Reliability and validity of the Child Behavior
Checklist Obsessive-Compulsive Scale
Eric A. Storcha,b,*, Tanya K. Murphyb,
Daniel M. Bagnerc, Natalie B. Johnsb,
Audrey L. Baumeisterb, Wayne K. Goodmanb,
Gary R. Geffkenb
aDepartment of Pediatrics, University of Florida, Gainesville, FL 32610, USA
bDepartment of Psychiatry, Box 100234, University of Florida, Gainesville, FL 32610, USA
cDepartment of Clinical and Health Psychology, University of Florida, Gainesville, FL 32610, USA
Received 7 March 2005; received in revised form 6 June 2005; accepted 16 June 2005
This study examined the psychometric properties of the Obsessive-Compulsive Scale
(OCS) of the Child Behavior Checklist (CBCL). Participants included 48 youth with
obsessive-compulsive disorder (OCD), 41 with a non-OCD internalizing disorder, and 101
with an externalizing disorder. Confirmatory factor analysis of the 8-item OCS did not
result in an adequate fit. Exploratory factor analysis identified a 1-factor model consisting
of 6 items. Adequate internal consistency for the revised OCS (OCS-R) was obtained, and
convergent validity was supported by moderate relationships with other OCD indices. The
OCS-R had stronger associations with measures of OCD symptoms than with measures of
depression and externalizing behaviors. Youth with OCD had significantly higher OCS-R
scores than those with internalizing and externalizing disorders. Suggestions for cutoff
scores are provided using results from ROC analyses. Overall, these findings suggest that
the OCS-R is a reliable and valid instrument for the assessment of pediatric OCD.
# 2005 Elsevier Inc. All rights reserved.
Keywords: Obsessive-compulsive disorder; Children; Validity; Obsessive-Compulsive Scale –
Revised; Child Behavior Checklist
20 (2006) 473–485
* Corresponding author. Tel.: +1 352 392 3611; fax: +1 352 846 1455.
E-mail address: firstname.lastname@example.org (E.A. Storch).
0887-6185/$ – see front matter # 2005 Elsevier Inc. All rights reserved.
Previously thought rare, recent research has identified pediatric obsessive-
compulsive disorder (OCD) as one of the most common childhood psychiatric
illnesses with a point-prevalence rate between 2 and 4% (Douglass, Moffitt, Dar,
McGee, & Silva, 1995; Maina, Albert, Bogetto, & Ravizza, 1999; Rapoport &
Inoff-Germain, 2000). Symptoms frequently begin in childhood (DeVeaugh-
Geiss et al., 1992), with insidious onset, and pursue a protracted yet fluctuating
course (Murphy et al., 2004). Not surprisingly, pediatric OCD is related to
significant functional impairment within academic, family, and social domains
largely due to distress and frequency of ritual engagement (Piacentini, Bergman,
Keller, & McCracken, 2003).
Advances in psychological and pharmacological interventions strongly
suggest that early detection and treatment can improve prognosis (Leonard
et al., 1993; Pediatric OCD Treatment Study Team, 2004). Yet, the vast majority
of children with OCD do not receive appropriate, complete intervention (Heyman
et al., 2001). One likely explanation for this finding is the lack of standardized
assessment instruments appropriate for large-scale screenings that sufficiently
capture the broad phenomenology of symptoms. Only two self-report instruments
appear suitable for such screenings, namely the Leyton Obsessional Inventory-
the Children’s Florida Obsessive-Compulsive Inventory (C-FOCI; Storch et al.,
is the inability of each to capture the broad range of symptoms often present.
Rather, the instruments focus on presence and severity of specific symptoms with
relatively high base rates (e.g., checking rituals, germ obsessions), thereby
neglecting less frequently occurring obsessions and/or compulsions (e.g.,
ritualized eating, sexual obsessions, horrific images).
The Child Behavior Checklist (CBCL; Achenbach, 1991) is a widely used
parent-report questionnaire designed to assess the behavioral problems and
social competencies of children 4–18 years of age. Recently, Nelson et al.
