session six is very high (70%); this day Thomas was
taking an asthma medication that was sedating and
this likely influenced his behavior. During the first
treatment period (Social Story), we see that
Thomas’s time sitting at circle increased dramati-
cally. During the reversal period (the second base-
line period), his percentage of time sitting at circle
decreased dramatically. When treatment was rein-
troduced (Social Story), his behavior once again
improved. The reversal period was inserted
between the two treatment periods to enable the
investigators to demonstrate a causal relationship
between the intervention and Thomas’s behavior.
One difficulty with the ABAB procedure is
that withdrawing treatment could pose some
ethical problems. However, the seriousness of this
issue depends on the specific targeted behaviors
involved. For example, withdrawing a treatment
that markedly reduces serious self-harm in a child
might be seen as unethical and unnecessary.
Multiple Baseline Designs. In some cases, it is
impossible to use a reversal period. As we have
noted, there may be ethical constraints. Also, in
clinical research settings, therapists may be unwill-
ing to have their clients reexperience situations that
could reinstate the very behaviors they are seeking
to eradicate. In such cases, investigators may use a
multiple baseline design. Here, two or more behaviors
are chosen for analysis. Perhaps an institutionalized
patient has severe problems behaving in a responsi-
ble manner. He does not take care of his room, fails
to follow good personal hygiene, or does not show
up on time for work assignments. Baseline data are
collected for his behavior in both personal and
work settings. Next, immediate rewards are intro-
duced whenever he behaves responsibly in personal
settings but not in work settings. Then, after a pre-
scribed period of time, measurements of behavior in
both settings are once again collected.
The last phase involves rewards for responsible
behavior in both settings. If responsible behavior
increases in the personal setting following reward
but not in the work setting when there is no
reward, it may be possible that some unknown
and uncontrolled factor other than reward is
operative. But if reward is then shown to enhance
responsible behavior in the work setting also, it
seems very unlikely that any factors other than
reward are involved. The use of dual baselines
gives investigators increased confidence in their
manipulations.
A study by Moras, Telfer, and Barlow (1993)
exemplifies a variant of the multiple baseline design.
Their study is noteworthy for several reasons. First,
they applied a single-case methodology to a treat-
ment that is not primarily behavioral (unlike most
applications of the single-case methodology).
Second, they targeted a clinical condition that is
common—coexisting generalized anxiety disorder
and major depression—yet complex because of
the myriad of presenting symptoms. Finally, they
were able to test the hypothesis that each form of
treatment used in the combined treatment (anxiety
control treatment and interpersonal psychotherapy
for depression) would affect specifically those
symptoms for which the treatment was originally
developed.
Box 4-4 presents Moras et al.’s (1993) descrip-
tions of Case 2, a man in his 30s suffering from both
generalized anxiety disorder and major depression,
as well as his weekly scores on self-report measures
of anxiety and depression throughout the treat-
ment. As can be seen, Case 2’s levels of anxiety
(BAI scores) and depression (BDI scores) were ele-
vated and clinically significant before treatment was
started (at baseline weeks 1 and 2). Anxiety control
treatment (ACT) was administered first for six ses-
sions, followed by an assessment interview (Asmt).
Next, interpersonal psychotherapy for depression
(IDT) was administered for six sessions, followed
by another assessment interview.
Several features of the results for Case 2 are
noteworthy. First, his anxiety scores did drop sig-
nificantly over the course of ACT. However, so did
his depression scores. Further, both depression and
anxiety scores appeared to drop (although not as
dramatically) during the IPT phase of treatment.
Moras et al. concluded that this combined form of
treatment (ACT and IPT) seems potentially effica-
cious for patients with generalized anxiety disorder
and major depression. However, contrary to theRESEARCH METHODS IN CLINICAL PSYCHOLOGY 115