As regular visitors to the Ministry will know, I written a couple of posts on the thorny subject of rape statistics in the last 18 months, one looking primarily at prevalence and under-reporting, the other at data relating to the overall performance of the criminal justice system, which covers the always contentious question of conviction and attrition rates.
As a general rule – and this is something I picked up from Ben Goldacre in a discussion on the subject of cancer statistics – the best way to think of rape statistics, particular when you see them quoted in the media, is that data we have to work with doesn’t tend to provide much in the way of ‘right answers’, only different degrees of wrong, a lesson that Amanda Bancroft (aka @_millymoo on Twitter) would have done well to heed before penning this article for Comment is Free.
Myths about rape conviction rates are putting people off going to the police
Media reports about rape do injustice to our legal system – it has improved and offers more protection than often thought
The conviction rate for rape is 58%. That bears repeating. The conviction rate for rape, is 58%.
Now repeat after me… the conviction rate for rape is 33.34% in cases which lead to a prosecution.
The conviction rate for any offence in cases in which rape is primary charge put to the court is 58.58%, or at least it was for 2010/11, but if we’re to stick only to cases which result in a prosecution then the full breakdown of case outcomes by the most serious offence for which a conviction was obtained actually looks like this.
- Acquittal – 41.42% (and I’ll get a breakdown of the acquittal statistics in due course)
- Rape convictions – 33.34%
- Conviction for a lesser sexual offence – 14.14%
- Conviction for a violent offence -5.06%
- Conviction for an indictable offence* – 5.06%
- Conviction for a summary offence* – 1.02%
* An indictable offence is any offence that can be tried, on indictment, by a Crown or High Court while a summary offence can be tried only in a Magistrates’ court unless its an ‘either-way offence’ in which the defendant can opt for a jury trial.
Amanda is, it seems, a former barrister and naturally brings a barrister’s mindset to her arguments, which means that she sees any kind of conviction as a result but personally I’m going to take a bit of punt here and suggest that women who’ve been through the criminal justice system in the hope of seeing their rapist convicted of rape are unlikely to see things in quite the same way if all they get out of their experience is a conviction for a lesser offence.
Call me picky if you like but I’m fairly sure that what rape victims want from the criminal justice system is justice, not a consolation prize that keeps the bean counters at the Crown Prosecution Service in bullshit statistics.
The conviction rate for reportable crimes of all types is 57%. I know you will have heard the figure of 6%. Everyone has. That figure is actually an attrition rate, not a conviction rate, and even as an attrition rate it is wrong – the attrition rate for rape is in the region of 12%.
An attrition rate is the amount of convictions resulting from reports of a crime, and is not routinely calculated for any crime other than rape. Therefore without manually undertaking the exercise, it is impossible to compare the (true) attrition figure for rape with other crimes. A conviction rate is the amount of convictions following a trial, and is calculated for all reportable crimes.
13% is audited figure for 2007 for any conviction arising from a reported rape case, the ‘attrition rate’ for this same dataset is the percentage of cases which didn’t result in conviction, i.e. 87%, a figure that is, perhaps, best thought of as a ‘crude attrition rate’ as it provides no indication of where in the system cases fell by the wayside, nor the reasons that these cases failed to generate any kind of conviction.
Look, if we’re going to do this properly we need to set a few basic ground rules as follows.
1. Conviction rates measure the number of cases which result in a conviction.
2. Attrition rates measure the number of cases which don’t result in a conviction.
3. For any given statistic you need to specify both the numerator and the denominator, so…
The 58% conviction rate is rate for any conviction for any offence relative to the number of cases prosecuted where rape was the primary offence.
The 13% conviction rate is the rate for any conviction for any offence relative to number of rape complaints recorded by the police in the audited dataset from which these figures were taken while the attrition rape for the same dataset is the number of cases for which no conviction was obtained relative to the number of rape cases recorded by the police in the dataset.
