A construct validation of the egocentricity index and the form-dimensional and vista type responses of Exner's comprehen...

MISSING IMAGE

Material Information

Title:
A construct validation of the egocentricity index and the form-dimensional and vista type responses of Exner's comprehensive Rorschach system
Physical Description:
Book
Creator:
Unger, Michael Andrew
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 21058717
oclc - 16927337
System ID:
AA00020396:00001

Full Text






A CONSTRUCT VALIDATION OF THE
EGOCENTRICITY INDEX AND THE
FORM-DIMENSIONAL AND VISTA TYPE
RESPONSES OF EXNER'S COMPREHENSIVE
RORSCHACH SYSTEM
















By

MICHAEL ANDREW UNGER


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



UNIVERSITY OF FLORIDA


1985


I II





































To my parents with much love
and devotion











ACKNOWLEDGEMENTS


A seed is planted. If the soil is fertile and the environment is

optimal for growth, the seed will blossom and flourish and all will bene-

fit from its beauty. In a similar manner, several individuals contri-

buted to the development and fruition of this project in a number of

ways. Dr. Eileen Fennell, as committee chairman, was a constant source

of support, guidance, and understanding. Like a gardener, she was there

whenever needed, yet fully encouraged the independent growth of my ideas

and skills. An idea acts as a seed in any study and Dr. Hugh Davis was

influential on my thinking through his conveyance of his appreciation and

understanding of the Rorschach and human behavior. Dr. Russell Bauer

offered his insights and ideas which were appreciated both on and off of

the field. The contribution of Dr. Randy Carter cannot be underesti-

mated. His ability to communicate the mind-bogging complexities of

statistics in a manner that is comprehensible and readily grasped is

truely appreciated. Dr. Lawrence Siegal was a steadying force and his

ideas contributed much to the methodological procedures. Much thanks are

extended to Dr. Nathan Perry who filled the vacated position of a

committee member at a late date in the project.

The actual running of the experiment would not have been possible

without the help of Dr. Richard Griggs who provided the subject popula-

tion through the psychology department subject pool. My fellow experi-

menters Arthur Brand, Mary Morris, Ernie Bordini, Ann Freund, Julie

Lipovsky, Leslie Fried, Marika Spevck, Linda Abeles, Karen Froming, and

Patricia Leavy provided invaluable assistance. I am deeply indebted


-iii-









to them for their time and devotion. Finally, I would like to thank my

family whose love and devotion, like the sunshine, fueled my spirit and

provided me with strength.


-iv-
















TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS . . . . . . . . . .... iii

LIST OF TABLES . . . . . . . . . . . vii

ABSTRACT . . . . . . . . . . ix

CHAPTER

I LITERATURE REVIEW . . . . . . . .. 1

Introduction . . . . . . . . . 1
The Rorschach Test. . . . . . . . 4
The Form-Dimensional Response: (FD) . . . .. 8
Reflection and Pair Responses: (R) and (2). . . 13
The Shading-Dimensionality Response: (Vista) . . 25
Statement of the Problem . . . . . . 30
Hypotheses . . . . .... . . . . . 33

II METHOD . . . . . . . #. . . . . 36

Subjects . . . . . . . . . . 36
Assessment Instruments . . . . . . . 37
The Self-Focus Sentence Completion Test . .. 37
The Self-Consciousness Questionnaire . .. 39
The Rosenberg Self-Esteem Scale ........43
The Profile of Mood States.... . . . 45
The Rorschach Inkblot Test. . . . . . 48
Experimental Procedure . . . . . . .. 51
Experimenters . . . . . . . . 51
Setting . . . . . . . .. 51
Original Contact . . . . . ..... 52
Assignment to Groups . . . . ..... 52
Experimental Session. ..... . . . . 53
Interrater Reliability of the Projective
Instruments . . . . . . ... 55
Rorschach Inkblot Test . .. . . . 55
Self-Focus Sentence Completion Test . . . 56
Hypotheses and Analyses . . . . . . ... 57










TABLE OF CONTENTS
(Continued)


CHAPTER Page

III RESULTS . . . . . . . . . . . 60

Evaluation of Possible Examiner and Test
Order Influences . . . ..... . . . 60
Interrater Reliability . . . . . .. 61
The Rorschach Inkblot Test. .. . . . 61
The Self-Focus Sentence Completion Test . . 61
Evaluation of the Hypotheses ..... . . . .. 61
The Form-Dimensional Response . . . . . 61
The Vista Response . . . . . . . 67
The Egocentricity Index . . . . 74
Further Exploration of the Egocentricity Index,
Form-Dimensional, and Vista Categories. . .... 84
Correlation Matrix ..... ........ . ... 89
Self-Esteem . . .......... . . . 89
Self-Focus . . . . . . . . ... 89
Point Biserial Correlation . . . . . . .. 91
The Form-Dimensional Response . . ... 91

IV DISCUSSION. . . . . . . . . . . 92

The Form-Dimensional Scoring Category . . . ... .93
The Vista Scoring Category ....... . . . 97
The Egocentricity Index . . . . . . .. 101
Further Examination of the Rorschach Responses . . 105
Evaluation of the Design . . . . . . .. 106
Implications for Future Research . . . . . 109

APPENDIX

INFORMED CONSENT TO PARTICIPATE IN RESEARCH. . . .111

REFERENCES. . . . . . ... . . . . . ... 116

BIOGRAPHICAL SKETCH . . . . . . . . . . .. 121


-vi-
















LIST OF TABLES


Table Page

1 Means and Standard Deviations of the
Form-Dimensional Groups. . . . . . . . 63

2 Summary of the Form-Dimensional Group
Classification by the Discriminant Function
Analysis . . . . . . .. . . . . 64

3 Stepwise Selection of Variables for the
Form-Dimensional Groups . . . . . ..... 66

4 Summary of the Chi-Square Analyses using the
Form-Dimensional Groups as the Classificant. . . 68

5 Means and Standard Deviations of the Vista Groups. 70

6 Summary of the Vista Group Classification by the
Discriminant Function Analysis . . . . . 71

7 Stepwise Selection of Variables for the
Vista Groups . . . .. . . . . . ... 73

8 Summary of the Chi-Square Analyses using the
Vista Groups as the Classificant . . . .. .. 75

9 Means and Standard Deviations of the Egocentricity
Index Groups . . . . . . . . . . 78

10 Summary of the Egocentricity Index Groups
Classification by the Discriminant Function
Analysis . . . . .. . . . . . . 79

11 Backward Stepwise Selection of Variables for the
Egocentricity Index Groups . . . . . . 81

12 Summary of the Egocentricity Index Group
Classification by the D and SN Scores of the
SFSC and the Anger Subscale. . . . . . .. 82

13 Summary of the Chi-Square Analyses using the
Egocentricity Index Groups as the Classificant . 85


-vii-










LIST OF TABLES
(Continued)

Table Page

14 Means for each of the Egocentricity Index, Form-
Dimensional and Vista Category Groups. . . . 88

15 Pearson Product Moment Correlations between
all Variables. . . . . . . . . 90


-viii-


















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


A CONSTRUCT VALIDATION OF THE
EGOCENTRICITY INDEX AND THE
FORM-DIMENSIONAL AND VISTA TYPE
RESPONSES OF EXNER'S COMPREHENSIVE
RORSCHACH SYSTEM

By

Michael Andrew Unger

August 1985

Chairman: Eileen Fennell

Major Department: Clinical Psychology

The Rorschach Inkblot Test has stimulated great interest, extensive

use, and a considerable amount of research. A number of Rorschach sys-

tems have been developed with respect to administrative procedures, scor-

ing criteria, and the interpretive hypotheses that they generate. The

present study focused on the Comprehensive Rorschach System developed by

Exner. Specifically, three scoring categories (determinants) introduced

by Exner were examined. The validity of these determinants was examined

by evaluating the extent or degree that they individually and/or collec-

tively measure the construct or psychological process that they hypothe-

tically represent.

The three scoring categories that were explored are those that de-

signate the form-dimensional response, the vista response, and the


-ix-










egocentricity index. Each scoring category is hypothesized to reflect a

psychological process or perceptual-cognitive process that is used to

create the response. The form-dimensional response is proposed to re-

flect a nonemotional introspective process. The egocentricity index is

hypothesized to represent a self-focusing process characteristic of ego-

centricity. The vista response is hypothesized to represent a painful

introspective process that is associated with depressive features.

The current study was conducted to further examine the psychological

processes represented by these variables. A nonpatient population of

undergraduate students was used. Four questionnaires were used that were

thought to represent the psychological processes hypothesized to be re-

flected by each variable. The Self-Focus Sentence Completion Test was

used to assess the tendency to be self or other-centered. The Profile of

Mood States was selected to provide an estimate of affective state. The

Self-Consciousness Scale was used to assess the process of self-focused

attention. The Rosenberg Self-Esteem Scale was employed to measure self-

regard.

Discriminant function, stepwise discriminant, backward stepwise

logistical regression, univariate (analysis of variance) and Chi-Square

analyses were utilized. These analyses failed to support a hypothesis of

convergent validity for the vista and form-dimensional responses. Public

self-consciousoness, self-esteem, and the mood state of anger were found

to be significantly related to the egocentricity index. These results

are discussed in terms of implications for future research.















CHAPTER I
LITERATURE REVIEW

Introduction


The Rorschach is a test that has stimulated great interest, exten-

sive use, and a considerable amount of research. First introduced in a

posthumously published monograph (Rorschach, 1921), the Rorschach has

since become an important instrument for the assessment of personality.

The test consists of ten inkblots, and subjects are required to verba-

alize what they perceive them to represent.

Rorschach himself only studied the inkblots for a brief period of

time prior to his premature death. He offered a variety of postulates

concerning the specific features of the test, but did not formulate a

global theory (Rorschach, 1921). This challenge was accepted by Beck,

Beck, Levitt, and Molish (1961), Hertz (1951), Klopfer, Ainsworth,

Klopfer and Holt (1954), Piotrowski (1957), and Rapaport, Gill, and

Schafer (1946), each of whom developed his own Rorschach system. The

five Rorschach systems all focus on the verbalizations of the subjects

whether generated during the free association or the inquiry as their

basic data. These are always interpreted regardless of the system

employed. In addition, all of the systematizers emphasize a global

approach to the interpretation of the test and require that all of the

data be integrated before drawing conclusions. Hypotheses are formulated

and either accepted or rejected on the basis of both the structural and

content components of the protocol.


-1-








-2 -


Many of Rorschach's original postulates form the basis on which the

five Rorschach systems have been developed. However, each system differs

significantly from the others with respect to the administrative proce-

dures, scoring, and the interpretative hypotheses that it generates.

These fundamental differences have made systematic research efforts of

the test difficult. For instance, researchers often fail to indicate

which system they are using and/or use more than one system when

conducting their studies. They also tend to interpret their results

under the assumption that they are valid across all systems regardless of

which systems) was used in conducting the study. This practice

precludes the usefulness of their findings since the responses scored on

the basis of one system do not necessarily coincide with those scored by

the others due to their different scoring criteria. This, in turn,

affects the descriptive statements generated since each of the

systematizers offer interpretive hypotheses based on the scoring

categories that are specific to their system. This implies that many of

the interpretive hypotheses are specific to one particular system and do

not necessarily have an equivalent in the others.

Confusion has also found its way into clinical practice.

Practitioners are often faced with the problem of becoming proficient in

the use of one Rorschach system at the cost of disregarding what might be

worthwhile in another. Those clinicians with a knowledge of more than

one system may choose to disregard intersystem differences and intermix

them. Exner and Exner (1972) conducted a survey and found that

75 percent of those clinicians who score subjects' responses admitted

to intermixing the scores from different systems and/or adding







-3 -


scores derived from personal experiences. Exner (1974) states that while

this methodology has been adapted by the practitioner, it is not a truely

standardized methodology. He has called for a complete standardization

of the Ror. aach, a process that had been impeded because the research

conducted has concerned not one, but several tests which share in common

the same stimulus figures. This, in turn, has precluded the compilation

of adequate reliability and validity data.

Exner (1974) developed a Comprehensive Rorschach System which he

believes will make the test more accessible to the researcher and more

productive to the practitioner by providing a standardized base from

which the test can be used, researched, and fully developed. His system

was designed to incorporate the best aspects of the other systems, to

include developments exclusive of any of the five systems, to have high

reliability and validity, and to provide descriptive information. Like

those before it, the Comprehensive System focuses on procedures, scoring,

interpretation, and research problems and methodologies. It borrows

liberally from the other systems and is intended to provide those who use

it with a common language and methodology.

Exner (1974) required that the Comprehensive Rorschach System have

high validity in response to criticisms against the Rorschach concerning

its validation. Goldfried, Stricker, and Weiner (1971) believe that the

question "Is the Rorschach valid?" should be replaced in favor of the

question "What is the Rorschach valid for?" Blatt (1975) points out that

in clinical practice, the Rorschach has typically been used to make

inferences about certain psychological constructs and Meehl (1959)

recommends a construct validation approach as the appropriate model for








-4 -


conducting research related to the Rorschach if it is to yield clinically

useful information. The current study was conducted in an effort to

further establish the construct validity of certain aspects of the Com-

prehensive Rorschach System. Specifically, three scoring categories

(determinants), introduced by Exner (1974), were examined. The extent or

degree that these determinants individually and/or collectively measure

the psychological construct or process that they hypothetically represent

was evaluated.

The Rorschach Test

The data of the Rorschach is a verbal report from the subject of

what he has seen or imagined when confronted with the ten ambiguous

inkblots (Exner, 1974). Exner (1978) has hypothesized that there are at

least four interrelated processes involved in each verbal response given

by the subject. First, the subject must acknowledge that the task is to

see something other than an inkblot and to categorize the many objects

which the inkblots or its parts might resemble. Second, the subject rank

orders the possible responses that he has detected. This process is

influenced by the individual's concern for perceptual accuracy, needs,

and internal sets. These needs and internal sets may be integral parts

of the individual's personality structure, manifest in the form of his

basic response style, or be more situational and transitory. Third, the

subject must feel or believe that his response will be acceptable. The

last factor hypothesized to influence subject's response is the manner by

which he verbally articulates his answer.

There are few rules and limited structure provided by the Rorschach

test for the individual to use to guide his behavior. He is forced to








-5 -


use his psychological resources and the behaviors which, in his judgment,

will produce the most acceptable response. From these responses,

descriptive statements can be made by the clinician that focus on such

features of the individual as response styles, affectivity, cognitive

operations, motivations, preoccupations, and interpersonal-interenviron-

mental perceptions and response tendencies (Exner, 1974).

To assist in making descriptive statements, each Rorschach system

has a scoring process that converts the individual's verbal responses

into a logical and systematic format so that they can be studied

individually and collectively. This provides the structural data whereas

the verbal responses provide the content data from which hypotheses are

generated. The current study was concerned only with the structural

data.

Rorschach (1921) originally devised a system of symbols and

abbreviations to code each verbal response into a language which

identifies its characteristics and components. His scheme for scoring

consisted of four basic categories: 1) Location (the area of the inkblot

where the response occurred); 2) Determinant(s) (the stimulus features of

the blot that contributed to the formation of the percept); 3) Content

(the content of the response); and 4) Popularity (whether the response is

common to the general population) (Exner, 1974). While all of the

systematizers have maintained this format in their respective systems,

Hertz (1940), Beck (1937), and Exner (1974) have each added a scoring

category designated as "Organizational Activity," to represent the extent

to which the individual has organized various features of the inkblot in

a meaningful way (Exner, 1974).








