Journal of Anxiety Disorders 24 (2010) 700–708

Contents lists available at ScienceDirect

Journal of Anxiety Disorders

The impact of experiential avoidance and obsessive beliefs

on obsessive-compulsive symptoms in a
severe clinical sample
Rachel C. Manos a,c,∗ , Shawn P. Cahill a,1 , Chad T. Wetterneck b,2 , Christine A. Conelea a,c,3 ,
Ashley R. Ross c,4 , Bradley C. Riemann c,5
a
Department of Psychology, University of Wisconsin-Milwaukee, PO Box 413, Milwaukee, WI 53201, USA
b
Department of Psychology, University of Houston – Clear Lake, 2700 Bay Area Blvd., Houston, TX 77058, USA
c
Rogers Memorial Hospital, 37000 Valley Road, Oconomowoc, WI 53066, USA

a r t i c l e i n f o a b s t r a c t

Article history: The present study sought to replicate and extend a study by Abramowitz et al. (2009), who examined how
Received 27 October 2009 well experiential avoidance (EA) and obsessive beliefs predicted obsessive-compulsive (OC) symptoms
Received in revised form 5 May 2010 in a non-clinical sample. The current study utilized a severe, clinical, treatment-seeking sample (N = 108),
Accepted 5 May 2010
and examined how well EA and obsessive beliefs predicted changes in OC symptoms from pre- to post-
treatment. Findings were generally consistent with Abramowitz et al. EA was generally not related to
Keywords:
OC severity and did not add significantly to the prediction of OC symptom domains above and beyond
Obsessive-compulsive disorder
depression or general anxiety, whereas obsessive beliefs did. Pre- to post-treatment change in one type of
Experiential avoidance
Obsessive beliefs
obsessive belief (perfectionism/certainty), but not change in EA, predicted global change in OC severity.
Results suggest that EA as it is measured currently may not play a significant role in OC severity or changes
in OC severity across treatment.
© 2010 Elsevier Ltd. All rights reserved.

Experiential avoidance (EA) is defined as an unwillingness to
experience or remain in contact with unpleasant private experi-
ences (e.g., emotions, thoughts, images, etc.) through attempts to
escape from or avoid these experiences (Hayes, Wilson, Gifford,
Follette, & Strosahl, 1996). EA is a primary target of acceptance and
commitment therapy (ACT; Hayes, Strosahl, & Wilson, 1999) and
has received considerable research attention due to the prolifera-
tion of ACT research in the past decade. EA is associated with higher
levels of general psychopathology, depression, anxiety, trauma-
related symptoms, and lower quality of life (Hayes et al., 2004).
There is evidence that decreases in EA are associated with improve-
ments in psychopathology and that focusing treatment on EA may
∗ Corresponding author at: Department of Psychology, University of Wisconsin-
Milwaukee, PO Box 413, Milwaukee, WI 53201, USA. Tel.: +1 414 229 5078;
fax: +1 414 229 5219.
E-mail addresses: rcmanos@uwm.edu (R.C. Manos), cahill@uwm.edu
(S.P. Cahill), wetterneck@uhcl.edu (C.T. Wetterneck), cconelea@uwm.edu
(C.A. Conelea), ARoss@rogershospital.org (A.R. Ross), BRiemann@rogershospital.org
(B.C. Riemann).
1
Tel.: +1 414 229 5099.
2
Tel.: +1 281 283 3364.
3
Tel.: +1 414 229 6078.
4
Tel.: +1 262 646 1159.
5
Tel.: +1 262 646 1388.
lead to symptom improvements within an anxiety-disordered sam-
ple (Chawla & Ostafin, 2007).
It has recently been suggested that ACT interventions target-
ing EA may be a preferable alternative to exposure and response
prevention (ERP), the traditional cognitive-behavioral treatment
method utilized in the treatment of obsessive-compulsive dis-
order (OCD; Eifert & Forsyth, 2005; Twohig, Hayes, & Masuda,
2006). OCD is characterized by obsessions (i.e., intrusive, unrea-
sonable thoughts, images, fear, or worry) and compulsions or
rituals used to reduce the anxiety associated with the obsessions
(American Psychiatric Association, 2000). Eifert and Forsyth (2005)
have suggested that compulsions function as attempts at EA, in that
individuals with OCD engage in compulsions in order to “. . .control
or reduce their unwanted thoughts because they want to reduce
the negative affect associated with them (p. 58).” In support of this
view, an eight-session ACT intervention devoid of within-session
exposure in a sample of four participants led to reductions in EA
and reductions in OCD symptoms from pre- to post-treatment, but
this relationship could not be explored in depth due to the nature
of the study design (Twohig et al., 2006). Although treatment did
not explicitly involve therapist aided exposure techniques, a com-
ponent of treatment involved having patients make “behavioral
commitments” to engage in values-guided behavior that seemed
to have led to ERP-like activities in some instances. For example,
Twohig et al. (2006) list examples of behavioral commitments,

