DEPRESSION

AND

ANXIETY 28 : 863869 (2011)

Research Article
ETIOLOGY OF OBSESSIVECOMPULSIVE SYMPTOMS AND
OBSESSIVECOMPULSIVE PERSONALITY TRAITS:
COMMON GENES, MOSTLY DIFFERENT ENVIRONMENTS
Steven Taylor, Ph.D.,1 Gordon J.G. Asmundson, Ph.D.,2 and Kerry L. Jang, Ph.D.1

Background: Little is known about the etiologic relationship between obsessive

compulsive (OC) symptoms and traits of OC personality disorder. The traits
include perfectionism and rigidity. Some theorists have proposed that OC
personality disorder is one of several disorders falling within an OC spectrum.
This implies that OC personality traits and symptoms should have etiologic
factors in common, and this should not be simply because symptoms and traits are
both shaped by nonspecific etiological influences, such as those shaping negative
emotionality (neuroticism). Methods: To investigate these issues, a community
sample of 307 pairs of monozygotic and dizygotic adult twins provided scores on
six types of OC-related symptoms, two markers of negative emotionality, and a
measure of OC personality traits. Results: Analyses indicated that symptoms and
traits arose from a combination of genetic and nonshared environmental factors.
A matrix of genetic correlations was computed among the variables, which
represented the correlations between the genetic components of pairs of variables.
A matrix of environmental correlations was similarly computed. Each matrix
was factor analyzed. One genetic factor was obtained, indicating that all
variables were influenced by a common genetic factor. Three environmental
factors were obtained, with salient loadings on either (a) all six OC symptoms,
(b) negative emotionality and obsessing, or (c) OC personality traits and
ordering. Conclusions: OC symptoms and traits were etiologically related
primarily because they are shaped by the same nonspecific genetic factor that
influenced negative emotionality. Implications for the concept of the OC spectrum
are discussed. Depression and Anxiety 28:863869, 2011.
r 2011 Wiley-Liss, Inc.
Key words: behavioral genetics; compulsions; obsessions; obsessivecompulsive
spectrum; obsessivecompulsive personality; twins

INTRODUCTION

Two major approaches the classification of psycho-

pathology involve lumping, that is, grouping disorders into broad categories, and splitting, which is
the dividing of disorders into subtypes. There are
merits to both approaches. Lumping involves the
grouping of disorders that seem to share similar
phenotypes. For example, grouping together obsessivecompulsive disorder (OCD) and the putative

Department of Psychiatry, University of British Columbia,

Vancouver, British Columbia, Canada
2
Department of Psychology, University of Regina, Regina,
Saskatchewan, Canada

r 2011 Wiley-Liss, Inc.

obsessivecompulsive (OC) spectrum disorders, as

described later in this article. The lumping of such
disorders together into a single diagnostic group
The authors disclose the following financial relationships within
the past 3 years: Contract grant sponsor: Canadian Institutes of
Health Research Institute of Neurosciences, Mental Health, and
Addiction; Contract grant number: PTS-63186.
Correspondence to: Steven Taylor, Department of Psychiatry,
University of British Columbia, 2255 Wesbrook Mall, Vancouver, BC
V6T 2A1, Canada. E-mail: taylor@unixg.ubc.ca

Received for publication 4 April 2011; Revised 25 May 2011;

Accepted 28 May 2011
DOI 10.1002/da.20859
Published online 18 July 2011 in Wiley Online Library (wiley
onlinelibrary.com).

864

Taylor et al.

