26 Feb 2005 19:22 AR AR240-CP01-03.

tex XMLPublishSM (2004/02/24) P1: JRX

10.1146/annurev.clinpsy.1.102803.143810

Annu. Rev. Clin. Psychol. 2005. 1:67–89

doi: 10.1146/annurev.clinpsy.1.102803.143810
Copyright c 2005 by Annual Reviews. All rights reserved
First published online as a Review in Advance on October 11, 2004

CLINICAL JUDGMENT AND DECISION MAKING∗

Howard N. Garb
Wilford Hall Medical Center, Lackland Air Force Base, San Antonio,
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

Texas 78236-5300; email: [email protected]

Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

Key Words personality assessment, psychological assessment, psychodiagnosis,

behavioral prediction, statistical prediction
■ Abstract When clinical psychologists make judgments, are they likely to be cor-
rect or incorrect? The following topics are reviewed: (a) methodological advances in
evaluating the validity of descriptions of personality and psychopathology, (b) recent
findings on the cognitive processes of clinicians, and (c) the validity of judgments
and utility of decisions made by mental health professionals. Results from research
on clinical judgment and decision making and their relationship to conflicts within the
field of clinical psychology are discussed.

CONTENTS
CLINICAL JUDGMENT AND DECISION MAKING . . . . . . . . . . . . . . . . . . . . . . . 67
METHODOLOGICAL ADVANCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
COGNITIVE PROCESSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Cognitive Heuristics and Biases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Diagnosis and Psychology of Categorization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
VALIDITY OF JUDGMENTS AND UTILITY OF DECISIONS . . . . . . . . . . . . . . . 71
Description of Personality and Psychopathology . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Case Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Behavioral Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Treatment Decisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
THE ROMANTIC AND EMPIRICIST TRADITIONS IN CLINICAL
PSYCHOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

CLINICAL JUDGMENT AND DECISION MAKING

Upon receiving the Bruno Klopfer Distinguished Contribution Award from the
Society for Personality Assessment, Caldwell (2004) gave an example of his suc-
cess interpreting the Minnesota Multiphasic Personality Inventory (Hathaway &
McKinley 1943):

∗
The U.S. Government has the right to retain a nonexclusive, royalty-free license in and to
any copyright covering this paper.

1548-5943/05/0427-0067$14.00 67
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

68 GARB

We got a severe 4-6-8 profile on a young woman. I looked at the tortured

implications of the pattern and somehow said, “She will have something like
cigarette burn scars on her hands, where her father prepared her to steel herself
to the suffering of life.” The round burn marks were on her hands and extended
a little way up her arms (Caldwell 2004, p. 9).
This is a remarkable interpretation. However, it is anecdotal in nature. Caldwell
did not present scientific data to support his style of test interpretation. Instead, he
selectively related his experiences.
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

Clinical psychologists are likely to have one of two reactions to the claims made
by Caldwell, depending on whether they are influenced by romantic or empiricist
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

ideas (Wood et al. 2003, pp. 92–94). Romantics are likely to be greatly impressed
and inspired by Caldwell’s accomplishments. Those from the empiricist tradition
are likely to have a more skeptical attitude. Empiricists insist that grand claims be
scientifically tested. Their point of view is exemplified by a new journal, The Scien-
tific Review of Mental Health Practice, which is devoted to the objective investiga-
tion of controversial and unorthodox claims in clinical psychology, psychiatry, and
social work. The distinction between romantics and empiricists seems especially
apt, given the title of Caldwell’s address: “My Love Affair with an Instrument.”
In this article, research on clinical judgment and its relationship to the romantic
and empiricist traditions in clinical psychology is examined. First, the following
topics are covered: (a) methodological advances in studying the validity of descrip-
tions of personality and psychopathology, (b) the cognitive processes of clinicians,
and (c) the validity of clinical judgments and the utility of treatment decisions.

METHODOLOGICAL ADVANCES

Important research has been conducted on clinical judgment and decision making,
but it has focused on issues that are relatively easy to address. For example, some
studies have described interrater reliability. Other studies have been conducted to
determine if judgments based on a small amount of information (e.g., results from
a single test) agree with results from a large set of information (e.g., judgments
based on an interview and history information).
Some important questions seldom have been addressed. For example, when a
clinical psychologist describes a client’s personality, it is difficult to know how
to determine if the psychologist is correct. The description can be compared to
other fallible indicators (e.g., self-report ratings, peer ratings), but it is also dif-
ficult to determine the validity of these indicators. Similarly, when assessing a
child, a clinician can systematically collect information from the child, parents,
teachers, and peers. However, what if the information from these different sources
is contradictory, as is often the case?
Two new approaches to evaluating the validity of descriptions of personality
and psychopathology can be described. One approach is to have clients make rat-
ings every day. For example, in a superb study, Wu & Clark (2003) constructed
a behavior record form to measure daily activities. Participants were instructed to
indicate whether they had performed a particular behavior on a given day (e.g.,
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

CLINICAL DECISION MAKING 69

got into an argument, lost their temper). This particular measure was developed to
provide measurements of aggression, impulsivity, and exhibitionism. The advan-
tage of having participants make self-ratings every day is that they do not have to
rely on memory. Also, by making ratings for specific behaviors, they do not have
to make judgments that require more than a low degree of inference. If we were
interested in learning whether psychologists in clinical practice can report whether
their clients are aggressive, impulsive, or exhibitionistic, then it would be of value
to compare their judgments to results from this behavior record.
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

A second approach involves improvements in research design and statistical

analysis. Kraemer et al. (2003) have proposed an elegant solution for what has
long been a recalcitrant problem. Psychologists routinely collect information from
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

more than one source. For example, children can make self-ratings and can be
described by parents, teachers, and peers. When these descriptions are not in
agreement, it may be unclear which description is valid.
According to Kraemer et al. (2003), an informant’s report reflects the influences
of (a) the actual characteristics that we want to measure (e.g., traits, symptoms,
competencies), (b) the context in which the subject is observed, (c) the perspec-
tive of the informant, and (d) the error of measurement. Context refers to the
setting/place and circumstance in which the rater knows the person being rated
(e.g., home versus outside of home). Perspective refers to characteristics of the
informant that influence his or her assessment of the trait (e.g., someone who is
making self-ratings will have a different perspective than will other people). Given
this framework, our goal is to obtain a measure of a trait that is relatively free from
variance attributable to context (or setting) and perspective (biases of the infor-
mants). To do this, one must gather ratings from multiple informants who know
the client in different contexts and who have differing perspectives. For example,
ratings in the context of school could be made by a teacher and a child. Ratings in
the context of home could be made by a parent and the child. Perspective would
vary as a function of self (child) and other (teacher, parent). All of the trait ratings
could then be analyzed using a principal components analysis. If one is successful,
the principal components analysis will yield separate factors for trait(s), context,
and perspective. If this occurs, then one will have obtained a measure of trait(s)
that is closer to the gold standard than the individual ratings made by the infor-
mants because one will have removed variance due to context and perspective. This
would represent a significant advancement in being able to evaluate the validity of
a clinician’s ratings of traits and symptoms.

COGNITIVE PROCESSES
Describing the cognitive processes of clinical psychologists could prove beneficial
for improving training and clinical practice. It would be helpful if we could compare
the cognitive processes of trainees with those of expert clinicians. This would allow
supervisors to give empirically guided feedback to the students and trainees, e.g.,
on what information to attend to and what questions to ask. Remarkably, this rarely
has been done. In fact, little is know about individual differences among clinicians.
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

70 GARB

Research on cognitive processes is also important because computers and ar-

tificial intelligence are likely to play increasingly large roles in our lives (Garb
2000). The more we know about the cognitive processes of clinicians, the more
intelligently we will be able to harness the power of artificial intelligence for
making judgments and decisions. Although we are far from reaching our goals,
advances continue to be made. Recent work on heuristics and biases are mentioned
below, followed by a description of work on the psychology of categorization and
psychodiagnosis.
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

Cognitive Heuristics and Biases

Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

Cognitive heuristics and biases, as formulated by Tversky & Kahneman (1974),

have been frequently used to describe how clinical psychologists and other people
make judgments (e.g., Arkes 1981, Dawes 1986, Garb 1998, Kayne & Alloy 1988,
Turk & Salovey 1988, Wedding & Faust 1989). Thus, progress made in studying
heuristics and biases is likely to inform research on clinical judgment. With this in
mind, it is important to note that in the opinion of Kahneman (2003, p. 703), the
formulation of the affect heuristic is “probably the most important development
in the study of judgment heuristics in the past few decades.”
The affect heuristic was introduced and described by Slovic et al. (2002):
Affective responses occur rapidly and automatically—note how quickly you
sense the feelings associated with the stimulus words treasure or hate. We
argue that reliance on such feelings can be characterized as the affect heuristic
(p. 397).
One can think of many instances when affect is likely to guide clinical judgment
and decision making. On the one hand, the affect heuristic may have positive effects
and be related to clinical intuition and the setting of meaningful treatment goals.
However, it may also yield negative effects and be related to biases (e.g., race
bias and gender bias). It is not yet clear if the research methods that have been
employed in studying the affect heuristic in everyday judgment making can be
usefully employed in the study of clinical judgment.
Research on the affect heuristic may have important implications for a long-
standing issue in clinical psychology: the issue of whether computers or clinical
judges should make judgments and decisions (Meehl 1954). To the extent that clin-
ical psychologists can be positively guided by their feelings, they can be expected
to do well compared to computers. However, in comparisons of these two meth-
ods, clinicians have rarely been more accurate than computers (Grove et al. 2000),
which suggests that the affect of the clinician may have limited value for making
accurate judgments and decisions. Research is needed to directly address this issue.

