For months, Tortoise has been investigating the sudden or unexplained deaths of women who are thought or known to have experienced domestic abuse, or where investigative failings mean families will never know for sure if their loved one has been killed. It is harrowing work and has uncovered terrible cases – of lost or damaged evidence, killers missed by police, inadequate inquiries by coroners, grieving parents ignored or rebuffed.

There is a familiar statistic: that two women every week in England and Wales are killed by a partner or former partner. But is it accurate? These are just the killings we know about because police have investigated them for manslaughter or murder, and the killers convicted.

Academics, family support organisations and activists who study this issue are certain that homicides are being missed, justice denied – and more women put at risk. 

We launch our investigation at a time of heightened concern about domestic abuse under lockdown. The National Domestic Abuse Helpline reported a 65 per cent increase in calls and contacts between April and June 2020, when the pandemic really took hold in the UK, compared with the first three months of the year. In April, a painstaking project called Counting Dead Women reported that the rate of women being killed by men in the UK had almost trebled during the first week of lockdown alone. 

We will tell some of the personal stories in detail elsewhere in our investigation, but here we want to try to read the landscape. We need to understand the data to sense what is going on. What is being collected and where?

What is the scale of domestic abuse?

The problem with domestic abuse data is that it always comes with a caveat. In England and Wales, the Office for National Statistics (ONS) says that domestic abuse affected 1.6 million women and 786,000 men in 2019, but we think the number is actually much higher – because we also know that only 17% of women report to police. Of the 2.4 million cases that are reported, fewer than one in 10 end in conviction. 

Our investigation has a narrow focus: we are trying to understand how many deaths there are where it is suspected by families or friends that the woman involved has a history of being abused by their partner or ex. These deaths are known as intimate partner homicides.

Professor Jane Monckton-Smith, a forensic criminologist who has spent years researching homicide and coercive control, puts it simply: “We don’t know how big the problem is.” Monckton-Smith works to understand cases where families strongly believe that the sudden death of their loved one – a daughter, a sister, an aunt – is suspicious. Time and again, she has helped these families appeal to the police after dubious investigations, and in doing so she has identified a pattern: women’s killings may well be going uncounted in official homicide statistics. 

In latest figures, the Femicide Census counts 91 women killed in 2018 by current or former male intimate partner; Prof Monckton-Smith believes the real figure could be twice that. If we follow that estimate, and based on 2018 figures, a woman is dying every other day at the hands of her current or former partner.

The UK government is a world leader in open data, and home to some groundbreaking charitable efforts to collect information on domestic abuse. And yet, we don’t know exactly how many women are really being killed.

To understand where things are going wrong, we need to understand the data. Our investigation involved a deep-dive into what is being collected, where, and by whom.

So what data do we have?

Let’s start by looking globally: The United Nations Office on Drugs and Crime’s Global Study on Homicide records international homicide counts, rates, trends and patterns, including leading research on gender-based killing. Looking at all of the global data, men are most commonly the killers and the victims, except when it comes to intimate partner homicides. In these cases, women are overwhelmingly the victims, at 82%.

“None of that implies that women don’t kill their partners – they do,” says criminologist Sandra Walklate. But, she says, the pattern – globally – is clear.

But this data has limitations. The UN relies on information sourced from governments, most of which have different legal definitions of what intimate partner homicide actually is. Some, including many countries in South America, have laws around femicide – when a man kills a woman or girl. The task is made more difficult because some countries don’t record whether the killer is a man, while others don’t collect the data at all – nations in Africa and Asia account for the biggest gaps.

But the available numbers are enough for the UN to produce regional estimates. According to their most recent report, 82 women a day were killed by their current or former intimate partner in 2017 – the most recent estimate. These numbers represent a fifth of global homicides.

So what do we know – in the UK?

Understanding the context of intimate partner homicides has been central to our research – and Prof Monckton-Smith’s. “We are not properly tracking at a government level the amount of homicides there are, and in what specific contexts,” she says. 

