Behavioral Analysis

An Alternative to Diagnostic Classification

FREDERICK H. KANFER, PhD, AND GEORGE SASLOW, MD, PORTLAND, ORE

DuringURING the past decade attacks on conven- permit prediction of responses to social stresses,
life crises, or psychiatric treatment. This limita¬
tional psychiatric diagnosis have been so wide-
spread that many clinicians now use diagnostic tion obviates anything but a crude and tentative
labels sparingly and apologetically. The con- approximation to a taxonomy of effective indi¬
tinued adherence to the nosological terms of the vidual behaviors.
traditional classificatory scheme suggests some Zigler and Phillips,28 in discussing the re¬
utility of the present categorization of behavior quirement for an adequate system of classifica¬
disorders, despite its apparently low reliabil- tion, suggest that an etiologically-oriented closed
ity 1,21; its limited prognostic value 7,26; and its system of diagnosis is premature. Instead, they
multiple feebly related assumptive supports. In believe that an empirical attack is needed, using
a recent study of this problem, the "symptoms broadly defined as meaningful and
symptom
patterns of carefully diagnosed paranoid schizo- discernible behaviors, as the basis of the classifi-
phrenics were compared. Katz et al12 found catory system" (p 616). But symptoms as a
considerable divergence among patients with class of responses are defined after all only by
the same diagnosis and concluded that "diag- their nuisance value to the patient's social en¬
nostic systems which are more circumscribed in vironment or to himself as a social being. They
their intent, for example, based on manifest are also notoriously unreliable in predicting the
behavior alone, rather than systems which at- the patient's particular etiological history or his
tempt to comprehend etiology, symptom patterns response to treatment. An alternate approach
andprognosis, may be more directly applicable lies in an attempt to identify classes of
to currentproblems in psychiatric research" dependent variables in human behavior which
(p202). would allow inferences about the particular con¬
We propose here to examine some sources of trolling factors, the social stimuli, the physio¬
dissatisfaction with the present approach to logical stimuli, and the reinforcing stimuli, of
which they are a function. In the present early
diagnosis, to describe a framework for a be-
havioral analysis of individual patients which stage of the art of psychological prognostication,
implies both suggestions for treatment and out- it appears most reasonable to develop a program
come criteria for the single case, and to indicate of analysis which is closely related to subsequent
the conditions for collecting the data for such treatment. A classification scheme which implies
a program for behavioral change is one which
an analysis.
has not only utility but the potential for experi¬
I. Problems in Current Diagnostic Systems mental validation.
Numerous criticisms deal with the internal The task of assessment and prognosis can
therefore be reduced to efforts which answer
consistency, the explicitness, the precision, and the following three questions : (a) which specific
the reliability of psychiatric classifications. It
seems to us that the more important fault lies
behavior patterns require change in their fre¬
in our lack of sufficient knowledge to categorize quency of occurrence, their intensity, their dura¬
behavior along those pertinent dimensions which tion or in the conditions under which they occur,
Submitted for publication Oct 9, 1964.
{b) what are the best practical means which can
From the University of Oregon Medical School. produce the desired changes in this individual
Reprint requests to 3181 SW Sam Jackson Pk Rd, (manipulation of the environment, of the be¬
Portland, Ore 97201 (Dr. Saslow). havior, or the self-attitudes of the patient), and

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pittsburgh User on 05/01/2016

