Content uploaded by Elizabeth F Loftus
Author content
All content in this area was uploaded by Elizabeth F Loftus
Content may be subject to copyright.
Prevention & Treatment, Volume 5, Article 41, posted October 18, 2002
Copyright 2002 by the American Psychological Association
Commentary on "On The Social Psychology of the Psychological Experiment: With
Particular Reference to Demand Characteristics and Their Implications"
Demand Characteristics, Treatment
Rationales, and Cognitive Therapy for
Depression
Jonathan W. Kanter, Robert J. Kohlenberg, and Elizabeth F. Loftus
Department of Psychology, University of Washington
ABSTRACT
In research design, the term demand characteristics refers to the sum total of cues that convey an
experimental hypothesis to subjects and influence their behavior. In psychotherapy, the term may
refer to the sum total of cues that convey the therapist's wishes, expectations, and worldviews to
clients and influence their behavior. Psychotherapeutic demand may play a role in dissociative
identity disorder, in the repressed memory controversy, and during the delivery of treatment
rationales, a practice common to manual-guided psychotherapies. An exploration of demand
characteristics during the presentation of the treatment rationale in one such psychotherapy—
cognitive therapy for depression—is offered, and data from an analogue study of demand
responding to this rationale are presented.
Jonathan W. Kanter is now at the Department of Psychology, University of Wisconsin—Milwaukee; Elizabeth F.
Loftus is now at the Department of Psychology and Social Behavior and the Department of Criminology, Law, and
Society, University of California, Irvine.
Correspondence concerning this article should be addressed to Robert J. Kohlenberg, Box 351525, Department of
Psychology, University of Washington, Seattle, Washington 98195.
E-mail: fap@u.washington.edu
Mary was suffering from depression and anger and sought therapy from a cognitive therapist.
Following the standard protocol, the therapist presented the therapy treatment rationale during the
first session. He explained to Mary that when certain events occurred in her life, she first had
thoughts about those events that then, in turn, produced her negative feelings. When Mary
returned for the second session, she told him that she'd gone to her mother's for dinner, and when
a plate of broccoli was placed on the table (a vegetable she never had liked), she noticed that she
first thought, "She doesn't care about me" and then noticed that she felt angry and depressed. Did
Mary's experience have anything to do with the explanation given to her in her initial therapy
session? Mary's case is hypothetical but allows us to illustrate a case we will make, namely, that
the therapist's explanation may well have influenced what Mary experienced and what she
reported during her next therapy session
In this article we apply the concept of demand characteristics to the treatment rationales
presented to clients receiving psychotherapy. First, several aspects of the definition of demand
characteristics are highlighted. Then, the notion of demand characteristics is extended to
psychotherapy, with attention to dissociative identity disorder and the repressed memory
controversy. In these two areas, the role of demand characteristics already has been investigated,
and the notion has been raised that rigid adherence to a particular conceptual model of
psychopathology may lead to demand responding. We then suggest that treatment rationales can
b
e viewed as a package in which demands often are neatly wrapped and explicitly articulated. We
review research on treatment rationales for systematic desensitization that has demonstrated the
influence of demand characteristics, and describe a preliminary study investigating demand
characteristics in treatment rationales for cognitive therapy for depression. Cognitive therapy is
presented as an example of strong treatment that nonetheless may be a vehicle through which the
unintended influence of demand characteristics may occur.
Definitional Issues
In his seminal work, Orne (1962b) defined demand characteristics as the sum total of cues that
convey an experimental hypothesis to subjects and influence their behavior. The most basic
example of a demand characteristic is an experimental procedure that involves a pretest,
immediately followed by experimental treatment, which in turn is followed by a posttest using
the same measure as did the pretest. About this situation, Orne (1962b) wrote, "even the dullest
college student is aware that some change is expected, particularly if the test is in some obvious
way related to the treatment" (p. 779). Orne further specified that if demand characteristics are
present and discernable, many subjects—influenced by the authority of the experimenter and the
role of the "good subject" who desires to contribute meaningfully to science—subsequently will
respond in a manner that will support the hypotheses being tested. As such, demand
characteristics are a primary threat to the internal validity of an experiment—an influence on
dependent variables, a source of bias—to be isolated from the true effects of independent
variables.
Aspects of the definition suggest that for demand characteristics to have their influence, the
subject must consciously apprehend the true purpose of the experiment. Orne conveys this
necessity in the quote above ("even the dullest college student . . .") and also later in his original
(1962b) exposition of the term: "One of the basic characteristics of the human being is that he
will ascribe purpose and meaning. . . . In an experiment where he knows some purpose exists, it
is inconceivable for him not to form some hypothesis as to the purpose" (p. 780). The necessity
of conscious apprehension also is presumed by the primary procedure suggested by Orne for
determining the presence of demand characteristics in an experiment: a postexperimental inquiry
that involves asking the subject what he or she perceived and believed about the experiment.
Another procedure suggested by Orne, a role-playing procedure in which new or the same
subjects are told the purpose of the experiment and asked to behave as if they were real subjects,
similarly presumes that responding under the influence of demand characteristics involves
conscious apprehension. In essence, this view of demand characteristics posits a subject
consciously, actively deducing the hypothesis of the experiment in order to determine how to
behave in the experiment.
