University of Southern California Dissertations and Theses

An Investigation Of The Spoken Language Of Chronic Schizophrenics As A Function Of Behavioral Adjustment

(USC Thesis Other)

An Investigation Of The Spoken Language Of Chronic Schizophrenics As A Function Of Behavioral Adjustment

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content This dissertation has been
microfilmed exactly as received 66-5493
REID, P h ilip , 1 9 3 0 -
AN INVESTIGATION O F THE SPOKEN LANGUAGE
O F CHRONIC SCHIZOPHRENICS AS A FUNCTION
O F BEHAVIORAL ADJUSTM ENT.
U n iv e r sity o f Southern C aliforn ia, P h .D ., 1966
Sp eech
University Microfilms, Inc., Ann Arbor, Michigan
AN INVESTIGATION OF THE SPOKEN LANGUAGE OF
CHRONIC SCHIZOPHRENICS AS A FUNCTION OF
BEHAVIORAL ADJUSTMENT
by
Philip Reid
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Speech)
January 1966
UNIVERSITY O F S O U T H E R N CA LIFO RN IA
THE GRA DUATE SCH O O L
U N IV ER SITY PARK
L O S A N G ELES, C A LIFO R N IA 9 0 0 0 7
This dissertation, written by
...............PMLip„Reid„................
under the direction of Ms Dissertation Com
mittee, and approved by all its members-, has
been presented to and accepted by the Graduate
School, in partial fulfillment of requirements
for the degree of
D O C T O R OF P H IL O S O P H Y
Dean
Date £anuar.yx_.1.966
ACKNOWLEDGEMENTS
The author wishes to extend his sincerest gratitude
to the many professional people without whose support this
study could not have been completed.
The Veterans Administration Hospital at Sepulveda,
California, provided the facilities and subjects to carry
out this study. A special word of gratitude is due
Dr. Harold J. Segel, Clinical Psychologist, for :his fre
quent informal comments at various stages of the study.
Thanks is given also to Dr. John H. McCormack, Chief,
Department of Psychology, Dr. Anthony J. Brunse, Chief,
Department of Psychiatry, Dr. Tibor J. Nagy, Chief, Depart
ment of Physical Medicine and Rehabilitation Services, Dr.
Edward F. Gocka, Chief Systems Research Analyzer of the-
Western Research Support Center, and Mrs. Betty Abramson,
Medical Librarian.
TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS........................................ ii
LIST OF T A B L E S ............. v
Chapter
I. INTRODUCTION............................ . . 1
Statement of the Problem
Questions and Hypotheses
Importance of the Study
Remainder of the Study
II. REVIEW OF LITERATURE . . . . ................ 7
The Quantification of the Structure of
Schizophrenic Language
Schizophrenic Rigidity and Verbal Behavior
III. METHODS' AND PROCEDURES........................ 16
Experimental Design
Subjects
Establishing Group Parameters
Statistical Treatment of Data
IV. RESULTS AND DISCUSSION.........................32
Results
Discussion
Chapter Page
V. SUMMARY, CONCLUSIONS AND IMPLICATIONS .... 44
Summary
Conclusions
Implications
APPENDIXES............................................ 52
BIBLIOGRAPHY............................... 64
LIST OF TABLES
Table Page
1. Table Showing Means, F Scores, and Significance
Levels of Differences for Group L (Low in
Behavioral Adjustment) and Group H (High in
Behavioral Adjustment) in TTR Variables Before
and After Adjusting Groups for Intelligence . 33
2. Table Showing Matrix of Intercorrelations of
all Measures of all 59 Subjects................ 35
3. Table of Means, Standard Deviations, Standard
Error of Means, Maximum and Minimum Scores
and Ranges of all Variables on all 59
Subjects..........................................36
4. Table Showing Means, F Scores and Significance
Levels of Differences for Group L (Low in
Behavioral Adjustment) And Group H (High in
Behavioral Adjustment) in Grammatical Cate
gory Variables Before and After Adjusting
Groups for Intelligence ....................... 38
v
CHAPTER I
INTRODUCTION
In studying the literature concerning comparisons
of the language patterns of psychotic, neurotic and normal
persons, the relationship of verbal utterance and overt
behavioral adjustment emerged as an area meriting further
investigation. It was felt that the analysis of spoken
language, when related to degree of behavioral adjustment,
could yield useful clinical insights. If significant re
lationships could be shown to exist, further impetus would
be given to the application of. linguistic techniques to the
study of mental illness. The basis for the present study
is the research accomplished by previous investigators re
garding relationships of the spoken language of various
clinical groups.
Psychiatrists and clinical psychologists have
attempted to differentiate the mentally disturbed individ
ual on the basis of form and content of spoken language.
Lorenz and Cobb (25, 26, 27) have demonstrated that various
clinical groups have characteristic thematic and structural
features in their speech. They cite as an example the
spoken language of the manic individual which is character
ized by verbosity and variability while the language of the
depressive is frequently marked by relative moroseness and
tends to be curt and pithy.
The two processes, language and thought, although
not identical, still are closely related. Rabin and King
(33) referred to language as the "instrument for the ex
pression of thought." Further supporting this point, New
man and Mather (32) and Sanford (37) suggest that verbal
expression is not an autonomous category of behavior but
rather one of several behavioral forms providing external
symbols of human functioning. Sanford, reflecting the
attitude of other researchers, cited the need for more
extensive study in the covarying aspects of verbal expres
sion and behavioral adjustment.
Statement of the Problem
In order to establish the relationship between
spoken language and behavioral adjustment, schizophrenia
was chosen for investigation because it embodies idiosyn-
eratic language, disturbances of thought and deviant be
havioral adjustment. The intent of this study was to in
vestigate the relationship between degrees of behavioral
adjustment and degrees of verbal rigidity in chronic
schizophrenics in order to evaluate language and schizo
phrenia. Verbal rigidity was defined.as the number of dif
ferent words employed and as the variety of frequencies of
grammatical constructions in elicited samples of spoken
language.
Although not patently stated, the results of studies
by Fairbanks (13), Mann (3), Feldstein and Jaffe (14) and
Hammer and Salzinger (21) suggest th'e possibility of util
izing Type Token Ratio (TTR) as a means of measuring verbal
rigidity. This method is discussed in detail on page 22.
The two methods of TTR most used in other studies of lan
guage structure are the "Overall TTR" and the "Segmental
TTR." The Overall TTR indicates the number of different
words or types in a language sample containing a specified
number of words, or tokens. The Segmental TTR shows the
ratio of types to tokens within equally fractionated seg
ments of a total language sample.
Hypotheses regarding group differences in variabil
ity in word utterances and usage of grammatical construe-
tions were suggested by Fairbanks (13) , Lorenz and Cobb
(28) and Ruesch (36). Their evidence and speculation
suggest that when comparing spoken language samples of a
group of schizophrenics and a group of normals, one would
find that the schizophrenics employ more adjectives, ad
verbs, verbs, pronouns and self reference words and fewer
nouns than the normals. Because of the obvious diversity in
behavioral adjustment between groups of schizophrenics and
groups of normals and because of the wide continuum of
behavioral adjustment which could exist within the non
specific group called "normals," it was assumed by the
present investigator that an investigation of spoken lan
guage of the single clinical group of schizophrenia,
divided in terms of behavioral adjustment, would yield
additional useful information concerning the relation be
tween language and adjustment. Further, it seemed reason
able to assume that differences in usage of grammatical
categories between these groups would take the same direc
tion as those reported by the investigators cited above.
Questions and Hypotheses
Arising from the general problem of the relation
ship between language and adjustment were two basic ques-
tions and three hypotheses. The questions were: (1) Does
the flexibility of word utterance of the chronic schizo
phrenic vary as a function of behavioral adjustment?
(2) Do certain grammatical constructions used by the chron
ic schizophrenic vary as a function of behavioral adjust
ment?
The hypotheses were: (1) The Overall TTR will vary
as a function of the behavioral adjustment of the chronic
schizophrenic. That is, the more poorly adjusted chronic
schizophrenic as contrasted with the better adjusted schiz
ophrenic will have a lower Overall TTR. (2) The Mean
Segmental TTR will vary as a function of the behavioral
adjustment of the chronic schizophrenic. That is, the more
poorly adjusted chronic schizophrenic as contrasted with
the better adjusted schizophrenic will have a lower Mean
Segmental TTR. (3) Usage of certain grammatical construc
tions will vary as a function of the behavioral adjustment
of the chronic schizophrenic. That is, based upon the
findings and speculations of other investigators, the more
poorly adjusted chronic schizophrenic as contrasted with the
better adjusted schizophrenic will employ, in a language
sample of a specific length, (a) more adjectives, (b) more
adverbs, (c) more verbs, (d) fewer nouns, (e) more pronouns,
and (f) more self reference words.
Importance of the Study
The value of this study lies (1) in the need to
refute or confirm in part the notion that particular verbal
behavior reflects level of behavioral adjustment, (2) in
the fact that positive findings would open the door to
similar or parallel kinds of research utilizing different
linguistic categories, and (3) in the fact that such stud
ies provide objective means (e.g., clues to a behavioral
index) as opposed to arbitrary means of defining chronic
schizophrenia. Because of the vagueness of arbitrary
methods, objective means would be preferable.
Remainder of the Study
Chapter II contains a survey of the theory and re
search upon which the problem of this study is based.
Methods and procedures followed in the execution of the
study are outlined in Chapter III. Chapter IV comprises a
discussion of the results while Chapter V contains a sum
mary, conclusions and implications for related research.
CHAPTER II
REVIEW OF LITERATURE
The literature surveyed in this study was selected
from the following vantage points: (1) a review of pre