0887-6185/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.janxdis.2010.05.001
 R.C. Manos et al. / Journal of Anxiety Disorders 24 (2010) 700–708
including not engaging in rituals on campus, removing hoarded
material from one’s home or storage units, and spending time in
public locations that provoked anxiety or were previously avoided.
Thus, it seems possible that these behavioral commitments may
have led to symptom reduction via ERP strategies.
Whereas ACT interventions aim to reduce EA and increase
acceptance and willingness to experience unpleasant private expe-
riences (Hayes et al., 1999), cognitive-behavioral interventions
work to decrease and change maladaptive interpretations of those
private experiences (e.g., Abramowitz, 1997; Rachman, 1997).
This focus on changing cognitive distortions is beneficial when
used in conjunction with ERP (Abramowitz, 1997), and may
lead to a lower dropout rate than ERP without cognitive tech-
niques (Whittal, Robichaud, Thordarson, & McLean, 2008). The
Obsessive-Compulsive Cognitions Working Group (OCCWG; 2005)
identified three domains of beliefs related to the maladaptive
interpretation of thoughts by OCD sufferers, which are assessed
by the Obsessive Beliefs Questionnaire-44 (OBQ44; OCCWG,
2005): responsibility/threat estimation, perfectionism/certainty,
and importance/control of thoughts.
In an attempt to explore the relationship between EA and
obsessive-compulsive (OC) symptoms, Abramowitz, Lackey, and
Wheaton (2009) examined whether EA predicts OC symptoms
in a non-clinical sample and compared this to how well obses-
sive beliefs (as measured by the OBQ44) predict OC symptoms.
Abramowitz et al. (2009) found that EA and obsessive beliefs were
not significantly related to one another, and that EA did not sig-
nificantly add to the prediction of OC symptom dimensions above
and beyond depressive symptoms and obsessive beliefs. Obses-
sive beliefs, however, contributed to prediction of OC checking and
obsessing symptoms after controlling for EA and depression. They
concluded that perhaps the construct of EA was too general to sig-
nificantly add to the prediction of OC symptoms over more specific
constructs such as obsessive beliefs. Self-identified limitations of
this study include use of a non-clinical and non-treatment-seeking
sample, measuring the variables of interest at a single time point,
and use of a general measure of EA that may not discriminate
between various avoidance strategies. Furthermore, Abramowitz
et al. (2009) controlled for depressive symptoms in their analy-
ses but did not control for general anxiety. Finally, we note that
the Obsessive-Compulsive Inventory – Revised (OCI-R; Foa et al.,
2002) was used to measure severity of OC symptoms, whereas the
Yale – Brown Obsessive-Compulsive Scale (Y-BOCS; Goodman et
al., 1989) is the most widely used measure of OC symptom severity
(Björgvinsson & Hart, 2007).
The current paper attempts to replicate and extend results of
Abramowitz et al. (2009) while addressing their identified limi-
tations, with the exception of the measure of EA as more specific
measures of EA have not yet been validated. Specifically, the present
study examines the impact of EA and obsessive beliefs on OC symp-
toms in two ways. The first is by conducting analyses similar to
those in the Abramowitz et al. (2009) study among EA, obses-
sive beliefs, and OC symptoms (including use of the self-report
Yale-Brown Obsessive-Compulsive Scale [Y-BOCS-SR; Baer, Brown-
Beasley, Sorce, & Henriques, 1993]) above and beyond depression
and general anxiety in a severe, clinical, treatment-seeking sam-
ple. The second is by examining relationships among EA, obsessive
beliefs, and OC symptoms with respect to changes occurring from
admission to discharge. Based on the Abramowitz et al. (2009)
findings, it is expected that EA will not significantly predict admis-
sion level OC symptoms over and above depression, anxiety or
obsessive beliefs, whereas obsessive beliefs will predict OC symp-
toms over and above depression, anxiety, and EA. It is further
predicted that change in obsessive beliefs will predict change in
OC symptoms, whereas change in EA will not predict change in OC
symptoms.
701
1. Method
1.1. Participants and procedures
Participants included 108 adults (59 females) with a primary
diagnosis of OCD who were admitted to either a residen-
tial OCD treatment program (n = 88) or an intensive outpatient
OCD treatment program (n = 20) at Rogers Memorial Hospital in
Oconomowoc, WI. Before admission to one of the OCD treat-
ment programs, all prospective patients undergo a 90-min phone
assessment that includes the Y-BOCS and other anxiety measures.
The information is then reviewed by Dr. Bradley Riemann, Clin-
ical Director of the Obsessive-Compulsive Disorder Center and
Cognitive-Behavioral Therapy Services, who determines whether
the person is appropriate for one of the OCD treatment programs
at Rogers Memorial Hospital, and, if so, recommends which pro-
gram would be best for the patient. Then, each patient has an
initial assessment with a psychiatrist experienced in the assess-
ment and treatment of OCD and who made the official diagnosis
and up to four comorbid psychiatric diagnoses. Most (54%) of the
current sample had at least one additional diagnosis, with the most
common comorbid diagnosis being an affective disorder. Of those
with psychiatric comorbidity, 36% had a secondary diagnosis of
an affective disorder. For the present sample, admissions dates
ranged from May 2003 to August 2007. The average length of stay
was 52 days (SD = 26.85; range = 6–143). Participants ranged in age
from 18 to 65 (M = 32.10 years, SD = 12.44). Upon admission to the
hospital, each participant completed an admissions packet consist-
ing of various questionnaires. Measures of interest for the current
study were grouped into three categories (measures of OCD sever-
ity, predictor variables, and control variables) and are described in
Section 1.2.
Treatment consisted of intensive cognitive-behavioral therapy,
with an emphasis on ERP with adjunctive cognitive restructuring
and medication management. Weekday treatment programming
on the residential unit included 30 min of a homework review
group, 180 min of therapist aided and self-directed exposure,
60 min of therapist aided cognitive restructuring, 60 min of recre-
ational therapy, 45 min of self-directed cognitive restructuring and
90 min of additional self-directed exposure. Residential partici-
pants were also offered 120 min of voluntary group therapy per
week. Required weekend programming per day included 30 min of
homework review, 120 min of self-directed exposures, and 90 min
of cognitive restructuring. Intensive outpatient program patients
attended treatment programming three hours per day Monday
through Thursday. Programming included a 15-min homework
review group at the beginning and end of each day, and 180 min
of a combination of self-directed and therapist aided exposure and
cognitive restructuring. Programming totaled 12 h per week with
an additional required two hours of treatment-related homework
every day. Patients in both programs received one-on-one time
with behavioral therapists each treatment day.
1.2. Measures
1.2.1. OCD severity
Two measures were used to assess severity of OCD. The self-
report version of the Yale-Brown Obsessive-Compulsive Scale
(Y-BOCS-SR; Baer et al., 1993) contains five items that rate sever-
ity of obsessions and five that rate severity of compulsions on a
scale ranging from 0 to 4, for a total score ranging from 0 to 40
(0–7 = subclinical, 8–15 = mild, 16–23 = moderate, 24–31 = severe,
and 32–40 = extreme). Although the Y-BOCS 10-item interview ver-
sion is the most comprehensive assessment available for OCD,
Baer et al. (1993) found that scores on the self-report version and
the interview version were very highly correlated (r = .97). The
702 R.C. Manos et al. / Journal of Anxiety Disorders 24 (2010) 700–708
Y-BOCS-SR has acceptable internal consistency (˛ = .78) for OCD
samples and adequate test-retest reliability over a one-week inter-
val (r = .79) in a non-clinical sample (Steketee, Frost, & Bogart, 1996).
Internal consistency for the current sample was unavailable, as only
the subscale and total scores for the Y-BOCS-SR were included in
the database.
The Obsessive-Compulsive Inventory – Revised (OCI-R; Foa et
al., 2002) is an 18-item measure of distress associated with obses-
sions and compulsions. Total scores range from 0 to 72 with high
scores indicating greater distress. The OCI-R has six subscales:
washing (WA; concern with contamination), checking (CH; check-
ing locks, etc., checking if one has made a mistake), ordering (OR;
ordering and arranging objects), obsessing (OB; intrusive violent,
sexual, blasphemous thoughts), hoarding (HO; collecting and sav-
ing objects), and neutralizing (NE; counting, mental counting).
Subscale scores range from 0 to 12. For the current study, items
from the original Obsessive-Compulsive Inventory (OCI; Foa, Kozak,
Salkovskis, Coles, & Amir, 1998) distress scale were used to create
subscale scores of the OCI-R by selecting relevant items. All sub-
scales demonstrated acceptable internal consistency in the current
sample (WA ˛ = .89, CH ˛ = .82, OR ˛ = .93, OB ˛ = .82, HO ˛ = .82, NE
˛ = .81).
1.2.2. Predictor variables
Two predictor variables were evaluated: OCD specific beliefs
and EA. The Obsessive Beliefs Questionnaire – 44 (OBQ44; OCCWG,
2005) is a 44-item self-report measure of obsessive belief domains
associated with OCD: responsibility/threat estimation (referred
to as OBQ44-RT; “I often think things around me are unsafe”),
perfectionism/certainty (OBQ44-PC; “If I’m not absolutely sure of
something, I’m bound to make a mistake”), and importance/control
of thoughts (OBQ44-ICT; “For me, having bad urges is as bad
as actually carrying them out”). Tolin, Worhunsky, and Maltby
(2006) reported good internal consistencies for the three subscales
(OBQ44-PC ˛ = .93, OBQ44-ICT ˛ = .90, OBQ44-RT ˛ = .89). All three
subscales demonstrated good internal consistency in the current
sample (OBQ44-PC ˛ = .86, OBQ44-ICT ˛ = .82, OBQ44-RT ˛ = .87).
The Acceptance and Action Questionnaire (AAQ; Hayes et al.,
2004) is a 9-item scale designed to measure EA. Items are rated
on a 7-point scale with total scores ranging from 9 to 63. Higher
scores on the AAQ indicate greater levels of EA. The AAQ has been
found to possess adequate internal consistency (˛ = .70; Hayes et
al., 2004); however, in the current sample, internal consistency
at admission was lower (˛ = .58). Previous research on the AAQ in
OCD samples has also found lower than adequate internal consis-
tency (Sylvester et al., 2007). Hayes et al. (2004) found that higher
scores on the AAQ significantly correlated with higher levels of gen-
eral psychopathology, depression, anxiety, trauma, various specific
fears (e.g., agoraphobia, blood/injury phobia) and lower quality of
life, and found a test-retest reliability of r = .64 over a four-month
period.
1.2.3. Control variables
Two measures were included to control for general psychi-
atric distress. The Beck Depression Inventory – II (BDI-II; Beck,
Steer, & Brown, 1996) is a 21-item scale that assesses depressive
symptoms. Items are rated on a 4-point scale with higher scores
indicating greater depression (range: 0–63). The BDI-II has been
found to have high one-week test-retest reliability (r = .93; Beck et
al., 1996) and high internal consistency (˛ = .91–.93 among college
students, ˛ = .92 for outpatients; Beck et al., 1996; Dozois, Dobson, &
Ahnbert, 1998; Steer, Kumar, Ranieri, & Beck, 1998). The BDI-II and
the Hamilton Rating Scale for Depression (HRS-D; Hamilton, 1960,
1967) are significantly correlated with r = .71 (Beck et al., 1996).
Internal consistency for the current sample was unavailable, as only
the total score for the BDI-II was included in the database.
The Beck Anxiety Inventory (BAI; Beck, Epstein, Brown, & Steer,
1988) is a self-report measure developed to distinguish cogni-
tive and somatic symptoms of anxiety from cognitive and somatic
symptoms of depression. Participants are instructed to rate how
much they have been bothered by a variety of anxiety symptoms
during the previous month. The BAI has 21 items, and each item is
rated from 0 (“not at all”) to 3 (“severely – it bothered me a lot”;
range of scores = 0–63), with higher scores indicating greater anx-
iety. The BAI has demonstrated high internal consistency (˛s = .92
and .94) and test-retest reliability over a 1-week interval (rs = .67
and .75; Beck et al., 1988; Fydrich, Dowdall, & Chambless, 1992). The
BAI has been shown to be correlated with the modified Hamilton
Rating Scale of Anxiety (r = .51, Beck et al., 1988) and the Trait and
State scales from the State-Trait Anxiety Inventory-Form Y (rs = .58
and .47, respectively; Fydrich et al., 1992). Internal consistency for
the current sample was unavailable, as only the total score for the
BAI was included in the database.
1.3. Analytic plan
Analyses were conducted first with admission data, and second
by examining change from admission to discharge in variables of
interest and how these variables related. In both cases, interactions
between variables of interest and treatment program (residen-
tial or intensive outpatient) were examined in order to determine
whether data from these two programs could be collapsed or
needed to be examined separately.
For the admission data, we started by examining Pearson’s cor-
relations. For significant correlations, multiple regression analyses
were conducted. Depression scores (BDI-II) were entered in Step
1, EA scores (AAQ) were entered in Step 2, and obsessive belief
scores (OBQ44) were entered in Step 3, with the Y-BOCS-SR and
the six OCI-R subscale scores as separate dependent variables. If
the AAQ significantly added to the prediction of the dependent
variable, regression analyses were repeated with Steps 2 and 3
reversed in order to determine if results changed. Pending no sig-
nificant changes, only the first set of regression analyses (with the
AAQ in Step 2 and the OBQ44 in Step 3) was included. In addi-
tion to these analyses, we extended Abramowitz et al.’s (2009)
analyses by completing an additional set of regressions in which
general anxiety (BAI) was entered in Step 1 rather than depres-
sion scores. In order to examine relationships among changes in
variables of interest from admission to discharge, residual gain
scores were calculated, which corrects for problems in using raw
scores (see Steketee & Chambless, 1992). Next, Pearson’s corre-
lations in residual gain scores for all variables of interest were
examined. For variables found to be significantly correlated, mul-
tiple regression analyses were completed in the same manner
as described above, with the difference being that in all cases
residual gain scores were used rather than scores from admis-
sion.
2. Results
2.1. Admission data
All participants had usable data for the OCI-R, AAQ, and BAI. Data
were either missing or unusable for the Y-BOCS-SR (1 participant),
BDI-II (1 participant), and OBQ44 (4 participants for each of the PC
and ICT subscales, 2 participants for the RT subscale).
2.1.1. Differences by treatment program
A preliminary analysis of demographic data and admission data
was conducted to compare participants from the residential pro-
gram with participants from the intensive outpatient program. A
 R.C. Manos et al. / Journal of Anxiety Disorders 24 (2010) 700–708 703