However, at other times, Orne seems to suggest that demand responding involves something
other than a conscious process, although the nature of the process is not clearly articulated. For
example:
It is not conscious deception by the subject which poses the problem here. That occurs only
rarely. Demand characteristics usually operate subtly in interaction with other experimental
variables. They change the subject's behavior in such a way that he is often not clearly aware of
their effect. (Orne, 1969, p. 148)
The treatment of demand responding as conscious compliance at times and as a nonconscious
process at other times parallels a shifting emphasis on two different sets of variables presumed to
motivate demand responding: the striking degree of compliance exhibited by experimental
subjects and the influence of the role of the "good experimental subject." On the one hand, Orne
(1962b) emphasized the experimenter as an authority figure who fosters compliance and the
experiment as a unique situation that can generate unusual degrees of compliance; witness the
compliance one may obtain from a simple request when it is couched as an experiment versus
when it is not. On the other hand, Orne also emphasized the influence of the role of the good
subject—a subject who values science, wants to participate in a meaningful experiment and
hopes to contribute data that will be beneficial. Orne posits that the good subject role is well
understood in our culture, particularly by the college students, who are the modal subject.
Further, he associates conscious responding with compliance and nonconscious responding with
the role of the good subject:
It became clear from extensive interviews with subjects that response to the demand
characteristics is not merely conscious compliance. When we speak of "playing the role of a good
experimental subject," we use the concept analogously to the way in which Sarbin (1950)
describes role playing in hypnosis: namely, largely on a nonconscious level. The demand
characteristics of the situation help define the role of "good experimental subject," and the
responses of the subject are a function of the role that is created. (Orne, 1962b, p. 779)
Nonetheless, subsequent researchers, whether focused on demand characteristics as a topic of
methodological importance in its own right or as a means to invalidate research done in other
content areas, tended to muddy this distinction and, if anything, characterized demand responding
as a simple process of conscious compliance. For example, many textbooks on research methods
define demand characteristics in this fashion (e.g., Shaughnessy & Zeichmeister, 1997). Perhaps
this was done because in the experimental world, conscious compliance can be studied more
easily (e.g., through postexperimental inquiry) than can nonconscious processes.
A definition of demand characteristics in psychotherapy, paralleling the language of Orne's
original definition, might be stated as: The sum total of cues that convey the therapist's wishes,
expectations, and worldviews to a client and become significant influences on the client's
behavior, specifically influencing the client to confirm the therapist's wishes, expectations, and
worldviews with subsequent behavior. Several differences between experimental and
psychotherapeutic characteristics are worth noting. First, in an experiment, the environmental
cues, experimenter behaviors, and other factors that comprise demand characteristics can be
isolated and controlled, albeit not completely but oftentimes rather easily. For example, a double-
blind design will go far in solving many of the basic dilemmas posed by demand characteristics
and expectancy effects. In addition, clever researchers can disguise the true purpose of an
experiment, and postexperimental inquiry and other procedures can be undertaken to further
minimize and isolate demand effects. Much has been written on these techniques (e.g., Bernstein
& Nietzel, 1977).
In psychotherapy, the rules and priorities are different. Although many clinicians are exposed
through training in ethical guidelines to the power relations between themselves and their clients,
and we may assume that many are sensitive to the potential impact of direct advice and opinions,
there are no formal procedures or other efforts to identify and minimize more subtle demand
effects.
There is no doubt that a clinician is in a position of authority relative to most clients. Early
research on the operant conditioning of verbal behavior demonstrates subtle, nonconscious
influences of authority. For example, Greenspoon (1954, 1955) demonstrated that subtle
contingent responses from the experimenter, such as a grunt or an "um-hum" with an affirmative
inflection, influenced graduate students to use more plural nouns in their speech, even though the
graduate students were unaware of the process of influence. This study sparked a series of
subsequent studies, reviewed by Frank (1961) and, briefly and more recently, by Leigland (1995).
Four general findings noted in Frank's (1961) review are worth mentioning: (a) Subjects'
verbalizations are more easily influenced by persuaders with power or prestige; (b) compliant and
highly anxious subjects are more easily influenced than are less compliant or less anxious
subjects; (c) subtle signs of approval are just as effective in conditioning verbal behavior as are
obvious signs of approval; and (d) subjects do not need to be aware of the contingent responding
for it to be effective, nor does the subject need to be aware that a conditioning process is
involved. Frank (1961) concluded, "This much, at least, seems safely established: One person can
influence the verbalizations of another through very subtle cues, which may be so slight that they
never come to the center of awareness" (p. 108).
Similarly, research on interpersonal expectancy effects, which can be regarded as a subset of
demand responding,1 provides a wealth of data on the influence of authority as it pertains to the
psychotherapeutic relationship. Originally experimenter expectancy effects (Rosenthal, 1963), the
term interpersonal expectancy effects has come to represent the general process by which an
authoritative person's expectations about another person come to influence that person and bring
about the expected behavior. Considerable research has investigated the specific behaviors of the
authoritative person that result in the expected behaviors of the subject. In a meta-analytic review
of this research, Harris and Rosenthal (1985) summarized the behaviors that have been shown to
result in expectancy-confirming responses, which include the following: creating a less negative
climate (e.g., not behaving in a cold manner), maintaining closer physical distances, providing
more input by introducing more material or more difficult material, creating a warmer climate,
exhibiting less off-task behavior, having longer interactions, having more frequent interactions,
asking more questions, encouraging more, engaging in more eye contact, smiling more, praising
more, accepting the subject's ideas by modifying, acknowledging, summarizing, or applying what
he or she has said, providing more corrective feedback, nodding more, and waiting longer for
responses. This list suggests that the behaviors and cues involved may be quite subtle and operate
outside of the conscious awareness of the subject. Notably, the problem of experimenter
expectancy effects is effectively solved for researchers by the process of keeping experimenters
who interact with subjects blind to the experimental condition of the subject. There is no such
solution for clinicians; in fact, the behaviors listed above comprise many of the primary tools of
the clinician, and clinicians are taught to use them contingently.
