To paraphrase Brendan Behan, there is no problem so bad that the arrival of Miranda Devine cannot make it worse. When La Devine turned her attention to the subject of domestic violence/partner violence (DV/PV hereafter) she did not disappoint in disappointing, sheeting at least some of the blame home to you know who:

“If you want to break the cycle of violence, end the welfare incentive for unsuitable women to keep having children to a string of feckless men.”

Nasty stuff, but not unusual. The right can’t pretend DV/PV doesn’t exist, so they get angry with those who keep reminding us of problems deep in the social fabric. Yet part of the reason Devine can get some traction is because there has been a near-ban among many progressives about talking about social factors such as class and race in the occurrence of DV/PV.

Countless news and op-ed pieces over the past year have asserted a simple model of DV/PV — that it is based on power and control along the line of gender only. Yet it is only one among several relatively unquestioned assumptions that have powered the debate and become received fact. The danger of this is that bad reasoning will take us along the wrong path in dealing with the issue — and that, after a couple of years of op-eds, reports and grandstanding, we’ll be right back where we started from. And by then, public attention will have moved on.

So, over the next couple of days, it seems worth considering some of these assumptions one by one, not with the aim of producing another neat theory, but seeing if dominant assumptions stand up to scrutiny.

1. Is violence against women on the rise? Is there an ‘epidemic’ of domestic violence/partner violence?

The notion of an “epidemic” of DV/PV, and a rise in its occurrence, has become a standard observation on the problem, with the word “epidemic” much used — all suggesting that the rates of occurrence, rather than the rates of reporting, have increased; the term “epidemic” suggests that something has blown out to massively greater numbers than was the case. In some reports, first-person evidence of women’s refuges having unprecedented demand is added. The dominant idea is that some real social event has occurred, which many people appear to be taking as the cause of new attention being paid to the issue.

Yet the plain fact is, we have no real way of knowing whether the rate of DV/PV has increased, decreased or stayed the same from such stats. We certainly don’t now whether it has gone up or down over decades, since public attitudes and police practice towards it have changed so much, but even recent year-on-year statistics don’t tell us much, beyond an increase in reporting itself. The problem is well-illustrated in the Quentin Bryce report Not Now, Not Ever, for the Queensland government, which notes the increase in reporting in recent years:

“In Queensland, reported occurrences of domestic and family violence have increased:

  • In 2010-2011 there were 52,889 occurrences, a 7.0% increase on the previous year
  • In 2011-2012 there were 57,963 occurrences, a 9.6% increase on the previous year
  • In 2012-2013 there were 64,258 occurrences, a 10.8% increase on the previous year
  • In 2013-2014 there were 66,016 occurrences, a 2.7% increase on the previous year.” (p.46)

Further on, however, the Bryce report notes that police and legal approaches changed with the introduction of a new law in 2012. This considerably broadened the definition of violence to include, among other things:

“The Act defines the conduct of domestic violence as including physical, sexual, emotional, psychological, and economic abuse or any other threatening, coercive, or controlling behaviour which causes the victim to fear for their safety or wellbeing or that of someone else. Examples of this type of behaviour include:

  • Causing physical injury;
  • Threatening physical injury or death whether towards the primary victim or others, including pets;
  • Coercing or forcing the victim to engage in sexual activity or attempting to do so;
  • Threatening to, or depriving a person of, their liberty;
  • Damaging a person’s property or threatening to do so;
  • The perpetrator threatening to self-harm or suicide for the purpose of tormenting, intimidating or frightening the person to whom the behaviour is directed;
  • Conducting unauthorised surveillance of the victim (may include following or tracking the victim, monitoring telephone calls, text messages or email) or unlawfully stalking the victim;
  • Controlling or withholding the family assets and income which denies the victim economic or financial autonomy or the ability to pay the reasonable living expenses for the family; and
  • Tormenting, intimidating or harassing the victim (may include repeatedly following or contacting the victim without consent, derogatory taunts, withholding medication, disclosing the victim’s sexual orientation without consent).” (p. 68)

It is the inclusion of these last few points — threatening self-harm, repeated contact, disclosing sexual orientation — as violence that is worth more investigation. All hostile and menacing behaviours to be sure, but their inclusion as violence may well have caused an increase in incident reporting, which could have altered the nature and volume of call-outs and write-ups.

But there is no clear system of categorising police call-outs, even though these have become the standard by which we talk about domestic violence. A report in The Age attempting to break down police call-outs and reporting into categories demonstrated the problem, for while it could distinguish between call-outs that occasioned a safety notice being issued, intervention order applied for, or charges laid, the nature of the events themselves — from physical assault to an out-of-control argument, which, 10 years ago, would not have met the standard for DV/PV — remain unknown.

