The Limits of Child Effects: Evidence for Genetically Mediated Child
Effects on Corporal Punishment but Not on Physical Maltreatment
Sara R. Jaffee
University of Pennsylvania and Institute of Psychiatry, King’s
Avshalom Caspi and Terrie E. Moffitt
Institute of Psychiatry, King’s College London and University
Institute of Psychiatry, King’s College London
Thomas S. Price
Wellcome Trust Centre for Human Genetics, Oxford University
Institute of Psychiatry, King’s College London
Research on child effects has demonstrated that children’s difficult and coercive behavior provokes harsh
discipline from adults. Using a genetically sensitive design, the authors tested the limits of child effects
on adult behavior that ranged from the normative (corporal punishment) to the nonnormative (physical
maltreatment). The sample was a 1994–1995 nationally representative birth cohort of 1,116 twins and
their families who participated in the Environmental Risk Longitudinal Study. Results showed that
environmental factors accounted for most of the variation in corporal punishment and physical maltreat-
ment. However, corporal punishment was genetically mediated in part, and the genetic factors that
influenced corporal punishment were largely the same as those that influenced children’s antisocial
behavior, suggesting a child effect. The authors conclude that risk factors for maltreatment are less likely
to reside within the child and more likely to reside in characteristics that differ between families.
For much of the last century, research on parent disciplinary
practices has described how parents influence their children’s
development but has failed to consider how children’s behavior
might simultaneously influence the nature of their interactions
with parents. Only in the past 30 years have researchers begun to
explore the bidirectional nature of parent–child relations and con-
cluded that much of what parents do is a response to children’s
behavior (e.g., Anderson, Lytton, & Romney, 1986; Bell & Chap-
man, 1986; Maccoby & Martin, 1983; O’Connor, 2002; Patterson,
Reid, & Dishion, 1992). More recently still, some behavioral
geneticists have questioned whether parents have any influence on
their children’s behavior beyond that which is transmitted genet-
ically (Harris, 1998; Rowe, 1994). This article asks the following
questions: What are the limits of children’s influence on adults’
behavior? and To what degree do children’s coercive and disrup-
tive behaviors elicit responses from adults that range from physi-
cally punitive to abusive?
A number of well-designed prospective studies have found that
children who are physically disciplined or maltreated are at in-
creased risk of engaging in violent antisocial behavior in childhood
and adulthood (Cicchetti & Manly, 2001; Gershoff, 2002; Lans-
ford et al., 2002; Widom, 1989). However, other studies have
shown that children whose behavior is difficult and coercive elicit
more harsh and inconsistent disciplinary responses from adults,
suggesting a reciprocal cycle in which children’s antisocial behav-
ior elicits a punitive response from adults that, in turn, increases
the likelihood that the child will engage in further antisocial
behavior (Cohen & Brook, 1995; Kandel & Wu, 1995; Patterson et
al., 1992). Thus, the association between children’s experience of
corporal punishment or maltreatment and their antisocial behavior
may reflect an effect of punitive parenting practices on child
Sara R. Jaffee, Department of Psychology, University of Pennsylvania,
and Social, Genetic, and Developmental Psychiatry Centre, Institute of
Psychiatry, King’s College London, London, England; Avshalom Caspi
and Terrie E. Moffitt, Social, Genetic, and Developmental Psychiatry
Centre, Institute of Psychiatry, King’s College London, London, England,
and Department of Psychology, University of Wisconsin—Madison;
Monica Polo-Tomas and Alan Taylor, Social, Genetic, and Developmental
Psychiatry Centre, Institute of Psychiatry, King’s College London, Lon-
don, England; Thomas S. Price, Bioinformatics and Statistical Genetics,
Wellcome Trust Centre for Human Genetics, Oxford University, Oxford,
We are grateful to the Environmental Risk (E-Risk) Longitudinal Twin
Study mothers and fathers, the twins, and the twins’ teachers for their
participation. Our thanks to Michael Rutter and Robert Plomin for their
contributions; to Thomas Achenbach for kind permission to adapt the Child
Behavior Checklist; to Hallmark Cards for their support; and to members
of the E-Risk team for their dedication, hard work, and insights. The
E-Risk Study is funded by Medical Research Council Grant G9806489.
Terrie E. Moffitt is a Royal Society–Wolfson Research Merit Award
Correspondence concerning this article should be addressed to Sara R.
Jaffee, Department of Psychology, University of Pennsylvania, 3720 Wal-
nut Street, Philadelphia, PA 19104. E-mail: email@example.com
2004, Vol. 40, No. 6, 1047–1058
Copyright 2004 by the American Psychological Association
behavior, an effect of child behavior on parenting practices, or a
reciprocal process of mutual influence.
Behavioral geneticists have expanded research on child effects
to show that it is often the heritable characteristics of children that
elicit particular disciplinary responses from adults. Evidence for
child effects on parenting comes from twin and adoption studies
that differ with respect to the age of the sample, the measure of
parenting, and the design of the study but show consistently that a
wide range of parenting behaviors, as reported by children, par-
ents, and observers, are heritable (for reviews, see O’Connor,
2002; Plomin & Bergeman, 1991; Reiss, Neiderhiser, Hethering-
ton, & Plomin, 2000). In the language of quantitative behavioral
genetics, parenting practices are said to be “genetically mediated”
by heritable characteristics of the child.1If, for example, monozy-
gotic twins are disciplined more similarly than dizygotic twins, it
means that genetically influenced characteristics of the child ex-
plain why some children are disciplined more often than others.
For instance, a child whose genotype directly or indirectly (e.g., in
combination with environmental stressors) increases the likelihood
that she or he will engage in disruptive oppositional behavior may
be more likely to elicit a punitive response from adults than a child
who does not share that genotype (or the environmental stressors
that potentiate the behavioral expression of that genotype). In their
seminal work on the interplay between genes and environments,
Scarr and McCartney (1983) referred to this process as an “evoc-
ative” or “active” gene–environment correlation whereby children
evoke a response from and thereby create their own environments.
As may be clear, these findings from behavioral genetic studies
showing genetic influences on parenting practices are a special
case of the child effects developmental psychologists have been
discussing for several decades.
Although a number of studies have demonstrated genetic influ-
ences on parent control strategies, relatively few of these have (a)
identified what it is that children do to evoke punitive responses
from adults, (b) shown that the genetic influences on children’s
behaviors are the same as the genetic influences on adults’ control
strategies, or (c) probed the limits of children’s influence on
adults’ behavior toward them. Exceptions include findings from
two adoption studies (Ge et al., 1996; O’Connor, Deater-Deckard,
Fulker, Rutter, & Plomin, 1998). In a study of 45 adolescent
adoptees and their families, Ge and colleagues (Ge et al., 1996)
focused on adoptees who were presumed to be at genetic risk for
antisocial behavior on the basis of their biological parents’ psy-
chopathology. They found that these children elicited more harsh
and inconsistent discipline and less nurturant and involved parent-
ing from their adoptive parents compared with adoptees whose
biological parents did not have a history of disorder. Although the
data were cross-sectional, this study also found evidence of recip-
rocal parent–child effects: Adoptees’ antisocial behavior was in-
fluenced by and was an influence on mothers’ (but not fathers’)
negative parenting. In a study of 88 adoptive families followed
longitudinally from middle childhood to early adolescence,
O’Connor and colleagues (O’Connor et al., 1998) found that
children’s genetically influenced externalizing problems elicited
negative control from parents (i.e., guilt induction, hostility, with-
drawal from relationship). However, the magnitude of genetic
mediation was modest, and evidence was also found for a non-
genetically-mediated effect of children’s externalizing behavior on
parents’ negative control. In addition, analyses of twin data have
shown that the association between parents’ conflictual negative
behavior toward their children and children’s antisocial behavior
was primarily accounted for by genetic factors, even controlling
for the continuity of parent and child behavior over time (Neider-
hiser, Reiss, Hetherington, & Plomin, 1999; Pike, McGuire, Reiss,
Hetherington, & Plomin, 1996).