-6 -


An investigation of each of these scoring categories would surely

have been fruitful for research, but was beyond the scope of this paper.

The current study was concerned solely with the determinant scores of the

Rorschach. The scoring of determinants provides information concerning

the perceptual-cognative process which has produced the response (Exner,

1974). In the Comprehensive System, the many stimulus characteristics of

the inkblots fall into nine determinant scoring categories: 1) form, 2)

movement, 3) color (chromatic), 4) color (achromatic), 5) texture

(shading), 6) general shading, 7) dimensionality (form), 8) dimension-

ality (shading), 9) pairs and reflections (Exner, 1974). There has been

comparatively little research conducted to confirm which psychological

constructs the last three of these nine determinants are hypothesized to

represent. It was therefore decided to further limit the focus of the

current study to these scoring categories.

The determinants which Exner (1974) introduced in his comprehensive

system and which were investigated in this study have historically been

scored as "light-dark" determined responses by other systematizers. The

scoring of these "light-dark" determined responses is an area about which

Rorschach wrote very little and, as a result, each systematizer has

developed his own criteria (Exner, 1974). For example, Beck et al. (1961)

scored the "vista" response as one in which the variations in shading gave

a three-dimensional effect, as of something seen in perspective. His

criteria for content responses scorable as vista were "the distant,

usually landscapes, sometimes aerial views, occasionally astronomical

percepts, the heights in landscapes; architecture, usually as height and

less often as being distant; insularity, usually land in water, but








-7 -


sometimes water surrounded by land; depths, usually water, but also caves;

reflections, usually in water, sometimes in mirrors" (Beck et al., 1961,

pg.109). For a response to be scored as vista, the individual's verbali-

zations were required to refer to distance, height, depth or reflection.

Beck hypothesized that the vista response represented "sensed pain" or the

character trait of inferiority feelings (Beck et al., 1961).

Klopfer, Ainsworth, Klopfer and Holt (1954) used the symbol "FK" to

designate those responses verbalized by an individual where a

three-dimensional or depth impression is combined with definite form

perception, giving perspective to the content. The symbol "Fk" is used

to designate a response where the use of shading gives the impression of

a three-dimensional plane involving definite shape. "K" or "kF" are

scored when the individual uses shading to give a diffuse, unstructured,

but three-dimensional effect to the content of their response. Klopfer

et al. (1954), like Beck et al. (1961), did not always require that the

individual directly refer to the light-dark features (shading) of the

inkblot in his verbal response or during inquiry and believed that it is

the concept, rather than the explicit statement, which gives the clue as

to the use of shading. The Fk score is also used to refer to reflections

and symmetry responses. However, the form determinant, and not Fk, is

scored in those instances where a reflection response is solely based on

the symmetry of the inkblot and there is no evidence that shading was

used to create a three-dimensional effect. The form determinant is also

endorsed when perspective is thought to be linear and the shading is not

used to create a three-dimensional effect. The FK scoring category and

its varients are hypothesized to psychologically represent an attempt by







-8 -


the individual to cope with affection and anxiety by introspective

efforts (Klopfer et al., 1954). In other words, they are interpreted to

represent an attempt by the individual to cope with the anxiety generated

by his needs for security, affection and belongingness by putting his

problem at a distance from himself so that he can view it more dispassion-

ately. Klopfer et al. (1954) believed that the FK type response appeared

to be characteristic of those individuals who are unsuccessful in coping

with their anxiety by more intellectual means and lack the emotional in-

sight needed to cope effectively with their problems.

Exner (1974) has introduced into the Comprehensive Rorschach System

four determinant scoring categories to designate those responses that Beck

et al. (1961) would score as vista and Klopfer et al. (1954) would score

as a FK type response. These determinants will be discussed in the

following passages with respect to their scoring criteria and the research

conducted thus far to determine the psychological processes that they

reflect.

The Form-Dimensional Response: (FD)

Exner (1974) introduced the Form-Dimensional (FD) determinant to

designate those verbal responses to the inkblots which include perspective

or dimensionality based exclusively on form, interpreted by size or in

relation to other blot areas. Although there is no scoring category

compatible to it in the other Rorschach systems, Klopfer et al. (1954)

have included such responses in the FK category (even though no shading is

involved) arid Beck et al. (1961) have scored these responses as vista when

the unarticulated use of shading seems very likely. Form-dimensional

responses have been observed to occur in responses to each of the inkblot







-9 -


cards (Exner, 1974). Although the criteria for scoring FD have not been

finalized, it is scored when the dimensionality or perspective is based on

the size of the blot area used or the relationship between blot areas.

The usefulness of FD as a separate scoring determinant emerged from

a study of sixty Rorschach records obtained from subjects who, during a

period of ninety days following testing, made a suicide attempt, eighteen

of which were successful (Exner, 1974). Examination of these protocols

revealed that on the average, they contained a significantly greater

number of form determined dimensional-perspective responses (X = 3.1) as

compared to a nonpsychiatric group (X = 0.8). Exner also observed that

the FD type of response occurred with considerable frequency in the

records of a variety of subjects whether they were from psychiatric or

nonpsychiatric populations. Using this particular scoring criterion,

243 FD responses were tallied from 250 Rorschach records whereas only 186

vista responses were tallied as scored by the Beck et al. (1961) Rorschach

system (Exner, 1974). The mean number of FD type responses and the

frequency with which it occurs increases consistently with the increase

in a person's age (Exner, 1978). Beginning at age eight, this type of

response has been observed to appear at least once in more than half of

all individuals tested.

It was first believed that the FD response was related to the

depressive features characteristic of suicidal individuals. Further

examination of various Rorschach protocols showed that an outpatient

reference group gave an average of .23 FD type responses and a schizo-

phrenic reference group gave an average of 0.5 FD responses per record.

Nonpatients were found to give 0.92 of these responses per record








-10-


(Exner, 1974). It was reasoned that since outpatients are encouraged to

be introspective during therapy, and depressives, by definition, are

stimulated toward a form of introspection, the FD response may be re-

lated to introspective activity. Exner has offered some tentative re-

search looking at this relationship.

Exner (1974) cites three experiments that were conducted to evaluate

the hypothesis that the FD response is related to an introspective

activity. In the first study, a total of 495 Rorschach records were

collected from four groups of subjects: nonpsychiatric, outpatient

nonpsychotic, inpatient nonschizophrenic, and inpatient schizophrenic.

It was assumed that an introspective person would be more prone to

internalization and less prone toward affective display. Since the "M"

determinant is thought to reflect an internalization process and "C" a

proneness toward affective display, subjects were sorted into three

groups based on the number of these responses. Only two groups, sum M

greater than sum C and sum M less than sum C, were used to compare the

number of FD responses produced. It was found that the sum M greater

than sum C group gave a greater than average number of FD responses.

This provided some support for the notion that FD is related to

internalization. However, it may be the case that FD simply represents a

delay of affect. The question of introspectiveness per se had yet to be

addressed.

The second study was designed to further evaluate the process of

introspectiveness as related to the FD response. Forty subjects were

selected from a "waiting list" of a psychiatric facility and randomly

assigned to four equal groups. They were instructed that they would meet









-11-


in "holding groups" to initiate their treatment plans and goals prior to

being assigned to an individual therapist. During these meetings,

trained observers scored the verbalizations of each subject to determine

if they were directed toward the "self" or others and whether their

statements pertained to past, present, or future. Rorschachs were given

to all subjects prior to their first group meeting. They were then

assigned to either an "upper" or "lower" group based on a median split of

the total number of FD responses given. It was assumed that more intro-

spective subjects would make statements that focused more on themselves

than others. A comparison of the two groups found that the upper FD

group gave significantly more "self" verbalizations during the first

three group sessions. A comparison of the number of statements made

directed towards others was not reported. Although it was predicted

that more introspective individuals would be more concerned with the

present than with the past or future, no differences were found between

groups on these variables. The only clear result of this study was that

high FD individuals made more self-referenced statements than individuals

who gave a fewer number of FD responses. No control group was used to

compare these results to a group of individuals who gave an average

number of FD responses.

It was next decided to test whether the self-oriented responses

given by the "high" FD group were a manifestation of egocentricity rather

than self-examination. To rule out this possibility, subjects were also

administered a Self-Focus Sentence Completion Test (Exner, 1973) as an

index for measuring egocentricity. Three subscale scores can be obtained

from this test: 1) the Self-Focus subscale score, 2) the External-World







-12-


Focus subscale score, and 3) a D score representing the difference

between them. Using this latter score to compare groups, it was found

that the high FD group was slightly more oriented toward the external

world than the lower FD group. Exner concluded that this would seem to

eliminate the possibility that the verbalizations of this group were

simply manifestations of egocentricity. No comparison of the groups on

the basis of the self-focus subscale score was reported. The reader is

left wondering if the high FD group endorsed more self-focusing item

stems than the low FD group. No interpretation was offered to account

for the results of the low FD group. Based on the author's criteria, it

would appear that these individuals are more egocentric than the high FD

group. Once again, comparisons to a control group were not reported.

Working from the conclusion that the FD response represents a pro-

cess different from egocentricity, Exner next tested its relationship to

introspection per se. It was decided to study this relationship within

the context of psychotherapy. It was assumed that subjects who are

trained to examine themselves during the early stages of treatment would

be more introspective and give a greater than average number of FD

responses. The Rorschach was administered to fifteen nonpsychotic

patients at the onset of individual psychotherapy, the tenth treatment

session, and termination. The mean number of FD responses given at these

times was .06, with a range of 0 to 4; 3.11, with a range from 0 to 6;

and 1.24, with a range of 0 to 3; respectively. It was believed that

these results provided some evidence that the FD type response is related

to a psychological activity involving self-inspection or examination or

at least self-awareness.








-13-


From these studies, Exner (1974) concluded that if the FD response

is also interpreted as a type of "form" response, as having affect-delay

operations, then the self-examination may be a form of affective-delay

through a type of reasoning which centers on the self, its assets, and

its limitations. It represents a "nonemotional introspective" process.

A number of criticisms can be raised against these studies. As al-

ready mentioned, the high and low FD groups were not compared to a

control group. Therefore, these subjects may not be more or less "intro-

spective" than the general population. A second criticism concerns the

third study. A number of variables other than introspectivenss per se,

such as level of anxiety, may have influenced the number of FD responses

given. The conclusion that the results were due to an introspective pro-

cess cannot be reached unless a more stringent criterion is used to test

this relationship. Introspection can be defined as an examination of

one's own thoughts and feelings (Duval and Wicklund, 1973). None of the

studies cited above directly tested whether subjects who give more FD

responses are more aware of their own thoughts and feelings than the

general population. In addition, none of the studies tested directly

whether an affective component is involved or absent from the FD

response. Therefore, the interpretive conclusions reached by Exner with

regard to this response still seem tentative and more the result of in-

ference than direct empirical testing.

Reflection and Pair Responses: (R) and (2)

The reflection and pair determinants are not scored separately in

any of the other Rorschach systems. Beck et al. (1961) and Klopfer et al.

(1954) each suggested that the reflection type of response be scored







-14-


as a type of light-dark response. Beck et al. (1961) score the reflec-

tion response that uses the light-dark features of the blot as a shading

vista response, and those based solely on the symmetry of the blot as a

form determinant response. Beck et al. assume that the same perceptual-

cognitive process is involved in the reflection type of response as that

involving depth and/or dimensionality. Klopfer et al. (1954) used FK to

designate the reflection response based on blot symmetry and if the midline

of the ink blot is noted as a shading differentiation.

Like the FD responses, the reflection (R) response is based on the

form of the inkblot. Exner designates two scoring categories for the

reflection response. The first, the reflection-form response (rF), is

scored for those responses in which "the symmetry features of the blot

are primary in determining the answer, and form is used nonspecifically,

or ambiguously, as an object being reflected" (Exner, 1974, p. 101). Re-

flection responses scored in this manner always involve content with non-

specific form requirements (i.e., clouds, shadows, rocks). The form-

reflection response (FR) is used to designate those responses where "the

form of the blot is used to identify specific content, which is inter-

preted as reflected because of the symmetry of the blot" (Exner, 1974,

p. 101). The reflection response, whether rF or Fr, is based on the

symmetry of the inkblot. The reflection response which does not use

symmetry and uses the light-dark features of the inkblot is scored

separately as either a vista or as a shading response. The pair (2)

response is based solely on the form properties of the inkblot and is

scored when "the perceived object is reported as two objects because of

the card symmetry" (Exner, 1974, p. 102).








-15-


The rationale for creating two separate scoring categories for re-

flection and pair responses emerged in a study originally designed to

examine narcissism and acting-out behavior (Exner, 1969a). The study was

based on a premise of narcissm expounded by Kohut (1966). Kohut de-

scribes a "narcissistic balance" which provides the impetus toward drive

satisfaction, but which is also tempered by social reality. When this

balance is disrupted, two forms of pathology may occur. The first

involves an excessive preoccupation with the self and the immediate gra-

tification of impulses which occurs in the majority of sexual perversions

and character disorders. The second takes the form of withdrawal which

occurs in many of the depressive reactions and some schizophrenia. The

study thus included three equal groups of homosexuals, sociopaths, and

depressives as their behavior is, in theory, narcissistic. A control

group was also included. All subjects were administered Rorschachs. When

their records were scored by the Beck et al. (1961) method, each group was

found to give a high frequency of vista responses. This was not the case

when a separate scoring category was used to designate reflection

responses. The homosexual and character disorder groups gave signifi-

cantly more reflection responses than the other two groups. The depres-

sive and control groups were not found to differ significantly in

this regard. On the basis of these findings, it was reasoned that the

scoring of the reflection response warranted a separate category for

diagnostic appraisal (Exner, 1969). This logic was also applied to the

pair (2) response which was also recorded on the assumption that they

might represent a more subtle or controlled form of reflection type

response (Exner, 1969). Pair responses occurred in almost all of the








-16-


records. Significant differences were found using t-tests to compare

groups. The homosexuals, as a group, gave more of these responses than

all other groups. The depressives gave fewer pair responses than all

other groups (Exner, 1969).

In light of these findings, Exner (1969) then decided to investigate

the relationship between the reflection and pair responses to the

concept of narcissism. A sample of 1,547 males composed of college

students and industrial workers was administered a 50-item sentence com-

pletion test originally designed by Watson (1965), and subsequently modi-

fied by Exner (1969), to assess the characteristic of narcissism. A

random sample of 750 sentence completion tests were scored and used in

the study. From this population, two groups of forty subjects were

selected to form a high and low narcissistic group. The high narcissistic

group was composed of those subjects who had high self-focus scores and

low external-world scores. The low narcissistic group was comprised of

those subjects who had low self-focus scores and high external-world focus

scores. Whether the difference between the self-focus and external-world

focus scores was significant was not reported. The high and low narcis-

sistic groups were compared using t-tests. The high narcissistic group

was found to average significantly more reflection and pair responses than

the low narcissistic group. No control group was used to determine if the

number of pair and reflection responses given differed significantly from

a "normal" population. Exner (1969) formulated the three following

postulates based on these findings. First, those individuals who are

highly narcissistic will manifest this trait in their perception and

verbalization of reflection or mirror-image type responses on the








-17-


Rorschach. Second, those individuals who are low on the narcissistic

trait will not exhibit reflection type responses and will have a

restricted number of pair responses. Third, persons demonstrating a

reasonable narcissistic balance in their adjustment can be expected to

verbalize a reasonable number of pair responses and few, if any, reflec-

tion responses. Exner further concluded that these findings seem consis-

tent with the psychoanalytic conception of narcissism as "self-love"

(Freud, 1931/1961).

