Australian male violence against women: what the statistics say (and media should report)

Amid Australia’s justified concern over male violence against women, it seems worth  keeping in mind our achievements. Femicide, in particular, has more than halved in the past three decades.


Prologue: Violence against women is a bad thing, and it’s still bad even when, as the article below points out, it used to be far worse. We should be trying hard to lower rates of violence, by finding good solutions and implementing them with urgency. As part of this, we should understand just what we’re dealing with – which is what this summary tries to do.

The issue of violence against women is in the news right now. Here’s a short summary of what we know about the issue in Australia.

  • Before we say anything else, we need to acknowledge this: a really accurate picture of violent crime is hard to draw. Of the several factors clouding our vision, one stands out: most police-gathered crime figures are very unreliable. That goes double for violence against women. We can’t just hang that on the police, either: many crimes never get reported, or the police don’t find enough evidence to charge anyone, or judges and juries don’t convict. And all of these things change over time, as society changes.
    • It’s hard to exaggerate what a problem this data unreliability poses when we try to find out about crime. My strong impression is that most of the public and many commentators expect official crime statistics will tell us everything we need to know. They never do.
    • How bad is the problem? One typical analysis claims that “about 70% of domestic violence is never reported to the police.” You can probably come up with plenty of reasons why this figure is so high.
    • Rates of reporting, charging and convicting thus affect the figures far more than do underlying changes in the actual level of violence in Australia.
  • The result of all that is that most experts don’t trust all the official statistics to give them an accurate read on what’s happening. Instead they look for the most reliable figures – which are, necessarily, the figures that will suffer least from under-reporting. That leads them to the figures for homicides. These suffer less from under-reporting, simply because it’s hard to avoid people noticing when someone dies.
  • And so to women. The homicide indicators suggest Australian femicide – homicide of women – has fallen over the past three decades at a speed that might surprise many people. Among the most reliable indicators is intimate partner homicide; female victims are down 60+% in the 33 years to 2023-23. See the graph below. (Source: Australian Institute of Health and Welfare, based on figures from the Australian Institute of Criminology’s National Homicide Monitoring Program)

    Intimate partner homicide, 1989-90 to 2022-23
    Rate per 100,000 population aged 18 years and over

    Figure from AIC NHMP

    • This slow collapse is honestly amazing, because it came after four decades of rising homicide rates, including a big rise in the 1970s. If you had told me in 1990 that the future of intimate partner homicide would be the sort of decline pictured above, I would have been pretty sceptical, but also excited that things were going to get so much better.
  • Here’s where the statistical analysis gets more complicated. The Australian Institute of Criminology has released new figures recently (since this post went public, in fact – thanks to Jenna Price for alerting me to this). These figures bring our data up to June 2023.
    • These new figures show an uptick in 2022-23. You might take this as saying that intimate partner femicides have been rising. On the other hand,  a look at the graph above suggests that bigger jumps in the rate occurred in 90-91, 92-93, 94-95, 2001-02, 05-06, 07-08 and 11-12. This data is just jumpy from year to year, because we’re dealing with quite small numbers by statistical standards. In 2022-23, intimate partner femicide claimed 34 victims – few enough to complicate any year-to-year analysis.
    • To quote Ben Spivak at the Swinburne Centre for Forensic Behavioural Science : “We have to be careful in drawing conclusions about year-to-year changes in homicide. Given the small underlying numbers, small changes in the number of homicides can be made to look larger than they are when reported as percentage change.”
    • This caution obviously applies even more so to the four months at the start of 2024, where it is occasionally claimed that femicide has reached epidemic levels. News reports on the Anzac Day 2024 weekend featured a figure of “26 women allegedly killed by men in the first 115 days of the year”.That figure presumably includes both intimate partner femicide and other killings of women.
    • I don’t know how reliable that number is. But on my initial maths, if that rate of homicide continued right through 2024, it would mean 83 female victims of homicide for Australia this year – a rate of .59 female homicides per 100,000 women. Those 83 deaths – worse than one every five days – are way too high. They are not notably out of line with other recent years, such as 2020-21’s 69; indeed, they are consistent with a continued gradual fall. The .59 rate would have been a record low just a few years ago. And even if the figures jump again this year, it’s not clear what we should make of it: for instance, the figures leapt for a couple of years in a row in the early 2000s, and then just resumed their slide.
  • When we step back and look at other countries, Australia seems to have done well by global standards at reducing violence against women. The femicide rate is falling in many places around the world (as shown in the graph below). But not that many places have bettered Australia’s rate of change over the past 15 years. At the same time, we shouldn’t pat ourselves on the back too hard – because we really just don’t know what causes violent crime numbers to move around over the decades in any country, ours included. (Note that this graph goes only to 2021 and does not include the updated figures described above.)

