It had to be something that was thought through carefully, done properly, and with a sound foundation. The programme had to have a holistic approach and to include developmental/awareness work, groupwork, one-to-one work, and a helpline. This would require a mixture of staffing and the recognition of difference within the service offered. The implications for resources raised by such a programme were acknowledged. People involved would also need to be highly skilled and very committed.

Women’s organisations experienced in domestic violence work needed to be involved at all levels, including in management, design and delivery.

Careful thought needed to be given to the venue/building since the stigma attached to it might discourage many men. There were particular issues for some minority ethnic communities.

It had to have the involvement of all key agencies and to develop effective inter-agency systems but, at the same time, be cautious about who was involved.

Support and safety for women and children needed to be given priority and to be in place – some respondents suggesting access to emergency accommodation for women and children or panic buttons linked to the police. The importance of being realistic with women was emphasised.

Any initiative had to be properly funded and not create competition for existing organisations.

The voluntary nature of the programme (seen as very positive generally) could pose issues about maintenance of attendance.

Issues of ethnicity and equality should not be an add-on but an integral part of the whole. There was a view that there should be one overall project within which there was a recognition that services needed to be delivered differently to different communities. It was suggested that any new initiative should employ developmental workers to work with communities.

While perpetrator work was not an area of expertise for the voluntary sector in Coventry, it was felt that the programme should be led by a voluntary organisation, though with the involvement of other agencies.

All the pro-feminist projects interviewed had been established in an inter-agency context and most continued to consider this essential, despite the time and effort it consumed. Bodies were represented on the management or executive committee of a project because they supported the work, made referrals to it, helped to fund it, or all of these. One of the key attributes of a committee member was seen as the ability and willingness to attend regularly so as to carry forward the work.

Organisations which recurred on virtually all perpetrator project management committees included the police, probation, social services and Women's Aid or other local women's organisations. A programme serving a clear-cut locality (one city, as opposed to the whole of London, for example) typically also involved the local council. Bodies which emerged from the study as also important to think about include those with responsibility for primary health care and mental health. Individuals with private sector experience of financial management could play a useful role. There were also organisations of particular local relevance. In Coventry, this would include those representing minority ethnic populations.

Links with any local domestic violence forum or partnership were considered crucial. The potential project, if established, would need to become an active member of relevant local umbrella groups.

It was seen as necessary to have a formalised agency commitment at senior management level if resources or staff time were being contributed to a project. Projects had prospered less, particularly over time, if they were dependent on individual interest and goodwill. It was equally regarded as crucial for all the agencies involved in managing a project to agree about, and to be accountable for, its purpose, philosophy and direction. Agreed definitions and priorities needed to govern the work undertaken, with the paramountcy of the safety of women and children at the centre of these; men come to a project to be challenged, not counselled (with the exception of TEMPER).

Protocols were important. For example, an arrangement whereby the police were informed of the address from which a man was attending the group, and the project was notified of any call outs to that address, was not foolproof in one project because some officers had fears about data protection. Yet the agreement a man signed when he joined the programme included consent to his details being communicated to other agencies. The bottom line of a group for non-convicted men was said to be that all criminal justice agencies must agree that it would never be used as an alternative to prosecution or as any other form of diversion. No referral should be accepted while a prosecution was pending, lest it influence the outcome.

Definitions

An agreed definition of domestic violence, rooted in gender awareness, was seen as an essential starting point. All interviewees agreed that groups should be confined to those who had been abusive to a partner or ex-partner. Although all interpersonal violence should be challenged, for example in other family relations, the groups could only work where there was a tight commonality in the men's behaviour and attitudes.

Venue and location

It was the norm for voluntary sector perpetrator projects to have their own base, in rented premises. As the work is stressful, it helped the workers if they had dedicated space for their own offices as well as for the groupwork sessions, free of interruptions. One worker commented that it was important to 'keep grounded and focused'  and that the men tried hard enough to deflect the workers from challenging their behaviour, without other distractions such as having to fight for space in a multi-use office base. Also, in borrowed space, the nature of the parent organisation had sometimes conveyed the wrong messages; for example, this programme had been seen as a probation project both by referrers and referred, yet it was not.

It had been of great assistance where the local council had made premises available at reduced cost, for example in a hard-to-let building. The area of the City in which the group was held affected the attendance to some extent (different socio-economic and ethnic groups were attracted in different areas) but did not, we were told, make the group unviable whatever area it was in, provided it sought city-wide referrals. TEMPER has been accepting Coventry referrals but would probably concentrate more on Warwickshire if a new group was established in Coventry. Referrals from Coventry had, in any case, been steeply declining.

It was helpful for the men to have an entrance which was not too public as they tended to be hesitant (anxious, ashamed, ambivalent) about attending. If any women were interviewed in the same building, however, there were different considerations, with good lighting, visibility and safety over-riding other factors. A project with its own entrance preferred this on grounds of confidentiality and the avoidance of anxiety to other occupants.

