlations for S-I D and S-O D, however, were less appar-

ent.

Hypothesis 2

The second hypothesis postulated that reducers will

have low, and augmenters high, RS scores. An inspection of

Table 7 shows that correlations obtained between %OE and RS

were generally low and, contrary to hypothesis, negative.

Consequently, Hypothesis 2 received disconfirming evidence

of an expected congruence between an augmenting and sensi-

tizing tendency.


Table 7.


Correlations between %OE and RS


Order Subgroup r*


AD -.048

EAD -.298

PAD -.017

DA -.129

EDA -.157

PDA -.190


*N1one of the correlations reached statistical signifi-
cance.









Hypothesis 3

The third hypothesis states that perceptual reactance

will have more predictive power for self-ideal and self-

other discrepancies than anxiety. Table 6 shows that for

the EAD subgroup %OE was a stronger predictor of self-other

discrepancy than was anxiety, as well as of the tendency to

use extreme response categories (ER and TOTE), while anxiety

proved a better predictor of self-ideal discrepancy. Among

the other three subgroups (PAD, EDA, and PDA), %OE predicted

better than anxiety the use of extreme response categories,

but not self-ideal nor self-other discrepancies. Hypothesis

3, therefore, received partial support.

Hypothesis 4

The fourth hypothesis states that perceptual reactance

will have greater predictive power for self-ideal and self-

other discrepancies than social competence. Table 6 shows

(columns AD and DA) that the prisoners in the present study

tended to use extreme response categories in answering

self-image questionnaires and tended to have high self-ideal

and self-other discrepancies. Subject status constituted a

better predictor of all four criteria than %OE. Thus,

Hypothesis 4 received no support from our data.

One thing to be noted is that it was the prisoners and

not the students who tended to have the higher self-image

discrepancies. This result is at least partly inconsistent








with the theory of Achenbach and Zigler which states that

those who are high in social competence should have high

self-ideal discrepancies. Students were presumed in this

study to have higher social competence than prisoners, yet

they tended to have lower self-ideal and self-other discrep-

ancies than prisoners.

Relative predictive power of .OE, anxiety, RS,
and status e..'ith reward to Hypotneses 1-4

Table 8 shows intercorrelations between predictor

variables. This table makes clear both the independence


Table 8.


Intercorrelations of predictor variables


Variable

%OE
ACL
RS
St


%OE

1.000


%OE

1.000


%OE
ACL
RS
St


AD order
ACL RS

.130 -.018
1.000 .418**
1.000





DA order
ACL RS

.126 -.129
1.000 .685**
1.000


St

-.177
-.301*
-.366**
1.000




St

-.026
-.329*
-.580**
1.000


(low correlations) of %OE from the other variables and the

dependence (high correlations) of the other variables on

each other. Consequently, to investigate relative predictive









power for the criteria of self-image discrepancies and

tendencies to use of extreme response categories, it was

necessary to partial out the interdependency of the predictor

variables. This was done by multiple regression analysis.

Tables 9 and 10 show the regression equations which

describe the relationship to the criteria of the predictor

variables in the AD and DA orders. It will be noted in

Table 9 that %OE occupies the second rather than the third

column in prediction of S-O D. This advance is due to the

partialling out of the interdependence of ACL, RS, and St.

Again, in the DA order, in prediction of ER, %OE is in the

second rather than the fourth column, a change due to

partiallina out the interdependence of the other variables.

As was noted previously, St had such a strong moderat-

ing effect that the four subgroups were also analyzed by

regression analysis. Table 11 shows the same pattern between

ACL and RS (high positive correlations) and virtual inde-

pendence of %OE from ACL and RS.

The effect of partialling out the interdependency of

the predictor variables is summarized in the regression

equations of Tables 12-15. The position of %OE in the EAD

subgroup (Table 13) becomes stronger in the prediction of

self-ideal discrepancy once the dependence of RS and ACL is

accounted for. A glance at Tables 12-15 also shows how

consistently %OE is a predictor of high ER's and TOTE's in

every subgroup, although it accounts for only a small propor-

tion of the variance in each case except the EAD (Table 13).





















$^ -
0co
I.44

U) II
a)






0)



0 0



U) -I
Su1
-i)--4

I 0
m
4 C



U) -4
4-)

*0 ,4



o
0



0
r(5
*n






Cto
0o
0




m- 0




e C
a(>
-4

0

rI U)

U) 1


0) 3





-w(

:-


0
O 0











0
1--1 I












0







O












CI,-
0















4.-1 -1

.-i -,-
04





C
0
O 0

CW




0




LO



0.4





-4




r-
o
un
O- Q


m
ou



co


0 0

I u I
U) U)


O

m 00
*
r- o oa' to
- r co -
o i'0 o (N

I I






*) Cl *





+ I
.