vious studies pertinent to the quantification of the struc
ture of schizophrenic language and (2) a survey of the
literature concerning the concept of schizophrenic rigidity
and its relationship to verbal behavior.
The Quantification of the Structure
of Schizophrenic Language
The use of analytic methodology to study the rela
tion of language patterns to mental attitudes and behavior
dates back to the early 1920's. Because of the two basic
types of variables under investigation in the present
study, TTR measures and grammatical construction categor
ies, this section will be divided into two segments. The
first will be given to a survey of the literature concern
ing the history and use of the Type Token Ratio as a
measure of the flexibility of schizophrenic language and
the second to a review of approaches to the study of gram
matical usage of language with particular attention to
schizophrenic language.
TTR and the Flexibility
of Spoken Language
The introduction of the TTR, an instrument which
pruports to measure the flexibility or variability of lan
guage, is credited to Zipf (42) and was expanded by John
son (24). The TTR is a quantitative measure of language
and is the proportion of different words or types out of
the total number of words or tokens. While Zipf had uti
lized this method (without labelling it) for investigating
the various attributes of language per se, Johnson applied
the TTR to the study of the relationship between personal
ity patterns and language patterns. This language measure
has subsequently been employed in studies comparing the
written and/or spoken language of many different popula
tions. A discussion follows of those studies most perti
nent to the present study. These are the studies by Fair
banks (13), Feldstein and Jaffe (14) and the works cited by
Mowrer (31).
Fairbanks, contrasting nontranquilized schizophren-
ics with superior college freshmen, found that schizo
phrenic language "differs from the language of 'normal'
persons in being . . . less highly differentiated in struc
ture— the ratio of different words (types) to total words
(tokens) is lower . . (13:37). Feldstein and Jaffe
(14), attempting "to delineate further the sensitivity of
the TTR measure," matched a group of schizophrenics with a
group of non-psychiatric patients in terms of age, educa
tion and I.Q. level. They found that their results did not
distinguish a significant vocabulary diversity between the
groups. This study differed from Fairbanks' study in that
the schizophrenic subjects in Feldstein and Jaffe's study
were receiving various tranquilizing drugs and smaller
speech segments were used for analysis (25-word segments
rather than 100-word segments). In addition, whereas
Fairbanks used fourteen proverbs to elicit a speech sample
from each subject, Feldstein and Jaffe utilized four sets
of five pictures that told a story. These methodological
differences could account for the contradictory results
between studies. Also if the intra-group behavioral ad
justment level of Feldstein and Jaffe's schizophrenic group
was high enough, this could explain the lack of differen-
' tiation between the TTR's of the schizophrenic and normal
groups.
Sanford (37), in a discussion of the concept of the
relationship between verbal utterance and other forms of
behavior, asserted that an individual's semantic habits may
well reflect the nature of his adjustment. Roshal's work
(35) lends support to the concept of vocabulary flexibility
as an index of relative behavioral adjustment. Testing her
hypothesis that successful psychotherapy would be accom
panied by an increase in vocabulary variability, Roshal
calculated the TTR for samples of the first and last inter
views of "successful" psychotherapeutic treatment cases.
She found an increase in the ratio. Assuming that there
was improvement in the behavioral adjustment of these
cases, it may be inferred that a language pattern of many
different signs with few repeated signs is indicative of a
high degree of variability and flexibility in language
usage— and greater behavioral adjustment. Conversely,
monotony and repetitiveness of language habits and less
behavioral adjustment would be reflected by a pattern of
fewer different signs with frequent recurrence of any one
sign. Similar results were reported by Grummon (19) and
Horvitz (23).
11
Grammatical Usage
Sixty-nine years ago, John Dewey (7) remarked that
scholars knew very little about the psychology of an adverb
or a noun or other parts of speech. A significant interest
in the area of psychogrammar was first demonstrated in the
early 1920's. Busemann (5) employed his verb/adjective
ratio in an attempt to evaluate emotional stability. He
used the number of verbs in a speech sample as the numer
ator and the number of adjectives as the denominator.
Busemann asserted that this yielded an "action quotient"
and that the presence of a small number of verbs, and con
sequently a low ratio, is indicative of emotional stabil
ity. Boder (3) simply reversed the items in the quotient,
calling his measure the adjective/verb ratio. He did not
attempt to relate the adjective/verb quotient to emotional
stability but used it to investigate linguistic aspects of
literature.
Various researchers have discussed the grammatical
imbalance of schizophrenic language structure as compared
with that of normals. Fairbanks (13) reported a greater
number of verbs and pronouns and fewer nouns and articles
in the speech of schizophrenics than in the spoken language
of the normal group. Lorenz and Cobb (28) also found this
12
to be true. Mabry (29), comparing the language character
istics of "scattered" and "nonscattered" schizophrenics
with normals on the basis of responses to a sentence com- •
pletion test, reported no significant differences in the
use of different parts of speech between the groups. A
"scattered" schizophrenic was one who utilized obviously
deviant language. Content of ideation was not taken into
account in making such a classification.
Ruesch (36) wrote that the schizophrenic, rather
than grouping his thoughts around nouns, emphasizes that
which is expressed by adjectives and verbs. He explained
that tendency by utilizing the concept of the developmental
arrest of the schizophrenic at the non-verbal level. In
referring to the schizophrenic, Ruesch wrote, "he experi
ences very much like an infant. More affinity to sensa
tions and movement than what it is that does this" (36:
205). He went on to explain that while the perception of
*
color, shape and movement can be experienced directly, the
identity of an object is something that has to be expressed
in words and that this is a process which can be learned
only from others. Because the schizophrenic tends to with
draw, "the predicate mode prevails in his experience and
expression" (36:205).
13
The persistent use of self reference words by-
schizophrenics has often been pointed out in the litera
ture. Lorenz (25) mentioned the frequent "reflexive" em
ployment of pronouns in schizophrenic language. Fairbanks
(13) wrote of the spoken language of the schizophrenic as
being filled with more self reference terms of the first
person singular pronoun class than is the spoken language
of the normal individual. ^ Grummon reported "a marked ten
dency for a high frequency of pronouns to be associated
with poor adjustment" (19). Poor adjustment was assumed by
Grummon to be existent in his subjects prior to psycho
therapy.
It is evident from the available literature that
conflicting evidence exists concerning the language pat
terns of schizophrenics as compared with normals. These
conflicts arise from the lack of homogeneity between stud
ies cited in terms of the characteristics of the subjects
employed and the stimuli utilized to elicit spoken lan
guage. In the present study rigid selection criteria have
been employed and language investigation has been limited
to one clinical group. Precise linguistic data from an
intensive investigation of this clinical group are needed
to help clarify the relation between the condition of
14
schizophrenia and the spoken language of the schizophrenic.
Schizophrenic Rigidity and Verbal Behavior
Many researchers have discussed the behavioral
rigidity of the schizophrenic. For example, Rashkis (34)
pointed out that the schizophrenic is notoriously unable to
vary his frame of reference. Hanfmann (22) and Goldstein
(18) discussed the concrete attitude of the schizophrenic
and his apparent inability to shift voluntarily from one
aspect of a situation to another. This impairment of
functioning emphasizes the behavioral rigidity of the
schizophrenic. Further, this behavioral rigidity may be
manifested by specific language habits. Stout (38) viewed
language as an instrument by which the subject, examining
the world about him, gives direction to his thought.
Buhler (4) specified verbal expression a major function of
language since verbally articulated thought is the mode of
expression of psychological orientation. In a study of the
language patterns of neurotics and psychotics, Lorenz and
Cobb wrote, "Habits of speech seem to reflect habits of
orientation toward the self, external objects and reality.
As a consequence, groups of persons who show certain
psychological traits strongly will, when they speak, show
15
similar tendencies in their use of language" (28:671). Some
examples of this concept are Busemann's (5) and Ruesch's
(36) notions regarding the schizophrenic's use of verbs
referred to earlier. Fromm-Reichmann (16), in an article
concerning the psychotherapy of schizophrenia, described
the schizophrenic as "self-engulfed and withdrawn" as a
result of "narcissistic regression." Weigert (40) dis
cussed the psychotic's "maintenance of the isolation of a
defensive egocentricity" due to the presence of inordinate
anxieties, pressing for discharge, which disturb the use*
of spoken language in effective communication.
The works cited above constitute a sampling of a
generally prevailing opinion both in clinical observation
and research that a close relationship obtains between
various facets of verbal behavior of schizophrenics and
verbal rigidity. Since such relationships have been sug
gested it may be inferred that some specific dimensions of
language behavior and rigidity in schizophrenia correlate.
CHAPTER III
METHODS AND PROCEDURES
Experimental Design
The design of this study is experimental. Level of
behavioral adjustment constitutes the independent variable
and is manipulated by selection. Controls to be discussed
later in this chapter were rigorously observed. Two TTR
measures and six grammatical construction categories com