In this sense, when applied to psychotherapy, a shift occurs in the conceptualization of demand
characteristics from "nuisance variables" (Bernstein & Nietzel, 1977) with effects on dependent
variables that must be ruled out and isolated from the effects of independent variables to variables
that have a profound and lasting influence on behavior, that is, the independent variables
themselves. In psychotherapy, the environmental cues and therapist behaviors that would be
considered demand characteristics in an experiment may be the very essence and foundation of
the psychotherapeutic process. This conceptualization of psychotherapy as infused with demand
characteristics has a long list of proponents and opponents in the guise of the debate on specific
versus common factors in psychotherapy (e.g., Ilardi & Craighead, 1994; Weinberger, 1995); in
fact, Horvath (1984) describes demand characteristics, expectancies, and placebo effects together
as equivalent to "nonspecific variables" or "common factors." Some (e.g., Frank, 1961) go so far
as to say that the entire process of psychotherapy is one of persuasion and influence (cf. Abroms,
1968); few doubt that persuasion and influence occur at some level.
Another important difference between experimental and psychotherapeutic demand
characteristics exists. In experiments, the demand is for hypothesis-confirming behavior. Once
the experiment is over, there is no need for maintenance of the demanded behavior, and there is
no continuing relationship with the experimenter through which behavioral changes may be
reinforced. In most experimental situations, emphasis is on the immediate results of the
experiment, and after the experiment is completed efforts are made through debriefing to ensure
that the participant has not been changed in any psychologically meaningful way (longitudinal
and treatment outcome research are notable exceptions). In fact, the primary function of human
subjects review committees is to ensure that permanent meaningful changes in psychological
experience do not occur. In this sense, it is assumed that any remaining effects of demand
characteristics are transient, meaningless psychologically, and not worthy of attracting further
attention.
In psychotherapy, the priorities are of a different sort than are those of experimental situations.
Permanent, meaningful changes in behavior are the primary outcome of interest. A client is told,
in essence, "Do this and you will get better." The demand is for permanent behavior change, and
this may be reinforced by a belief or awareness that one is improving or that one is pleasing the
therapist; the therapist, in turn, may directly reinforce such demanded behavior when it occurs or
may reinforce talk about such behavior occurring in daily life.
Demand Characteristics in Psychotherapy
Dissociative Identity Disorder and False Beliefs
Orne and colleagues (Orne & Bates, 1992; Orne & Bauer-Manley, 1991) developed a theory of
the development of dissociative identity disorder (DID; formerly multiple personality disorder
[MPD] when Orne published this theory) that emphasized the influence of demand
characteristics. In brief, they suggested that a therapist who believes it is possible for multiple
personalities to exist within the skin of a single individual may be likely to view inconsistency in
behavior as evidence of multiple personalities. This conceptual model will influence how
therapists conceptualize, treat, and respond to their patients, resulting in therapists looking for
distinct personalities in their patients, selectively reinforcing clients when those clients view their
own inconsistent or unacceptable behaviors as produced by different personalities, further
solidifying the reality of split personalities by colluding with one of the personalities against the
others, and so forth. Note that this view of the iatrogenic development of DID makes more of
demand characteristics than simple compliance or conscious striving to please the therapist; it
relies heavily on the concept of nonconscious role playing used by Orne to account for hypnotic
states (Orne, 1962a). The exact nature of the process was unspecified, but the implication is that
the therapist is fundamentally changing the client through demand. Orne and Bates (1992) left no
doubt that they viewed the process as more than simple compliance: "The resulting disorder is
real, but its manifestations are largely determined by the cultural context in which they occur and
the type of therapy received" (p. 249). Similarly, Orne and Bauer-Manley (1991) wrote:
We see, then, that the therapist's conceptual model may have serious consequences for his or her
patient that go well beyond academic or theoretical concerns. The demand characteristics that can
result from a conceptual model of the self whose metaphors are reified and taken to be literally
true can, in patients capable of dissociation, help to produce the serious clinical condition we call
MPD. (p. 103)
Perhaps nowhere is the power of demand characteristics in psychotherapy clearer than in the case
of the repressed memory controversy. Consider the false memory or false belief position on this
controversy. This position focuses on the possibility that completely false memories may be
planted in clients by therapists who do not appreciate the processes of persuasion and influence—
subtle demands placed on clients—that occur in psychotherapy. Basic research has shown how
beliefs and memories can be distorted, manipulated, and explicitly planted in specific situations.
For example, people have been induced to recall nonexistent broken glass and tape recorders, a
clean-shaven man as having a mustache, straight hair as curly, and a barn in a bucolic scene that
contained no buildings at all (for a review, see Loftus & Ketcham, 1991). Hypnotic suggestion
(Laurence & Perry, 1983) and simply asking participants to imagine and describe loud noises
(Weekes, Lynn, Green, & Brentar, 1992) have resulted in later memories of noises that never
occurred. Children have been induced to believe that they had witnessed assaultive behavior
when they had not (Haugaard, Reppucci, Laird, & Nauful, 1991). Adults have been induced to
believe that they had been lost in a mall as children when they had not (Loftus, 1997), had been
hospitalized overnight when they had not, had had an accident at a family wedding when no such
accident had occurred (Hyman, Husband, & Billings, 1995), and had witnessed demonic
possession as children (Mazzoni, Loftus, & Kirsch, 2001).