Diagnosis and Psychology of Categorization

Research methods used by cognitive psychologists to study categorization have
been usefully employed to study clinical judgment. Results suggest that
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

CLINICAL DECISION MAKING 71

psychologists frequently make diagnoses by comparing clients to prototypes (e.g.,

Cantor et al. 1980; also see Blashfield et al. 1985, 1989; Evans et al. 2002; Garb
1996). A prototype is a clinician’s conception of a hypothetical client who best
exemplifies a particular disorder. The results are important because they suggest
that interrater reliability will be low when psychologists do not share the same
prototypes. The results also suggest that agreement between clinicians’ diagnoses
and diagnoses based on the Diagnostic and Statistical Manual of Mental Disor-
ders (DSM; American Psychiatric Association 1994) criteria will be affected by
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

the dissimilarity of the clinicians’ prototypes to the criteria.

Recent results indicate that mental health professionals also make diagnoses
by forming causal theories (Kim & Ahn 2002; also see Wakefield et al. 1999). In
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

five outstanding but complex experiments, Kim & Ahn (2002) had psychologists
and psychology graduate students make ratings for a series of tasks. Some of
the ratings took from 2.5 to 6 hours to complete. For one of the tasks, the DSM
criteria and the associated symptoms listed in the DSM manual were presented
to judges. Participants were to draw arrows indicating cause-and-effect relations
between any of the symptoms that they felt were causally connected. They were also
instructed to describe the strength of the postulated causal relations. Psychologists
and psychology graduate students provided detailed causal explanations. Most of
these causal theories were quite complex. A composite of the drawings made by
psychologists and psychology graduate students for a specific phobia is presented
in Figure 1. Notice that in this example, the DSM criterion “Person recognizes
fear as excessive” was not perceived as being part of a causal relation with other
symptoms. Kim & Ahn then gave the judges additional tasks. They were able to
determine that the criteria and symptoms that had been described as forming part
of a causal relation were weighed more heavily than were other criteria when they
made diagnoses. For example, for the diagnosis of specific phobia, the criterion
“Person recognizes fear as excessive” was given a mean diagnostic importance
rating of only 75.5; ratings for the other DSM criteria ranged from 87.3 to 96.3.
This is of interest because when using the DSM, clinicians are supposed to weigh
each criterion equally.
Other results reported by Kim & Ahn (2002) are also of interest. For one
of the tasks, psychologists and psychology graduate students were instructed to
read a set of case studies. When they were later asked about the clients, they
recalled causally central symptoms more often than causally peripheral symptoms
and isolated symptoms. In fact, the results revealed a bias to falsely recognize
symptoms that were causally central to clinicians’ theories of different disorders.

VALIDITY OF JUDGMENTS AND UTILITY OF DECISIONS

When clinicians describe clients, are they likely to be correct or incorrect? Below,
results are described for (a) the description of personality and psychopathology,
(b) diagnosis, (c) case formulation, (d) prediction, and (e) decision making.
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

72 GARB
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

Figure 1 Composite of participants’ drawings of their causal theories of specific phobia.

The American Psychiatric Association’s Diagnostic and Statistical Manual of Mental
Disorders-Fourth Edition (1994) diagnostic criteria are shown in bold. From NS Kim & W
Ahn (2002). Copyright American Psychological Association. Reprinted with permission.

Description of Personality and Psychopathology

Practicing psychologists seem to be less sophisticated at describing personality
traits than at making diagnoses. At least for diagnoses, there are explicit rules (DSM
criteria), and we can hope to have an idea of what types of judgments clinicians are
making. Although we can be certain that clinical psychologists attend to symptoms
and consider whether a client has a personality disorder when they evaluate and
treat a client, it is frequently not clear if they spend much time thinking about the
client’s personality traits. For example, research suggests that the evaluation of a
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

CLINICAL DECISION MAKING 73

client’s normal-range personality traits (e.g., persistence, conscientiousness) can

be helpful for treatment planning (Garb 2003), but it is not clear that clinicians
consider these traits when thinking about a client. This is a complicated issue
because cognitive processing frequently occurs outside of one’s awareness (e.g.,
Kihlstrom 1999). Perhaps a clinician will consciously think about a trait if the client
falls toward the end of the dimension. Similarly, a large body of research supports
the Big Five Model of personality (Costa & Widiger 2002), but results from this
area of research do not guide most psychologists. Instead, psychologists may be
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

guided by implicit models of personality that owe more to everyday experience

than to scientific findings. In other words, this area of research is so primitive, we
are not even sure what types of personality impressions clinicians form.
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

In reviewing the literature, it is clear that questions about the validity of de-
scriptions of personality and psychopathology are usually sidestepped or answered
indirectly. Remarkably, in these studies, clinicians virtually never make their judg-
ments based on all of the information that is usually available in clinical practice.
This is not true of studies on diagnosis, prediction of behavior, and treatment
decision making. Important research has been conducted on the description of
personality and psychopathology, but it has focused on answering questions that
are easily answered.
A popular approach to studying clinical judgment in this area is to examine
interrater reliability: One can see if different clinicians make similar ratings when
evaluating the same set of clients. Overall, interrater reliability has varied widely
for describing personality traits, but has often been good for describing psychiatric
symptoms (Garb 1998, pp. 10–14).
Recent work on interrater reliability has focused on the description of defense
mechanisms. Overall, results from two studies suggest that psychologists should
not be describing defense mechanisms unless they receive extensive training. In
both studies, clinicians made ratings using the Defense Functioning Scale, a scale
that was introduced for further study in the DSM-IV (American Psychiatric As-
sociation 1994). Perry et al. (1998, p. 56) concluded, “Our findings indicate that
the defense axis is a feasible, acceptably reliable, and nonredundant addition to
DSM-IV,” but their results can also be interpreted as indicating that there is little
support for adding a defense mechanism axis to the next edition of the DSM. For
example, they reported that median kappa reliabilities were 0.42 for describing
individual defenses. A kappa value less than 0.40 typically indicates that interrater
reliability is poor. A median kappa value of 0.42 suggests that almost half of the
kappa reliabilities were below 0.40, so interrater reliability was presumably poor
in many cases. In the other study (Hilsenroth et al. 2003), positive results were
obtained: The mean intraclass correlation coefficient for ratings of six defense
levels was 0.59. However, the study was not conducted under typical clinical con-
ditions: ratings were made after a 2-hour semistructured clinical interview and a
1- to 1.5-hour interpretive/feedback interview.
Studies on validity have also been conducted. Typically, one indicator of a
construct has been related to another indicator of the construct (e.g., ratings based
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

74 GARB

on test results have been compared to ratings based on an interview). This is an

appropriate, although limited, strategy. The results illustrate difficulties in making
clinical judgments. For example, research generally suggests that it is difficult to
draw valid inferences from nonverbal behavior (Ambady & Rosenthal 1992, Garb
1998). Also, psychologists frequently are not more accurate than graduate students
(Garb 1989, 1998). Difficulty in learning from clinical experience is a topic that is
discussed at the end of this article.
Recent work has evaluated the ability of clinicians to detect lying. In most stud-
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

ies, people have not been accurate at this task (DePaulo 1994). However, by coding
videotapes using the Facial Action Coding System (Ekman & Friesen 1978), one
can detect some individuals who are lying (Frank & Ekman 1997). Furthermore,
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

judges have been able to detect lying after viewing videotapes of interviews under
the following condition: when the experimenters select videotapes that differ in the
expression of emotion for subjects who lied and subjects who told the truth (Ekman
& O’Sullivan 1991, Ekman et al. 1999). Thus, by carefully selecting videotapes to
show clinicians, one can increase the likelihood that the clinicians’ ratings will be
valid. This result is of theoretical interest: When individuals show facial muscular
movement that indicates they are lying, clinicians’ ratings will have modest va-
lidity. However, the result is of limited clinical significance for two reasons. First,
individuals who lie do not always show facial muscular movement that is associ-
ated with lying. Second, considering that the chance level of accuracy was 50%, a
majority of the psychologists obtained only a modest level of accuracy. For exam-
ple, for “regular” clinical psychologists, 4% achieved 0%–30% levels of accuracy,
52% achieved 40%–60% accuracy, and 44% achieved 70%–100% accuracy.

