Do domestic violence perpetrator programmes work in reducing violence and abuse?

No, says Julie Bindel.

Yes, says the University of Durham

Rehabilitation programmes for domestic violence perpetrators can work (12 January 2015)

The vast majority of men who abuse their partners stop their physical and sexual violence if they attend a domestic violence perpetrator programme, according to new research. The research, led by Durham and London Metropolitan universities, suggests domestic violence perpetrator programmes (DVPPs) could play an important role in the quest to end domestic violence.

Steel yourself or take a seat – Julie Bindel is absolutely right. I agree with her. Cherish the moment, even if we have come to the same conclusion from very different directions.

A lot of us have been sceptical of the effectiveness of Duluth model interventions for a long time. Whether or not one questions the theory underpinning the model, it remains the case that the published evidence for the effectiveness of such courses is sparse almost to the point of non-existence.

Yes, there are evaluations which show improvements in men’s (only ever men’s) abusive and violent behaviour, (see here ) but the big question was always: compared to what? Abusive behaviour is not constant. In most abusive people it diminishes with age, and many abusive people will reach a point in their lives when they will become less violent, less controlling, less abusive. Serious abusers tend to be referred for help (or self-refer) at a time when their offending behaviour is close to its peak. So if you have 1000 abusers and monitor their behaviour over a year or two, a small number might get worse but a much larger number will improve a bit while some will stop the behaviour altogether. That happens even if you don’t intervene at all.

This is really basic social science. We should not ask if a group of men on a perpetrators’ course become less violent.

We should ask if a group of men on a course become less violent than they would have done if they had not done the course.

Even better, we should ask if men on this course become less violent than they would have done if they had done that course instead, or served a different judicial sentence or whatever.

Now, go searching in the archives for that research and you’ll find the cupboard is pretty much bare.

To be fair, there are some very good reasons for that gap in the literature. It is an exceptionally difficult question to answer for reasons of safety, practicalities and above all, ethics. Evaluating the various court-mandated perpetrator programmes around the world is hard enough. Evaluating community-based treatment programmes, where most of the participants are there voluntarily and so might be presumed to be more motivated to succeed – or can simply drop out if they are failing to improve – well, such research is almost drowned in methodological issues.

So when the University of Durham announced they had demonstrated that such courses “can work”, (and annoyed Julie along the way), I eagerly asked the University for a copy of the paper and they kindly obliged.

It is the final report of a three year initiative called Project Mirabal, authored by Professors Liz Kelly and Nicole Westmarland. It does not appear in a journal but is listed as being published by and copyrighted to the University of Durham. There is no mention of peer review.

The introduction includes an exciting surprise. I will now quote from the paper with only a few minor edits for brevity and a few words of commentary.

**

In order to demonstrate that change was due to the DVPP intervention with men, the research design involved creating a matched comparison group of women receiving support about domestic violence but in an area where there was no community based DVPP.

Excellent!

Freedom Programmes were chosen for this, since they are widely available and work only with victims. The quantitative data collection followed the current orthodoxy of taking women’s accounts as the most reliable in terms of men’s use of violence and abuse: women whose partners were on a DVPP were designated the intervention group, and those on Freedom Programmes the comparison group. Data analysis revealed the two groups were comparable across basic demographics, length of relationship and baseline levels of violence and abuse. Where they did not match was that comparison group women were more likely to have children who had no contact at all with their father (40% in comparison versus 16% in intervention group). Where there was no contact, this was primarily because either the child or the perpetrator did not want it in the comparison group, whereas in the intervention group where contact was limited this was more likely to be the result of decisions by the family court or Children’s Services.

Well that’s not too serious -the groups are slightly different but that was inevitable…

Most crucially, we found that women in the intervention group were far more likely to still be with the man who had abused them: nearly half were together before the man started on a programme and over a third were still together 15 months on. This was the case for hardly any of the women in the comparison group – just 13 per cent at first interview and 9 per cent at 15 months.

OK this is more worrying, but they must have known this was going to happen from first interview and decided to proceed?

This finding suggests that women are in contact with Freedom Programmes and DVPPs at different points in the process of dealing with domestic violence. Thus whilst we do have comparison group data (which largely found there to be no significant differences in reductions in violence and abuse), the fact that they are not an equivalent comparison group rendered the comparative data difficult to interpret in a way where we could be sure of our explanations. For this reason we do not report this data here.

WHOAH. Stop right there…. the study found no significant differences in reductions in violence and abuse between the comparison group and the intervention group?

Let that sink in for a moment. The study found no significant differences in reductions in violence and abuse between the comparison group and the intervention group!

Let us be charitable and go along with the authors’ interpretation. It was simply not possible to compare the intervention group with any kind of control group. (The more cynical interpretation would be that they did do exactly such a comparison and found that the intervention did not work, whereupon they changed the whole nature of their study.) Where does this leave us? It leaves us exactly where we were last week – there is still no credible evidence that Duluth-inspired domestic violence perpetrator programmes actually work any better than doing nothing at all.

Yes, the paper goes on at some length (52 pages, to be exact) with qualitative evidence of effectiveness, quotes from women who felt it was effective on their partners, evidence from staff involved in delivering the programme, etc etc etc. Some of which is quite compelling. But that does not alter the bottom line – there remains no evidence that the courses work any better than doing nothing at all.

Does this matter? Yes, a lot. Firstly, quite obviously, because it may be that financial resources which are being spent on perpetrator programmes might be used more effectively on victim support, shelters or whatever else.

Even more troubling, however, is that when it comes to interventions with perpetrators, Duluth model programmes are pretty much the only game in town. A charity called Respect has been tasked by the government with accrediting (effectively licensing) all perpetrator programmes in the UK and they will only accredit courses based on the Duluth model. There is (limited but growing) evidence from around the world that other approaches (eg cognitive behavioural therapies, or the types of restorative justice and healing programmes detailed in the wonderful work of Linda G. Mills) might be effective, at least with some offenders. These are, to all intents and purposes, forbidden in the UK, to protect a patriarchal model for which there is no evidence of effectiveness at all. As a consequence, efforts to research and develop perpetrator programmes which might actually work are being stifled and stymied in favour of unevidenced woo.

That is not just a disgrace. It should be a national scandal.