a Non-punitive Paradigm of Probation Practice: Some Sobering Thoughts L.R. SINGER
Lawrence Singer is an Assistant Chief Probation Officer with Merseyside Probation Service.
Beforehand he worked as a probation officer for Berkshire Probation Service and later as a research officer for Somerset Probation Service where he undertook the research reported here.
The paradigm which is described originates from doctoral research into generic probation practice and has been presented in a book Adult Probation and Juvenile Supervision: Beyond the Care-Control Dilemma (1989, Gower).
SUMMARY
The Government White Paper Crime, Justice and Protecting the Public proposes a radical change for probation practice.
Henceforth probation work will no longer be an alternative to punishment but an alternative form of punishment.
This article challenges the conceptual and practical appropriateness of applying the notion of punishment to the probation context.
Probation officers, it is argued, are concerned with establishing rules for probationers and not inflicting pain on them.
An alternative non-punitive paradigm of probation practice is described.
The paradigm is based on the joint supervisory aims of support and surveillance which are implemented through the methods of counselling, assisting, directing, and monitoring.
To illustrate the paradigm, reference is made to an alcohol education course developed for offenders in trouble through drink.
The efficacy of the alcohol project and the non-punitive paradigm is demonstrated by the results reported from a three year evaluation study covering 152 offenders.
The article concludes with an affirmation of the disciplined but non-punitive character of current probation work and argues that, to successfully protect the public and reintegrate offenders into the community, the probation officer's employment of established social work knowledge, skills, and values is a strength not a weakness.
The aim of this article is to challenge the assertion contained in the White Paper Crime, Justice and Protecting the Public (Home Office, 1990a) that probation practice can be conceived of as a form of punishment in the community.
Issue will be taken with the appropriateness of applying the concept of punishment to the probation context through reference to a demanding and successful probation project concerned with offenders in trouble through drink.
In describing the project, a non-punitive model for understanding probation practice will be developed.
This non-punitive paradigm will provide a framework for contrasting the efficacy of current practice with future plans inappropriately to transform probation work into ' a new approach to punishment in the community ' (p. 18, para. 4.3).
The punitive approach anticipated for the Probation Service in the White Paper is underpinned by a definition of punishment that resides' in the restrictions on liberty and in the enforcement of the orders'.
The White Paper goes on to emphasize: ' It is the loss of liberty involved in carrying out the terms of the order rather than the activities carried out during the order which is the punishment ' (p. 18, para. 4.4).
Notwithstanding the popular appeal and political capital to be derived from such a humane-sounding statement, this definition is nevertheless at variance with the traditional understanding of the meaning of punishment.
Discourses on punishment and dictionary definitions of the term clearly specify the central feature of punishment as being the infliction of pain rather than the loss of liberty per se (Hudson, 1987).
The conceptual and practical implications of this discrepancy will be examined in the third and final section of this article.
Beforehand, however, consideration will be given first to the probation project and second to its evaluation.
In these sections punishment will be understood to mean the infliction of pain and the review of the project will seek to demonstrate that the probation practice described is not only more demanding but also more effective than this traditional understanding of punishment.
At a time when the Home Secretary is inviting the Probation Service to ' move centre stage in the Criminal Justice System ' (NACRO, 1990, p. 17) it may seem misguided and unnecessarily pedantic to take issue with the White Paper's notion of punishment.
Punishment is just a word after all: nothing more than convenient political rhetoric meriting only lip-service from probation managers and practitioners poised at the ' window of opportunity '.
However, this fails to acknowledge a fundamental axiom of social work practice; namely that language not only reflects but also shapes social reality.
As Davies (1981) has pointed out, speech is the basic tool of the social worker.
What we say influences what we think and in turn, what we do.
If punishment is assimilated into the probation glossary it will inevitably influence the sort of practice the Service undertakes.
In the sections to follow the stark question is therefore put ' Is the Probation Service in the business of inflicting pain? ' and the answer which unfolds may be summarized ' No, because it neither aspires nor wishes to! '
THE PROJECT: MORE DEMANDING THAN PUNISHMENT
Government concern about the problem of drink-related offending is as old as the Probation Service.
Indeed, as Bochel (1976) has shown, the establishment of a Probation Service in England and Wales was inextricably linked with work undertaken with offenders in trouble through drink.
Almost a century later the new approach to punishment in the community outlined by the White Paper once again presents alcohol misuse as a key problem requiring community-based intervention.
' Alcohol ', the White Paper states, ' is implicated in a wide variety of crimes, drunken driving, offences of public disorder and domestic violence ' (p. 21, para. 4. 14).
In describing its vision of effective supervision the White Paper goes on to note:
offenders take part in alcohol education programmes so that they can learn to drink sensibly.
They learn about the relative strength of different drinks, the amount which is safe to consume, and the pitfalls to avoid (e.g. competitive drinking and buying ' rounds') (p. 37, para. 7.11).
As a summary of alcohol education programmes, this is something of a simplification.
However it does provide a useful introduction to the origins of the Alcohol Education Course reported here.
Furthermore, it highlights the suitability of referring to this type of project for testing out the appropriateness of applying the notion of punishment to probation practice.
In 1985 the management of Somerset Probation Service was faced with the fact that nearly one in three of the offenders supervised in their area were either known to be or suspected of misusing alcohol (Singer, 1985).
Aware of practitioners' frustration in not being able to work effectively with these problematic clients, Probation management, in partnership with the voluntary sector, determined to develop an Alcohol Education Course (AEC).
By 1986, when social inquiry report monitoring information covering 1,257 offenders indicated that the problem of drink-related offending was virtually as great for pre- as for post-sentence clients, that is, alcohol was cited as a pre-offence problem in twenty-nine per cent of cases, a specialist project worker had been appointed.
Towards the end of 1986 an AEC was launched and has been running ever since.
The aim of the course is to educate offenders about alcohol to facilitate a lasting change in their subsequent use of this drug.
This aim is underpinned by three interrelated objectives and it is not without significance how they complement and serve to implement the White Paper's own trinity of ' aims of supervision ', that is, protection of the public, prevention of reoffending and successful reintegration of the offender in the community (p. 35, para. 7.3).
These were:
1.
To increase knowledge about alcohol so that future alcohol use can be more informed and therefore more prudent.
2.
To instil or reinforce responsible drinking attitudes and behaviour.
3.
To encourage the maintenance of, or, where appropriate, a reduction to a personally and socially safe level of alcohol unit intake.
To meet these objectives the AEC has a structured programme involving group and individual contact, that is, six group sessions sandwiched between one-to-one interviews held immediately before and after the group con act.
The initial one-to-one meeting, held with one of the two group leaders, serves to clarify what participants can expect from the course and what is expected of them.
A standard contract which covers conditions of attendance and an agreement to complete various assignments between group sessions is signed by each participant.
The final interview entails a personalized review of the course, as well as the formulation of a follow-up plan to be undertaken with the offender's supervising officer.
The course employs a variety of materials and techniques including the use of checklists, magazine and newspaper clippings, quizzes, videos, brainstorm exercises, discussions, drink diaries, and role plays (Menary, 1986).
The AEC is primarily intended for young adults subject to a probation order with a special condition of attendance (Schedule 11 4A(II) Criminal Justice Act 1982) and reflects the fact that most Somerset offenders in trouble through drink commit serious offences and/or possess a substantial number of previous convictions.
In short, they are seriously at risk of committing further offences and becoming trapped in the revolving door of custody.
The development of the Somerset AEC both accords with and constitutes a practical example of the Home Office's recent guidance to sentencers:
In recent years the probation service has devoted increasing resources to supervising the more serious and persistent offenders.
Probation orders which are combined with constructive, disciplined and effective programmes of supervision are increasingly being used for offenders who stand to receive custodial sentences (1990b, p. 35).
From the description of the AEC already given the demanding and disciplined nature of the course can be appreciated.
However, the course is not punitive; it is not concerned with the infliction of pain.
Instead, it is concerned with the establishment of rules.
This is plainly evident from the section of the contract signed by each participant where what is expected of them is spelt out:
You will be expected to keep a drink diary throughout the course, and make attempts at reducing the amount you drink if it is greater than accepted healthy drinking levels.
You must be willing to be helped and help others  your experiences will be useful to other group members.
You must attend 8 sessions (these will consist of: an introductory session, six group sessions and a final assessment interview).
You must not drink before the session.
You must be on time  lateness will count as an absence.
Unlike the prison inmate, AEC participants can not sit back and passively do their time.
Probation practice is not, however, solely concerned with the establishment of rules.
As has been demonstrated elsewhere with regard to the generic probation setting (Singer, 1989), the supervision of offenders involves two overarching aims: namely support and surveillance.
Support refers to anything which is said or done to help the offender solve or mitigate their personal or social problems: surveillance to anything which is said or done to induce the offender to conform to socially acceptable standards of behaviour.
These aims are implemented through the practice of four distinguishable but related methods.
In terms of support, there is counselling and assisting: in terms of surveillance, there is directing and monitoring.
Each of these four methods are employed during the eight sessions.
Although summarized as an alcohol education course, the sessions are not just concerned with giving information but also with enabling participants to make practical use of it.
In the first group session the counselling method is registered in the course tutor's notes (Menary, 1986) by the explanation: ' This session aims to get participants to consider the subject of alcohol in ways they had not thought of in the past ' (p. 3).
During this session the importance of monitoring is also emphasized:
Drinks exercises introduce the notion of measurement, concentration of alcohol content; information which will be necessary to carry out the recording of alcohol consumption throughout the duration of the course (p. 3).
In the fifth group session entitled ' Focus on the Individual ', the counselling and the monitoring combine with the directing and assisting methods.
At the start of the session the tutor's notes make clear that directing extends beyond enforcing the conditions of attendance:
This session allows for an in depth examination of the drinking diaries over the last five weeks  identifying the risky circumstances or situations....
The aim at the end of the session is to encourage participants to make rules for their future drinking which will lead to a lessening of identified problems (p. 4).
Similarly, the assisting extends beyond simply providing handouts of useful information:
It is important that participants are provided with the skills to reduce their drinking levels or handle more effectively ' pub centred ' problems.
The use of role play, incorporating previously discussed difficulties (i.e. homework assignment session 4) in ' real ' or set up situations provides an opportunity for testing these skills.
Probably the best illustration of the assisting method in action during the course comes in the fifth session when Somerset practitioners take the logic of working with the problem in the community to its ultimate conclusion.
In the second half of the session, course leaders sponsor participants to an evening at the pub for the purpose of practising drinking skills and purchasing non-alcoholic drinks.
The counselling, assisting, directing, and monitoring on the AEC give substance to the aims of support and surveillance and provide the basis for establishing a model of probation practice.
Punishment has no place in this model.
Instead the AEC can best be understood as an example of a non-punitive paradigm of probation practice.
Twenty years ago the Advisory Council on the Penal System observed that notions of punitiveness do not accord with ' the British concept of probation as an alternative to, rather than a suspension of, punishment ' (quoted in Bochel, 1976, p. 238).
Twenty years later the AEC demonstrates that this remains no less true now than it was then.
But does the AEC work?
THE EVALUATION: MORE EFFECTIVE THAN PUNISHMENT
According to Home Office statistics, probation practice is more efficient and no less effective than imprisonment.
At 1988/89 rates (Home Office, 1990b, p. 110), it cost on average 288 per week to keep an offender in custody, compared with 19 per week to supervise an offender on probation; yet the Home Office acknowledge: ' It is hard to show any effect that one type of sentence is more likely than any other to reduce the likelihood of reoffending, which is high for all ' (p. 7).
If anything can be shown, it would appear that probation is more effective than prison in reducing the risk of further offending.
In a review of reconviction rates, for example, a Home Office report concluded: ' Those given probation orders were therefore less likely to be reconvicted within two years of being given probation than those released from custody ' (Home Office, 1986, p. 5).
Methodological objections could be made to comparing the two groups because of the different periods of time monitored, that is, for prisoners the period of time refers to after the sentence has been served whereas for probationers it is after the sentence has been imposed.
Moreover, it could be argued that prisoners are more at risk of further offences than probationers because of, first, their commission of relatively more serious offences, and second, their more extensive and therefore more entrenched criminal careers.
This may be so for the population of probationers generally but does not apply to the AEC sample.
As noted earlier, the AEC was designed for offenders at risk of further offending and/or custody.
Between 1986 and 1988 152 offenders undertook the AEC while simultaneously being monitored to assess the impact of the course upon them.
As a sample of offenders, the course participants have more in common with prisoners than probationers.
In Table 1 below a comparison of the offences committed by the sample and those committed by persons commencing probation or entering custody nationally is given.
Whereas for theft and fraud type offences there were almost twice as many commencing probation nationally as amongst the AEC participants, the reverse was true for violence, burglary and criminal damage type offences.
The similarity between the AEC participants and offenders entering custody  rather than persons commencing probation nationally  is clearly shown by the identical proportions each group recorded for violent offences.
Furthermore, the greater proportion of other indictable offences for the AEC sample, compared with probationers nationally, reflected the large number of drink-driving offenders, that is, 25.
Given this array of serious offences committed by the AEC sample it is not surprising that the average number of previous convictions was seven, with only 10 per cent having no previous convictions compared with 28 per cent who had ten or more.
In this connection it is relevant to note that 84 per cent of the sample had consumed alcohol prior to the commission of their current offences and three out of four of the convictions for previous offences had been drink-related.
The possibility that the sample's previous convictions simply reflected persistent petty offending can be discounted because over a third had already received a custodial sentence (37 per cent); a further fifth community service (20 per cent); and nearly another fifth a probation or supervision order (17 per cent).
Most of the sample were male (94 per cent), aged between seventeen and twenty-four (61 per cent) and the risk of further offending and custody is highlighted by the fact that over half of the sample had committed an offence within a year of the current offence.
In Somerset the management of the Probation Service has reached an agreement with the local and neighbouring courts that Schedule 11 probation orders, that is, orders containing day centre or specified activity requirements, are operated as a direct alternative to custody.
In this light the at-risk status of the sample is further indicated by the fact that 91 per cent of the sample were subject to Schedule 11, with only six per cent being on ordinary probation and three per cent on statutory licence following release from custody, that is, parole and youth custody licence.
The large proportion of schedule 11 cases in the AEC sample contrasts dramatically with the proportion commencing probation nationally in 1988, that is, 63 per cent of the AEC's probationers were subject to a day centre requirement compared with only six per cent nationally; similarly, 31 per cent of the AEC probationers were subject to a specified activity requirement compared with only six per cent nationally (Home Office, 1990c, p. 46).