Diagnosis
Acceptable levels of interrater reliability have been reported for psychodiagnoses in
field trials for both DSM-III (American Psychiatric Association 1980, pp. 470–471)
and for the tenth revision of the International Classification of Diseases (Sartorius
et al. 1993, 1995). However, it is likely that these results were obtained because
the clinicians participating in the field trials were familiar with and adhered to
diagnostic criteria. In clinical practice, many mental health professionals frequently
do not adhere to diagnostic criteria (e.g., Blashfield & Herkov 1996, Davis et al.
1993, Ford & Widiger 1989, Morey & Ochoa 1989); therefore, one might expect
that diagnoses made in routine clinical practice will be dissimilar to diagnoses
based on semistructured interviews. Semistructured interviews are used to ensure
that diagnoses are based on specific criteria and rules. The scientific literature
reveals that agreement is generally poor between diagnoses made in routine clinical
practice and diagnoses based on semistructured interviews (Shear et al. 2000,
Strakowski et al. 1997; also see Garb 1998, pp. 53–54).
Several factors indicate that semistructured interviews are more valid than diag-
noses made in clinical practice. First, clinicians do not always ask about important
symptoms. In one study (Miller et al. 2001), clinicians evaluated only about 50%
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

CLINICAL DECISION MAKING 75

of the key criteria that were assessed using semistructured interviews. Of course,
some clinicians will evaluate all of the relevant criteria, but the research suggests
that many do not. Second, it can be noted that interrater reliability is better for
semistructured interview diagnoses than for clinical diagnoses. Finally, agreement
with diagnoses made by expert clinicians has been better for diagnoses based
on semistructured interviews than for diagnoses made in routine clinical practice
(Basco et al. 2000, Miller et al. 2001). In response to these findings, Widiger &
Samuel (2004) recommended that clinicians administer a self-report inventory to
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

alert themselves to the potential presence of maladaptive personality traits followed

by a semistructured interview to verify their presence.
At the same time, we must remain aware of the limitations of semistructured
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

interviews. Semistructured interviews clearly do not represent a gold standard.

Improvements in methodology need to be applied to further evaluate their validity,
e.g., by collecting longitudinal data (Spitzer 1983). In fact, it is possible that in
some circumstances, diagnoses based on therapy sessions will be more accurate
than diagnoses based on semistructured interviews, because therapists have the
opportunity to observe and interact with clients over time. With other clients,
it is possible that self-report inventories will be more valid than semistructured
interviews (Widiger & Samuel 2004). Perhaps the key to using semistructured
interviews is to recognize that they should not be done in rote fashion and that
interviewers should be careful in making inferences from the assessment data. In
particular, clients should feel that there is a therapeutic element to the interviews
and that they can elaborate on their answers.
Interestingly, Westen (1997) has argued against using semistructured interviews
for diagnosing personality disorders. He concluded that no personality-disorder
semistructured interview or questionnaire has “demonstrated acceptable evidence
that it validly assessed the constructs it purported to assess” (p. 895), but he sup-
ported his conclusion by citing a review by Perry (1992). The review by Perry
(1992) covered results from only nine studies. A more recent review (Widiger
2002) covered results from 35 studies on the diagnosis of personality disorders.
Widiger (2002, p. 463) found that “convergent validity generally improves as the
degree of structure increases.” For example, the worst convergent validity coef-
ficient was obtained in the only study to have used unstructured interviews by
practicing clinicians.
Clinicians make fewer diagnoses than do research investigators using semistruc-
tured interviews, presumably because they inquire about fewer symptoms (Basco
et al. 2000, Zimmerman & Mattia 1999). For example, in one study (Basco et al.
2000), the semistructured interview method identified 96 current comorbid mental
disorders in a sample of 200 patients. Clinicians using routine diagnostic methods
identified only 35 current comorbid diagnoses. Not only do clinicians make fewer
diagnoses than research investigators, research indicates that nearly all comorbid
mental disorders are underdiagnosed in routine clinical practice (Garb 1998). For
example, the mental disorders of mentally retarded clients are frequently under-
diagnosed (Jopp & Keys 2001). Similarly, evidence indicates that mental health
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

76 GARB

professionals underdiagnose substance abuse disorders among psychiatric inpa-

tients and mental disorders among substance abuse patients (e.g., Hansen et al.
2000, Kirchner et al. 1998). To describe the underdiagnosis of comorbid disor-
ders, the term “diagnostic overshadowing” is frequently used (Reiss et al. 1982).
An advantage of using a semistructured interview is that the questions one asks
will not vary as a function of patient characteristics such as race. In fact, to learn
if diagnoses made by clinicians are biased, clinicians’ diagnoses are frequently
compared with diagnoses made by research investigators. If differences between
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

the two sets of ratings are related to race, gender, or other variables, then one
can conclude that one set of diagnoses is biased. Bias has been observed for
diagnosis. For example, for diagnoses made in clinical practice and diagnoses
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

based on semistructured interviews, agreement is higher for white patients than

for nonwhite patients (Strakowski et al. 1997), and schizophrenia is frequently
overdiagnosed in African American patients (e.g., Whaley 2001).
Controversies exist over whether some mental disorders are overdiagnosed.
Historically, schizophrenia was overdiagnosed in the United States, at least until
the advent of DSM-III in 1980 when the criteria for schizophrenia were made
more stringent (Cooper et al. 1972). More recently, questions have been raised
about the possible overdiagnosis of dissociative identity disorder (formerly multi-
ple personality disorder; Lilienfeld et al. 1999) and attention deficit/hyperactivity
disorder (ADHD; LeFever et al. 2003). Surprisingly, post-traumatic stress disorder
(PTSD), after a period of wide acceptance, has become marked by controversy. In
fact, in March 2004, Congress held a hearing on the possible overdiagnosis of this
disorder. The controversy was instigated in part by the finding that although only
15% of the men who served in Vietnam were assigned to combat units, the Na-
tional Vietnam Veterans Readjustment Study yielded a lifetime PTSD prevalence
rate of 30.9% (Burkett & Whitley 1998, McNally 2003a). In addition, after ex-
amining military records via the Freedom of Information Act, Burkett & Whitley
(1998) estimated that about 75% of veterans receiving PTSD compensation are
pretenders. According to Burkett & Whitley (1998), the high prevalence rate of
PTSD reported for the National Vietnam Veterans Readjustment Study may have
occurred because military records were not checked to verify statements made by
veterans. Congress has recently become involved because they want to know how
many soldiers are likely to return from Iraq with PTSD. It is also of interest to note
that delayed onset PTSD has not been observed to occur in longitudinal studies of
individuals suffering from exposure to noncombat trauma.

Case Formulation
In addition to describing personality and psychopathology and making diagnoses,
psychologists also try to understand the causes of their clients’ behaviors and
symptoms. As already mentioned, even when they make other types of judgments
(e.g., diagnoses), they frequently consider causal factors (Kim & Ahn 2002). Thus,
case formulation is a key task for clinical psychologists—one that is likely to affect
how they perform on other tasks.
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

CLINICAL DECISION MAKING 77

Unfortunately, making causal judgments can be extremely difficult, and is per-

haps more difficult than other tasks facing mental health professionals. It is simply
more difficult to explain things than to describe them. In fact, for my book on clin-
ical judgment (Garb 1998), I did not locate a single study on causal judgments in
which validity was good or excellent for individual clinicians, regardless of whether
the specific task was related to behavioral, cognitive behavioral, or psychodynamic
assessment.1 In reviewing research on psychological assessment, I concluded:
Relatively little research has been conducted to evaluate the reliability and
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

validity of case formulations. The research that has been conducted suggests
that the validity of case formulations is often poor (Garb 1998, p. 100).
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

Recent studies have obtained similar findings. In particular, interrater reliability

continues to be poor. For example, psychologists and psychology graduate students
in one study (Daleiden et al. 1999) made judgments about the reinforcement func-
tions of children who refuse to attend school. Case descriptions of school-refusing
children were obtained from an anxiety disorder clinic. Clinicians were to make
ratings for different types of negative reinforcers that could be maintaining the
children’s behavior. Judgments made by individual clinicians were of question-
able interrater reliability (e.g., values of kappa ranged from 0.12 to 0.17). Relia-
bility was acceptable when judgments were aggregated across clinicians (values
of kappa were as large as 0.70). However, in clinical practice, case formulations
are typically made by a single clinician, not by averaging ratings across clini-
cians. Similarly, in a second study (Persons & Bertagnolli 1999), mental health
professionals made cognitive behavioral formulations after listening to recordings
of initial interviews with three depressed women. Interrater reliability for schema
ratings (ratings of underlying cognitions) was poor for individual clinicians (mean
intraclass correlation coefficient, or ICC, equal to 0.37), but good when schema
ratings were averaged over five judges (ICC = 0.72).
Psychologists and other mental health professionals, and even individuals in-
volved in public mental health policy, frequently make causal judgments that are
not supported by empirical research. Instead, judgments frequently are based on
informal observations or clinical lore. A recent example was given by Baumeister
et al. (2003, p. 1):
Teachers, parents, therapists, and others have focused efforts on boosting
self-esteem, on the assumption that high self-esteem will cause many positive
outcomes and benefits. . . . Our findings do not support continued widespread
efforts to boost self-esteem in the hope that it will by itself foster improved
outcomes.
At one time, the state of California had funded a task force on self-esteem
because of the belief that “raising self-esteem would help solve many of the state’s

