Sticks, Stones, and Hurtful Words: Relative Effects of Various Forms of Childhood Maltreatment

Abstract

Childhood maltreatment is an important psychiatric risk factor. Research has focused primarily on the effects of physical abuse, sexual abuse, or witnessing domestic violence. Parental verbal aggression has received little attention as a specific form of abuse. This study was designed to delineate the impact of parental verbal aggression, witnessing domestic violence, physical abuse, and sexual abuse, by themselves and in combination, on psychiatric symptoms.
Symptoms and exposure ratings were collected from 554 subjects 18-22 years of age (68% female) who responded to advertisements. The Verbal Abuse Questionnaire was used to assess exposure to parental verbal aggression. Outcome measures included dissociation and symptoms of "limbic irritability," depression, anxiety, and anger-hostility. Comparisons were made by using effect sizes.
Verbal aggression was associated with moderate to large effects, comparable to those associated with witnessing domestic violence or nonfamilial sexual abuse and larger than those associated with familial physical abuse. Exposure to multiple forms of maltreatment had an effect size that was often greater than the component sum. Combined exposure to verbal abuse and witnessing domestic violence had a greater negative effect on some measures than exposure to familial sexual abuse.
Parental verbal aggression was a potent form of maltreatment. Exposure to multiple forms of abuse was associated with very large effect sizes. Most maltreated children had been exposed to multiple types of abuse, and the number of different types is a critically important factor.

Full text

Am J Psychiatry 163:6, June 2006 993
Article
ajp.psychiatryonline.org
Sticks, Stones, and Hurtful Words: Relative Effects
of Various Forms of Childhood Maltreatment
Martin H. Teicher, M.D., Ph.D.
Jacqueline A. Samson, Ph.D.
Ann Polcari, R.N., C.S., Ph.D.
Cynthia E. McGreenery
Objective: Childhood maltreatment is
an important psychiatric risk factor. Re-
search has focused primarily on the ef-
fects of physical abuse, sexual abuse, or
witnessing domestic violence. Parental
verbal aggression has received little atten-
tion as a specific form of abuse. This study
was designed to delineate the impact of
parental verbal aggression, witnessing do-
mestic violence, physical abuse, and sex-
ual abuse, by themselves and in combina-
tion, on psychiatric symptoms.
Method: Symptoms and exposure ratings
were collected from 554 subjects 18–22
years of age (68% female) who responded
to advertisements. The Verbal Abuse
Questionnaire was used to assess expo-
sure to parental verbal aggression. Out-
come measures included dissociation and
symptoms of “limbic irritability,” depres-
sion, anxiety, and anger-hostility. Compar-
isons were made by using effect sizes.
Results: Verbal aggression was associ-
ated with moderate to large effects, com-
parable to those associated with witness-
ing domestic violence or nonfamilial
sexual abuse and larger than those associ-
ated with familial physical abuse. Expo-
sure to multiple forms of maltreatment
had an effect size that was often greater
than the component sum. Combined ex-
posure to verbal abuse and witnessing
domestic violence had a greater negative
effect on some measures than exposure
to familial sexual abuse.
Conclusions: Parental verbal aggression
was a potent form of maltreatment. Ex-
posure to multiple forms of abuse was
associated with very large effect sizes.
Most maltreated children had been ex-
posed to multiple types of abuse, and
the number of different types is a criti-
cally important factor.
(Am J Psychiatry 2006; 163:993–1000)
Childhood abuse has been the focus of increasing
concern, particularly when it is sexual or severely physical.
This has been further reinforced by studies that suggest
that abuse produces enduring effects on brain develop-
ment (1). While sexual abuse has come under intense
scrutiny as a psychiatric risk factor, emotional maltreat-
ment may be a more elusive and insidious problem. Emo-
tional abuse encompasses several forms of childhood
maltreatment, such as the witnessing of domestic violence
and exposure to verbal aggression (2). Generally, exposure
to verbal aggression has received little attention as a spe-
cific form of abuse, although it may be at least as impor-
tant as witnessing domestic violence. In one large national
study, Vissing et al. (3) found that 63% of American parents
reported one or more instances of verbal aggression, such
as swearing at and insulting their child. Children who were
the target of frequent verbal aggression exhibited higher
rates of physical aggression, delinquency, and interper-
sonal problems than other children (3).
Maternal verbal abuse during childhood has been asso-
ciated with a markedly higher risk for development of bor-
derline, narcissistic, obsessive-compulsive, and paranoid
personality disorders (4). These associations remained
significant after control for temperament, physical abuse,
sexual abuse, neglect, parental psychopathology, and co-
occurring psychiatric disorders (4). Verbal abuse may also
have more lasting consequences than other forms of
abuse (5) and, in combination with physical abuse and ne-
glect, produce the most dire outcome (6). However, child
protective service agencies, doctors, and lawyers are most
concerned about the impact and prevention of physical or
sexual abuse (7, 8).
Co-occurrence of multiple types of childhood abuse is
known to be common, with reported rates of co-occur-
rence ranging from 3% to 55% (9). When the effects of ex-
posure to multiple forms of adversity are considered, it ap-
pears that the greater the number of forms experienced,
the more severe the subsequent pathology (9, 10).