In the UK, the ONS provides an annual homicide report for England and Wales, while Scotland has its own similar dataset. But these figures only scratch the surface.

The ONS finds that an average of 85 women a year were killed by a partner or ex-partner in England and Wales over the past decade – that’s 44% of all homicides against women. In Scotland, that percentage is 49%.

“The [ONS] report is actually quite vague about categories of homicide,” says Monckton-Smith. While government data tells us the number of victims, their gender, and their relationship to the perpetrator, there is no further information around the crimes and their nature. Some cases may also be lost because the killer’s gender is not noted.

Crucially, there is no information about the perpetrator’s history of domestic abuse. This makes it hard to understand the relationship between domestic abuse and homicide, even on the most basic level. 

Counting Dead Women

The UK’s most significant dataset on the killing of women began as an online blog. Eight women were killed in the first three days of 2012, and, that same year, the Chief Executive of domestic violence charity nia, Karen Ingala Smith, began to count and name them on her WordPress page, Counting Dead Women. She trawled through articles, police reports and domestic homicide reviews to collect and memorialise the cases. In 2015, Ingala Smith and Clarissa O’Callaghan launched the Femicide Census, following their work on the count. 

Their 10-year report, released in November 2020, paints a stark picture of homicide against women in the UK. According to their report, there has been no improvement – women are being killed by men at the same rate that they were a decade ago, averaging 143 deaths a year. This number includes all killers, not just intimate partners.

The trends are largely in line with those of Scotland, England and Wales but we can account for a number of small differences: though the census is UK-wide, it only takes into account cases where a man has been charged, whereas the others include unsolved crimes. The Femicide Census also excludes trans women, whereas in government data the sex of a homicide victim is determined by the police force that records the crime, which may account for some discrepancies.

The Femicide Census provides crucial context for each killing, providing data on everything from the location, to the method of killing, to the perpetrator’s history of abuse. 

By collecting on every case, we get closer to the heart of the problem.

The 10-year report finds that, over the past decade, 62% of the cases they counted were women who died at the hands of an intimate partner.

Nearly two-thirds of perpetrators were currently or had previously been in an intimate relationship with the victim…

… while 72% of female homicide victims die in their homes.

The Census also begins to link domestic abuse and femicide – 59% of cases involve a history of coercive control or violence, while almost half the perpetrators were known to have histories of abuse against women. This data is collected manually from dozens of sources, including police reports, domestic homicide reviews and news articles. It’s the same method Ingala Smith used when she began blogging nearly a decade ago.

But that first number is likely to be even higher. According to the Femicide Census, “it is likely that the victim was subjected to violent or controlling behaviour by the perpetrator in the overwhelming majority of cases prior to the femicide”. Several academic studies have shown a clear link between a history of domestic abuse and intimate partner homicide, but this is not reflected in the numbers because data sharing on this issue is so poor. 

These breaks in the chain have become a central concern in our investigation. 

“The first time the criminal justice system might know about violence in the person’s life is the fatal violence,” says Walklate, the University of Liverpool criminologist. It’s an uncomfortable reality, she says, but information about domestic abuse is rarely passed on to the police. “Family know, friends often know, or the neighbors might know,” she says. 

This information is sometimes held by the healthcare system, too, which might, for example, observe recurring patterns of physical violence or mental distress in victims and suspect domestic abuse is involved. None of this is shared with the criminal justice system, so “having that complete picture is a little more complicated”. Without a wider view, it’s hard to find an intervention point and save lives.

Some steps have been taken to improve the situation. In 2015, the government launched the Domestic Violence Disclosure Scheme, known more commonly as ‘Clare’s Law’. Named after Clare Wood, who was killed in 2009 by a former partner with a history of violence, the scheme allows potential victims of domestic abuse, or their friends and relatives, to obtain information from the police about a person’s history of domestic abuse offences and convictions. The police also have a right to disclose that information to a person they perceive to be at risk, such as a woman entering a relationship with a known or suspected abuser. The hope is that the scheme will help protect potential victims, but the reality can be far less simple. The victim may fear it is dangerous to put in an application for information, out of fear of retaliation. Or, as we keep finding, the data is simply not available on many of the perpetrators, because they were never reported to the police.   