1
However, interrater reliability was sometimes fair when conditions did not resemble clinical
practice, e.g., when ratings were averaged across clinicians.
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

78 GARB

problems, including crime, teen pregnancy, drug abuse, school underachievement,

and pollution” (Baumeister et al. 2003, p. 3). Research has not supported this belief.
Case formulations based on informal observations and clinical lore are often
based on the reports of clients and their families. However, research indicates that
memories are biased in systematic ways (for numerous examples, see McNally
2003b). In general, clients tend to remember their past behavior and mental status
in ways that conform with their own implicit theories of psychopathology. For
example, in one study (Rueter et al. 2000), adolescents completed forms indicating
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

whether they had engaged in behaviors related to conduct disorder during the
previous year. They completed the forms annually, starting in some instances at
age 12 and in other instances at age 14, and ending at age 15. Retrospective ratings
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

were made at age 19, at which time they were asked what criterion behaviors
they had engaged in prior to age 15. Participants who met antisocial criteria at
age 19 overreported having had behavior problems related to conduct disorder,
whereas participants who experienced a decline in antisocial behavior by age 19
underreported having had behavior problems related to conduct disorder in the
past. Thus, memories were distorted to conform to present behavior.
To increase the likelihood of making valid causal judgments, psychologists
should be more systematic in collecting history data. A well-replicated finding
is that many clinicians do not regularly ask clients about important events, e.g.,
if they have ever been sexually or physically abused (Garb 1998, pp. 88–89).
Similarly, Widiger & Clark (2000) implied that clinicians do not rigorously inquire
about the course of a client’s psychopathology. They recommended that DSM-V
require that clinicians systematically learn about the “life-span history of a patient’s
symptomatology . . . by recording, for example, age of onset, lifetime history of
disorders, and their longitudinal course” (p. 956). Finally, to decrease reliance
on retrospective reports of specific behaviors, clients can carry small handheld
electronic diaries that will allow for describing those behaviors in real-time and in
real-world contexts (Shiffman et al. 2002).
In summary, psychologists should be cautious when explaining behavior. Even
case formulations that seem reasonable may be incorrect. Psychologists should
also become more systematic and rigorous in collecting history data.

Behavioral Prediction
Behavioral prediction is one of the most exciting areas of clinical judgment. Al-
though progress has been slow for improving the validity of predictions for some
tasks (e.g., prediction of course of mental disorder, prediction of suicide), exciting
advances have been made for the prediction of violence and the prediction of re-
cidivism of sexual offenses. For these latter two tasks, new assessment instruments
are being adopted, and statistical prediction rules are having their greatest impact.

PREDICTION OF VIOLENCE Although early studies on the long-term prediction

of violence indicated that clinicians are wrong twice as often as they are correct
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

CLINICAL DECISION MAKING 79

(Monahan 1981), results from subsequent studies indicate that predictions of vi-
olence made by clinicians are often valid (Mossman 1994). For example, in a
well-known study on the long-term (six-month) prediction of violence (Lidz et al.
1993), clinicians made correct predictions for 58% of the patients who did not
become violent and 60% of those who did.
Although mental health professionals are able to predict violence with moder-
ate validity, race bias and gender bias have been documented. Black psychiatric
inpatients and black prison inmates often are predicted to be more violent than
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

are white psychiatric inpatients and white prison inmates, even when race is not
significantly related to the occurrence of violence (Garb 1998, pp. 113–114). This
finding continues to be replicated. In a recent study (Hoptman et al. 1999), psy-
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

chiatrists at a forensic psychiatric hospital were to predict which patients would

become assaultive during a three-month period. Fifty-seven percent of the patients
were African American. They committed 55% of the assaults. The psychiatrists
had predicted that they would commit 67% of the assaults.
With regard to gender bias, violence tends to be overpredicted for male patients
and underpredicted for females (Garb 1998, p. 115). A recent study (Elbogen et al.
2001) suggests that this is related to the gender of the mental health professionals. In
this study, male and female clinicians agreed upon general levels of dangerousness
for female patients, but female clinicians viewed male psychiatric patients as more
dangerous than did male clinicians.
Advances have occurred for the prediction of violence. The importance of eval-
uating psychopathy when predicting violence has been established (e.g., Mona-
han et al. 2001). Forensic psychologists, but not other licensed psychologists, are
generally careful to evaluate this construct when predicting violence (Tolman &
Mullendore 2003). In addition, a number of assessment instruments and aids have
gained acceptance among forensic psychologists, including measures of psychopa-
thy (e.g., the Hare Psychopathy Checklist-Revised, PCL-R; Hare 1991), statistical
prediction rules (e.g., the Violence Risk Appraisal Guide, VRAG; Quinsey et al.
1998), and rules and guidelines designed to structure the process for evaluating
risk for violence (e.g., the HCR-20, a 20-item checklist to assess the risk for future
violent behavior in criminal and psychiatric populations; Webster et al. 1998).
These advances have had a large impact on forensic psychology but apparently
not on clinical and counseling psychology. In a recent survey (Tolman & Mullen-
dore 2003), forensic psychologists, but not other licensed psychologists, reported
being aware of the scientific literature on predicting violence. Both groups of psy-
chologists also described the assessment instruments that they use. The PCL-R and
the VRAG were among the top five instruments used by forensic psychologists to
evaluate psychopathy. This was not true for the other licensed psychologists. Given
the controversy that has surrounded the use of the Rorschach inkblot test (Hunsley
& Bailey 1999, Lilienfeld et al. 2000), it is of interest to note that the Rorschach was
among the top five instruments used by nonforensic psychologists for conducting
general evaluations and for evaluating psychopathy. For forensic psychologists, the
Rorschach was not among the top five tests for either task. Similarly, in another
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

80 GARB

survey (Lally 2003), a majority of diplomates in forensic psychology deemed the

Rorschach to be unacceptable for a range of tasks including the evaluation of (a)
risk for violence, (b) risk for sexual violence, (c) competency to stand trial, (d)
competency to waive Miranda Rights, and (e) malingering.

PREDICTION OF SEXUAL CRIME Advances also have occurred for the prediction
of sexual recidivism. Recidivism rates for criminals committing sex offenses range
from about 35% to 55% (Hanson et al. 2003). In recent years, research on the actu-
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

arial prediction of sexual recidivism has increased dramatically: Only one finding
was identified in a review of the literature published before 1996 (Hanson &
Bussiere 1998), whereas more than 50 findings have been subsequently described
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

(Hanson et al. 2003). Actuarial measures that have been used to predict sexual
recidivism include the Violence Risk Appraisal Guide (VRAG; Harris et al. 1993),
the Sex Offender Risk Appraisal Guide (SORAG; Quinsey et al. 1998), the Rapid
Risk Assessment for Sex Offender Recidivism (RRASOR; Hanson 1997), and the
Static-99 (Hanson & Thornton 1999). Positive results have been obtained (Hanson
et al. 2003). When results from different studies have been pooled, the accuracy
of actuarial predictions has been statistically significant, medium in effect size
(d = 0.68, 95% confidence interval of 0.62 to 0.42), and significantly more accu-
rate than unstructured clinical predictions (d = 0.28, 95% confidence interval of
0.14 to 0.42). Thus, compared to clinical prediction, the use of actuarial scales is
a major improvement for predicting sexual recidivism.

Treatment Decisions
Treatment decisions often are based on interviews, medical records, and, in some
cases, test results. Research topics include (a) the utility of assessment instruments,
(b) the utility of standardized versus tailored treatment plans, (c) the appropriate-
ness of evaluations of competency to consent to treatment, and (d) the use of
evidence-based recommendations for treatment.
Nelson-Gray (2003) described a method for studying the utility of assessment
instruments. In the following example, she related using the method to evaluate a
semistructured interview:
If treatment was more successful for individuals whose diagnosis had been
established by a set of assessment devices that included the semistructured
interview than by a set of assessment devices that excluded the semistructured
interview, then . . . the treatment utility of the semistructured interview would
be demonstrated (p. 524).
This method is rarely used, so little is directly known about the utility of as-
sessment instruments such as semistructured interviews and psychological tests
(Hayes et al. 1987, Meehl 1959, Nelson-Gray 2003).
A treatment plan that is formulated on an individual basis is not necessarily
better than a standardized one. In one study (Schulte et al. 1992), 120 phobic
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

CLINICAL DECISION MAKING 81

clients were randomly assigned to one of three groups: (a) a control group, (b) a
standardized therapy group (exposure in vivo), and (c) a group receiving individ-
ually planned treatment. For the clients receiving individually planned treatment,
therapists were allowed to use all of the therapeutic methods commonly employed
in behavior therapy and cognitive therapy. Standardized treatment proved to be the
most successful, which suggests that in some instances the utility of standardized
treatment is greater than the utility of tailored treatment plans.
The most vigorous area of research on clinical decision making in the past six
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

to seven years has been on the use of evidence-based treatment recommendations.