One question left unanswered by this study concerned the depressives.

If depressed individuals are, in theory, narcissistic (Kohut, 1966), why

do they not verbalize as many reflection responses as the homosexual

or sociopathic groups? It was found that the depressive group gave a

significantly greater number of vista responses than any of the other

groups (exclusive of reflection). If this is so, what differentiates the

pair and reflection type responses from the vista response?

Exner reported that because of the complexity of the concept of nar-

cissism, it seemed "more appropriate" to use the concept of "egocentri-

city" or "self-focus" to interpret these data (Exner, 1969). No further

explanation was offered for this change in stance. Exner (1973) subse-

quently replicated the aforementioned study using a nonpsychiatric

population of college students. The sentence completion blank was

revised so as to include only 30 items and a more comprehensive scoring

system. Sixty subjects were selected to form two groups on the basis of

their extremely high or extremely low self-focus scores. Each subject

was subsequently administered the Rorschach. A comparison of the two

groups revealed that the high self-focus group produced nearly twice as








-18-


many pair and reflection responses. A comparison between groups on the D

score (self-focus score minus external-world focus score) was not

reported although this was designated as a criterion for narcissism in

the first study and egocentricity in a previously cited study investiga-

ting the FD response (Exner, 1974). No control group was used to

determine if the frequency of the pair and reflection type responses

differed from a normal population. Nevertheless, these results were

considered by Exner to provide some support for the notion that the

reflection and pair responses represent a self-focusing process.

In one of the few studies that was proposed to study the pair and

reflection response to the process of egocentritity itself, Exner (1974)

administered several psychological tests, including the Rorschach, to 21

candidates applying for a junior engineering position. Prior to their

interview, each candidate was brought into an office which contained a

two-way mirror and candidly filmed for a ten minute interval. Trained

raters scored the film for the amount of time each candidate spent

viewing himself in a mirror. A median split was used to divide subjects

into high and low groups based on their mirror viewing time. Significant

differences were found between the groups on their Rorschach productions.

The subjects in the upper half group gave significantly more reflection

and pair responses than the lower group. It is not clear how egocentri-

city is being defined here or whether it can be said to be represented

here at all.

In a separate paper, Exner also reported on the use of the Self-

Focus Sentence Completion Test to compare high and low mirror viewing

groups (Exner, 1973). The high mirror viewing group was found to score







-19-


significantly higher on the self-focus subscale than the low mirror

viewing group. The relationship between these variables with the number

of reflection and pair responses produced was not reported. Based on

other findings (Exner, 1974), it may be inferred that those subjects

who stared in the mirror for a greater length of time, and gave more

self-focused responses, would give more reflection and pair-type

responses than those individuals who spent little time looking in the

mirror. This is only an inference, and it is surprising that Exner did

not report on this relationship as supporting or contradicting his

hypothesis.

Raychauduri and Mukerji (1971) found that reflection responses

appear very infrequently among adult protocols except for those indivi-

duals with serious sex role confusion, those who are prone to antisocial

behavior, and schizophrenics. They conclude that the reflection response

is a "primitive" form of self-centeredness.

Based on the findings of the above studies, Exner (1974) proposed

that reflection and pair responses are related to an egocentric or self-

focusing process. He then developed an egocentricity index (3r + (2)/R)

to designate self-concern. It was decided to multiply the number of

reflection responses by three because of the relative infrequency with

which it occurs in all populations. For nonpatients this ratio ranges

from 0.25 to 0.40 responses per protocol. A number of studies were sub-

sequently carried out to determine the validity of the egocentricity

index. One such study was based on the assumption that children are very

self-centered, especially during the early developmental years. Exner

and Weiner (1982) found that reflection responses appear in half of the








-20-


protocols of children ages five through eight. The proportion of

Rorschach protocols containing reflections begins to diminish at age

nine. However, fifteen-year olds give a large proportion of these

responses. The pair responses were found to occur in the protocols of

children at every age group. With the exception of the children at the

five year age level, no statistically significant differences for the

mean number of pair responses given were found between age groups. The

3r (2)/R index was found to decline consistently as age increases

(Exner and Weiner, 1982).

Exner, Kuhn, Schumacher and Fishman (1975) examined the relationship

between the egocentricity index and the field dependence-independence

phenomenon or locus of control (I-E) variable. A rod and frame apparatus

was used to measure field dependence and Rotter's I-E scale to measure

locus of control. The subjects were 38 patients and 32 nonpatients.

Rank order correlations were used to test the relationship. A low

positive, but not significant relationship was found between locus of

control and the egocentricity index. The finding was the same for the

relationship between field dependence-independence and the egocentricity

index.

Exner and Weiner (1982) introduced a special Rorschach scoring

category, "Personal" (PER), to designate reference to personal knowledge

or experience in direct relation to a response. "Ordinarily, the

personal pronouns I, me, my, or we will be present in the PER response,

but it is not sufficient to justify the scoring. Rather the verbaliza-

tion to be scored implies a clear justification to the percept" (p. 321).

Exner and Weiner (1982) selected a group of 20 subjects from a pool

of 225 college students based on their responses to the Geough Adjective








-21-


Checklist. This scale is frequently used as an index of self-esteem

although it was not designed to do so. Hence, its validity and relia-

bility are highly suspect (Wylie, 1974). The twenty subjects selected

had endorsed the greatest number of positive adjectives and the fewest

number of negative adjectives. A sample of 20 subjects was selected

from the remaining pool to act as a control group. All were administered

the Rorschach. While the number of PER responses given per group differed

significantly, the groups were not found to differ significantly with

respect to their mean egocentricity index scores.

These findings are of interest for a number of reasons. The PER

response is not unlike the self-focus response of the Self-Focus Sentence

Completion Test developed and used by Exner in his 1969 and 1973 studies.

Both responses contain personal references. In this study, the relation-

ship between the number of PER responses and the egocentricity index was

not compared directly. If the egocentricity index does represent a self-

focusing process, subjects who endorsed the greatest number of positive

adjectives and gave a high number of PER resonses should have had a

significantly higher mean egocentricity index score than the control

group. The fact that they did not calls further attention to the point

made earlier with regard to the absence of control groups in many of

Exner's studies and sheds some doubt on the findings from previous

studies.

This was the first and only study by Exner relating self-esteem to

the egocentricity index. Although no significant relationship was found

between the indices, the questionable validity of the questionnaire used

still leaves this relationship open to further testing. In fact, Exner








-22-


(1978) stated that "although the egocentricity index is essentially

related to a self-focusing process, the issue of self-esteem is also

probably manifest, at least when the index is low" (p. 71). This state-

ment was based on the relatively high frequency of depressed subjects

and suicide prone patients who give a low index. Exner (1978) cautions,

however, that a high egocentricity index should not be interpreted to

indicate a form of "self-glorification." An excessive self-focus may be

related to low self-esteem and serve as a defense against feelings of poor

self-worth. In fact, higher than average egocentricity indices occur in the

Rorschach records of one in every five effective suicide cases.

A number of studies have looked at the relationship between the

egocentricity index (and its constituents) and psychopathology. Exner

(1974) studied three psychiatric groups: 75 inpatient schizophrenics, 74

inpatient depressive reactions, and 31 inpatient character disorders.

Three psychologists or psychiatrists used the Inpatient Multidimensional

Psychiatric Scale and a "significant other" used the Katz Adjustment

Scale-Form R to evaluate subjects pre- and posttreatment. Subjects were

divided into those who had changed significantly and those who did not

change significantly. The findings showed that improved schizophrenics

and character disorders, as a group, gave significantly fewer reflection

responses at posttreatment. In contrast, the depressives gave a greater

number of these responses at posttreatment as compared to pretreatment.

No differences were found for the number of reflection responses given

pre- and posttreatment by the unimproved group. With respect to pairs,

the improved schizophrenics gave significantly fewer of these responses

at posttreatment, the improved depressives gave significantly more, and








-23-


the improved character disorders showed no differences. From these

findings, Exner (1974) postulated that if reflection and pair responses do

represent a form of self-focus or egocentricity, too much or too little

may characterize pathological conditions. Posttreatment improvement is

characterized by a change toward a level of egocentricity similar to that

which is characteristically found in nonpatient records.

Exner, Wylie, and Kline (1977) found that the egocentricity index

changes in patients treated in therapy. Four hundred and thirty volunteer

subjects, ages 20 through 36, were solicited from private practitioners,

clinics and hospitals and assigned to one of seven modes of treatment.

The treatment modalities were 1) a psychoanalytic, uncovering form of

psychotherapy, 2) gestalt individual psychotherapy, 3) a modeling form of

treatment with the major focus on interpersonal patterns, 4) assertive-

ness treatment, 5) systematic desensitization, 6) transactionally oriented

group therapy, and 7) a biofeedback form of intervention. No effort was

made to study therapists or to match subjects for pathology. All subjects

were administered a Rorschach at pretreatment, eight to nine months after

the onset of treatment, sixteen to eighteen months after the onset of

treatment, and 27 to 29 months after the onset of treatment. The

psychological status of each subject was evaluated concurrently with each

testing by a significant other and the patient's therapist. Ratings were

made using Form R of the Katz Adjustment Scale.

The outpatient study was not designed to compare therapy effective-

ness but, rather, to determine what changes would occur in the Rorschach

under different modes of therapy where the patient selected the type of

therapy. A change in the egocentricity index from either too high or too







-24-


low to the "normal range" was found to occur significantly more often for

those outpatients rated as improved than for those rated as not improved

after at least twenty-nine months of therapy. Those subjects who entered

treatment with an egocentricity index in the average (normal) range were

rated by significant others as having a higher success rate than those

subjects who entered treatment with index scores outside that range.

Although one might conclude from this study that a high or low ego-

centricity index score is associated with psychopathology, it should be

noted that a number of patients who fell into the normal range on this

index at the beginning of treatment were seeking therapy for problems in

living. In addition, because there was no effort made to match subjects

on type or degree of pathology, those subjects who exhibited "normal" ego-

centricity indices may not have had problems or difficulties as chronic or

severe as the other two groups.

High egocentricity indices may be associated with normal adjustment

in certain populations. Exner and Murillo (1973, 1977) found that a sub-

stantial number of their subjects who fell into high and low egocentricity

index groups improved in psychotherapy even though they did not change

significantly on this index. Two groups of chronic inpatient

schizophrenics were studied. One group was treated with a modified

version of electroconvulsive therapy with psychotherapy and the second

group was treated with medication and psychotherapy. Of those subjects

who progressed sufficiently to be discharged from hospitalization, those

treated with ECT showed a significant reduction in the mean egocentricity

index from pretreatment, while the drug-treated subjects who were

discharged still showed a comparatively high index. A comparison of the








-25-


discharged ECT and drug-treated groups on ratings of social behavior one

year later yielded a significant difference favoring the ECT group. This

difference was not apparent three years following discharge. It should

be noted that their mean egocentricity indices did not change significant-

ly during this time. The postdischarge success of the drug-treated group

points to an instance where an excessively high egocentricity index does

not appear to be a handicap.

A study by Exner and Murillo (1975) indicates that this might not be

the case with those subjects discharged with below average egocentricity

indices. They found that 16 of 22 nonschizophrenics who relapsed had ego-

centricity indices of less than average at discharge. Of the fifty-five

subjects from this group who did not relapse, only four had egocentricity

indices below the "normal" range. When viewed in conjunction with the

findings of the other studies relating the egocentricity index to psycho-

pathology, it appears that there are no hard and fast rules for defining

this relationship.


The Shading-Dimensionality Response: (Vista)


Rorschach and Oberholzer (1923) observed that individuals sometimes

used the light-dark features of the inkblot to create a percept of

dimensionality. They postulated that this response represented a cautious

affectivity with depressive nuances. Klopfer et al. (1954) was the first

systematizer to formally score this type of response. He also used this

category (Fk), to designate diffuse shading and reflection responses.

Klopfer et al. (1954) thought that the Fk type response represented a type

of introspective process whereby the individual copes with anxiety by








-26-


distancing himself emotionally from his problems in order to view them more

objectively. Beck et al. (1961) also incorporated a formal scoring

category (Vista), into their system to account for responses that use the

shading characteristics of the inkblot to alter the flat perspective into a

three-dimensional percept. Beck included reflection repsonses in this

category. His scoring criteria did not require that the subject articulate

the shading features of the inkblot. Beck et al. (1961) postulated that

this type of response represented a process of self-appraisal that involved

feelings of depression and inferiority.

Exner (1974) adopted the symbols representing vista type responses (V,

VF, and FV), from Beck et al. (1961) to designate those responses involving

depth or dimensionality based on the light-dark features of the blot.

Exner (1974) scored an answer as a pure vista response (V), when the

subject reported "depth or dimensionality based exclusively on the shading

characteristics of the blot, with no form involvement" (p. 92). A

vista-form response (VF) was scored when the subject's answer included a

"primary emphasis on the shading features to represent depth or dimension-

ality" and incorporated "the form features of the blot for clarification

and/or elaboration" (p. 92). The form-vista response (FV) is scored when

the subject "uses form as the primary feature and the shading component is

used to represent depth or dimensionality for purposes of clarification

and/or elaboration" (p. 92).

Exner (1974) found that the vista response occurs with the lowest

means and frequencies of all the determinants for all adult populations

(i.e., inpatients, outpatients and normals). Exner and Weiner (1982) found

that none of the different vista type responses occur in the records of







-27-


six, seven, and nine year old children but do occur with considerable fre-

quency at the age of twelve. The frequency of these repsonses then de-

clines as age increases.

A number of investigations were conducted to determine the psychologi-

cal significance of this type of response. Exner (1974) evaluated the

Rorschach records of sixteen subjects who made suicide gestures within 55

days after testing, only two of which were successful. These subjects were

from a larger pool of 64 subjects diagnosed as suffering from reactive de-

pression, neurotic depression, involutional reaction, or depressive psycho-

sis. None of the subjects had a history of prior suicide attempts. A com-

parison of the group of 16 patients who attempted suicide to the non-

attempting group revealed that those who attempted suicide had significantly

more vista type responses. Both groups were found to give significantly

more vista responses than a randomly selected group of outpatients and

nonpatients. The results suggested that the vista response may be associa-

ted with depressive features and the painful introspective process that is

common to these patients. This interpretation is not unlike that postulated

by Beck et al. (1961).

The vista response was also investigated in a study reviewed earlier

that examined the relationship between the introspective process and the FD

response (Exner, 1974). Subjects were divided into high and low groups

based on a median split of the number of FD responses and placed in waiting

groups. The high FD group was rated as making more self-focusing verbaliza-

tions than the low FD group. It was also found that the high FD group gave

significantly more vista answers (X = 1.41) than the low FD group (X = 0.52).