prise the dependent variables. The design utilizes prior
hypotheses^ based upon the research and speculation of other
investigators. Hypotheses deal with the effects of varying
the degree of behavioral adjustment upon the language struc
ture of the subjects.
Subjects
Fifty-nine hospitalized veterans participated in
this study. The sample was drawn from the in-patient popu
lation of the Psychiatric Service of the Sepulveda Veterans
Administration Hospital. With the aid of hospital records,
16
subjects were selected on the basis of the following cri
teria: (1) All subjects were male. (2) All were Caucas
ian. (3) All were native born speakers of "American" .
English. (4) All were selected from the age range twenty-
six through fifty. This range was consistent with the age
range of the total veteran patient population at the
Sepulveda Veterans Administration Hospital at that time.
(5) All were medically diagnosed as chronic schizophrenics.
Such a diagnosis was established by a hospital staffing of
each patient. Present at each staffing were a psychia
trist, a psychologist, ward nurses and various rehabilita
tion therapists. On the basis of information presented
from these sources, the psychiatrist assigned a diagnosis.
(6) All had a history of hospitalization in a psychiatric
ward of at least one year and current hospitalization of at
least three months. (7) All were non-mute patients.
(8) No subject had experienced psychosurgery. (9) Only
those subjects were employed for whom previous neurological
examination had not indicated the presence of organicity.
(10) No subject had received electro- or insulin therapy
for six months prior to the experimental procedure.
(11) All were on maintenance levels of tranquilizing drugs
of the phenothiazine group. Maintenance level was defined
18
as the optimum level at which the patient is relatively
free from overwhelming anxiety and overt physical symptoms.
This level is determined by the patient's psychiatrist
through trial and error. When the optimum level is de
cided it is maintained for an indefinite period of time.
(12) Voluntary participation of each subject was a pre
requisite for use in this study.
Establishing Group Parameters
After a sample of sixty subjects meeting the fore
going criteria were selected, the sample was dichotomized
on the basis of the median MACC Behavioral Adjustment
scores. This instrument is attached as Appendix A and
will be described and discussed later. Because one sub
ject was found to have an extremely low behavioral adjust
ment score, this subject was eliminated from the study so
as not to bias the data. Thirty subjects comprised the
group which was low in behavioral adjustment (Group L);
the group high in behavioral adjustment (Group H) had
twenty-nine subjects.
In addition, determinations of other differences
than behavioral adjustment were considered and it was found
that no significant differences obtained between the groups
19
on the basis of age as established by a non-significant F
ratio in the main body of the statistical procedure and on
the basis of level of educational attainment as established
by _t scores using the procedure discussed by Blalock (1).
(See Appendix B.)
No effort was made to equate the groups for intel
ligence in the initial selection procedure since this would
minimize verbal differences— differences that are being
sought in this study. However, adjustment of the data was
made which indicated what results would obtain if the ef
fects of the difference on intelligence between the groups
were removed. Intelligence and verbal ability have long
been recognised as being intimately related. This may be
seen in the great degree of weight generally accorded
verbal items in tests of intelligence. For this reason
particular attention was given to the evaluation of these
variables. Language usage of the experimental groups was
compared prior and consequent to balancing the groups for
intelligence. Analysis of covariance is a statistical
technique used to neutralize sample differences in a con
comitant variable conjointly with comparing the average
values of the variable of interest. Its application con
sists of comparing the means of independent samples after
20
statistically eliminating the random sampling differences
of a variable x, concomitant to the main variable under
study. In addition, the variation in £ is reduced by the
elimination of the effect of the variation in x. within a
sample. This increases the possibility of establishing
statistical significance among the means of ^ if a true
di f ference exi s ts.
Test Used to Establish Intelligence
Levels of Groups
A conclusion reached by Fairbanks (13) was that a
positive correlation probably exists between TTR and in
tellectual level. Zipf (43) implied that such a correla
tion exists. To determine the operability of this factor
in terms of the sample under investigation in the present
study, provisions were made in the statistical design to
determine the relatedness of intellectual level and TTR and
to make necessary adjustments so that the experimental out
come would not be affected. This procedure is described
above. The Wonderlic Personnel Test was employed to es
tablish the intelligence levels of the groups.
The Wonderlic Personnel Test was originally de
signed for testing adults employed in business and indus
try. In recent years this instrument has been utilized in
21
the Veterans Administration as a rapidly administered test
of general intelligence of psychiatric patients. The gen
erality of the test may be seen in the varity of items it
employs in categories such as analogies, analysis of geo
metric figures, arithmetic problems, disarranged sentences,
similarities, logic and spatial relations.
The Wonderlic was selected as the measure of intel
ligence for this study because it is a standardized test
that may be administered with brevity. A table indicating
the distribution of the population employed by Wonderlic is
presented in the test manual (41). An N of 53,864, in
cluding all educational levels of both sexes from grade
twelve to the senior level of college showed a mean of 22.7
points out of a possible 50 points with a standard devia
tion of 7.5, indicating a normal distribution. Wonderlic
norms used for this study have been established by Giedt
(17). When the low score skewness of the distribution from
a sample of psychotics is corrected by using the method
described by Dixon and Massey (8:58) and transforming the
scores to T_ scores, the new mean is 88. This is closely
comparable with means of other tests, namely: the Wechsler
Adult Intelligence Scale, 91.; the Shipley Hartford
Vocabulary Test, 86.; and the Cattell Culture Free Test,
22
85. Brevity of the test prevented the development of
ambient reactions which is often the case when longer test
times are required. Tests were administered as part of the
hospital's psychometric program and administration and
scoring were accomplished by trained personnel. Instruc
tions in the Wonderlic Test Manual (41) were carefully
followed.
Recognizing the limitations of a brief paper and
pencil estimate of intelligence/ it was felt that the
Wonderlic Test was acceptable for defining the population
of this study. While it was necessary to control for in
telligence, intelligence per se was not the crucial vari
able under consideration. Owing to the utility found for
this test in applied situations and its aforementioned
correlation with more extensive tests of intelligence, the
Wonderlic was deemed satisfactory for use in the present
study.
Test Utilized to Establish Group
Parameters (Independent Variable)
The MACC Behavioral Adjustment Scale (MACC) was
developed by Ellsworth (11) for evaluating overt behavioral
effects observed by ward nurses and aides. (See Appendix
A.) The measure consists of sixteen five-point linear
23
scales which yield four cluster scores and which purported
ly describe the following categories of observable behav
ior: mood, cooperation, communication and social contact.
The sum of the cluster scores is the total adjustment
score. In order for a group of items to be labeled as a
"cluster," the item intercorrelations of the clustered
items must be greater within the cluster than with items
outside the cluster. This arrangement is arrived at by a
process called cluster analysis.
Ellsworth standardized the MACC on functionally
psychotic hospitalized patients. Studies of concurrent
validity, a measure of the correspondence between test
results and the present status of individuals, were under
taken by Ellsworth (10) and Ellsworth and Clayton (12).
Product moment correlations were computed in these studies.
The external criteria of these studies were (1) the ability
of the MACC to differentiate closed and open ward patients
and (2) the ability to corroborate independently diagnosed
cases of psychopathology. In the former, the MACC showed
that the open and closed groups were differentiated by a
significant score and in the latter correlation coeffic
ients of .69 and .74, both significant beyond .01, were
obtained. The Ellsworth (10) and Ellsworth and Clayton
Filmed as received
without page(s)___
UNIVERSITY MICROFILMS, INC.
25
(12) studies also reported inter- and intra-rater relia
bility coefficients all better than .80.
Within twenty-four hours of the elicitation of a
language sample of a subject, ratings of the behavioral
adjustment of that subject were accomplished by a team of
two psychiatric nurses. All nurses who participated in
this study had been trained by psychiatrists and psycholo
gists in the use of the MACC. Each team member rated the
subject independently and in order to qualify as a rater
of a particular subject, she must have had the subject in
her care for not less than three months prior to rating
him. A total of seven nurses participated in this study,
and, working in various pairings, comprised seven teams.
A relability study combinbing all pairs of rater combina
tions for all subjects yielded a correlation of .88 for the
total adjustment score. Individual inter-rater reliabili
ties were obtained for each individual scale of the MACC
and also for the total adjustment score of the MACC. These
were averaged using Fisher's (15) Z prime transformation
for each individual correlation. Since the total adjust
ment score reliability is most germane to this study only
this value is reported.
26
Methods Utilized to Elicit the Language Sample
(Dependent Variables)
The language sample of each subject was elicited by
the same experimenter with the instructions and stimulus
presented in Appendix C. Subjects were requested to ex
plain and give an example of each proverb. The physical
