You know what... fair enough. Retracted.
(I still stand by the example making a valid point though, fictitious though it may be)
False Rape Accusations Far Less Common Than Believed
- Thread starter luchog
- Start date
You know what... fair enough. Retracted.
(I still stand by the example making a valid point though, fictitious though it may be)
Information Analyst
Penultimate Amazing
I don't see anyone claiming that 5% of reported rapes means that 5% of "all" rapes are false reports. We can only apply that 5% to those actually reported, but it isn't valid to reduce that 5% to a much smaller percentage of the estimated total number of actual rapes, an estimate which may or may not include potential false reports.
If, rather than having only 10% reporting, there was in fact 50% reporting, it would seem exceptionally unlikely that the actual number of false reports would remain the same or little more than that seen with 10% reporting. The opposite would seem inevitable, due the the various ways in which a report may be determined as being false.
Okay seriously unless we assume there is some ultimate power with the magical ability to determine guilt or innocence (other related concepts like honest and accuracy) with 100% precision all of these numbers have to be a best guest.
Belz...
Fiend God
No need to be so dismissive. Illustrating argument in ways easy to understand and remember has been a thing for thousands of years. No need to be flustered just because.... because what, really? What set you off this time?
Dave Rogers
Bandaged ice that stampedes inexpensively through
The post I originally replied to seemed to be claiming pretty much that.
Nor is it valid, really, to increase it, or keep it the same. In fact, we have no data to work from on the dependence of the rate of false reporting on the rate of true reporting. That's a major problem, of course, if we want to put in place measures that will increase true reporting without encouraging large numbers of false accusations.
Dave
Ziggurat
Penultimate Amazing
- Joined
- Jun 19, 2003
- Messages
- 60,896
Unfortunately, no. The problem is by its nature almost impossible to measure precisely. The 5% rate from that source is most likely a lower bound.
There's a bit of a Schrodinger's cat aspect to this too: if we were able to determine the truthfulness of falseness of a claim reliably, that itself would serve to reduce the number of false claims, since the probability of passing off a false claim can affect the willingness of a person to make a false claim. So if you invented a method of reliably identifying false rape reports, then measured the false report rate precisely, you couldn't extrapolate that rate back to before you had that method.
This. Far too much of this discussion is based on us assigning a value to things which are defined as unvalued.
We don't know how many of "Undetermined Value X" is determined to be "Value Y" because that's what undetermined means.
Information Analyst
Penultimate Amazing
I suppose that what this comes down to is that we currently live in societies in which the majority of people who are raped are extremely reluctant to report it, but of the minority that do report 5% are false. Five percentage are still confident enough that their claim is valid, or that nobody will discover a malicious false claim. It would be fantastic if our societies were such that every person who had been raped - or at least the vast majority - felt confident enough to report it, but it follows that if that was possible, it would be also be possible for more people to be either mistaken or malicious.
theprestige
Penultimate Amazing
The problem with the article cited in the OP is that it's not a best guess, it's quite likely the *worst* guess.
Belz...
Fiend God
Actually we do. Statistics is exactly the tool we have to do this.
Exactly. And if we reduced the burden of proof to 'listen and believe' because the rate of false reports is low, you can bet some real money that it would go way up.
Dave Rogers
Bandaged ice that stampedes inexpensively through
Statistical analysis is useless without data. What data do we have on how the rate of false accusations changes when the rate of reporting changes? You can't extrapolate a trend from a single point.
Dave
Ziggurat
Penultimate Amazing
- Joined
- Jun 19, 2003
- Messages
- 60,896
Statistics won't really work here. The problem is that you cannot assume that the population of reported cases is a representative sample of the total population of cases. The fact that they are reported means they went through a selection process, one which is definitely not random. We should expect that the population of cases which are reported is going to be different than the population of cases which are not reported, but precisely because they aren't reported, we don't have any good handle on the characteristics of the unreported case. And since these populations are probably different, the percentage of false cases within them could easily be different as well, possibly very different. But we don't have enough information to determine how they are different. So we cannot use statistics to extrapolate the characteristics of one population to another population that we should expect is different but in unknown ways.
Belz...
Fiend God
You're right but the point that statistics is how we'd find a reasonable rate stands. Dave can't just say it's zero.