The more highly convicted and at-risk profile of the AEC sample was also reflected by the fact that whereas nationally as many as 80 per cent of the persons commencing probation in 1988 had no additional requirements made to their probation orders, only six per cent of the probationers undertaking the AEC received this traditional type of probation order from the courts.
The evaluation of the AEC was based upon a comprehensive assessment of the extent to which the course was able to achieve its stated objectives.
This involved the collection and analysis of five different types of information: an alcohol knowledge test, a drink profile questionnaire, a drink diary, a consumer questionnaire and reconviction rate statistics.
The nature of the method, its relation to course objectives and the results obtained are described below.
As will gradually become evident, the AEC has proven itself to be an unambiguous success.
ALCOHOL KNOWLEDGE TEST
The alcohol knowledge test (AKT) is a twenty item multiple choice test of alcohol knowledge.
Each question, like the course overall, is practical in focus and concerned with information about the relative strength of different alcoholic drinks, detoxification times, legal limits and medical effects.
The purpose of the test is to provide an objective basis for assessing the nature and extent of participants' alcohol knowledge before and after doing the course.
Accordingly, the AKT was administered during the first and last of the group sessions and constituted a straightforward method for evaluating the success of the first course objective: specifically, to increase participants' alcohol knowledge so that future alcohol use can be more informed and therefore more prudent.
It will probably not surprise the reader to learn that at the outset of the course serious gaps in participants' knowledge were identified.
For example, over half of the participants (56 per cent) did not know how many beers they could drink before reaching the legal limit for driving and well over two-thirds of them knew neither how long it took for alcohol to be burned up by the body (71 per cent) nor what, if anything, was an effective way of sobering up (71 per cent).
However, by the conclusion of the course over three-quarters of the sample knew the beer equivalent to the legal limit (77 per cent); the correct detoxification time (78 per cent); and the fact that there is no effective way to accelerate the detoxification process (79 per cent)  including a cold shower or a cup of strong coffee!
When the sample's average score out of twenty was calculated, it revealed an increase from 9.5 to 14.3 out of twenty.
This means that there was an overall increase in the participants' alcohol knowledge between the start and the finish of the course of 50.5 per cent: not abstract knowledge, but knowledge that is relevant to drinking with associates in the pub or elsewhere.
To test whether each of the participants' before and after scores were simply a fluke occurrence, as distinct from a genuine increase in alcohol knowledge, a Wilcoxon Matched Pairs Test was undertaken using SPSS and the result proved statistically significant, that is, Z= - 10. 1839, 2-tailed, P=0.0000.
DRINK PROFILE QUESTIONNAIRE
It does not necessarily follow that knowing more about alcohol will automatically result in more responsible drinking attitudes and behaviour, that is, the second course objective.
Participants could, for example, attempt to make their attitudes consonant with their recently acquired knowledge only to find themselves unable to break from their established pattern of behaviour (Festinger, 1957).
To measure the extent to which the AEC was able to successfully anticipate and overcome this potential problem a drink profile questionnaire (DPQ) was devised and administered before and after the course.
The DPQ is a Likert-style schedule comprising five possible answers attracting a score of between one and five points for twenty-five different questions concerned with key aspects of drinking attitudes and behaviour.
When completing the DPQ participants were told that, unlike the AKT, there were no right answers.
Regarding attitudes, participants were invited to rank from strongly agree (five points) to strongly disagree (one point) their level of agreement with statements such as' To enjoy yourself, it is necessary to have something to drink ' and ' If you limit the amount you drink, it spoils your enjoyment '.
Regarding behaviour, participants were asked to indicate from always (five points) to never (one point) the frequency of instances occurring where, for example, they got drunk even when they did not intend to; their drinking left them short of cash; they drank to help themselves deal with a crisis; and they missed meals because they wanted to drink instead.
The participants' answers were scored and their grand totals then checked against where they fell on a continuum of five risk levels ranging from very low to very high.
Whereas in the AKT an upward trend is indicative of success, the reverse is true for the DPQ, that is, the lowering of risk level represents a shift towards having instilled or reinforced responsible drinking attitudes and behaviour.
It was precisely this type of downward shift that was recorded.
Between the first and the second DPQ over three-quarters of the participants either reduced to (72 per cent) or remained at (six per cent) a medium or low level of risk.
To assess whether this before  after difference was merely a fluke occurrence the Wilcoxon Test was undertaken and the result proved statistically significant, that is, Z=-7.7565, 2-tailed, P=0.0000.
Bearing in mind the course leaders' concern to instil or reinforce responsible drinking attitudes and behaviour, it is worth noting that before the course 77 per cent of the participants agreed with the statement ' To enjoy yourself, it is necessary to have something to drink '; yet by the end of the course 88 per cent disagreed with this attitude.
Similarly, whereas over half of the sample (57 per cent) admitted to always or mostly continuing to drink even when they knew they were intoxicated, at the end of the course 60 per cent said this rarely or never occurred.
DRINK DIARY
The third objective of the AEC is to encourage the maintenance of, or, where appropriate, a reduction to a personally and socially safe level of alcohol unit intake.
Determining how much is too much is debatable but authoritative (if somewhat tough) guidelines have been set by Alcohol Concern in consultation with the Health Education Council and the medical professions (Alcohol Concern, 1986, p. 5).
Three levels of alcohol consumption are specified: low, moderate, and high.
Males consuming over 36 units of alcohol a week and females consuming more than 22 units have a high level and are advised to cut down immediately.
AEC participants were required to keep a weekly drink diary after each of the group sessions to monitor their drinking levels.
Although it must be acknowledged that there is a danger of deliberate under-recording, the emphasis placed by the course on the value and importance of being honest with oneself appears to have resolved this problem.
The non-punitive approach of the course is reflected in its policy neither to reward low scores nor to punish underscoring.
This is because the aim of the exercise is to provide the participants with a practical tool for monitoring their drinking: if they want to cheat, they will only be cheating themselves.
In this connection the comments from two of the participants were typical.
The first time I went he got us to do a drink diary and I know I didn't tell the truth.
I put down a bit less than I had been drinking.
I don't know why I did that.
I thought ' What's the other person going to put down, is he going to do the same?
Well I don't want to be seen drinking more than he is! '
As the group went on I was being truthful, I put down what I was drinking.
The two lads that were with me at the end, they were being truthful.
Theirs went down but mine went up.
These drink diaries are excellent, brilliant way as long as nobody cheats.
You're only cheating yourself when you cheat.
They were easy to do.
Information from the drink diaries, based on a comparison of the averages for the first and the last three weeks of the course, revealed a shift from high to moderate drinking levels for males, that is, from 40 to 32 units per week.
The nine females who undertook the course remained within the high zone but nevertheless registered a reduction from an average of 39 units to 28 units per week.
Overall, almost three-quarters of the participants either reduced their intake (61 per cent) or remained at the same level (11 per cent) between the first and the last three weeks of the course.
Using the Wilcoxon test this downward trend was again found to be statistically significant, that is, Z=  4.9595, 2  tailed, P= 0C0000.
CONSUMER FEEDBACK
After completing the AEC, participants were interviewed to provide them with an opportunity to reflect and comment on the course.
Interviews were conducted by the research officer or the alcohol project worker (instead of the course leaders) to ensure confidentiality; and a semi-structured interview guide was used.
The guide covered questions about the participants' expectations, experience and evaluation of the course.
The interview transcripts run to some 121 pages and can only be superficially summarized here.
However, five points of direct relevance to the success of the course and the applicability of the non-punitive paradigm warrant a brief mention.
First, out of all the responses made, only two participants stated that their attendance on the course had not been worthwhile and none said that the course should be discontinued.
Second, the vast majority of participants were able to indicate the practical, as distinct from abstract, usefulness of the course.
With regard to drinking and driving for example:
I've learnt a lot, my job depends on me driving and it's made me realise how much I can drink, how much I can't drink, what I can drink and realising about road accidents.
The amount you drink does affect your driving quite considerably
Third, different aspects of the course appeared to cater for different needs.
For some the counselling was particularly appreciated:
A lot more knowledge, it made you think a lot more about what you were doing really.
For others it was the assistance that was especially useful:
When we all went to the pub and some of them went up and got non-alcoholic drinks and that, and they all said they couldn't go into a pub and do that and they done it, no problem.
I think it was useful going to the pub.
Fourth, many participants frankly admitted that they had very low expectations at the outset: specifically that the course was simply designed to punish them, in return for which they would give the minimum amount of attention possible.
All agreed that this was not
Their experience as the following comments clearly demonstrate:
I thought they were out to get me to be honest with you.
I thought they were out to get me as a punishment.
But I don't think so now.
He was just telling us and you learn.
You've got ta face it haven't you, you've got ta face it whatever.
If you don't face anything, you don't get anywhere.
I was frightened.
I thought it was going to be worse than it was.
It was good once it got going.
I didn't think we would have the opportunity to say things  we all got a chance to speak.
I really thought it was going to be sort of ' You must not drink ever again' and ' You naughty boy, you mustn't do it ever again' sort of thing.
I had the idea that it was going to be ' Do this, do that ' but it wasn't like that.
Finally (and in direct accord with the non-punitive approach of the course), the emphasis upon participants by the end of the course making their own rules  as opposed to them being made by others  was greatly appreciated.
My rules are to cut down drinking, control my temper if I am drinking, not to drink in a such a large group and not to waste much money.
It's a good idea, you should set yourself your own targets so you can know.
RECONVICTION RATE STATISTICS
Positive test results and favourable feedback might be sufficient evaluation criteria outside of the Criminal Justice System but reconviction rates for probation practice must remain the acid test.
Before and after course results are important indices of immediate impact, but did participants manage to keep out of trouble with the law twelve months after completing the course?
Thanks to the assistance of the statistics department of Avon and Somerset Police, it was possible to answer this question and provide a fifth and final indicator of success.
In spite of the high risk profile of the participants, that is, an average of seven previous and often recent convictions (see above), 95 or 63 per cent of the sample were not recorded as having committed further offences during the twelve month follow-up.
In other words, almost two out of three of the participants had kept away from further trouble through drink.
As one participant observed:
Yea, drinking's always got me into trouble in the past, I mean now I've thought about it, I've thought what a stupid idiot I'd be, so you know if it wasn't for the course, then I probably would have gone and done it again, you know, without a second thought.
Without the benefit of the results from a random allocation study based on experimental and control groups, second thoughts about how much of the two out of three success rate can be directly attributed to the AEC are inevitable.
However, Phillpotts and Lancucki's (1979) authoritative study for the Home Office on reconviction rates offers a basis for provisional optimism.
In their six year follow-up study of 5,000 offenders convicted of standard list offences in January 1971, they reported that:
Offenders who were reconvicted within the follow-up period were likely to have their first reconviction fairly soon after sentence (or, for those given custodial sentences, after discharge).
Offenders with a large number of previous convictions were not only more likely to be reconvicted but were reconvicted sooner than those with few previous convictions (1979, p. 41).
What is of particular relevance to the AEC sample is Phillpott and Lancucki's finding that 53 per cent of all male offenders with five or more preconvictions were reconvicted within a year (1979, p. 24).
If the AEC had made no difference to the male participants with five or more preconvictions we would expect over half of them to have been reconvicted.
However, as might be anticipated from the progress registered earlier in connection with improvements regarding alcohol knowledge, attitudes and behaviour, instead of 53 per cent only 42 per cent of the ' five or more ' male offenders were reconvicted.
DISCIPLINE: MORE APPROPRIATE THAN PUNISHMENT
The idea espoused in the White Paper that restrictions on liberty will become ' the connecting thread in a range of community penalties as well as custody ' (p. 19, para. 4.5) is conceptually flawed for two basic reasons.
First, if restriction on liberty is the defining feature of punishment, what are we to make of the experience of children who have to go to school and adults who have to work for a living?
Is their liberty not being restricted in terms of both intention and effect?
For example, employers and teachers keep registers of attendance and prescribe as well as supervise the activities of their employees/pupils, that is, restriction by intent.
Moreover, children attending school and adults turning up for work inevitably have their liberty restricted: simply because the engagement in one activity necessarily restricts their opportunity to participate in other activities, that is, restriction by effect.
It is patently absurd to regard attendance at school or work per se as a punishment.
Second, the notion of a continuum of restriction ranging from community penalties to custody  neat as it sounds  fails to acknowledge the unique and painful experience of imprisonment.
As Sykes (1971) has shown, confinement to and within a prison entails a number of pains.
These involve physical, financial, and psychological hardships including the loss of liberty, goods and services, heterosexual relationships, autonomy, security and life chances post-custody.
Lest anyone doubt the validity of these observations in England and Wales, consider the following admission about a custodial sentence from the Home Office:
It is likely to affect an offender's personal or financial prospects.
There is the possibility that serious hardship will be inflicted on the offender's family, particularly those responsible for the care of young children....
On release from custody, it can be very difficult for the offender to be reintegrated into the community.
Furthermore, tor the young or inexperienced offender, mixing with more sophisticated inmates may deepen his anti-social attitudes and increase his criminal skills (1990b, pp. 8C9).
Probation practice can not legitimately be placed on this sort of continuum.
The description and evaluation of the AEC clearly demonstrates that, in their particular sphere of constructive influence, probation officers, like employers and teachers, are concerned in part with the establishment of rules and not the infliction of pain.
Put another way, it is discipline, not punishment, which constitutes the connecting thread running through the diverse range of probation practice.
This point was made by one commentator during the first review of the ' Probation System ' at the beginning of this century: his comments remain pertinent for the Home Office in their current review at the end of the century.
Probation... is a test or trial of the character of a convicted offender under suspension of judgment, in order that the court may determine if the probationer be fit to retain his or her place as a helpful member of society, or being unfit, must be deprived of his or her liberty as a menace to society.
It is established for the purpose of correction and oversight.
It is a conditional freedom.
It recognises that the newer and better way to govern the law-breaker is not through the expiring system of the cell and the lash, but rather through the mind and the heart (Bridgwater, 1909, p. 9).
The punitive calculus, with its just deserts gloss, may have political appeal and offer an antidote to the soft image that has beset the Service since its inception.
However, this is to mistake not only the demanding nature of current practice but also the non-punitive knowledge, skills, and values which provide for its success.
The AEC is but one example of a number of fruitful initiatives undertaken by practitioners in and through the medium of counselling, assisting, directing and monitoring.