Table 1
Descriptive statistics at admissions and comparison by treatment type.

Variable Overall Residential Partial Statistic (df), p

Demographic variables
Gender (female) 59 (54.6%) 47 (53.4%) 12 (60.0%) Fisher’s Exact, ns
Age 32.10 (12.44) 32.31 (12.55) 31.20 (12.20) t(106) = −.358, ns
Measures of OCD severity
Admit Y-BOCS-SR 26.67 (6.66) 27.46 (6.37) 23.25 (6.96) t(105) = −2.620, .010
Admit OCI-R Total 26.23 (13.41) 27.02 (13.73) 22.75 (11.60) t(106) = −1.290, ns
Admit OCI-R WA 4.58 (4.34) 4.70 (4.40) 4.05 (4.11) t(106) = −.602, ns
Admit OCI-R CH 3.98 (3.32) 4.08 (3.37) 3.55 (3.12) t(106) = −.643, ns
Admit OCI-R OR 4.58 (3.86) 4.67 (3.96) 4.20 (3.47) t(106) = −.490, ns
Admit OCI-R OB 6.91 (3.71) 7.17 (3.67) 5.75 (3.73) t(106) = −1.557, ns
Admit OCI-R HO 3.02 (3.04) 3.15 (3.09) 2.45 (2.84) t(106) = −.993, ns
Admit OCI-R NE 3.16 (3.58) 3.25 (3.62) 2.75 (3.49) t(106) = −.561, ns
Predictor variables
Admit OBQ44-RT 62.94 (18.63) 63.42 (19.43) 60.90 (14.91) t(104) = −.543, ns
Admit OBQ44-PC 64.63 (18.11) 65.13 (18.62) 62.42 (15.90) t(102) = −.587, ns
Admit OBQ44-ICT 47.44 (13.60) 47.53 (14.00) 47.05 (12.07) t(102) = −.141, ns
Admit AAQ 40.55 (5.78) 40.86 (6.09) 39.20 (4.01) t(106) = −1.159, .ns
Control variables
Admit BDI-II 27.75 (11.22) 28.76 (10.98) 23.38 (11.52) t(105) = −1.960, ns
Admit BAI 22.94 (11.94) 23.49 (11.86) 20.50 (12.27) t(106) = −1.011, ns

Note: Y-BOCS-SR, Yale-Brown Obsessive-Compulsive Scale Self Report; OCI-R, Obsessive-Compulsive Inventory – Revised; WA, washing; CH, checking; OR, ordering; OB,
obsessing; NE, neutralizing; OBQ44, Obsessive Beliefs Questionnaire 44; RT, responsibility/threat estimation; PC, perfectionism/certainty; ICT, importance/control of thoughts;
AAQ, Acceptance and Action Questionnaire; BDI-II, Beck Depression Inventory – II; BAI, Beck Anxiety Inventory.