The present study is concerned with the degree to which heri-
table characteristics of children—specifically, their antisocial be-
havior—are associated with adults’ behaviors that range in sever-
ity from corporal punishment to physical maltreatment. Corporal
punishment is defined as “the use of physical force with the
intention of causing a child to experience pain but not injury for
the purposes of correction or control of the child’s behavior”
(Straus, 1994, p. 4). It differs from maltreatment because it does
not result in significant physical injury (Gershoff, 2002). Whereas
child maltreatment is relatively uncommon, corporal punishment is
a normative disciplinary practice among parents of young children,
with 94% of American parents reporting that they had spanked
their children by the time the children were 3 or 4 years old (Straus
& Stewart, 1999) and over 90% of United Kingdom parents
reporting that they had physically disciplined their children by the
time the children were 11 years old (Nobes, Smith, Upton, &
Heverin, 1999). Corporal punishment is administered most com-
monly among children ranging in age from 5–8 years (Gallup,
Moore, & Schussel, 1995); children studied for this article were 5
Whereas children’s coercive and disruptive behavior may pro-
voke normative control strategies like corporal punishment, it is
less likely to elicit physical maltreatment, risk for which is most
strongly influenced by characteristics of the perpetrators and by
the social context in which families reside (Azar, 2002; Belsky,
1993; Jaffee, in press). Consistent with the possibility that char-
acteristics of adults and family environments, but not of children,
influence the likelihood that children will be maltreated, a twin
study showed that similarities between children in their experience
of maltreatment were not accounted for by their genetic similarity
(Jaffee, Caspi, Moffitt, & Taylor, 2004). This finding suggested
that children’s heritable characteristics do not provoke abuse. The
current study extends previous work by (a) investigating genetic
1When estimates of genetic effects on parenting practices are based on
how parents or observers report that parents treat their twin children,
genetic influences on the parenting measure are interpreted as reflecting
heritable child characteristics. Likewise, when estimates of genetic effects
on parenting practices are based on how twin children report that they
themselves have been treated, genetic influences on the parenting measure
also reflect children’s heritable perceptual biases. In contrast, when esti-
mates of genetic effects on parenting are based on how twin adults describe
the way they treat their own children, genetic influences on the parenting
measure are interpreted as capturing heritable characteristics of parents
(Plomin, Reiss, Hetherington, & Howe, 1994). In our examples (and in our
data), reports of parenting are provided by parents or observers with
reference to twin children. Thus, genetic influences on corporal punish-
ment or maltreatment are interpreted as reflecting heritable child charac-
teristics and not heritable characteristics of parents. However, to the degree
that there is a diluted hereditary resemblance between parents and children
(i.e., they share 50% of their genes, on average), genetic influences on
parenting measures like corporal punishment or maltreatment may also
reflect the parents’ genotype to some degree.
JAFFEE ET AL.
and environmental influences on physical maltreatment as well as
on normative control strategies like corporal punishment, (b) by
identifying specific child behaviors that are associated with cor-
poral punishment and physical maltreatment and that may account
for genetic influences on those adult behaviors, and (c) by speci-
fying which etiological factors corporal punishment and physical
maltreatment share in common and which are distinctive. This
investigation explores the limits of child effects on adults’
On the basis of our review of the literature, we predicted that the
association between children’s experience of corporal punishment
and their antisocial behavior would be genetically mediated. Sup-
port for this hypothesis would be found if (a) we detected signif-
icant genetic influences on corporal punishment and (b) we found
that the genetic factors that influenced children’s antisocial behav-
ior were the same as those that influenced corporal punishment and
accounted for a large portion of the phenotypic association be-
tween the two. On the basis of our hypothesis that children do not
influence adults’ behavior outside the normal range, we predicted
that although corporal punishment and physical maltreatment
would have some common causes, heritable characteristics of
children would influence the former but not the latter.
The Environmental Risk (E-Risk) Longitudinal Twin Study
Participants are members of the E-Risk Longitudinal Twin Study, which
investigates how genetic and environmental factors shape children’s de-
velopment. The sampling frame from which the E-Risk families were
drawn was two consecutive birth cohorts (1994 and 1995) in a birth register
of twins born in England and Wales (Trouton, Spinath, & Plomin, 2002).
Of the 15,906 twin pairs born in these 2 years, 71% joined the register.
The E-Risk Study probability sample was drawn using a high-risk
stratification strategy. High-risk families were those in which the mother
had her first birth when she was 20 years of age or younger. We used this
sampling (a) to replace high-risk families who were selectively lost to the
register via nonresponse and (b) to ensure sufficient base rates of problem
behavior, given the low base rates expected for 5-year-old children. Age at
first childbearing was used as the risk-stratification variable because it was
present for virtually all families in the register, it is relatively free of
measurement error, and early childbearing is a known risk factor for
children’s problem behaviors (Maynard, 1997; Moffitt & E-Risk Study
Team, 2002). The high-risk sampling strategy resulted in a final sample in
which one third of study mothers constitute a 160% oversample of mothers
who were at high risk on the basis of their young age at first birth (15–20
years), whereas the other two thirds of study mothers accurately represent
all mothers in the general population (aged 15–48 years) in England and
Wales in 1994–1995. Demographic estimates are derived from the General
Household Survey (Bennett, Jarvis, Rowlands, Singleton, & Haselden,
1996). To provide unbiased statistical estimates from the whole sample that
can be generalized to the population of British families with children born
in the 1990s, the data reported in this article were corrected with weighting
to represent the proportion of maternal ages in that population.
The E-Risk Study sought a sample size of 1,100 families to allow for
attrition in future years of the longitudinal study while retaining statistical
power. An initial list of families who had same-sex twins was drawn from
the register to target for home visits, with a 10% oversample to allow for
nonparticipation. Of the 1,203 families from the initial list who were
eligible for inclusion, 1,116 (93%) participated in home-visit assessments
when the twins were age 5 years, forming the base sample for the study;
4% of families refused, and 3% were lost to tracing or could not be reached
after many attempts. With parent’s permission, questionnaires were mailed
to the children’s teachers, and teachers returned questionnaires for 94% of
cohort children. Written informed consent was obtained from mothers. The
E-Risk Study has received ethical approval from the Maudsley Hospital
Zygosity was determined using a standard zygosity questionnaire, which
has been shown to have 95% accuracy (Price et al., 2000). Ambiguous
cases were zygosity-typed using DNA. The sample includes 55% monozy-
gotic (MZ) twins and 45% dizygotic (DZ) twins. Sex is evenly distributed
within zygosity (49% male).
Corporal punishment and child physical maltreatment were assessed
separately for each twin by interviewing mothers with the standardized
clinical interview protocol from the Multi-Site Child Development Project
(Dodge, Bates, & Pettit, 1990; Dodge, Pettit, Bates, & Valente, 1995;
Lansford et al., 2002). The same set of questions about nonphysical
discipline, corporal punishment, and physical maltreatment was asked
about each twin, and the interviews about each twin were separated by 1.5
hr of questions on other topics. The interview protocol was designed by
Dodge and colleagues (Dodge et al., 1990, 1995; Lansford et al., 2002) to
enhance mothers’ comfort with reporting valid discipline and child mal-
treatment information while also meeting researchers’ legal and ethical
responsibilities for reporting. Under the United Kingdom Children Act
(Department of Health, 1989), our responsibility was to secure intervention
if maltreatment was current and ongoing. At the start of the interview about
discipline and maltreatment, the interviewer explained to the mother that if
she reported maltreatment that had occurred in the child’s first 4 years and
was not ongoing, that information could remain confidential. However, if
she reported maltreatment that occurred in the year prior to the interview
and the risk to the child was ongoing, the E-Risk Study would be under
legal obligation to assist the family to get help. Thus, when mothers gave
informed consent to proceed with the interview, they understood that a
report of recent ongoing maltreatment would constitute a request for help
(if the maltreatment was not already known to authorities). The interview
did not ask directly about the timing of incidents and therefore mothers
who wished to report maltreatment while avoiding intervention could have
opted to describe maltreatment as happening in the past.
We interviewed mothers about maltreatment instead of ascertaining
cases from Child Protective Service registers for three reasons. First,
official record data identify only a small proportion of cases, which may be
a biased unrepresentative subset (Walsh, McMillan, & Jamieson, 2002;
Widom, 1988). Second, because of time delays in detection, investigation,
and legal proceedings against perpetrators, official record data sources tend
not to record children as confirmed cases until older ages and the children
in our sample were 5-year-olds. Third, searching child protection records
for this sample would have required parental consent, placing record data
at the same potential risk of parental concealment as mothers’ reports.