+n r I n (N


(N (N
SE CP
0
SE-i


W
0
o CL)

S*

o (N o *;3
o 0o
0 m 0 m

+ +





U O


















04 i C0
*

.o .C o0 r


+ i








- U
K < !

m () co f0
o m a' m
* .
+ +






4J 4.
U) 1 U) I
-K
-


n ro co r


(N


c
0
-4


04
(1)


u


0


No
(N 0

* *


(N
CA
CC

0\
r-


o o


v v


c*l



















9 II
O
0


0 -r.-




mE

cp 0

04-
0
Ec



C) 4-


c m
n,



:>
-4 PW*4

u



n,0
'0

OnO



C
0 0

0


0>

-4
O
tn 4J)
Ml U







o 0
'-4

0) 3
a m





.0


-4


0
.C4J
*J L)
4 -4
0 0)











0
-4 u
OQ





0
o













mu
'0 -r-


O4
cI














0
U








-4
0
-4





1
44 C









U)





r-
0
u-





c*
0
4l


0 0


I I
cM UM
>-i cc


MN U (N

0E
H


0 0 U 0





I +I







U U. U) U)

Oa Ln Ln LO i -
o co rm n LoN 10 n L n













* * *r *
+ I + +







41 u 0 U
U) *

0o Co m 0) I'D LA LA
o c CD m -A C (n4

i
11110


4)
,.'
(N 0

r- N

o
N
I


oo

v v


*I1
OO





















En


a
3




O

14





0


4



0



0
4-





S0
g4
H 0


0
-41
0


a


4-
0









-4
-4
-




(1


oP
r-4




U
0
OP


U U)


m o
ro n o
r-i **' 0


SC






0









o
o
OO <









co1
0 0















So o









co0 0
O c 0 0

I --



I~ 0-


*9c




Ir-
0 r< 0









o
< H C C 0
Q U o o
S ). 0
-I






0
u 0
0 o







I -I





















0 4 .
c
















00




O o


O U 4
n o








0 U E
0.0 I









Table 12. Regression equations for predicting self-image
measures for education students aiven the AD order
of presentation of stimuli (Group EAD, N = 24)


First Second Third
Criterion Constant predictor predictor predictor


S-I D 30.370 +2.137 ACL .126 %OE +2.137 RS
R2 .349** .434** .434**
S-O D 25.106 .132 %OE + .703 ACL .041 RS
R2 .433** .549 .555**
ER 7.069 .066 %OE + .287 ACL + .015 RS
R2 .208* .264* .266
TOTE 34.100 .256 %OE +1.565 ACL .161 RS
R2 .239* .308* .321*

*p < .05.
.**p < .01.


Table 13. Regression equations for predicting self-image
measures for prisoners given the AD order of presen-
tation of stimuli (Group PAD, N = 24)


First Second Third
Criterion Constant predictor predictor predictor


S-I D 12.476 +1.203 ACL +.200 RS +.117 %OE
R2 .210* .286* .318*
S-0 D 30.857 .762 ACL -.103 RS -.068 %OE
R2 .049 .088 .109
ER 6.660 + .103 %OE +.026 RS -.119 ACL
R2 .141 .146 .152
TOTE 24.497 + .276 %OE +.428 ACL +.011 RS
R2 .160 .172 .172


*p < .05.


.:E









Table 14. Regression equations for predicting self-image
measures for education students given the DA order
of presentation of stimuli
(Group EDA, N = 24)


First Second Third
Criterion Constant predictor predictor predictor


S-I D 16.403 +.277 RS -1.164 ACL -.021 EOE
R2 .487** .569** .571**
S-O D 16.122 +.141 R3 .059 ACL .-.008 %OE
R2 .187* .189 .190
ER 4.173 +.053 %OE .294 ACL +.051 RS
R2 .045 .050 .072
TOTE 14.011 +.205 %OE .859 ACL +.195 RS
R2 .033 .049 .068


.05.
.01.


Table 15. Regression equations for predicting self-image
measures for prisoners given the DA order of
presentation of stimuli
(Group PDA, N = 24)


First Second Third
Criterion Constant predictor predictor predictor


S-I D 25.158 +.422 RS .274 ACL -.006 %OE
R2 .149 .151 .151
S-O D 17.710 +.303 RS -1.004 ACL +.042 %OE
R2 .059 .124 .130
ER 14.562 -.041 %OE + .014 RS +.013 ACL
R2 .027 .029 .030
TOTE 51.965 -.120 %OE .049 RS +.211 ACL
R2 .022 .023 .024


*p <
**p <








Hypotheses 5 and 6

Hypothesis 5 suggested that underestimators who are

highly anxious will have lo..'er self-image discrepancy scores

than underestimators who are less anxious. Hypothesis 6

suggested that overestimators who are highly anxious will

have higher self-image discrepancy scores than those who are

less anxious. Table 16 shows that results were obtained

in the AD condition of the present sample almost directly

contrary to expectation. Highly anxious underestimators had

higher scores than low-anxious underestimators, while highly

anxious overestimators had, in general, lower discrepancy

scores than low-anxious overestimators.

Nonetheless, although the direction of difference

obtained was opposite to that anticipated, it was opposite

on both hypotheses. Therefore, analysis of variance tests

were run to determine the strength of the interaction between

%OE and ACL. Because of the unequal cell frequencies, a

special Multivariate Analysis of Variance computer program

was used. Table 17 presents the results for the AD order.

It will be noted that the interaction of %OE and ACL is

highly significant on S-I D, S-0 D,and TOTE.

The DA order (the less reliable of the two orders),

on the other hand, did not produce such clear-cut results.