    Female homicide rate, 1990 to 2021, per 100,000 women, for Australia and comparator nations

    Figure from Our World In Data

  • This graph of the female homicide rate also underlines another point: it is possible to push the rate much lower than Australia’s currently is. Singapore has done it. One question is how much we would be willing to change the nation’s culture to replicate Singapore’s performance. (That country’s law enforcement regime is … tough. That said, my own view is that an obvious place to toughen Australia’s regime would be in enforcement of various court orders around men’s violence against women.)
  • We do have one more fairly reliable source of domestic violence data – the Personal Safety Survey done every five years by the Australian Bureau of Statistics. This tries to avoid the police crime figures issues by coming at the problem from the other end: it asks members of the general public what crimes they have experienced. As Swinburne’s Ben Spivak notes, the survey numbers suggest at worst that overall domestic violence rates are not rising, and at best that they are falling over time. For instance, the rate of cohabiting partner violence against women reported in the Personal Safety Survey fell from 1.7% in 2016 to 0.9% in 2021-22. (We don’t have more recent numbers; it’s a five-year survey.)
  • Some people cite police crime data to argue that the analysis above goes astray. This argument says that rises over the past decade in rates of male offending in categories like “sexual assault and related offences”, “abduction and harassment” and “acts intended to cause injury” suggest these types of male violence against women are moving up, even as homicide rates move down. That is, they argue that changes in reporting rates don’t explain these numbers. They also point to the past 16 months of intimate partner homicides, where the trend seems to have been up.
    • It’s possible that they’re right. But it does not seem all that likely to me that medium-term homicide has detached itself so thoroughly from other medium-term violence indicators.
    • The most likely explanation for rises in these categories seems to me likely to be that the official levels of these non-homicide offences are rising because our efforts to raise reporting, charging and conviction rates are actually bearing fruit – that is, less crimes are slipping through the cracks. Most analysts of crime statistics are wary of police-generated statistics for just this reason. But it is always possible that some of the figures reflect real trends in underlying crime. This is a point that I want to explore further and if necessary revise in this post.

To the extent that the homicide indicators a) indicate actual crime levels and b) are at odds with people’s perceptions, commentators and the media should work to make both the figures and people’s perceptions more accurate. The stories we tell about crime rates have a real impact on people’s lives. As crime academics Terry Goldsworthy and Gaell Brotto have noted, a person’s fear of current crime levels can be influenced by a number of things, including media exposure. Don Weatherburn, former head of the NSW Bureau of Crime Statistics and Research, has complained: “The female homicide rate is much lower now than it was 20 years ago. The media never report this.”

All the above parsing of crime statistics may seem like nit-picking, or even like an apology for domestic violence. (I’ve been accused of both since writing the original version of this post.)

But such dismissals are foolishness. Crime statistics represent our best way to find out what’s going on in the world. The alternative is to try to discuss crime by having people talk only about what their friends down the pub reckon (“mate, this whole violent crime thing is a made-up problem/out of control”)  or what they saw on a TV show last week (“look, it’s obvious the real problem is immigrants/keeping the dole too low/30 years of letting crime run rampant”). Crime statistics matter every time some politician or commentator, left or right, makes a claim about the violent crime rate. They matter even more when policymakers start to talk of changing laws.

Yes, it’s possible that the male-on-female violence trend has turned for  the worse in the past 18 months. But on these numbers, that is not yet obvious. We still seem to me to be in a new era of lower crime. The debate should recognise that fact. And at the same time we should work towards the next era, when crime rates can be lower still.

*  The author studied criminal statistics at the University of Adelaide and has dealt with statistics and their presentation in various roles for more than 30 years.
This post has been update several times since its first posting, as new data has come to hand.
Thanks to Dr Ben Spivak for checking over some of my conclusions; any errors remain my own. Please let me know if you spot any.

