Dempublicents1 wrote:The polygraph itself may not necessarily influence the decision, but the implications of having it as standard policy for rape cases says something about the police in question. I highly doubt that they employed the polygraph for every report of a crime, yet they made it standard practice in rape cases. This suggests that they already believed that a large number of rape cases were false reporting - that they were already biased to see the alleged victims as liars.
And that can be true even if the polygraph isn't employed. There are all sorts of ways in which a police officer can indicate to a victim that he is not inclined to believe her and that can lead to someone recanting.
Note: I'm not saying every case was a false recantation or anything like that.
I would be surprised if more than a comparatively small fraction of recantations were false.
It's actually not easy to admit that you fabricated everything.
But I am disputing the idea that this precinct was somehow an ideal pick for such a study. And using a university setting afterward - a setting which is undoubtedly a different demographic from the population as a whole - isn't any better.
Well, it may not be ideal, but it controls for a number of "known" problems that people have brought up in trying to deal with official data, such as police departments being overworked, understaffed, and considering rape reports a low priority, or declaring them false reports based on what some internal "expert" thinks (as in some of the other studies cited in the 2006 lit review).
Are there still uncertainties in what's going on? Yes, and many of these problems are linked to the fact that it's basically a review of police paperwork and claimed policies, not involving on-the-ground reviews of what officers are actually doing.
As I pointed out in my first post on the article, many of their "typical" cases don't seem to have been people who intended to go to the police. Many of them claimed that a rape had happened to a doctor or parent or other authority figure and were then brought to the police by said authority figure. It sounds like a lie that got out of their control, rather than an intent to report to the police.
Yeah. That fits with what I'm saying, doesn't it? Most of the benefits you can reap out of claiming rape don't actually require police involvement. You might find yourself obliged to file a report in order to maintain credibility among the people you're trying to convince.
In addition, many of the "typical" cases were women who did not name any specific assailant.
Yup. But what can happen with those is that you get unlucky and match the description fabricated.
Back at Appalachian, I remember the day that campus got locked down because someone didn't want to pay for a door. See, they damaged the door on their apartment right behind the convocation center, and didn't want it coming out of their deposit, so they told the landlord that their room had just been broken into by some dude with a gun. They offered a detailed description - IIRC, it was something like "wearing a Pink Floyd t-shirt, blue jeans, and one red shoe" or something like that, who then took off running out of the complex (i.e., towards campus).
This is reported to the police, who then put out an alert to be on the lookout for such-and-such. Someone calls in a sighting - someone matched that description and was sighted on campus. Campus is locked down for most of the day while the police poke around campus. Students are freaking out. (Virginia Tech is about a 3 hour drive from Appalachian State, and there had been a rather major shooting spree there not long before.) All because this totally fabricated crime happened to line up neatly with the description of someone walking around campus.
Like I said, I haven't really seen any good evidence that false reporting in rape cases happens any more or less often than similar false reporting of other crimes. I'm just as skeptical of claims that there are far fewer cases of false reporting in rape. I understand the reasons that people believe either to be true, but reasons that make sense aren't actual hard evidence.
Read the 2006 overview. Other crime statistics on false reporting are if anything even sparser, but the FBI has released both figures for comparison for a class of "unfounded" crime reports (8% for rapes, 2% for other crimes on average; the FBI 'unfounded' category isn't actually quite exactly the same thing Kanin's talking about.)
I think equating "not guilty" with "innocent" is an incorrect characterization of the justice system. "Not guilty" means that a jury was not convinced beyond a reasonable doubt, not that they were convinced that the accused was innocent, much less that the person was in fact innocent. Rape is, by it's very nature, going to be incredibly difficult to prove beyond a reasonable doubt - regardless of whether or not it happened.
What happens in a trial?
OK, there's the standard of beyond a reasonable doubt - but there's a fuck-ton of evidence-gathering and argument involved in a case that goes to trial, and the jury needs to come to a unanimous decision. Sure, there are some cases where a jury would find probable cause but not beyond a reasonable doubt, but we often get to the end of a trial feeling pretty sure about what actually happened. We did DNA testing, we reviewed alibis, we cross-examined witnesses of all kinds, we've eliminated as much uncertainty as we could, and then you lock twelve people in a room together and try to figure out whether or not the accused is guilty, holding out for a unanimous decision in order to reduce the chance of random error and to make sure that any cases that seem uncertain get thrown into a retrial.
