Behavioral Inhibition in Preschool Children At Risk Is a
Specific Predictor of Middle Childhood Social Anxiety: A Five-
Dina R. Hirshfeld-Becker, PhD, Joseph Biederman, MD, Aude Henin, PhD,
Stephen V. Faraone, PhD, Stephanie Davis, BA, Kara Harrington, MS, Jerrold F. Rosenbaum, MD
ABSTRACT: Objective: Behavioral inhibition (BI) to the unfamiliar represents the temperamental tendency to
exhibit fearfulness, reticence, or restraint when faced with unfamiliar people or situations. It has been hypoth-
esized to be a risk factor for anxiety disorders. In this prospective longitudinal study, we compared the
psychiatric outcomes in middle childhood of children evaluated at preschool age for BI. Method: The baseline
sample consisted of 284 children ages 21 months to 6 years, including offspring at risk for anxiety (children of
parents with panic disorder and/or major depression) and comparison offspring of parents without mood or
major anxiety disorders. They had been assessed for BI using age-specific laboratory protocols. We reassessed
215 of the children (76.5%) at 5-year follow-up at a mean age of 9.6 years using structured diagnostic inter-
views. Results: BI specifically predicted onset of social anxiety. The rate of lifetime social anxiety (DSM-IV social
phobia or DSM-III-R avoidant disorder) was 28% versus 14% (odds ratio [OR] ? 2.37; 95% confidence interval
[CI]: 1.10–5.10) in inhibited versus noninhibited children. BI significantly predicted new onset of social phobia
among children unaffected at baseline (22.2% vs 8.0% in inhibited versus noninhibited children (OR ? 3.15,
95% CI: 1.16–8.57). No other anxiety disorders were associated with BI. Conclusion: BI appears to be a
temperamental antecedent to subsequent social anxiety in middle childhood. Children presenting with BI
should be monitored for symptoms of social anxiety and may be good candidates for preventive cognitive
(J Dev Behav Pediatr 28:225–233, 2007) Index terms: behavioral inhibition, social anxiety, high risk, childhood anxiety.
Identifying childhood precursors to anxiety disorders
can inform both our understanding of the cause of these
disorders and the selection of candidates for preventive
intervention. One precursor that has been studied in
detail is behavioral inhibition (BI) to the unfamiliar.1BI
represents a temperamental tendency, observable as early
as toddlerhood, to show fearfulness and restraint in re-
sponse to novel people and stimuli. Inhibited toddlers
respond to novelty with fretting, clinging to mother, and
avoidance, whereas older preschoolers respond with ret-
icence and somber affect. Pilot studies conducted by our
group in the late 1980s suggested that BI represented a
marker of general anxiety proneness. We found elevated
rates of BI in offspring of parents with panic disorder,2
elevated rates of anxiety disorders in parents of toddlers
selected as inhibited,3and elevated rates of anxiety dis-
orders in inhibited children.4,5
This work led us to launch a longitudinal study of
offspring of parents with panic disorder contrasted with
offspring of parents with depression and of parents with-
out anxiety or mood disorders.6,7At baseline, children
ages 21 months to 6 years underwent laboratory assess-
ments of BI (N ? 284), and then were assessed by diag-
nostic interview (wave 1), either concurrently (for the
older children) or when they reached the age of 5 to 6
years. Wave 1 data8suggested that the risk associated
with BI might be specific to social anxiety. At a mean age
of 6 years, the 216 inhibited children old enough to be
assessed psychiatrically had significantly higher rates of
social anxiety than noninhibited children (17% vs 5%
meeting criteria for social phobia or DSM-III-R childhood
avoidant disorder), yet the two groups did not differ on
any other anxiety disorders. These results accorded with
other longitudinal studies9,10that also found specific links
to social anxiety. They were also consistent with “bottom-
up” studies suggesting elevated rates of social anxiety in
parents of inhibited or shy children.3,11,12
Social phobia is an anxiety disorder characterized by
fear and avoidance of social situations and of scrutiny by
others, that causes marked distress or interference with
social or academic functioning.13To be diagnosed in a
child, it must persist for 6 months or longer, and the child
must show appropriate relatedness with family members
or familiar peers. It afflicts an estimated 7% to 13% of
adults14,15with estimated prevalences in childhood and
adolescence of 1.6% to 5.1%.16Social phobia may be
characterized by fears of discrete performance situations
Harvard Medical School/Massachusetts General Hospital, Boston, Massachusetts.
This study was supported by NIMH grant R01 MH047077-13.
Received March 2006; accepted August 2006.
Address for reprints: Dina R. Hirshfeld-Becker, Ph.D., Massachusetts General Hos-
pital Clinical and Research Program in Pediatric Psychopharmacology, 185 Alewife
Copyright © 2007 Lippincott Williams & Wilkins
Vol. 28, No. 3, June 2007
(e.g., speaking or writing in front of the class, participat-
ing in sports or musical performances), or by general
fears of social interaction. The modal age of onset is in
mid-adolescence,17,18but in many cases, the onset is ear-
lier. In DSM-III and III-R, childhood social anxiety was
diagnosed as “avoidant disorder,” characterized by in-
tense fear and avoidance of unfamiliar peers, without
reference to social-evaluative concerns; in DSM-IV,
avoidant disorder became subsumed under social phobia.
Generalized social phobia is thought to have an earlier
age of onset (mean age, 11 years) and to be more highly
inherited.19Children with social phobia often fear meet-
ing new children, talking to unfamiliar people, attending
parties, and interacting with groups of peers. The disor-
der is associated with social isolation, academic dysfunc-
tion, and in some cases school refusal, as well as with the
development of comorbid depression and substance
abuse.20,21It can be chronic if untreated22and can lead to
difficulty establishing relationships and occupational dys-
Other studies, however, have suggested that BI is also
linked more broadly to panic disorder and depres-
sion.2,6,23,24Studies of offspring of parents with panic and
depression have found elevated rates of BI.6,23,25Retro-
spective studies of adults with panic disorder and depres-
sion have found BI to be elevated in both,24and concur-
rent studies of young adolescents have found self-
reported BI tobe associated
depression, worry, panic disorder, generalized anxiety
disorder, separation anxiety disorder, and obsessive-com-
pulsive disorder. One possibility is that BI is associated
with social anxiety in early childhood and with other
outcomes in later childhood or adolescence.
