Alcoholism, selves and others

MISSING IMAGE

Material Information

Title:
Alcoholism, selves and others
Physical Description:
viii, 106 leaves : ill. ; 28 cm.
Language:
English
Creator:
Ward, David Andrew, 1943-
Publication Date:

Subjects

Subjects / Keywords:
Self-perception   ( lcsh )
Alcoholism -- Psychological aspects   ( lcsh )
Social perception   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Includes bibliographical references (leaves 102-105).
Statement of Responsibility:
by David Andrew Ward.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000207453
notis - AAX4250
oclc - 04064175
System ID:
AA00003532:00001

Full Text











ALCOHOLISM, SELVES AND OTHERS


By

DAVID ANDREW WARD

















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




UNIVERSITY OF FLORIDA


1975




























Copyright by
David Andrew Ward
1975














ACKNOWLEDGMENTS


Many people, either directly or indirectly, have been

helpful in the writing of this dissertation. I will not be

able to adequately thank them all, but for all of the assis-

tance I have received I am deeply grateful.

I especially want to thank Dr. Charles Frazier for his

abiding interest and concern from beginning to end. His

genuine scholarship and great tolerance during the number of

revisions are incalculable and for which I am deeply thankful.

A debt of gratitude is owed to Dr. Benjamin Gorman for his

keen insight and general assistance along the way. To

Dr. Robert Ziller, I owe thanks for the tremendous influence

he has had on my thinking. I have been fortunate to know

him as a teacher and as a friend.

Also, I want to express my deep appreciation for the

personal support and intellectual stimulation that Dr. Grace

Henderson has given me. I will always be greatful for her

willingness to share with me her ideas which have been very

helpful in my own development. Finally, I want to thank

Dr. Felix Berardo for tolerating me throughout my entire

graduate training. His sound scholarship has influenced me

greatly. For all his personal advice and help, again, I am

thankful.


iii









The use of the data for this dissertation was made

possible by Dr. George Warheit. I want to thank him for his

general assistance in making the data available.

In addition to the people mentioned above, there are

many others who have helped me in various ways, including

the hundreds of students that I have taught while serving

as teaching assistant. I will not attempt to mention names

but to simply express my sincere gratitude to all of them.

To my family I owe the most. To my brother, Bud, who

has been with me through it all, I express my love. My

Dad, who is gone now, gave me the spirit to fight amidst the

many storms of life. I will forever remember his encourag-

ing words. To Lynn, who is my partner and love, I thank

you with all my heart. Finally, anything that I have been

able to accomplish, I owe to my Mom. It has been her love

and sacrifice that has made the writing of this disserta-

tion possible. Mom, you know how much I love you.














TABLE OF CONTENTS


Page

ACKNOWLEDGMENTS . . iii

ABSTRACT . ... .iv

CHAPTER

I. INTRODUCTION . . 1

II. THEORY AND METHOD . 7

III. FINDINGS . . ... .28

IV. A DIFFICULTY IN THE LABELING APPROACH .. 59

V. SUMMARY AND CONCLUSIONS .. 75

APPENDICES . . ... .. .88

I. SELECTED SOCIODEMOGRAPHIC CHARACTERISTICS OF
THE STUDIED POPULATION COMPARED TO BREVARD
COUNTY MENTAL HEALTH CENTER ALCOHOLISM POPU-
LATION BY PERCENTAGE . .. 89

II. CROSS-TABULATION TABLES OF VARIABLES COMPRIS-
ING THE ZERO-ORDER HYPOTHESES AND THEIR
DEGREE OF ASSOCIATION AS MEASURED BY GAMMA 90

III. A COMPARISON OF ORDINAL LEVEL MEASURES .. 101

REFERENCES . . .. .. 102

BIOGRAPHICAL SKETCH . ... 105









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

ALCOHOLISM, SELVES AND OTHERS

By

David Andrew Ward

August, 1975

Chairman: Felix M. Berardo
Major Department: Sociology

The primary concern of the study was to test a formal-

ized theory of perceived self-other concepts as they re-

late to alcoholic self-views and drinking patterns.

It was noted that in social-psychological studies the

dominant approach in accounting for the interrelationships

among others' appraisals and self-views has been to con-

ceptualize the self-concept as a unidimensional phenomenon.

In this study, a multiple-component approach to the self

was employed in an effort to account for the impact of the

perceived alcoholic evaluations of others on a subject's

alcoholic self-views and drinking patterns.

A formalization of dimensions of the self-concept and

perceived labeling constructs, in accordance with the Gibbs'

scheme of theory construction, produced eleven testable

hypotheses. The data bearing upon the zero-order hypotheses

supported the theory in every case.

In addition to testing the hypotheses derived from the

theory, an effort was made to empirically ascertain the









relative predictive ability of each concept of the theory

when drinking patterns was regarded as the dependent

variable. The joint use of Automatic Interaction Detector

(AID) and Multiple Classification Analysis (MCA) rendered
2
a R of 43 per cent. Alcoholic self-concept was found to

be the strongest predictor in the analysis. Also, the two-

construct pattern of self-esteem and self-complexity showed

to be of high importance in reducing the unexplained variance

in the dependent variable.

A search of the literature showed that much of the

theory and research dealing with the interrelationships among

perceptions of self and others stemmed from labeling theory

as an approach to deviance. Though this approach has been

able to account for deviant self-development and behavioral

patterns, it has been less than adequate in its ability to

explain differential responses to like labeling conditions.

This problem introduced an additional concern for the present

study. It was proposed that a conceptual framework of

selected self-other perceptions deriving from the formal

theory could provide a viable means for addressing this

problem when analyzed within a multivariate design. Sup-

port for the framework came from data showing that the

influence of the perceived alcoholic evaluations of signifi-

cant others on a subject's alcoholic self-concept was con-

tingent upon relative degrees of esteem and complexity.


vii














CHAPTER I

INTRODUCTION


This is a study in alcoholism. Its primary concern is

to empirically test a formalized theory of perceived self-

other concepts as these concepts relate to alcoholic self-

views and drinking patterns.

The symbolic interactionist tradition has stimulated

theory and research which suggest that the perceived labels

of others have consequences for self-conceptions. This has

been referred to as the "reflected appraisal" notion and

has been formalized by Kinch (1963). Though Kinch's theory

produces propositions associating others' perceived be-

haviors with one's self-concept, his formalization does not

produce propositions associating the perceived appraisals

of others with different dimensions of the self-concept;

for example, self-esteem and self-complexity.

The omission of various dimensions of the self-concept

when theorizing about self-other relations reflects a lack

of adequate conceptualization and measurement on the part

of students of the self. The vast amount of theory and re-

search on the self-concept shows that a unitary conceptuali-

zation of the construct has been the dominate approach. In

contrast to this unitary view, the present study posits that









a multifaceted conceptualization of the construct must be

adopted if it is to remain theoretically and empirically

viable. As the concept applies to this study, it is pro-

posed that a multidimensional as opposed to a unitary view

can most adequately account for the interrelationships among

perceived alcoholic evaluations of others and a subject's

alcoholic self-views and behavioral patterns.

Much of the literature dealing with the interrelation-

ships among the perceived appraisals of others and self-

conceptions stems from labeling theory as an approach to

deviance. A review of that literature indicates that tra-

ditionally the major emphasis has been to stipulate the

situational conditions under which labeling is likely to be

successful. Though this approach has been able to account

for deviant self-development and behavioral patterns, it

has been less than adequate in its ability to explain dif-

ferential responses to like labeling conditions. This

problem introduces an additional concern for the present

study. It is proposed that a conceptual framework of

selected self-other perceptions deriving from the formal

theory may be used to address this problem when analyzed

within a multivariate design.

The mode of theory construction employed in this study

is that which is set forth by Gibbs (1972). The Gibbsian

scheme of theory building is rather straightforward. The

constructs of the theory to be tested are defined. Each









construct is rendered empirically applicable by asserting

a relation between the construct and its referential for-

mula. The referentials are interrelated in accordance with

the "sign rule" of derivation. These logically interrelated

referentials render testable hypotheses. The scores

yielded by the application of referential formulas to data

is the point at which empirical applicability of the con-

struct is achieved. In accordance with some stipulated

testing procedure the score values rendered by the referen-

tial formulas are used to test the hypotheses.

In addition to testing the zero-order hypotheses de-

rived from the formal theory, multivariate analyses are

employed to ascertain which variables of the theory are

most important in predicting drinking patterns. The

analytical procedure for addressing this prediction problem

will be to designate drinking patterns (one of the con-

cepts of the formal theory) as the dependent variable and

all other variables as predictors. The partial beta co-

efficients rendered by the joint use of two multivariate

statistical procedures will indicate the relative importance

of each predictor in explaining variation in drinking

patterns. The final R2 will reflect the total amount of

variance explained by all predictors taken simultaneously.

The data for this study are the self-reports of sub-

jects who are considered alcoholics and are undergoing

treatment in a rehabilitation center. A more detailed









discussion of the data collection process is presented in

Chapter II. However, a few words concerning the applica-

bility of the data for testing the formal theory are in

order at this point. The thrust of the labeling perspective

as it applies to the self-conceptions and behavioral pat-

terns of the labeled person is that others' appraisals in-

fluence one's self-conceptions. This "reflected appraisal"

notion asserts a relation between the actual behavior of

others and one's consequent self-definitions. Since the

data for this study are from the respondents' self-reports,

the objective appraisals of others can only be inferred from

the subject's perceptions. Consequently, the variables re-

lating to others must necessarily be couched in perceptual

terms. Therefore, labeling constructs, for the purpose of

this study, are in the realm of symbolic interaction and

are measured by individual perceptions of "others'" reac-

tions.


General Outline of the Study

In Chapter II, the formal theory is set forth. The

theory is divided into two parts: the intrinsic and the

extrinsic. The first part is concerned with the proposi-

tions, transformational statements, and hypotheses to be

tested. The second (extrinsic) part specifies definitions

of constructs, referential formulas, and a stipulation of

the social units to which the theory is to be applied.

Also, in the extrinsic part, literature highlighting the









faceted nature of the self-concept in addition to literature

bearing upon selected social-psychological constructs em-

ployed in the formal theory is reviewed. Also, in Chapter

II, the problem of predicting drinking patterns from all

other variables of the theory is stated. Finally, the data

and method of analysis for testing the formal theory and

addressing the prediction problem are discussed.

Chapter III presents the findings relating to the

zero-order hypotheses derived from the theory. Also,

analyses that bear upon the prediction problem are included.

Drinking patterns is designated as the dependent variable

and two multivariate statistical procedures are employed to

ascertain the relative predictive ability of all other

variables when taken simultaneously. All of the data bearing

upon the testing of the formal theory of this study, in

addition to data regarding the problem of prediction are

presented in Chapter III.

The final section of Chapter III is concerned with

empirically addressing a potentially weak link in the formal

theory. The possibility of tautology deriving from the use

of self-report data when testing assertions relating per-

ceptions of self to perceptions of others' appraisals is

dealt with, and clues for a solution are explored.

In Chapter IV, the theoretical and research literature

surrounding a shortcoming in labeling theory is reviewed.

A conceptual framework deriving from the formal theory of






6


this study is employed in an attempt to remedy this short-

coming.

Finally, in Chapter V, a summary of the study and some

concluding remarks are presented. Also, suggestions for

future research are offered.














CHAPTER II

THEORY AND METHOD


In this chapter, the formal theory is presented. The

propositions and hypotheses are outlined in the intrinsic

part of the theory (Gibbs, 1972), and in the extrinsic part

of the theory, definitions of concepts, transformational

statements, and referential formulas are set forth. Also

in the extrinsic part, literature that highlights the di-

mensionality of the self is reviewed, and theory and re-

search surrounding selected concepts employed in the formal

theory are presented. Additionally, the problem of pre-

dicting differential drinking patterns from all other vari-

ables of the theory is stated. Finally, the data and

method of analysis for testing the formal theory and ad-

dressing the prediction problem are discussed.


The Formalized Theory: Intrinsic Part

The structure of the formalized theory is shown in the

pictograph on page 10. The intrinsic statements are inter-

related syntactically in accordance with the "sign rule" of

derivation (Gibbs, 1972:190-196).









PROPOSITIONS

P : Among individuals, the greater the "Perceived Alco-
holic Evaluations of Significant Others," the
greater the "Alcoholic Self-Concept."

P2: Among individuals, the greater the "Alcoholic Self-
Concept," the greater the "Drinking Patterns."

P3: Among individuals, the greater the "Perceived Alco-
holic Labeling Frequency," the greater the
"Alcoholic Self-Concept."

P4: Among individuals, the greater the "Perceived Alco-
holic Evaluations of Significant Others," the
greater the "Alcoholic Labeling Frequency."

P : Among individuals, the greater the "Perceived Alco-
holic Evaluations of Significant Others," the
greater the "Drinking Patterns."

P6: Among individuals, the greater the "Perceived Alco-
holic Labeling Frequency," the greater the "Drinking
Patterns."

P7: Among individuals, the greater the "Self-Esteem,"
the less the "Alcoholic Self-Concept."

P8: Among individuals, the greater the "Self-Complexity,"
the greater the "Alcoholic Self-Concept."

P9: Among individuals, the greater the "Self-Esteem,"
the less the "Self-Complexity."

P10: Among individuals, the greater the "Self-Esteem,"
the less the "Drinking Patterns."

P11: Among individuals, the greater the "Self-
Complexity," the greater the "Drinking Patterns."