encourages researchers to investigate whether such

disorders have common etiological mechanisms and
treatments. Conversely, the splitting of disorders, such
as the splitting of OCD into subtypes, encourages
researchers to investigate why, for example, a given
person exhibits one kind of OC symptom (e.g.,
checking) instead of some other (e.g., washing), and
how these differences might be translated into different
approaches to treatment. Lumping and splitting
approaches to classification are consistent with one
another because subtypes can be nested within
disorders which, in turn, can be nested within broader
categories. Moreover, behavioral genetic (twin) studies
suggest that the etiology of psychopathology is
organized, such that there are broad (disorder nonspecific) etiological factors as well as narrow (disorder
or subtype specific) factors.[1] The purpose of this study
was to investigate this issue with regard to OC
symptoms and traits of OC personality disorder. The
latter is said to fall within the OC spectrum. Unlike
conventional approaches to classification, we used an
etiology-based approach; that is, by using behavioral
genetic methodology to assess whether OC subtypes
and personality traits have etiological factors (genetic
or environmental) in common.
Hollander et al.[2,3] argued that for DSM-V, OCD
should be moved out of the anxiety disorders category
and into a new category of OC spectrum disorders (also
called OC-related disorders). The composition of this
category is currently unclear. Hollander et al.[3] proposed that the category should include OCD, OC
personality disorder, Tourettes syndrome, grooming
disorders (trichotillomania, chronic skin picking, or
nail biting), body dysmorphic disorder, and eating
disorders. In other publications, Hollander et al. have
argued that the spectrum should include a broader
range of disorders, including hypochondriasis, pathological gambling, depersonalization disorder, and autism.[2] The putative OC spectrum disorders were
thought to be etiologically related to OCD for various
reasons. For example, some of the symptoms of
spectrum disorders are phenotypically similar to
obsessions and compulsions.[3]
The concept of the OC spectrum is controversial,
partly because it has yet to be demonstrated that the
features of these disorders are etiologically related to
one another and to OC symptoms.[4] In this study, we
examined whether OC personality traits can be usefully
regarded as falling with the OC spectrum. This study is
a secondary analysis of a previously reported twin study
of OC symptoms.[5] That study was particularly
concerned with the interrelationships among OC
symptom subtypes. This investigation reports new
findings on the etiologic relationship between these
symptoms and OC personality traits.
The core traits of OC personality disorder, as
identified by empirical investigations, are excessive
perfectionism (e.g., preoccupation with details and
orderliness, excessive work devotion) and rigidity
Depression and Anxiety

(reluctance to delegate, hypermorality, excessive need

for control over ones environment).[6] These traits are
egosyntonic; that is, they are congruent with the desires
and goals of the individual. People with OC personality
disorder typically experience some degree of satisfaction or even enjoyment while engaging in perfectionistic behaviors, even though such behaviors have
deleterious consequences, such as impaired social and
occupational functioning.[7] In comparison, the symptoms characteristic of OCD are typically egodystonic;
that is, obsessions and compulsions are in conflict or
dissonant with the desires and goals of the individual.[4]
Although OCD and OC personality disorder may cooccur with one another, it is unknown whether the two
are etiologically related.
If putative OC symptom subtypes and OC personality traits are etiologically related to one another, then
the question arises as to whether this is something
specific to these phenomena or whether this is due to
broad etiologic factors that influence many other
disorders. Broad factors may be the genetic and
environmental determinants of negative emotionality
(neuroticism). The latter is characterized by a tendency
to experience a range of negative emotions (e.g.,
anxiety, irritability, depression) in response to threat,
frustration, or loss.[8] Neuroticism (or more accurately,
the constellation of genetic and environmental determinants that shape neuroticism) is considered to be a
broad etiologic factor that influences many different
kinds of psychopathology.[8] Accordingly, we investigated the etiologic relationship between OC symptoms, OC personality traits, and trait anxiety and
affective lability. The latter two variables are major
facets or markers of neuroticism.[9,10] The concept of
an OC spectrum would not be supported if we found
that spectrum conditions and OC symptoms were
etiologically related merely because both were influenced by some factor that plays a role in many other
emotional disorders. Such a finding would not provide
evidence of a special etiological relationship between
OC symptoms and personality traits.
In summary, the purpose of this study was to conduct
an etiology-based (behavioral genetic) investigation of
whether OC personality traits are etiologically related
to OC symptoms. If the traits do belong in the OC
spectrum then they should have genetic and/or
environmental etiological factors in common with the
OC symptom subtypes, and this relationship should
not simply be due to shared etiology with two markers
of neuroticismtrait anxiety and affective lability.