The discipline and maltreatment interview protocol has (a) good con-
current validity as evidenced by correlations above .60 with mothers’
reports of their child-directed aggression using the Conflict Tactic Scales
(Dodge et al., 1990; Straus & Gelles, 1988), (b) good interreporter reli-
ability as evidenced by a correlation of .60 between mothers’ and fathers’
reports in 396 couples (Dodge et al., 1995), and (c) good predictive validity
as evidenced by significant 12-year prediction from preschool maltreat-
ment to outcomes in Grade 11, including increased violence, school ab-
senteeism, anxiety and depressive symptoms, and posttraumatic stress
disorder symptoms, controlling for a variety of social and family risk
factors (Lansford et al., 2002).
Corporal punishment was assessed by asking mothers whether the child
had experienced a variety of disciplinary practices in the past year, some of
which assessed nonphysical discipline (e.g., “isolation,” “withdrew privi-
THE LIMITS OF CHILD EFFECTS
leges”) and three of which assessed corporal punishment (“grabbing or
shaking,” “smacking or hitting,” or engaging in “other physical disci-
pline”). A score of 1 was assigned if the mother reported that a particular
disciplinary practice had been used with the child, and a score of 0 was
assigned if the disciplinary practice had not been used. If mothers reported
any form of corporal punishment, they were then asked how often in the
past year the child was physically punished, with responses ranging from
0 (never) to 5 (daily). Children who did not receive any corporal punish-
ment were assigned a score of 0 on the frequency variable. Mothers were
then asked these same questions about the variety of disciplinary practices
used with the child and the frequency with which corporal punishment was
used during the child’s first 4 years. We created variety and frequency
variables based on mothers’ reports of corporal punishment across the
children’s first 5 years. The variety variable indexes how many different
kinds of physical disciplinary practices were used (M ? 1.63, SD ? 0.82).
The frequency variable indexes how often, on average, corporal punish-
ment was used with the child (M ? 1.59, SD ? 1.15). The variety and
frequency scores across the child’s first 5 years were highly correlated (r ?
.65, p ? .01). To create the corporal punishment composite variable used
in our analyses, we standardized and summed the variety and frequency
scores (M ? 0.03, SD ? 1.78). Eighty-seven percent of children had
experienced corporal punishment at least once in their first 5 years
(weighted and unweighted prevalences were identical).
Following the questions about nonphysical discipline and corporal pun-
ishment, mothers were asked standardized probe questions about potential
physical maltreatment, including “When [name] was a toddler, do you
remember any time when she/he was disciplined severely enough that
she/he may have been hurt?” and “Did you worry that you or someone else
[such as a babysitter, a relative, or a neighbor] may have harmed or hurt
[name] during those years?” (1% of mothers declined to answer the
questions). Questions were carefully worded to avoid implying that the
mother was the perpetrator so that mothers might feel more willing to
report that a child had been maltreated. In cases where mothers reported
any physical maltreatment, interviewers probed mothers for details about
the incident and recorded notes. Interviewers coded the likelihood that the
child had been maltreated on the basis of the mothers’ narratives. This
classification showed 90% agreement between coders in our study and
adequate intercoder agreement (? ? .56) in Dodge et al.’s study (Dodge,
Pettit, & Bates, 1994; Dodge et al., 1995). The 10% of codes that disagreed
tended to reflect uncertainty about whether physical maltreatment was
probable or definite. All children in the maltreatment group experienced
some sort of physical maltreatment. Many children undoubtedly had ex-
perienced additional sexual abuse, psychological abuse, or neglect, but this
was not systematically ascertained. On the basis of the mother’s report of
the severity of discipline and the interviewer’s rating of the likelihood that
the child had been physically maltreated, children were coded as having not
been, possibly been, or definitely been physically maltreated. Examples of
possible physical maltreatment in our sample (n ? 273 children) included
instances in which the mother reported that she smacked the child harder
than intended and left a mark or bruise or cases in which social services
were contacted by schools, neighbors, and/or family members out of
concern that the child was being physically maltreated. Examples of
definite physical maltreatment included children who were beaten by a
teenaged stepsibling, punished by being burned with matches or thrown
against doors, had injuries (e.g., fractures or dislocations) from neglectful
or abusive care, or were formally registered with a social services child
protection team. The prevalence of such definite serious physical maltreat-
ment as defined in this sample was 1.5% (n ? 34 children). For the
purposes of our analyses, the child physical maltreatment variable was
recoded into a dichotomous variable representing children who experi-
enced no physical maltreatment (unweighted, prevalence ? 86%; weighted
to represent the population, prevalence ? 88%) versus a combined group
of children who experienced possible or definite physical maltreatment
(unweighted, prevalence ? 14%; weighted to represent the population,
prevalence ? 12%). The prevalence of physical maltreatment was similar
among MZ (11%) and DZ (14%) twins. Our combined prevalence of 12%
resembles the 15% prevalence estimate reported by Dodge and colleagues
(Dodge et al., 1990), whose measurement protocol we used. Our preva-
lence rate of 1.5% for definite physical maltreatment is consistent with
physical maltreatment estimates of 1.5% and 2.6% from population sur-
veys in North America (Bland & Orn, 1986; Egami, Ford, Greenfield, &
Children’s antisocial behavior was assessed at 5 years with the Achen-
bach family of instruments (Achenbach, 1991a, 1991b). The Aggression
and Delinquency scales were supplemented with Diagnostic and Statistical
Manual of Mental Disorders, Fourth Edition (American Psychiatric Asso-
ciation, 1994) items assessing conduct and oppositional defiant disorder.
Mother and teacher reports of antisocial behavior correlated .29 (p ? .01),
which is typical of interrater agreement about behavioral problems (Achen-
bach, McConaughy, & Howell, 1987). Mother and teacher reports of
children’s behavior problems were summed (items scored from 0 to 2).
Scores ranged from 0 to 130 (M ? 21.17, SD ? 16.27). The internal
consistency of the combined score was .94.
The results are presented in six sections. In the first section, we
present prevalence rates of corporal punishment and physical mal-
treatment and describe the association between the two as well as
their association with children’s antisocial behavior. In the second,
third, and fourth sections, we test whether children’s experience of
corporal punishment is genetically mediated by their antisocial
behavior. In the remaining sections, we test these hypotheses about
children’s experience of physical maltreatment, and we determine
what accounts for the association between corporal punishment
and physical maltreatment.
Prevalence Rates of Corporal Punishment and
Corporal punishment was far more common among families
than maltreatment; 87% of children had experienced corporal
punishment by the time they were 5 years old, but only 12% had
been possibly or definitely maltreated. Among the children who
had ever received corporal punishment, 87% had never been
maltreated. Among the children who were maltreated, only 6% had
never experienced corporal punishment (suggesting that among the
6%, the abuse perpetrator was not the parent). Although the vast
majority of children who experienced corporal punishment had
never been maltreated, having been physically disciplined in-
creased the odds of having been maltreated by 2.5 times (odds
ratio ? 2.56, 95% confidence interval [CI] ? 1.27 to 5.19, p ?
.01). All associations among corporal punishment, maltreatment,
and child antisocial behavior were statistically significant. The
Pearson correlation between corporal punishment and child anti-
social behavior was .27 (p ? .01). Children who had been phys-
ically maltreated (M ? 29.16, SD ? 19.35) had significantly
higher antisocial behavior scores than children who had not been
maltreated (M ? 20.05, SD ? 15.47; B ? 9.11, SE ? 1.41, ? ?
.18, p ? .01).