About David Walker

David Walker runs editorial consultancy Shorewalker DMS (shorewalker.net), editing and advising business and government on reports and other editorial content. Newsletter: https://shorewalker.net/subscribe . Among other roles, David has edited the award-winning Acuity and INTHEBLACK magazines, been chief operating officer of online publisher WorkDay Media, held senior policy and communications roles at the Committee for Economic Development of Australia and the Business Council of Australia and run the website for online finance start-up eChoice. He is a former economics writer for The Age and News Ltd. He has qualifications in law and corporate finance.
This entry was posted in Gender, Interesting Graphs, Media, Social Policy. Bookmark the permalink.
Subscribe
Notify of
guest

21 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Nicholas Gruen
Admin
2 months ago

Thanks for all the work in this David. Just what I’ve been wanting as I’ve asked myself the obvious questions about whether this thing is worse or better than before.

johnrwalker
johnrwalker
2 months ago
Reply to  David Walker

Look forward to updates
Could the shortage of housing be a factor?
Surely if leaving the , increasingly, violent partner in practice means, sleeping in your car etc then women who say ten years ago would have simply left a fairly bad situation and moved on, might currently stay on just a bit too long ?
Not sure but am not confident that homicide rates are a good proxi, murder and grevious bodily harm rates in the Braidwood region have been very low for decades ,( in twenty years ive not heard of even one murder within say fifty ks of here )however i gather from ecumenical sources that DV in general has at the same time been common and serious in consequences for victims, for decades.

johnrwalker
johnrwalker
2 months ago
Reply to  David Walker

Beautiful by the fire today🙂

johnrwalker
johnrwalker
2 months ago
Reply to  David Walker

PS i recently read that in the current total population of Australia, broken down by their age, the
biggest single group is those aged 32( from memory) and in general people in thier thirties are now a bigger percentage of the total population. And its in your thirties that many try to settle down have a family , a project that at the moment must be unusually stressful for many.
Wondering if that increase in population aged thirty something might be a factor

John
John
1 month ago
Reply to  David Walker

Hi David wondering if you have any updates etc?

John
John
29 days ago
Reply to  David Walker

“not clear how causation runs here or if it runs at all” exactly
I guess if there were stats for women needing significant medical treatment for DV related injuries, that would help?

Peter Lozo, BSc, PhD
Peter Lozo, BSc, PhD
1 month ago
Reply to  David Walker

The 2022-2023 data on the rate of the female intimate partner homicide is 28% higher than the rate for 2021-2022. However, one still cannot conclude from that 28% rise whether or not there is a national crisis in the DV against women!
The Commonwealth Government has an ambitious plan to reduce the female IPH by 25% over the next few years. If we compare their target against the recent 28% rise we can conclude tthat the Government won’t meet its target (and is about 200% away from the desired target) Based on that analysis, we can conclude that there is a significant problem but it still doesn’t tell you whither there is a national crisis in DV against women.

I am happy to expand on this further given that I analysed the data several days ago. I will provide the links to several Government documents:

1. https://www.dss.gov.au/ending-violence

2. https://www.aic.gov.au/media-centre/news/australia-sees-rise-female-intimate-partner-homicide-new-research-report

Peter Lozo, BSc, PhD
Peter Lozo, BSc, PhD
1 month ago

CORRECTION TO MY LAST COMMENT: That should be 25% per year over the next few years (rather than 25% over the next few years).

Last edited 1 month ago by Peter Lozo, BSc, PhD
R. N. England
R. N. England
2 months ago

David Walker has gently reminded us that the media (here especially the ABC) habitually distort the facts. It would be interesting to see some data on violent crime committed by females. Years ago, Stephen Scarlett, the Sydney Children’s Court magistrate commented, “Girls have gone berserk”, which suggests violent crime by females might be increasing. There are some that might regard that as an achievement in the struggle for equality.

R. N. England
R. N. England
2 months ago
Reply to  David Walker

Domestic and public violence are different areas. Public violence often involves gangs, some of it female-against female, which the media can’t resist. The gang scene is overwhelmingly male, but there may be a slight change going on. Women are increasingly encouraged to participate in rough team sport and militarism, so it would not be a surprise if they also turned up in less organised forms of rough, competitive teamwork.