We sought to answer two questions. First, is there a dis-
cernible impact of exposure to childhood verbal aggres-
sion in the absence of physical abuse, sexual abuse, or ex-
posure to domestic violence? Second, what are the relative
psychiatric consequences of childhood exposure to verbal
aggression, witnessing domestic violence, physical abuse,
and sexual abuse, experienced either alone or in combina-
tion? Dissociation and “limbic irritability” were selected as
two primary variables for analysis, as previous research
had shown robust correlations between dissociation and
hippocampal size (11) and between limbic irritability and
blood flow to the cerebellar vermis as assessed with func-

994 Am J Psychiatry 163:6, June 2006
CHILDHOOD MALTREATMENT
ajp.psychiatryonline.org
tional magnetic resonance imaging (12). We also assessed
the effects of exposure on symptoms of depression, anxi-
ety, and anger-hostility.
Method
Subjects
Detailed ratings of symptoms and exposure history were col-
lected from 554 young adults who responded to advertisements
requesting healthy subjects or individuals with a history of an un-
happy childhood. They ranged in age from 18 to 22 years (mean=
19.8, SD=1.4 years); 378 were women, and 176 were men. After
complete description of the study to the subjects, written in-
formed consent was obtained. Most of the study subjects (73%)
were white, 6% were black, 6% were Hispanic, 10% were Asian, 1%
were Native American, and 4% were from other ethnic groups.
Assessments
Exposure to verbal aggression. The Verbal Abuse Question-
naire consists of 15 items that cover the key components of verbal
abuse—scolding, yelling, swearing, blaming, insulting, threaten-
ing, demeaning, ridiculing, criticizing, belittling, etc. In a separate
group of 48 college students, the questionnaire showed high in-
ternal consistency as applied to both maternal and paternal be-
haviors (Cronbach alphas, 0.98 and 0.94, respectively). Scores
from the questionnaires for maternal and paternal verbal abuse
were averaged as a measure of childhood verbal abuse. More de-
tailed discussion of scale development and psychometric proper-
ties is available from the authors.
The Verbal Abuse Questionnaire provides a continuous mea-
sure of exposure, which correlated strongly with all of our symp-
tom measures. However, for comparison with other forms of mal-
treatment that were dichotomized into “presence” or “absence,”
we selected a cutoff score to identify subjects exposed to a sub-
stantial degree of verbal aggression. From the subjects with no
history of physical or sexual abuse a cutoff score was selected that
designated the top 10% of scores. We used this score (>40) to de-
lineate a level of verbal aggression that was unusually high but
not rare. This upper 10% criterion had face validity, as this also
corresponded in this group to the percentage of subjects who en-
dorsed a history of exposure to domestic violence. The effect sizes
for the Verbal Abuse Questionnaire as a continuous measure were
always greater than the effect sizes for the dichotomous variable,
but cutoff scores greater than 40 provided effect sizes that were in
close agreement (70%–80% as large).
Exposure to other abuse and trauma. History of exposure to
physical abuse was obtained by self-report in response to the fol-
lowing question: “Have you ever been physically hurt or attacked
by someone such as husband, parent, another family member, or
friend (for example, have you ever been struck, kicked, bitten,
pushed, or otherwise physically hurt)?” If so, the subjects were
asked to provide information on their relationship to this individ-
ual, the number of times they were hurt, ages at initiation and ter-
mination of these episodes, whether the abuse received or should
have received medical attention, and whether the abuse resulted
in permanent injuries or scars. An individual was classified as
having experienced physical abuse if he or she reported any epi-
sode of inflicted physical injury that received or should have re-
ceived medical treatment or resulted in permanent injury or if
there were at least four reported episodes of what he or she felt
were less serious attacks.
Individuals were classified as having experienced sexual abuse if
they responded affirmatively to the following question: “Have you
ever been forced into doing more sexually than you wanted to do or
were too young to understand? (By “sexually” we mean being forced
against your will into contact with the sexual parts of your body or
his or her body.)” They were also asked to provide information on
their relationship to this individual, the number of times they were
forced, ages at first and last abuse, and whether or not they felt terri-
fied or had their life or another person’s life threatened.
History of exposure to domestic violence was assessed by using
the question “Have you ever witnessed serious domestic violence?”
Limbic irritability. The Limbic System Checklist-33 (13) was
created to evaluate the frequency with which subjects experience
symptoms often encountered as phenomena of ictal temporal
lobe epilepsy, as described by Spiers et al. (14). These items consist
of paroxysmal somatic disturbances, brief hallucinatory events,
visual phenomena, automatism, and dissociative experiences.
Psychometric studies showed that the Limbic System Checklist-33
has high test-retest reliability (r=0.92, N=16) (13). Scores were low
in normal comparison subjects (<10) and higher in patients with
documented temporal lobe epilepsy (>23). Scores on the Limbic
System Checklist-33 are dramatically influenced by abuse history
(13), more so than any other variable we have examined (12).