Table 18 shows that direction of difference was much more

variable. Table 19 shows the interaction of %OE and ACL

to be nonsignificant on all criteria.











W
El


4J


0
M 10
U







0
O)











-4
E
0 .
do


-4 *







cE
C)1










-A:
C













0-
un tp















0 n
LOA:
Or










C E







S-4
IHn
















0 -H
CO



LnU












4 -
W O
-4t
V)






.0
0 4-




4 -



0-







(13
E-t
Xro )


CO
r40
0.-4
- 4-1
fo (1 0
-u
.0 -'-


0
A:P












0


r0
a, -4

* U
.Q-I .
-a,
. --4
0Ot













(~ I i0 o o D m .M

II r ..
z m (N- ,-.-4


in o m o 0 o0 m
(N mn o 0 0 a
2- -I ,-4 1N .-4 I-4


n -4 mn -T

(N -ZT (N


V
/





N a'
m (





2 I) I

mm
n


H
L0


o % 0 n %D 10

(N m [0 -4 m
(N MN (N -4



a a a .


a 0

O0 E
1 0
E M E-


II









Table 17. Analyses of variance for the AD order



Source Sum of squares d.f. Mean square F


S-I D


%OE (A)
Anxiety (B)
Status (C)
Ax B
A x C
B x C
within cell


35.9
676.2
206.3
1491. 3
337.0
97.4
9033.0


35.9
676.2
206. 3
1491.3
337.0
97.4
225.8


0.159
2.994*
0.913
6.604***
1.492
0.431


S-O D


%OE (A)
Anxiety (B)
Status (C)
Ax B
Ax C
B x C
within cell


682.3
252.1
664.7
563.3
55.7
92.8
2730.8


682.3
252.1
66-1.7
563.3
55.7
92.8
68.3


9.994 **
3.693***
9.736***
8.251***
0.816
1.359


%OE (A)
Anxiety (B)
Status (C)
Ax B
Ax C
B x C
within cell


16.2
37.3
1239.8
36.1
2.2
27. 8
933.8


16.2
37. 3
1239.8
36.1
2.2
27.8
23.3


0.694
1.599
53.106***
1.545
0.095
1.192


TOTE


%OE (A)
Anxiety (B)
Status (C)
Ax B
Ax C
B x C
within cell


*p
**p
* p


19. 5
700.1
5945.8
2331.4
1419.2
52.4
11164.7


49.5
700.1
5945.8
2331.4
1419.2
52.4
279.1


0.177
2.508
21.302***
8.353***
5.085**
0.188


.10.
.05.
.01












.CU




OL-

0






m 0
:



ow





c -1 0
O n





,n 0

-4 n (
t
O n rdl




E E
Cl
E E

C


0 C
C ) 0

-P C


A C Cr

0a 0 Co
In r
U) 0

rdl (




-4 -1-I
r-4
0 OC





CU


(V
Oa

CO
u)O






ra



r-4
-Q
'U
E-


c
aV
CO
CT 0
OC 4J














H C













C


*.l U
r-"
o'o

HU
a


0 r


U



0
4-4
4-4
.-14
'a


0
U
c
C


'4-4
4-4
o



-1
U


3
O

gl











Ln


U') C14 () C4 v
-m (N 0 N ; L C n
m 4 ( N i 'H


In
% D 0 r-q kD 11 0 rM

4o D Ln 4 r











II . . . .
L ( (N N m N (N
:7 N (N N N H


0

0 co m CN m m aC






S c c c; *
Z 4 (N ( (N





f aa fo m m



H O
1-4 00
EI I









Table 19. Analyses of variance for the DA order



Source Sum of squares d.f. Mean square F


S-I D


%OE (A)
Anxiety (B)
Status (C)
A x B
A x C
B x C
within cell


395.2
2314.9
2101.8
320.6
19.6
4.7
12694.0


395.2
2314.9
2101.
320.6
19.6
4.7
314.4


1.245
7.294***
6.623***
1.010
0.062
0.015


S-O D


%OE (A)
Anxiety (B)
Status (C)
A x B
Ax C
B x C
within cell


29.3
39.4
870.2
28.5
16.0
11.5
5538.7


29. 3
39.4
970.2
28.5
16.0
11.5
138.5


0.212
0.284
6.285**
0.206
0.116
0.083


%CE (A)
Anxiety (B)
Status (C)
Ax B
Ax C
B x C
within cell


0.1
6.0
1600.5
3. 4
23.9
27.8
885.6


0.1
6.0
1600.5
3.4
23.9
27.8
22.1


0.007
0.270
72. 289***
0.155
1.078
1.254


TOTE


%OE (A)
Anxiety (B)
Status (C)
A x B
Ax C
B x C
within cell


**p
***p


41.3
27.0
5171.9
337. 8
192.0
294.6
15130.7


42.3
27.0
5171.9
337.8
192.0
294.6
378.3


C .112
0.071
13.673***
0.893
0.508
0.779


< .05.
< .01.









Additional Analysis Relative to the Prisoners'
Stimulus-Governed Tendency

Table 20 points out two interesting results relative

to Petrie's findings that delinquents tend to be more stimu-

lus governed than normals. The first result is that both


Table 20. Scores of subgroups by series


Series
Sub- A%OE D%OE
Order groups Mean S.D. Mean S.D. p

AD EAD 59.0 31.3 51.1 36.1 n.s.