But before this body of research is discussed, one other study is first described.
Although we have known for years that many psychiatrists frequently do not attend
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

to all of the necessary legal criteria when seeking civil commitments for psychiatric
hospitalizations (e.g., Bagby et al. 1991, Garb 1998), a recent study raises questions
about how they decide if a patient is competent to agree to electroconvulsive therapy
(Kitamura & Kitamura 2000). In this study, 176 members of the Japanese Society
of Psychiatry and Neurology read transcripts of competency interviews for five
patients who were recommended to undergo electroconvulsive therapy because
drugs had failed to improve their mental status. When indicating if a patient was
competent to make this treatment decision, interrater reliability among clinicians
was poor. The intraclass correlation coefficient was only 0.31.
The treatment of mental disorders may be transformed by the adoption of
evidence-based practice recommendations. For example, the Agency for Health
Care Policy and Research and the National Institute of Mental Health established
the Schizophrenia Patient Outcomes Research Team (PORT) to develop and dis-
seminate recommendations for treating schizophrenia based on scientific evidence.
Their recommendations addressed the use of antipsychotic agents as well as the use
of psychological, family, and other interventions (Lehman et al. 1998b). With the
publication of treatment recommendations, investigators have been able to evaluate
treatment decisions made by mental health professionals to learn if they are consis-
tent with evidence-based practice. In general, results indicate that implementation
of the treatment recommendations would transform clinical practice.
For the treatment of schizophrenia, mental health professionals frequently de-
viate from recommendations for evidence-based practice. For example, the PORT
(Lehman et al. 1998a) surveyed a stratified random sample of 719 individuals
diagnosed with schizophrenia in inpatient and outpatient settings. Treatment was
in conformance with the treatment recommendations less than half of the time,
although it was better for pharmacological than for psychosocial treatments. Treat-
ment recommendations were followed less often for minority patients. For exam-
ple, dosage of antipsychotic medication for an acute episode should be in the
range of 300–1000 chlorpromazine equivalents per day for a minimum of six
weeks. Lehman et al. (1998a) found that only 62.4% of inpatients received the
recommended dosage at discharge. Minority patients were more likely to be on
a high dose (>1000 chlorpromazine equivalents) than were Caucasian patients
(27.4% and 15.9%, respectively).
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

82 GARB

Since the publication of the PORT findings and recommendations, a large num-
ber of studies have been conducted to determine how well mental health profes-
sionals follow evidence-based practice when treating schizophrenia. In general,
half of the patients with schizophrenia frequently do not receive antipsychotic
medications within recommended dosages; black patients frequently receive ex-
cessively high dosages of antipsychotic medicine, and they are less likely to be
tried on atypical psychotics (Chen et al. 2000, Covell et al. 2002, Daumit et al.
2003, Dickey et al. 2003, DosReis et al. 2002, Herbeck et al. 2004, Kreyenbuhl
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

et al. 2003, Owen et al. 2001, Sohler et al. 2003, Valenstein et al. 2001, Valenti
et al. 2003, Walkup et al. 2000; also see Botts et al. 2003, Buchanan et al. 2002,
Remington et al. 2001, Weissman 2002, Woods et al. 2003). Conformance with
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

psychosocial treatment recommendations was studied less frequently, although, to

give just one example, Lehman et al. (1998) reported that patients were referred
for family therapy only 13.9% to 26.8% of the time, even though research sup-
ports the use of specified forms of family therapy in conjunction with psychotropic
medicine for the treatment of schizophrenia.

THE ROMANTIC AND EMPIRICIST TRADITIONS IN

CLINICAL PSYCHOLOGY
As noted above, two traditions in clinical psychology can be described. Both em-
piricists and romanticists base their judgments on a combination of scientific find-
ings, informal observations, and clinical lore, although empiricists place a greater
emphasis on scientific findings. A key distinction is that those in the romantic
tradition are likely to accept findings based on clinical validation, whereas those
in the empiricist tradition are likely to maintain a skeptical attitude. Thus, when
Caldwell (2004) interpreted the Minnesota Multiphasic Personality Inventory and
inferred that his client would have scars from being burned with cigarettes, he was
offering a clinical validation of his method for interpreting the inventory. Empiri-
cists insist upon empirical validation: They would insist that a study be conducted
with appropriate controls. That is, they would insist that an expert make judgments
for a series of clients and that a research investigator record the validity of the judg-
ments. It may seem harsh to question an anecdote told by an expert clinician, but
when expert psychologists have been the objects of study, their claims frequently
have not been supported (Garb 1989; Wood et al. 2003, pp. 136–142).
Silver (2001), upon receiving the American Psychological Association’s Award
for Distinguished Contributions to Applied Psychology, made an argument that
exemplifies the romanticist tradition. He minimized the importance of scientific
research and praised the process of clinical validation. Silver related that he first
forms an impression of a client using biographical and interview data. He then
predicts how a client will perform on a test. If his predictions turn out to be
accurate, he concludes that his impression of the client and his use of the test
are likely to be accurate. If his predictions are inaccurate, he collects additional
information. Silver gave the following example:
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

CLINICAL DECISION MAKING 83

I even use psychological tests whose validity is in question. For example, I will
administer the Draw-A-Person Test (Machover 1949). Again, I make specific
predictions. Without so doing, I would not use this test because of the weak
support for its validity (Silver 2001, p. 1009).
Just as Caldwell (2004) referred to clinical experience to validate his use of a test,
Silver does the same. The problem with Silver’s approach is that a psychologist’s
prediction of test scores will sometimes be correct by chance, not because the
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

psychologist’s overall impression of the client is accurate. When this occurs, the
psychologist will be misled into believing that his impression of the client and his
use of the test have been validated.
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

Another example illustrates the romantic and empiricist traditions in clinical

psychology (additional examples are given in Garb et al. 2004). As noted above,
psychologists frequently make diagnoses by comparing clients to prototypes. That
is, they frequently do not adhere to the DSM criteria, but instead form an impres-
sion of how similar a client is to their conception of the prototypical client with
a particular disorder. Shedler & Westen (2004) have recommended that the DSM
diagnostic criteria for personality disorders be revised so that mental health profes-
sionals simply rate how similar a client is to a prototype. To generate prototypes,
Shedler & Westen asked experienced psychiatrists and psychologists to describe
their conceptions (prototypes) of personality disorders. Shedler & Westen reported
that the prototypes “were clinically richer than the DSM descriptions” (p. 1350).
The approach recommended by Shedler & Westen (2004) falls into the romantic
tradition because it relies on the “combined experience of seasoned clinical prac-
titioners” while using statistical methods to aggregate the ratings made by those
clinicians (p. 1364). Using this approach, the DSM criteria would not be revised
on the basis of research studies on the etiology, nature, and course of a mental
disorder, but instead on the basis of clinicians’ observations.
Research on clinical judgment and decision making supports the empiricist tra-
dition. A large body of research indicates that it can be surprisingly difficult to
learn from informal observations, both because clinicians’ cognitive processes are
fallible and because accurate feedback on the validity of judgments is frequently
not available in clinical practice (Chapman & Chapman 1969; Garb 1989, 1998).
Furthermore, when clinical lore is studied, it often is found to be invalid. For
example, according to clinical lore, the comprehension and picture-arrangement
subtests of the Wechsler intelligence tests are sensitive measures of social judg-
ment. However, when this was examined empirically (Lipsitz et al. 1993), scores
on these Wechsler subtests were not significantly related to a clinician-rated mea-
sure of overall social competence. Furthermore, there is reason to believe that most
inferences based on Wechsler subtest variation “are either untested by science or
unsupported by scientific findings” (Kamphaus 1998, p. 46; also see Watkins 2003).
In conclusion, psychologists should reduce their reliance on informal observation
and clinical validation when (a) choosing an assessment instrument or treatment
intervention, (b) revising diagnostic criteria, and (c) making clinical judgments or
test interpretations.
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