MATERIALS AND METHODS

PARTICIPANTS
A community sample of 167 monozygotic (MZ) and 140 dizygotic
(DZ) twin pairs was recruited across Canada as part of the University
of British Columbia Twin Project.[11] A community sample was
selected because obsessions and compulsions are common in the
general population, even though such symptoms may not be

Research Article: OC Symptoms and Personality Traits

865

sufficiently severe to merit a diagnosis of OCD.[12] These nonclinical OC symptoms, compared to obsessions and compulsions in
people diagnosed with OCD, tend to be less frequent, shorter in
duration, and associated with less distress. However, nonclinical OC
symptoms have similar form and content to obsessions and
compulsions found in OCD.[12] Moreover, taxometric research
generally suggests that OC symptom severity is more likely to be
continuous (dimensional) than taxonic (categorical).[13,14] Similarly,
OC personality traits vary on a continuum from mild to clinically
severe, with the milder forms being common in community samples
and similar in form to the more severe forms.[15] Accordingly, studies
of nonclinical samples, such as this investigation, are relevant for
understanding putative OC subtypes and spectrum conditions.
The sample consisted of 33 MZ malemale pairs, 134 MZ
femalefemale pairs, 14 DZ malemale pairs, 86 DZ femalefemale
pairs, and 40 DZ malefemale pairs. Most (78%) were women and
the mean age was 40 years (SD 5 15 years). Most (68%) were
employed full- or part-time, with the remainder being full-time
students (7%), full-time homemakers (7%), retirees (7%), or people
subsisting on disability or unemployment benefits (10%). For further
sample details, see ref. 5.

cannot tolerate a mess). The remaining OC personality items

referred to the egosyntonic desire to perform things perfectly (e.g.,
I do everything to the best of my ability) or refer to feelings of
enjoyment or satisfaction when tasks or ones time are perfectly
organized (e.g., I like attending to small details, I am happiest
when my time is carefully organized). This scale does not contain
items pertaining to hoarding, and so does not overlap with the OCIR hoarding scale. Although DSM-IV suggests that hoarding can be a
feature of OC personality disorder, research indicates that hoarding is
not a core feature of OC personality disorder.[6] Based on their
research, Hummelen et al.[22] concluded that hoarding should not be
included in the diagnostic criteria for OC personality disorder.

MEASURES

STATISTICAL PROCEDURES

Zygosity was determined by means of a highly accurate

questionnaire,[16,17] along with an examination of recent color
photographs. The questionnaire has an accuracy of 9395% in
establishing zygosity, compared with DNA testing.[16,18]
Participants also completed the ObsessiveCompulsive InventoryRevised (OCI-R),[19] which consists of six three-item subscales that
assess the major types of putative OC symptom subtypes: checking,
neutralizing (cognitive rituals), washing, obsessing, hoarding, and
ordering.1 For each item, respondents rated on a five-point scale on
how much each symptom had distressed or bothered them during the
past month. The scale ranged from 0 (not at all) to 4 (extremely).
The following OCI-R items illustrate what each scale measured.
Checking: I check things more often than necessary; neutralizing:
I feel I have to repeat certain numbers; washing: I sometimes have
to wash or clean myself simply because I feel contaminated;
obsessing: I am upset by unpleasant thoughts that come into my
mind against my will; hoarding: I have saved up so many things that
they get in the way; and ordering: I need things to be arranged in a
particular order. Scores on the OCI-R have been shown to have
good reliability and validity in clinical and nonclinical samples.[1921]
The Dimensional Assessment of Personality Pathology (DAPP)[15]
was also completed by participants. Scores on the DAPP have been
shown to have good reliability and validity in clinical and nonclinical
samples.[15] Three DAPP subscales were used for the purpose of this
study. Two were measures of trait anxiety (anxiety proneness) and
affective lability (moodiness or liability to experience negative
emotions), which are both markers for neuroticism or negative
emotionality.[9,10] Each subscale consists of 16 statements rated on a
five-point scale (1 5 very unlike me, 5 5 very like me). The third
DAPP subscale measured the OC personality traits of perfectionism
and rigidity (constructed by the scale developers to be measured
together rather than a subscale for each trait). For the purpose of this
study, 3 items from this 16-item scale were omitted because they
overlap with the OCI-R ordering subscale. The omitted DAPP items
were those referring to distress or intolerance of disorder (e.g., I

