Content uploaded by Frank Ascione
Author content
All content in this area was uploaded by Frank Ascione on Nov 30, 2016
Content may be subject to copyright.
10.1177/0886260505283341Journal of Interpersonal ViolenceDadds et al. / Childhood Cruelty to Animals
Associations Among
Cruelty to Animals,
Family Conflict, and
Psychopathic Traits in Childhood
Mark R. Dadds
University of New South Wales
Clare Whiting
Griffith University
David J. Hawes
University of New South Wales
Previous research has produced mixed findings on the role of child and family
factors in the genesis of childhood cruelty. The authors examined the relation-
ships of cruelty to animals to a range of child and family factors. First, the
authors test the idea that cruelty is a callous aggression that will be more
strongly associated with psychopathic (callous or unemotional, CU) traits than
general externalizing problems. Second, the authors operationalize family
problems as open conflict rather than parenting problems as used earlier. Re-
sults indicated that for both genders, CU traits were associated strongly with
cruelty. For boys, externalizing problems also added prediction in regression
analyses. Family conflict was not associated with cruelty for either. These
results suggest that cruelty to animals may be an early manifestation of the sub-
group of children developing conduct problems associated with traits of low
empathy and callous disregard rather than the more common pathway of exter-
nalizing problems and parenting problems.
Keywords: child; aggression; cruelty; animals; psychopathy; family; violence
Few single childhood behaviors carry with them the ominous and perni-
cious associations that childhood cruelty to animals (CCA) has acquired
in popular psychology. The relationship of CCA to violent behavior in adult-
411
Journal of Interpersonal Violence
Volume 21 Number 3
March 2006 411-429
© 2006 Sage Publications
10.1177/0886260505283341
http://jiv.sagepub.com
hosted at
http://online.sagepub.com
Authors’ Note: The authors gratefully acknowledge the support of Griffith University, the Na-
tional Health and Medical Research Council of Australia, and several Queensland Independent
schools for this research.
hood has been the focus of a significant body of forensic literature, and recent
editions of the Diagnostic and Statistical Manual of Mental Disorders
(DSM; American Psychiatric Association, 1994, 2000) have recognized
CCA in the criteria for conduct disorder. However, despite much interest in
CCA as a risk factor for later violence, little research into its early manifes-
tations has been undertaken, and little is known about the relationship of
CCA to the various antisocial pathways articulated in recent developmental
literature.
A wealth of research within forensic settings has demonstrated dispropor-
tionately high rates of CCA in the histories of violent adult offenders (see
Dadds, Turner, & McAloon, 2002, for a review). On the positive side, CCA is
typically measured in forensic research via intensive qualitative interviews,
and interpersonal outcomes are usually at the severe end, such as murder,
interpersonal violence, and sexual offending. Here the wealth of evidence
suggests that CCA is an early marker of a propensity for violence (see
Arluke, Levin, Luke, & Ascione, 1999; Dadds et al., 2002; Felthous &
Kellert, 1987; C. Miller, 2001). On the negative side, most of these studies
are cross-sectional or retrospective, and longitudinal studies that show CCA
is a unique antecedent to violence have not been conducted.
In contrast, the scant research into CCA in the mental health sciences typi-
cally relies on a one-item measure of CCA from a behavior checklist, or
as a one-off symptom derived from a diagnostic interview, and examines it
against criteria of diagnostic outcomes. For example, Loeber, Keenan,
Lahey, Green, and Thomas (1993) found that a variety of early symptoms
discriminated severity and chronicity of conduct problems. CCA discrimi-
nated between milder and more severe forms of conduct problems but was
not unique in this regard. As exemplified by Loeber et al. (1993), clinical
research has been careful to evaluate CCA against the backdrop of a range of
behavior problems that the person may exhibit. This approach has not sup-
ported the idea that CCA has unique predictive power. On the negative side,
the measures used clinically may not operationalize CCA according to its
most pernicious features (e.g., intentionality, motives, secrecy). Furthermore,
DSM diagnostic criteria may be too blunt a tool to allow the potential power
of CCA to be detected. That is, the forensic literature indicates CCA may be
associated with severe acts of predatory violence; DSM diagnoses such as
conduct disorder are far broader than this and cover a range of antisocial
behavior of which severe violence is only a small part.
Despite this, there is some clinical research that hints at the importance of
CCA. Frick et al.’s (1993) meta-analysis of 60 studies, albeit using the Child
Behavior Checklist one item, found that cruelty to animals was one of a small
number of symptoms useful in discriminating between children with severe
412 Journal of Interpersonal Violence
conduct problems (destructive subtype) and mild conduct problems (non-
destructive subtype). Luk, Staiger, Wong, and Mathai (1999) used the one
item of the Child Behavior Checklist (Achenbach, 1991) to measure CCA
and found that, in a sample of clinic-referred conduct problem children,
those described as cruel to animals by their parents differed from those not so
described on a number of variables. The CCA children showed a more severe
pattern of conduct problems and had higher self-esteem than did those in the
non-CCA group. They did not differ on impulsivity, internalizing symptoms,
or family problems. Luk et al. (1999) concluded that CCA may be associated
with a psychopathic subtype of conduct disorder marked by more severe
behavior problems but high self-esteem and callous disregard for others.
Notwithstanding these suggestive findings, it is clear that research may
benefit from measures that capture the complex and often secretive nature of
cruelty. Very recently, researchers have built on Ascione, Thompson, and
Black’s (1997) dimensional model of CCA (including frequency, recency,
type of animals, context, regret), and a theoretically driven measure that is
brief and reliable using either child or parent report has recently appeared
(the Cruelty to Animals Inventory; Dadds, Whiting, Bunn, Fraser, &
Charlson, 2004; Guymer, Mellor, Luk, & Pearse, 2001). Dadds et al. (2004)
showed that the Cruelty to Animals Inventory has excellent reliability and
validity, including convergence with independent observations of children’s
cruel versus nurturant behavior when in free play with companion animals.