TRANSFORMATIONAL STATEMENTS

T1: Among individuals, the greater the "Perceived Alco-
holic Evaluations of Significant Others," the
greater the SIGOTHER.

T2: Among individuals, the greater the "Perceived Alco-
holic Labeling Frequency," the greater the FQLABEL.

T3: Among individuals, the greater the "Alcoholic Self-
Concept," the greater the ALCOSELF.










T4: Among individuals, the greater the "Drinking Pat-
terns," the greater the PATTERNS.

T5: Among individuals, the greater the "Self-Esteem,"
the greater the ESTEEM.

T6: Among individuals, the greater the "Self-Complexity,"
the greater the COMPLEX.


THEOREMS

Thl: Among individuals, the greater the SIGOTHER, the
greater the FQLABEL (from T1, CI, P1, C2' T2).

Th2: Among individuals the greater the FQLABEL, the
greater the ALCOSELF (from T2, C2, P2' C3, T3).

Th3: Among individuals, the greater the ALCOSELF, the
greater the PATTERNS (from T3, C3, Pg, C4, T4)

Th4: Among individuals, the greater the ESTEEM, the less
the PATTERNS (from T5, C5, P10, C4' T4).

Th5: Among individuals, the greater the ESTEEM, the less
the COMPLEX (from T5, C5, P6, C6, T6).


greater

greater

greater

greater

greater

greater

greater

greater

greater

greater

greater


the

the

the

the

the

the

the

the

the

the

the


HYPOTHESES

SIGOTHER, the greater the ALCOSELF.

ALCOSELF, the greater the PATTERNS.

FQLABEL, the greater the ALCOSELF.

SIGOTHER, the greater the FQLABEL.

SIGOTHER, the greater the PATTERNS.

FQLABEL, the greater the PATTERNS.

ESTEEM, the less the ALCOSELF.

COMPLEX, the greater the ALCOSELF.

ESTEEM, the less the COMPLEX.

ESTEEM, the less the PATTERNS.

COMPLEX, the greater the PATTERNS.


I:

II:

III:

IV:

V:

VI:

VII:

VIII:

IX:

X:

XI:


The

The

The

The

The

The

The

The

The

The

The







10











+w



IL





+$

uu

























94J


-t








00






4-1
U- -Z'
0 I r r Z, 0
>
0 Cd-i;





o c
C6









The Formalized Version: Extrinsic Part

Unit Term: An individual is any human agent possessing lin-
guistic skills. This is consistent with the symbolic inter-
actionist view that self-conception is possible only through
communication in an interactional context.

Substantive Terms: Concepts. Six of the terms in the in-
trinsic statements are concepts, each defined as follows:

Alcoholic Self-Concept: The extent to which a person
views himself as an alcoholic and perceives his
drinking as excessive and as wrong in comparison with
others.

Self-Esteem: The extent to which a person feels a
sense of self-worth.

Perceived Alcoholic Labeling Frequency: The perceived
frequency with which a person has been either formally
or informally sanctioned for his drinking alcohol.

Perceived Alcoholic Evaluations of Significant Others:
The extent to which others, whose opinions are impor-
tant to the subject, are perceived by the subject as
viewing him as an alcoholic and his drinking as ex-
cessive and wrong in comparison with others.

Self-Complexity: The number of unlike self-cognitions
or the number of self-cognitions which are different
ends of the same dimension.

Drinking Patterns: The frequency or quantity of a
person's drinking or the extent of one's preoccupation
with alcohol.


Substantive Terms: Referentials

All of the following acronyms designate calculational pro-
cedures which require self-report data only.

SIGOTHER designates the following calculational pro-
cedure: The subjects are read the following statement and
then asked to respond to a series of questions.

"Here are a few statements that a person's family
and friends might make about him. Which of these
do you think your family and friends would make
about you?"









1. They think I drink too much.

2. They think I am an alcoholic.

3. They think I am a problem drinker.

4. They think I am a recovered alcoholic.

In addition to these four items, the following was in-
cluded:

5. Do other people think your drinking is out of
line?

To each of these items, the subject is asked
to respond "yes" or "no." The sum of the sub-
ject's responses comprises the SIGOTHER score,
with Item 4 reversed.

FQLABEL designates the following calculational pro-
cedure: The subjects are asked to respond to the following:

1. Has your spouse ever left you or threatened to
leave you if you did not cut down or quit drinking?

2. Has your spouse or other family member ever com-
plained that you spend too much money for alco-
holic beverages?

3. Have you ever been arrested or picked up by the
police for intoxication or other charges in-
volving alcoholic beverages?

4. Has a physician ever told you that drinking was
injuring your health?

5. Have you ever been to a hospital, clinic, center,
or gone anywhere for help with a problem related
to drinking?

To each of these items, the subject is asked
to respond "yes" or "no." The sum of the sub-
ject's responses comprises the FQLABEL score.

ALCOSELF designates the following calculational pro-
cedure: The subjects are read the following statement and
then asked to respond to a series of questions.

"What kind of person would you say you are (when
you are sober)?"









1. Would you say you are a recovered alcoholic?

2. Do you think you drink too much?

3. Would you say you are a problem drinker?

4. Do you think you are an alcoholic?

5. Do you think there is anything wrong with the way
you drink?

To each of these items, the subject is asked
to respond "yes" or "no." The sum of the sub-
ject's responses constitutes the ALCOSELF score,
Item 1 reversed.

PATTERNS designates the following calculational pro-
cedure: The subjects are read the following statement and
then asked to respond to a series of questions.

"Which of these statements describes your usual
drinking pattern?"

1. I worry about not being able to get a drink when
I need one.

2. I try to stay high or keep a "glow" on during
most of the day.

3. I awaken the next day not being able to remember
some of the things I had done while I was
drinking.

4. I stay intoxicated for several days at a time.

5. I neglect my regular meals when I am drinking.

6. I drink mainly for the effect of the alcohol--
with little attention to type of beverages or
brand names.

7. I take a drink the first thing when I get up in
the morning.

8. Once I start drinking it is difficult for me to
stop before I become completely intoxicated.

To each of these items, the subject is asked
to respond "never," "sometimes," "frequently."
The sum of the items constitutes the PATTERNS
score.









ESTEEM designates the following calculational proce-
dure: The subjects are read the following statement and
then asked to respond to a series of questions.

"What kind of person would you say you are (when
you are sober)?"

1. Are you generally the kind of person you would
like to be?

2. Do you ever think that the best solution to your
problems would be to be dead?

3. Are you generally satisfied with yourself?

4. Do you have much confidence in yourself?

To each of these items, the subject is asked
to respond "yes" or "no." The sum of the sub-
ject's responses comprises the ESTEEM score, with
Item 2 reversed.

COMPLEX designates the following calculational pro-
cedure: The subjects are asked to respond to the following
items.

1. Do you feel in good spirits?

2. Do you sometimes wonder if anything is worthwhile
anymore?

3. For the most part, do you feel healthy enough to
carry out the things that you would like to do?

4. Have you ever failed to do some of the things you
should do?

5. Do you ever feel weak all over?

6. Are you shy around other people?

7. Do you feel at ease when you are with other
people?

To each of these items, the subject is asked
to respond "yes" or "no." The sum of the re-
sponses constitutes the COMPLEX score.









Self Theory: The Emergence of the
Multiple Component Approach

In this section, the literature supporting a multi-

dimensional approach to the self-concept is reviewed. The

position taken in this study is that if the self-concept is

to be employed successfully in explaining interrelationships

among perceptions of others and perceptions of self, a multi-

faceted approach must be adopted. The tendency to avoid a

multidimensional approach has been noted by Ziller:

Although early theorists referred to a variety of
components of the self concept, such as complexity
of the self concept, self-esteem, and self-
centrality, most investigators have focused on the
concept of self-esteem. This concept has presented
a sufficient number of difficulties to the investi-
gators, however, that extentions into new regions
may have been discouraged. (1973:xv)

In view of what has been said regarding the assumed efficacy

of a multiple component approach in accounting for the in-

terrelationships among perceptions of self and others, a

review of the literature on the self as it highlights the

faceted nature of the concept is a necessary prelude to the

presentation of the formal theory to be tested in this

study.

A search of the self-theory literature reveals that the

self has often been equated with self-esteem (Wylie, 1961).

This seems to have been the outcome of a combination of in-

adequate conceptualization and operationalization. Recent

theorizing in the area of the self has led to the multiple

component view. In contrast to the global, unitary notion










of self, this approach points to the multidimensional nature

of the concept. Dimensions such as self-esteem, self-

consistency, self-complexity, etc., are considered crucial

for adequate conceptualization.

The viability of a multiple component approach to the

self has been demonstrated in a variety of substantive

areas. Brookover, Erickson and Joiner (1967) have shown

that one dimension of the global self-concept, the "Self-

Concept-of-Academic Ability," was predictive of academic

achievement even when I.Q. was controlled. Their reasoning

for the development of the Self-Concept-of-Academic-Ability

(S.C.A.A.) is depicted in the following:

Perhaps the best description of a large part of the
self-concept literature is that it is verbally re-
dundant or synonymous but nonreplicative. Literally
hundreds of studies have been done on [the] self-
concept and reported in the educational socio-
logical and psychological literature. Yet few of
these studies can be replicated because of either
poor methodology or unclear conceptualization or
usually both if one controlled for the aca-
demic dimension of self-concepts, the association
between general measures of self-concept and G.P.A.
will drop to zero. (1967:24)

Ziller and Golding's (1969:287-300) theoretical develop-

ment of "The Alienation Syndrome" was one of the most sophis-

ticated attempts to integrate various dimensions of the

self. Taking Seeman's (1959) lead, Ziller presented a theore-

tical systhesis of three components of alienation. His phenomi-

nological orientation led him to explore the more general

alienation syndrome through the analysis of three dimensions

of the self. Specifically, he hypothesized that the









combination of self-esteem, self-centrality, and social

interest captured the general concept of alienation:

The inability to reconcile opposing guidance systems
is assumed to lead to reduced social interest or
"disengagement." Since self-esteem is socially de-
rived and reinforced, reduced social interest also
is associated, in turn, with reduced self-estimate,
which, in turn, is associated with increased self-
centrality and reduced social interest. The pro-
cess is cyclical and degenerative. Increased
self-centrality tends to lead to further reduction
in social interest and so on, as a function of the
situation. The cycling and interaction effects of
low self-esteem, low social interest, and high self-
centrality is assumed, here, to be the essence of
the concept of alienation. (1973:65)

Therefore, "Seeman's (1959) components of self-estrangement

and normlessness may be associated with self-esteem and

social interest, and isolation may be associated with self-

centrality" (Ziller, 1973:65). Empirical support for

Ziller's theoretical synthesis has come from studies as

varied as those of behavior problem children, neuropsychi-

atric patients, the aging, and American Blacks.

Moorse and Gergen (1967) were able to show that sub-

jects tend to compare themselves with others according to

their degrees of self-consistency which is one dimension of

the self. More specifically, in an attempt to test

Festinger's "Social Comparison" theory, they demonstrated

that others' characteristics influenced a subject's self-

esteem and that this influence was contingent on self-

consistency. The implication is that if self-consistency

were held constant, others' characteristics would have a

uniform impact on esteem. This research shows that both









self-esteem and self-consistency are critical self-

dimensions that must be taken into account in explaining

behavior.

Ziller and Golding's (1969) now classic study of the

"Political Personality Syndrome" showed that self-esteem

and self-complexity are crucial self-dimensions in pre-

dicting a candidate's success or failure. It was the di-

mension of self-esteem in combination with the dimension of

self-complexity that enabled the researchers to adequately

account for a candidate's political success.

As the previous studies indicate, the multidimensional

approach to the self is both theoretically and empirically

more tenable than a global view. It is suggested that

propositions associating perceptions of others with self

perceptions can most adequately be accounted for with a

multiple component approach.


Theory and Research Surrounding Constructs
Used in the Formal Theory

This section provides a brief overview of the self-

concept literature as it bears upon the conceptualization

of the present study. It will serve as the theoretical and

empirical basis for the usage of selected constructs em-

ployed in the formal theory of this study.

In Ziller and Golding's previously cited work on the

political personality, the somewhat nonobvious finding that

persons low in self-esteem had a higher probability of









winning elections is instructive. The notion is that those

who are low in self-esteem tend to monitor their social en-

vironments more and are therefore attuned to their con-

stituency. Conversely, those who are high in self-esteem

tend to be field-independent and are less influenced by

others' opinions. Also, self-complexity is associated with

relative field-dependence which, in turn, is predictive of

success. Regarding the concept of self-complexity, Ziller

wrote:

Thus, it is hypothesized that the complex indi-
vidual is more inclined toward assimilation of self
and others or perceiving similarities between self
and others, whereas the simplex individual is in-
clined toward contrasting self and others. In
general then, it is proposed that persons with more
complex self-concepts attend to a broader range of
social stimuli, perceive more similarities between
self and others, and are more responsive to others
[emphasis mine]. (1973:79)

Ziller has not been the only theorist to construe com-

plexity as responsiveness. Direct support for the hypothe-

sis associating complexity of the self-concept and re-

sponsiveness to others is provided by Thompson (1966). It

has been shown that those with complex, as opposed to simplex

self-concepts, were able to perceive strangers twenty years

their senior as being more similar to themselves. Com-

plexity of the self-concept was measured by counting the

number of adjectives checked as descriptive of the self.

