So what is child abuse?

A 12-YEAR-OLD girl from an ethnic minority family writes in a school essay that she sleeps in the same bed as her father. A boy at infant school says he has been watching naughty videos with his Dad. A three- year-old girl at nursery school talks of nightmares about sticky snakes; she rolls Plasticine into a shape which her teacher thinks looks like a penis, and she draws a picture of two people with a protrusion between them which could be either a long finger or an erection.

Are these children being sexually abused? A few years ago, at the height of professional concern about the prevalence of child abuse, culminating in the Cleveland panic in 1987, many social workers would probably have presumed yes. But after the launch of research commissioned by the Department of Health after Cleveland - which calls for a change in social work away from child protection investigations and towards preventative family support - the answer should be maybe.

How to define what is child abuse and what is "normal" family behaviour is one of the core questions at the heart of the Government's attempts to develop a better child care system for the future. Hugging, kissing, bathing, even sleeping with your children - are these natural patterns of parental behaviour or are they inappropriate, over-sexualised acts of abuse?

And what are normal childish pastimes? When children draw pictures of witches and snakes, does this mean they are symbols of frightening, abusive events? Or can the images have a simpler, innocent explanation? These are fundamental questions with which teachers, social workers and other professionals involved in caring for children frequently have to grapple.

Supposed inappropriate behaviour by parents and the misinterpretation of what children say and draw are parallel lines of inquiry which have featured in some of the most notorious cases of false allegations of abuse, such as Rochdale and Orkney which led the Government to commission research into child protection practices.

In an overview of more than 20 research papers , Child Protection: Messages from Research, published last month, the Department of Health makes clear it wants a shift away from firefighting to fire prevention.

The research concludes that too many families are drawn into the child protection system and too few get adequate family support. Of 160,000 investigations annually, only 15 per cent of children are considered sufficiently at risk to be placed on a child protection register. Essentially the department is now saying that child care professionals should avoid seeing child abuse demons everywhere. And if suspicions do arise, the accused should be treated as innocent until proven guilty. In recent years the opposite has applied.

The overview sets some exercises for social workers to consider. The first two questions posed above are hypothetical; the one involving the three-year-old is real. The Foster family (all names have been changed) of Tower Hamlets in London's East End, claim their case is a classic example of the over-anxious attitude of the child care professions obsessed with child abuse.

The Fosters have two daughters, Amy, now aged four, and Geraldine, aged five. John, the father, who works as a maintenance manager, was first suspected of abusing Amy last December after she was reluctant to go to nursery school.

Her mother, Marie, says she had been ill with a cough and congestion and was given medicine to clear the symptoms. She wasn't sleeping properly, was cranky, over-tired and clingy. The Fosters were called to the school and Amy's teacher said she had been talking of nightmares about a sticky snake which hissed.

Notes made later show the family's health visitor and their GP warned that Amy could have been affected by the medication, which sometimes had side-effects, including nightmares and hallucinations.

However, unbeknown to the Fosters, after they visited the school the head teacher called Tower Hamlets social services and asked for a child protection referral. But the preliminary file written by the duty social worker went missing and social services took no action for three months.

Meanwhile the school conducted its own investigations. Amy's form teacher began to collect a file of her drawings which she felt had sexual overtones. She made a Plasticine sausage shape which the teacher thought resembled a penis and she talked of a "thing" squirting white stuff. Then Geraldine's teacher reported that Geraldine told her that her uncle was horrible to her and so was her Dad.

The first time that Marie Foster knew that her husband was under suspicion was when a social worker and female police officer called at the house in March and presented the file collated by the school. No other investigation had been conducted. But the social worker and police officer suggested to Marie that John should leave the house. The couple refused to be split up but agreed John would not be left alone with his daughters until the investigation was complete.

The girls' parents have innocent explanations for the children's behaviour. They say Amy's sticky snake was a snake-shaped jellied sweet given to her by a neighbour; the thing that squirted white stuff was a nasal spray for congestion; Geraldine's uncle used to tease her mercilessly, for instance messing up her carefully laid-out arrangements of paper shapes, and her Dad would sometimes join in.