By the same token, 33.34% is the estimated (i.e. non-audited) conviction rate for rape offences only for prosecutions in which rape was the primary offence for 2010/11, the year for which I’ve been working on the figures, while a full breakdown of the attrition rates relative to prosecutions looks like this…
- Acquittals – 16.4%
- Unreliable witness – 4.93%
- Conflict of evidence – 4.24%
- Essential legal element missing – 3.75%
- Evidence of victim does not support case – 3.5%
- Other (unspecified) reasons – 3.31%
- Victim retraction – 3.25%
- Administrative finalisation – 1.1%
- Effect on victim/witness mental health – 0.5%
- Victim non-attendance – 0.4%
This is all useful information, of course, but we’re still on dealing with case outcomes at the prosecution stage – none of tis tells us much of value about the 72.9% of reported rape cases (estimated) which failed to spawn a prosecution – that’s the attrition rate for reported rapes at the police investigation stage and, sadly, we don’t have any reliable data on the reasons behind this high attrition rate, not least because the police don’t currently work to any kind of consistent recording/reporting standards.
Why is this important? Because the rhetoric of rape, which largely propounds myths, puts survivors off reporting the crimes committed against them, making them perceive that the system designed to assist them is actually wholly against them.
Last week, Mumsnet released a survey of its users as part of its We Believe You rape awareness campaign. Sixty-eight per cent of respondents said low conviction rates would make them hesitate to report a rape due to low conviction rates – clearly they had heard the 6% figure too.
Clearly? Not necessarily.
The Mumsnet survey is not a scientific poll, nor is it based on a nationally representative sample, so we have to be rather careful in drawing any concrete inferences from its results.
For example, of the 1609 women (presumably) who completed the survey, only 1294 answered the question which asked which factors would make them feel hesitant about reporting a rape to the police, if they were unsure as to whether or not they would make such a report.
19.6% of survey respondants didn’t answer that question from which we can infer what? That all these women were sure that they’d report a rape to the police, no matter what, or may that all these women would not report a rape at all, or maybe a mix of both.
We can’t actually be sure, although a separate question did reveal that amongst the women who filled in the survey who had been raped, at some point, 83% had not reported the rape to the police, a figure which is consistent with the figures for under reporting you get when you compare the data dfrom the British Crime Survey to the figures for reported rape cases.
Moreover, conviction rates were not the only factor identified in the survey as – maybe – having a significant influence of whether or not women would report a rape to the police. The 68% who appear to find conviction rates somewhat off-putting was the highest response for any of the survey options but that doesn’t mean we can disregard the fact that 61% expressed concerns about having to reveal personal details in court, while 59% were seeming worried about the potential impact on their family and 50% were worried that family members might think that the rape was their own fault [sometimes words fail me]. Then there’s the 53% who cited embarrassment or shame as a reason they’d hesitate over reporting a rape, while 47% appear to feel that the police’s attitude leaves much to be desired and 46% were none too impressed with the charge rates, all of which tend to suggest that many of the women who completed this survey were rather more clued in on some of the nuances about conviction and charge rates that Amanda assumes.
An over-concentration in the media and in anti-rape campaigns on the conviction rate relative to reported cases rather to prosecutions may well create an unhelpful impression of the overall performance of the criminal justice system, but swapping out one bad statistic for another isn’t a magic bullet to fix the system – even allowing for the limitations of the Mumsnet survey, the evidence it does provide clearly supports the view that there a number of factors that influence women when it comes to decided whether or not to report a rape to the police.
Again, we come back to HL Mencken’s maxim that…
Explanations exist; they have existed for all time; there is always a well-known solution to every human problem — neat, plausible, and wrong.
Sod ‘Keep Calm and Carry On’, if ever there was anything that should be appear on a wall poster in every single office in Whitehall, its that statement.
