Rape, Responsibility and Bad Statistics

Lots of ground to cover here so I’ll straight down to business.

First up Tim Worstall notes an interesting statistical claim from Julie Bindel in today’s Grauniad, which he files under the category ‘idiotarians’:

Violence against women is the cause of more deaths and disability around the world in 15- to 44-year-olds than cancer, malaria, traffic accidents or war.

Tim would like to know where the evidence is the for this, which as good question as figures from the World Health Organisation put deaths from cancer at 7 miiion a year (12.5% of all deaths) and malaria at around 1 million and year while traffic accidents rack up 1.2 million deaths and anything from 20-50 million injuries/disabilities each year.

Given the scale of the problem that Bindel claims exists, you might expect this to be mentioned by the WHO in their new ‘landmark’ study on domestic violence around the world – yet neither the summary report nor the full report makes any such claims – in fact the WHO report, in calling for more research, explicitly notes that:

In some places few [sic] data on violence against women are available. More research on the magnitude and nature of the problem of violence against women, and its costs, in given countries or settings is therefore urgently needed to provide a stronger basis for advocacy and action.

Which seems to confirm that Bindel’s figure derive from that time-honoured journalistic tradition more commonly refered to as ‘pulling numbers out of your ass’.

Meanwhile, Amnesty International’s survey on attitudes toward rapes, reported here by the Times, has been rattling around the blogosphere and spawning much interesting commentary.

As things go, Andrew at Non-trivial Solutions, with a bit of an assist from Jarndyce in the comments, has pretty much the right kind of take on things in noting that ‘blame’ and ‘responsibility’ are very different things and that there seems to be a fair bit of public confusion between notions of responsibility and risk within the survey answers which is reflected in some of the commentary on this study.

I think the sensible thing for Amnesty to do is commission a follow-up study to unpick this whole blame-responsibility-risk issue and provide a more nuanced assessment of public attitudes.

My main practical criticism of this study, however, is reserved for its seeking to draw conclusions about the public’s understanding of rape and the issues it raises – which is ‘sold’ as ‘public ignorance’ – from their apparent lack of awareness of current crime statistics and conviction rates. Andrew rightly points out that aspects of the survey’s design will tend to introduce a degree of cognitive bias into the results.

What the survey and many of those commenting on it fail to address, however, is the question of why its assumed that the public should possess an encyclopedic knowledge of crime statistics and conviction rates for rape, or any other offence for that matter, in the first place?

I think it a near certainty that the same kind of survey questions, if applied to other types of criminal offence, would turn up much the same the results – i.e. that the public don’t really know but about half of them are willing to take a guess based on the multiple-choice options put in front of them.

As it happens, had I been one of those surveys, I would have got the questions on rape statistics and conviction rates pretty much right – but only because I’ve had cause – for professional reasons – to look these things up and have, in any case, read something in a newspaper on just this subject fairly recently.

Ask me the same questions about fraud, street robbery or pretty much any other crime and I really couldn’t tell you what the figures are, not because I’m ignorant but simply because I’ve had no particular reason to look them up. If I should find myself needing to know the numbers, well the I do know where to look for them and can track them down pretty quickly if and when I need them.

All of which makes me wonder whether, if I rang a few journalists and asked them to tell me, off the top of their heads and without looking it up, the gross domestic product of Lesotho and the countries per capita expenditure on toilet paper and then stuck out a press release claiming that journalists know nothing about Africa, would that get me an op-ed piece in the Grauniad?

Still, this is an issue which looks set to rumble on for a bit, not least courtesy of the High Court judge who ruled that consent given when pissed is still consent, at the conclusion of a rape trial which ended with the judge instructing the jury to render an acquittal.

This has sparked something of a controversy as the alleged victim claims to have been unconscious at the time the rape took place, making it appear that the judge has overturned both part of the recent Sexual Offences Act 2003 and a hundred year-old principle in common law that consent cannot be given – quite obviously – when unconscious, even though the law professor drafted in by the times to comment on this ruling notes that:

“If a woman has sex with a man just because she is drunk, and with whom she would not have had sex otherwise, then she cannot say she did not really consent because she was under the influence of drink.â€?

