‘Children Referred to Closed Units’ by Pat Cawson and Mary Martell

Pat Cawson and Mary Martell (1979) Children referred to closed units DHSS Research Report No. 5 London: Her Majesty’s Stationery Office ISBN 0 11 320684 4

Though published at the end of the 1970s after most of the other research covered in this series, it was commissioned much earlier, during a period of increasing interest in research into child care, and was one of the first government sponsored pieces of research into a child care topic. Governments had long relied on advisory committees but from the 1960s increasingly began to turn to researchers for answers to their dilemmas.

Unfortunately for the government, and the newly created Social Services Departments, when Pat Cawson and Mary Martell completed their research in the mid-1970s, the Department of Health and Social Security was so embarrassed by it that Norman Tutt, then Principal Social Services Officer, approached Spencer Millham to ask him if he would do it all again (presumably in the hope that he would get a more DHSS friendly report); in the event, Millham et al. (1978) came to much the same conclusions and, when Cawson and Martell published their much delayed report, they were able to incorporate into it the findings of later research.

The reaction of the DHSS was to ignore the research and go their own muddled way (Blumenthal, 1985) with consequences which Alan Johnson was to spell out all too clearly nearly thirty years later:

for many of the 60,000 children who are in care at any one time, childhood and adolescence are often characterised by insecurity, ill health and lack of fulfilment. This is terribly sad. And we can hardly be surprised that it results in many children in care underachieving educationally and getting nowhere near fulfilling their potential as adults (Department for Education and Skills, 2006, p. 3).

Key Ideas

– The 1930s and 1940s saw the lowest use of closed accommodation for children and young people since it had been introduced as a legal penalty.

– Closed provision within the Approved School sector was first discussed in 1951 and recommended following the Carlton School disturbances in 1959.

– Secure Units rarely accommodated the children for whom they were planned.

– Most developments were driven by events and decisions outside the control of the relevant ministry.

– The ‘deterioration of children’ theory has a long history and no evidential base.

– Many decisions taken lacked an evidential or ethical base.

– Though Youth Treatment Centres were established to treat young people with mental health rather than behaviour disorders, there were no significant differences between the children referred to Secure Units and Youth Treatment Centres.

– There was no obvious reason why most children had been selected for referral when compared with the rest of the Approved School and local authority

– Many children had had multiple short placements in the period before referral; in other words, their difficult behaviour was more often a response to repeated failures to meet their needs within existing facilities than to any underlying problem.

– Proximity to one of the existing facilities increased the likelihood of referral.

– Though there were ‘official’ reasons for rejection, no reasons were given for acceptance.

– It was not clear how acceptance related to behaviour, though there was a tendency for children who were a nuisance rather than a security risk to be admitted to the Youth Treatment Centre.

– The subsequent delinquency of those accepted was higher than that of those rejected.

– Admission appeared to be harmful for the less criminally sophisticated.

– The ‘early diagnosis’ approach was not supported by the results of placement.

– The capacity for rejection in the child care system seems to be endless.

– Violent behaviour in care is almost always an institutional problem.

– Change will only come with changes in the open child care system.


In the Introduction they note that Kingswood, Redhill and Red Bank Secure Units were opened between 1964 and 1966; they were intended for boys but girls were also admitted to Red Bank. In July 1971 the first Youth Treatment Centre, St Charles, had been opened as a mixed unit and a second had been opened in Birmingham during the course of their research into referrals from 1971 to 1974.

In Chapter 1 Development of closed units and its research implications, they outline the history of closed units for children from the 1908 Children Act which abolished imprisonment for children in general (though it was retained for serious crimes) to the opening of the Youth Treatment Centres in the 1970s.

Borstals, formally established under the 1908 Prevention of Crime Act, were for ‘unruly and depraved young persons’ of 16-21 and, following the merger of Reformatories and Industrial Schools under the 1933 Children and Young Persons Act, the next fifteen years represented the lowest level of closed provision for young offenders since custodial institutions had first been introduced as a legal penalty.

The 1948 Criminal Justice Act brought in Detention Centres for 14-21 year olds which were intended as an alternative to the birch for adolescents without a previous custodial sentence. In 1951, following representations for a closed school or a closed block in an open school from the heads of Approved Schools, the Franklin Report recommended that ‘persistently troublesome’ adolescents in Remand Homes be transferred to Detention Centres.