Eighteen of the twenty high FD Rorschach records contained vista responses,

whereas these responses appeared in only seven of the 20 low FD protocols.








-28-


The high frequency of occurrence of both the vista response and the FD

response was thought to offer support for the hypothesis that the vista

response is an introspective process. There was no mention of the relation-

ship between the number of vista responses given and the number of self-

focusing statements made. It may be that these subjects are more self-

focused than introspective. It may also be asked whether the high occurrence

of vista with FD responses can be accounted for by variables other than an

introspective process per se. The ensuing study was conducted by Exner to

address this issue.

Exner (1974) compared the Rorschach records of 74 inpatients who were

tested at pretreatment and posttreatment. All of the subjects were diagnosed

as suffering from some form of affective disorder. At least one vista re-

sponse appeared in 62 of the 74 Rorschach records at the onset of treatment

(X = 1.61). At posttreatment these responses were found to occur in only 29

of the 74 protocols (X = 0.59). The number of FD responses did not change

significantly from pre- to posttreatment (2.1 VS. 2.3). An explanation for

the results was offered that suggested that while the painful characteristics

of the introspective process diminished, the tendency to introspect remained

high. Exner is implying that the vista response contains an affective com-

ponent, while the FD response does not. Yet, if the introspective process

remains prominent, then the vista response should also remain high. It may

be that this response represents a certain type of introspection. Exner

(1974) suggests that the vista response reflects a painful introspective pro-

cess rather than a more dispassionate view of oneself. There are a number of

studies which provide some support for this hypothesis.








-29-


The Rorschach protocols of those individuals who participated in the out-

patient study described earlier (Exner, Wylie, and Kline, 1977) were evalua-

ted for their vista responses. Subjects participated in one of seven differ-

ent modes of psychotherapy. Rorschachs were administered to all subjects at

four different time intervals during the course of treatment. An examination

of the Rorschach protocols of those subjects treated by dynamic, gestalt, and

group psychotherapy revealed that at the first retest, the frequency of vista

responses nearly doubled. The frequency of these responses at other intervals

was not reported. It was suggested that the increase in the number of vista

responses given by these subjects was due to the fact that the types of

therapy that they entered taught them to introspect and required that they

focus on "negative qualities." The authors suggested that the results re-

affirmed the hypothesis that the vista response reflects a form of emotional

experience associated with self-examination or introspection although no data

were presented on changes in FD or on other measures of self-examination or

introspection.

Exner and Wylie (1977) reported the use of a constellation of eleven

Rorschach variables, eight or more of which appear in a large percentage of

records collected from effected suicide cases within 60 days prior to the

attempt. They examined the Rorschach records of 59 subjects, ages 18 to 57,

that were administered within 60 days of an effected suicide. "A discriminant

functions analysis was calculated to determine whether any one of the eleven

variables in the constellation should be weighted more heavily than the

others" (p. 206). The variable FV + VF + V + FD > 2 was found to correlate

positively and significantly (0.38) with effected suicide. The authors

concluded that the FD response seems to be related to introspection and that







-30-


the vista response is related to a negative and painful internal experience

evolving from self-inspection; the individual does not like what he sees. It

is unclear what the authors mean by painful or negative internal experience

when they describe the vista response. No measure is used to directly assess

whether the vista response contains an affective component or which affect it

represents. Similarly, no study has directly assessed whether the FD response

is void of an affective component. The relationship between the FD and the

vista response to the process of introspection also needs to be assessed more

directly. Finally, if self-esteem is related to the individual's internal

experiences, then one could also speculate that it might contribute to the

variance of the vista response.


Statement Of The Problem



Exner (1974) has proposed three broad scoring categories to designate

those responses that Beck et al. (1961) catagorize as vista type responses

and Klopfer et al. (1954) score as FK. The form-dimensional category was

introduced to designate those responses which include perspective or

dimensionality based exclusively on form, interpreted by size or in relation

to other blot areas. The reflection type response was proposed to denote

those responses where the symmetry and form features of the blot determine the

answer. The pair response was introduced to score those instances where the

object is reported as two objects because of the card symmetry. A vista

scoring category was also introduced to score those responses that used the

light-dark features of the inkblot to create a percept of dimensionality.

The Rorschach determinants reflect the perceptual-cognitive processes

used to create the response (Exner, 1974). Exner has hypothesized which








-31-


psychological processes each of these categories represent and has conducted

and cited a number of studies that test these hypotheses. The FD response has

been proposed to reflect a nonemotional introspective process. The basis for

this hypothesis resulted from observations of the frequency of these responses

given by various patient populations. The studies that were conducted to

explore this hypothesis did so by relating the frequency of FD responses to

indices of affective delay, the tendency to verbalize self-focused

statements, an egocentricity questionnaire, and the psychotherapeutic process.

A number of these studies did not include a control group. None of the studies

tested directly whether an affective component contributed to the variance of

this response or whether subjects who give a greater than average frequency of

these responses are indeed more aware of their own thoughts and feelings.

The egocentricity index (3r + (2)/R) was hypothesized to represent a

self-focusing process characteristic of egocentricity. The pair and reflec-

tion responses were first studied in relation to patient populations that were

thought, in theory, to be narcissistic. They were also studied in relation to

subjects who attained high and low scores on a pen and pencil test of narcis-

sism (Watson, 1965), modified by Exner (1973). Exner decided that the concept

of narcissism was too complex and the data were subsequently interpreted as re-

flecting a self-focusing process or egocentricity. These determinants were

then studied in relation to time spent viewing oneself in a mirror. Although

the author concluded that reflection and pair responses represent a self-

focusing process, they did not adequately demonstrate that this self-focusing

process is at the expense of a concern for, or awareness of, others which the

concept of egocentricity implies. Similarly, it is not clear how self-focus-

ing is defined. One is not sure whether it entails a focus on one's appear-

ance, one's thoughts and feelings, or one's self in relation to others.







-32-


The egocentricity index (3r + (2)/R) has been investigated as represent-

ing a self-focusing process by observing the frequency of its occurrence in

child and patient populations. This relationship has also been explored by

using Locus of Control and Field Dependence-Independence paradigms, and "PER"

responses on the Rorschach. These studies have met with mixed results. The

egocentricity index has also been investigated in relation to self-esteem.

Because of the highly questionable validity of the self-esteem measure used,

further testing still needs to be conducted in this area. A review of the

studies that have explored the relationship between the egocentricity index

and psychopathology have met with mixed results and suggest that no hard and

fast rules apply to this relationship.

The vista response has been hypothesized to represent a painful

introspective process that is associated with depressive features. The person

does not like what he sees. This hypothesis was inferred on the basis of

observation of the frequency of this response in patient populations,

particularly depressives. The relationship of this response to depressive

features, affect, and the introspective process was tested by studying the

frequency of its occurrence in relation to the FD response either separately

or over the course of therapy. However, a number of variables are prevalent

in the therapeutic process other than the introspective process and painful

affect that were not controlled for and may have contributed to the variance

of both the FD and vista response. No study was directly tested whether

subjects who give a higher than average number of these responses are indeed

more aware of their own thoughts and feelings nor has a definition of the

painful or depressive features that the vista response is hypothesized to

represent been offered. Similarly, no direct test was made to determine








-33-


whether an affective variable contributes to the variance of the vista

response. It is not clear how Exner and Weiner (1982) conclude that the

individual who gives a high frequency of vista responses does not like what he

sees. It could then be asked whether the subject's self-esteem influences the

frequency with which this variable occurs.

It is clear that a number of questions remain unanswered concerning the

construct validity of the FD, vista, and the egocentricity index. The current

study was conducted to answer some of these questions in an effort to further

examine the psychological processes represented by these variables. A

non-patient population of subjects was used in order to control for the

influence that pathology may have. Four questionnaires were used that were

thought to represent the psychological processes hypothesized by Exner to be

reflected by each Rorschach variable. A series of stepwise discriminant

analyses were used to assess the relationship between the probability of a

high or low frequency of each of the three Rorschach variables and scores

obtained on the questionnaires.


Hypotheses


One hypothesis of the study was that those individuals who tend to be

more self-centered than other-centered would show a low frequency of form

dimensional and vista responses and a high frequency of egocentricity

index type responses on the Rorschach. The Self-Focus Sentence Completion

Test (Exner, 1973) was used to assess the subject's tendency to be

self-centered or other-centered (see method section).

A second hypothesis of the study was that those individuals with high

depression, anxiety, and anger subscale scores would show a high frequency








-34-


of vista responses but not either the FD or egocentricity index variables.

The Profile of Mood States (McNair, Lorr, and Droppleman, 1971) was used to

provide an estimate of the subjects' affective state (see method section). It

was decided to use the anger, depression, and anxiety subscales of the profile

on the basis that these affective states are typically labeled as painful and

dysphoric and are, in theory, characteristic of depressives (Freud, 1957;

Rado, 1927/1956; and Fenichel, 1945).

The Self-Consciousness Scale (Fenigstein, Scheier, and Buss, 1975) was

used to assess the individual's tendency to direct attention inward or out-

ward and the process of self-focused attention. Two subscales of this measure

are of particular interest to the current study. The Private Self-Conscious-

ness Subscale is an index of the degree to which an individual is concerned

with attending to his/her inner thoughts and feelings. The Public Self-

Consciousness Subscale measures the extent to which the individual has a

general awareness of himself as a social object that has an effect on others.

It was hypothesized that high scores on both the Public and Private Self-

Consciousness Subscales would be associated with a high frequency of

form-dimensional and vista responses. Both subscales represent a process of

self-focused attention. As such, it was also hypothesized that a high score

on either the private self-consciousness or public self-consciousness subscale

would be predicted to a probability of a high frequency of occurrence of those

responses comprising the egocentricity index.

Exner and Weiner (1982) postulated that the individual who gives a high

frequency of vista responses is introspective and does not like what he sees.

Rosenberg's Self-Esteem Scale (Rosenberg, 1965) was used to assess the








-35-



individual's self-regard. A hypothesis of the current study was that scores

on the Rosenberg's Self-Esteem Scale (RSES) that indicate low self-esteem

would be associated with a high frequency of the vista type responses on the

Rorschach. The literature on the egocentricity index indicates that low

self-esteem may be represented by either a high or low score on this variable.

Therefore, a final hypothesis of the study was that low self-esteem scores

would be associated with a high frequency of the egocentricity index-type

responses.















CHAPTER II
METHOD

Subjects


One hundred and fifty subjects (60 males and 90 females) from the Uni-

versity of Florida participated in the study. Their ages ranged from eighteen

to thirty-six with a mean age of twenty. They were recruited from a subject

pool comprised of students enrolled in introductory experimental psychology

classes that required hour per hour credit for participation in various

studies.1 The subjects were obtained via psychology classes, but were found

to be majoring in a number of areas: 1) computer related (8); 2) engineering

(14); 3) business related (12); and health related (61). Fifty-five were as

yet undecided. To recruit subjects, a sign-up sheet was posted that read as

follows:

A normative study is being conducted to collect data on Uni-
versity of Florida students for a number of psychological
tests. As a participant of this research project, you will be
asked to complete four questionnaires designed to reflect the
thoughts and feelings that college students have. In addition,
you will be administered the Rorschach Inkblot Test which is a
test that is also designed to assess student's thoughts and
feelings in a less direct manner. The total duration of this
procedure will be two hours.

To ensure confidentiality, each subject who participated in the study was

assigned a specific number. This number, rather than their name or other

identifying information, was used to identify test materials and in the anal-

yses conducted. Only the author and his supervisor had access to the sub-

ject's name.


'Due to time constraints and difficulty obtaining subjects from the subject
pool, fifteen female volunteer occupational therapy students also participa-
ted in the study.


-36-








-37-


Assessment Instruments


The Self-Focus Sentence Completion Test

Questionnaire format and scoring

This questionnaire was first developed by Watson (1965) as an operational

criterion of narcissism. Sentence completion responses were scored for three

categories: 1) narcissistic fantasy; 2) real world anchorage; and 3) object

cathexis. Exner (1973) observed that the majority of items used by Dr. Watson

to discriminate high and low narcissistic individuals from a larger popula-

tion contained self-reference statements (i.e., I, me, or my). He subsequent-

ly consulted with Dr. Watson to develop a new sentence completion blank which

he termed the Self-Focused Sentence Completion Test (SFSC), (Exner, 1973).

The present format of the SFSC uses thirty of the original forty-five

sentence stems (see Exner, 1973). The subject is instructed to complete each

sentence so that it expresses a complete thought. Responses are judged on a

"yes-no" basis to meet the scoring criteria of one of four mutually exclusive

categories. The self-focus category (S) designates those responses that

clearly focus on the self with little or no regard for the external world.

The external-world focus category (E) designates those responses which clear-

ly manifest concern for real things or people. The ambivalence response cate-

gory (A) denotes those sentences which clearly contain both self-focused and

external-world focused statements, either of which could be scored separately.

A neutral response category (0) designates those responses which do not meet

any of the above criteria.

Scores are obtained for each of the scales by simply counting the number

of responses judged to be representative of each category. Sample responses

are provided in the manual for scoring. Normative data are provided for male








-38-


(N=622) and female (N=426) college students for each of the scales. In addi-

tion to the scale scores, a "D" score can be obtained by subtracting the total

number of external-world focusing responses from the total number of self-

focusing responses given by a subject. A separate score (SN) can also be cal-

culated to represent the self-focusing statements which are judged to be nega-

tive in content by simply adding the total number of these responses. Norma-

tive data for both the "D" and SN scores are also provided.

For purposes of the current study, the SN score, the self-focus scale,

and the "D" score were used to predict to a probability of frequency of the

Rorschach variables of interest. They were also used as correlates to other

measures.

Reliability studies

High interscorer reliability has been found for Ph.D. clinical psycholo-

gists who were asked to study the scoring criteria and then independently

score twenty records (Exner, 1973). The reliability coefficients obtained

were S=.94, E=.93, A=.92, Sn=.90, Ea=.88.

Validity studies

Exner (1973) reports several studies that have been conducted to investi-

gate the concurrent validity of the SFSC. In one such study, it was found

that a greater than average number of self-focusing responses (S) related

significantly to those psychiatric classifications that are considered to be

excessively self-centered (i.e., homosexuals and psychopaths). A second study

found that an above average number of "S" responses correlated highly with

Rorschach reflection responses. In a third study, trainees who failed to

complete Peace Corps training or who performed inadequately in the service

were found to give an above average number of "S" responses.








-39-


Two additional studies cited by Exner made use of an interview situation

to investigate those individuals who gave a proportionately greater number of

self-focused than exernal-world focused responses (Exner, 1973). These indi-

viduals tended to verbalize more self-focusing statements during their inter-

view. They were also found to spend more time viewing themselves in a mirror

prior to their interview.

Exner (1973) observed that the psychiatric diagnosis of depression was

often associated with those individuals who gave a substantially greater

number of external-world than self-focused type responses. It was also ob-

served that when the self-focus to external-world focus ratio was dispropor-

tionate regardless of the direction, individuals exhibited less effective

and more pathological behaviors. A pre- and posttreatment change from a dis-

proportionate ratio to one which is more balanced has been found for schizo-

phrenics and "acting-out" adolescents who manifest symptom remission and no

apparent relapse (Exner, 1973). These pre- and posttreatment studies suggest

that this form of response style can and does change.