The non-punitive approach of probation officers is a strength rather than a weakness.
Some years ago the ' Nothing Works' school of thought dominated the Probation Service.
Now it seems that some things work, especially those like the AEC that are associated with what has come to be described as the problem-focused/task-centred approach (Roberts, 1990).
These new ways of working are simply a consolidation of the established pattern among officers of personalizing their practice, that is, carefully matching their particular social work knowledge and skills to the unique needs of the individual offender (Singer, 1989).
This trend reflects a developing and successful specificity of probation practice: the pity is that this specificity of practice has not been complemented by a more appropriate specificity of language from the authors of the White Paper.
The Role of Residential and Nursing Homes in the Last Year of People's Lives ANN CARTWRIGHT
Ann Cartwright, B.Sc., Ph.D., is Director of the Institute for Social Studies in Medical Care
SUMMARY
Almost a quarter of a random sample of adults dying in 1987 had spent at least part of the last year of their lives in a residential or nursing home; one in eight spent all of that year in one.
The old, the unmarried.
those with difficulty looking after themselves, the confused, the incontinent, the blind, the deaf, and the bad tempered were more likely to be in such homes than others.
With their greater dependency.
people living in residential homes received rather more consultations and home visits  although not night calls  from their general practitioners.
But they did not get more help from community nurses and they were less likely to be admitted to, or to die in, a hospital or hospice.
According to the assessments of relatives, friends, and neighbours who answered the questions, the quality of life of people who had been in a residential home for a year or more before their death was similar to that of others who died but had never been in such a home.
It was those who were admitted in the year before they died who were perceived as having a relatively poor quality of life in that year.
Increasingly, the quality of life during the year before death is going to depend on the attributes of residential homes and their staff.
The marked rise in the number of residential homes, particularly the recent growth in the private sector (DHSS, 1987; Bebbington and Tong 1986; Laing, 1985) has drawn attention to the somewhat uncertain role of these homes (Judge and Sinclair, 1986).
Two areas of concern are first, who should have residential care and second, the relationship between residential and community care.
A recent study of people's lives in the year before their death makes it possible to examine the part residential and nursing homes play in caring for people during the twelve months before their death.
This paper then looks at the characteristics of the people who spent some or all of their time in such homes and at the care they got from general practitioners, community nurses and hospitals.
Care and conditions in the homes and the extent of visiting are described and a final section considers the quality of people's lives as perceived by their relatives and others.
METHODS
The data come from a study of a random sample of adult (aged 15 and over) deaths in ten areas of England in October and November 1987.
For these deaths we tried to identify and then interview the people who could tell us most about the last twelve months of the lives of the people who died.
Someone was interviewed about 639 of the deaths, 80 per cent of the original sample of 800.
Husbands or wives were seen about 36 per cent of the 639, other relatives about 43 per cent, friends or neighbours about 10 per cent, staff of institutions about 10 per cent and other officials about one per cent.
Staff of institutions responded about 18 per cent of those who had been in residential homes for less than a year, 47 per cent of those in one for longer.
In general, the answers from different types of respondent did not differ all that widely and the variations that emerged seemed, for the most part, understandable in terms of the characteristics and circumstances of the people they were telling us about, although there did seem to be a tendency for the staff of homes to put a more rosy interpretation on situations.
For an analysis of these differences, and for further details about the study, see Cartwright and Seale (1990).
TYPES OF HOMES AND NUMBERS RECEIVING CARE
The ' homes' covered in this study include residential homes for older people, nursing homes and other institutions excluding hospitals or hospices.
For convenience they are referred to here as residential homes.
Fourteen per cent of the sample of deaths occurred in such homes; this compares with a much lower proportion, 5 per cent in 1969 (Cartwright et al.,
1973).
A further 9 per cent of the people who died spent at least part of the last twelve months of their lives in such homes; so almost a quarter, 23 per cent, were in residential homes at some stage in the last year of their lives.
The length of time they spent there and the types of homes involved are shown in Table 1.
Those classed as' other homes' included convalescent homes, a rehabilitation centre, an assessment centre, a holiday home for the disabled and rest homes which may in practice have been homes for older people (the classification was made on the basis of the information given by the person interviewed).
A relatively high proportion of those in other types of home were there for less than a month.
Sixty-two per cent of those in old people's homes had been there for a year or more compared with 47 per cent of those in nursing homes, but this difference might have occurred by chance.
Of the old people's homes, 56 per cent were said to be run by the Council, 6 per cent by a charity and 39 per cent were private.
In contrast the great majority of nursing homes, 83 per cent, were privately run.
Nearly a quarter of those in nursing homes had not had to pay towards the cost of living there compared with only 7 per cent of those in old people's homes.
The reason for the difference seemed to be that the DHSS paid for nursing home treatment in a number of instances: ' They took the pension and attendance allowance and the DHSS paid the rest. '
Nevertheless the proportion for whom payment was thought to have been something of a problem was 7 per cent of all those in old people's homes (15 per cent for private ones, 6 per cent for others) and 17 per cent of those in nursing homes  further differences which did not reach statistical significance.
WHO SPENDS ALL OR ANY PART OF THE LAST YEAR OF THEIR LIFE IN A RESIDENTIAL HOME?
AGE, SEX AND FAMILY?
People who lived longer were more likely to have spent some or all of the last year of their lives in residential homes.
People have been counted as spending all of the last year of their lives in residential homes if they had lived in one for a year or more.
(In fact, as will be seen later, a number of them were admitted to hospital during the last twelve months of their lives and some died in hospital.)
Few under 75 had spent any time in such homes but more than half those dying when they were aged 85 or more had spent at least some of the last year of their lives in one.
The figures are in Table 2 which also shows expected age related differences with sex and marital status.
Further analyses by age and marital status showed that married people less often spent any time in residential homes than single or previously married people of the same age.
The proportions among those aged 75  84 were 8 per cent of the married, 34 per cent of the others, and for those aged 85 or more 25 per cent compared with 61 per cent.
Neill et al.
(1988) found that few married people applied for residential care and those who did so were less likely to be admitted than applicants who were single or widowed.
There were no significant differences in the present study between the never and the previously married although it might be expected that people with children might be less likely to go into a home than those without any and the widowed, divorced, or separated are much more likely to have children than the single: 75 per cent compared with 6 per cent.
An analysis by marital status and number of children showed that for the married the main difference was between those with no children, 17 per cent of whom had been in a residential home, and those with one or more, for whom the proportion was 5 per cent.
For the widowed or divorced there was no difference between those with and without children nor any significant trend with the number of children.
So it would seem that children may assist one parent to care for the other who might otherwise go into a home, but apart from this they have to residential homes for some while caring for and thus avoiding the admission of others, and that these two influences cancel each other out in the statistics.
An analysis by the sex of children showed no significant differences between those with a son or sons only and those with only a daughter or daughters.
An earlier study (Cartwright et al.,
1973) found that those with a daughter or daughters were more likely to die in their own homes than those with just a son or sons.
Data in Table 2 suggest that those with living brothers or sisters may be less likely to be admitted to a home.
Part of this is an age effect: the proportion with such relatives falling from 86 per cent of those aged 45  54 to 41 per cent of those aged 85 or more.
Among those aged 85 or more a similar proportion of those with and without living siblings had been admitted to a home at some stage but among those aged 75  84 more of those without any such relatives had been admitted: 30 per cent compared with 17 per cent.
These variations reflect the fact that siblings themselves age and die and therefore become less of a source of support as people grow older.
CAUSE OF DEATH
People who died of cancer were less likely than those dying from other causes to have been in such homes, 7 per cent against 29 per cent.
The difference remains when the age at which they die is taken into account.
(Cause of death was taken from the death registration form as coded by OPCS.)
When the type of home was taken into account, those dying of respiratory disease were more likely to have been in nursing homes  14 per cent compared with 6 per cent for those dying of other conditions  whereas those dying of stroke were more likely to have been in old people's homes: 27 per cent against 12 per cent.
SELF CARE
The proportions reported to have had difficulty with various aspects of caring for themselves (getting in and out of a bath or shower, dressing and undressing, going to the toilet, washing and shaving, feeding themselves, making a hot drink, or needing help at night) for a year or more before death was 87 per cent of those who had been in a residential home for a year or more, 60 per cent of those in for a shorter time and 25 per cent of those who had not been in such a home at all.
(People who had been in a hospital for all the year before they died have been excluded from these comparisons.)
But many of the symptoms asked about were reported for similar proportions of those who had been in residential homes and those who had not; these were pain, trouble with breathing, vomiting or feeling sick, drowsiness, sleeplessness, dry mouth or thirst, depression, loss of appetite, difficulty swallowing, dizziness, bed sores, an unpleasant smell, and backache.
However, for a number of these symptoms, more people in residential homes had had them for a year or more: the proportions were 31 per cent against 20 per cent for drowsiness, 22 per cent against 14 per cent for dizziness, 19 per cent against 11 per cent for loss of appetite and 4 per cent against one per cent for bedsores.
Only one symptom, a persistent cough, was reported more often for those who had not been in a residential home  24 per cent compared with 10 per cent of those who had spent some time in such a home.
When deaths from cancer were excluded the difference remained: 22 per cent against 9 per cent.
SYMPTOMS
The symptoms reported more frequently for those in residential homes are shown in Table 3.
Several of them  constipation, confusion, urinary incontinence and difficulty seeing and hearing  have been shown elsewhere to increase with age (Cartwright, 1990a).
Some of them, such as mental confusion, incontinence and difficulty seeing and hearing, may have contributed to them being admitted to such homes.
They can certainly make it more difficult for people to look after themselves.
And among those people with confusion, constipation and incontinence, those in residential homes were more likely to have had the symptoms for a year or more.
The figures were 81 per cent compared with 39 per cent for confusion, 78 per cent and 52 per cent for constipation, and 54 per cent against 36 per cent for incontinence.
Bad temper might result from being in such a home.
Another explanation for the relatively high proportion of those in residential homes reported to be bad tempered could be that relatives, friends or neighbours were less likely to report or perceive the person who died as being bad tempered than staff of institutions.
Among those in residential homes the frequency with which bad temper was reported was 31 per cent when the respondent was a relative, friend or neighbour, 45 per cent when he or she was a staff member  a difference which did not reach statistical significance.
But if the comparison between those who had been in a home and others is confined to those for whom relatives, neighbours or friends were interviewed, there is still a difference: 31 per cent against 19 per cent.
Bad temper could be either a cause or effect of being admitted to a residential home.
Other studies have found that caring relatives are more likely to favour residential care as a solution for the elderly confused if the old person exhibits difficult behaviour (Levin et al.,
1989).
So the evidence suggests that because of their age and frailty those living in residential homes will have needed more care during the last year of their lives than other people who died.
CARE FROM GENERAL PRACTITIONERS
People in residential homes remain under the care of general practitioners.
The numbers of consultations, home visits and night calls reported for those in such homes for a year or more before they died are compared with those for others under the care of a general practitioner in Table 4.
As expected, those in residential homes had more consultations and more visits from their general practitioners than others.
But, if anything, they had rather fewer night calls.
Of course night calls are usually made in a crisis and the help normally available in residential homes may make it easier to cope with a minor one without calling out a doctor.
In the minority of homes (a quarter) where all the residents had the same doctor, visiting rates were higher (16.8 a year on average) than in the majority (three-quarters) where the residents could have different doctors (9.5).
However, the proportions for whom the general practitioner was described as reluctant to visit, or for whom it was felt it would have been helpful if the doctor had visited the person who died more often, did not vary significantly between the groups.
So it would seem that people in homes where they were free to have a doctor of their own choice were not seen as being more deprived of visits in spite of having fewer of them than people in homes where everyone had the same doctor.
Those in homes where everyone had the same doctor might receive less individual attention at visits if general practitioners saw several patients on the same visit.
The mixture of praise and criticism for general practitioners' care of people in residential homes was similar to comments about their care of others during the last year of their lives.
Although most respondents praised, or were satisfied with, this care the study identified a number of inadequacies.
Home visits were much appreciated when they happened but failure to visit was one of the more common and significant criticisms (Cartwright, 1990b).
The matron of a residential home, talking about a doctor who she felt was reluctant to visit said it would have been helpful if the person who died had been visited more often: ' It would have given her a lot more comfort and the staff a lot more confidence. '
When asked why she thought the doctor had not come more often, she replied, ' The doctors feel the residents are in a rest home to save them working. '
The doctor had not been asked to come more often ' because we're resigned to the fact that, unless it's an emergency, he won't come out. '
The matron also told us that this patient had to go into a nursing home ' towards the end ' because ' she needed morphine injections which the doctor wouldn't give otherwise she wouldn't have had to move to somewhere strange.
She had to leave her room, her furniture: that's why she gave up. '
CARE IN HOSPITAL
People who had lived in old people's homes were less likely than others who died to have gone into a hospital or hospice during the last year of their lives or to die in one.
This can be seen from the figures in Table 5 which also suggest that the longer the time they had lived in an old people's home the less likely they were to go into hospital.
But those least likely to have spent time in a hospital or hospice or to die in one were those who had lived in a nursing home for a year or more.
This lower use of hospital inpatient facilities was not accounted for by the older age of the people in residential homes although fewer of all those aged 85 or more had spent time in hospital  64 per cent against 80 per cent of those under 85.
Among this older group 46 per cent of those who had been in a residential home for a year or more were admitted, 75 per cent of the others.
There were similar differences when people dying of respiratory disease and of cerebrovascular accident were considered separately: the proportions admitted to hospital were 25 per cent and 47 per cent of those in a residential home for a year or more, 76 per cent and 89 per cent of others among the two causes of death groups respectively.
The numbers dying of other causes were too small to analyse separately.
The only source of information from this study of any possible reluctance to admit people from residential homes to hospital relates to those who were admitted and respondents' views on whether they should have been admitted earlier.
This showed that in practice relatively few, one in ten, of those who had spent any time in residential homes were thought to have been admitted ' too late '; this proportion was a quarter for those who had not spent any time in a residential home.
Residential homes, and long stay nursing homes in particular, appear to cut down the use of hospital care during the last year of people's lives.
What of the nursing care given by district and other nurses to those in old people's homes?
HELP FROM VISITING NURSES
Just over a third, 35 per cent, of those who had lived in an old people's home for a year or more had had help from a nurse who visited the home in the year before they died.
This was similar to the proportion receiving help from district and other nurses among those who had not spent any time in such homes once sudden deaths and those who had spent all their last year in hospital have been excluded.