Those who had experienced both kinds of arrangement said that they preferred to have the women's workers based in the same premises as the men's workers, as opposed to located elsewhere. It aided effective communication and integrated the work more closely. Wherever the workers were based, the women's work tended to take place elsewhere. There were advantages in running women's groups in a multi-use building which did not identify the women as experiencing domestic violence. One project, for exampe, used a building alongside a playgroup, luncheon club and so on, but with a separate entrance. There had to be good childcare facilities where women's groupwork was undertaken. If women did come to workers' offices for one-to-one interviews, it was normal not to have male and female service users on the premises at the same time or, if this was inevitable (where the woman worked and could only be seen in the evening in a project running men's groups every night, for example), we were advised that she should never be there at the same time as her abuser and only with her full understanding and express consent. However this was managed, it was considered essential to have good security installed on all premises that were utilised for domestic violence work.

Workers

Respondents said that workers recruited to the programme needed to have relevant experience topped up by initial training and consultancy. Working with domestic violence perpetrators is skilled, demanding and difficult work. It is possible to do more harm than good and it is dangerous for people to be 'learning on the job'.

One project interviewed had started life with probation officers and social workers seconded as the workers. This had laid the project open to the vagaries of change in policy priorities (both locally and nationally) and had caused some problems, including a current insecurity because of the shift in probation towards working only with convicted men.

Other projects employed designated workers and had succeeded in attracting dedicated, committed and skilful people. One respondent remarked that recruitment could be a problem, given the high level of experience that was required. We were also told that it could assist in handling ethnicity issues in the group if the workers were not all white, even if they were not necessarily of the same ethnicity as group members.

Referrals

Those interviewed typically worked both with convicted and non-convicted men, but were aware of plans within the National Probation Directorate that might threaten this joint model. None of the projects we talked to thought there would be any difficulty in attracting sufficient referrals on non-convicted men. All spoke of having too many men for the places available: 'There is never a shortage. The problem is always dealing with the numbers'  (groupworker). The number of groups running at any one time had risen for some, to try and cope with the demand.

Referrals of non-convicted men were not less serious in nature overall; they included worrying cases of psychological abuse and dangerousness. Some men did appear to come through the group at an earlier stage in the abuse, however. One respondent found them sometimes more readily aware of emotional abuse than convicted men, the latter having progressed through the criminal justice system largely for physical violence.

Practice varied as to whether a waiting-list was operated. One project had found a high drop-out rate amongst those who had been on such a list and now preferred to ask men to re-refer themselves after an interval of a month or two. Another retained men's interest by running a monthly introductory afternoon to acclimatise them to the project while waiting. Projects varied in whether they considered that waiting to get onto a group was good or bad. One project minimised the wait so as not to lose people, while another considered it an advantage that motivation was tested during this time.

More than one project mentioned GP surgeries as a good source of partner/self-referrals. Either the man or his partner might have approached the doctor for help with depression or other problems associated with the violence and the group might have been mentioned or there might have been a poster on display. Another project said that health professionals, including in mental health, had themselves referred only a few men.

Other important sources of referrals included men's partners, the police (notably DVUs), probation, social services (the latter said to vary according to local levels of awareness) and Relate. One project had had an excellent response following newspaper coverage. Another had found that men referred through Relate did not have a good understanding about the challenging model involved in the programme; these men had expected something more along counselling lines and had not been prepared to accept such a rigorous focus on their own behaviour. Clearly, then, clarity amongst referring agencies as to what was involved was essential. Once a group was available, agencies could place higher expectations on men to change their behaviour. In the case of Relate, for example, its national policy is not to work with a couple where there is current violence; in one area, because there was now a perpetrators' programme, it was possible for Relate counsellors to give the men the leaflet about the programme and expect them to go through the group before returning to work on the relationship.

Cases from social services constituted a third to a half of the referrals in several groups we interviewed or heard about. Social services referrals tended to be made where there were current child protection concerns. The man's attendance at a group might be written into the care plan and, consequently, feedback would be required in the form of written progress reports. Reports might also be required for child protection conferences and court proceedings. A service level agreement could help to offer some security of funding or, alternatively, charges might be made on a case-by-case basis. In Peterborough, where DVIP had recently established a programme, Social Services were holding a fortnightly panel with other relevant agencies to consider all cases with a domestic violence element where workers were unsure what to do. Women's Aid was able to offer services in some cases and DVIP offered the men's programme where appropriate.

There was agreement amongst those interviewed that, whoever made the original enquiry (another agency, the man's partner), he would not be accepted for assessment for the programme until he rang in himself to request an appointment. This was seen as providing at least a small indication of motivation. There was such a problem with 'no shows' that accepting referrals on trust, without the man's own involvement, did not seem sensible. Also, he might be making promises to his partner or others that he did not in fact intend to carry through.