The development of this measure has potential for more precisely testing the
relationship of CCA to broader aspects of developmental psychopathology,
in particular the idea that there may be different and measurable pathways to
violent behavior. In this study, we conduct the first test of CCA associations
with two putative pathways to violent behavior: the common pathway asso-
ciated with general externalizing behavior problems and coercive parenting
and the trait or psychopathy pathway held to be driven by aspects of the
child’s temperament that are problematic for the development of empathy
and conscience, the putative temperamental aspect of psychopathy.
A common model of chronic antisocial behavior emphasizes early esca-
lating externalizing behavior in the context of disrupted, conflicted family
relationships and poor parenting. In terms of externalizing behavior, there is
no doubt that rates of CCA are higher in clinic-referred externalizing chil-
dren; however, the concordance is only moderate. For example, 28% of con-
duct disordered children showed CCA compared to approximately 3% in a
comparison group (Luk et al., 1999). Similar or higher percentages are found
in adjudicated forensic samples (e.g., Felthous & Kellert, 1987; K. S. Miller
& Knutson, 1997). Clearly, CCA does not characterize the histories of all
people with aggressive, antisocial behavior.
Dadds et al. / Childhood Cruelty to Animals 413
Similarly, evidence for a relationship between CCA and parenting and
family problems is mixed and grows substantially weaker as improved
designs are used. Correlations can be found between self-reported CCA and
a history of severe parental punishment in college students (Flynn, 1999).
Descriptive studies consistently document that domestic violence and physi-
cal abuse are frequent among participants with animal cruelty histories (e.g.,
Gelles & Straus, 1988; Kellert & Felthous, 1985). Severe cruelty to animals
is consistently found in families receiving services for domestic violence
(Boat, 1995; Deviney, Dickert, & Lockwood, 1983). However, K. S. Miller
and Knutson’s (1997) study of a large sample of adjudicated adult offenders
found no relationship between CCA and adverse family experiences,
although their study is unusual in combining both observed and perpetrated
CCA. Similarly, the small number of studies using more experimental
designs and community or child clinical samples have not supported this
relationship. The Luk et al. (1999) study using parent reports failed to find
convincing differences on family functioning between CCA and non-CCA
conduct problem children. Dadds et al. (2004) assessed child behavioral dis-
turbance and parenting style in a sample of 1,333 children aged 3 to 9 years.
The hypothesis that parenting style would be predictive of child cruelty was
not supported, with the relationship between positive and negative parenting
and CCA disappearing once the level of behavioral disturbance in the child
was controlled. Thus, the relationship of parenting and family conflict to
CCA in children remains unclear.
An alternative to the externalizing behavior or poor parenting pathway to
antisocial behavior emphasizes temperamental factors in the child associated
with disruptions to development of empathy and conscience. Children who
possess highly emotional or arousable temperaments are thought to be more
susceptible to the emotions of others and discomfort triggered by transgres-
sions from behavioral standards. Therefore, these children respond to lower
levels of punishment and learn to avoid particular behaviors quickly, includ-
ing negative emotions in others (Dadds & Salmon, 2003; Hare & Schalling,
1978). However, for children possessing a temperament that is less sensi-
tive to others’ emotions and punishment and more prone to impulsive thrill
seeking and anger, discipline-related socialization may be less effective
(Kochanska, 1993), and the child may fail to learn appropriate inhibitory
control and concern for others.
The paired notions of lack of inhibitory control (impulsivity) and lack of
empathic or emotional arousal (callous or unemotional traits, CU) are synon-
ymous with the traditional two-factor conceptualization of psychopathy in
adults (e.g., Hare, Hart, & Harpur, 1991) and more recently in children and
adolescents (Frick & Ellis, 1999). As shown in adult data, high scores on the
414 Journal of Interpersonal Violence
CU factor are associated with more severe and proactive aggression, less
responsiveness to cues of punishment, disrupted processing of emotional
cues, a reward-dominant style, and less anxiety. These factors are thought to
impede the development of conscience (Barry et al., 2000; Blair, Colledge,
Murray, & Mitchell, 2001; Christian, Frick, Hill, Tyler, & Frazer, 1997;
Lynam, 1997; O’Brien & Frick, 1996). Further, there is tentative evidence to
suggest that conduct problems in high CU children are less likely to have
been directly influenced by a family background characterized by poor par-
enting practices and low socioeconomic status (SES; Christian et al., 1997).
In summary, a number of authors have suggested that there may be multi-
ple developmental pathways to violent and antisocial behavior. In this article,
we test the idea that CCA may be a marker of early manifestation of the early
psychopathy pathway described above. We aimed to investigate the relation-
ship between cruelty and CU personality traits. It has been shown that chil-
dren are more likely to develop both conduct problems and cruelty to animals
in abusive or high conflict family environments. However, a divergent etiol-
ogy in the development of conduct problems, whereby some children de-
velop conduct disorder primarily because of high CU traits, regardless of
the family environment, is possible (Christian et al., 1997; Frick, O’Brien,
Wootton, & McBurnett, 1994; Lynam, 1997; Wootton, Frick, Shelton, &
Silverthorn, 1997). This suggests there may be divergent or interacting path-
ways in the development of abusive behavior toward animals whereby chil-
dren are differentially influenced by conflictive or abusive family systems,
on the one hand, and temperamental CU traits on the other. This investigation
aimed to explore CCA by investigating competing correlates of children’s
cruelty to animals: the family environment, children’s externalizing behav-
ior, and CU traits. The externalizing behavior or poor family stress or par-
enting model would predict that high levels of family conflict and the child’s
externalizing behavior would contribute to the child’s cruelty to animals re-
gardless of the child’s temperament. On the contrary, the early psychopathy
model would predict that relationships of externalizing and family or parent-
ing to CCA would reduce to zero once measures of the CU traits in the child
were entered into the equation.