Research on false memories has generated controversial conclusions about the nature of
reportedly repressed memories of extremely traumatic events—such as childhood sexual abuse—
uncovered during some forms of psychotherapy (e.g., Loftus & Ketcham, 1994). For obvious
ethical reasons, it has not been possible to experimentally plant false beliefs of childhood sexual
abuse. However, traumatic childhood memories of a lesser sort have been planted. Recent
research on false memories planted during a therapy-like dream interpretation is illustrative
(Loftus & Mazzoni, 1998; Mazzoni, Lombardo, Malvagia, & Loftus, 1999). Undergraduate
participants initially were asked to report how confident they were that various life events had or
had not happened to them before age 3. Participants were selected for the next phase of the study
if they reported with confidence that they specifically had not been lost or in dangerous situations
before age 3. Participants then were randomly assigned to one of two alternative dream
interpretation conditions. Each participant met privately with a clinical psychologist, under
separate pretense, for a brief dream interpretation. During this meeting, half were told that their
dreams suggested that they had been lost before age 3 (the "lost" condition), and the other half
were told that they had been in a dangerous situation before age 3 (the "danger" condition). Four
weeks later, participants again were asked to report how confident they were that the same life
events had or had not happened to them before age 3. Results indicated that the dream
interpretation influenced participants' beliefs about childhood events; those in the lost and danger
conditions, respectively, reported with more confidence that they had been lost or in danger
before age 3.
This research is suggestive but offers no evidence that memories are planted in actual clinical
practice. However, evidence from a variety of nonexperimental sources—therapist accounts of
the therapy process, client accounts of the therapy process, sworn statements of clients and
therapists during litigation, and taped interviews of therapy sessions—suggests that it is possible
for completely false memories of childhood sexual abuse to be planted during certain forms of
psychotherapy (Loftus, 1993). This evidence includes multiple examples of therapists who
explicitly suggest the possibility that childhood sexual abuse occurred despite repeated client
claims that it did not. It also includes examples of therapists encouraging clients with no
memories of abuse to guess or tell stories about childhood abuse, interpreting dreams that are
neither sexual nor abusive in content as signs of abuse, recommending books about survivors of
abuse to clients with no memories of abuse, and strongly reinforcing what begins as mere
suspicions of abuse in clients. In these cases, the demands placed by therapists are not subtle, nor
are the effects.
Demand Characteristics and Treatment Rationales
These positions on DID and false memories suggest a demand characteristics interpretation of
psychotherapeutic change. At issue is a subculture of therapists who hold particular conceptual
views of psychopathology, do not emphasize alternative interpretations of client presentations
and changes, are explicit with clients about their conceptual model and expectations, and are
quite persistent in application of the model. These behaviors may place strong demands on
clients, resulting in new personalities and new memories.
Although it may be tempting to disclaim these therapist behaviors as uncharacteristic of most
practicing clinicians, in fact these behaviors may characterize, to an extent, all clinicians who
conform to a particular conceptual model of psychopathology to the inclusion of others, and who
present this model in the form of a treatment rationale. In practice, demand characteristics often
are wrapped into a neat package and explicitly articulated in the form of a treatment rationale.
The process of providing a treatment rationale is considered to be a common characteristic of the
beginning of nearly all cognitive and behavioral psychotherapies, and Frank (1961, 1971, 1982),
in his early writings on common factors in psychotherapy, identified the treatment rationale as
one of the six features common to all psychotherapies. According to Frank (1971), the treatment
rationale "includes an explanation of the cause of the patient's distress and a method for relieving
it" (p. 356). More recently, Addis and Carpenter (2000) defined the treatment rationale
functionally: "Its overarching purpose is to provide clients and therapists with a model of etiology
(Why is this person having this problem?) and treatment (What should we do to change it?)" (p.
147). Demand characteristics may be particularly salient in treatment rationales for manual-
guided treatments, as these manuals generally explicate a particular conceptual model, do not
emphasize alternatives, and encourage the persistent application of the model.
Systematic desensitization. Researchers fully explored the issue of demand characteristics and
treatment rationales in the 1970s with respect to systematic desensitization (e.g., Bernstein &
Nietzel, 1973; Marcia, Rubin, & Efran, 1969; McReynolds & Tori, 1972; Oliveau, Agras,
Leitenberg, Moore, & Wright, 1969; Smith, Diener, & Beaman, 1974; Wilkins, 1971). Oliveau et
al. (1969) found that providing a therapeutic rationale significantly influenced the outcome of
systematic desensitization. After systematic desensitization training, those given the rationale
demonstrated significantly more approach
b
ehavior to a feared object (a snake) than did those not
given a rationale. This study showed an incremental effect specifically for a therapeutic rationale.
Also, Marcia et al. (1969) showed that providing a rationale designed to produce a positive
expectation for change combined with a procedure containing none of the theoretically crucial
elements of desensitization produced as much behavior change as did an actual desensitization
procedure, suggesting that expectancy-laden rationales can result in behavior change at
therapeutic levels. Similarly, McReynolds and Tori (1972) found that a desensitization procedure
designed for phobia reduction produced changes in frustration tolerance if the rationale
theoretically linked the procedure to frustration tolerance and predicted that changes would occur.
These studies demonstrate specifically the notion that a rationale that is credible, scientific, and
suggestive of an innovative and powerful therapy will generate subtle expectancies and demands
for treatment-related behavior changes.
The treatment rationale in cognitive therapy for depression. We have been studying possible
demand characteristics present during the presentation of treatment rationales similar to those
offered in cognitive therapy for depression (CT; Beck, Rush, Shaw, & Emery, 1979) using an
analogue design and undergraduate participants. In this section we briefly review the CT
rationale and then describe some preliminary data from a study of its potential impact.