To further explore the specificity of the risk conferred
by BI, we examined its psychopathologic outcomes in
our longitudinal sample of children at risk at 5-year fol-
low-up (wave 2) in middle or late childhood. To our
knowledge, this is the first study to examine longitudinal
outcomes of BI among a large controlled sample of chil-
dren at high risk for anxiety disorders.
This study is a follow-up of a sample originally re-
cruited between 1993 and 1998. As described in earlier
reports,6,7we had recruited three groups of parents: (1)
131 parents treated for panic disorder and their 227
children (of these parents, 113 either had comorbid panic
and depression [n ? 102] or a spouse with depression [n
? 11]); (2) 39 parents with major depression without
panic disorder or agoraphobia and their 67 children; and
(3) 61 comparison parents without major anxiety or
mood disorders and their 119 children. Of these 413
children, 284 were in the age range at baseline (ages 21
months to 6 years) to undergo a laboratory assessment of
BI (see Fig. 1 for an overview of the study design).
Parents with panic and depression had been recruited
from hospital outpatient and health maintenance organi-
zation settings and advertisements and were included if
they met full DSM-III-R criteria for panic disorder or major
depressive disorder on the Structured Clinical Interview
for DSM-III-R (SCID)26and had been treated for these
disorders. Parents who were acutely psychotic or suicidal
were excluded from the study; however, parents in the
panic and depression groups were not excluded on the
basis of any comorbid disorders. Comparison parents
were recruited through advertisements to hospital per-
sonnel and in the community and were included only if
both parents did not meet DSM-III-R criteria on the SCID
for any major anxiety (panic disorder, agoraphobia, social
phobia or OCD) or mood disorders (major depressive
disorder, bipolar disorder, or dysthymia) and had never
received psychiatric treatment.
The institutional review board at Massachusetts Gen-
eral Hospital approved all study protocols. At each study
wave, after complete description of the study to the
subjects, informed consent was obtained. All parents
signed written consent for themselves and their children.
Children assented to study procedures, as follows. Chil-
dren 7 and older signed written assent forms, which were
read to them. Young children undergoing laboratory as-
sessments were told verbally in developmentally appro-
priate terms that they would be playing games with a
woman they did not know and that their mother would
be present the whole time. The mothers were instructed
that if they thought their child was “too unhappy” during
the assessment or that the examiner should skip over a
procedure, they should let the examiner know.
Assessment of Behavioral Inhibition
As described previously,6,8children ranging in age
from 21 months to 6 years had been assessed for BI at
baseline during a single visit to the Harvard Infant Study
Laboratory under the direction of Jerome Kagan and
Nancy Snidman. Three different age-specific laboratory
observational protocols (for children ages 21 months, 4
years, or 6 years, respectively) were used.6For each, the
child came to the laboratory, accompanied by mother,
and interacted with unfamiliar toys and adult examiners
and the child’s behavior was videotaped and quantified.
Three definitions based on different summary indexes
were used to aggregate children’s inhibited behaviors,
and children were considered behaviorally inhibited if
131 Parents with
39 Parents with Major
61 Parents without Mood or
413 children total
284 children ages 21 mos.-6 yrs assessed at Baseline for behavioral inhibition (BI)(90 had BI)
3 children excluded who also showed
signs of behavioral disinhibition
281 children (87 had BI)
215 children (mean age 10) assessed at Wave 2
for psychiatric diagnoses (67 had BI)
20 children did not
participate in Wave 2
65 children not assessed for psychiatric
diagnoses in Wave 1
216 children assessed in Wave 1 (at mean
age 6) for psychiatric diagnoses (64 had BI)
46 children did not
participate in Wave 2
170 children reassessed
in Wave 2
Figure 1. Schematic drawing of the study.
Behavioral Inhibition and Social Anxiety
Journal of Developmental & Behavioral Pediatrics
they met criteria for two or more definitions: (1) Dichot-
omous BI: According to this a priori definition based on
earlier work by Kagan and colleagues, 21-month-olds
were categorized as inhibited if they exhibited four or
more fears during the course of the battery or were rated
as showing “minimal” vocalization or smiling. Four- and
6-year-olds were rated inhibited if the number of sponta-
neous comments and smiles they made were both in the
lowest 20th percentile of the comparison children of
parents without anxiety or depressive disorders in their
respective age groups. (2) Global BI: A 4-point rating of
the child’s inhibition across the entire battery was made.
Children who received ratings of 3 (more inhibited than
not) or 4 (extremely inhibited) were considered inhib-
ited. (3) A summary score was derived from a principal
factors factor analysis with varimax rotation of all the
behavioral variables rated, computed separately for chil-
dren in each of the three age groups. Children were rated
as inhibited if their summary inhibition score fell in the
upper 20th percentile for the comparison children of
parents without anxiety or mood disorders, assessed with
the same age-specific protocol.
In a subsequent analysis, children who had shown
signs during the laboratory assessments of extreme ap-
proach or disinhibition were classified as behaviorally
disinhibited.27Three of the 284 children assessed showed
signs of both inhibition and disinhibition and therefore
were excluded from the present analysis. Therefore, the
present follow-up focused on the 281 children who could
be rated unequivocally as inhibited or not.