She added than any case where a woman was proved to be unconscious would be a long way down the line from that.

All of which explains why the judge directed the jury as he did when you not the operative phrase “any case where a woman was proved to be unconscious” – and its also worth noting that he made his comments in response to a request from Huw Rees, the prosecution barrister, that the case be abandoned “in light of the evidence revealed in cross-examinationâ€?.

Essentially what happened here is that the prosecution were unable to establish with corroborating evidence and beyond reasonable doubt that the alleged victim was actually unconscious and, therefore, incapable of giving consent, which – whether you like it not – means an acquittal.

So, the reality here is that the judge isn’t overturning either the Sexual Offences Act 2003 or common law after all as without proof that the ‘victim’ was unconscious, the question of consent become entirely moot.

All of whcih bring me to the last post I want to reference, which is this one from Gendergeek, and specifically this from Emmy’s actual post:

Amnesty’s Kate Allen believes that the government must take action to combat this ‘blame culture’ and publicise the true prevalence of rape and its appalling conviction rate in the UK. The results of this opinion poll indicate that current policy approaches are not working.

And this comment from Emma in the comments:

Your confidence in the British public is not borne out by the attrition rate for rape complaints. Only 4% are converted into convictions in Scotland.

Yes, it certainly is true that headline conviction rates are appalling low in rape cases – 5.6% overall being the figure most often quoted of late.

Except that this has very little to do with public attitudes to rape or ‘ignorance’ of crime statistics, as this recent Home Office report ‘A gap or a chasm? Attrition in reported rape cases‘ clearly demonstrates.

The 5.6% conviction rate that is widely reported is the headline rate for all reported rape cases. By way of contrast the actual conviction rate in cases which make it to the courtroom is actually 60% – 24% due to the defendant entering a plea of guilty and another 36% where a trial leads to a conviction on a guilty verdict – which amounts to 48% of all cases where a full jury trial takes place leading to convictions.

The real problem here lies in getting cases to court – only 14% of reported rapes result in trial proceedings but:

9% of reported rapes are designated false during investigations; the majority of which comprise reports from women in the 16-25 age group.

A full third of all reported rapes fail to proceed past the investigation stage due to ‘evidence issues’ – in many cases the issue here is one of ‘victim credibility’.

Another third are lost when the victim withdraws from the case, either for fear of being disbelieved or of the criminal justice system and court process.

So you tell me, where’s the problem here?

You can’t really blame ‘public attitudes’ when so few cases go to court in the first place, a jury trial being the only situation where such attitudes – as opposed to the victim’s own perception of what public attitudes may be or the attitudes their encounter from the police and CPS – are really put the the test.

All things being equal, a 48% conviction rate in trials is still pretty much on the low side compared to most other offences but even then, without studying and evaluating what goes in jury room in cases which result in an acquittal its still impossible to say how much of factor public attitudes are as opposed to the quality of evidence, despite the fact that the evidence from the US, where for example Alameda County, California (which includes Berkely and Oakland) has a conviction rate at trial of 90%, shows that the use of specialist rape investigation and prosecution teams increase the conviction rate in rape cases at trial to around 85%.

This all brings us back to a particular pernicious and disreputable subtext to the current law and order debate, one which masquerades as concern for ‘victim’s right’ but which is actually a coded argument in favour of articificially inflating conviction rates by the quick-fix method of denuding the fundamental principles of justice, the presumption of innocence and the requirement for ‘proof beyond reasonable doubt’.

Put simply, what the evidence says is that the way to improve conviction rates in rape and many other offences is through the use of highly-skilled specialist teams who can generate quality evidence – this is certainly true in the US in rape cases and equally true in the UK where the introduction of a specialist unit to tackle serious and complex fraud cases has successfully upped the conviction rate from around 63£ to 85%, albeit that this unit has also had the odd spectaclur and extremely expensive failure.