According to Beck et al. (1979), the cognitive rationale is presented to the patient during the first
and second sessions of therapy. During the first session the therapist defines cognition, describes
the relationship between cognition and affect, and provides examples that illustrate the
relationship. At the end of the first session, the patient is given the booklet Coping With
Depression (Beck & Greenberg, 1974/1995), which reiterates and expands on the cognitive
rationale. The patient is asked to read the booklet before the second session, at which time the
therapist is instructed to review the patient's responses to the rationale and the booklet.
The cognitive rationale has been represented as the A-->B-->C sequence, first presented by Ellis
(e.g., Ellis, 1973; Ellis & Harper, 1975). According to Beck (1967, p. 322), A represents an event
or stimulus, B represents cognition, and C represents affect. Simply put, the rationale presents a
specific model of the psychological world: Given a specific event or stimulus (A), one reacts
cognitively; that is, one interprets the event (B), and this interpretation influences how one
subsequently feels (C). Note that the cognitive rationale since has been elaborated by cognitive
scientists (e.g., Clark, Beck, & Alford, 1999) into a theory that is significantly more complex and
subtle than that presented in Coping With Depression, but this complexity and subtlety has not
parlayed into clinical practice, as the simpler theory expounded in Coping With Depression is
used with clients.
In a preliminary study, undergraduate participants were randomly assigned to receive one of two
booklets. The first booklet, the ABC booklet, is based on Coping With Depression (Beck &
Greenberg, 1974/1995). Unlike Coping With Depression, the ABC booklet does not discuss
depression and does not target a depressed population. Rather, it was designed to train
undergraduate participants in the cognitive rationale as it applies to general responding to stimuli.
The booklet presents the rationale as authoritative and correct, provides definitions of cognition
and affect that are consistent with it, and describes the basic ABC sequence. Two experienced
cognitive therapists, both trained at the Beck Institute, reviewed the booklet for consistency with
the cognitive model, and the booklet was revised to be consistent with their recommendations.
The second booklet, the ACB booklet, was designed as the reverse of the ABC booklet. It
presents an A-->C-->B sequence as authoritative and correct, and trains readers in an ACB
rationale. In this case the definitions of A, B, and C are the same (A represents an event or
stimulus, B represents cognition, and C represents affect), but the sequence of and relation
b
etween cognition and affect are reversed: Given a specific event or stimulus (A), one reacts with
an immediate emotional response (C), and this reaction subsequently influences how one
interprets the event (B). All attempts were made to use the same examples and sentences as in the
ABC booklet, but with the order of cognition and affect reversed to fit the different sequence.
However, sometimes these reversals did not result in content that was as logical or consistent as
in the ABC booklet, so the wording and examples were changed as appropriate. The two
experienced cognitive therapists judged that the booklet presented a reasonable and logical
alternative to the ACB booklet.
Before and after reading their assigned booklets, participants responded to a computerized image-
response task, designed to measure how participants experience the sequence of their thoughts
and feelings. Participants were shown a series of 60 images (ranging from low-arousal images,
such as a coffee cup or a flower, to high-arousal images, such as two individuals having sex or a
burnt, decaying animal) on a computer screen and were asked to press a key indicating whether
they experienced thoughts first or feelings first in response to each image. The percentage of
thoughts to total responses was computed for each participant's prebooklet and postbooklet
responses. Higher scores indicated more responding in a manner consistent with an ABC
sequence, in other words, reporting more thoughts first in response to the images. Because
participants were forced into responding with either a thought or a feeling, lower scores indicated
more responding in a manner consistent with an ACB sequence, that is, having more feelings first
in response to the images. We predicted that those who were "told" that thoughts come first
would report more instances in which they had thoughts first; conversely, those who were "told"
that feelings come first would report more instances in which feelings came first.
Results for 60 ABC and 60 ACB subjects are displayed in Figure 1.2 Participants receiving the
ABC booklet reported more thoughts in response to the images after reading the booklet, whereas
participants receiving the ACB booklet reported more feelings in response to the images after
reading the booklet. In the ABC group, participants claimed to have thoughts first 58% of the
time before reading the ABC booklet but did so 70% of the time afterward. Conversely, in the
ACB group, participants claimed to have thoughts first 55% of the time before reading the ACB
booklet, but this dropped to 44% of the time afterward.3 These results are consistent with a
psychotherapeutic demand characteristics interpretation, but note that although this study
purports to measure an analogue of psychotherapeutic demand characteristics, it also contains
experimental demand characteristics. In other words, subjects may have been influenced by the
content of the booklets (psychotherapeutic demand) and may have easily apprehended a purpose
of the experiment that would lead to hypothesis-confirming behavior (experimental demand).
Evidence for experimental demand comes from a postexperimental inquiry: Qualitative analyses
of open-ended responses suggested that most (61%) subjects guessed that a change on the
posttest measure was expected in response to the booklet. This was not the actual purpose (few
subjects guessed that there was a second booklet), but the demand placed by this understanding
would lead to hypothesis-confirming responses. The pretest–intervention–posttest design of the
study fostered easy understanding of its purpose, and no attempt was made to conceal it.
Evidence for psychotherapeutic demand comes from the same data (39% of subjects did not
believe that the purpose of the experiment was to test the influence of the booklets). Also, 89% of
subjects reported that they believed the booklet and 96% reported that it made sense to them,
suggesting the booklet also may have had some effect (however, these questions also may contain
experimental demand). The strong effect found in this study (only 24% of subjects did not change
in the hypothesized manner) suggests that both influences may have been operative. Furthermore,
this finding is consistent with earlier findings on the effects of treatment rationales for systematic
desensitization, showing large demand effects from believable rationales.