Psychiatric symptoms. The Dissociative Experience Scale (15)
consists of 28 questions that assess the frequency of various dis-
sociative experiences. Scores on each item range from 0 to 100,
and they are averaged to provide an index score. Total scores un-
der 20 capture most healthy subjects and patient groups with no
appreciable dissociative symptoms.
Self-ratings of other psychiatric symptoms were obtained by
using Kellner’s Symptom Questionnaire (16). This is a 92-item
yes/no questionnaire used to elicit ratings of depression, anxiety,
anger-hostility, and somatic complaints. It was developed to de-
tect response to psychotropic medications and is sensitive to sub-
tle differences from normal.
Data Analysis
The strength of the association between maltreatment history
and self-report symptom scores was assessed by calculating the
effect sizes and 95% confidence intervals (CIs) for the differences
between subjects who had no exposure to maltreatment and sub-
jects exposed to the different maltreatment categories. Effect size
is a more valuable measure for assessing the impact of an experi-
ence than the p value, which is strongly affected by group size.
Cohen’s d′ values of 0.2, 0.5, and 0.8 are usually interpreted to rep-
resent small, medium, and large effects, respectively (17).
Effect size was used to highlight the differences between sub-
jects exposed to different forms of abuse and healthy comparison
subjects. Analysis of variance provided a single statistical mea-
sure for the effect of exposure to the various categories of mal-
treatment on the dependent variables. Finally, a few planned sta-
tistical comparisons germane to the main focus of the study were
made in order to understand the significant omnibus F values.
These tests were done to address different specific types of expo-
sure: 1) emotional abuse versus sexual abuse, 2) emotional abuse
versus physical abuse, 3) verbal abuse versus witnessing domestic
violence, 4) the combination of verbal abuse and witnessing do-
mestic violence versus familial sexual abuse, and 5) multiple cat-
egories of abuse.
Results
Effects on Limbic Irritability
As seen in Figure 1, there were robust effects of each of
the five broad abuse categories (emotional only, sexual
only, physical only, any two, all three) on ratings on the
Limbic System Checklist-33 (F=21.46, df=5, 542, p<10
–15
,
n
2
=0.149). There was no effect of gender on this measure
(F=1.13, df=5, 542, p>0.30). Subjects in all of the broad

Am J Psychiatry 163:6, June 2006 995
TEICHER, SAMSON, POLCARI, ET AL.
ajp.psychiatryonline.org
abuse categories had ratings that were higher than those of
subjects who had never experienced maltreatment. Emo-
tional abuse had a moderately large effect on the rating of
limbic irritability (d′=0.703, 95% CI=0.548–0.858; t=4.95,
df=300, p<10
–5
). Subjects who were exposed only to emo-
tional abuse had ratings that were as high as those of sub-
jects who were exposed only to physical abuse (t=0.72, df=
85, p>0.40) or only to sexual abuse (t=–0.83, df=102,
p>0.40).
Subjects who were exposed to two different categories
of abuse had higher scores for limbic irritability than sub-
jects exposed to emotional abuse only (t=2.65, df=161,
p<0.01) or physical abuse only (t=3.07, df=144, p<0.003).
Subjects who reported exposure to all three categories of
abuse had significantly higher scores than those exposed
to any single category of abuse.
Limbic irritability was also significantly affected by ex-
posure to any of the more specific types of abusive experi-
ence within each broad abuse category (verbal, nonfamil-
ial sexual, etc.) (F=4.44, df=8, 372, p<10
–4
). Exposure to
verbal abuse had a relatively large effect size (d′=0.876,
95% CI=0.641–1.111), whereas witnessing domestic vio-
lence had a moderate effect size (d′=0.403, 95% CI=0.189–
0.617), although this difference could have occurred by
chance (t=1.49, df=42, p=0.15). It is interesting that com-
bined exposure to verbal abuse and witnessing domestic
violence had a very large effect size (d′=1.428, 95% CI=
1.063–1.793), which was essentially equal to the effect size
of familial sexual abuse (d′=1.333, 95% CI=1.043–1.623).
Effects on Dissociative Experiences
As seen in Figure 2, there were robust effects of the
broad abuse categories on dissociation (F=13.63, df=5,
540, p<10
–11
, n
2
=0.105), and they did not differ between
genders (F=1.88, df=5, 540, p<0.10). Emotional abuse had
a large effect on Dissociative Experience Scale ratings (d′=
0.910, 95% CI=0.753–1.067). Exposure to physical abuse
and sexual abuse had moderate effects (d′=0.533, 95% CI=
0.351–0.715, and d′=0.659, 95% CI=0.505–0.813, respec-
tively). Exposure to two different broad categories of
abuse was associated with a large effect size (d′=0.902,
95% CI=0.783–1.021) that was similar to the effect size for
FIGURE 1. Effects of Childhood Maltreatment on Limbic Ir-
ritability Scores of 554 Young Adults
a
Based on comparison of subjects with each category of maltreat-
ment and a comparison group (N=250) with no history of exposure
to abuse or other form of early adversity. The Limbic System Check-
list-33 (13) was created to evaluate the frequency with which sub-
jects experience symptoms often encountered as phenomena of ic-
tal temporal lobe epilepsy.