Additional support comes from Ziller et al. (1964) who

found that children with complex as opposed to simplex self-

concepts were more sociometrically popular.









Research which supports the implications of self-esteem

offered by Ziller (1973) shows that people who feel inade-

quate and inept, rely more heavily on the opinions of others

(field-dependence). Persons low in self-regard are more

easily persuaded (Janis, 1954), tend to be influenced more

by group criticism (Stotland et al., 1957), and conform

more (Coopersmith, 1967) than do those with high self-

esteem. Additionally, people low in self-esteem do not

have the independence necessary for creative work, which

would involve a relative degree of autonomy from the cri-

ticism of others (Coopersmith, 1967). Likewise, Terbovic

(1970) classified subjects into high and low self-esteem.

Then he placed the subjects in a teaching situation where

the "student" (a confidant of the experimenter) was only

allowed to turn on a red light to indicate his displeasure

or a green light to indicate his pleasure with the subject's

instruction. The instruction concerned directions for

drawing a figure. The findings showed that persons with low

as opposed to those with high self-esteem valued more highly

the feedback from the students. Again, the results support

the thesis that persons with low self-esteem are more re-

sponsive to others.

This brief review has provided theoretical and empirical

support for the usage of two major concepts employed in the

formal theory being tested, namely that self-complexity has

been associated with field-dependence or responsiveness.









Additionally, those high in self-esteem manifest high de-

grees of field-independence as opposed to those low in

esteem who are responsive to others' appraisals.


The Formalized Theory as a Predictor of
Differential Drinking Patterns

The formalized theory of this study is both a theo-

retical advancement in self-other concepts as they relate to

behavioral patterns and one from which predictions can be

made. The prediction problem was to ascertain which vari-

ables are most important in determining differential drinking

patterns. This was accomplished by regarding drinking pat-

terns as the dependent variable and all other variables in

the formal theory as predictors. Multiple regression statis-

tical techniques (nonlinear) rendered beta coefficients

which indicated the relative predictive power of each vari-

able in the prediction equation. The final R2 indicated the

variation explained in drinking patterns by all other vari-

ables in the formal theory.


Data and Method

The data for this study were collected at the Alcoholic

Rehabilitation Centers in the South Central Florida counties

of Lake, Sumter, and Polk. The subjects came to the reha-

bilitation centers in a number of ways. Some were committed

on an involuntary basis while others came voluntarily.

The various ways that people came to the centers re-

flect in a large way the diversity that exists among a group









of individuals considered alcoholic. Some of the people

were long-term drinkers while others were more sporadic in

their drinking behavior. With some, drinking had begun to

affect their families and their ability to hold a job.

Others had been arrested for driving while intoxicated, and

were bound over to the centers by the courts, but showed no

difficulties with family relations or with occupational

stability. It is this diversity within a group of people

who are most often considered alike that allows for the

empirical testing of a theory that attempts to explain

variations in terms of alcoholic self-views and drinking

patterns.

Prior to release the subjects were interviewed by one

of the staff members at the respective rehabilitation cen-

ters. An extensive interview schedule was administered

which contained socio-demographic items and items regarding

drinking behaviors as these behaviors related to the sub-

jects, their families, and wider interpersonal relation-

ships. Additionally, items tapping perceptual phenomena

like "selves" and "others" were included. These perceptual

variables are emphasized in this study since it is the

writer's assumption that people's subjective views of self

and others are critical in predicting behavior. Interviews

were completed on 298 subjects which comprised the total

population of this study.









Since the subjects were not drawn on a probability

basis from any specified population, tests of significance

are statistically inappropriate, thus rendering inferences

beyond the subjects analyzed necessarily speculative. How-

ever, to the extent that the subjects studied are similar in

selected socio-demographic characteristics to other known

alcoholic populations, tentative inferences can be offered.

In order to test the zero-order hypotheses derived from

the formal theory, the gamma statistic was used. Gamma is

amenable to ordinal level data and lends itself directly to

a proportional-reduction-in-error interpretation. Also,

gamma reflects the directionality of the relationship which

is necessary in testing the formalized theory.

For the prediction problem, which was to ascertain the

relative importance of the independent variables taken

simultaneously in explaining variation in drinking patterns,

two multivariate statistical procedures were employed. The

first procedure is entitled, Automatic Interaction Detector

(AID).

AID is something like a "step-wise regression program."

Regarding one of the variables as a dependent variable, the

analysis employs a nonsymmetrical branching process, based

on variance analysis techniques, to subdivide the sample

into a series of subgroups which maximize one's ability to

predict values of the dependent variable. The independent

variables ("predictors") need not be quantitative. One can









have either quantitative categories (as for age or income),

or qualitative categories (as for sex, marital status, cause

of death, or political preference). Also with AID, the

quantitative predictors can be categorized into intervals

of unequal length (e.g., incomes under $15,000, $15,000-

24,999, $25,000-49,999, etc.) and even into nonordinal

categories (e.g., ages 20-24, 25-54, or 55-64,and 65 and

older). Linearity and additivity assumptions inherent in

conventional multiple regression techniques are not required.

The AID program operates by finding that dichotomy,

based on any predictor, which gives the lowest within-group

sum of squared deviations for the dependent variable.

Essentially, this is the dichotomization which "accounts

for" more of the variance of the dependent variable (i.e.,

has a larger "correlation" with the dependent variable) than

any other dichotomization based on grouping the categories

of a single predictor into two groups.

Having established this first dichotomy, the AID pro-

gram then takes the "eligible" group with the largest within-

group sum of squared deviations for the dependent variable

and "splits" it in a similar manner. A Group is eligible

for splitting if it has a within-group sum of squared de-

viations at least as great as a specified proportion of the

original sum of squared deviations. Splits will be made

only if the within-group sum of deviations is reduced by

some minimum proportion of the total sum of squares. The









process of dichotomizing groups continues until there are no

eligible groups which can be split to yield the specified

minimum within-group sum of squared deviations reduction.

In the analysis of this study a group is "eligible" for

splitting if it has at least ten cases in both "subgroups"

resulting from that split, a within-group sum of squared

deviations at least 1 per cent as great as the original sum

of squared deviations, and if the within-group sum of

squared deviations of the group to be split is reduced by at

least 5 per cent of the total sum of squares.

After the initial interaction detector runs the joint

use of AID and Multiple Classification Analysis (MCA), a

nonparametric analogue to conventional regression analysis

is employed (Andrews et al., 1971). This procedure is now

described briefly.

Sonquist (1971) has extended the use of AID by coupling

it with MCA. The rationale for the joint use of AID and

MCA has been outlined extensively. Therefore, only the

basic reasoning is presented here. First, an initial AID

run is employed to identify interaction effects among pre-

dictors. From this analysis, if nonadditivity is serious,

interaction terms are constructed by combining the different

categories of the interacting predictors ("cross-products")

to create a new variable that is to be used in MCA. This is

referred to as the "complete model" by Blalock (1960). To

determine the seriousness of interaction, the predictors









chosen at each step in the partitioning process of AID are

compared with the alternative predictors as to the propor-

tion of unexplained variance reduced. If the alternative

predictors are high competitors, and if they contribute to

the establishment of symmetry in the AID "tree" format,

the seriousness of interaction is reduced (Sonquist, 1971:

21).

Conjointly, the gravity of interaction is assessed by

comparing the R2's of MCA with and without product terms.

If the total variance explained by the two models is not

significantly different, that is, if the cross-product terms

do not add significantly to the reduction of unexplained

variance, an additive model is assumed. Though there are

several separate runs from both procedures, the final model

is based on the MCA R2 and statistics analogous to zero-

order eta and partial beta coefficients. The final multiple

correlation coefficient then reflects the best combination

of self-other variables which maximize prediction in

drinking patterns.

In summary of this chapter, the formal theory of per-

ceived self-other concepts as these concepts relate to

alcoholic self-views and drinking patterns has been pre-

sented. Though previous theories have dealt with the inter-

relationships among perceptions of self and others, none

has been formalized so as to incorporate dimensions of the

self-concept. It is proposed that a multidimensional view









of the self can most adequately account for the interrela-

tionships among perceived alcoholic evaluations of others

and the subject's alcoholic self-views and behavioral pat-

terns. A review of the literature indicates that the

multiple component approach to the self is theoretically

and empirically more tenable than a unitary view. Further,

the literature surrounding selected concepts of the formal

theory was reviewed. This review provided support for the

usage of self-complexity as being associated with respon-

siveness to others, and self-esteem as being associated

with the lack of responsiveness. Additionally, the problem

of predicting drinking patterns from all other variables

of the theory was stated. Finally, the method of analysis

for testing the formal theory and for addressing the pre-

diction problem has been outlined.














CHAPTER III

FINDINGS

Zero-Order Findings

In this chapter, the findings that relate to the zero-

order hypotheses derived from the formal theory are pre-

sented. Along with the findings of this study, previous

theory and research consistent with the present framework

are also discussed. This will serve to show how seemingly

fragmented findings from diverse sources can be brought

together under a single, systematic framework. It high-

lights the positive aspects of formal theory construction.


Hypothesis I: The greater the SIGOTHER,
the greater the ALCOSELF

As may be seen in Table 1, the descriptive statement

relating the perceived alcoholic evaluations of others with

alcoholic self-concept is .50. This hypothesis is con-

sistent with the orientations of Mead (1934), Cooley (1902),

and other interactionists who posit the view that one's

self-conceptions are a function of the perceived appraisals

of others. In accordance with the interactionist framework,

the finding in this research shows that a subject's view of

himself as an alcoholic is heavily influenced by the per-

ceived alcoholic evaluations of significant others.














Table 1. Zero-Order Hypotheses Derived From the Formal
Theory and Descriptive Statements as Measured
by Gamma



Hypotheses Descriptive Statements


I: The greater the SIGOTHER, + .50
the greater the ALCOSELF

II: The greater the ALCOSELF, + .64
the greater the PATTERNS

III: The greater the FQLABEL, + .45
the greater the ALCOSELF

IV: The greater the SIGOTHER, + .49
the greater the FQLABEL

V: The greater the SIGOTHER, + .45
the greater the PATTERNS

VI: The greater the FQLABEL, + .36
the greater the PATTERNS

VII: The greater the ESTEEM, .20
the less the ALCOSELF

VIII: The greater the COMPLEX, + .38
the greater the ALCOSELF

IX: The greater the ESTEEM, .30
the less the COMPLEX

X: The greater the ESTEEM, .30
the less the PATTERNS

XI: The greater the COMPLEX, + .34
the greater the PATTERNS









Hypothesis II: The greater the ALCOSELF,
the greater the PATTERNS

As shown in Table 1 the hypothesis relating subject's

alcoholic self-view with the subject's drinking patterns is

the highest association obtained. The descriptive statement

is .64.

The symbolic interactionist framework offers the thesis

that one's behavior is contingent on one's self-views which

are a function of the perception of others' appraisals. The

first hypothesis was concerned with the second aspect of

this thesis. The present hypothesis deals with the first.

The interrelationships between the perception of others'

appraisals, self-views, and consequent behaviors have been

systematically stated by Kinch (1963:481-486). Two propo-

sitions from Kinch's formalization of Meadian self theory

are relevant here:

1. The individual's self-concept is based on his per-
ception of the way others are responding to him.

2. The individual's self-concept functions to direct
his behavior.

Proposition one relates to the first hypothesis in the for-

mal theory of this study. Kinch's second proposition is

identical to the second proposition of the present theory

and from which the second hypothesis is derived. The second

hypothesis of this study is supported by a moderately strong

association which lends confirmation to both the present

framework and Kinch's theoretic statement.









Hypothesis III: The greater the FQLABEL,
the greater the ALCOSELF

From Table 1 it is shown that the association between

the perceived frequency with which subjects are labeled

alcoholic and the subject's alcoholic self-concept is .45

and in the direction hypothesized.

The notion that the greater the perceived frequency

with which a person is labeled, the greater the probability

that the person will incorporate the label into his self-

definition, has intuitive appeal. Theory and research re-

lating to this hypothesis come from attitude change models.

Just as one holds attitudes toward objects, so can he hold

attitudes toward self. If a person holds one view of self

and perceives that others are frequently confronting him

with a view which is counter to his own, a state of cogni-

tive imbalance may occur. Following Heider (1946), in

order for a person to achieve a balanced state, he may bring

his self-conception into line with the perceived views of

others. Since the perceived appraisals of others are fre-

quent, a person cannot easily dismiss them in self-

assessment. Consequently, the perceived relative frequency

with which persons are labeled alcoholics should, according

to consistency theory, have corresponding effects on self-

conceptions. Of course, if there is no discrepancy between

self-view and the perceived labeling frequency of others at

the outset, then self-views are not changed but are merely

reinforced.









Hypothesis IV: The greater the SIGOTHER,
the greater the FQLABEL

As may be seen in Table 1 the descriptive statement

associating the perceived alcoholic evaluations of signifi-

cant others and the perceived frequency with which one is

labeled alcoholic is .49.

This hypothesis is derived from the sign rule men-

tioned in Chapter II. Despite the fact that the association

between the variables must be positive in accordance with

the sign rule of logic, the relationship is consistent with

the symbolic interactionist framework.