(2001) developed a factorally derived 8-item Obsessive-Compulsive Scale
(OCS) imbedded within the CBCL. Analyses for this study were conducted in a
sample of 73 youth with OCD, 73 with a non-OCD psychiatric disorder, and 73
healthy controls. The OCS had good internal consistency (a = .84) and
discriminated between youth with OCD and a mixed psychiatric diagnostic
group and non-psychiatric controls. Using cutoff scores at the 60th and 70th
percentiles, sensitivity was 75–85% and specificity was 82–93%. Other
measures of OCD (e.g., Children’s Yale-Brown Obsessive-Compulsive Scale
[CY-BOCS]) or impairment were not included to examine convergent or
divergent validity. A second relevant study was conducted to determine genetic,
age, gender, and environmental contributions to OCS scores (Hudziak et al.,
2004). Participants were a large twin sample taken from the Netherlands Twin
Registry and the Missouri Twin Study. Findings suggested that OCS scores
were highly heritable and influenced by genetic and unique environmental
factors in younger children. Genetic and environmental influences were
E.A. Storch et al./Anxiety Disorders 20 (2006) 473–485474
consistent across gender and age within the younger group, but were less for
Overall, a major strength of the OCS over existing OCD measures is that the
and characteristics that are not specific to a single symptom (e.g., ‘‘strange
ideas’’). A second strength is that the use of parent-ratings addresses the issue of
underreporting that is frequently characteristic in ego-syntonic OCD (Merlo,
Storch, Murphy, Goodman, & Geffken, in press). Additional advantages of the
OCS that are specific to the CBCL include the ease of administration and scoring,
translation into 43 languages, existence of parallel teacher and self-report
measures, and utility for assessing other psychiatric symptomatology. Despite
positive findings of Nelson et al. (2001) and Hudziak et al. (2004), however, there
are several compelling reasons to further evaluate the OCS. First, the OCS item
content was based on results from a principal factor analysis where factors with
eigenvalues greater than one were extracted. Use of this criterion alone to extract
factorsis susceptible toretainingtoo manyfactors (Velicer,Eaton,&Fava, 2000).
In contrast, utilizing algorithms (e.g., Glorfeld’s extension [Glorfeld, 1995] and
minimum average partials (MAP) method [Velicer, 1976]) has been emphasized
as the most accurate method of recovering the true number of factors (O’Connor,
2000). Thus, results must be replicated in an independent sample to ensure factor
reliability. Second, estimates of convergent and divergent validity of the OCS
were not obtained which limits understanding of the OCS clinical utility. Finally,
discriminant validitywas examinedthrough comparisons with a mixeddiagnostic
group and non-psychiatric control group. Examining discriminant validity with
such samples limits the ability to determine the extent to which instruments are
assessing specific or common features across child psychopathology (Schniering
between clinical disorders with similar nosological composition, namely other
Given these issues, the specific aims of this study were as follows: (1) to re-
of children and adolescents with OCD; (2) to examine further psychometric
properties of the OCS, including internal consistency, convergent and divergent
validity, and discriminant validity; and (3) to provide information about the
specificity and sensitivity of the OCS.
Participants were obtained from two sources: (a) consecutive patients with
OCD at an outpatient psychiatry clinic diagnosed with OCD by a board certified
child psychiatrist with 10 years experience in OCD and related disorders (TKM),
E.A. Storch et al./Anxiety Disorders 20 (2006) 473–485475
and (b) children and adolescents with a non-OCD psychiatric condition seen for
outpatient psychodiagnostic testing by a licensed clinical psychologist with 16
seen between March and November 2004. Psychodiagnostic assessments were
conducted between January 1998 and November 2004. In both instances,
diagnoses were based on a detailed clinical interview and all available clinical
information. In the Psychiatry clinic, diagnoses were confirmed by one of two
licensed clinical psychologists with extensive experience in pediatric OCD.
Written consent/assent to participate, approved by the University of Florida
Institutional Review Board (IRB), were obtained for patients attending the
psychiatry clinic. Approval from the University of Florida IRB was obtained to
conduct an archival records review of children and adolescents seen for
The final sample consisted of 190 children and adolescents (55 female and 135
male). Overall, the mean age of the samplewas 10.5 years (S.D. = 3.3, range = 4–
18 years). The ethnic composition was: 86% Caucasian, 8% African American,
2% Latin American, and 4% parent-identified as ‘‘Other.’’ No significant
differences were found in gender, age, or ethnicity across groups.