Are Children’s Experiences of Corporal Punishment
We conducted maximum-likelihood univariate model-fitting
analyses using Mx software (Neale, Boker, Xie, & Maes, 2002) to
JAFFEE ET AL.
determine the extent to which individual differences in children’s
experience of corporal punishment were accounted for by genetic
and environmental factors (Neale & Cardon, 1992). These behav-
ioral genetic model-fitting analyses rely on the different level of
genetic relatedness between MZ and DZ twin pairs to estimate the
contribution of genetic and environmental factors to individual
differences in a given phenotype. Phenotypes include any behavior
or characteristic that is measured separately for each twin, such as
each twin’s score on a behavior problem checklist or each twin’s
experience of corporal punishment. Population variation on any
phenotype may be partialed into an additive genetic component
and two types of environmental components by using the follow-
ing logic. First, MZ twins share all their genes, but DZ twins, like
all nonidentical siblings, share half of their polymorphous genes,
on average. Polymorphous genes are those associated with differ-
ences among people. For example, genes influencing eye color are
polymorphous, but genes determining that people have eyes are
not. As such, a genetic contribution to antisocial behavior or
corporal punishment is indicated when the similarity of MZ twins
is greater than the similarity of DZ twins. In model-fitting terms,
this yields a significant variance component called A (additive
genetic variance). Second, MZ twins’ genetic similarity is twice
that of DZ twins, and therefore, if nothing more than genes were
influencing behavior, then MZ twins should be at least twice as
similar as DZ twins. If not, this indicates that something more than
genes has made the twins similar (i.e., environments that the
siblings share in common must have enhanced their similarity). In
model-fitting terms, this yields a significant variance component
called C (common, or shared, environmental variance). Third, if
MZ twins, despite sharing all their genes, are not perfectly iden-
tical in their experience of corporal punishment or antisocial
behavior, this indicates that nonshared experiences unique to each
family member reduce their similarity. In model-fitting terms, this
yields a significant variance component called E (child-specific
environmental variance as well as measurement error).
Because the latent variables A, C, and E are unmeasured, they
do not have a natural scale and must be assigned a variance (i.e.,
the variance is fixed at 1.0). The goal of fitting different structural
equations to twin data is to account for the observed covariance
structure using the most parsimonious number of parameters. To
compare the fit of different models, we used two model-selection
statistics. The first was the chi-square goodness-of-fit statistic.
Large values indicate poor model fit to the observed covariance
structure. When two models are nested (i.e., identical with the
exception of constraints placed on the submodel), the difference in
fit between them can be evaluated with the chi-square difference,
using as its degrees of freedom the df difference from the two
models. When the chi-square difference is not statistically signif-
icant, the more parsimonious model is selected, as the test indi-
cates that the constrained model fits equally well with the data. The
second model-selection statistic was the root-mean-square error of
approximation (RMSEA), which is an index of the model discrep-
ancy, per degree of freedom, from the observed covariance struc-
ture (MacCallum, Browne, & Sugawara, 1996). Values less than
.05 indicate close fit and values less than .08 indicate fair fit to the
data (Browne & Cudeck, 1993). For detailed explanations of the
statistical methods that operationalize the logic behind behavior
genetic designs, see Plomin, DeFries, McClearn, and McGuffin
The top part of Table 1 presents the results of the univariate
model for corporal punishment. Columns 1 and 2 show the within-
pair twin correlations for children’s experiences of corporal pun-
ishment separately for MZ and DZ twins. The magnitudes of the
twin correlations for exposure to corporal punishment are consis-
tent with those reported by other investigators in both adoption and
twin studies (Deater-Deckard, Fulker, & Plomin, 1999; Wade &
Kendler, 2000). To test the hypothesis that children’s experience
of corporal punishment was genetically mediated, we compared
the full model with one in which genetic factors were constrained
to have no effect on individual differences in corporal punishment.
The constrained model fit significantly worse than the full model,
?2(1, N ? 1,096) ? 90.56, p ? .01, indicating that children’s
experience of corporal punishment was significantly genetically
mediated, in part. The best fitting model indicated that shared
environmental factors accounted for 66% and genetic factors ac-
counted for 25% of the variation in children’s experience of
corporal punishment, with nonshared environmental influences
and measurement error accounting for the remainder of the
Within-Pair Twin Correlations for Corporal Punishment, Maltreatment, and Child Antisocial Behavior and Results of Univariate
Population variance attributable Model fit statistics
A ? 0
A ? 0
A ? 0
.90*.79*.25 (.19–.31) .66 (.59–.71)
.77*.71* .07 (.00–.21)
.75* .30* .73 (.64–.76)
in which genetic factors are hypothesized to have no effect (A ? 0). MZ ? monozygotic; DZ ? dizygotic; A ? additive genetic effect; C ? common
environmental effect; E ? unique environmental effect; RMSEA ? root-mean-square error of approximation.
* p ? .01.
Numbers in parentheses represent the 95% confidence intervals. ?2
differenceis the difference in fit between the full model and the reduced model
THE LIMITS OF CHILD EFFECTS
Is Corporal Punishment Genetically Mediated by
Children’s Genetic Predisposition for Antisocial
The hypothesis that corporal punishment is genetically mediated
by children’s antisocial behavior requires that children’s antisocial
behavior be heritable. Previous work in this sample has shown that
the heritability of 5-year-old children’s antisocial behavior ranges
from 33% to 71%, depending on the informant, and that the
heritability of antisocial behavior that is pervasive across settings
is 82% (Arseneault et al., 2003). The bottom part of Table 1 shows
the estimates of genetic and environmental influences on the
measure of antisocial behavior used in this report. To test whether
common genetic factors accounted for the association between
children’s experience of corporal punishment and their antisocial
behavior, we specified a common factors model (Neale & Cardon,
1992). The common factors model tests whether there are genetic
and environmental factors that influence corporal punishment and
child antisocial behavior or whether the genetic or environmental
factors that influence corporal punishment are largely distinct from
those that influence children’s antisocial behavior.
To estimate the common factors model, a Cholesky decompo-
sition model was initially fit using Mplus software (Muthe ´n &
Muthe ´n, 1998). In the Cholesky model, variation within the first
phenotype (e.g., corporal punishment) is accounted for by latent
additive genetic, shared environmental, and nonshared environ-
mental factors. Variation within the second phenotype (e.g., anti-
social behavior) is accounted for by the same genetic and envi-
ronmental latent factors that influence corporal punishment as well
as another set of latent genetic and environmental factors that
account specifically for variation within antisocial behavior. The
resulting parameters were transformed into those shown in Figure
1 according to the algebraic formulae presented by Loehlin (1996).
Figure 1 presents the results of the common factors model. The
fit of the model was adequate, ?2(11, N ? 1,096) ? 16.67, p ? .12,
RMSEA ? .031, and the results supported the hypothesis that
corporal punishment is genetically mediated by children’s antiso-
cial behavior. That is, the phenotypic correlation between corporal
punishment and antisocial behavior can be recovered by following
tracing rules (Kenny, 1979) and summing the products of the paths
that connect the phenotypes (e.g., .492? .042? .192? .28). To
determine how much of the phenotypic correlation (r ? .28) is
accounted for by genetic factors, the product of the paths connect-
ing the two phenotypes via the latent genetic factors is calculated
and divided by the overall phenotypic correlation (e.g., .492/.28 ?
.86). Thus, genetic factors accounted for 86% of the covariation
between antisocial behavior and corporal punishment, with non-
shared environmental factors accounting for most of the remain-
der. To formally test the hypothesis that children’s experience of
corporal punishment is mediated by genetic factors that also in-
fluence their antisocial behavior, we set the paths from the com-
mon genetic factor to the phenotypes equal to zero. If this model
provided a good fit to the data, it would indicate that the associ-
ation between corporal punishment and child antisocial behavior
could be adequately described without reference to genetic factors
that influence them both. The fit of the reduced model was sig-
nificantly worse than the fit of the full model, ?difference
1,096) ? 49.42, p ? .01. Thus, to the degree that there are genetic
influences on variation in children’s experience of corporal pun-
(1, N ?
Parameter estimates are standardized and must be squared to obtain population variance in each phenotype
attributable to genetic, shared environmental, or nonshared environmental factors. Ac, Cc, and Ec? genetic,
shared environmental, and nonshared environmental factors (respectively) that are common to both corporal
punishment and antisocial behavior; Acpand Aasb? genetic factors that are specific to corporal punishment and
antisocial behavior, respectively; Ccpand Casb? shared environmental factors that are specific to corporal
punishment and antisocial behavior, respectively; Ecpand Easb? nonshared environmental factors that are
specific to corporal punishment and antisocial behavior, respectively.