significant difference was found in Y-BOCS-SR, with residential par-
ticipants scoring significantly higher (more severe; see Table 1).
In order to determine whether it was appropriate to collapse res-
idential and intensive outpatient data, treatment program was
dummy coded and separate multiple regression analyses were
performed for all predictor and control variables. With Y-BOCS-
SR severity at admission as the dependent variable, regressions
were completed with the treatment program entered in Step 1,
the predictor or control variable entered in Step 2, and the treat-
ment program × predictor variable interaction entered in Step 3.
The interaction term in all regressions was not significant, indicat-
ing that level of treatment did not moderate correlations between
predictor or control variables and Y-BOCS-SR severity. Therefore,
data across treatment programs were collapsed for all subsequent
analyses examining admissions data.
2.1.2. Pearson’s correlations
Intercorrelations among all of the questionnaires are presented
in Table 2. Of note, there were significant moderate correlations
between the two measures of OCD severity, with correlations
between the Y-BOCS-SR and OCI-R subscales ranging between
r = .207 and .352. In addition, there was a significant moderate cor-
relation between the AAQ and OBQ44-ICT subscale (r = .257), with
correlations between the AAQ and OBQ44-RT and PC subscales not
attaining significance (rs = .188 and .189, respectively). All subscales
of the OBQ44 had significant correlations with the Y-BOCS-SR (r
values ranging between .202 and .278), and all but five correla-
tions between the OBQ44 subscales and the OCI-R subscales were
significant (significant r values ranging between .212 and .416).
Specifically, the OCI-R OR subscale was not correlated with the
RT and PC subscales of the OBQ44 and the OCI-R NE subscale was
not significantly correlated with any of the OBQ44 subscales. The
AAQ did not significantly correlate with the Y-BOCS-SR (r = .159)
or any of the OCI-R subscales (r values ranging between −.051 and
.113). Both control variables were significantly correlated with the
Y-BOCS-SR (r values of .401 and .387 for the BDI-II and BAI). The
BDI-II was also significantly correlated with three of the six OCI-
R subscales (CH, OR, and OB; r values ranging between .172 and
.373); the BAI was significantly correlated five of the OCI-R sub-
scales (all but HO; significant r values ranging between .250 and
.412). Both control variables were significantly correlated with the
AAQ (r values = .343 and .236, for the BDI-II and BAI respectively)
and all subscales of the OBQ (r values ranging between .320 and
.460).
2.1.3. Multiple regression analyses
Results for the regression analyses when the BDI-II was entered
first, the AAQ second, and the OBQ44 subscales third are presented
in the center columns of Table 3. These analyses revealed that
neither the AAQ nor any of the OBQ44 subscales remained sig-
nificant predictors of the Y-BOCS-SR beyond BDI-II scores. With
respect to OCI-R subscales, the analyses revealed that the AAQ
did not predict any subscales beyond the BDI-II. The OBQ44-RT
subscale remained significantly associated with the WA subscale,
the OBQ44-PC subscale remained significantly associated with the
OB subscale, and the OBQ44-ICT subscale remained significantly
associated with the OR and HO subscales. In addition, the OBQ44-
RT subscale was found to be a significant negative predictor of
the HO subscale after first controlling for the BDI-II and the AAQ,
and simultaneously controlling for the OBQ44-ICT and PC sub-
scales.
Tolerance for all regressions ranged from .152 to .849, with num-
bers approaching zero indicating a problem with multicollinearity.
As the reversal of the direction of the relationship between the
OBQ44-RT and the OCI-R HO subscale was unexpected, additional
multiple regression analyses were conducted to better understand
when the reversal occurred. When the BDI-II was entered at Step 1,
the AAQ entered at Step 2, and the OBQ44-RT subscale alone at Step
3, the result was a significant positive relationship for the OBQ44-RT
(ˇ = .230, t = 2.119, p = .037). However, a significant negative rela-
tionship was obtained when the OBQ44-RT subscale was entered
simultaneously with the OBQ44-ICT subscale (ˇ = −.427, t = −2.121,
p = .036) and a non-significant negative relationship was obtained
when the OBQ44-RT subscale was entered simultaneously with the
OBQ44-PC subscale (ˇ = −.148, t = −.719, p = .474).
Results for the regression analyses when the BAI was entered
first instead of the BDI-II (see right-hand columns of Table 3)
followed by the AAQ revealed the AAQ was not a significant
predictor of the YBOCS-SR or any OCI-R subscale. By contrast,
the OBQ44-ICT remained a significant predictor of the Y-BOCS-
SR and the CH, OR, and HO subscales of the OCI-R. In addition,
the OBQ44-RT subscale was found to be a significant nega-
704 R.C. Manos et al. / Journal of Anxiety Disorders 24 (2010) 700–708

Table 2
Correlations at admission.

OCI-R WA OCI-R CH OCI-R OR OCI-R OB OCI-R HO OCI-R NE OBQ 44-RT OBQ 44-PC OBQ 44-ICT AAQ BDI-II BAI
** ** ** ** * * * * ** **
Y-BOCS-SR .352 .329 .258 .268 .207 .243 .220 .202 .278 .159 .401 .387**
OCI-R WA 1.0 .319** .060 .229* −.055 .146 .321** .212* .265** −.051 .153 .250**
OCI-R CH 1.0 .487** .151 .363** .308** .391** .400** .413** .113 .312** .389**
OCI-R OR 1.0 .105 .552** .540** .154 .121 .296** .039 .172* .270**
OCI-R OB 1.0 .011 .266** .375** .416** .304** .034 .373** .412**
OCI-R HO 1.0 .394** .236* .319** .358** .055 .130 .151
OCI-R NE 1.0 .082 .184 .190 −.036 .084 .259**
OBQ44-RT 1.0 .875** .885** .188 .460** .397**
OBQ44-PC 1.0 .840** .189 .431** 356**
OBQ44-ICT 1.0 .257** .423** 320**
AAQ 1.0 .343** .236*
BDI-II 1.0 .588**
BAI 1.0

Y-BOCS-SR, Yale-Brown Obsessive-Compulsive Scale Self Report; OCI-R, Obsessive-Compulsive Inventory – Revised; WA, washing; CH, checking; OR, ordering; OB, obsessing;
NE, neutralizing; OBQ44, Obsessive Beliefs Questionnaire 44; RT, responsibility/threat estimation; PC, perfectionism/certainty; ICT, importance/control of thoughts; AAQ,
Acceptance and Action Questionnaire; BDI-II, Beck Depression Inventory – II; BAI, Beck Anxiety Inventory.
*
p < .05.
**
p < .01.

tive predictor of the HO subscale. Tolerance for all regressions
ranged from .150 to .888. Again, as reversal of the direction
of the relationship between the OBQ44-RT and the OCI-R HO
subscale was unexpected, additional multiple regression analy-
ses were conducted. When the BAI was entered at Step 1, the
AAQ entered at Step 2, and the OBQ44-RT subscale entered alone
at Step 3, the result was a non-significant positive relationship
for the OBQ44-RT (ˇ = .206, t = 1.960, p = .053). However, a sig-
nificant negative relationship was obtained when the OBQ44-RT
subscale was entered simultaneously with the OBQ44-ICT subscale
(ˇ = −.480, t = −2.357, p = .020) and a non-significant negative rela-
tionship was obtained when the OBQ44-RT subscale was entered
simultaneously with the OBQ44-PC subscale (ˇ = −.202, t = −1.012,
p = .315).

Table 3
Summary of the final step of significant regression equations predicting the Y-BOCS-SR and OCI-R subscale scores (BDI-II or BAI entered in Step 1, AAQ in Step 2, and OBQ44
subscales in Step 3).