The sample was classified into three diagnostic groups: Group I included 48
subjects with a primary diagnosis of OCD (OCD Group); Group II included 41
participants with internalizing disorders other than OCD (e.g., generalized
anxiety disorder, major depression; Internalizing Group); and Group III included
101 subjects with externalizing disorders (e.g., conduct disorder, oppositional
defiant disorder, attention-deficit hyperactivity disorder; Externalizing Group).
The children in Group II had the following primary diagnoses: generalized
anxietydisorder(n = 8),posttraumaticstressdisorder(n = 4),anxietydisordernot
otherwise specified (NOS) (n = 3), social phobia (n = 1), specific phobia (n = 3),
major depression (n = 10), depressive disorder NOS (n = 6), dysthymic disorder
(n = 6). The children in Group III had the following primary diagnoses:
oppositional defiant disorder (n = 41), conduct disorder (n = 6), disruptive
behavior disorder NOS (n = 4), attention deficit hyperactivity disorder, combined
type (n = 35), attention deficit hyperactivity disorder, inattentive type (n = 7),
attention deficit hyperactivity disorder, not otherwise specified (n = 6), and
alcohol abuse (n = 2). No Group II or III children had comorbid diagnoses of
1.2.1. Child Behavior Checklist
The Child Behavior Checklist (Achenbach, 1991) is an empirically derived
behavior rating scale, appropriate for children and adolescents between the ages
of 4 and 18. Parents rate items on a 3-point scale: 0 = not true, 1 = somewhat or
sometimes true, and 2 = very true or often true. The CBCL is widely used and has
established psychometric properties. Mean test-retest reliabilities have been
E.A. Storch et al./Anxiety Disorders 20 (2006) 473–485476
reportedto range from0.95 to1.00, andinternalconsistencyhas ranged from 0.78
to 0.97 (Achenbach, 1991).
1.2.2. Children’s Yale-Brown Obsessive-Compulsive Scale
The Children’s Yale-Brown Obsessive-Compulsive Scale (Scahill et al., 1997)
and compulsions over the past week. The CY-BOCS is internally consistent,
1997; Storch et al., 2004a). Cronbach’s a for the Total Score was .87.
1.2.3. Children’s Depression Inventory – Short-Form
Presence and severity of depressive symptoms were assessed using the
Children’s Depression Inventory – Short-Form (CDI-S; Kovacs, 1992). The CDI-
S is a 10-item self-report measure adapted from the original 27-item version. The
child endorses one of three statements that best describe his or her cognitive,
affective, or behavioral symptoms of depression during the previous 2 weeks.
Psychometric studies of the CDI-S, within clinical and non-clinical populations,
suggest that the measure has relatively high levels of internal consistency and
convergent and divergent validity (Kovacs, 1992). Cronbach’s a in this sample
1.2.4. Tourette’s Disorder Scale – Parent Rated
a 15-item parent-rated scale designed to measure a broad range of symptoms
common to Tourette’s disorder, including tics, obsessions, compulsions,
inattention, hyperactivity, aggression, and mood disturbances. Factor analysis
has identified four factors, namely OCD, Tics, Aggression, and ADHD (Shytle
et al., 2003; Storch et al., 2004b). Factor items are summed to derive factor scale
scores. A Total Score is computed by summing all items. Support for the
convergent and divergent validity of the TODS-PR scores has been reported
Score, OCD, Tics, Aggression, and ADHD factors were .92, .72, .78, .93, and .90.
