Phenotypic and genetic differentiation of anxiety-related behaviors in middle childhood.
ABSTRACT Anxiety-related behaviors (ARBs) are commonly observed during typical development, yet few studies have investigated their etiology in middle childhood. This study aimed to examine both the phenotypic and genetic differentiation of ARB subtypes within the general population at age 7 and 9. It constituted a follow-up to an earlier study of ARBs in preschool children.
We investigated the phenotypic structure of ARBs in a large population-based twin sample, comprising 7,834 twin pairs at age 7 and 3,644 twin pairs at age 9. Quantitative genetic modeling techniques were then used to determine the relative influences of genetic and environmental factors upon different types of ARB and upon the covariation between them.
Factor analysis supported the presence of five ARB factors at both ages: negative cognitions, negative affect, fear, obsessive-compulsive behaviors, and social anxiety. Multivariate genetic analyses revealed significant genetic effects and a small but significant influence of shared environment for all ARB subtypes. There was a moderate level of genetic specificity for each subtype as well as some shared genetic effects. Shared environmental influences correlated highly across all types of ARB, whereas nonshared environmental effects were largely subtype specific.
The current results suggest that ARBs can be differentiated both phenotypically and genetically within middle childhood, with subtypes reflecting symptom groupings of diagnosable disorders but also aspects of temperament. Although some etiological risk factors lead to a generalized vulnerability to anxiety, others may serve to differentiate between different types of ARBs.
- SourceAvailable from: Alisa N Almas[Show abstract] [Hide abstract]
ABSTRACT: Anxiety disorders are prevalent throughout childhood and adolescence. As such, identifying the factors and mechanisms that precede, maintain, or exacerbate anxiety disorders is essential for the development of empirically based prevention and intervention programs. The current review focuses on child temperament (i.e., behavioral inhibition) and the child's environment, including parenting, childcare, and peer relationships, as these factors have been linked to internalizing problems and anxiety diagnoses. Research programs are needed that examine the associations between the environment and anxiety in temperamentally at-risk populations. In order to be successful, early intervention and prevention programs require a more detailed analysis of the interplay between various environmental contexts, both distal and proximal to the child, and the child's temperamental reactivity to novelty and threat. Furthermore, conducting these investigations across multiple levels of analysis in large-scale, longitudinal samples would be an important addition to the literature on the developmental psychopathology of anxiety.Journal of Child Psychology and Psychiatry 02/2010; 51(4):497-517. · 5.42 Impact Factor
- PLoS ONE 01/2013; 8(6). · 3.53 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: We aimed to characterize multiple psychotic experiences, each assessed on a spectrum of severity (ie, quantitatively), in a general population sample of adolescents. Over five thousand 16-year-old twins and their parents completed the newly devised Specific Psychotic Experiences Questionnaire (SPEQ); a subsample repeated it approximately 9 months later. SPEQ was investigated in terms of factor structure, intersubscale correlations, frequency of endorsement and reported distress, reliability and validity, associations with traits of anxiety, depression and personality, and sex differences. Principal component analysis revealed a 6-component solution: paranoia, hallucinations, cognitive disorganization, grandiosity, anhedonia, and parent-rated negative symptoms. These components formed the basis of 6 subscales. Correlations between different experiences were low to moderate. All SPEQ subscales, except Grandiosity, correlated significantly with traits of anxiety, depression, and neuroticism. Scales showed good internal consistency, test-retest reliability, and convergent validity. Girls endorsed more paranoia, hallucinations, and cognitive disorganization; boys reported more grandiosity and anhedonia and had more parent-rated negative symptoms. As in adults at high risk for psychosis and with psychotic disorders, psychotic experiences in adolescents are characterized by multiple components. The study of psychotic experiences as distinct dimensional quantitative traits is likely to prove an important strategy for future research, and the SPEQ is a self- and parent-report questionnaire battery that embodies this approach.Schizophrenia Bulletin 09/2013; · 8.80 Impact Factor
DEPRESSION AND ANXIETY 26:316–324 (2009)
PHENOTYPIC AND GENETIC DIFFERENTIATION
OF ANXIETY-RELATED BEHAVIORS IN MIDDLE
Victoria Hallett, MSc.,1?Angelica Ronald, Ph.D.,2Fruhling Rijsdijk, Ph.D.,1and Thalia C Eley, Ph.D.1
Background: Anxiety-related behaviors (ARBs) are commonly observed during
typical development, yet few studies have investigated their etiology in middle
childhood. This study aimed to examine both the phenotypic and genetic
differentiation of ARB subtypes within the general population at age 7 and 9. It
constituted a follow-up to an earlier study of ARBs in preschool children.