By 1959 closed provision was being seen as a solution to both absconding and disruptive behaviour and the report into the Carlton School disturbances, which had been provoked by the actions of the Deputy Manager, recommended:

  • one or more schools with closed facilities
  • detention rooms in all senior Approved Schools
  • improved staffing/regime.

A 1960 Home Office Working Party identified

– persistent absconders

– exceptionally unruly and uncooperative boys

– exceptionally disturbed boys requiring psychiatric care

– ‘medical misfits’ such as epileptics and diabetics

as potential candidates for closed conditions, though the Approved Schools Central Advisory Committee said the first two were the problem, as they could deal with the last two. Among the papers submitted at this time was a paper from the Heads Association questioning the ‘rotten apples’ diagnosis and suggesting improvements to conditions of service and to staff training, with perhaps a temporary closed school while conditions were improved. In part they were questioning the proposed approach because of their experience of boys who had been to detention centre. In the end, the final report recommended

  • ‘one or two strong detention rooms’ for emergency use in each senior and intermediate school
  • the establishment at each of the three Classifying Schools of closed blocks for 20 boys over 13 years old.

For the girls’ schools, they recommended improving relationships.

In February 1961 a Home Office Inspectorate Study Group suggested the units should be for 14-16 year old absconders and disruptive pupils and should involve:

  • a ‘brisk regime’
  • no home leave
  • a quick return to open school.

Kingswood School, near Bristol, initially refused to accept a closed unit because it wanted to be allowed to run its own regime – which it was – and units opened at Kingswood in 1964, Redhill, Surrey, in 1965 and Red Bank, Newton-le-Willows, in 1965.

When a Working Party reviewed the units in 1966/7, they found they were rarely admitting boys for disruptive behaviour, mostly for absconding, and those admitted were mostly older boys, not the 14-15 year olds originally expected except at Kingswood. Moreover, instead of boys returning to their old school, the schools were developing their own treatment plans including preparation for release home. The Working Party considered the possibility of accepting non-offenders and “boys who needed intensive staff help who might not need the security” but recognised the impact this might have on places for children needing secure provision and suggested semi-secure units or hostels to enable boys to move on. However, this Working Party report was never written up.

Meanwhile, as the Ministry of Health was moving away from locked wards, a 1967 Working Party on Severely Disturbed Children and Young People in Approved Schools had considered provision for children with psychiatric needs and recommended that the Secure Units return to their original purpose while a 1967 Inspectorate Report on difficult and disturbed girls in Approved Schools had looked at the possibilities of a girls closed unit.

When the 1966/7 Working Party was convened in 1969 it decided that, though the Secure Units were not meeting their original purpose, they were satisfying a continuing need and recommended the establishment of Youth Treatment Centres for very disturbed children. Their deliberations were interrupted by the trials of Peter E and Mary Bell and the need to have some provision for young children convicted under §53 of the 1933 Children and Young Persons Act. (These two cases also led to the dropping of a clause in the 1969 Children and Young Persons Bill which would have led to the repeal of §53).

As a result in 1971 St Charles Youth Treatment Centre comprising a maximum security house, a semi-secure house and an open house was opened. A 1972 DHSS report Development of secure provision in community homes identified the development of an ‘extreme psychodynamic viewpoint’ (Cawson and Martell, 1979, p. 35) in which terms such as ‘central secure units,’ ‘intensive care units,’ ‘separation rooms’ and ‘maximally secure rooms’ (that is, padded rooms) were being used.

Meanwhile, Clarke and Martin (1971), in their study of absconding, had disproved the ‘rotten apples’ theory. Later Millham et al. (1975) and Martin (1977) were to show that “troublesome behaviour in an institutional setting is produced by the institution rather than being an attribute of the individual” (Cawson and Martell, 1979, p. 37) . They note that the theme of the ‘deterioration of children’ can be traced in the literature back to the nineteenth century.

They query the ‘need for containment’ pointing out that there was never any consideration of how containment was expected to reduce aggression and cite Millham et al. (1976), who suggest that problems of aggression become greater in family style units with closer personal relationships. They also note that, while girls actually presented more difficult behaviour in Approved Schools and Borstals than boys, there was no pressure for a closed unit. Moreover, the Youth Treatment Centres were to be mixed because that was seen as better for girls but there was no discussion as to why it was necessary to lock up non-offenders.

In Chapter 2 Referrals to the units, they describe how, prior to 1971, referrals were made through the Classifying Schools and the Home Office; on 1 January 1971 the DHSS and local authorities became responsible. Kingswood and Redhill continued to use the central referral system through the DHSS while Red Bank created their own system; however, there was no systematic register of referrals until 1972. St Charles started with a standard referral procedure though this was not always followed by local authorities.