Method
Participants
Participants were 131 children aged 6 to 13 years and their parents, re-
cruited from five independent schools within southeast Queensland, Austra-
lia. The participant pool included 67 males and 64 females (age males M=
Dadds et al. / Childhood Cruelty to Animals 415
10.0, SD = 2.2; age females M= 10.1, SD = 1.1). The five schools were from
several different locations including rural, coastal, and suburban areas (ap-
proximately 33% each). All schools were located in or around Brisbane, a
city with a population of 1.3 million that contains a mix of European, Asian,
and Indigenous cultural backgrounds but is predominantly Caucasian. Of the
child participants in this study, 44.8% had one sibling, and 27.6% had two
siblings (number of siblings M= 1.51). A total of 81.8% of the parents were
married or de facto, and 15.1% were single parents. The average number of
pets owned by the families was 2.02, with 73.6% of families owning four or
fewer pets. Not all families had a pet currently in the home, but all had some
contact with companion animals. Of the fathers, 31.8% had completed ter-
tiary study, 24.5% had completed college or obtained a trade certificate,
15.2% had completed high school, and 26.4% had completed grade 10. For
the mothers, 34.2% had completed tertiary study, 24.2% had attended col-
lege or trade school, 15.2% had finished high school, and 26.1% had com-
pleted grade 10. Therefore, parents in this study were generally educated,
middle class, and Caucasian and had above average SES.
Measures
The test battery comprised an information and consent sheet, the experi-
mental measures, a return envelope, and a raffle ticket to enter a drawing for a
prize. The completed questionnaires were returned in the postage-paid en-
velope, with return rates ranging from 30.5% to 79.2% across schools. The
return rate was not associated with level of CCA, childhood adjustment prob-
lems, or family conflict reported across schools.
The Children and Animals Inventory (CAI; Dadds et al., 2004) includes
parent and child self-report forms based on the Children and Animals As-
sessment Instrument (CAAI; Ascione et al., 1997), a semistructured inter-
view for children. Nine theory-driven dimensions of cruelty are assessed as
follows: severity (based on degree of intentional pain and injury caused to an
animal), frequency (the number of separate acts of cruelty), duration (period
of time during which the cruel acts occurred), recency (the most recent acts),
diversity across and within categories (number of animals abused from dif-
ferent categories and the number of animals harmed from any one category),
sentience (level of concern for the abused animal),1covertness (child’s at-
tempts to conceal the behavior), isolation (whether the cruelty occurred
alone or with other children or adults), and empathy (the degree of the child’s
remorse for the cruel acts). Items are worded to exclude accidental cruelty by
using words such as on purpose or deliberate. Each item offers a negative
response, such as, “I have never hurt an animal,” to allow a total score of 0 for
416 Journal of Interpersonal Violence
children who reported never having displayed intentional cruelty to animals.
In addition to the nine Likert-type items (two items assessed frequency), a
free-response question (item 10) asks the reporter to describe an incident or
pattern of cruelty that can be used qualitatively. Responses to this item are
scored from 0 to 3 according to a specified coding system to obtain a score
for severity. Total possible scores for the CAI range from 0 (no instances of
animal cruelty) to 39 (severe, chronic, and recent cruelty to a range of ani-
mals with the child showing no empathy). It should be noted however that the
distribution of CAI scores is inherently skewed in nonclinic samples, with
the majority of participants scoring 0.
Internal consistency of the CAI is greater than .9 for both child and parent
versions, and test-retest is in the range of r= .75 to .80. Convergence for par-
ent and child reports has been in the range of .3 to .6, with greater agreement
for females and younger children. Finally, the CAI reliably predicts actual
independent observed levels of nurturing versus cruel behavior in boys play-
ing with a companion animal (Dadds et al., 2004).
Children’s externalizing behavior was measured using the Child Behav-
ior Checklist–Revised–Parent Form (Achenbach, 1991). Only the 33-item
externalizing scale was used. Each item in this checklist is composed of a
statement regarding the child and a range of answers of 0 (not true), 1 (some-
what or sometimes true), or 2 (very true). The Child Behavior Checklist–
Revised is widely used and has reliability, validity, and extensive normative
data for both normal and clinical populations (Achenbach, 1991).
The Antisocial Process Screening Device (APSD; Frick & Hare, 2001)
was used to assess for the presence of CU traits. This is a 20-item behavior
rating scale that was completed by the child’s primary caregiver. Each item is
scored either 0 (not al all true), 1 (sometimes true), or 2 (definitely true).
Frick et al. (1994) found that the APSD contained two factors: a 6-item CU
factor and a 10-item impulsivity or conduct problems factor. Recent findings
have cast some doubt on the stability of the impulsivity or conduct problems
when used in community samples (Frick, Bodin, & Barry, 2000). However,
in both clinic and community samples, the CU factor remains separate from
the impulsivity or conduct problems and narcissism factors, is only weakly
associated with conduct problems in general, and represents the unique char-
acteristic of psychopathy. This structure has recently been confirmed in a
large Australian sample (Dadds, Frost, Fraser, & Hawes, 2005). Thus, only
the 6-item CU factor was used for this study to identify children with a CU
personality style. Both internal consistency and stability of this subscale are
adequate in this age group. Alpha reliability coefficient for the CU scale in
this sample was .70.
Dadds et al. / Childhood Cruelty to Animals 417
The conflict subscale of the Family Environment Scale (Moos & Moos,
1981) was used to measure family conflict. This nine-item subscale mea-
sures the amount of openly expressed anger, aggression, and conflict among
family members. Items are answered true or false, and the score is added.
