Sei sulla pagina 1di 21

Clinical Child and Family Psychology Review, Vol. 3, No.

2, 2000

Children Who Murder: A Review

David M. Shumaker1 and Ronald J. Prinz1,2

Despite considerable research on juvenile homicide, pre-adolescent homicide offenders have


received less attention. This paper reviews the existing literature on preteen murderers in
order to characterize the current state of research knowledge about this population, and
draws on some of the work on adolescent homicide as well. The analysis of this literature
considers historical context, methodological issues, previous attempts to classify youthful
homicide offenders, and predictors of preteen homicidal behavior. While there is a high
degree of heterogeneity within this population, several developmental similarities emerged
across cases that were associated with the perpetration of homicide by preteens. A high
percentage of preteen homicide offenders come from homes characterized by physical abuse,
domestic violence, poor or absent parenting, and overall instability. Gun availability may
have been a facilitating factor. Support for different etiologies of preteen versus adolescent
homicide is weak. Recommendations for future research directions are offered.
KEY WORDS: juvenile murder; youth crime; violence; delinquency.

INTRODUCTION (Howell, Krisberg, & Jones, 1995; Zahn & McCall,


1999). Third, the precursors and processes inherent
Though homicidal youth have received consider- in such extreme misconduct by preteens are impor-
able attention in the media and in the social sciences, tant to understanding developmental trajectories and
children under age 13 who commited homicide are avenues of potential prevention.
understudied. Research on youth homicide focuses This review began with studies published in 1974
more on adolescents than preteens, most likely be- or later because a comprehensive review by Adams
cause of the differential offense rate. In 1996, 934 (1974) adequately covered the earlier work. An ex-
youth between 13 and 16 years old, as compared to ception to the 1974 cutoff was made for the compari-
only 16 preteens, were arrested for homicide and son of case studies in order to boost the database.
non-negligent manslaughter (which are both defined Studies for the current review were identified via
as the willful killing of one human being by another) searches of the psychological, social work, and crimi-
(FBI, 1997). nal justice databases, as well as the Social Science
Despite the low base rate, preteen homicide of- Citation Index. Studies were included if it was clear
fenders should be studied for several reasons. First, that there were at least some preteen homicide of-
young murderers (i.e., children who commit homicide fenders in the sample. Caveats are noted when studies
or non-negligent manslaughter) challenge longstand- with mixed samples (i.e., teen and pre-teen offenders)
ing conceptions of child development and create a were considered. Media accounts and books in the
serious dilemma for the justice system (Rowley, popular press were excluded. Several studies were
Ewing, & Singer, 1987). Second, juvenile homicide excluded from consideration either because the
rates in the U.S. have doubled since the mid-1980s youth were 13 years old or older, or because the
age range of the sample was ambiguous (e.g., Darby,
1
Allan, Kashani, Hartke, & Reid, 1998; Sendi & Blom-
University of South Carolina, Columbia, South Carolina.
2
To whom correspondence should be addressed at Department
gren, 1975; Toupin & Morrissette, 1990). However,
of Psychology, University of South Carolina, Columbia, South data on homicide by adolescents were pertinent to
Carolina 29208. the paper in the following ways: (1) trends for pre-

97
1096-4037/00/0600-0097$18.00/0 2000 Plenum Publishing Corporation
98 Shumaker and Prinz

teens with respect to arrests for homicide, violence decreased steadily for youth 1316 years old and
in general, and property offenses were placed in con- remained low during the same period for youth 12
text by considering the corresponding trends for ado- years old and younger (see Fig. 1). Unfortunately,
lescents; and, (2) family, environmental, and child UCR data providing a breakdown of homicide arrests
variables profiled in preteen case studies were simi- for 12 and under versus 1316 was not available prior
larly profiled in adolescent case studies, again to pro- to 1994. However, to place these trends in context,
vide a developmental context. it is still helpful to examine violent offenses and prop-
The review examines classification schemes and erty offenses for the two age groups over a 10-year
typologies of youthful homicide, predictors of homi- span. In the 13- to 16-year-old group, arrests in-
cidal behavior in children, and how childhood behav- creased both for offenses against person (which sub-
ioral characteristics of adult murderers, particularly sumes homicide) and for offenses against property
adult serial-killers, might bear on the study of pre- from 1987 to 1997 by 109% and 21%, respectively,
teen homicide. while for the 12-and-under group, arrests for offenses
against person increased by 134% and offenses
against property increased 10% (see Figs. 2 and 3).
BACKGROUND The population data suggest that although preteens
showed increasing rates of arrest for violent offenses
Recent Trends in Youth Homicide from 1987 to 1997, their arrest rates for homicide
offenses remained low. It is not clear however
Though considerable data are available on re- whether preteen involvement in homicide offenses
cent trends in juvenile homicide (Howell et al., 1995), remained constant, or alternatively preteens were un-
fewer statistics on preteen homicide are reported. likely to be charged with homicide despite increased
The National Center for Health Statistics (NCHS) involvement in violence. It was also apparent that
mortality files and two FBI resources, the Uniform crime trends for the 12-and-under group essentially
Crime Reports (UCR) (Federal Bureau of Investiga- paralleled that of the 13- to 16-year-old group.
tion, 1997) and an annual Supplemental Homicide The SHR, which aggregates the 10- to 17-year-
Report (SHR) (Snyder & Finnegan, 1998), provide old age range, nonetheless offers some indication of
summary statistics on youth homicide. NCHS statis- gender trends for reported homicide arrests among
tics, limited to victim characteristics, are seldom ref- youth. For the period of 1991 to 1996 based on 16,175
erenced in psychological studies of young perpetra- reported juvenile arrests for homicide, males used
tors (Riedel, 1999). The more commonly used UCR guns more frequently than females (81% versus 40%),
and SHR statistics appear to have moderate reliabil- males involved one or more accomplices more often
ity and validity, but do not allow for precise estimates than females (54% versus 43%), males targeted vic-
of homicide by preteens (Flewelling & Williams, tims their age or older more often than females (98%
1999; Heide, 1993; Reidel, 1999; Zahn & McCall, versus 80%), and females more likely than males tar-
1999). In addition, caution is recommended when geted a family member (31% versus 5%) and less
interpreting data based on arrests instead of convic- likely targeted a stranger (16% versus 33%).
tions because differences between the two categories
might be due in part to factors such as ethnicity, age,
and SES. Historical Context
Throughout the 1980s and 1990s, less than 100
preteens per year were arrested for murder and non- Research during the 1940s through 1960s on
negligent manslaughter in the United States, account- youthful homicide offenders suffered from limita-
ing for less than 0.1% of all arrests for murder. De- tions due to small sample sizes, broad age ranges
spite fluctuations, no time trend has emerged. During of perpetrators, and a tendency to focus on one
the years 1991 to 1996, preteen males were approxi- particular type of homicide (e.g., parricide cases
mately nine times more likely to be arrested for mur- only) (Adams, 1974). For example, in the best
der or nonnegligent manslaughter than preteen fe- known of the early research efforts, Bender and
males, while approximately 60% of all preteen arrests Curran (1940) focused on the case histories of an
for homicide involved African American youth 8-year-old female and a nine-year-old male who
(Snyder & Finnegan, 1998). had each killed another child, nine highly aggressive
For 1994 to 1997, arrests for homicide offenses children (seven of whom were preteens), and four
Children Who Murder 99

Fig. 1. Number of arrests for homicide offenses by youth 1316 years of age and 12 years of age and under during the period 1994 to
1997 in the United States.

adolescents charged with homicide. Bender and evaluation did not address potential preteenteen
Curran (1940) concluded, based on their limited differences.
sample, that preteens lacked a real understanding Research during the 1960s on homicidal youth,
of the immutability of death in their victims, while again based solely on case studies, nonetheless ex-
adolescents grasped the finality of the act but gave panded the range of etiologies being considered. In-
a superficial impression of not being emotionally vestigators attributed youth homicidal behavior to
affected. In subsequent work with a larger sample, maladaptive familial patterns (Easson & Steinhilber,
Bender (1959) proposed several new hypotheses to 1961; Hellsten & Katila, 1965), to organically based
the field. Bender (1959) analyzed a sample of 21 psychological disorder (Woods, 1961), or to a severe
preteens and 12 adolescents, all of whom committed lack of impulse control (Smith, 1965). A common
murder prior to the age of 16, and found that: (1) assumption among investigators was that a homicidal
preteens were more likely to drown, or set fire to, child was acting on behalf of an adult who had hostile
their victims, while adolescents were more likely feelings towards the victim (Easson & Stenhilber,
to use guns; (2) half of the entire sample was 1961; Hellsten & Katila, 1965; Sargent, 1962), which
psychiatrically evaluated prior to their homicidal may have been an artifact from focusing predomi-
acts, and several were diagnosed with schizophrenia, nantly on parricidal cases.
epilepsy, chronic brain syndromes without epilepsy, In the 1970s, several new hypotheses about
or depression; and, (3) other significant factors youth homicide emerged regarding self-destructive
included compulsive fire-setting, retardation, unfa- impulses (Greenberg & Blank, 1970; Malmquist,
vorable home environments, and personal life expe- 1971), psychopathic tendencies (Tooley, 1975), lan-
rience with a violent death. Most of Benders (1959) guage and educational deficits (King, 1975), and ma-
100 Shumaker and Prinz

Fig. 2. Number of arrests for offenses against person by youth 1316 years of age and 12 years of age and under during the period 1987
to 1997 in the United States.