PAD 72.1 29.1 60.3 30.4 <.10

p <.10 n.s.

DA EDA 43.5 29.0 44.6 27.6 n.s.

PDA 54.3 34.2 32.4 26.4 <.01

p n.s. <.10


prisoner groups tended to make their estimates on the initial

series which they were given more strongly in accord with the

direction of stimulus presentation than did students. Thus,

while both EADs and PADs overestimated on the A Series

(the one presented to them first), the PADs did so to a

significantly greater degree. Likewise, while both EDAs

and PDAs underestimated on the D Series, the PDAs under-

estimated to a greater degree.

The second result of interest from Table 20 is that

both prisoner groups were more strongly influenced by the









change of direction of stimulus presentation than were

students. Thus, while neither of the student groups showed

significant differences on %OE between the A and D Series,

both prisoner groups did. Therefore, this study tends to

confirm Petrie's findings of a stimulus-governed character

among delinquents.


Summary of Results

Two orders of presentation of the perceptual reactance

stimulus materials were given, an ascending-descending (AD)

and descending-ascending (DA) one. It was found both that

order of presentation greatly influenced percent cverestimation

(%OE) and that the AD order was more reliable. It was also

found that.students and prisoners differed greatly on the

personality measures of anxiety (ACL) and repression-

sensitization (RS). Therefore, the data were analyzed within

order and subject status.

Percent overestimation of the perceptual reactance

materials for students in the AD order was found to be an

effective predictor of the discrepancy between perceived self

and self-as-thought-perceived-by-others (S-O D), and also of

the tendency to use extreme response categories in descrip-

tions of the self (ER and TOTE), even though the correlations

were in the direction opposite to that which had been

hypothesized. Prisoners in the AD order also showed a

relationship between %OE and TOTE. However, subject status

acted as such a strong moderator variable on %OE that the









directions of the correlations obtained on the TOTE variable

were opposite for prisoners and students.

The results for the DA order were less firm.

The perceptual reactance measure in the AD order pre-

dicted S-O D and the tendency to use extreme response cate-

gories better than the Byrne Repression-Sensitization Scale,

but predicted less well than the RS on the self-ideal

discrepancy measure. The expected significantly positive

relationship between %OE and RS was not obtained; for that

matter, it was negative rather than positive as had been

predicted.

With respect to the differential predictive power of

perceptual reactance and anxiety, the former in the AD order

did show some promise in the field of though vs. Lhought-

perceived-by-others discrepancy, as well as in prediction of

use of extreme response categories, with the caution, however,

that social status acts as a strong moderator variable. With

respect to prediction of self-ideal discrepancy, however,

anxiety was a better predictor.

Social competence, as defined as student or prisoner

within the limits of this experiment, predicted self-image

discrepancy far better than perceptual style, although this

finding should be taken with caution.

Anxiety and %OE among AD Ss were found to relate inter-

actively to the criteria used in this study. Among underesti-

mators, low anxious Ss tended to have lo'..'er S-I. D, S-O D and

TOTE scores than high anxious Ss, while among overestimators





60



low anxious Ss tended to have higher scores on these variables

than high anxious Ss. The results obtained for the less

reliable DA order were not as clear.

Finally, it was noted that the data also support Petrie's

suggestion that prisoners tend to be more stimulus-governed

than students.













CHAPTER IV

DISCUSSION


The results of the present experiment point to several

suggestions. Perceptual reactance needs to be understood

not only in terms of over- or underestimation, but also as

a response to set. Moreover, the defensive styles of

repression or sensitization seem unrelated to perceptual

reactance or style. On the other hand, perceptual reactance

as measured in the present experiment constitutes a good

predictor of self-other discrepancy, at least among student

women, and the meaning of such a variable needs discussion.

It was also found that the prisoners in the present

sample are atypically more anxious than the students, and the

prisoners in general tend to be more stimulus-governed than

students, as Petrie has found.

Finally, comparison of two very different populations

seems to lead to very different and at times diametrically

opposed conclusions regarding the subjective meaning of

perceptual stimulation for students and prisoners.


Perceptual Reactance and Set

One of the essential arguments underlying Petrie's work

is that certain individuals have a tendency to reduce, others

to augment, incoming perceptual stimuli subjectively. The









present experiment produces evidence to show that besides

subjective reduction and augmentation there is also a "set"

effect which should be taken into account in judging strength

of perceptual reactance.

Those given the AD order began with a tendency to move

upward in their estimation of the size of the squares.

Their initial stimuli with the A Series were 1.2, 1.6, and

2.0 inches, etc. When they began the D Series, the established

set changed slowly to a "decreasing set," and they continued

to overestimate while the stimuli decreased in size, until

a new set was established. The DAs, on the other hand,

began with an initial set that moved downwards in the D Series.

When they came to the A Series, they too changed slowly to an

"increasing set." The net effect of the slowness of change

would be an overall higher %OE over 50% by the ADs and an

overall lower %OE on the part of the DAs.