We have already seen that the perceived evaluations of

significant others have a heavy influence on a subject's

alcoholic self-conception. As the descriptive statement

indicates, this association is moderately strong. Inter-

actionist theory suggests that it is the perceived appraisals

of significant others that have the most important influence

on self-constitution. Once the self becomes somewhat

crystalized, an individual tends to selectively perceive the

frequency of appraisals consistent with the established

self-definition (Kinch, 1973). Therefore, if one's alco-

holic self-view is strong due to the alcoholic evaluations

of significant others, we should expect that the person

would be more receptive to a greater number (relative fre-

quency) of alcoholic appraisals from diverse sources than

one who does not hold a strong alcoholic self-view. Since

the measure of frequency of alcoholic evaluations is based









on perceived labels from others who range in intimacy from

"spouse" to "hospital," one who has a strong alcoholic self-

image due to the alcoholic evaluations of significant others

should select out a greater number of alcoholic labels,

which is consistent with his strong alcoholic self-view.

This is predicted by the sign graph relating these three

variables and is supported by the data relating to the

present hypothesis.


Hypothesis V: The greater the SIGOTHER,
the greater the PATTERNS

As may be seen in Table 1, the descriptive statement

relating the perceived alcoholic evaluations of significant

others with subject's drinking patterns is .45, as mea-

sured by gamma. This is in the direction hypothesized by

the formal theory.

A third proposition from Kinch's (1963) formalization

of the interactionist framework is derived from the two

propositions stated when discussing the relationships be-

tween alcoholic self-conception and drinking patterns:

3. The way the individual perceives the responses of
others toward him will influence his behavior.

The present hypothesis links others' perceived behavior to

one's consequent response, and is identical to Kinch's de-

rived proposition. The paradigm becomes fully interrelated

proposing that one's perception of others' responses









influences one's self-views which influence one's behavior.

The finding in this study supports both Kinch's formalization

and the present theory.


Hypothesis VI: The greater the FQLABEL,
the greater the PATTERNS

As is shown in Table 1 the descriptive statement re-

lating the perceived labeling frequency of others and

drinking patterns is .36.

It is the association between perceived labeling fre-

quency and drinking patterns that completes the sign graph

relating drinking patterns to perceived labeling frequency,

perceived alcoholic evaluations of significant others, and

one's alcoholic self-concept. The two concepts, perceived

alcoholic evaluations of significant others, and alcoholic

self-view, in addition to the "corollary" concept (Kinch,

1973) perceived labeling frequency, serve to theoretically

link the perception of others' appraisals with the sub-

ject's consequent drinking patterns.


Hypothesis VII: The greater the ESTEEM,
the less the ALCOSELF

From Table 1 it may be seen that the descriptive

statement associating subject's self-esteem with one's

alcoholic self-view is -.20, as measured by gamma.

This hypothesis stems from the conceptual usage and

empirical findings in regard to self-esteem that was dis-

cussed in Chapter II. The inference from the finding re-

lating to this hypothesis is that those high in self-









estimates are field-independent and hence somewhat insulated

from the perceived evaluations of others; consequently they

do not develop alcoholic self-views. Theoretical and em-

pirical support for this inference can be found in works

previously cited (Ziller, 1973; Janis, 1954; Coopersmith,

1967).


Hypothesis VIII: The greater the COMPLEX,
the greater the ALCOSELF

From Table 1 it is shown that the descriptive state-

ment relating self-complexity with one's alcoholic self-

conception is .38. Though this association is moderate,

it is in the direction hypothesized by the formal theory.

The inference stemming from this finding is that those who

are high as opposed to those who are low in self-complexity

show a greater tendency to respond to the perceived alco-

holic evaluations of others and hence to develop alcoholic

self-concepts in accordance with those appraisals. The

association of self-complexity with responsiveness to

others' appraisals is consistent with the findings of Ziller

(1973) and Thompson (1966).


Hypothesis IX: The greater the ESTEEM,
the less the COMPLEX

As may be seen from Table 1 the descriptive statement

relating self-esteem with self-complexity is -.30.

Though this is a derived hypothesis, there is a vast

amount of research supporting the relation of these two









self-dimensions. In Chapter II, we found that a number of

studies support this hypothesis. With regard to esteem, it

was found by Ziller (1973), Coopersmith (1967), and others

that the concept is associated with field-independence.

The findings associating self-complexity with responsiveness

or field-dependence were cited in the preceding discussion

of Hypothesis VIII.


Hypothesis X: The greater the ESTEEM,
the less the PATTERNS

In Table 1 it may be seen that the association between

self-esteem and drinking patterns is -.30 and in the direc-

tion hypothesized by the theory.

The inference from this finding is that subjects high

in self-estimates are relatively insulated from the per-

ceived alcoholic evaluations of others and hence, do not

develop strong alcoholic self-views. The data have already

supported the hypothesis associating alcoholic self-views

and drinking patterns. If the inference is accepted, the

low alcoholic self-concepts among subjects high in self-

esteem result in lower patterns of drinking. The inference

has empirical support from the negative association between

the variables of the present hypothesis.


Hypothesis XI: The greater the COMPLEX,
the greater the PATTERNS

As may be seen in Table 1 the association between self-

complexity and drinking patterns is .34.









The basis for this hypothesis comes from the relative

responsiveness of subjects high in self-complexity. Those

high in self-complexity are responsive to the perceived

alcoholic appraisals of others. This leads to the develop-

ment of an alcoholic self-image which, in turn, leads to

higher patterns of drinking. As previously found, the

association between alcoholic self-concept and drinking

patterns was .64. Unlike those high in self-esteem who are

associated with a resistance to the perceived alcoholic

appraisals of significant others, those high in self-

complexity manifest the opposite pattern. They are vul-

nerable to the perceived alcoholic labeling of others. The

positive association between complexity and drinking pat-

terns in the present hypothesis supports this inference.

A few words by way of summary for the first section

of this chapter are in order. The data in this study

support the formalized theory in every instance. The

highest association was between subject's alcoholic self-

views and drinking patterns. The phenomenological orienta-

tion that guides this study suggests that one's "self" is

a crucial object taken into account when actors are

assessing probable lines of conduct. This orientation is

supported in a strong way by the high correspondence between

one's conception of himself as an alcoholic and his con-

sequent drinking behaviors.









Automatic Interaction Detector Findings

In this section of Chapter III, the problem of pre-

dicting differential drinking patterns is now addressed.

Analysis indicates which variables of the theory are most

important in predicting variations in drinking patterns.

The analysis of the data is based on a procedure developed

by Sonquist and Morgan (1964), Automatic Interaction De-

tector (AID). In keeping with the criterion of AID, that

the dependent variable must be an interval scale or a

dichotomy (Andrews et al., 1971:08), the variable drinking

patterns was dichotomized, for the purposes of this

analysis, into a group of 129 subjects having scores ranging

from 8 to 14 and a group of 169 subjects having scale values

ranging from 15 to 24. The group of subjects with scale

scores ranging from 8 to 14 comprised the "lows" on drinking

patterns, and the group whose scale scores ranged from 15

to 24 constituted the "highs." The 17 categories ranging

from 8 to 24 were established by multiplying the 8 drinking

patterns items by their 3 response categories, "never,"

"sometimes," and "frequently," which were coded 1 through

3 respectively.

Figure 2 presents the "tree" resulting from the AID

splitting process when attempting to predict values of

drinking patterns.

Of the independent variables used, alcoholic self-

concept was the strongest predictor with respect to reducing









































































Un


0
-4 .





8
--(









the unexplained variance in drinking patterns. The split

for the entire population of subjects by alcoholic self-

concept occurred between categories five through eight and

categories nine and greater. Those subjects high in

alcoholic self-conceptions manifested far higher drinking

patterns (X=1.8) than those low in alcoholic self-views

(X=1.4).

Of the subjects high in alcoholic self-concept

those who perceived more alcoholic evaluations from signifi-

cant others manifested higher drinking patterns (X=1.8).

This partition occurred between categories five through

seven and categories eight through ten, the mean drinking

pattern scores being 1.6 and 1.8 respectively. Further, of

the subgroup high in perceived alcoholic evaluations by

significant others, those high in self-complexity manifested

higher patterns of drinking (X=1.9). Therefore, those sub-

jects who were high in alcoholic self-views and high on

perceived alcoholic evaluations by significant others, and

those who scored high on self-complexity, scored the highest

on drinking patterns.

Returning to the original partition, one may trace

those subgroups that were the lowest on patterns of drinking.

The top "twig" of the AID tree shows the subgroups that are

highest in the dependent variable, those who are lowest in

drinking patterns are depicted by the bottom "twig" of the

format.









As previously mentioned, alcoholic self-concept was

the strongest predictor of drinking patterns. As Figure 2

shows, those who were lower on this variable manifested

lower scores on the dependent variable (X=1.4). Of the

subjects who were low on drinking patterns as a result of

the first split by the predictor, alcoholic self-concept

those who were high on self-esteem showed a lower mean score

on drinking patterns (X=1.3). This partition occurred be-

tween categories four through six and categories seven and

eight, the means being 1.6 and 1.3 respectively. Further,

of the group high on self-esteem (group 5 in the bottom

twig) those who scored lower on alcoholic self-concept

manifested lower drinking patterns. Finally, group 2 showed

that, again, subjects high in self-estimates scored lower

on drinking patterns (X=l.l). Therefore, it was those who

were low in alcoholic self-concept and relatively high in

self-estimates that showed the lowest degree of drinking

patterns.

The previous discussion was designed to determine how

the partitioning process of AID might aid the researcher in

empirically identifying various subgroups in a population

that is relatively high andlow on the dependent variable,

in this case, drinking patterns. Though this is a fruitful

use of the AID tree format insofar as one may be interested

in pinpointing extreme problem drinkers and their correlates

(predictors), it is somewhat secondary to AID's major use.









The major use of the algorithm is to detect interaction

effects among predictors. In order to exploit this feature

of AID, reference will be made to Table 2, in which both the

predictors and their alternatives are presented. An analysis

of the interior of the tree format, as well as the top and

bottom twigs, reveals interaction effects and their extent.

The seriousness of nonadditivity may be ascertained by

assessing the extent to which the tree departs from symmetry

and by a comparison of the R2's rendered by Multiple Classi-

fication Analysis (MCA) with and without interaction terms.

If the addition of cross-products does not add to the ex-

plained variation in drinking patterns, an additive model

is assumed.

As shown in Table 2 and indicated by the analysis of

the AID partitioning process undertaken in the first part of

this section, alcoholic self-conception was the predictor

making the first split. If the effects of the predictors

were additive, one would expect that the resulting groups

(groups 2 and 3) would be split by the same variable. As

Figure 2 illustrates, this is not the case. Group 2 is

partitioned by the predictor, perceived alcoholic evalua-

tions of significant others, while group 3 is split by self-

esteem. However, examination of Table 2 shows that a close

alternative predictor for group 3 is the perceived evalua-

tions of significant others. This indicates that, though

parent groups 2 and 3 were not split by the same predictor,















Table 2.


Proportion of Variance Explained by Predictors of
Drinking Patterns at the Time of Their Selection
Compared to Best Alternative Predictorsa


Split Chosen Predictors Best Alternative Predictors
Group BSSx/TSSix BSSx/TSSix
x. Jx X IX


ALCOSELF


SIGOTHER


ESTEEM


COMPLEX


ALCOSELF


ESTEEM

ESTEEM


COMPLEX

ALCOSELF

COMPLEX

SIGOTHER

FQLABEL


12.9


1.7


7.2


3.5


3.9


3.2

2.5


1.5


SIGOTHER
FQLABEL

COMPLEX
FQLABEL

SIGOTHER
ALCOSELF

ESTEEM
SIGOTHER

SIGOTHER
FQLABEL

COMPLEX

COMPLEX
FQLABEL

FQLABEL

SIGOTHER

FQLABEL


1.0


ESTEEM


11.1
9.0

1.5
1.4

6.7
6.6

1.6
1.3

3.0
1.9

1.3

1.7
1.6

.8


aWhere no alternative predictor is presented its
BSS /TSSix was not significant enough to record.
x ix









they almost were. Since the perceived evaluations of

significant others was a high alternative for partitioning

group 3, the gravity of nonadditivity at this point is re-

duced significantly.

Group 9 did not meet the split reducibility criterion

and was therefore not partitioned. Group 8 was split by

self-complexity between categories three through six with a

mean of 1.8 and categories nine and greater with a mean of

1.9 (group 12 in the tree). Table 2 shows that parent group

4 was split on self-complexity and parent group 5 parti-

tioned by alcoholic self-conception. This, at first glance,

appears to be the first significant departure from symmetry.

However, examination of the alternative predictors for

groups 4 and 5 shows that, in both cases, the perceived

evaluations of significant others is a viable alternative

predictor; though it was the best alternative for group 5,

it was the second best for group 4. Since the perceived

evaluations of significant others fared well as an alter-

native predictor for both groups, interaction effects in

this part of the tree were considered less serious than

initially suspected. However, a first-order interaction

term was constructed using the variables self-complexity and

self-esteem in the cross-product term. This same kind of

analysis was extended throughout the entire partitioning

process of the AID algorithm. On the basis of this analysis,

two first-order interaction terms and one second-order were









constructed. These cross-products were employed in MCA

analyses to ascertain if the addition of interaction terms

would add significantly to the final R2. If the R2 of MCA

with interaction terms did not significantly reduce unex-

plained variance in the dependent variable, then an additive

model would be assumed and the mild asymmetry of the AID

tree format would be attributed to the arbitrary selection

of the split reducibility criterion and the minimum obser-

vations criterion discussed in Chapter II.