Methods: We investigated the phenotypic structure of ARBs in a large
population-based twin sample, comprising 7,834 twin pairs at age 7 and
3,644 twin pairs at age 9. Quantitative genetic modeling techniques were then
used to determine the relative influences of genetic and environmental factors
upon different types of ARB and upon the covariation between them. Results:
Factor analysis supported the presence of five ARB factors at both ages: negative
cognitions, negative affect, fear, obsessive–compulsive behaviors, and social
anxiety. Multivariate genetic analyses revealed significant genetic effects and a
small but significant influence of shared environment for all ARB subtypes.
There was a moderate level of genetic specificity for each subtype as well as some
shared genetic effects. Shared environmental influences correlated highly across
all types of ARB, whereas nonshared environmental effects were largely subtype
specific. Conclusions: The current results suggest that ARBs can be
differentiated both phenotypically and genetically within middle childhood, with
subtypes reflecting symptom groupings of diagnosable disorders but also aspects
of temperament. Although some etiological risk factors lead to a generalized
vulnerability to anxiety, others may serve to differentiate between different types
of ARBs. Depression and Anxiety 26:316–324, 2009.
r 2009 Wiley-Liss, Inc.
Key words: anxiety; anxiety disorders; child; twin; genetics; environment
Anxiety-related behaviors (ARBs) are commonly
observed during typical child development, with over
70% of children reporting symptoms of anxiety during
middle childhood.Clinical anxiety disorders are
thought to affect up to 30% of children and
adolescents, although in most cases, ARBs diminish
during development.Despite their prevalence, the
phenotypic structure and developmental etiology of
Published online 4 February 2009 in Wiley InterScience (www.
Received for publication 6 February 2008; Revised 25 September
2008; Accepted 30 September 2008
Contract grant sponsor: UK Medical Research Council.
Abbreviations: ARB, anxiety-related behavior; OCB, obsessive–
compulsive behavior; TEDS, Twins Early Development Study.
?Correspondence to: Victoria Hallett, Social, Genetic, and
Developmental Psychiatry Centre, Institute of Psychiatry, De
Crespigny Park, Box P080, London, SE5 8AF, UK.
1Social, Genetic and Developmental Psychiatry Centre, In-
stitute of Psychiatry, King’s College, London, UK
2Brain and Cognitive Development Centre, School of Psychol-
ogy, Birkbeck College, London, UK
rrrr 2009 Wiley-Liss, Inc.
subclinical ARBs are relatively poorly understood. This
study incorporated phenotypic and quantitative genetic
analyses to examine the relative contributions of
genetic and environmental influences upon different
types of ARB during middle childhood.
PHENOTYPIC STUDIES OF ANXIETY
Previous phenotypic studies have used a number of
approaches to differentiate between subtypes of ARBs
within earlyand middle childhood.[4,5]First, sub-
clinical symptoms of anxiety have often been divided
into the same categories as anxiety disorders, including
separation anxiety, social phobia, and generalized
anxiety.[6,7]Secondly, other studies have suggested that
anxiety symptoms are more closely associated with
aspects of temperament including behavioral inhibi-
affectivity.[11,12]Some studies have suggested that the
structure of ARBs remains consistent during child-
hood,whereas others have proposed changes in the
presentation of anxious and depressive symptoms as
emotions and cognitions develop.[13,14]
QUANTITATIVE GENETIC STUDIES OF
Previous twin studies have analyzed the etiological
risk factors for specific anxiety symptoms including
separation anxiety,[15,16]social anxiety,[3,7]fears/pho-
bias,[17,18]and obsessive–compulsive behaviors (OCB).[3,19]
In general, they have reported that ARBs are moder-
ately to highly heritable, with a significant effect of
both nonshared environmental factors (which serve to
make children in the same family more different) and
shared environmental factors (which make children in
the same family more similar).Quantitative genetic
studies of depression have generally indicated a
moderate heritability, with a negligible role of shared
Previous research has reported a high degree of
covariation between different ARB subtypes in children
and adolescentsand significant heterotypic conti-
nuity between them. This suggests that children may
experience different ARBs at different times during
development.[24,25]However, few multivariate studies
have investigated the genetic and environmental
relationship between different ARB groups within the
First, a preschool investigation of 4,564 twin pairs
found modest levels of genetic overlap between five
subtypes of ARB: general distress, separation anxiety,
fear, OCB, and shyness.Shared environmental
influences were largely similar across subtypes, whereas
nonshared environmental factors were more variable
ARBs in middle childhood, focusing on the covariation
between specific phobia, separation anxiety, and
social phobia at age 6 in a sample at high risk for
individuals who met criteria for a ‘‘symptom syn-
drome,’’ fulfilling the criteria of the DSM-IV, regard-
lessofthe degree of
shared environmental overlap was observed between
separation anxiety along with a significant familial
correlation between specific phobia and social phobia
With respect to middle childhood and adolescence,
one twin study investigated the association between
symptoms of overanxious disorder, simple phobias,
separation anxiety, and depression in girls aged 8–13
and 14–17.The results showed shared genetic
and depression in middle childhood, with some
additional overlap in their shared environmental risk
The current analysis represented a follow-up to the
preschool study,investigating symptoms in the same
sample at age 7 and 9. We hypothesized that ARBs in
middle childhood would cluster into subtypes reflect-
ing both clinical anxiety disorders and also aspects of
temperament. In particular, we predicted both overlap
and specificity in the etiological influences affecting
each ARB subtype, including a moderate degree of
genetic independence between the ARB subtypes and a
higher level of overlap between their shared-environ-
Data came from the Twins Early Development Study (TEDS), a
population-representative sample of all twins born in England and
Wales between the years 1994 and 1996.The sample included 7,834
twin pairs at age 7 and 3,644 pairs at age 9 (see Table 1 for further
details). Twin zygosity was determined using parent ratings of
similarity, supplemented by DNA genotyping.