Initial Diagnostic Assessments
As reported earlier,8children who reached the age of
5 years by the end of the first 5-year grant period had
been assessed during the first wave of the study (wave 1),
based on interviews with the mother using the DSM-III-
R-based Kiddie Schedule for Affective Disorders and
Schizophrenia, Epidemiologic Version (K-SADS-E).28
Wave 2 Follow-up Assessment Procedures
Wave 2 follow-up assessments occurred a median of 5
years after baseline BI assessments (mean, 5.4 ? 1.4
years). Psychiatric assessments of the children relied on
the DSM-IV–based K-SADS-E.28,29We interviewed the
mothers about all children and directly interviewed chil-
dren older than 12 years (n ? 45). Because child and
mother reports do not always agree30,31and because
parent and child interviews may provide complementary
data (since children may report symptoms of which par-
ents are unaware, and parents may report symptoms that
children may not wish to admit), we followed the com-
monly used “or” rule and considered a disorder positive if
diagnostic criteria were unequivocally met in either the
parent or child interview. This a priori procedure for
combining information from direct and indirect inter-
views has been used in all our previous studies7,32Al-
though an analysis of the concordance rates for the 45
parent-child pairs is beyond the scope of this report, to
give an indication of the differences, the presence of any
lifetime anxiety disorder was reported for 26 children: 11
by mother only, six by child only, and nine by both; social
anxiety was reported for seven children: five by mother
only, and two by child only. Lifetime rates of disorder
Mothers and children were assessed by interviewers
blind to the child’s temperamental status and to all pre-
vious information about the child and family. Interview-
ers had a bachelor’s or master’s degree in psychology and
were trained to high levels of interrater reliability. They
underwent a training program that required them to (1)
master the diagnostic instruments, (2) learn about DSM-IV
criteria, (3) watch training tapes, (4) participate in inter-
views performed by experienced raters, (5) rate several
subjects under the supervision of the project coordinator,
(6) undergo continued supervision of their assessments
by senior staff, and (7) audiotape all interviews for later
random checking. All interviews were then presented for
review to a committee of board-certified child and adult
psychiatrists and licensed psychologists who were blind
to the subject’s ascertainment status, referral source, and
all other data to resolve diagnostic uncertainties. Diag-
noses were considered positive only if a consensus was
achieved that criteria were met to a degree that would be
considered clinically meaningful.
We computed kappa coefficients of agreement for all
diagnoses by having experienced, board-certified child
and adult psychiatrists and licensed clinical psychologists
diagnose subjects from audiotaped interviews made by
the assessment staff. Based on 500 assessments from in-
terviews of children and adults, the median kappa coef-
ficient was 0.98. Kappa coefficients included ADHD
(0.88), conduct disorder (CD) (1.0), oppositional defiant
disorder (ODD) (0.90), major depression (1.0), mania
(0.95), separation anxiety (1.0), agoraphobia (1.0), panic
(0.95), obsessive-compulsive disorder (OCD) (1.0), gen-
eralized anxiety disorder (GAD) (0.95), specific phobia
(0.95), and social phobia (1.0). These measures indicated
excellent reliability between ratings made by the noncli-
nician raters and experienced clinicians. In addition, to
estimate the reliability of the diagnostic review process,
we computed kappa coefficients of agreement between
clinician reviewers. The median reliability between indi-
vidual clinicians and the diagnoses assigned by the review
committee was 0.87. Kappa coefficients included atten-
tion-deficit/hyperactivity disorder (ADHD) (1.0), CD
(1.0), ODD (.90), major depression (1.0), mania (0.78),
separation anxiety (0.89), agoraphobia (.80), panic (.77),
OCD (0.73), GAD (0.90), specific phobia (0.85), and so-
cial phobia (0.90).
We first conducted univariate comparisons to deter-
mine whether demographic variables differed between
groups of children who were and were not reassessed or
children with or without BI. We classified a demographic
variable as a potential confound if it was significantly
associated (at p ? .10) with the predictor variable. We
then used the potential confound as a covariate in the
models used to assess the link between BI and child
Vol. 28, No. 3, June 2007
© 2007 Lippincott Williams & Wilkins
For binary outcomes (e.g., presence or absence of
disorder), we used logistic regression, and for continuous
outcomes, we used linear regression. Multiple siblings in
a single family cannot be considered independent be-
cause they share genetic, cultural, and social risk factors.
Therefore, we used Huber-White robust estimates of vari-
ance so that p values would be accurately estimated, as
implemented in STATA.33For some comparisons, sparse
data in some cells prevented the fitting of a standard
logistic regression model using maximum likelihood in-
ference. In these cases, we employed an alternate logistic
model that uses conditional exact inference, as imple-
mented in LogXact.34This method accommodates situa-
tions when the statistical assumptions of standard logistic
regression are violated; however, it cannot derive odds
ratios (ORs). For standard logistic regression models, we
report the asymptotic test statistics and p values. For
outcomes that could only be fit with conditional exact
inference, we report p values as provided by LogXact. All
tests were two tailed with alpha at .05.
Attrition and Demographic Characteristics
Of the 281 children assessed for behavioral inhibition
(BI) at baseline, 77% were reassessed diagnostically at
wave 2 follow-up. Children who were (n ? 215) and
were not (n ? 66) reassessed did not differ on baseline
demographic variables (mean ? SD age: 4.18 ? 1.29 vs
4.24 ? 1.47 years, ?2
Hollingshead class, 2.04 ? 0.98 vs 2.15 ? 1.01, ?2
0.47, p ? .49; family intactness: 86% in both groups,
p ? .68), temperament (BI: 31% vs 30%, ?2
.90), or parental diagnoses (panic disorder: 51% vs 62%,
p ? .36). At follow-up, children ranged in age from 6 to
15 years (mean ? SD, 9.61 ? 1.92).
As seen in Table 1, inhibited (n ? 67) and noninhibited
or race, but did differ at trend on SES and age. In addition,
the groups differed at trend on parental diagnosis (73% BI
and 60% non-BI children had parents with lifetime major
rental psychopathology, and SES are known risk factors for
2? 3.87, p ? .14; sex: 46% vs 36%
1? 1.99, p ? .16; socioeconomic status [SES]:
1? 0.020, p ? .88; race: 92% vs 94% white, ?2
1? 0.020, p ?
1? 1.48, p ? .22; depression: 65% vs 57%, ?2
1? 3.50, p ? .061). Because older age, pa-
child psychopathology and because we were interested in
assessing the independent risk conferred by BI, these vari-
ables were covaried in all comparisons.
Psychiatric Disorders in Offspring
As seen in Figure 2, inhibited and noninhibited chil-
dren did not differ on overall rates of any mood, disrup-
tive behavior, or anxiety disorders. Similarly, they did not
differ on the proportion of children who met more than
one disorder (42% in both groups, z ? ?0.75, p ? .46) or
two or more anxiety disorders (34% in inhibited and 31%
in noninhibited children, z ? ?0.19, p ? .85). (Note that
the high rate of psychopathology in general in this sample
of high-risk offspring35is consistent with other studies of
offspring at high risk for anxiety and depression36,37).
