Digging Up The Roots of Violence (1999)

Written by: 

George Hosking OBE

01 January 1999

Violent behaviour is learned very young and improving parenting skills may be the way to prevent it.

I USED to be bothered by reports of child abuse.  It seemed wrong to me that children should suffer excessive violence at the hands of parents, who (I presumed) lost their tempers and went a step too far in taking it out on their children.  I assumed the NSPCC was doing good work in addressing the problem. Earning good money as a senior business executive, I occasionally contributed money to them.

Then, about eight years ago, my life was changed.  Two separate cases of child murder by parents or step-parents filled the newspapers, one quickly after the other.  What caught my attention was not the deaths.  It was that for a year or two before their deaths these children suffered systematic torture at the hands of their parents, so extreme that death itself must have been a release.  I was shocked to the core.  I had not previously realised that children could (and often do) suffer so at the hands of their parents.  At that moment I made a decision – more accurately, a decision made me – that I could not live in a world where such things occurred while I did nothing about it.

The question was: what could I do?  I had qualified some years before as a psychologist, and counselled business managers on career development and personal issues, but my prime occupation consisted of advising large multinational companies on the formation of international strategies and acting as a company doctor to make sick companies well again.  I had considerable success at this, enabling subsidiaries of many large corporations to add millions of pounds per annum to their profitability.

Key to success in both fields was that I really understood the warp and weft of business, having spent many years in a large multinational at senior levels.  I knew that to change how businesses perform it was crucial to understand the root causes of their cost and profit structures.  Most managers attempt to improve their businesses by attacking symptoms.  When work is invested in really understanding root causes, a potent change in strategy becomes feasible, beyond anything previously imagined to be possible. 

Could the same approaches be applied to the problem of child abuse?  My first thought was, yes, it could.  I had experience of highly effective international political campaigns, for example in relation to UNICEF's global immunisation campaign, and knew that commitment plus a powerful strategy could produce real changes in society; and strategy was my stock in trade.  But before I could contribute I faced two challenges: I did not understand child abuse the way I understood business and I had no idea what the root causes of such behaviour were – indeed I couldn't begin to imagine what would lead parents to abuse, damage and torture their children as I was beginning to learn does occur.

Learning Curve

I then began a voyage of exploration to understand the root causes of child abuse.  For some weeks I trawled the internet seeking advice, opinions, evidence, references, experts and material relevant to my search.  One of my first discoveries was that there were almost as many theories about what causes child abuse as people to hold them.  Diet, alcohol, drugs, genetics, overcrowding, poverty, envy, television, schools, parents, grandparents, single mothers, stepfathers, atheism, religion, a culture of violence.  I was bombarded with ideas by people, every one convinced they held the sole true understanding of the problem.  I resolved to base my conclusions only on verifiable scientific evidence. 

During the internet exchanges I found half a dozen people who shared my goal and joined in the search.  Our initial research showed a two-way link: a prime consequence of child abuse was that many sufferers became violent adults; a prime cause of violent behaviour in adults was having been abused when young.  At this point we broadened our concern to cover all violence and formed an international organisation which we called WAVE(World-wide Alternatives to ViolencE) with two goals:

  • to build an international network dedicated to understanding the root causes of violence; and 
  • to find and promote effective means of preventing/reducing violence.

Within months we had members from the UK, Sweden, Germany, Switzerland, Israel, Liberia, Australia, New Zealand, Canada and the USA, including therapists, psychologists, child and prison psychiatrists, criminologists, policemen, doctors, lawyers, teachers, mothers, fathers, poets, writers, musicians, builders, computer specialists, businessmen and women, prison visitors and former violent offenders. 

After early months in which debates about US gun control and media violence generated much heat, we set our initial priority: to understand the root causes of violence.  This study has now been taking place for three years, drawing on research from around the world.  Already a trained counsellor, during the course of this study I qualified as a clinical criminologist and for the past two years have been working with violent offenders and paedophiles.  Practical experience from therapists has supplemented our review of academic research.

Our interim conclusions are summarised below.  They are not 'final'.  While evidence for them is strong, we will continue studying and learning for years to come.  New data may change our views.  WAVE will always be open to a change of heart when new evidence is provided.  We are currently reviewing a place in our conclusions for issues relating to trauma and attachment, and exploring some issues relating to diet.