ternal dominance (Walshe-Brennan, 1977). Formula- ily on case studies, as noted in a review by Adams
tions began to include the consideration of risk factors. (1974). In the past 20 years, preteen homicide re-
For example, Duncan and Duncan (1971) proposed search has begun to move away from reliance on
seven specific risk factors: (1) the intensity of a childs case-study methodology. Of the 23 reviewed studies
hostile reactions; (2) the degree of control a child has (19751998), only 3 (13%) were case studies. Most
over impulses; (3) a childs ability to formulate alterna- of the studies used either passive-observational de-
tive solutions to difficult life situations; (4) the provoc- signs (8, 35%), one-group nonarchival descriptive
ativeness of the intended victim; (5) the degree of help- analysis (7, 30%), or archival designs (5, 22%).
lessness of the intended victim; (6) the availability of For the studies using passive-observational de-
weapons; and (7) a history of homicidal threats made signs or one-group descriptive analyses, samples
by the would-be perpetrator. All of these factors per- ranged in size from 7 to 72 homicidal youth (mean
tain to the childs cognitive, emotional, and coping N 32.1). However, it was often unclear how many
mechanisms or to the victims characteristics. The con- youth in each sample were below the age of 13. Only
ceptualization by Duncan and Duncan (1971) did not the three case studies (Adam & Livingston, 1993;
share a single predictor in common with that of Bender Paluszny & McNabb, 1975; Tooley, 1975) and one
(1959) who focused exclusively on biological and fa- other study (Petti & Davidman, 1981) consisted ex-
milial predictors of homicide in children. clusively of preteen homicide offenders. The other
19 studies mixed teens and preteens.
Methodological Considerations Some samples included homicidally aggres-
sive children who had not actually committed a ho-
Much of the work on preteen homicide pub- micide (Lewis et al., 1985; Lewis, Sahnok, Gran, &
lished in the 1950s, 1960s, and early 1970s relied heav- Ritvo, 1983; Mezzich, Coffman, & Mezzich, 1991;
Children Who Murder 101

Fig. 3. Number of arrests for offenses against property by youth 1316 years of age and 12 years of age and under during the period 1987
to 1997 in the United States.

Petti & Davidman, 1981; Pfeffer, 1980). For example, simply owing to chance (Myers, Scott, Burgess, &
Lewis et al. (1983) included children whose precipitat- Burges, 1995). Such attacks were subsumed under the
ing offense reportedly ranged from relatively less term homicidally aggressive acts but investigators
serious behaviors (e.g., threatening and assaultive varied somewhat in what other kinds of serious vio-
behaviors) to more dangerous behaviors (e.g., at- lence were considered homicidally aggressive acts.
tempting to set afire or strangle victims). The hetero- In addition, studies of adolescent homicide per-
geneity of these samples is problematic for drawing petrators offer a framework for conceptualizing pre-
reliable inferences. On the other hand, studying chil- teen homicide. Many of the methodological issues
dren who commit attempted murder or other violent confronted by teen homicide researchers are also per-
acts just short of homicide may shed light on key tinent to the preteen work. Nonetheless, caution is
etiological and circumstantial variables that relate required when making comparisons between the pre-
to youth at risk to commit homicide. Charting the teens and either homicidally aggressive children or
progression of aggressive behavior over childhood teen homicide offenders. Certainly, as the age-gap
and adolescence on the part of subsequently- or offense-characteristics widens between the groups,
homicidal adolescents has proven to be informative many more differentiating variables emerge that re-
about pre-homicide adjustment (King, 1975; Myers, duce the power of such a comparison. To complicate
1994). matters, age is not always synonymous with develop-
The consideration of homicidally aggressive acts mental level because some teen murderers exhibit
is further complicated by differing thresholds in de- arrested development, and some preteens reflect ex-
fining these. For example, a homicidal act is typically posure to events and experiences well beyond their
defined as an attack on an individual that was not fatal years.
102 Shumaker and Prinz

CLASSIFICATION SYSTEMS is much less common. Such broad psychiatric ap-


AND TYPOLOGIES proaches have been replaced by attempts to classify
this population using standard diagnostic criteria
Researchers studying youth who murder have (Mezzich et al., 1991; Myers & Kemph, 1990). With
constructed various classification systems or typolog- a sample of 14 youth ages 1017 who had committed
ies of homicide offenders. Clearly defined subgroups murder (N 10) or a homicidal act (N 4),
of youth at risk to commit homicide would presum- Myers and Kemph (1990) applied the DSM-III-R
ably aid in studying significant antecedent variables version of the Diagnostic Interview for Children
(Flewelling & Williams, 1999). Generally, ap- and Adolescents as the primary screening tool and
proaches to classifying youthful homicide offenders confirmed conduct disorder (CD) for 86% of the
are based on three frameworks: (1) psychiatric/psy- sample. Though there was no evidence of psychotic
chological constructs, (2) specific characteristics of symptomatology, half of the youth qualified for
the crime, and (3) the victim-offender relationship. diagnoses of anxiety and substance-use-dependence
The utility of each approach is considered. disorders, while a few qualified for oppositional
defiant disorder, attention-deficit-hyperactivity dis-
order (ADHD), major depression, and functional
Classification Based on Psychiatric/ enuresis. No preteen-teen differences were noted.
Psychological Constructs In a study of 135 violent delinquent adolescents
(VDAs) between the ages of 12 to 19 years old who
Stearns (1957) was one of the first to argue for had committed either homicide, rape, arson, robbery,
the presence of a biological or clinical syndrome that armed robbery, aggravated assault, or assault, Mez-
explained the dynamics behind seemingly senseless zich, Coffman, and Mezzich (1991) identified sub-
homicides. Though many of his assertions have been groups based on cluster analysis of DSM-IV disorders
questioned (Easson & Steinhilber, 1961), others have and historical variables. The nine input variables in-
joined Stearns in conceptualizing many youth mur- cluded: (1) Axis I diagnoses, (2) Axis II diagnoses,
ders from a medical-model perspective consistent (3) Axis III physical disorders, (4) Axis IV severity
with a dispositional approach. For example, Miller of psychosocial stressors, (5) Axis V highest level of
and Looney (1974) pointed to two dispositional fac- adaptive functioning during the past year, (6) IQ
tors, dehumanization and episodic dyscontrol, as key level, (7) an index of mental illness and criminality
facets of murder syndromes, and described cold in family, (8) chronicity of delinquent history, and
and non-empathic children in their sample who em- (9) severity and number of offenses. This approach
braced violence and seemed to place little value on yielded four primary clusters of VDAs whom Mez-
human life. Similarly, Tooley (1975) described two zich et al. (1991) labeled as Stable Behavioral Handi-
6-year-old children who had made murderous as- capped Offenders (n 44), Brighter and Reactive
saults on their siblings as cool, canny, and pos- Later Starters (n 27), Early Start Frequent and
sessing a sense of control and self-sufficiency far be- Serious Offenders (n 45), and Physically Ill Of-
yond their chronological age. Zenoff and Zients fenders (n 19). In general, each cluster possessed
(1979) made additional arguments for the presence its own distinct familial, cognitive, criminal, and psy-
of predatory offenders, coining the term nonem- chiatric profile, leading the authors to suggest multi-
pathic to refer to this sub-group. Sorrells (1980) factorial etiologies. Unfortunately, the breakdown of
drew a distinction between nonempathic and two preteens and adolescents was not provided for this
other categories of youth homicide perpetrators that study.
he labeled pre-psychotic and neurotically fear- Myers and Kemph (1990) and Mezzich et al.
ful. From these early works emerged a typology of (1991) were among the first to apply a standardized
offender who lacked empathy, would murder either classification system of youthful homicide perpetra-
in a calculated fashion or a fit of rage, did not suffer tors. It is not clear, however, why a psychiatric nosol-
from obvious psychotic symptomatology, and had a ogy, such as the DSM system, has or may have utility.
biological or syndromal predisposition to exhibit Other researchers (e.g., Malmquist, 1996; Yarvis,
aggressive/violent behavior, all characteristics com- 1991) have conceptualized Axis I and II diagnoses
monly associated with psychopathy. as proximal causal factors involved in the perpetra-
In the past 10 years, conceptualization of pre- tion of this offense, instead of diagnostic categories,
teen and teen homicide as a psychopathic syndrome to inform a typology of youthful homicide. Recent
Children Who Murder 103