Statistical analyses proceeded in three steps. First, we computed

the heritabilities of all variables, where each variable was assessed by
one of the above-mentioned scales. Second, if a variable was
heritable, we then computed its genetic and environmental correlations with other heritable variables. Third, we assembled two
matricesa matrix of genetic correlations and a matrix of environmental correlationsand analyzed each matrix by means of
exploratory factor analyses to determine whether the variables have
genetic or environmental factors in common. Details of the data
analytic methods are as follows.
Each variable was defined, by means of structural equation
modeling, as a standardized latent variable (i.e., with M 5 0,
SD 5 1) in which its respective items served as indicators. For each
variable, the item loadings for MZ twins were equated with those of
DZ twins, so that the latent variables for MZ twins represented the
same constructs as those of DZ twins. Latent variables were
computed by means of Mplus, version 5,[23] in which covariance
matrices were analyzed by means of robust Maximum Likelihood
estimation. Heritability, genetic, and environmental correlations were
calculated by the same procedures.
Heritability estimates were based on the within-pair similarities of
MZ pairs compared to those of DZ pairs. In general, larger MZ than
DZ correlations indicate the presence of genetic effects (that is,
effects due to segregating genes) because the greater MZ similarity is
attributed to the two-fold greater genetic similarity of MZ than DZ
twins. Biometric models of genetic and environmental influences
were fitted to the variancecovariance matrices of the latent variables.
For each variable, MZ and DZ within-pair variancecovariance
matrices were decomposed by means of biometric structural equation
modeling into variance components attributable to genetic factors
(A), shared environmental factors (i.e., factors common to both
members of a twin pair; C), and nonshared environmental factors
(i.e., factors not shared by members of a given pair; E).[24] Shared
environmental factors include family influences common to both
members of a twin pair, such as parental instruction or modeling
(observational learning). Nonshared environmental factors include
events affecting one twin but not the other. Nonshared environmental factors also include measurement error, although error was
minimized as far as possible by using psychometrically sound
instruments and by specifying the variables in this study as latent

The debate about whether hoarding phenomena should be classified

as OC symptoms is beyond the scope of this article. This issue is
discussed in a previous study based on the same dataset as the present
investigation Ref. 5.

PROCEDURE
Twin pairs were recruited through newspaper advertisements print,
radio media stories, and twin club registries. Inclusion criteria
consisted of fluency in written and spoken English. Written informed
consent was obtained and participants were mailed a packet of
questionnaires. Participants were asked to independently complete
the questionnaires in a nondistracting setting. Each participant
received an honorarium for being involved in the study.

Depression and Anxiety

866

Taylor et al.

Sex, coded as 0 5 male and 1 5 female, was correlated

from .02 to .13 with these variables. The magnitude of
correlations was very small, falling below what Cohen[26]
classified as a small correlation. Accordingly, we did
not adjust for age or sex in the remaining analyses.
For each variable, the best fitting model, as assessed by
Akaikes Information Criterion, consisted of variance
components attributable to genetic and nonshared
environmental factors. All variables were significantly
heritable. As previously reported for this dataset, genetic
factors accounted for 4055% variance for OC symptoms and 5356% variance for negative emotionality.
The remaining variance was due to nonshared environment.[5] For OC personality traits, which were not
analyzed in our previous study, the proportions of
variance (and 95th percentile confidence intervals) for
genetic (A) and nonshared environment (E) were,
respectively, .45 (.33.58) and .55 (.42.67). In other
words, 45% of phenotypic variance in scores on OC
personality traits was due to genetic factors and the
remainder to nonshared environment. For descriptive
purposes, genetic and environmental correlations among
all the variables are shown in Tables 2 and 3, respectively.

variables, with questionnaire items as indicators. With this approach,

error variance was modeled separately from the true score variance
attributable to the latent variables.
Like most twin studies,[1] we did not have sufficient statistical
power to disentangle the effects of additive, dominance, and epistatic
genetic factors. Accordingly, our estimate of A (defined above) refers
to broad heritability, representing the proportion of variance
attributable to all forms of genetic influence on the variables.
For each of these variables, the best fitting model was identified by
progressively dropping elements of the model and comparing the
relative goodness of fit; that is, comparing the following models:
ACE, AE, CE, and E. The last-mentioned model is the most basic,
consisting of unique environmental effects and residual error
variance. Goodness of fit of these models was assessed by the
Akaikes Information Criterion (AIC), defined as w22df. The bestfitting model was that with the smallest AIC value.
Choleskys decomposition was used to compute the genetic and
environmental correlations among the variables.[24] Genetic correlations represent the correlations among the genetic components of the
variables. Environmental correlations are the correlations among the
components of the variables due to environmental factors. This
enabled us to assemble, for all variables, a matrix of genetic
correlations and a matrix of environmental correlations. Although
we refer to these matrices as correlation matrices, they are also
variancecovariance matrices, because the latent variables were scaled
to have a mean of 0 and standard deviation of 1.
Each matrix was analyzed by means of exploratory factor analysis,
using Maximum Likelihood factor analysis with oblique (Oblimin)
rotation. The number of factors to extract was determined by parallel
analysis using mean and 95th percentile simulated Eigenvalues. Here,
1,000 data matrices, based on the same sample size and number of
variables as in this study, were randomly generated and the resulting
mean and 95th percentile Eigenvalues were computed. These values
were compared against the Eigenvalues obtained in each exploratory
factor analysis. The number of factors to be retained for the latter was
determined by the number of observed Eigenvalues that exceeded the
simulated Eigenvalues.[25] We conducted an exploratory factor
analysis instead of a confirmatory analysis because the genetic and
environmental factor structures were unknown.