An inspection of Petrie's data indicates that the direc-

tional effect of presentation of stimuli is also present in

her experiments. It will be recalled that she gave her pro-

cedures in two parts, a Large-Block Stimulation and a Small-

Block Stimulation. In the Large-Block Stimulation series

the stimulus block was 2 1/2 inches wide, while the comparison

block was 1 1/2 inches wide. In the Small-Block Stimulation

series the subject had a 1-inch stimulus block and then a

2-inch comparison block.









In her sample of Wellesley women (Petrie, 1967, p. 133)

there were none who augmented in the Large-Block Stimulation

(i.e., when the direction was downward in size), and all but

two augmented in the Small-Block Stimulation (where the

direction was upward in size). Thus, she too found a

directional effect of presentation of stimuli.


Perceptual Style and Repression-Sensitization

The defensive styles of repression and sensitization

seem unrelated in the present sample to perceptual styles.

However, two cautions on accepting the generalizability of

such findings might be sounded.

a) The first caution is that the present sample consists

only of females. Merbaum and Badia (1967) found a distinct

difference between males and females in the tolerance of

repressors and sensitizers to noxious electric shock. While

female sensitizers and repressors did not differ in either

tolerance for electric shock nor in their sensitivity to the

onset of aversiveness, male sensitizers tolerated less shock

and showed greater sensitivity to the onset of aversiveness

than did male repressors. Since Petrie (1967) has shown

tolerance for pain to be an important dimension relative to

perceptual reactance, and since Merbaum and Badia's female

sensitizers and repressors did not differ in their tolerance

for painful electric shock while their males did, there is

aroused in us at least the suspicion that a male sample might








show.: more interesting results in regard to the relationship

between perceptual reactance and repression-sensitization

than would a female sample.

b) The RS scores obtained in the present study tend to

be greatly different from the norms published by Byrne et al.

(1963). The mean they obtained for males was 42.25, for

females 42.68. Becker (1967) found females to have lower

RS scores than males, but still his males had a mean score

of 48.42, his females, 43.30. In the present study the

students had a mean RS of 33.33, the prisoners, 56.38. Pre-

suming that students have greater verbal ability than

prisoners, a lower RS score for them than for prisoners might

be expected on the basis of findings by Clark (1969). He

found that male student repressors have better verbal ability,

social intelligence, sex knowledge and scholastic goals.

However, granted that those with verbal abilities do have

lower RS scores, the means obtained in the present study

for both groups are too different from the norms found by

Byrne et al. to allow us to reach firm conclusions about the

relationship RS has with %OE.


Overestimation and Self-Other Discrepancy

In the present study a high correlation was found between

%OE and self-other discrepancy among student women given the

AD order (Group EAD). The meaning of a discrepancy between

the way I see myself and the way I think others see me has









not been seriously investigated before, and thus the relation-

ship obtained between these two variables is all the more

interesting.

A high felt discrepancy between what I see myself to

be and the way that I think others view me would seem to be

associated with a fairly high degree of subjectively felt

misunderstanding by others. The individual who sees him-

self as greatly different from the way he thinks others see

him is an individual for whom others may well be strangers,

and perhaps even somewhat hostile and unfriendly in their

"otherness." He would not be too prone, therefore, to risk

himself to personal exposure in that world of strangers, and

will much prefer the safe world of non-exposure to failure.

The person, on the other hand, who does not see others'

views of himself as widely discrepant from his own may

possibly be an individual who does not regard others as

strangers, and is not threatened by them. He will, on such

a view of things, more easily risk personal exposure, there-

fore, to the world outside.

The element of risk can be thought of as present, however,

not just in self-other discrepancies, but also in judgments

an individual makes of perceptual stimuli measured against

a yardstick he might be given. In the present experiment

individuals given the AD order were asked to make judgments

of sizes of squares larger than 1 inch against a 1-inch

reference square. Those students who tended to underestimate









the size of the squares generally also showed high self-other

discrepancies, while those who overestimated the size of the

squares had a low self-other discrepancy. Underestimation

among the EADs might be regarded as a form of caution, of

unwillingness to take risks, of "being on the safe side."

Conversely, overestimators among the EADs would be those

who throw caution to the winds and plunge into the task of

estimating with a freer approach and less of a concern for

exactitude.

If the suggested explanation is valid, it would mean that

a generalized trait of caution might be at work which affects

not only an individual's judgments of his own sensory percep-

tions, but also his judgments of his social perceptions.

This could be stated by saying that a person's fear of

making mistakes on perceptual judgments may well spring from

the same source in the personality as does his fear and

suspicion of other people.


The Anxiety of the Prisoners

Contrary to expectations, the prisoners were significantly

more anxious than the students. Such expectations had been

founded on the presumption that a prisoner population would

contain more psychopaths and externalizing individuals than

a student population. Lykken (1957) and Skrzypek (1969)

both found a negative relationship between anxiety and

psychopathy.










Several explanations may be sought for the findings in

the present study of a positive relationship between anxiety

and status. One might be that the measures used in the

various studies (Lykken's, Skrzypek's, and the present one)

are not comparable. Lykken's Activity Preference Schedule

is a measure of social anxiety used to differentiate psycho-

paths, Skrzypek employed an Anxiety Differential constructed

along semantic differential lines, and the Zuckerman Affect

Adjective Check List, in the "Today" form employed in this

study,is a measure of situationally determined and transient

anxiety. Consequently, the anxiety as measured here may be

only marginally related to more permanent ongoing maladjust-

ment. Nonetheless, it is precisely in the situational

aspects of anxiety that one might expect psychopaths to be

differentiated from normals.