Following obtainment of parental consent and child assent for the children and
adolescents diagnosed with OCD whowere seen in the psychiatry clinic, families
were administered the CBCL, CY-BOCS, TODS-PR, and CDI-S. Experienced
clinicians (either a postdoctoral clinical psychology fellow or psychiatric nurse)
administered the CY-BOCS to both the child and parent jointly. Training
consisted of an instructional meeting about the CY-BOCS content and structure
with the first or second author, two practice interviews, and two directly observed
interviews. Children and parents completed their respective measure(s) in a
E.A. Storch et al./Anxiety Disorders 20 (2006) 473–485 477
TODS-PR, and CDI-S were not collected for the non-OCD participants. Data on
the CY-BOCS, TODS-PR, or CDI-S were not included for six youth with OCD
due to various reasons (e.g., incomplete, not collected due to time constraints by
1.4. Analytic plan
To examine the factor analytic structure of the OCS in the current sample, we
performed a confirmatory factor analyses (CFA) using LISREL 8.53 on the OCD
group (Jo ¨reskog & So ¨rbom, 2002). The measurement and structural models were
evaluated with the following fit indices: x2, the Goodness of Fit Index (GFI), the
comparative fit index (CFI), the incremental fit index (IFI), Normed Fit Index
(NFI), Relative Fit Index (RFI), and the root mean square residual (RMR). GFI,
CFI, IFI, NFI, and RFI fit indices range from 0 to 1, with values of .95 or higher
indicating an adequate fit between the observed model and the theoretical model
(Chou & Bentler, 1993). For the RMR, values below .05 indicate a good fit and
the data did not adequately fit the hypothesized model, we planned to conduct an
exploratory factor analysis using algorithms such as Glorfeld’s extension and
minimum average partials method.
The internal consistency of the OCS was evaluated using Cronbach’s a
coefficient (Cronbach, 1951). Pearson product-moment correlations were
conducted to examine the relations between the OCS and clinician-ratings of
OCD severity, self-reported depression, and parent-reported OCD symptoms,
behavior, and tics. One-way analysis of variance (ANOVA) was calculated to
investigate differences in the OCS across diagnostic groups (OCD vs.
Internalizing vs. Externalizing). A statistically significant ANOVA was further
examined using Tukey honestly significant difference (HSD) follow-up tests for
pairwise comparisons. We also computed sensitivity and specificity against the
Internalizing and Externalizing Groups to assess the diagnostic accuracy of cut-
2.1. Factor analysis
Based on the previously found 1-factor structure of the OCS (Nelson et al.,
2001),alleightitemswere specifiedasloading on thesinglefactor.TheGoodness
of Fit for this model was poor, x2(21, N = 48) = 97.18, P <.001, with fit indices
corroboratingthisfinding,GFI = .59,RMR = .28,NFI = .26,CFI = .26,IFI = .31,
and RFI = .009.
E.A. Storch et al./Anxiety Disorders 20 (2006) 473–485478
Giventhis, a principal axis exploratory factor analysis (EFA) was performed to
identify a model that fits the current sample data. Criteria for identifying the
factors were based on: (1) Glorfeld’s version of parallel analysis with a sample
size of N = 48 and k = 8 variables (the eigenvalue must be greater than 1.663
eigenvalues for the first component, 1.388 for the second component, and 1.20 for
the third component, using the 95th percentile and 1000 replications); (2) the
minimum average partials method (Velicer, 1976); and (3) the scree plot (see
Fig. 1). A 1-factor solution meeting these criteria and accounting for 38% of the
variance was identified (eigenvalue = 3.066 for the factor). With the use of
available syntax (O’Connor, 2000), Velicer’s MAP test was conducted and
indicated one component as well. Items 32 (feels he/she has to be perfect) and 84
(strange behaviors), however, loaded on a second factor and were subsequently
dropped from the scale because the second factor was not sufficiently strong to be
retained. The eigenvalue was only 1.273, which was less than the required value
(1.388) from Glorfeld’s version of parallel analysis. In addition, the MAP method
and scree plot are evidence for retaining only the first factor. Thus, our revised
version of the OCS (OCS-R) contains six items. See Table 1 for factor loadings.
E.A. Storch et al./Anxiety Disorders 20 (2006) 473–485479
Fig. 1. Scree plot for the OCS-R exploratory factor analysis.
Exploratory factor analysis of the OCS-R
Item no.Item Factor loading
Feels he/she might think or do something bad
Feels too guilty
Can’t get his/her mind off certain thoughts; obsessions
Repeats certain acts over and over; compulsions
2.2. Reliability and convergent validity
Cronbach’s a for the OCS-R was .75. The OCS-R correlated strongly with the
OCS (r = .96, P <.001). Table 2presents correlations among the OCS-R, original
OCS, and measures of OCD, depression, tics, and behavior. Overall, correlations
between the OCS-R and measures of OCD (CY-BOCS and TODS-PR OCD
factor) were highest. The OCS-R was also correlated with the CDI Total Score,
TODS-PR Aggression factor, TODS-PR ADHD factor, and TODS-PR Total
Score with correlations of a moderate effect size. The OCS-R was not
significantly related to the TODS-PR Tic factor. The relations between the
original OCS and CY-BOCS and TODS-PR OCD factor were slightly lower as
compared tothe OCS-R. Additionally,associations betweenthe original OCS and
the CDI Total Score, TODS-PR Aggression factor, TODS-PR ADHD factor, and
TODS-PR Total Score were slightly higher than those with the OCS-R.