FACTOR ANALYSES
Factor analysis of the matrix of genetic correlations
yielded a one-factor solution, which accounted for 64%
of variance. The Eigenvalues were 5.76, 1.04, 0.64,
0.60, 0.32, 0.26, 0.18, 0.16, and 0.05. The pattern of
factor loadings (Table 4) shows that the single-factor
solution represents a general distress factor, with the
markers of negative emotionality having the highest
loadings.
For the matrix of environmental correlations, a
three-factor solution was obtained, accounting for
80% of variance. The Eigenvalues were 4.37, 1.67,
1.16, 0.63, 0.48, 0.25, 0.20, 0.18, and 0.08. Table 4
shows that the factors had small correlations among
one another, according to Cohens[26] scheme for
classifying the magnitude of correlations. The pattern
of factor loadings (Table 5) indicates that the factors
can be labeled as follows: (I) OC personality traits and
ordering, (II) negative emotionality and obsessing, and
(III) OC symptoms. With the exception of two
variables with cross-loadings, the pattern of factor

RESULTS
PRELIMINARY ANALYSES
All the phenotypic (observed) correlations among OC
symptoms and personality traits were statistically
significant (Po.05). These correlations are presented
in Table 1 for descriptive purposes. Age was correlated
from .16 to .10, with the variables listed in Table 1.
TABLE 1. Phenotypic (observed) correlations

Obsessing Neutralizing Checking Washing Hoarding Ordering Affective lability Trait anxiety
Neutralizing
Checking
Washing
Hoarding
Ordering
Affective lability
Trait anxiety
Obsessivecompulsive personality traits
All Po.05.
Depression and Anxiety

.63
.58
.56
.51
.46
.66
.70
.20

.76
.71
.57
.64
.49
.47
.27

.79
.56
.68
.42
.49
.35

.47
.66
.38
.42
.28

.49
.47
.45
.12

.42
.43
.59

.87
.32

.35

Research Article: OC Symptoms and Personality Traits

867

TABLE 2. Genetic correlations

Obsessing Neutralizing Checking Washing Hoarding Ordering Affective lability Trait anxiety
Neutralizing
Checking
Washing
Hoarding
Ordering
Affective lability
Trait anxiety
Obsessivecompulsive personality traits

.64
.69
.72
.64
.58
.75
.81
.35

.71
.66
.51
.51
.56
.57
.17

.63
.60
.68
.60
.62
.35

.44
.60
.62
.71
.33

.60
.55
.65
.22

.61
.61
.68

.89
.24

.34

All Po.05.

TABLE 3. Environmental correlations

Obsessing Neutralizing Checking Washing Hoarding Ordering Affective lability Trait anxiety
Neutralizing
Checking
Washing
Hoarding
Ordering
Affective lability
Trait anxiety
Obsessivecompulsive personality traits

.58
.42
.44
.37
.38
.55
.54
.28

.76
.65
.51
.73
.28
.23
.33

.84
.41
.71
.21
.26
.28

.36
.59
.22
.19
.20

.35
.29
.22
.02

.28
.26
.49

.86
.33

.31

Environmental correlations pertain to nonshared environment; that is, environment not shared within each twin pair. All Po.05, except for the
correlation between hoarding and obsessivecompulsive personality traits.

TABLE 4. Factor loadings on the genetic factor

Variable

TABLE 5. Factor loadings on environmental factors and

correlations among these factors
Loading

Trait anxiety
Obsessing
Affective lability
Washing
Checking
Ordering
Neutralizing
Hoarding
Obsessivecompulsive personality traits

.91
.88
.87
.78
.76
.73
.70
.70
.50

Bold: Salient (Z.30) loadings.

loadings indicates that the three classes of variables

(OC personality traits, negative emotionality, and OC
symptoms) can be clearly etiologically distinguished in
terms of environmental factors.