84 GARB

The Annual Review of Clinical Psychology is online at

http://clinpsy.annualreviews.org

LITERATURE CITED
Ambady N, Rosenthal R. 1992. Thin slices of chotic pharmacy in the ambulatory care set-
expressive behavior as predictors of interper- ting, 1993–2000. Psychiatr. Serv. 54:1086
sonal consequences: a meta-analysis. Psy- Buchanan RW, Kreyenbuhl J, Zito JM, Lehman
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

chol. Bull. 111:256–74 A. 2002. The Schizophrenia PORT Pharma-

American Psychiatric Association. 1980. Diag- cological Treatment Recommendations: con-
nostic and Statistical Manual of Mental Dis- formance and implications for symptoms and
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

orders. Washington, DC: Am. Psychiatr. As- functional outcome. Schizophr. Bull. 28:63–
soc. 3rd ed. 73
American Psychiatric Association. 1994. Diag- Burkett BG, Whitley G. 1998. Stolen Valor:
nostic and Statistical Manual of Mental Dis- How the Vietnam Generation Was Robbed
orders. Washington, DC: Am. Psychiatr. As- of Its Heroes and Its History. Dallas, TX:
soc. 4th ed. Verity
Arkes HR. 1981. Impediments to accurate clin- Caldwell AB. 2004. My love affair with an in-
ical judgment and possible ways to mini- strument. J. Personal. Assess. 82:4–10
mize their impact. J. Consult. Clin. Psychol. Cantor N, Smith EE, French R, Mezzich
49:323–30 J. 1980. Psychiatric diagnosis as proto-
Bagby RM, Thompson JS, Dickens SE, No- type categorization. J. Abnorm. Psychol. 89:
hara M. 1991. Decision making in psychiatric 181–93
civil commitment: an experimental analysis. Chapman LJ, Chapman JP. 1969. Illusory cor-
Am. J. Psychiatry 148:28–33 relation as an obstacle to the use of valid
Basco MR, Bostic JQ, Davies D, Rush AJ, psychodiagnostic signs. J. Abnorm. Psychol.
Witte B, et al. 2000. Methods to improve 74:271–80
diagnostic accuracy in a community mental Chen RS, Nadkarni PM, Levin FL, Miller
health setting. Am. J. Psychiatry 157:1599– PL, Erdos J, et al. 2000. Using a computer
605 database to monitor compliance with phar-
Baumeister RF, Campbell JD, Krueger JI, Vohs macotherapeutic guidelines for schizophre-
KD. 2003. Does high self-esteem cause bet- nia. Psychiatr. Serv. 51:791–94
ter performance, interpersonal success, hap- Cooper JE, Kendell RE, Gurland BJ, Sharpe L,
piness, or healthier lifestyles? Psychol. Sci. Copeland JRM, et al. 1972. Psychiatric Di-
Public Interest 4:1–44 agnosis in New York and London. London:
Blashfield RK, Herkov MJ. 1996. Investigating Oxford Univ. Press
clinician adherence to diagnosis by criteria: Costa PT, Widiger TA. 2002. Personality Dis-
a replication of Morey and Ochoa 1989. J. orders and the Five-Factor Model of Person-
Personal. Disord. 10:219–28 ality. Washington, DC: Am. Psychol. Assoc.
Blashfield RK, Sprock J, Haymaker D, Hodgin 2nd ed.
J. 1989. The family resemblance hypothesis Covell NH, Jackson CT, Evans AC, Essock SM.
applied to psychiatric classification. J. Nerv. 2002. Antipsychotic prescribing practices in
Ment. Dis. 177:492–97 Connecticut’s public mental health system:
Blashfield RK, Sprock J, Pinkson K, Hodgin rates of changing medications and prescrib-
J. 1985. Exemplar prototypes of personal- ing styles. Schizophr. Bull. 28:17–29
ity disorder diagnoses. Compr. Psychiatry Daleiden EL, Chorpita BF, Kollins SH, Drab-
26:11–21 man RS. 1999. Factors affecting the reliabil-
Botts S, Hines H, Littrell R. 2003. Antipsy- ity of clinical judgments about the function
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

CLINICAL DECISION MAKING 85

of children’s school-refusal behavior. J. Clin. sonality disorders. J. Consult. Clin. Psychol.

Child Psychol. 28:396–406 57:301–5
Daumit GL, Crum RM, Guallar E, Power Frank MG, Ekman P. 1997. The ability to de-
NR, Primm AB, et al. 2003. Outpatient tect deceit generalizes across different types
prescriptions for atypical antipsychotics for of high-stakes lies. J. Personal. Soc. Psychol.
African Americans, Hispanics, and whites 72:1429–39
in the United States. Arch. Gen. Psychiatry Garb HN. 1989. Clinical judgment, clinical
60:121–28 training, and professional experience. Psy-
Davis RT, Blashfield RK, McElroy RA. 1993. chol. Bull. 105:387–96
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

Weighting criteria in the diagnosis of person- Garb HN. 1996. The representativeness and
ality disorder: a demonstration. J. Abnorm. past-behavior heuristics in clinical judgment.
Psychol. 102:319–22 Prof. Psychol. Res. Pract. 27:272–77
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

Dawes RM. 1986. Representative thinking in Garb HN. 1998. Studying the Clinician: Judg-
clinical judgment. Clin. Psychol. Rev. 6:425– ment Research and Psychological Assess-
41 ment. Washington, DC: Am. Psychol. Assoc.
DePaulo BM. 1994. Spotting lies: Can humans Garb HN. 2000. Computers will become in-
learn to do better? Curr. Dir. Psychol. Sci. creasingly important for psychological as-
3:83–86 sessment: not that there’s anything wrong
Dickey B, Normand S-LT, Hermann RC, Eisen with that? Psychol. Assess. 12:31–39
SV, Cortes DE, et al. 2003. Guideline rec- Garb HN. 2003. Incremental validity and the as-
ommendations for treatment of schizophre- sessment of psychopathology in adults. Psy-
nia: the impact of managed care. Arch. Gen. chol. Assess. 15:508–20
Psychiatry 60:340–48 Garb HN, Wood JM, Lilienfeld SO, Nezworski
DosReis S, Zito JM, Buchanan RW, Lehman MT. 2005. Roots of the Rorschach contro-
AF. 2002. Antipsychotic dosing and concur- versy. Clin. Psychol. Rev. In press
rent psychotropic treatments for Medicaid- Grove WM, Zald DH, Lebow BS, Snitz BE,
insured individuals with schizophrenia. Nelson C. 2000. Clinical versus mechanical
Schizophr. Bull. 28:607–17 prediction: a meta-analysis. Psychol. Assess.
Ekman P, Friesen WV. 1978. The Facial Ac- 12:19–30
tion Coding System. Palo Alto, CA: Consult. Hansen SS, Munk-Jorgensen P, Guldbaek B,
Psychol. Press Solgard T, Lauszus KS, et al. 2000. Psy-
Ekman P, O’Sullivan M. 1991. Who can catch choactive substance use diagnoses among
a liar? Am. Psychol. 46:913–20 psychiatric in-patients. Acta Psychiatr.
Ekman P, O’Sullivan M, Frank MG. 1999. Scand. 102:432–38
A few can catch a liar. Psychol. Sci. 10: Hanson RK. 1997. The Development of a Brief
263–66 Actuarial Scale for Sexual Offense Recidi-
Elbogen EB, Williams AL, Kim D, Tomkins vism. Ottawa, Ont.: Dept. Solicitor Gen.
AJ, Scalora MJ. 2001. Gender and per- Hanson RK, Bussiere MT. 1998. Predicting re-
ceptions of dangerousness in civil psychi- lapse: a meta-analysis of sexual offender re-
atric patients. Leg. Criminol. Psychol. 6: cidivism studies. J. Consult. Clin. Psychol.
215–28 66:348–62
Evans DL, Herbert JD, Nelson-Gray RO, Gau- Hanson RK, Morton KE, Harris AJR. 2003.
diano BA. 2002. Determinants of diagnos- Sexual offender recidivism risk: what we
tic prototypicality judgments of the per- know and what we need to know. Ann. NY
sonality disorders. J. Personal. Disord. 16: Acad. Sci. 989:154–66
95–106 Hanson RK, Thornton D. 1999. Static-99: Im-
Ford MR, Widiger TA. 1989. Sex bias in the proving Actuarial Risk Assessments for Sex
diagnosis of histrionic and antisocial per- Offenders. Ottawa, Ont.: Dept. Solicitor Gen.
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

86 GARB

Hare RD. 1991. The Hare Psychopathy Check- A Synthesis, ed. LY Abramson, pp. 295–365.
list-Revised (PCL-R). Toronto, Ont.: Multi- New York: Guilford
Health Syst. Kihlstrom JF. 1999. The psychological uncon-
Harris GT, Rice ME, Quinsey VL. 1993. Vi- scious. In Handbook of Personality, ed. LR
olent recidivism of mentally disordered of- Pervin, O John, pp. 424–42. New York: Guil-
fenders: the development of a statistical ford. 2nd ed.
prediction instrument. Crim. Justice Behav. Kim NS, Ahn W. 2002. Clinical psychologists’
20:315–35 theory-based representations of mental dis-
Hathaway SR, McKinley JC. 1943. The Min- orders predict their diagnostic reasoning and
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

nesota Multiphasic Personality Inventory. memory. J. Exp. Psychol.: Gen. 131:451–76