Emotional
Only (N=52)
Domestic
violence (N=24)
Verbal
abuse (N=20)
Both (N=8)
Physical
Only (N=35)
Nonfamilial (N=7)
Familial (N=18)
Both (N=10)
Sexual
Only (N=53)
Nonfamilial (N=40)
Familial (N=13)
Any Two
Categories (N=111)
All Three
Categories (N=53)
Size of Effect (Cohen's d′)
on Composite Score on Limbic
System Checklist-33 (±95% CI)
a
0.5 1.5
Type of Childhood
Maltreatment
2.00.0 1.0
FIGURE 2. Effects of Childhood Maltreatment on Dissocia-
tive Experience Scale Scores of 554 Young Adults
a
Based on comparison of subjects with each category of maltreat-
ment and a comparison group (N=249) with no history of exposure
to abuse or other form of early adversity.
Emotional
Only (N=52)
Domestic
violence (N=24)
Verbal
abuse (N=20)
Both (N=8)
Physical
Only (N=35)
Nonfamilial (N=7)
Familial (N=18)
Both (N=10)
Sexual
Only (N=53)
Nonfamilial (N=40)
Familial (N=13)
Any Two
Categories (N=111)
All Three
Categories (N=53)
Size of Effect (Cohen's d′)
on Score on Dissociative
Experience Scale (±95% CI)
a
1.0 4.00.0 2.00.5 3.51.5
Type of Childhood
Maltreatment

996 Am J Psychiatry 163:6, June 2006
CHILDHOOD MALTREATMENT
ajp.psychiatryonline.org
emotional abuse. Exposure to all three categories was as-
sociated with a very large effect size (d′=1.431, 95% CI=
1.269–3.024) that was greater than the effect of exposure to
physical abuse (t=2.97, df=86, p<0.005) or sexual abuse (t=
2.55, df=104, p<0.02) and marginally greater than the ef-
fect of emotional abuse alone (t=1.92, df=103, p=0.06).
Exposure to verbal abuse alone and witnessing of do-
mestic violence had moderately strong effects on dissoci-
ation ratings (d′=0.690, 95% CI=0.456–0.924, and d′=0.564,
95% CI=0.349–1.343, respectively). Subjects exposed to
both verbal abuse and domestic violence (but no other
form of maltreatment) had Dissociative Experience Scale
scores 4.5 times as high as those of the nonabused sub-
jects (d′=3.719, 95% CI=3.324–4.114). In this limited study
group, the effect of combined exposure to verbal abuse
and witnessing domestic violence had more impact on
dissociation ratings than exposure to familial sexual abuse
(t=2.42, df=19, p=0.02).
Effects on Anxiety Symptoms
As seen in Figure 3, there were robust effects of the
broad abuse categories on anxiety (F=14.36, df=5, 541,
p<10
–12
, n
2
=0.109), which did not differ between genders
(F=1.53, df=5, 541, p<0.20). Emotional abuse alone and
sexual abuse alone had moderate effects (d′=0.510, 95%
CI=0.356–0.664, and d′=0.684, 95% CI=0.530–0.838, re-
spectively). Exposure to physical abuse alone exerted only
a small and nonsignificant effect (d′=0.158, 95% CI=–
0.023–0.339; t=0.81, df=283, p=0.16). Exposure to two dif-
ferent broad categories of abuse was associated with a
large effect size (d′=0.832, 95% CI=0.714–0.950), and expo-
sure to all three abuse categories was associated with a
very large effect size (d′=1.429, 95% CI=1.267–1.591) that
was significantly greater than the effect of exposure to any
single category of abuse.
Exposure to verbal abuse alone and witnessing do-
mestic violence had relatively weak effects on anxiety
ratings (d′=0.422, 95% CI=0.189–0.655, and d′=0.158,
95% CI=–0.056–0.372, respectively). Combined exposure
to verbal abuse and witnessing domestic violence had a
greater than additive effect (d′=1.718, 95% CI=1.351–
2.085). Subjects exposed to both verbal abuse and do-
mestic violence (but no other forms) had anxiety scores
that were 2.2 times as high as those of the nonabused
subjects (t=4.76, df=256, p<10
–6
). The effect of combined
exposure to verbal abuse and witnessing of domestic vi-
olence was as great as the effect of exposure to familial
sexual abuse (d′=1.204, 95% CI=0.915–1.493).
FIGURE 3. Effects of Childhood Maltreatment on Ratings of Anxiety, Depression, and Anger-Hostility of 554 Young Adults
a
Based on comparison of subjects with each category of maltreatment and a comparison group (N=250) with no history of exposure to abuse
or other form of early adversity. Kellner’s Symptom Questionnaire (16) is a yes/no self-report questionnaire that is sensitive to subtle differ-
ences from normal.