The Many Causes of Violence

There is no single cause of violence.  Convincing arguments can be made for increased levels of violence being due to disrupted upbringing, situational stress, economic inequality, deprivation of justice, extreme poverty, low wages, unemployment, social isolation, overcrowding and poor housing.  We accept that all of these are contributory factors.  We are also persuaded that violence levels are increased by a culture of violence, absence of moral or spiritual teaching, exposure to violence in the media, and ready availability of firearms. Concerning diet and alcohol, we have reviewed conflicting evidence and not yet formed an opinion. Genetic influences can be shown to play a part for some people, but these are really triggered by environmental factors. 

While accepting that all these factors contribute to violence, it is our view that their impact is dwarfed by the impact of factors relating to the family.

Family Violence 

Levels of violence in a family setting are high.  The general public is rightly concerned to protect its children from paedophiles, yet the bulk of sexual abuse is perpetrated by family members.  The 1993 British Crime Survey estimated over half a million domestic assaults occurred in Britain in 1991.  Browne1 estimates that 10 percent of elderly people in Britain are subject to abuse and neglect.  Forty-two percent of murder or manslaughter cases in Britain involve a domestic dispute and one third of domestic victims are children.2  The NSPCC claims that three or four children die per week in the UK at the hands of their parents.3  It has been stated that the greatest risk of suffering violence, emotional abuse, sexual assault and murder for people in western society occurs in the home, at the hands of other family members.4  Violence in the family is more common than love. 

We believe the prime root cause of violent behaviour is early life experience within the family – specifically, receiving harmful parenting.  Harmful parenting can mean neglect, or direct physical violence, but also includes shouting, emotional abuse, and use of violent discipline.  It can be perpetrated by parents who mean well, and believe they are doing what is best for their children, but whose interactions with their children unwittingly cause lasting damage to the quality of lives their offspring are able to lead.

Within this overall pattern child abuse has a central role.  Despite decades of attention to this issue, child protection agencies report that abuse today is just as serious a problem as it was 50 years ago.  One factor often cited by workers in the field as the cause of its persistence is the transmission of patterns of violence and abuse from one generation to another – the cycle of violence.5  Many of you will be familiar with the famous remark of Cambridge University professor of psychological criminology David Farrington, following his study of South London males from the age of 8 to 32: "Anti-social children grow up to become anti-social adults who go on to raise anti-social children".6

Parents Train Violence

We believe that parental behaviour unwittingly 'trains' children to be violent.  Parents model abusive behaviour by using violence to control the child; the child thus learns that violence pays.  Children learn the techniques of being violent from their parents, and also the moral justification for violence. 

Last year I was for the first time in my adult life the recipient of violence.  While driving home from work one evening I made an error of judgement which resulted in another car ploughing into mine.  I believe the accident was my fault, though its severity owed something to the other driver's speed.  Sitting shocked in my seat after the accident, I saw the other driver – a man of about 30 (I am 55) leap out of his car and run to my door.  He hauled the door open, punched me twice, hard, once in the eye and once in the mouth, then grabbed me around the throat and began to throttle me (my throat took a month to recover).  He then hauled me out of the car and threw me to the ground.  He was about to jump on me when, fortunately, some other motorists pulled him away.  A couple of weeks later I spoke to the young man and asked him why he did what he did.  His reply was clear and simple: "You did wrong, so I gave you a slap".  That told me much about the rules he absorbed as a child. 

I am familiar with the phrase "just gave him a slap" from my work with violent offenders, when they describe incidents in which they have busted the jaws and ribs of people who crossed them.  A great deal of severe child abuse is similarly described as a slap or a tap by abusing parents.  Child abuse is usually the end of a continuum that began with "punishment".  NSPCC reports show that frequent punishments by parents include shaking, throwing, freezing baths, pulling hair, biting, scalding and the Chinese burn.  Three quarters of babies are hit before they are one year old, and more than one third of children are hit with an implement. 

Parents justify this abusive punishment as being "for the child's own good" (if you have not done so already, please read Alice Miller7).  Analysis of parents' thought patterns when committing abuse shows that they often have unrealistic expectations, requiring children to show understanding at a level 12 months or more beyond what is appropriate for their age.  They also make wrong interpretations – for example, that the child deliberately seeks to annoy. 