At assessment, limitations on accepting men into the group might relate to mental health issues, alcohol or drugs, or lack of real motivation to undertake the work. Close working links with community mental health teams, alcohol and drug teams and projects, GPs and others were important, both to assist in assessing whether a man was likely to be able to benefit from the programme, and in referring on those men who needed to address their other problems before joining the group. It was seen as a question of balance. If a man had delusional ideas or was using stimulants heavily, he was unlikely to be able to participate effectively or to face this very challenging kind of group without the crutch of an artificial protection against emotional pain. If, on the other hand, the man was actively addressing his problem and was able to get to the group sober and drug-free, he might be accepted.

Projects agreed, too, that the initial assessment was crucially important in screening referrals and that, the more skilled workers became at assessment, the more completion rates could be improved: 'We were not very selective at first'  (former groupworker). A common feature, for example, was for a man to seek help at a point of crisis in his relationship, often to persuade his partner to stay or return. He was seen as being more likely to succeed in the programme if he had thought about it more deeply and if he had other reasons to want to change.

Well established projects did not necessarily need ongoing publicity, such was the volume of potential work. Groups were typically running with about eight or nine members, perhaps recruiting slightly more initially (ten or twelve) to allow for drop-outs.

Non-completion rates

All respondents referred to high non-completion rates. In one mixed group, for example, 50 per cent of men on probation but only 15-20 per cent of non-convicted men who had got as far as attending the first evening completed the whole group. Another project said that non-convicted men were more motivated at the beginning but had fewer sanctions to keep them attending. Everywhere, many men pulled out during assessment or were listed as 'no shows' at the first meeting. Keeping men in the group for long enough to do any real work with them might therefore be the issue for a group of non-convicted men. They had far greater flexibility about how much of a group they chose to attend. They might drop out early on, or cease attending near the scheduled end of the group (two respondents said this was common) because they felt they had done enough, whereas convicted men were required to see it through. On the other hand, two respondents said non-convicted men were more likely to choose to undertake follow-up work. To summarise, attrition rates in the proposed group are likely to be very high but there may be good outcomes for those who persist.

Ideas for improving completion rates included: more rigorous assessment of motivation, operating a waiting-list to test motivation beyond initial remorse or ultimata from partners (though another project emphasised keeping in touch to hold men’s interest), charging men for attendance (see section on funding) and including a block week early in the programme to build group cohesion (see below).

Practical ways of mitigating the effects on staff of attrition in numbers included a drop-in afternoon to give information about the group, thus avoiding multiple missed appointments with individuals, and a second entry point halfway through the group so that additional men could be recruited to take the places of those who had left. The latter idea had the  additional advantage of mixing new members with more experienced ones who could challenge denial of responsibility for the abuse and minimisation of its effects.

There could be an emotional impact on staff when non-convicted men dropped out of the group and disappeared because there was no follow-up or tracking of them as there was when a man was on probation. This could be worrying for the groupworkers who heard nothing further about what had happened, unless the woman kept in touch with the parallel service.

Some practice issues

Most of those interviewed were working to the minimum standards and good practice guidelines (Respect, 2000) developed by the National Practitioners' Network, now relaunched as Respect, (Appendix 1). TEMPER was one obvious exception to this. The respondent from there did not agree with the Respect ethos. There was evidence from interviews with potential referrers, however, that they would prefer a programme which did meet these standards.

Some projects offered more than initial assessment and groups. One  favoured a period of one-to-one work before the group. This was used to move the individual beyond initial resistance by helping him to open up about what he had done and to start taking responsibility for it. The worker, in turn, was able to begin building up a detailed picture of how the individual created and escalated abuse, and did not have to deal with total denials in the group where the level of challenge became greater. The endpoint of the one-to-one work was an individually tailored safety plan, whereby the man was aware of how he thought, reacted and interpreted, and had strategies to behave safely. Offering one-to-one work meant that men could engage in some work instead of just being on a waiting-list. It was important not to let them think this was the whole story and that they 'knew it all' before the group, however. Having the one-to-one option was important for those who could not handle the group for practical or emotional reasons, but was seen as a 'second best' in terms of a dynamic of change. A further use of one-to-one work was as a 'refresher' for men who recontacted this project a year or so after finishing the group, because they recognised that they were slipping and wanted to do some more work.

Another variation on the traditional groupwork model was in Leicester where the group spent a block week together (four and a half full days of intensive work involving inputs from survivors, child welfare professionals and an interactive theatre company) before switching into twice-weekly and then monthly meetings. This was said to aid group cohesion, and potentially completion rates, as well as the intensity of the work that could be achieved. Participants were said to value the week as powerful, positive and energising, while staff regarded it as a strength of their model of working.