Emotional
Only (N=52)
Domestic
violence (N=24)
Verbal
abuse (N=20)
Both (N=8)
Physical
Only (N=35)
Nonfamilial (N=7)
Familial (N=18)
Both (N=10)
Sexual
Only (N=53)
Nonfamilial (N=40)
Familial (N=13)
Any Two
Categories (N=111)
All Three
Categories (N=53)
–0.5 0.5 1.5
Size of Effect (Cohen's d′) on Score on Measure From Kellner's Symptom Questionnaire (±95% CI)
a
Anxiety Depression Anger-Hostility
2.50.0 1.0 2.0
–0.5 0.5 1.5 2.50.0 1.0 2.0 –0.5 0.5 1.5 2.50.0 1.0 2.0
Type of Childhood
Maltreatment

Am J Psychiatry 163:6, June 2006 997
TEICHER, SAMSON, POLCARI, ET AL.
ajp.psychiatryonline.org
Effects on Depression Symptoms
There were robust effects of the broad abuse categories
on depression (F=15.91, df=5, 541, p<10
–12
, n
2
=0.121), and
they were consistent across genders (F=0.58, df=5, 541,
p>0.70). Emotional abuse alone and sexual abuse alone
had large effects (d′=0.804, 95% CI=0.648–0.960, and d′=
0.971, 95% CI=0.815–1.127, respectively). Exposure to
physical abuse alone exerted a moderate effect (d′=0.452,
95% CI=0.271–0.633). Exposure to all three categories ex-
erted a very large effect (d′=1.696, 95% CI=1.530–1.862),
which was significantly greater than the effect of exposure
to any single category of abuse.
Exposure to verbal abuse alone and witnessing of do-
mestic violence had moderately strong effects on depres-
sion (d′=0.730, 95% CI=0.496–0.964, and d′=0.463, 95% CI=
0.249–0.677, respectively). Combined exposure to verbal
abuse and witnessing domestic violence had a greater
than additive deleterious effect (d′=2.042, 95% CI=1.672–
2.412). Subjects who were exposed to this combined type
of emotional abuse had depression scores that were 2.8
times as high as those of the nonabused subjects (t=5.66,
df=256, p<10
–8
). The effect of combined exposure to verbal
abuse and witnessing domestic violence was greater than
or equal to the effect of exposure to familial sexual abuse
(d′=1.494, 95% CI=1.202–1.786) or exposure to all three
categories of abuse (d′=1.696, 95% CI=1.530–1.862).
Effects on Symptoms of Anger-Hostility
There were robust effects of the broad abuse categories
on anger-hostility (F=15.18, df=5, 541, p<10
–12
, n
2
=0.119).
Emotional abuse alone and sexual abuse alone exerted
moderate effects (d′=0.630, 95% CI=0.475–0.785, and d′=
0.520, 95% CI=0.367–0.673, respectively). Physical abuse
alone exerted only a weak and not statistically significant
effect (d′=0.131, 95% CI=–0.049–0.311; t=0.73, df=283,
p>0.40) and was eclipsed by the effects of exposure to
emotional abuse (t=2.08, df=85, p=0.04). Combined expo-
sure to any two categories of abuse had a large effect size
(d′=0.851, 95% CI=0.733–0.969), and exposure to all three
categories had an even larger effect size (d′=1.121, 95% CI=
0.963–1.279).
Exposure to verbal abuse alone had a moderately strong
effect on anger-hostility (d′=0.758, 95% CI=0.523–0.993),
and these subjects had scores that were 58% greater than
those of the nonabused individuals (t=3.25, df=268,
p<0.002). Witnessing domestic violence had a relatively
weak effect on these symptoms (d′=0.327, 95% CI=0.113–
0.541), and the scores of subjects with this childhood ex-
perience were only 25% higher than those of the compari-
son subjects (t=1.53, df=272, p=0.13). Combined exposure
to verbal abuse and witnessing domestic violence had an
additive deleterious effect (d′=1.156, 95% CI=0.793–1.519).
The effect of combined exposure to verbal abuse and do-
mestic violence was at least as great as the effect of expo-
sure to familial sexual abuse (d′=0.427, 95% CI=0.142–
0.712) and was comparable to the effect of exposure to all
three types of abuse (d′=1.121, 95% CI=0.963–1.279).
History of Treatment
The impact of multiple exposures to different forms of
maltreatment was also mirrored in the percentage of sub-
jects reporting a past history of psychiatric treatment (χ
2
=
40.4, df=5, p<10
–7
). Previous psychiatric care was reported
by 3% of the subjects with no history of maltreatment, 0%
of those with a history of physical abuse alone, 8% of those
who were exposed to emotional maltreatment, and 9% of
those who were exposed to sexual abuse. In contrast, 18%
of the subjects exposed to any two types of maltreatment
and 25% of those exposed to all three categories reported a
past history of psychiatric care.
Discussion
Childhood exposure to parental verbal aggression was
associated, by itself, with moderate to large effects on
measures of dissociation, limbic irritability, depression,
and anger-hostility. Exposure to verbal aggression was as-
sociated with numerically larger effects on scores on the
Limbic System Checklist-33 and Kellner Symptom Ques-
tionnaire than was exposure to domestic violence, al-
though these differences could have occurred by chance.