The Self-Consciousness Questionnaire

Questionnaire format and scoring

The Self-Consciousness Questionnaire is a trait measure that was designed

to assess the tendency of a person to direct his or her attention inward or

outward (see Fenigstein, Sheier, and Buss, 1975). This conception of

self-consciousness as a trait is different from the concept of self-awareness

which refers to a state and is the result of either transient situational

variables, chronic dispositions, or both.

A number of procedures were followed in the development and construction

of the Self-Consciousness Scale. First, behaviors that were thought to







-40-


represent the domain of self-consciousness were selected. The following

behavior classification was constructed: "1) preoccupation with past, present,

and future behavior; 2) sensitivity to inner feelings, 3) recognition of one's

positive and negative attributes; 4) introspective behavior, 5) a tendency to

picture or imagine oneself; 6) awareness of one's physical appearance and pre-

sentation; and 7) concern over the appraisal of others" (Fenigstein et al.,

1975, p. 527). Thirty-eight items were selected to sample this domain and the

scale was administered to undergraduate students. A factor analysis yielded

three factors: 1) private self-consciousness, 2) public self-consciousness,

and 3) social anxiety. The first two factors refer to a process of self-

focused attention while the third refers to a reaction to this process. The

scale was then revised to eliminate those items that loaded highly on more

than one factor, those that were endorsed either too frequently or infrequent-

ly, and those that were described by subjects as too ambiguous. A five-point

Likert-type scale was included to rate each item. The revised scale was ad-

ministered to a sample of college students. A factor analysis yielded the

same three factors. All items loaded above 0.40 with their appropriate

factor.

The final version of the scale is composed of three subscales, two of

which appear to be separate aspects of self-consciousness: the social anxiety

subscale, the private self-consciousness subscale, and the public self-

consciousoness subscale. The private self-consciousness subscale items re-

flect the process of attending to one's inner thoughts and feelings. It is

defined by a cognitive, private mulling over about the self. This dimension of

self-consciousness is postulated to be similar to the Jungian concept of

introversion (Jung, 1933), in that the individual is generally oriented toward







-41-


the inner world of ideas and concepts. However, private self-consciousness is

more specific. The focus is on thoughts and reflections that deal solely with

the self. The public self-consciousness subscale is composed of items that

define a general awareness of self as a social object that has an effect on

others. This process is thought to be related to that proposed by Mead (1934)

who argued that consciousness of self comes about when the person becomes

aware of another's perspective after which he can view himself as an object.

The social anxiety subscale is composed of items reflecting discomfort felt in

the presence of others.

The scale is composed of twenty-three items each of which is rated on a

scale of zero (extremely uncharacteristic), to four (extremely characteristic).

This yields low scores for a low level of self-consciousness and high scores

for a high level of self-consciousness. Subscale means and standard devia-

tions are available for male and female undergraduate students. For purposes

of the current study, only the private and public self-consciousness subscale

scores will be used to predict to a probability of frequency of the Rorschach

variables of interest.

Reliability studies

Fenigsteink et al. (1975) reported high test-retest reliability over a

two-week period. The correlation coefficients for the subscales were private

self-consciousness 0.84, public self-consciousness 0.79, social anxiety 0.73,

and 0.80 for the total scale score. Similar findings have also been reported

by Heineman (1979) in a German replication, and Vleeming and Engelse (1981) in

a Dutch replication of the original study conducted by Fenigstein et al. (1975).

Construct validity for the self-consciousness scale has been provided by

Fenigstein et al. (1975), Heineman (1979), and Vleeming and Engelse (1981), in








-42-


their replication of the original study. In each study, a factor analysis

yielded the same three factors and the same correlation pattern between

the subscales. Public self-consciousness correlated moderately with both

private self-consciousness (0.23) and social anxiety (0.21). Both the

German (Heineman, 1979) and Dutch (Vleeming and Engelse, 1981) replica-

tions found that certain items (2, 6, 7, 8, and 22) loaded on factors

other than were intended. It is not sure whether this finding was due to

the translation of the items into a different language.

Discriminant validity for the self-consciousness scale has been pro-

vided by Carver and Glass (1976). These authors correlated the private

and public self-consciousness subscales to variables that were thought to

potentially influence these variables. The subscales were tested in

relation to intelligence, the need for achievement, test anxiety, activity

level, and sociability. The private and public self-consciousness sub-

scales were found to be relatively free from associations with these other

variables.

A number of studies have also been conducted to investigate the

concurrent validity of the self-consciouosness scale. Women who obtained

high scores on the public self-consciousness subscale were found to be

more sensitive to rejection by a peer group, as measured by their

subsequent attraction toward the group and willingness to reaffiliate

with it, than were women who obtained low scores on this subscale

(Fenigstein, 1974). Turner, Carver, Scheier, and Ickes (1978) found

significant correlations between private self-consciousness and the

Guilford Zimmerman Thoughtful-ness Scale and the Pavio Imagery Scale.

They also found that each self-consciousness subscale correlated








-43-


significantly with the self-monitoring scale. Scheier, Fenigstein, and

Buss (1974) divided subjects into high and low groups on the basis of

their private self-consciousness subscale scores. Subjects in the high

group were more aware of their internal affective state and became more

aggressive following provocation by an angry investigator. Buss and

Scheirer (1976) demonstrated that individuals who obtain a high score on

the private self-consciousness subscale make more self-attributions than

individuals who obtain low scores on this scale.

The Rosenberg Self-Esteem Scale

Questionnaire format and scoring

The Rosenberg Self-Esteem Scale (RSES) was developed as a measure

of global self-regard (Rosenberg, 1965). In constructing the scale,

Rosenberg (1965) was guided by the following practical and theoretical

considerations: (1) the ease of administration, 2) economy of time, 3)

unidimensionality (be able to rank people along a continuum), and 4) face

validity. A Guttman scale procedure was used to ensure that the last two

considerations were adequately met. Positive and negative items were

presented alternatively to reduce the effect of a response set (see

Rosenberg, 1965).

The RSES consists of ten items that are rated on a four-point Likert-

type scale. A scoring procedure designed by Rios-Garcia and Cook (1975)

will be employed as it is less complicated and more easily interpretable

than that proposed by Rosenberg (1965). The scale is scored by sub-

tracting the sum of those items reflecting poor self-regard from those

items reflecting self-satisfaction. A constant of twenty-five is then

added to this score. The higher the score, the lower the self-esteem.








-44-


High self-esteem, as reflected in the RSES items, expresses the feeling

that one is "good enough" or feelings of self-acceptance. He does

not consider himself to be superior to others but, rather, simply believes

that he is a person of worth and he respects himself. A low self-esteem

score implies self-rejection, self-dissatisfaction, and self-contempt. He

lacks respect for the self that he observes. Normative data are provided

for undergraduate psychology students (Rios-Garcia and Cook, 1975).

Reliability studies

Rios-Garcia and Cook (1975) obtained a test-retest reliability coeffi-

cient of 0.79 for a sample of college students using their scoring

procedure.

Validity studies

A number of studies have been conducted using selected items of the

RSES either alone or in conjunction with items from other scales:

(Beckman 1970; Kaplan and Pokorny 1969, 1970, 1971; Nocks and Bradley

1969). The self-esteem measure that was used in this study contains the

RSES original items.

The RSES was designed to represent a unidimensional scale. However,

an internal factor analysis performed by Kaplan and Porkorny (1969) pro-

duced two uncorrelated factors which accounted for forty percent of the

total variance. Their factor one scores, which they termed self-deroga-

tion, correlated in a similar fashion to those obtained by Rosenberg

(1965) with the total scale. Their second factor was thought to reflect

"a posture of conventional defense of individual worth, a stance which is

comparable with either high or low scores on the self-derogation factor"

(p. 425). This finding has not been replicated.








-45-


The RSES was originally designed to be used with an adolescent popu-

lation. However, a number of studies have used this measure with adult

populations. Silber and Tippett (1965) provided some evidence for the

convergent validity of the RSES using a college population. The RSES was

correlated against three other measures of self-esteem: the Kelly Reper-

tory Test (r=0.67), the Health Self-Image Questionnaire r=0.83, and inter-

viewers' ratings of self-esteem (r=0.56). These convergent validities were

among the highest observed in cross-instrument correlations (Wylie, 1974).

Some construct validity of the measure has been provided in studies con-

cerning the relationship between RSES scores and other variables which

self-esteem is theoretically expected to be related. Rosenberg (1965)

found that among 50 subjects at the National Institute of Mental Health

who were rated by nurses, those with high RSES scores were significantly

more often rated as gloomy and disappointed. They were also rated as

touchy, easily hurt, and not well liked. Kaplan and Pokorny (1969) found

a significant correlation between their self-derogation factor from the

RSES and Rosenberg's Guttman scale of depressive affect. Most of the

other studies were conducted using adolescent populations. Rios-Garcia

and Cook (1975) used their modified version of the RSES to investigate the

relationship between self-acceptance and psychological adjustment and

several measures of defensiveness. Low self-esteem scores correlated

positively with repressive-sensitization, ego control, and trait anxiety,

and negatively with general defensiveness.

The Profile of Mood States

Questionnaire format and scoring

The Profile of Mood States (POMS) is a factor analytically derived

inventory which attempts to measure six identifiable mood or affective








-46-


states (McNair, Lorr, and Droppleman, 1971). These include tension-

anxiety, depression-dejection, anger-hostility, vigor-activity, fatigue-

inertia, and confusion-bewilderment. The total inventory is composed of

sixty-five adjectives that are rated on a five-point intensity scale (0 =

not at all, 4 extremely). Subjects are asked to endorse each item so

that it describes how they have been feeling during the past week includ-

ing the present day (see McNair, Lorr, and Droppleman, 1971).

The POMS can be completed in about five minutes, and it is easily ad-

ministered, scored, and interpreted. To obtain a raw score for each mood

factor, one simply sums the responses (taking into account negative

weights) for the adjectives defining each subscale. Raw scores can then

be plotted on profile sheets that convert them into T-scores. A profile

sheet based on a normative sample of college students is available.

In this study, only the tension-anxiety, depression-dejection, and

anger-hostility subscale T-scores will be used to predict to a frequency

of the Rorschach variables of interest and as correlates to other

measures. The tension-anxiety subscale is composed of nine adjective

scales descriptive of heightened musculoskeletal tension. The adjectives

refer to reports of somatic tension that may not be overtly observable

(tension, on edge), as well as observable psychomotor manifestations

(shaky, restless). They also refer to vague, diffuse anxiety states

(anxious, uneasy). The depression-dejection subscale is composed of

fifteen adjective scales. The subscale score represents the mood of de-

pression accompanied by a sense of personal inadequacy. The items com-

posing this subscale refer to "feelings of personal worthlessness (un-

worthy), futility regarding the struggle to adjust (hopeless, desperate),








-47-


a sense of emotional isolation from others (blue, lonely, helpless,

miserable), sadness (sad, unhappy), and guilt (guilty, sorry for the

things done)" (p. 7). The anger-hostility subscale is composed of twelve

adjective scales. The subscale score represents the mood of anger and

antipathy towards others. The items comprising this scale describe feel-

ings of intense, overt anger (angry, furious), milder feelings of hos-

tility (grouchy, annoyed), and the more sullen and suspicious components

of hostility (resentful, spiteful).

Reliability studies

All six POM factors have been found to be internally consistent. The

authors cite data from two studies that have shown all the reliabilities

to have been highly satisfactory. All the indices of the extent to which

the individual items within the six mood scales measure the same factor

were near 0.90 or above.

Test-retest reliability has been provided by correlations between the

POMS-scores at intake and pretreatment. The median time period was 20

days with a range of 3-110 days. The reliability estimates for this

period ranged from 0.65 for vigor to 0.74 for depression. The test-retest

correlations between intake score and scores following six weeks of psy-

chiatric treatment ranged from 0.43 for vigor and 0.53 for anger. The

obtained stability coefficients are lower than the 0.80 to 0.90 levels

expected of measures of stable personality characteristics. A mood,

however, is a relatively transient state that would not be expected to

reach the levels required of personality traits.

Validity studies

The authors cite six factor analytic replications that were conducted

in the development of the POMS as evidence for the factorial validity of







-48-


the six mood factors. The results of each were highly similar despite

using different populations normalss vs. patients), rating time periods,

and rating scales. In addition, the individual items defining each mood

scale and loading on a particular factor have, upon examination, face or

content validity.

The POMS is supported by several studies that have provided evidence

for predictive and construct validity. Lorr, McNair, Weinstein, Michaux,

and Raskin (1961) used the vigor, tension-anxiety, depression-dejection,

anger-hostility, and fatigue subscales of the POMS to compare psycho-

therapy alone to four other treatment groups. Outpatients showed improve-

ment on all subscales except vigor. A number of other studies have been

conducted that used the POMS to assess or predict treatment outcome: Lorr,

McNair, and Weinstein (1964), Haskell, Pugtch, and McNair (1969), McNair,

Goldstein, Lorr, Cibelli, and Roth (1965), and McNair, Kahn, Droppelman,

and Fisher (1967, 1968). The POMS has also been used to assess affective

states in studies that assess response to emotion-inducing conditions:

Pillard and Fisher (1967), Pillard, Atkinson and Fisher (1967), and

Pillard and Fisher (1970). Some concurrent validity has also been re-

ported for the inventory. For instance, the tension-anxiety subscale has

been found to correlate highly (.80) with the Taylor manifest anxiety

scale.

The Rorschach Inkblot Test

Questionnaire format and scoring

The Rorschach Inkblot Test consists of ten inkblots and subjects are

required to verbalize what they perceive each blot, or its various compo-

nents, to represent (Rorschach, 1921). The subject's responses are re-

corded verbatim. After the subject has responded to the ten different








-49-


inkblots, he is presented with each one a second time so that the examiner

can inquire about what the subject perceived when he gave his response

(see Exner, 1974).

Exner's Comprehensive Rorschach System was utilized for scoring re-

sponses in the present study (Exner, 1974). The form-dimensional deter-

minant (FD) designates those verbal responses to the inkblots which in-

clude perspective or dimensionality based exclusively on form, inter-

preted by size or in relation to other blot areas (Exner, 1974). The

reflection response is also based on the form features of the inkblot.