DISCUSSION
Hollander et al.[2,3] argued that OCD should be
moved out of the anxiety disorders category and placed
in a new DSM-V category comprising OCD, OC
personality disorder, Tourettes syndrome, grooming
disorders (trichotillomania, chronic skin picking, or
nail biting), body dysmorphic disorder, eating disorders,
and possibly other disorders. Despite suggestions that
some symptoms of the putative OC spectrum disorders

Variable

Factor I

Obsessivecompulsive
personality traits
Ordering
Affective lability
Trait anxiety
Neutralizing
Checking
Washing
Obsessing
Hoarding
Factor correlations
I
II
III

Factor II

Factor III

.96

.14

.11

.38
.11
.11
.10
.07
.01
.10
.21

.00
.96
.84
.02
.08
.04
.46
.19

.70
.02
.06
.82
.94
.87
.43
.48

.08
.13

.27

Bold: Salient (Z.30) loadings.

are phenotypically similar to obsessions and compulsions,[3] there is also evidence of important divergence;
for example, the symptoms of OC personality disorder
and grooming disorders seem egosyntonic, whereas
those of OCD are egodystonic.[27] It remains to be
demonstrated that the features of the putative OC
spectrum disorders are etiologically related to one
another and to obsessions and compulsions.[27]
This study used behavioral genetic methods to
address this issue by investigating the nature of the
Depression and Anxiety

868

Taylor et al.

etiologic relations between OC symptoms and OC

personality traits. Both specific and broad etiologic
factors, the latter reflected in measures of neuroticism,
were investigated. OC symptoms, OC personality traits
and negative emotionality were influenced by a
common (general distress) genetic factor, but largely
differed in their environmental etiologic factors. Thus,
OC symptoms and OC personality traits were etiologically related primarily because both were influenced
by a broad genetic factor that plays a role in general
distress (negative emotionality). Negative emotionality
has been implicated in many different forms of
psychopathology, including anxiety, mood, and many
other disorders.[8] Thus, our results do not support the
view that OC personality traits and OC symptoms have
some special etiologic relationship. Such a relationship
would be expected to be found if OC personality traits
belong in the OC spectrum.
The categorization of symptoms or disorders into
OC subtypes or spectrum conditions, as proposed by
Hollander et al.[2,3] is a rough classification that does
not appear to neatly carve nature at its joints. This is
most clearly illustrated by our findings for ordering and
obsessing symptoms, which have mixed etiologies, as
indicated by the cross-loadings in Table 5. Ordering
was etiologically related to other types of OC
symptoms (neutralizing, checking, washing, obsessing,
and hoarding) and to OC personality traits; they shared
a common genetic factor and there was some degree of
overlap in environmental influences. Obsessing was
shaped by genetic and environmental etiologic factors
that shaped other OC symptoms, as well as an
environmental factor that shaped negative emotionality.
It is also noteworthy that all variables were influenced
by variable-specific genetic and environmental factors.
This is indicated by the fact that none of the factor
loadings in Tables 4 and 5 were 1.00.
We have argued elsewhere[27,28]reorganization
along the lines suggested by Hollander et al. will have
little, if any, impact on clinical practices or treatments
and that such reorganization is lacking in empirical
support. The proposed OC spectrum disorders category does not fully capture the etiological complexities
of OC symptoms and OC personality traits identified in
this study. A classification system, such as that proposed
by Hollander et al., is predicated, in part, on the notion
that there is a special genetic link between OCD and
the putative OC spectrum conditions. These findings
suggest that, instead, OCD and the putative OC
spectrum conditions or at least OC personality traits
share a broad genetic factor that becomes expressed as a
function of unique environmental influences.
A further problem is that the diagnostic grouping of
OCD with OC personality disorder has little or no
clinical utility regarding the selection of treatment. In
order for a diagnostic grouping to have value to the
practicing clinician, it should have clear treatment
implications. That is, disorders within a diagnostic group
(e.g., a group of putative OC spectrum disorders) should
Depression and Anxiety