Table 3 presents the variables used by AID in pre-

dicting values of the dependent variable, drinking patterns.

The multiple correlation coefficient shows that 43 per cent

of the variation in drinking patterns was explained by the

tree.

Of the variables chosen, alcoholic self-concept was

the most important predictor. This is consistent with the

zero-order findings mentioned earlier in this chapter.

Following the variable, alcoholic self-concept, the two-

variable pattern of self-esteem and self-complexity were

next in importance as reflected by their partial beta co-

efficients. After this dyadic pattern, the perceived alco-

holic evaluations of significant others surpasses perceived

labeling frequency having betas .022 and .005 respectively.


Multiple Classification Analysis Findings

In the preceding section of this chapter, analysis was

based on Automatic Interaction Detector, a multivariate














Table 3. Predictors Explaining Variance in Drinking
Patterns Using Automatic Interaction Detector


Predictor


ALCOSELF

ESTEEM

COMPLEX

SIGOTHER

FQLABEL


Partial Beta Coefficient
B x2


.230

.140

.032

.022


.005

R2 = BSST/TSST = .429


Z TSSi Z TSS.j
i TSST j TSST


= Partial Beta Coefficient


ix = Parent groups split by predictor x.
jx = New groups formed by splitting a parent group on
predictor x.

TSS = Total Sum of Squares.
BSS = Best Sum of Squares possible in dichotomizing
predictor x.
2
R = NP
Z x2 = BSST = Total Proportion of Variance
i TSST Explained by Tree

NP = Number of predictors used in the analysis.


sX2 =


--









procedure based on an interactive model. Sonquist (1971)

has extended the use of AID by coupling it with Multiple

Classification Analysis (MCA). The rationale for the joint

use of AID and MCA was outlined in Chapter II.

From the AID analysis in the preceding section, three

interaction terms were constructed. An extensive examina-

tion of the AID tree format combined with the alternative

predictor table, indicated that interaction effects among

predictors was minimal. However, cross-products were con-

structed to serve as a check against this indication. As

mentioned previously, if the additive assumption suggested

by the AID tree analysis is valid, the inclusion of multi-

plicative terms into MCA should not increase its ability to

account for unexplained variance. An initial MCA run em-

ploying the cross-products rendered a multiple correlation

coefficient of .46. The R2 and partial beta coefficients

yielded by MCA without interaction terms are presented in

Table 4, which shows that the total amount of variation ex-

plained by all variables taken simultaneously is 43 per

cent. Based on the fact that the AID partitioning process

(when alternative predictors are considered) approached

symmetry, and the MCA R2 without interaction terms accounted

for approximately the same amount of variation in the

dependent variable, an additive model is assumed.

Additional evidence for the additivity assumption comes

from a comparison of the relative predictive ability of each














Table 4. Proportion of Variation in Drinking Patterns Ex-
plained by Using Multiple Classification Analysis



Predictor (MCA)b (MCA)b
Eta2 Beta2


ALCOSELF .34 .200

COMPLEX .20 .043

ESTEEM .09 .033

FQLABEL .18 .026

SIGOTHER .23 .024

MCA R2 = .43

bBeta2: Beta measures the ability of the predictor to
explain variation in the dependent variable after adjusting
for the effects of all other predictors. Beta2 for pre-
dictor i = Di/T, where Di = sum of squares based on adjusted
deviations for predictor i, and T = total sum of squares.

Eta2: Eta-squared is the proportion of variance in the
total group explainable by each variable taken by itself.









variable by the AID and MCA algorithms. From Table 4 it may

be seen that alcoholic self-conception is the most important

predictor of drinking patterns in the MCA procedure. This

was also the case with the AID procedure, the findings from

which are shown in Table 3. With MCA, the two-construct

vulnerability pattern of self-esteem and self-complexity

emerges as an important predictor of the dependent variable.

As previously mentioned, this pattern was important in ex-

plaining variance in drinking patterns when AID was used.

Finally, the variables perceived evaluations of significant

others and subject's perceptions of labeling frequency

proved to be less predictive than the other variables for

both multivariate procedures.

Some concluding statements for this section of Chapter

III are in order. In addition to testing the hypotheses

derived from the formal theory of this study, the problem

of prediction was addressed. The problem was to ascertain

which variables were most important in predicting drinking

patterns. This was accomplished by designating drinking

patterns as the dependent variable and employing two multi-

variate procedures that took all other variables into

account simultaneously. An initial AID run was designed to

detect interaction effects among predictors. Interaction

terms were constructed and employed in the MCA procedure.

It was found that the multiplicative terms did not increase

the total variance explained when compared with the MCA









multiple correlation coefficient excluding cross-products.

Therefore, the final MCA run, assuming additivity, rendered

a R2 of 43 per cent. Of the predictor variables, alcoholic

self-concept was found to be most important in reducing un-

explained variance as reflected by its partial beta in both

the AID and MCA procedures. Further, the two-variable pat-

tern of self-esteem and self-complexity was relatively im-

portant in its predictive ability in both algorithms.


A Potential Theoretical Difficulty

There was a potentially weak link in the formal theory

deriving from the methodology of this study. As mentioned

in Chapter I there was the possibility of tautology stemming

from the use of self-report data. In this section of

Chapter III, data analysis is undertaken that provides some

clues for its solution.

The potential tautological problem in hypotheses re-

lating perceptual variables like "selves" and "others"

ascertained from self-reports, stems from the possibility

that the reports about self and others may yield data which

reflect a single phenomenon. If the relationships are

tautological due to some cognitive process like selective

perception, then subjects will perceive only those evalua-

tions of others that are consistent with self-evaluations,

irrespective of who the others are. Even though the others

may vary in terms of their assumed importance, subjects will

select out only those statements consistent with their own









self-views, notwithstanding variations in the others' social-

psychological significance. Therefore, when the researcher

correlates two variables of an hypothesis relating percep-

tions of self and others, a very high correspondence is ob-

tained. If the assertion is tautological, one might ideally

expect a perfect correspondence.

The analytical procedure for addressing this problem is

elaboration analysis (Rosenberg, 1968). Two of the con-

structs in the formal theory, namely, self-esteem and self-

complexity, were combined into a two-variable vulnerability

pattern. More specifically, subjects who were low in self-

esteem and high in self-complexity were high in vulner-

ability, and those who were high in self-esteem and low in

self-complexity were low in vulnerability. The two cate-

gories of vulnerability served as the "test factor" in the

elaboration process. In addition to the stipulation that a

subject must meet the joint condition of esteem and com-

plexity (whether low esteem and high complexity or high

esteem and low complexity) he must also meet the condition

that the perceived evaluations of friends are not the same

as the perceived appraisals of a person to whom he feels

closest. This last condition is crucial in testing the

problem of tautology. What is needed is to show that there

is a differential impact by the perceived alcoholic evalua-

tions of friends and closest person on alcoholic self-

conception according to varying degrees of vulnerability.









If this can be shown, it is evidence that self-perceptions

and perceptions of others are two separate variables. Of

those subjects in the high vulnerability category, 26 met

the added condition that there is a difference between the

perceived appraisals of friends and closest person, while 29

subjects met the condition in the low vulnerability category.

The variables correlated within each category of vul-

nerability were the perceived appraisals of others and alco-

holic self-concept. In the initial measure constituting

the perceived alcoholic evaluations of significant others,

only the perceived evaluations of the subject's friends

were measured. It was assumed that the subject's friends

comprised significant others in their lives. In the "par-

tial" correlations computed within each category of vulner-

ability, perceived alcoholic evaluations of the subject's

friends and the perceived alcoholic evaluations of a person

to whom he felt "closest" (when the two are different) were

used to explore the hypothesis that high vulnerable' alco-

holic self-conceptions are influenced by all others (extreme

field-dependence) while low vulnerable' alcoholic self-

views tend to be influenced more by the perceived appraisals

of those to whom they feel closest. If it could be shown

that there was a differential influence by perceived friends

and closest person on alcoholic self-concept according to

relative degrees of vulnerability, then the "specifying"

effect of the vulnerability pattern would be evidence that









perception of others is not identical to self-perception.

In a word, the tautology would then be dampened in a signifi-

cant way by:

1. the fact that there is not a perfect correlation
between the perceived alcoholic evaluations of
others and alcoholic self-perception, and

2. the fact that there is a differential impact by
the perceived alcoholic evaluations of friends and
closest person on alcoholic self-concept within
each category of vulnerability, which is evidence
that self-perceptions and perceptions of others are
two separate variables.

Table 5 presents the relationships between the per-

ceived alcoholic evaluations of the subject's friends,

closest person and consequent alcoholic self-concept within

the high vulnerability category. It was hypothesized that

subjects scoring extremely high on self-complexity and ex-

tremely low on self-esteem would be responsive to the

appraisals of all others (extreme field-dependence). The

association between perceived alcoholic evaluations by per-

sons to whom the subjects felt closest and the subject's

alcoholic self-concept is .62, as measured by gamma. This

moderately strong relationship indicates the responsiveness

associated with those high on the vulnerability pattern.

Also, from Table 5, it may be seen that the association be-

tween perceived evaluations by the subject's "friends" and

consequent alcoholic self-conception is .79. Again, the

strong association shows that the two-construct pattern is

indicative of responsiveness to others' appraisals. There-

fore, the hypothesis that extreme vulnerable are responsive














Table 5. Relationships Between Others' Evaluations and
Alcoholic Self-Concept Within High Vulnerability
Category


ALCOSELF


FRIENDS
COUNT I
ROw PCT
COL PCT
TOT PCT 1.00T .2.001. 3,001
..-.I .--- I- I.....-. I.......,
1,00 I 14 I 1 I 0 I
I 93.3 I 6,7 I 0.0 I
I P7.5 I 20,0 I 0,0 I
I. 53.8 I 3 8 .1 0.0. I
SI- ---....--- I -- -- I---. -
2.00 I I 2 1 4 I
I 14.3 28.6 1 57.1 I
I 6.3 I 40,0 I 80,0 I
I .3.8. I 7,7 15 4 I
-I -,....I----.....I--,---* I
3.00 I .1 T 2 1 1 I
I 25.0 I 50,0 I 25,0 I
I 6.3 T 40:0 I 200 I
38 I 7,7 .1. 3,8
SI ------ --- I--- ....I ----* 7
COLUMN 16 5 5
TOTAL 61.5 19,2 19,2


Gamma = .79.


ALCOSELF


CLOSEST
COUNT CLEST
ROW PCT I
CnL PCT I
TOT PCT I 1.001 .2,001. 3,001
--.--.... --....... I-.....-1.. .......- I
1.00 I 13 I 1 I .1 I
I 86.7 I 6,7 I 6.7 I
I 76.5 I 1 ,3 I 50.0 I
I 50.0 T 3.8 I 3,8 I
-I------- ---.---I.-------
2,00 I .2 I 5 I 0 I
I 28.6 I 71,4 I 0.0 I
I 11.8 7 71. I 0,0 I
I 7.7 T 19.2 I 00 I
1-j-- ---- 1 ---- -- j- -, T
3,00 I .2 I 1 1 I
I 50.0 I 25,0 1 25.0 I
I 11.8 I 1 .3 I 50.0 I
I 7,7 3,8 I 3.8 I
.I ------. I ------------. -- --- I
COnLI:UN 17 7 .2
TOTAL 65,. 26,9 7,7


Gamma = .62.


SROW
TOTAL

15



26,s


4
15,4


26
100,0


.ROW
TOTAL

15
57,7



26.,


26
100,0









to the perceived alcoholic evaluations of all others is

supported.

The relationships between the perceived alcoholic

evaluations of the subject's friends, closest person, and

consequent alcoholic self-conception within the low vulner-

ability category are presented in Table 6. It was hypothe-

sized that subjects scoring low on the vulnerability pattern

would be less responsive to the perceived appraisals of

others, and, if responsive, more so to those to whom the

subjects felt closest. As the data show, the hypothesis was

confirmed. The relationship between alcoholic self-concept

and perceived alcoholic evaluations from the subjects'

friends is .00, as measured by the gamma statistic. The

association between the perceived alcoholic evaluations of

persons to whom the subjects felt closest and alcoholic

self-conception is .32, indicating that if low vulnerable

respond to others' opinions, the others must be highly

significant.

In summary, the data that bear upon the potential

problem of tautology outlined in this section provide evi-

dence which mitigates considerably the fears of circularity.

The problem was to show that self-perceptions and percep-

tions of others are two distinct variables. The analysis

showed that for high vulnerable the impact of the perceived

evaluations of friends and closest person on alcoholic self-

view was similar. As hypothesized, this was not the case













Table 6. Relationship Between Others' Evaluations and
Alcoholic Self-Concept Within Low Vulnerability
Category


ALCOSELF


FRIEND
COUNT I
ROw PCT I
COL PCT
TOT PCT I 1.00T .2,001 3.001
S**..*** I .I.....--I ..-.... I ... ---I
1.00 I 10 I 2 I 3 I
I 66.7 1 13.3 I 20.0 I
I 62. I 50.0 I 500
1 34.5 1 6.9 I 10 3 I
-I---- -I-I---- I ------- .
2.00 I .8 I 2 I 3 I
I 61.S I 15,4 I 23.1 I
5 5o. 1 50 0
I- -li I
3.00 I II 0 0
I 100.0 I 0.0 1 0,0 T
1 5.3 I 0.0 I 0,0 I
I 3.4 0.0 I. 0.0 I
-I ... .. ** ........ ........
COLUM 19 4 .6
TOTAL 65.5 13,8 20.7


Gamma = .00.