The research sample was taken from 202 central referrals (that is, not including Red Bank) from June 1972 to July 1974 and 88 Youth Treatment Centre referrals from June 1971 to July 1974 of which nineteen were referred to both types of establishment. The age range was predominantly younger, with two thirds 13-14 year olds, reflecting the original expectations of the Home Office and the population of residential care; the average IQ was below the national average but 10% had IQs of 110 and over while 9%, mostly referred to the Youth Treatment Centre, had been diagnosed with brain damage or epilepsy.

Over half had been in or been referred to a psychiatric treatment unit and nearly two thirds had been treated in a Child Guidance Clinic; they were more likely to be at the Youth Treatment Centre. A quarter of the Youth Treatment Centre and a fifth of the Secure Unit children had self-harmed or threatened self-harm; the girls were more likely to be suicidal and to come from broken homes. Many children had been referred for treatment but not received it.

Though the boys were more likely to come from a criminal family, they were less likely to do so than Approved School boys generally. 39% had no known family problem.

There was therefore no evidence for two ‘types’ of children, one suitable for Secure Units and one for Youth Treatment Centres; indeed, the differences between the boys and girls in the sample were generally greater than the differences between the sample and the Approved School population.

In Chapter 3 Behaviour prior to referral, they note that their evidence came from requests for admission which would be more likely to stress difficult behaviour. Boys in Secure Units were more likely to be offenders and girls less likely whether in Secure Units or the Youth Treatment Centre. The offending of those in Secure Units was comparable with the Approved School population but the offending of boys in the Youth Treatment Centre was lower. There was a similar pattern for absconding. Nearly all referrals to Secure Units mentioned offending or absconding but 31% of Youth Treatment Centre referrals did not mention them at all.

12% of referrals concerned children in voluntary care and one third were non-offenders; indeed the proportion of non-offender girls referred to the Youth Treatment Centre was higher than the proportion in Approved Schools.

Though the boys were more delinquent than the girls, they were less so than children referred to Secure Units previously, who tended to be older, and a substantial minority had no indication of prior delinquency.

Though 63% were referred as persistent absconders, mostly to Secure Units, 35% did not abscond. Nor did children’s actual violent behaviour correlate with the assessments of aggression given; girls were more likely to be reported as actively violent while boys referred to Youth Treatment Centre were more likely to have been labelled ‘aggressive’ but not to have been violent. Only 16% of the children had been involved in violence outside an institution and those referred for violence were mostly from large institutions, not small children’s homes. Boys were more likely to have been violent against other children inside than against staff.

Children referred to secure units were less likely than those in Youth Treatment Centres to have been described as unruly or subversive; only 9% were referred for this reason but nearly half were reported as such once they had been admitted raising questions as to the real reason for referral.

Children’s anti-authority attitudes were not related to coming from a criminal family; only 9% of violent outbursts were described as inexplicable. There was little evidence of psychopathy, and risky sexual behaviour was more likely to be seen as an issue for girls than for boys.

Among the children risky behaviour tended to peak at 13-14 rather than going up with age but it did correlate with previous court appearances. Nuisance behaviour tended to be higher in the Youth Treatment Centre but followed the same trend.

In summary:

– Three quarters of Youth Treatment Centre boys were offenders but less than one third were referred for this reason.

– Half of secure unit boys had been recommended for psychiatric treatment or oversight but only 5% had been referred for this reason.

– One third of Youth Treatment Centre children were referred for good quality care, not for security, especially girls.

– Overall, the children referred were less of a problem than those referred to Approved Schools ten years earlier.

– There was no obvious reason why most children were selected for referral.

In Chapter 4 Institutionalisation and previous placement histories, they report that the children came predominantly from delinquency establishments with a few from children’s homes and psychiatric units. 75% had spent less than a quarter of their lives in institutions; only 3% their whole lives. To put it another way, most children were not seriously institutionalised and a substantial minority were reacting to their first experience of care.

23% had had at least one prior placement of over two years, 39% at least one of six months to two years and the rest placements of less than six months. However, only 3% had been in one institution and 12% in two while 16% had been in seven or more. The boys had been in an average of five and girls an average of four institutions and most of those who had had seven or more placements had had them in the year before admission, not to meet long term needs but to deal with short term situations.

17% had been refused admission to one or more Approved Schools and 18% to one or more other types of placement.