According to the eminent psychologist Dr Jerry Patterson (who carried out pioneering work on the contribution of coercive family processes to child behaviour problems) and his colleagues,8 the cycle of violence persists because violent family members directly train their children in antisocial behaviours.  Such parents use little positive reinforcement, while effective punishment for deviant behaviours is missing or erratic, and "dozens of daily interactions" reinforce coercive behaviour.  Some reinforcement is positive (for instance laughing and paying attention), but the most important is escape-conditioning: the child uses aversive behaviours to terminate unwelcome intrusions.  Examples include toddler temper tantrums, screaming fits and hitting siblings or the mother to avoid having to do something they don't want to do.  These behaviours are reinforced if they produce the result the child desires.  Unfortunately unskilled parents often themselves respond by shouting or hitting.  As the training continues, children and parents gradually escalate the intensity of their coercive behaviours, in an upward spiral of aggressive interactions, often leading to high-amplitude behaviours such as mutual hitting and physical attacks.  In this training, the child eventually learns to control other family members through coercive means.

Harsh Discipline

The use of harsh discipline may be the core practice within families which fosters the development of future violence in the recipient.  Harsh parenting frequently runs in families, as people tend, consciously or unconsciously, to copy the parenting styles of their own parents.  There are many examples showing how families may continue a pattern of chronic and very severe child maltreatment over many generations, despite substantial intervention.5 

Among many studies indicating that harsh discipline leads to violence and criminality in children are those9,10showing poor or explosive discipline practices of grandparents correlated with antisocial behaviour by both parents and grandchildren. 

In 1990 epidemiologist Patricia Cohen and her colleagues,11 published a long-term study of children in New York state, which had run from when they were aged between 5 and 10 till they reached 13-18.  They measured 16 factors relating to family structure and relationships, environmental context and parenting, seeking the best predictors at the ages of 5 to 10 of future adolescent antisocial or delinquent behaviour.  Power-assertive punishment by parents was the most potent.  The second most potent was parents' anti-social personality, which is related to the punishment practices (on the parents) by the grandparents.12  Professor Farrington identified "harsh authoritarian discipline" applied between the ages of 8 and 10 as a significant predictor of adult criminal behaviour. 

The harmful effects of receiving harsh discipline are not limited to future criminal behaviour.  Harsh parenting has been associated with mental illness (and parental mental illness was Cohen's fourth ranking major factor associated with the development of antisocial behaviour in children).  Men who experienced a violent childhood are more likely to assault their wives; children observing such abuse are more likely to assault their own spouse later in life. 

My own counselling experience with business executives has shown me that, even among highly successful people who are not violent, harsh emotional parenting does severe and lasting damage to the quality of life the recipient experiences for decades thereafter.  This can show as fear of authority figures, a near-pathological need to control others (for fear of being controlled), need to dominate staff, lack of trust, inability to empathise with the feelings of others, and being 'cut off' from one's own emotions.  Often the person most aware of the symptoms is the executive's spouse or partner, who tends to suffer the consequences.

Reducing Violence: Cure-Based Approaches

People who are violent can be reformed.  Those who work with violent offenders are aware of the cognitive distortions common amongst such individuals.  Cognitive behavioural therapy is widely used in the UK in the probation and prison services in dealing with violent offenders, and there is evidence that it is effective in reducing re-offending.  The Prison Service is currently carrying out a large scale evaluation of the effectiveness of such approaches, which for reasons of cost tend to be conducted in a group setting. 

WAVE is also exploring approaches which supplement a cognitive behavioural base with attention to the emotional origins of violence.  One practitioner of this approach, who has achieved striking results with extremely violent offenders, including murderers, describes it in this way :

Any adult showing terror, rage or violence is at that moment operating infant-logic and infant-strategies in an adult context, and deep inside would rather not behave this way since basically such strategies do not work, i.e., do not allow the perpetrators to achieve what they want. 

"Trauma in childhood – physical, emotional or sexual – can 'freeze' the individual at that age – commonly three to eight years – and produce the reaction, 'If that is what reality is like, I don't want to know any more...' "13

This therapist first creates an intensely supportive context in which the violent individual can safely re-examine the way his personality was put together, replacing it with a more realistic, up-to-date adult version.  He sees his task as being to persuade the individual that the trauma, and its terror, can safely be confronted and thereby be perceived to be out of date, over, and never to recur.