formulations by Myers (Myers & Blashfield, 1997; Relying upon data from 72 children between the
Myers, Burgess, & Nelson, 1998; Myers et al., 1995) ages of 12 to 18 years old charged with homicide and
seem to recognize the limitations of such a system a control group of 35 adolescents between the ages
and rely more heavily on behavioral and offense char- of 15 to 18 years old charged with larceny, Cornell
acteristics instead of on diagnoses. The Mezzich et et al. (1987b) argued that there was sufficient evi-
al. (1991) study better integrates psychiatric, physical, dence to support use of their three-group typology of
cognitive, adaptational, familial, and historical vari- juvenile homicide offenders. Overall, very few youths
ables. Preteen homicide classification systems based were classified as psychotic at the time of the offense
on DSM constructs may increase in utility by adding (N 5), while the crime group (N 37) and conflict
other factors (Malmquist,1996), particularly when group (N 30) were more heavily represented. In
contextual issues are taken into account. comparing the crime and conflict groups, it was dis-
covered that members of the crime group were more
likely to murder strangers, have an accomplice, flee
Classification Based on Specific Characteristics of the crime scene, and be intoxicated at the time of
the Offense the offense. Whereas members of the conflict group
were more likely to murder family members, act
Recent attempts to classify youthful homicide alone, use a weapon (usually a gun), and be caught
perpetrators on the basis of specific characteristics of at the crime-scene. Though far more speculative, the
the offense are found in the typologies developed by authors found preliminary evidence to suggest that,
Cornell (Cornell, 1989; Cornell & Benedek, 1987b) as opposed to the conflict group, members of the
and Myers (Myers et al., 1995). Specifically, Cornell crime group were more likely to have had a history of
(Cornell, 1989; Cornell & Benedek, 1987b) proposed delinquency and prior criminal activity, poor school
that the majority of efforts in this area have not con- adjustment, and substance abuse, but a lower fre-
sidered the reliability of proposed typologies, and quency of stressful life events prior to the offense.
Myers et al. (1994) maintained that no adequate clas- No distinctions between preteen and adolescent
sification systems have yet emerged. Both research youth were noted.
groups developed typologies that shift the focus from Myers et al. (1995) examined a classification sys-
personal factors about the perpetrator to contextual tem based on the FBI Crime Classification Manual
factors associated with the crime (e.g., criminal enter- (CCM) (Douglas et al., 1992) in an attempt to im-
prise versus domestic situation, type of weapon, prov- prove upon the Cornell (Cornell, 1989; Cornell &
ocational circumstances). Benedek, 1987b) typology. The CCM provided a
Prior to introducing their classification system, standard terminology for the classification of murder
Cornell (Cornell, 1989; Cornell & Benedek, 1987b) and included several categories of offense-types to
highlighted two important advantages inherent to a allow for a more precise classification of a particular
typology based on circumstances surrounding the of- offender. In terms of specific structure, the CCM is
fense. To begin, he argued that such a classification divided into four major categories (24 subcategories)
system would have greater relevance to legal deci- according to motive: (1) criminal enterprise (eight
sion-making by virtue of its direct link to an illegal subcategories); (2) personal cause (nine subcategor-
act. In addition, he maintained that it would be easier ies); (3) sexual homicide (four subcategories); and
to validate the typology using factors that had tradi- (4) group cause (three subcategories). In the study,
tionally served as the diagnostic criteria for psychiat- the authors interviewed a sample of 25 youth between
ric classification systems (e.g., developmental and the ages of 7 to 17 years old who had committed a
prior adjustment factors). In terms of the actual clas- homicide (N 21) or a homicidal act (N 4) in
sification system studied, he proposed three groups of an attempt to assess the utility of the CCM as a
offenders: (1) psychoticencompassing individuals classification system for juvenile murders.
who presented clear psychotic symptoms at the All of the youth were classifiable according to
time of the offense; (2) conflictencompassing non- CCM criteria and fell into the criminal enterprise
psychotic individuals who were engaged in an inter- (36%) or personal cause (64%) categories (Myers et
personal conflict with the victim; and (3) crime al., 1995). Closer analysis of children in the criminal
encompassing non-psychotic individuals who enterprise category revealed that the majority of ho-
committed the offense in the course of committing micidal acts took place during commission of another
another crime. felony. Children in the personal cause category com-
104 Shumaker and Prinz

mitted a greater range of crimes with the following teen homicide. The researchers compared the per-
sub-categories represented: spontaneous domestic sonality, familial, and environmental characteristics
homicide (8% of entire sample), staged domestic ho- of 10 youth charged with parricide to 10 youth
micide (8%), argument murder (12%), conflict mur- charged with murdering a relative (other than their
der (12%), revenge murder (4%), and mixed sexual mother or father) or a close acquaintance and 10
homicide (20%). Myers et al. (1995) concluded that youth charged with murdering a stranger. Common
the CCM was a promising classification method, but to the three groups was a history of family disorgani-
called for improvements in the instrument that would zation characterized by marital conflict, economic in-
allow for more precise classifications of intrafamilial, security, parental brutality, and a lack of social or
conflict, and psychotic homicides. Unfortunately, the community ties. Children charged with parricide,
authors did not provide information about the type however, were more socially isolated and had fewer
of offenses committed by younger versus older chil- previous instances of aggressive behavior or lapses
dren in the sample. of impulse control than the comparison groups. Cor-
Overall, Cornell (Cornell, 1989; Cornell & mier et al. (1978) argued in favor of not only discrimi-
Benedek, 1987b) and Myers et al. (1995) found that nating intra-familial from other types of juvenile ho-
homicidal youth were readily classifiable within a micide, but also making a further distinction within
framework based on specific characteristic of the of- the sub-group of intra-familial offenders between
fense, with roughly half of their samples committing children who commit matricide (killing of ones
murder in the course of another crime and half during mother) versus patricide (killing of ones father) ver-
interpersonal conflicts. In addition to the advantages sus fratricide (killing of ones brother).
cited by Cornell (Cornell, 1989; Cornell & Benedek, Cornell et al. (1987a) and Rowley, Ewing, and
1987b), Myers et al. (1995) noted that this classifica- Singer (1987) have also argued for classification
tion method uses standardized terminology that may based on the victim-offender relationship. Cornell et
facilitate better communication among researchers, al. (1987a) argued that intrafamilial youthful homi-
mental health professionals, law enforcement offi- cide offenders are more likely to have extended his-
cials, and the courts. By virtue of its extensive focus tories of interpersonal conflict with the victim and
on the characteristics of the crime, however, this ap- are more likely to use a gun in the commission of
proach ignores key developmental and psychological the offense. Rowley et al. (1987) posited significant
differences among youthful homicide offenders. psychological differences between intra- versus extra-
familial juvenile homicide perpetrators. The most
compelling support for this classification method can
Classification Based on the be found in the extensive research conducted by
Victim-Offender Relationship Heide (Heide, 1999a; Heide, 1997; Heide, 1996;
Heide, 1992). Relying on comprehensive clinical in-
Classification based on the victim-offender rela- terviews of 59 homicidal youth between the ages of
tionship is considered separately here because the 12 to 17 years old, Heide (1992) found seven different
distinction between familial and non-familial homi- types of offenders could be clinically identified and
cide may be important (Cormier, Angliker, Gagne, & reliably rated. Of the small group of children in her
Markus, 1978; Cornell, Benedek, & Benedek,1987a; sample who committed parricide (N 7), all but
Ewing, 1997; Heide, 1999). Early work examined chil- one fit the profile of the situationally trapped kid
dren who committed intra-familial homicide sepa- typology of offendercharacterized by a history of
rately from other youthful homicide perpetrators severe abuse, an extreme sense of desperation, a gen-
(Adelson, 1972; Carek & Watson, 1964; Duncan & erally passive approach to life, and a typically good
Duncan, 1971; Sadoff, 1971; Scherl & Mack, 1966; prognosis.
Tucker & Cornwall, 1977). Corder, Ball, Haizlip, Rol- Findings from research on victimoffender rela-
lins, and Beaumont (1976) and Cormier et al. (1978), tionship suggest the utility of the intrafamilial versus
however, conducted the first systematic attempts to extra-familial distinction. The pattern of differences
compare intra-familial versus extra-familial youthful between these sub-groups parallels that observed be-
homicide perpetrators. Although the Corder et al. tween the conflict and crime typologies previously
(1976) sample included only 13- to 18-year-olds, the discussed. The main risk associated with this type of
study merits review because of its influence on the classification system is a tendency to examine samples
conceptualization and categorization of preteen and comprised exclusively of one type of offender (i.e.,
Children Who Murder 105

either intrafamilial or extra-familial perpetrators). ical disturbance has been found (Corder et al., 1976,
This would not allow for cross-typology comparisons Cornell et al., 1987; Walshe-Brennan, 1974). Busch
that are necessary for the field to continue to test the et al. (1990) did find a higher rate of alcohol abuse
validity of classification systems based on the victim- by homicidal youth compared with non-violent of-
offender relationship. fenders (38% versus 24%), and others have similarly
found substance abuse to be common in juveniles
who commit homicide (Cornell et al., 1987; Corder
PUTATIVE PREDICTORS OF HOMICIDE et al., 1976; Myers & Kemph, 1990; Myers et al., 1995).
Furthermore, three studies reported that anywhere
Many variables have been considered as putative from 2550% of homicidal youth were intoxicated at
predictors of homicide by preteens, albeit sometimes the time of the offense (Cornell, 1989; Sorrells, 1977;
post hoc. Putative predictor variables are considered U.S. Department of Justice, 1987), although it is un-
at the individual, familial, environmental/contextual, clear whether preteen homicidal offenders show the
and offense-specific levels. same pattern.
Across several studies (Myers 1994; Myers &
Kemph, 1988; Myers & Kemph, 1990; Myers and
Individual Level Mutch, 1992; Myers et al., 1995), Myers and col-
leagues have found a high incidence of CD (8488%),
Psychological as have others (Cornell et al., 1987; Lewis et al., 1983;
Malmquist, 1996; Petti & Davidman, 1981; Rosner et
In a review of 17 studies examining the psycho- al., 1978; Scott, 1999; Yates, Beutler, & Crago, 1984).
logical characteristics of homicidal youth, Ewing Similarly, adult serial killers often had a history of
(1990) concluded that the majority of offenders were childhood CD (Myers, 1994; Ressler, Burgess, &
not psychotic at the time of the offense, but non- Douglas, 1988). Myers et al. (1995) also found a high
psychotic disorders may or may not have been pres- rate of ADHD in their sample, which is consistent
ent. Findings by Busch, Zagar, Hughes, Arbit, and with other work on early identification of psycho-
Russell (1990) and Zagar, Arbit, Sylvies, Busch, and pathic characteristics in children (Frick, OBrien,
Hughes (1990) were consistent with Ewing (1990). Wootton, & McBurnett, 1994; Lyman, 1996). Lyman
Busch et al. (1990) compared 71 homicidal youth (1996) has argued that the co-occurrence of child-
between 10 and 17 years old with 71 nonviolent delin- hood CD and ADHD is a primary risk factor for
quent youth and found psychotic symptoms (i.e., hal- adult psychopathic tendencies. This finding links well
lucinations) in 3% and 0% of the groups, respectively, conceptually to the lack of empathy often found
and Zagar et al. (1990) found similar results with a among homicidal youth. Beyond CD, Malmquist
separate sample. However, Myers et al. (1995) found (1996) found a high rate of depressive symptomatol-
that although no youth in their sample (N 21) ogy in his sample and suggested that a sense of hope-
met DSM-III-R diagnostic criteria for a psychotic lessness in severely depressed children was a key
disorder, 71% had a history of psychotic symptoms, risk factor for future homicide or suicide. Malmquist
including paranoid ideation (67%), delusional think- (1996) also has argued that an early history of impul-
ing (10%), auditory hallucinations (29%), gustatory sivity coupled with separation anxiety disorder in
hallucinations (5%), and derealization (5%). Taken these children may have foreshadowed the develop-
together with other studies (Lewis et al., 1988; Lewis ment of borderline personality disorder.
et al., 1985; Rosner, Weiderlight, Rosner, & Wiecz- Overall, recent investigations suggest the pres-
orek, 1978), Myers et al. (1995) suggest that homicidal ence of CD and ADHD, two disorders with strong
youth may have psychotic features that do not reach connections to psychopathic behavior, in addition to
a diagnostic threshold of psychosis. poly-substance abuse, depression, separation anxiety
In some samples, homicidal youth have shown disorder, and enuresis.
a high rate of psychopathic characteristics (Green-
berg & Blank, 1970; Hellsten & Katilla, 1965; Sorrells,
1977; Tooley, 1975; Zenoff and Zents, 1979) as well Cognitive
as impulse-control deficiencies (Cornell, 1989; Mc-
Carthy, 1978; Miller & Looney, 1974; Smith, 1965; In a review of studies on homicidal youth (mixed
Woods, 1961), while in other samples little psycholog- ages), Ewing (1990) found conflicting evidence re-
106 Shumaker and Prinz