ALCOSELF


COINT CLOSEST
ROW PCT I
COL PCT I
TnT PCT I 1.00T 2.001 3.00T
--.-.. ..I-T ...-------- I.------. I
1.00 I 11 I 3 I I T
I 73.3 T 20.0 I 6.7 I
I 57.9 I 60.0 I 20.0 I
I 37.9 I 10.3 I 3.4 T
-----------------I-----TI
2;00 I 7 1 21 4
I 53. I 15.a I 30,8 I
I 36.8 T 40n. I 80,0 I
I 21.1 T 6.: I. 13.8 1
-I .. 7- ^-....- --.-.. I ...--..I
3,00 1 1 I 0 I 0 T
I 100.0 '. 0,0 1 0.0 f
5.3 T 0,0 0.0 I
.. 3 4. I 0.0 I 6.0 I
CfI I5 17 1 9 15 5
T" TAL 65,5 17'2 17,2


Gamma = .32.


ROW
TOTAL

15
51,7


13


1
3.4


29
100,0


Rnw
TOTAL

51
51,7


3,.


29
100,0









for the low vulnerable. In order for others' perceived

alcoholic appraisals to have an influence on a subject's

alcoholic self-concept within the low vulnerability cate-

gory, the "other" had to be one who was considered closest

to the subject. The differential impact of the perceived

alcoholic evaluations of friends and closest person on a

subject's alcoholic self-views among those subjects who were

relatively insulated from others' appraisals (low vulner-

ability) is supporting evidence that self-perceptions and

perceptions of others are two separate variables.

In summary of the entire chapter, the zero-order hy-

potheses derived from the formal theory presented in Chapter

II were supported by the data. The strongest association

obtained was between the subject's alcoholic self-concept

and drinking patterns. This is consistent with the posi-

tion taken in this study that one's self-views are important

in predicting behavior. When all variables of the theory

were taken simultaneously to predict drinking patterns, the

R2 was .43. Again, alcoholic self-conception proved impor-

tant in accounting for variations in patterns of drinking.

After the prediction problem was addressed, the problem of

tautology was outlined and empirically investigated. The

elaboration analysis provided data which considerably

lessened the fear of tautology. The data showed that there

was differential impact by "friends" and by a person the

subject deemed as "closest" on alcoholic self-view among





58



the two categories of vulnerability. This is taken as evi-

dence that self-perceptions are not identical to perceptions

of others.














CHAPTER IV

A DIFFICULTY IN THE LABELING APPROACH


In developing the formal theory of this study, it was

found that much of the literature bearing upon the inter-

relationships among perceptions of others and perceptions of

self stemmed from labeling theory as an approach to devi-

ance. A search of the theoretical and research literature

on labeling theory brought attention to a shortcoming of the

approach, that is the inability of the approach to account

for differential responses to similar labeling situations.

In this chapter, a framework comprised of selected concepts

from the formal theory are used to address this point.

Before outlining the framework and the method of analy-

sis for testing it, an overview of labeling theory is

needed to illuminate the shortcoming to which the framework

is addressing itself. Labeling theory is traced from early

research to the works of contemporary theorists, and the

emphasis on situational factors in accounting for self-

reconstitution and consequent behavioral patterns is noted.

This emphasis on situational variables has resulted in the

approach's inability to account for unlike responses to

like labeling situations. Attempts have been made to remedy

the shortcoming by invoking social-psychological constructs,









but an overview of these studies shows that the usage of

the social-psychological constructs are inconsistent, prin-

ciply because of inadequate conceptualization and measure-

ment. Additionally, many of the approaches have been a

collection of disconnected hypotheses with no empirical

testing. It is suggested that the proposed framework can

add further to the understanding and possible solution of

the shortcoming of the labeling approach.


Labeling Theory

Labeling theory or the societal reaction approach as

an approach to deviance can be traced to the writings of

Durkheim (Erikson, 1966). Durkheim emphasized the "nor-

mality" of deviance as a boundary maintaining function.

Mead (1918:585-592), developed the same thesis when he

argued the functions of "punitive justice." However, this

emphasis is on the implications of societal reaction for

the revival of social solidarity. The consequences of

labeling for the self-concept of the labeled were initially

pointed out by Tannenbaum (1938) with his notion of the

"dramatization of evil." The role of labeling in the making

of the deviant is captured in the following:

The process of making the criminal, therefore, is a
process of tagging, defining, identifying, segre-
gating, describing, emphasizing, making conscious
and self-conscious; it becomes a way of stimulating,
suggesting, emphasizing and evolving the very traits
that are complained of. If the theory of relation
of response to stimulus has any meaning, the entire
process of dealing with the young delinquent is









mischievous in so far as it identifies him to him-
self or to the environment as a delinquent person.

The person becomes the thing he is described as
being. Nor does it seem to matter whether the
valuation is made by those who would punish or by
those who would reform. (1938:19-20)

In this statement, it is apparent that for Tannenbaum the

response of others or "tagging" is the critical factor in

reconstituting self-definitions and shaping deviant behavior

patterns.

Another theorist who has emphasized the impact of the

negative reactions of others on the self-concept is

Lemert (1951), whose most notable contribution has been his

distinction between primary and secondary deviance. For

the immediate discussion, it is the idea of secondary devi-

ation that bears relevance:

As an illustration of this sequence the behavior of
an errant schoolboy can be cited. For one reason or
another, let us say excessive energy, the schoolboy
engages in a class-room prank. He is penalized for
it by the teacher. Later, due to clumsiness, he
creates another disturbance and, again he is repri-
manded. Then, as sometimes happens the boy is
blamed for something he did not do. When the teacher
uses the tag "bad boy" or "mischief maker" or other
invidious terms, hostility and resentment are ex-
cited in the boy, and he may feel that he is
blocked in playing the role expected of him. There-
after, there may be a strong temptation to assume
his role in the class as defined by the teacher,
particularly when he discovers that there are re-
wards as well as penalties deriving from such a
role. (1951:77)

Further, Lemert notes that "there must be a spreading cor-

roboration of a sociopathic self-conception and societal re-

inforcement at each step in the process" (1951:77). Stated









differently, one's altered self-conception is contingent

upon societal reaction which, in turn, leads to role per-

formance congruent with the reconstituted self-definition.

Becker (1963) has provided one of the most popular ex-

positions on the way that deviant self-conceptions are

shaped by interaction with others. He says: ". being

caught and branded has important consequences for one's fur-

ther social participation and one's self-image" (1963:31-32).

In terms of further social participation, the labelers tend

to generalize the initial act so that "people automatically

assume" that the labeled person possesses other undesirable

attributes. In other words, recognition by others of the

original deviant act evokes the imputation of a generalized

deviant label (what Becker refers to as a "master status"),

which facilitates acceptance of more deviant acts and a

deviant self-image by the person being labeled. One first

becomes identified as a deviant then other attributions are

made. As Becker puts it:

Treating a person as though he were generally rather
than specifically deviant produces a self-fulling
prophecy. It sets in motion several mechanisms
which conspire to shape the person in the image
people have of him. (1963:34)

The contributions of Lemert and Becker appear to be

amplifications or simply restatements of Tannenbaum's origi-

nal thesis (Frazier, 1973:29). The basis of labeling theory

draws on a familiar theme in symbolic interaction theory:

One's self-concept is influenced by the responses of others.









A Shortcoming of the Labeling Approach

Though the labeling theorists have attended to the im-

pact of the label on the self-concept, they have given less

theoretical and empirical attention to social-psychological

constructs that would specify who will and will not succumb

to the label. The major emphasis has been to stipulate the

situational conditions under which labeling is likely to be

successful. For example, Lofland suggests that "when low

education and youth or old age are combined in the same

Actor it seems apparent that he becomes especially

vulnerable to escalation [to a deviant identity]" (1969:181).

Likewise, Becker (1963:17) notes that by and large the poor,

racial minorities, women, and the young in American society

tend to have rules imposed on them by the more powerful

segments of society--that is, are more susceptible to being

labeled deviant. However, a critical problem remains: How

can the labeling approach account for differential re-

sponses to similar labeling conditions? Though this is a

legitimate concern and one to which the approach must neces-

.sarily address itself, it is not an indictment of the

approach per se. The point is that the approach has empha-

sized structural and situational factors which instigate

successful labeling and has payed less attention to social-

psychological constructs that might explain variations in

response to like labeling circumstances.









Differentials in Vulnerability: Attempts
at Addressing the Problem

Notwithstanding the paucity of theorizing and research

regarding social-psychological constructs that might account

for differential responses to similar labeling conditions,

there have been some noteworthy contributions for attempting

to address this problem.

One theoretical treatment employing social-psychological

constructs in an attempt to explain varying responses to

like labeling circumstances can be found in the writings of

Lofland (1969), who maintains that the escalation to deviant

identity is facilitated by a simplex cognitive system. Low

complexity (for Lofland, a small number of cognitive cate-

gories) renders one more vulnerable to the deviant imputa-

tions of others, and conversely, persons are less vulner-

able if they are high in cognitive complexity. "They will

have less doubt about themselves in response to negative

imputations by Others" (1969:179). Though Lofland's intent

is not explicit, it appears that he has construed cognitive

complexity as making one less vulnerable to the responses

of others since complexity is positively related to educa-

tion, and education is inversely related to rates of devi-

ance according to official statistics. Using education as

an indicator of social class, he concluded that "[t]he well-

known differentials among the social classes in rates of

officially recorded apprehension and conviction or commit-

ment can be at least partially understood in these terms"










(1973:179). The association of low cognitive complexity

with responsiveness to others' appraisals is contrary to

the vast amount of research findings regarding complexity

and responsiveness (Ziller, 1973; Thompson, 1966; Ziller et

al., 1964). Since Lofland has neither operationalized the

concept nor empirically tested his hypothesis, his assertion

remains conjectural.

Goffman (1961a) tended to emphasize situational factors

in identity transformation in what has come to be known as

"psychologically compelling" situations (Middlebrook, 1974).

Goffman wrote:

Persons who become mental-hospital patients vary
widely in the kind and degree of illness that a
psychiatrist would impute to them, and in the attri-
butes by which laymen would describe them. But
once started on the way, they are confronted by
some importantly similar circumstances and respond
to these in some importantly similar ways. Since
these similarities do not come from mental illness,
they would seem to occur in spite of it. It is
thus, a tribute to the power of social forces
(societal reaction) that the uniform status of men-
tal patient can not only assure an aggregate of
persons a common character, but that this social
reworking can be done upon what is perhaps the most
obstinate diversity of human materials that can be
brought together by society [parenthesis and empha-
sis mine]. (1961a:129)

Though Goffman's emphasis here is on the external con-

ditions instigating self-reconstitution, he does not abandon

the possibility of variability in response altogether. In

the context of this discussion, Goffman seems to posit the

notion that resistance to labeling can come from conflicting

situational definitions by others (Goffman, 1959:254) or









when the actor holds multiple self-definitions (Goffman,

1961b:132). The idea of "a simultaneous multiplicity of

selves" is a social-psychological concept which could

account for variability in response to like labeling condi-

tions. Goffman implies that this concept serves as a poten-

tial insulator from others' appraisals. An alternative

view is that multiple self-definitions are associated with

self-uncertainty and hence, with vulnerability.

The concept of self-uncertainty has been employed by

Schwartz and Stryker (1970). They define the construct as

being the number and degree of variability in one's concep-

tions of self. They associate self-uncertainty with lack

of "commitment" which instigates receptivity of others'

appraisals. Lofland's cognitive complexity and Goffman's

idea of multiple self-definitions appear to bear similari-

ties at the abstract level to the "uncertainty" notion.

However, they associate their respective concepts with lack

of responsiveness while Schwartz and Stryker associate varia-

tions in conceptions of self with vulnerability to outside

appraisals. Since these theorists are positing social-

psychological constructs that could account for variations

in response to similar labeling situations, the conflicting

hypotheses need to be resolved empirically.

Becker (1963), another theorist who has attempted to

account for differences in response to similar labeling,

invoked the notion of choice when he wrote:









Apprehension may not lead to increasing deviance if
the situation in which the individual is apprehended
for the first time occurs at a point where he can
still choose between alternative lines of action.
Faced for the first time, with the possible ultimate
and drastic consequences of what he is doing, he may
decide that he does not want to take the deviant
road, and turn back. If he makes the right choice,
he will be welcomed back into the conventional
community; but if he makes the wrong move, he will
be rejected and start a cycle of increasing devi-
ance. (1963:37)

Like Lofland and Goffman, Becker's theoretical bent

seems to shun, at least tentatively, the kind of operation-

alization that would render his theoretic statements empiri-

cally testable in terms of the Gibbsian imagery of theory

construction.

Reckless et al. (1957) supplied an early attempt to

empirically test a framework that incorporated social-

psychological constructs in accounting for differences in

vulnerability. More specifically, Reckless et al. were

interested in the self-concept as a potential insulator

against juvenile delinquency in high delinquency areas. In

this study, the researchers were able to discriminate poten-

tial delinquents and nondelinquents on the basis of a par-

ticular dimension of the self-concept, namely, the extent

to which one views himself as engaging in or preferring

socially acceptable attitudes and behaviors. The authors

noted that ". a socially appropriate or inappropriate

concept of self and other is the basic component that steers

the youthful person away from or toward delinquency ."