ARBs were measured using 25 parental questionnaire items at age
7 and 9 (see Table 2), answered using a three-point Likert scale
(never:0; sometimes:1, often:2). Thirteen items, presented in bold,
(ARBQ),[3,26]including items from the Emotional Scale of the
Strengths and Difficulties Questionnaire.Twelve additional items
were included to assess symptoms of depression, fear, and OCB. The
items had good construct validity and a high internal consistency
(0.85). Of note, the measure serves to assess individual differences in
ARBs within the general population. It does not constitute a
diagnostic tool for internalizing disorders, but shows strong
similarities with screening questionnaires such as the Screen for
Child Anxiety-Related Emotional Disordersand the Spence
Children’s Anxiety Scale.
317Research Article: Differentiating Anxiety Symptoms at 7–9 Years
Depression and Anxiety
Principal components analysis (PCA) was used to examine the
factor structure of the ARB items at both ages. An oblique rotation
was used (Direct Oblimin command) and the number of factors was
determined by using eigenvalues greater than 1. Composite scores
were calculated using a scale mean replacement technique; with one
missing item allowed per scale. Transformations were performed to
minimize skew in the totals (optimizing minimal skew using the
lnskew0 command in Stata 9.1).
QUANTITATIVE GENETIC ANALYSES
The twin method allows us to determine the relative contribution
of three latent components to each ARB subscale: additive genetic
influences (A), shared environment (C), and non-shared environ-
TABLE 2. Means and factors loadings from principal components analysis
1 Often makes comments critical of him/herself
2 Tends to blame him/herself
3 Has low self-confidence
4 Many worries, often seems worried
5 Anxious that bad things will happen
6 Asks for reassurance that s/he is OK
7 Often unhappy, downhearted or tearful
8 Does not enjoy him/herself
9 Seems keyed up, on edge, tense
10 Does something over and over again
11 Complains or whines
12 Often complains of headaches, stomach
aches, or sickness
13 Is often extremely upset or distressed when
14 Has twitches/mannerisms or tics of the
15 Is afraid of small closed spaces, heights,
water, or the dark
16 Many fears, easily scared
17 Is afraid of animals or insects
18 Resists sleeping alone
19 Is afraid of medical procedures such as going to
the doctor or dentist
20 Tends to check that things are done exactly
22 Tends to be shy or timid
23 Is afraid in social situations
24 Takes a long time to warm to strangers
25 Nervous or clingy in new situations, often loses
0.31 (age 9)
0.44 (age 9)
0.44 (age 9)
0.35 (age 7) 0.31 (age 7)
0.38 (age 7)
0.20 (age 7)
0.32/0.320.32 (age9)0.32 (age 7)
0.32/0.380.31 (age 9)
0.35 (age 9)
0.43 (age 9)
Values are provided for the 7-year and 9-year results respectively (age7/age9). Items were free to load upon more than one factor, however, only
the highest factor loadings and other loadings40.3 are presented here. Values in bold represent the highest loading for each item. The item titles
in bold refer to items that were included in the preschool study.
TABLE 1. The sample size at age 7 and 9, separated by sex and zygosity
Number of twin pairs in each sex and zygosity group (after exclusions)
Number of excluded pairsa
MZMDZMMZF DZF DZOSFinal total
MZM, monozygotic males; DZM, dizygotic males; MZF, monozygotic females; DZF, dizygotic females; DZOS, dizygotic opposite sex pairs.
aTwin pairs were excluded due to extreme prenatal or perinatal difficulties or the presence of a severe medical condition in one or both of the twins.
318 Hallett et al.
Depression and Anxiety
mental factors (E). It relies on the assumption that monozygotic (MZ)
twins share all of their genetic information, whereas dizygotic (DZ)
pairs share on average only 50% of their genes. MZ and DZ pairs are
assumed to be equally similar with regard to shared environmental
influences. Further details of this method are provided elsewhere.