They also received similar numbers of visits.
Given their age, symptoms and frailty, it might have been thought that the residents of old people's homes would receive more care from visiting nurses than people in the community, as they received more visits from general practitioners.
But some of the less skilled tasks undertaken by district and other nurses for people living in their own homes were probably done by staff in the residential homes.
The next section looks at respondents' views of the care given to people in residential and nursing homes and the conditions there.
CARE AND CONDITIONS IN THE HOMES
Almost nine tenths, 88 per cent, of the people who had spent some time in a residential home were said to have had access to a telephone where they could make calls and people could telephone them.
Among those who did not have this facility, two-fifths were thought to have wanted it.
Ninety-two per cent of those in old people's homes compared with 82 per cent of those in other sorts of homes had this amenity  a difference which fell short of statistical significance  but more of those in old people's homes had a room of their own, two-thirds against half.
Overall the proportion with a room of their own was three-fifths and again two-fifths of those without one were thought to have wanted one.
If the respondent was a relative, friend, or neighbour, they were asked to say whether they thought various aspects of the home were good or not so good.
The replies of those who felt able to make these assessments are shown in Table 6.
Clearly their views may have been different from those of the people who died, but respondents are likely to have taken into account the reactions of their relatives and friends, and how these respondents felt about the care in the homes is also of interest and concern.
The most frequent criticism was of the food, and over this there was a difference between old people's homes for which 26 per cent thought the food ' not so good ' and nursing homes for which it was 7 per cent.
There were no differences between Council homes and private ones on any of these assessments, nor in the amenities available.
And the overall assessments of the homes were similar in the two groups and between nursing homes and old people's homes.
Altogether 50 per cent were described as' excellent ', 31 per cent as' good ', 9 per cent as' fair ', 6 per cent as' poor ' and one per cent as' dreadful ', with 3 per cent making other comments.
A number of criticisms were exacerbated by feelings that people were in the wrong type of institution.
A son whose father had Parkinson's disease and had been in an old people's home commented:
They were not equipped to deal with a progressive condition, they were not that kind of establishment.
The degree of mental stimulation was absent.
It was a very distressing time for him.
The whole point of his being there was because of his condition.
He could have been in a hotel, had he not needed care and supervision which he blatantly did not get.
The daughter of a woman with Huntingdon's chorea who had been in a nursing home described the care she got there as
Limited  not because they didn't want to care, they just didn't know how to care for this type of patient....
It should not be registered as a nursing home.
The one ' dreadful ' home was thought to have contributed to the death of the man admitted there:
It was an awful place: dreary, badly furnished and he wasn't as clean as he'd been in hospital  food all down his clothes and he hadn't shaved, he really looked unkempt.
I don't think the staff cared.
I 'm sure the fact of moving to the home contributed to his death.
He was so well when he left hospital... and the deterioration within a week was unbelievable.
I just feel he gave up when he got in that home.
(Died of Parkinson's disease and pneumonia aged 73 within 24 hours of being readmitted to hospital after being in an old people's home for less than a month.)
A number of respondents commented on the lack of stimulation: ' They sat in the front room and didn't move.
Just sat there, no t.v. and no-one seemed to talk to them. '
When asked about the other residents in the homes, respondents' views were roughly equally divided about whether the person who died had on balance found them helpful (25 per cent) or tiresome (22 per cent).
Most of them, however, 53 per cent, made other comments, generally to the effect that the person had had little to do with other residents: ' Her mind had gone so she was not able to socialise.,
Many of the relatives and friends who rated the homes as an excellent place for the person who died to live towards the end of his or her life made rather perfunctory comments: ' I can't fault the home '; ' It was very nice, no qualms about it. '
Others were rather more forthcoming:
I would say that, instead of ' spoon-feeding ' the residents, they tried to make them independent.
The staff were always there to help but didn't impose themselves.
This was very good.
And they kept her there when they could have sent her to hospital.
(Friend of woman who had had a number of strokes.)
A wife whose husband had had multiple sclerosis and went to a holiday home for the disabled for two weeks to give her a rest said:
It's a good idea, gets them away from home to meet new people.
They deserve a holiday like everybody else.
He had a good time, said he really enjoyed it and would like to go again, and I had all the washing done for me  he didn't bring anything home dirty except the previous day's clothes.
Sometimes relatives and patients had obviously had different views of the home.
One daughter-in-law who rated the home as excellent went on to explain that her mother-in-law:
hated everything about everything.
We thought they were angels in the home and that everything was marvellous.
The staff told us she was the most difficult lady  as far as she was concerned nothing was right.
She had to share (the bathroom and lavatory) much to her annoyance, but they were just next to her room.
You get what you pay for.
The views of people who had previously looked after those who died sometimes reflected their gratitude and admiration for the people who took on this task.
A friend of 30 years standing who had previously cared for a woman with senile dementia described how:
I thought she was just being awkward until she slowly went downhill and became incontinent.
I got help from the social services for this: they brought sheets and pillowcases from the hospital.
The crux came when she became doubly incontinent and her mind had got to the point where she had forgotten how her body functioned.
It was winter and I was changing her clothes three times a day and my self-control snapped and I shook her very hard to stop her sitting down without her pants and was so upset by my behaviour and frightened by what I had done that they agreed to put her in a home.
The young nurses were very kind and would kiss and cuddle her.
VISITING
How much contact do people in residential homes have with relatives, friends, and neighbours outside?
This question is considered for people who had been in such homes for a year or more.
The number of different visitors they were reported to have had during the last twelve months averaged 5.0.
Seven per cent had none, 52 per cent had one to four, 33 per cent five to nine, and 8 per cent ten or more.
The number of visits was much greater: 11 per cent had 300 or more  almost one a day, 33 per cent between 100 and 300  two or more a week on average, 19 per cent 50  99, between one and two a week, 4 per cent 20C49  between one a fortnight and one a week, 14 per cent 10C19  less than one a fortnight, 12 per cent had less than 10 or less than one a month.
The number of visits was strongly related to the number of visitors: of those with less than five visitors, a quarter had a hundred or more visits, but this proportion was three-quarters for those with five or more visitors.
Of those who were visited, 44 per cent had only relatives to see them, 11 per cent were visited just by friends or neighbours and 45 per cent by both relatives and friends; the most frequent visitor was a relative for 78 per cent and a friend or neighbour for 22 per cent.
Three of the five who had no visitors were said by the people we interviewed to have had no relatives.
Neither of the other two who had no visitors had any children; one had elderly relatives abroad, the other, in the opinion of the deputy officer in charge of the home, only had herself to thank:
Because of her awkward manipulative ways she just didn't get any visitors.
She did get visits from other visitors but they wore off because of her awkwardness.
 She went out of her way to be unpleasant to people.
Isaacs and his colleagues (1972) found that relatives' failure to provide basic care for elderly people was often associated with their previous rejection by the elderly person.
However, for one in five of the people in residential homes for a year or more, our respondents said that there were relatives who they felt might have visited more often.
This compares with one in four of those living in the community who were felt to have relatives who might have helped more  a difference which might well occur by chance.
QUALITY OF LIFE
Respondents were asked whether they would describe the quality of life of the person in the year before they died as good, fair, or poor.
Two things appeared to affect this assessment independently.
One was whether the respondent was a staff member or a relative.
Those spending any time in a residential home were less often described as having a poor quality of life if the respondent was a staff member than if he or she was a relative: 10 per cent compared with 42 per cent.
This probably reflects the perceptions of the respondents rather than a real difference in the actual quality of life of the people who died, although staff members may have been more willing to act as respondents for residents they had got on well with, and those residents may have had a better quality of life because of their relationship with the staff.
The other factor that was related to assessments of quality of life was whether people had been in a residential home for a year or more or for a shorter time.
Taking all types of respondents together their ratings were similar for people who had not been in a residential home at all and those who had been in one for a year or more, while those who had only spent part of the last year of their lives in a residential home were generally felt to have had a worse quality of life: for 39 per cent of them it was rated as poor compared with 27 per cent of the other two groups.
The difference between those spending all and those spending some of the last twelve months in a home persisted when the analysis is confined to those for whom relatives responded.
This can be seen in Table 7.
In addition, among those for whom a staff member responded only one in twelve of those admitted during the last year of the life was felt to have had a good quality of life during that time compared with nearly half (46 per cent) of those who had been in a home for a year or more.
One possible explanation for these findings is that people take time to adjust to living in residential homes.
Another possibility is that those going into residential homes shortly before their death do so in a crisis situation.
Weaver and his colleagues (1985) found that residents most able to come to terms with admission were those who had exercised some degree of control or choice in entering residential care.
Findings from the present study suggest that once people had become established in the homes their quality of life, in the opinion of their relatives, was similar to that of other people who died and had never been in such homes.
DISCUSSION
Most of the people, just over two thirds, who spent all the last year of their life in a residential or nursing home were 85 or more, and it is this age group which is predicted to increase most rapidly in the next twenty years (Central Statistical Office (CSO), 1989) so it is likely that increasing numbers will spend the last year of their lives in such homes.
But this increase may not be as great as the rise in the numbers of older people if there is an improvement in their state of fitness as Jefferys and Thane (1989) suggest.
The symptoms that were more common among those in residential homes, particularly confusion, incontinence, and bad temper, indicate some of the problems in making these homes attractive and stimulating places to live.
The description of one man of ninety-four illustrates many of the difficulties:
He wasn't ill, it was just general deterioration.
He was still quite active for his age, but his mental deterioration was the problem  he'd lost his sense of direction.
He found it difficult to accept his own failings.
Memory loss was the worst aspect, the fact that he was unable to complete a crossword puzzle in ten minutes as he once could troubled him.
He was suffering from frustration.
He was worried because he was deaf, worried because of his bowels.
He worried a lot.
He was, I think, unhappy and tired.
He often said he'd had enough of living and was' ready to go ' was how he put it.
(Died from falling down a flight of steps and fracturing his skull.)
That man was admitted to hospital after his fall and died there within a week without regaining consciousness.
But in general people living in residential homes were less likely than others either to die in, or to be admitted to, hospital during the last year of their lives.
Residential homes seem to cut down the use of hospital care; they probably also reduce the need for it, but this is less certain.
They may also reduce the demand and need for night calls from general practitioners and, taking into account the age and frailty of residents, the pressures on the district nursing service.
Because of the nature of the sample this study could not ascertain the views of residents themselves.
Others who have done so have argued that elderly people are often faced with a choice between an unpleasant battle to survive in their own homes and an equally unpleasant enforced dependence in the institution (Wilkin and Hughes, 1987).
In this study relatives and friends generally thought conditions in the home were ' good ' rather than ' not so good ', although a number of their comments and descriptions indicate that circumstances were sometimes far from ideal and analysis of respondents' views of the quality of life suggest that the initial period after admission may be particularly stressful for residents.
While most residents were in fairly frequent contact with relatives and friends from outside the home, one in five of those who spent all the last twelve months of their lives in a home had either no visitors at all or less than one visit a month.
Becoming old and frail can be distressing; to be also isolated from relatives and friends seems dire.
The task of relieving avoidable distress and creating a caring and heartening environment in residential and nursing homes is going to be of increasing importance in determining the quality of the last year of people's lives.
How can we Measure the Effects of Client Participation on the Effectiveness of Social Work Intervention?
ALAN S. YORK and HAYA ITZHAKY
Alan York was born in England and received his BA (in law) from Trinity College, Oxford in 1962.
In 1963 he emigrated to Israel, took a certificate in Community Social Work in Tel Aviv and worked as a community worker in an immigrant development town.
He returned to England in 1970 and took his MA in Social Work at Leicester in 1972.
He received his doctorate (in sociology) at Bar-Ilan University, Ramat-Gan, Israel in 1979.
Haya Itzhaky was born and brought up in Israel and graduated from the School of Social Work at Bar-Ilan University in 1973.
She worked as a community worker in an immigrant village, and took her MA (in social work) and her doctorate (in sociology) in Bar-Ilan.
Both Dr York and Dr Itzhaky teach in the School of Social Work at Bar-Ilan University, where they are in charge of the community work specialization.
SUMMARY
This paper examines the concepts of client or citizen participation in community work and of effectiveness in social work and other related fields.
It surveys and summarizes the literature in these two fields, focusing on how the two key concepts are empirically measured.
After examining the literature on the effects of citizen participation on community work intervention, the paper puts forward an operative model for both researchers and practitioners to measure how the participation of clients can make community work more effective and productive.
The underlying assumption (or, more accurately, hypothesis) of this paper is that client participation makes for a better social work intervention, and thus the higher the degree of client participation the more effective the intervention will be (Kurzman and Solomon, 1970; Freedberg, 1989; York, 1989).
In Bernstein's words:
we find that imposing, telling or giving orders do not work well.
Only as the client is thoroughly involved and comes to accept on deepening levels the process of change can our methods be effective in relation to our goals (Bernstein, 1960, p. 8).
Most of the debate on the validity and primacy of client self-determination has taken place within the casework context (Bernstein, 1960; McDermott, 1975; Freedberg, 1989) where the elements of eligibility, statutory rights, social control and professional authority may be of central importance (Rothman, 1989; Spicker, 1990).
This paper will consider the issue in the context of community work, in which the elements accentuated by Rothman are of less consequence and client participation has indisputably remained a primary value.
We are currently engaged on a research project that seeks to test empirically whether client participation does indeed increase the effectiveness of community work intervention.
We found much literature on client participation and even more on effectiveness, but there is considerable controversy on how to measure these variables.
Thus the scarce research work on the effects of participation on effectiveness is further limited by its inability clearly to define (and to measure) what it is studying.
In this paper, we shall review the relevant literature, including that from other applied social sciences, and show the operative implications for research and practice.
The first section will review the literature on client participation and its measurement, and the second section will summarize the ' state of the art ' in the measurement of effectiveness, stressing evaluation of community work intervention.
The third section will review the published research linking participation and effectiveness in social work and other human welfare interventions and programmes.
If these researchers had succeeded in measuring operatively and empirically what degree of participation was achieved and how this affected the degree of effectiveness, then our search would have been finished.
The reader will have guessed, no doubt, that the quest was unsuccessful, and so the final section will suggest an operative model with practical implications for community work practitioners and researchers.
WHAT IS CLIENT PARTICIPATION AND HOW CAN WE MEASURE IT?
Traditionally social workers have espoused client self-determination, the right of the clients to participate in decisions affecting their lives and interests (Kurzman and Solomon, 1970; Gilbert and Specht, 1976).