(1957:569).









In a later work, Reckless and Dinitz (1967) reaffirmed

the efficacy of the "socially desirable" self-concept as an

insulator against delinquent influences and, as a reasonable

inference, from negative social labels. They concluded:

In our quest to discover what insulates a boy
against delinquency in a high delinquency area, we
believe we have some tangible evidence that a good
self-concept, undoubtedly a product of favorable
socialization, veers slum boys away from delin-
quency, while a poor self-concept, a product of
unfavorable socialization, gives the slum boy no
resistance to deviancy, delinquent companions, or
delinquent subculture. We feel that component of
the self-strength, such as a favorable concept of
self, acts as an inner buffer or inner containment
against deviance, distraction, lure, and pressures.
(1967:523)

Schwartz and Stryker (1970) have developed a conceptual

framework dealing with variations to like labeling condi-

tions. Their theoretical discussion and empirical investi-

gation is one of the most elaborate to date. They employed

the semantic differential technique developed by Osgood,

Suci, and Tannenbaum (1957) to measure conceptions of self.

Of the four dimensions of the self extracted by factor

analyzing a list of self-regarding adjectives, the evalua-

tive dimension was held to index self-esteem. They found

that it was not only those who were high in self-esteem that

tended to be insulated from delinquency, but also those who

showed a low degree of variability on esteem and other self-

dimensions such as potency, activity, and interpersonal

quality.









Schwartz and Stryker (1970) maintained that commitment

to a nondeviant identity reduces the probability of one's

becoming deviant. Commitment is assumed to reflect the de-

gree of certainty one has about a particular identity. There-

fore, it is self-certainty (among other things) which insu-

lates one from the delinquent influences of peers and hence

from a delinquent identity. Conversely, it is those who are

less certain regarding their self-conceptions who are more

prone to the delinquent influences of their peers in high

delinquency areas. Self-uncertainty indicates responsive-

ness to others. They wrote:

Shifting our attention now to the specifics of the
comparison of "good" and "bad" boys, all of whom
are lower-class, our argument, to repeat, is that
"good" boys ought to show lower levels of intra-
individual variability or self-certainty because
they are more committed to their good-boy role
(or conversely, less committed to deviant roles)
than are the "bad" boys, who are vulnerable to
deviance precisely because their search for a
stable identity may yet go in either direction
[emphasis mine]. (1970:87)

Self-uncertainty is measured by the degree of variability

one expresses regarding conceptions of self. The less the

variability (measured by the standard deviation of responses

to adjectives relating to the self), the greater the self-

certainty, and conversely, the greater the variation in

self-regarding attitudes, the greater the self-uncertainty.

Schwartz and Stryker's method of measuring self-

uncertainty is similar to Ziller's (1973) self-complexity

scale. Ziller's measure consists of a checklist of









adjectives relating to the self. The greater the number of

adjectives one uses to describe self, the greater the self-

complexity. As previously mentioned, both Ziller and

Schwartz and Stryker agree on the implications of com-

plexity, that is, Ziller's "complexity" and Schwartz and

Stryker's "uncertainty" are considered to be indicative of

one's greater vulnerability to outside definitions of self.

It is this hypothesis that is employed in the present study.

This overview has served to highlight a shortcoming in

the labeling approach and to call attention to some attempts

to remedy the problem. It is felt that the findings re-

garding the interrelationships among perceptions of self and

others from the formal theory of this study provide insights

which can add to the understanding gained by previous

attempts to remedy the shortcoming.

In the literature review in the extrinsic part of the

formal theory, theoretical and empirical support was found

for the usage of two self-dimensions of the formal theory,

namely, self-esteem and self-complexity. Previous theory

and research supported the association of self-esteem with

field-independence or lack of responsiveness to others'

appraisals while self-complexity was associated with re-

sponsiveness to definitions outside self. The findings from

the formalized theory provided further support for the above

usage of the respective concepts.









On the basis of the above discussion, "partial correla-

tions" were used to examine the data of this study, and to

explore the hypothesis that the influence of the perceived

alcoholic evaluations of significant others on the subject's

alcoholic self-view is contingent on a two-variable pattern

of esteem and complexity. More specifically, subjects who

were low in self-esteem and high in self-complexity com-

prised those who were high in vulnerability. And, those who

were high in esteem and low in complexity constituted those

who were low in vulnerability. These low and high categories

of the two-variable pattern served as the "test factor" in

the elaboration process. The variables that were correlated

within each category of vulnerability were the perceived

alcoholic appraisals of significant others and alcoholic

self-conception. If it could be shown that the relationship

between the perceived alcoholic evaluations of others and

alcoholic self-concept is contingent on the two-construct

pattern of vulnerability, then support for the thrust of

the framework would be provided.

In accordance with the analysis procedure delineated

above, elaboration analysis yielded the results shown in

Table 7. As may be seen, the relationship between the per-

ceived alcoholic appraisals of others and subject's alco-

holic self-concept within the high vulnerability category

is .70, as measured by gamma. The influence of the per-

ceived appraisals of others on alcoholic self-view is










Table 7. Relationships Between Perceived Alcoholic Evalua-
tions of Significant Others and Alcoholic Self-
Concept Within the Two Categories of Vulnerability

High Vulnerability Category


ALCOSELF


ALCOSELF


FRIENDS
CnUNT I
ROw PCT I
COIL PCT I
TOT PCT I 1.001 .2.001
--.-..---I-----.---I---....
1.00 I 10 I 12 I
I 05.5 I 5 ,5 I
I 18,9 I 16,0 I
I 7.8 I 9,4 I
-I -------- -------I
2.00 I 41 I 6 I
I 87.2 I 1?,8 I
I 77.4 I 8P I
I 32.0 I 07 I
-I -P.M -- ----.- I
3.00 I .2 I 57 I
I 3.a I 96.6 I
I 3.8 I 76.0 I
I 1.6 T 1a.5 I
-I------.. T--"-".-I
COiuIMN 53 75
TOTAL 41. 58,6

Gamma = .70.


Low Vulnerability Category

FRHIEND3
COUNT I
ROW PCT I
COl PCT I
TOT PCT I .00I .2001
1.00 I 8 I 11 I
I a2.1 I 57.9 I
1 53.3 I 37.9 I
I 18.2 I 25.0 .1
-I -------- T ------. I
2.00 I 7 I 17 I
I 29.2 I 70.8 I
I 46.7 T 58.6 I
I 15.'? T 38.6 I
-I- --- --..I. .. -I
3.00 I 0 I I
I 0.0 I 100.0 I
I 0.0 I 3,4 I
I 0.0 T 2.3 I
'"I---< ~"---- I -~- ~---1I
COI ULN 15 29
TnTAL 3u.1 65.9


Gamma = .32.


ROW
TOTAL

22
17.2



36.7


59
A6,1


128
100.0


ROW
TOTAL

19
43.2


24
54,5



2.5


44
100,0









further shown by the fact that 46.1 per cent of the subjects

scored highest on the alcoholic self-concept measure. When

the same variables were correlated within the low vulner-

ability category, the association was .32, as also is shown

in Table 7. While only 2.3 per cent scored highest on the

alcoholic self-concept measure within the low vulnerability

category, 43.2 per cent scored lowest. These data show that

subjects high in self-esteem and low in self-complexity are

relatively insulated from the perceived alcoholic evalua-

tions of significant others.

The above findings lend empirical support for the

framework. However, a problem remains. Since the theory

purports to account for differential responses to similar

labeling situations, the distributions of the labeling vari-

able (in this case, the perceived alcoholic evaluations of

significant others) should be similar for both categories

of vulnerability. As may be seen from Table 7, the distri-

bution for both the high and low vulnerability categories

are negatively skewed. Though the relative frequency dis-

tributions are not identical in terms of absolute numbers,

the distributions are similar with regard to percentages.

For the high vulnerable, the column percentages show that

41.4 per cent of the total were low on the labeling variable

while 58.6 per cent were high. For the low vulnerable, the

column percentages for the perceived labeling variable are

34.1 per cent for the low category and 65.9 for the high









category. Ideally, the distributions should be identical.

Given the working conditions of the field, it appears that

this is only attainable in a controlled experimental design.

Notwithstanding the shortcomings of the present design, the

findings from the multivariate analysis are held to be sub-

stantial support for the present framework. In other words,

variations in response to similar labeling situations are,

in large part, best explained in terms of the two self-

concept dimensions of self-esteem and self-complexity.














CHAPTER V

SUMMARY AND CONCLUSIONS


This study on alcoholism has been primarily concerned

with empirically testing a formalized theory of perceived

self-other concepts as those concepts relate to alcoholic

self-views and drinking patterns.

In social-psychological studies, the dominant approach

in accounting for the interrelationships among others'

appraisals and self-definitions has been to conceptualize

the self-concept as a unidimensional phenomenon. In this

study, a multiple component approach to the self-concept was

employed in an effort to account for the impact of the

perceived alcoholic evaluations of others on the subject's

alcoholic self-conceptions and drinking patterns.

A formalization of dimensions of the self-concept and

perceived labeling constructs, in accordance with the Gibb's

scheme of theory construction, produced eleven testable

hypotheses. The data bearing upon the zero-order hypotheses

supported the theory in every case. The findings demon-

strated how the use of various dimensions of the self-concept

could enhance the theoretical and empirical tenability of

the interactionist approach to self-development and be-

havioral patterns.









In addition to testing the zero-order hypotheses de-

rived from the formal theory, an attempt was made to em-

pirically ascertain the relative predictive ability of each

concept in the formal theory when drinking patterns was

designated as the dependent variable. Two multivariate

statistical procedures were employed. Automatic Interaction

Detector (AID) was used to ascertain the gravity of inter-

action effects among predictors. An extensive analysis of

the AID "tree" format and an examination of the alternative

predictors for each split indicated that interaction was

minimal. To check against this indication cross-products

were constructed and incorporated into Multiple Classifica-

tion Analysis (MCA), a multivariate procedure which assumes

additivity. A comparison of the R2's of AID and MCA (with-

out cross-products) showed that they were both in the range

of 43 per cent. Additionally, the predictive ability of

the concepts of the formal theory in explaining variance

of the dependent variable was essentially in the same order

of importance for both algorithms. In each procedure, the

two-construct pattern of self-esteem and self-complexity was

found to be important in reducing unexplained variance as

reflected by their partial betas.

A potentially weak link in the theory deriving from the

use of self-report data was examined. Two important vari-

ables, alcoholic self-concept and alcoholic evaluations by

others, were measured by individual perceptions. This left









open the possibility that the reported perceptions of self

and others' evaluations may reflect a single phenomenon.

To address this potential problem, multivariate analysis

was employed. Esteem and complexity were combined to yield

a two-variable pattern of vulnerability which was dichoto-

mized and designated as the "test factor" in the elabora-

tion process. More specifically, those low in esteem and

high in complexity were the high vulnerable, and con-

versely, those high in esteem and low in complexity were the

low vulnerable. This allowed exploration of the hypothesis

that there would be a differential impact by the perceived

alcoholic evaluations of friends and closest person on

alcoholic self-conception according to differing degrees of

vulnerability. The hypothesis was confirmed by the data

which reduced the suspicion of tautology considerably. The

differential influence on alcoholic self-conception by the

perceived alcoholic evaluations of friends and closest per-

son under varying conditions of vulnerability is evidence

that self-perceptions and perceptions of others are, at

least in part, separate variables.

A search of the literature showed that much of the

theory and research dealing with the interrelationships

among perceptions of self and others stemmed from labeling

theory as an approach to deviance. A review of that litera-

ture indicated that the major emphasis traditionally, has

been to stipulate the situational conditions under which

labeling is likely to be successful. Though this approach









has effectively accounted for deviant self-development and

behavioral patterns, it has been less than adequate in ex-

plaining differential responses to like labeling conditions--

a problem of additional concern in the present study. It

was proposed that a conceptual framework of selected self-

other perceptions deriving from the formal theory might

provide a viable means for addressing this problem when

analyzed within a multivariate design.

Support for the framework came from data showing that

the influence of the perceived alcoholic evaluations of

significant others on subject's alcoholic self-concept was

contingent upon relative degrees of esteem and complexity.

Esteem and complexity were combined into a two-variable

pattern of vulnerability. Dichotomizing the vulnerability

pattern served as the test factor in the elaborating pro-

cess. Correlating the perceived alcoholic evaluations of

significant others with alcoholic self-concept within the

low and high categories of vulnerability yielded strong

support for the conceptual framework. In the high vulner-

ability category the association between the two variables

was .70, as measured by gamma. In the low category, the

association was .32, indicating that subjects high in esteem

and low in complexity were relatively insulated from the

perceived alcoholic appraisals of others. Though strong

support was found for the framework, a problem remained.

The framework, at the outset, purported to explain









differences in response to similar labeling situations;

therefore, the distributions of perceived labeling within

each category of vulnerability should be identical. Though

they were not identical, they were similar. Both distribu-

tions were negatively skewed and the column percentages

were similar enough between categories of vulnerability to

accept the data as supportive of the conceptual formulation.


Suggestions for Future Research

In this study, some difficulties arose which suggested

improvements for future studies in the area of alcoholism,

particularly when these studies deal with self-other con-

structs.