2.3. Criterion validity
The overall model examining differences on the OCS-R among the three
diagnostic groups was significant, F(2, 187) = 11.87, P <.001. Tukey’s post hoc
analyses suggest that youth diagnosed with OCD (M = 6.48, S.D. = 3.52) had
significantly higher scores on the OCS-R than youngsters with internalizing
(M = 4.75, S.D. = 3.17) and externalizing behavior disorders (M = 3.69,
S.D. = 3.46).
E.A. Storch et al./Anxiety Disorders 20 (2006) 473–485480
Pearson product moment correlations for various measures of psychological functioning
N = 42
N = 42
N = 42
N = 42
N = 42
N = 42
N = 42
N = 42
(2) CY-BOCS Total
(3) CDI Total
(5) TODS-PR ADHD
(6) TODS-PR OCD
(7) TODS-PR Tic
(8) TODS-PR Total
Note: Correlations between the original OCS and respective indices are presented in parentheses in
column 1. OCS-R = Obsessive-Compulsive Scale – Revised; CY-BOCS Total = Children’s Yale-
Brown Obsessive-Compulsive Scale Total Score; CDI Total = Children’s Depression Inventory Total
Score;TODS-PRAggression = Tourette’sDisorderScale–ParentRatedAggressionfactor;TODS-PR
ADHD = Tourette’s Disorder Scale – Parent Rated ADHD factor; TODS-PR OCD = Tourette’s
Disorder Scale – Parent Rated OCD factor; TODS-PRTic = Tourette’s Disorder Scale – Parent Rated
Tic factor; TODS-PR Total = Tourette’s Disorder Scale – Parent Rated Total Score.
Using ROC analysis, we examined cutoff values that distinguished individuals
with OCD from those with other internalizing disorders and those with
externalizing disorders. First, an ROC analysis was performed on the OCD group
(n = 48) and the internalizing disorders group (n = 41). The area under the curve
(AUC) was .656 and was significant versus the chance or random ROC line
(P <.01). As shown in Table 3, an OCS-R score of 4.5 provided the optimum
balance between sensitivity and specificity. That is, 69% of individuals with OCD
were correctly classified and 56% of the individuals with other internalizing
disorders were correctly classified. A second ROC analysis was conducted on the
OCD group (n = 48) and the externalizing disorders group (n = 101). The AUC
was .732 and was significantly different from the random ROC line (P <.0001).
As shown in Table 3, sensitivity and specificity were maximized at a cutoff value
of 3.5. The majority of the OCD participants (77%) were correctly classified, and
59% of the externalizing group was correctly classified.
The importance of detecting and monitoring obsessive-compulsive symptoms
in pediatric populations highlights the need for a valid, reliable, and sensitive
instrument that can be used in routine clinical practice or as a screening tool. Two
measures have been developed in an attempt to achieve this goal, yet the utility of
both are limited by the heterogeneous nature of pediatric OCD symptoms.
Recently,Nelson et al.(2001) developed the OCS, a subsetof questions contained
within the CBCL that are hypothesized to assess shared phenomenological
elements across diverse pediatric OCD clinical presentations. Although initial
psychometric properties of the OCS were promising, a number of psychometric
results needed replication or initial examination, including the factor structure,
convergent and divergent validity, and criterion validity. Accordingly, each of
these was addressed within this study.
E.A. Storch et al./Anxiety Disorders 20 (2006) 473–485 481
OCS-R cutoff scores and their sensitivity and specificity values
OCD vs. internalizing disorders
OCD vs. externalizing disorders
The OCS factor structure was examined using confirmatory and exploratory
factor analytic techniques. A CFA of the model found by Nelson et al. (2001)
resulted inarelativelypoorfitinthe presentsample.AsubsequentEFArevealeda
similar 1-factor model that retained six of the eight original items (items 32 (feels
he/she has to be perfect) and 84 (strange behavior) were deleted). Several reasons
may account for differences between these results and those of Nelson et al.