The trouble with this approach as that, politically-speaking, it requires both a significant cash investment to put such units together and takes time to deliver concrete results – if government were to make such an investement today it could take anything from three to five years for such units to start to make a significant dent in conviction rates – allowing both for the time it takes to get units set up, trained and resoourced and the time it takes to investigate cases, get them to trial and see their outcomes reflected in official statistics.

As is the case with many things, doing the job properly can be an expensive and time-consuming proposition, even if that’s the right way to go.

Unfortunately, politicians have a second option; tinkering around the edges of the law to create the impression that they’re doing something, which comes considerably cheaper and is more ‘immediate’ in political terms as it appears to generate results – new and amended laws – the actual effectiveness of which can be nigh on impossible to evaluate for several years.

For politicians this can all too easily appear a win-win strategy. you tweak the law so that, in theory, it should make getting a conviction bit easier and dish out stiffer and more punitive sentences – all of which plays well with the likes of the Sun, Mail and Express – and then a few years down the line when it turns out to have made little or no difference at all you’ll either be out of office and out of the firing line or you can just pass the buck and blame judges for not interprting the law the way you wanted them to, tell everyone that ‘public opinion’ supports more changes to the law and start the whole dumb process all over again.

This is where the supporter of this current campaign to tackle violence against women, and women’s groups in particular, need just to be a touch careful and not lose sight of the nuances of the debate.

What we should all be demanding is that government tackle the difficult things here, the things that require long-term and sustained investment in education – to try and change attitudes, particularly amongst men – in better policing and specialist investigation teams and prosecutors and in the co-ordinated provision of support and after-care services for women who are victims of rape, domestic violence and violent attacks in general.

What needs to be avioded, though, is falling into the trap laid by the notional idea of ‘victim’s rights’ and that, somehow, changing the law in ways which, so the government claims, rebalances the system towards the victim but which, in reality, eat away at the fundamental principles of Britis justice, which seems to promise much but invariably delivers very little.

We should avoid the easy option here, the one which takes away more and more of our precious rights and freedoms not just because its wrong in principle – although it is – but because if offers the least effective solution here adn if we genuinely want a first-class response from government on this issue we should make it clear from the outset that we’re not going to be suckered into taking second, third or fourth best just to cover their collective arses with a quick fix before the next election.

Update: It turns out the statistics that Bindel uses (badly) in her article are taken from 1993 World Development Report, which is produced by the World Bank and, as Jo Salmon has pointed out in comments, is in widespread use by both Oxfam and the US Congress.

None of which validates the manner in which these statistics are used in the article.

For one thing the statistics in question relate to morbidity not mortality, which is assessed in terms of an estimate of the years of healthy life (1 in 5) lost to women due to violence – which is not an assessment of cause at all but an indication of the extent to which such violence is a significant contributory factor in women’s health problems.

This notional idea of ‘healthy years lost’ due to the effects of violence would certainly include situations where there is a direct causal link, where the violence results in death or injury, but also many more indirect effects, for example those arising from drug or alcohol dependancy which can certainly be linked to violence but cannot definitively be said to be caused by violence as there are a wide range of other factors which must be taken into account.

In addition, the actual WDR estimates being cited place healthy years lost due to violence above those lost to either breast cancer or cervical cancer but make no claim at all in relation to all cancers – in fact breast and cervical cancers combined make up about a third of all cancers in women in this particular age group.

In essence there’s nothing particularly wrong with the statistics themselves but the manner in which they’re being used here is problematic and their significance is overplayed – as is all too often the case when sucvh things turn up in newspaper articles.

  • Jo

    FYI, Oxfam use the stat.

  • Jo
  • Jo
  • wiggles

    Thank you for a well balanced piece on this difficult subject. One more stat that I have read on a BBC News report (but can’t say how authentic it is) is the police’s belief that only 15% of rapes are actually reported. If nearly 6 out of every 7 rape victims won’t come forward, for whatever reason, then I think we have a long way to go before we can realistically say that we are solving the problems of sex crimes in the UK.