But after a case conference in March, the children were placed on the Tower Hamlets "at risk register". Their parents are now fighting to clear their names. They are also objecting to what they claim are clear breaches of procedure by social services staff - for instance assigning an unqualified social worker to conduct a videotaped interview with the children - although this has not yet taken place. They complain that from the outset the social worker failed to keep an open mind, refused to listen to any alternative explanations and assumed the children were victims of child abuse.

John used to bath with his daughters when they were younger, but had already stopped before the child abuse suspicions first surfaced. The social worker made it plain she disapproved of such "inappropriate behaviour" but now, he says, he is afraid even to show his daughters affection. At four Amy is still clingy with her mother and likes to get into her parents' bed - on her mother's side they insist. Would this be frowned on too?

The controversial new study, to be published later this year, Normal Sexuality and Sexual Knowledge in Children, concludes that social workers "tend to overestimate the levels of problem behaviour and sometimes misinterpret what is perfectly normal". The Department of Health's overview says these findings "expose the gap between popular images of family life and what is actually occurring".

The researchers, Marjorie Smith and Margaret Grocke, of the Institute of Child Health at the University of London, focused on a random sample of children aged between four and sixteen who had not been abused.

Asked about their children's behaviour, 63 per cent of parents reported their children had "definitely or probably" touched the mother's breasts; 12 per cent had touched father's genitalia; 67 per cent had been seen masturbating; 31 per cent had seen simulated sexual intercourse on television, 77 per cent bathed with their parents, and 70 per cent had come into their parents' bed at least weekly.

Challenging conventional views on abuse, the study concludes: "Although maltreatment can occur in a variety of settings, the evidence of normal behaviour is important in defining what is abnormal or abusive.

"Excessive masturbation, over-sexualised behaviour, an extensive sexual curiosity or sexual knowledge and genital touching are thought to be indicators of sexual abuse. But since these behaviours were found to be common within a community sample, they were not in themselves sufficient to suggest abuse.

"Family nudity was a pointer to other sexual behaviours within the families. It was also clear that a small but significant number of children who had not been abused had been in situations or had access to sources which could have given them explicit sexual knowledge."

Paradoxically, despite the fact that in society in general there is more awareness and openness about sexual matters, since the child abuse panics of Cleveland, Rochdale and Orkney there is anecdotal evidence that many families now fear that walking round the house naked, or bathing with children, could be misinterpreted as abuse.

Parents Against Injustice, a self-help group which advises thousands of families who claim they have been falsely accused, has noticed that many fathers are now reluctant to change nappies of baby girls or bath their daughters. Sue Amplett, its director, is aware of one father who was told by a social worker that bathing with his four-year-old daughter was "inappropriate, over-sexualised behaviour". Taking holiday photographs of their children with no clothes on was sometimes regarded with suspicion.

She said: "The research is saying there is more nudity in families, more openness and awareness of sexual matters, through education and television. What many social workers have failed to recognise is what impact this has, for instance on children's play. Social workers need to be more aware of the realities of normal family life".

So, in the light of the Government's planned new direction for social work, how should social workers react to the hypothetical questions posed above? David Niven, chairman of the British Association of Social Workers, counsels caution. Unable to comment on the Foster case, he instead discussed the case of the 12-year-old sleeping in her father's bed. Because of their statutory responsibilities the social workers would have to act, he said, assuming the case was referred by a concerned teacher. Before Cleveland the social worker might have assumed abuse and "gone in" to confront the family. But now, he said, social workers were better trained to try to establish the facts.

Now, the social worker would ask what was the context apart from the school essay. It's not unusual for a child to jump into its parents' bed. But was the mother also in bed? Or was she alone with her father? Was it a family of eight living in two rooms? Was sharing the parents' bed part of the culture? Or was the child uncomfortable and unhappy with the relationship?

"You wouldn't go rushing in, but you couldn't ignore it. You mustn't accuse anyone of anything until you've got a reason to believe the child is at risk."

Mr Niven, a social worker for 20 years, welcomes publication of the latest research. "What is needed is a national debate about the protection of children and about what we as a society think a child's rights are. When is a child safe or unsafe? What is child abuse?"