There are smaller schemes, too, like Operation Encompass, which allows data sharing between the school system and the police to ensure that children who live in homes where there is domestic violence receive adequate support. The scheme currently works with several police forces, and is a first step towards vital data being habitually shared.

Walklate acknowledges the value of such programmes, but also points out their shortcomings. “It’s all very piecemeal,” she says. Most data sharing schemes focus exclusively on the police, but what’s really needed is standardised national policies which set out what data police and others must share. Otherwise, “how do you stitch the whole picture together?”

Dead, but not counted

Even a study as scrupulously put together as the Femicide Census cannot capture or confirm every domestic abuse death. It is, inevitably, an underestimation. 

The Census only includes cases when it is legally clear that a man is the perpetrator, either because they were charged with murder or manslaughter, or because an inquest reached an unlawful killing outcome. We know from the ONS that a third of all homicides against women go unsolved. 

But we also know that there are femicides that are solved, but even then are not counted in official statistics. 

Throughout our investigation, we have heard the stories of families who suspected they lost their relatives to domestic abuse, and are now fighting for justice. Julie, whose daughter Katie Wilding was found dead beside her ex-partner Mitchell, has been fighting for justice for the past four years. Katie’s death was recorded as a drug overdose, despite Mitchell’s known history of violence against her – according to Julie, he had even threatened to kill her using drugs. Despite this, and despite her not being a habitual drug user, the role of domestic abuse in her death was never acknowledged.

There are others like Katie. Women whose families don’t want to go public. Women like Susan Nicholson, who was found dead of asphyxiation after suffering violent domestic abuse, and whose death police wrote off as an accident, even though her partner’s previous girlfriend had also died in very similar circumstances. Or Diane Stewart, found dead in the garden of her home in 2010; an inquest concluded she had suffered sudden unexpected death from epilepsy. Diane’s husband is now serving a life sentence for the murder of another woman – his next wife-to-be, the author Helen Bailey. Ten years after Diane’s death, he has just been charged with her murder, too.

“The number of cases that are being presented to me – it’s a lot,” says Prof Monckton-Smith.  “It’s not just one or two. These aren’t outliers. This is happening week in and week out. 

“As far as actual numbers, we’re guessing. It could be worse even than I think.”

What is the number?

At Tortoise, we decided to start building a count, to acknowledge the lives of these women, and to recognise the role domestic abuse played in their deaths.

We asked every police force in the UK for data on the deaths of women with a background of serious domestic abuse. 

The hardest thing about drafting these questions was understanding how the police keep records about domestic abuse. Do they keep a record against the victim? The perpetrator? The person who phoned in the incident? The address? We spent weeks trying to figure it out. We spoke to former police officers, domestic abuse charities, academics, who were all unsure, or gave different answers, until one person told us with confidence: there is no national system for domestic abuse flagging; each police force does it differently. So we had to ask questions wide enough to capture all these different records of domestic violence, but precise enough, too, that the police would not immediately reject our request. Precise enough so that we could get data on as many women as possible.

When our first response came in a few days later, it was clear, even then, that the police records were not straightforward. They asked us to refine our questions. “We do not flag victims or suspects with markers for Domestic Abuse,” one response said. “Can you clarify if the requestor would be happy for where the victim has been assessed as vulnerable due to Domestic Abuse?” It was an early indication that getting this data might be harder than we imagined. 

Ten days after we sent the questions, our first full response came in. The force did not answer our central question, saying it was too complicated for them to pull out the information. They did, however, offer us one statistic: the number of female homicide victims in their area – they called this “a gesture of goodwill”.