A simpler and preferable explanation would be that

expectation of psychopathy in large numbers among prisoners

is unwarranted. A prison population might be expected to be

made up of some mental retardates, borderline psychotics,

and neurotics (as well as some normals who fell afoul of the

law!) along with "true" psychopaths.

Whatever the explanation,it seems that the present

prisoner group exhibits a much different anxiety-response

structure from the students studied. The prisoners may have

been in part responding to the unfamiliarity of the testing

situation, apprehension about mysterious uses to which








"psychological tests" might be put, and so on, all conditioned

by reinforcement within the prison which would make'close

guarding of one's inner thoughts imperative to survival.


The Personality of the Stimulus-Governed

The stimulus-governed tendency of prisoners which has

been suggested by the data of the present experiment is a

tendency which Petrie (1967) found also among juvenile delin-

quents. Eleven out of 48 prisoners (23%) were stimulus-

govern.d, whereas only 6 out of the 48 students (13%) showed

such a tendency, and the overall difference between A%OE and

D%OE was significant for the prisoners in general (cf. Table

20). However, Petrie found the stimulus-governed tendency

among delinquents associated with a high number of delin-

quencies. An inspection of records for prior arrests among

the prisoner group in the present sample revealed an opposite

trend. Three out of the 11 (27%) stimulus-governed prisoners

had had five or more arrests, while 16 out of the remaining

37 (43%) had had five or more arrests. Petrie had found

comparable percentages of 47% and 19%, respectively.

Nor does the early beginning age for crime for the two

groups in our sample agree with hers. The average age of

first arrests for the stimulus-governed Ss was 21.2 as opposed

to 20.2 for the rest of the prisoners. Therefore, the

stimulus-governed in our sample seem to constitute a group

essentially undistinguishable from the others by recidivism

or age-of-first-arrest criteria.









An inspection of their records on type of crime committed,

however, does show that 4 of the 11 (36%) stimulus-governed

Ss were in prison for various types of narcotics violations,

as contrasted with only 7 of the remaining 37 (19%) who were

in jail for narcotics violations. Hence, a brief inspection

of the relationship obtained in the data between narcotics

violations and perceptual anchoring was warranted.

Of the 11 narcotics violators, 4 had been given the AD

and 7 the DA order of presentation. The data show that none

of the ADs exhibited stimulus-governed tendencies, but that

all except 1 of the DAs exhibited either actually more than

50% overestimation in the A and less than 50% (i.e., were

underestimators) in the D Series, or else varied widely

between the two series but were either above or below 50% on

both series. The mean difference between A%OE and D%OE for

the narcotics violators in the PDA group was 41, for the rest

of the PDA subsample, 20.1. Hence, it could be surmised that

narcotics violators tend-to be more stimulus-governed.

Whether it be curiosity or an immature inability to control

the environment, there seems to be a tendency for narcotics

violators to go far more easily with the drift of the stimulus,

just as they probably tend to seek stimulation by drug-induced

means rather than by other means.

Petrie views the stimulus-governed as essentially immature

individuals, and has data to support such a developmental

view. She sees them as those for whom the external environ-

ment is contracting and expanding in size and intensity,









depending on external chance events, and sees this as coming

more under control with increasing age. This suggests that

the stimulus-governed narcotics violator, therefore, might

be viewed as an individual who had never developed the mature

perception of a stable external environment.


Situational Factors and Perceptual Style

Comparison of two populations, students and prisoners,

which differ greatly along not only personality lines and

also by environmental situations, suggests that both person-

ality and living conditions play their part in the perceptual

phenomena. Not only are the prisoners more stimulus-governed

than the students, but their overall overestimation scores

seem to relate to self-image measures (and especially to use

of extreme terminology regarding the self) in a fashion which

is diametrically opposed to those of students. Thus, in the

AD condition (where there is greater consistency between

A and D Series and therefore which may be looked at as a more

accurate estimate of the phenomenon we are attempting to

measure), it is the student "reducer" but the prisoner

augmenterr" who shows linear relationships with the criteria.

Thus, the student who (in the evolving terminology of the

present study) moves but little from the anchor tends to have

both a high self-ideal and self-other discrepancy and to use

extreme response categories in describing himself, his ideal,

and what he thinks others perceive him to be.









Although the data in the present study give us no clue

as to why there should be this discrepancy, two lines of

speculation suggest themselves. One is the extremely low

self-esteem in which female prisoners may hold themselves.

Thus, Bertrand (1967) notes that'

Criminal women and delinquent girls perceive them-
selves much more often than their non-delinquent peers
(and more than delinquent males) as compelled, driven,
"broken in initiative." Their deviance from their
socially ascribed role is so great that it calls for
a reflection, in the others' eyes, of abnormality and
even of illness. Having failed to contribute to the
social function of "pattern maintenance" entrusted to
women in our western societies, having offended the
family values of which they have been constituted the
guardian, criminal women and delinquent girls must be
regarded as "sick" and/or as pitiful, victimized and
manipulated beings. In four out of seven of our groups,
the self-percept of female delinquents went so far in
the Object direction as being one of serious psycho-
social pathology (p. 2771).