However, as seen in Figure 3, within the anxiety dis-
orders, inhibited children differed selectively on social
anxiety. For the analyses that follow, we classify children
as having social anxiety if they met criteria for either
DSM-III-R avoidant disorder or DSM-IV social phobia. In-
hibited children were significantly more likely than non-
inhibited children to manifest social anxiety (28% vs 14%,
OR ? 2.37, 95% confidence interval [CI]: 1.10–5.10,
1? 4.84, p ? .028).
Because social anxiety often has its onset in late child-
hood or early adolescence, we also examined whether BI
predicted new onset of social anxiety. Of the 215 chil-
dren we reassessed, 170 had had a wave 1 diagnostic
assessment8; of these, 153 had had no social anxiety.
Children who had not had social anxiety at wave 1 were
significantly more likely to show new onset: inhibited
versus noninhibited, 28.6% versus 10.8%; OR ? 3.17, 95%
CI: 1.19–8.48, ?2
children were significantly more likely to show new on-
set specifically of DSM-IV social phobia (22.2% vs 8.0%;
OR ? 3.15, 95% CI: 1.16–8.57, ?2
To further examine the data longitudinally, we com-
pared the degree of concurrence of diagnoses from wave
1 to wave 2. Of the 11 inhibited children (17%) found at
wave 1 to have social anxiety, six (55%) reported the
lifetime diagnosis at wave 2 follow-up, and one (9%) was
lost to follow-up. Of the 53 inhibited children who did
not meet criteria for social anxiety at wave 1, 12 (23%)
were found to have an onset of social anxiety, whereas 11
(21%) were lost to follow-up. If we considered as having
lifetime social anxiety any child for whom social phobia
1? 5.28, p ? .022). In addition, BI
1? 5.05, p ? .025).
Demographic Characteristics at Follow-up of Children with and without Behavioral Inhibition
Children with BI (N ? 67)Children without BI (N ? 148)Significance
MeanSDMeanSDWald ?2, p
1? 2.69, p ? .10
1? 3.27, p ? .070
Wald ?2, p
Child’s sex (% female)
Family intactness (% intact)
Race (% white)
1? 1.37, p ? .24
1? 0.010, p ? .93
1? 0.34, p ? .56
BI, behavioral inhibition. Socioeconomic status refers to the Hollingshead Four Factor Social Status class (ranging from 1, highest, to 5, lowest). Families were consid-
ered intact if the child’s biological parents were living together and were not divorced or separated.
Behavioral Inhibition and Social Anxiety
Journal of Developmental & Behavioral Pediatrics
or avoidant disorder had ever been endorsed (either at
wave 1 or 2), the overall rate in inhibited versus nonin-
hibited children was 34% versus 17% (OR ? 2.52, 95% CI:
1.18–5.38, z ? 2.40, p ? .017).
Interestingly, of the 45 children assessed at follow-up
that had not been assessed for psychopathology in wave
1 (i.e., 21 month olds assessed for BI at baseline who did
not reach the age of 5 by the end of wave 1), only a small
proportion developed social anxiety (one in 24 inhibited
children, or 4%). This suggests that BI measured with our
21-month protocol is not as good a predictor of social
anxiety as BI measured at older ages. To explore this
possibility, we examined the rates of lifetime and new-
onset social anxiety among children whose BI was mea-
sured at ages 21 months, 4 years, and 6 years (Table 2). It
was evident that BI measured at ages 4 or 6 years was
associated with two- to threefold elevations in the rate of
lifetime and new-onset social anxiety, whereas BI mea-
sured at 21 months was not. Although the associations for
children assessed at age 6 did not attain full significance
because of the small cell sizes, the ORs were nearly
identical to those for children assessed at age 4.
This study provides prospective evidence that behav-
ioral inhibition (BI) observed in the laboratory in early
childhood represents a specific risk factor for social anx-
iety during middle childhood among offspring at risk for
anxiety disorders. The risk conferred by BI appears spe-
cific to social anxiety since no other anxiety, mood, or
behavior disorders were significantly elevated among in-
hibited children. Most telling, BI predicted new onset of
social phobia within the 5-year follow-up period.
Our findings are consistent with those of our earlier
study finding BI associated with social anxiety in this
sample at a mean age of 6 years.8Our study identifies BI
mainly as an indicator of risk among children already at
risk because of family history. However, other studies
have found links between BI and social anxiety among
children not selected with reference to parental psycho-
pathology. These studies found associations between lab-
oratory-observed BI in toddlerhood and increased gener-
alizedsocial anxiety with
adolescence,9self-rated (retrospective) BI in early adoles-
cence and new onset of social phobia during high
school,10and self-rated BI and concurrent social anxi-
ety.38,39This is the first study to examine prospective
outcomes of laboratory-observed preschool BI using psy-
chiatric diagnoses derived from structured diagnostic in-
terviews in a sample of offspring at risk for anxiety dis-
Although our results suggest specificity in the link
between BI and social anxiety, other studies suggest that
the picture may be more complex. BI shows a familial
association with parental panic disorder2,6,25,40and major
have also linked BI to major depression24,41and panic
disorder.24Moreover, in a study that identified the corti-
cotropin-releasing hormone (CRH) locus as a gene asso-
ciated with BI among children from our sample, the
association between the CRH-linked locus and BI was
particularly strong among the offspring of parents with
panic disorder.42Since social phobia is known to have an
earlier age of onset than either major depression43or
panic disorder,44it is possible that the inhibited children
Percent of Children Affected
Figure 2. Rates of disorders at follow-up in children with and without
behavioral inhibition. Any anxiety disorder refers to presence of any of the
following: panic disorder, agoraphobia, social phobia, specific phobia, obses-
sive-compulsive disorder, generalized anxiety disorder, separation anxiety
disorder, or DSM-III-R avoidant disorder. Any mood disorder refers to the
presence of any of the following: major depressive disorder, bipolar I
disorder, bipolar II disorder, or dysthymia. Any disruptive behavior disorder
refers to the presence of any of the following disorders: attention-deficit/
hyperactivity disorder, oppositional defiant disorder, or conduct disorder. The
differences shown in this figure are not statistically significant.