Many and varied reasons have been put forward for this basic element of practice, but they can be summarized under three headings: ideological reasons (the right of a person to decide for himself or the recognition of this as an existential reality), educational and developmental reasons (participation as part of human growth and development, for the individual, the family, the group, the organization or the community), and political reasons (participation as a means to enfranchisement and empowerment, an antidote to alienation) (Ross, 1960; Bernstein, 1960; Katan, 1980; Freedberg, 1989; Rothman, 1989).
Freedberg (1989) has shown how social workers have applied the principle in different ways at different times.
Bernstein (1960) puts forward limitations on self-determination, and shows how it stems from ' the supreme social work value... human worth ' (p. 7).
Thus self-determination is not an absolute value, but it is' supremely important.
Only through the rich utilization of this concept can we fully honor the human-worth value ' (p. 8).
Rothman (1989) extends Bernstein's critique and suggests that many social workers use manipulative and directive techniques under the guise of self-determination.
He suggests the construction of a continuum of categories, each with its set of conditional factors, in which the practitioner could decide upon the degree of directiveness or non-directiveness needed in a specific situation.
Almost all the examples he gives for a restriction of client participation are atypical of a community work situation.
Defined simply (even simplistically) client participation means that the client system, as a whole or in part, is always part of the action system.
How can we measure client participation in community work?
What are the dimensions that can be observed empirically?
The literature is well developed and offers a wealth of suggestions.
Arnstein (1969), Hollnsteiner (1977) and Wandersman (1979a) base typologies on the distribution of power between the establishment and the clients, drawing similar but slightly different levels between the two extremes, and considering other relevant dimensions such as roles, representation, influence, and the stages of the process.
Mogulof (1974) sees a dialectic in which the establishment and the change agent system determine how much participation they will allow, while the client system decides how strong is its desire or demand for participation.
The level of participation becomes, therefore, a function of the motivation of the partners.
Susskind and Elliot (1981) take a similar line, seeing the relationship between the establishment and the client system as an evolution from paternalism to co-operation via conflict.
Others prefer a multiplicity of dimensions.
Cole (1974) describes two major dimensions: the scale of participation and its range.
The former delimits the activity involved, the latter the number of participants.
Wandersman (1979b) outlines five dimensions: the setting and the scale of the activity, the stage in the process at which participation takes place, the identity and the representativeness of the participants, the level of participation (as described above in Wandersman, 1979a), and the different techniques used by the change agent in achieving participation.
Finally, Liron and Spiro (1988) choose three multi-faceted dimensions to examine the participation of clients in the Israeli Urban Renewal Project: authority (how central, final, and comprehensive are the decisions of the participatory steering committees, and how complete is their domination over resources), the influence of the client representatives vis  vis the establishment and the professionals, and the representativeness of these representatives.
All of the dimensions from the literature are measurable, and many of the authors present operative techniques for their measurement.
Thus we shall conclude this first section by summarizing six criteria put forward in the literature reviewed above, together comprising a multidimensional and empirical test of the extent of client participation at the community level.
1.
LEVELS OF PARTICIPATION
The levels will reflect the influence and authority of the different partners, and the distribution of power among them.
2.
TYPE AND SCALE OF ACTIVITIES
In what types of activities or interventions are the levels of participation applied?
The scale of the activities can be inferred both from the activity type and the level of participation involved.
3.
REPRESENTATIVENESS OF CLIENT PARTICIPANTS
How far do the clients involved in participation represent the complete client system?
This question is important whether the client system be a social group, a committee of a functional community or a neighbourhood activist group.
In one way or another, the representatives will be compared with the total client system.
THE ORIGINS OF PARTICIPATION
Upon what authority do the clients participate and whose initiative was it that launched the participation?
Is participation voluntary and welcomed by the clients, or is it demanded of them by the action system, the change agent or even by some extraneous system, as in some parent-teacher associations?
TECHNIQUES OF PARTICIPATION
What methods are used to achieve client participation?
The different methods will reflect different levels of participation, and show how broad the participatory front is.
ATTITUDES TO CLIENT PARTICIPATION
It seems clear to us that all of the above criteria will be enhanced or negated by the way in which the change agent or the action system reacts to participation: what stereotypes do they have of client representatives? do they believe in participation? what expectations do they have of participation and client representatives?
Similarly, parallel questions will be asked of the participants, for the clients also bring expectations and stereotypes (of the workers, administration, etc. and of their fellow clients) to participation, as well as a wide gamut of often conflicting ideologies.
WHAT IS EFFECTIVENESS AND HOW CAN WE MEASURE IT?
Social work has begun to evaluate its effectiveness in the last fifty years, and Sheldon (1986) has described two separate waves of reviews and researches.
From the 1940s till the early 1970s the question asked was' does social work work? ', and the answers given, for example by Mullen and Dumpson (1972), Fischer (1973 and 1976), and Wood (1978), were generally negative: social work intervention was apparently ineffective.
The second wave, from the 1970s to the present day, have limited their attention to the question ' how effective is a particular program with a particular type of client? '
(Reid and Hanrahan, 1982, p. 328), and the answers appear to be positive.
Reid and Hanrahan (1982), in a review of twenty-eight studies between the years 1973 and 1981, find that the objectives of the interventions were generally achieved, and so they conclude that ' intervention carried out by social workers can be effective ' (p. 333).
Of particular importance to the subject of this paper, they emphasize in their summary that the client in the studies reviewed ' is engaged as a voluntary collaborator ', and his' motivations were directed toward what the social workers were attempting to provide '.
This kind of ' motivational congruence between practitioners and clients' was generally missing from the studies in the first wave of effectiveness research, they suggest (p. 338).
This implies the connection between client participation and effectiveness that Bernstein (1960) (as quoted above) seems to take almost as axiomatic, and which we would like to demonstrate empirically.
Ruben (1985) reviews twelve more research studies between 1978 and 1983, and generally replicates the findings of Reid and Hanrahan.
(See also Brawley and Martinez-Brawley, 1988.)
How is effectiveness defined and measured in these studies?
Sheldon (1986) points out that ' effectiveness' is not absolute: for whom? in relation to what? are some of the limitations.
Thus his definition is' a much less ambitious affair ':
It implies describing carefully the nature, extent and implications of any problems prior to intervention; saying what it is you might do to alter these and why it would be beneficial to try; describing how best this might be brought about; saying in advance what kinds of public test might be applied to support the view that something worthwhile has been achieved; pursuing a solution, and then defending the results against criticism from peers (Sheldon, 1986, p. 224).
However, the literature so far described is concerned almost entirely with social work at the micro level: individuals, families and small groups, much of it using techniques that are of little relevance to community work, for example behaviour modification.
Sheldon (1986) suggests that one implication of the research findings is that social workers are more effective on ' smaller-scale problems' than ' large-scale experiments on problems of genuine concern to the community '.
Jenkins (1987) describes effectiveness research so far as a ' limited concept, primarily relating to the intervention process' and with little validity in the ' macro area '.
Rees (1987) maintains that effectiveness is meaningless without usefulness, and this implies, among other things, political constraints.
Thus our review of literature must now address those studies specifically concerned with the evaluation of effectiveness in community work.
Key and his associates (1976), in an attempt to decide how to evaluate their community project, make a thorough review of evaluation research literature.
They emphasize qualitative and quantitative approaches, methods ranging from the rational to the intuitive, and the role of the clients in the evaluative process.
Hunt (1987) states the dilemma more briefly: the scientific approach may be quite unrealistic.
Control groups are not relevant, and important variables can seldom be specified in advance owing to the complexity and dynamic nature of the processes involved.
Objectives are usually very general to begin with, and the non-directive nature of the work means that they are defined as the work goes on.
Finally, she asks, ' when is an outcome not an outcome? ', that is, how are outcomes defined and for whom: funders? community workers? people/clients? researchers?
These studies (and earlier studies by the team of Epstein, Fellin and Tripodi: Tripodi et al.,
1971; Epstein et al.,
1973) lay a basis for the study of effectiveness in community work, but, instead of giving clear, operative and replicable indications of how to measure it, emphasize the problematical nature of the task.
Thus, in the spirit of the culture of openness and coherence proposed by Rees (1987), let us consider what might be learned from other applied social sciences, particularly management science.
Lewin and Minton (1986) in their comprehensive survey of the literature find no less than thirteen orientations to management from the year 1911 onwards, each with its own typical effectiveness attributes.
Steers (1975), summarizing seventeen research studies, finds various criteria of effectiveness, but the three most frequent ones are adaptiveness, productivity, and participant satisfaction (meaning, of course, worker, not client participation).
He concludes that the effectiveness of a given system should be based on its ability to fit in with the external systems making up its environment on the one hand, and, on the other hand, on its competence in allowing its own sub-systems to fit in with each other.
This model is clearly process oriented.
Pennings and Goodman (1977) present a systems model that also stresses internal processes: they see effectiveness as the result of coalitions and arrangements between dominant and interest groups, and the satisfaction of the participants from the efficiency of these inner processes.
Miles (1980) presents an ' ecology model ', not dissimilar to the systems models.
Dividing previous schools of thought into goal attainment models and systems models (and criticizing both of them), Miles defines organizational effectiveness as' the ability of the organization to minimally satisfy the expectations of its strategic constituencies' (Miles, 1980, p. 375), thus demanding both sets of criteria.
However, the criteria of effectiveness will change with the different constituencies: different sub-systems will define effectiveness differently, some in terms of goal attainment (say the management and the stock holders) and some in terms of process (the workers, for example).
This means that ' organizational assessment is best conceived of not as a one-shot affair, but as an ongoing process' (Miles, 1980, p. 380).
Campbell (1977) too divides organizational effectiveness into two categories: ' goal-centred ' and ' natural system ', and he lists no less than thirty criteria ' that have been proposed seriously as indices of organizational effectiveness' (p. 36).
Probably, therefore, Cameron (1986) is right when, after summarizing five areas of agreement on effectiveness and three areas of conflict, he concludes that ' agreement about effectiveness is mainly an agreement to disagree ' (p. 544).
His suggestion is to see competing criteria as both compatible and congruent and to accept that ' organizational effectiveness is inherently paradoxical.
To be effective, an organization must possess attributes that are simultaneously contradictory, even mutually exclusive, (Cameron, 1986, pp. 544C5).
Thus successful organizations will have a high degree of role specialization (reinforcing expertise and efficiency) together with a high generality of roles (encouraging flexibility and interdependency), a continuity of leadership alongside of infusion of new leaders, and so on.
Our conclusion must be that there are many criteria of organizational effectiveness, some of which are mutually exclusive, yet, paradoxically, these seemingly conflicting attributes are able to co-exist and to work together.
Most of these criteria are measurable and have been measured in management research in quantifiable terms.
It remains to be seen to what extent they can be applied to social work.
We propose two composite criteria for the measurement of effectiveness, both using variables used in previous research, in accordance with the specific demands and context of the researcher or practitioner.
1.
CLASSIFICATION OF EFFECTIVENESS
Using the terms of Miles (1980), we classify effectiveness by ' goal attainment ' (that is goal achievement, productivity, etc.) and ' systems processes' (adaptiveness, participant satisfaction, inner processes, systems resources, etc.).
The assessor may use either of these classifications (with whatever sub-groups he or she prefers) or both of them.
Both measurements of effectiveness are valid, and the one used will depend upon the purpose of the assessment, the customer ordering the assessment and the audience at which the assessment is aimed.
An evaluation of effectiveness of a community economic enterprise in terms of productivity and profit is legitimate in one set of circumstances; the same evaluation of the same enterprise in terms of citizen satisfaction and decision-making processes is equally legitimate in another set of circumstances.
If possible, the evaluation will include both types of effectiveness measurement.
2.
EFFECTIVENESS EVALUATED BY WHOM?
Clearly not just the type of evaluation of effectiveness but its essence will be affected by who evaluates.
In the example of a factory, completely different assessments may be expected from the board of directors, shareholders, management, workers, and shop-stewards.
DOES PARTICIPATION LEAD TO EFFECTIVENESS?
A REVIEW OF LITERATURE
The amount of published research on this topic is not extensive.
The management research field includes many studies on effectiveness, but the only type of participation to be found in them is worker participation, quite different from our topic of client participation.
Some of the conclusions from this literature do seem applicable to our topic, and they will be considered at the end of this section.
Research on the participation of the public in general or clients in particular tends to centre around the form of participation involved (inputs) and its effect on the change agent and action systems (processes, ' thruputs'), but little attention is paid to the results of participation (outputs) in general, and in achieving the specific goals of the activity involved.
Moreover, the conclusions of the studies which do consider the effectiveness of participation are by no means clear or unanimous.
Mazmanian (1976), in his evaluation of projects carried out by the Corps of Engineers, and Gilbert and Specht (1977), in their classic evaluation of the Model City programme between the years 1967  71, appear to reach a similar conclusion: participation helps process but not goal attainment.
Mazmanian finds participation effective in such process goals as leading to public appreciation of the Engineers' work, but not in what he calls' product ' goals, such as increasing public support for the projects or recommending courses of action.
Gilbert and Specht show that project workers used client participation effectively in the planning stage, where the process was the main consideration, work was non-directive, and the major skills needed of the worker were interactional in nature.
But when they reached the implementation stage, in which their orientation was to concrete tasks and the skills required of them were chiefly technical, then they became more directive and used client participation less.
These conclusions appear to have been replicated by York (1989), who finds that community workers express more satisfaction and effectiveness in their short-term projects, which are more directive and task-oriented, than in their long-term projects.
However, the conclusions of Gilbert and Specht are indirect and do not necessarily show that client participation in implementation stages is ineffective.
As York finds, workers prefer not to involve their clients in short-term task processes, but this may be for a multitude of reasons other than that of effectiveness: lack of time (client participation is generally time consuming), agency policy and pressure (paternalism), professional socialization and norms, and so on.
Only Mazmanian, so far, clearly states that client participation leads to ineffectiveness, and he does so only for ' product goals'.
We should like to know more about how participation was achieved in these projects and what was the nature of the tasks involved.
In any case, it is clear that Mazmanian does not deal with social work programmes, and the change agents involved were not social workers.
Gittell (1980), in a research study of school-related citizen organizations in three American cities, finds that both lower- and middle-class parent organizations had little effect on educational decision-making, their declared goal.
In our terms, client participation was not effective.
However, she shows how imperfect was the participation, particularly in lower-class organizations, and also makes clear that some of the objectives were achieved.