garding rates of intellectual deficiency, learning disa- educational difficulties and neurological abnormalit-
bilities, and neurological abnormalities. Several stud- ies; however, considerable variability and inconsis-
ies found significant cognitive impairment (Busch et tency with respect to intellectual functioning was re-
al., 1990; Lewis, Lovely, et al., 1988; Lewis et al., 1985; ported.
Zagar et al., 1990), while other studies found low
rates of cognitive problems (Petti & Davidman, 1981;
Russell, 1986; Walshe-Brennan, 1977). Familial Level
In a slight majority of the studies examining in-
telligence, homicidal youth scored below normal on For both intra- and extra-familial murders, the
Full Scale IQ, with considerable variation (Busch et majority of youthful homicide (both preteen and ado-
al., 1990; Lewis, Lovely, et al., 1988; Hays et al., 1978; lescent) perpetrators have adverse family histories
Solway Richardson, Hays, & Elion, 1981; Zagar et (Crespi & Rigazio-DiGilio, 1996; Duncan & Duncan,
al., 1990), and in a sample of homicidal preteens 1971; Ewing, 1990/1999; Heide, 1992; McCarthy,
(N 9), Petti and Davidman (1981) reported Full 1978; Sargent, 1962). In a review of 18 studies since
Scale IQ scores ranging from 73 to 106 with a mean 1974 that examined familial factors, eight adverse
score of 89. Other researchers, however, found little familial factors emerged in the histories of youthful
evidence of below average intelligence (Bender, murderers, including physical abuse, sexual abuse,
1959; Malmquist, 1971; Walsh-Brennan, 1975). instability of caretaker situation/residence, paternal
The majority of studies on learning disabilities absence, parental alcohol/drug abuse, parental psy-
for homicidal youth offered evidence of either spe- chiatric history, parental criminal background, and
cific learning difficulties (Bender, 1959; Busch et al., violence in the home.
1990; Lewis et al., 1988; Myers, 1995) or general Child physical abuse and witnessing violence in
school adjustment problems (Cornell et al., 1987; the home were the most frequently cited familial
Sendi & Blomgren, 1975). Busch et al. (1990) found factors. For example, Busch et al. (1990) found that
that 10- to 17-year-old youth were significantly more 25% of their large sample had experienced physical
likely to have severe educational difficulties (e.g., abuse, which was significantly higher than the 20%
retardation, lower perceptual IQ scores, and lower rate found in the nonviolent offender group. Physical
full IQ scores) than non-violent adolescent offenders. abuse and exposure to violence may have affected
Myers and colleagues (Myers, 1995; Myers & Mutch, the youth in two distinct ways. First, children who live
1992) found high rates of language/learning disorders in homes where abuse and violence are commonplace
and school retention. In the Myers (1995) sample, may fail to experience adequate nurturance and so-
76% of homicidal youth had a learning disability, and cialization from caregivers (Corder et al., 1976). Over
86% had failed at least one school grade. time, these children may seek out sources of immedi-
Although some studies did not find neurological ate gratification, resulting in periodic aggressive out-
abnormalities in this population (e.g., Petti & bursts when thwarted. Second, violence in the home
Davidman, 1981; Walshe-Brennan, 1975), the major- may produce modeling and normalizing effects
ity did, including a high rate of neuropathy, including (Sorrells, 1977). Lewis (1983) expanded this position,
evidence of epilepsy, serious head traumas, abnormal arguing that a combination of modeling and the pres-
EEG findings, and other neurological deficits (Bailey, ence of child neurological vulnerabilities increase the
1996; Busch et al., 1990; Lewis, Pincus et al., 1988; risk of murder in cases where that child becomes
Myers et al., 1995). Lewis et al. (1988) found homi- overwhelmed with the frustration of witnessing re-
cidal juveniles were significantly more likely to have peated violence in the household.
neurological impairments than non-violent juvenile Less is known about the relationship between
offenders. However, neurological impairment did not preteen homicide and other adverse familial variables
significantly differentiate homicidal juvenile offend- besides the ones discussed above. Lewis et al. (1983)
ers from youths charged with other violent offenses. suggested that a parental criminal, psychiatric, and/
Further, Busch et al. (1990) found that homicidal or substance abuse history is indicative of a familial
youth in their sample were significantly more likely predisposition towards maladaptive behavior. Youth
to suffer from epilepsy or CNS conditions than non- who murder are significantly more likely than either
violent offenders (7% versus 1%). Overall, studies non-violent or otherwise violent offenders to have a
investigating the cognitive characteristics of homi- criminally violent family member (Busch et al., 1990).
cidal youth found considerable evidence for learning/ Another argument is that these variables are indica-
Children Who Murder 107

tors of an unstable family environment and poor par- youth feels obligated to use a gun in the commission
enting practices that yield an inadequate socialization of a crime in order to prove their prowess to the
process (Corder et al., 1976). The more extensive adults; (4) youth who are the victim of abusive treat-
study of familial risk factors associated with juvenile ment feel the need to retaliate after arming them-
offending may provide clues regarding the familial selves with a gun; (5) youth carry guns as signs of
dynamics common to preteen homicide (Capaldi & status but in the course of an ordinary disagreement
Patterson, 1996; Farrington, 1995; Hoge, Andrews, & use the weapons to settle disputes; and, (6) in cases
Leschied, 1996; Loeber & Dishion, 1983; Patterson, where a gun is readily available in the home, an
DeBaryshe, & Ramsey, 1989). emotionally disordered child might react to a family
argument in an unthinking and impulsive manner.
The frequency with which each of these typologies
Environmental/Contextual Level occurs in preteen homicide however is unclear.
Gang involvement is also frequently cited as a
Three relevant contextual variables are gun risk factor in youth homicide (Maxson, Woods, &
availability, gang activity, and media portrayal of vio- Klein, 1995; NYGC, 1997). On a national level, the
lence, all of which are frequently referenced in discus- UCR classified 43% of all murders committed by
sions of homicidal youth. To date, however, relatively juveniles in 1994 (N 1,157) as juvenile gang kill-
little systematic research has been conducted in this ings (FBI, 1997). This finding has been confirmed
area. This could be due in part to the difficulties in a series of regional (city-wide) investigations of
inherent in finding a way to measure the impact of gang-related homicides (e.g., Block & Block, 1993;
environmental factors on individual behavior. Rogers, 1993). Bush et al. (1990), for example, classi-
As previously discussed, guns are by far the fied 41% of 71 juvenile murders as having a gang-
weapon of choice in juvenile homicide, with 81% of related motivation.
the males and 41% of the females in the previously Gang involvement can increase youth risk for
discussed SHR sample using this weapon during the homicide in a variety of ways. Ewing (1990) proposed
commission of their murders (Snyder & Finnegan, four categories, including drug-related homicides, ho-
1998). Little systematic research, however, has been micides committed in the course of other crimes (e.g.,
conducted on the prevalence of gun usage in preteen rape and robbery), gang-on-gang member/internal
homicide. When considering research suggesting in- strife homicides, and unintentional deaths of inno-
creased rates of gun ownership with age (Bjerre- cent bystanders. In addition, some authors have
gaard & Lizotte, 1995; Blumstein, 1995; Fox, 1996), focused more specifically on the relation between
it would be inappropriate to assume equivalent rates illegal drug trade and gang-related homicides
of gun usage in preteen and teen homicide. Conceptu- (Blumstein, 1995), while others have argued that ter-
ally, guns are the only variable that has been defined ritorial turf battles between rival gangs account
as both a contextual risk factor and a tool used to for a higher percentage of gang-related murders
carry out the offense. Cook and Moore (1999), among (Block & Block, 1993; Goldstein, 1991). Alterna-
others, maintain that less availability of guns would tively, Bjerregard and Lizotte (1995) have argued
reduce youth homicide. that gang-members are more likely to own a gun and
Sparse research on gun use in preteen homicide be involved in a range of illegal activities, two risk
has led to hypotheses based solely on clinical observa- factors that in combination greatly increase the likeli-
tion or archival data. Cornell (1993), who has con- hood of homicide and other serious crimes.
ducted the most extensive investigation on this topic, The few authors who have carefully considered
argues for at least six circumstances in which access the ages of gang-affiliated homicide offenders found
to handguns places a youth (both preteens and ado- that the vast majority of homicides are committed
lescents) at increased risk for homicidal behavior: (1) by relatively older juveniles. For example, Maxson,
criminally motivated children, emboldened by car- Gordon, and Kleins (1996) recent study of Los
rying a handgun, graduate from lower to higher Angeles gangs found the mean age of homicide sus-
crimes; (2) youth gangs seek superiority over rival pects in 1994 was approximately 20 years old. Spergel
gangs and an arms race eventually escalates into (1983) found that only 2.2% of the gang-related homi-
a shooting war; (3) adult criminals encourage children cides in Chicago between the years 1978 to 1981 were
to carry weapons so that the adults will not face committed by children under 14 years old. Though
criminal charges if confronted by the police, and the this latter study runs a risk of being out of date, as
108 Shumaker and Prinz

recently as 1995, Spergel argued that gang violence ever, suffers from a lack of supporting evidence. For
is perpetrated primarily by youth between 15 and 24 example, Gadow and Sprafkin (1993) conducted a
years of age. While the exact frequency of gang- series of studies on the viewing habits of a sample
related preteen homicide is unknown, these data sug- of children with emotional and behavioral disorders
gest that gangs play a relatively greater role in teen and found that, in comparison to a group of non-
homicide than in preteen homicide. emotionally disturbed children, emotionally dis-
Media violence and youth homicide might be turbed children watched more hours of aggressive
related. Although an exhaustive review of this topic television, were more likely to prefer aggressive char-
is beyond the scope of this paper, some general obser- acters, had more difficulty comprehending the unre-
vations are in order. Undoubtedly, viewing television ality of the television portrayals of violence, and were
violence increases a childrens likelihood of acting more willing to hurt other children following expo-
in an aggressive manner in laboratory situations sure to aggressive content in laboratory situations.
(e.g., Friedrich-Coffer & Huston, 1986; Roberts & Surprisingly, however, the researchers found that,
Macoby, 1985). Some experimental evidence even in real-world settings, both violent and nonviolent
suggests that exposure to television violence in- television were equally likely to induce antisocial be-
creases aggressive behavior in real-world situations havior in the subjects as compared to no television
(e.g., Wood, Wong, & Chachere, 1991). Furthermore, at all, leading the authors to conclude that television
longitudinal studies conducted in real-world settings is, at best, a secondary risk factor associated with
have shown relatively high correlations between interpersonal conflicts in children.
childhood television violence viewing and later ag- Considering these findings, it could be that expo-
gressive and/or criminal behavior (e.g., Huesmann, sure to media violence should be conceptualized as an
Eron, Lefkowitz, & Walder, 1984; Singer, Singer, & environmental risk factor that operates in a manner
Rapacznski, 1984). Thus, a substantial amount of data approaching that of the physical abuse and witnessing
indicates that exposure to media violence can have of violence variables discussed in an earlier section.
a causal effect on aggression in children under specific This argument has been previously put forth with
conditions and may correlate with criminal behavior respect to youth homicide (Hardwick & Rowton-
in adulthood. Lee, 1996; Sorrells, 1977).
What has been more difficult to show, however,
is a causal connection between television violence
and childrens violent criminal behavior. Paik and Offense-Specific Prediction
Comstock (1994) conducted a meta-analysis of 217
studies that investigated the relationship between vi- Because of the relative lack of systematic re-
olent television viewing and a range of antisocial be- search on preteen homicide, the discussion of of-
havior in children. The researchers found small-to- fense-specific prediction will only draw a basic dis-
medium overall effect sizes relating consumption of tinction between intra-familial versus extra-familial
televion violence in simulated aggressive behavior homicide. The research literature does not yet sup-
and minor aggressive behavior (R2 .10), illegal port differential patterns of predictors for preteen
activities (R2 .03), and interpersonal violence versus adolescent homicide with respect to discrimi-
(R2 .01). This finding supports other research which natig intra-familial and extra-familial homicide.
suggests a lack of a substantial relationship between Overall, four factors appear to discriminate intra-
exposure to media violence and real-world, violent familial from extra-familial youth homicide. Specifi-
behavior in children (Freedman, 1984; McGuire, cally, intra-familial murders are more often charac-
1986). Many youth are exposed to large amounts of terized by: (1) access to a gun in the home; (2) exten-
television violence, but only a minuscule proportion sive physical, emotional, and/or sexual abuse and
commit homicide. interpersonal conflict in the household (that is com-
While viewing television violence by itself may paratively more severe than that of extra-familial
not have a substantial impact generally on violent murderers); (3) few alternative modes of reaction to
behavior in children, another possibility is that this stress available to the child; and, (4) a notable in-
factor may have a contributory and interactional ef- crease in adverse conditions in the home immediately
fect on the violent behavior of children already suffer- prior to the murder (Corder et al., 1976; Cornell et
ing from emotional, aggressive, and/or cognitive al., 1987; Heide, 1992; Post, 1982). These factors seem
problems (Lande, 1993). Even this hypothesis, how- to be related more to parricide than to fratricide.
Children Who Murder 109

Given that preteens are physically smaller than teens, and 57% of the adolescent cases (16 out of 28) in this
they might target siblings more readily than parents sample are intra-familial potentially obscures the true
(Ewing, 1990). Much less is known, however, about picture of the modal preteen murderer. Ewing (1990)
the predictors of fratricide in particular. Dynamics estimated that 10 to 20% of all murders by youth
of patricide may be somewhat different than what are intrafamilial.
might be operating in fratricide. For example, escape A second caution concerns the considerable
from the family might be a motive for patricide but variability across case studies regarding the focus and
not fratricide. descriptive detail. Some authors focused more inten-
Youth including preteens who murder acquain- sively on individual factors while others emphasized
tances or strangers appear to have more extensive familial variables. In addition, several of the case
histories of aggressive and violent behavior and per- reports were short in length and provided only mini-
haps are more likely to have hyperactivity and im- mal detail about history of the offense. While these
pulse-control deficits (Corder et al., 1976; Cornell et considerations almost ensure that additional salient
al., 1987). They also are more likely to have engaged factors have been omitted in some of the cases, we
in other criminal activity and to have been removed have attempted to reduce the impact of this issue by
from the home, which would reduce the risk to family reporting the presence of a particular factor when it
members (Corder, 1976). Based upon these factors, is suggested in the text.
extra-familial preteen murderers would be at in- It was desirable to have some common pool of
creased risk for psychopathy in adulthood and contin- variables on which to compare the case studies. In
ued involvement in serious violent crime. the absence of any other obvious strategy, the chosen
approach involved focusing on variables found to be
prevalent in the childhood backgrounds of adults who
COMPARISON OF CASE STUDIES committed the most egregious homicides because
there is concern that youth who kill might have been
Using case reports derived from several pub- heading for an extremely troubled adulthood. The
lished articles, the present section compared the be- specific behavioral, psychiatric, and familial variables
havioral, psychiatric, and familial predispositions of selected for comparison were ones that appeared in
several preteens who either committed homicide at least 50% of the 38 cases of sexual and serial killers
(N 10) or a homicidally aggressive action (N 1) described by Ressler et al. (1988). Four other vari-
with a larger sample of adolescents who also either ables were added: deviant sexual experiences, cruelty
committed homicide (N 22) or a homicidally ag- to animals, truancy, and ruminations about murder
gressive act (N 6). Eight articles contributed cases prior to the event.
to the pool of preteen offenders: Paluszny and Several variables cited by Ressler et al. (1988)
McNabb (1975) (one case), Bender and Curran did not appear at all in any of the case studies and
(1940) (two cases), Bernstein (1979) (one case), so were excluded from further consideration (com-
Sargent (1962) (one case), Greenberg and Blank pulsive masturbation, nightmares, destroying prop-
(1970) (one case), Tooley (1975) (two cases), Adam erty, poor body image, voyeurism, pornography, fe-
and Livingston (1993) (one case), and Easson and tishism, rape fantasies admitted, and consenting sex).
Steinhilber (1961) (two cases). Nine articles contrib- Using the case studies, a series of chi square analyses
uted cases to the pool of adolescent homicide offend- were conducted comparing the performance of pre-
ers: Bender and Curran (1940) (two cases), Myers teens and adolescents across 18 selected variables
(1994) (three cases), Holmes and Holmes (1994) (one and are summarized in Table I. Overall, both age-
case), Woods (1961) (two cases), Smith (1965) (three groups indicated a high presence of adverse family
cases), Hellsten and Katila (1965) (four cases), Rus- and individual variables, with preteens and adoles-
sell (1979) (five cases), Russell (1985) (two cases), cents reporting an average of 7.5 and 6.5 risk factors
and Easson and Steinhilber (1961) (six cases). per case, respectively. In addition, both groups exhib-
There are obvious limitations to case-study ited a wide distribution of risk factors. With the ex-
methodology even though this area of inquiry fre- ception of truancy, unhealthy sexual experiences, and
quently depends on case data as a primary method ruminations about murder, each of the remaining risk
of study. Case studies have over-represented intra- factors occurred at least once in each age group.
familial homicides committed by preteens. Indeed, Of particular note, youth in both groups reported
the fact that 63% of the preteen cases (7 out of 11) especially high levels of physical/emotional abuse
110 Shumaker and Prinz

Table I. Comparisons of Homicidal Preteens and Adolescents


Preteens (N 11) Adolescents (N 28) Chi-square
Variable number (%) indicated number (%) indicated significance
Familial/environmental variables
Negative relationship with male caretaker 10 (91%0 14 (50%) p .05
Family psychiatric problems 6 (55%) 11 (39%)
Child perceives unfair treatment 4 (36%) 9 (32%)
Family alcohol/drug problems 4 (36%) 13 (46%)
Unstable residency/home environment 7 (64%) 20 (71%)
Child physical/emotional abuse 9 (82%) 21 (75%)
Domineering mother 5 (45%) 14 (50%)
Child variables
Cruelty to children 9 (82%) 6 (21%) p .01
Unhealthy sexual experiences 0 (0%) 12 (43%) p .01
Lying 4 (36%) 2 (7%) p .05
Fire-setting 3 (27%) N1 (4%) p .05
Ruminations about murder 0 (0%) 9 (32%) p .05
Truancy 0 (0%) 8 (29%) p .05
Oppositional behavior 8 (73%) 12 (43%)
Fantasy/day dreaming 1 (9%) 8 (29%)
Stealing 2 (18%) 9 (32%)
Cruelty to animals 1 (9%) 2 (7%)
Enuresis 3 (27%) 8 (21%)

and instability in the home living environment. Al- of physical and emotional abuse reported in this age
though youth appear to have family backgrounds sim- group. The majority of the preteens households can
ilar to Resslers adult sample, caution is recom- best be described as unpredictable, non-empathic en-
mended in assuming that the youth were similar to vironments where the child was consistently at risk
the adult murderers in other respects. for witnessing or experiencing violence, usually at
Shifting focus, there were several key differences the hand of their primary male caretaker. Post (1982)
between preteens and adolescents in this sample. Re- and Heide (1992) have argued that the intense psy-
ports about preteens reflected greater incidence of chological pressure with little opportunity for release
lying, cruelty/violence with other children, fire set- common to these environments, results in highly ag-
ting, and a negative relationship with primary male gressive behavior under certain conditions.
caretaker. By contrast, adolescents reflected more While Post (1982) and Heide (1992) have argued
reporting of truancy, deviant sexual experiences (in- that psychological pressure is a factor in intra-familial
cluding sexual abuse), and rumination about murder. murder, this factor may play a role in extra-familial
Developmental factors could be operating here. For murder as well. In the collective sample of case stud-
example, sexual development and associated experi- ies, no significant differences between intra- and ex-
ences accelerate in adolescence, and so with the ex- trafamilial homicidal youth were observed with the
ception of sexual victimization might not be as promi- exception of the lying risk factor. This suggests that,
nent for children. contrary to previous assumptions (Heide, 1992),
One of the most intriguing findings concerned there were more similarities than differences in the
the high rate (82%) of preteens who had engaged in developmental backgrounds of these two groups.
cruel behavior towards other children, compared The other significant differences between pre-
with 21% for adolescents. It is possible that more teens and adolescents can be interpreted in a variety
adolescents had a childhood history of cruelty than of ways. For example, it could be that older children,
reported. Nonetheless, early evidence of cruel and by virtue of their increased level of independence,
violent behavior towards peers seems to engender have a greater opportunity to be truant from school,
risk for future violence. thus explaining their higher rates of this behavior. In
The finding that preteens were significantly more addition, by the time a child reaches adolescence, the
likely to have experienced negative relationships with scope of troublesome behavior might have increased
their male caretakers can best be interpreted when to such an extent that researchers are less likely to
viewed in conjunction with the similarly high levels focus on relatively more minor behavior problems
Children Who Murder 111

such as lying; thus, explaining the lower rate of this In terms of predisposing and predictive factors,
behavior found in the teenage sub-group. preteen homicide offenders were likely to have men-
Overall, the case study analysis revealed that tal health histories characterized by high rates of CD
both children and adolescents who murder reflect and ADHD, disorders which have clear links to adult
similar backgrounds to Resslers adult sample, a find- psychopathy. Though less certain, there is also evi-
ing that has ominous prognostic implications for dence pointing to elevated rates of subtle psychotic
youthful murderers. Either the absolute number or symptomatology, separation anxiety disorder, poly-
a particular combination of these risk factors has substance abuse, enuresis, and depression. In terms
a strong bearing on whether a youth will commit of cognitive functioning, while there are conflicting
additional violent actions in adulthood. Alterna- data regarding the intellectual characteristics of this
tively, there could be a host of unknown differences population, there is substantial evidence indicating
between child and adult murderers that mitigate the an elevated rate of neurological abnormalities and
impact of these adverse factors and make it harder to learning disabilities. Perhaps the clearest finding,
predict future dangerousness. It appears that Hagan however, pertains to family background. A high per-
(1997) has conducted the only systematic longitudinal centage of preteen homicide offenders come from
investigation to date into the recividism rates of youth homes characterized by abuse, domestic violence,
incarcerated for homicide. While Hagan (1997) was poor or absent parenting, and overall instability.
unable to document additional homicides upon re- There is qualified support for gun availability as a
lease, he found a majority of subjects engaged in predictor. It is important to note, however, that many
criminal activity and, furthermore, a sizable portion if not most pre-adolescent youth who are exposed to
of these crimes were committed against persons. Un- family violence, inadequate parenting, and related
til longitudinal studies that follow the progress of adversity do not commit murder. Similarly, the vast
preteen murderers are conducted, the true recidivism majority of children exposed to such prior conditions
rates in this population will continue to be a matter who do commit assaultive and otherwise violent acts
of speculation. However, the present case study com- do not kill their victims. So the aforementioned ad-
parison, in combination with Hagans findings, should verse and predisposing conditions, including gun
alert clinicians, researchers, and policy-makers that availability, may facilitate the development of homi-
there is cause for concern when attempting to predict cidal behavior but may not tell us why one child
future aggression and psychopathic behavior in this commits homicide while many others do not.
population. Overall, while this pool of predictors closely re-
sembles that of adolescent homicide offenders, it
would be incorrect to assume an identical etiology
CONCLUSION between preteen and juvenile homicide. What ap-
pears to distinguish preteens from this older age-
The incidence of homicide by preteens in com- group is not the range of risk factors that can contrib-
parison to homicide by adolescents has remained low ute to their offense but, instead, the sheer number
and does not appear to be on the increase, at least and intensity of risk factors that impact an individual
as characterized by homicide arrest rates. Perhaps prior to the murder. For example, by virtue of their
because of the low base rate, there is relatively little more limited physical and cognitive abilities, it is clear
research devoted to this population. Most studies in that preteens have a higher threshold for committing
this area have been forced to rely on small samples, murder than adolescents. In a select few cases, how-
which has limited findings and in some cases spawned ever, this threshold will be reached when a certain
inaccurate assumptions about this population. amount, combination, or intensity of conditions is
A number of studies attempted to classify youth- present. The case study comparison provided some
ful homicide on the basis of either psychiatric con- preliminary support for this hypothesis in light of
structs, specific characteristics of the offense, or the slightly elevated number of risk factors in preteen
victimoffender relationship. Each of these ap- cases as well as differences in the prevalence of cer-
proaches has some utility. However, classification sys- tain predictors. The challenge for researchers is to
tems based on specific characteristics of the offense begin to tease apart the specific combination(s) of
have the advantages of allowing consideration of a risk factors that will cause a preteen to murder (or
greater range of developmental factors and facilitat- engage in violent or pre-violent behavior that is a
ing interdisciplinary communication. prelude to homicide, such as using a weapon or
112 Shumaker and Prinz

threatening to kill someone) and to develop models Ethnographic and qualitative methodologies are rec-
of offending that coherently explain these dynamics. ommended in conjunction with quantitative ap-
While there are most likely key differences be- proaches so that valid conclusions can be reached.
tween preteen murderers and their peers who are In closing, the most striking impression about
involved in less serious crimes, youth homicide re- the literature on children who murder is the high
searchers perhaps have assumed too many differ- degree of individual, familial, and environmental dif-
ences between these populations. Alternatively, it ficulties and stressors present across cases. Indeed,
might be useful to examine continuities and disconti- the popular notion that extreme conditions will give
nuities of behavior to better understand what leads rise to extreme behavior appears to ring true in the
children to escalate from relatively minor to more phenomenon of preteen homicide. Invariably, envi-
serious offenses, including marked violence. Loebers ronmental conditions, which are potentially mallea-
(and colleagues) work on developmental trajectories ble, set the stage for this most tragic of youth be-
is particularly pertinent for this issue (Loeber, DeLa- haviors.
matre, Keenan, & Zhang, 1998; Loeber, Farrington,
Stouthamer-Loeber, Moffitt, & Caspi, 1998). Eventu-
ally, through a combination of studies that focus ex- REFERENCES
clusively on preteen homicide offenders and those
Adam, B. S., & Livingston, R. (1993). Sororicide in preteen girls:
that look at juvenile offending in general, the field A case report and literature review. Acta Paedopsychiatrica,
should reach a point where comprehensive models 56, 4751.
of preteen homicide offending are proposed akin to Adelson, L. (1972). The battering child. The Journal of the Ameri-
can Medical Association, 222, 159161.
that of Harwick and Rowton-Lees (1996) recent Adams, K. A. (1974). The child who murders: a review of theory
model of juvenile homicide. and research. Criminal Justice and Behavior, 1, 5161.
The review suggested four additional recom- Bailey, S. (1996). Adolescents who murder. Journal of Adoles-
cence, 19, 1939.
mendations for future research: Bender, L. (1959). Children and adolescents who have killed.
1. Though statistics suggest that there are only American Journal of Psychiatry, 116, 510513.
a handful of murders committed by preteen females Bender, L., & Curran, F. J. (1940). Children and adolescents who
kill. Journal of Criminal Psychopathology, 1, 297322.
each year in the United States, there is some evidence Bjerregaard, B., & Lizotte, A. J. (1995). Gun ownership and gang
indicating differences in the types of homicides fe- membership. Journal of Criminal Law and Criminology,
males commit as compared to males. This would sug- 86, 3758.
Bernstein, J. I. (1979). Premeditated murder by an eight-year old
gest etiological differences between male and female boy. International Journal of Offender Therapy and Compara-
preteen homicide offenders that warrant closer exam- tive Criminology, 23, 4756.
ination in future research efforts. Block, R., & Block, C. R. (1993). Street gang crime in Chicago.
Washington, DC: U.S. Department of Justice, National Insti-
2. The review found sufficient evidence to sup- tute of Justice.
port continued investigation into whether (and how) Blumstein, A. (1995). Youth violence, guns, and the illicit-drug
psychological assessment measures can be used to industry. Journal of Criminal Law and Criminology, 86, 1036.
Busch, K. G., Zagar, R., Hughes, T. M., Arbit, J., & Bussell,
predict vulnerability to extreme aggression in pre- R. E. (1990). Adolescents who kill. Journal of Clinical Psy-
teens. Neither of these recommendations, however, chology, 46, 472485.
precludes the study of contributing environmental Capaldi, D. M., & Patterson, G. R. (1996). Can violent offenders
be distinguished from frequent offenders: Prediction from
conditions, including easy access to lethal weapons. childhood to adolescence. Journal of Research on Crime and
3. Given how little we still know about preteen Delinquency, 33, 206231.
homicide, it is important to study preteens who en- Carek, D. J., & Watson, A. S. (1964). Treatment of a family
involved in fratricide. Archives of General Psychiatry, 16,
gage in attempted homicide or similar potentially 533542.
lethal acts that do not result in homicide, to overcome Cook, T. D., & Campbell, D. T. (Eds.). (1979). Quasi-experimenta-
some of the base rate limitations. The distinction tion: Design and analysis issues for field settings. Chicago:
Rand McNally.
between attempted and completed homicide may be Cook, P. J., & Moore, M. H. (1999). Guns, gun control, and homi-
somewhat arbitrary and more a function of circum- cide: A review of research and public policy. In M. D. Smith &
stances than of significant subgroup characteristics. M. A. Zahn (Eds.), Homicide: A sourcebook of social research.
(pp. 277296). London: Sage Publications.
4. Little is known about how ethnic and cultural Corder, B. F., Ball, B. C., Haizlip, T. M., Rollins, R., & Beaumont,
variables interact with homicidal behavior by chil- R. (1976). Adolescent parricide: a comparison with other ado-
dren. The base rate problem obstructs the study of lescent murder. American Journal of Psychiatry, 133, 957961.
Cormier, B. M., Angliker, C. C., Gagne, P. W., & Markus, B.
this issue as well. Including attempted homicide (1978). Adolescents who kill members of the family. In J.
should greatly increase the size of available samples. Eekerlaar and S. Katz (Eds.), Family violence: an international
Children Who Murder 113

and interdisciplinary study (pp. 466478). Toronto: Butter- characteristics of juvenile murderers versus status offenders.
worth. Psychological Reports, 43, 8082.
Cornell, D. G. (1989). Causes of juvenile homicide: A review of Heide, K. M. (1992). Why kids kill parents. Columbus, Ohio: Ohio
the literature. In E. P. Benedek & D. G. Cornell (Eds.), State University Press.
Juvenile homicide. (pp. 336). Washington, DC: American Heide, K. M. (1993). Juvenile involvement in multiple offender
Psychiatric Press. and multiple victim parricides. Journal of Police and Criminal
Cornell, D. G. (1993). Juvenile homicide: a growing national prob- Psychology, 9, 5364.
lem. Behavioral Sciences and the Law, 11, 389396. Heide, K. M. (1996). Why kids keep killing: The correlates, causes,
Cornell, D. G., Benedek, E. P., & Benedek, D. M. (1987). Charac- and challenges of juvenile homicide. Stanford Law and Policy
teristics of adolescents charged with homicide: a review of 72 Review, 7, 4349.
cases. Behavioral Sciences and the Law, 5, 1123. Heide, K. M. (1997). Juvenile homicide in America: How can we
Cornell, D. G., Benedek, E. P., & Benedek, D. M. (1987). Juvenile stop the killing? Behavioral Sciences and the Law, 15, 203220.
homicide: Prior adjustment and a proposed typology. Ameri- Heide, K. M. (1999a). Young killers: The challenge of juvenile
can Journal of Orthopsychiatry, 57, 383393. homicide. Thousand Oaks, CA: Sage Publications.
Crespi, T. D., & Rigazio-DiGilio, S. A. (1996). Adolescent homi- Heide, K. M. (1999b). Youth homicide: An integration of psycho-
cide and family pathology: Implications for research and treat- logical, sociological, and biological approaches. In M. D.
ment with adolescents. Adolescence, 31, 353367. Smith & M. A. Zahn (Eds.), Homicide: A sourcebook of social
Darby, P. J., Allan, W. D., Kashani, J. H., Hartke, K. L., & Reid, research. (pp. 221238). London: Sage Publications.
J. C. (1998). Analysis of 112 juveniles who committed homi- Heide, K. M. (1993). Weapons used by juveniles and adults to kill
cide: Characteristics and a closer look at family abuse. Journal parents. Behavioral Sciences and the Law, 11, 397405.
of Family Violence, 13, 365375. Heide, K. M. (1994). Evidence of child maltreatment among ado-
Douglas, J. E., Burgess, A. W., & Burgess, A. G. (1992). Crime lescent parricide offenders. International Journal of Offender
Classification Manual. New York: Lexington Books of Mac- Therapy and Comparative Criminology, 38, 151162.
millan. Heide, K. M. (1999). Youth homicide: An integration of psycholog-
Duncan, J. W., & Duncan, G. M. (1971). Murder in the family: ical, sociological, and biological approaches. In M. D. Smith &
A study of some homicidal adolescents. American Journal of M. A. Zahn (Eds.), Homicide: A sourcebook of social research.
Psychiatry, 127, 7478. (pp. 221238). London: Sage Publications.
Easson, W. M., & Steinhilber, R. M., (1961). Murderous aggression Hellsten, P., & Katila, O. (1965). Murder and other homicide by
children under 15 in Finland. Psychiatric Quarterly, Suppl.
by children and adolescents. Archives of General Psychiatry,
39, 5474.
4, 19.
Hoge, R. D., Andrews, D. A., Leschied, A. W. (1996). An investiga-
Ewing, C. P. (1990). When children kill: The dynamics of juvenile
tion of risk and protective factors in a sample of youthful
homicide. Lexington, MA: Lexington Books.
offenders. Journal of Child Psychology and Psychiatry, 37,
Ewing, C. P. (1997). Fatal families: The dynamics of intrafamilial
419424.
homicide. Thousand Oaks, CA: Sage Publications. Holmes, R. M., & Holmes, S. T. (1994). Murder in America. Lon-
Farrington, D. P. (1995). The development of offending and antiso- don: Sage Publications.
cial behaviour from childhood: Key findings from the Cam- Howell, J. C., Krisberg, B., & Jones, M. (1995). Trends in juvenile
bridge Study in delinquent development. Journal of Child crime and youth violence. In J. C. Howell, B. Krisberg, J. D.
Psychology and Psychiatry, 360, 929964. Hawkins, & J. J. Wilson (Eds.), Serious, Violent, & Chronic
Federal Bureau of Investigation (1997). Uniform Crime Reports Juvenile Offenders: A Sourcebook (pp. 135). Thousand Oaks,
1996. Washington DC: Author. CA: Sage Publications.
Flewelling, R. L., & Williams, K. R. (1999). Categorizing homi- Huesmann, L. R., Eron, L. D., Lefkowitz, M. M., & Walder,
cides: The use of disaggregrated data in homicide research. L. O. (1984). Stability of aggression over time and generations.
In M. D. Smith & M. A. Zahn (Eds.), Homicide: A sourcebook Developmental Psychology, 20, 11201134.
of social research. (pp. 96106). London: Sage Publications. King, C. H. (1975). The ego and the integration of violence in
Fox, J. A. (1996). Trends in juvenile violence. Washington, D.C.: homicidal youth. American Journal of Orthopsychiatry, 45,
Bureau of Justice Statistics, Department of Justice. 134145.
Freedman, J. L. (1984). Effect of television violence on aggressive- Lande, R. G. (1993). The video violence debate. Hospital and
ness. Psychological Bulletin, 96, 227246. Community Psychiatry, 44, 347351.
Frick, P. J., OBrien, B. S., Wootton, J. M., & McBurnett, K. (1994). Lewis, D. O., Lovely, R., Yeager, C., Ferguson, G., Friedman, M.,
Psychopathy and conduct problems in children. Journal of Sloane, G., Friedman, H., & Pincus, J. H. (1988). Intrinsic
Abnormal Psychology, 103, 700707. and environmental characteristics of juvenile murderers. Jour-
Friedrich-Cofer, L., & Huston, A. C. (1986). Television violence nal of the American Academy of Child and Adolescent Psychi-
and aggression: The debate continues. Psychological Bulletin, atry, 27, 582587.
100, 364371. Lewis, D. O., Moy, E., Jackson, L. D., Aaronson, R., Ritvo, U.,
Godow, K. D., & Sprafkin, J. (1993). Television violence and Settu, S., & Simons, A. (1985). Biopsychosocial characteristics
children with emotional and behavioral disorders. Journal of of children who later murder: a prospective study. American
Emotional and Behavioral Disorders, 1, 5463. Journal of Psychiatry, 142, 11611166.
Goldstein, A. P. (1991). Delinquent gangs: a psychological perspec- Lewis, D. O., Pincus, J. H., Bard, B., Richardson, E., Prichep,
tive. Champaign, Illinois: Research Press. L. S., Feldman, M., & Yeager, C. (1988). Neuropsychiatric,
Greenberg, H. R., & Blank, R. H. (1970). Murder and self-destruc- psychoeducational, and family characteristics of 14 juveniles
tion by a 12-year-old boy. Adolescence, 5, 391396. condemned to death in the United States. American Journal
Hagan, M. P. (1997). An analysis of adolescent perpetrators of of Psychiatry, 145, 584589.
homicide and attempted homicide upon return to the commu- Lewis, D. O., Shanok, S. S., Gran, M., & Ritvo, E. (1983). Homicid-
nity. International Journal of Offender Therapy and Compara- ally aggressive young children: neuropsychiatric and experien-
tive Criminology, 41, 250259. tial correlates. American Journal of Psychiatry, 140, 148153.
Hardwick, P. J., & Rowton-Lee (1996). Adolescent homicide: To- Loeber, R., DeLamatre, M. S., Keenan, K., & Zhang, Q. (1998).
wards assessment of risk. Journal of Adolescence, 19, 263276. A prospective replication of developmental pathways in dis-
Hays, J. R., Solway, K. S., & Schreiner, D. (1978). Intellectual ruptive and delinquent behavior. In R. B. Cairns & L. R.
114 Shumaker and Prinz

Bergman (Eds.), Methods and models for studying the individ- Patterson, G. R., DeBaryshe, B. D., & Ramsey, E. (1989). A
ual (pp. 185218). Thousand Oaks, CA: Sage. developmental perspective on antisocial behavior. American
Loeber, R., & Dishion, T. (1983). Early predictors of male delin- Psychologist, 44, 329335.
quency: A review. Psychological Bulletin, 94, 6899. Petti, T. A., & Davidman, L. (1981). Homicidal school-age chil-
Loeber, R., Farrington, D. P., Stouthamer-Loeber, M., Moffitt, dren: cognitive style and demographic features. Child Psychia-
T. E., & Caspi, A. (1998). The development of male offend- try and Human Development, 12, 8289.
ing: Key findings from the first decade of the Pittsburgh Pfeffer, C. R. (1980). Psychiatric hospital treatment of assaultive
Youth Study. Studies on Crime and Crime Prevention, 7, 141 homicidal children. American Journal of Psychotherapy, 34,
171. 197207.
Lyman, D. R. (1996). Early identification of chronic offenders: Post, S. (1982). Adolescent parricide in abusive families. Child
Who is the fledgling psychopath? Psychological Bulletin, Welfare, 61, 445455.
120, 209234. Reidel, M. (1999). Sources of homicide data: A review and compar-
Malmquist, C. P. (1971). Premonitory signs of homicidal aggression ison. In M. D. Smith & M. A. Zahn (Eds.), Homicide: A
in juveniles. American Journal of Psychiatry, 128, 9397. sourcebook of social research. (pp. 7595). London: Sage Pub-
Malmquist, C. P. (1996). Homicide: A psychiatric perspective. lications.
Washington, DC: American Psychiatric Press. Ressler, R. K., Burgess, A. W., & Douglas, J. E. (1988). Sexual
Maxson, C. L., Woods, K. J., & Klein, M. W. (1995). Street gang homicide: Patterns and motives. New York: The Free Press.
migration in the United States. Los Angeles: University of Rice, M. E., & Harris, G. T. (1995). Violent recidivism: Assessing
Southern California, Center for the Study of Crime and Social predictive validity. Journal of Consulting and Clinical Psychol-
Control, Social Science Research Institute. ogy, 63, 737748.
Maxson, C. L., & Klein, M. W. (1996). Defining gang homicide: Roberts, D. F., & Macoby, N. (1985). Effects of mass communica-
an updated look at member and motive approaches. In C. R. tion. In G. Lindzey & E. Aronson (Eds.), Handbook of social
Huff (Ed.) Gangs in America: 2nd Edition (pp. 320). London: psychology, Vol. 2 (pp. 539598). New York: Random House.
Sage Publications. Rogers, C. (1993). Gang-related homicides in Los Angeles County.
McCarthy, J. B. (1978). Narcissism and the self in homicidal adoles- Journal of Forensic Sciences, 38, 831834.
cents. American Journal of Psychoanalysis, 38, 1929. Rosner, R., Weiderlight, M., Rosner, M. B. H., & Wieczorek,
McGuire, W. J. (1986). The myth of massive media impact: Savag- R. R. (1978). Adolescents accused of murder and manslaugh-
ings and salvagings. In G. Comstock (Ed.), Public communica-
ter: A five year descriptive study. Bulletin of the American
tion and behavior: Vol. 1, (pp.173257). San Diego, CA: Aca-
Academy of Psychiatry and Law, 23, 6572.
demic Press.
Rowley, J. C., Ewing, C. P., & Singer, S. I. (1987). Juvenile homi-
Mezzich, A. C., Coffman, G., & Mezzich, J. E. (1991). A typology
cide: The need for an interdisciplinary approach. Behavioral
of violent delinquent adolescents. Journal of Psychiatry and
Sciences and the Law, 5, 110.
Law, 19, 6378.
Miller, D., & Looney, J. (1974). The prediction of adolescent Russell, D. H. (1979). Ingredients of juvenile murder. International
homicide: Episodic dyscontrol and dehumanisation. American Journal of Offender Therapy and Comparative Criminology,
Journal of Psychoanalysis, 34, 187198. 23, 6572.
Myers, W. C. (1994). Sexual homicide by adolescents. Journal of Russell, D. H. (1986). Girls who kill. International Offender Ther-
the American Academy of Child and Adolescent Psychiatry, apy and Comparative Criminology, 30, 171176.
33, 962969. Sadoff, R. L. (1971). Clinical observations on parricide. Psychiatric
Myers, W. C., & Blashfield, R. (1997). Psychopathology and per- Quarterly, 45, 6569.
sonality in juvenile sexual homicide offenders. Journal of Sargent, D. (1962). Children who killa family conspiracy? Social
the American Academy Psychiatry and the Law, 25, 497 Work, 17, 3542.
508. Scherl, D. J., & Mack, J. E. (1966). A study of adolescent matricide.
Myers, W. C., Burgess, A. W., & Nelson, J. A. (1998). Criminal Journal of the American Academy of Child Psychiatry, 5
and behavioral aspects of juvenile sexual homicide. Journal 569593.
of Forensic Sciences, 43, 340347. Scott, C. L. (1999). Juvenile violence. Forensic Psychiatry, 22,
Myers, W. C., & Kemph, J. P. (1988). Characteristics and treatment 7183.
of four homicidal adolescents. Journal of the American Acad- Sendi, I. B., & Blomgren, P. G. (1975). A comparative study of
emy of Child and Adolescent Psychiatry, 27, 595599. predictive criteria in the predisposition of homicidal adoles-
Myers, W. C., & Kemph, J. P. (1990). DSM-III-R Classification cents. American Journal of Psychiatry, 132, 423428.
of murderous youth: Help or hindrance? Journal of Clinical Singer, J. L., Singer, D. G., & Rapaczynski, W. (1984). Family
Psychiatry, 51, 239242. patterns and television viewing as predictors of childrens
Myers, W. C., & Mutch, P. J. (1992). Language disorders in disrup- beliefs and aggression. Journal of Communication, 34, 7389.
tive behaviour disordered homicidal youth. Journal of Foren- Smith, S. (1965). The adolescent murderer: A psychodynamic in-
sic Sciences, 37, 919922. terpretation. Archives of General Psychiatry, 13, 310319.
Myers, W. C., Scott, K., Burgess, A. W., & Burgess, A. G. (1995). Snyder, H., & Finnegan, T. A. (1998). Easy access to the FBIs
Psychopathology, biopsychosocial factors, crime characteris- Supplementary Homicide Reports, 19801996. [Data presen-
tics, and classification of 25 homicidal youths. Journal of the tation and analysis package]. Pittsburgh, PA: National Center
American Academy of Child and Adolescent Psychiatry, 34, for Juvenile Justice.
14831489. Snyder, H., & Sickmund, M. (1995). National report on juvenile
National Youth Gang Center. (1997). 1995 National youth gang offending and victimization. Washington, DC: Office of Juve-
survey. Washington, DC: Office of Juvenile Justice and Delin- nile Justice and Delinquency Prevention.
quency Prevention. Solway, K. S., Richardson, L., Hays, J. R., & Elion, V. H. (1981).
Paik, H., & Comstock, G. (1994). The effects of television violence Adolescent murderers: Literature review and preliminary re-
on antisocial behavior: A meta-analysis. Communication Re- search findings. In J. R. Hays, T. K. Roberts & K. S. Solway
search, 21, 516546. (Eds.), Violence and the violent individual (pp. 193210). New
Paluszny, M., & McNabb, M. (1975). Therapy of a 6-year old who York: SP Medical and Scientific Books.
committed fratricide. Journal of the American Academy of Sorrells, J. (1977). Kids who kill. Crime and Delinquency, 23,
Child Psychiatry, 14, 319336. 312320.
Children Who Murder 115

Sorrells, J. (1980). What can be done about juvenile homicide? Walshe-Brennan, K. S. (1977). A socio-psychological investigation
Crime and Delinquency, 26, 152161. of young murderers. British Journal of Criminology, 17, 5863.
Spergel, I. A. (1983). Violent gangs in Chicago: Segmentation and Wood, W., Wong, F., & Chachere, J. (1991). Effects of media
integration Unpublished manuscript, University of Chicago. violence on viewers aggression and unconstrained social in-
Spergel, I. A. (1995). The youth gang problem: a community ap- teraction. Psychological Bulletin, 109, 371383.
proach. New York: Oxford University Press. Woods, S. M. (1961). Adolescence violence and homicide. Archives
Stearns, A. W. (1957). Murder by adolescents with obscure motiva- of General Psychiatry, 5, 3844.
tion. American Journal of Psychiatry, 114, 303305. Yates, A., Beutler, L. E., & Crago, M. (1984). Characteristics
Tooley, K. (1975). The small assassins: Clinical notes on a subgroup of young, violent offenders. Journal of Psychiatry and Law,
of murderous children. Journal of the American Academy of 16, 137149.
Child Psychiatry, 14, 306318. Yarvis, R. M. (1991). Homicide: Causative factors and roots. Lex-
Toupin, & Morissette, (1990). Juvenile homicide: A case control ington, MA: Lexington Book Company.
study. Medicine & Law, 9, 986994. Zagar, R., Arbit, J., Sylvies, R., Busch, K., & Hughes, J. (1990).
Tucker, L. S., & Cornwall, T. P. (1977). Mother-son folie a deux: Homicidal adolescents: A replication. Psychological Reports,
A case of attempted patricide. American Journal of Psychia- 67, 12351242.
try, 134, 11461147. Zahn, M. A., & McCall, P. L. (1999). In M. D. Smith & M. A.
U.S. Department of Justice. (1987). Survey of youth in custody Zahn (Eds.), Homicide: A sourcebook of social research. (pp.
(Bureau of Justice Statistics special report). Washington, DC: 923). London: Sage Publications.
Government Printing Office. Zenoff, E. H., & Zients, A. B. (1979). Juvenile murderers: should
Walshe-Brennan, K. S. (1975). Children who have murdered. Med- the punishment fit the crime? International Journal of Law
ico Legal Journal, 43(1), 2024. and Psychiatry, 2, 533553.
Copyright of Clinical Child & Family Psychology Review is the property of Springer Science & Business
Media B.V. and its content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email articles for
individual use.

Potrebbero piacerti anche