Groups needed to be long. One groupworker reported that his project had graduated from running three month-long groups to their present 15-month model, encompassing 10 one-to-one sessions followed by 48 weeks in the group. Of the former, he commented that 'It felt like the work was just starting' since 'You're talking about making substantial changes in their whole philosophy to life' . He placed this in a context where the man's day-to-day environment did not support the change and only the group was really there to reinforce the progress he might be trying to make. Despite the length of this programme, it was not unknown for men to opt to come back and go through again or, more commonly, to have some individual refresher sessions. Men were given feedback when they completed the group, which sometimes stated that they ideally still needed to do more work. American research (Gondolf, 1995) suggests that practitioners can become very skilled at gauging what levels of change men have achieved in a group.

Even though participants had not been convicted, groups still imposed sanctions for non-compliance with rules contained in the signed agreement, for example missing sessions or behaving unacceptably in the group. These included exclusion from the group and a requirement to repeat parts of the materials. (Valid absences could be dealt with by making up those sessions subsequently.) There was awareness that excluding a man from the group was an ironic sanction since he then escaped being held accountable and his partner was left to cope with his dangerousness. Convicted men would be handed back to the criminal justice system at this point and could have their order breached and/or be resentenced but non-convicted men are left to their own devices.

Ethnicity and anti-racism

One of the projects interviewed was operating in a predominantly white area and had had only a handful of Black participants, with no clear policy on how to work with them. When a Philippino man had attempted to attribute his behaviour to culturally accepted gender norms, for example, the workers had not been certain how to deal with this. A probation officer working in perpetrators' groups in the West Midlands, on the other hand, was very firm that 'culture isn't a justification for violence'. He and his colleagues worked hard to challenge such excuses and to retain the focus on the man's responsibility for his own behaviour. This had been an issue with some Asian men who had said that the workers did not understand their culture. However, the workers maintained a focus on issues that are the same in all cultures (though they are manifested differently) around who has authority, who has privilege, who has power.

Other respondents talked about the importance of white workers naming racism, both at assessment and in the group, and making links between different forms of oppression and prejudice as a learning tool, while not allowing men to use their own experience of oppression as an excuse for oppressing others. At the same time, there was sensitivity to the resonances of a professional white man telling a working class Black man that he was an oppressor. Keeping the focus on the detail of his behaviour was seen as the answer to this.

One project tended to hold on to referrals of Asian men until there were sufficient to make them a majority or sizeable minority in a particular group and certainly more than an isolated one or two - say, four or five out of ten. This made it easier to deal with ethnic and racial issues, to illustrate that culture is not homogenous, and to make comparisons between different life experiences. A different problem then presents itself, however, in that Asian men might not wish to be visible to members of their own communities and might prefer to be small in number within a mixed group. This was discussed at assessment, as was the question of whether the man is comfortable talking about racism with a white worker.

Where English was not the man's first language, there had to be a judgement taken as to whether he could participate adequately in the group, perhaps with additional one-to-one help with written materials. One respondent felt that it might be worth struggling a little, and perhaps slowing down the pace to ensure comprehension, rather than denying some men the opportunity to participate. However, only one project talked about offering work in another language. Even then, this was not the groupwork programme and had to be undertaken through a worker who was not a domestic violence specialist.

Reaching minority ethnic communities had necessitated particular efforts at outreach in order to raise the profile of domestic violence. This could be a slow process but we heard that it could result in slowly rising numbers of both men and women from minority communities using the project's services. Even work focused on just one minority ethnic community could improve practice across the board since awareness levels rose within the project and all agencies and communities observed the visible efforts being made. It could also make a real difference to the level of referrals. For example, Everyman raised awareness amongst the local community in Brixton and managed to achieve 25-30 per cent of referrals of Black men. This was done with posters in community centres and talks to local groups. Leicester Against Domestic Violence was beginning to use Asian television and radio but had not yet recruited more than small numbers.

Age

Groups spanned the 18+ age group. Men in the West Midlands groups tended to range from their early 20s to their 40s. DVIP, which reported a few men in their 50s and the bulk from their mid-twenties to 40, had found those in the 18-22 range to be the hardest to engage, perhaps because they did not relate to the idea of being attached in a stable relationship. There might be something to be said for a separate group for younger men but not as a first step since it might well be harder to run. Other than this, there was no problem in putting all ages together.

Socio-economic background

Probation referrals were said to be predominantly of working class men. Other referrals, however, came from right across the social spectrum, ranging in London from executives working in the City to unemployed men. All groups taking non-probation referrals mentioned men from amongst the professional classes and more affluent social groups. This is borne out by the literature on the prevalance and incidence of domestic violence which shows no appreciable class differences. As one groupworker remarked: 'The group is a great leveller'.  In terms of group process, the men got on well together and there were no apparent tensions resulting from the social class mix. The sites used for publicity and awareness-raising about the project, as well as the project's own location had affected who had come forward. Advertising in GP surgeries, for example, had reached a good socio-economic mix.