e t c
l a i co
d r os iD
r ey t e i xn
d i o
s iDd r o
o i t a r ap
a i bohPc i f i cepS
a i bohP l a i c
a i bohpa r ogA
r ed r o s iDc i n
) 841=n ( IB
n ( IB
d e f fAn e r d l i hC f o t ne c r eP
Figure 3. Rates of anxiety disorders at follow-up in children with and
OCD, obsessive compulsive disorder, GAD, generalized anxiety disorder;
Separation, separation anxiety disorder. Any social anxiety disorder refers to
the presence of either social phobia or avoidant disorder.
aOdds ratio (OR) ? 2.16, 95% confidence interval (CI):
1? 3.33, p ? .068;bOR ? 3.79, 95% CI: 1.13–12.67,
Vol. 28, No. 3, June 2007
© 2007 Lippincott Williams & Wilkins
in our sample may still emerge with one or both of these
disorders. Continuing follow-up of the sample will enable
us to test these hypotheses.
Although BI increased the statistical risk of developing
social anxiety, the majority of inhibited children were not
affected. With the caveat that the children were still
within the age of risk for onset of social anxiety,44it is
possible that for some children, early BI either improves
over time or else remains a behavioral style that may
influence the individual’s reaction to novelty or choice of
activities,45but not necessarily lead to distress or im-
Our data suggest that BI observed at ages 4 and 6 years
but not at 21 months appears most associated with later
social anxiety. These findings contrast with those from
the cohorts recruited by Kagan and colleagues1from
birth records, in which toddlers identified as inhibited at
age 21 or 31 months were found to be significantly more
likely to develop generalized social anxiety by age 13.9
One explanation for this discrepancy concerns the differ-
ent ways in which inhibition was assessed at 21 months.
In the Kagan et al earlier study, the 21-month assessment
included tasks slightly different from those used with our
sample (including separation from the mother) and was
scored differently (with nine inhibited behaviors re-
quired, including long latencies to interact as well as signs
of distress/avoidance and low vocalization).46The cohort
assessed at 31 months was rated as inhibited based on a
peer-play session with an unfamiliar child. Additionally,
children classified as inhibited at 21 months in the earlier
study of Kagan et al1were reassessed for BI at ages 4, 5,
and 7 years and found to have moderate stability of their
inhibition, with those who were most stable showing the
highest rates of personal and familial social anxiety.47In
contrast, in the present study, the 21-month assessment
relied on a single assessment with an adult examiner,
with fewer inhibited behaviors (four fears or minimal
vocalizations or smiles) needed to classify a child as in-
hibited. It is possible that in this context, ratings of
inhibition (fear and quietness) at 21 months may be
confounded by residual, developmentally normative sep-
aration or stranger anxiety in some children or by matu-
rational variability in the cognitive ability to process novel
This pattern of associations might raise the question of
whether the inhibition we observed might be learned
reactions to having a parent with depression/anxiety
rather than a manifestation of temperament. However, in
a previous study, we rated the children from this sample
at baseline according to the degree to which they were
exposed to parental depression and anxiety disorders.49
Despite ample power to detect medium effect sizes, we
found no association between any of our definitions of BI
and exposure to parental anxiety or depression. This is in
accord with studies50,51that have found high heritability
Our results differed from the earlier assessment wave
in that BI no longer showed an inverse association with
disruptive behavior disorders. At a mean age of 6 years,
the inhibited children appeared protected against the
development of attention-deficit/hyperactivity disorder
(ADHD) and oppositional defiant disorder,8but by mean
age 10, they had rates of these disorders not discernibly
different from their noninhibited peers. Many of the in-
hibited children were offspring of depressed parents; as
such, they carried an increased vulnerability for ADHD
and other disruptive behavior disorders.7,52It is possible
that BI delays the onset of disruptive behavior disorders,
but does not reduce their risk in middle childhood.
This study was limited by reliance on parental report
for lifetime diagnoses in children younger than age 12.
We did not directly interview younger children because
they are limited in their expressive/receptive language
abilities, lack the ability to map events in time, and have
limited powers of abstraction. Several studies53,54have
documented poor reliability of reports of psychopathol-
ogy by children younger than 12 years of age and Breton
et al55found that children ages 9 to 11 years understood
only about 40% of the questions from the Diagnostic
Interview Schedule for Children. In contrast, maternal
reports of psychopathology have shown high reliability,56
even over a 1-year period.57Future study waves, in which
all the children will be old enough to be directly inter-
viewed about their lifetime symptoms, will enable us to
at Ages 21 Months, 4 Years, and 6 Years
Rates of Social Anxiety (Social Phobia and/or Avoidant Disorder) at 5-Year Follow-up Among Children Assessed for Behavioral Inhibition
Lifetime social anxiety
Assessed at 21 mo (n ? 36)
Assessed at age 4 yr (n ? 124)
Assessed at age 6 yr (n ? 55)
New onset social anxiety
Assessed at 21 mo (n ? 7)
Assessed at age 4 yr (n ? 96)
Assessed at age 6 yr (n ? 50)
OR ? 0.56 (95% CI: 0.058–5.33); z ? ?0.51, p ? .61
OR ? 2.77 (95% CI: 1.12–6.85); z ? 2.20, p ? .028
OR ? 3.50 (95% CI: 0.86–14.25); z ? 1.75, p ? .080
OR ? 1.50 (95% CI: 0.043–52.93); z ? 0.22, p ? .82
OR ? 3.32 (95% CI: 1.02–10.80); z ? 1.99, p ? .047
OR ? 3.67 (95% CI: 0.78–17.19); z ? 1.65, p ? .099
BI, behavioral inhibition; OR, odds ratio; CI, confidence interval. Social anxiety refers to presence of either DSM-IV social phobia or DSM-III-R avoidant disorder. Be-
cause of the small cell sizes, tests were done without covarying potential confounds. When socioeconomic status, child age, and parental diagnosis was covaried, the values
were as follows: lifetime social anxiety, with BI assessed at age 4 years: OR ? 2.96 (95% CI: 1.07– 8.18); z ? 2.10, p ? .036; at age 6 years: p ? .58 by LogXact. New onset
social anxiety with BI assessed at age 4 years: p ? .17 by LogXact; at age 6 years: p ? .58 by LogXact.