Pearson’s correlations were calculated, showing the association
between Twin 1 and Twin 2 for each trait (see Table 3). This allowed
estimation of A, C, and E parameters and identification of
dominance, contrast or sex effects within the data. Univariate model
fitting was carried out to estimate these parameters more accurately,
using the structural equation modeling package Mx.Models
were fit to the raw data using age and sex regressed residual
subscale totals. The univariate results are available from the
first author on request. Phenotypic correlations were also calculated
to determine the overlap between ARB subtypes within an individual.
Cross-trait cross-twin intraclass correlations were also calculated,
establishing the relationship between Trait 1 in Twin 1 and Trait 2 in
TABLE 3. Phenotypic correlations and Pearson’s twin correlations
Negative cognitionsNegative affect Fear Social anxietyOCBa
Descriptive statistics based upon raw scores (age 7/age 9)
No of items
Means (standard deviation)
6/6 7/65/52/2 4/4
Phenotypic correlationsa(age 7 correlations are presented above the diagonal; age 9 correlations below the diagonal)
MZ twin correlations
0.39 (0.36–0.42)0.30 (0.28–0.32)0.30 (0.26–0.34)
Negative affect 0.32 (0.30–0.34)
0.28 (0.25–0.31) 0.28 (0.26–0.32)
Fear0.25 (0.23–0.27)0.24 (0.22–0.26)
Social anxiety 0.25 (0.22–0.28)0.22 (0.19–0.25)0.26 (0.23–0.29)
DZ twin correlations
0.25 (0.22–0.28) 0.21 (0.17–0.23) 0.20 (0.17–0.23)
Negative affect 0.22 (0.20–0.24)
0.22 (0.19–0.25)0.21 (0.21–0.24)
Fear0.18 (0.16–0.20)0.21 (0.19–0.23)
Social anxiety0.14 (0.12–0.16) 0.17 (0.15–0.19)0.17 (0.15–0.20)
Phenotypic correlations: age 7 correlations are presented above the diagonal, age 9 correlations below the diagonal. These refer to the relationship
between Trait 1 and Trait 2 within the same individual. Twin Correlations: in the lower half of the table, values in bold along the diagonals
represent the univariate Pearson’s twin correlations (age7/age9), obtained using a constrained saturated script in Mx. Confidence intervals (95%)
for all correlations are displayed in parentheses. Cross-twin cross-trait correlations are presented below the diagonal (at age 7) and above the
diagonal (age 9).
aThe OCB subscale was excluded from multivariate analyses and so cross-twin cross-trait correlations are not presented here.
319Research Article: Differentiating Anxiety Symptoms at 7–9 Years
Depression and Anxiety
Figure 1 shows the three multivariate models that were tested. The
correlated factors solution incorporated different parameter estimates
for each subscale with correlations between the genetic (rG),
nonshared (rE), and shared (rC) environmental influences. These
correlations range from 0 (no overlap between factors) to 1 (complete
overlap)). The independent pathway model and common factor
models represent more restricted representations of the data, testing
the hypothesis that there are common genetic and environmental
influences underlying the different subscales, see Neale and
Cardonfor further explanation.
Table 2 summarizes the results from the PCA. Six
underlying factors had eigenvalues greater than 1. The
sixth factor contained only two items with low loadings
and consequently, we decided to constrain the PCA to
include only five factors. Items were allowed to load on
more than one factor. The highest factors loadings are
presented in Table 2, along with additional loadings
greater than 0.3. The full PCA is available from the
first author on request. The same factor structure was
observed at age 7 and 9 years, with only small
differences in the factor loadings at each age. The
factors were named: negative cognitions, negative
affect, fear, OCB, and social anxiety. There were
significant yet modest correlations between these
factors (0.11–0.29). Negative cognitions accounted for
the greatest proportion of the variance at age 7 (21%)
and age 9 (23%).
Composite subscale scores were created, using items
with a loading of greater than 0.3. As a result, ‘‘Has
twitches/mannerisms or tics of the hands/body’’ was
excluded at both ages and ‘‘Complains of headaches/
stomach aches and sickness’’ was excluded at age 7.
Summary statistics and subscale reliabilities are pro-
vided in Table 3. Females had a higher mean score for
fear and social anxiety at age 7 and negative affect and
social anxiety at age 9.
The univariate twin correlations (shown in bold
in the bottom half of Table 3) indicated strong
additive genetic influences with some involvement of
shared environment. The univariate models did not
indicate sex differences in the data and consequently
sex differences were not incorporated into the multi-
variate models. The high, positive DZ correlations
were not indicative of contrast effects or genetic
dominance. The OCB composite score showed a
categorical distribution and did not appear to give a
comprehensive representation of OCB symptoms.