Demand characteristics in the CT treatment manual. The evidence from the analogue study is
suggestive but does not address the possibility that demand characteristics may be present during
the presentation of the treatment rationale in CT—as practiced in the community—that influence
the nature of clients' psychological experiences. Concerning the possibility of real influence in
CT as practiced, a useful parallel may be that of the false belief paradigm: Is it possible that
cognitive therapists could be planting their model of the nature of psychological experience for
some clients just as false memories can be planted? Simply put, is it possible that cognitive
therapists are planting cognitions? As Loftus (1993) built a case for the possibility of planted
false memories in clients from nonexperimental sources such as therapist accounts of the therapy
process, so too may a case be made for the possibility of planted cognitions in CT from an
analysis of the CT manual (Beck et al., 1979).
Several sections of the CT manual (Beck et al., 1979) that suggest how therapists present the
cognitive rationale to clients are noteworthy. First, a therapist strictly following the manual will
not find encouragement to consider alternative conceptualizations of depression in which
distorted cognitions are not primary. The manual suggests that early in treatment, the cognitive
rationale should be presented to all clients and the role of thoughts in the determination of
feelings should be repeatedly demonstrated (pp. 142–148). If a client suggests that he or she is
not having specific cognitions in response to an event, the recommendation is to view the client
as unaware of the cognitions (Beck & Greenberg, 1974/1995, p. 12) and for continued practice
and completion of homework assignments aimed at bringing cognitions into awareness. The
manual (Beck et al., 1979) states, "Some patients may have difficulty identifying dysfunctional
thoughts or images, or they may not see the connection between thoughts and feelings. . . . In
either case it is desirable to demonstrate to the patient the presence of cognitions" (p. 149). If a
patient disagrees with the rationale, this disagreement is interpreted as distorted thinking to be
challenged and corrected as treatment evolves (p. 90).
An interesting facet of CT is the attempt to frame the client–therapist relationship as one of
mutual, scientific collaboration, in which the client and therapist are working together as
scientists out to discover the truth, whatever it may be. For example, the manual suggests
presenting homework assignments as experiments: "A high degree of compliance with homework
can be achieved if the therapist presents the assignment as an experiment" (Beck et al., 1979, p.
274). To illustrate, Beck et al. (1979, pp. 274–275) described an exchange between a patient and
a therapist in which the patient initially rejects the homework assignment because he does not
think it will be helpful. The therapist then reframes the assignment to the patient as a hypothesis
to test out—"I have a hypothesis that it will help you and you have a hypothesis that it won't. I
don't know for sure who is right. Do you?"—and the client concedes to do the assignment. In this
fashion, the manual suggests that therapists persist in covert application of the cognitive model
while presenting a stance of experimentation and openness to the client and asking that the client
do the same. This transformed therapy-as-science-experiment now bears much resemblance to
typical research experiments, in which the researcher has a hypothesis and attempts to conceal it
from the participant. However, in CT, the therapist has a certain degree of control over the
interpretation of the results of these experiments. Here, unwittingly, Beck et al. may most
powerfully invoke the high degree of control and demands for hypothesis-confirming behavior
well researched through basic investigations into demand characteristics in experimental settings.
Consider a hypothetical depressed client. Perhaps, as is often the case, this client has entered
treatment after exploring other options and feels that this is her only hope for recovery. However,
as is often the case, the client is not aware that there are different schools of thought about the
nature and treatment of depression. She looks to the therapist for guidance and hope, assumes that
the therapist is knowledgeable, authoritative, and working from a scientifically informed
position—all of which, in part, are valid assumptions. The therapist, without offering alternatives,
provides an explanation for her depression that is plausible, unfalsifiable, and linked to specific
treatment techniques. In addition, the therapist, without explicit consent or acknowledgment that
this is happening, begins training the client to identify thoughts that presumably intervene
between the event and the affect. This training occurs through the presentation of a rationale,
examples, demonstrations, homework assignments, and feedback, and it is conducted in the spirit
of mutual collaboration. The therapist and client are two investigators, together out to discover
the truth. However, if the client disagrees with the utility of this approach or suggests that it is not
appropriate, the therapist responds with reassurance that learning to recognize thoughts is
difficult and takes time, and uses the disagreement as an additional example of the veracity of the
model—the disagreement is itself a distorted cognition. The therapist then assigns additional
homework assignments aimed at reinforcing the cognitive model and continues assigning
homework assignments until the goal of awareness of distorted cognitions is achieved, at which
point the therapist responds favorably to the client.
Summary
In this article, we have attempted to build a case for the possibility of demand characteristics in
psychotherapy rationales, using examples from DID, the repressed memory controversy,
systematic desensitization for phobias, and CT for depression. We do not wish to suggest that all,
or even any, CT therapists practice CT in the manner suggested herein. There is no definitive
proof that CT therapists do practice CT in this manner or use and follow the CT manual explicitly
and directly in their practice (Goldfried & Wolfe, 1998), it is an assumption based on a variety of
plausible inferences. This analysis of the CT manual is not intended as a critique; CT is an
effective therapy (Hollon, Shelton, & Davis, 1993), although its mechanisms of action are
disputed (Barber & DeRubeis, 1989; Beckham & Watkins, 1989).
The provision of the treatment rationale may be a point early in therapy when clients are
particularly vulnerable to unintended influences, and treatment protocols that adhere to a
particular model to the exclusion of others may extend this vulnerability to other points in
therapy, when opportunities to assess a client's fit with the model are appropriate but not utilized.
Orne recognized this possibility. Orne and Bates (1992) described how a rationale can function as
a demand characteristic:
Our patients are highly motivated to accept the formulations we give them, because the process
of labeling and describing a condition implies that the disorder is understood and instills hope
that it can be successfully treated. As a therapeutic rationale is gradually imparted, patients come
to define themselves in terms of it. Hence, patients of psychoanalysts eventually develop Oedipal
conflicts, patients of cognitive–behaviorists predictably evidence cognitive distortions, patients o
f
existential therapists reliably display profound existential anxiety, and so on. In a true sense, the
therapist is correct in each of these situations, for there is evidence in the therapeutic material to
support each distinct model of treatment. (p. 250)
Researchers and clinicians have not addressed the possibilities raised by Orne and Bates (1992).