One figure that gives some possible counter-evidence to notions of a rise in occurrence is the rate of female victims of homicide. According to Destroy The Joint’s “Counting Dead Women” website, the number of female victims of homicide this year to date is in the mid-60s. With around 280-300 total murders in Australia per year, and women representing one-third of the victims (both fairly stable figures — though the overall murder rate has dropped because of improvements in emergency medicine), that would suggest a total figure for the year in the mid-80s, and thus no change in that figure, year-on-year. (As Gay Alcorn noted going over some of this territory some months ago, one problem with the Counting Dead Women list in talking about DV/PV and gender is that it includes murder by strangers and women killing women and omits women killing men or children.)

That’s hardly conclusive since the ratio of non-lethal to lethal violence may shift over time, and non-lethal violence might have increased. Most researchers see the ratio of DV/PV murder to a wider rate of non-lethal violence to be stable over time, but it might well have altered in recent years. It’s very difficult to tell.

The absence or presence of an “epidemic” or a rise in DV/PV shouldn’t alter whether increased money and energy are devoted to public campaigns — but it does alter our idea of what we’re up against. If there has been an increase in DV/PV, it may come from any number of causes — the rise of an overt masculinist ideology (the so-called “pick-up artist” and “men’s rights activist” movements), the spread of easily available violent pornography, or further shifts in class, culture and economy that have affected gender relations. But if there hasn’t, then we are dealing with something far more resistant to change and reform — a baseline level of violence that has been with us for years or decades. And that would demand a substantially different approach to incident reduction.

2. Is there a single root cause of DV/PV?

When DV/PV began to emerge as a public, criminal act and a focus of substantial research and action in the 1970s, a great deal of research focused on socioeconomic status and social power or lack thereof — and also on the “system dynamics” of the families in question.

By the late 1970s, this approach came under question from a number of researchers who argued that such approaches ignored the primacy of gender, and the uses to which acts of violence were put. They felt that the sociological approaches, if applied as a form of intervention, removed any notion of accountability by violent men. One model of intervention developed in Minnesota in the early 1980s put “power and control” at the core of a range of branching behaviours by violent men. This concept — the Duluth model — diagrammed out as the power/control wheel, became first the dominant and then the all-but-exclusive mode of understanding DV/PV in large sections of the English-speaking West. It’s reproduced in the Bryce report. It dominates many of the dozens of op-eds that have appeared on the topic in past months. It appears to dominate the Australian imagination of DV/PV.

There are two problems with this. First the Duluth model schema was developed as a framework for intervention, a way of simplifying a complex reality during the state-managed encounter with violent men. Yet in its reproduction over the decades it has become a full social theory of domestic/family relations, and of gender relations more generally. What was designed to guide intervention has become a purported map of “really” occurring processes, which excludes vast areas of social life.

The second problem is the Duluth model does not appear to work particularly well. Multiple studies of its efficacy have found varying degrees of success with the model, from middling to worse compared with other approaches — with Babcock et al’s meta-analysis in 2004 constituting a pretty comprehensive refutation of the Duluth model’s claim to efficacy.

Though there has been some indication that it is useful in some situations — and many practitioners use it as one approach among many — its dominance in the public mind, and its shaping of public understanding of DV/PV, is out of all proportion to its usefulness or application. In the Bryce report it is presented as simple fact, the sole model by which DV/PV occurs.

How did the Duluth model acquire such power? There are many reasons. Though it appears to be a radical approach, its implicit model of action is individualist and reductive. Any social categorical explanation as to how men might act violently is rejected as an excuse, and violent action is constructed as an individual choice. The aim of the intervention is to make men own it as a choice — the effect of it as a social theory is to construct it as always having been a choice. The Duluth model cut with the grain of the traditional morality of the Reagan-Thatcher era, and the idea that there was “no such thing as society”, i.e. that explanation at a social and collective level had no meaning.

The model was popular in an era when feminism was still on the march, while other movements — particularly anti-capitalism and anti-racism — had stalled. It has been popular with social work/care professionals, and it appealed to politicians (illustrated by Martin McKenzie-Murray’s illuminating inside story of its persistence in the Victorian government in The Saturday Paper) because it allowed them to talk tough and avoid terrible news stories about violent men getting soft treatment from other approaches, etc, etc. The Duluth model may well be steering us away from talking about class, race and many other factors playing a key role in DV/PV, and leading us down a false path. A theory of social action — and one that not well-confirmed by evidence — is being presented as more-or-less empirical fact in documents like the Bryce report.

So, to sum up, there’s no clear evidence of the two pillars of current discussion of DV/PV: no clear picture of a new DV/PV epidemic, and no evidence that our single omnipresent “power/control” approach works. Which suggests — as we’ll see tomorrow — that many of the calls to action may well be leading us in exactly the wrong direction.

Tomorrow, let’s look at what we’re not looking at in DV/PV.