Consequently, this subscale was excluded from multi-
MULTIVARIATE GENETIC ANALYSIS
Table 3 presents the phenotypic correlations at age 7
(0.17–0.43) and 9 (0.18–0.48). The highest phenotypic
correlation was between negative cognitions and
negative affect at both ages. The lower half of Table 2
also shows the pairwise cross-trait cross-twin correla-
tions. These were indicative of genetic influence upon
the covariation between subscales.
Fit statistics from the multivariate models are
presented in Table 4. Lower AIC values indicated a
better fit to the data.
The correlated factor solution provided a signifi-
cantly better fit than either the independent pathway or
common factor models at both ages. The parameter
estimates from this model are presented in Table 5.
In Table 5, the values in bold along the diagonals
represent the proportions of the variance of each scale
that were attributable to additive genetic factors
(h2: top third of the table), shared environmental
influences (c2: middle third) and nonshared environ-
ment (e2: bottom third). Negative affect had the lowest
heritability estimate at both ages (0.50 at age 7, 0.46
at 9). The highest heritability estimates were found
for social anxiety at age 7 (0.61) and fear at age 9 (0.58).
Nonshared environment accounted for around a
third of the variance of each scale. Shared environment
accounted for the lowest percentage of the variance for
each scale (0.09–0.21 at age 7 and 0.11–0.23 age 9).
The values below the diagonal represent the correla-
tions between the genetic influences (top third of the
table), shared environmental factors (middle third), and
Figure 1. (a) Correlated factor solution of the Cholesky decom-
position. (b) Independent pathway model. (c) Common factor
model.?P: Common Factor.
320Hallett et al.
Depression and Anxiety
nonshared environmental influences (bottom third)
upon each scale. The highest genetic overlap was
between Negative Affect and Negative Cognitions
Scales (0.38 at age 7, 0.53 at age 9) and the lowest
between negative affect and both social anxiety (0.13
at 7, 0.25 at 9) and fear (0.15 at age 7, 0.17 at age 9).
Conversely, the shared environmental correlations
were high across all the subscales, particularly between
social anxiety and both fear (0.98 at 7, 0.74 at age 9)
and negative affect (0.98 at age 7, 0.78 at age 9). In
general, there were lower levels of overlap between
nonshared environmental factors (0.19–0.36 at age 7
and 0.14–0.42 at age 9).
Finally, the values above the diagonals in Table 5
refer to the proportions of the phenotypic correlation
between two subscales that were attributable to shared
mental factors (middle third), and nonshared environ-
mental influences (bottom third). A large proportion
of the phenotypic overlap between negative cognitions
and each of the other subscales (42–57%) was
attributable to genetic factors. Conversely, the correla-
tion between negative affect and both social anxiety
and fear was most strongly influenced by shared
environmental factors, accounting
46–57% of the phenotypic overlap. The smallest
proportion of thephenotypic
attributable to nonshared environmental influences
This study found that ARBs could be divided into
distinct, yet correlated phenotypic subtypes in middle
childhood, influenced by common and specific etiolo-
gical factors. We carried out phenotypic and quantita-
tive genetic analyses at two time points, providing a
follow-up to a previous analysis of the same sample at a
Five ARB subtypes were identified: negative cogni-
tions, negative affect, fear, OCB, and social anxiety.
Significant genetic influences contributed to individual
differences for each subtype. We also investigated the
different scales, revealing both shared and specific
genetic and environmental influences.
between the etiological factors affecting
PHENOTYPIC STRUCTURE OF ANXIETY-
The phenotypic structure of ARBs reflected clinical
subtypes of anxiety disorder (including fear and social
anxiety) but also aspects of temperament (negative
cognition and negative affect). The same factor
structure was observed at both ages 7 and 9, supporting
previous reports of stability in these symptoms over
time.The significant phenotypic correlations be-
tween the subscales suggested that typically developing
children might experience more than one type of ARB
concurrently. This is in keeping with clinical studies,
which have reported high rates of comorbidity between
Two interesting comparisons can be made with the
earlier preschool study.First, separation anxiety
traits no longer formed a distinguishable ARB subscale
in middle childhood. Although separation anxiety
behaviors are normal and adaptive in early develop-
ment, these symptoms may become increasingly
Our result was in contrast to previous studies in middle
childhood, which have
anxiety as a distinct subscale within population-based
Secondly, negative cognitions and negative affect
formed distinct subscales in middle childhood. This
reflects the inclusion of more varied ARB items at ages
7 and 9, but also points to the development of negative
thinking and introspection in this age group. The
negative cognitions scale shows similarities with the
personality trait of ‘‘neuroticism,’’ which has been
shown to share phenotypic and genetic influences with
both anxiety disorders and depression.[35,36]Similarly,
previous studies have proposed that ‘‘negative affectiv-
ity’’ may constitute a commonality between anxiety and
TABLE 4. Multivariate fit statistics at age 7 and 9
Independent pathway model
Common factor model
Independent pathway model
Common factor model
Values on the left of the table show the difference between each of the multivariate models and the saturated model.
aIndependent pathway and common factor models are compared to the Cholesky decomposition on the right of the table.