Combined exposure to verbal abuse and witnessing of do-
mestic violence was associated with extraordinarily large
adverse effects, particularly on dissociation. This finding
is consonant with studies that suggest that emotional
abuse may be a more important precursor of dissociation
than is sexual abuse (18).
These findings raise the possibility that exposure to verbal
aggression may be a stressor that affects the development of
certain vulnerable brain regions in susceptible individuals,
resulting in psychiatric sequelae (1). Alternatively, exposure
to verbal aggression in childhood may put into force a pow-
erful negative model for interpersonal communication,
which is then incorporated as a behavioral response in fu-
ture relationships. Toth and Cicchetti (19) proposed a cas-
cade of interpersonal events in maltreated children that be-
gins with insecure attachment relationships, moves to
negative representational models of the self and of the self in
relation to others, and eventuates in impaired perceived
competence, poorer social functioning, and lowered self-es-
teem. Similarly, Crittenden (20) found that exposure to
abuse and neglect affects attachment patterns and coping
strategies. These possibilities are not mutually exclusive,
and each could contribute in important ways.
In the present study group, with our definitions of abuse
it appeared that emotional maltreatment was more closely
associated with psychiatric sequelae than was physical
abuse. It is possible that we would have observed a more
deleterious effect of physical abuse if we had adopted a
stricter definition. We are conducting further studies with
additional measures to test this possibility.

998 Am J Psychiatry 163:6, June 2006
CHILDHOOD MALTREATMENT
ajp.psychiatryonline.org
Combined exposure to different categories of abusive
experiences often equaled or exceeded the impact of ex-
posure to familial sexual abuse. This is of great importance
as it suggests that combined exposure to less blatant
forms of abuse may be just as deleterious as the most egre-
gious acts we confront. Fifty-nine percent of the subjects
in the present study with a history of maltreatment had
been exposed to more than one type of abuse.
These findings are concordant with the results of a large-
scale epidemiological study designed to assess the preva-
lence and health impact of early trauma experiences, the
Adverse Childhood Experiences Study (21). All of the publi-
cations resulting from this work point to a relationship be-
tween self-reported early exposure to adversity and subse-
quent problems, including depression (9), attempted
suicide (10), substance abuse, and an array of medical dis-
orders. A dose-response relationship was observed such
that the greater the number of childhood adverse experi-
ences, the greater the risk for a negative health outcome in
adulthood. Macfie et al. (22) also reported that exposure to
multiple forms of trauma, along with severity and chronic-
ity, was predictive of subsequent psychopathology.
The mechanisms underlying the additive or synergistic
effects of exposure to different types of abuse are un-
known. It is possible that an additive or greater effect may
emerge because abuse at home could prevent a child from
seeking help or reassurance from his or her parents when
confronted with abuse outside the home. Alternatively, in-
dividuals exposed to different types of abuse may experi-
ence them at different developmental stages, which would
increase the likelihood that abuse occurred during key
sensitive periods. A third hypothesis is that exposure to
multiple types of abuse increases the frequency of expo-
sure and that, in addition to genetic factors, a certain min-
imal number of exposures is necessary for the develop-
ment of an adverse outcome (23).
The present study is limited by our reliance on self-re-
ports. How do we know if the participants’ abuse histories
are valid? We are sensitized to this issue by the debate sur-
rounding false or repressed memories. However, the issue
of repressed memories is unlikely to be a significant factor
in this research protocol as the events reported were cur-
rent memories and the study provided no incentive for the
subjects to fabricate a history of abuse, as they were never
informed of our screening criteria. While we are con-
cerned about potential fabrications, research suggests
that the overall bias is in the opposite direction—individu-
als are more likely to minimize or deny their adverse child-
hood experiences (24). Positive reports of maltreatment
can be corroborated (25). While retrospective self-report
studies constitute the vast bulk of the literature on the ef-
fects of early abuse on adults, prospective studies that
confirm its impact are emerging (23, 26).
We cannot exclude the possibility that individuals who
have a relatively high degree of current psychiatric symp-
toms may report aspects of their childhood in a more neg-
ative light than do individuals who are free of such symp-
toms (27). It is also possible that exposure to familial
emotional, physical, or sexual abuse is highest in families
with mental illness, and thus, genetic factors could con-
tribute to the higher symptom scores we observed in our
subjects with exposure to familial abuse. Studies of twins
discordant for childhood sexual abuse provide a poten-
tially powerful means of assessing the respective contribu-
tion of childhood sexual abuse while controlling for ge-
netic vulnerabilities and shared environment. Kendler et
al. (28) reported that the twins with childhood sexual
abuse had an overall increased risk for major depression
and a substantially increased sensitivity to the depres-
sogenic effects of other stressful life events.
We have used the term “association” to describe the re-
lationship between symptom ratings and retrospective
self-reports of abuse. While we hypothesize that there may
be a causal relationship, there are other legitimate ways to
interpret the data. Indeed, it may be the case that the rela-
tionship between abuse reports and symptoms is due to a
combination of direct effects of early stress, recall bias,
and increased genetic load.