It isn’t just conviction and attrition rates that are an issue and which impact on rape reporting. The whole discourse about rape is mired in myths and half-truths, which leads survivors to believe the system is against them and that they are unlikely to be believed.
Funny you should mention that, given that the CPS gave the following explanation for their acquittal rates at trial.
The rise in jury acquittals tends to suggest that more work may be needed with partners to address public awareness and challenge myths and stereotypes, which have traditionally led to high jury acquittal rates in sexual cases.
And the reliance on an appeal to tradition suggest that the CPS are making bullshit excuses in the interests of political expediency when they should be looking to conduct substantive longitudinal research to identify where, exactly, the problems are.
To be clear on this point, myths and stereotypes may well be an issue but, overall, the research evidence in this area is rather inconclusive, contradictory and, in some cases, counterintuitive. One of my favourite papers on the subject of jury attitudes comes from the US, where juror selection is a significant and hotly contested part of the trial process, and contrary to what many people might expect the study found that, in rape cases, the ideal juror – from a prosecution standpoint – is a middle aged male Italian-American, and preferably a family man with at least one daughter, while young female professional are best avoided if you want to secure a conviction.
If you can’t figure out why, given a choice between Giuseppe and Jennifer, rape prosecutors in the US will take Giuseppe every time, then drop me a note in comment and I’ll explain.
Last week, the court of appeal heard the appeal against conviction in what has become known as Sarah’s case. In 2010, ‘Sarah’ reported that her husband had raped her. After the prosecution had started, she retracted that allegation; then subsequently withdrew the retraction. Sarah was convicted of perverting the course of justice, and sentenced to eight months in prison.
In respect of the sentence the court of appeal moved quickly, and Sarah had her sentence reduced to a community sentence after she had served three weeks in prison. The director of public prosecutions, Keir Starmer, made it clear he was appalled Sarah had been prosecuted, and swiftly changed the guidance so that in future prosecutions can only be brought in such cases with his permission.
Sadly, the way Sarah’s case was commonly reported last week can only be described in some instances as hysterical. There were the expected headlines of “rape victim sent to prison for withdrawing allegation” with little explanation that such a thing is highly unlikely to happen again, owing to the change in guidance, and that the court of appeal had very quickly ordered Sarah’s release.
I asked Vera Baird, the former solicitor general and one of the commentators on Sarah’s case, to make it clear that we were unlikely to see this situation repeated because of the change in guidelines, in order to ensure survivors would not fear that what happened to Sarah would happen to them. Sadly, she responded that I was “optimistic” and that the police would simply use the charge of wasting police time in order to get around the guidelines, and that the Crown Prosecution Service would not change the guidelines for that offence to match the change made to perverting the course of justice. As the law stands, the guidelines on wasting police time are the same as perverting the course of justice, in that prosecutions require the consent of the DPP – it is unfortunate the former solicitor general did not seem to know this.
Hard cases… eh?
I don’t want to get into debating the detail of this case as it been more than adequately covered by legal bloggers, rather there’s a more general point to be made about victim withdrawals and that’s that this issue, like so many others that contribute to the high attrition rates we’re seeing at the investigation stages, is woefully under-researched and poorly understood as a result.
In only one-fifth (21%, n=66) of cases did the police pro formas provide any details on the reasons for victim withdrawals, although it is notable that, compared to no crimes and insufficient evidence, there were far fewer references to either victim characteristics or evidential problems here. Where there is information, for half of the complainants (n=33) fear of the court process was the deciding factor, 15 refer to complainants’ uncertainty about whether they were raped, including some thinking it may have been consensual, seven state that the person ‘wants to put the event behind them’, and another seven complainants are reported to have made the decision after consultation with their family. From the police information, in at least 13 of these cases the assailant was a current/ex-partner, which may have played a part in the decision (in six cases it clearly did).
Early victim withdrawal, at the police investigation stage, accounted for 17% of the cases that failed to generate a prosecution in this study, and yet in only a fifth of these cases was any data on the complainant’s reasons for withdrawing their allegation collected.