Behavioral Inhibition and Social Anxiety
Journal of Developmental & Behavioral Pediatrics
confirm our current findings. The study was also limited
by the low representation of minority families. Further
studies of more ethnically diverse samples are needed to
extend the generalizability of the findings.
Another limitation concerns the way in which BI was
assessed. The batteries developed by Kagan and col-
leagues and used in this study focus mainly on the child’s
behavior during interactions with unfamiliar examiners
(exposure to social novelty). Although in some of the
batteries, behaviors with unfamiliar objects were as-
sessed, these samples of behavior were not sufficient to
quantify reactions to nonsocial as distinct from social
novelty. Some have argued that inhibition to nonsocial
stimuli may predict different outcomes,58perhaps predis-
posing to different types of phobic disorders. Future stud-
ies that differentially measure social and nonsocial inhibi-
tion and their outcomes could shed further light on the
specificity of the link between BI and social anxiety. In
addition, we classified children as inhibited based on
their meeting two out of three of our definitions. Al-
though this is a conservative approach, we had no exter-
nal validation that the children classified based on this
single assessment really had stable temperamental inhibi-
tion. Moreover, practitioners observing BI in the consult-
ing room will have to draw on methods other than stan-
dardized laboratory observations (e.g., global impression
and parent report). Further studies are needed to address
these measurement differences.
Our negative findings should be interpreted with cau-
tion. Although we had a power of 80% or greater to
detect medium effect sizes, smaller differences would
have escaped detection. Finally, our use of the “or” rule in
classifying children as socially anxious (based on meeting
either avoidant disorder or social phobia) may have over-
estimated the prevalence of this problem. We used this
more general classification because we believe that pedi-
atricians should be attentive to social anxiety in general,
whether it manifests in fearful avoidance of new peers or
in full DSM-IV social phobia criteria.
Despite these limitations, our findings suggest that
preschool-age BI identifies a subset of offspring of parents
with panic disorder and/or depression who are at ele-
vated risk to develop social anxiety by early to middle
childhood. Clinically, these results suggest that young
children presenting with BI should be monitored closely
for signs that their tendency to display inhibition in social
settings may be leading to distress, avoidance, or social or
academic dysfunction. Careful attention to family history
is also important since parents of inhibited children are
likely to have high rates of anxiety disorders them-
selves.3,6Because parental psychopathology may influ-
ence a child’s adaptation, referral of the parents for treat-
ment of their own symptoms may be indicated.
Our study raises the hypothesis that young inhibited
children may benefit from preventive cognitive-behav-
ioral interventions to reduce social anxiety and improve
coping.59A cognitive-behavioral approach to treating
anxiety in young children60,61consists of a short-term
intervention that emphasizes (1) psychoeducation, (2)
coping strategies taught to the child, (3) parent-guidance
about how to contingently reinforce the child’s adaptive
coping and not inadvertently reward anxious responses,
and (4) exposure exercises to extinguish fear associa-
tions. Pilot data from our group62support the hypothesis
that these strategies, used beneficially with older chil-
dren,63–66can be adapted for use with younger children
as well. Rapee and colleagues have demonstrated that a
brief parent group intervention can reduce anxiety dis-
tress among preschoolers with BI.62
Short of referring to a cognitive-behavioral specialist,
pediatricians might also provide practical advice to par-
ents of young inhibited children. Parents might be coun-
seled to empathize with their child’s discomfort with
unfamiliar people, yet not to allow the child to avoid
them. Inhibited children might benefit from opportuni-
ties to habituate to new settings (e.g., advance visits to a
new classroom or dentist’s office, playdates with new
classmates in advance of beginning school), as well as
from graduated exposure to feared situations, facilitated
by immediate rewards.
Future studies are needed to explore the effects of
parental treatment on child outcome among children at
risk, as well as to evaluate in randomized, controlled trials
the effects of parental treatment, parental guidance, and
parent-child cognitive-behavioral interventions on the de-
velopment of social anxiety disorders among children
The authors gratefully acknowledge Drs. Jerome Kagan and
Nancy Snidman of the Department of Psychology, Harvard Univer-
sity, for their assistance on this project.
1. Kagan J, Reznick JS, Snidman N. Biological bases of childhood
shyness. Science. 1988;240:167–171.
2. Rosenbaum JF, Biederman J, Gersten M, et al. Behavioral
inhibition in children of parents with panic disorder and
agoraphobia: a controlled study. Arch Gen Psychiatry. 1988;45:
3. Rosenbaum JF, Biederman J, Hirshfeld DR, et al. Further evidence
of an association between behavioral inhibition and anxiety
disorders: Results from a family study of children from a non-
clinical sample. J Psychiatr Res. 1991;25:49–65.
4. Biederman J, Rosenbaum J, Hirshfeld D, et al. Psychiatric
correlates of behavioral inhibition in young children of parents
with and without psychiatric disorders. Arch Gen Psychiatry.
5. Biederman J, Rosenbaum JF, Bolduc-Murphy EA, et al. A 3-year
follow-up of children with and without behavioral inhibition.
J Am Acad Child Adolesc Psychiatry. 1993;32:814–821.
6. Rosenbaum JF, Biederman J, Hirshfeld-Becker DR, et al. A
controlled study of behavioral inhibition in children of parents
with panic disorder and depression. Am J Psychiatry. 2000;157:
7. Biederman J, Faraone S, Hirshfeld-Becker D, et al. Patterns of
psychopathology and dysfunction in a large sample of high-risk
children of parents with panic disorder and major depression: a
controlled study. Am J Psychiatry. 2001;158:49–57.
8. Biederman J, Hirshfeld-Becker DR, Rosenbaum JF, et al. Further
evidence of association between behavioral inhibition and social
anxiety in children. Am J Psychiatry. 2001;158:1673–1679.