Another limitation is that our probe question for expo-
sure to domestic violence was very broad, and so the defi-
nition of domestic violence may not be comparable to def-
initions in studies that specifically focused on observations
of mothers being battered. We know from more detailed
assessment of 54 individuals (16 men and 38 women,
mean age=20.8 years, SD=1.1) who responded positively to
this general probe question on the witnessing of serious
domestic violence that 65% had witnessed their mothers
being threatened or assaulted, 43% had witnessed siblings
Patient Perspectives
Angela was an 18-year-old college freshman who en-
rolled in this study after spotting our advertisement on a
subway car. Following a thorough telephone intake inter-
view, she was mailed a study booklet for completion. The
booklet contained questions concerning sociodemo-
graphic status, as well as standardized scales and ques-
tions concerning physical abuse, sexual abuse, domestic
violence, verbal aggression, psychiatric symptoms, and
parenting behaviors. The booklet concluded with a com-
ments section, in which she wrote, “This survey took a lot
longer to fill out than I thought it would. This is the first
time I have thought about these things in years and the
first time I have talked about it.”
We followed up with a personal contact to explore her
history further. We suggested to Angela that her ability to
reveal these events in her responses might be an indica-
tor that she was ready now to talk about them, and we
gave her the contact information for counseling services
at her university, as well as an invitation to be connected
with one of our clinicians.

Am J Psychiatry 163:6, June 2006 999
TEICHER, SAMSON, POLCARI, ET AL.
ajp.psychiatryonline.org
being threatened or assaulted, and 17% had witnessed
threats or assaults of their fathers. Further, 24% had wit-
nessed the severe beating of the person involved. The ef-
fect of exposure to domestic violence might have been
greater had we used a more specific and limited definition.
Finally, representative sampling techniques were not
used. This means that the impact of maltreatment ob-
served may not generalize to other groups. Sixty-five sub-
jects from the group were recruited for additional studies
and went through structured diagnostic interviews, neu-
ropsychological testing, and imaging protocols. They had
an average Hollingshead two-factor socioeconomic status
of 4.0 (SD=0.8) (upper middle class).
In all likelihood, the overall degree of psychopathology
was probably lower in our healthy, predominantly colle-
giate study group than would be found in a representative
sample, although the relative effects of exposure to differ-
ent forms of abuse should generalize to other populations.
Verbal abuse was associated with effect sizes that were
numerically greater than those associated with witnessing
domestic violence or familial physical abuse. However, wit-
nessing domestic violence and physical abuse can qualify
as a category A(1) traumatic event necessary for the DSM-
IV diagnosis of posttraumatic stress disorder (PTSD), while
exposure to verbal abuse cannot. We wonder, particularly
in children, if threats to one’s mental integrity and sense of
self can also be traumatizing. Bremner et al. (29) made an
interesting observation: “Surprisingly, emotional abuse
items, such as being often shouted at, appeared to have se-
vere consequences in terms of risk for PTSD.” The specific
role of verbal abuse in the development of PTSD has yet to
be determined. Research is needed to evaluate whether
such exposure is causative or whether it contributes to the
development of PTSD by amplifying the effects of exposure
to traumatic events.
It will likely come as no surprise to clinicians that paren-
tal verbal aggression is associated with psychiatric symp-
toms. The potential effects of exposure to verbal abuse, by
itself and in combination with other forms of abuse, need
to be carefully considered in research studies focusing on
the effects of early experience. Individuals interested in
the welfare of maltreated children should not underesti-
mate the consequences of verbal abuse. Finally, careful at-
tention should be given to the number of different types of
traumatic experiences a child was exposed to, as this may
be even more critical than the specific type of abuse.
Received Feb. 21, 2005; revision received Oct. 6, 2005; accepted
Jan. 27, 2006. From the Department of Psychiatry, Harvard Medical
School, Boston; and the Developmental Biopsychiatry Research Pro-
gram, McLean Hospital. Address correspondence and reprint re-
quests to Dr. Teicher, Developmental Biopsychiatry Research Pro-
gram. McLean Hospital, 115 Mill St., Belmont, MA 02478;
martin_teicher@hms.harvard.edu (e-mail).
Supported by RO1 grants to Dr. Teicher from NIMH (MH-53636, MH-
66222) and the National Institute on Drug Abuse (DA-016934, DA-
017846).