Two scoring categories are used to designate this type of response. The

reflection-form response (rF) is scored for those responses in which "the

symmetry features of the blot are primary in determining the answer, and

form is used nonspecifically, or ambiguously, as an object being reflect-

ed" (Exner, 1974, p. 101). Reflection responses scored in this manner

always involve content with nonspecific form requirements (i.e., clouds,

shadows, rocks). The form-reflection response (Fr) is used to record

those responses where "the form of the blot is used to identify specific

content, which is interpreted as reflected because of the symmetry of the

blot" (Exner, 1974, p. 101). The reflection response, whether rF or Fr,

is based on the symmetry of the inkblot. The pair response (2) is based

solely on the form properties of the inkblot and is scored when "the per-

ceived object is reported as two objects because of the card symmetry"

(Exner, 1974, p. 102). Exner (1974) constructed an egocentricity index

comprised of both the pair and reflection type responses (3r + (2)/R). It

was decided to multiply the number of reflection responses by three

because of the relative infrequency with which it occurs in all popula-

tions. The vista determinant is used to record those responses involving


F







-50-


depth or dimensionality based on the light-dark features of the blot. The

pure vista response (V) is scored when the subject's response involved

"depth or dimensionality based exclusively on the shading characteristics

of the inkblot, with no form involvement" (Exner, 1974, p. 92). The vista-

form response (VF) is scored when the response includes a "primary empha-

sis on the shading features to represent depth or dimensionality" and in-

corporates "the form features of the blot for clarification and/or ela-

boration (Exner, 1974, p. 92). The form-vista response (FV) is scored

when the subject "uses form as the primary feature and the shading

component is used to represent depth or dimensionality for purposes of

clarification and/or elaboration" (Exner, 1974, p. 92). Normative data

for adult nonpatients plus four psychiatric groups are provided for the

egocentricity index, the form-dimensional, and the vista type responses

(Exner, 1978).

Reliability studies

Some studies that examined the test-retest reliability of the ego-

centricity index and the vista response are available. In each of these

studies the vista response has been combined with all of the shading

determinants to form one category (SH). Exner, Leura, Armbruster, and

Viglione (1977) conducted a test-retest study using 100 nonpatient adults

to determine the temporal consistency of the structural data over a three

year period. Correlations for SH and the egocentricity index were r=0.66

and r=0.87, respectively. While it would appear that the egocentricity

index was relatively stable over time, the shading scores seemed to follow

a more transient pattern. Similar findings were also found for 25

volunteer nonpatients retested after one week (Exner and Bryant, 1974).








-51-


The correlation coefficient for SH was 0.51 and 0.91 for the egocentri-

city index. A test-retest study using 25 different nonpatient volunteers

over a 60 day period yielded similar results: SH, r=0.59; egocentricity

index, r=0.85 (Exner, Armbruster, and Leura, 1975).

Validity studies

The studies which have investigated the construct validity of the

egocentricity index, form-dimensional, and vista type response have been

described previously in the literature review section of this paper.


Experimental Procedure


Experimenters

The author and ten clinical psychology predoctoral graduate students

acted as experimenters. Each experimenter had received a semester of

formal instruction in the administration, scoring, and interpretation of

the Rorschach using Exner's Comprehensive System. Each experimenter had

also been supervised in the administration, scoring and interpretation of

at least ten Rorschach protocols in an outpatient clinic. All

experimenters were blind to the hypotheses of the study except for the

author. It was originally intended that no exmainer would test more than

fifteen subjects and none less than ten. Due to experimenter attrition

and time constraints placed on the use of the subject pool, the author

tested eighty-one subjects while sixty-nine subjects were tested by the

other ten experimenters.

Setting

All subjects were tested in one of five rooms located in the depart-

ment of clinical psychology at Shands Teaching Hospital. Four of the







-52-


rooms were generally used for the purpose of conducting formal

psychological assessments while the fifth was used for conferences. It

was believed that this setting would provide a sense of authenticity. Each

room contained a table that provided ample room for both the examiner and

subject to sit side by side. Precautions were taken to cover all one-way

mirrors by drawing the curtains in those rooms that contained them. This

was done to avoid the possible reactive effects that mirrors have on

self-awareness (Duval and Wicklund, 1973), and the frequency of reflection

type responses given on the Rorschach (Exner, 1973).

Original Contact

Once a subject had signed up to participate in the study, he/she was

contacted by telephone and given directions to the clinic. He or she was

also informed which examiner would greet them. Each examiner had speci-

fied certain hours of the day that they would be available for testing

subjects. These hours were posted on the sign-up sheet. Subjects were

therefore assigned to the examiner whose time slot they had endorsed.

Assignment to Groups

All subjects were tested individually. They were randomly assigned

to one of two groups based on the order of test presentation. Half of the

subjects completed the Self-Focus Sentence Completion Test, the RSES, the

POMS, and the Self-Consciousness Questionnaire prior to having the

Rorschach administered to them. The other half completed these question-

naires following the administration of the Rorschach. Subjects were

divided into these two groups to test for possible ordering effects and

the reactivity of self-focus on the frequency and type of Rorschach

responses given in the subsequent analyses.







-53-


Experimental Session

Upon their arrival in the testing room each subject was seated and

presented with an informed consent form to review and sign (Appendix F).

Once he/she agreed to participate in the study, he/she was then read the

following statement:

You are currently participating in an experiment designed
to collect normative data on University of Florida stu-
dents for a number of psychological tests. Your responses
will not be examined individually but, instead, will be
pooled with other subject data. It is important that you
respond as accurately and honestly as possible to ensure
that the normative information is, in fact, representative
of University of Florida students. The data will also be
used to further validate certain aspects of the Rorschach
Inkblot Test which will also be administered to you. If
you should have any questions at any time during the pro-
cedure, feel free to ask them.

These instructions were given in accordance with a procedure sug-

gested by Wylie (1974) to reduce social desirability and elicit a genuine

response set from the subject on self-report tests by depersonalizing the

situation.

Prior to meeting the subject, each examiner was given a packet by the

author containing all testing materials. The subject was read the pre-

ceding instructions and then administered the tests in the order that they

were presented in the packet. All tests were self-explanatory and self-

administered except the Rorschach. The Self-Focus Sentence Completion

Test depicted the following instructions:

Listed below you will find thirty incomplete sentences.
Please complete each sentence so that it expresses a com-
plete thought. There are no right or wrong answers for
this test.

The instructions of the Self-Consciousness Questionnaire were as follows:

The following statements reflect thoughts that students
have. Next to each statement is a rating scale that







-54-


ranges from "0" (extremely uncharacteristic), to "4" (ex-
tremely characteristic). Please rate each statement by
checking one answer so that it best describes your
thoughts.

The Rosenberg Self-Esteem Scale presented the following instructions:

The following statements reflect thoughts that students
have. Next to each statement is a rating scale that
ranges from "1" (strongly agree), to "4" (strongly dis-
agree). Please rate each statement by checking each
answer so that it best describes your thoughts.

The Profile Of Mood States depicted the following instructions:

Below is a list of words that describe feelings people
have. Please read each one carefully. Then fill in one
space under the answer to the right which best describes
how you have been feeling during the past week including
today.

The Rorschach Inkblot Test was administered in accordance with the stan-

dard guidelines of Exner's Comprehensive Rorschach System. All examiners

were seated beside the subject. Subjects were read the standard introduc-

tion to the test:

Now we are going to do the inkblot test. Have you ever
heard of it or have you taken it before? If the subject
had not heard about the test, the examiner informed him/
her that it is just a series of inkblots that I'll show
you and I want you to tell me what they look like to you.
(Exner, 1974, p. 30)

The examiner then handed the subject the first Rorschach card and asked

"What might this be?" Each response given by the subject to each card was

recorded verbatim. A prompt for more than one response was given to the

first card only. After the subject had responded to all ten cards, he/she

was required to review all ten cards again so that the examiner could in-

quire about and insure the accurate scoring of the responses given in the

free association. The inquiry was introduced with the following state-

ment:








-55-


Now I want to go back through the cards again. It will not
take very long. I will read what you told me that you saw
and then I want you to help me to see it as you did. In
other words, I want to know where you saw it, and what it
is there that makes it look like that, so that I can see
it just like you did. (Exner, 1974, p. 37)

All questions by the exmainer in the inquiry were those recommended by

Exner (1974): "I'm not sure I see it as you do" (p. 38), "Could you tell

me what makes it look like that to you" (p. 38), and "I'm not sure I know

where you see it" (p. 37).

All questions concerning the Rorschach were answered after the entire

experimental procedure was completed. If they asked what it was being

used for, they were informed:

This is another measure that we are using to find out
about college students' thoughts and feelings. Unlike
other tests, it does this in an indirect way.

After all the questionnaires had been completed and the Rorschach had

been administered, the subject was read a debriefing statement.

There has been no deception in this study. Your answers are
being used to collect normative data and to further validate
certain aspects of the Rorschach Inkblot Test. Thank you
very much for your cooperation and participation.


Interrater Reliability Of The Projective Instruments


Rorschach Inkblot Test


All 150 Rorschach protocols were scored by an expert Rorschach

examiner who was paid four dollars per protocol for his services.2 To

ensure a high consistency of scoring between subjects and to reduce halo



2"Expert" is being defined here as someone who has been instructed and
supervised in the administration, scoring, and interpretation of one
hundred or more Rorschach protocols in accordance with the guidelines
of Exner's Comprehensive System in a clinic or work-related area.








-56-


effects, responses were scored card by card rather than subject by subject

for groups of ten. That is, card I was scored for ten subjects prior to

scoring card II.

Interrater reliability was obtained for the first thirty Rorschach

protocols. The author (an expert Rorschach examiner) served as the

second rater. Reliability data were gathered for every fifth response in

the order scored by the first examiner for three groups of ten subjects.

Each response was scored for 1) location, 2) determinants, 3) content,

and 4) form quality. Form quality was scored on the basis of whether the

response met the criteria of poor form quality (-), or good form quality

(+). The weak and ordinary form quality responses were classified into

these categories. The popularity of a response and the organizational

quality were not scored. Reliability was calculated for an overall

Rorschach response and for each of the aforementioned categories by divid-

ing the number of respective Rorschach responses scored correctly by both

examiners by the total number of responses scored. A percentage score for

interrater agreement was obtained by multiplying this number by one hun-

dred. It was decided that a minimum of eighty percent interscorer agree-

ment would be an acceptable cut-off point to ensure scoring accuracy.

Self-Focus Sentence Completion Test

All 150 Self-Focus Sentence Completion Tests were scored by a clini-

cal psychology graduate student who was trained by the author according to

the test manual for scoring and interpretation (Exner, 1973). He was paid

fifty cents per completion test for his services. Fifteen completion

tests were randomly selected for the purpose of obtaining interscorer re-

liability. The author acted as the second scorer. All responses for a








-57-


test that was selected were scored. Interscorer agreement was calculated

by the same procedure that was employed with the Rorschach.


Hypotheses And Analyses


Exner proposed that the form-dimensional response represents a non-

emotional intropsective process. To test this hypothesis controlling for

other variables, a stepwise discriminant analysis was performed. The form-

dimensional response was dichotomized into a low and high frequency group

based on the mean number of these responses given by the 150 subjects in

the study and acted as the dependent variable. The "D" and self-focus sub-

scale scores of the Self-Focus Sentence Completion Test, the private and

public self-consciousness subscale scores, and the anger, depression, and

anxiety subscale scores acted as the independent variables. It was ex-

pected that 1) the different mood subscale scores would not predict to

frequency of this response, 2) a high score on the self-focus subscale and

high scores on the private and public self-consciousness subscales would be

predictive of a high frequency of this response, and 3) the D score would

be predictive of a low frequency of the FD response.

Exner postulated that the egocentricity index reflects a self-

focusing process. To test this hypothesis and to investigate whether self-

esteem is related to the egocentricity index controlling for other vari-

ables, a stepwise discriminant analysis was conducted. The egocentricity

index was dichotomized by the same procedure employed with the FD response

and acted as the dependent variable. The independent variables were the

"D", SN, and self-focusing subscale scores of the Self-Focus Sentence Com-

pletion Test; the private and public self-consciousness subscale scores;







-58-


the anger, anxiety and depression subscale scores; and the RSES. It was

predicted that a high D score as well as high scores on the self-focusing

and the private and public self-consciousness subscales would be predicted

to a high frequency of the egocentricity index. No relationship between

the egocentricity index and the different mood subscale scores was ex-

pected to be found. It was expected that both a high score on the RSES

(low self-esteem) and SN subscale would predict to both a high and low

frequency of the egocentricity index.

The vista response is proposed to represent a painful introspective

process that is associated with depressive features and a displeasure for

what one sees. To test this hypothesis controlling for other variables, a

stepwise discriminant analysis was conducted. The vista response was

dichotomized into a high and low frequency group in the manner described

for the FD response and acted as dependent variable. The independent

variables were the private and public self-consciousness subscale scores;

the SN, and the self-focusing subscale scores of the Self-Focus Sentence

Completion Test; the anger, anxiety, and depression subscale scores; and

the RSES. It was expected that high scores on all of these variables would

be predicted to a high frequency of the vista type response. A significant

relationship was not expected to be found between the D score of the SFSC

and the vista response.

In addition to the planned analyses noted earlier, one additional

post hoc analysis was also conducted. "Pathological" groups were created

by dividing subjects into high and low groups based on their scores fall-

ing one standard deviation above and below the means provided for each

questionnarie and subscale. The frequency of occurrence of the egocentri-

city index, the form-dimensional, and vita type responses for these groups








-59-




were compared to the frequency of their occurrence in the "normal" popula-

tion. Comparisons were made using Chi-Square analyses.
















CHAPTER III

RESULTS

Evaluation Of Possible Examiner And Test Order Influences


The author and ten predoctoral clinical psychology graduate students

acted as examiners and administered the Rorschach Inkblot Test. All exam-

iners other than the author were blind to the study. This created the

possibility that the author elicited a different number of the Rorschach

variables of interest as compared to the other examiners. To test for the

possibility of examiner bias, t-tests were performed to compare the ten

examiners pooled together to the author with respect to the frequency of

the egocentricity index, form-dimensional, and vista type responses elic-

ited. No significant differences were found: egocentricity index (t=0.78,

p <0.44), form-dimensional (t=0.32, k <0.75), and vista (t=1.21, 2 <0.23).

It was also possible that the order of test presentation influenced

the frequency of the Rorschach responses elicited. The study was designed

so that this source of bias could be tested for. Subjects were assigned

to one of two test order groups defined by whether the Rorschach was

administered first or last in the test battery. To compare the test order

groups with respect to each of the Rorschach variables of interest, t-tests

were performed. No significant differences between the two test order

groups were found: vista (t=0.21, p <0.83), form-dimensional (t=0.71, <

0.48), and egocentricity index (t=0.53, k <0.60).


-60-








-61-


Interrater Reliability

The Rorschach Inkblot Test

Interrater reliability was calculated for an overall Rorschach

response and for each of its different constituents. Percent agreement

between the author and the designated scores was used as the index of re-

liability. The interrater reliability percentage scores obtained were

1) overall response (92%), 2) location (98%), 3) determinants (94%), and

form quality (93%). Interscorer agreement for each analysis met the pre-

determined minimum requirement of eighty percent for ensuring accurate

scoring.

The Self-Focus Sentence Completion Test

Percent agreement between the author and designated scorer was used

as the index of interrater reliability. An interrater agreement of eighty-

seven percent was found and was considered to be adequate for ensuring

accurate scoring of the test.


Evaluation Of The Hypotheses


The first series of analyses indicated that there were no examiner or

test order biases with respect to the frequency of the Rorschach responses

elicited. They also indicated that the scores obtained from the Self-

Focus Sentence Completion Test and the Rorschach were an accurate repre-

sentation of the subjects' responses. In light of these findings, all

subsequent analyses testing the hypotheses were conducted without con-

trolling for examiner or test-order influences.