environment employed was a small room in which were a
table, two chairs and a tape recorder. In each case, the
experimenter sat across the table from the subject and the
tape recorder was placed at the side of the experimenter
and in full view of the subject.
During language sample elicitation, the experiment
er attempted to keep at a minimum any verbal interaction
other than that required by the instructions. In accord
ance with the instructions, questions asked by the subject
regarding the stimulus were not answered and responses
given by. the subject were not rewarded by the experimenter.
Only requests to repeat a proverb were honored. When it
became apparent to the experimenter that the subject had
completed a response, the next proverb was given. The
presentation of stimulus materials was terminated if the
subject requested this. A transcription was made from a
tape recording of the first 1000 words uttered by the sub-
27
ject. The speech sample was then submitted to analyses.
Measures Employed in Analyzing
the Language Sample
The TTR is a quantitative measure of language and,
as was mentioned previously, is the proportion of differ
ent words or types out of the total number of words or
tokens.. It is a measure of vocabulary flexibility or
variability. Carroll's research (6) demonstrated that be
cause the number of different words decreases as successive
increments are added to a language sample, the number of
tokens employed in computing the TTR must be kept constant
in order to assure comparability from one sample to
another. In the present study, fifty words are arbitrarily
used as the standard number of tokens, each language sam
ple being divided into twenty consecutive fifty word seg
ments. For purposes of this study "Mean Segmental TTR,"
the arithmetic mean of the Segmental TTR's for a given sub
ject, will be used. The Overall TTR of the 1000 word
sample was computed as was the TTR for each of the fifty
word segments and the Mean Segmental TTR for each subject.
Following division of the fifty-nine subject sample on the
basis of behavioral adjustment ratings, mean Overall TTR's
and Segmental TTR's were computed for each group.
28
The following rules, essentially the same rules
followed by Fairbanks (13) and Mann (30), were observed in
order to determine what constituted a word for purposes of
TTR analysis: (1) Whenever a particular word was used as a
different part of speech it was counted as a different
word. For example, will as a verb and will as a noun.
(2) A common noun and a properrnoun having the same spell
ing were counted as one word. (3) A hyphenated word was
tabulated as one word. Authority for proper hyphenation
was ascribed to Webster's New International Unabridged
Dictionary (39). (4) Contractions were tabulated as the
two separate words represented. For example, didn* t was
changed to did not. This was done to conform with previous
studies of a similar nature. (5) A number was counted as
one word regardless of length. (6) Abbreviations such as
M.D. or Ph.D. were tabulated as one word. (7) The tabula
tion of neologisms or coined words was accomplished accord
ing to the parts of speech each seemed to functionally
£
assume in a sentence. (8) The pause filling utterance uh
was not considered a word and was not tabulated.
The 1000 word language sample of each subject was
analyzed to determine type-frequency characteristics in
terms of grammatical constructions. Mean type-frequencies
of the parts of speech under investigation were computed
for each group. Attention was given to the following gram
matical categories: (1) adjectives; (2) adverbs;
(3) verbs; (4) nouns; (5) pronouns; and (6) self reference
words (I,. me, mine, my, myself, own name).
The rules previously presented for TTR analysis
were applied to the grammatical analysis. Identification
of parts of speech were made on the basis of the following
rules: (1) Adjectives— all regularly known adjectives and
any verb form (i.e., participle) which the dictionary (39)
recognizes as an adjective. (2) Adverbs— regular classifi
cation. (3) Verbs— simple verbs, participles plus auxil
iaries, unless the dictionary recognizes them as nouns and
adjectives, as the case may be. (4) Nouns— all regularly
known common and proper nouns and gerunds which the dic
tionary recognizes as nouns. (5) Pronouns— all personal
and indefinite forms, such as my, our, your, their, etc.
Also all demonstrative, relative, and interrogative pro
nouns such as this, those, who, whom, where, etc.
Statistical Treatment of Data
The statistical design employed was multivariate
analysis of covariance permitting the computation and test
30
ing of the means of several variables between groups.^
Utilizing this design it was possible to treat the depen
dent variables either singly or jointly. Results are ex
pressed in terms of an F ratio which is a value used in
determining whether the difference between groups is
statistically significant. This is true for both the uni
variate F obtained by considering each dependent variable
singly and the multivariate F obtained by considering
dependent variables jointly. The multivariate F ratio is
a function of the determinant of the matrix of main effects
divided by the determinant of the matrix of error variance.
The statistical method utilized also shows the important
property of covariance. This covariance property, dis
cussed earlier in this chapter, allows for the manipulation
of one or more experimental variables, otherwise generally
difficult to control. This is done on the basis of the
obtained correlations of these covariates with the depend
ent variables. The effects of these covariates are removed
^The statistical design was based upon a program
written by Hall and Cramer (20), in turn based upon the
work of Bock (2). The author acknowledges with gratitude
the assistance he received from the Western Research Sup
port Center, Sepulveda Veterans Administration Hospital,
Sepulveda, California.
statistically in the calculation of the F values.
CHAPTER IV
RESULTS AND DISCUSSION
The outcome of the statistical manipulation of the
data is shown in terms of its significance to the hypo-
these stated in Chapter I. The results will be discussed
first reviewing the TTR measures and.second the indexes
arising from the parts of speech in the subjects' proto
cols.
Results
TTR Measures
The Overall TTR, as shown in Table 1, displays a
difference between Groups L (lower behavioral adjustment)
and H (higher behavioral adjustment) significant at the 5
per cent level, Group L having the lower mean of 0.263 as
compared with 0.283 for Group H and showing an F of 4.89.
In contrast the two groups differ at the 1 per cent level
of significance in respect to the Mean Segmental TTR. In
32
TABLE 1
TABLE SHOWING MEANS, F SCORES, AND SIGNIFICANCE LEVELS OF DIFFERENCES FOR
GROUP L (LOW IN BEHAVIORAL ADJUSTMENT) AND GROUP H (HIGH IN BEHAVIORAL ADJUSTMENT)
IN TTR VARIABLES BEFORE AND AFTER ADJUSTING GROUPS FOR INTELLIGENCE
Non-Adjusted Adiusted
Variable
Means
F Score
With IQ Significance
Group L
(N = 30)
Group H
(N = 29)
F Score
CL, 57)
Significance Level
of Difference v,
Between the Groups
as Co-
variate
(1. 56)
Level
Adjusted for
IQ
Overall TTR 0.263 0.283 : 4.89 .05a 2.13 N.S.C
Mean Segmental TTR 0.65 0.69 9.52 .oib 5.46 .025
I.Q. 92.50 98.86 4.47 .05
- - - - - - - - - -
a
Call a significant difference any comparison with an F value >4.01 at the .05 level.
Call a significant difference any comparison with an F value >7.10 at'the .01 level.
c
N.S. = Non Significant
u>
u>
34
this case it may be seen that.Group L still maintains the
lower mean of 0.65 as contrasted with 0.69 for Group H,
producing an F of 9.52.
k »
When the data are adjusted to reduce the differenc
es between Group L and H in respect to intelligence by
using the covariance technique discussed in Chapter III,
the Overall TTR no longer significantly differentiates the
two groups as is indicated by an F of 2.13 and as shown in
the adjusted column of Table 1. The Mean Segmental TTR,
while it is attenuated to the .025 level, still signifi
cantly differentiates Group L and Group H. When results
are pooled and intercorrelations are drawn on all fifty-
nine subjects, a correlation between the two variables,
Overall TTR and Mean Segmental TTR, of 0.85 obtains as may
be seen in Table 2. In addition to this high correlation,
the small and closely related standard deviations of the
two variables were noticed to have high within-group homo
geneity. The test for homogeneity of variance as described
by Edwards (9:105-106) was employed to validate this obser
vation. The overall norm data for each variable is shown
in Table 3. Also, when the groups are pooled, some posi
tive correlations are seen between Overall TTR and Mean
Segmental TTR and Adjectives and Nouns. This is depicted
TABLE 2
TABLE SHOWING MATRIX OF INTERCORRELATIONS OF ALL MEASURES OF ALL 59 SUBJECTS*
Overall
TTR
Mean
Segmental
TTR
Adjec
tives
Adverbs Verbs Nouns Pronouns
Self
Reference
Words
Overall TTR
Mean
Segmental 0.8505
TTR
Adjectives 0.4999 0.4855
Adverbs -0.0474 0.0571 -0.1308
Verbs -0.6308 -0.5195 -.3752 0.1156
Nouns 0.6033 0.4332 0.3586 -0.1805 -0.7360
Pronouns -0.4413 -0.3536 -0.4013 0.0219 0.5557 -0.732
Self
Reference -0.0547 -0.0272 -0.1196 -0.0443 0.4291 -0.3788 0.5815
Words
*An r value of .257 needed for a correlation to be significantly different from .0 at the
.05 level.
u>
Ln
TABLE 3
TABLE OF MEANS, STANDARD DEVIATIONS, STANDARD ERROR OF MEANS, MAXIMUM AND MINIMUM
SCORES AND RANGES OF ALL VARIABLES ON ALL 59 SUBJECTS
Variable Mean
Standard
Deviation
Standard
Error
of Mean
Maximum Minimum Range
Overall TTR 0.2728 0.0358 0.0047 0.3890 0.1910 0.1980
Mean Segmental TTR 0.6706 0.0436 0.0057 0.7980 0.5720 0.2260
Adjectives 73.8814 13.1949 1.7178 117.0000 50.0000 67.0000
Adverbs 88.0339 14.7285 1.9175 131.0000 60.0000 71.0000
Verbs 237.6610 23.3346 3.0379 282.0000 173.0000 109.0000
Nouns 151.8136 23.6311 3.0765 226.0000 107.0000 119.0000
Pronouns 177.1695 31.6005 4.1140 276.0000 106.0000 170.0000
Self Reference Words 41.4237 25.9291 3.3757 116.0000 3.0000 113.0000
Wonderlic Scores
(I.Q.)
95.6271 11.9018 1.5495 130.0000 70.0000 60.0000
co
37
in Table 2. Specifically, Overall TTR has a correlation
of 0.50 with Adjectives and 0.60 with Nouns; Mean Segmental
TTR has a correlation of 0.49 with Adjectives and 0.43 with
Nouns. All other correlations are extremely low or nega
tive .
Grammatical Construction Categories
None of the grammatical construction categories
produced statistically significant differences between the
two groups as seen in Table 4. Furthermore, after adjust