all respond to similar sorts of treatments, and these

treatments should be different from disorders classified
within some other diagnostic group (e.g., psychotic
disorders). This is a further reason for not grouping
OCD and OCPD into the same diagnostic group; the
treatments of choice for OCD are serotonergic medications and cognitivebehavioral interventions entailing
exposure and response prevention.[4] These are not the
treatments of choice for OCPD. In fact, little is known
about the optimal treatment of OCPD, although there is
some support for the efficacy of supportiveexpressive
psychotherapy.[29]
Like most investigations, this study has a mix of
strengths and limitations. The strengths include the
use of psychometrically sound instruments, powerful
(latent variable based) statistical methods, and a
methodology that permitted inferences about etiology.
The question arises as to whether results from twins
can be generalized to the population at large. Our
previous twin studies, based on other twin samples
collected by our research group, have found that twins
are representative of the general population in terms of
demographics, personality traits, and various clinical
variables.[11] However, this was not explicitly examined
in this study.
A further limitation includes the use of a nonclinical
sample. It remains to be seen whether our findings can
be generalized to clinical samples, such as people
diagnosed with OCD or OC personality disorder.
Another limitation is that we did not fully assess the
range of clinical conditions that are thought to fall
within the OC spectrum. Our analyses were limited to
OC personality traits. Further research is needed, in
which the entire range of putative OC spectrum
conditions are assessed, in order to determine whether
they are etiologically related to OC symptoms.
CONCLUSION
OC symptoms and OC personality traits were related
because they were both etiologically influenced by a
nonspecific genetic factor; that is, the factor associated
with negative emotionality. This does not support the
view that OC personality traits are part of the OC
spectrum. Future research is also needed to delineate the
nature of the genetic and environmental factors identified
in this study. Little is known about the genes associated
with OC symptoms or OC personality traits. Candidate
genes for OCD have been identified in preliminary
studies,[30] but replication is required because replication
failures are almost the norm when one attempts to
identify genetic markers of a given disorder.[30,31]
Further research is also need to delineate the
environmental factors that contribute to OC symptoms
and to OC personality traits. The presence or severity of
OC symptoms is correlated with particular environmental experiences, such as childhood trauma, especially
sexual abuse.[32] It has also been speculated that
particular learning experiences, such as those that teach

Research Article: OC Symptoms and Personality Traits

the child to acquire an inflated sense of personal

responsibility, are conducive to the development of
OC symptoms.[33] More research is needed to determine
whether such environmental events play a causal role.
The role of geneenvironment interactions in the
development of OC symptoms and OC personality
traits also remains to be investigated.

REFERENCES
1. Kendler KS, Prescott CA. Genes, Environment, and Psychopathology: Understanding the Causes of Psychiatric and
Substance Use Disorders. New York: Guilford; 2006.
2. Bartz JA, Hollander E. Is obsessivecompulsive disorder an
anxiety disorder? Prog Neuropsychopharmacol Biol Psychiatry
2006;30:338352.
3. Hollander E, Braun A, Simeon D. Should OCD leave the anxiety
disorders in DSM-V? The case for obsessivecompulsive related
disorders. Depress Anxiety 2008;25:317329.
4. Abramowitz JS, Taylor S, McKay D. Obsessivecompulsive
disorder. Lancet 2009;373:489497.
5. Taylor S, Jang KL, Asmundson, G. Etiology of obsessions and
compulsions: a behavioral-genetic analysis. J Abnorm Psychol
2010;119:672682.
6. Pinto A, Eisen JL, Mancebo MC, Rasmussen SA. Obsessive
compulsive personality disorder. In: Abramowitz JS, McKay D,
Taylor S, editors. ObsessiveCompulsive Disorder: Subtypes and
Spectrum Conditions. Amsterdam: Elsevier; 2008:246270.
7. Livesley WJ. Practical Management of Personality Disorder.
New York: Guilford; 2003.
8. Lahey BB. Public health significance of neuroticism. Am Psychol
2009;64:241256.
9. Larstone RM, Jang KL, Livesley WJ, et al. The relationship
between Eysencks P-E-N model of personality, the five-factor
model of personality, and traits delineating personality dysfunction. Pers Indiv Differ 2002;33:2537.
10. Schroeder ML, Wormworth JA, Livesley WJ. Dimensions of
personality disorder and their relationships to the Big Five
dimensions of personality. Psychol Assess 1992;4:4753.
11. Jang KL, Taylor S, Livesley WJ. Personality is something and
personality does something: the University of British Columbia
Twin Project. Twin Res Hum Genet 2006;9:739742.
12. Gibbs N. Nonclinical populations in research on obsessive
compulsive disorder. Clin Psychol Rev 1996;16:729773.
13. Haslam N, Williams BJ, Kyrios M, et al. Subtyping obsessive
compulsive disorder: a taxometric analysis. Behav Ther 2005;36:
381391.
14. Olatunji BO, Williams BJ, Haslam N, et al. The latent structure
of obsessivecompulsive symptoms: a taxometric study. Depress
Anxiety 2008;25:956968.
15. Livesley WJ, Jackson DN. Manual for the Dimensional
Assessment of Personality Pathology. Port Huron, MI: Sigma
Assessment Systems; 2009.
16. Kasriel J, Eaves LJ. The zygosity of twins: further evidence on
the agreement between diagnosis by blood groups and written
questionnaires. J Biosoc Sci 1976;8:263266.