Disability

Established projects had worked with men with disabilities and had few issues to report about this, beyond ideally needing accessible premises. Reading ability was relevant, we were told, because men were expected to do written tasks between meetings and received hand-out materials, as well as having the written agreement covering participation in the programme to understand and sign. All of this was said to be handled by undertaking additional work with individuals, where necessary, outside of group sessions, reading over the materials together to make sure that participants fully comprehended them. If a man could not write, he was seen half an hour early in one project, to dictate his homework ready for the group. If an individual still could not participate fully, he would be offered one-to-one work instead of the group. Literacy problems were said to be quite common in the groups so this extra time has to be allowed for when calculating staffing levels.

Sexuality

There was widespread agreement that it would not be appropriate to include gay men in a predominantly heterosexual group of abusive men. The issues in same sex and male/female violence, though they overlap around the unacceptability of interpersonal abuse, are not identical and it was also considered important to be able to retain the focus in the group on men's violence to women. Hence it was thought more helpful to undertake one-to-one work with gay men who had abused other men (or theoretically to offer a separate group, if numbers were sufficient, with a specialist focus). Homophobic and heterosexist attitudes still needed to be challenged in the groups, of course, especially as abusers ground their behaviour in particularly rigid gender attitudes.

Evaluation

Several of the projects interviewed had had external evaluations conducted, as well as gathering their own data. There was agreement that feedback from partners and ex-partners was the most important source of information on whether men had changed. This is fully borne out by the available research. Partner feedback covered whether the woman had noticed changes in her partner's behaviour towards her and also whether his attitudes towards women had changed. Feedback from women could be difficult to obtain since some had separated and others were reluctant. Feedback from men was seen as unreliable because of minimisation and denial, while reconviction rates revealed only a fraction of repeat attacks. Data on attendance and completion rates were also being gathered.

Women's support services

It was seen as crucially important that local women's organisations should be involved from the very beginning. Examples were given of formal accountability to women, and of women being involved in the management, design and delivery of the work. Respondents stressed the need to put an effective level of resources into women's services, to maintain close links with other women's organisations and to integrate the men's and women's work closely within one project.

One project had moved from having the women's support work separately located and monitored, to both men's and women's workers sharing a building, having joint supervision and maintaining a two-way flow of relevant information. They had found that this made it far easier to help individual women keep safe and to challenge men with renewed abuse that they were not disclosing. Specific safety strategies were worked out through joint case discussions and through raising emergency concerns, with a high level of mutual trust and respect between the workers in both halves of the project enhanced through joint supervision and training. All budgeting and charging for perpetrator work fully integrated both elements of the service (i.e. for both perpetrator and partner) so that external agencies and funders automatically bought into the integrated service. There was also some additional funding for the women's support work from local authority and crime reduction sources, to top up joint government and charitable funding. Staffing needs were felt to be similar in both halves of the project.

They described two key aspects to the work with women: first, contacting the man's partner or ex-partner to tell her about the group and its ethos (notably that the abuse she had experienced was the man's responsibility and not hers), telling her that there was no guarantee that he would change, and keeping her informed about his progress through the group so that she could make realistic decisions. Secondly, some projects offered services aimed at empowering women. There was a high level of proactivity involved in contacting each woman and ensuring that she was given 'at least information on every move he makes' (groupworker). This covered the fact that the man had joined the group, notification if he failed to attend, or if he dropped out or was excluded. The Respect standards cover in detail what women should be able to expect, including in terms of safety and confidentiality.

Around the country, we heard about a range of services offered to women in connection with perpetrators' groups, including: telephone, face-to-face and groupwork advice and support, a helpline, a joint project with an outreach scheme, group information sessions for women prior to each group starting, regular drop-in sessions, themed workshops and empowerment groups. Many of the women contacted had not used any other women's services prior to this and some could be in crisis. They might need referring on, or someone to act as an advocate for them. There was an educational element to the work as well as emotional support, so that women could understand more about domestic violence and the risks it posed to them and their children.

We were told that linked support work required specialist training. It was different to refuge work or other services where it was the woman who had initiated contact. There were very specific issues involved in the women's worker making the first call to a woman safely (what to do if the perpetrator answered the telephone or was present in the room), and in understanding the men's programme and how the man's progress was being monitored through it. There was also a focus on safety planning with the woman and on her having space to consider all her options realistically.

Some women were known to have used the opportunity of the man's participation in the group to opt out of the relationship and those interviewed considered this a success because the woman was able to do so safely while her abuser's behaviour was being closely monitored. One groupworker cited as arguably her greatest success a man who had probably not changed at all but where close interagency working had effectively managed the risk he posed so that there had been no further harm to his partner.