(2001). First, we used stringent criteria for factor extraction which may be most
appropriate for the research questions addressed in this study (O’Connor, 2000).
Second, our sample may have differed from Nelson et al. (2001) in terms of
demographics and illness presentation (e.g., comorbidity). The questions deleted
may be more applicable for adolescents than child; thus, the relatively younger
age of this sample may have influenced our findings. Although our sample was
somewhat smaller than that of Nelson et al. (2001), our sample size conforms to
accepted guidelines for factor analysis (Hair, Anderson, Tatham, & Black, 1998).
Reliability and validation analyses were generally supportive for the utility of
the OCS-R. The internal consistency was adequate, although slightly lower than
Nelson et al. (2001). However, the OCS-R correlated strongly with the OCS,
suggesting that the original version may not provide additional information
beyond the OCS-R. Convergent validity was supported vis-a `-vis moderate
relationships with parent-rated and clinician-rated OCD symptoms. The OCS-R
was also positively and moderately associated with depression, aggressive
behavior, and ADHD symptoms. Intuitively, this may reflect the comorbidity
inherent in pediatric OCD. For example, studies have documented high
comorbidity with depression (26%; Swedo, Rapoport, Leonard, Lenane, &
Cheslow, 1989), disruptive behavior (53%; Geller, Biederman, Griffin, Jones, &
two items may provide a more efficient symptom measurement as the OCS-R
showed relatively higher relationships with measures of OCD and slightly lower
associations with measures of behavior than the original OCS.
The OCS-R successfully discriminated between youth with OCD and those
with an internalizing or externalizing disorder. ROC analyses demonstrated that
the OCS-R had acceptable sensitivity and specificity in classifying youth with
OCD from those with an internalizing or externalizing disorder. Use of cutoff
scores should be based on the goals of the measurement. For example, if one
wishes to capture all youth with OCD and has little concern for false positives, a
cutoff score of 2.5 would correctly classify 90% of cases with OCD, but only
exclude 24% of youth with a different internalizing disorder and 49% of those
the use of a sample with high external validity (e.g., multiple comorbid disorders,
attending a outpatient university psychology clinic) and the shared symptom
presentation between OCD and other internalizing disorders. On balance,
however, sensitivity and specificity figures were lower than Nelson et al. (2001)
and not uniformly high for any particular cutoff value suggesting the limits of
relying on the OCS-R alone as a screening instrument. Administering adjunctive
E.A. Storch et al./Anxiety Disorders 20 (2006) 473–485 482
measures concurrently, such as the LOI-CV and C-FOCI, is recommended and
may improve the accuracy of detection rates.
These findings should be interpreted in the context of several limitations. First,
as this is only the second psychometric examination of the OCS and our analysis
favored a slightly revised scale, numerous empirical questions remain. For
example, is the OCS-R factor structure stable, reliable between raters and over
time, sensitive to treatment effects, and uniform across age and gender? Although
Hudziak et al. (2004) provides convincing evidence for the stability of the OCS
across gender and age, this should be examined in the revised measure. Second,
we did not examine specificity and sensitivity in a sample of children without a
psychiatric diagnosis. Specificity and sensitivity estimates may be greater in such
a sample. Third, given that many children disguise or hide their symptoms,
parents may under-report symptoms on the OCS-R. Despite this limitation, the
OCS-R is currently the onlyknownparent-reportinstrument forthe assessment of
pediatric OCD a fact that takes on increased salience given that many youth will
misrepresent symptom severity. On balance, the OCS should not be used in
isolation as parents may not be aware of certain symptoms, particularly those
which are embarrassing to the youth. Finally, although both attending providers
have considerable clinical experience and are board certified or licensed (and
OCD diagnoses were confirmed by an independent clinician), diagnoses were
made on the basis of an unstructured clinical interview and thus, may reflect some
In sum, the content of the OCS-R is similar to the OCS, yet results in a briefer,
psychometrically sound measure. Given the wide use, strong psychometric
properties, and ease of administration and scoring of the CBCL, the OCS-R
provides a method of screening for obsessive-compulsive symptoms in a manner
that is less influenced by symptom heterogeneity than symptom specific
measures. Although these data support the reliability and validity of the OCS-R,
we highlight the need for future studies to examine additional psychometric
questions (e.g., factor structure, inter-parent reliability, and temporal stability).
