Obsessive– Compulsive Disorder 471
Given the recent development of the CBOCI, investigations into its psychometric
properties are limited. The original validation study revealed that the Obsessions, Com-
pulsions, and Total Score have high internal reliability, factorial validity, and moder-
ate to strong convergence with other OCD symptom measures like the YBOCS (D.A.
Clark, Antony, et al., 2005). Moreover, criterion validity is strong, with OCD patients
scoring significantly higher on all three scales than nonobsessional anxious, depressed,
and nonclinical comparison groups.^1 Like all OCD symptom measures, the CBOCI has
lower discriminant validity as indicated by its moderate correlations with anxiety and
depression measures. The treatment sensitivity of the measure has not yet been investi-
gated and its test– retest reliability has not been determined in an OCD sample. Analysis
based on the validation sample indicated that a cutoff score of 22 on the CBOCI Total
Score yielded high sensitivity (90%) and specificity (78%) for distinguishing OCD from
student controls (D. A. Clark, 2006b). The CBOCI and manual are available from Pear-
son Assessment at pearsonassess.com.
Other OCD Symptom Measures
Three other OCD symptom measures are frequently employed to assess the frequency
and severity of obsessive and compulsive symptoms. The 42-item Obsessive– Compulsive
Inventory (OCI) was developed by Foa and colleagues to assess the frequency and distress
of seven symptom domains of OCD (Foa, Kozak, Salkovskis, Coles, & Amir, 1998). The
measure consists of separate frequency and distress scales for washing, checking, doubt-
ing, ordering, obsessing, hoarding, and mental neutralizing. Psychometric evidence for
the instrument is strong, although the discriminate validity of the hoarding subscale
is questionable (Foa et al., 1998). A revised brief version of the OCI was developed
that consists of 18 items comprised of six distress scales: washing, checking, ordering,
obsessing, hoarding, and neutralizing (Foa, Huppert, et al., 2002). The new OCI-R
subscales were highly correlated with the old OCI subscales (rs = .92) and had sound
psychometric characteristics. Two recent clinical studies confirmed a six- factor struc-
ture that corresponded to the OCI-R subscales, good convergent validity with other
OCD measures (although correlations with the YBOCS are rather weak), more mod-
est discriminant validity (i.e., moderate association with anxiety and depression mea-
sures), and consistent ability to discriminant OCD from nonobsessional anxious groups
(Abramowitz & Deacon, 2006; Huppert et al., 2007; see similar findings for nonclinical
samples reported by Hajcak, Huppert, Simmons, & Foa, 2004). Abramowitz and Dea-
con (2006) found that the OCI-R obsessions and checking subscales, in particular, had
modest relations with OC cognition measures and that an OCI-R Total cutoff score of
14 provided the best differentiation of OCD from other anxiety disorders. A significant
limitation of the OCI-R is its underrepresentation of obsessions (only three items) and
disproportionate weight toward compulsive symptoms (Grabill et al., 2008). The OCI-R
is published as an appendix to Foa et al. (2002).
The Padua Inventory (PI) is a 60-item questionnaire originally developed on an Ital-
ian sample to assess the distress associated with common obsessive and compulsive phe-
(^1) Gabrill et al. (2008) arrived at an erroneous conclusion about the instrument’s criterion- related validity
based on a misreading of the significant group differences reported in the validation study.