Children

Work on parenting was an essential part of some of the groupwork curricula we heard about, both because it is common for perpetrators to expose children to danger and distress (involving children in their abuse of women and often directly abusing children), and because they typically have much to learn about responsible parenting and age-appropriate, non-violent discipline. Where fathers were still actively involved with their children (particularly where social services referrals were concerned), parenting was a vitally important part of the agenda, though it might not feel relevant for all participants. Men without children could contribute in the group by thinking about their own childhoods and the way their fathers treated them. Respect standards require that all projects have a clear child protection policy.

The aim of women's support services, to increase woman's safety and confidence, was considered also likely to improve the welfare of her children. Childcare facilities were provided when women's groups were running.

Children also have their own needs. These can be divided into safety work and recovery work. Most North American groups for child witnesses of women abuse combine both, but will not take children until the violence is thought to have ended. In the UK, almost the entire focus is on child protection and only a handful of projects (other than through refuge childwork) offer direct services to children who have lived with domestic violence. The Children's Society in Birmingham offers groups and there is a specialist children's counselling service in Warwickshire. Meeting Coventry's needs in this regard could ideally be the subject of a separate development. Certainly, a complete raft of services would offer intervention with men and also direct work with children alongside a full range of services for women.

Finance

Funding was thought likely to be the most difficult problem facing a potential group in Coventry - both raising the money initially and sustaining it over time. Voluntary sector projects talked about being prey to insecure funding and finding themselves distracted into fund-raising every two to three years.

The total cost of running a project, encompassing both a men's programme and women's support work, based on three actual sets of costs from different parts of the country, ranges from £90,000 to £115,000 a year (see Appendix for the budget break-down of the most detailed and most expensive of the three). Staffing is the largest single item in all three, at around £60,000 to £80,000 per annum. All have subsidised premises. The most expensive of the three projects raises its income largely from a combination of £87,000 in grants and £31,500 through a partnership contract.

Income

Funding sources we heard about varied between projects, within the same project over time, and sometimes between the men's and women's work in any one project. They included local statutory agencies such as social services, probation and health (health authorities and trusts, health promotion bodies), particularly outside London where there was greater co-terminosity between projects and local authorities. Grants had come from the Lottery, Sure Start, diocesan funding and the Chuch Urban Fund, amongst many other sources. Insecure funding and the need to devote time to writing grant bids were a problem for projects based in the voluntary sector. In London, it was proving extremely difficult to get men's work funded. Women's services were somewhat easier because they fitted into a more accepted image of domestic violence work. One respondent thought that difficulties and insecurity in accessing funds would be the biggest threat to the survival of a new project.

As was mentioned above, under 'Referrals', a service level agreement is a possibility. Otherwise, agencies wishing to refer men to a project can be charged a fee. An assessment for suitability to enter one particular programme, for example, was charged at £300 per individual, with attendance at 32 sessions costing just under £1,500. A 15-month programme charged £2,500 per perpetrator. (It should be noted that TEMPER cites a unit cost of only £250 per participant but that it does not operate to Respect standards with, for example, a much shorter programme, an ungendered analysis and only the barest of information to partners.) We were firmly warned by well established projects that undertaking risk assessments required several years experience and that this should not be regarded as a suitable source of an additional income stream in the short term. In contact cases, for example, the report had to be defended in court and the author could be cross-examined on his or her credentials for undertaking the work.

Another possible source of funding was to impose a sliding scale of charges on men attending the group or one-to-one sessions. These might start at zero for unwaged men, through to full cost (i.e. the same as an agency would pay to send someone - around £45-£50 per session) for those earning £30,000 and above. Charging is a contentious issue. Arguments we heard in favour included increased motivation and investment in the group: 'It's psychological to value what you pay for' and  making the man pay for what he has done and for the effort being put in with him: 'As a worker, you're struggling with an ungrateful bloke on £40k a year'. One respondent said that wealthier men had sometimes offered to pay. Arguments against charging a fee focused on some men's inability to pay, though unwaged men were not charged at all in one project, and the possible impact on the family (if he used money from the household budget to pay the fees), though women using parallel services did not seem to have reported this. A moral point was raised about convicted perpetrators getting a groupwork service for free while others had to pay. One practitioner wondered whether this might lead to a two-tier service.

It is unlikely that charges would contribute significantly to funding. One very busy project, running nightly groups, estimated the total income from fees paid by perpetrators at only £7,000 per annum. Certainly, collecting fees was said to add considerably to the workload. Good advice from one respondent was to charge three months in advance. This reduced the administration costs, might prevent 'no shows' and increase motivation to complete, and prevented arguments about the man avoiding paying for any individual session when he might have been ill or had had some other reason not to attend.'I don't want that as part of my role - "I was on holiday that week" or "My car broke down". It was seen as essential that an administrator should handle invoicing.