The conflict subscale has strong internal consistency (Cronbach’s α= .75),
2-month test-retest reliability (.85), and 12-month test-retest reliability (.76;
Moos & Moos, 1981).
Results
The distribution of CAI scores on both the parent and child report forms
of the CAI are presented for each gender in Table 1. It can be seen that chil-
dren self-reported higher levels of cruelty than indicated by the reports of
their parents, and as expected, males had higher rates of cruelty than did
females, and the majority of participants scored 0 on both the parent and
child forms of the measure. Both these findings are consistent with our use of
the CAI in larger normative samples (Dadds et al., 2004). For each gender,
two regressions were conducted. In the first, the dependent variable was the
parent reports of cruelty to animals (CAI parent total score), and in the sec-
ond, child report was used. Although a composite of the two scores would be
advantageous in reducing the total number of analyses conducted, we chose
to use them separately. Combining scores assigns the unique variance con-
tained in each measure to error (Keiley, Lofthouse, Bates, Dodge, & Petit,
2003). Unique variance may be especially important for a low-rate secretive
behavior such as cruelty, and previous studies have shown that parent reports
may underestimate secretive cruelty (Dadds et al., 2004). Predictors were
added in the following order: block 1—number of pets and child age; block
2—family variables including father’s education, mother’s education, and
418 Journal of Interpersonal Violence
Table 1
Distribution of Children and Animals Inventory (CAI) Scores
for Parent and Child Reports
Parent Report Child Report
CAI Score Males % Females % Males % Females %
0 70.1 75.0 47.8 62.5
1-10 15.0 9.4 22.5 20.4
11-20 15.0 15.6 19.5 17.3
21-30 — — 9.0 —
31-35 — — 1.5 —
conflict; block 3—externalizing behavior and CU traits; block 4—CU
Traits ×Externalizing Behavior. It is important to note that cruelty occurs at a
low rate in community samples, with only 30.6% of participants scoring
greater than 0 on the CAI (i.e., reporting any cruel behavior). Such a skewed
distribution presents a number of options for analysis, each with advantages
and limitations. Although a linear regression is capable of explaining the full
range of CAI scores observed, such an analysis assumes normality of distri-
bution and may therefore be biased in this application. An alternative would
be to reduce the continuous scores on the CAI to a categorical variable and
conduct a logistic regression. For a logistic regression, participants would be
coded as cruel (scoring 1 or greater) or not cruel (scoring 0), with the analysis
used to determine the predictive effects of the independent variables on chil-
dren falling into either of these categories. However, although this method
would be more statistically sound, it would limit prediction to the same kind
of data produced by single-item measures of cruelty (e.g., the respective item
on the Child Behavior Checklist). It was decided that the inclusion of both
analyses would therefore provide the best representation of the data.
Results from the linear regression in which CAI scores were treated as a
continuous variable are presented in Table 2. For parent report of males on
the CAI, neither block 1 (age or number or pets) nor block 2 (parent educa-
tion, family conflict) resulted in a significant increase in prediction of cru-
elty. The addition of block 3 led to a significant increase in prediction to
25.3% of variance in CAI scores, whereas the interaction in block 4
explained no additional variance. In the final equation, 25.3% of the variance
had therefore been explained, F(8, 58) = 2.45, p< .05. Significant predictors,
once all had been entered, were mother’s education and externalizing behav-
ior, such that boys exhibiting higher levels of externalizing behavior, and
whose mothers had lower levels of education, were more likely to be reported
as cruel by their parents.
For the females, using parent reports, block 1 resulted in no significant
prediction of cruelty. The addition of blocks 2 and 3 however, resulted in sig-
nificant increases in prediction to 18.6% and 38.6% of the variance, respec-
tively. Block 4 failed to add any significant prediction, with a total of 38.9%
of the variance explained in the final equation, F(8, 55) = 4.38, p< .01. Sig-
nificant predictors, once all had been entered, were age, father’s education,
and CU traits, such that girls exhibiting higher levels of CU traits, whose
fathers had lower levels of education, were more likely to be reported as cruel
by their parents.
Using child scores on the CAI for males, the addition of each of the 4
blocks added significantly to the prediction of cruelty. Block 1 explained
16.1% of the variance, followed by an increase to 26.4% with the addition of
Dadds et al. / Childhood Cruelty to Animals 419
block 2 and increases to 37.1% and 47.4% for blocks 3 and 4, respectively. In
the final equation, 47.5% of the total variance for cruelty had therefore been
explained, F(8, 58) = 6.53, p< .01. Significant predictors, once all had been
entered, were mother’s education and the interaction between CU traits and
externalizing behavior, with lower levels of maternal education and an esca-
lating CU-externalizing interaction associated with higher levels of CCA.
This interaction is presented in Figure 1. Female child reports of cruelty were
not predicted by either block 1 or 2; however, block 3 was significant in pre-
dicting an extra 11.2% of variance in cruelty. Block 4 failed to add any signif-
icant prediction. The only significant predictor shown, once all predictors
had been entered, was CU traits, with girls exhibiting higher levels of CU
traits more likely to report cruelty.
Results of the logistic regressions conducted using the same sets of vari-
ables are presented in Table 3. For the males, using parent scores on the CAI,
block 1 resulted in no significant prediction for age or number of pets. The
420 Journal of Interpersonal Violence
Table 2
Linear Regression Analysis Summary for
Prediction of Child Cruelty as Reported by Parents and Children
CAI Child Report CAI Parent Report
R2rβR2rβ
Males
Age .16** .32 .24* .04 –.20 –.18
No. of pets .28 .27* –.02 –.01
Father education .26* .15 — .14 –.08 .04
Mother education –.34 –.33** –.29 –.30*
Family conflict .16 –.04 –.02 –.06
Externalizing behavior (EB) .37* .30 –.06 .25* .34 .37*
Callous or unemotional (CU) traits .12 — –.05 –.12
CU Traits ×EB .47** .41 .54** .25 .16 –.03
Females
Age — –.06 –.15 .04 –.14 –.28*
No. of pets — –.04 –.13 –.14
Father education .06 .08 .14 .18* .19 .27*
Mother education .01 .09 –.15 –.10
Family conflict .20 .09 .18 –.00
EB .17 .17 .04 .38** .36 .17
CU traits .33 .50** .43 .51**
CU Traits ×EB .18 .15 –.19 .38 .33 –.08
Note: CAI is the Children and Animals Inventory. Beta weights shown are those once all vari-
ables have been entered.
*p< .05. **p< .01.
addition of block 2 variables resulted in no improvement in prediction, but
mother’s education level (p< .01) was individually significant in predicting
whether a male child was cruel or not cruel. The addition of block 3 variables
resulted in a significant equation, χ2(3) = 14.51, p< .001, with 89.4% and
60.0% of participants classified correctly into not cruel and cruel groups,
respectively. Individual variables that were significant in predicting cruelty
were age (p< .05), mother’s education (p< .05), externalizing behavior (p<
.01), and CU traits (p< .05). Thus, using parent reports of cruelty, older
males with low-educated mothers, high externalizing behavior, and high CU
traits were more likely to be cruel to animals. The interaction term in block 4
added no improvement in prediction of cruelty. For the females, using parent
reports, both blocks 1 and 2 resulted in no significant prediction of cruelty,
although age (p< .05) and mother’s education (p< .05) were significant pre-
dictors. Block 3 was significant, χ2(2) = 21.16, p< .001, accurately catego-
rizing 89.8% of children as not cruel and 73.3% as cruel. Significant individ-
ual variables were age (p< .01), number of pets (p< .01), and CU traits (p<
.01). Based on parent reports, these results suggest that older females with
low-educated mothers, high numbers of pets, and high CU traits are more
likely to be cruel. No improvement in prediction was seen following the
addition of block 4.
Dadds et al. / Childhood Cruelty to Animals 421
CU Category
HighLow
Mean CAI: Child Report
20
10
0
Externalizing
Low
High
Figure 1
Interaction Between Callous or Unemotional Traits and Externalizing
Scores on the Cruelty to Animals Inventory for Male Self-Report
Using child scores on the CAI for males, neither block 1 nor block 2 pro-
vided significant prediction of children’s cruelty. However, block 3 produced
a significant equation, χ2(2) = 9.97, p< .01, accurately categorizing 65.6% of
males as not cruel and 74.3% as cruel. The only significant predictive vari-
able was CU traits (p< .05), indicating that, using child reports of cruelty,
males with high CU traits were more likely to be cruel than were those males
low on CU traits. Block 4 added no improvement in the prediction of cruelty.
Female child reports of cruelty were not predicted by either block 1 or block
2, but block 3 was significant in predicting cruelty, χ2(2) = 16.48, p< .001,
accurately classifying 90% of children who were not cruel and 56.5% of chil-
dren who were cruel. The only significant variable was CU traits (p< .05),
indicating that CU traits are significant in designating cruel from not cruel
422 Journal of Interpersonal Violence
Table 3
Logistic Regression Analysis Summary for Prediction of Child
Cruelty as Reported by Parents and Children
CAI Child Report CAI Parent Report
%% %%
Correct Correct Correct Correct
Not Wald Not Wald
Cruel Cruel (z)* Cruel Cruel (z)*
Males
Age 40.6 65.7 0.11 100.0 — 4.55*
No. of pets 2.94 2.81
Father education 62.5 51.4 0.07 87.2 30.0 0.82
Mother education 0.81 5.62*
Family conflict 0.13 0.02
Externalizing behavior (EB) 65.6 74.3 2.20 89.4 60.0 8.06*
Callous or unemotional (CU) traits 5.57* 3.79
CU Traits ×EB 68.8 74.3 — 91.5 55.0 —
Females
Age 100.0 — 2.09 95.8 18.8 7.69*
No. of pets 0.36 5.28*
Father education 92.5 16.7 0.25 89.6 12.5 0.06
Mother education 0.66 1.11
Family conflict 0.32 1.01
EB 92.5 54.2 0.20 93.8 68.8 1.78
CU traits 8.85* 8.61*
CU Traits ×EB 90 62.5 — 93.8 68.8 —
Note: CAI is the Children and Animals Inventory. Wald (z) weights shownare those once all vari-
ables have been entered at the last significant block.
*p< .05. **p< .01.
females when using child reports. Finally, the interaction term in block 4
added no improvement in prediction of cruelty.
There is a possibility that the relationships found between CU traits and
cruelty to animals in the above regressions are inflated because of the CAI
containing items measuring concern and level of empathyfor mistreated ani-
mals. To check that CU traits were indeed associated with cruel behavior and
not just empathic feelings, CAI scores for both parents and children were
recalculated with the sentience and empathy items omitted, and the regres-
sions were run again. No differences in results were found; in fact, there was
a tendency for the associations between CU and cruelty to strengthen.
Discussion
This study examined CCA as a correlate of alternative developmental
pathways to antisocial behavior. The first of these describes antisocial behav-
ior resulting from a combination of externalizing problems in the child and
abusive or high conflict family environments. The second, early psycho-
pathy pathway, implicates CU personality traits as a key etiological factor in
the development of antisocial behavior. The results of the current study
clearly show that cruelty to animals is a consistent correlate of the latter
model, in which the temperamental characteristics of a child are associated
with their cruel behavior. Evidence that family problems are associated with
cruelty was not found to be impressive. Generally, these results support a
model of cruelty associated with child characteristics placing the child at risk
for failing to learn adequate empathic, conscience-driven behavior. Consis-
tent with previous attempts to delineate specific pathways by gender, the
associations were different for males and females. General externalizing
problems were associated with cruelty for males only.
Moderate support was found for a relationship between parent education
(examined as an indicator of SES) and cruelty. Low education in parents was
significantly associated with cruelty to animals in both males and females
when the full range of variance was examined in a linear regression. By con-
trast, family conflict did not appear to contribute to the cruelty reported for
males or females. This is surprising given the extensive literature on the rela-
tionship of family dysfunction and violence and cruelty to animals in chil-
dren. However, it is consistent with previous studies specifically designed to
address this relationship (Dadds et al., 2004; Luk et al., 1999; K. S. Miller &
Knutson, 1997).
There may be important reasons for this inconsistency. First, previous
studies finding a positive relationship are limited to clinical samples where
Dadds et al. / Childhood Cruelty to Animals 423
extremes of child disturbance and/or family violence are the rule. In commu-
nity samples such as these, population variance in both the predictor vari-
ables and cruelty levels are greater. If a significant relationship between fam-
ily dysfunction and cruelty exists only at extremes, it may not be picked up in
a community sample. Although their use of a measure of CCA that combined
perpetrated and observed CCA is unusual, it should be noted that K. S. Miller
and Knutson (1997) used an incarcerated violent sample and also failed to
find a relationship. Second, the conflict subscale of the Family Environment
Scale may not measure the critical aspects of family dysfunction that predict
cruelty. Further research may benefit from expanding measurement of fam-
ily variables, especially to include measures that tap more extremes of vio-
lence. However, it should be noted that the negative findings found here for
family conflict support the results of earlier research by Dadds et al. (2004).
This study found that measures of parenting behaviors that have previously
been shown to predict aggressive behavior (monitoring, positive parent-
ing, inconsistent and extreme punishment; Dadds, 1995) also failed to show
unique associations with CCA.
Notwithstanding the methodological issues, it appears that family conflict
(and parenting problems) are associated with general behavior problems but
have no unique relationship to specific acts of cruelty. The research reviewed
earlier showing relationships between family problems and cruelty typically
utilized clinical and forensic samples in which the predominant problems
were antisocial and aggressive behavior. It is possible that the relationship
between family conflict and cruelty is really just a side effect of the (well-
established) relationship of family conflict to general antisocial and aggres-
sive behavior in these samples. Cruelty, although overlapping with antisocial
and aggressive behavior problems, may have distinct developmental path-
ways. This is supported by the significant results obtained for the child vari-
ables as predictors.
Analyses of how cruelty is associated with different patterns of child and
family functioning showed important differences according to the source of
information and the gender of the child. For males, parent reports showed
that cruelty was consistently associated with high externalizing problems
and low parental education. However, for child reports, CU traits featured as
the main correlate, as a main effect according to the logistic regression and in
interaction with externalizing problems for the linear analysis. This is consis-
tent with the early psychopathy model where CU traits and externalizing
behavior are believed to interact and magnify each other in the development
of antisocial behavior (e.g., Frick & Ellis, 1999). These differences between
findings resulting from child and parent reports reinforce the importance of
424 Journal of Interpersonal Violence
considering informants separately. It is likely that the self-reports of children
contain unique information about any secretive activities, which will by defi-
nition be missing from parent reports. The performing of aggressive acts in
a manner that allows for concealment indicates a strong degree of behav-
ioral control, unlike that associated with the impulsive and/or emotional out-
bursts characterizing typical externalizing behavior. Such secretive aggres-
sion toward animals can be seen as consistent with the construct of CU traits
that emphasize a cold, dispassionate style with behavior driven by purpose-
ful sensation seeking.
For females, CU traits were associated with cruelty according to both par-
ent and child reports and were operated as a main effect. That is, in contrast to
males, externalizing problems did not feature in the cruelty. This may have
important implications for developmental models of antisocial behavior in
females. Recent research has shown that antisocial behavior in females may
not typically emerge until the adolescent years (Silverthorn & Frick, 1999).
Its emergence at this time shows a clear relationship to factors such as family
conflict and adversity; however, it is difficult to find relationships of family
environment factors to conduct problems earlier in the developmental trajec-
tory for females. Alternatively, the relationship of cruelty to CU traits may
mark a pattern of problems that are prognostic in females. This is the third
study to report that CU traits are more likely to operate as a main effect in
characterizing the development of problems in young females (see also
Dadds et al., 2005; Frick et al., 2003). That is, they do not operate in interac-
tion with general levels of externalizing problems to predict outcomes early
in female trajectories. Such findings are contrasted with the preliminary evi-
dence for the early psychopathy model in which CU traits are associated with
the growth of antisocial behavior only in boys with concurrent externalizing
behavior (Dadds et al., 2005; Frick et al., 2003). Problems of autonomy and
control that emerge for the first time in females in adolescence may occur
more strongly for high CU females who are less influenced by parental con-
trol and discipline. These may become manifest as both antisocial behavior
and family conflict in the teenage years. Notwithstanding these speculations,
it is clear that CU traits appear to be solid bases for early cruelty in females.
It should be noted that this study relied on self- and parent-report data.
Previous work has shown that the CAI does converge with independent
observations of child cruelty and nurturance; however, an improved method-
ology would incorporate some behavioral validation of the constructs under
scrutiny and the collection of collateral data such as teacher report of child
externalizing behavior and CU traits. In addition, the middle-class nature of
the sample should be noted in terms of generalizability of the results. The
Dadds et al. / Childhood Cruelty to Animals 425
findings would benefit by some replication in a larger and more diverse sam-
ple of children and their parents. Further, our measure of family conflict may
not have adequately measured the construct of dysfunction and violence as
hypothesized, and further research may benefit from improving the specific-
ity and coverage of this measurement. Finally, because of the skewed distri-
bution observed in CAI scores, two alternative forms of regression analyses
were conducted. The decision was made to perform both in an attempt to best
address the theoretical or statistical limitations associated with each respec-
tive method. Although both analyses produced consistent results, little is
known about the relationship between specific levels of cruelty and various
developmental outcomes such as later antisocial behavior. The continuation
of such research using larger samples could therefore allow for the identifi-
cation of cutoff points that are meaningful in relation to such outcomes.
The associations between CU personality traits and cruelty to animals and
other aggressive behaviors warrant further research attention. Psychopathy
is a putative construct consisting of both behavioral and trait factors. The trait
component is difficult to measure objectively. Cruelty to animals may be an
important behavioral manifestation of this trait, thus allowing for more accu-
rate identification of early signs of risk for psychopathy in children. Cruelty
can also be studied more experimentally through laboratory work and inter-
vention studies that focus on reducing it. Future research could expand on the
humane education program run in schools by Ascione and Weber(1996) that
has shown generalized improvements in children’s empathy as a result of
training in animal relationships. The potential of these programs for children
with generalized behavior problems is unknown but warrants attention.
In summary, the current study has shown that cruelty to animals in chil-
dren shows different patterns of association with child adjustment according
to the developmental model of antisocial behavior adopted, the gender of the
child, and the specific informant used to collect the information. Further, the
pattern of these differences is meaningful and theoretically important. That
is, cruelty in boys was associated with an early psychopathy pathway charac-
terized by an interaction of externalizing behavior and CU. The latter traits
featured more prominently in child reports, possibly because of their associ-
ation with the likelihood of secretive, purposeful cruelty. Consistent with
contemporary models of gender differences in pathways to antisocial behav-
ior, we found that externalizing behavior did not feature in females’cruelty
but was associated with CU traits independently from the children’s general
adjustment and family problems. Cruelty to animals may have potential as an
early indicator of trait factors placing children at risk for the development of
ongoing problems.
426 Journal of Interpersonal Violence
Note
1. Ascione, Thompson, and Black (1997) use sentience to refer to the animal’s level of sen-
tience; the Children and Animals Inventoryuses the term to refer to the child’s empathic concern
for the animal.
References
Achenbach, T. M. (1991). The Child Behavior Checklist—1991. Burlington: University of
Vermont.
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disor-
ders (Rev. 3rd ed.). Washington, DC: Author.
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disor-
ders (4th ed.). Washington, DC: Author.
Arluke, A., Levin, J., Luke, C., & Ascione, F. (1999). The relationship of animal abuse to violent
and other forms of antisocial behavior. Journal of Interpersonal Violence,14, 963-975.
Ascione, F. R., Thompson, T. M., & Black, T. (1997). Childhood cruelty to animals: Assessing
cruelty dimensions and motivations. Anthrozoos,10(4), 170-177.
Ascione, F. R., & Weber, C. V. (1996). Children’s attitudes about the humane treatment of animals
and empathy: One-year follow up of a school-based intervention. Anthrozoos,9(4),188-195.
Barry, C. T., Frick, P. J., DeShazo, T. M., McCoy, M. G., Ellis, M., & Loney, B. R. (2000). The
importance of callous/unemotional traits for extending the concept of psychopathy to chil-
dren. Journal of Abnormal Psychology,109, 335-340.
Blair, R. J. R., Colledge, E., Murray, L., & Mitchell, D. G. V. (2001). A selective impairment in
the processing of sad and fearful expressions in children with psychopathic tendencies. Jour-
nal of Abnormal Child Psychology,29, 491-498.
Boat, B. W. (1995). The relationship between violence to children and violence to animals: An
ignored link? Journal of Interpersonal Violence,10, 229-235.
Christian, R. E., Frick, P. J., Hill, N. L., Tyler, L.,& Frazer, D. R. (1997). Psychopathy and con-
duct problems in children: II. Implications for subtyping children with conduct problems.
Journal of the American Academy of Child and Adolescent Psychiatry,36, 233-241.
Dadds, M. R. (1995). Families,children,and the development of dysfunction. Thousand Oaks,
CA: Sage.
Dadds, M. R., Frost, A., Fraser, J., & Hawes, D. (2005). Disentangling conduct problems and
early psychopathic traits in children: A longitudinal community study. Journal of Consulting
and Clinical Psychology,73, 400-410.
Dadds, M. R., & Salmon, K. (2003). Learning and temperament as alternate pathways to punish-
ment insensitivity in antisocial children. Clinical Child and Family Psychology Review,6,
69-86.
Dadds, M. R., Turner, C., & McAloon, J. (2002). Developmental links between cruelty to animals
and human violence. Australian and New Zealand Journal of Criminology,35, 363-382.
Dadds, M. R., Whiting, C., Bunn, P.,Fraser, J., & Charlson, J. (2004). Measurement of cruelty in chil-
dren: The Cruelty to Animals Inventory. Journal of Abnormal Child Psychology,32, 321-334.
Deviney, E., Dickert, J., & Lockwood, R. (1983). The care of pets within child abusing families.
International Journal for the Study of Animal Problems,4, 321-329.
Felthous, A. R., & Kellert, S. R. (1987). Childhood cruelty to animals and later aggression
against people: A review. American Journal of Psychiatry,144, 710-717.
Dadds et al. / Childhood Cruelty to Animals 427
Flynn, C. P. (1999). Exploring the link between corporal punishment and children’s cruelty to
animals. Journal of Marriage and the Family,61, 971-981.
Frick, P. J., Bodin, S. D., & Barry, C. T. (2000). The psychopathy screening device in community
and clinic samples. Psychological Assessment,12, 382-393.
Frick, P. J., Cornell, A. H., Bodin, S. D., Dane, H. E., Barry, C. T., & Loney, B. R. (2003).
Callous-unemotional traits and developmental pathways to severe conduct disorder. Devel-
opmental Psychology,39, 246-260.
Frick, P. J., & Ellis, M. (1999). Callous/unemotional traits and subtypes of conduct disorder.
Clinical Child and Family Psychology Review,2, 149-168.
Frick, P. J., & Hare, R. D. (2001). The Psychopathy Screening Device. Toronto, Canada: Multi-
Health Systems.
Frick, P. J., O’Brien, B. S., Wootton, J. M., & McBurnett, K. (1994). Psychopathy and conduct
problems in children. Journal of Abnormal Psychology,103, 700-707.
Frick, P. J., van Horn, Y., Lahey, B. B., Christ, M. A., Loeber, R., Hart, E. A., et al. (1993).
Oppositional defiant disorder and conduct disorder:A meta-analytic review of factor analy-
ses and cross-validation in a clinic sample. Clinical Psychology Review,13, 319-340.
Gelles, R. J., & Straus, M. A. (1988). Intimate violence. New York: Simon & Schuster.
Guymer, A. C., Mellor, D., Luk, E. S. L., & Pearse, V. (2001). The development of a screening
questionnaire for childhood cruelty to animals. Journal of Child Psychology and Psychiatry,
42, 1057-1063.
Hare, R. D., Hart, S. D., & Harpur, T. J. (1991). Psychopathy and the DSM-IV criteria for antiso-
cial personality disorder. Journal of Abnormal Psychology,100, 391-398.
Hare, R. D., & Schalling, D. (Eds.). (1978). Psychopathic behavior: Approaches to research.
New York: John Wiley.
Keiley, M. K., Lofthouse, N., Bates, J. E., Dodge, K. A., & Petit, G. S. (2003). Differential risks
of covarying and pure components in mother and teacher reports of externalising and inter-
nalising behavior across ages 5 to 14. Journal of Abnormal Child Psychology,31, 267-283.
Kellert, S. R., & Felthous, A. R. (1985). Childhood cruelty to animals among criminals and non-
criminals. Human Relations,38, 1113-1129.
Kochanska, G. (1993). Toward a synthesis of parental socialisation and child temperament in
early development of conscience. Child Development,64, 325-347.
Loeber, R., Keenan, K.,Lahey, B. B., Green, S. M., & Thomas, C. (1993). Evidence for develop-
mentally based diagnoses of oppositional defiant and conduct disorder. Journal of Abnormal
Child Psychology,21, 377-410.
Luk, E. S. L., Staiger, P. K., Wong, L., & Mathai, J. (1999). Children who are cruel to animals: A
revisit. Australian and New Zealand Journal of Psychiatry,33, 29-36.
Lynam, D. R. (1997). Pursuing the psychopath: Capturing the fledgling psychopath in a
nomological net. Journal of Abnormal Psychology,106, 425-438.
Miller, C. (2001). Childhood animal cruelty and interpersonal violence. Clinical Psychology
Review,21, 735-749.
Miller, K. S., & Knutson, J. F. (1997). Reports of severe physical punishment and exposure to
animal cruelty by inmates convicted of felonies and by university students. Child Abuse and
Neglect,21, 59-82.
Moos, R. H., & Moos, B. S. (1981). The Family Environmental Scale manual. Palo Alto, CA:
Consulting Psychologists Press.
O’Brien, B. S., & Frick, P. J. (1996). Reward dominance: Associations with anxiety, conduct prob-
lems, and psychopathy in children. Journal of Abnormal Child Psychology,24, 223-239.
Silverthorn, P., & Frick, P. J. (1999). Developmental pathways to antisocial behaviour: The
delayed-onset pathway in girls. Development and Psychopathology,11, 101-126.
428 Journal of Interpersonal Violence
Wootton, J. M., Frick, P. J., Shelton, K. K., & Silverthorn, P. (1997). Ineffective parenting and
childhood conduct problems: The moderating role of callous/unemotional traits. Journal of
Consulting and Clinical Psychology,65, 301-308.
Mark R. Dadds is professor of psychology at the University of New South Wales, Sydney,
Australia, and senior research fellow of the National Health and Medical Research Council of
Australia. He directs several national intervention programs for children, youth, and their fami-
lies at risk for mental health problems. He has been national president of the Australian Asso-
ciation for Cognitive and Behavioural Therapy, director of research for the Abused Child Trust
of Queensland, and a recipient of several awards including an Early Career Award from the
Division of Scientific Affairs of the Australian Psychological Society, a Violence Prevention
Award from the federal government via the Institute of Criminology, and the APS Ian Matthew
Campbell Award for excellence in clinical psychology. He has authored 4 books and more than
100 articles on child and family psychology.
Clare Whiting completed her bachelor of psychology with honours at the University of New
England in New South Wales. Although always interested in child psychology, she became curi-
ous about the mechanisms surrounding antisocial behavior and callousness. In her honours the-
sis, she explored the relationship between parenting style and children’s tendency to bully or be a
victim of bullying. In 1999, she commenced her doctorate in clinical psychology where she fur-
ther explored her interest in child psychology, antisocial behavior, and family dynamics. After
the completion of her doctorate in 2001, she worked in an adult psychiatric hospital for 3 years
before taking up a position as school psychologist at a girls’ school in Brisbane where she cur-
rently works under her married name of Clare Rosoman.
David J. Hawes holds a bachelor of psychology (honours) from GriffithUniversity anda Ph.D.
in clinical psychology from the University of New South Wales, where he is currently employed
as a research associate. He has published research in early intervention for conduct problems in
children at risk for chronic and severe antisocial behavior and clinical assessment of childhood
psychopathology and parenting practices. His research has been presented at international con-
ferences in the United States, Europe, and Australia.
Dadds et al. / Childhood Cruelty to Animals 429