Thus Gittell's data are not conclusive.
Rosener (1978) finds much consideration in the literature of participation as an end in itself, and this is relatively easy to evaluate: how many clients participated? what kinds of people? how frequently? how much time did they invest? what were their attitudes (and those of the workers) to participation?
However, participation is seldom evaluated as a means to an end, that is effectiveness, she maintains.
To do this, two difficult questions must be answered: what are the criteria of effectiveness (goals and objectives) according to the different participants (citizens, officials, workers)? and how clear is the cause/effect relationship between these goals and participation?
Thus Rosener suggests a matrix between ' knowledge of a cause/effect relationship between participation program and achievement of specified goals and objectives' (complete or incomplete) and ' agreement on program goals and objectives, whose they are, and the criteria to measure success and failure ' (yes or no) (1978, p. 459).
But, as Rosener herself points out, examples of complete knowledge and perfect agreement are rare indeed.
We learn, therefore, two important lessons: deciding before intervention what our goals and objectives are, if possible a common decision of all the participants (though separate, if not conflicting decisions would be legitimate); and setting criteria and measurement indicators to show the relationship between participation and effectiveness.
The first point is undeniable, and few field workers and even fewer researchers would object to it a priori (though its applicability may be complicated by the dynamic development of a project).
The second point is more difficult, for how often in the social sciences do we have ' complete knowledge ' of a cause/effect relationship?
The aspiration towards some sort of rational understanding of the causal relationship is probably the best we can achieve.
Although management science literature on worker participation is of limited relevance to the topic of client participation, some of the recent research on the relationship between worker participation and productivity sheds light, in our opinion, on the wider topic of how participation in general leads to effectiveness.
Erez et al.
(1985) suggest that participation per se does not cause better performance; they hypothesize that participation in decision making strengthens the level of goal acceptance, and this leads to better performance.
Thus the causal model is a two-step one.
Their two empirical studies, one laboratory and one in the field, partially support their hypotheses: increased goal acceptance leads to more effective performance, and participation is an effective strategy to enhance goal acceptance.
They suggest that this may be through the perceived control an individual has over his goals, as compared to that when goals are externally assigned.
Miller and Monge (1986), in their ' meta-analytic review ' of participation, satisfaction, and productivity, in forty-seven articles and book chapters, find two possible models to explain how participation and productivity are linked.
The cognitive model proposes a direct causal link: participation brings about enhanced information on goals, and so performance is improved.
The affective model, like that of Erez and her associates, is multi-stage: participation fulfils needs, these lead to satisfaction which strengthens motivation, and thus workers' productivity is improved.
Both of these models are supported by their analysis.
They suggest that the models may be linked, and further suggest that the direction of the causal link between participation and satisfaction is not entirely clear: possibly satisfaction leads to participation and not vice versa.
Both of these management research studies stress the complexity of the relationship between participation and effectiveness and of the variables involved.
Thus we can use the existing research to help us find our empirical measures for the effects of client participation on effectiveness in social work, but we have not yet achieved our aim.
None of the studies presents an operative model (Rosener, in our opinion, comes closest, but her model, as she shows herself, is more heuristic than empirical), and so in the next section a model will be set out and explained in the light of the evaluation criteria for participation and effectiveness put forward in the first two sections of this paper.
We believe the model can be used to answer the question posed in the title of this paper.
PARTICIPATION AND EFFECTIVENESS: AN OPERATIVE MODEL
In the evaluation of the effectiveness of social services and interventions, we have shown the importance of who evaluates, and our model identifies three different systems (on the basis of Gilbert and Specht, 1979, and Pincus and Minahan, 1973) that could be involved in the process:
1.
Change Agents  professional and para-professional workers in the change agent system;
2.
Administration, Political and Administrative Establishment  within the change agent system and in the larger systems of which the change agent system is a sub-system, for example politicians and bureaucrats in local or national government within which the change agent system is located;
3.
Clients  individuals and groups within the client system, those that were part of the action system and those that were not.
These, then, are the three groups that will evaluate the effectiveness of client participation, each in its own way and using its own peculiar criteria.
Each of the three groups will use the double classification of goal attainment (goal achievement, productivity, etc.) and systems processes (adaptiveness, participant satisfaction, inner processes, systems resources, etc.), as explained above in the summary of the section on effectiveness.
Thus the model includes six categories of effectiveness evaluation, two each for all three groups of evaluators.
They are summarized in Table 1 with examples given of the types of outputs to be expected in each of the categories.
The change agents, in the light of their professional values, get satisfaction out of client participation (Bernstein, 1960, although the findings in York, 1989, referred to above, in the third section, may throw some doubt on this).
Goal attainment is also improved in the long run: leadership development among the clients is a goal by itself, and the strengthening of client independence will lead to the better preservation of any other outputs achieved.
The clients get satisfaction from participation (or, perhaps, participate as a result of their satisfaction), take part in a learning process which enriches them, and identify with a process in which they themselves are involved.
They achieve declared or undeclared goals: involvement, legitimization, recognition, prestige, all of which increase their influence among the others in the client system and those in external systems, thus increasing the chances of further and better outputs in the future.
The administration and the external political and administrative establishment may see participation as a political end in itself, as well as a means to building coalitions for the political good of their establishment.
If participation saves resources (jobs carried out by volunteers, for example, instead of paid workers), then this is an important output, along with the public legitimization and support which are likely to accompany client participation.
This, then, is the model which seems to us most appropriate for attempting empirically to answer our question: How can we measure the effects of client participation on the effectiveness of social work intervention?
Using the six measurement criteria summarized at the end of the section on client participation, we are able to measure the extent to which our clients are participating in the community work project (the independent variable), and the model presented in this section allows us to evaluate the effectiveness of the project (the dependent variable) in a number of different ways.
CONCLUSION
In this article we have reviewed the extensive literature on client participation and effectiveness and the limited literature in which the effect of the former on the latter was examined.
With the help of this previous research, we propose criteria for the measurement of these two important variables and an operative model to illustrate and to measure how participation can lead to effectiveness.
We hope that improved methodology will allow a better evaluation of the relationship between two extremely important values in the social worker's day to day work, and will show that the participation of our clients is not just ideologically based but practically worthwhile.
General Practice, Social Work, and Mental Health Sections: The Social Control of Women MICHAEL SHEPPARD
Michael Sheppard trained in social work at the University of East Anglia and was a social worker in Norfolk Social Services Department.
His publications include Mental Health: The Role of the Approved Social Worker (University of Sheffield Press, 1990) and Mental Health Work in the Community: Theory and Practice in Social Work and Community Psychiatric Nursing (Falmer, 1991).
SUMMARY
A survey was undertaken of all referrals for compulsory admission received by a city mental health centre over a one year period.
The centre receives the overwhelming majority of all these referrals in the city.
Referrals from GPs were compared with other referrals focusing particularly on women.
The results indicate GPs discriminated against women, referring considerably more women than men with less emphasis on major (psychotic) mental illness.
The involvement of approved social workers (ASWs) in the assessment process was associated with diversion of many of the women away from compulsory admission, although even they appeared affected by patriarchal assumptions.
GPs behaviour with regard to sections is consistent with reports about sexist practice in other areas of work.
This, however, is particularly grave with sections because of civil liberties implications.
The article concludes that ASWs need be aware of potential sexist GP practice, that their psychosocial perspective is critical to assessment and that ASW training should include gender issues.
Researchers are beginning to unravel some of the factors significant in the process of assessment for compulsory admission (section assessments) under the 1983 Mental Health Act.
Studies have focused on Approved Social Workers (ASWs) (Sheppard, 1990; Barnes et al.,
1990) and the police (Bean et al.,
1989; Bean, 1990; Rogers, 1989).
One study has used theoretical and empirical research to provide the foundations for a social work knowledge base assessing social risk (Sheppard, 1990).
There has, however, been no focus on the position of GPs and their interaction with ASWs in the assessment process.
This article examines general practitioners (GPs) as referrers, and women as the subject of referral for section under the 1983 Act.
It does so by examining the process of assessment in which the ASW plays a critical role between the initial referral from the GP and the decision whether or not to section the woman.
This is interesting, both because of the serious consequences of a mental health section, and because of concern expressed about sexist practice by GPs in other areas of their work (Barrett and Roberts, 1978; Cooperstock, 1978).
GPs occupy a pivotal position in relation to mental health.
Because they occupy key positions in primary care, they are for most people the most accessible caring profession, and psychiatry is considered by doctors to be primarily their occupational territory.
Hence they deal with high levels of psychiatric morbidity (Jenkins et al.,
1988; Goldberg and Huxley, 1980).
They are likely, therefore, to be the first profession approached by, for example, relatives, for assessment.
In many circumstances GPs are likely to be the first professional to decide a section is necessary and initiate the section assessment process subsequently involving ASWs.
It is widely recognized that the proportion of women who suffer mental disorders  particularly depression  exceeds that of men (Cochrane, 1983).
A variety of possible contributing factors have been put forward: that biological factors make women more vulnerable than men; that women are socially disadvantaged by the roles they are expected to perform, and psychologically disadvantaged by socialization preparing them for these roles; that women express their emotions more readily; and that health professionals are more likely to equate feminine characteristics with ill health (Weissman and Klerman, 1977; Penfold and Walker, 1984; Corob, 1987).
Feminists in particular make much of the social disadvantage under which women suffer.
In view of this disadvantage, and possible GP sexism, the study of GP section referrals and the influence of ASWs in the subsequent process of assessment is significant.
Compulsory admission in England and Wales is carried out under the Mental Health Act 1983.
This is primarily carried out under sections 2, 3, and 4, although section 5 provides for doctor and nurse holding powers, police powers exist under section 136 and courts have powers to direct individuals into hospital (for example, section 37).
Section 2 involves admission for assessment, section 4 for emergency admission for assessment and section 3 for admission for treatment.
Admission under these sections, in brief, may be made when a person is suffering from a mental disorder, and such admission is in the interests of the person's (defined as patient) own health or safety or for the protection of other persons.
The duration of section 2 may be up to 28 days and of section 3 up to six months, although it is renewable.
Section 4 may last up to 72 hours, and may be made only when admission is of urgent necessity and complying with the provisions of section 2 would cause undesirable delay.
Application may be made by an ASW or nearest relative although the draft code of practice (Department of Health, 1989) states the former should apply whenever possible.
Sections 2 and 3 require two medical recommendations, one by an approved doctor, while section four requires only one.
Individuals may also be admitted informally (section 131) or not admitted at all.
There is, therefore, a variety of possible outcomes following a referral.
METHOD
The research was undertaken on referrals received by the Psychiatric Advisory Service (PAS) at the Nuffield Mental Health Centre in a medium sized city of about 240,000, with a very small ethnic minority population.
The Centre is situated in a central location with good transport links to all parts of the city.
The PAS is well known in the city and receives the overwhelming majority of the section assessment requests received during normal working hours.
It is staffed by a team of community psychiatric nurses and a team of social workers together with psychiatrists approved under the Act.
It is open from 9 a.m. to 5 p.m.
Referrals at other times are received by out of hours social workers.
Referrals received from all sources requesting a section assessment over a one year period were included in this study.
When the ASWs responded to a section referral, whatever the source, they were asked to complete fully structured questionnaires.
These had been developed over a one year period and had been fully piloted.
Detailed analysis both of the questionnaires and piloting is provided elsewhere (Sheppard, 1990; 1991).
The research instruments involved questions about the source of referral, section requested and psychiatric state suggested, if any, in the referral.
They involved demographic characteristics such as age, sex, and marital status; information on psychiatric state and psychiatric disorder; and social problems.
Finally, information was provided on who in the patient's social circle was interviewed as part of the assessment process, factors considered in assessment, and outcome.
These data, it should be remembered, were provided by the ASW, and in some respects, such as the presence of social problems, will represent their perspective on the patient.
In other respects, such as outcome, we have unambiguously objective information.
These questionnaires were completed within the week following assessment.
Agency records, which recorded all referrals, were checked weekly to confirm that ASWs had completed questionnaires where relevant.
RESULTS
MALE AND FEMALE REFERRALS
Altogether 120 referrals were made during the study year.
Forty (33 per cent) were by GPs.
Of the other 80 referrals 17 were by the family, 13 by psychiatrists, 17 other health referrals, 15 police, nine social workers and nine other referrals.
The overall rate of referral was 49.9 per 100,000 population, and for GPs, 16.6 per 100,000.
These figures relate to referrals during working hours only, and compare with overall referral rates varying between 19.9 and 102.7 per 100,000 in 42 social services departments studied by Barnes et al.
(1990).
Although, therefore, quite high, the rate of referral in this study was well within the range for different areas.
Seventy-five referrals (62 per cent) were women.
This is consistent with trends identified by Barnes et al.
(1990), and raises issues of what, if any, social processes are occurring which leave women more frequently identified as candidates for sections.
The social profile of the women differed markedly from that of the men in a number of respects.
Men tended to be younger than women: 58 per cent of men compared with 49 per cent of women were aged under 45.
A number of differences are significant.
Table 1 shows significant differences in marital status.
Markedly more men than women were single, while considerably more women were married or cohabiting.
More women than men had children, although the small proportion of men with children confirms their limited involvement in family relationships.
Men were more disadvantaged in other ways.
Only 43 per cent of women were unemployed or in families where no member was employed compared with 56 per cent of men (p=0.024).
Sixty-seven per cent of men compared with 42 per cent of women were on state benefit excluding child benefit and retirement pension (universal to parents and the retired elderly) (p=0.013).
Housing condition also showed greater, but not significant, male disadvantage.
When divided into two  stable residence consisting of council and owner occupied housing, and unstable consisting of private rented, lodging, and hostel accommodation and no fixed abode  76 per cent of women lived in stable accommodation compared with 63 per cent of men.
Marked differences, however, are not evident in the ' processing, of these people.
Eighty-nine per cent of men and 84 per cent of women were previously known to the psychiatric services, indicating longstanding problems.
In just over half male and female referrals no specific mental disorder was identified.
Following the section assessment 84 per cent of males and 95 per cent of females were considered to suffer some kind of mental disorder.
Over 70 per cent of both groups were considered psychotic, with rather more men schizophrenic and rather more women suffering affective psychosis.
Frequently, then, an unspecific referral was subsequently considered a psychosis.
On average the ASWs identified more social problems for women (2.12 per case) than men (1.6 per case).