 (c) what factors are currently maintaining it into a two-step process. A statistical procedure
and what are the conditions under which this is suggested to facilitate the primary or diag¬
behavior was acquired. The investigation of the nostic labeling process. However, the choice of
history of the problematic behavior is mainly treatment depends not only on the diagnosis
of academic interest, except as it contributes proper. Treatment decisions are also influenced
information about the probable efficacy of a by the moral, ethical, social, and economic con¬
specific treatment method. ditions of the individual patient, his family and
Expectations of Current Diagnostic Sys¬ the society in which he lives. The proper assign¬
tems.—In traditional medicine, a diagnostic ment of the weight to be given to each of these
statement about a patient has often been viewed values must in the last analysis be left to the
as an essential prerequisite to treatment because physician's judgment (Ledley and Lusted14).
a diagnosis suggests that the physician has some The Ledley and Lusted model presumes avail¬
knowledge of the origin and future course of able methods for the observation of relevant
the illness. Further, in medicine diagnosis fre¬ behavior (the symptom complex), and some
quently brings together the accumulated knowl¬ scientific knowledge relating it to known ante¬
edge about the pathological process which leads cedents or correlates (the disease process). Con¬
to the manifestation of the symptoms, and the temporary theories of behavior pathology do not
experiences which others have had in the past in yet provide adequate guidelines for the observer
treating patients with such a disease process. to suggest what is to be observed. In fact,
Modern medicine recognizes that any particular Szasz 25 has expressed the view that the medicai
disease need not have a single cause or even a model may be totally inadequate because psy¬
small number of antecedent conditions. Never¬ chiatry should be concerned with problems of
theless, the diagnostic label attempts to define at living and not with diseases of the brain or other
least the necessary conditions which are most biological organs. Szasz 25 argues that "mental
relevant in considering a treatment program. illness is a myth, whose function it is to disguise
Some diagnostic classification system is also and thus render more potable the bitter pill of
invaluable as a basis for many social decisions moral conflict in human relations" (p 118).
involving entire populations. For example, plan¬ The attack against use of the medical model
ning for treatment facilities, research efforts in psychiatry comes from many quarters.
and educational programs take into account the Scheflen 23 describes a model of somatic psy¬
distribution frequencies of specified syndromes chiatry which is very similar to the traditional
in the general population. medical model of disease. A pathological process
Ledley and Lusted 14 give an excellent con¬ results in onset of an illness; the symptoms are
ception of the traditional model in medicine by correlated with a pathological state and represent
their analysis of the reasoning underlying it. our evidence of "mental disease." Treatment
The authors differentiate between a disease com¬ consists of removal of the pathogen, and the
plex and a symptom complex. While the former state of health is restored. Scheflen suggests that
describes known pathological processes and this traditional medical model is used in psy¬
their correlated signs, the latter represents par¬ chiatry not on the basis of its adequacy but
ticular signs present in a particular patient. The because of its emotional appeal.
bridge between disease and symptom complexes The limitations of the somatic model have
is provided by available medical knowledge and been discussed even in some areas of medicine
the final diagnosis is tantamount to labeling the for which the model seems most appropriate.
disease complex. However, the current gaps in For example, in the nomenclature for diagnosis
medical knowledge necessitate the use of proba¬ of disease of the heart and blood vessels, the
bility statements when relating disease to symp¬ criteria committee of the New York Heart Asso¬
toms, admitting that there is some possibility for ciation 17 suggests the use of multiple criteria for
error in the diagnosis. Once the diagnosis is cardiovascular diseases, including a statement of
established, decisions about treatment still de¬ the patient's functional capacity. The committee
pend on many other factors including social, suggests that the functional capacity be ". esti¬
..

moral, and economic conditions. Ledley and mated by appraising the patient's ability to per¬
Lusted14 thus separate the clinical diagnosis form physical activity" (p 80), and decided

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pittsburgh User on 05/01/2016

 largely by inference from his history. Further,17 lead to similar symptoms and respond to similar
". .
(it) should not be influenced by the char¬ treatment. This dimension of diagnosis is con¬
.

acter of the structural lesion or by an opinion siderably more fruitful when dealing with be¬
as to treatment or prognosis" (p 81). This ap¬ havior disorders which are mainly under control

proach makes it clear that a comprehensive as¬ of some biological condition. When a patient is
sessment of a patient, regardless of the physical known to suffer from excessive intake of alcohol
disease which he suffers, must also take into his hallucinatory behavior, lack of motor co¬
account his social effectiveness and the particular ordination, poor judgment, and other behavioral
ways in which physiological, anatomical, and evidence disorganization can often be related
psychological factors interact to produce a par¬ directly to some antecedent condition such as the
ticular behavior pattern in anindividual patient. toxic effect of alcohol on the central nervous
Multiple Diagnosis.—A widely used practical system, liver, etc. For these cases, classification
solution and circumvention of the difficulty in¬ by etiology also has some implications for prog¬
herent in the application of the medical model nosis and treatment. Acute hallucinations and
to psychiatric diagnosis is offered by Noyes other disorganized behavior due to alcohol
and Kolb.18 They suggest that the clinician con¬ usually clear up when the alcohol level in the
struct a diagnostic formulation consisting of blood stream falls. Similar examples can be
three parts : ( 1 ) A genetic diagnosis incorporat¬ drawn from any class of behavior disorders in
ing the constitutional, somatic, and historical- which a change in behavior is associated pri¬
traumatic factors representing the primary marily or exclusively with a single, particular
sources or determinants of the mental illness; antecedent factor. Under these conditions this
(2) A dynamic diagnosis which describes the factor can be called a pathogen and the situation
mechanisms and techniques unconsciously used closely approximates the condition described by
by the individual to manage anxiety, enhance the traditional medical model.
self-esteem, ie, that traces the psychopathologi¬ Utilization of this dimension as a basis for
cal processes; and (3) A clinical diagnosis which psychiatric diagnosis, however, has many prob¬
conveys useful connotations concerning the re¬ lems apart from the rarity with which a speci¬
action syndrome, the probable course of the dis¬ fied condition can be shown to have a direct
order, and the methods of treatment which will "causal" relationship to a pathogen. Among the
most probably prove beneficial. Noyes' and current areas of ignorance in the fields of psy¬
Kolb's multiple criteria 18 can be arranged along chology and psychiatry, the etiology of most
three simpler dimensions of diagnosis which common disturbances probably takes first place.