321Research Article: Differentiating Anxiety Symptoms at 7–9 Years
Depression and Anxiety
The moderate to high heritabilities of ARB subtypes
in this study supported the findings of previous twin
studies in middle childhood.In line with previous
multivariate studies, our results showed both shared
and specific genetic effects upon the ARB sub-
scales.[3,18,26]There was a high level of genetic
independence between the subscales and they were
not influenced by a single underlying factor. However,
despite this genetic specificity, the negative cognitions
subscale did correlate reasonably highly with the other
subscales, particularly negative affect. This suggests
that the genes associated with negative cognitions may
confer a more general vulnerability to anxiety sympto-
matology as a whole.
The ‘‘generalist genes’’ theory has suggested that
certain genes may influence a child’s predisposition to a
broad range of developmental disorders.Conver-
sely, environmental factors may serve to differentiate
the particular presentation of anxious or depressed
symptoms. However, our results suggest that genes are
involved in both general predisposition to anxiety and
the specification of symptom subtypes. These findings
are in agreement with the preschool study, and
previous work that has differentiated psychiatric
disorders on the basis of their genetic influences.
The heritability of anxiety subtypes appears to
remain fairly consistent during middle childhood.
However, our findings cannot speak to whether the
same genes are involved at age 7 and age 9. Recent
findings suggest that etiological influences on child-
namic.A study by Kendler et al., showed that
certain genes are associated with anxiety symptoms
throughout development, whereas others are specific to
particular age groups.Further longitudinal investi-
gation is needed to determine whether there is
continuity in the genetic influences upon ARBs within
TABLE 5. The correlated factors solution of the Cholesky decomposition
Age 7Age 9
Genetic influences (A)
Shared environment (C)
Nonshared environment (E)
Values in bold on the diagonal: the proportion of the variance of each scale attributable to genetic influences (h2: top third) shared environmental
influences (c2: middle third), and nonshared environmental influences (e2: bottom third). Below the diagonal: values represent the genetic
correlations (top third), shared environmental correlations (middle third), and nonshared environmental correlations (bottom third) between the
variables. Above the diagonal: values show the percentage of the phenotypic correlation accounted for by additive genetic factors (top third),
shared environment (middle third), and nonshared environment (bottom third).
322 Hallett et al.
Depression and Anxiety
Our results suggest that shared environment is
influential in the development and maintenance of
ARBs in middle childhood, particularly fear and
negative affect. This finding is in keeping with two
previous multivariate studies of ARBs in the general
population, which reported modest but significant
shared environmental influences.[3,18]It also supports
the multivariate study of symptom syndromes in a
population-based sample, which showed significant
shared environmental influence on the overlap between
specific phobia and separation anxiety.
The role of shared environment within anxiety is
interesting, as these factors are rarely implicated in
adult anxiety disorders or other forms of childhood
psychopathology. Aspects of home environment and
parental care may impact upon the security of parent–
child attachment and affect a child’s ability to regulate
emotions effectively.Caregivers can model both
adaptive and maladaptive coping strategies for their
children and constitute an important source of support
and security during times of stress.Interestingly,
although shared environment made the smallest con-
tribution to the variance of each trait individually, these
influences were largely shared among the subtypes. As
a consequence, the identification of shared environ-
mental risk factors could be broadened to a wide range
Effects of nonshared environment were highly
specific for each subscale. Thus, it appears that shared
and nonshared environmental influences may play
different roles in the emergence of ARBs. Although
shared environment may affect the development of
anxiety symptoms in general, nonshared environmental
influences may help to differentiate anxiety subtypes in
genetically predisposed individuals.
Nonshared environmental influences are likely to
include experiences outside the home and at school. A
recent study investigated discordant MZ twin pairs and
suggested that negative school experiences and peer
rejection may constitute child specific influences upon
anxiety at age 7.Furthermore, parents do not always
treat their children in the same way, potentially leading
to differences in the development of anxiety.
STRENGTHS AND LIMITATIONS
The large sample size was a key strength of this
study, providing power to determine a small but
significant influence of shared environment. The
questionnaire items were based upon a previously
published scaleand this study also benefited from
using empirical factor analytic techniques to determine
the structure of these items.
Certain limitations need to be considered. First, the
study relied on parent-rated measures of ARBs. Future
replications would benefit from the inclusion of
multiple informants or observational data to reinforce
our findings. Secondly, there has been also some
attrition of the TEDS sample over timeand there
are some limitations inherent to the twin method.