The approach opens for violent offenders the possibility that they too can become civilised, sociable, responsible citizens, which is what they want, deep down. 

In my first year of working with violent offenders I used purely cognitive behavioural methods.  Psychometric tests showed good results from these.  However, coming across convincing proponents of the trauma-based approach, as described above, I resolved to add this dimension to my work.  I had previously been trained in how to access and integrate both cognitive and emotional aspects of childhood traumatic memories in adult clients, using modified flooding techniques.  This includes activating intense feelings coupled with creating the understanding that the feelings are linked mainly to past events.  Adding this dimension to my therapeutic work has deepened my understanding of the origins of violence in my clients and resulted in improved results both as reported by clients and as measured by psychometric tests.  Issues we have been able to address through these techniques include all-pervading fear of loneliness, rage (for example, at having been abused), terror (of being locked in a dark cupboard), depression, sadness, fear, being bullied, being a victim, grief and guilt.  The memories accessed may be absent from day-to-day consciousness, but are fully accessible – that is, they are not hidden memories. 

Trauma-removing approaches tend to be one-to-one and so are time-intensive.  Current resource allocations within the prison and probation service do not favour such approaches.  A pity.  It would be good to confirm on a larger scale that the most violent offenders in our society could be cured. 

The question of the link between trauma in childhood and subsequent adult violence is one currently being explored by WAVE.  One writer who has pulled together much of the background on this subject is psychotherapist, psychiatrist and biologist Felicity de Zulueta,14 head of the Traumatic Stress Unit at the Maudsley Hospital in South London and a member of WAVE's advisory panel.

Prevention or Cure? 

If there is a limit on resources channelled into attempts to cure violence once it is instilled in people, what about efforts to prevent it developing in the first place?  There is no question as to the current emphasis.  At a recent Royal Society of Medicine Conference on the roots of violence in children, John Bright from the Government's Cabinet Office stated that current government spending puts nearly 300 times as much resources into dealing with the consequences of crime as it does to prevention.15  Given that most of us would agree that this is wrong, where should efforts at prevention be directed?

Parental Competence

If inappropriate parenting is a prime cause of violence, a critical focus of any organisation committed to reducing violence should be to improve parental competence.  A study in Massachusetts, which followed children from ages 10-15 through to 45-53, found that juvenile deviance was related to parental competence, and that delinquency and criminality were both correlated with mothers' competence, fathers' interaction with the family and family control methods.16  Professor Farrington has also identified poor parental child rearing as a prime cause of anti-social behaviour.  Other research has found that unskilled parenting plays a key role in young males subsequent aggression towards their intimate partners.17 

Patterson and his colleagues believe parental discipline may be the principal cause of inter-generational transmission of criminality.  They cite a growing consensus that parents of antisocial children lack parenting skills.18,19,20  But change can be effected.  Patterson's training studies show that changes in parental discipline lead to significant reductions in child antisocial behaviour.

The First Three Years Are Crucial

As damage to children is usually done early in life – many parents hitting, smacking and abusing very young children from the ages of nought to three, and three-quarters of babies in Britain being smacked before they are one year old – priority should be given to improving parenting of the very young. 

Evidence of how root causes occur at an early age comes from the Dunedin Study in New Zealand, in which it was found that nurses could predict future criminal tendencies 18 years in advance.21 

Every child born in Dunedin in 1972 has been followed from birth, and assessed every two to three years on a variety of health, social, behavioural and environmental measures.  At age three, an 'at risk' group of children was identified by nurses on the basis of 90 minutes' observation.  These children were restless and negative, and lacked persistence and attention. 

At age 21, males in the 'at risk' group were compared with other 21 year olds: 47% abused their partners (compared with 9.5% of others); three times as many had anti-social personality; two and a half times as many had two or more criminal convictions; and 55% of 'at risk' offences were violent (compared with 18% of others).  Offences committed by the 'at risk' group were of the much more serious type, such as robbery, rape and homicide.  And in the 'at risk' group as a whole there was much more history of violence, illiteracy, poverty, shouting and lack of love for parents.