If this is so, it could be expected that a perceptual style

such as the one being investigated, which is presumably less

subject to change than are changes in the personality, could

cease being so directly related to self-image discrepancy

measures in a linear fashion once the self-image has come

under the weight of society-imposed guilt. Consequently, in

this explanation the constancy of perceptual anchoring/

reactance mechanisms would be viewed as remaining essentially

intact no matter what variation took place in the external

situation of the individual, but that once that external

situation had distorted an individual's self-image too badly,

perceptual styles would cease to be related to self-image

discrepancies.








A second line of speculation to explain these differ-

ences would be on the basis of divergent reinforcement

histories. Students might be thought of as generally en-

couraged to seek out new modes of thinking and acting, to be

daring in their attempts to find truth. Prisoners could be

viewed as finding nonconformity punishable and being more

prone to remain anchored to the habitual. If a student

would not feel free to move from a safely anchored position,

i.e., if he were to have some rigidity, he would be also

more than likely to see himself as not measuring up to the

ideal of the student set before him by the society in which

he lives, as being misunderstood by his peers, and'would

probably use more extreme expressions to describe himself,

his ideal, and his belief of what others think of him. The

prisoner, on the other hand, who would not feel free to move

from an anchor perceptually would find that what for the

student would be rigidity is for him a comfortable conformity

to the role of prisoner expected of him. Moreover, he pre-

sumably has been punished for stepping outside the social

norm, and that lesson would be present to him by his very

presence in prison. He would, therefore, feel less distant

from his ideal, and use fewer extreme expressions to describe

himself, his ideal, and his view of what others see him to be.!

Whatever the explanation, it will remain for future

investigation to find the variables which influence such

disparate findings for the two populations on the relation

between perceptual reactance and self-image discrepancies.









Lines of Suggested Future Research

Several avenues of fruitful research seem opened by the

present study. One is that future investigation on the

perceptual reactance measure employed here may wish to refine

it. This could be done in several ways.

First, it seems fairly evident that it is the A Series

by itself which holds the most promise for reliability and

usefulness in self-image work. On the other hand, it is the

DA order which turned up the greatest number of stimulus-

governed individuals. Hence, the future investigator may

wish to choose a particular order and/or series for discover-

ing particular types of individuals.

Secondly, more attention needs to be paid to the point

in the total range at which the anchor stimuli are presented,

the discrepancy between the gaps at the beginning of the

A vs. the D Series, and the non-equality of the two series.

Third, the comparability to Petrie's work could be made

more exact by creating some kind of spatial or non-verbal

estimator, rather than a written task as in the present

investigation, to judge the size of the squares. This might

be a device by which the subject would adjust a square with

movable sides to the size of the one that is shown on the

screen.

Another area for investigation would be to inspect the

relation between the measure used here and the individual's

tolerance for pain and/or sensory deprivation.








These relationships, it will be remembered, were found very

early by Petrie in connection with her Block Comparison Test.

It would be well to sort out whether Petrie's work genuinely

found a completely different sort of thing from the present

one with regard to sensation seeking and avoidance, or

whether both deal with anchoring effects which are related in

a more complex fashion to avoidance behavior.

A third area of research would be into the environmental

conditions surrounding two groups so diverse as students and

prisoners. Much of the difficulty in interpreting the results

of the present experiment stemmed from obtaining opposite

results from the two groups. Consequently, more information

might be solicited concerning the defining characteristics

of their two environments, as well as what sorts of things

are reinforcing to each.

And finally, more research is needed into the subjective

meaning of self-other discrepancy. Very little seems to have

been done to date on this particular aspect of self-image

discrepancy, as more of the research has been concentrated on

self-ideal discrepancy or on self-vs.-the-way-others-see-me

discrepancy (i.e., based on others' actual reports of what

they think of me). The interestingly high relationship which

was obtained in the present sample between visual perceptual

overestimation and self-other discrepancy would indicate that

the subjective meaning of the latter should be more fully

investigated.














CHAPTER V

SUMMARY


This study set out to test the hypothesis that discrep-

ancies between perceived and ideal self, and between per-

ceived and thought-perceived-by-others self,are related not

only to anxiety, defensive style, and social competence, but

also to such a conflict-free ego dimension as perceptual

reactance.

Ninety-six white women, half of them graduate and

undergraduate students at the University of Florida, and

half of them prisoners in a Florida institution of maximum

security, served as subjects for the study.

Subjects were administered Achenbach and Zigler's

30-item questionnaire under three sets of instructions:

self as seen by self, self as ideally desired, and self as

thought perceived by others. The predictor variables were

a Squares Estimation Task, Zuckerman's Affect Adjective

Check List, Byrne's Revised Repression-Sensitization Scale,

and a dichotomization of the sample by student vs. prisoner.

The Squares Estimation Task consisted of 150 black squares

on a white background, projected on a screen. Half of the

presentations progressed in ascending, half in descending,

order, and order of presentation of the two series was

counterbalanced for the sample. Since the two orders of









presentation were found to have widely divergent reliabili-

ties, and since subject status was found to reflect two very

different sets of personality variables, the data were

analyzed within these variables rather than as independent

main effects.