Vol. 28, No. 3, June 2007
© 2007 Lippincott Williams & Wilkins
9. Schwartz C, Snidman N, Kagan J. Adolescent social anxiety as an
outcome of inhibited temperament in childhood. J Am Acad
Child Adolesc Psychiatry. 1999;38:1008–1015.
10. Hayward C, Killen J, Kraemer K, et al. Linking self-reported
childhood behavioral inhibition to adolescent social phobia. J Am
Acad Child Adolesc Psychiatry. 1998;37:1308–1316.
11. Rickman M, Davidson R. Personality and behavior in parents of
temperamentally inhibited and uninhibited children. Dev
12. Cooper PJ, Eke M. Childhood shyness and maternal social phobia:
a community study. Br J Psychiatry. 1999;174:439–443.
13. APA. Diagnostic and Statistic Manual of Mental Disorders, 4th
Edition. Washington, DC: American Psychiatric Association, 1994.
14. Kessler RC, McGonagle KA, Zhao S, et al. Lifetime and 12-month
prevalence of DSM-III-R psychiatric disorders in the United States.
Arch Gen Psychiatry. 1994;51:8–19.
15. Egger HL, Angold A. Common emotional and behavioral disorders
in preschool children: presentation, nosology, and epidemiology.
J Child Psychol Psychiatry. 2006;47:313–337.
16. Costello EJ, Angold A. Epidemiology. In: March JS, ed. Anxiety
Disorders in Children and Adolescents XVI. New York, NY:
Guilford Press, 1995:109–124.
17. Ost L-G. Age of onset in different phobias. J Abnorm Psychol.
18. Merikangas KR, Avenevoli S. Epidemiology of mood and anxiety
disorders in children and adolescents. In: Tsuang MT, Tohen M,
eds. Textbok in Psychiatric Epidemiology, 2nd ed. New York,
NY: John Wiley & Sons, 2002.
19. Mannuzza S, Schneier FR, Chapman TF, et al. Generalized social
phobia: reliability and validity. Arch Gen Psychiatry. 1995;52:
20. Beidel DC, Turner SM, Morris TL. Psychopathology of childhood
social phobia. J Am Acad Child Adolesc Psychiatry. 1999;38:643–
21. Mancini C, Van Ameringen M, Bennett M, et al. Emerging
treatments for child and adolescent social phobia: a review.
J Child Adolesc Psychopharmacol. 2005;15:589–607.
22. Noyes RJ, Holt CS, Woodman CL. Natural course of anxiety
disorders. In: Mavissakalian MR, Prien RF eds. Long-term
Treatment of Anxiety Disorders XII. Washington, DC: American
Psychiatric Association, 1996:1–48.
23. Kochanska G. Patterns of inhibition to the unfamiliar in children
of normal and affectively ill mothers. Child Dev. 1991;62:250–
24. Reznick JS, Hegeman IM, Kaufman E, et al. Retrospective and
concurrent self-report of behavioral inhibition and their relation
to adult mental health. Dev Psychopathol. 1992;4:301–321.
25. Battaglia M, Bajo S, Strambi LF, et al. Physiological and behavioral
responses to minor stressors in offspring of patients with panic
disorder. J Psychiatr Res. 1997;31:365–376.
26. Spitzer R, WIlliams J, Gibbon M, et al. Structured Clinical
Interview for DSM-III-R-Non-Patient Edition (SCID-NP, Version
1.0). Washington, DC: American Psychiatric Press, 1990.
27. Hirshfeld-Becker DR, Biederman J, Faraone S, et al.
Temperamental correlates of disruptive behavior disorders in
young children: preliminary findings. Biol Psychiatry. 2002;50:
28. Puig-Antich J, Ryan ND. Schedule of Affective Disorders and
Schizophrenia for School-Aged Children (Kiddie-SADS).
Pittsburgh, PA: Western Psychiatric Institute and Clinic, 1986.
29. Orvaschel H, Puig-Antich J. Schedule for Affective Disorders and
Schizophrenia for School-Age Children-Epidemiologic, 4th
version. Ft. Lauderdale: Nova University, Center for Psychological
30. Comer J, Kendall PC. A symptom-level examination of parent-
child agreement in the diagnosis of anxious youths. J Am Acad
Child Adolesc Psychiatry. 2004;43:878–886.
31. Safford SM, Kendall PC, Flannery-Schroeder E, et al. A
longitudinal look at parent-child diagnostic agreement in youth
treated for anxiety disorders. J Clin Child Adolesc Psychol. 2005;
32. Biederman J, Faraone S, Milberger S, et al. A prospective 4-year
follow-up study of attention deficit hyperactivity and related
disorders. Arch Gen Psychiatry. 1996;53:437–446.
33. Stata. Stata Reference Manual: Release 5. College Station, TX:
Stata Corporation, 1997.
34. Cytel. LogXact For Windows: Logistic Regression Software
Featuring Exact Methods User Manual. Cambridge MA: Cytel
Software Corporation, 1996.
35. Biederman J, Petty C, Hirshfeld-Becker DR, et al. A controlled
longitudinal five-year follow-up study of children at high and low
risk for panic disorder and major depression. Psychol Med. 2006;
36. Weissman M, Leckman J, Merikangas K, et al. Depression and
anxiety disorders in parents and children: results from the Yale
Family Study. Arch Gen Psychiatry. 1984;41:845–852.
37. Beidel D, Turner S. At risk for anxiety: I. Psychopathology in the
offspring of anxious parents. J Am Acad Child Adolesc
38. Mick M, Telch M. Social anxiety and history of behavioral
inhibition in young adults. J Anxiety Disord. 1998;12:1–20.
39. Muris P, Merckelbach H, Wessel I, et al. Psychopathological
correlates of self-reported behavioural inhibition in normal
children. Behav Res Ther. 1999;37:575–584.
40. Manassis K, Bradley S, Goldberg S, et al. Behavioural inhibition,
attachment and anxiety in children of mothers with anxiety
disorders. Can J Psychiatry. 1995;40:87–92.