In fact, with the notable exception of exposure to the literatures on DID and false memories, such
as described above, there is little to no training in or appreciation for the role of demand
characteristics in psychotherapy rationales. Sensitivity to the influence of demand characteristics
has yet to be achieved, even in the best of our empirically based treatments, like CT for
depression. Indeed, demand characteristics may flourish unchecked in psychotherapy.
References
Abroms, G. M. (1968). Persuasion in psychotherapy. American Journal of Psychiatry, 124,
1212–1219.
Addis, M. E., & Carpenter, K. M. (2000). The treatment rationale in cognitive behavioral
therapy: Psychological mechanisms and clinical guidelines. Cognitive and Behavioral Practice,
7, 147–156.
Barber, J. P., & DeRubeis, R. J. (1989). On second thought: Where the action is in cognitive
therapy for depression. Cognitive Therapy and Research, 13, 441–457.
Beck, A. T. (1967). Depression: Clinical, experimental and theoretical aspects. New York:
Harper & Row.
Beck, A. T., & Greenberg, R. L. (1995). Coping with depression. Bala Cynwyd, PA: The Beck
Institute. (Originally published 1974)
Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression. New
York: Guilford Press.
Beckham, E. E., & Watkins, J. T. (1989). Process and outcome in cognitive therapy. In A.
Freeman, K. Simon, L. Beutler, & H. Arkowitz (Eds.), Comprehensive handbook of cognitive
therapy (pp. 61–81). New York: Plenum Press.
Bernstein, D. A., & Nietzel, M. T. (1973). Procedural variation in behavioral avoidance tests.
Journal of Consulting and Clinical Psychology, 41, 165–174.
Bernstein, D. A., & Nietzel, M. T. (1977). Demand characteristics in behavior modification: The
natural history of a "nuisance." In M. Hersen, R. M. Eisler, & P. M. Miller (Eds.), Progress in
behavior modification (Vol. 4, pp. 119–162). New York: Academic Press.
Clark, D. A., Beck, A. T., & Alford, B. A. (1999). Scientific foundations of cognitive theory and
therapy of depression. New York: Wiley.
Ellis, A. (1973). Humanistic psychotherapy: The rational-emotive approach. New York:
McGraw Hill.
Ellis, A., & Harper, R. A. (1975). A new guide to rational living. Englewood Cliffs, NJ: Prentice
Hall.
Frank, J. D. (1961). Persuasion and healing. Baltimore: Johns Hopkins University Press.
Frank, J. D. (1971). Therapeutic factors in psychotherapy. American Journal of Psychotherapy,
25, 350–361.
Frank, J. D. (1982). Therapeutic components shared by all psychotherapies. In J. H. Harvey & M.
M. Parks (Eds.), Psychotherapy research and behavior change (pp. 5–37). Washington, DC:
American Psychological Association.
Goldfried, M. R., & Wolfe, B. E. (1998). Toward a more clinically valid approach to therapy
research. Journal of Consulting and Clinical Psychology, 66, 143–150.
Greenspoon, J. (1954). The effect of two nonverbal stimuli on the frequency of members of two
verbal response classes. American Psychologist, 9, 384.
Greenspoon, J. (1955). The reinforcing effect of two spoken sounds on the frequency of two
responses. American Journal of Psychology, 68, 409–416.
Harris, M. J., & Rosenthal, R. (1985). Mediation of interpersonal expectancy effects: 31 meta-
analyses. Psychological Bulletin, 97, 363–386.
Haugaard, J. J., Reppucci, N. D., Laird, J., & Nauful, T. (1991). Children's definitions of the truth
and their competency as witnesses in legal proceedings. Law and Human Behavior, 15, 253–271.
Hollon, S. D., Shelton, R. C., & Davis, D. D. (1993). Cognitive therapy for depression:
Conceptual issues and clinical efficacy. Journal of Consulting and Clinical Psychology, 61, 270–
275.
Horvath, P. (1984). Demand characteristics and inferential processes in psychotherapeutic
change. Journal of Consulting and Clinical Psychology, 52, 616–624.
Hyman, I. E., Husband, T. H., & Billings, F. J. (1995). False memories of childhood experiences.
Applied Cognitive Psychology, 9, 181–197.
Ilardi, S. S., & Craighead, W. E. (1994). The role of nonspecific factors in cognitive-behavior
therapy for depression. Clinical Psychology: Science and Practice, 1, 138–156.
Laurence, J. R., & Perry, C. (1983). Hypnotically created memory among highly hypnotizable
subjects. Science, 222, 523–524.
Leigland, S. (1995). An experimental analysis of ongoing verbal behavior: Reinforcement, verbal
operants, and superstitious behavior. Analysis of Verbal Behavior, 13, 79–104.
Loftus, E. F. (1993). The reality of repressed memories. American Psychologist, 48, 518–537.
Loftus, E. F. (1997). Creating false memories. Scientific American, 277, 70–75.
Loftus, E., & Ketcham, K. (1991). Witness for the defense: The accused, the eyewitness, and the
expert who puts memory on trial. New York: St Martin's Press.
Loftus, E. F., & Ketcham, K. (1994). The myth of repressed memory. New York: St. Martin's
Press.
Loftus, E. F., & Mazzoni, G. A. L. (1998). Using imagination and personalized suggestion to
change people. Behavior Therapy, 29, 691–706.