Common factors model of the association between corporal punishment and antisocial behavior.
JAFFEE ET AL.
ishment, these also account for individual differences in children’s
Addressing Issues of Temporality in the Association
Between Corporal Punishment and Children’s Antisocial
The univariate and bivariate analyses of corporal punishment
showed that there were significant genetic influences on corporal
punishment and that genetic factors accounted for most of the
covariance between corporal punishment and children’s antisocial
behavior. However, these results do not allow us to plausibly test
the hypothesis that children’s antisocial behavior evokes physical
discipline because children’s antisocial behavior was not measured
prior to the assessment of corporal punishment. A partial solution
to this problem was to examine only those instances of corporal
punishment that occurred in the year before children turned 5 years
of age so that our measure of children’s antisocial behavior would
at least be concurrent with (if not temporally prior to) the measure
of corporal punishment. As reported by their mothers, 78% of
children were physically disciplined in the year before they turned
5, but only 14% of those children had ever been physically
maltreated. Similar to the approach described in the Method sec-
tion, we standardized scores representing the variety of disciplin-
ary practices used in the past year (ranging from 1 to 3; M ? 0.82,
SD ? 0.47) and scores representing the frequency of corporal
punishment in the past year (ranging from 0 to 5; M ? 1.61, SD ?
1.26). The Pearson correlation between past-year corporal punish-
ment variety and frequency scores was .64 (p ? .01). The stan-
dardized frequency and variety scores were summed to create a
measure of total past-year physical discipline (M ? –0.02, SD ?
Univariate analyses of past-year corporal punishment replicated
the findings reported in Table 1. Genetic factors accounted for
21% of the variation in past-year corporal punishment (95% CI ?
.13 to .29), with shared environmental factors accounting for an
additional 62% (95% CI ? .54 to .69) and nonshared environmen-
tal factors accounting for the remaining 17% (95% CI ? .15 to .19)
of the variance. Genetic influences on corporal punishment were
significantly different from zero, as indicated by the significant
decrement in model fit when the influence of genetic factors was
hypothesized to be zero, ?difference
The phenotypic correlation between children’s antisocial behav-
ior and past-year corporal punishment was .28 (p ? .01). Results
of the bivariate model replicated those reported in Figure 1. Ge-
netic factors that were common to past-year corporal punishment
and children’s antisocial behavior accounted for 76% of the phe-
notypic correlation between the two. The fit of the bivariate model
deteriorated significantly when common genetic influences on
past-year corporal punishment and antisocial behavior were hy-
pothesized to be zero, ?difference
(1, N ? 1,104) ? 29.78, p ? .01.
(1, N ? 1,104) ? 32.10, p ?
Are Children’s Experiences of Physical Maltreatment
Because maltreatment was a dichotomous variable, analyses of
genetic and environmental influences on maltreatment were con-
ducted using contingency table data and liability threshold models
in Mx (Neale et al., 2002). Although liability threshold models
provide unbiased estimates of genetic and environmental variance
components when the prevalence of the dichotomous outcome is
as low as 10% (prevalence rates of physical maltreatment in our
sample were 12%), they are underpowered to detect genetic and
environmental effects relative to models that account for variation
in continuously distributed outcomes (Neale, Eaves, & Kendler,
1994). To demonstrate that genetic influences on maltreatment are
significant, we must have sufficient power to reject a model of
purely environmental effects on maltreatment (i.e., a CE model) in
favor of a model that freely estimates genetic and environmental
influences on maltreatment. We conducted power analyses follow-
ing recommendations by Neale and colleagues (Neale et al., 1994)
and determined that our model had power of .78 to reject a model
of purely environmental influences on maltreatment (i.e., a CE
model) in favor of a model that hypothesized genetic and envi-
ronmental influences on maltreatment (i.e., an ACE model) if true
heritability was as low as .20. However, power was poor (power ?
.40) to reject a CE model in favor of an ACE model if true
heritability was as low as .10.
The middle part of Table 1 shows genetic and environmental
influences on children’s experience of maltreatment. To test the
hypothesis that children’s experience of maltreatment is geneti-
cally mediated, we compared the fit of the full model (ACE) to one
in which genetic factors were constrained to have no effect on
individual differences in exposure to maltreatment (CE). The full
model did not fit significantly better than a more parsimonious
model in which genetic influences on children’s experience of
maltreatment were constrained to be zero, ?2(1, N ? 1,115) ?
1.72, ns. Thus, unlike children’s experience of corporal punish-
ment, genetic influences were not needed to account for individual
differences in children’s experience of maltreatment. The best-
fitting model indicated that shared environmental factors ac-
counted for 94% of the variance in children’s experience of mal-
measurement error accounting for the remainder. We must note,
however, that given the 95% CI around our point estimate for
genetic influences on maltreatment, it is likely that our model was
underpowered to reject the CE model in favor of a model speci-
fying both genetic and environmental influences on maltreatment.
Identifying Common Precursors of Corporal Punishment
and Physical Maltreatment
The results of the univariate analyses showed that variation in
children’s experience of corporal punishment was influenced pri-
marily by genetic and shared environmental factors and that vari-
ation in children’s experience of maltreatment was influenced
primarily by shared and nonshared environmental factors. These
results alone suggest that influences on corporal punishment and
physical maltreatment differ (e.g., genetic factors influence the
former but not the latter). However, these forms of physical
2Results were unchanged regardless of whether mother or teacher
reports of children’s antisocial behavior were analyzed separately. For
example, the phenotypic correlation between teacher reports of children’s
antisocial behavior and corporal punishment was .15 (p ? .01). Genetic
factors accounted for 80% of this association.
THE LIMITS OF CHILD EFFECTS
discipline are related, as demonstrated by the fact that in our
sample, children who received corporal punishment were at in-
creased risk of experiencing maltreatment. This suggests that cor-
poral punishment and maltreatment must have some similar ori-
gins (or that one causes the other). To formally test for etiological
similarities, we estimated a common factors model of the associ-
ation between corporal punishment and maltreatment. A Cholesky
decomposition model was initially fit to the data using Mplus
software (Muthen & Muthen, 1998). Tetrachoric correlations were
used to estimate the model because of the categorical nature of the
maltreatment variable. The resulting parameters were transformed
into the parameter estimates in Figure 2 using the algebraic for-
mula presented by Loehlin (1996). The fit of the model was
adequate, ?2(14, N ? 1,096) ? 15.26, p ? .36, RMSEA ? .013.
The results demonstrated that 74% of the covariation between
corporal punishment and maltreatment was accounted for by
shared environmental factors that were common to both. To test
whether genetic factors were needed to account for the covariation
of corporal punishment and maltreatment, we set all common
genetic influences equal to zero. The constrained model did not fit
significantly worse than the full model, ?difference
3.36, ns. Thus, corporal punishment and maltreatment co-occur
because of factors that differ between families but not because
both are genetically mediated by child behavior.
(1, N ? 1,096) ?
Implications for Theory
This study showed that the reason some children are more likely
than others to be physically maltreated is largely explained by
factors that differ between families. Similarly, the fact that some
children are more likely than others to be physically disciplined is
predominantly explained by factors that differ between and within
families. These findings highlight the limits of child effects in
much the same way that behavioral genetics research on the family
has highlighted the “limits of family influence” (Rowe, 1994).
Unquestionably, children play a significant role in shaping their
environmental experience (Scarr, 1992; Scarr & McCartney,
1983). Indeed, our results showed that genetic factors played a
significant role in explaining why some children were more likely
than others to experience normative physical disciplinary practices
like spanking. However, children—particularly young children—
are also subjected to certain experiences like maltreatment, over
which they have little control and which have important implica-
tions for their developmental outcomes. We do not suggest that
children can control whether they are spanked or smacked in
response to their misbehavior—only that there is likely to be a
contingent, genetically mediated relation between children’s dif-
ficult coercive behavior and their experience of corporal punish-
ment, whereas there is less likely to be such a contingent relation
between children’s behavior and their experience of maltreatment.