Minneapolis: Univ. Minn. Press Kirchner JE, Owen RR, Nordquist C, Fischer
Hayes SC, Nelson RO, Jarrett RB. 1987. The EP. 1998. Diagnosis and management of sub-
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

treatment utility of assessment: a functional stance use disorders among inpatients with
approach to evaluating assessment quality. schizophrenia. Psychiatr. Serv. 49:82–85
Am. Psychol. 42:963–74 Kitamura T, Kitamura F. 2000. Reliability
Herbeck DM, West JC, Ruditis I, Duffy FF, of clinical judgment of patients’ compe-
Fitek DJ, et al. 2004. Variations in use of tency to give informed consent: a case vi-
second-generation antipsychotic medication gnette study. Psychiatr. Clin. Neurosci. 54:
by race among adult psychiatric patients. 245–47
Psychiatr. Serv. 55:677–84 Kraemer HC, Measelle JR, Ablow JC, Essex
Hilsenroth MJ, Callahan KL, Eudell EM. 2003. MJ, Boyce WT, Kupfer DJ. 2003. A new
Further reliability, convergent and discrimi- approach to integrating data from multiple
nant validity of overall defensive function- informants in psychiatric assessment and re-
ing. J. Nerv. Ment. Dis. 191:730–37 search: mixing and matching contexts and
Hoptman MJ, Yates KF, Patalinjug MB, Wack perspectives. Am. J. Psychiatry 160:1566–
RC, Convit A. 1999. Clinical prediction of 77
assaultive behavior among male psychiatric Kreyenbuhl J, Zito JM, Buchanan RW, Soeken
patients at a maximum-security forensic fa- KL, Lehman AF. 2003. Racial dispar-
cility. Psychiatr. Serv. 50:1461–66 ity in the pharmacological management of
Hunsley J, Bailey JM. 1999. The clinical utility schizophrenia. Schizophr. Bull. 29:183–93
of the Rorschach: unfulfilled promises and an Lally SJ. 2003. What tests are acceptable for use
uncertain future. Psychol. Assess. 11:266–77 in forensic evaluations? A survey of experts.
Jopp DA, Keys CB. 2001. Diagnostic over- Prof. Psychol. Res. Pract. 34:491–98
shadowing reviewed and reconsidered. Am. LeFever GB, Arcona AP, Antonuccio DO.
J. Ment. Retard. 106:416–33 2003. ADHD among schoolchildren:
Kahneman D. 2003. A perspective on judg- evidence of overdiagnosis and overuse of
ment and choice: mapping bounded rational- medication. Sci. Rev. Ment. Health Pract. 2:
ity. Am. Psychol. 58:697–720 49–60
Kamphaus RW. 1998. Intelligence test inter- Lehman AF, Steinwachs DM, Co-Investigators
pretation: acting in the absence of evidence. of the PORT Project. 1998a. Patterns of usual
In WISC-III Clinical Use and Interpreta- care for schizophrenia: initial results from the
tion: Scientist-Practitioner Perspectives, ed. Schizophrenia Patient Outcomes Research
A Prifitera, DH Saklofske, pp. 39–57. New Team (PORT) client survey. Schizophr. Bull.
York: Academic 24:11–20
Kayne NY, Alloy LB. 1988. Clinician and pa- Lehman AF, Steinwachs DM, Co-Investigators
tient as aberrant actuaries: expectation-based of the PORT Project. 1998b. Translating
distortions in assessment of covariation. In research into practice: the Schizophrenia
Social Cognition and Clinical Psychology: Patient Outcomes Research Team (PORT)
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

CLINICAL DECISION MAKING 87

treatment recommendations. Schizophr. Bull. diagnosis of the DSM-III personality disor-

24:1–10 ders. J. Personal. Disord. 3:180–92
Lidz CW, Mulvey EP, Gardner W. 1993. The Mossman D. 1994. Assessing predictions of
accuracy of predictions of violence to others. violence: being accurate about accuracy. J.
JAMA 269:1007–11 Consult. Clin. Psychol. 62:783–92
Lilienfeld SO, Lynn SJ, Kirsch I, Chaves JF, Nelson-Gray RO. 2003. Treatment utility of
Sarbin TR, et al. 1999. Dissociative identity psychological assessment. Psychol. Assess.
disorder and the sociocognitive model: re- 15:521–31
calling the lessons of the past. Psychol. Bull. Owen RR, Feng W, Thrush CR, Hudson TJ,
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

125:507–23 Austen MA. 2001. Variations in prescribing

Lilienfeld SO, Wood JM, Garb HN. 2000. practices for novel antipsychotic medications
The scientific status of projective techniques. among Veterans Affairs Hospitals. Psychiatr.
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

Psychol. Sci. Public Interest 1:27–66 Serv. 52:1523–25

Lipsitz JD, Dworkin RH, Erlenmeyer-Kimling Perry JC. 1992. Problems and considerations in
L. 1993. Wechsler Comprehension and Pic- the valid assessment of personality disorders.
ture Arrangement subtests and social adjust- Am. J. Psychiatry 149:1645–53
ment. Psychol. Assess. 5:430–37 Perry JC, Hoglend P, Shear K, Vaillant GE,
Machover K. 1949. Personality Projection in Horowitz M, et al. 1998. Field trial of a diag-
the Drawing of the Human Figure. Spring- nostic axis for defense mechanisms for DSM-
field, IL: Thomas IV. J. Personal. Disord. 12:56–68
McNally RJ. 2003a. Progress and controversy Persons JB, Bertagnolli A. 1999. Interrater reli-
in the study of posttraumatic stress disorder. ability of cognitive-behavioral case formula-
Annu. Rev. Psychol. 54:229–52 tions of depression: a replication. Cogn. Ther.
McNally RJ. 2003b. Remembering Tra- Res. 23:271–83
uma. Cambridge, MA: Belknap Press of Har- Quinsey VL, Harris GT, Rice ME, Cormier
vard Univ. Press CA. 1998. Violent Offenders: Appraising and
Meehl PE. 1954. Clinical Versus Statistical Pre- Managing Risk. Washington, DC: Am. Psy-
diction: A Theoretical Analysis and a Review chol. Assoc.
of the Evidence. Minneapolis: Univ. Minn. Reiss S, Levitan G, McNally RJ. 1982. Emo-
Press tionally disturbed mentally retarded people:
Meehl PE. 1959. Some ruminations on the val- an underserved population. Am. Psychol. 37:
idation of clinical procedures. Can. J. Psy- 361–67
chol. 13:102–28 Remington G, Shammi CM, Sethna R,
Miller PR, Dasher R, Collins R, Griffiths P, Lawrence R. 2001. Antipsychotic dosing pat-
Brown F. 2001. Inpatient diagnostic assess- terns for schizophrenia in three treatment set-
ments: 1. Accuracy of structured vs. un- tings. Psychiatr. Serv. 52:96–98
structured interviews. Psychiatry Res. 105: Rueter MA, Chao W, Conger RD. 2000. The ef-
255–64 fect of systematic variation in retrospective
Monahan J. 1981. The Clinical Prediction of conduct disorder reports on antisocial per-
Violent Behavior. Washington, DC: US Gov. sonality disorder diagnoses. J. Consult. Clin.
Print. Off. Psychol. 68:307–12
Monahan J, Steadman HJ, Silver E, Appelbaum Sartorius N, Kaelber CT, Cooper JE, Roper MT,
PS, Robbins PC, et al. 2001. Rethinking Risk Rae DS, et al. 1993. Progress toward achiev-
Assessment: The MacArthur Study of Mental ing a common language in psychiatry: re-
Disorder and Violence. New York: Oxford sults from the field trial of the clinical guide-
Univ. Press lines accompanying the WHO classification
Morey LC, Ochoa ES. 1989. An investigation of mental and behavioral disorders in ICD-
of adherence to diagnostic criteria: clinical 10. Arch. Gen. Psychiatry 50:115–24
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

88 GARB

Sartorius N, Ustun TB, Korten A, Cooper JE, tion of specialization? Prof. Psychol. Res.
van Drimmelen J. 1995. Progress toward Pract. 34:225–32
achieving a common language in psychiatry. Turk DC, Salovey P, eds. 1988. Reasoning, In-
II: Results from the international field trials ference, and Judgment in Clinical Psychol-
of the ICD-10 diagnostic criteria for research ogy. New York: Free Press
for mental and behavioral disorders. Am. J. Tversky A, Kahneman D. 1974. Judgment un-
Psychiatry 152:1427–37 der uncertainty: heuristics and biases. Sci-
Schulte D, Kunzel R, Pepping G, Schulte- ence 185:1124–31
Bahrenberg T. 1992. Tailor-made versus Valenstein M, Copeland L, Owen R, Blow F,
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

standardized therapy of phobic patients. Adv. Visnic S. 2001. Delays in adopting evid-
Behav. Res. Ther. 14:67–92 ence-based dosages of conventional antipsy-
Shear MK, Greeno C, Kang J, Ludewig D, chotics. Psychiatr. Serv. 52:1242–44
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