Control variable = BDI-II Control variable = BAI

R 2
ˇ t p R2 ˇ t p

DV = Y-BOCS-SR Total .208 <.001 .202 .001

Control variable .391 3.615 <.001 .359 3.506 .001
AAQ −.004 −.037 ns .031 .321 ns
OBQ44-RT −.215 −.918 ns −.234 −.992 ns
OBQ44-ICT .394 1.889 ns .471 2.243 .027
OBQ44-PC −.101 −.502 ns −.122 −.603 ns
DV = OCI-R WA .131 .018 .170 .003
Control variable .065 .577 ns .214 2.063 .042
AAQ −.129 −1.243 ns −.152 −1.540 ns
OBQ44-RT .505 2.061 .042 .434 1.823 ns
OBQ44-ICT .092 .421 ns .144 .674 ns
OBQ44-PC −.322 −1.524 ns −.347 −1.728 ns
DV = OCI-R CH .212 <.001 .265 <.001
Control variable .136 1.260 ns .314 3.224 .002
AAQ −.047 −.480 ns −.070 −.756 ns
OBQ44-RT .039 .168 ns −.110 −.492 ns
OBQ44-ICT .370 1.778 ns .405 2.016 .047
OBQ44-PC −.008 −.042 ns .043 .228 ns
DV = OCI-R OR .089 .015 .119 .013
Control variable .087 .794 ns .186 1.853 ns
AAQ −.060 −.581 ns −.068 −.689 ns
OBQ44-ICT .264 2.470 <.015 .254 2.517 .013
DV = OCI-R OB .288 <.001 .272 <.001
Control variable .330 3.222 .002 .292 3.006 .003
AAQ −.109 −1.155 ns −.077 −.836 Ns
OBQ44-RT .052 .235 ns .090 .405 Ns
OBQ44-ICT −.307 −1.555 ns −.211 −1.055 Ns
OBQ44-PC .535 2.794 .006 .441 2.347 .021
DV = OCI-R HO .189 .001 .196 .001
Control variable .028 .259 ns .114 1.116 Ns
AAQ −.070 −.701 ns −.084 −.860 Ns
OBQ44-RT −.556 −2.348 .021 −.634 −2.701 .008
OBQ44-ICT .672 3.187 .002 .675 3.212 .002
OBQ44-PC .234 1.145 ns .274 1.385 Ns

Note: ˇ, standardized regression coefficient; Y-BOCS-SR, Yale-Brown Obsessive-Compulsive Scale Self Report; OCI-R, Obsessive-Compulsive Inventory – Revised; WA, washing;
CH, checking; OR, ordering; OB, obsessing; NE, neutralizing; OBQ44, Obsessive Beliefs Questionnaire 44; RT, responsibility/threat estimation; PC, perfectionism/certainty;
ICT, importance/control of thoughts; AAQ, Acceptance and Action Questionnaire; BDI-II, Beck Depression Inventory – II; BAI, Beck Anxiety Inventory.
 R.C. Manos et al. / Journal of Anxiety Disorders 24 (2010) 700–708 705

Table 4
Variables of interest from admissions to discharge.

Measure Admissions Discharge Change (full Test of significance Change Change Test of significance
sample) (admissions to discharge) (residential) (partial) (residential compared to
partial)
M (SD) M (SD) M (SD) t(df), p M (SD) M (SD) t(df), p

Y-BOCS-SR 26.67 (6.66) 15.24 (7.03) 11.43 (7.30) 16.195 (106), <.001 11.66 (6.92) 10.45 (8.90) −.664 (105), ns
OCI-R WA 4.58 (4.34) 2.04 (2.82) 2.54 (3.38) 7.814 (107), <.001 2.47 (3.29) 2.85 (3.84) .450 (106), ns
OCI-R CH 3.98 (3.32) 2.17 (2.22) 1.81 (2.64) 7.151 (107), <.001 1.72 (2.63) 2.25 (2.69) .816 (106), ns
OCI-R OR 4.58 (3.86) 2.79 (3.15) 1.80 (2.74) 6.804 (107), <.001 1.78 (2.85) 1.85 (2.25) .097 (106), ns
OCI-R OB 6.91 (3.71) 4.38 (3.15) 2.53 (2.84) 9.250 (107), <.001 2.39 (2.85) 3.15 (2.80) 1.086 (106), ns
OCI-R HO 3.02 (3.04) 1.85 (2.46) 1.17 (2.31) 5.278 (107), <.001 1.20 (2.34) 1.05 (2.19) −.259 (106), ns
OCI-R NE 3.16 (3.58) 1.45 (2.40) 1.70 (2.82) 6.287 (107), <.001 1.63 (2.77) 2.05 (3.07) .607 (106), ns
OBQ44-RT 62.94 (18.63) 48.11 (18.24) 14.81 (16.13) 9.457 (105), <.001 14.15 (15.57) 17.65 (18.51) .873 (104), ns
OBQ44-PC 64.63 (18.11) 50.50 (18.36) 14.06 (16.39) 8.749 (103), <.001 13.55 (15.94) 16.32 (18.55) .663(102), ns
OBQ44-ICT 47.44 (13.60) 36.69 (13.77) 10.61 (10.82) 10.001 (103), <.001 10.26 (10.52) 12.10 (12.17) .683 (102), ns
AAQ 40.55 (5.78) 37.60 (5.36) 2.95 (6.03) 5.080 (107), <.001 3.09 (6.04) 2.35 (6.11) −.490 (106), ns
BDI 27.75 (11.22) 11.33 (10.38) 16.45 (12.01) 14.166 (106), <.001 17.16 (11.57) 13.38 (13.67) −1.274 (105), ns
BAI 22.94 (11.94) 13.53 (9.99) 9.40 (11.28) 8.060 (107), <.001 9.38 (10.89) 9.50 (13.21) .042 (106), ns

Note: Y-BOCS-SR, Yale-Brown Obsessive-Compulsive Scale Self Report; OCI-R, Obsessive-Compulsive Inventory – Revised; WA: washing; CH: checking; OR: ordering;
OB: obsessing; NE: neutralizing; OBQ44: Obsessive Beliefs Questionnaire 44; RT: responsibility/threat estimation; PC: perfectionism/certainty; ICT: importance/control
of thoughts; AAQ: Acceptance and Action Questionnaire; BDI-II: Beck Depression Inventory – II; BAI: Beck Anxiety Inventory.

