Biopsychosocial Characteristics of Children Who Later Murder: A Prospective Study

Abstract

The authors document the childhood neuropsychiatric and family characteristics of nine male subjects who were clinically evaluated as adolescents and were later arrested for murder. Those subjects are compared with 24 incarcerated delinquents who did not go on to commit violent offenses. The future murderers displayed a constellation of biopsychosocial characteristics that included psychotic symptoms, major neurological impairment, a psychotic first-degree relative, violent acts during childhood, and severe physical abuse. The authors relate this combination of factors to prediction of violence and discuss ethical issues that are involved in intervention to prevent violence.

Murder is the most serious of all crimes. Therefore, knowledge that can help to predict and thus potentially prevent extraordinary violence is worth reporting. The purpose of this paper is to document the psychological, neurological, and experiential factors that consistently appeared in the cases of nine adolescents who later committed murder and to compare these adolescents with a sample of 24 delinquents who did not go on to commit violent acts within 6 years of discharge from a juvenile correction facility. With the exception of a single paper describing one young murderer (1), to the best of our knowledge all of the literature on the neuropsychiatric antecedents of homicide is retrospective.

There is a controversy in the literature regarding the prevalence and severity of mental illness in murderers. McKnight et al. (2) found that 77% of 100 murderers received psychiatric diagnoses (schizophrenia, especially paranoid, manic-depression, psychopathic personality, and epilepsy). In fact, 55% were unfit to stand trial and 27% were found not guilty by reason of insanity. That study was, however, skewed by the fact that the sample was taken from a prison hospital. Several other authors (3–7) have found murderers to have a higher prevalence of psychoses, especially the schizophrenias, than the general population. Wolfgang (8), on the other hand, found that only 3% of the murderers in Philadelphia prisons were insane. Gillies (9) found 90% of his Scottish sample of murderers to be normal and free of psychiatric diagnoses. Once again, results vary (2, 10). Wong and Singer (11) found that only 7% of the 621 murderers they studied in Hong Kong were mentally ill or not guilty by reason of insanity.

The literature on homicidal aggression from the 1940s, 1950s, and early 1960s focused primarily on social and psychodynamic factors to the almost complete exclusion of neurobiologic factors (12–18). Sargent (13) described five murderous children and hypothesized a family conspiracy in which the child who killed acted out an unconscious parental wish. Easson and Steinhilber (12) presented eight cases of attempted homicide, none of which was successful. Their conclusion was similar to Sargent's: “All cases demonstrate that one or both parents had fostered and had condoned murderous assault.” Michaels (19), using Easson and Steinhilber's clinical data, highlighted factors that Easson and Steinhilber had minimized, namely, history of enuresis, epilepsy, and abuse. Martin (14) evaluated the psychodynamics of two adolescent murderers but failed to elaborate on several symptoms of organic brain impairment, which he mentioned only in passing. Smith (15), after evaluating eight young murderers, concluded that they suffered from early experiences of deprivation which resulted in underdeveloped egos and vulnerability to outbursts of violent aggression. Miller and Looney (18) theorized that adolescents who tended to dehumanize others were at greatest risk of committing murder when their wishes were thwarted. In 1978 McCarthy (20) discussed the psychodynamics of 10 adolescent murderers and found that “narcissistic disturbances, particularly an impaired capacity for self-esteem regulation and underlying narcissistic rage, were related to homicidal behavior.” Malmquist (21) suggested that homicide “can serve the illusory function of saving one's self and ego from destruction by displacing onto someone else the focus of aggressive discharge.”

Much has been written about the association of parental brutality and homicidally aggressive behavior (3, 16–18, 22–24). King (17) noted that the nine adolescent murderers he studied were “often singled out for abuse.” The presence of repeated violence and abuse in the environment of many adolescent murderers led Pfeffer (25) to view much of the adolescent's assaultive and homicidal behavior as an attempt to master the trauma he has experienced by controlling and victimizing others. Sendi and Blomgren (3) noted that an exposure to extreme violence or murder differentiated their group of 10 murderers from a control group. They also reported that seduction by or sexual perversion of a parent was associated with homicidal youth.

Duncan and Duncan (22) and Lander and Schulman (26), in studies based on small numbers of cases, described destructive and nonnurturing parent-child relationships and hypothesized that these may have led to acting out of intense hostility and homicidal behavior. McCarthy (20) asserted that early deprivation is often associated with homicidal behavior. Corder et al. (23, 24) reported severe psychopathology in parents of homicidal adolescents. They described “extreme maladjustment, e.g., chronic alcoholism, repeated hospitalization for psychosis, and many incarcerations for criminal acts.”

There is a continuing debate in the literature about whether psychomotor epilepsy and violence are associated with each other. Although some authors (27–34) have called attention to an association between the two, others (35–39) have questioned the relationship.

King (17) and others (40, 41) have reported extensive information pointing to the presence of a continuum of learning deficits and neurological problems associated with youthful homicide. Lower intelligence and mental retardation have also been reported to play a role in homicide (2, 6, 10, 42). King (17) and Lion (43) studied the effects of cognitive and language communication deficits in homicidal youths. Their studies showed that these adolescents may have an “inability” or “disinclination” to master language and basic communication skills.

The studies we have cited can be characterized in two ways. First, each study tends to focus on a single dimension of behavior (e.g., psychodynamic, neurological, experiential). Second, all of the studies of more than one subject are retrospective. That is, they reconstruct childhood factors after murder has occurred. For these reasons, we considered it important to report the neuropsychiatric and psychosocial characteristics of nine murderers examined before their commission of homicidal acts.

Method

Samples

Our subjects consisted of two groups. The first was a sample of nine boys who had been neuropsychiatrically evaluated when they were between the ages of 12 and 18 years (mean=14.5 years, median=14.0 years) and who subsequently were charged with murder. Five had been evaluated while incarcerated in a juvenile correctional facility, one had been evaluated at a juvenile court clinic, one had been evaluated during his late teens while in prison for robbery, and two had been evaluated during psychiatric hospitalizations before committing murder.

To determine whether these nine subjects differed in any identifiable ways from ordinary incarcerated delinquents who did not go on to commit violent acts as late adolescents and young adults, we compared the nine subjects with 24 incarcerated delinquent boys from an earlier clinical study who, at the time of follow-up 6 years after evaluation, had no known arrests for serious felonies (class A or B) according to state police records. They had been neuropsychiatrically evaluated when they were between the ages of 10 and 16 years (mean=15.0 years, median=15.5 years) and were incarcerated in a juvenile correctional facility. This comparison group was composed of youngsters whose offenses resulting in juvenile incarceration ranged from nonviolent acts, e.g., burglary, breach of the peace, violation of probation (N=9), to serious violence, e.g., assault with a weapon, rape (N=15).

Sources of Clinical Data

Our clinical data for the nine subjects who murdered consisted of all available neuropsychiatric records that predated each subject's act of murder. Comprehensive psychiatric and neurological evaluations were available on eight of these subjects, including detailed medical and family histories. In the ninth case, the results of a preliminary psychiatric assessment and information from a parental interview were available; however, no neurological data or data regarding physical abuse were available.

The clinical data for the 24 comparison subjects consisted of psychiatric evaluations performed 6 years before follow-up, which included detailed medical and family histories. In 22 cases, the results of neurological evaluations were also available. The nature of these neuropsychiatric evaluations has been described elsewhere (44).

The use of records antedating each subject's act of murder has advantages and disadvantages. Because the data on the murderers were gathered at different institutions (e.g., court clinic, psychiatric hospital) and under different conditions and were not intended specifically for the purpose of studying antecedents to homicide, they are not uniform. In the case of the 24 comparison subjects, the data were gathered in the course of a clinical study of neuropsychiatric and family characteristics of incarcerated delinquents and were more uniform. However, in both samples, because the evaluations from which data were gathered had been, in effect, prospective and not retrospective, they had been unbiased by prejudices which might have been associated with an examiner's knowledge that a subject had committed murder.

Presence or absence of the following neuropsychiatric signs and symptoms was recorded: auditory, visual, or olfactory hallucinations; loose, rambling, illogical, or concrete thought processes; paranoid delusions; extreme sadness or depression; suicidal behavior; cruelty to animals; inability to stop fighting; frequent headaches; loss of consciousness; and abnormal EEG, other evidence of seizures, or other signs of major or minor neurological abnormalities. Symptoms and signs were considered positive if mentioned explicitly in the records and an example was given of them. Because symptoms may exist but not be noted, this method would underestimate the extent of psychopathology. We also recorded relevant medical data, including evidence of perinatal problems and a history of any illnesses or accidents affecting the CNS.

Information regarding the subjects' families included family constellation, occupations, and specific psychosocial problems, such as psychiatric hospitalizations, extreme violence within the family, and child neglect or abuse. A child was considered to have been abused by his parents or guardians if he had been punched; beaten with a stick, board, pipe, or belt buckle; beaten with a belt or switch other than on the buttocks; or deliberately cut, burned, or thrown downstairs or across a room. A child was considered not to have been abused if he had been struck with an open hand or beaten with the leather part of a belt or switch on the buttocks only. In addition, we recorded the timing and nature of the subjects' previous aggressive acts and delinquent offenses.

Follow-Up Data

For the nine murderers follow-up data were obtained as follows: in six cases data came from police arrest records, in two cases it came from newspaper reports, and in one case it came from a telephoned report by a corrections officer to the correctional school from which the subject had absconded. All follow-up data on the comparison sample came from state police arrest records. Of note, the commission of adult violent crimes in the comparison group is undoubtedly an underestimate, since it was known that some subjects had moved from the state and their offenses therefore might not have been known to the state in which they were incarcerated as adolescents.

Results

The nature of the homicidal acts of the nine murderers and the quality of their juvenile violent behaviors are summarized in table 1. As can be seen, all nine had manifested extreme violence as children and adolescents. In most cases there was evidence that extreme violence had occurred several years before the commission of murder. For example, one boy had committed multiple sexual assaults at least 5 years before committing murder, another robbed at knifepoint 10 years before committing murder, and another assaulted a woman with a knife and robbed her 3 years before committing murder. At least four of the boys who later murdered were extraordinarily violent in early childhood: one burned his bed when he was 4 years old, another was too violent to be permitted into grade school and required instruction at home starting at age 10, another threatened a teacher with a razor at age 10, and another choked a bird at age 2 and threw a dog out of a window at age 4.

TABLE 1.

Early Violence and Later Murderous Acts of Nine Male Subjects

	Early Violence		Later Murderous Acts
Subject	Age	Act	Age	Act
1	4 years	Burned bed	19 years	Charged with felony murder and assault with dangerous weapon
	Early adolescence	Multiple sexual assaults
	14 years	Raped young boy
2	16 years	Robbery at knifepoint	26 years	Stabbed bus driver in altercation over fare
	Adolescence	Multiple assaultive acts
3	Fifth grade	Too aggressive to attend school, was instructed at home	24 years	Shot girlfriend and two of her relatives, killing one and wounding the other
	13 years	Robbed gas station
	14 years	Assaulted younger boy
4	14 years	Robbery at knifepoint	20 years	Charged with felony murder, robbery, kidnapping, and sexual assault
	16 years	Assaulted and kidnapped two homosexual men
5	13 years	Assaulted 85-year-old woman, resulting in her breaking a hip	15 years	Shot butcher with whom he was not acquainted, no apparent motive
	14 years	Robbery at gunpoint; assaulted teacher with hammer
6	School age	Threw caustic liquid in child's eyes	20 years	Shot two young children when unable to locate their uncle, with whom he was angry
	14 years	Threw hammer at peer
7	2 years	Choked bird	18 years	Raped and stabbed woman 13 times
	4 years	Threw dog out window
	Middle childhood	Broke sibling's arm
	16 years	Assaulted and raped girl
8	10 years	Attacked female teachers; threatened teacher with razor	17 years	Charged with felony murder and arson
9	14 years	Assaulted sister with knife	18 years	Bludgeoned teenager to death after altercation in arcade
	15 years	Robbery

When one considers the nature of the homicidal acts the subjects committed as older adolescents and young adults, one is impressed by the fact that in most cases the acts seem to have been mindless, impulsive, and unpredictable. Indeed, in five of the six cases in which information regarding the victims was available, the victim was unknown to the assailant until just before the murder (e.g., a bus driver, a butcher, two children). Thus, many of the acts seemed spontaneous rather than premeditated. Whether or not they occurred during psychotic episodes, episodes of organically influenced dyscontrol, or even seizures could not be determined from the available data.

The psychiatric, neurological, and family characteristics of each of the nine murderers are summarized in table 2. As can be seen, psychotic symptoms, especially paranoid ideation, were prevalent in this sample, and three of these subjects had been psychiatrically hospitalized during early adolescence. Severe neurological impairment was especially prevalent. Of the eight for whom neurological information was available, three had histories of grand mal seizures and abnormal EEGs, one was macrocephalic and had an abnormal EEG, and three others had demonstrable lapses of fully conscious contact with reality witnessed by others and a variety of psychomotor epileptic symptoms, including metamorphopsias, frequent déjà vu, and olfactory hallucinations. Of note, six had received severe head injuries in childhood, including falls from roofs and car accidents resulting in loss of consciousness.

TABLE 2.

Clinical and Family Characteristic of Nine Male Murderers

Subject	Psychotic Symptoms	Major Neurological Impairment and Head Injuries	Psychosis and/or Psychiatric Hospitalization Among Relative	Physical Abuse
1	Paranoid ideation; illogical, concrete thinking	Multiple psychomotor epileptic symptoms, seizure disorder suspected by neurologist; face and head beaten	Mother and maternal grandfather hospitalized	Father and mother beat him with belts and sticks in face and head
2	Paranoid ideation; bizarre behaviors; loose, illogical thinking; visual and auditory hallucinations	Seizures in infancy; abnormal EEG; multiple psychomotor epileptic symptoms	Natural mother hospitalized and given ECT at age 14, while pregnant with subject	Adoptive parents denied abuse
3	Paranoid behavior; refusal of all verbal communication with examiner; bizarre, violent drawings	No data on neurological status; fall from porch at age 3 while riding bicycle, loss of consciousness	Mother hospitalized; father hospitalized and diagnosed as schizophrenic	–
4	Paranoid ideation	Grand mal seizures in childhood; abnormal EEG; fall from second story at age 9, loss of consciousness	Mother hospitalized; father brain damaged, epileptic, possibly psychotic	Father extremely violent, beat son and was jailed for beating wife
5	Previous psychiatric hospitalization; paranoid ideation; rambling, illogical thinking; auditory hallucinations	Lapses of consciousness; head injury in car accident at age 9; loss of consciousness due to hitting left temporal region with hammer	Sister hospitalized; father thought to be psychotic	Father beat him severely
6	Loose, rambling, illogical thinking; visual hallucinations; suicide attempt by jumping out window	Grand mal seizures in infancy and childhood; abnormal EEG; fall from roof at age 8	Father paranoid schizophrenic	Father beat him and took him to church several times a day to exorcise devil and cure seizures
7	Paranoid ideation; auditory hallucinations; bizarre behaviors; diagnosis of paranoid schizophrenia; inserted thermometer into bladder and did not remember it	Seizure disorder suspected by neurologist; multiple psychomotor symptoms; car accident at age 13, loss of consciousness	Mother hospitalized; father expelled from police force	Mother beat him, broke finger
8	Previous psychiatric hospitalization; paranoid ideation; auditory hallucinations; loose, illogical thinking; attempt to burn self with matches	Macrocephaly (arrested hydrocephalus); abnormal EEG; multiple psychomotor epileptic symptoms	Sister and maternal grandfather hospitalized	Mother beat him with belt buckle, cord, and wire while he was naked
9	Previous psychiatric hospitalization; paranoid ideation; concrete thinking; bizarre behaviors	Normal EEG, normal results on neurological examination; fall from bus at age 13, loss of consciousness	Father psychotic	Father beat him, tried to kill him and his brother

As can also be seen in table 2, all nine of the murderers had a first-degree relative who had been psychiatrically hospitalized and/or demonstrably psychotic. Five had mothers who had been psychiatrically hospitalized, whereas of four fathers known to have been psychotic, only one was known to have been hospitalized. Given the fathers' violent behaviors, it is likely that they were incarcerated rather than hospitalized.

That abuse and violence were rampant in the households of most of these boys is also illustrated in table 2. Of the eight for whom information on abuse was available, seven had been severely abused by one or both parents. Belt buckles, cords, broom handles, sticks, and shoes were used to inflict punishment. The mother of one boy broke his finger while she was under the influence of alcohol. The hospital record of another boy revealed that his father had attempted to kill him and his brother on several occasions. Records revealed that six of the boys had witnessed extreme violence within their households. They had observed their parents assaulting each other and other members of the family. One boy had been present at the stabbing of his uncle. The father of another youth was constantly involved in fights outside the family, during one of which he inflicted brain damage on his victim by bludgeoning him with a metal pipe.

Thus, all nine boys were severely neuropsychiatrically impaired, and all had close relatives who were psychotic and/or had been psychiatrically hospitalized. Most had both witnessed and been the victims of violence. Furthermore, all had been extraordinarily violent long before they committed murder.

Comparison of Murderers With Ordinary Delinquents

The results of the comparison of the nine murderers and the 24 incarcerated delinquents who did not go on to commit violent crimes can be seen in table 3. The most powerful individual variables distinguishing the groups were psychotic symptoms and neurological impairment. The difference between the groups in the proportion with first-degree psychotic relatives approached significance. It is somewhat surprising that neither early violence alone nor a history of abuse strongly distinguished the groups from each other. Rather, the combination of all five variables most strongly distinguished the murderers from the ordinary delinquents.

TABLE 3.

Early Neuropsychiatric, Family, and Experiential Characteristics of Children Who Later Murdered and of Incarcerated Delinquents Who Did Not Murder

	Murderers			Ordinary Delinquents
		With Variable			With Variable		Chi-Square Analysis (with Yates' correction)
Variable	With Data Available	N	%	With Data Available	N	%	χ2 (df=1)	p
Psychotic symptoms	9	9	100	24	12	50	5.1	<.05
Major neurological impairment	8	7	88	22	6	27	6.4	<.02
Physical abuse	8	7	88	24	14	58	1.2	n.s.
Psychiatric hospitalization or psychosis among first-degree relatives	9	9	100	24	14	58	3.6	<.10, >.05
Serious violence as juvenile	9	9	100	24	15	62	2.9	<.10, >.05
All five variable	8	6	75	22	2	9	9.9	<.005

Discussion

Individually, each of the factors associated with the extreme violence of the nine homicidal subjects often exists to a greater or lesser extent in the backgrounds of other aggressive individuals. It was the coexistence of all of these neuropsychiatric and family factors and their severity that distinguished these youngsters most clearly from the comparison group of ordinary delinquents.

It seems that severe CNS dysfunction, coupled with a vulnerability to paranoid psychotic thinking, created a tendency for the nine homicidal subjects to act quickly and brutally when they felt threatened. Living within psychotic households, they were frequently the victims of and witnesses to psychotic parental rages, experiences that undoubtedly further exacerbated their tendencies toward the physical expression of violence. Whether or not these youngsters dehumanized their victims, as Miller and Looney (18) suggested, could not be determined. If, indeed, they did, these objectively identified neuropsychiatric vulnerabilities undoubtedly contributed to their distortions of reality.

The finding that the parents of these youngsters were often both psychotic and violent may shed light on earlier studies purporting to document a genetic predisposition to violence (45). Previous studies of large samples of antisocial individuals have revealed that their offspring are significantly more likely to become antisocial, even when raised apart from the biological parents. The parents of these nine murderers were violent and, although police records were not available, we suspect that many had come in conflict with the law. So, indeed, had their children. Were one to limit one's investigation to issues of violence, it would be easy to hypothesize an inherited predisposition to violence on the basis of the family violence of these murderous youngsters. Our greater knowledge regarding the neuropsychiatric characteristics of our subjects and their families suggests an alternative hypothesis. That is, if there was an inherited predisposition to maladaptive antisocial behavior in these nine boys, the inheritance probably was a predisposition to psychosis and/or neurological dysfunction. These vulnerabilities in the context of violent households manifested themselves in uncontrolled violence. Indeed, in the case of our one adopted subject, his adoptive parents were not psychotic. His biological mother, however, was psychotic and was hospitalized and undergoing ECT at the time of his birth.

The constellation of neuropsychiatric impairment, parental psychosis, a history of physical abuse, and prior acts of violence in delinquents who later murder raises the question of prediction. Does the finding of this constellation of factors in a given child justify a prediction of future violence and therefore a need for intervention? As Monahan (46) has shown in his extensive review of the literature, clinicians are notoriously poor at accurately predicting violence. The major drawback to the use of clinical predictors is the tendency to overpredict and therefore to stigmatize needlessly. This danger is especially true of studies purporting to predict violence early in life before it manifests itself in dangerous ways (47–49).

Our research, in contrast to earlier clinical studies, identifies clinical and family characteristics of youngsters who have already demonstrated extreme aggression. One boy kidnapped while threatening his victims with a pitchfork, two had raped and physically assaulted their victims, one had menaced his victims with guns and knives, one had committed armed robbery, one was so violent in childhood he could not attend school, and one threw caustic liquid in the face of another child. Surely, when the constellation of psychotic symptoms, severe neuropathology, and the existence of psychotic, abusive family members is identified in an already violent youngster, responsible intervention is justified. Although it is impossible to predict that a youngster with these attributes will commit murder in the future, it seems safe to conclude that in the absence of appropriate intervention he will commit further acts of violence. Moreover, were the question of future violence not an issue, the serious psychopathology, neurological impairment, and family violence identified would justify intervention. This kind of youngster has an immediate need for help regardless of considerations of prediction. The appropriate interventions to meet the current needs of these youngsters and adolescents are also those most likely to diminish future violence. Thus, they have value for prevention.

To date, the most frequently cited studies of the youthful characteristics of seriously antisocial adults (50, 51) focus primarily on the number and nature of previous antisocial acts as predictors of future maladaptive behaviors and, in essence, point only toward a policy of preventive detention. Our findings suggest that violence alone is not as good a predictor of future aggression as is violence coupled with neuropsychiatric vulnerabilities, parental brutality, and parental psychosis.

Each of the clinical factors identified in this study has implications for potentially therapeutic interventions. A recognizable constellation of biopsychosocial factors such as that described signals a need for specific medical, psychiatric, and social assistance. In this way, these findings differ from those of Wolfgang et al. (51), who used immutable factors such as number and nature of past offenses and race for predictive purposes.

Murder is the most serious of crimes. It is, therefore, the one we most need to learn to prevent. The finding of a constellation of identifiable factors characteristic of violent youngsters before their commission of homicidal acts and the ability to differentiate them from ordinary delinquents challenges us to develop programs to recognize and treat these multiply handicapped children.