869

17. Nichols RC, Bilbro WC. The diagnosis of twin zygosity. Acta
Genetica et Statistica Medica 1966;16:265275.
18. Reed T, Plassman BL, Tanner CM. Verification of self-report
of zygosity determined via DNA testing in a subset of the
NAS-NRC twin registry 40 years later. Twin Res Hum Genet
2005;8:362367.
19. Foa EB, Huppert JD, Leiberg S, et al. The obsessivecompulsive
inventory: development and validation of a short version. Psychol
Assess 2002;14:485496.
20. Abramowitz JS, Deacon BJ. Psychometric properties and
construct validity of the obsessivecompulsive inventory-revised:
replication and extension with a clinical sample. J Anxiety Disord
2006;20:10161035.
21. Huppert JD, Walther MR, Hajcak G, et al. The OCI-R:
validation of the subscales in a clinical sample. J Anxiety Disord
2007;21:394406.
22. Hummelen B, Wilberg T, Pedersen G, Karterud, S. The quality
of the DSM-IV obsessivecompulsive personality disorder
construct as a prototype category. J Nerv Ment Dis 2008;196:
446455.
23. Muthen LK, Muthen BO. Mplus Users Guide. 5th ed. Los
Angeles, CA: Author; 2007.
24. Neale MC, Maes H. Methodology for Genetic Studies of Twins
and Families. Dordrecht, The Netherlands: Kluwer; 2004.
25. OConnor BP. SPSS and SAS programs for determining the
number of components using parallel analysis and Velicers MAP
test. Behav Res Meth Ins C 2000;32:396402.
26. Cohen J. Statistical Power Analyses for the Behavioral Sciences.
2nd ed. Hillsdale, NJ: Erlbaum; 1988.
27. Abramowitz JS, Storch EA, McKay D, et al. The obsessivecompulsive spectrum: a critical review. In: McKay D,
Abramowitz JS, Taylor S, Asmundson G, editors. Current
Perspectives on the Anxiety Disorders: Implications for DSM-V
and Beyond. New York: Springer; 2009:329352.
28. Taylor S, Asmundson G, Abramowitz JS, McKay D. Classification of anxiety disorders for DSM-V and ICD-11: issues,
proposals, and controversies. In: McKay D, Abramowitz JS,
Taylor S, Asmundson G, editors. Current Perspectives on the
Anxiety Disorders: Implications for DSM-V and Beyond.
New York: Springer; 2009:481511.
29. Crits-Christoph P, Barber JP. Psychological treatments for
personality disorders. In: Nathan PE, Gorman JM, editors.
A Guide to Treatments That Work. 3rd ed. New York: Oxford
University Press; 2007:641658.
30. Pauls DL. The genetics of obsessivecompulsive disorder: a
review. Dialogues Clin Neurosci 2010;12:149163.
31. Hemmings S, Stein DJ. The current status of association studies
in obsessivecompulsive disorder. Psychiatr Clin North Am
2006;29:411444.
32. Cath DC, van Grootheest DS, Willemsen G, et al. Environmental factors in obsessivecompulsive behavior: evidence from
discordant and concordant monozygotic twins. Behav Genet
2008;38:108120.
33. Salkovskis PM, Shafran R, Rachman S, Freeston MH. Multiple
pathways to inflated responsibility beliefs in obsessional problems: possible origins and implications for therapy and research.
Behav Res Ther 1999;37:10551072.

Depression and Anxiety