Antonios Sarhanis
Admin
2 months ago

One of the unfortunate side effects of the broader understanding of what constitutes violence or assault is that it also gets harder to determine what we’re dealing with.
What I mean is that I believe just a threat that gets reported to police is considered a case of domestic violence. Again, not saying that it isn’t a serious matter, but bucketing that into the same undifferentiated group with beatings and murder can make things difficult to compare.
Sexual assault also has the same kind of statistical difficulties. Sexual assault can also be a verbal threat or an unwanted fondle, which can get lumped into the same category as the worst of depravities and add to the count in the same way.
And while I’m not much of a Waleed Aly fan, this is a really good article!
https://www.theage.com.au/politics/federal/holding-all-men-responsible-for-a-violent-minority-has-failed-to-keep-women-safe-20240501-p5fo82.html

Mark Hayes
Mark Hayes
2 months ago

News reports on the Anzac Day 2024 weekend featured a figure of “26 women allegedly killed by men in the first 115 days of the year”. 

The number seems to accurate as far as it goes.
https://www.facebook.com/people/Counting-Dead-Women-Australia/100063733051461/ (click on the ‘see more’ button under the heading ‘Counting Dead Women Australia 8 January’·)
It details all (now) 28 female deaths at the hands of men.
But….
The number includes the five women killed in Bondi Junction plus one or two other deaths which were clearly not DV or intimate partner related. As usually happens with these things, the accurate number is quoted but its not accurately described.

Yes 28 women killed by men. No there weren’t 28 DV murders.

johnrwalker
johnrwalker
2 months ago

Obvious questions are :
Is the recent uptick in DV murders an passing upward blip in a overall downward trend, or a sign of something new?
And
Have reporting rates changed?

Pip Cornall
1 month ago

Thank you David, I agree, getting an accurate picture of intimate partner violence, is most challenging. As a former teacher and Juvenile Justice convener in Australia and America, I have studied and taught about gender violence for many decades. Only recently I became aware of the work of Nicola Graham-Kevan PhD, is a Professor of Criminal Justice Psychology and Director of the Centre for Criminal Justice Research and Partnership at the University of Central Lancashire and Michael P Johnson, Emeritus Professor of Sociology, Women’s Studies, and African and African American Studies at Penn State, He is an internationally recognized expert on domestic violence. Johnson identified common couple violence which is unique directional, and intimate (patriarchal) terrorism describes the more entrenched power over situation. Nicola Graham–Kevan builds upon Johnson’s work and believes the courts need to be aware of the various types of intimate violence. Their work resonates with me and my experience in the field. Best regards, Pip Cornall

Peter Lozo, BSc, PhD
Peter Lozo, BSc, PhD
1 month ago

The critical question is whether one can arrive at a sufficiently reliable prediction about the RATE OF VIOLENCE against women (perpetrated by their current or former intimate partner) from the data about the RATE OF HOMICIDE against women (perpetrated by their current or former intimate partner), and therefore whether one can
(i) reliably assess the Commonwealth Government’s claim that there is a CRISIS in violence against women (by their current or former intimate partner)
AND
(ii) whether Social and Welfare services that are available to women who want to escape from domestic violence are comfortably able to accommodate the demad for help.

I just do not see how any analysis of the data on female intimate partner homicide over the past 20 or so years can predict whether there is currently a crisis in the domestic violence against women (or crisis in general violence again women).
It’s my opinion that the best predictor of whether or not there is a crisis in DM violence against women is to actually look at the data from women’s shelters around the country to see where there had been a surge over the past 18 months in the number of women who are seeking emergency shelter in order to escape from domestic violence, and whether the demand was met or whether the system is under stress.

Various CEO’s of Women’s Shelter are reporting a stressed system.

A stressed situation in women’s shelters around the country for 12 or so months is a very strong indicator that there is a national crisis.
For example:

1. CEO of a domestic violence support service in Perth says the extent of the crisis is the worst she has ever seen.https://www.abc.net.au/news/2024-04-29/perth-women-turned-away-refuge-shelters-violence-against-women/103780236

2. “Last year, we extended the number of safe beds that we were able to offer women and children by 30 per cent, and the number of people we had to turn away from our network, even considering that increase, grew by 92 per cent.
So that sas a lot about the demand that is out there.
The year before last, we had about 600 people that we were unable to accommodate, and in the last year that’s grown to well and truly over a thousand.” https://www.sbs.com.au/news/article/i-support-women-fleeing-violence-this-is-what-needs-to-happen-now/k333l1euo