Over the following days, the refusals continued to roll in, with each police force maintaining that it would be too time-consuming to extract the data. Many explained that they would have to manually search their records, going case by case in order to determine whether domestic abuse was involved. Others said that they had no way of cross-referencing the data we asked for, saying they could not think of a way that we could narrow down our question to get the response we need. Some provided no explanation at all. 

By this point, it was clear that there is no national system for recording domestic abuse. Each force has its own. People are being subjected to violence on a daily basis, but what this process told us was, we have no national consensus on how to count what’s being done to them, by whom, or the outcomes. 

Our first glimmer of hope came in the form of an email from one English police force. They’d answered all of our questions, year by year, with no need for clarification. According to their data, one woman with a history of domestic abuse had died in sudden or unexplained circumstances, in the past five years: her death was investigated as a homicide, but no charge was brought.

That response came in fairly quickly. That implied the force in question had ready, easy access to the data on their recording system. If they could do it, we thought, why couldn’t others? 

Perhaps their data system was just better. Or, maybe they simply recorded their data in a way that was more in line with the questions we’d asked. With no other forces to compare the data with, it was hard to know. 

In the following weeks, among all of the rejections, seven more forces sent through complete answers. But it wasn’t always a smooth process. Two of them, for example, sent single answers to multi-part questions, meaning we had to send dozens of clarification emails in order to get what we asked for. Another sent through data for the wrong dates. But eventually, we had eight complete responses – something that had felt impossible just two weeks earlier.

We are still waiting to hear back from 12 forces out of the original 46 – they missed their deadlines, or extended them because of multiple clarification requests. Our original goal – counting how many hidden homicides may actually be occurring, may not yet be achievable, at least not precisely as we’d first hoped, but the numbers start to paint a picture of the problem.

So what have we learned so far? We are not identifying individual police forces at this stage – the reason for that is the numbers are relatively low and there is a risk that individual cases where legal issues are unresolved could be easily identified.

But we have partial data from 13 forces, which identifies 77 cases where women were killed by current or former intimate partners – and in 34 of them, police were aware of earlier domestic abuse allegations.

Separately, eight of the forces recorded 30 cases where women with a history of domestic abuse died suddenly. Fifteen of those cases were in a single police force area. Of the 30 cases, 14 were investigated as homicides and 12 led to arrests. Criminal charges were brought in five cases. To date, none have led to convictions.

That means 16 of these women’s deaths were not investigated as a homicide.

Finally, the incomplete data records nine women who are currently missing, and who have a history of domestic abuse. Six of them are in a single police force area.

What is this telling us? Three significant things:

• Early signs of domestic abuse are being logged – at least sometimes, in some places – but in dozens of cases have not kept women safe. Could lives be saved if better, standardised records were kept?
• Are sudden deaths of women with a history of domestic abuse always treated with the same investigative urgency – and if one force has 15 cases, how many are not being acknowledged in the other 45 forces?
• Is the evidence gathering robust in such cases – what could be learned by close analysis of all sudden death cases where there is a background of domestic abuse?

It is worth underlining once again: this is a partial picture from only 13 of 46 police forces, none of them include the major city forces across the UK.

If these indicative figures extend across the country, it suggests most forces will have questions to answer: about the way they record information; the proactive steps they take to safeguard women with a history of domestic abuse; the protocols they follow when a woman with such a history dies suddenly.

Time for change

The data is a mess. We want to fix that. As a first step, we are asking the government to mandate that this data gets collected. We want to make public the total number of women known to the police to have been subjected to domestic abuse or coercive control who have died in sudden or unexplained circumstances, and what the inquest verdict, and criminal justice process and outcome was in each case.

Then, we can begin to build a better picture of the problem, and press for better data sharing around domestic abuse. 

This is an investigation about accurate numbers. But, of course, really it is about people – women and their families. It’s about making sure their lives count – in every sense.