maladaptive behaviors that can be targeted for inter-
vention and may also help the clinician anticipate
which other symptoms (other than the target beha-
viors) may change as a result of treatment. These
predictions are based on the established covariation
patterns among the disorder’ssymptoms.Forexam-
ple, increasing the amount of social interaction
engaged in by a depressed patient may also result in
fewer reports of depressed mood. Although more
research is needed regarding the covariation among
problematic behaviors (Kazdin, 1985), the criteria sets
(symptom lists) for mental disorders at least give us
initial hypotheses about what behaviors may or may
not change as a result of treatment.
Finally, it is noteworthy that cognitive phenom-
ena and processes are now considered more legitimate
subjects for behavioral assessment and behavioral
intervention. In behavioral assessment, not only are
behaviors, antecedent/stimulus conditions, and conse-
quences sampled but so are “organismic”variables
(Goldfried & Davison, 1994). These organismic vari-
ables may include a variety of physiological factors,
but many cognitive variables are assessed as well. In
particular, client expectations are regarded as impor-
tant. The client’s expectations concerning the nature
and meaning of the presenting problem, the minimal
standards of success that the client sets, and the client’s
expectations from behavior therapy are just a few of
the variables that are assessed in contemporary behav-
ioral assessment. Behavior therapists appear to have
found that an overly rigid adherence to learning mod-
els that do not incorporate organismic variables is too
constraining.
However, this does not leave the door wide
open for anykind of physiological or cognitive
measure. Rather, these measures and methods
must satisfy the same rigorous standards set forth
for the more traditional behavioral methods. Valid-
ity must be demonstrated, not assumed. For exam-
ple, a cognitive measure purported to be related to
panic disorder (e.g., beliefs of“uncontrollability”)
must be correlated with other behavioral measures
of panic disorder symptoms, and changes in these
beliefs should result in some improvement in other
panic disorder symptoms and lead to a better out-
come in the future. Through these and other pro-
cedures, the concurrent and predictive validity of a
measure can be established as well as its treatment
utility (Hayes, Nelson, & Jarrett, 1987; Kazdin,
1985).CHAPTER SUMMARY
Behavioral assessment differs from traditional assess-
ment in several fundamental ways. Behavioral assess-
ment emphasizes direct assessments (naturalistic
observations) of problematic behavior, antecedent
(situational) conditions, and consequences (reinforce-
ment). By conducting such a functional analysis, clin-
icians can obtain a more precise understanding of the
context and causes of behavior. It is also important to
note that behavioral assessment is an ongoing process,
occurring at all points throughout treatment.
We have surveyed some of the more common
behavioral assessment methods. Behavioral interviews
are used to obtain a general picture of the presenting
problem and of the variables that seem to be
maintaining the problematic behavior. Observation
methods provide the clinician with an actual sample
(rather than a self-report) of the problematic behav-
ior. Observations can be made in naturalistic condi-
tions (as behavior typically and spontaneously occurs)
or under more controlled conditions (in simulated
or contrived situations or conditions). Behavioral
assessors may also have clients self-monitor (“self-
observe”) their own behaviors, thoughts, and
emotions. A variety of factors can affect both the
reliability and validity of observations, including the
complexity of the behavior to be observed, how
observers are trained and monitored, the unit of anal-
ysis chosen, the behavioral coding system that is used,
reactivity to being observed, and the representative-
ness of the observations.BEHAVIORAL ASSESSMENT 279