The authors would like to thank Pam Allen for her contributions to this study.
Achenbach, T. M. (1991). Manual for the Child Behavior Checklist/4-18 and 1991 profile. Burlington,
VT: University of Vermont, Department of Psychiatry.
Berg,C. Z.,Whitaker,A.,Davies,M.,Flament,M.F.,& Rapoport,J.L.(1988).Thesurveyformofthe
Leyton Obsessional Inventory – Child Version: norms from an epidemiological study. Journal of
the American Academy of Child and Adolescent Psychiatry, 27, 759–763.
E.A. Storch et al./Anxiety Disorders 20 (2006) 473–485 483
Chou, C., & Bentler, P. M. (1993). Invariant standardized estimated parameter change for model
modification in covariance structure analysis. Multivariate Behavioral Research, 28, 97–110.
Cronbach, L. (1951). Coefficient alpha and the internal consistency of tests. Psychometrika, 16, 297–
DeVeaugh-Geiss, J., Moroz, G., Biederman, J., Cantwell, D., Fontaine, R., Greist, J. H., et al. (1992).
Clomipramine hydrochloride in childhood and adolescent obsessive–compulsive disorder – a
multicenter trial.Journalofthe AmericanAcademyofChildandAdolescentPsychiatry,31,45–49.
Douglass,H. M., Moffitt,T. E., Dar, R., McGee, R., & Silva, P. (1995). Obsessive-compulsivedisorder
in a birth cohort of 18-year-olds: prevalence and predictors. Journal of the American Academy of
Child and Adolescent Psychiatry, 34, 1424–1431.
Geller, D. A., Biederman, J., Griffin, S., Jones, J., & Lefkowitz, T. R. (1996). Comorbidity of juvenile
obsessive-compulsive disorder with disruptive behavior disorders. Journal of the American
Academy of Child and Adolescent Psychiatry, 35, 1637–1646.
Glorfeld, L. W. (1995). An improvement on Horn’s parallel analysis methodology for selecting the
correct number of factors to retain. Educational and Psychological Measurement, 55, 377–393.
Goodman, W. K., Price, L. H., Rasmussen, S. A., Mazure, C., Fleischman, R. L., Hill, C. L., Heninger,
G. R., & Charney, D. S. (1989a). The Yale-Brown Obsessive Compulsive Scale: I. Development,
use, and reliability. Archives of General Psychiatry, 46, 1006–1011.
Goodman, W. L., Price, L. H., Rasmussen, S. A., & Mazure, C. (1989b). The Yale-Brown Obsessive
Compulsive Scale (Y-BOCS): Validity. Archives of General Psychiatry, 46, 1012–1016.
Hair, J. F., Anderson,R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis(5th ed.).
Upper Saddle River, NJ: Prentice Hall.
Heyman, I., Fombonne, E., Simmons, H., Ford, T., Meltzer, H., & Goodman, R. (2001). Prevalence of
obsessive-compulsive disorder in the British nationwide survey of child mental health. British
Journal of Psychiatry, 179, 324–329.
Hudziak, J. J., Van Beijsterveldt, C. E., Althoff, R. R., Stanger, C., Rettew, D. C., Nelson, E. C., et al.
(2004). Genetic and environmental contributions to the Child Behavior Checklist Obsessive-
Compulsive Scale: a cross-cultural twin study. Archives of General Psychiatry, 61, 608–616.
Jo ¨reskog, K. G., & So ¨rbom, D. (2002). LISREL 8.53 user’s reference guide. Lincolnwood, IL:
Kovacs, M. (1992). The Children’s Depression Inventory (manual). Toronto, Canada: Multi-Health
Leonard, H. L., Swedo, S. E., Lenane, M. C., Rettew, D. C., Hamburger, S. D., Bartko, J. J., et al.
(1993). A 2- to 7-year follow-up study of 54 obsessive-compulsive children and adolescents.
Archives of General Psychiatry, 50, 429–439.