The Form-Dimensional Response

Exner proposed that the form-dimensional response (FD) represents a

nonemotional introspective process. To test this hypothesis, a discriminant








-62-


analysis was conducted followed by a stepwise discriminant analysis. For

these analyses, the FD response was dichotomized based on the mean number

of these responses given by the 150 subjects who participated in the

study and acted as the dependent variable. The mean was found to be 0.54

and the subjects were categorized according to whether they did or did

not verbalize this type of response (present n=58, absent n92). The

independent variables for both analyses were the "D" score and

self-focusing subscale score of the Self-Focus Sentence Completion Test,

the private and public self-consciousness subscale scores,3. and the

depression, anger, and anxiety subscale scores. Table 1 presents the

means and standard deviations for each of these variables.

Discriminant function analysis

The results of the discriminant function analysis indicated that

there was tremendous overlap between the present and absent form-dimen-

sional categories with respect to the independent variables. The linear

combination of the variables was unable to discriminate between these two

groups (Table 2).

Stepwise discriminant function analysis

The next analysis tested the relationship of each independent vari-

able as it individually related to the FD groups. Testing at the .1

significance level, the assumption of equal covariance matrixes could not

be rejected and Fisher's discriminant analysis was conducted using a



3"Due to a typing error, one item was not included in the private self-
consciousness subscale leaving nine items instead of ten. To correct
for this, a weighted score was added individually to each subscale
score. The weighted score was determined by averaging the nine items
of each subscale. The hypotheses were evaluated using both the
corrected and noncorrected subscale scores. Only the analyses using
the corrected subscale scores are reported as the results obtained
using either subscale score were not significantly different.








-63-


Table 1

Means and Standard Deviations of the Form-Dimensional Groups

Form-Dimensional Present


Variable

D Score of the SFSC

Self-Focus Subscale

Private Self-Consciousness

Public Self-Consciousness

Anger Subscale

Depression Subscale

Anxiety Subscale


N

58

58

58

58

58

58

58


Mean

16.48

9.93

26.67

18.81

47.83

46.84

48.98


Standard Deviation

5.57

3.39

4.55

3.61

8.21

6.77

9.23


Form-Dimensional = Absent


D Score of the SFSC

Self-Focus Subscale

Private Self-Consciousness

Public Self-Consciousness

Anger Subscale

Depression Subscale

Anxiety Subscale


15.43

9.47

25.83

19.74

50.25

48.23

50.42


6.02

3.03

5.34

4.88

10.06

8.99

10.22








-64-


Table 2



Summary of the Form-Dimensional Group Classification
by the Discriminant Function Analysis





Number of Observations and Percents
Classified into the Form-Dimensional Categories


Form-Dimensional Response Present Absent Total

Present 12 46 58
20.69 79.31 100.00

Absent 13 79 92
14.13 85.87 100.00

Total 25 125 150

Percent 16.67 83.33 100.00


Prior Probability 0.39 0.61


Specificity (percent correctly classified as absent) = 0.85

Sensitivity (percent correctly classified as present) = 0.21

Total Correct Classification = 0.61

Percent Correctly Classified By Chance Given The Prior Probability =0.52

False Positive Rate (percent classified as present but were absent) = 0.52

False Negative Rate (percent classified as absent but were present) = 0.37







-65-


stepwise procedure (Kendall and Stuart, 1977). The results indicated

that none of the independent variables were individually related to the

FD groups at the .1 significance level and none were able to individually

discriminate between the two groups (Table 3).

Analysis of variance

As an alternate way of testing for differences between the two

groups, a series of one-way ANOVAs was conducted to compare the form-

dimensional groups with respect to the mean scores of each subscale.

The aforementioned seven independent variables now acted as the dependent

variables and the dichotomized FD response acted as the independent vari-

able. No significant differences were found. F-values for each of the

ANOVAs were 1) the D score of the SFSC (F (1,148)=1.14, p <0.29); 2) the

self-focusing subscale of the SFSC (F (1,148)=0.76, <0.39); 3) the

private self-consciousness subscale (F (1,148)=1.00, p <0.32); 4) the

public self-consciousness subscale (F (1,148)=1.56, p <0.21); 5) the

anger subscale (F (1,148)=2.38, p <0.13); 6) the depression subscale (F

(1,148)=1.01, p <0.32); and 7) the anxiety subscale (F (1,148)=0.76, p <

.38).

Chi-square analysis

A series of Chi-Square analyses was performed to test whether the

frequencies observed were different from that which would be predicted by

chance. Three subject populations were created on the basis of their

scores on each of the aforementioned subscales. High and low population

groups were defined by those subjects who scored either one standard de-

viation above or below the subscale means obtained from the study sample.

Those subjects whose scores fell within one standard deviation of the

mean were considered to represent an intermediate group. The Chi-Square








-66-


Table 3



Stepwise Selection of Variables for the Form-Dimensional Groups


Variable

D score of the SFSC


Self-Focus Subscale


Private Self-Consciousness


Public Self-Consciousness


Anger Subscale


Depression Subscale


Anxiety Subscale


F

1.14


0.76


0.99


1.56


2.37


1.01


0.76


Prob
0.29


0.39


0.32


0.21


0.13


0.32


0.38


Tolerance

1.00


1.00


1.00


1.00


1.00


1.00


1.00








-67-


analyses compared the three groups for each subscale to the dichotomized

form-dimensional groups. The analyses did not yield significant results

(Table 4).

The Vista Response

The vista Rorschach response is proposed to represent a painful in-

trospective process that is associated with depressive features and a

displeasure for what one sees. To test this hypothesis, a discriminant

analysis was conducted followed by a stepwise discriminant analysis. For

these analyses, the vista response was dichotomized in the same manner as

the form-dimensional response and acted as the dependent variable. The

mean was found to be 0.48 and the subjects were categorized accordingly

(present n=51, absent n=99). The independent variables for both analyses

were the private and public self-consciousness subscale scores, the "D"

and "SN" scores of the Self-Focus Sentence Completion Test, the self-

focus subscale score of the SFSC, the Rosenberg Self-Esteem Scale, and

the anxiety, depression, and anger subscale scores. Table 5 presents

the means and standard deviations for each of these variables.

Discriminant function analysis

The results of the discriminant function analysis indicated that

there was a great deal of overlap between the present and absent vista

groups with respect to the independent variables. The linear combination

of the independent variables was unable to discriminate between these two

populations (Table 6).

Stepwise discriminant function analysis

The next analysis tested the relationship of each independent vari-

able as it individually related to the vista categories. Testing at the

.1 significance level, the assumption of equal covariance matrixes could







-68-


Table 4

Summary of the Chi-Square Analyses using the
Form-Dimensional Groups as the Classificant

Form-Dimensional Groups---Self-Focusing Subscale


High I

Present Frequency 12
Column Percent 50.00

Absent Frequency 12
Column Percent 50.00

x2 = 2.00 p <0.37

Form-Dimensional Groups---Self-Focusing Subscale


itermediate

38
38.00

62
62.00


Present Frequency
Column Percent

Absent Frequency
Column Percent

X2 = 2.63 <0.27

Form-Dimensional Groups---Private


High

12
46.15

14
53.85


Intermediate

41
39.81


Low

5
23.81

16
76.19


62
60.19


Self-Consciousness Subscale

High Intermediate


Present Frequency 8 42
Column Percent 38.10 41.18

Absent Frequency 13 60
Column Percent 61.90 58.82

X2 1.20 k <0.55

Form-Dimensional Groups---Public Self-Consciousness Subscale

High Intermediate

Present Frequency 6 46
Column Percent 23.08 42.59

Absent Frequency 20 62
Column Percent 76.92 57.41

X2 = 3.38 p <0.19


Low

8
29.63

19
70.37






Low

6
37.50

10
62.50


Low

8
30.77

18
69.23







-69-


Table 4-continued


Form-Dimensional Groups---Anxiety Subscale

High

Present Frequency 9
Column Percent 37.50

Absent Frequency 15
Column Percent 62.50

X2 = 0.22 p <0.90

Form-Dimensional Groups---Depression Subscale

High

Present Frequency 4
Column Percent 20.00

Absent Frequency 16
Column Percent 80.00

X2 = 3.67 p <0.16

Form-Dimensional Groups---Anger Subscale

High

Present Frequency 5
Column Percent 27.78

Absent Frequency 13
Column Percent 72.22

X2 = 1.56 p <0.46


Intermediate

41
39.81

62
60.19






Intermediate

51
42.15

70
57.85






Intermediate

49
41.18

70
58.82


Low

8
34.78

15
65.22






Low

3
33.33

6
66.67






Low

4
30.77

9
69.23








-70-


Table 5

Means and Standard Deviations of the Vista Groups

Vista = Present


Variable

D score of the SFSC

Self-Focus Subscale

SN score of the SFSC

Rosenberg Self-Esteem Scale

Private Self-Consciousness

Public Self-Consciousness

Anxiety Subscale

Depression Subscale

Anger Subscale


N

51

51

51

51

51

51

51

51

51


Mean

16.02

9.43

4.35

17.20

26.61

19.59

50.10

48.12

48.94


Standard Deviation

5.15

2.71

2.63

4.56

4.85

4.30

8.75

8.01

8.39


Vista = Absent


D score of the SFSC

Self-Focus Subscale

SN score of the SFSC

Rosenberg Self-Esteem Scale

Private Self-Consciousness

Public Self-Consciousness

Anxiety Subscale

Depression Subscale

Anger Subscale


15.75

9.76

3.99

17.17

25.92

19.27

49.75

47.47

49.51


6.21

3.40

2.18

4.37

5.16

4.53

10.39

8.34

9.96







-71-


Table 6

Summary of the Vista Groups Classification
by the Discriminant Function Analysis


Number of Observations and Percents
Classified into the Vista Groups


Vista Present Absent Total

Present 28 23 51
54.90 45.10 100.00

Absent 13 86 99
13.13 86.87 100.00

Total 41 109 150

Percent 27.33 72.67 100.00


Prior Probability 0.34 0.66


Specificity (percent correctly classified as absent) = 0.87

Sensitivity (percent correctly classified as present) = 0.28

Total Correct Classification = 0.73

Percent Correctly Classified By Chance Given The Prior Probability =0.55

False Positive Rate (percent classified as present but were absent) = 0.32

False Negative Rate (percent classified as absent but were present) = 0.21







-72-


not be rejected and Fisher's discriminant analysis was conducted using a

stepwise procedure. The results indicated that none of the independent

variables were individually related to the vista categories at the .1

significance level and none were able to individually discriminate

between the two populations (Table 7).

Analysis of variance

As an alternate method of testing for differences between the two

groups, a series of one-way ANOVAs were conducted to compare the vista

categories with respect to the mean scores of each subscale. The afore-

mentioned nine independent variables now acted as the dependent variables

and the dichotomized vista response acted as the independent variable.

There were no significant differences found between the absent and

present vista groups. The F-values for each of the ANOVAs were the "D"

score of the Self-Focus Sentence Completion Test (F (1,148)=0.07,

p <0.79); the public self-consciousness subscale (F (1,148)=0.17, p <

0.68); the private self-consciousness subscale (F (1,148)=0.62, P <0.43);

the self-focus subscale (F (1,148)=0.35, <0.55); the "SN" score

of the SFSC (F (1,148)=0.81, p <0.37); the Rosenberg Self-Esteem Scale (F

(1,148)=0.02, <0.97); the anxiety subscale (F (1,148)=0.04, p <0.84;

the depression subscale (F (1,148)=-.21, p <0.65); and the anger subscale

(F (1,148)=0.12, p <0.73).

Chi-square analyses

As was done with the form-dimensional response, a series of Chi-

Square analyses were conducted to test whether the frequencies observed

were different from that which would be predicted by chance. Three








-73-


Table 7

Stepwise Selection of Variables for


D score of the SFSC


Self-Focus Subscale


SN score of the SFSC


Rosenberg Self-Esteem Scale


Private Self-Consciousness


Public Self-Consciousness


Anxiety Subscale


Depression Subscale


Anger Subscale


F


0.07


0.35


0.81


0.01


0.62


0.17


0.04


0.21


0.12


the Vista Category


Prob

0.79


0.55


0.37


0.97


0.43


0.68


0.84


0.65


0.73


Tolerance


1.00


1.00


1.00


1.00


1.00


1.00


1.00


1.00


1.00








-74-


subject groups for each of the independent variables were created in the

manner explained earlier and compared to the vista groups. The analyses

did not yield significant results (Table 8).

The Egocentricity Index

Exner postulated that the egocentricity index reflects a self-focus-

ing process. To test this hypothesis and to determine if self-esteem

and/or an affective state is also related to the egocentricity index, a

discriminant analysis was conducted followed by a backward stepwise lo-

gistical analysis. The egocentricity index was dichotomized in the manner

described for the form-dimensional response and acted as the dependent

variable for these analyses. The mean was found to be 0.39 and the sub-

jects were grouped accordingly (present n=73, absent n=77). The inde-

pendent variables for these analyses were the private and public self-

consciousness subscale scores, the self-focus subscale, the "D" and "SN"

scores of the Self-Focus Sentence Completion Test, the Rosenberg Self-

Esteem Scale, and the anxiety, and depression subscale scores.

Table 9 presents the means and standard deviations for each of the in-

dependent variables entered into the analyses.

Discriminant function analysis

The results of the discriminant function analysis indicated that

there was considerable overlap between the present and absent egocentri-

city index populations with respect to the independent variables.

However, the linear combination of these variables was able to discrimi-

nate between these two populations (Table 10).