ments were made to remove the effects of intelligence be
tween groups with the aforementioned covariation technique,
still no significant differences emerged. A trend in the
direction of significance was seen in the category Nouns
for which the F score increased from 1.85 to 3.23 between
the non-adjusted and adjusted scores.
When the groups are pooled and tabulated, the data
indicate curves approximating the normal distribution with
the exception of the category Self Reference Words which
shows a high degree of variability through its large
standard deviation. This variable served to show a large
difference between group means. However, when compared
against the large resultant standard error of the difference
TABLE 4
TABLE SHOWING MEANS, F SCORES AND SIGNIFICANCE LEVELS OF DIFFERENCES FOR
GROUP L (LOW IN BEHAVIORAL ADJUSTMENT) AND GROUP H (HIGH IN BEHAVIORAL ADJUSTMENT)
IN GRAMMATICAL CATEGORY VARIABLES BEFORE AND AFTER ADJUSTING GROUPS FOR INTELLIGENCE
Non-Adi usted Adjusted
Means
F Score
with IQ Significance
Significance Level as Co- Level
Group L Group H . F Score of Difference variate Adjusted for
(N = 30) (N = 29) (1, 57) Between the Groups (1, 56) IQ
Adjectives 73.00 74.79 0.27 N.S.3 0.01 N.S.a
Adverbs 87.10 89.00 0.24 N.S. 0.63 N.S.
Verbs 235.86 239.52 0.36 N.S. 1.25 N.S.
Nouns 155.90 147.59 1.85 N.S. 3.23 N.S.
Pronouns 176.03 178.34 0.08 N.S. 1.46 N.S.
Self Reference
Words
37.93 45.03 1.11 N.S. 2.45 N.S.
I.Q. 92.50 98.86 4.47
__b
.05
---- ----
aN.S. = Non Significant
Call a significant difference any comparison with an F value >4.01
OJ
00
39
of the means, it proved non-significant. Previously men
tioned was the fact that in Table 2, Nouns and Adjectives
correlated positively with the TTR measures. It may be
noted also that Nouns and Adjectives correlate positively
with one another.with a coefficient of 0.36, while other
categories, Adverbs,' Verbs, Pronouns and Self Reference
Words which show low or negative correlations with the TTR
measures, show some positive correlations with one another,
namely: Verbs and Pronouns, .68; Verbs and Self Reference
Words, .43; Pronouns and Self Reference Words, .58.
Discussion
TTR Measures
The most significant single datum to emerge is the
change that takes place when the groups are adjusted to
minimize intelligence differences. As previously brought
out in Chapter III, intelligence measures generally predict
verbal measures with accuracy. Because both verbal meas
ures, the Overall TTR and the Mean Segmental TTR, ordinarily
reflect vocabulary extent and proficiency in usage, one
would not expect either measure to continue to show sig
nificant differences after adjustments have been made
removing the effects of intelligence. However, such is not
40
the case since the Mean Segmental TTR belies this prediction
and shows that the groups differ significantly in this re
gard. In review, the Overall TTR indicates average varia
bility of the whole language sample for each individual and
the Mean Segmental TTR shows the sequential variability of
that sample. It seems apparent from the results that the
Overall TTR is a function of the individual's intelligence.
At least, this is the primary inference that may be drawn
from the differences between the non-adjusted and adjusted
data seen in Table 1.
The Mean Segmental TTR, on the other hand, does not
show this functional covariability with intelligence and
therefore it is necessary to attribute the group differ
ences to another source. Carroll (6), following Zipf (42),
has suggested that in successive increments of a language
sample we see a deceleration of new word usage, i.e., as
the language sample increases, the number of new words
employed decreases. The inference being, in this case,
that the subject will expend his vocabulary fairly rapidly
at first and as a consequence utilize an increasing number
of repetitions. It may be seen that the Mean Segmental TTR
is a measure of the degree of variability in language usage
and therefore a measure of verbal rigidity. On these bases
it may be said that the group defined by the MACC as those
subjects exhibiting the poorest behavioral adjustment is
also the group showing the greatest verbal rigidity. The
findings of the present study confirm those made by Roshal
(35). When she calculated the TTR for samples in the first
and last interviews of cases in successful psychothera
peutic treatment, she found an increase in the TTR indicat
ing that a greater degree of variability and flexibility
in language usage is concomitant with better behavioral
adjustment. The subject's verbal rigidity seen in the more
poorly adjusted schizophrenic's constriction of language
usage, demonstrated in the present study by constricted
Mean Segmental TTR scores compared to those of better ad
justed schizophrenics, reinforces considerably the sugges
tion by Hanfmann (22) and Goldstein (18) that the severely
involved schizophrenic suffers an inability to shift from
one aspect of a situation to another. On the other hand,
Feldstein and Jaffe (14), in a study similar to the pres
ent study, but utilizing different groups, namely normals
and schizophrenics, found that the Mean Segmental TTR
scores were not significantly different for the two groups.
This discrepancy in findings may be attributed to two
refinements made in the present study. The first was the
42
use of the single clinical group of schizophrenia as com
pared with the contrasting of schizophrenics with normals
performed by Feldstein and Jaffe. The second refinement
was that the groups in the present study were chosen on the
more discrete basis of level of behavioral adjustment.
In summary, on the basis of the data collected and
statistically treated, Hypothesis 1, concerning Overall
TTR, is rejected. Hypothesis 2, concerning Mean Segmental
TTR, is confirmed.
Grammatical Construction Categories
No statistically significant differences were found
between groups in terms of grammatical construction cate
gories. As far as frequency of appearance of categories
are concerned, other findings, principally those of Mabry
(29) and Mann (30), showed grammatical categories to be
insensitive to differences between schizophrenic groups and
schizophrenics and normals in written language. In con
trast, studies by Fairbanks (13) and Lorenz and Cobb (28),
investigating spoken language, showed that schizophrenics
used a proportionately greater number of verbs and pronouns
and fewer nouns than normals. The discrepancy in findings
between the present study and those by Fairbanks and Lorenz
43
and Cobb may well be due to differences in subjects em
ployed and stimulus materials utilized.
It is interesting to note that the data for gram
matical construction categories, although indicating non
significant differences between the groups, do show results
in a directly opposite direction to that hypothesized.
Statistical non-significance precludes the necessity for
speculation regarding this finding in that statistical non
significance implies that a less than chance relationship
exists between groups and that scores may be expected to
fall in any direction.
On the basis of the findings of this study, Hypo
thesis 3, pertaining to grammatical construction categories,
is rejected. There emerged no relation among the data to
suggest that analyses of grammatical categories would in
any way be sensitive to differentiating the two groups in
the present study, i.e., grammatical categories extracted
from their context on the basis of frequency alone do' not
serve as an index portraying characteristic differences
between schizophrenic groups of different behavioral
adjustment.
CHAPTER V
SUMMARY, CONCLUSIONS AND IMPLICATIONS
Summary
The Problem
The purpose of this study was to investigate the
relationship between degrees of behavioral adjustment and
degrees of verbal rigidity in chronic schizophrenics by
tapping some structural aspects of spoken language. Aris
ing from the general problem of the relationship between
language and adjustment were two basic questions and three
hypotheses. The questions were: (1) Does the flexibility
of word utterance of the chronic schizophrenic vary as a
function of behavioral adjustment? (2) Do certain gram
matical constructions used by the chronic schizophrenic
vary as a function of behavioral adjustment? The hypo
theses were: (1) The Overall TTR will vary as a function
of the behavioral adjustment of the chronic schizophrenic.
That is, the more poorly adjusted chronic schizophrenic as
44
45
contrasted with the better adjusted schizophrenic will have
a lower Overall TTR* (2) The Mean Segmental TTR will vary
as a function of the behavioral adjustment of the chronic
schizophrenic. That is, the more poorly adjusted chronic
schizophrenic as contrasted with the better adjusted
schizophrenic will have a lower Mean Segmental TTR.
(3) Usage of certain grammatical constructions will vary as
a function of the behavioral adjustment of the chronic
schizophrenic. That is, the more poorly adjusted chronic
schizophrenic as contrasted with the better adjusted
schizophrenic will employ, in a language sample of a spec
ific length, (a) more adjectives, (b) more adverbs, (c) more
verbs, (d) fewer nouns, (e) more pronouns and, (f) more
self reference words.
The Research Design and Method
Fifty-nine males with the diagnosis of chronic
schizophrenic, patients of the Psychiatric Service of the
Sepulveda Veterans Administration Hospital, were rated in
terms of behavioral adjustment and concomitantly their
spoken language was sampled. Two groups were delineated by
dividing the fifty-nine subject sample at the median be
havioral adjustment score. Thirty subjects comprised the
46
group which was low in behavioral adjustment (Group L)
while the group high in behavioral adjustment (Group H)
had twenty-nine subjects. Language samples were subjected
to Type-Token Ratio Analysis and analysis to determine
type-frequency characteristics of selected grammatical■con
struction categories. The research design of this study
was experimental, the independent variable being degree of
behavioral adjustment which was manipulated by selection;
the dependent variables comprised two TTR measures and six
grammatical construction categories. The statistical de
sign utilized was a multivariate analysis of covariance.
Results
Prior to adjusting the data to reduce the differ
ences between groups in respect to intelligence, F ratios
for both TTR measures (Overall TTR and Mean Segmental TTR)
significantly differentiated the groups. After corrections
were made for intelligence differences, Overall TTR no
longer significantly differentiated the groups while Mean
Segmental TTR, although attenuated, still displayed an F