may have some practical value to the clinician : No specific family environment, no dramatic
(1) etiological, (2) behavioral, and (3) predic¬ traumatic experience, or known constitutional
tive. The kind of information which is conveyed abnormality has yet been found which results in
by each type of diagnostic label is somewhat the same pattern of disordered behavior. While
different and specifically adapted to the purpose current research efforts have aimed at investi¬
for which the diagnosis is used. The triple-label gating family patterns of schizophrenic patients,
approach attempts to counter the criticism aimed and several studies suggest a relationship be¬
at use of any single classificatory system. Con¬ tween the mother's behavior and a schizophrenic
fusion in a single system is due in part to the process in the child,10 it is not at all clear why the
fact that a diagnostic formulation intended to presence of these same factors in other families
describe current behavior, for example, may be fails to yield a similar incidence of schizo¬
found useless in an attempt to predict the re¬ phrenia. Further, patients may exhibit behavior
sponse to specific treatment, or to postdict the diagnosed as schizophrenic when there is no
patient's personal history and development, or evidence of the postulated mother-child relation¬
to permit collection of frequency data on hos¬ ship.
pital populations. In a recent paper Meehlle postulates schizo¬
Classification by Etiology.—The Kraepelinian phrenia as a neurological disease, with learned
system and portions of the 1952 APA classifi¬ content and a dispositional basis. With this array
cation emphasize etiological factors. They share of interactive etiological factors, it is clear that
the assumption that common etiological factors the etiological dimension for classification would

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pittsburgh User on 05/01/2016

 at best result in an extremely cumbersome sys¬ Descriptive classification is not always this
tem, at worst in a useless one. deceptive, however. Assessment of intellectual
Classification by Symptoms.—A clinical diag¬ performance sometimes results in a diagnostic
nosis often is a summarizing statement about the statement which has predictive value for the

way in which a person behaves. On the assump¬ patient's behavior in school or on a job. To
tion that a variety of behaviors are correlated date, there seem tobe very few general state¬
and consistent in any given individual, it be¬ ments about individual characteristics, which
comes more economical to assign the individual have as much predictive utility as the IQ.
to a class of persons than to list and categorize A second limitation is that the current ap¬
all of his behaviors. The utility of such a sys¬ proach to diagnosis by symptoms tends to center
tem rests heavily on the availability of empirical on a group of behaviors which is often irrelevant
evidence concerning correlations among various with regard to the patient's total life pattern.
behaviors (response-response relationships), These behaviors may be of interest only because
and the further assumption that the frequency of they are popularly associated with deviancy and
occurrence of such behaviors is relatively inde¬ disorder. For example, occasional mild delusions
pendent of specific stimulus conditions and of interfere little or not at all with the social or
specific reinforcement. There are two major occupational effectiveness of many ambulatory
limitations to such system. The first is that
a patients. Nevertheless, admission of their occur¬
diagnosis by symptoms, as we have indicated in rence is often sufficient for a diagnosis of
an earlier section, is often misleading because psychosis. Refinement of such an approach be¬
it implies common etiological factors. Freed¬ yond current usage appears possible, as shown
man
7
gives an excellent illustration of the dif¬ for example by Lorr et al15 but this does not
ferences both in probable antecedent factors and remove the above limitations.
subsequent treatment response among three Utilization of a symptom-descriptive ap¬
cases diagnosed as schizophrenics. Freedman's proach frequently focuses attention on by-prod¬
patients were diagnosed by at least two psy¬ ucts of larger behavior patterns, and results in
chiatrists, and one would expect that the tradi¬ attempted treatment of behaviors (symptoms)
tional approach should result in whatever which may be simple consequences of other im¬
treatment of schizophrenia is practiced in the portant aspects of the patient's life. Emphasis
locale where the patients are seen. The first pa¬ on the patient's subjective complaints, moods
tient eventually gave increasing evidence of an and feelings tends to encourage use of a syn¬
endocrinopathy, and when this was recognized drome-oriented classification. It also results fre¬
and treated, the psychotic symptoms went into quently in efforts to change the feelings,
remission. The second case had a definite history anxieties, and moods (or at least the patient's
of seizures and appropriate anticonvulsant medi¬ report about them), rather than to investigate
cation was effective in relieving his symptoms. the life conditions, interpersonal reactions, and
In the third case, treatment directed at an un¬ environmental factors which produce and main¬
covering analysis of the patient's adaptive tech¬ tain these habitual response patterns.
niques resulted in considerable improvement in Classification by Prognosis.—To date, the
the patient's behavior and subsequent relief from least effort has been devoted to construction of
psychotic episodes. Freedman7 suggests that a classification system which assigns patients to
schizophrenia is not a disease entity in the sense the same category on the basis of their similar
that it has a unique etiology, pathogenesis, etc, response to specific treatments. The proper ques¬
but that it represents the evocation of a final tion raised for such a classification system con¬
common pathway in the same sense as do head¬ sists of the manner in which a patient will react
ache, epilepsy, sore throat, or indeed any other to treatments, regardless of his current behavior,
symptom complex. It is further suggested that or his past history. The numerous studies at¬
the term "schizophrenia has outlived its useful¬
ness and should be discarded" (p 5). Opler
19·20 tempting to establish prognostic signs from
has further shown the importance of cultural projective personality tests or somatic tests
factors in the divergence of symptoms observed represent efforts to categorize the patients on
in patients collectively labeled as schizophrenic. this dimension.