Outgoings

It was emphasised that there are no cheap options in this work. Outgoings mentioned to us included staffing (practitioners, administration, possibly a manager), payments to sessional workers, premises, office equipment and running costs, travel, paid supervision every week or fortnight (@ £80 per hour in one project), training (especially in the first year), recruitment, catering, publicity and outreach. Even low rent premises may generate cleaning, energy and insurance costs. Groups and childcare need materials. Staff have to go to conferences and national network meetings. One respondent said that one-off costs like publicity leaflets had often been absorbed by a participating agency. When a funding bid had meant that costs were no longer hidden but had to be exposed and totalled, the sums involved looked astronomical; then, when the bid had failed, people had lost heart in their ability ever to raise such a large sum. Other projects had kept applying to various bodies, but no one found it easy.

Two projects put the 'bottom line' to get through the first two years at three salaried workers, one each for the men's and women's work and an overall co-ordinator, plus sessional workers (who could perhaps be seconded) and administrative support. Volunteers were seen as a possibility for sessional work but only if they were able to offer consistency and sufficient time, attended all training and supervision, and met a high threshold of skill and awareness of domestic violence issues. Pay levels for staff needed to be high enough to recruit sufficiently experienced staff, with realistic advertising and relocation costs included in the first year's budget. It was said always to be harder to recruit to men's programmes than to women's services, probably for the obvious reason that the work is challenging, stressful and carries fewer rewards. It requires a high level of skill already to be in place and co-workers have to be able to rely totally on each other. Most groups ran with two co-workers. West Midlands Probation used three.

Practitioners explained that they needed to be allowed sufficient time for preparation, debriefing, supervision, undertaking and offering training, promotional work for the project, inter-agency liaison, taking referrals, intake work, one-to-one work and administration, as well as running the groups themselves. It was important that they did not become embroiled in fund-raising, management of the building, and so on. Some established projects explained that they had a manager to undertake this role, together with much of the external liaison.

Local work with convicted men

A representative of the West Midlands area of the National Probation Service  described running 14 domestic violence programmes a year, each of 75 hours over nine months. Respect standards were adhered to. Theoretically, a group for convicted men could be run in Coventry if a majority of participants lived there but, in practice, this did not happen. There were two from Coventry out of eight in one current group, for example. Men did not appear to object to travelling to Birmingham (their fares were paid), although it could have been an unacknowledged reason for drop-outs, but there was some feeling that the distance might make probation officers in Coventry assume 'there's nothing available in Coventry' when, technically, this was not the case. Whatever the reason, numbers from Coventry had been relatively small. There had not been any men from the Asian community, at least in recent memory.

No problem was seen with a man, having gone through one of the actual and potential groups, later going through the other (i.e. as non-convicted and later convicted, or vice versa), provided that the criminal justice system did not blur the boundaries. It was important to ensure that the establishment of a new project for non-convicted men did not result in any man having charges against him dropped or reduced. The project itself would need to be very clear about these boundary issues, as would the police, the Crown Prosecution Service and the courts.

Probation West Midlands was funding four partnership projects to offer support and empowerment work with women. A combined sum of £50,000 per year to provide this work was going to women's organisations in Birmingham, Wolverhampton, Dudley and Sandwell, and the three domestic violence support services in Coventry. There were still said to be issues of mistrust as to whether perpetrators' groups could be effective, so communication was being worked on, according to a Probation respondent.

Potential criticisms (with possible responses in brackets)

Projects had had to overcome a certain amount of hostility, at various stages. Issues that recurred in the interviews are listed below, together with the responses projects had tried to make or might have made according to their working ethos.

i) There's too much focus on men. (Men are the vast majority of serious and damaging domestic violence perpetrators. It is to women's advantage if these men's behaviour can be changed. An associated service for women is an integral part of the project. There is an absolute commitment not to compete for funding with emergency and outreach services for women.)

ii) It is anti-men because it names perpetrators as responsible for abuse and is confrontational in style. (It' s a way of helping men to have better lives - therefore it's not anti-men at all.'  It is perfectly possible to be challenging while also acknowledging the pain in a man's own life. 'We're not completely dismissive of him.' )

iii) What about women who are violent to men? (There are some but they are a small minority and they typically inflict less damage. Many are actually acting in self-defence. All violence in personal relationships is unacceptable but this project is specifically about changing men's behaviour.)

iv) What about violent sons, uncles, brothers? (This project will only work if, within the group, there is a tight focus on one kind of perpetrator. This need not prevent other kinds of work being undertaken by other projects, or intervention by the criminal justice system. Also, the project may undertake some wider awareness-raising and educational work over time.)