2.2. Change over the course of treatment
2.2.1. Descriptives
See Table 4 for descriptive data on variables of interest as well as
results of t-tests for correlated samples for the change from admis-
sion to discharge. All variables examined were significantly lower
(i.e., less severe) at discharge than at admission. Furthermore, t-
tests for independent samples comparing admissions to discharge
change scores across the two treatment programs found no signif-
icant differences in the amount of improvement for participants in
the residential program compared to intensive outpatient.
2.2.2. Differences by treatment program
To examine the hypothesized relationship between changes in
predictor variables (i.e., EA, obsessive beliefs) and changes in OC
severity, we calculated residualized gain scores for the Y-BOCS-
SR and AAQ total scores and the OCI-R and OBQ44 subscales in
order to account for shortcomings of using raw change scores (see
Steketee & Chambless, 1992). In order to determine whether it was
again appropriate to collapse data from the residential and inten-
sive outpatient programs, treatment program was dummy coded
and multiple regressions were performed for all predictor change
variables (BDI-II, BAI, AAQ, and OBQ44 subscales), wherein the
treatment program was entered in Step 1, the residual gain score
of the predictor variable was entered in Step 2, and the treatment
program by predictor change variable (residual gain score) inter-
action was entered in Step 3. For all cases, the interaction term was
not significant, indicating that treatment level did not moderate
correlations between predictor and outcome variables. Therefore,
data were collapsed across treatment programs for all subsequent
analyses.
2.2.3. The relationship of change in predictor variables and
outcome measures
Pearson correlations were calculated between the residualized
gain scores for all variables of interest (see Table 5). Changes in
obsessive beliefs, as measured by the OBQ44 subscales, were sig-
nificantly correlated with changes in OC symptoms, as measured by
the Y-BOCS-SR and OCI-R subscales (r values ranging between .380
and .609, median r value = .480, all p values < .001). Change in EA
was significantly correlated with change in one facet of OC sever-
ity, as indicated by OCI-R OR subscale (r = .227, p = .018). Change in
EA was not significantly correlated with change in OC severity on
the Y-BOCS-SR or any other OCI-R subscale.
Admissions to discharge residualized gain scores were used in
multiple regression analyses in which changes in Y-BOCS-SR and
OCI-R subscale scores were dependent variables, changes in BDI
or BAI were entered in Step 1, changes in EA were entered in
Step 2, and changes in obsessive beliefs (OBQ44-RT, OBQ44-ICT,
and OBQ44-PC) were entered in Step 3 (see Table 6). As the AAQ
was never significant when entered in Step 2, additional analyses
with the OBQ subscales entered at Step 2 and the AAQ entered at
Step 3 were not conducted. All seven regressions were significant.
Change on the OBQ44-PC was a significant predictor of change on
the Y-BOCS-SR and OCI-R WA subscale when the BDI-II was entered
at Step 1, but not when the BAI was entered at Step 1, whereas
change on the OBQ44-PC was a significant predictor of OCI-R NE
when the BAI was entered at Step 1 but not when the BDI-II was

Table 5
Pearson correlations between admissions to discharge standardized residual predictor and outcome variables.

Predictor variables Other variables

Y-BOCS-SR OCI-R WA OCI-R CH OCI-R OR OCI-R OB OCI-R HO OCI-R NE

** ** ** ** ** **
OBQ44-RT .609 .396 .437 .537 .589 .452 .480**
OBQ44-PC .608** .444** .429** .522** .585** .463** .510**
OBQ44-ICT .499** .380** .381** .478** .550** .428** .480**
AAQ .092 .109 .085 .227* .100 .123 .044
BDI-II .659** .384** .411** .354** .493** .465** .421**
BAI .589** .379** .395** .351** .495** .403** .354**

Note: Y-BOCS-SR: Yale-Brown Obsessive-Compulsive Scale Self Report; OCI-R: Obsessive-Compulsive Inventory – Revised; WA: washing; CH: checking; OR: ordering;
OB: obsessing; NE: neutralizing; OBQ44: Obsessive Beliefs Questionnaire 44; RT: responsibility/threat estimation; PC: perfectionism/certainty; ICT: importance/control of
thoughts; AAQ: Acceptance and Action Questionnaire.
*
p < .05.
**
p < .001.
706 R.C. Manos et al. / Journal of Anxiety Disorders 24 (2010) 700–708

Table 6
Summary of the final step of significant regression equations predicting change in Y-BOCS-SR and OCI-R subscale scores from admission to discharge standardized residual
gain scores (BDI-II or BAI entered in Step 1, AAQ in Step 2, and OBQ44 in Step 3).

Control variable = BDI-II Control variable = BAI

R2 ˇ t p R2 ˇ t p

DV = Y-BOCS-SR Total .561 <.001 .489 <.001

Control variable .490 5.731 <.001 .364 3.847 <.001
AAQ −.009 −.129 ns −.087 −1.130 ns
OBQ44-RT .265 1.755 ns .271 1.646 ns
OBQ44-ICT −.218 −1.610 ns −.076 −.517 ns
OBQ44-PC .293 2.050 .043 .243 1.544 ns
DV = OCI-R WA .251 <.001 .216 <.001
Control variable .239 2.139 .035 .197 1.706 ns
AAQ −.005 −.055 ns −.028 −.297 ns
OBQ44-RT −.114 −.577 ns −.073 −.360 ns
OBQ44-ICT −.007 −.042 ns .041 .224 ns
OBQ44-PC .435 2.335 .022 .359 1.849 ns
DV = OCI-R CH .262 <.001 .238 <.001
Control variable .224 2.022 .046 .174 1.533 ns
AAQ −.042 −.451 ns −.068 −.730 ns
OBQ44-RT .219 1.121 ns .250 1.253 ns
OBQ44-ICT −.112 −.636 ns −.063 −.348 ns
OBQ44-PC .250 1.353 ns .195 1.016 ns
DV = OCI-R OR .287 <.001 .295 <.001
Control variable .065 .595 ns .030 .270 ns
AAQ .082 .884 ns .064 .714 ns
OBQ44-RT .278 1.449 ns .264 1.375 ns
OBQ44-ICT −.013 −.077 ns .014 .080 ns
OBQ44-PC .219 1.203 ns .246 1.336 ns
DV = OCI-R OB .424 <.001 .406 <.001
Control variable .246 2.511 .014 .252 2.510 .014
AAQ −.028 −.334 ns −.055 −.675 ns
OBQ44-RT .137 .795 ns .149 .843 ns
OBQ44-ICT .104 .667 ns .171 1.072 ns
OBQ44-PC .266 1.627 ns .177 1.047 ns
DV = OCI-R HO .292 <001 .249 <.001
Control variable .284 2.611 .010 .179 1.582 ns
AAQ .001 .007 ns −.035 −.378 ns
OBQ44-RT .071 .373 ns .117 .590 ns
OBQ44-ICT .017 .096 ns .079 .439 ns
OBQ44-PC .248 1.365 ns .205 1.080 ns
DV = OCI-R NE .280 <.001 .301 <.001
Control variable .171 1.564 ns .133 1.222 ns
AAQ −.096 −1.035 ns −.137 −1.542 ns
OBQ44-RT .030 .156 ns −.030 −.158 ns
OBQ44-ICT .068 .391 ns .149 .861 ns
OBQ44-PC .343 1.878 ns .380 2.073 .041

Note: ˇ, standardized regression coefficient; Y-BOCS-SR, Yale-Brown Obsessive-Compulsive Scale Self Report; OCI-R, Obsessive-Compulsive Inventory – Revised; WA, washing;
CH, checking; OR, ordering; OB, obsessing; NE, neutralizing; OBQ44, Obsessive Beliefs Questionnaire 44; RT, responsibility/threat estimation; PC, perfectionism/certainty;
ICT, importance/control of thoughts; AAQ, Acceptance and Action Questionnaire; BDI-II, Beck Depression Inventory – II; BAI, Beck Anxiety Inventory.