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pittsburgh User on 05/01/2016

 Windle 26 has called attention to the low de¬ II. A Functional (Behavioral-Analytic)
gree of predictability afforded by personality Approach
(projective) test scores, and has pointed out the The growing literature on behavior modi¬
difficulties encountered in evaluating research in fication procedures derived from learning
this area due to the inadequate description of the theory s,e,ii,i3,27 SUggests that an effective diag¬
population sampled and of the improvement nostic procedure would be one in which the
criteria. In a later review Fulkerson and Barry 8 eventual therapeutic methods can be directly re¬
came to the similar conclusion that psychological lated to the information obtained from a con¬
test performance is a poor predictor of outcome tinuing assessment of the patient's current
in mental illness. They suggest that demographic behaviors and their controlling stimuli. Ferster ß
variables such as severity, duration, acuteness has said "... a functional analysis of behavior
of onset, degree of precipitating stress, etc, ap¬ has the advantage that it specifies the causes of
pear to have stronger relationships to outcome behavior in the form of explicit environmental
than test data. The lack of reliable relationships events which can be objectively identified and
between diagnostic categories, test data, demo¬ which are potentially manipulable" (p 3). Such
a diagnostic undertaking makes the assumption
graphic variables, or other measures taken on the
patient on the one hand, and duration of illness, that a description of the problematic behavior,
response to specific treatment, or degree of re¬ its controlling factors, and the means by which
covery, on the other hand, precludes the con¬ it can be changed are the most appropriate
struction of a simple empiric framework for a "explanations." It further makes the assump¬
diagnostic-prognostic classification system based tion that a diagnostic evaluation is never com¬
only on an array of symptoms. plete. It implies that additional information
None of the currently used dimensions for about the circumstances of the patient's life
diagnosis is directly related to methods of pattern, relationships among his behaviors, and
modification of a patient's behavior, attitudes, controlling stimuli in his social milieu and his
response patterns, and interpersonal actions. private experience is obtained continuously
Since the etiological model clearly stresses until it proves sufficient to effect a noticeable
causative factors, it is much more compatible change in the patient's behavior, thus resolving
with a personality theory which strongly em¬ "the problem." In a functional approach it is
phasizes genetic-developmental factors. The necessary to continue evaluation of the patient's
classification by symptoms facilitates social-ad¬ life pattern and its controlling factors, concur¬
ministrative decisions about patients by provid¬ rent with attempted manipulation of these
ing some basis for judging the degree of variables by reinforcement, direct intervention,
deviation from social and ethical norms. Such or other means until the resultant change in the

a classification is compatible with a personality patient's behavior permits restoration of more

theory founded on the normal curve hypothesis efficient life experiences.
and concerned with characterization by com¬ The present approach shares with some psy¬
parison with a fictitious average. The prog¬ chological theories the assumption that psycho¬
nostic-predictive approach appears to have the therapy is not an effort aimed at removal of
most direct practical applicability. If continued intrapsychic conflicts, nor at a change in the
research were to support certain early findings, personality structure by therapeutic interactions
it would be indeed comforting to be able to of intense nonverbal nature, (eg, transference,
predict outcome of mental illness from a pa¬ self-actualization, etc). We adopt the assump¬
tient's premorbid social competence score,28 or tion instead that the job of psychological treat¬
from the patient's score on an ego-strength ment involves the utilization of a variety of
scale,4 or from many of the other signs and methods to devise a program which controls the
single variables which have been shown to have patient's environment, his behavior, and the
some predictive powers. It is unfortunate that consequences of his behavior in such a way that
these powers are frequently dissipated in cross the presenting problem is resolved. We hypothe¬
validation. As Fulkerson and Barry have indi¬ size that the essential ingredients of a psycho¬
cated,8 single predictors have not yet shown therapeutic endeavor usually involve two
much success. separate stages : ( 1 ) a change in the perceptual

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pittsburgh User on 05/01/2016

 discriminations of a patient, ie, in his approach the pertinent controlling stimuli in his social and
to perceiving, classifying, and organizing sen¬ physical environment, and the sociological limi¬
sory events, including perception of himself, and tations within which both patient and therapist
(2) changes in the response patterns which he have to operate. Therefore, the initial formula¬
has established in relation to social objects and tion will constantly be enlarged or changed, re¬
to himself over the years.11 In addition, the sulting either in confirmation of the previous
clinician's task may involve direct intervention therapeutic strategy or in its change.
in the patient's environmental circumstances, A Guide to a Functional Analysis of Indi¬
modification of the behavior of other people vidual Behavior.—In order to help the clinician
significant in his life, and control of reinforcing in the collection and organization of informa¬
stimuli which are available either through self- tion for a behavioral analysis, we have con¬
administration, or by contingency upon the be¬ structed an outline which aims to provide a
havior of others. These latter procedures working model of the patient's behavior at a
complement the verbal interactions of traditional relatively low level of abstraction. A series of
psychotherapy. They require that the clinician, questions are so organized as to yield immediate
at the invitation of the patient or his family, par¬ implications for treatment. This outline has been
ticipate more fully in planning the total life found useful both in clinical practice and in
pattern of the patient outside the clinician's teaching. Following is a brief summary of the
office. categories in the outline.*
It is necessary to indicate what the theoretical 1. Analysis of a Problem Situation: f The patient's
view here presented does not espouse in order to major complaints are categorized into classes of be¬
havioral excesses and deficits. For each excess or deficit
understand the differences from other proce¬ the dimensions of frequency, intensity, duration, ap¬
dures. It does not rest upon the assumption that propriateness of form, and stimulus conditions are

(a) insight is a sine qua non of psychotherapy, described. In content, the response classes represent
( b ) changes in thoughts or ideas inevitably lead the major targets of the therapeutic intervention. As
an additional indispensable feature, the behavioral
to ultimate changes in actions, {c) verbal thera¬
assets of the patient are listed for utilization in a
peutic sessions serve as replications of and therapy program.
equivalents for actual life situations, and {d) 2. Clarification of the Problem Situation : Here we
a symptom can be removed only by uprooting consider the people and circumstances which tend to
its cause or origin. In the absence of these as¬ maintain the problem behaviors, and the consequences
of these behaviors to the patient and to others in his
sumptions it becomes unnecessary to conceptual¬ environment. Attention is given also to the conse¬
ize behavior disorder in etiological terms, in
quences of changes in these behaviors which may
psychodynamic terms, or in terms of a specifi¬ result from psychiatric intervention.
able disease process. While psychotherapy by 3. Motivational Analysis : Since reinforcing stimuli
verbal means may be sufficient in some instances, are idiosyncratic and depend for their effect on a num¬

the combination of behavior modification in life ber of unique parameters for each person, a hierarchy
of particular persons, events, and objects which serve
situations as well as in verbal interactions serves
to extend the armamentarium of the therapist. *A limited supply of the full outline is available
and copies can be obtained upon request from us.
Therefore verbal psychotherapy is seen as an
t For each patient a detailed analysis is required.
adjunct in the implementation of therapeutic be¬ For example, a list of behavioral excesses may include
havior changes in the patient's total life pattern, specific aggressive acts, hallucinatory behaviors, cry-
not as an end in itself, nor as the sole vehicle ing, submission to others in social situations, etc. It is
for increasing psychological effectiveness. recognized that some behaviors can be viewed as
excesses or deficits depending on the vantage point
In embracing this view of behavior modifica¬
from which the imbalance is observed. For instance,
tion, there is a further commitment to a constant excessive withdrawal and deficient social responsive-
interplay between assessment and therapeutic ness, or excessive social autonomy (nonconformity)
strategies. An initial diagnostic formulation and deficient self-inhibitory behavior may be comple-
seeks to ascertain the major variables which can mentary. The particular view taken is of consequence
be directly controlled or modified during treat¬ because of its impact on a treatment plan. Regarding
certain behavior as excessively aggressive, to be re-
ment. During successive treatment stages addi¬ duced by constraints, clearly differs from regarding
tional information is collected about the patient's the same behavior as a deficit in self-control, subject
behavior repertoire, his reinforcement history, to increase by training and treatment.

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pittsburgh User on 05/01/2016

as rein forcers is established for each patient. Included The formulation of a treatment plan follows
in this hierarchy are those reinforcing events which from this type of analysis because knowledge
facilitate approach behaviors as well as those which, of the reinforcing conditions suggests the moti¬
because of their aversiveness, prompt avoidance re¬
vational controls at the disposal of the clinician
sponses. This information has as its purpose to lay
for the modification of the patient's behavior.
plans for utilization of various reinforcers in pre¬
scription of a specific behavior therapy program for The analysis of specific problem behaviors also
the patient, and to permit utilization of appropriate provides a series of goals for psychotherapy or
reinforcing behaviors by the therapist and significant other treatment, and for the evaluation of treat¬
others in the patient's social environment.
ment progress. Knowledge of the patient's bio¬
4. Developmental Analysis : Questions are asked
about the patient's biological equipment, his sociocul¬ logical, social, and cultural conditions should
tural experiences, and his characteristic behavioral help to determine what resources can be. used,
development. They are phrased in such a way as (a) and what limitations must be considered in a
to evoke descriptions of his habitual behavior at treatment plan.
various chronological stages of his life, ( ) to relate The various categories attempt to call atten¬
specific new stimulus conditions to noticeable changes tion to important variables affecting the patient's
from his habitual behavior, and (c) to relate such
altered behavior and other residuals of biological and current behavior. Therefore, they aim to elicit
sociocultural events to the present problem. descriptions of low-level abstraction. Answers
5. Analysis of Self-Control : This section examines to these specific questions are best phrased by
both the methods and the degree of self-control describing classes of events reported by the
exercised by the patient in his daily life. Persons,
patient, observed by others, or by critical inci¬
events, or institutions which have successfully rein¬ dents described by an informant. The analysis
forced self-controlling behaviors are considered. The
deficits or excesses of self-control are evaluated in does not exclude description of the patient's
relation to their importance as therapeutic targets habitual verbal-symbolic behaviors. However, in
and to their utilization in a therapeutic program. using verbal behaviors as the basis for this
6. Analysis of Social Relationships : Examination analysis, one should be cautious not to "explain"
of the patient's social network is carried out to evalu¬ verbal processes in terms of postulated internal
ate the significance of people in the patient's environ¬
ment who have some influence over the problematic
mechanisms without adequate supportive evi¬
behaviors, or who in turn are influenced by the patient dence, nor should inference be made about
for his own satisfactions. These interpersonal relation¬ nonobserved processes or events without corrob¬
ships are reviewed in order to plan the potential par¬ orative evidence. The analysis includes many
ticipation of significant others in a treatment program, items which are not known or not applicable
based on the principles of behavior modification. The
review also helps the therapist to consider the range of
for a given patient. Lack of information on some
actual social relationships in which the patient needs items does not necessarily indicate incomplete¬
to function. ness of the analysis. These lacks must be noted
7. Analysis of the Social-Cultural-Physical En¬ nevertheless because they often contribute to the
vironment : In this section we add to the preceding better understanding of what the patient needs
analysis of the patient's behavior as an individual, to learn to become an autonomous person. Just
consideration of the norms in his natural environment.
as important is an inventory of his existing
Agreements and discrepancies between the patient's
idiosyncratic life patterns and the norms in his en¬ socially effective behavioral repertoire which
vironment are defined so that the importance of these can be put in the service of any treatment pro¬
factors can be decided in formulating treatment goals cedure.
which allow as explicitly for the patient's needs as for This analysis is consistent with our earlier
the pressures of his social environment.
formulations of the principles of comprehensive
The preceding outline has as its purpose to medicine 9*22 which emphasized the joint opera¬
achieve definition of a patient's problem in a tion of biological, social, and psychological
manner which suggests specific treatment op¬
factors in psychiatric disorders. The language
erations, or that none are feasible, and specific and orientation of the proposed approach are
behaviors as targets for modification. There¬ rooted in contemporary learning theory. The
fore, the formulation is action oriented. It can conceptual framework is consonant with the
be used as a guide for the initial collection of view that the course of psychiatric disorders
information, as a device for organizing avail¬ can be modified by systematic application of
able data, or as a design for treatment. scientific principles from the fields of psychology

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pittsburgh User on 05/01/2016

 and medicine to the patient's habitual mode of home or having had homosexual experiences)
living. lies in its relation to the independent variables
This approach is not a substitute for assign¬ which can be identified as controlling the cur¬
ment of the patient to traditional diagnostic rent problematic behavior. The observation that
a patient has hallucinated on occasions may be
categories. Such labeling may be desirable for
statistical, administrative, or research purposes. important only if it has bearing on his present
But the current analysis is intended to replace problem. If looked upon in isolation, a report
other diagnostic formulations purporting to about hallucinations may be misleading, result¬
serve as a basis for making decisions about spe¬ ing in emphasis on classification rather than
cific therapeutic interventions. treatment.
In the psychiatric interview a behavioral-
III. Methods of Data Collection for a analytic approach opposes acceptance of the
Functional Analysis content of the verbal self-report as equivalent to
Traditional diagnostic approaches have actual events or experiences. However, verbal
utilized as the main sources of information the reports provide information concerning the pa¬
tient's verbal construction of his environment
patient's verbal report, his nonverbal behavior and of his person, his recall of past experiences,
during an interview, and his performance on and his fantasies about them. While these self-
psychological tests. These observations are
sufficient if one regards behavior problems only descriptions do not represent data about events
as a property of the patient's particular pattern
which actually occur internally, they do repre¬
sent current behaviors of the patient and indi¬
of associations or his personality structure. A
cate the verbal chains and repertoires which the
mental disorder would be expected to reveal it¬
self by stylistic characteristics in the patient's patients has built up. Therefore, the verbal be¬
behavior repertoire. However, if one views be¬ havior may be useful for description of a pa¬
havior disorders as sets of response patterns tient's thinking processes. To make the most of
which are learned under particular conditions such an approach, variations on traditional in¬
and maintained by definable environmental and terview procedures may be obtained by such
internal stimuli, an assessment of the patient's techniques as role playing, discussion, and in¬
behavior output is insufficient unless it also de¬ terpretation of current life events, or controlled
free association. Since there is little experi¬
scribes the conditions under which it occurs.
mental evidence of specific relationships be¬
This view requires an expansion of the
tween the patient's verbal statements and his
clinician's sources of observations to include the
nonverbal behavioral acts, the verbal report
stimulation fields in which the patient lives, and
alone remains insufficient for a complete analy¬
the variations of patient behavior as a function
sis and for prediction of his daily behavior.
of exposure to these various stimulational
variables. Therefore, the resourceful clinician
Further, it is well known that a person responds
to environmental conditions and to internal cues
need not limit himself to test findings, interview
which he cannot describe adequately. There¬
observations in the clinician's office, or referral
histories alone in the formulation of the specific fore, any verbal report may miss or mask the
most important aspects of a behavioral analysis,
case. Nor need he regard himself as hopelessly
ie, the description of the relationship between
handicapped when the patient has little observa¬ antecedent conditions and subsequent behavior.
tional or communicative skill in verbally recon¬
In addition to the use of the clinician's own
structing his life experiences for the clinician.
Regardless of the patient's communicative skills person as a controlled stimulus object in inter¬
the data must consist of a description of the pa¬ view situations, observations of interaction with
tient's behavior in relationship to varying en¬ significant others can be used for the analysis of
vironmental conditions. variations in frequency of various behaviors as
A behavioral analysis excludes no data re¬ a function of the person with whom the patient

lating to a patient's past or present experiences interacts. For example, use of prescribed stand¬
as irrelevant. However, the relative merit of ard roles for nurses and attendants, utilization
any information (as, eg, growing up in a broken of members of the patient's family or his

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pittsburgh User on 05/01/2016

 friends, may be made to obtain data relevant to brief samples of patient behaviors in work
the patient's habitual interpersonal response pat¬ situations can be obtained by minimally trained
tern. Such observations are especially useful if observers.
in a later interview the patient is asked to The patient himself may be asked to provide
describe and discuss the observed sessions. samples of his own behavior by using tape re¬
Confrontations with tape recordings for com¬ corders for the recording of segments of inter¬
parisons between the patient's report and the actions in his family, at work, or in other
actual session as witnessed by the observer may situations during his everyday life. A television
provide information about the patient's per¬ monitoring system for the patient's behavior is
an excellent technique from a theoretical view¬
ception of himself and others as well as his
habitual behavior toward peers, authority point but it is extremely cumbersome and ex¬
figures, and other significant people in his life. pensive. Use of recordings for diagnostic and
Except in working with children or family therapeutic purposes has been reported by some
units, insufficient use has been made of material investigators.2,5·24 Playback of the recordings
obtained from other informants in interviews and a recording of the patient's reactions to the
about the patient. These reports can aid the ob¬ playback can be used further in interviews to
server to recognize behavioral domains in which clarify the patient's behavior toward others and
the patient's report deviates from or agrees with his reaction to himself as a social stimulus.
the descriptions provided by others. Such in¬ Psychological tests represent problems to be
formation is also useful for contrasting the pa¬ solved under specified interactional conditions.
tient's reports about his presumptive effects on Between the highly standardized intelligence
another person with the stated effects by that tests and the unstructured and ambiguous pro¬
person. If a patient's interpersonal problems jective tests lies a dimension of structure along
extend to areas in which social contacts are not which more and more responsibility for pro¬
clearly defined, contributions by informants viding appropriate responses falls on the pa¬
other than the patient are essential. tient. By comparison with interview procedures,
It must be noted that verbal reports by other most psychological tests provide a relatively
informants may be no more congruent with greater standardization of stimulus conditions.
actual events than the patient's own reports and But, in addition to the specific answers given on
need to be equally related to the informant's intelligence tests or on projective tests these
own credibility. If such crucial figures as tests also provide a behavioral sample of the pa¬
parents, spouses, employers can be so inter¬ tient's reaction to a problem situation in a rela¬
viewed, they also provide the clinician with some tively stressful interpersonal setting. Therefore,
information about those people with whom the psychological tests can provide not only quanti¬
patient must interact repeatedly and with whom tative scores but they can also be treated as a
interpersonal problems may have developed. miniature life experience, yielding information
Some observation of the patient's daily work about the patient's interpersonal behavior and
behavior represents an excellent source of in¬ variations in his behavior as a function of the
formation, if it can be made available. Observa¬ nature of the stimulus conditions.
tion of the patient by the clinician or his staff In this section we have mentioned only some
of the numerous life situations which can be
may be preferable to descriptions by peers or

supervisors. Work observations are especially evaluated in order to provide information about
important for patients whose complaints include the patient. Criteria for their use lies in econ¬
difficulties in their daily work activity or who omy, accessibility to the clinician, and relevance
describe work situations as contributing factors to the patient's problem. While it is more con¬
to their problem. While freer use of this tech¬ venient to gather data from a patient in an
nique may be hampered by cultural attitudes office, it may be necessary for the clinician to
toward psychiatric treatment in the marginally have first-hand information about the actual
adjusted, such observations may be freely ac¬ conditions under which the patient lives and
cessible in hospital situations or in sheltered works. Such familiarity may be obtained either
work situations. With use of behavior rating by utilization of informants or by the clinician's
scales or other simple measurement devices, entry into the home, the job situation, or the

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pittsburgh User on 05/01/2016

 social environment in which the patient lives. The last section of the paper deals with
Under all these conditions the clinician is effec¬ methods useful for collection of information
tive only if it is possible for him to maintain a necessary to complete such a behavior analysis.
This paper was written in conjunction with Re¬
nonparticipating, objective, and observational search
grant MH 06921-03 from the National Institutes
role with no untoward consequences for the pa¬ of Mental
Health, United States Public Health Service.
tient or the treatment relationship.
The methods of data collecting for a func¬
tional analysis described here differ from tradi¬ REFERENCES
tional psychiatric approaches only in that they
require inclusion of the physical and social
stimulus field in which the patient actually
operates. Only a full appraisal of the patient's
living and working conditions and his way of
life allow a description of the actual problems
which the patient faces and the specification of
steps to be taken for altering the problematic
situation.
Summary
Current psychiatric classification falls short
of providing a satisfactory basis for the under¬
standing and treatment of maladaptive behavior.
Diagnostic schémas now in use are based on
etiology, symptom description, or prognosis.
While each of these approaches has a limited
utility, no unified schema is available which per¬
mits prediction of response to treatment or
future course of the disorder from the assign¬
ment of the patient to a specific category.
This paper suggests a behavior-analytic ap¬
proach which is based on contemporary learning
theory, as an alternative to assignment of the
patient to a conventional diagnostic category. It
includes the summary of an outline which can
serve as a guide for the collection of informa¬
tion and formulation of the problem, including
the biological, social, and behavioral conditions
which are determining the patient's behavior.
The outline aims toward integration of informa¬
tion about a patient for formulation of an action
plan which would modify the patient's problem¬
atic behavior. Emphasis is given to the partic¬
ular variables affecting the individual patient
rather than determination of the similarity of
the patient's history or his symptoms to known
pathological groups.

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pittsburgh User on 05/01/2016