Interestingly, a comparison with the numbers Approved Schools thought they might recommend for closed units in 1960 with the numbers actually sent in 1971/4 showed that many schools who thought they would send a lot did not send any in 1971/4 and vice versa.

One third of the children were being admitted directly from their first placement, not as a last resort. Absconding and risky behaviour tended to be associated with admission but not nuisance behaviour scores. Children assessed as ‘unruly’ had been in more institutions but otherwise were no different and there was no correlation between dangerous offending and having an ‘unruly’ certificate. There was a correlation between drug-offending and absconding.

Most referrals to Secure Units came from London and the Home Counties while Red Bank got half its referrals from the North West suggesting that availability led to referrals. The Inner London boroughs which provided the highest number of referrals had fewer children in care than the North West and West Midlands combined while the Outer London boroughs which had the lowest figures for children in care also provided a high number of referrals.

While higher referring local authorities tended to refer children for offending or absconding, lower referring local authorities were more likely to have done so under judicial pressure but, in the end, there were no differences between the children based on where they came from. The key factors appeared to be: court appearance, age and offences while absconding.

In summary: most children had passed through a limited range of establishments but there had been no long term effort to deal with their problems and most of the admissions related to a small number of previous establishments.

In Chapter 5 Selection and acceptance, they note that the 1960 estimates were much too large; the estimates used in 1967/68 were based on transfers from Approved School to Borstal or mental hospital along with the extrapolation calculated by Mason (1968) which had led to a recommendation for 200 places. The Secure Units took admissions from a list of referrals; so an applicant could wait a long time if they were not a priority. The Youth Treatment Centre decided if an application met the criteria (74% were rejected) and then took them on a ‘first come, first served’ basis. Some referrals lapsed because they could not wait to have their needs met and alternative arrangements were made. Psychiatrists, local authorities, heads of the child’s existing unit and courts might seek to exert pressure to secure an admission.

Though there were ‘official’ reasons for refusals, it was not possible to elicit reasons for acceptance. In practice, if a Secure Unit refused an application because it was a low priority, a local authority could look elsewhere whereas, if the Youth Treatment Centre accepted, the local authority had to wait for a discharge.

However, there was a tendency to exclude under 12 and 15 year olds and the Youth Treatment Centre rarely took 14 year olds; some were refused because, by the time a vacancy occurred, they would be too old.

The Youth Treatment Centre tended to refuse children of low IQ but these were often accepted in Secure Units which were also more likely to accept psychiatric cases even though the Youth Treatment Centre had been set up specially for them.

Some children were refused because too few alternatives had been tried but many were also accepted. The authors questioned the ethics of placing children who had been given three to six months in a Detention Centre in secure accommodation.

It was unclear how behaviour related to acceptance; 22% of those accepted by both the Youth Treatment Centre and Secure Units were non-offenders. There was a tendency to accept children who were a nuisance rather than a security risk at the Youth Treatment Centre; indeed the strongest predictor of acceptance was a psychiatric referral and the second was age, not behaviour prior to admission. However, the Youth Treatment Centre was not relieving Secure Units of the more disturbed children.

In summary: there needs to be greater clarification of the age, IQ and psychiatric history criteria, and there is a obviously a conflict between secure provision being ‘last resort’ and the idea of ‘early diagnosis and treatment.’

In Chapter 6 Admission and outcome, the authors note that, though not part of the study, Red Bank was taking a wider age range of children and sticking to the rule that there had to be at least two previous placements; however, there was no difference in the children’s family backgrounds. But Kingswood was developing more like a Youth Treatment Centre.

Comparing the data on Kingswood and Redhill in 1964-68 and 1972-74, there had been

  • a drop in age range
  • a narrowing of the IQ range
  • an increase in broken families
  • fewer children from criminal families
  • many fewer with psychiatric problems
  • fewer court appearances
  • a possible decrease in absconding
  • fewer prior placements.

The subsequent delinquency of those admitted was much higher than that of children not admitted though re-offenders were more likely to have a history of offending and absconding. Re-offending was associated with admission to a drug unit and the young person’s risk behaviour score whether admitted or not but those with records of violent behaviour were less likely to re-offend if they were not admitted.

Overall, admission appeared to be harmful for the less criminally sophisticated and the results did not support the ‘early diagnosis’ approach to admission.

In Chapter 7 Issues in the future development and use of closed units, the authors conclude that the continuing use of the Secure Units opened in 1964 was now being justified by continuing demand and that the ‘capacity for rejection in the child care system … seem[ed] to be … endless’ (Cawson and Martell, 1979, p. 215) .