Frank E, et al. 2000. Diagnosis of nonpsy- Valenti AM, Narendran R, Pristach CA. 2003.
chotic patients in community clinics. Am. J. Who are patients on conventional antipsy-
Psychiatry 157:581–87 chotics? Schizophr. Bull. 29:195–99
Shedler J, Westen D. 2004. Refining person- Wakefield JC, Kirk SA, Pottick KJ, Hsieh D.
ality disorder diagnosis: integrating science 1999. Disorder attribution and clinical judg-
and practice. Am. J. Psychiatry 161:1350– ment in the assessment of adolescent antiso-
65 cial behavior. Soc. Work Res. 23:227–38
Shiffman S, Gwaltney CJ, Balabanis MH, Liu Walkup JT, McAlpine DD, Olfson M, Labay
KS, Paty JA, et al. 2002. Immediate an- LE, Boyer C, et al. 2000. Patients with
tecedents of cigarette smoking: an analysis schizophrenia at risk for excessive antipsy-
from ecological momentary assessment. J. chotic dosing. J. Clin. Psychiatry 61:344–
Abnorm. Psychol. 111:531–45 48
Silver RJ. 2001. Practicing professional psy- Watkins MW. 2003. IQ subtest analysis: clin-
chology. Am. Psychol. 56:1008–14 ical acumen or clinical illusion? Sci. Rev.
Slovic P, Finucane M, Peters E, MacGregor Ment. Health Pract. 2:118–41
DG. 2002. The affect heuristic. In Heuris- Webster CD, Douglas KS, Eaves D, Hart SD.
tics and Biases: The Psychology of Intuitive 1998. The HCR-20: Assessing Risk for Vio-
Judgment, ed. T Gilovich, D Griffin, D Kah- lence (Version 2). Odessa, FL: Psychol. As-
neman, pp. 397–420. Cambridge, UK: Cam- sess. Resourc.
bridge Univ. Press Wedding D, Faust D. 1989. Clinical judgment
Sohler NL, Walkup J, McAlpine D, Boyer and decision making in neuropsychology.
C, Olfson M. 2003. Antipsychotic dosage Arch. Clin. Neuropsychol. 4:233–65
at hospital discharge and outcomes among Weissman EM. 2002. Antipsychotic prescrib-
persons with schizophrenia. Psychiatr. Serv. ing practices in the Veterans Healthcare
54:1258–63 Administration–New York metropolitan re-
Spitzer RL. 1983. Psychiatric diagnosis: Are gion. Schizophr. Bull. 28:31–42
clinicians still necessary? Compr. Psychiatry Westen D. 1997. Divergences between clinical
24:399–411 and research methods for assessing person-
Strakowski SM, Hawkins JM, Keck PE, McEl- ality disorders: implications for research and
roy SL, West SA, et al. 1997. The effects of the evolution of Axis II. Am. J. Psychiatry
race and information variance on disagree- 154:895–903
ment between psychiatric emergency service Whaley AL. 2001. Cultural mistrust and the
and research diagnoses in first-episode psy- clinical diagnosis of paranoid schizophre-
chosis. J. Clin. Psychiatry 58:457–63 nia in African American patients. J. Psy-
Tolman AO, Mullendore KB. 2003. Risk evalu- chopathol. Behav. Assess. 23:93–100
ations for the courts: Is service quality a func- Widiger TA. 2002. Personality disorders. In
26 Feb 2005 19:22 AR AR240-CP01-03.tex XMLPublishSM (2004/02/24) P1: JRX

CLINICAL DECISION MAKING 89

Handbook of Assessment and Treatment versial Inkblot Test. San Francisco: Jossey-
Planning for Psychological Disorders, ed. Bass
MM Antony, DH Barlow, pp. 453–80. New Woods SW, Sullivan MC, Neuse EC, Diaz E,
York: Guilford Baker CB, et al. 2003. Racial and ethnic ef-
Widiger TA, Clark LA. 2000. Toward DSM- fects on antipsychotic prescribing practices
V and the classification of psychopathology. in a community mental health center. Psy-
Psychol. Bull. 126:946–63 chiatr. Serv. 54:177–79
Widiger TA, Samuel DB. 2005. Evidence based Wu KD, Clark LA. 2003. Relations between
assessment of personality disorders. Psychol. personality traits and self-reports of daily be-
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

Assess. In press havior. J. Res. Personal. 37:231–56

Wood JM, Nezworski MT, Lilienfeld SO, Zimmerman M, Mattia JI. 1999. Psychiatric di-
Garb HN. 2003. What’s Wrong With the agnosis in clinical practice: Is comorbidity
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

Rorschach? Science Confronts the Contro- being missed? Compr. Psychiatry 40:182–91
P1: JRX
February 26, 2005 19:35 Annual Reviews AR240-FM

Annual Review of Clinical Psychology

Volume 1, 2005

CONTENTS
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

A HISTORY OF CLINICAL PSYCHOLOGY AS A PROFESSION IN AMERICA

(AND A GLIMPSE AT ITS FUTURE), Ludy T. Benjamin, Jr. 1
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

STRUCTURAL EQUATION MODELING: STRENGTHS, LIMITATIONS,

AND MISCONCEPTIONS, Andrew J. Tomarken and Niels G. Waller 31
CLINICAL JUDGMENT AND DECISION MAKING, Howard N. Garb 67
MOTIVATIONAL INTERVIEWING, Jennifer Hettema, Julie Steele,
and William R. Miller 91
STATE OF THE SCIENCE ON PSYCHOSOCIAL INTERVENTIONS FOR
ETHNIC MINORITIES, Jeanne Miranda, Guillermo Bernal, Anna Lau,
Laura Kohn, Wei-Chin Hwang, and Teresa La Fromboise 113
CULTURAL DIFFERENCES IN ACCESS TO CARE, Lonnie R. Snowden
and Ann-Marie Yamada 143
COGNITIVE VULNERABILITY TO EMOTIONAL DISORDERS,
Andrew Mathews and Colin MacLeod 167
PANIC DISORDER, PHOBIAS, AND GENERALIZED ANXIETY DISORDER,
Michelle G. Craske and Allison M. Waters 197
DISSOCIATIVE DISORDERS, John F. Kihlstrom 227
THE PSYCHOBIOLOGY OF DEPRESSION AND RESILIENCE TO STRESS:
IMPLICATIONS FOR PREVENTION AND TREATMENT,
Steven M. Southwick, Meena Vythilingam, and Dennis S. Charney 255
STRESS AND DEPRESSION, Constance Hammen 293
THE COGNITIVE NEUROSCIENCE OF SCHIZOPHRENIA, Deanna M. Barch 321
CATEGORICAL AND DIMENSIONAL MODELS OF PERSONALITY
DISORDER, Timothy J. Trull and Christine A. Durrett 355
THE DEVELOPMENT OF PSYCHOPATHY, Donald R. Lynam
and Lauren Gudonis 381
CHILD MALTREATMENT, Dante Cicchetti and Sheree L. Toth 409
PSYCHOLOGICAL TREATMENT OF EATING DISORDERS, G. Terence Wilson 439
GENDER IDENTITY DISORDER IN CHILDREN AND ADOLESCENTS,
Kenneth J. Zucker 467

vii
P1: JRX
February 26, 2005 19:35 Annual Reviews AR240-FM

viii CONTENTS

THE DEVELOPMENT OF ALCOHOL USE DISORDERS, Kenneth J. Sher,

Emily R. Grekin, and Natalie A. Williams 493
DECISION MAKING IN MEDICINE AND HEALTH CARE, Robert M. Kaplan
and Dominick L. Frosch 525
PSYCHOLOGY, PSYCHOLOGISTS, AND PUBLIC POLICY,
Katherine M. McKnight, Lee Sechrest, and Patrick E. McKnight 557
COGNITIVE APPROACHES TO SCHIZOPHRENIA: THEORY AND THERAPY,
Access provided by University of California - San Francisco UCSF on 12/11/14. For personal use only.

Aaron T. Beck and Neil A. Rector 577

STRESS AND HEALTH: PSYCHOLOGICAL, BEHAVIORAL, AND
BIOLOGICAL DETERMINANTS, Neil Schneiderman, Gail Ironson,
Annu. Rev. Clin. Psychol. 2005.1:67-89. Downloaded from www.annualreviews.org

and Scott D. Siegel 607

POSITIVE PSYCHOLOGY IN CLINICAL PRACTICE, Angela Lee Duckworth,
Tracy A. Steen, and Martin E. P. Seligman 629

INDEX
Subject Index 653