182 CHAPTER 5
Exclusion Criteria
In most RCTs, including the TDCRP, researchers investigate treatments for patients
who have a “pure” form of the DSM-IV-TR disorder of interest (Goldfried & Wolfe,
1998)—andonly that one disorder. The decision to exclude people with an addi-
tional disorder or disorders makes many patients ineligible to participate. In reality,
however, most patients have more than one disorder, which raises external validity
questions about research using this exclusion criteria. In fact, after treatment, people
with multiple disorders do not usually make the gains made by people who have
only a single disorder (Coryell et al., 1988).Researching Treatment
One problem for researchers studying a particular type of therapy is that a therapist
who claims to provide a particular type of therapy may in fact provide treatment that
is remarkably similar to that of a therapist using a different theoretical orientation
(Lambert & Bergin, 1994). This makes comparisons among types of therapies dif-
fi cult. Suppose that some self-defi ned CBT therapists focus on relationship issues, and
some self-defi ned IPT therapists give concrete suggestions for improving relationships.
If there were no differences in outcomes between the two types of treatment, it could
be because therapists in the two groups were all doing much the same thing!
In an effort to address this problem, RCTs generally require therapists to base
their treatments on detailed manuals that provide session-by-session guidance and
specify techniques to be used with patients. RCTs provide brief therapy, typically
from 6 to 16 sessions. Different therapeutic approaches use different manuals; the
most common theoretical approach for manual-based therapy is CBT. Manual-
based treatment ensures that all therapists who use one particular approach provide
similar therapy that is distinct from other types of therapy (Kendall, Holmbeck, &
Verduin, 2004; Nathan, Skinstad, & Dolan, 2000).
Such research has led to more than 108 different treatment
manuals for adults with 51 different DSM-IV-TR disorders
(Beutler & Harwood, 2002; Chambless & Ollendick, 2001). Re-
search aimed at identifying the unique specifi c components of a
successful therapy has progressed by adding, subtracting, or other-
wise changing specifi c components of a particular treatment manual,
and then comparing the results from the two variations. For
example, one RCT treatment group might receive CBT for depres-
sion, whereas another group would receive the same CBT plus an
interpersonal component of treatment (similar to IPT). If patients
in both groups fare the same, then the interpersonal component
cannot be considered an active ingredient in this sort of treatment.
Conversely, if the dual-component group fares better after treat-
ment, then it can be inferred that the added interpersonal compo-
nent made the treatment more effective. Similarly, to study your
grief box therapy, you might have one group in which the members
both make the boxes and reach out to friends and family and another
group in which the members would only make the boxes without
an emphasis on reaching out to friends and family.
In studies using this approach, the same therapist may perform both variants
of a treatment (in contrast to earlier RCTs, in which a different group of therapists
performed each type of therapy). By having each therapist perform different vari-
ants of treatment, researchers attempt to control for possible therapist variables
that might affect treatment.Allegiance Effect
Another issue in RCT research is the allegiance effect, in which studies conducted
by investigators who prefer a particular theoretical orientation tend to obtain data
that support that particular orientation (Luborsky et al., 1999). Specifi cally, RCT
investigators who support one type of treatment tend to have patients who do better
with that type of treatment, whereas patients in the same study (using the sameIn order to ensure that the particular type of
therapy under investigation is, in fact, the type
of therapy that is provided in the session,
researchers may have therapists use manuals
that specify in detail the goals of each session
and the methods to be used.Ted Szczepanski/Worth Publishers
Allegiance effect
A pattern in which studies conducted
by investigators who prefer a particular
theoretical orientation tend to obtain data
that supports that particular orientation.