Fewer of the females became conduct-disordered but where they did, 30% of the 'at risk' group had teenage births (the others had none) and 43% were in violent, abusive relationships.22  This means that immature mothers with no strong parenting skills and violent partners have already borne the next generation of identifiably 'at risk' children.  Professor Farrington's famous remark is as true in South Island New Zealand as in South London.

Parent Training/Coaching

WAVE proposes that a key step to reduce violence in our society is to improve parenting, especially parenting of the very young.  Professor Farrington's recommendations for reducing violence include parent management training.  Internationally known American family violence expert Professor Richard Gelles in 1993 recommended five steps for preventing family violence, one being to teach parents alternatives to violence as a way of controlling children.23  A 1997 review by Kevin Browne, Professor of Criminological Psychology at the University of Birmingham, and Martin Herbert, Professor of Clinical and Community Psychology at the University of Exeter, concluded that: "...parent training programmes are our main hope in work with families with highly aggressive children..."2 

Methods of parent training differ.  Patterson and his colleagues at the Oregon Social Learning Centre document success with over 200 families with extremely aggressive antisocial children by using a strictly social learning approach.24  Dr Carolyn Webster-Stratton, Child Psychologist and Professor of Nursing at the University of Washington and Professor Martin Herbert 25,26 integrate a cognitive social learning method with humanistic elements that emphasise the importance of a non-judgmental, co-operative, listening relationship between therapist and parents, modelling the basis of a successful relationship between parents and problem child.  They prefer to call their approach 'parent coaching', emphasising the collaborative relationship as opposed to one of expert therapist to inadequate parent. 

Parent training is cost effective.  The RAND Corporation in the US made a study27 of the comparative cost effectiveness, in reducing crime, of five approaches :

  • 'Three strikes' then prison 
  • Home visits for the very young 
  • Parent training (parents of five to seven year olds) 
  • Graduate incentives 
  • Supervision early in teenage delinquency

In my view this study took a narrow view of the benefits of both parent training and home visits.  It put no value on the prevention of child abuse (which it did not treat as a crime) and ignored benefits other than crime reduction (for instance, better educational achievements, employment and emotional development).  Also the parent training it evaluated was carried out after children were beyond the crucial age of three.  Even so, it found parent training was a highly cost effective approach for preventing serious crime, costing only £4,000 per serious crime prevented (compared with £9,000 for both teenage supervision and prison).  The costs of crime are many times higher.

A National Campaign

It is possible to reduce violence in society to a fraction of current levels.  James Gilligan, former Director of Mental Health for the Massachusetts prison system, and Director for the Study of Violence at Harvard Medical School, puts it this way: "It is really quite clear that we can prevent violence, and it is also clear how we can do so, if we want to ... my own work over the past 25 years, in violence prevention programs with the most violent homicidal and suicidal men that our unusually violent American society produces ... has convinced me that it is possible to eliminate most of the violence that now plagues us if we really want to."28 

To this end, WAVE is calling for a national campaign to improve parenting skills.  Training should be made available to all teenagers, before they become parents, to all pregnant women and their partners, and then for all parents (especially those with children under three),not just those whose children are most at risk.  Such a programme would be both an investment in the quality of life of tomorrow's children and a major contribution to creating a world with less violence.  It still leaves all the other causes, such as poverty and poor housing,to be dealt with : these need not be alternatives in a caring,loving society.

A Wider Perspective: Relationship Training

Professor Kevin Browne of Birmingham University, one of the UK's most assiduous investigators of family violence, whose work I have cited, proposes that there are multiple causes of family violence, including external and structural stresses and parental factors.  His multifactor model29 encompasses all of these, but identifies the quality of family relationships as the key determining factor in levels of family violence.  This broadens the emphasis from the parent to child relationships to include other relationships within the family. 

The Chairman of Relate, Ed Straw, claims that successful parenting requires relationship skills, and that the single most important determinant of successful child rearing is lack of conflict between the parents.  Helping marriage survive is not only a desirable end in itself; it also helps to reduce violence.  In the New York study by Cohen that I referred to, the third most potent predictor of adolescent antisocial behaviour was a mother-stepfather home. 