The results gave some support to the predictive power

of perceptual reactance for self-image discrepancies. Those

given the Descending Series of perceptual stimuli first

produced less reliable results, and hence straightforward

interpretation of that part of the study was made difficult.

Those given the Ascending Series first, on the other hand,

were consistent on the total series and showed some

interesting results.

The students in Ascending Series first who underestimated

were significantly more prone than their overestimating

counterparts to use extreme response categories in respond-

ing, and to have a high self-other discrepancy. Among this

subgroup perceptual reactance constituted a better predictor

of these criteria than did anxiety or defensive style. On

the other hand, perceptual reactance did not predict self-

ideal discrepancy with any significance, although the

underestimator was prone to have a high self-ideal discrepancy.

Among the prisoners given the Ascending Series first,

on the other hand, it was the overestimator who tended to

use extreme response categories. No strong relationship

in the prisoner group, however, was found to exist between

perceptual reactance and the self-image discrepancy measures.









The influence that order of presentation had on per-

formance was discussed in terms of a need to modify percep-

tual reactance theory to include the concept of set. The

influence that subject status had on performance was

discussed in terms of environmental and social history.

































APPENDIX A

Self-Concept Measures













Questionnaire R


For each of the following statements, indicate which of the
following answers (a, b, c, d, e, f) indicates how true the
statement is.

Answers:
a. This is very true of me
b. This is quite true of me
c. This is slightly true of me
d. This is slightly untrue of me
e. This is quite untrue of me
f. This is very untrue of me


Statements:

1. I often let myself go when I am angry.

2. I can do certain things much better than the average
person.

3. I like to take life easy.

4. I am willing to help other people who are in trouble.

5. I feel out the opinions of others before making a decision.

6. I enjoy myself at parties or other social gatherings.

7. I have a sense of responsibility about my duties. I do
what is expected of me.

8. I enjoy a good hot argument.

9. My likes and dislikes change frequently.

10. I have a good sense of humor.

11. I am ready to stand up for my rights.

12. I prefer easy tasks to difficult ones.

13. I go my own way regardless of the opinion of other people.

14. I am shy with men.









15. I get bored rather easily.

16. I depend a lot upon the judgment of my friends.

17. When I meet a stranger I think he is a better man than
I am.

18. I don't mind having jokes played on me.

19. I make enemies without realizing it.

20. I feel uncomfortable if I have to be by myself any length
of time.

21. I often think about how I look and what impression I am
making upon other people.

22. I can work at a difficult task for a long time without
getting tired of it.

23. I bear grudges.

24. I often imitate or agree with someone whom I consider a
superior person.

25. I do favors for friends whenever I can.

26. I sometimes act on the spur of the moment, without
stopping to think.

27. I have great faith in my own ideas.

28. I am able to keep working day in and day out without
getting tired or bored.

29. I often feel annoyed with myself.

30. I make friends easily.














Answer Sheet for Questionnaire R


NAME


DATE


1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.


16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.


L---~--














Questionnaire I


For each
which of
how true

a.
b.
c.
d.
e.
f.


of the statements on Questionnaire R, indicate
the following answers (a, b, c, d, e, f) indicates
the statement is.

I would like this to be very true of me
I would like this to be quite true of me
I would like this to be slightly true of me
I would like this to be slightly untrue of me
I would like this to be quite untrue of me
I would like this to be very untrue of me














Answer Sheet for Questionnaire I


NAME


DATE


1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.


16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.


--




84







Questionnaire O


For each of the statements on Questionnaire R, indicate
which of the following answers (a, b, c, d, e, f) indicates
how true the statement is.


a. Others hold this to be very true of me
b. Others hold this to be quite true of me
c. Others hold this to be slightly true of me
d. Others hold this to be slightly untrue of me
e. Others hold this to be quite untrue of me
f. Others hold this to be very untrue of me














Answer Sheet for Questionnaire O


NAME


DATE


1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.


16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.































APPENDIX B

Squares Estimation Test













Answer Sheet for Schematizing Test
(the proper responses have been filled in)


1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.


1.2

1.6

2.0

2.4

2.8

2.0

1.6

2.0

1.2

2.4

1.2

2.8

2.0

2.4

1.6

2.0

2.4

2.8

3. 2

3.8

2.8


23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.


3.8

2.0

3.2

2.0

3.8

2.8

3.2

2.4

2.8

3.2

3.8

4.6

5.5

3. 8

3.2

5.5

2.8

4.6

2.8

5.5

3.8


45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

-63.

64.

65.


3.2

3.8

4.6

5.5

6.6

7.9

5.5

4.6

7.9

3.8

6.6

3.8

7.9

5.5

6.6

4.6

5.5

6.6

7.9

9.5

11.4


67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.


22. 2.4


S88. 9.5


6.6

11.4

5.5

9.5

5.5

11.4

7.9

9.5

6.6

13.7

11.4

9.5

7.9

6.6

9.5

7.9

13.7

6.6

11.4

6.6

13.7


44. 4. 6


66. 7.9




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs