You are on page 1of 14

Adv. Behav. Res. Ther. Vol. 1, pp. 217-230 0146-6402/78/1201-0217$05.

00/0
@PergamonPressLtd. 1978.Printed in Great Britain.

THE IMPORTANCE OF BEING THEORETICAL:


A COMMENTARY ON BANDURA’S
“SELF-EFFICACY: TOWARDS A UNIFYING THEORY
OF BEHAVIORAL CHANGE”

G. Terence Wilson
Graduate Schoolof Applied and ProfessionalPsychology,RutgersUniversity,Piscataway,NJ08854.U.S.A.
(Received My 1978)

Abstract - Theory is an important but relativelyneglectedaspecl of contemporary behavior


therapy. To be sckntifially usefi& a theory should be capable of integrating available facts in an
internally consistentmanner, amenabk to experimentaltests,heuristicin directingthe searchfor
new knowledge.and comparefavorablywith existingtheories.ImprovedthzOryin behaviortherapy
would seemto be especiallyhelpful in infusing the field with freshconceptsand proceduresand in
Possiblybringinggreaterorder to the proliferation of incrcasin~ydivergentconceptsand methods.
Sclf-&icacytheoryisevaluatedin the light of thesedifTcrentcriteria.It isconcludedthat sclf-cffkacy
theory is testable and rich in heuristicvalue. Its implicationsfor broader conceptual issues, spccifx
researchquestions,behavioralassessment,and the generalizationand maintenanceof treatment
produced change are briefly discussed. Finally, self-efficacytheory is compared to alternative
conditioning explanationsof fear reduction methods.

Behavior therapy was once defined as the application of “modem ‘learning theory” to
clinical disorders (Eysenck, 1959). Critics soon delighted in pointing out that there was no
single, unified learning theory; that the links between laboratory conditioning research and
clinical practice were often tenuous at best; and that the use of the terminology of
conditioning theory to describe procedures and explain causal processes was more
metaphorical than paradigmatic (e.g. Breger and McGaugh, 1965; London, 1972).
Although some defenses about behavior therapy being derived from conditioning principles
were offered (e.g. Eysenck, 1970; Wolpe, 1969), the field ‘developed and diversified to the
point where to&y simplistic definitions in terms of “learning the&y” are outdated and
widely rejected. Contemporary behavior therapy is marked by a lively diversity of views, a
broad range of heterogeneous procedures with different theoretical rationales, different
research strategies, and open debate about conceptual bases, methodological requirements
and evidence of efficacy. In short, there is no clearly agreed upon or commonly accepted
definition of behavior therapy. Broadly conceived of, the different approaches in
contemporary behavior therapy include applied behavior analysis, a neobehavioristic
mediational model, social learning theory, cognitive behavior modification, and
multimodal behavior therapy (Agras et al., 1977; Kazdin and Wilson, 1978).
In the 20 years since Eysenck (1959) and Wolpe (1958) formally derived behavior therapy
from learning theory, the field has experienced remarkable growth. Once a beleaguered
minority movement, behavior therapy has been elevated to an accepted part of the
therapeutic establishment (e.g. American Psychiatric Association, 1973; Risley, 1977).
Initially applied to a limited range of disorders, behavioral interventions have been extended
to a bewildering array of diverse psychological, psychiatric, medical, educational, industrial
and community settings (e.g. Franks and Wilson, 1973, 1976, 1977; Kazdin and Wilson,
1978; Leitenberg, 1976b; Williams and Gentry, 1977) and have shown impressive efficacy in
many of these instances (e.g. Kazdin and Wilson, 1978; Rachman and Wilson, (in press)).
This growth has been most evident in the development of new techniques and the continuing
refinement of existing strategies in the clinicalpractice and evaluation of behavior therapy
(e.g. Azrin, 1977; Goldfried and Davison, 1976). There can be little doubt that the
development of well-defined, innovative therapeutic techniques for specific problems has
contributed greatly to whatever clinical popularity behavior therapy enjoys. Methodo-
logically, .the evaluation of the outcome of behavioral methods has become increasingly
sophisticated. New applied research strategies have been pioneered (e.g. Hersen and
217
218 G. T. Wilson

Barlow, 1976; Kazdin, (in press)) while the amount of research continues to burgeon.
Behavioral research and therapy are alive and well.
But what about theory? In part because of the often justifiable criticism of narrow
conceptions of behavior therapy in terms of conditioning principles and in the course of the
technological expansion and broadening applicability of behavioral methods, the
importance of theory has often been lost sight of, ignored or even denounced as the clinically
irrelevant plaything of the ivory-tower-dwelling academic. A commitment to a squarely
pragmatic, technically eclectic therapeutic stance with little if any concern for theoretical
niceties characterizes the positions of so varied a group of influential figures in behavior
therapy as Azrin (1977), Lazarus (1977) and Marks (1976), among others. Unquestionably,
the therapeutic yield from this sort of pragmatic approach has often been rich, Given the
current incomplete state of our art, the practitioner frequently has little choice but to adopt a
flexible, pragmatic approach. However, acknowledging the exigencies of clinical practice
must not undermine a recognition of the potentially important role theory might play in
directing clinical research and in improving upon current therapeutic strategies.

THE IMPORTANCE OF THEORY

Why is a concern about theory important in contemporary behavior therapy? Theory


provides a means of integrating and interpreting available information and influences the
search for new knowledge. Confronted with the increasingly apparent multiplicity of
methods and concepts encompassed by the term “behavior therapy” (cf. Wilson, (1978), a
useful means of ordering the evidence, establishing priorities and filtering the good from the
bad would be most useful. Whether viewed as loose metaphors or scientific guidelines, it is
now widely accepted that the principles of classical and operant conditioning have been
extremely successful in stimulating a wealth of experimental/clinical research and
innovative therapeutic techniques over the past two decades. However, it can be argued that
the heuristic value of this learning theory has reached asymptote. There is a sense of
conceptual staleness in the still often useful but frequently repetitive or trivial extension of
existing conditioning methods to an ever wider range of different behaviors. Replication and
the empirical consolidation of advances already made are a sine qua non of an applied
science. But a cognitive map that promises to open up new territory is needed to replace the
once vibrant but now creaky guideposts of “modem learning theory.”
Scientific theory is also invaluable in making explicit the conceptual biases of a particular
approach. Formulated so that its assumptions are testable and its procedures replicable, a
theory makes it possible to refine, modify or scrap pet therapeutic notions. Ideally, rival
theoretical positions would be unnecessary if there were consensus on experimentally
validated methods and outcomes. But this is a distant goal. In the meantime, unable to rely
exclusively upon an empirically validated body of effective techniques, the therapist has to
draw treatment strategies from some additional source. * In short, the appeal is either to an
organized conceptual framework or a melange of personal preference, intuition and
subjective judgement. Although the latter option will undoubtedly be helpful to some clients
in certain situations, behavior therapy has always been an attempt to improve upon this
state of affairs by developing an applied science of clinical behavior change. Treatment
strategies must be distinguishable from the therapists who apply them and must be
communicable to other therapists for whom they are also effective. If not, it is impossible to
determine what the critical therapeutic elements are and what is irrelevant (or even harmful)
and better discarded. The self-corrective process is the very essence of the scientific
approach and is greatly facilitated by systematically relating treatment strategies to a
consistent theoretical framework.
Some behaviorists, notably applied behavior analysts, have characterized behavior
modification primarily as a methodologicai approach to personal and social problems (e.g.

*The banality of the frequently heard clinician’s retort, “I use whatever works”, should be apparent from this
circumstance.
Towards a Unifying Theory of Behavioral Change 219
Baer et al., 1968). It is doubtful, however, that methodological criteria alone can define
behavior therapy or provide a basis for the choice of therapeutic strategies in ylinical
practice. One need only examine the theoretical differences between different approaches
within the current compass of behavior therapy. Although they share common
methodological priorities and prize rigorous experimental evaluation of methods, applied
behavior analysis and social learning theory, for example, differ with respect to their
conceptualizations of a model of human nature, causal processes in psychological
functioning, and their implications for a technology of behavior change (e.g. Bandura, 1976,
1977b, (in press); Catania, 1975, 1976: Goldiamond, 1976; Mahoney, 1974). It is a fact that
theoretical issues and postures help defme behavior therapy and influence basic and applied
research. The inescapable influence of theory (explicit or otherwise) on therapy is noted by
Franks (1969): “How the behavior therapist practices (including his choice of techniques, his
approach to the problems of genera/strategy and his specific relationships with his patient) thus
depen& both upon his explicit theoretical orientation and upon his impIicit philosophical and
cultural milieu” (p. 21).
Despite what at first blush appears to be an anti-theoretical orientation, even the doctrine
of technical eclecticism is far from unfettered by implicit theoretical constraints. Thus
Lazarus (1971) has noted that the therapist’s choice of techniques is determined by “well-
conceived guiding principles and general theories of human behavior” (p. 39). The problem
with this position is that it is not always clear what these guiding principles are (cf. Franks
and Wilson, 1975,1976). Similarly, Azrin (1977), in making the case for the insufficiency of
reinforcement concepts as the basis for an outcome-oriented therapeutic approach, goes
beyond the strict operant conditioning approach in advocating the use of “emergent
principles” of behavior change. Precisely how these principles emerge from reinforcement
theory and what their boundary conditions are, among other questions, remain to be shown.
The clarification of these theoretical issues obviously has direct relevance to therapeutic
practice. With current behavioral approaches increasingly embracing what at times appear
to be divergent methods, conflicting concepts, rival rationales and apparently lacking a
uniform philosophy, one is reminded of the adage that there is nothing quite so practical as a
good theory that might bring order to what may at times appear to be confusion.
Although he accepts the scientific importance of theory in the general case, London
(1972) nevertheless advocated the end of ideology in behavior modification. His premise
was twofold: (a) current theory in behavior modification was not Very good, and (b) no
theory was useful for the development of the field at that time. In contrast to London’s
(1972) thesis the present paper proposes that Bandura’s (1977a) self-efficacy theory is an
example of the sort of theory that is (a) scientifically valuable; and (b) is extremely timely
and useful for the development of present-day behavior therapy.

CRITERIA FOR A SCIENTIFIC THEORY

Briefly, a scientific theory should (a) accommodate and integrate the currently known
facts of behavior change; (b) be stated so that it is testable - experiments can support, or
more importantly, d&confirm it; (c) it should be heuristic in stimulating novel research and
prompting therapeutic innovations; and (d) it should compare favorably with existing
theoretical alternatives. The following is a summary evaluation of Bandura’s (1977a) self-
efficacy theory in terms of each of these criteria.

Does it integrate known facts?

The current conceptual models of behavior therapy can each 3e said to place primary
emphasis on one dimension of psychological functioning to the relative neglect of the others.
Thus applied behavior analysis focuses almost exclusively on observable, overt behavior as
typified by the Journal of Applied Behavior Analysis. In his neo-behavioristic, S-R
220 G. T. Wilson

counterconditioning approach, Wolpe (1976) emphasizes emotional habits or conditioning


of the autonomic nervous system. Most recently, cognitive (behavior) therapy has focused
pre-eminently on the causal role of maladaptive thought patterns in behavioral and
emotional disorders (e.g. Beck, 1976; Mahoney and Arnkoff, (in press); Meichenbaum,
1977). One of the advantages of Bandura’s (1977b) social learning approach in general and
self-efficacy theory in particular is that they integrate these three regulatory systems of
antecedent, consequent, and mediational influence in a comprehensive framework (see
Wilson, (in press) (b)).
Lazarus (1976) too has proposed a framework that goes beyond a primary emphasis on
one dimension of psychological functioning. In multimodal behavior therapy clinical
problems are analyzed in terms of seven separate but interactive modalities: Behavior,
Affect, Sensation, Imagery, Cognition, Interpersonal behavior and Drugs. Therapy then
focuses on specific problems within each modality. But what are the criteria for the choice of
different techniques under specifiable conditions? In the absence of the necessary guiding
principles diverse methods from disparate theoretical systems are likely to be selected
arbitrarily on the basis of subjective judgment. Lazarus (1976) states that these seven
modalities are “interdependent and interactive” and “not every case requires attention to
each modality”. There can be little disagreement on this point but the critical questions that
remain to be answered are what is the nature of the interactive process among different
modalities and what are the ramifications for maximally efficient therapeutic intervention?
It is suggested here that the answers are most likely to be developed within the context of a
consistent theoretical framework that yields precise predictions that are amenable to
operational measurement and evaluation. self-efficacy theory provides one such framework
for elucidating the interdependence between cognitive and behavioral change. The inclusion
of appropriate psychophysiological measures in subsequent’studies of self-efficacy theory
should indicate how this dimension of psychological functioning interacts with cognitive
and behavioral determinants.
Self-efficacy theory exemplifies the integrative power of the social learning approach in
several ways that may be mentioned briefly. Whereas self-efficacy theory postulates a
cognitive mechanism to explain the effects of diverse psychological procedures on the
initiation, generalization and maintenance of therapeutic change, the most powerful means
of producing that change are performance-based methods. The importance of this
distinction between treatment procedure and theoretical process must be underscored. By
ignoring the role of cognitive processes, strict behavioristic approaches needlessly delimit
the scope and efficacy of their treatment methods. By failing to emphasize the superior
efficacy of behavioral methods in altering cognitive processes, cognitive therapists run the
risk of overlooking the most effective therapeutic techniques yet developed and returning
therapy to a verbal, interview-based model of treatment in opposition to which behavior
therapy was originally developed. Although the more cautious cognitive therapists are
aware of this danger (e.g. Mahoney and Amkoff, (in press)), recognition of the primacy of
behavioral methods is far from evident in much of the rapidly emerging literature on
cognitive (behavior) therapy. In Elfis’s (1970) rational emotive therapy (RET), for example,
abnormal behavior is viewed as nothing but a question of irrational, maladaptive
cognitions. Moreover, the treatment methods emphasized most heavily in RET are
cognitive in nature - verbal persuasion, rational argument and logical reasoning. Cognitive
explanations of behavior that bear some similarity to the notion of self-efficacy are not new
in psychology. The consistent flaw in these previous cognitive theories has been the failure to
detail a comprehensive causal analysis of behavior. Cognitive constructs have been
postulated to control behavior but the determinants of those cognitive constructs have not
been adequately specified. In contrast to these incomplete causal analyses, Bandura (1977a)
suggests what the determinants of self-efficacy are. Accordingly, this allows for their direct
modification by behavioral means. The interactive process between these determinants and
self-efficacy is one of reciprocal determinism (see below). Bandura’s (1977a) theory enjoys
another advantage. Unlike previous cognitive explanations, self-efficacy theory does not
postulate a static or global personality construct. Consistent with recent conceptualizations
of personality (cf. Mischel, 1973), efficacy expectations may be situation specific and covary
with what the person does in particular life situations.
Towards a Unifying Theory of Behavioral Change 221

Is it testable?

The theory generates specific predictions concerning the initiation, generalization and
maintenance of behavior change. Testing the theory requires measurement of self-efftcacy
expectations that is independent of behavioral performance. Subjects’ self-reports of
perceived efficacy define this independent measure. It is in this connection that problems in
testing the generalizability of the theory to different behaviors across different conditions
might arise. *
In testing predictions from self-efficacy theory, as Bandura (1977a) observes, it is
important that subjects have a clear understanding and full knowledge of their performance
requirements and the situational circumstances. To the extent that these are ill-defined,
discrepancies between behavior and efficacy expectations are likely to occur. The tightly
controlled laboratory conditions under which Bandura has tested the theory satisfy these
conditions of testing. Subjects knew exactly what they had to do under predictable
circumstances, i.e. they were knowledgeable about the nature of the feared object (the
snake). Even here, however, Bandura et al. (1977) note that while efficacy expectations were
accurate predictors of approach behavior, there was more of a discrepancy between
expectation and behavior during the generalization test with the “unfamiliar snake” than
the test with the familiar snake.
Interestingly, from a measurement point of view, this issue highlights the whole question
of obtaining valid self-reports, an increasingly important concern as we acknowledge the
necessity of multiple response measures instead of a narrow reliance on non-verbal overt
behavior (e.g. Rachman and Hodgson, 1974). At a minimum, accurate self-report will
depend upon a knowledge of the specific cues that govern the target behavior and the
perceived consequences that the report has on future behavior. This knowledge will
necessarily be based on exposure to the relevant cues. While this can often be accomplished
symbolically by having the person imagine the feared situation, there will be times when
only actual exposure to the feared situation will suffice. Beck (1976), for example, in
discussing what he calls the client’s “subjective estimate of the probabilities of harm”, notes
that they shift as a function of exposure to the feared situation. Under protective
circumstances removed from the threatening situation (e.g. the trusted therapist’s “safe”
offrce) the client can readily tigree that the probability of harm in approaching the feared
situation is minimal., Even imagery of experiencing the threatening situation might evoke
little or no anxiety. However, the closer the client comes to making actual contact with the
feared situation, the greater the self-reported estimates of fear.
To summarize, testing the predictions of self-efficacy theory will require specific measure
of well-defined responses in relation to particular situations. While practical difficulties
might be encountered in meeting these conditions, the task does not appear to be
insuperable. After all, this is what in principle behavioral assessment is all about.

Is it heuristic?

Self-efficacy theory has implications for broader conceptual schemes of human behavior,
basic and applied research, and clinical practice. It is in this domain of stimulating and
guiding the discovery of novel facts and treatment strategies that Bandura’s (1977a) theory
is likely to prove invaluable.

Conceptual considerations. Conceptually, self-efficacy theory provides a specific example


of the reciprocal determinism model of causal processes in human behavior (cf. Bandura, (in
press)). In rejecting the psychoanalytic model of largely autonomous psychic forces, some
behavior modifiers sought refuge in a similarly unidirectional causal model of another kind
- radical behaviorism.

*This is not a criticism of so-called “analogue ” research, the value of which is discussed by Kazdin and Wilson
(1978). Controlled laboratory research of the sort reported by Bandura (1977a) is an appropriate means of pilot
testing and developing a new theory of behavior change.
222 G. T. Wilson

The radical behaviorism of the Skinnerian or strict operant approach is essentially a


unidirectional causal model because the environment is held to be the ultimate determinant
of behavior. As Skinner (1971) bluntly put it, “A person does not act upon the world the world
acts upon him” (p. 211). Although the operant approach is often described as an
interactionist position, true interdependence among person variables, behavior, and the
environment requires that the person be the agent as well as the object of environmental
change. The importance of self-directed change is a distinguishing characteristic, of the
reciprocal determinism model. In the unidirectional operant analysis there cannot be self-
control, only situational control of apparent self-directed acts (Rachlin, 1974; Stuart, 1972).
Self-efftcacy theory is an illustration of a social learning model in which the influence of
environmental events on behavior is partly mediated by internal cognitive mechanisms.
These cognitive processes help determine what is attended to, how it is centrally processed,
whether it will be remembered, and how it affects efficacy and outcome expectations. The
causal process in psychological functioning involves a continuous interdependence among
behavior, mediational processes, and environmental factors. Thus a person’s efficacy
expectations will affect how (s)he behaves and the environmental consequences of that
behavior will in turn influence personal expectations and behavior. There is no inevitable
prime mover in behavior although the relative influence of these separate but interactive
regulatory factors will vary across individuals and situations. As a model of causal processes
in human behavior, reciprocal determinism, aside from being more palatable than
environmental tyranny be it benign or not, promises to be more heuristic in making sense of
the complex interpersonal interactions that constitute psychotherapy (cf. Wilson and
Evans, 1976, 1977).

Research ramifications. Theoretical views, whether macfe explicit or not, exercise a


powerful influence on what phenomena are investigated and how they are researched.
Relegating private events to the role of epiphenomena or non-causal correlates of overt
behavior, radical behaviorists have shied away from the study of cognitive processes despite
increasing evidence of their seminal influence on behavior (e.g. Bandura, 1977b; Bower,
1977; Brewer, 1974; Mahoney, 1974; Meichenbaum, 1977; Mischel, 1973). Self-efftcacy
theory should help redress this reciprocal interaction between behavior and one class of
cognitive events - expectations.
Although controversial, the concept of “expectancy” is hardly new in behavior therapy.
A prominent example of this has been the continuing debate over the role of expectations in
the process responsible for the success of systematic desensitization (e.g. Borkovec, 1973;
Davison and Wilson, 1973; Kazdin and Wilcoxon, 1976). As a whole, expectancy-inducing
instructions have had little systematic effect on the outcome of systematic desensitization
(Franks and Wilson, 1976; Lick and Bootzin, 1975). However, the conceptual analysis that
has directed this body of research has been lacking. Thus Bandura (1977a) points out that
expectations cannot be analyzed globally as a static, unidimensional factor. Account must
be taken of the fact that expectations might vary across dimensions such as magnitude,
generality and strength. In calling for a more specific yet multidimensional analysis of the
effects of expectations, Bandura (1977a) is bringing research on expectations up-to-date
with the more sophisticated sort of analysis that is now customary with other classes of
psychological variables in behavior therapy. The consequences promise to be empirically
rewarding.
One of the ways in which Seligman’s (1975) theory of learned helplessness has proved
useful has been in sparking an explosion of derivative experimental research. Self-efficacy
theory should serve a similar function. Indeeed, it will be interesting to determine the
possible impact of self-efficacy theory on the concept of and the research on learned
helplessness. Seligman (1975) has cited Melges and Bowlby’s (1969) views on hopelessness in
psychopathological processes as similar to his own theory of depression. In fact, although
they never developed the distinction, Melges and Bowlby (1969) distinguished between what
one could call efficacy and outcome expectations: “Our thesis is that while a depressed
patient’s goals remain relatively unchanged his estimate of the likelihoodof achieving them and
his confidence in the efficacy of his own skilledactions are both diminished. . . ” (underscoring
added).
Towards a Unifying Theory of Behavioral Change 223
In Bandura’s (1977a) experiments with phobics, information about the feared situation
reliably altered efficacy expectations in a predictable manner. But what if objectively
dependable information is not appraised realistically or accurately? Depression, for
example, is a disorder in which the central problem may be the client’s consistently negative
interpretation of life’s experiences. In Beck’s (1976) cognitive theory of depression the
various emotional, motivational and behavioral phenomena of depression are the direct
result of negative self-evaluations. In terms of Bandura’s (1977a) theory, the cognitive
processing of efficacy information has gone awry. The nature and functions of this cognitive
processing remain to be delineated. For example, are there critical events that increase or
decrease efficacy expectations? Is it easier to restore rather than create efftcacy expectations
in clients? Are faulty interpretations of experience a matter of cognitive deficits or
distortions? Attributional processes, to which Bandura (1977a, p. 201) briefly alludes, and
related lines of research from cognitive social psychology are likely to feature prominently in
subsequent investigations of these issues. Happily, despite our lack of understanding of
these basic cognitive processes, Beck and his colleagues have developed therapeutic
strategies for the treatment of depression that have yielded unprecedented success (Rush et
al., 1977). This treatment method, a behaviorally-based program tailored to modifying the
way in which clients perceive themselves and their respective worlds, dovetails with the
assumptions and therapeutic implications of the self-efficacy theory.

Clinical implications. A comprehensive behavioral assessment of clinical problems should


include a systematic analysis of the client’s efficacy expectations. One possibility is that
estimates of self-efficacy might provide a useful predictor variable of therapeutic outcome.
As a fairly typical case consider the behavior treatment of obesity. Despite intensive
research efforts, a reliable predictor of successful outcome has yet to be identified (Stunkard
and Mahoney, 1976; Wilson, (in press) (c)). Since behavioral treatment programs for obesity
are predicated on the development of improved self-control over excessive eating behavior,
frequent attempts have been made to relate outcome to personality variables such as
Rotter’s (1966) locus of control. Not surprisingly in view of the fate of most global
personality concepts, these attempts have been unsuccessful (e.g. Tobias and MacDonald,
1977). Yet the not implausible notion persists that the client’s causal attributions about
therapeutic progress may be‘a decisive influence on maintenance of treatment produced
improvement (e.g. Jeffrey, 1974). Both Green (1976) and Stuart ‘and Guire (1978), for
example, reported significant positive correlations between measures of clients’ perceived
control over coping with food and weight loss. The greater specificity of efficacy
expectations compared with the locus of control construct and the fact that they tap
personal efftcacy rather than performance-outcome control (see Reid and Ware (1974) for a
related distinction) may increase the chances of discovering a predictor variable.
Perhaps the most important prediction that derives from Bandura’s (1977a) theory is that
the strength of efftcacy expectations will determine whether the client will initiate coping
behavior, what form it will take, and how long it will be maintained in the face of obstacles
and adverse experiences. As such, this prediction is addressed directly to the Achilles heel of
behavior therapy - the generalization and maintenance of behavior change. An example of
how self-efftcacy theory may bear on the question of long-term maintenance of therapeutic
improvement is to be seen in connection with the well-known phenomenon of relapse in
clients with addictive disorders. As I have discussed more fully elsewhere (cf. Wilson, (in
press) (a)), Bandura’s (1977a) theory provides a plausible and testable framework for
reconceptualizing some of the processes governing relapse phenomena and a basis for
effective intervention with lasting therapeutic effects.
In the case of alcohol abuse for instance, most treatment programs are steeped in the
traditional disease model alcoholism. The expectations that are created by this sort of
therapeutic philosophy can vitally affect relapse. These include the belief that the alcoholic
has an irreversible disease, that (s)he is qualitatively different from non-alcoholics, that (s)he
is uniquely vulnerable to the effects of alcohol, and that (s)he will never be able to exercise
voluntary control over consumption once drinking has been initiated. Recast in terms of
self-efficacy theory, the alcoholic’s efficacy expectations about coping with alcohol are
deliberately minimized. The outcome expectations that are developed emphasize the
224 G. T. Wilson

certainty of a return to uncontrollable drinking after a single drink. As a result, the often
demoralizing abruptness with which a client who has been abstinent even for a lengthy
period of time can revert to addictive drinking is explicable on grounds other than the
biological inevitability of “first drink then drunk”. Firstly, the fact that the client has been
abstinent even for a period of years does not necessarily mean that (s)he has developed a
sense of self-efficacy about alcohol. Secondly, the corollary of minimal or nonexistent
efficacy expectations is that coping behavior in the face of the inevitable difficulties the client
experiences in attempting to remain sober in the natural environment will be easily
extinguished. Finally, the outcome expectation of “first drink then drunk’* can function as a
self-fulfilling prophecy. Of course, expectations alone cannot account for the reasons for
relapse in all alcoholics, A complete social learning analysis requires consideration, among
other factors, of the individual’s incentives for sobriety and his or her requisite behavioral
skills for coping with the interpersonal difficulties of returning to continued sobriety: Given
the necessary behavioral skills and adequate incentives, however, efficacy expectations may
be a major determinant of the maintenance of treatment-produced improvement.
As for treatment, both symbolic and performance-based methods can be used to develop
realistic efficacy expectations that will minimize the likelihood of a return to addictive
drinking following the occasional “slip.” The treatment procedure would not be dissimilar
to the participant modeling or in viva exposure methods used with considerable success in
the treatment of phobics (Bandura, 1977a) and obsessive-compulsive disorders (Rachman
and Hodgson, (in press)). In this procedure the client is systematically exposed to the full
range of stimuli that elicit the problem behavior to extinguish maladaptive avoidance
responses (i.e. excessive drinking) and to substitute constructive coping behaviors. More
controversially the logic of this approach dictates what Marlatt (in press) has called a
“programmed relapse”. * With appropriate preparation and instruction, the client would
deliberately take a drink in the presence of the cues that were formerly discriminative stimuli
for alcohol abuse. The goal is to establish self-directed coping skills and efficacy
expectations that reduce the probability of any subsequent transgression serving as a
reaffirmation of a perceived inability to remain sober, a sense of helplessness and a stimulus
for further drinking.

Is it superior to alternative theories?

The adequacy of the self-efficacy theory can be analyzed from several different
perspectives. For example, it can be compared with Seligman’s (1975) theory of learned
helplessness in accounting for a wide range of different behaviors. Such an extended analysis
is beyond the confines of the present paper which is addressed to the more limited question
of how well Bandura’s (1977a) theory compared to conditioning theories in explaining the
effects of specific fear reduction techniques in behavior therapy.

The neobehaviorikic S-R conditioning model. The proof of the pudding, of course, is in the
eating. Since Bandura (1977a) reports the results of only two studies that were designed to
evaluate the self-efficacy theory any conclusions about the superiority of the theory must be
regarded as tentative at best. With this reservation in mind, one can look at Bandura and
Adams’ (in press) findings indicating the superior explanatory power of the self-efficacy
theory compared to two-factor conditioning theory.? In fact, the conditioning theory on

*Marlatt (in press) has drawn upon cognitive dissonance and attribution theory in proposing a cognitive
behavioral analysis of relapse phenomena in addictive behaviors. The practical implications of Marlatt’s analysis
and that derived from efficacy expectations are essentially similar. It remains to be seen whether the self-efftcacy
analysis summarized here will confer any theoretical or therapeutic advantages over Marlatt’s cognitive
explanation.
iIn the Bandura and Adams’ (in press) study, extinction of anxiety was defined in terms of subjects’ completion
of the imaginal desensitization hierarchy. Critics might argue that extinction necessarily entails the elimination of
conditioned autonomic responsa to the fear-provoking stimuli. However, self-report of anxiety is the way in which
extinction is determined in the routine practice of systematic desensitization. Secondly, reduction of autonomic
arousal is not necessary for the extinction of avoidance behavior in animals (Rescorla and Solomon, 1967) or
phobic clients (Leitenberg, 1976). Autonomic arousal appears to be a correlated coeffect of avoidance behavior as
opposed to a direct cause as one believed.
Towards a Unifying Theory of Behavioral Change 225

which fear reduction methods such as systematic desensitization and flooding have been
based has been under fire for some time now. Thus Rachman (1977), confirming
considerable previous criticism, has detailed how the conditioning theory of fear
acquisition, a fundamental facet of behavior therapy from the start (Wolpe, 1958), is clearly
inadequate. Similarly, traditional two-factor theory faces seemingly insurmountable
problems in accounting for the maintenance and modification of neurotic anxiety
conditions (e.g. Eysenck, 1976b; Rachman, 1976b; Seligman and Johnston, 1973). A major
theoretical difficulty has been the striking resistance to extinction that phobics show despite
nonreinf’orced exposure to fear evoking stimuli.
Attempts have been made to save the theory by suggesting that the haphazard and non-
systematic exposure to feared events that phobics experience is not the stuff of which
extinction is made. Although this might be a factor, it is doubtful that it can fully explain the
persistence of neurotic behavior. Taking another tack, Borkovec (1975) has proposed a
cognitive extension of Mowrer’s (1960) two-factor theory in which mere physical exposure
to the external fear stimuli is insufficient to ensure extinction. Rather,fwrctionalexposure to
the fear stimuli may be the critical ingredient, and this can be negated or interfered with by
cognitive avoidance responses. This position is supported by data from Borkovec’s (1972,
1973) avoidance response control condition. In this condition snake phobic subjects were
imploded until they signalled anxiety, at which point they visualized an avoidance response
they might have typically made in the actual situation. In marked contrast to both the
desensitization and implosion therapy treatments in this study, the avoidance control
condition resulted in neither improvement on the behavioral avoidance test nor reductions
in heart rate during the therapy session.
Borkovec’s (1975) hypothesis merits serious consideration. Like all other versions of
conditioning theory, it attributes variability in outcome of fear reduction methods to the
quantity or overall amount of non-reinforced exposure. One of the innovative and distinctive
features of self-efficacy theory is the emphasis on the quality of the non-reinforced exposure.
The dependability of the information derived from the exposure that alters efficacy
expectations and the reliability of the cognitive processing of that information are the
critical mechanisms. This appeal to the quality as opposed to the quantity of exposure is
saved from being a post hoc tour de force by assessing quality (i.e. derivative efficacy
estimates) independently of the behavior it is said to explain.
Although non-reinforced exposure has been consistently shown to be a necessary
condition for therapeutic success (cf. Davison and Wilson, 1973; Leitenberg, 1976a;
Rachman, 1972; Wilson and Davison, 1971), the critical process mechanisms in exposure
methods have not been elucidated. In particular, the variability of performance under
similar conditions of exposure remain to be explained. In their papers on desynchrony in
fear and avoidance, Hodgson and Rachman (1974) and Rachman and Hodgson (1974)
outline several hypotheses about the observed patterns of variability in outcome of fear
reduction techniques. In one of their hypotheses, for example, they predict that the degree of
desynchrony that results from a therapeutic intervention will be a function of the particular
technique employed. Thus they suggest that in flooding, avoidance behavior will change
first with self-report and physiological arousal lagging behind. In systematic desensitization
it is suggested that behavioral change will follow reductions in subjective and affective
anxiety, whereas modeling is said to produce little desynchrony across response systems.
Hodgson and Rachman (1974) have thoughtfully set out the various clinical and
experimental observations that have been reported. However, an explicit theory to explain
the findings is still needed. Self-efficacy theory is one alternative. Instead of ordering the
desynchrony data in terms of the effects of different techniques, a finer-grained analysis
might generate predictions on the basis of the strength and generality of the efficacy
expectations induced by the different methods. Inevitably, it will require a sophisticated and
complex analysis of this sort in accounting for the highly variable patterns of behavior that
are observed in therapeutic outcomes. Among other items, this self-efficacy analysis might
explain the now familiar fact that it is not necessary to elicit maximal anxiety during
flooding in order to achieve therapeutic success (e.g. Leitenberg, 1976a; Mills et al., 1973;
Rachman, 1976b). In terms of the self-efficacy theory it is not the intensity of drive or level of
anxiety that is elicited but the informative function of the cues the client is exposed to that
determines outcome. The latter may or may not correlate highly with the former.
226 G. T. Wilson

Rejecting the conditioning theories of Watson and Mowrer as experimentally invalidated


and Skinner’s as tautological, Eysenck (1976b) has proposed a new learning theory model of
neurotic reactions, incorporating more recent concepts such as the incubation of fear and
biological preparedness. While acknowledging the importance of cognitive factors in
learned fear responses, Eysenck (1976a) suggests that his new conditioning theory and a
more cognitive theory will have few differential consequences. It is too soon to tell, but if
Bandura and A&m’s (in press) findings are any indication, the two theories will generate
different predictions. One reason is that in the new conditioning model fear is still
conceptualized as a drive that motivates phobic behavior. In the self-efftcacy theory the
informational function of fear-producing emotional arousal is emphasized.* Among other
advantages, the latter notion may provide a better explanation of the variability in treatment
outcome as noted above. The modifications Eysenck has made in his theoretical system
improve upon previous conditioning models in more satisfactorily accounting for the
resistance to extinction problem, but it is far from clear how it can accommodate findings of
the sort reported by Bandura (1977a).
There are other difficulties for the new conditioning model. If cognitive factors are
important as Eysenck concedes, then how are they to be integrated into the theory? Eysenck
(1976a) suggests that Pavlov’s second signalling system might accommodate the necessary
cognitive extensions of conditioning principles. However, this simply restates the problem
as it exists since the second-signalling system is a tenuous construct at best that is itself in
need of explanation. Related to this point is the relatively minor emphasis Eysenck (1976b)
places on vicarious learning despite the overwhelming evidence on the powerful influence of
vicarious learning on the development, maintenance, and modification of a wide range of
human behavior including avoidance responses (e.g. Bandura, 1977b; Rachman, 1976a;
Rosenthal and Bandura, (in press)). Not surprisingly, the role of vicarious learning in self-
efficacy theory is logically consistent and empirically supported.
The operant model. The operant model, with its many advantages of simplicity, parsimony
and theoretical consistency cannot be easily swept aside as tautological (Eysenck, 1976b).
However, Bandura’s (1977a) discussion of self-efftcacy theory must now rank with the issue
of the much debated nature of self-control processes as the cutting edge of the increasingly
apparent conceptual split between social learning theory and applied behavior analysis.
Bandura’s (1977a) paper affirms a model of behavior that is distinctly at odds with the
operant view in emphasizing cognitions as causal events, advocating an authentic reciprocal
determinism, and containing comments such as that in which differential reinforcement is
interpreted as a “special case of observational learning” (p. 192) - heady stuff, that is
certain to send a shiver down the operant spine. Specifically in regard to the data that are
presented, the challenge to the operant approach is the evidence that indicates that past
behavior is inferior to personal expectations of self-efficacy (derived from that past
experience) as a predictor of future behavior.
In their lucid analysis of cognition and behaviorism, Rachlin and Lacey (1977) distinguish
between methodological and radical behaviorism. The methodological behaviorist is said to
accept generalizations that relate one behavior to another “where theformer is a symptom of
the cause, rather than the cause of the latter*‘. For example, the verbal expression of perceived
self-efficacy may be reliably related to subsequent avoidance behavior but it cannot be a
cause (controlling variable) of the avoidance behavior. The radical behaviorist only accepts
generalizations in which behavior is related to environmental variables. According to this

*Social learning theory actually encompasses conditioned fear responses as a component process in the
acquisition and modification of phobic behavior. One has to ask whete the physiological arousal that serves a
source of information that affects efficacy expectations comes from. In his earlier analysis. Bandura (1969) stated
that emotional arousal is a product of two different sources of stimulation: “One is the emotional arousal serf-
generated by symbolic activities in theform of emotion-provoking thoughts aboutfrightening events. Thesecondis the
response evoked directly by conditioned aversive stimuli” (p. 364). Rachman (1977) has reached a similar conclusion
in his recent analysis of the different pathways to fear. The novel aspect of self-efftcacy theory is that it provides a
mechanism whereby emotional arousal, however it is elicited, mediates behavioral avoidance. It is the individual’s
cognitive appraisal of arousal that determines whether it motivates phobic behavior. In two-factor theory
conditioning the Eysenck’s new model of neuroses fear is directly linked to phobic behavior in a causal fashion. The
problems the desynchrony data pose for this causal sequence have already been mentioned (Bandura, 1977b;
Rachman and Hodgson, 1974).
Towards a Unifying Theory of BehavioralChange 227

view, for every generalization in which cognition is related to behavior, there must exist a
corresponding behavior-environment generalization. In this manner cognitive variables
can be overlooked without missing any lawful psychological relationships. Bandura’s
(1977a) theory challenges both these forms of behaviorism in proposing that a cognitive
variable (self-efficacy) is superior to a simple behavior-environment generalization in both
predicting and controlling behavioral outcome.
In accounting for these results, the applied behavior analyst may point out that cognitive
variables can be dispensed with in favor of the person’s past history of environmental
interaction (cf. Rachlin and Lacey, 1977). Simple performance on an item of a behavioral
avoidance hierarchy might be said to be too circumscribed a sample of “past behavior”.
Indeed, with thii view of antecedent events and a revised law of effect that is based not on
immediate or temporally contiguous consequences but an aggregate effect over time (Baum,
1973), explanation of almost any behavior change method seems possible. The more
thoughtful operant conditioner such as Rachlin (1977) will acknowledge that concepts such
as “history of environmental interaction” and a “correlation based law of effect” are
unabashedly inferential. (S)He would add, however, that they constitute inference without
mediation. Rachlin (1977) opts for inferences about past events as opposed to current
cog&ions in explaining behavior, arguing that both are speculative but the latter are at least
potentially observable and “suggest metho& of observation and modification short of
surgery” (p. 374). Yet in practice it does not appear (or at least has not yet seemed to be the
case) that the “history of environmental interaction” is any more observable than present
cognitions. Moreover, the focus on a cognitive construct such as self-efficacy demonstrably
leads to methods of indirect assessment and modification that are not surgical and, more to
the point, were not prompted by the operant conditioning model.
This is the critical point. It is suggested here that utility or heuristic value is the criterion
that provides the acid test of the relative merits of competing theories in contemporary
behavior therapy. The seminal influence of operant conditioning on the development of
behavior therapy need hardly be emphasized (e.g. Kazdin, 1975; Krasner and Ullmann,
1965; Leitenberg, 1976b; Ullmann and Krasner, 1965). It continues to direct a useful albeit
limited technology of behavior change. As mentioned earlier, however, the theoretical and
clinical potential of classical and operant conditioning learning principles appear to have
been mined to the hilt. Fresh theoretical leads are required To take a few examples, the
demonstrably successful use of vicarious observation methods, including covert, symbolic
and live modeling, were derived not from operant conditioning, but social learning theory.
The fact that some procrustean manouver may be engaged in to accommodate these
methods within the post hoc parsimony of an operant framework is less important than the
failure of this conceptual scheme to generate such techniques.

CONCLUSION

For the most part, behavior therapy has been marked by bold and imaginative
innovations, by a vigorous diversity of viewpoints and a healthy aptitude for self-criticism.
In breaking new ground that will influence theory, research and clinical practice in behavior
therapy, self-efficacy theory promises to help sustain and nourish this impetus towards a
better understanding and more effective treatment of psychological disorders. Regardless of
its ultimate fate, the heuristic value of Bandura’s (1977a) theory is bound to enrich our
knowledge.
There will be those who will complain that we can ill-afford to pursue what Grossberg
(1977) has arbitrarily rejected as “the fools’ gold of cognitive self-control constructs*‘. The
sense of the present paper can be summarized by stating that we cannot afford not to extend
the clinical and experimental foundations of behavior beyond the relatively narrow
conceptual boundaries within which we have thus far grown fat. The inferential leap must be
made and Bandura (1977a) has indicated where we might look in leaping. Extending the
analysis of self-efficacy theory to different problems beyond the comforting confines of the
research laboratory will take a combination of methodological rigor, clinical skill and a little
luck. The progress of the last 20 years suggests that this is not an unrealistic expectation.
228 G. T. Wilson

REFERENCES
Agras, W. S.. Axrin, N., George, A., Kaxdin, A. E.. Mischel, W., Rachman, S. and Wilson, G. T. Behavior
Therapy: An Evabtarion. Unpublished manuscript, Center for Advanced Study in the Behavioral Sciences,
Stanford, Cahiomia.
American Psychiatric Association (1968) Daignostic andStotistica1 MonualofMentolDisorders(2nd edn) American
Psychiatric Association, Washington, D.C.
Axrin, N. H. (1977) A strategy for applied research: Learning based but outcome oriented Am. Psychol. 32,
140449.
Bandura, A. (1976) Self-reinforcement: Theoretical and methodological considerations. Behaviorism 4, 135-155.
Bandura, A. (1977a) Sclf-efticacy: Toward a unifying theory of behavioral change. Psycho!. Rev. 84, 191-215.
Bandura, A. (1977b) Sociof Learning 2%eory. Prentice-Hall, Englewood Cliffs, NJ.
Bandura, A. (1978) The self system in reciprocal determinism. Amer. Psychol. 33.344-358.
Bandura, A. and Adams, N. E. Analysis of self-efficacy theory of behavioral change. Cogn. Ther. Res. (in press).
Bandura, A., Adams, N. E. and Beyer, J. (1977) Cognitive processes mediating behavioral change. J. Personafity
sac. Psychol. 35, 125-139.
Baum, W. M. (1973) The correlation-based law of effect. J. exp. Analysis Behov. 20, 137-153.
Beck, A. T. (1976) Cognitive Therapy and the Emotionol Disorders. International Universities Press, New York.
Borkovec, T. D. (1972) Effects of expectancy on the outcome of systematic desensitization and implosive
treatments for analogue anxiety. Behav. Ther. 3.294.
Borkovec, T. D. (1973) The role of expectancy and physiological feedback in fear research: A review with special
reference to subject characteristics. Behav. Ther. 4,491-505.
Borkovec, T. D. (1974) Heart-rate process during systematic desensitization and implosive therapy for analogue
anxiety. Behav. Ther. 5.636-641.
Borkovec, T. D. (1975) Cognitive extensions of two-factor theory. Paper presented as Association for
Advancement of Behavior Therapy, San Francisco, California. December.
Bower, 8. (1977) Cognitive psychology and behavior modification. Invited address, Association for Advancement
of Behavior Therapy, Atlanta, Georgia, 9 December.
Bnger, L. and McGaugh, J. L. (1966) A critique and reformulation of “learning theory” approaches to
psychotherapy and neurosis. Psychol. Bull. 65, 170-173.
Brewer. W. F. (1974) There is no convincing evidence for operant or classical conditioning in adult humans.
Iti W. B. Weimer and D. S. Palermo (Eds) Cognition and the symbolic Processes. Lawrence Erlbaum
Associates, HiBsdale, NJ.
Catania, A. C. (1975) The myth of self-reinforcement. Behaviorism 3. 1921199.
Catania, A. C. (1976) Self-reinforcement revisited. Behaviorism 4, 157-162.
Davison, G. C. and Wilson, G. T. (1973) Processes of fear reduction in systematic desensitixation: Cognitive
and social reinforcement factors in humans. B&v. Ther. 4, I-2I.
Ellis. A. (1970) The Essence of Rationol Psychotherapy: A Comprehensive Approach to Treatment. Institute for
Rational Living, New York.
Eysenck, H. J. (1959) Learning theory and behaviour therapy. J. Ment. Sci X&,61-75.
Eysenck, H. J. (1970) Behavior therapy and its critics. .I. Behav. Ther. exp. Psychiat. 1, 5-15.
Eysenck, H. J. (1976a) Behaviour therapy - dogma or applied science? In: P. Feldman and A. Broadhurst
(J&Is) The Experimental Bases of Behaviour Therapy. Wiley, New York.
Eysenck, H. J. (1976b) The learning theory model of neurosis - a new approach. Behav. Res. Ther. 14.251-268.
Franks, C. M. and Wilson, G. T. (1973) Annual Review of Behavior Therapy: Theory ond Practice, Vol. I.
Brtmner/Maxel, New York.
Franks, C. M. and Wilson, G. T. (1976) Annual Review of Behavior Therapy: Theory and Practice, Vol. IV.
Bnmner/Maxel, New York.
Franks, C. M. and Wilson, G. T. (1977) Annual Review of Behavior Therapy: Theory and Practice, Vol. V.
Brunner/Maxel, New York.
Goldfried, M. R. and Davison, G. C. (1976) CIinicol Behavior Theropy. Holt, Rinehart & Winston, New York.
Goldiamond, I. (1976) Self-reinforcement as an explanatory tiction. J. appl. Behav. Analysis 9, 509-Z 14.
Green, L. (1976) The temporal and stimulus dimensions of self-monitoring in the behavioral treatment of
obesity. Unpublished doctoral dissertation, Rutgers University.
Grosberg. J. M. (1977) The British viewpoint in behavior therapy. Contemp. Psychol. 22.308-310.
Hersen, M. and Barlow, D. H. (1976) Single Case ExperimentolDesigns: Strotegiesfor Studying Behavior Change.
Pergamon Press, New York.
Hodgson. R. and Rachman, S. (1974) Desynchromy in measures of fear. II. Behav. Res. Ther. 12.319-326.
Jeffrey, D. B. (1974) A comparison of the effects of external control and self-control on the modification and
maintenance of weight. J. abnorm. Psychol. 83.404-410.
Kaxdin. A. E. (1975) Behavior Modt~cation in Applied Settings. Dorsey Press, Homewood. II.
Kaxdin, A. E. Research Design in CIinicaI Psychology. Lippincott Philadelphia, PA. (In press.)
Kaxdin, A. E. and Wilcoxon, L. A. (1976) Systematic desensitization and nonspecific treatment effects: A
methodological evaluation. Psycho!. Bull. 83.729-758.
K&in, A. E. and Wilson, G. T. (1978) Evaluation of Behavior Theropy: Issues, Evidence andResearch Strategies.
BaBinger, Cambridge, Mass.
Lazaras. A. A. (1976) MuItimoaW Behavior theropy. Springer, New York.
Laxarus, A. A. (1977) Has behavior therapy outlived its usefulness? Am. Psycho! 32,550-554.
Leitenberg, H. (1976a) Behavioral approaches to treatment of neuroses. In: H. Leitenberg (Ed) Handbook of
Behavior Moctifiation and Behavior Therapy. Prentice-Hall, Englewood Cliffs, NJ.
Leitenberg, H. (1976b) Handbook of Bekavior Modt$cotion and Behavior Therapy. Prentice-Hall. Em&wood
Cliff.., NJ.
Lick, J. and Bootxin. R. (1975) Expectancy factors in the treatment of fear: Methodological and theoretical
issuea. Psychof. Bull. 82,917-93 1.
London, P. (1972) The end of ideology in behavior modification. Am. Psychol. 27.913-926.
Mahoney. M. J. (1974) Cognition and Behavior ModtJicotion. Bahinger, Cambridge, Mass.
The Importance of being Theoretical 229

Mahoney, M. J. and Arnkoff, D. Cognitive and self-control therapies. In: S. L. Garfield and A. E. Bergin (Eds)
Handbook of Psychotherapy and Behavior Change (2nd ed.) Wiley, New York. (In press.)
Marks, I. M. (1976) The current status of behavioral psychotherapy: Theory and practice. Am. J. Psychiar. 133,
253-261.
Marlatt, G. A. Craving for alcohol, loss of control and relapse: A cognitive-behavioral analysis. In: P. E. Nathan
and G. A. Marlatt (Eds) Experimental andBehavioral Approaches lo Alcoholism. Plenum Press, New York.
(In press).
Meichenbaum, D. (1977) Cognitive Behavior Modijication. Plenum Press, New York.
Melges, F. T. and Bowlby, J. (1%9) Types of hopelessness in psycho-pathological process.Arclrs. gen. Psychiat.
20.690699.
Mills, H. L., Agras, W. S., Barlow, D. H. and Mills, J. R. (1973) Compulsive rituals treated by response
prevention. Archs. gen. Psychiat. 2&524-529.
Mischel, W. (1973) Toward a cognitive social learning reconceptuahxation of personality. Psychol. Rev. 80,
252-283.
Mowrer, 0. H. (1960) Learning Theory and the SymboIic Processes. Wiley, New York.
Rachhn, H. (1974) Self control. Behaviorism 2, 94-107.
Rachlin, H. (1977) A review ofM. J. Mahoney’s “Cognition and Behavior Modification”/. appl. Behav. Analysis 10,
369-374.
Rachlin, H. and Lacey, M. H. (1977) Behavior, cognition and theories of choice. Unpublished manuscript, State
University of New York at Stony Brook.
Rachman, S. (1972) Clinical applications of observational learning, imitation and modeling. Behav. Ther. 3,
379-397.
Rachman, S. (1976a) Observational learning and therapeutic modeling. In: M. P. Feldman and A. Broadhurst
(Ed@ Theoretical and Experimental Bases of the Behavior Therapies. Wiley, New York.
Rachman, S. (1976b) The passing of the two-stage theory of fear and avoidance: fresh possibilities. Behav. Res.
Ther. 14, 125-131.
Rachman, S. (1977) The conditioning theory of fear-acquisition: A critical examination. Behav. Res. Ther. 15,
375-388.
Rachman, S. and Hodgson, R. (1974) Synchrony and desynchrony in fear and avoidance: I. Behav. Res. Ther. 12,
311-318.
Rachman, S. and Hodgson, R. Obsessions and Compulsions. Prentice-Hall. Englewood Cliffs. NJ. (Inpress).
Reid, D. W. and Ware, E. E. (1974) Multidimensionahty of internal versus external control: Addition of a third
dimension and nondistinction of self versus others. Can. J. Behav. Sci. 6, 131-142.
Rescorla. R. A. and Solomon, R. L. (1967) Two-process learning theory: Relationships between Pavlovian
conditioning and instrumental learning. PsychoI. Rev. 74, I5 1-182.
Risley, T. (1977) Presidential address, Association for Advancement of Behavior Therapy, Atlanta, Georgia
10 December.
Rosenthal, T. L. and Bandura, A. Psychological modeling Theory and practice. In: S. L. Garfield and A. E.
Bergin (Eds) Handbook of Psychotherapy and Behavior Change (2nd edn). Wiley, New York. (In press.)
Rotter, J. B. (1966) General&d expectancies for internal versus external control of reinforcement. Psychor
Monogr. 80, (1, Whole No. 609).
Rush, A. J., Beck, A. T., Kovacs, M. and Hollon, S. (1977) Comparative efficacy of cognitive therapy and
pharmacotherapy in the treatment of depressed out-patients. Cogn. Ther. Rei 1. 17-37.
Seligman, M. E. P. (1975) Helplessness. Freeman, San Francisco, CA.
Seligman, M. E. P. and Johnston, J. (1973) A cognitive theory of avoidance Iearning. In: J. McGuigan and
B. Lumsden (Eds) Contemporary Approaches IO Conditioning and Leamhg. Wiley, New York.
Stokes, T. F. and Baer, D. M. (1977) An implicit technology of generalization. J. appl. Behav. Analysis 10,
349-368.
Stuart, R. B. (1972) Situational versus self-control. In: R. D. Rubin, H. Fenaterheim, J. D. Henderson and
L. P. UUmann (Eds) Advances in Behavior IXerapy. Academic Press, New York.
Stuart, R. B. and Guire, K. (1977) Some correlates of the maintenance of weight lost through behavior
modification. Int. J. Obesity, 2, 225-236.
Stunkard, A. J. and Mahoney, M. J. (1976) Behavioral treatment of the eating disorders. In: H. Leitenberg (Ed.)
Handbook of Behavior Mniijlcarion and Behavior Therapy. Appleton Century Crofts, New York.
Tobias, L. L. and MacDonald, M. L. (1977) Internal locus of control and weight loss: An insufftcient condition.
J. consult. c/in. Psychol. 45, 647-653.
Uhmann, L. P. and Krasner, L. (1965) CaseStudies inBehaviorModificat. Hoh, Rinehart & Winston, New York.
Williams, R. B. and Gentry, W. D. (1977) Behavioral Approaches lo Medica/ Treatment. Ballinger, Cambridge,
Mi3SS.
Wilson, G. T. (1978) On the much discussed nature of the term ‘behavior therapy.” Behav. Ther.
9.89-98.
Wilson, G. T. (a) Booze, beliefs, and behavior: Cognitive processes in alcohol use and abuse. In: P. E. Nathan
and G. A. Marlatt (Eds) Experimental andBehaviorai Approaches to Alcoholism. Plenum Press, New York.
(In press.)
Wilson, G. T. (b) Cognitive behavior therapy: Paradigm shift or passing phase? In: J. P. Foreyt and D. Rathjen
(Eds) Cogairive Behavior Therapy: Research and Application. Pknum Press, New York. (In press.)
Wilson, G. T. (c)Methodological considerations in treatment outcome research on obesity. J. consult. c/ix. Psychol.
(In press.)
Wilson, G. T. and Davison, G. C. (1971) Processes of fear reduction in systematic desensitization: Anid
studies. Psycho/. Bull. 76, 1-14.
Wilson, G. T. and Evans, I. M. (1976) Adult behavior therapy and the therapist-client relationship. In:
C. M. Franks and G. T. Wilson (Eds) Annual Review of Behavior Therapy: Theory and Practice, Vol. IV.
Brunner/Maxel, New York.
Wilson, G. T. and Evans, I. M. (1977) The therapist-client relationship in behavior therapy. In: R. S. Gutman
and A. M. Raxin (Eds) The Therapist’s Contribution to Effective Psychotherapy: An Empirical Approach.
Pergamon Press, New York.
230 G. T. Wilson

Wolpe, J. (1958) Psychotherapy by Reciprocal Inhibition. Stanford University Press, Stanford.


Wolpe, J. (1969) The Practice of Behuvior Therapy Pergamon Press, New York.
Wolpe, J. (1976) Behavior therapy and its malcontents - II. Multimodal eclecticism, cognitive exclusivism and
“exposure” empiricism. J. Behov. Ther. exp. Psychbr. 7, 109416.

You might also like