ratio which significantly differentiated the groups. No
statistically significant differences were found between
the groups in terms of grammatical construction categories
47
either antecedent or consequent to correcting for intelli
gence differences. The findings were discussed in the
light of the questions posed by the problem.
Conclusions
The following conclusions seem warranted in light
of the findings of this study:
1. The language variability measured by the Over
all TTR is not related to behavioral adjustment
in chronic schizophrenics. Neither constraint
nor expansibility in language usage of chronic
schizophrenics measured by the Overall TTR is
related to behavioral adjustment.
2. The language variability measured by the Mean
Segmental TTR is related to behavioral adjust
ment.
a. Rigidity in language usage measured by the
Mean Segmental TTR appears as a function of
the poorer behavioral adjustment of the
chronic schizophrenic.
b. Flexibility in language usage measured by
the Mean Segmental TTR appears as a function
of the better behavioral adjustment of the
48
chronic schizophrenic.
3. Language variability measured by frequency of
occurrence of the following grammatical con
struction categories is not a function of be
havioral adjustment in chronic schizophrenics:
(a) Adjectives? (b) Adverbs; (c) Verbs?
(d) Nouns; (e) Pronouns; and (f) Self Reference
Words.
Implications
The results of this study have shown that verbal
behavior as measured by the variability of word usage with
in a total 1000-word language sample does not reflect the
differing levels of behavioral adjustment of two groups
of chronic schizophrenics when the groups are equated for
intelligence. This was demonstrated by the utilization of
the statistical technique of covariance with which the
effects of the variable of intelligence were neutralized
between groups. Overall TTR emerged as simply a measure of
vocabulary breadth. On the other hand, variability or
flexibility of language as a function of behavioral adjust
ment was reflected by the mean TTR of 50-word segments
within a 1000-word language sample after the groups were
,49
balanced in terms of intelligence.
The question may be asked whether the differences
in Mean Segmental TTR results are attributable to differ
ences between the groups iri terms of behavioral adjustment
or to differences in schizophrenic thought processes or to
both. Because this study did not contrast behavioral
adjustment in non-schizophrenic groups with behavioral
adjustment in schizophrenic groups, it was not possible to
ascertain the degree to which schizophrenic thinking, as
distinct from behavioral adjustment, may have contributed
to the differences that were found.
The Mean Segmental TTR has the advantage of being
objective as contrasted with a measure of manifest content.
By dealing with the structure rather than the content of
language, the Mean Segmental TTR taps that part of speech
which is probably less under the volitional control of the
subject than is content. Through quantifying the language
pattern of a subject without regard to the much more sub
jective ramifications of what he is expressing, a greater
degree of objectivity is gained in terms of the manner in
which he is actually functioning on his hospital ward.
Future research with the Mean Segmental TTR could
be directed toward determining its use as an index of
50
behavioral adjustment with clinical entities other than
schizophrenia. The Mean Segmental TTR also could be em
ployed to measure language modifications following electro
shock and insulin therapy, and psychosurgery. Contrasting
the Mean Segmental TTR of tranquilized patients with non-
tranquilized individuals within the same clinical group
could provide significant information relating language
usage with differential effects of chemotherapy. Another
tack of investigation associated with chemotherapy could
involve an analysis of variance of Mean Segmental TTR re
sults after the administration of various tranquilizing
agents to patients of differing behavioral adjustment.
The present study has demonstrated that frequency
characteristics of certain grammatical categories extracted
from their context do not differentiate between schizo
phrenic groups of different behavioral adjustment levels.
Grummon's (19) findings, based on his study using only four
subjects, indicated that the number of clauses contained in
a language sample of a given length discriminated between
subjects in terms of overt adjustment status. While the N
of four renders this relatively valueless for predictive
purposes, Grummon is suggesting an area that merits explor-
tion. The quantification of the number, type and length of
51
clauses in a language sample of a specific length, using
samples containing greater numbers of subjects, is indi
cated in light of Grummon1s results. An implication drawn
from Grummon's work is that sentence analysis, quantifying
complexity in terms of number of clauses and elaboration
of individual clauses, might yield significant findings re
garding individuals of varying degrees of behavioral
adjustment. For example, one might consider the basic ele
ments of a clause to be the subject and verb. The manner
in which an individual embellishes any agreed upon basic
unit could be quantified.
Based upon the findings of the present study, it is
recommended that further investigation utilizing the single
word as the basic unit of communication in the study of the
relationship of grammar to behavioral adjustment in chronic
schizophrenic be abandoned and that combined units of var
ious lengths be studied. The findings strongly suggest
that diverse combinations of units of communication rather
than the units themselves may provide a more productive
means of research into the language anomalies of schizo
phrenia heretofore not fully exploited.
A P P E N D I X E S
APPENDIX A
MACC BEHAVIORAL ADJUSTMENT SCALE
5L VUCC&Ui^. I ^ d lc lf u d tm e n t S c c iie
FORM SI
O b je c tiv e ~ d b p p ro a c h
to th e
^ d v a iu a tio n d 3 e h a .v io r a $ ^ d d ju 5 tm e n td
By
ROBERT B. ELLSWORTH, Ph. D.
Published by
O T S
EaabUiKtd 1948
WESTERN PSYCHOLOGICAL SERVICES
PUBLISHERS • DISTRIBUTORS
BOX 775, BEVERLY HILLS, CALIFORNIA
D ate: Place: W ard No. Activity Rated M F
Patient's N am e:
Rater's Name:
Total Adjustment Score Centile Score
Remarks: (for example, relatively low or high areas of behavior, etc.)
Evaluation:
CLINICIAN OR RATER:
W - 8 la
C opyright © 1962 by WESTERN PSYCHOLOGICAL SERVICES
N ot to be reproduced in w hole o r in p o rt w ithout w ritten perm ission o f copyright ow ner.
All rights reserved* Printed in U.S.A,
IMPORTANT: To ell raters: Be sure to familiarize yourself with the Directions For Rating found on p a g e four of this Scale,
before you begin to make your ratings.
1. Is he pleasant, never seems to be irritable or grouchy?
1 2 3 4 5
Usually very
grouchy
Most often ' •
irritable
Sometimes pleasant Most frequent
pleasant
Always pleasant
2. Does he generally cooperate, “ go along" with things asked of him?
1 2 3 4 5
Almost never
cooperates
Most frequently
resistive
Resistive rather
often
G oes along with
requests most of
the time ’
Always does w hat
is asked
3. Does he take p art in sensible “ back and forth" conversation, listening as well as talking to you
to your questions, but a “ give an d tak e" conversation?
, not just short answers
1 2 3 4 5
Never back
an d forth
conversation
Occasional back
and forth
conversation
Fairly often
“ give and tak e"
conversation
Usually good
“ back and forth"
talk
Almost always
listens an d talks
realistically
4. How many FRIENDS does he have (patients he talks to, spends time with, AND w ho w ant to be with him?
1 2 3 4 5
No friends, very
hostile to others or
alw ays by himself.
Shows interest when
approached by other
patients but rarely
talks with them
Usually tries to be
around other
patients but is
passive socially
Has one or two
friends with whom he
spends some time
Spends considerable
time with other
friends
5. Is he sullen?
1 2 3 4 5
Always sullen Most often sullen Sometimes sullen Rarely sullen Always pleasant^
6. Does he seem resistlive?
1 2 3 4 5
Very resistive Most often resistive Sometimes resistive Rarely resistive Never resistive
7. If asked a question, does he respond in such a w ay th a t he is understood, using words th at i are understandable?
1 2 3 4 5
Mute or talks
“ jibberish"
Answers make
little sense
Response often
sensible
Usually
sensible
Almost alw ays
sensible
8. Does he easily enter into “ give and take" conversation with other PATIENTS?
1 2 3 4 5
Rarely says anything
to others, even
when approached
Talks with others
but only when
approached
Talks willingly to
other patients but
depends on them to
“ keep the conversation
going"
Enters easily into
conversation, keeps
his end of the
conversation gbing
Talks with many
different patients
often stimulates
conversation with
other patients
9. Is he bitter?
1 2 3 4 5
Always bitter Usually bitter Sometimes bitter Rarely bitter Never bitter
10. In the things th at are expected of him to do, does he go ah e ad and do them on his own without having to b e told how
an d when to do it, or must he be directed an d encouraged to do them?
1 2 3 4 5
No initiative Occasionally
acts “ on his ow n"
Fairly often goes
ah e ad “ on his ow n"
Usually shows
initiative
Almost alw ays goes
a h e ad “ on his ow n"
11. Does he seem accessible, easy to “ get through" to, ab le to understand you when you talk to him?
1 2 3 4 5
Like talking
to a “ brick w all"
O ccasionally
“ get through"
O ften accessible Almost always
accessible
Very easy to “get
through" to him
12. Does he spend his time alone?
I 2 3 4 5
Always alone, pays
no attention to
others
Always clone but is
alert and aw are of
others about him,
rarely participates
in activity
Reluctant but can be
drawn into activity
when approached
Usually will take
part in activity
when invited to do so
O ften doing things
with other patients
13. Is he angry and hostile?
1 2 3 4 5
Always angry Usually angry Sometime friendly Usually friendly Almost alw ays friendly
14. In tasks assigned to him, can he “stay w ith"
“ lost"?
the task w ithout frequent redirection, without becoming preoccupied and
1 2 3 4 5
Almost alw ays be
comes preoccupied
quickly
Rarely “stays
w ith" tasks
Continues tasks
fairly long
Usually “ stays
w ith" it
Almost alw ays
completes tasks
15. Does he quickly grasp an d understand w hat is told him, w ithout having to explain things three
passively listening, or paying attention, but grasping easily w hat you w ant?
or four times, not just
1 2 3 4 5
Never really
com prehends