The jury is supposed to presume innocence until proven guilty, but they make both kinds of mistakes. Letting the guilty go free is supposed to happen substantially more often than false convictions. Risinger makes a case using DNA evidence for about 5%. False positives in DNA testing actually happen more often than you might expect, and we can expect to see a number of wrongful convictions made using DNA matches; I would be totally unsurprised to see the false positive rate for convictions at around 10%.
Let me construct a simple model for you. Let's say that fully three quarters of the people brought to trial are actually guilty. Let's say that jurors decide whether they think the accused is guilty or innocent, then negotiate extensively and rarely hang (in practice, hung juries happen on the order of 1%), with the minority eventually being brought on board with the majority, and an initial 50-50 tie being split down the middle. It's a nice simplifying assumption, leaning on Condorcet's jury theorem. We could make things more complicated, but it'll illustrate how juries are supposed to work, and the sort of relationship we expect to see between false positives and true negatives.
Now, suppose an innocent person has a 25% chance of being considered guilty by an individual juror, and this goes up to 60% for a guilty party. Thanks to the accuracy amplification of Condorcet's Jury Theorem (i.e., the binomial distribution), we then would have a 3.4% chance of convicting someone given that they were innocent, and a 25% chance of failing to convict a guilty individual. Overall, that would give a 57% conviction rate, with a 6% false conviction rate and 57% of those being found not guilty actually being innocent.
Those are the sort of figures we're looking at on the conviction rate and false conviction rate - you can play around with parameters as you like, but if our convicted population contains around a 5% innocent subpopulation and a >50% conviction rate, our exonerated population really should contain a very significant innocent portion.
IMO, probably a majority for crimes in general.
Not all that surprising. I don't know about the UK, but I know that prosecutors in the US are not required to actually pursue every possible case. Most won't even bother with a case unless they're pretty certain they can win it.
As you mentioned, this is actually a big part of the problem with many of the statistics out there. They report "unfounded" cases, but these include the cases in which the prosecutor (or the police) didn't find enough evidence to pursue a prosecution.
Much as rape is, by its very nature, difficult to prove beyond a reasonable doubt, it is also, by its very nature, difficult to quantify accurately.
I'm not sure that fits. If that's the definition of the "unfounded" cases, because the FBI reported that only 8% of rape cases were "unfounded," that would mean the prosecutor felt there was enough evidence to proceed with a prosecution for 92% of the reports. Which would in turn mean 92% were able to be brought to trial. Which seems rather high.
I found statistics for Ireland first. 1 in 3 reported rapes are brought to trial. If you like, look for ones for the US / UK. I know that in the US, we're going to expect to see widely divergent rates from locality to locality, because local DAs are often elected and have highly variable approaches to choosing what cases to prosecute.
Now, to go back to the article I linked to initially... if you report a very low conviction rate, it discourages reporting actual rapes, because victims believe the police won't do anything anyway (low return for reporting), but doesn't discourage false accusations (the motives for which aren't particularly tied to rape conviction). This then leads to a higher ratio of false reports to true reports, which signals to the police that rape victims are likely to be lying. This signal leads them to treat victims distrustfully, which leads to lower follow-up rates, discouraging in blanket fashion both actual victims and false accusers from continuing what has become an even less pleasant process, closing a feedback loop.
I think many of the ways that people are choosing to measure this sort of thing are just serving a political agenda. Some groups have decided to paint the picture as horrifying as possible. Some have decided to try to pretend there is no rape problem. I think a lot of it comes down to the same thing as Risinger's concern with Scalia's comments: People are picking conveniently large [small] denominators that don't actually represent the population that they seem to be claiming it does. From reports to convictions, for example, we have a significant amount of chaff in the false accusations and in, as well, forgiving victims. Except for statutory rape cases (which aren't actually prosecuting the same crime), the victim usually has the choice of whether or not to press charges, and may choose to decline to do so for reasons of his or her own.
If we want to consider the true odds of obtaining a rape conviction, we're interested in the population that not only is willing to report the rape to the police, but also:
-Have told the truth in doing so.
-Will continue cooperating with police/lawyers/et cetera and press charges through to the conclusion if the DA is willing to take those charges to trial.
This is a rather smaller population than those making reports [and rather larger than those actually making it to a trial]. If 6% of reports lead to a conviction, that doesn't quite tell us what percentage of reports are, in fact, true reports, and what percentage of victims are actually willing to prosecute. Most know their attacker, and know, suspect, or will learn during the process that rape is considered a very serious crime, and pushing charges through will change the lives of not only the attacker, but many other people that they know. The percentage of people who are willing to inflict those consequences and are making a true accusation who obtain a conviction will be substantially larger, and seems more relevant.