In a recent pamphlet30 Ed Straw called us to action, proposing a massive national campaign, as sustained as the Victorian drive for universal schooling, Roosevelt's New Deal, or the 30 year campaign against drink-driving.  He wrote :

"This paper proposes a comprehensive and sustained programme to improve people's relationship and parenting skills in Britain.  To succeed, this will demand the vision and commitment that launched the National Health Service.  Half a century on from the creation of the NHS there is an overwhelming need for a national relationship and parenting service." (his italics) 

"Educational attainment, employability, health, income and future relationships are all affected significantly by the quality of parenting and our experience of other close relationships.  We pick up the costs of poor parenting and poor relationships throughout the social service system – in the financial and social costs of benefits, insurance, criminal justice and drug and alcohol abuse." 

"Public expenditure on benefit payments arising from broken homes approaches £4 billion annually.  We already know a great deal about how to make a significant improvement in people's ability to forge good long-term relationships and become effective parents.  But government spends less than one thousandth of this sum on programmes to spread and apply this knowledge to prevent problems occurring in the first place." (his italics)

In his excellent, pragmatic pamphlet, which I encourage everyone sympathetic to the points in this article to read, Ed Straw set out a 9-point programme to transform British society.  I hope his brother, Home Secretary Jack Straw, is able to give these proposals his full backing.

Government and Local Initiatives

The government is not lacking in initiatives.  Its National Childcare Strategy requires the formation of Early Years Development and Childcare Partnerships with responsibility for providing support to parents and informal carers within early years development plans for every local authority area. 

In 1998 the government set up Sure Start,31 a new strategy to help children under four and their families, particularly in disadvantaged areas.  Sixty Sure Start 'trailblazer' districts were announced and the plan is to have at least 250 programmes running by the end of this Parliament.  Programmes will include: outreach and home visiting; support to families and parents; support for good quality play, learning and childcare experiences; and health care and advice about child health and development. 

The aim of Sure Start is to invest in the crucial early years to prevent social exclusion in later life.  Initiatives will include support for community based organisations which provide parenting education, and organisations which provide training to health visitors to take on relationship counselling.  Laudable though its initiatives are, WAVE is not sure the government is aware of the opportunity to reduce violence in our society through parent and relationship training.  If it did, initiatives would be on a larger scale and be more vigorously promoted. 

At the Royal Society of Medicine Conference I mentioned earlier, the Cabinet Offices' John Bright said that what was missing in our national design for reducing violence was local initiatives.  A bold and innovative lead in this has already been taken by the people of Totnes in Devon.32  There, responding to the energy of just a few committed individuals, the community (including local schools and churches, the police, doctors, nurses and local government) has got together for an ambitious project to provide parenting and relationship skills training throughout the community.  All teenagers are receiving relationship training.  Parents, and couples entering relationships or contemplating marriage are being offered parenting and/or relationships training.  Weekend marriage enrichment programmes are being offered.  All new parents of children at the 10 local primary schools are invited to go on a parent-school partnership programme.  An annual conference every February assesses progress to date and makes plans for the future.  Totnes is not waiting for others to resolve the problems of broken marriages, failed relationships and aggressive, violent children.  What if each of us reading this article took a similar lead in our own communities?

WAVE's Key Findings:

  • A major cause of violence is early life experience in the family. 
  • A principal method of instilling violence is parental discipline practices. 
  • Priority for corrective action should be to improve parenting skills – especially for parents of the very young.


Browne, K D (1989). 'Family violence: spouse and elder abuse.' In K Howells and C Hollin (eds.), Clinical Approaches to Violence. Wiley, Chichester.

Browne, K D and Herbert M (1995). Preventing Family Violence. Wiley, Chichester.

National Society for the Prevention of Cruelty to Children (1985). 'Child abuse deaths.' Information Briefing No. 5. NSPCC, London.

Gelles, R J and Cornell C P (1990). Intimate Violence in Families. Sage, Newbury Park. 

Buchanan, A (1996). Cycles of Child Maltreatment: Facts, Fallacies and Interventions. Wiley, Chichester.

Farrington, D P (1995). 'The development of offending and anti-social behaviour from childhood: key findings from the Cambridge study in delinquent development.' Journal of Child Psychology and Psychiatry, 36. 

Miller, A (1987). For Your Own Good: The Roots of Violence in Childhood. Virago, London. 

Patterson, G ., DeBaryshe, B D and Ramsey, E (1989). 'A developmental perspective on antisocial behaviour.'American Psychologist, 44, 2, 329-335. 