-75-


Table 8

Summary of the Chi-Square Analyses using
the Vista Groups as the Classificant

Vista---Self-Focus Subscale of the SFSC


Present Frequency
Column Percent

Absent Frequency
Column Percent

X2 = 1.39 p. <0.50

Vista---D score of the SFSC


Present Frequency
Column Percent

Absent Frequency
Column Percent

X2 = 1.24 p <0.54

Vista---Private Self-Consciousness


High

6
25.00

18
75.00


High

7
26.92

19
73.08



Subscale

High


Present Frequency 6
Column Percent 28.57

Absent Frequency 15
Column Percent 71.43

X2 = 0.74 p <0.69

Vista---Public Self-Consciousness Subscale

High

Present Frequency 8
Column Percent 30.77

Absent Frequency 18
Column Percent 69.23

X
2 = 0.24 p <0.89


Intermediate

37
37.00

63
63.00


Intermediate

38
36.89

65
63.11






Intermediate

37
36.27

65
63.73






Intermediate

38
35.19

70
64 81


Low

8
30.77

18
69.23


Low

6
28.57

15
71.43


Low

8
29.63

19
70.37






Low

5
31.25

11
68.75








-76-


Table 8-continued


Vista---SN Score of the SFSC


High

Present Frequency 10
Column Percent 45.45

Absent Frequency 12
Column Percent 54.55

X2 = 1.6 <0.45

Vista---Rosenberq Self-Esteem Scale


Intermediate

37
32.46

77
67.54


Present Frequency
Column Percent

Absent Frequency
Column Percent

X2 = 0.11 p <0.95

Vista---Anxiety Subscale


Present Frequency
Column Percent

Absent Frequency
Column Percent

X2 = 2.21 <0.33

Vista---Depression Subscale


Present Frequency
Column Percent

Absent Frequency
Column Percent

X2 = 2.55 <0.28


High

7
36.84

12
63.16


High

10
41.67

14
58.33


High

8
40.00

12
60.00


Intermediate

33
34.02

64
65.98


Intermediate

36
34.95

67
65.04


Intermediate

38
31.40

83
68.60


Low

4
28.57

10
71.43


Low

11
32.35

23
67.65


Low

5
21.74

18
78.26


Low

5
55.56

4
44.44








-77-


Table 8-continued


Vista--Anger Subscale


Frequency
Column Percent

Frequency
Column Percent

X2 = 2.8 p <0.25


High

5
27.78

13
73.22


Intermediate

44
36.97

75
63.03


Present


Absent


Low

2
15.38

11
84.62








-78-


D sc

Priva

Publ

Ange2

Anxi<

Depr

Self-

SN sc

Rosei


Table 9

Means and Standard Deviations of the Egocentricity Index Groups

Egocentricity Index=Present

Variable N Mean Standard Devid

ore of the SFSC 73 15.41 5.73

ite Self-Consciousness 73 26.10 4.93

ic Self-Consciousness 73 19.40 3.68

r Subscale 73 48.00 8.26

ety Subscale 73 49.75 9.11

session Subscale 73 47.77 8.04

-Focus Subscale 73 9.32 3.52

core of the SFSC 73 4.48 2.26

berg Self-Esteem Scale 73 17.62 4.08


ition


Egoce,

D score of the SFSC

Private Self-Consciousness

Public Self-Consciousness

Anger Subscale

Anxiety Subscale

Depression Subscale

Self-Focus Subscale

SN score of the SFSC

Rosenberg Self-Esteem Scale


itricity Index

77

77

77

77


= Absent

16.25

26.21

19.36

50.57

49.97

47.62

9.96

3.77

16.77


5.98

5.20

5.08

10.32

10.54

8.41

2.80

2.37

4.70








-79-


Table 10

Summary of the Egocentricity Index Group
Classification by the Discriminant Function Analysis


Number of Observations and Percents
Classified into the Egocentricity Index Groups


Egocentricity Index Present Absent Total

Present 57 16 73
78.08 21.92 100.00

Absent 22 55 77
28.57 71.43 100.00

Total 79 71 150

Percent 52.67 47.33 100.00


Prior Probability 0.49 0.51


Specificity (percent correctly classified as absent) = 0.71

Sensitivity (percent correctly classified as present) = 0.78

Total Correct Classification = 0.75

Percent Correctly Classified By Chance Given The Prior Probability = 0.50

False Positive Rate (percent classified as present but were absent) = 0.28

False Negative Rate (percent classified as absent but were present) = 0.23








-80-


Backward stepwise logistical regression analyses

A second analysis was conducted to test the relationship of each

independent variable as it individually related to the egocentricity

index groups. Testing at the .1 significance level, the assumption of

equal covariance matrixes was rejected and a backward stepwise logistical

regression analysis was performed (Table 11). The results indicated that

the private and public self-consciousness subscales, the Rosenberg

Self-Esteem Scale, the anxiety and depression subscales, the D score of

the SFSC, and the self-focusing subscale were not individually related to

the egocentricity index groups at the 0.05 significance level and were

unable to discriminate between the two populations. Significant results

were obtained for the "SN" score of the Self-Focus Sentence Completion

Test and the anger subscale partialing out the effects of all other

independent variables. Individuals in the "present" egocentricity index

group endorsed significantly more negative self-statements as reflected by

their SN scores (X=4.5), than the "absent" egocentricity index group

(X=3.8), (p <0.01). Individuals in the "present" egocentricity index

group endorsed feeling significantly less angry (X=48) than those in the

"absent" egocentricity index group (X=50.6), (p <0.03). Table 12 depicts

the ability of the linear combination of these variables and the D score

of the SFSC to classify subjects into the two egocentricity index groups.

While there was significant separation between the two egocentricity index

populations with respect to these independent variables, there was also

considerable overlap and the separation was not very clear or well

defined.







-81-


Table 11

Backward Stepwise Selection of Variables for
the Egocentricity Index Groups


Variable


Depression Subscale


Public Self-Consciousness Subscale


Rosenberg Self-Esteem Scale


Self-Focus Subscale


Private Self-Consciousness Subscale


Tension Subscale


SN score of the SFSC


Anger Subscale


D score of the SFSC


Chi-Square


0.36


0.16


0.38


0.23


0.34


1.39


6.75


4.75


3.11


P Value


0.55


0.69


0.54


0.63


0.56


0.24


0.01


0.03


0.08








-82-


Table 12


Summary of the Egocentricity Index Group
Classification by the D and SN Scores of the
SFSC and the Anger Subscale

Predicted


Absent



Present


Total


Absent

46



33


Present

27



40


Total

73



73


Sensitivity (percent correctly classified as present) = 0.55

Specificity (percent correctly classified as absent) =0.63

Total Correct Classification = 0.59

False Positive Rate (percent classified as present but were absent) = .40

False Negative Rate (percent classified as absent but were present) = .42


True


------------------------------------------







-83-


Analysis of variance

As an alternate method of testing for differences between the two

groups, a series of one-way ANOVAs was conducted to compare the egocen-

tricity index categories with respect to the mean scores of each sub-

scale. The aforementioned nine independent variables now acted as the

dependent variables and the dichotomized egocentricity index acted as the

independent variable. No significant results were found for the D score

of the Self-Focus Sentence Completion Test (F (1,148)=0.76, <0.38), the

private self-consciousness subscale (F (1,148)=0.02, p <0.89), the public

self-consciousness subscale (F (1,148)=0.01, p <0.96, the anxiety

subscale (F (1,148)=0.02, p <0.89), the depression subscale (F

(1,148)=0.01, k <0.92), the self-focus subscale (F (1,148)=1.56, 2 <.21),

the SN score of the SFSC (F (1,148)=3.54, 2 <.06) and the anger subscale

(F (1,148)=2.83, <0.09).

Chi-square analyses

A series of Chi-Square analyses was conducted to test whether the

frequencies observed were different from that predicted by chance. Three

subject groups for each of the independent variables were created on the

basis of their scores on each of the aforementioned subscales. High and

low population groups were defined by those subjects who scored either

one standard deviation above or below the subscale means obtained from

our sample. Those subjects whose scores fell within one standard devia-

tion of the mean were considered to represent an intermediate group. The

Chi-Square analyses compared the three groups for each subscale to the

egocentricity index categories. The analyses did not yield significant

results for the self-focus subscale, the D score of the Self-Focus Sen-

tence Completion Test, the private self-consciousness subscale, the







-84-


Rosenberg Self-Esteem Scale, the anxiety subscale, the depression sub-

scale, and the anger subscale. A significant relationship was found for

the SN score of the Self-Focus Sentence Completion Test and the public

self-consciousness subscale (Table 13).


Further Exploration of the Egocentricity Index
Form-Dimensional and Vista Categories



Exner's Comprehensive Rorschach System includes separate scoring

criteria for designating egocentricity-index, form-dimensional, and vista

type responses. These responses are subsumed predominantly under the

heading of the Fk response in the Klopfer system (Klopfer et al., 1954),

and the vista type response in the Beck Rorschach System (Beck et al.,

1961). A series of one-way ANOVAs and subsequent analyses using Duncan's

Multiple Range Test with a significance level set at 0.05 was therefore

conducted to further explore the psychological constructs represented by

the different combinations of these responses. Eight groups were created

to represent all of the possible combinations of the dichotomized Rorschach

responses and acted as the independent variables for these analyses. The

dependent variables for these analyses were the private and public

self-conscousness subscales, the "D" and "SN" scores of the Self-Focus

Sentence Completion Test, the Self-Focus Subscale of the SFSC, the

Rosenberg Self-Esteem Scale, and the anger, depression, and anxiety

subscales. Table 14 presents the means for each of these scores and

subscales. No significant relationships were found between the eight

groups and the independent variables. The F-values for each of the ANOVAs

were as follows: 1) the D score of the SFSC (F (7,138)=0.76, p <0.63);








-85-


Table 13




Summary of the Chi-Square Analyses using the
Egocentricity Index Groups as the Classificant



Egocentricity Index---Self-Focus Subscale


Present Frequency
Column Percent

Absent Frequency
Column Percent


High

11
45.83

13
54.17


Intermediate

46
46.00

54
54.00


X2 = 2.09 k <0.35



Egocentricity Index---D score of the SFSC


Present Frequency
Column Percent

Absent Frequency
Column Percent


High

10
38.46

16
61.54


Intermediate

53
51.46

50
48.54


Low

10
47.62

11
52.38


X2 = 1.41 k <0.49



Egocentricity Index---Private Self-Consciousness Subscale


Present Frequency
Column Percent

Absent Frequency
Column Percent


High

10
47.62

11
52.38


Intermediate

49
48.04

53
51.96


X2 = 0.14 p <0.93


Low

16
61.54

10
38.46


Low

14
51.85

13
48.15








-86-


Table 13-continued


Egocentricity Index--Public Self-Consciousness Subscale


High

Present Frequency 8
Column Percent 30.77

Absent Frequency 18
Column Percent 69.23

X2 = 7.33 <0.03

Egocentricity Index---SN score of the SFSC


Intermediate

60
55.56

48
44.44


High


Intermediate


Present Frequency 15 56
Column Percent 68.18 49.12

Absent Frequency 7 58
Column Percent 31.82 50.88

X2 = 9.99 <0.01

Eogocentricity Index--Rosenberg Self-Esteem Scale

High Intermediate

Present Frequency 13 45
Column Percent 68.42 46.39

Absent Frequency 6 52
Column Percent 31.58 53.61


X2 = 3.45 p <0.18

Egocentricity Index--Anxiety Subscale

High

Present Frequency 11
Column Percent 45.83

Absent Frequency 13
Column Percent 54.17

X2 = 0.46 2 <0.79


Intermediate

52
50.49

51
49.51


Low

5
31.25

11
68.75


Low

2
14.29

12
85.71


Low

15
44.12

19
55.88


Low

10
43.48

13
56.52








-87-


Table 13-continued


Egocentricity Index---Depression Subscale


High

Present Frequency 10
Column Percent 50.00

Absent Frequency 10
Column Percent 50.00

X2 = 0.08 2 <0.96

Egocentricity Index---Anger Subscale


Intermediate

59
48.76

62
51.24


Present Frequency
Column Percent

Absent Frequency
Column Percent

X2 = 3.90 k <0.14


High

6
33.33

12
66.67


Intermediate

58
48.74

61
51.26


Low

4
44.44

5
55.56


Low

9
69.23

4
30.77








-88-


Table 14

Means for each of the Egocentricity Index,
Form-Dimensional and Vista Category Groups


Variable Group 1


Private

Self

D

Public

SN

Anxiety

Anger

Depression

RSES


25.64

7.64

15.00

17.27

4.64

49.18

45.82

47.55

17.64


Group 2 Group 3 Group 4 Group 5 Group 6 Group 7 Group 8


26.00

9.25

14.69

19.72

3.53

48.94

51.28

48.28

16.75


24.90

9.45

15.70

19.32

4.45

50.71

49.45

47.87

17.87


27.25

10.45

15.75

19.00

3.80

46.95

46.65

45.60

16.15


26.64

10.00

17.14

19.29

4.42

51.29

50.29

47.43

17.86


27.82

8.55

14.36

22.45

5.64

52.73

48.45

51 64

19.55


26.69

11.00

18.15

19.31

3.77

49.46

48.69

47.54

17.54


26.14

10.36

16.64

19.57

3.79

49.50

51.07

46.93

15.00


Note: Private = Private Self-Consciousness Subscale;
of the SFSC; D = D Score Of The SFSC; Public =
Subscale; SN = SN Score Of The SFSC; Anxiety =
Anger = Anger Subscale Of The POMS; Depression
POMS; RSES = Rosenberg Self-Esteem Scale


Group 1 =

Group 2 =

Group 3 =

Group 4 =

Group 5 =

Group 6 =

Group 7 =

Group 8 =


Vista

Vista

Vista

Vista

Vista

Vista

Vista

Vista


Self = Self-Focus Subscale
Public Self-Consciousness
Anxiety Subscale Of The POMS;
= Depression Subscale Of The


Present, Form-Dimensional = Present, Egocentricity Index = Present

Absent, Form-Dimensional = Absent, Egocentricity Index = Absent

Absent, Form-Dimensional = Absent, Egocentricity Index = Present


Absent,

Absent,

Present,

Present,

Present


Form-Dimensional =

Form-Dimensional =

Form-Dimensional =

Form-Dimensional =

Form-Dimensional =


Present, Egocentricity Index = Present

Present, Egocentricity Index = Absent

SAbsent, Egocentricity Index = Present

SPresent, Egocentricity Index = Absent

SAbsent, Egocentricity Index = Absent







-89-


2) the private self-consciousness subscale (F (7,138)=0.61, 2 <0.75); 3)

the public self-consciousness subscale (F (7,138)=1.20, k <0.31); 4) the

self-focus subscale of the SFSC (F (7,138)=1.64, p <0.13); 5) the SN

score of the SFSC (F (7,138)=1.28, p <0.27)1 6) the RSES (F (7,138)=1.33,

p <0.24); 7) the anxiety subscale (F (7,138)=0.50, k <0.84; 8) the

depression subscale (F (7,138)=0.57, p <0.78); and 9) the anger subscale

(F (7,138)= 0.76, p <0.63).


Correlation Matrix


A Pearson product-moment correlation matrix was computed to investi-

gate the relationships among the different questionnaires and subscales

(Table 15). Some of the correlations were of particular interest to the

study and are reported below.

Self-Esteem

A significant relationship was found between the "SN" score of the

Self-Focus Sentence Completion Test and the Rosenberg Self-Esteem Scale

(r=.31, p <.001). These variables were also found to be significantly

related to the following subscales: 1) the anxiety subscale (r=.31, <

.001), (for SN), (r=.23, <.004), (for RSES); 2) the depression subscale

(r=.37, p <.001), (for SN), (r=.32, p <.001), (for RSES); 3) the anger

subscale (r=.20, p <.02) (for SN).

Self-Focus

A relationship was found for the "D" score of the Self-Focus

Sentence Completion Test and the following variables: 1) the private self-

consciousness subscale (r=.15, <.06); 2) the self-focus subscale of the

















1) D


2. Private

3) Public


4) Self

5) SN

6) RSES


7) Anxiety

8) Depression

9. Anger

10) CPrivate I


Table 15
Correlation Matrix

Pearson Product Moment Correlations Between All Variables


2 3 4 5 6 7 8 9 10


.15 -.01


.75*** .38***


.19* .03


.24** .05


-.08


.20* -.01

-.17* -.02


-.12 -.13


.31*** .31*** .37*** .20*


.23** .32*** .14


.70*** .61***


.62*** .16*

.02


Note: N 150; D = The D Score Of The SFSC; Private = The Private Self-Consciousness
Subscale; Public = The Public Self-Consciousness Subscale; Self = Self-Focus
Subscale Of The SFSC; SN = The SN Score of the SFSC; RSES = The Rosenberg Self-
Esteem Scale; Anxiety = The Anxiety Subscale Of The POMS; Depression = The De-
pression Subscale Of The POMS; Anger = The Anger Subscale Of The POMS; CPrivate =
The Corrected Private Self-Consciousness Subscale.


*p
*p
***p


.05
.01
.001