What I would like to see is openness about rape. The myths, the half-truths, and the nowhere near the truths do not help survivors come forward, never mind helping them come forward with any confidence. I would like to see the full picture being presented. I accept there is still huge scope for improvement in the criminal justice system in the way victims are treated. But change has happened, and will continue to happen, and we need to get that message out there, too.
Of course we should be more open with women about rape, the question is what are we supposed to open with when the availability of decent quality research evidence is somewhere close to the annual sales figure for rocking horse shit?
I should point out here that Amanda’s profile at the Guardian states that…
She blogs at Beneath the Wig and is putting together the 58% Campaign, aiming to educate on the whole picture of rape.
And clarify matters further, here’s all the ‘putting together’ that she’s managed so far…
I want to see a campaign that yells loudly about the 58% figure. One that says ‘if there is evidence to take this to trial, you are more likely than not to see your offender convicted’ – but also one that is honest about the difficulties in getting a case to trial – not to put people off making a complaint, but to enable them to make a proper, informed and educated assessment of whether they want to make a complaint.
I guess we should just ignore the fact that ‘if there is evidence to take this to trial’ is a pretty big ‘if’ – almost 16,000 rapes were reported to the police in 2010/11 of which, based on current CPS performance data, only a little over a quarter stand any chance of ever seeing the inside of a courtroom, and only 1 in 3 of the cases that do make it all the way to court will result in an actual conviction for rape.
If anyone wants to campaign for something here, how about we campaign for some good quality research – at least then we’d have something worth being open with.
22 thoughts on “Myths about myths about rape conviction rates”
An incisive article highlighting how vested sector interests can coopt vital data gathering and defer the public good – via dam lies and selective statistics
What I am interested in is how the rape statistics here compare with other serious offences in terms of conviction and attrition rates. By way of example, what are the comparative statistics for, say, infliction of grievous bodily harm with intent? Although the rape statistics are depressing to say the least, how do they compare with other offences into which similar levels of investigative resource could be expected? Are these conviction and attrition rates unique to rape, as is implied? That really is the key question. If the answer is yes, there is some serious work to be done!
For murder and related crimes: conviction at trial rates are much higher, attrition before trial rates are much lower.
Conviction at trial rates are higher than 58% for almost every crime, though. (CPS guidance on when to prosecute of course makes any rate lower than 50% for any crime extremely unlikely, though rape was in that position until fairly recently)
Attrition is harder to judge: rape prosecution has the significant advantage over many serious crimes that the question of “was it the suspect or someone else who committed this crime?” is usually fairly easy. On the other hand compared with GBH or even murder there will on average be fewer witnesses excluding the victim and perpetrator(s). So direct comparisons with other crimes – even other serious crimes – aren’t worth that much, I think.
If you compare region-by-region police and CPS performance it’s fairly clear that the attrition rate is considerably higher than it theoretically could be, with an attrition rate around 50% – including conviction for lesser offences, admittedly – to be expected for a hypothetical police/CPS region that performed in all areas equivalent to the current best police/CPS region in each area. (and I don’t think – and I expect the best regions would agree on this or they wouldn’t have got as far as they had – that current best practice is equal to possible best practice)
So actually it looks as if the important statistic is the attrition rate in rape cases – if they go to court, things are not necessarily as bad as is sometimes thought – it’s the fact that so few ever get anywhere near a jury which should be the first focus of action perhaps?
The problem is that CPS under general guidelines must only send a case to court if they believe there’s at least a 50% chance of conviction.
That the total rate is only 58% – given that about a third of those are guilty pleas for which the evidence must often have been at well over 50% strength – suggests that CPS are already sending more to trial than they normally would prefer to. Focusing on sending them more cases will just mean that they lose more cases.
Under FOI 10% minima figures were returned by various constabularies for false/malicious rape allegations.