Huesmann, R R, Eron, L D, Lefkowitz, M N and Walder, L O (1983, April). The stability of aggression over time and generations. Paper presented at the meeting of the Society for Research in Child Development, Detroit, MI. 

Patterson, G R and Dishion, T J (1988). 'Multilevel family process models: traits, interactions, and relationships.' In R Hinde and J Stevenson-Hinde (eds), Relationships Within Families: Mutual Influences (pp 283-310). Clarendon Press, Oxford. 

Cohen, P, Brook, J. S, Cohen, J, Velez, C N and Garcia M (1990). 'Common and uncommon pathways to adolescent psychopathology and problem behaviour', in L Robins and M Rutter (eds) Straight and Devious Pathways from Childhood to Adulthood. Cambridge University Press, Cambridge. 

Patterson, G R (1982). A Social Learning Approach: 3. Coercive Family Process. Castalia, Eugene, OR. 

Johnson, R (1996). Violence – causes and cures (unpublished paper delivered to Alternatives to Violence Programme Annual Conference, 1996), summarised in Quaker AVP Newsletter, No 13, August 1996, Religious Society of Friends (Quakers), London. 

de Zulueta, F.(1993). From Pain to Violence: The Traumatic Roots of Destructiveness. Whurr, London. 

Bright, J (1999). Social Exclusion Unit, The Cabinet Office. Unpublished paper presented to the Royal Society of Medicine Conference, The roots of violence in children and young people, March 1999. 

McCord, J (1990). 'Long-term perspectives on parental absence', in L Robins and M Rutter (eds), Straight and Devious Pathways from Childhood to Adulthood,. Cambridge University Press, Cambridge. 

Capaldi D M and Clark, S (1998). 'Prospective family predictors of aggression toward female partners for young, at risk males.' Developmental Psychology, 34, 1175-1188. 

Kazdin, A E (1985). Treatment of Antisocial Behaviour in Children and Adolescents. Dorsey Press, Homewood, IL. 

Loeber, R and Stouthamer-Loeber, M (1986). 'Family factors as correlates and predictors of juvenile conduct problems and delinquency.' In M Tonry and N Morris (eds) Crime and Justice: An Annual Review of Research, Vol VII, 7, 29-149. University of Chicago Press, Chicago, IL. 

Rutter, M, and Giller, H (1983). Juvenile delinquency: Trends and Perspectives. Penguin Books, New York. 

Caspi, A, Moffitt, T E, Newman, D L and Silva, P A (1996). 'Behavioural observations at age 3 years predict adult psychiatric disorders', Archives of General Psychiatry, 53, 1033-1039, American Medical Association. 

Moffitt, T E and Caspi, A (1998). 'Annotation: implications of violence between intimate partners for child psychologists and psychiatrists.' Journal of Child Psychology and Psychiatry, 39, 2, 137-144, Cambridge University Press, Cambridge. 

Gelles, R J (1993). 'Family violence.' In R L Hampton, T P Gullotta and other (eds) Family Violence: Prevention and Treatment, 1-25. Sage, Newbury Park, CA. 

Patterson, G R (1986). 'Performance models for antisocial boys.' American Psychologist, 41, 432-444. 

Webster-Stratton, C and Herbert, M (1993). 'What really happens in parent training?' Behavior Modification, 17, 407-456. 

Webster-Stratton, C and Herbert, M (1994). Troubled Families – Problem Children: Working with Parents – A Collaborative Process. Wiley, Chichester. 

Greenwood, P W, Model, K E, Rydell, C P and Chiesa, J (1996). Diverting Children from a Life of Crime: Measuring Costs and Benefits, RAND, California. 

Gilligan, J (1996). Violence – Our Deadly Epidemic and its Causes. Grosset/Putnam. New York. 

Browne, K D (1988). 'The nature of child abuse and neglect', in Early Prediction and Prevention of Child Abuse (eds) K.Browne, C. Davies and P. Stratton, Wiley, Chichester. 

Straw, E (1998) Relative Values: Support for Relationships and Parenting. Demos, London. 

Department of Education and Employment (1999). Sure Start: Making a Difference for Children and Families.

Grimshaw, C (1998). Totnes Community Family Policy Project. Totnes Community Family Policy Trust, Totnes