Thirdly, future studies could aim to measure OCB
symptoms more comprehensively and to investigate
prior suggestions that OCB may be genetically
independent from other ARB subtypes. Finally, it is
possible that the items in our measure did not give an
accurate reflection of separation anxiety symptoms in
middle childhood. Items such as ‘‘resists sleeping
alone’’ may not have been suitable for this age group,
potentially influencing the factor structure. Future
work could assess separation anxiety using a wider
range of age-appropriate items.
Our results suggest that subclincial anxiety symp-
toms are phenotypically distinguishable in middle
childhood, with subtypes reflecting both clinical
aspects of anxiety and aspects of temperament and
remaining consistent across a 2-year period. Although
some genes and shared-environmental factors may
predispose individuals to anxious symptomology, it
seems that others serve to differentiate particular
anxiety subtypes. Further longitudinal follow-ups are
now required to determine whether the same influ-
ences are implicated at different stages in development.
Career Development Award from the UK Medical
Research Council. The authors thank Derek Bolton,
Thomas G. O’Connor, and Sean Perrin, along with all
the families that have participated in TEDS.
Thalia C Eley is funded by a
1. Muris P, Meesters C, Merckelbach H, Sermon A, Zwakhalen S.
Worry in normal children. J Am Acad Child Adolesc Psychiatry
2. Costello EJ, Mustillo S, Erkanli A et al. Prevalence and
development of psychiatric disorders in childhood and adoles-
cence. Arch Gen Psychiatry 2003;60:837–844.
3. Eley TC, Bolton D, O’Connor TG et al. A twin study of anxiety-
related behaviours in pre-school children. J Child Psychol
4. Mellon RC, Moutavelis AG. Structure, developmental course,
and correlates of children’s anxiety disorder-related behavior in a
Hellenic community sample. J Anxiety Disord 2007;21:1–21.
5. Spence SH. Structure of anxiety symptoms among children: a
confirmatory factor-analytic study. J Abnorm Psychol 1997;106:
6. Spence SH, Barrett PM, Turner CM. Psychometric properties of
the Spence Children’s Anxiety Scale with young adolescents.
J Anxiety Disord 2003;17:605–625.
7. Ogliari A, Citterio A, Zanoni A et al. Genetic and environmental
influences on anxiety dimensions in Italian twins evaluated with
the SCARED questionnaire. J Anxiety Disord 2006;20:760–777.
323 Research Article: Differentiating Anxiety Symptoms at 7–9 Years
Depression and Anxiety
8. Gladstone GL, Parker GB, Mitchell PB et al. Relationship
between self-reported childhood behavioral inhibition and life-
time anxiety disorders in a clinical sample. Depress Anxiety
9. Stevenson J, Batten N, Cherner M. Fears and fearfulness in
children and adolescents: a genetic analysis of twin data. J Child
Psychol Psychiatry 1992;33:977–985.
10. Thapar A, McGuffin P. A twin study of antisocial and neurotic
symptoms in childhood. Psychol Med 1996;26:1111–1118.
11. Clark LA, Watson D. Tripartite model of anxiety and depression:
psychometric evidence and taxonomic implications. J Abnorm
12. Turner CM, Barrett PM. Does age play a role in the structure of
anxiety and depression in children and youths? An investigation
of the tripartite model in three age cohorts. J Consult Clin
13. Cole DA, Truglio R, Peeke L. Relation between symptoms of
anxiety and depression in children: a multitrait-multimethod-
multigroup assessment. J Consult Clin Psychol 1997;65:
14. Christensen H, Jorm AF, Mackinnon AJ et al. Age differences in
depression and anxiety symptoms: a structural equation model-
ling analysis of data from a general population sample. Psychol
15. Bolton D, Eley TC, O’Connor TG et al. Prevalence and genetic
and environmental influences on anxiety disorders in 6-year-old
twins. Psychol Med 2006;36:335–344.
16. Feigon SA, Waldman ID, Levy F et al. Genetic and environ-
mental influences on separation anxiety disorder symptoms
and their moderation by age and sex. Behav Genet 2001;31:
17. Lichtenstein P, Annas P. Heritability and prevalence of specific
fears and phobias in childhood. J Child Psychol Psychiatry
18. Silberg JL, Rutter M, Eaves L. Genetic and environmental
influences on the temporal association between earlier anxiety
and later depression in girls. Biol Psychiatry 2001;49:1040–1049.
19. Bolton D, Rijsdijk F, O’Connor TG et al. Obsessive–compulsive
disorder, tics and anxiety in 6-year-old twins. Psychol Med
20. Gregory AM, Eley TC. Genetic influences on anxiety in
children: what we’ve learned and where we’re heading. Clin
Child Fam Psychol Rev 2007;10:199–212.