entered at Step 1. The OBQ44-RT and OBQ44-ICT subscales did not
significantly predict any of the OCI-R subscales or the Y-BOCS-SR
regardless of whether the BDI-II or BAI was entered at Step 1. Tol-
erance for all regressions ranged from .198 to .917.
3. Discussion
3.1. Admission data
Our results indicated that EA had limited association with mea-
sures of OCD severity and did not add significantly to prediction
of OC symptom dimensions above and beyond obsessive beliefs
and depression. By contrast, we found that moderate associations
between a range of obsessive beliefs and OCD severity, and cer-
tain obsessive beliefs continued to predict specific OCD domains
above and beyond EA and depression. These findings broadly repli-
cate and extend findings of Abramowitz et al. (2009). Interestingly,
the specific types of obsessive beliefs that predicted specific symp-
tom dimensions differed from Abramowitz et al.’s (2009) findings.
For example, Abramowitz et al. found that, after controlling for
depression and EA, importance/control of thoughts predicted OB
and perfectionism/certainty predicted CH. By contrast, compara-
ble analyses in our study found that when first controlling for
depression and EA, perfectionism/certainty predicted OB and nei-
ther importance/control of thoughts nor perfectionism/certainty
predicted CH. We also found several associations between obses-
sive beliefs and OCD domains after first controlling for depression
and EA that were not found in the Abramowitz study. Specifically,
importance/control of thoughts predicted OR and HO and respon-
sibility/threat estimation predicted WA, OR, and HO. Notably in
the case of HO, although the zero-order relationship was positive,
the relationship became negative in multiple regression analyses in
which depression or anxiety, EA, and at least one additional OBQ44
subscale were included. Our results also differ from findings by
Myers, Fisher, and Wells (2008), who reported that after controlling
for worry, responsibility/threat estimation significantly predicted
all OCD domains on the original OCI (Foa et al., 1998), perfection-
ism/certainty significantly predicted WA, Doubting, and OR, and
importance/control of thoughts significantly predicted CH, OB, and
NE in a student sample.
It is not clear why these differences exist, although it may be
that relationships among these variables differ for severe, clini-
 R.C. Manos et al. / Journal of Anxiety Disorders 24 (2010) 700–708
cal samples, as in the current study, compared to student samples
as in Abramowitz et al. (2009) and Myers et al. (2008), although
this would not account for differences between the Abramowitz
and Myers studies. Further research is needed to examine why the
pattern, and not just strength, of these relationships differs among
these investigations.
We also investigated these relationships when controlling for
general anxiety. For the most part, the pattern remained the same
as when we controlled for depression with respect to which obses-
sive belief domains predicted OC severity. The three exceptions
were that importance/control of thoughts significantly predicted
OC severity on the Y-BOCS-SR and OCI-R CH, and that perfection-
ism/certainty did not significantly predict OCI-R NE.
3.2. Change data
We also investigated whether changes in obsessive beliefs and
EA were associated with changes in OC severity from admission to
discharge. Change in all three categories of obsessive beliefs was
associated with change in OCD severity on the Y-BOCS-SR and all
OCI-R subscales. Change in EA was associated with change on only
the OCI-R OR subscale. However, our multiple regression analy-
ses revealed that change in EA did not continue to predict OCI-R OR
scores above changes in anxiety or depression, whereas perfection-
ism/certainty remained a significant predictor of total Y-BOCS-SR
scores and OCI-R WA when the BDI-II was the control variable
and perfectionism/uncertainty predicted OCI-R NE scores when
the BAI was the control variable. None of the other OBQ44 sub-
scales (i.e., importance/control of thoughts or responsibility/threat
estimation) significantly predicted changes in OCD severity.
Overall, our pattern of results supports the notion that EA, at
least as measured by the AAQ, may not play a significant role in
OC severity in a severe, clinical sample. Alternatively, it may be the
case that changes in OC severity over treatment were not reflected
by changes in EA because treatment did not directly target EA.
The treatment used, ERP plus thought challenging and other cogni-
tive techniques, is thought to produce change through mechanisms
that may be distinct from EA (e.g., habituation to anxiety, cognitive
change). From the ACT perspective, cognitive change strategies may
be viewed as further attempts to control and modify OC symptoms,
which conflicts with the notion of accepting thoughts as thoughts
(i.e., as part of “cognitive defusion”) while behaving in accor-
dance with one’s values. Had treatment been focused on processes
emphasized in ACT, such as defusion, acceptance, mindfulness, self
as context, contact with the present moment, and values-directed
behavioral commitments, then perhaps changes in EA may have
had a stronger relationship with changes in OC symptoms.
In contrast to acceptance theory, cognitive theory would predict
strong associations between specific obsessive beliefs and specific
OC symptom domains and contend that treatment brings about
symptom change through changes in beliefs. Accordingly, changes
in obsessive beliefs should be strongly correlated with changes
in OC symptoms. Broadly speaking, our results provide modest
support for the tenants of cognitive theory, in that: (1) obses-
sive beliefs were correlated with OC symptoms at admission, (2)
there was some evidence for specificity between specific belief
domains and specific OC symptoms, and (3) change in obsessive
beliefs was in some cases predictive of changes in OC symptoms.
We noted in our introduction that, at least in some cases, treat-
ment with ACT may result in interventions that appear to overlap
with ERP procedures. To the extent that ACT is efficacious in the
treatment of OCD, perhaps similar processes are occurring as those
hypothesized by cognitive theory. In other words, rather than pro-
moting acceptance and values-based action, interventions based
on ACT principles may actually promote symptom change through
habituation or cognitive changes. Further research directly com-
707
paring mechanisms of change in CBT and ACT is needed to examine
whether change in obsessive beliefs will continue to have a sig-
nificant impact on change in OC symptoms in other treatment
modalities such as ACT, or if other treatment-specific mechanisms
(e.g., change in EA) will play a significant role.
It is also important to note that measures of depression (BDI-
II) and general anxiety (BAI) performed as well, and in some cases
better, than the AAQ and OBQ44 in several analyses. In the exam-
ination of admission scores, the BDI-II was significantly correlated
with half of the OCI-R subscales, and the BAI was significantly cor-
related with almost all OCI-R subscales. Changes in BDI-II and BAI
scores were significantly correlated with changes in the Y-BOCS-SR
and changes in all OCI-R subscales. One reason that these mea-
sures may have “outperformed” the AAQ and OBQ44 scores may
be that depression and general anxiety are more relevant to OCD
symptoms and severity than the theoretical constructs of obses-
sive beliefs and EA. On the other hand, it might be that obsessive
beliefs and/or EA are quite relevant but that their correspond-
ing measures are inadequate assessment tools. Indeed, the AAQ
has been criticized for measurement problems, such as low inter-
nal consistency, which was a concern in the current study. It has
also been suggested that the AAQ may be too general to function
as an appropriate measure of disorder-specific EA (Hayes et al.,
2004). To date, numerous versions of the AAQ have been devel-
oped targeting specific conditions, such as smoking (Gifford et al.,