Research on child effects has shown that children’s difficult and
coercive behavior often elicits harsh discipline from parents in the
form of shouting, threatening, or smacking (Reid, Patterson, &
Snyder, 2002). Consistent with these findings, our results showed
that genetic factors accounted for variation in children’s experi-
ence of corporal punishment. Assuming that there is not a “cor-
poral punishment gene,” genetic influences on physical discipline
must reflect heritable characteristics of children. In analyses that
included past-year measures of corporal punishment that were
estimates are standardized and must be squared to obtain population variance in each phenotype attributable to
genetic, shared environmental, or nonshared environmental factors. The effects of Amand Acpare not
statistically significant. Ac, Cc, and Ec? genetic, shared environmental, and nonshared environmental factors
(respectively) that are common to both corporal punishment and maltreatment; Acpand Am? genetic factors that
are specific to corporal punishment and maltreatment, respectively; Ccpand Cm? shared environmental factors
that are specific to corporal punishment and maltreatment, respectively; Ecpand Em? nonshared environmental
factors that are specific to corporal punishment and maltreatment, respectively.
Common factors model of the association between corporal punishment and maltreatment. Parameter
JAFFEE ET AL.
temporally concurrent with measures of child antisocial behavior
and in analyses that included measures of corporal punishment that
spanned the children’s first 5 years of life, we found that children’s
oppositional, aggressive, and coercive behavior was associated
with the use of more frequent and more varied forms of physical
discipline and that the genetic factors that influenced variation in
corporal punishment also accounted for variation in child antiso-
cial behavior. This provides suggestive evidence that heritable
characteristics of children influence parenting in the normal range.
Our finding that both genetic and shared environmental factors
accounted for individual differences in corporal punishment is
consistent with one other study in which adult female twins and
their parents were asked to report retrospectively on how often the
parents had used corporal punishment to discipline each twin when
the twins were children (Wade & Kendler, 2000). Their results
showed that genetic factors accounted for a significant 19% and
shared environmental factors accounted for 66% of the variation in
mother-reported corporal punishment, whereas our results yielded
estimates of 25% and 66%, respectively. However, our research is
the first to explore the genetic and environmental origins of mal-
treatment and to show that corporal punishment and maltreatment
are associated with children’s antisocial behavior for different
reasons. Whereas our findings suggest that children’s genetically
influenced aggressive and difficult-to-manage behavior may pro-
voke incidents of corporal punishment, maltreatment is not genet-
ically mediated by children’s behavior. In contrast, maltreatment
has been shown to influence children’s antisocial behavior directly
(Jaffee et al., 2004) and to exacerbate children’s genetic risk for
antisocial behavior (Caspi et al., 2002; Jaffee et al., in press).
Implications for Research
Not all child victims of maltreatment are abused by their par-
ents, particularly their biological parents (Daly & Wilson, 1988).
However, research suggests that, in many cases, maltreatment
results from a parent losing control in the course of ordinary
discipline (Gershoff, 2002). Our findings suggest that researchers
who study the processes by which such disciplinary episodes
become abusive episodes should look to aspects of the familywide
environment as potential moderators of adults’ behavior. A num-
ber of studies have shown that demographic risk factors (e.g., low
education, single parenthood, neighborhood poverty), familial risk
factors (e.g., marital conflict, parent psychopathology, parents’
negative emotionality), and parenting risk factors (e.g., low paren-
tal involvement) predict maltreatment (Azar, 2002; Belsky, 1993;
Jaffee, in press).
Our findings on corporal punishment, based on our United
Kingdom sample of 5-year-old twins, are likely to generalize to
other Western developed countries because they replicate research
conducted with a U.S. sample of adult twins and their parents, who
provided retrospective reports of the twins’ experiences of child-
hood corporal punishment (Wade & Kendler, 2000). Nevertheless,
the results of this study must be understood within the confines of
First, unidirectional models of parent–child relations, be they
models of parent effects or child effects, cannot demonstrate the
degree to which parents and children mutually influence one
another over time (Caspi & Moffitt, 1995; Patterson, DeBaryshe,
& Ramsey, 1989). In all likelihood, children’s difficult and coer-
cive behavior elicits punitive responses from adults, which, in turn,
exacerbate the child’s misbehavior. Although our data showed that
genetic factors that influenced children’s difficult and coercive
behavior also explained why some children were more likely than
others to be physically disciplined, we were not able to model the
very likely process by which corporal punishment has a reciprocal
influence on children’s antisocial behavior. Longitudinal data in
which punitive discipline and child antisocial behavior are mea-
sured repeatedly over time are needed to model such a pattern of
reciprocal effects. Such fine-grained temporal assessment of child
and parent behavior is a strength of observational studies (e.g.,
Reid et al., 2002), although researchers interested in child effects
on corporal punishment or physical maltreatment per se might be
hindered by the prohibitively low base rates of these practices
during the observation frame. It is also possible that corporal
punishment moderates children’s genetic risk for antisocial behav-
ior, although the existence of such gene–environment interactions
would not preclude the gene–environment correlation demon-
strated in our analyses (Kendler & Eaves, 1986).
Second, although we failed to find that heritable characteristics
of children elicited maltreatment from adults, the possibility re-
mains that nonheritable child characteristics provoke abusive ep-
isodes. Although the child characteristics that have been identified
by researchers as potential elicitors of maltreatment (e.g., disrup-
tive behavior or mental retardation; Kadushin & Martin, 1981) are
moderately to highly heritable, heritability is a population estimate
that does not reflect any given child’s level of genetic risk. For
example, in their study of adoptees, O’Connor et al. (1998) noted
that parents reacted negatively to any child who engaged in coer-
cive and difficult behavior and not only to those who were also at
Third, although environmental factors accounted for most of the
variation in corporal punishment and physical maltreatment, ge-
netic influences on the former, but not the latter, were significantly
different from zero. On the basis of this finding, we concluded that
heritable characteristics of children (specifically, their antisocial
behavior) may provoke corporal punishment but are less likely to
provoke physical maltreatment. We must qualify this conclusion in
two ways. First, the heritability of maltreatment was .07. Power in
our sample to detect a heritability of this magnitude was relatively
poor. Thus, our finding that corporal punishment was genetically
mediated but maltreatment was not may have been due to poor
power to detect very small genetic influences on maltreatment.
However, we note that had our estimate of genetic influences on
maltreatment been as large as the heritability estimate for corporal
punishment (h2? .25), we would have had adequate power to
detect it. Second, we note that because the 95% CIs around the
heritability estimates for corporal punishment and physical mal-
treatment overlapped to some degree (see Table 1), these estimates
were at best marginally different from one another.
Fourth, because our group of maltreated children was small in
number and because physical maltreatment was the only type of
maltreatment that was formally assessed, it was not possible to
compare subtypes of maltreatment or to compare groups according
to severity, chronicity, or precise developmental period of mal-
treatment, although all cases were necessarily confined to the
THE LIMITS OF CHILD EFFECTS
infancy–toddlerhood and preschool years (Barnett, Manly, & Cic-
chetti, 1993). Whether some subtypes of maltreatment are more
heritable than others is an empirical question that we could not
address, although there is little theoretical rationale to suggest
Fifth, although our measures of corporal punishment and mal-
treatment have shown good interreporter agreement in other sam-
ples, in our sample they came from only the mother. Ideally, our
results should be replicated in studies that use independent ratings
of corporal punishment and maltreatment from other sources. On
the one hand, the possibility that mothers concealed instances of
corporal punishment or maltreatment would have exerted a con-
servative effect on our findings by causing us to misclassify
maltreated children as nonmaltreated or children who were phys-
ically disciplined as not physically disciplined (although most
children in the sample were physically disciplined at some point in
their first 5 years). On the other hand, it is possible that mothers’
concealment of maltreatment in cases in which either one or both
twins had been maltreated may have artifactually inflated the
estimate of shared environmental influence on maltreatment (by
making most twins in the sample concordant for maltreatment or
for nonmaltreatment). That said, strenuous efforts were made to
enhance mothers’ comfort in reporting instances of maltreatment,
the maltreatment protocol we used has been validated in other
samples, and the prevalence rates of maltreatment in our sample
match those in other epidemiological samples, including ones in
which a different maltreatment measure was used. Thus, it is
unlikely that mothers concealed maltreatment for a great many
Sixth, further research is necessary to determine whether find-
ings from studies of twins generalize to studies of singletons.
However, the rates of maltreatment and corporal punishment found
in our sample of twins are similar to those found in singleton
samples, and studies of children’s antisocial behavior have shown
that rates do not differ in twin and singleton samples (Simonoff et
al., 1997; van den Oord, Koot, Boomsma, Verhulst, & Orleveke,
1995). In addition, genetically sensitive designs that include non-
twin kinship data may provide more accurate estimates of shared
Children who are physically punished and children who are
maltreated are at risk for antisocial behavior in childhood and
adulthood. Our findings minimize the possibility that children who
are the victims of maltreatment provoke abusive incidents as a
function of their difficult and coercive behavior, particularly when
that behavior is strongly genetically influenced. Difficult children
may provoke corporal punishment, but the factors causing child
abuse are more likely to be found within the family environment or
the adult abuser.
Achenbach, T. M. (1991a). Manual for the Child Behavior Checklist/4–18
and 1991 profile. Burlington, VT: University of Vermont, Department of
Achenbach, T. M. (1991b). Manual for the Teacher’s Report Form and
1991 profile. Burlington, VT: University of Vermont, Department of
Achenbach, T. M., McConaughy, S. H., & Howell, C. T. (1987). Child/
adolescent behavioral and emotional problems: Implications of cross-
informant correlations for situational specificity. Psychological Bulletin,
American Psychiatric Association. (1994). Diagnostic and statistical man-
ual of mental disorders (4th ed.). Washington, DC: Author.
Anderson, K. E., Lytton, H., & Romney, D. M. (1986). Mothers’ interac-
tions with normal and conduct-disordered boys: Who affects whom?
Developmental Psychology, 22, 604–609.
Arseneault, L., Moffitt, T. E., Caspi, A., Taylor, A., Rijsdijk, F. V., Jaffee,
S. R., et al. (2003). Strong genetic effects on cross-situational antisocial
behavior among 5-year-old children according to mothers, teachers,
examiner–observers, and twins’ self-reports. Journal of Child Psychol-
ogy and Psychiatry and Allied Disciplines, 44, 832–848.
Azar, S. T. (2002). Parenting and child maltreatment. In M. H. Bornstein
(Ed.), Handbook of parenting: Vol. 4. Social conditions and applied
parenting (pp. 361–388). Mahwah, NJ: Erlbaum.
Barnett, D., Manly, J. T., & Cicchetti, D. (1993). Defining child maltreat-
ment: The interface between policy and research. In D. Cicchetti & S. L.
Toth (Eds.), Child abuse, child development, and social policy (pp.
7–73). Norwood, NJ: Ablex.
Bell, R. Q., & Chapman, M. (1986). Child effects in studies using exper-
imental or brief longitudinal approaches to socialization. Developmental
Psychology, 22, 595–603.
Belsky, J. (1993). Etiology of child maltreatment: A developmental–
ecological analysis. Psychological Bulletin, 114, 413–434.
Bennett, N., Jarvis, L., Rowlands, O., Singleton, N., & Haselden, L. (1996).
Living in Britain: Results from the General Household Survey. London:
Her Majesty’s Stationery Office.
Bland, R., & Orn, H. (1986). Family violence and psychiatric disorder.
Canadian Journal of Psychiatry, 31, 129–137.
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model
fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation
models (pp. 136–162). Newbury Park, CA: Sage.
Caspi, A., McClay, J., Moffitt, T. E., Mill, J., Martin, J., Craig, I. W., et al.
(2002, August 2). Role of genotype in the cycle of violence in maltreated
children. Science, 297, 851–854.
Caspi, A., & Moffitt, T. E. (1995). The continuity of maladaptive behavior:
From description to understanding in the study of antisocial behavior. In
D. Cicchetti & D. J. Cohen (Eds.), Developmental psychopathology (1st
ed., pp. 472–511). New York: Wiley.
Cicchetti, D., & Manly, J. T. (Eds.). (2001). Operationalizing child mal-
treatment: Developmental processes and outcomes [Special issue]. De-
velopment and Psychopathology 13(4).
Cohen, P., & Brook, J. S. (1995). The reciprocal influence of punishment
and child behavior disorder. In J. McCord (Ed.), Coercion and punish-
ment in long-term perspective (pp. 154–164). Cambridge, England:
Cambridge University Press.
Daly, M., & Wilson, M. (1988, October 28). Evolutionary social psychol-
ogy and family homicide. Science, 242, 519–524.
Deater-Deckard, K., Fulker, D. W., & Plomin, R. (1999). A genetic study
of the family environment in the transition to early adolescence. Journal
of Child Psychology and Psychiatry and Allied Disciplines, 40, 769–
Department of Health. (1989). The Children Act. London: Her Majesty’s
Dodge, K. A., Bates, J. E., & Pettit, G. S. (1990, December 21). Mecha-
nisms in the cycle of violence. Science, 250, 1678–1683.
Dodge, K. A., Pettit, G. S., & Bates, J. E. (1994). Socialization mediators
of the relation between socioeconomic status and child conduct prob-
lems. Child Development, 65, 649–665.
Dodge, K. A., Pettit, G. S., Bates, J. E., & Valente, E. (1995). Social
information-processing patterns partially mediate the effect of early
JAFFEE ET AL.
physical abuse on later conduct problems. Journal of Abnormal Psy-
chology, 104, 632–643.
Egami, Y., Ford, D., Greenfield, S., & Crum, R. (1996). Psychiatric profile
and sociodemographic characteristics of adults who report physically
abusing or neglecting their children. American Journal of Psychiatry,
Gallup, G. H. J., Moore, D. W., & Schussel, R. (1995). Disciplining
children in America: A Gallup poll report. Princeton, NJ: The Gallup
Ge, X., Conger, R. D., Cadoret, R. J., Neiderhiser, J. M., Yates, W.,
Troughton, E., et al. (1996). The developmental interface between nature
and nurture: A mutual influence model of child antisocial behavior and
parent behavior. Developmental Psychology, 32, 574–589.
Gershoff, E. T. (2002). Corporal punishment by parents and associated
child behaviors and experiences: A meta-analytic and theoretical review.
Psychological Bulletin, 128, 539–579.
Harris, J. R. (1998). The nurture assumption. New York: Free Press.
Jaffee, S. R. (in press). Family violence and parent psychopathology:
Implications for children’s socioemotional development and resilience.
In S. Goldstein & R. Brooks (Eds.), Handbook of resilience in children.
New York: Kluwer.
Jaffee, S. R., Caspi, A., Moffitt, T. E., Dodge, K. A., Rutter, M., Taylor, A.,
et al. (in press). Nature ? Nurture: Genetic vulnerabilities interact with
child maltreatment to promote behavior problems. Development and
Jaffee, S. R., Caspi, A., Moffitt, T. E., & Taylor, A. (2004). Physical
maltreatment victim to antisocial child: Evidence of an environmentally
mediated process. Journal of Abnormal Psychology, 113, 44–55.
Kadushin, A., & Martin, J. (1981). Child abuse: An interactional event.
New York: Columbia University Press.
Kandel, D. B., & Wu, P. (1995). Disentangling mother–child effects in the
development of antisocial behavior. In J. McCord (Ed.), Coercion and
punishment in long-term perspectives (pp. 106–123). Cambridge, En-
gland: Cambridge University Press.
Kendler, K. S., & Eaves, L. J. (1986). Models for the joint effect of
genotype and environment on liability to psychiatric illness. American
Journal of Psychiatry, 143, 279–289.
Kenny, D. A. (1979). Correlation and causality. New York: Wiley.
Lansford, J. E., Dodge, K. A., Pettit, G. S., Bates, J. E., Crozier, J., &
Kaplow, J. (2002). Long-term effects of early child physical maltreat-
ment on psychological, behavioral, and academic problems in adoles-
cence: A 12-year prospective study. Archives of Pediatrics and Adoles-
cent Medicine, 156, 824–830.
Loehlin, J. C. (1996). The Cholesky approach: A cautionary note. Behavior
Genetics, 26, 65–69.
MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power
analysis and determination of sample size for covariance structure mod-
eling. Psychological Methods, 1, 130–149.
Maccoby, E. E., & Martin, J. A. (1983). Socialization in the context of the
family: Parent–child interaction. In E. M. Hetherington (Ed.), Handbook
of child psychology: Socialization, personality, and social development
(pp. 1–102). New York: Wiley.
Maynard, R. A. (1997). Kids having kids: Economic costs and social
consequences of teen pregnancy. Washington, DC: Urban Institute
Moffitt, T. E., & E-Risk Study Team. (2002). Teen-aged mothers in
contemporary Britain. Journal of Child Psychology and Psychiatry and
Allied Disciplines, 43, 727–742.
Muthe ´n, L. K., & Muthe ´n, B. O. (1998). Mplus: The comprehensive
modeling program for applied researchers [Computer software]. Los
Neale, M. C., Boker, S. M., Xie, G., & Maes, H. H. (2002). Mx: Statistical
modeling (6th ed.). Richmond, VA: Virginia Commonwealth University,
Department of Psychiatry.
Neale, M. C., & Cardon, L. R. (1992). Methodology for genetic studies of
twins and families. Dordrecht, the Netherlands: Kluwer.
Neale, M. C., Eaves, L. J., & Kendler, K. S. (1994). The power of the
classical twin study to resolve variation in threshold traits. Behavior
Genetics, 24, 239–258.
Neiderhiser, J. M., Reiss, D., Hetherington, E. M., & Plomin, R. (1999).
Relationships between parenting and adolescent adjustment over time:
Genetic and environmental contributions. Developmental Psychology,
Nobes, G., Smith, M., Upton, P., & Heverin, A. (1999). Physical punish-
ment by mothers and fathers in British homes. Journal of Interpersonal
Violence, 14, 887–902.
O’Connor, T. G. (2002). Annotation. The “effects” of parenting reconsid-
ered: Findings, challenges, and applications. Journal of Child Psychol-
ogy and Psychiatry and Allied Disciplines, 43, 555–572.
O’Connor, T. G., Deater-Deckard, K., Fulker, D., Rutter, M., & Plomin, R.
(1998). Genotype–environment correlations in late childhood and early
adolescence: Antisocial behavioral problems and coercive parenting.
Developmental Psychology, 34, 970–981.
Patterson, G. R., DeBaryshe, B. D., & Ramsey, E. (1989). A developmental
perspective on antisocial behavior. American Psychologist, 44, 329–
Patterson, G. R., Reid, J. B., & Dishion, T. J. (1992). A social learning
approach. IV. Antisocial boys. Eugene, OR: Castalia.
Pike, A., McGuire, S., Reiss, D., Hetherington, E. M., & Plomin, R. (1996).
Family environment and adolescent depressive symptoms and antisocial
behavior: A multivariate genetic analysis. Developmental Psychology,
Plomin, R., & Bergeman, C. S. (1991). The nature of nurture: Genetic
influence on “environmental” measures. Behavioral and Brain Sciences,
Plomin, R., DeFries, J. C., McClearn, G. E., & McGuffin, P. (2001).
Behavioral genetics (4th ed.). New York: Worth.
Plomin, R., Reiss, D., Hetherington, E. M., & Howe, G. (1994). Nature and
nurture: Genetic contributions to measures of the family environment.
Developmental Psychology, 30, 32–43.
Price, T. S., Freeman, B., Craig, I., Petrill, S. A., Ebersole, L., & Plomin,
R. (2000). Infant zygosity can be assigned by parental report question-
naire data. Twin Research, 3, 129–133.
Reid, J. B., Patterson, G. R., & Snyder, J. (2002). Antisocial behavior in
children and adolescents: A developmental analysis and model for
intervention. Washington, DC: American Psychological Association.
Reiss, D., Neiderhiser, J. M., Hetherington, E. M., & Plomin, R. (2000).
The relationship code: Deciphering genetic and social influences on
adolescent development. Cambridge, MA: Harvard University Press.
Rowe, D. C. (1994). The limits of family influence. New York: Guilford
Scarr, S. (1992). Developmental theories for the 1990s: Development and
individual differences. Child Development, 63, 1–19.
Scarr, S., & McCartney, K. (1983). How people make their own environ-
ments: A theory of genotype–environment effects. Child Development,
Simonoff, E., Pickles, A., Meyer, J. M., Silberg, J., Maes, H. H. M.,
Loeber, R., et al. (1997). The Virginia Twin Study of Adolescent
Behavioral Development. Archives of General Psychiatry, 54, 801–808.
Straus, M. A. (1994). Beating the devil out of them: Corporal punishment
in American families. New York: Lexington.
Straus, M. A., & Gelles, R. J. (1988). How violent are American families?
Estimates from the National Family Violence Resurvey and other stud-
ies. In G. T. Hotaling & D. Finkelhor (Eds.), Family abuse and its
consequences: New directions in research (pp. 14–36). Thousand Oaks,
Straus, M. A., & Stewart, J. H. (1999). Corporal punishment by American
parents: National data on prevalence, chronicity, severity, and duration
THE LIMITS OF CHILD EFFECTS
in relation to child and family characteristics. Clinical Child and Family
Psychology Review, 2, 55–70.
Trouton, A., Spinath, F. M., & Plomin, R. (2002). Twins Early Develop-
ment Study (TEDS): A multivariate, longitudinal genetic investigation
of language, cognition, and behaviour problems in childhood. Twin
Research, 5, 444–448.
van den Oord, E. J. C. G., Koot, H. M., Boomsma, D. I., Verhulst, F. C.,
& Orleveke, J. F. (1995). A twin–singleton comparison of problem
behaviour in 2–3 year olds. Journal of Child Psychology and Psychiatry
and Allied Disciplines, 36, 449–458.
Wade, T. D., & Kendler, K. S. (2000). The genetic epidemiology of
parental discipline. Psychological Medicine, 30, 1303–1313.
Walsh, C., McMillan, H., & Jamieson, E. (2002). The relationship between
parental psychiatric disorder and child physical and sexual abuse: Find-
ings from the Ontario Health Supplement. Child Abuse and Neglect, 26,
Widom, C. S. (1988). Sampling biases and implications for child abuse
research. American Journal of Orthopsychiatry, 58, 260–270.
Widom, C. S. (1989, April 14). The cycle of violence. Science, 244,
Received May 2, 2003
Revision received May 11, 2004
Accepted June 11, 2004 ?
Call for Nominations
The Publications and Communications (P&C) Board has opened nominations for the editorships
of Clinician’s Research Digest, Emotion, JEP: Learning, Memory, and Cognition, Professional
Psychology: Research and Practice, and Psychology, Public Policy, and Law for the years
2007–2012. Elizabeth M. Altmaier, PhD; Richard J. Davidson, PhD, and Klaus R. Scherer, PhD;
Thomas O. Nelson, PhD; Mary Beth Kenkel, PhD; and Jane Goodman-Delahunty, PhD, respec-
tively, are the incumbent editors.
Candidates should be members of APA and should be available to start receiving manuscripts in
early 2006 to prepare for issues published in 2007. Please note that the P&C Board encourages
participation by members of underrepresented groups in the publication process and would partic-
ularly welcome such nominees. Self-nominations also are encouraged.
Search chairs have been appointed as follows:
Clinician’s Research Digest: William C. Howell, PhD
Emotion: David C. Funder, PhD
JEP: Learning, Memory, and Cognition: Linda P. Spear, PhD, and Peter Ornstein, PhD
Professional Psychology: Susan H. McDaniel, PhD, and J. Gilbert Benedict, PhD
Psychology, Public Policy, and Law: Mark Appelbaum, PhD, and Gary R. VandenBos, PhD
Candidates should be nominated by accessing APA’s EditorQuest site on the Web. Using your
Web browser, go to http://editorquest.apa.org. On the Home menu on the left, find Guests. Next,
click on the link “Submit a Nomination,” enter your nominee’s information, and click “Submit.”
Prepared statements of one page or less in support of a nominee can also be submitted by e-mail
to Karen Sellman, P&C Board Search Liaison, at firstname.lastname@example.org.
The deadline for accepting nominations is December 10, 2004, when reviews will begin.
JAFFEE ET AL.