Maina, G., Albert, U., Bogetto, F., & Ravizza, L. (1999). Obsessive-compulsive syndromes in older
adolescents. Acta Psychiatrica Scandinavia, 100, 447–450.
Merlo, L. J., Storch, E. A., Murphy, T. K., Goodman, W. K., Geffken, G. R. (in press). Assessment of
pediatric obsessive-compulsive disorder: a critical review of current methodology. Child Psy-
chiatry and Human Development.
Murphy, T. K., Sajid, M., Soto, O., Shapira, N., Edge, P., Yang, M., et al. (2004). Detecting pediatric
autoimmune neuropsychiatric disorders associated with streptococcus in children with obsessive-
compulsive disorder and tics. Biological Psychiatry, 55, 61–68.
Nelson, E. C., Hanna, G. L., Hudziak, J. J., Botteron, K. N., Heath, A. C., & Todd, R. D. (2001).
Obsessive-Compulsive Scale of the Child Behavior Checklist: specificity, sensitivity, and pre-
dictive power. Pediatrics, 108, 14 [Electronic version].
O’Connor, B. P. (2000). SPSS and SAS programs for determining the number of components using
parallel analysis and Velicer’s MAP test. Behavior Research Methods, Instruments & Computers,
Pediatric Obsessive-Compulsive Disorder Treatment Study (POTS) Team. (2004). Cognitive-beha-
vioral therapy, sertraline, and their combination for children and adolescents with obsessive-
E.A. Storch et al./Anxiety Disorders 20 (2006) 473–485484
compulsive disorder: the Pediatric Obsessive-Compulsive Disorder Treatment Study (POTS) Download full-text
randomized controlled trial. Journal of the American Medical Association, 292, 1969–1976.
Piacentini, J., Bergman, R. L., Keller, M., & McCracken, J. T. (2003). Functional impairment in
children and adolescents with obsessive-compulsive disorder. Journal of Child and Adolescent
Psychopharmacology, 13S-1, 61–S69.
Rapoport, J. L., & Inoff-Germain, G. (2000). Treatment of obsessive-compulsive disorder in children
and adolescents. Journal of Child Psychology and Psychiatry, 41, 419–431.
Scahill, L., Riddle, M. A., McSwiggin-Hardin, M., Ort, S. I., King, R. A., Goodman, W. K., et al.
American Academy of Child and Adolescent Psychiatry, 36, 844–852.
Schniering, C. A., & Rapee, R. M. (2002). Development and validation of a measure of children’s
automatic thoughts: the children’s automatic thoughts scale. Behaviour Research and Therapy, 40,
Shytle, R. D., Silver, A. A., Sheehan, K. H., Wilkinson, B. J., Newman, M., Sanberg, P. R., et al.
(2003). The Tourette’s Disorder Scale (TODS): development, reliability and validity. Assessment,
evaluation of the Children’s Yale-Brown Obsessive Compulsive Scale. Psychiatry Research, 129,
Storch, E. A., Murphy, T. K., Geffken, G. R., Soto, O., Sajid, M., Allen, P., et al. (2004b). Further
psychometric properties of the Tourette’s Disorder Scale – Parent Rated Version (TODS-PR).
Child Psychiatry and Human Development, 35, 107–120.
Storch, E. A., Murphy, T. K., Geffken, G. R., Soto, O., Sajid, M., Bagner, D. M., et al. (2005). Factor
Adolescent Psychology, 34, 312–319.
Swedo, S. E., Rapoport, J. L., Leonard, H., Lenane, M., & Cheslow, D. (1989). Obsessive-compulsive
disorder in children and adolescents: clinical phenomenology of 70 consecutive cases. Archives of
General Psychiatry, 46, 335–341.
Velicer, W. F. (1976). Determining the number of components from the matrix of partial correlations.
Psychometrika, 41, 321–327.
Velicer, W. F., Eaton, C. A., & Fava, J. L. (2000). Construct explication through factor of component
components.In: R. D. Goffin,& E. Helmes(Eds.), Problemsand solutionsin humanassessment: a
festschrift to Douglas Jackson at seventy (pp. 41–71). Hillsdale, NJ: Erlbaum.
E.A. Storch et al./Anxiety Disorders 20 (2006) 473–485485