Marcia, J. E., Rubin, B. M., & Efran, J. S. (1969). Systematic desensitization: Expectancy change
or counterconditioning? Journal of Abnormal Psychology, 74, 382–387.
Mazzoni, G. A. L., Loftus, E. F., & Kirsch, I. (2001). Changing beliefs about implausible
autobiographical events: A little plausibility goes a long way. Journal of Experimental
Psychology: Applied, 7, 51–59.
Mazzoni, G. A. L., Lombardo, P., Malvagia, S., & Loftus, E. F. (1999). Dream interpretation and
false beliefs. Professional Psychology: Research and Practice, 30, 45–50.
McReynolds, W. T., & Tori, C. (1972). A further assessment of attention-placebo effects and
demand characteristics in studies of systematic desensitization. Journal of Consulting and
Clinical Psychology, 38, 261–264.
Oliveau, D. C., Agras, W. S., Leitenberg, H., Moore, R. C., & Wright, D. E. (1969). Systematic
desensitization, therapeutically oriented instructions and selective positive reinforcement.
Behaviour Research and Therapy, 7, 27–33.
Orne, M. T. (1962a). Implications for psychotherapy derived from current research on the nature
of hypnosis. American Journal of Psychiatry, 118, 1097–1103.
Orne, M. T. (1962b). On the social psychology of the psychological experiment: With particular
reference to demand characteristics and their implications. American Psychologist, 17, 776–783.
Orne, M. T. (1969). Demand characteristics and the concept of quasi-controls. In R. Rosenthal &
R. L. Rosnow (Eds.), Artifact in behavioral research (pp. 147–181). New York: Academic Press.
Orne, M. T., & Bates, B. L. (1992). Reflections on multiple personality disorder: A view from the
looking-glass of hypnosis past. In C. Pierce, M. Greenblatt & A. Kales (Eds.), The mosaic of
contemporary psychiatry in perspective (pp. 247–260). New York: Springer Verlag.
Orne, M. T., & Bauer-Manley, N. K. (1991). Disorders of self: Myths, metaphors, and demand
characteristics of treatment. In J. Strauss & G. R. Goethals (Eds.), The self: Interdisciplinary
approaches (pp. 93–106). New York: Springer-Verlag.
Rosenthal, R. (1963). On the social psychology of the psychological experiment: The
experimenter's hypothesis as unintended determinant of experimental results. American Scientist,
51, 268–283.
Shaughnessy, J. J., & Zeichmeister, E. B. (1997). Research methods in psychology (4th ed.). New
York: McGraw-Hill.
Smith, R. E., Diener, E., & Beaman, A. L. (1974). Demand characteristics and the behavioral
avoidance measure of fear in behavior therapy analogue research. Behavior Therapy, 5, 172–182.
Weekes, J. R., Lynn, S. J., Green, J. P., & Brentar, J. T. (1992). Pseudomemory in hypnotized
and task-motivated subjects. Journal of Abnormal Psychology, 101, 356–360.
Weinberger, J. (1995). Common factors aren't so common: The common factors dilemma.
Clinical Psychology: Science and Practice, 2, 45–69.
Wilkins, W. (1971). Desensitization: Social and cognitive factors underlying the effectiveness of
Wolpe's procedure. Psychological Bulletin, 76, 311–317.
Wilkins, W. (1978). Expectancy effects versus demand characteristics: An empirically
unresolvable issue. Behavior Therapy, 9, 363–367.
Footnote
1 What is the relationship between demand characteristics and experimenter expectancy effects
(Harris & Rosenthal, 1985; Rosenthal, 1963)? Experimenter expectancy effects have been
defined in two ways. In some cases, and often in texts on research methods (e.g., Shaughnessy &
Zeichmeister, 1997), they are defined as a form of observer bias: Data are systematically biased
because of the observer's expectations about subjects' behavior, owing to knowledge of the
study's hypotheses. Defined thus, expectancy effects are distinct from demand characteristics:
Expectancy effects bias outcomes through the behavior of the experimenters; demand
characteristics bias outcomes because of the influence of contextual cues on the subject's
behavior. However, Rosenthal originally (1963) and consistently defined expectancies as subject,
not experimenter, variables, and considerable research has examined the role of subjects'
expectations on experimental outcomes. This research also has examined the environmental
(including experimenter) variables responsible for changing or enhancing the subjects'
expectancies. Defined as such, expectancy effects may be seen as cognitive events mediating the
effects of certain demand characteristics pertaining to the experimenter's behavior and
expectations. Wilkins (1978) noted that these definitional differences are purely theoretical and
empirically unresolvable and that the preference for an expectancy versus demand explanation
will "depend on the paradigmatic preferences of the theoretician to employ, respectively,
cognitive–motivational or environmental–behavioral models of human functioning" (p. 366).
Thus, the wealth of research on expectancy effects defined as subject variables adds considerably
to our understanding of demand characteristics (for a review, see Harris & Rosenthal, 1985), and
this research is subsumed under the general topic of demand characteristics for the purposes of
this article.
2 The data presented herein are preliminary and subject to change when this study is completed.
3 The difference in mean scores between groups at posttest (ABC: M = .70, SD = .17; ACB: M =
.44, SD = .18), with pretest scores (ABC: M = .58, SD = .12; ACB: M = .55, SD = .13) entered as
covariates, is significant, F(1, 117) = 74.03, p < .001.
Figure 1. Proportion of thoughts to total responses for both groups at pretest and posttest.
A preview of this full-text is provided by American Psychological Association.
Content available from Prevention & Treatment
This content is subject to copyright. Terms and conditions apply.