On average children had had four placements in the year before they were admitted; selection was inconsistent and there was a poor fit between social work assessments and the children’s perceptions of their behaviour. Most referrals involved no security risk; often the serious offenders were no danger to the community because their offences were domestic and they were the least likely to re-offend. They note that “violent behaviour in care, as distinct from violent offending, was almost entirely an institutional phenomenon” (Cawson and Martell, 1979, p. 220).

They conclude that, in order to understand why children are referred to closed units, you need to know why they are referred to open units. There is no support for a direct causal explanation; the ‘rotten apples’ theory has no support and containment is not needed for aggressive behaviour.

The differences between girls and boys are mainly in lower levels of delinquency and the units have tended to withdraw from coping with older, more delinquent young people.

In the end the difficulties have not gone away; rather the overall policy has failed and change will only come when there is change in the open child care system and in social work practice.


Along with the government, most professionals in the wider field of child care also ignored the results of this research because they misguidedly thought that it only applied to children needing secure care. So it is worth drawing out the general lessons from their work.

Firstly, the ‘deterioration of children’ theory was first documented in the Fertile Crescent over 3,000 years ago:

This youth is rotten from the very bottom of their hearts; the young people are malicious and lazy; they will never be as youth happened to be before; our today’s youth will not be able to maintain our culture (Anon. In Gibson, 1971, p. 549).

There is no evidence for it and it is high time it was abandoned.

Secondly, stereotyping boys as needing more control and girls as needing more care provides neither with what they really need.

Thirdly, secure provision is almost always a response to failures in open provision, not to the needs of the children themselves.

Fourthly, as Wiener and Wiener (1990) were to demonstrate in the research which they began around the same time, moves within care are more significant than length of time in care in adversely affecting care outcomes.

Fifthly, assessments of ‘aggression’ did not correlate with actual violent behaviour, most of which took place in large institutions and against other children rather than in the community or against staff. Similarly, there was no correlation between the issue of ‘unruly’ certificates and dangerous offending.

Sixthly, risky behaviour tends to peak around 13-14 years old.

Seventhly, the use of secure accommodation correlated with its availability and resulted in more adverse outcomes than when secure accommodation had not been used.

Over a decade before this report was published, when Lucy Faithfull was Children’s Officer for Oxford City, she used to argue that the way to reduce the number of people going into prison was to invest in quality child care. Yet as Alan Johnson (Department for Education and Skills, 2006) implicitly admitted for England and Wales and Barclay and Hunter (2008) demonstrate for Scotland, the lesson has yet to be learned.


Barclay, A. and L. Hunter (2008) Blurring the boundaries: the relationship between secure accommodation and ‘alternatives’ in Scotland In A. Kendrick (Ed.) Residential child care: prospects and challenges Research Highlights in Social Work 47, Chapter 12, pp. 166-180 London: Jessica Kingsley

Blumenthal, G. J. (1985) The development of secure units in child care Aldershot: Gower

Cawson, P. and M. Martell (1979) Children referred to closed units DHSS Research Report No 5 London: Her Majesty’s Stationery Office

Clarke, R. V. G. and D. N. Martin (1971) Absconding from approved schools Home Office Research Studies 12 London: Her Majesty’s Stationery Office

Department for Education and Skills (2006) Care matters: transforming the lives of children and young people in care Cm 6932 London: The Stationery Office

Gibson, R. (1971) The satchel and the shining morning face British Medical Journal (2):549-552

Martin, D. N. (1977) Disruptive behaviour and staff attitudes at the St Charles Youth Treatment Centre Journal of Child Psychology and Psychiatry 18(3):221-228

Mason, P. (1968) The nature of the approved school population and its implications for treatment In R. F. Sparks and R. G. Hood (Eds.) The residential treatment of disturbed and delinquent boys: papers presented at the Cropwood round table conference Cambridge: University of Cambridge Institute of Criminology

Millham, S., R. Bullock and P. Cherrett (1975) After grace, teeth: a comparative study of residential experience of boys in approved schools London: Human Context.

Millham, S., R. Bullock and K. Hosie (1976) On violence in community schools In

N. Tutt (Ed.) Violence London: Her Majesty’s Stationery Office

Millham, S., R. Bullock, and K. Hosie (1978) Locking up children: secure provision within the child care system Farnborough: Saxon House

Wiener, A. and E. Wiener (1990) Expanding the options in child placement Lanham MD: University Press of America

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.