Twenty women (27 per cent) but only five men (11 per cent) had marital problems, while only one man (2 per cent) compared with ten women (13 per cent) had child care problems, both differences show just below significance (taken as p&lt;0.05).
Women, then, had more social problems, particularly connected with familial responsibilities.
Outcome (Table 2) was generally similar for both groups, although considerably more men were placed on section 3 and rather more women on section 2.
The former, which is more draconian, may reflect the longer term perceived treatment needs of the men.
This may be related to a relative lack of social stability: being more frequently single and in less permanent accommodation.
Overall, then, the main differences were in the social profile, and a greater number of social problems in women.
Difference in outcome may be related to different social profiles.
This still leaves the issue of the process by which more women were referred.
WOMEN AND GP REFERRALS
The impact of GPs on the sex composition of referrals is profound (Table 3).
Exactly a third of referrals were made by GPs.
However, the overwhelming majority of GP referrals were women, compared with only just over half the other referrals.
Indeed GP referrals account for nearly all the excess of female over male referrals.
Examination of outcome, furthermore, is intriguing.
The difference between male and female GP referrals was noticeable: five of the eight males referred (63 per cent) were sectioned, compared with only 12 (38 per cent) of female referrals.
However, compared with the 70 per cent of women referred from other sources (39) who were sectioned, the number of female GP referrals sectioned was significantly lower (p=0.011).
This indicates a significantly greater diversion of women, following ASW assessment, generally with a psychiatrist, away from compulsory admission when referred by GPs.
An excess of GP referrals over other referrals of women informally admitted was matched by an excess of female over male GP referrals who were informally admitted.
Two initial observations may be made.
This combination of data  an excess of GP over other referrals of women, together with a significantly greater likelihood of diversion of women referred by GPs away from compulsory admission compared with other female referrals  suggests apparent discrimination on the part of GPs against women.
Second, the role of specialist professionals  ASWs and psychiatrists  appears significant in enabling women referred by GPs to avoid compulsory admission.
Outcome is shown in more detail in Table 4.
This shows a noticeable excess of women referred by GPs over other female referrals who were not admitted at all.
Informal admissions are also interesting.
Noticeably more women referred by GPs were admitted informally than either male GP referrals or other female referrals.
This group involves women who were considered to need assessment or treatment, but where compulsion was not considered necessary.
This suggests the significance of specialist workers in two possible ways: either preventing compulsion where there was insufficient risk to the patient or others, or persuading those who would otherwise have been sectioned to enter hospital informally.
Bean's (1980) typology of doctor  patient interaction showed that, under the 1959 Act, voluntary admission occurred when there was agreement between doctor and patient on the patient's sick role or where, although agreement did not exist, the patient acquiesced and accepted this role.
Compulsory admission occurred where the doctor ascribed the sick role but the patient did not accept it or did not react at all.
The nature of the ASW involvement suggests that they too would be involved in this process (Sheppard, 1990).
Referrals by GPs where outcome was informal admission were consistently characterized by familial disruption: indeed only two of these cases were referred where the woman's behaviour drew attention to her in the wider social environment outside the family.
The noticeable familial disruption occurring in other cases is particularly interesting because it reflects social role expectations of women, where the primary role traditionally is that of homemaker, family manager, and housewife (Oakley, 1981).
Thus, informal admissions were characterized by a combination of mental ill health and transgression of traditional social role expectations.
Three examples, drawn from the ASWs' account, indicate the different types of familial disruption.
The first type  active familial disruption  involved a woman in her thirties who was considered to be suffering from schizophrenia.
She was very overactive and had not slept for three nights.
She was pacing up and down in the house, was hostile to other family members and was shouting at voices that she (and no-one else) heard.
She was paying little attention to her normal ' household duties' such as child care, shopping, and looking after the house.
She had two young children, aged five and seven, and the ASW expressed concern for their welfare.
Her husband, who usually worked 200 miles away, had been forced to return to look after the children.
The second example  a woman considered to be depressed  involved passive withdrawal from her usual role.
In this case, a 45 year old woman with a large family had recently moved into the area.
She was described as having become ' withdrawn and introverted '.
She was not going out at all and was communicating very little with her immediate family.
She had two adult children who lived nearby, and had stopped visiting them as was her usual custom.
She had stopped undertaking her usual tasks within the home, was eating very little and had taken to her bed a few days previously and had not re-emerged.
The third example involved escape from familial circumstances.
This was a woman in her twenties who was considered to be schizophrenic.
There was, according to the ASW, a ' continuing crisis at home '.
She had suffered paranoid delusions for a number of months, that the house was electronically ' bugged ', that there was a camera in the television watching her and that she was being followed.
Her husband was described as having difficulty coping with her.
She had left home and was refusing to return because of the ' weird things' that were going on.
When her husband tried to reason with her she refused to accept what he was saying, and had begun to suspect that he was involved in some way (in the bugging).
The lack of section 4 admissions was due to the availability of an approved doctor at the Centre: the approved doctor is frequently the most difficult of the two doctors to get hold of under sections 2 or 3.
There was, amongst those sectioned, considerably less use of section 3 among women referred by GPs than either male GP referrals or other referrals of women.
It is not immediately obvious why this should be the case, but this provides evidence of a less restrictive response to GP referrals of women, since section 3 applies for six months and section 2 for 28 days.
Section 2 is for assessment and may more frequently be expected where an individual had no previous psychiatric history.
However, the proportion of men and women previously known to the psychiatric services was exactly the same (84 per cent) for both GP and other referrals.
Overall then, not only were female GP referrals sectioned less frequently than other female referrals, but when sectioned they were less frequently subject to the lengthier section 3.
PROBLEMS
We can explore these differences further in relation to problems.
Individuals' psychiatric state was divided into ' definite cases', ' borderline or threshold cases', and ' not cases', where no psychiatric disorder was present.
The ASW was asked which of these were considered applicable to the assessed individual.
Few differences emerged between the relevant groups: 88 per cent of male and 84 per cent of female GP referrals plus 86 per cent of other female referrals were considered definite.
There was a tendency, however, for female GP referrals to involve less major mental illness than others, although differences were not significant.
The individual diagnostic categories were broadly divided into three: neurotic, psychotic, and alcohol or drug problems, which could be neurotic or psychotic.
Of those considered mentally ill rather more female GP referrals (23 per cent) than other female referrals (8 per cent) and no male GP referrals were considered neurotic.
All male mentally ill GP referrals were psychotic, as were 82 per cent of other female referrals, compared with 77 per cent of GP female referrals.
The concept of primary problem was used by the ASW to identify, from the whole range of psychiatric and social problems noted in each case, which was considered to be the main or primary problem.
There was a slight tendency for GPs' female referrals to have fewer psychiatric primary problems: 78 per cent, compared with 86 per cent of other women and 88 per cent of male GP referrals.
Psychotic primary problems also had a slightly lower profile amongst female GP referrals: only 60 per cent came into this category, compared with 74 per cent of other females although similar to male GP referrals (62.5 per cent).
Social problems were divided into thirteen categories.
The ASWs were asked to indicate which of these each individual suffered and which of these they considered to be severe.
Female GP referrals averaged 2.28 social problems compared with 2 for other women and 1.87 for male GP referrals.
While, however, female GP referrals averaged more severe social problems per case (1.34) than male GP referrals (1.12), this was less than other female referrals (1.44).
Nonetheless, there appears to be a different ' threshold ' which triggers GP female referrals compared with other referrals, with greater emphasis on neurotic problems, fewer psychiatric primary problems and more social problems than other referrals.
GPs, therefore, more frequently resorted to section referrals for women with neurotic and social problems than is evident in other referrals.
The pattern of social and mental health problems indicates GPs were more sensitized to mental health and illness, and that this played a significant part in choice of referral.
The more frequent presence of social problems among female GP referrals is interesting here because these might be expected to raise the possibility of alternative referral (for example, to a social worker) and management.
However, these GPs seemed to be seeking a medical solution to these women's problems.
Second, the solution they sought had serious implications for civil liberties since sections would effectively deprive individuals of their liberty.
This search for a medical solution, then, may present a serious threat to civil liberties.
We may pursue this through a further examination of primary problem.
The detailed classifications of mental health and social problems were divided into six: neurotic, psychotic, alcohol  drugs for mental health; and practical, emotional and relationship, and physical ill health for social problems (Fitzgerald, 1978).
Amongst GP referrals, seven (22 per cent) of the women, but no men had emotional or relationship problems as primary.
These involved marital, loss or separation, social relations or isolation, and criminal behaviour problems.
Although not significant, only women had relationship primary problems.
However, this is particularly interesting when related to outcome.
No woman with emotional or relationship primary problems was sectioned.
Indeed, six of the seven women falling into this category were not admitted at all, and one was admitted informally.
In contrast only 15 per cent of all other GP referrals (and 12 per cent of other female referrals) were not admitted, while over half were sectioned.
This suggests GPs were attempting to deal, both medically and compulsorily, and only in the case of women, with problems which at subsequent assessment were considered to be primarily about social relationships.
CONSULTING OTHERS
One expectation of the ASW is that others in the patient's social network  relatives, friends, neighbours and so on  should be consulted to help build a picture of the individual and their behaviour (DHSS, 1983; Department of Health, 1989).
When women were referred by GPs, others were consulted significantly more frequently (88 per cent) than women who were referred by other referrers (61 per cent) (p=0.02).
This may be because the GP's previous contact with the family facilitated such contact or that a balance to the GP's assessment was sought.
In relation specifically to GP's referrals others were consulted with exactly the same frequency whether male or female referrals.
Consulting others, furthermore, had no impact on outcome, consultation occurring in 88 per cent of sectioned and 87 per cent of not sectioned GP referrals.
However, did the person consulted have an impact on outcome? in particular, what was the impact of male family members?
Table 5 shows significant differences in outcome for women according to whether or not a male family member was consulted by the ASW.
Women not admitted were markedly more likely to have a male family member consulted.
Where sections were undertaken, males were more frequently consulted where section 2 was undertaken than section 3.
The picture was similar in relation to sectioned women where female family members were consulted, but interestingly females were not generally consulted where women were not admitted.
Consulting both male and female family members was therefore associated with briefer section periods, but only males were associated with non admission.
Indeed, in relation specifically to women not admitted, the difference between male and female members consulted was significant (p=0.034).
There were also significant differences between GP and other referrals, specifically with regard to women not admitted.
Of the nine women referred by GPs who were not admitted, eight cases (89 per cent) involved consultation with a male family member.
Of the eight other women not admitted only two cases (25 per cent) involved consulting a male family member (Fisher exact probability test: one tail 0.013, two tail 0.015).
There was little difference, however, between GP and other referrals in the rate of male or female consultation in relation to section 2 and in female consultation with regard to non admissions.
These data indicate that consultation with either male or female family members can have the effect on women sectioned, of producing less lengthy compulsory admission.
It indicates also, however, that male family members can have an impact on non admission where females do not, and that this was significantly greater with GP referrals.
This suggests, perhaps, two possibilities.
The mere availability of a family male is reassuring, giving the ASWs and others assessing greater confidence that the woman can remain in the community.
Second, the male family members, because they are male, are given a credibility  and hence power  to influence non admission not given to female members.
However, this power is primarily relevant to GP referrals.
This strongly indicates that it is only males who have the status and power to provide, in the eyes of the relevant professionals, a credible challenge to the GP.
The Memorandum to the Act (DHSS, 1983) indicates a number of factors to be taken into account by ASWs.
The woman's past history of mental disorder, family relations and support and medical opinion were all considered both with GP and other referrals in a minimum of 93 per cent of cases.
The woman's financial and material circumstances were considered less frequently with GP referrals (28 per cent) than other referrals (49 per cent) as was the case with non familial social supports (GP referrals 53 per cent, other referrals 65 per cent).
The referred woman's wishes were, however, considered more frequently in GP (78 per cent) than other referrals (63 per cent).
Although these differences were not significant, there was less concentration on social factors relating to wider social supports and financial-material circumstances with GP referrals, suggesting, in the face of the GP's referral, and despite section diversions, less preparedness by ASWs to examine social factors which may challenge the GP's definition.
What power did the woman herself have to influence the course of events?
Table 6 shows the relationship between outcome and the extent to which their wishes were considered.
This shows significant differences: women's wishes were taken into account most frequently when they were informally admitted or not admitted.
Interestingly, women's wishes were more frequently considered with the longer section 3 than section 2 admissions (15 of 17 cases compared with 15 of 23 cases).
Taken together with consultation with family members, this suggests that, when sectioned, family members had a greater impact on the length of admission.
However, there was little difference between GP and other referrals where women were not sectioned.
Only one of the 20 women referred by GPs did not have her wishes considered.
The wishes of not admitted women referred by others were in all cases considered.
The trend was similar with men.
These data suggest that, for both men and women, outcome significantly depended on whether their wishes were considered, or, alternatively, that their wishes were taken more seriously in less serious cases.
In either situation the mental health state of the woman may be significant.
PROFILE OF FEMALES REFERRED
What of the social profile of women referred?
Women referred by GPs were generally younger than other women: 59 per cent of GP and 42 per cent of other referrals were aged between 16 and 44.
More women referred by GPs were married (47 per cent) or single (28 per cent) than other referrals (35 per cent married and 19 per cent single).
However, 22 per cent of GP referrals and 42 per cent of other referrals were widowed, divorced, or separated, and hence had lost their partner.
Rather more (28 per cent) women referred by GPs than other women referred (12 per cent) had children.
About three quarters of both groups lived in owner occupied or council housing.
Rather fewer GP referrals of working age (54 per cent) lived in a unit with no employed members than other referrals (65 per cent), while fewer GP (33 per cent) than other women referred (48 per cent) received state benefit other than child or retirement benefit.
In relation to GP referrals only, more women (59 per cent) than men (38 per cent) were aged 16 to 44 and more women (63 per cent) than men (43 per cent) were in more stable council or owner occupied housing.
More men (57 per cent) than women (28 per cent) were single, and more women (47 per cent) than men (29 per cent) were married.
Only one man (13 per cent) had children compared with nine women (28 per cent).
The proportion unemployed was similar.
Although differences are not significant the overall profile indicates that female GP referrals were more frequently married, aged between 16 and 44, had children, and lived in more stable housing.
Insofar as GPs were using section referrals as a method of social control, it was more frequently upon these women.
DISCUSSION
The difficulty of research in this area has meant that, with one notable exception (Barnes et al.,
1990) very large scale studies have not been undertaken, whether on the 1959 or 1983 Acts.