21. Eley TC. Depressive symptoms in children and adolescents:
etiological links between normality and abnormality: a research
note. J Child Psychol Psychiatry 1997;38:861–865.
22. Rice F, Harold GT, Thapar A. Assessing the effects of age, sex
and shared environment on the genetic aetiology of depression in
childhood and adolescence. J Child Psychol Psychiatry 2002;
23. Lieb R. Anxiety disorders: clinical presentation and epidemiol-
ogy. Handb Exp Pharmacol 2005;169:405–432.
24. Ferdinand RF, Dieleman G, Ormel J et al. Homotypic versus
heterotypic continuity of anxiety symptoms in young adolescents:
evidence for distinctions between DSM-IV subtypes. J Abnorm
Child Psychol 2007;35:325–333.
25. Gregory AM, Caspi A, Moffitt TE et al. Juvenile mental health
histories of adults with anxiety disorders. Am J Psychiatry
26. Eley TC, Rijsdijk FV, Perrin S et al. A multivariate genetic
analysis of specific phobia, separation anxiety and social phobia
in early childhood. J Abnorm Child Psychol 2008;36:839–848.
27. Oliver BR, Plomin R. Twins’ Early Development Study (TEDS):
a multivariate, longitudinal genetic investigation of language,
cognition and behavior problems from childhood through
adolescence. Twin Res Hum Genet 2007;10:96–105.
28. Goodman R. Psychometric properties of the strengths and
difficulties questionnaire. J Am Acad Child Adolesc Psychiatry
29. Birmaher B, Khetarpal S, Brent D et al. The Screen for Child
Anxiety Related Emotional Disorders (SCARED): scale con-
struction and psychometric characteristics. J Am Acad Child
Adolesc Psychiatry 1997;36:545–553.
30. Plomin R, DeFries JC, McClearn G et al. Behavioural Genetics.
New York: Worth Publishers; 2001.
31. Neale MC, Miller MB. The use of likelihood-based confidence
intervals in genetic models. Behav Genet 1997;27:113–120.
32. Neale MC, Cardon LR. Methodology for Genetic Studies of
Twins and Families. NATO ASI Series D: Behavioral and Social
Sciences, vol. 67. Kluwer Academic Publishers, B.V., Dordrecht,
The Netherlands; 1992.
33. Goisman RM, Goldenberg I, Vasile RG et al. Comorbidity of
anxiety disorders in a multicenter anxiety study. Compr
34. Kearney CA, Sims KE, Pursell CR et al. Separation anxiety
disorder in young children: a longitudinal and family analysis.
J Clin Child Adolesc Psychol 2003;32:593–598.
35. Kendler KS, Gardner CO, Gatz M et al. The sources of co-
morbidity between major depression and generalized anxiety
disorder in a Swedish national twin sample. Psychol Med 2007;
36. Hettema JM, Neale MC, Myers JM et al. A population-based
twin study of the relationship between neuroticism and inter-
nalizing disorders. Am J Psychiatry 2006;163:857–864.
37. Anderson ER, Hope DA. A review of the tripartite model for
understanding the link between anxiety and depression in youth.
Clin Psychol Rev 2008;28:275–287.
38. Eley TC. General genes: a new theme in developmental
psychopathology. Curr Dir Psychol Sci 1997;6:90–95.
39. Kendler KS, Walters EE, Neale MC et al. The structure of the
genetic and environmental risk factors for six major psychiatric
disorders in women. Phobia, generalized anxiety disorder, panic
disorder, bulimia, major depression, and alcoholism. Arch Gen
40. Bergen SE, Gardner CO, Kendler KS. Age-related changes in
heritability ofbehavioral phenotypes
young adulthood: a meta-analysis. T win Res Hum Genet 2007;10:
41. Kendler KS, Gardner CO, Lichtenstein P. A developmental twin
study of symptoms of anxiety and depression: evidence for
42. O’Connor TG, Croft CM. A twin study of attachment in
preschool children. Child Dev 2001;72:1501–1511.
43. Bogels SM, Brechman-Toussaint ML. Family issues in child
anxiety: attachment, family functioning, parental rearing and
beliefs. Clin Psychol Rev 2006;26:834–856.
44. Asbury K, Dunn J, Plomin R. The use of discordant MZ twins to
generate hypotheses about nonshared environmental influence
on anxiety in middle childhood. Soc Dev 2008;15:564–570.
45. Ronald A, Happe F, Plomin R. A twin study investigating the
genetic and environmental aetiologies of parent, teacher and
child ratings of autistic-like traits and their overlap. Eur Child
Adolesc Psychiatry 2008.
attenuation. Psychol Med
324 Hallett et al.
Depression and Anxiety