Understands some
after long
explanations
Gets most of it
with 1 or 2
explanations
Usually picks it up
fairly easily
G rasps right aw ay
w hat is told him
16. Is he well informed about OTHERS ON THE WARD?
1 2 3 4 5
Shows no evidence
of knowing any
patients by nam e
Sometimes knows to
whom you are
referring when you
use a patient's nam e
Usually knows to
whom you are
referring when you
use another
Knows and usually
calls by nam e most
of the personnel
Knows and usually
calls by nam e most
of the personnel
an d patients
patient's name
SCORING*
Mood C ooperation Communication Social
Contact
TOTAL
ADJUSTMENT
1. 2. 3. 4. (Add the sums of .
5, 6. 7. 8. Mood, Cooperation,
9.
>
10. 11. 12. Communication and
13. 14. 15. 16. Social Con act.)
SUM + + + =
*To obtain the patient's profile, transfer the summed scores for M ood, Cooperation, v .
Communication, Social Contact, an d Total Adjustment to the Profile Summary on Page 4.
DIRECTIONS FOR RATING
1. For each item, circle the number which most
characteristically describes the patient's
behavior.
2. Rate only patients you know through per
sonal observations or contact.
3. Be objective in your ratings. Rate the patient
as you saw him or her behave, not as you
wish or think the patient should behave. Do
not let your personal feelings about the p a
tient bias your ratings.
4. Do not give the same rating tor all items; use
your best judgment on each item. A patient
may be rated very low on one item even
though he may show a high level of adjust
ment on most items.
5. Rate quickly and do not hesitate to give ex
treme ratings if your observations point to
such ratings.
6. Practice rating several patients, then discuss
your ratings with one skilled in using this
scale. Rate the same patients on two different
occasions to improve the reliability of your
ratings.
DIRECTIONS FOR SCORING
1. Record the rating for each item on the bottom
of page 3.
2. Sum the scores for each behavior area and
plot the summed scores on the profile sheet
under the headings Mood, Cooperation, Com
munication, Social Contact.
3. Sum the 4 behavioral area scores to obtain
the Total Adjustment Raw Score.
4. Convert the Total Adjustment Raw Score into
a Centile Score on the Profile Sheet. The
Centile Score is then used in reports of p a
tient adjustment level, progress, etc.
PATIENTS'
RELATIVE
STANDING
VERY
HIGH
ABOVE
AVERAGE
AVERAGE
BELOW
AVERAGE
VERY
LOW
PROFILE SUM M ARY
Mood
— 20—
— 19—
-18-
- 17-
— 16 —
— 15—
— 14—
— 13-
— 12—
— 11 —
Coopera
tion
-20—
— 19—
— 18 -
— 17 —
— 15—
— 14 —
-13-
— 12 —
— 11 —
- 10-
— 10—
— 9— — 9—
— 8— — 8—
— 7— — 7—
-,-4-6__ — 4-6—
Communi
cation
-20—
— 19—
— 18—
- 17—
— 16—
-15-
- 4 -4 -
— 13—
— 12—
— 11—
— 1
—9 —
—8 —
—7 —
Social
Contact
— 20—
— 19—
-18-
— 17-
— 16 —
- 15—
-14—
-13 —
- 12—
- 11—
—10 —
— 9—
—8 —
— 7 —
Total Ad
justment
— 80—
— 79—
— 78—
— 77 —
— 76—
— 75—
— 74—
— 73—
— 72—
—71 —
—70 —
— 69—
— 68—
— 67 -
APPENDIX B
STATISTICAL DATA RELATING TO LEVEL OF
EDUCATIONAL ATTAINMENT OF GROUPS L AND H
APPENDIX B
STATISTICAL DATA RELATING TO LEVEL OF
EDUCATIONAL ATTAINMENT OF GROUPS L AND H
t : test calculations were based on the following
data (homogenous variances):
Variable Mean Standard Deviation Sample Size
Group H 11.8280 1.8529 29
Group L 11.4000 2.1107 30
t_ = 0.81252
d.f. = 57
2 2
S1 S2
Error term: Standard error of difference = — + —
nl n2
nl n2
59
APPENDIX C
INSTRUCTIONS AND STIMULUS MATERIAL
APPENDIX C
INSTRUCTIONS AND STIMULUS MATERIAL
The following instructions and stimulus material
were presented to each subject:
"I want to talk to you about some proverbs today.
You know what a proverb is. A proverb is a sentence that
teaches a lesson. I am going to read some proverbs to you,
and I want you to tell me what they mean. I also want you
to describe a situation in which each proverb would apply.
For example, the proverb 'Let sleeping dogs lie' means that
we should avoid stirring up old troubles and quarrels. An
example of a situation in which that proverb would apply
would be, for instance, if you and a friend had quarreled
over something several months ago, you should forget it and
be friends with him again instead of continuing to quarrel
with him each time that you see him. Do you understand
what I mean? When you begin to tell me what the proverb
means, I will not answer any questions or talk to you about
the proverbs. Now you tell me what this proverb means,
'The early bird catches the worm.'"
"Now give me an example illustrating that."
61
62
This procedure was continued with each of the fol
lowing proverbs:
Rome was not built in a day.
Better be happy than wise.
He who laughs last laughs best.
When the cat's away the mice will play.
A chain is as strong as its weakest link.
Where there's a will there's a way.
The devil finds work for idle hands.
The sun shines on all alike.
What you will sow you will reap.
Great bodies move slowly.
Like father, like son .
The wife is the key to the house.
Barking dogs never bite.
It is always darkest just before the dawn.
You can't touch pitch without being tarred.
Tell me the company you keep and I'll tell you
what you are.
Speech is the picture of the mind.
A fair face may hide a foul heart.
Deeds are males and words are females.
B I B L I O G R A P H Y
BIBLIOGRAPHY
Blalock, H. M.. Social statistics. New York: McGraw
Hill Book Co., Inc., 1960.
Bock, D. R. Programming univariate and multivariate
analysis of variance. Technometrics, 1963, 5_,
95-117.
Boder, D. P. The adjective-verb-quotient: a contribu
tion to the psychology of language. Psychol. Rec.,
1940, 3.* 310-343.
Buhler, K. Sprachtheorie. Jena: Gustave Fischer,
1934.
Busemann, A. Die Sprache der Jupend als Ausdruck der
Entwicklungsrhythmik. Jena: Gustave Fischer,
1925.
Carroll, J. B. Diversity of vocabulary and the har
monic series law. Psychol. Rec., 1938, 2_> 379-
396.
Dewey, J. The psychology of infant language. Psychol.
Rev., 1894, JL, 63—66.
Dixon, W. J. and Massey, F. J. Introduction to
statistical analysis. New York: McGraw Hill Book
Co., Inc., 1957.
Edwards, A. L. Experimental design in psychological
research. New York: Holt, Rinehart and Winston,
1960.
65
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
Eilsworth, R. B. The MACC behavioral adjustment
scale. Beverly Hills, Calif.: Western Psycholo
gical Services, 1957.
Ellsworth, R. B. The MACC behavioral adjustment
scale. Beverly Hills, Calif.: Western Psycholo
gical Services, 1961.
Ellsworth, R. B. and Clayton, W. H. Measurement of
improvement in mental illness. J. Consult.
Psychol., 1959, 2_3, 15-20.
Fairbanks, H. The quantitative differentiation of
samples of spoken language. Psychol. Monogr.,
1944, 2_, 19-38.
Feldstein, S. and Jaffe, J. Vocabulary diversity of
schizophrenics .and normals. J. Speech Hearing
Res. , 1962, 5., 76-78.
Fisher, R. A. On the "probable error" of a coeffi
cient of correlation. Metron, 1921, _1, Part 4,
1-32.
Fromm-Reichmann, F. Basic problems■in the psycho
therapy of schizophrenia. Psychiatry, 1958, 21,
1-6.
Giedt, F. H. Admissions testing program. A publica
tion of the Psychological Research Service,
Sepulveda Veterans Administration Hospital,
Sepulveda, California, 1959.
Goldstein, K. Methodological approach to the study of
schizophrenic thought disorder. In J. S. Kasanin
(Ed.), Language and thought in schizophrenia.
Berkeley: Univ. of Calif. Press, 1954.
Grummon, D. L. An investigation into the use of gram
matical and psychogrammatical categories of lan
guage for the study of personality and psycho
therapy. Unpublished Ph.D. dissertation, Univ. of
Chicago, 1950.
66
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Hall, C. E. and Cramer, E. M. A general purpose pro
gram to compute multivariate analyses of variance
on an I.B.M. 7090 computer. A publication of the
Biometric Laboratory, The George Washington
University, 1961.
Hammer, M. and Salzinger, K. Some formal character
istics of schizophrenic speech as a measure of
social deviance. Annals. N.Y. Acad. Sci., 1964,
105, 861-889.
Hanfmann, E. Analysis of the thinking disorder in a
case of schizophrenia. Arch. Neurol. Psychiat.,
1939, 41, 568-579.
Horvitz, V. S. An attempt to measure psycho
therapeutic progress by means of a frequency
analysis of the parts of speech. Unpublished B.A.
thesis, Univ. of Illinois, 1950.
Johnson, W. A program of research. Psychol. Monogr.,
1944, 2, 1-15.
Lorenz, M. Expressive behavior and language patterns.
Psychiatry, 1955, 1 8 , , 353-366.
Lorenz, M. and Cobb, S. Language behavior in manic
patients. A.M.A. Arch. Neurol. & Psychiat., 1952,
67, 763-770.
Lorenz, M. and Cobb, S. Language behavior in psycho
neurotic subjects. A.M.A. Arch. Neurol, and
Psychiat., 1953, 69,, 684T694.
Lorenz, M. and Cobb, S. Language patterns in psycho
tic and psychoneurotic subjects. A.M.A. Arch.
Neurol. & Psychiat., 1954, 72_, 665-673.
Mabry, M. Language characteristics of scattered and
non-scattered schizophrenics compared with nor
mals. J. Psychol., 1964, J57_, 29-40.
Mann, M. B. The quantitative differentiation ,of sam
ples of written language. Psychol. Monogr.,
1944, 56, 41-75.
67
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
Mowrer, O. H. Changes in verbal behavior during
psychotherapy. In O. H. Mowrer (Ed.)# Psycho
therapy theory and research. New Yorks The Ronald
Press, 1953.
Newman, S. S. and Mather, V. G. Analysis of spoken
language of patients with affective disorders.
flmer. J. Psychiat., 1938, 94, 913-942.
Rabin, A. I. and King, G. F. Psychological studies.
In L. Beliak, (Ed.), Schizophrenia. New York:
Logos Press, Inc., 1958.
Rashkis, H. A. Three types of thinking disorders.
J. Nerv. Ment. Pis., 1947, 106, 650-670.
Roshal, J. J. G. The type-token ratio as a measure of
change in behavior variables during psychotherapy.
In W. U. Snyder (Ed.), Group report of a program
of research in psychotherapy. Penn. State Col
lege, State College, Pa., 1953, 94-104.
Ruesch, J. Psychotherapy with schizophrenics. In A.
Auerback (Ed.), Schizophrenia. New York: Ronald,
Press, 1959.
Sanford, F. H. Speech and personality. Psychol.
Bull. . 1942, J39, 811.
Stout, G. F. Analytic psychology. Vol. 2. New York:
Macmillan, 1902.
Webster's New Twentieth Century Dictionary, Unabridged.
2nd ed. 1956.
Weigert, E. Problems of communication in psycho
therapy.. Psychiatry, 1958, 21. 245.
Wonderlic, E. F. Wonderlic personnel test manual.
Northfield, 111.: E. F. Wonderlic, 1961.
Zipf, G. K. The psychobiology of language. Boston:
Houghton Mifflin Co., 1935.
68
43. Zipf, G. K. Observations of the possible effect of
mental age upon the frequency-distribution of
words, from the viewpoint of dynamic philology.
J. Psychol., 1937, 4, 239-244.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