2002) and chronic pain (Westin, Hayes, & Andersson, 2008). To
help clarify the relationship between OCD and EA, further research
is needed using newer versions of the AAQ with better psycho-
metric properties, such as the AAQ-II (Bond et al., submitted for
publication), or to develop a more specific measure of EA in OCD.
For the latter suggestion, the MetaCognitions Questionnaire, which
assesses processes from a theory similar to ACT and also useful for
treating OCD (i.e., Metacognitive Therapy; cf. Fisher, 2009), may
serve as an example of how to design a measurement tool that
matches both the underlying theory and the specific aspects of the
disorder in question (cf. Solem, Haland, Vogel, Hansen, & Wells,
2009).
An important limitation of our results is that, although they are
broadly consistent with the cognitive theory of OCD, all of our mod-
els accounted for less than 50% of the variance in outcome, and most
accounted for substantially less. It is not clear to what extent this
reflects limitations in cognitive theory or limitations in the OBQ to
measure the hypothesized beliefs underlying OCD. Thus research
to further refine the OBQ or to test alternative models may help to
resolve this question.
References
Abramowitz, J. S. (1997). Effectiveness of psychological and pharmacological treat-
ments for obsessive – compulsive disorder: a quantitative review. Journal of
Consulting and Clinical Psychology, 65, 44–52.
Abramowitz, J. S., Lackey, G. R., & Wheaton, M. G. (2009). Obsessive-compulsive
symptoms: the contribution of obsessional beliefs and experiential avoidance.
Journal of Anxiety Disorders, 23, 160–166.
American Psychiatric Association. (2000). Diagnostic and statistical manual of men-
tal disorders IV-TR (4th ed.). Washington, DC: Author.
Baer, L., Brown-Beasley, M., Sorce, J., & Henriques, A. (1993). Computer-assisted
telephone administration of a structured interview for obsessive-compulsive
disorder. American Journal of Psychiatry, 150, 1737–1738.
Beck, A. T., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for measuring
anxiety: psychometric properties. Journal of Consulting and Clinical Psychology,
56, 893–897.
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the BDI-II. San Antonio, TX:
Psychological Corporation.
Björgvinsson, T., & Hart, J. (2007). Obsessive-compulsive disorder. In: M. Hersen, &
J. Rosqvist (Eds.), Handbook of assessment, conceptualization, and treatment. New
York: John Wiley & Sons.
Bond, F. W., Hayes, S. C., Baer, R. A., Carpenter, K. M., Orcutt, H. K., Waltz, T., et al. (sub-
mitted for publication). Preliminary psychometric properties of the Acceptance
and Action Questionnaire – II: a revised measure of psychological flexibility and
acceptance.
708 R.C. Manos et al. / Journal of Anxiety Disorders 24 (2010) 700–708
Chawla, N., & Ostafin, B. (2007). Experiential avoidance as a functional dimensional
approach to psychopathology: an empirical review. Journal of Clinical Psychology,
63, 871–890.
Dozois, D. J., Dobson, K. S., & Ahnbert, J. L. (1998). A psychometric evaluation of the
Beck Depression Inventory-II. Psychological Assessment, 10(10), 83–89.
Eifert, G. H., & Forsyth, J. P. (2005). Acceptance and commitment therapy for anxiety
disorders: a practitioner’s guide to using mindfulness, acceptance, and values-based
behavior change strategies. Oakland: New Harbinger.
Fisher, P. L. (2009). Obsessive compulsive disorder: a comparison of the CBT and
metacognitive approach. International Journal of Cognitive Therapy, 2, 107–122.
Foa, E. B., Huppert, J. D., Leiberg, S., Langner, R., Kichic, R., Hajcak, G., et al. (2002).
The Obsessive-Compulsive Inventory: development and validation of a short
version. Psychological Assessment, 14, 485–496.
Foa, E. B., Kozak, M. J., Salkovskis, P. M., Coles, M. E., & Amir, N. (1998). The valida-
tion of a new obsessive-compulsive disorder scale: the obsessive-compulsive
inventory. Psychological Assessment, 10, 206–214.
Fydrich, T., Dowdall, D., & Chambless, D. L. (1992). Reliability and validity of the Beck
Anxiety Inventory. Journal of Anxiety Disorders, 6, 55–61.
Gifford, E. V., Antonuccio, D. O, Kohlenberg, B. S., Hayes, S. C., & Piasecki, M. M. (2002).
Combining Bupropion SR with acceptance and commitment-based behavioral ther-
apy for smoking cessation: preliminary results from a randomized controlled trial.
Paper presented at the annual meeting of the Association for Advancement of
Behavioral Therapy, Reno, NV.
Goodman, W. K., Price, L. H., Rasmussen, S. A., Mazure, C., Fleischmann, R. L., Hill, C.
L., et al. (1989). The Yale-Brown Obsessive Compulsive Scale. Archives of General
Psychiatry, 46, 1006–1011.
Hamilton, M. (1960). A rating scale for depression. Journal of Neurology, Neurosurgery
and Psychiatry, 23, 56–62.
Hamilton, M. (1967). Development of a rating scale for primary depressive illness.
British Journal of Social and Clinical Psychology, 6, 278–296.
Hayes, S. C., Strosahl, K. D., & Wilson, K. G. (1999). Acceptance and commitment
therapy: an experiential approach to behavior change. New York: Guilford Press.
Hayes, S. C., Strosahl, K. D., Wilson, K. G., Bissett, R. T., Pistorello, J., Toarmino, D.,
et al. (2004). Measuring experiential avoidance: a preliminary test of a working
model. The Psychological Record, 54, 553–578.
Hayes, S. C., Wilson, K. W., Gifford, E. V., Follette, V. M., & Strosahl, K. (1996). Expe-
riential avoidance and behavioral disorders: a functional dimensional approach
to diagnosis and treatment. Journal of Consulting and Clinical Psychology, 64,
1152–1168.
Myers, S. G., Fisher, P. L., & Wells, A. (2008). Belief domains of the Obsessive Beliefs
Questionnaire-44 (OBQ-44) and their specific relationship with obsessive-
compulsive symptoms. Journal of Anxiety Disorders, 22, 475–484.
Obsessive Compulsive Cognitions Working Group. (2005). Psychometric validation
of the Obsessive Belief Questionnaire and Interpretation of Intrusions Inventory.
Part 2. Factor analyses and testing of a brief version. Behaviour Research and
Therapy, 43, 1527–1543.
Rachman, S. (1997). A cognitive therapy of obsessions. Behaviour Research and Ther-
apy, 35, 793–802.
Solem, S., Haland, A. T., Vogel, P. A., Hansen, B., & Wells, A. (2009). Change in metacog-
nitions predicts outcome in obsessive-compulsive disorder patients undergoing
treatment with exposure and response prevention. Behavior Research and Ther-
apy, 47, 301–307.
Steer, R. A., Kumar, G., Ranieri, W. F., & Beck, A. T. (1998). Use of the Beck Depression
Inventory-II with adolescent psychiaric outpatients. Journal of Psychopathology
and Behavioral Assessment, 20, 127–137.
Steketee, G., & Chambless, D. L. (1992). Methodological issues in prediction of treat-
ment outcome. Clinical Psychology Review, 12, 387–400.
Steketee, G., Frost, R., & Bogart, K. (1996). The Yale-Brown obsessive-compulsive
scale: interview versus self-report. Behavior Research and Therapy, 34, 675–684.
Sylvester, B. D., Entricht, T. L., Wetterneck, C. T., Hart, J. M., & Björgvinsson, T. (2007).
The relationships among acceptance, anxiety sensitivity, and depression in severe
OCD? Poster presented at the 41st Annual Convention of the Association for
Behavioral and Cognitive Therapies, Philadelphia, PA.
Tolin, D. F., Worhunsky, P., & Maltby, N. (2006). Are ‘obsessive’ beliefs specific to
OCD?: a comparison across anxiety disorders. Behaviour Research and Therapy,
44, 469–480.
Twohig, M. P., Hayes, S. C., & Masuda, A. (2006). Increase willingness to experience
obsessions: acceptance and commitment therapy as a treatment for obsessive
– compulsive disorder. Behavior Therapy, 37, 3–13.
Westin, V., Hayes, S. C., & Andersson, G. (2008). Is it the sound or your relationship
to it? The role of acceptance in predicting tinnitus impact. Behaviour Research
and Therapy, 46, 1259–1265.
Whittal, M. L., Robichaud, M., Thordarson, D. S., & McLean, P. D. (2008). Group and
individual treatment of obsessive-compulsive disorder using cognitive therapy
and exposure plus response prevention: a 2-year follow-up of two randomized
trials. Journal of Consulting and Clinical Psychology, 76, 1003–1014.