41. Caspi A, Moffitt TE, Newman DL, et al. Behavioral observations at
age 3 years predict adult psychiatric disorders. Arch Gen
42. Smoller JW, Rosenbaum JF, Biederman J, et al. Association of a
genetic marker at the corticotropin releasing hormone locus with
behavioral inhibition. Biol Psychiatry. 2003;54:1376–1381.
43. Kessler RC, Stang P, Wittchen HU, et al. Lifetime co-morbidities
between social phobia and mood disorders in the US National
Comorbidity Survey. Psychol Med. 1999;29:555–567.
44. Regier DA, Rae DS, Narrow WE, et al. Prevalence of anxiety
disorders and their comorbidity with mood and addictive
disorders. Br J Psychiatry. 1998;34(suppl):24–28.
45. Kagan J. Galen’s Prophecy: Temperament in Human Nature.
New York: Basic Books, 1994.
46. Garcia-Coll C, Kagan J, Reznick JS. Behavioral inhibition in young
children. Child Dev. 1984;55:1005–1019.
47. Hirshfeld DR, Rosenbaum JF, Biederman J, et al. Stable behavioral
inhibition and its association with anxiety disorder. J Am Acad
Child Adolesc Psychiatry. 1992;31:103–111.
48. Kagan J, Snidman N. The Long Shadow of Temperament.
Cambridge, MA: Belknap Press, 2004.
49. Hirshfeld-Becker DR, Biederman J, Faraone S, et al. Lack of
association between behavioral inhibition and psychosocial
adversity factors in children at risk. Am J Psychiatry. 2004;161:
50. Matheny AP. Children’s behavioral inhibition over age and across
situations: genetic similarity for a trait during change. J Pers.
51. DiLalla LF, Kagan J, Reznick JS. Genetic etiology of behavioral
inhibition among 2-year-old children. Infant Behav Dev. 1994;17:
Behavioral Inhibition and Social Anxiety
Journal of Developmental & Behavioral Pediatrics
405–412. Download full-text
52. Faraone SV, Biederman J. Do attention deficit hyperactivity
disorder and major depression share familial risk factors? J Nerv
Ment Dis. 1997;185:533–541.
53. Edelbrock C, Costello AJ, Dulcan MK, et al. Age differences in the
reliability of the psychiatric interview of the child. Child Dev.
54. Schwab-Stone M, Fallon T, Briggs M, et al. Reliability of diagnostic
reporting for children aged 6-11 years: a test-retest study of the
Diagnostic Interview Schedule for Children- Revised.
Am J Psychiatry. 1994;151:1048–1054.
55. Breton J, Bergeron L, Valla J, et al. Do children aged 9 through 11
years understand the DISC version 2.25 questions? J Am Acad
Child Adolesc Psychiatry. 1995;34:946–956.
56. Faraone SV, Biederman J, Milberger S. How reliable are maternal
reports of their children’s psychopathology? one year recall of
psychiatric diagnoses of ADHD children. J Am Acad Child
Adolesc Psychiatry. 1995;34:1001–1008.
57. Fallon T, Schwab-Stone M. Determinants of reliability in
psychiatric surveys of children aged 6-11 years: a test-retest study.
J Child Psychol Psychiatry. 1994;35:1391–1408.
58. Rubin K, Hastings P, Stewart S, et al. The consistency and
concomitants of inhibition: some of the children, all of the time.
Child Dev. 1997;68:467–483.
59. Hirshfeld-Becker DR, Biederman J. Rationale and principles for
early intervention with young children at risk for anxiety
disorders. Clin Child Fam Psychol Rev. 2002;5:161–172.
60. Hirshfeld-Becker DR. Cognitive-Behavioral Intervention for Young
Anxious Children: A Controlled Trial. Presented at: 52nd Annual
Meeting of the American Academy of Child and Adolescent
Psychiatry; 2005; Toronto, Canada.
61. Rapee RM, Kennedy S, Ingram M, et al. Prevention and early
intervention of anxiety disorders in inhibited preschool children.
J Consult Clin Psychol. 2005;73:488–497.
62. Hirshfeld-Becker DR, Masek B, Henin A, et al. Early CBT
intervention for young (4–7-year-old) anxious children. Presented
at: Proceedings of the Annual Meeting of the Anxiety Disorders
Association of America (ADAA); 2006; Miami, FL.
63. Kendall P, Chansky T, Kane M, et al. Anxiety Disorders in Youth:
Cognitive-Behavioral Interventions. Boston: Allyn and Bacon,
64. Kendall P, Kane M, Howard B, et al. Cognitive-Behavioral
Therapy for Anxious Children: Therapist Manual. Ardmore:
Workbook Publishing, 1992.
65. Kendall PC. Treating anxiety disorders in children: results of a
randomized clinical trial. J Consult Clin Psychol. 1994;62:100–
66. Kendall PC, Southam-Gerow MA. Long-term follow-up of a
cognitive-behavioral therapy for anxiety disordered youth.
J Consult Clin Psychol. 1996;64:724–730.
Three Kinds of Sons in Early Ireland
cases historical records provide a rough idea of what it may have been like. For example, surviving medieval Irish laws make
available some insight into varieties of family life.
There were three legally recognized kinds of sons living with their fathers: macc te ´ (warm son) was a dependent son who was
highly controlled by his father; macc u ´ar (cold son) who “has failed in his duty to provide filial service and obedience. . . He
cannot be harboured or protected by anyone;” and macc ailte (reared son), “who has been allowed his independence His father
has permitted him to choose whether he wishes to devote himself to a profession, or to husbandry.” However, the age range for
these distinctions is not stated.
The basis for and the consequences of these classifications are not recorded. “An additional difficulty is that the law has only
his wife’s kin.”
One can only imagine what it was like to be labeled legally and treated as warm, cold, or reared. Should we not be cautious
about labels we use?
Reference: Kelly F. A Guide to Early Irish Law. Dublin Institute for Advanced Studies. 1988:80-81.
Provided by Dr. John P. Carey, University of Cork
Submitted by William B. Carey, MD
Vol. 28, No. 3, June 2007
© 2007 Lippincott Williams & Wilkins