References
1. Teicher MH, Andersen SL, Polcari A, Anderson CM, Navalta CP, Kim
DM: The neurobiological consequences of early stress and child-
hood maltreatment. Neurosci Biobehav Rev 2003; 27:33–44
2. Bernstein DP, Ahluvalia T, Pogge D, Handelsman L: Validity of
the childhood trauma questionnaire in an adolescent psychiat-
ric population. J Am Acad Child Adolesc Psychiatry 1997; 36:
340–348
3. Vissing YM, Straus MA, Gelles RJ, Harrop JW: Verbal aggression
by parents and psychosocial problems of children. Child Abuse
Negl 1991; 15:223–238
4. Johnson JG, Cohen P, Smailes EM, Skodol AE, Brown J, Oldham
JM: Childhood verbal abuse and risk for personality disorders
during adolescence and early adulthood. Compr Psychiatry
2001; 42:16–23
5. Ney PG: Does verbal abuse leave deeper scars: a study of chil-
dren and parents. Can J Psychiatry 1987; 32:371–378
6. Ney PG, Fung T, Wickett AR: The worst combinations of child
abuse and neglect. Child Abuse Negl 1994; 18:705–714
7. Manning C, Cheers B: Child abuse notification in a country
town. Child Abuse Negl 1995; 19:387–397
8. Saulsbury FT, Campbell RE: Evaluation of child abuse reporting
by physicians. Am J Dis Child 1985; 139:393–395
9. Edwards VJ, Holden GW, Felitti VJ, Anda RF: Relationship be-
tween multiple forms of childhood maltreatment and adult
mental health in community respondents: results from the Ad-
verse Childhood Experiences Study. Am J Psychiatry 2003; 160:
1453–1460
10. Dube SR, Anda RF, Felitti VJ, Chapman DP, Williamson DF, Giles
WH: Childhood abuse, household dysfunction, and the risk of
attempted suicide throughout the life span: findings from the
Adverse Childhood Experiences Study. JAMA 2001; 286:3089–
3096
11. Stein MB: Hippocampal volume in women victimized by child-
hood sexual abuse. Psychol Med 1997; 27:951–959
12. Anderson CM, Teicher MH, Polcari A, Renshaw PF: Abnormal T2
relaxation time in the cerebellar vermis of adults sexually
abused in childhood: potential role of the vermis in stress-en-
hanced risk for drug abuse. Psychoneuroendocrinology 2002;
27:231–244
13. Teicher MH, Glod CA, Surrey J, Swett C Jr: Early childhood abuse
and limbic system ratings in adult psychiatric outpatients. J
Neuropsychiatry Clin Neurosci 1993; 5:301–306
14. Spiers PA, Schomer DL, Blume HW, Mesulam MM: Temporolim-
bic epilepsy and behavior, in Principles of Behavioral Neurol-
ogy. Edited by Mesulam MM. Philadelphia, FA Davis, 1985, pp
289–326
15. Bernstein EM, Putnam FW: Development, reliability and valid-
ity of a dissociation scale. J Nerv Ment Dis 1986; 174:727–735
16. Kellner R: A symptom questionnaire. J Clin Psychiatry 1987; 48:
268–273
17. Cohen J: Statistical Power Analysis for the Behavioral Sciences.
New York, Academic Press, 1988
18. Sar V, Tutkun H, Alyanak B, Bakim B, Baral I: Frequency of dis-
sociative disorders among psychiatric outpatients in Turkey.
Compr Psychiatry 2000; 41:216–222
19. Toth SL, Cicchetti D: Patterns of relatedness, depressive symp-
tomatology, and perceived competence in maltreated chil-
dren. J Consult Clin Psychol 1996; 64:32–41
20. Crittenden PM: Children’s strategies for coping with adverse
home environments: an interpretation using attachment the-
ory. Child Abuse Negl 1992; 16:329–343
21. Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM, Ed-
wards V, Koss MP, Marks JS: Relationship of childhood abuse
and household dysfunction to many of the leading causes of

1000 Am J Psychiatry 163:6, June 2006
CHILDHOOD MALTREATMENT
ajp.psychiatryonline.org
death in adults: the Adverse Childhood Experiences (ACE)
Study. Am J Prev Med 1998; 14:245–258
22. Macfie J, Cicchetti D, Toth SL: Dissociation in maltreated versus
nonmaltreated preschool-aged children. Child Abuse Negl
2001; 25:1253–1267
23. Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington H,
McClay J, Mill J, Martin J, Braithwaite A, Poulton R: Influence of
life stress on depression: moderation by a polymorphism in
the 5-HTT gene. Science 2003; 301:386–389
24. Brewin CR, Andrews B, Gotlib IH: Psychopathology and early
experience: a reappraisal of retrospective reports. Psychol Bull
1993; 113:82–98
25. Bifulco A, Brown GW, Lillie A, Jarvis J: Memories of childhood
neglect and abuse: corroboration in a series of sisters. J Child
Psychol Psychiatry 1997; 38:365–374
26. Noll JG, Trickett PK, Susman EJ, Putnam FW: Sleep disturbances
and childhood sexual abuse. J Pediatr Psychol 2005; 0(June
15):401 (epub ahead of print)
27. Pope HG Jr, Hudson JI: Does childhood sexual abuse cause
adult psychiatric disorders? essentials of methodology. J Psy-
chiatry Law 1995; 23:363–381
28. Kendler KS, Kuhn JW, Prescott CA: Childhood sexual abuse,
stressful life events and risk for major depression in women.
Psychol Med 2004; 34:1475–1482
29. Bremner JD, Vermetten E, Mazure CM: Development and pre-
liminary psychometric properties of an instrument for the
measurement of childhood trauma: the Early Trauma Inven-
tory. Depress Anxiety 2000; 12:1–12