One fundamental problem stems from the potentially con-

founding factor of social desirability response set when

self-dimensions such as esteem are measured with verbal de-

vices. This difficulty has been noted by Ziller (1973). If

a subject has a psychological readiness to respond in a

socially desirable manner, no known measurement device

couched in language can preclude it if the subject can

ascertain, either from the content of the questions or the

interview situation, what constitutes a socially desirable

response. The measure of self-esteem in this study is

particularly troublesome in this regard. Though one of the

major contentions of this study is that persons considered

alcoholic have been wrongly characterized as uniformly low

on this self-dimension, the negatively skewed distribution









of the esteem variable in this study would appear to be a

distortion in the opposite direction and may be an artifact

of response set. In light of this problem, Ziller's (1973)

nonverbal approach to measuring esteem would be an improve-

ment over the present measure for future research.

Another measurement difficulty stems from the small

number of items used to measure the constructs. For ex-

ample, the measure of self-complexity was not inclusive of

the many self-dimensions one might conceive of. The paucity

of items may not have fully discriminated the simplex from

the complex person. Subjects scoring somewhat high on a

7-item scale could conceivably be quite low on a more en-

compassing scale. This could be remedied by increasing the

number of items which represent a number of different self-

dimensions and different degrees of the same dimension.

Most probably, variability between and degrees of vari-

ability on dimensions of the self comprises the optimal con-

ceptualization of complexity. Ziller's (1973) measure of

complexity is recommended for future research since his

scale is constituted of a far greater number of dimensions

and different extremes on similar dimensions than the mea-

sure of this study.

Another methodological difficulty relates to the poten-

tial tautological problem addressed in Chapter III.

Ideally, when testing a theory such as the one in this

study, data regarding others' evaluations should be secured









from objective others. If this had been done, the possi-

bility of tautology could have been reduced considerably.

The lack of longitudinal data introduces an additional

problem. One of the concerns of this study was to test a

framework which would attempt to account for variations in

response to similar labeling conditions. This framework

stipulates that the self-dimensions which insulate one from,

or make one vulnerable to, the appraisals of others precede

others' appraisals in time. Therefore, longitudinal data

would allow for a more adequate testing of the framework.

The concepts of the paradigm for addressing the problem

of differential responses to similar labeling situations

came from the formal theory. In addition to the lack of

time-ordered data, shortcomings relating to the framework

itself warrant a brief discussion, particularly insofar as

these shortcomings suggest leads for future research.

In the testing of the framework, direct evidence came

from the elaboration analysis in which the relationship be-

tween the perceived alcoholic evaluations of others and

alcoholic self-concept was shown to be contingent upon

relative degrees of vulnerability. In the Gibbsian scheme,

this comes under the rubric of a third-order theory. Havens

and Gibbs wrote:

Variation of the empirical relation between two
variables may be due to measurement errors or
simply idiosyncratic factors, but there is always
the possibility that the relation is contingent on
other variables and hence varies in a systematic,
predictive way. If such a "contingent" variable









can be identified, it is possible to formulate a
third-order theory (Gibbs, 1972), meaning, one
which asserts that the relation between two vari-
ables, Y and Z, depends on the magnitude of or
variance in a third variable, X. (1975:261-262)

Following Gibbs (1972), the present attempt to account

for unlike responses to like labeling situations can be re-

stated into a third-order theory. Further, the third-order

theory can be stated as a fourth-order. The third- and

fourth-order theories are shown in the pictograph on page

83. The attempt, with the present framework, to account for

differential responses to like labeling conditions corre-

sponds to the fourth-order theory in the pictograph. Though

the framework examined in Chapter IV was empirically sup-

ported by the findings from elaboration analysis, it was not

stated formally. A formalization and empirical testing of

these concepts with longitudinal data suggest an area for

future study.


Generalizability of the Present Study

In the data and method discussion in Chapter II, it

was mentioned that the data for this study were not drawn on

a probability basis from any specified population; there-

fore, inferences beyond the studied subjects are statis-

tically inappropriate. However, this does not suggest that

the subjects used in this study are so unique that reason-

able generalizations beyond them cannot be made as long as

the nonprobability nature of the subjects is kept in mind.








83
































41:
C6
44 -
rl I I 8





















u :3.
ceu
> V
0 U L) L. ) >
tf)i
cc 93 U 0d








o br









In Chapter III, it was suggested that to the extent

that the studied subjects are similar in selected socio-

demographic characteristics to other known alcoholic popula-

tions, tentative generalizations could be made. In Appendix

I a comparison of selected background variables between the

studied population and a population of persons who under-

went treatment in Brevard County Mental Health Center during

1974 is made. With regard to age, it may be seen that in

both populations, the largest percentage of subjects occur

in the age category ranging from 30 to 44 years. The

characteristic, marital status, shows that in both popula-

tions the largest percentage of the subjects were married.

In terms of sex, males outnumbered females 3 to 1. Finally,

with regard to education, most of the subjects in both popu-

lations were high school graduates.

A comparison between the studied population and another

population of persons considered alcoholic and who under-

went treatment in a rehabilitation center in 1974, shows

that there is a good deal of similarity between the studied

population (which included subjects from Lake, Polk and

Sumter counties) and the population of Brevard County in

terms of socio-demographic characteristics. This similarity

provides a reasonable basis for generalizing the findings of

this study to other treatment populations.










A Concluding Remark

In these last paragraphs, an effort is made to trace

the way in which the study unfolded and the intellectual

process that accompanied it. Also, the implications of

this study for students of alcoholism are drawn. Finally,

it is suggested that the findings from this study offer an

alternative to a tendency in much of labeling theory to, at

least implicitly, view man in deterministic terms.

The most fundamental thesis upon which the study began

was that the self was an important concept in accounting for

alcoholism. It was felt that one of the main reasons people

drink excessively was because they see themselves as someone

who is an alcoholic. Taking the interactionist position, it

was also felt that other people's opinions were critical in

determining the extent to which people hold alcoholic self-

views. However, experience indicates that all persons

assumed to be alcoholic are not the same. In particular,

they do not all have similar alcoholic self-views, nor do

they all drink equally excessively. This led to thinking

about other dimensions of the self, such as self-esteem,

hoping to add to an understanding of the complex interrela-

tionships among self-other perceptions and behavioral pat-

terns. It is believed that the theory which emerged has

increased this understanding. Accordingly, the predictive

ability of the theory in a multivariate situation points

to the utility of frameworks employing the self-concept in









accounting for differential drinking patterns. The demon-

strated importance of the self-concept serves as a standing

corrective to approaches purporting to account for alcoholic

phenomena with socio-demographic variables to the exclusion

of social-psychological constructs.

Reading and thinking about the different self-other

constructs that might increase understanding about the web

of associations between self and others drew attention to

the problem of accounting for unlike responses to like

negative labels from others. As previously indicated, there

were shortcomings in the paradigm that emerged. But, it is

believed that there is tangible evidence that the combina-

tion of self-esteem and self-complexity accounts, at least

in part, for why some people do and others do not develop

alcoholic self-views in response to being branded alcoholic

by others.

In view of the preceding paragraph, a word about the

status of the labeling approach as a paradigm for explaining

deviant behavior in general is in order. The inability of

labeling theory to account for unlike responses to like

labeling situations, suggests that in much of labeling

theory, there is an implicit determinism. Schwartz and

Stryker put it this way:

That one can view another as nasty and yet behave
toward the other as though he were pleasant seems
not to be recognized. Nor is much said about the
conditions under which individuals do and do not
accept labels attached to them by others: there
is reason to believe that, at least in part, they









can psychologically control the effectiveness of
attempts to label them. The apparent determinism
of labeling theory is challenged by reference group
theory and symbolic interaction theory, all
of which force us to see people as active partici-
pants in the process by which they become "labeled."
(1970:16)

To add to the theories mentioned by Schwartz and

Stryker, a multidimensional self-theory is felt to challenge

the implicit determinism in much of labeling theory. Adop-

tion of such a formulation may, at least in part, account

for the resistance to labeling which is associated with a

less deterministic view of man. Stated differently, in the

proposed formulation an active resistance to the negative

labels of others is associated with those dimensions of the

self-concept that serve to insulate one from the personally

devastating imputations of others.






























APPENDICES













APPENDIX I


SELECTED SOCIODEMOGRAPHIC CHARACTERISTICS OF THE STUDIED
POPULATION COMPARED TO BREVARD COUNTY MENTAL HEALTH CENTER
ALCOHOLISM POPULATION BY PERCENTAGE


Brevard Studied
Characteristics C.M.H.C. Population


Under 30
30-44
45-60
60 +


9.9%
45.8
37.8
5.4
100.0%


10.8%
44.9
39.4
4.9
100.0%


Marital Status


Single
Married
Divorced
Separated
Widowed


Sex

Male
Female


Education

Grade school or less
Some high school
High school graduate
Some college
College graduate
Graduate school


6.3%
50.0
20.5
16.1
7.1
100.9%


7.4%
52.9
21.4
13.1
5.2
100.0%


75.7%
24.3
100.0%


71.3%
28.7
100.0%


13.5%
17.1
37.8
18.0
8.1
5.4
100.0%


17.4%
13.1
39.0
16.8
7.1
6.6
100.0%













4 I 4 I- lN F).. .0 00
t V) ) ~ j M 010 co
In.~ N '0* 0. 0.
C1- 0 3 V -in cNJ0

t-.


01 ON I



no o

F- ct oo l CISCO c ol ocoo
I N I 1 I 4 c 1








I --- -- -- -- -- -- --
C, I I I I I
U) c

H I tlnC0ICOOCICoolO~C0 ~
O- I I( I I C l
10




I I 0 I K S4 1
>I c I IN IN
CA
















U) I -'qN
PLO


I4 0.41 0f~~ JN i..I

> 00
MION. I- -.. I *rCJ I -.l I4
W hi r I I IC (

0 1 ONN0I I nn I q -M I UNQ 1 4
*W I I I I

wq 1- I I-- I --- I -0 I














uj

W-IT I -cc I ?-COQ .- I cool C N01 0 C
v I '1 I. I 4C 0 L>



















_j I A I
PQ vv I
I-I 0- 71 I




z Iy I IO71~ c I 1 I P-C~~I L




















p4 U)I-
x OU1t I r.I Ii I r

Ic Ot- I? .' I I *-01 I -
.1 I I I I


0) I)
01 IU I
E4 ~~-rl O iI'. 1 3~ *I .1 .. 2i*r
F9 V) J ,-- I 4qJ I N I N I (( I -i)













& v, ~~~~Lit ~ c
I-&~(N I It IC I

+4

OI I 7)P ( I I ( I

*/C I -io I o 0I0 0 1 0 0I.r

HI-I~C a. I.0I CC 01C0
p.4 :- I II

I--I Ic I Or I nc I Icl 0 (
H 0 0

0 C3
~ 4 ~.. 0 40I
LUCCI~U CA-






U) c
O L









91



















_0 N N 0





C, 0* o *




I









OO
*II 1 I IN C











I I



I II
II I
S 2 ** ** 1 o U
I Q 0 o C, O I c- I









SI I I



















F-I I I
LE IN- O'I. I .. *i I J
I *I I '1Ij o





I 41 1 I IU I
0 I I I I

oI Z'* I ***I ***I *0 -




I I I' I 0 -I .-. t I
I I I I I

UI I I
I f I



01 I I
-.N I '" I I- -u


Iu U)- pj *




CI 0
U I Oc.C'O I 'ICfl I I o ( );0Co) I -' c
(( .1. .I *







2UIU IC o 0 0 ZE.











IL
I-
1-

IX










92












































Ni O N. N I 40 # .... 'I0
4< 0. (e 0* 4* t o

-- 4 0




So100001 .)001 o0010000 10000 u000 I 00in

M- 0001 000 00'1 0')01 o00 0001 0j00 0




oi

S *I i ***I ** I** 0**. 1- .
I I I I I
I o I I I o I I



CO I o I

II
Si I
0O
"I 101 o iI g I 0, ) I 0 100001o 4



01
I 0 1 h I *'N I j ^ I I I ). o -







I 1 I *
I I I


n 0* -n I In i I *r I -1 -







0 ( I I I



LL o II l





t .n rfI u o u O o < *
I a I N I M I N I MM I N I I E








0 CCC I C 0 000I I- CiO I I
1 I I *I I *

I 04 I .. I -- I f I I -








IU I. ID o 10 I I I I -




0 CllO 0 U 0 0 0 0 014

1u00 0o a 0 0 0 ZU







t-
0

VI









93







































LU, CN V, V C NO
I'O -- -



; I I I I I


Of
Soo I 00 o o O0~ -iooo I -.o I i.lo'o I o oo -"I


S* ** I
SI 'I

000 11 0I I
01 I I I I


411 1 I NW I NN I ON I I I
o 0000 iOO IN..0 u o IJOt -- O -0coI 'Io '0i VO
ooo mOtI.l 0041 *of'*l N-1o ,1(0 0ooo 11.l
S I i I I I I *






1 I I* *** ** I I *I *



i --I rj-j I 0 I I I



nI I Ii
r 0f *** I ** *** 0 3 10 *** *- 0 *I.


NI 1 -I 'e l
I, I .. .. .. .. Io N I ........ N...o o I '... .-1






u I o
o ** .o *. Jr
-I -i O f n-I -ii In
CoII I I I I I I I


..I I I- I. I
g IC 4 p INNO IN StCl -f O2 'I NU ON 1 O I

0 000 -CrI OOCI COC- 00l ICC 000I 0. I
IU 0 I I

0 I I i i i i





LI I oI





-j




U.