v) It's doing work the statutory sector should be doing. (No one has been doing it up to now. Probation is not funded to work with unconvicted men.)

vi) It will deflect men from the criminal justice system or dilute responses to them. (We have policies in place to ensure this does not happen.)

vii) It doesn't work. (It appears to work with some men. It can also help women safely separate from men who may not change. It is worth the effort, even for this degree of gain.)

viii) It costs too much. (It won't be effective if done 'on the cheap'. It can be funded by agencies who would, in any case, pick up the costs of continuing violence, damage and distress.)

ix) We should be focusing on child abuse. (A high proportion of men who are violent to women also abuse children. The project works closely with social services and has a clear child protection policy. The group includes content on responsible parenting. Some children have died in cases where the abuse of women was not picked up as a relevant issue. We need integrated responses to both.)

x) It's not culturally appropriate for black or Asian men. (Men's violence and abuse occur in all cultures and all women deserve to be safe. The programme takes anti-racism and cultural appropriateness seriously and seeks to recruit an ethnically mixed membership to the group, working closely with local minority ethnic communities.)

xi) It's not appropriate for gay men. (Gay men can be accepted for one-to-one work. We would support and assist the development of more specialist services.)

xii) Men will use the group to put pressure on their partners to stay or to return. (Every effort is made to give partners a realistic idea of the programme, including low completion and success rates. Groups can actually provide some women with a safe opportunity to separate. Each man in the programme and each partner being supported has an individual safety plan.)

xiii) The group should be more therapeutic. (That can be offered elsewhere. This programme is about changing behaviour.)

xiv) There is a risk that workers will collude with the men. (Male/female co-working lessens this risk, as does regular supervision. The whole ethos of the group is about challenging men to take responsibility for their behaviour.)

Overall, projects which had established firm boundaries and a clear ethos, grounded in the safety of women and children, were able to win the arguments and often got their erstwhile critics on board to support the work.

Summary and key points

A group for non-convicted men in Coventry would be feasible but would require a great deal of hard work to establish, concerted action to raise the funds, and a commitment to the highest standards. Otherwise, there was a risk of doing more harm than good. The following points were emphasised by respondents.

The Respect standards (see Appendix 1) were seen as offering an essential baseline for good practice.

Projects operated in an interagency context, reflecting referral and funding sources, levels of local support and involvement, and the mix of the local community. Women's organisations needed to have a say in the management, design and delivery of services.

Health was emerging as an important potential player within this interagency context.

Formalised agreements covering resourcing, definitions, overall direction and detailed practice protocols were essential, particularly in a changing policy climate.

The definition adopted needed to be gendered and it was important only to work with men's abuse of women since group effectiveness depended on conveying the clearest of messages to  perpetrators who had a shared set of behaviours and attitudes.

Projects needed their own office base and preferred to have both men's and women's workers located there. Much of the women's work might take place elsewhere.

Both men's and women's workers needed considerable experience and also specialised training.

Recruiting workers to men's projects was not easy.

No one reported a problem with attracting referrals of non-convicted men and there was little need for ongoing publicity, except with minority ethnic communities where active outreach was advantageous. Waiting times were more likely to be a problem.

Perpetrators had to telephone the project themselves, even if someone else initially suggested they be referred.

Social services referrals, within care plans where there were child protection concerns, might fill a third or more of the places in a group with fees being paid accordingly.

Close links were needed with drug and alcohol and with health/mental health services.

Completion rates were a problem in all behaviour change projects, with variable feedback for non-convicted men. In one project, those who did complete were more likely to be motivated to continue with the work afterwards. Sanctions for non-compliance were needed.

Initial assessment was essential in establishing motivation and suitability.

Perpetrator groups could be mixed in age, ethnicity, ability levels and socio-economic background. They were not suitable for gay men (unless there was specialist provision) but one-to-one work could be provided. Active attention to anti-racist practice, while also challenging cultural excuses for violence, were essential.

Partner feedback was the most important aspect of evaluation.

Women's support services needed to be an integrated part of the same project.

Men's groups needed parenting content and women's services needed childcare support. Children also had their own needs, requiring separate service development.

Finance was likely to be the hardest problem to solve. Charging other agencies and individual men for services both worked well. However, there was a need to raise two or three times as much again through grant funding. These services did not come cheap, with overall annual costs for an integrated service running at £90,000 or more.

It was essential to draw clear boundaries with the criminal justice system so that no man was accepted into the programme while a prosecution was pending, and there was no diversion from, or dilution of, sentencing.

There was likely to be a degree of hostility to the establishment of a men's programme but a commitment to high standards and to the paramountcy of women's and children's safety were normally a basis for a shared way forward.

The programme must be seen as one part of a comprehensive raft of services, with no competition for funding with emergency and outreach services for women.