An Experimental Study Of Relationships Between Self-Concepts Of Fourth And Eighth Grade Stuttering And Non-Stuttering Boys

PDF

An Experimental Study Of The Effects Of Interaural Temporal Delays And Intensity Differences On Intracranial Localization Of Spondee Words

PDF

An Empirical Study Of Sex-Role Identification And Sex-Role Preference In A Selected Group Of Stuttering Male Children

PDF

Effects Of An Audience On Serial Association

PDF

Some Relationships Between Auditory Reaction Time And Number Of Oscillations Of Fixed Frequency Bekesy Tracings

PDF

An Experimental Investigation Of Repression Of The Auditory Perception Ofdisturbing Words As Indicated By Verbal And Electrodermal Responses

PDF

An Experimental Study Of Auditory Thresholds Of Adults For Warble Tone, Pure Tone, And Recorded Speech

PDF

Pitch Perturbation As A Function Of Subjective Vocal Constriction

PDF

Air Flow Rate As A Function Of Vocal Constriction In Normal Adult Males

PDF

An Experimental Investigation Of The Dichotic Application Of Delayed Sidetone And Masking Noise

PDF

A Normative Study Of Oral Sensation And Perception: Two-Point Discrimination, Form Identification, Tactile Pattern Recognition, And Mandibular Kinesthesia

PDF

An Investigation Of Attitude Differences In Parents Of Stutterers And Parents Of Non-Stutterers

PDF

An Investigation Of The Incidence Of Stuttering Among Elementary School Children In The Los Angeles City Schools

PDF

A Comparative Study Of Speech Patterns Of Selected Negro And White Kindergarten Children

PDF

An Experimental Investigation Of The Relationship Between Frequency Of Stuttering And Open Expression Of Aggression

PDF

An Experimental Study Of The Effect Of Punishment Of The Expectancy To Stutter On The Frequency Of Subsequent Expectancies And Stuttering

PDF

An Empirical Study Of The Behavioral Characteristics Of Sincere And Insincere Speakers

PDF

Teaching Operative Dental Technics Via Television: An Experimental Inquiry

PDF

An Experimental Study Of The Effect On Babbling Of Visual Interaction Between Mother And Infant

PDF

An Experimental Study Of Intelligibility Thresholds Of Hypacusic Adults Using Forced-Choice And Yes-No Techniques

Asset Metadata

Creator Reid, Philip (author)

Core Title An Investigation Of The Spoken Language Of Chronic Schizophrenics As A Function Of Behavioral Adjustment

Degree Doctor of Philosophy

Degree Program Speech

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,Speech Communication

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Perkins, William H. (committee chair), Dreher, John J. (committee member), Garwood, Victor P. (committee member), Ross, Albert E. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-194625

Unique identifier UC11360263

Identifier 6605493.pdf (filename),usctheses-c18-194625 (legacy record id)

Legacy Identifier 6605493.pdf

Dmrecord 194625

Document Type Dissertation

Rights Reid, Philip

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA