Adolescents murderers: abuse and adversity in childhood

Adolescents murderers: abuse and adversity in childhood

Journal of Adolescence 2002, 25, 221–230 doi:10.1006/jado.2002.0462, available online at http://www.idealibrary.com on Adolescents murderers: abuse a...

112KB Sizes 13 Downloads 200 Views

Journal of Adolescence 2002, 25, 221–230 doi:10.1006/jado.2002.0462, available online at http://www.idealibrary.com on

Adolescents murderers: abuse and adversity in childhood ANDREW J. HILL-SMITH, PIPPA HUGO, PATRICIA HUGHES, PETER FONAGY AND DAVID HARTMAN

With rising rates of juvenile violence, an improved understanding of its causes is much needed. The relative absence of controlled studies based on British populations of murderers further increases the need for information in this area. This case control study examines a group of 21 imprisoned males, convicted of murder whilst aged 18 years or less, and compares them on a range of psychosocial variables to 21 convicted non- violent burglars. The groups were matched for age at interview and race. Results showed that lower socio-economic status, harsh parenting from both mother and father, and exclusion from school were significantly more common for murderers. Harsh parenting from the mother than from the father appeared to contribute more strongly. When factors were combined in a general adversity index, a significant difference was found between the two groups. The study confirms that multiple environmental factors are associated with murderous behaviour in young men. r 2002 The Association for Professionals in Services for Adolescents. Published by Elsevier Science Ltd. All rights reserved

Introduction As in most western countries, violence among youth in the U.K. has been increasing over the last decade with an associated, but less dramatic rise in murder (Criminal Statistics, 1999). This has occurred despite a drop in overall crime rates. If we are to make a real impact on the problem, researchers must seek to understand the underlying contributory factors. There are few studies of adolescent murderers and the majority are largely observational and uncontrolled. Most have been conducted in the U.S.A. Inevitably in a research field studying rare events, studies have methodological weaknesses including reliance on retrospective data, and small sample numbers (Widom, 1989). There is therefore a need for up-to-date studies of British subjects that include the use of a control group. This study attempts to address this deficit.

Research on adolescent delinquency Research has found considerable overlap in factors that are associated with delinquency and those associated with violence. The West and Farrington classic longitudinal studies of 411 Camberwell boys found that vulnerability to delinquency at the age of 18 years was predicted if three out of five vulnerability factors were present at age 8–10. These included: low family income, large family size, convicted parents, low non-verbal intelligence, poor parental child-rearing behaviour (e.g. harsh parenting and poor supervision) (West and Farrington, 1973). Farrington subsequently added three further vulnerability factors at age 8–11: poor housing, Reprint requests and correspondence should be addressed to: Andrew J. Hill-Smith, Frimley Children’s Centre, Church Rd, Frimley, Camberley, Surrey, GU16 5AD, U.K. 0140-1971/02/$3500+000

# 2002 The Association for Professionals in Services for Adolescents. Published by Elsevier Science Ltd. All rights reserved

222

A. J. Hill-Smith et al.

antisocial child behaviour (high daring, lack of caring, troublesomeness and discipline difficulty), and tallness (Farrington, 1989). Other research has confirmed the importance of multiple factors. These include temperamental characteristics, notably hyperactivity and impulsivity (Babinski et al., 1999); social deprivation; parenting practices, especially poor supervision (Snyder et al., 1986), poor attachment (Rankin and Wells, 1990), physical maltreatment (Kratcoski, 1982); antisocial peer relationships (Dodge, 1986) and educational failure (Sturge, 1982). Other factors associated with aggression, include the experience of physical abuse amongst young children (Dodge et al., 1990) and explosive parenting (Patterson, 1982; Patterson et al., 1989). Physical aggression has been noted to be a stable characteristic of children, but not a fixed feature (Rutter, 1989). Studies of the genetics of childhood antisocial behaviour have reported a heritability contribution for antisocial behaviour ranging from 69% (Silberg et al., 1996) to 7% (Lyons et al., 1995) based on conduct disorder symptomatology. These findings seem to be influenced by design, notably reporting source for symptomatology (greater rates from parental report) and whether hyperactivity/inattention symptoms are included. Rutter et al. (1999) concluded that the genetic component for aggression has been consistently shown to be higher than that for stealing.

Research on adolescent murderers Factors which are associated with adolescents who have killed include large family size, and family disruption (Fiddes, 1981), family and paternal violence (Bailey, 1996), sexual abuse (Boswell, 1995; Bailey, 1996), alcohol abuse (Myers and Kemph, 1990; Labelle et al., 1991) and neuropsychological abnormalities (Woods, 1961; Zagar et al., 1990; Bailey, 1996). Methodologically, strong studies from the U.S. have found some similar factors. The Dorothy Lewis et al. (1988) controlled prospective study of young adolescents of whom 13 subsequently became juvenile murderers, found that major neurological impairment, episodic psychotic symptoms, family violence and physical abuse were significantly more common for murderers compared to non-violent delinquents. In Busch’s study of 1956 adolescent delinquents (Busch et al., 1990), he compared 71 homicide perpetrators with a matched group of 71 non-violent delinquents. The homicide perpetrators were significantly more likely to have a criminally violent family member (relative with a criminal record of violence), gang membership, severe educational difficulties (IQ less than 70) and alcohol abuse (by young person). These factors predominantly show greater adversity for homicidal adolescents, but there are exceptions, such as a finding of higher school drop out rates, and previous psychiatric treatment amongst controls (burglars) (Cornell et al., 1987). Girls tend to respond to biopsychosocial adversity differently (Lewis et al., 1991), show lower rates of conduct disorder and may follow different paths to conduct disorder (Von Knorring et al., 1987). The incidence of murder by female adolescents is considerably lower, for example in the U.K. in 1996 there were 22 convictions of males under 19 years for murder and one for females (data from Homicide Index, Home Office). The paths to aggression, violence, criminality and murder are varied and involve a complex interaction between biological and social influences. It is likely that similar findings will be found in a U.K. sample. This study aims to assess whether adolescents convicted of murder had experienced more adversity and physical violence during childhood than a matched group of non-violent burglars.

Adolescents murderers

223

Methods Subjects Subjects were imprisoned males convicted of murder and detained under the Criminal Justice Act, who were under 19 years at the time of the offence. The age cut-off was chosen in order to ensure an adequate sample size, whilst adhering to adolescent years. We excluded adolescents convicted of manslaughter, those who were not fluent in English, or who scored very low on the National Adult Reading Test (IQ equivalent o70). As participants were selected after conviction, our sample was defined by the rigorous process of criminal conviction. Females were excluded because they are few in number and follow different routes to criminal conviction and violence (Von Knorring et al., 1987) so their inclusion would weaken the methodology.

Controls Controls were non-violent imprisoned males convicted of burglary matched for race and age at interview who otherwise met the above criteria. The choice of non-violent burglars was mindful of previous findings that few differences exist between violent offenders and murderers (Lewis et al., 1988; Toupin and Morissette, 1990).

Procedure The study was conducted at three Young Offenders Institutions. Permission for the study was obtained from the Home Office Health Research Ethics Committee and from the Governor of each establishment. Subjects gave informed written consent to be interviewed. Consent was obtained from parents or guardians where subjects were younger than 18 years. Prison records of potential subjects at the participating institutions were examined to find age and race matched subjects convicted of burglary. They were then screened for involvement in serious violent acts. Those that revealed evidence of serious violence were excluded. Serious violence was defined as causing a broken bone (other than a nose), stabbing, knocking out the victim, causing injury requiring hospitalization or significant medical attention, a history of excessive fighting, violent crime, or arson. Both subject and control adolescents were seen in the young offender institutions in which they were detained. Twenty-seven subjects who were convicted of murder and who met our age criteria were identified and initially interviewed. Three were in the process of appealing against conviction and were excluded. A further 3 (11?0%) refused to participate, leaving 21 adolescents who consented to take part. To find appropriate controls, 87 age-matched adolescents convicted of burglary were screened for involvement in serious violent acts. Fifty-four (62?1%) were deemed to be too violent for the purposes of the study and a further 12 (13?8%) refused to participate. Twentyone controls were interviewed. One control participant refused to continue mid-way through the interview. Available data from this subject were included in the analysis.

Measures The following measures were used: (1)

Sixteen-item semi-structured interview assessing demographic and psychosocial disadvantage.

A. J. Hill-Smith et al.

224

(2)

(3)

(4) (5)

The Family structure and function section of the Child and Adolescent Psychiatric Assessment Interview, version 2?0 (Angold et al., 1995). The time frame for enquiry about family functioning used was ‘‘when living with your parents/parental figures in the last 5 years prior to imprisonment’’. History of experience, perpetration and attitude towards violence. A 16 question 40 min semi-structured interview was designed for the study. The interview was tape recorded and transcribed. Relevant components are history of committing violent acts (used to confirm non-violence among controls), and history of physical mistreatment by family members. Educational attainment and reading difficulty. National Adult Reading Test supplemented by the Schonell reading test if required by a low scoring response (Nelson, 1982).

Harsh parenting. Harsh parenting was coded separately and was measured from items 2 and 3 in the assessment. It was defined as physical violence directed to the subject by a parent, which was usually repeated and likely to cause injury, such as bruising. A single incident may be included provided it was of sufficient severity. Any threats to life, threats including use of a weapon, use of an implement that may cause injury, or hitting to the face, were included. Being slapped or hit with an object where there was no possibility of causing injury was not included.

General adversity score. A score was devised for general adversity using factors from the West and Farrington vulnerability index (1973): low family income (social class 4, 5 or 6), large family size (>5 children), convicted parents, low non-verbal IQ (below 80 on the NART), poor parental child-rearing behaviour (harsh parenting, as defined above). The presence or absence of each of these factors was given a score of 0 or 1 and summed to give a total score.

Statistical analysis Data were coded and analysed using SPSS for Windows version 8?0. For between-group categorical comparisons, two-tailed Fisher exact test was used throughout. General Adversity continuous scores were compared between the two groups using the Mann–Whitney U test. A further analysis was conducted dividing cases categorically into high (3 or above) and low scorers (below 3), and then analysed using the two-tailed Fisher’s exact test.

Characteristics of the sample The mean age at interview was 19?7 years (range 17?4–21?5) for murderers and 19?3 years (range 17?0–21?0) for burglars. Age at offence was 16?1 years (range 12?8–19?2) for murderers and 18?4 years (range 16?6–20?0) for burglars (p=o0?0001, two-tailed t-test). There was no significant difference in NART scores (99?4 for murderers and 98?9 for burglars). The paternal socio-economic group was significantly lower for the murderers (10?4 vs. 7?9) (two-tailed t-test p = 0?014), and shows a non-significant trend for mothers (11?4 vs. 9?9). Rates of parental separation or divorce on one or more occasion were lower for murderers: 57?1% vs. 70?0%. Because of difficulty in identifying enough race-matched

Adolescents murderers

225

burglars there was one more ethnic minority subject in the murderer group: 6 (28?6%) murderers were from ethnic minority groups vs. 5 burglars (19?0%). Subjects in different racial groups were not distinguishable on any of the above variables. Three murderers and 3 burglars reported being in a cohabiting relationship prior to imprisonment, and 3 burglars and 4 murderers had children. More than 75% of both subjects and controls experienced at least one change in parental arrangement, although there was no significant difference between the groups (median 3, range 0–12 for murderers and median 3 and range 0–16 for burglars). Despite our attempts to exclude violent controls, the violence interview transcripts identified 5 burglars who reported violence, which they had not revealed at screening. Inclusion of data from these individuals did not significantly affect the findings, except for one variable (paternal imprisonment, see below), therefore data from all 21 control adolescents are included in the results.

Results A reported history of harsh parenting from both mothers and fathers was higher in the subject group, and approached statistical significance for mothers (p = 0.096) (Table 1). Murderers were significantly more likely to have experienced harsh parenting from both parents (p = 0?019). Conversely, murderers showed lower rates of neither parent being harsh (p = 0?32). In comparison to the general population, in both murderer and burglar groups, rates of paternal convictions and imprisonment were very common. The same applied to criminal conviction amongst other parents (non-resident biological parents) (Table 2). The rate of criminal involvement amongst the parents of murderers was higher for both mothers and fathers, especially for imprisonment (p = 0?079) and arrests (p = 0?054). When the violent burglars were excluded from the analysis, the difference between groups became significant for paternal imprisonment (p = 0?024) (Table 2). Drug and alcohol use by mothers was 15?0% compared with 5?0% among controls; drug and alcohol use by fathers was 14?3% compared with 5.0% among controls. Twenty-five per cent of mothers and 20?0% of fathers of murderers were reported to have had psychological problems, compared with 40?0% of mothers and 10?5% of fathers of burglars. Parental discord was common in both groups (38?1% and 47?4%). However, none of these differences reached statistical significance.

Table 1

Harsh parenting. Significance is based on two-tailed Fisher’s exact test Murder (%) (N)*

Harsh mother Harsh father Both parents harsh Neither parent harsh

50?0 52?6 38?9 31?6

(10 of 20) (10 of 19) (7 of 18) (6 of 19)

Burglars (%)(N) 19?0 31?6 5?3 52?6

(4 of 21) (6 of 19) (1 of 19) (10 of 19)

Significance (p) 0?096 0?32 0?019 0?32

Numbers vary slightly in the assessments because some adolescents did not have one parent present for much of their childhood or were unable to remember details.

A. J. Hill-Smith et al.

226

Table 2

Parental criminality. Significance is based on two-tailed Fisher’s exact test Murderers (%) (N)

Mother ever arrested Father ever arrested Mother convictions Father convictions Father imprisoned Other parent conviction

15?0 75?0 15?8 68?4 47?4 80?0

(3 of 20) (15 of 20) (3 of 19) (13 of 19) (9 of 19) (4 of 5)

Burglars (%) (N) 9?5 42?1 9?5 38?9 16?7 20?0

(2 (8 (2 (7 (3 (1

of of of of of of

21) 19) 21) 18) 18) 5)

Significance (p) 0?66 0?054 0?62 0?103 0?079 0?206

The murderer group showed up to a three-fold increase in the rates of psychosocial adversity for other family background and educational factors, compared to the control adolescents. This was particularly strong for educational attainment, notably for frequency of exclusion from school (murderers’ mean 4?1 (range 0–15) vs. burglars’ mean 1?65 (range 0–6)); p = 0?020 Mann–Whitney U test), history of reading problems, history of special schooling and percentage of schooling attended after age 10 (Table 3). Selective negative view of the child (less positive about index child than about siblings) from both parents were similar in subjects and control adolescents (35?0% vs. 44?4% for mothers and 47?4% vs. 35?5% for fathers). Overinvolvement (4?8% vs. 5?9% for mothers and 9?5% vs. 5?9% for fathers) and inadequate supervision/control items (95?2% vs. 100?0% for mothers and 94?4% vs. 94?4% for fathers) showed scores at the extreme end of their range for both groups suggesting that they are not very discriminating for this population. When multiple adversities (see above) were condensed into a general adversity index, significant between-group differences were found (murderers’ mean 2?50 (range 0–4) vs. burglars’ mean 1?72 (range 0–3); p = 0?018 Mann–Whitney U test). The difference remained significant if groups were categorized into presence or absence of vulnerability to delinquency, i.e. scores of 3 or more out of 5 (9 high scorers for murderer group and 2 for controls. Fisher exact test N = 36, p = 0?027).

Table 3

Educational difficulties. Significance is based on two-tailed Fisher’s exact test Murderers (%) (N)

Burglars (%) (N)

History of reading problem Frequency of exclusion from school Frequency of expulsion from school Percentage of schooling attended after age 10

33?3 (7 of 21) Mean 4?1 range (0–15) Mean 1?52 (range 0–13) Mean 58?4 (3–100)

10?0 (2 of 20) Mean 1?65 range (0–6) Mean 0?5 (range 0–2) Mean 72?7 (range 20–100)

Mean age of school leaving

14?3 years (range 11–16)

15?2 (range 7–17) years

History of special schooling Antisocial reason for school leaving

47?6 (10 of 21) 72?2 (13 of 18)

30?0 (6 of 20) 50 (10 of 20)

Significance (p) 0?08 0?020 Mann–Whitney U 0?14 Mann–Whitney U 0?12 Independent samples t test 0?22 Independent samples t test 0?34 0?20

Adolescents murderers

227

Discussion Comparison of harsh parenting between studies is complicated by differing definitions. We used a narrow definition, so may have underestimated the extent of parental violence experienced by our subjects. Offender’s accounts were not supplemented by official records, which may have further compounded the situation. However, the rates of harsh parenting that we found are similar in direction and magnitude to those reported from some studies (Lewis et al., 1988; Myers, 1995), and higher than in others (Zagar et al., 1990; Boswell, 1995). Our study was unusual in finding that subjects reported that their mothers had also been harsh, since harsh parenting has more usually been attributed to fathers (Bailey, 1996). In addition, we found that the murderers compared to the non-violent burglars reported a significantly higher rate of harsh treatment from both parents. This has not previously been reported. Parental conviction, and in particular paternal conviction, is emerging as an important factor associated with extreme antisocial behaviour in adolescent males. Zagar and Busch (Zagar et al., 1990) found rates of parental criminal conviction similar to ours. Farrington’s study (1989) reported lower rates of criminal conviction of fathers than we did, but the direction and magnitude of their findings were similar. Their lower rates may be explained by their use of a community sample and ours of an imprisoned cohort. The difference we found between subject and control adolescents did not reach significance, possibly due to our small sample size. We can speculate that the children of violent parents may have an inherited predisposition to aggression, or may have learned violent behaviour from frequent exposure to it from parents. They may have suffered physical trauma leading to neurological damage, or the prolonged experience of fear in early childhood may have caused pathological prefrontal development associated with poor impulse control and explosive behaviour (Kraemer, 1992). In addition, while parental conviction and imprisonment may be associated with parental violence, imprisonment will mean separation from the parent, further neglect, and fall in family income. Although temperament and neurological abnormalities have been recognized as possible contributing factors (Woods, 1961; Lewis et al., 1988; Zagar et al., 1990; Bailey, 1996), we did not assess them. Temperament is notoriously difficult to measure other than by direct observation in infancy, and unlikely to yield useful information at this stage in the subjects’ lives. We would have liked to assess brain function and language disorders (Myers and Mutch, 1992), but the complexity of arranging necessary investigations for this population was beyond the scope of the study. It remains an open question as to whether any abnormalities demonstrated would have been genetic or acquired as a result of violence and deprivation. We should keep in mind the possibility that these adolescents could have been temperamentally more difficult, and may have or have had genetic or acquired neurological pathology (Miller, 1988) which predispose to hostility from parents, conduct disorder, and impulsive and aggressive behaviour. Whether or not predisposing biological factors are present, the broad range of adversity experienced by the majority of the adolescent murderers in their early years is striking. It should be kept in mind that our comparison is with an already highly deprived population. Our finding of a significant difference for the general adversity index is therefore particularly pertinent, and underlines the severity of disadvantage in the backgrounds of most of these young murderers. Even when compared to a group of youths with criminal convictions these

228

A. J. Hill-Smith et al.

young men were more likely to have lower socio-economic status, a greater history of reading problems and experience of special schooling, more frequent exclusion from school, more harsh parenting, and a high rate of criminally convicted parents. There appears to be no area of their lives whether family, school or the wider society where they were less disadvantaged. Poverty, neglectful and violent parenting would predispose to early conduct disorder, poor school performance and onwards to a downward spiral of failure, disruptive behaviour, exclusion from mainstream school and any experience of success. For most of the murderer group it was hard to identify any protective factors. Factors that might have enhanced the development of strength or resilience or good self-esteem are notably scarce. For example, the high rate of harsh treatment from both mothers and fathers is not only an adversity but is also an absence of protective parenting. We found that only one burglar reported harsh treatment from both parents, compared to six in the murderer group, which at least increases the possibility of some affection and protection within the family. The strengths of the study include the case-control design, and the direct collection of detailed data from interviewing subjects, rather than from case notes or questionnaires. In addition, we used a well-defined sample that was not subject to the potential biases of ongoing clinical or court intervention. However, the study had some weaknesses, including small sample size and reliance on a single source of data, largely in the form of historical recall. Although the groups were well matched for age at the time of interview, the significant difference in age of offence meant that the murderers had been in prison for longer than the burglars, and this could have led to difference in how families were viewed or in accuracy of recall. The relative rarity of conviction for murder among adolescents made it difficult to find exact matches for race, and also meant that we included in our sample murders that were different in character. For example, the killing of a parent and a murderous attack on another youth both represent extreme violence but are likely to have different motivation. This study has met its aim of assessing whether adolescents convicted of murder are more likely to experience greater adversity and physical violence during childhood than a matched group of non-violent burglars. Whatever the cause of their murderous behaviour, many had demonstrated social and educational problems, and yet therapeutic and social intervention had been unable to prevent further deterioration. Children and adolescents who display severe behavioural problems are expensive educationally, medically and socially. They lie on the boundaries between medical care, social services and the law. There is still much to be learned about why children and adolescents commit violent crime. Whilst the media often portray such young people as intrinsically evil, it behoves the professions involved to take the lead in identifying early markers of aberrant behaviour and to encourage government to provide the resources to explore possible routes to prevention.

Acknowledgements Our sincere thanks go to the young men who participated in the research, to the staff of the Young Offenders Institutions, and to Fiona Reid of the Department of Public Health Sciences, St George’ Hospital Medical School, for statistical advice. The study was partly funded by the NHS Executive South Thames Research and Development Directorate.

Adolescents murderers

229

References Angold, A., Prendergast, M., Cox, A., Harrington, R., Simonoff, E. and Rutter, M. (1995). The child and adolescent psychiatric assessment (CAPA). Psychological Medicine, 25(4), 739–753. Babinski, L. M., Hartsough, C. S. and Lambert, N. M. (1999). Childhood conduct problems, hyperactivity–impulsivity, and inattention as predictors of adult criminal activity. Journal of Child Psychology and Psychiatry, 40(3), 347–355. Bailey, S. (1996). Adolescents who murder. Journal of Adolescence, 19, 19–39. Boswell, G. (1995). Violent Victims. Prince’s Trust. Busch, K., Zagar, R., Hughes, J. R., Arbit, J. and Bussell, R. E. (1990). Adolescents who kill. Journal of Clinical Psychology, 46(4), 472–485. Cornell, D. G., Benedek, E. P. and Benedek, D. M. (1987). Characteristics of adolescents charged with homicide: review of 72 cases. Behavioural Sciences and the Law, 5(1), 11–23. Criminal Statistics (1999). Criminal Statistics: England and Wales, 1998. London: HMSO. Dodge, K. A. (1986). Social information-processing variables in the development of aggression and altruism in children. In Altruism and Aggression: Biological and Social Origins, Zahn-Waxler, C., Cummings, E., Mark, and Ianotti, R. (Eds). Cambridge: Cambridge University Press, pp. 280–302. Dodge, K. A., Bates, J. E. and Pettit, G. S. (1990). Mechanisms in the cycle of violence Science, 250, 1678–1683. Farrington, D. P. (1989). Early predictors of adolescent aggression and adult violence. Violence and Victims, 4(2), 79–100. Farrington, D. P. (1990). Implications of criminal career research for the prevention of offending. Journal of Adolescence, 13, 93–113. Fiddes, D. O. (1981). Scotland in the seventiesFadolescents in care and custody. A survey of adolescent murder in Scotland. Journal of Adolescence, 4, 47–65. Kraemer, G. W. (1992). A psychobiological theory of attachment. Behavioral and Brain Sciences, 15, 493–541. Kratcoski, P. C. (1982). Child abuse and violence against the family. Child Welfare, 61, 435–444. Labelle, A., Bradford, J. M., Bourquet, D. and Carmichael, M. (1991). Adolescent murderers. Canadian Journal of Psychiatry, 36, 583–587. Lewis, D. O., Lovely, R. and Yeager, C. (1988). Intrinsic and environmental characteristics of juvenile murderers. Journal of the American Academy of Child an Adolescent Psychiatry, 27(5), 582–587. Lewis, D. O., Yeager, M. A., Cobham–Portoreal, C. S., Klein, N., Showalter, B. A. and Anthony, A. (1991). A follow-up of female delinquents: maternal contribution to the perpetuation of deviance. Journal of the American Academy of Child an Adolescent Psychiatry, 30(2), 197–201. Lyons, M. J., True, W. R., Eisen, S. A., Goldberg, J., Meyer, J. M., Faraone, S. V., Eaves, L. J. and Tsuang, M. T. (1995). Differential heritability of adult and juvenile antisocial traits. Archives of General Psychiatry, 52, 906–915. Miller, L. (1988). Neurological perspectives on delinquency. Behavioural Sciences and the Law, 6, 409–428. Myers, W. C. and Kemph, J. P. (1990). DSM-III-R classification of murderous youth: help or hindrance? Journal of Clinical Psychiatry, 51, 239–242. Myers, W. C. and Mutch, P. J. (1992). Language disorders in disruptive behaviour disordered homicidal youth. Journal of Forensic Sciences, 37(3), 919–922. Nelson, H. E. (1982). National Adult Reading Test (NART): Test Manual. Windsor, Berkshire: NFERNELSON. Patterson, G. R. (1982). Coercive Family Process. Castalia: Eugene, Oregon. Patterson, G. R., DeBaryshe, B. D. and Ramsay, E. (1989). A developmental perspective on antisocial behaviour. American Psychologist, 44, 329–335. Rankin, J. H. and Wells, L. E. (1990). The effect of parental attachment and direct controls on delinquency. Journal of Research in Crime and Delinquency, 27(2), 140–165. Rutter, M. (1989). Pathways from childhood to adult life. Journal of Child Psychology and Psychiatry, 30, 23–51. Rutter, M., Silberg, J., O’Connor, T. and Simonoff, E. (1999). Genetics and child psychiatry: II empirical research findings. Journal of Child Psychology and Psychiatry, 40(1), 19–55.

230

A. J. Hill-Smith et al.

Silberg, J. L., Rutter, M. L., Meyer, J., Maes, H., Hewitt, J. K., Simonoff, E., Pickles, A., Loeber, R. and Eaves, L. (1996). Genetic and environmental influences on the covariation between hyperactivity and conduct disturbance in juvenile twins. Journal of Child Psychology and Psychiatry, 37(7), 803–816. Snyder, J., Dishion, T. J. and Patterson, G. R. (1986). Determinants and consequences of associating with deviant peers during preadolescence and adolescence. Journal of Early Adolescence, 6(1), 29–43. Sturge, C. (1982). Reading retardation and antisocial behaviour. Journal of Child Psychiatry and Psychology, 23, 21–31. Toupin, J. and Morissette, L. (1990). Juvenile homicide: a case control study. Medical Law, 9, 986–994. Von Knorring, A., Anderson, O. and Magnusson, D. (1987). Psychiatric care and course of psychiatric disorders from childhood to early adulthood in a representative sample. Journal of Child Psychology, 28, 329–341. Widom, C. S. (1989). Does violence beget violence? A critical examination of the literature. Psychological Bulletin, 106(1), 3–28. Woods, S. M. (1961). Adolescent violence and homicide. Archives of General Psychiatry, 5, 38–44. West, D. J. and Farrington, D. P. (1973). Who Becomes Delinquent? London: Heinemann Educational Books Ltd. Zagar, R., Arbit, J., Sylvies, R., Busch, K. and Hughes, J. (1990). Homicidal adolescents: a replication. Psychological Reports, 67 (3 pt. 2), 1235–1242.