It is also relevant that these were referrals to one mental health centre, although the lack of ASWs in district teams meant that these represented practically all the daytime referrals to ASWs.
Although we must recognize the limits of a study focusing on one city, the results are, perhaps, startling, and of considerable interest.
The pattern of results raises major concerns about the role of GPs in using section referrals as a method of social control of women.
Its significance lies in the seriousness of the consequences should a woman be inappropriately sectioned.
Enforced hospital entry would be accompanied by a loss of normal civil liberties.
Under section 2 or 3 this involves detention respectively up to 28 days or six months.
They may be forced to take medication (mind altering drugs), although consent or a second opinion is required three months after the beginning of treatment (section 58).
Although some protection is provided by appeal to the Mental Health Review Tribunal or nearest relative right to discharge, in practice both are rarely undertaken or successful, relative to the total sectioned.
Referral for compulsory admission is clearly an issue of power.
It involves generally male professionals (GPs) raising the issue of deprivation of liberty for women, frequently married and aged under 45, and, compared with other referrals, likely to have children.
The excess of these women in GP referrals, it might be argued, represents the excess of women with children seen in general practice (Shepherd et al.,
1966).
However, this fails to account for various other elements in the process of assessment.
Following assessment, involving ASWs and, at times, approved doctors, fewer women then men referred by GPs or other women referred were actually hospitalized.
The incidence of informal admission was also higher in GP referred women.
Although considered in need of assessment or treatment, these women did not require compulsory admission, with the attendant limits on civil liberties.
We may legitimately ask why, given GPs' professional concern with mental health, this should be the case to a greater extent than with other  frequently lay  referrals.
It strongly suggests that GPs' concern for civil liberties was considerably outweighed by their concern for treatment: at best a conceptual inability to distinguish between treatment need and treatment requirements combined with curtailment of civil liberties; at worst something more sinister.
The data indicate a connection between adequate assessment of social factors and civil liberties, commented on elsewhere (Sheppard, 1990), and considered further later.
They ascribe a significant role in this respect to mental health specialists, particularly ASWs.
The distribution of problems is also revealing.
Women referred by GPs were more neurotic and less psychotic than the comparison groups, and had more social problems.
GPs showed a different threshold triggering their female compared with their male referrals.
More telling still is that where ASWs considered relationship problems primary no woman referred by GPs was sectioned.
A further important aspect of the assessment process was the significance of ASW consultation with male family members, associated with section 2 more than section 3 and to a greater degree with GP than other referrals with non admission.
This creates a strongly patriarchal image: of women as the hapless subjects where, in the face of the (professionally powerful) GP referral, it was through ASW consultation with males that a potential section was averted, or the duration of a section was shortened.
A possible (relatively) benign explanation might be presented for GPs' failure to pursue a non compulsory route: that women, whose conventional role involves running the home and family, would be less willing to go into hospital precisely because of their feelings of responsibility for others in the home.
However, the evidence contradicts this.
First, of those not sectioned, a considerable proportion were not hospitalized at all: whether or not mentally ill, they were not, following assessment, seen to fit the criteria for compulsory admission.
Second, there was a marked trend for (female) referrals by GPs subsequently considered to be primarily about social relationships not to be sectioned.
Indeed, overall women referred by GPs showed greater evidence of neurotic, rather than psychotic, problems and a greater predominance of social problems.
GPs were clearly treating women differently and utilizing a different threshold or trigger for female referrals.
Finally, of those informally admitted, the majority of women were no longer contributing to the running of the household.
They were either disruptive, or withdrawn, or simply not around to be involved in familial or household responsibilities.
This is, of course, a not surprising by-product of some forms of mental illness and some women, despite this, might still argue that they were needed at home.
However, a sympathetic and patient GP would not have required great imagination to point out the disparity between such claims by these women and their ability to perform family and household tasks based on the evidence of their behaviour.
Certainly, if this was what was necessary, it did not subsequently prove beyond the wit of the ASW or psychiatrist to point this out.
As one of the ' specialists' involved in assessment subsequent to GP referral, ASWs clearly played a significant part in diverting these women from possible compulsory admission.
To what extent, however, does this reflect the work specifically of the ASW rather than the psychiatrist?
First, ASWs took responsibility for decisions diverting individuals from compulsory admission.
Where individuals referred by GPs were subsequently informally admitted or not admitted at all, in no case would the ASW have preferred compulsory admission.
There is, furthermore, considerable evidence in this study of social factors being of some significance.
This is most clearly the case in referrals where relationship problems were considered primary and where no compulsory admission took place.
Although we can not be categorical in relation to the cases in this study, other research has indicated that psychiatrists rarely trouble themselves with the social circumstances of individuals referred, and are rather preoccupied with evidence of mental disorder (Bean, 1980).
In this research social information, as opposed to psychiatric diagnosis, was not considered by psychiatrists to be central to their task, although they were aware of the influence of social factors on psychiatric symptomatology.
They therefore made little attempt, for example in relation to family history, income, education, and criminal history, to collect information on social circumstances.
Although this research involved the 1959 Act, unlike social workers, additional training has not been required of psychiatrists under the 1983 Act.
It is reasonable to suppose, therefore, that in relation to social factors influencing decision making, the ASW played an important part.
Third, a number of significant dimensions analysed reflect, specifically, the ASW rather than the medical role.
This is particularly the case in relation to social and familial factors considered, and the members of the individual's social network interviewed, as part of the assessment process.
Section 13 (2) of the Act specifically requires the ASW to consider ' all the circumstances of the case ' and these factors were specifically mentioned in the Memorandum (DHSS, 1983) as a guide to ASW assessment.
Where, for example, members of the patient's social network were interviewed as part of the assessment process, these interviews were specifically carried out by the ASWs.
Finally the research is based on information collected from the ASWs themselves.
Thus, for example, the definition of primary problem in each case comes from the ASWs, reflecting their perspectives on individuals assessed.
To that extent the results reflect the basis upon which ASWs made decisions, which frequently involved diversion of GP referrals from compulsory admission.
However, the ASW role was not wholly positive: ASWs were significant in diverting women from compulsory admission, and injecting a psychosocial perspective which appears to have contributed to this diversion.
Nonetheless, the significance of ASW consultation with male relatives, and a tendency to examine social factors which may have challenged GP definitions less than with other referrals are rather less positive.
These serious findings may be linked with other aspects of GP behaviour.
The first may be related specifically to GPs' training as medical practitioners.
Huntington (1981) identifies two significant elements of GP occupational culture: a biophysical (rather than psychosocial) frame of reference, and an action (rather than holding) orientation.
Although medical education may pay some attention to the social sciences in the pre-clinical curriculum, the highest priority is still overwhelmingly given to the so called ' hard sciences'.
This bias is reflected in practice, where medicine's primary concern is with the biophysical consequences of diagnosis and treatment, not its social causes and consequences (Freidson, 1970).
The result is an emphasis on the subject matter of general practice as a biomedical phenomenon, concomitantly ignoring or minimizing the need for psychosocial knowledge or skills.
This, Huntington (1981) thinks, is allied to an action orientation: a predilection for ' doing ' traceable back to teaching hospitals, where the dominant image of the emergency prevails.
The legacy in the GP is a tendency to take action for its own sake  doing something is better than doing nothing (Freidson, 1970).
The potential for ill considered actions is great with sections.
Viewed from the subsequent assessment, GPs mostly correctly identified individuals with mental illness.
However, many were referred for compulsory admission where they were subsequently not admitted, or admitted informally.
In the latter case, could the GP not have discussed this with the individual first? why resort to possible compulsory powers?
Second, by resorting to a medical route (and a compulsory one at that) for dealing with problems of women, frequently having a major psychosocial element, their behaviour indicated a limited frame of reference for viewing these problems  a predominantly biophysical perspective.
Bean (1980), in relation to the 1959 Act, makes some points about psychiatrists which may be extended to GPs.
He suggests that their orientation is towards the perceived therapeutic needs of the patient rather than their legal rights.
This does not mean that legal rights are entirely ignored, but it does mean the scope for making decisions primarily based on illness is increased.
Sheppard (1990) has described a ' mental health orientation ' in section assessments, with which the behaviour of these GPs is consistent.
This involves the balance of emphasis placed on the mental health status of the patient versus the social risk they present (to the health or safety of the patient or protection of others).
A mental health orientation places greater emphasis on the former to the detriment of the latter.
This mental health orientation leads to a more pessimistic outlook on patient needs  they not only require treatment, but it is assumed that it must be compulsory  and there is a tendency towards a ' presumption of risk '  that the presence of mental illness makes the situation inherently risky.
This is consistent with the behaviour of these GPs: placing mental illness to the fore when the problem may be primarily social, and seeking a compulsory rather than voluntary route.
When ASWs made assessments, social factors frequently took a high profile, as social rather than mental health problems were considered primary.
Where relationship problems were primary no women were sectioned, whether or not a mental health problem was present.
The injection of a more psychosocial perspective, therefore, diverted some women away from sections, and even informal hospitalization.
There are also characteristics which show specific disadvantage to women.
It is the involvement of the ASW, and possibly the approved doctor, which rescues some women from possible compulsion.
Why should GPs show discrimination against women? and why emphasize married women with children?
There is evidence of sex role stereotyping by mental health professionals, and GPs in particular.
Broverman et al.
's (1970) well known study showed that sex role stereotyping placed women in a disadvantageous position.
Characteristics considered healthy for an adult (sex unspecified) were primarily those considered healthy for men rather than those considered healthy for women.
The healthy person was also the healthy man.
This leaves women in a catch 22 ' whether to exhibit those positive characteristics considered desirable for men and adults and have their ' femininity' ' questioned... or to behave in the prescribed feminine manner [ and ] accept second class adult status' (Broverman et al.,
1970, p. 6).
By implication, the healthy female is not the healthy adult.
Some characteristics Broverman et al.
(1970) found identified by clinicians as characteristic of healthy women were: being more submissive, less adventurous, less objective, more emotional, more excitable in minor crises  strange characteristics for a so called healthy person!
Yet these are reflections of popular expectations.
Penfold and Walker (1984) comment that a girl is trained to consider her primary achievement is to be a wife and mother.
She is led to believe that she will be happy and fulfilled if she serves others, devoting herself to the needs of her husband and children.
She provides for the physical needs of the family, she is its emotional organizer, mediator of feelings, with a responsibility to dampen and absorb confrontations and generally keep the family running smoothly.
The GP in this social world  far from being an expert  is no more than a layperson, with the same prejudices as others.
Barrett and Roberts (1978) showed how GPs tended to see middle aged women patients primarily in terms of their family, and men primarily in terms of their occupation.
Where a woman fails to perform her expected role properly the GP tends to see her ' either in terms of the conventional models of their own social background (she must adjust to her situation), or in terms of individual personal inadequacy (that she must be referred to the psychiatrist) ' (Barrett and Roberts, 1978, p. 46).
However, extensive research shows that the stereotypically female role of mother, as well as other social disadvantage, is actually associated with vulnerability to depression and broader distress (Brown and Harris, 1978; Richman, 1974, 1976; Moss and Plavis, 1977).
The relative emphasis on neuroses in female GP referrals is interesting in this context.
The problem appears in medicine's successful expansion of its empire.
Doctors are consulted on environmental, technical, religious, or psychiatric matters, which are frequently beyond the scope of medical education.
Since medical advice has the force of moral imperative, health and illness are concepts expanding to ever increasing parts of human existence (Zola, 1975).
The effect is the creation of doctors as social control agents.
Two levels may be identified from previous discussions.
First, there is advice maintaining an unrewarding and demoralizing existence.
Barrett and Roberts (1978) found GPs extensively ' smoothed away surface anxieties' (p. 42) and adjusted women to a life located totally in the home.
In so doing they legitimated and endorsed the status quo, and fulfilled an ideological function of agent of disguised social control.
Doctors frequently used their medical-moral language to offer a set of prescriptions based ' on the conventional wisdom of their own social milieu ' (p. 42).
The second level of social control relates to the use of drugs.
The use of psychotropic drugs is a pervasive dimension of the medicalization of problems of living.
These drugs are prescribed in vast quantities, and far more frequently for women than men (Cooperstock, 1978).
Valium is a case in point (Koumjian, 1981): an extensively prescribed drug to deal with anxiety and stress.
It can, however, have severe withdrawal symptoms, including dizziness, nausea, vomiting, panic, anxiety and depression.
Valium prescription is frequently related to an individual's social functioning: the person's ability to live up to social ideals, social expectations and the physician's judgement are factors in a pseudo objective diagnosis.
Valium, therefore, is frequently prescribed for problems such as loneliness and marital discord, the latter interestingly considerably more evident in female than male section referrals in our study.
Valium use, furthermore, focuses on individual malfunctioning: social and economic problems are dealt with in a framework of a medical model of relief of individual distress, alleviating symptoms rather than operating in a social context which may require familial or wider social change.
Indeed, the conditions of medical practice support these approaches.
The lack of emphasis on psychosocial factors in training leaves GPs without sensitivity or skills in these areas.
Second, lack of time is a factor.
There is clear advantage  in terms of time use  in prescribing drugs rather than engaging in more time consuming counselling, or confronting the social origins of the problems.
Anyway, GPs are not trained for either of these.
Finally, advertising of tranquillizers has provided an explicit message: to use drugs in controlling conflict, for example producing a less demanding and more compliant patient, or for dealing with life crises or social problems.
A third, and more sinister, level of social control has been identified here: the use of compulsory admission.
This takes medical ideology to an extreme point.
There is an unequal knowledge relationship between GP and patient  the GP is expert and ' definer ' of the patient's problem.
The balance of power is also unequal in that GPs exert pressure for compulsory admission of women who could otherwise be admitted informally.
This involves a medical solution in so far as GPs seek medical care in a hospital environment.
Finally, such medical care will generally involve invasive drug therapy.
Our evidence suggests section referrals may be used by GPs as an extreme form of social control of women.
It also indicates that specialists, particularly ASWs, play a significant role in preventing unnecessary sections, hence guarding women's civil rights.
Even they, however, were not unaffected by patriarchal assumptions.
We have here further evidence of the need for ASWs to assume a social risk rather than mental health orientation (Sheppard, 1990).
ASWs are, by their training, the best equipped to examine social dimensions of section assessments.
Given the apparent pervasiveness of GPs' behaviour, it is vital, given their primary training in the complexity of the social world, that, first, ASWs are aware of issues of gender and sexism, and second that they are prepared to challenge medical power threatening women's civil rights.