Ironic.
Dave Rogers
Bandaged ice that stampedes inexpensively through
Not really, since I was never actually arguing for any particular rate of false reporting if the rate of reporting changed. My apologies if that wasn't clear.
Dave
theprestige
Penultimate Amazing
I think one thing that might help keep the discussion clear is a bit of standardized terminology. Here's how I think of the problem:
On the one hand, we have claimed incidents. Maybe a person never tells anyone other than themselves that the incident happens. That's a claim. Maybe they tell their parents, or their therapist, or their BFF, or a sociological survey, or their university administration. These are all claims. Depending on who receives the claim, it may or may not be tested with some unknown degree of rigor. Those are the claimed incidents, or CLAIMS
On the other hand, we have incidents that are reported to the police. These are the REPORTS, and they are a subset of the CLAIMS.
The REPORTS are tested, with varying degrees of rigor, for a variety of reasons. In one subset of REPORTS that are tested with a particular degree of rigor, we find them false at a rate of about 5% for that one subset of REPORTS.
We don't know what the falsehood rate is for the complete set of REPORTS, because outside of that one subset, we're not testing REPORTS with an adequate degree of rigor.
And we don't know the falsehood rate of CLAIMS in general, because we're not actually testing them very well at all.
And we can't extrapolate from CLAIMS to INCIDENTS, because we have no idea how many false CLAIMS there really are. Because we're not properly testing CLAIMS at all - we're only testing a subset of a subset, in a way that doesn't really admit extrapolation to the complete set.
On the one hand, we have claimed incidents. Maybe a person never tells anyone other than themselves that the incident happens. That's a claim. Maybe they tell their parents, or their therapist, or their BFF, or a sociological survey, or their university administration. These are all claims. Depending on who receives the claim, it may or may not be tested with some unknown degree of rigor. Those are the claimed incidents, or CLAIMS
On the other hand, we have incidents that are reported to the police. These are the REPORTS, and they are a subset of the CLAIMS.
The REPORTS are tested, with varying degrees of rigor, for a variety of reasons. In one subset of REPORTS that are tested with a particular degree of rigor, we find them false at a rate of about 5% for that one subset of REPORTS.
We don't know what the falsehood rate is for the complete set of REPORTS, because outside of that one subset, we're not testing REPORTS with an adequate degree of rigor.
And we don't know the falsehood rate of CLAIMS in general, because we're not actually testing them very well at all.
And we can't extrapolate from CLAIMS to INCIDENTS, because we have no idea how many false CLAIMS there really are. Because we're not properly testing CLAIMS at all - we're only testing a subset of a subset, in a way that doesn't really admit extrapolation to the complete set.
Aren't we in regards to justice only concerned with the reported to authorities rapes? (Not that the rate of rape is irrelevant or we shouldn't be looking at what is happening outside reports but as a society we can only deal with what we can know.)
Therefore if the rate is 5% is that high enough that we need to alter how our justice system usually works (I. E. Concealing identities of those being prosecuted)? To me it doesn't seem high enough yet of course that represents 5 lives out of a hundred that could be totally ruined by a false report. Perhaps we should look at this from the other end, following a false report and potential trial how can we restore someone's reputation?
Therefore if the rate is 5% is that high enough that we need to alter how our justice system usually works (I. E. Concealing identities of those being prosecuted)? To me it doesn't seem high enough yet of course that represents 5 lives out of a hundred that could be totally ruined by a false report. Perhaps we should look at this from the other end, following a false report and potential trial how can we restore someone's reputation?
Ziggurat
Penultimate Amazing
- Joined
- Jun 19, 2003
- Messages
- 60,896
The concept of "justice" extends beyond the formal legal system. It's not justice for someone's reputation to be unfairly ruined even if no legal sanctions are imposed. Furthermore, there are authorities other than the police. Campus administration can and do evaluate claims of rape and impose serious penalties as a result. But a rape reported to campus authorities and not to police is not considered "reported" by the paper referenced in the OP.
Seems to me that if you want to shield the identities of the accusers because of the sensitivity of the crime, you should shield the identities of the accused as well, at least absent a conviction, for the same reason.
And again, the 5% rate isn't the relevant rate for that question. It doesn't include claims which are false but haven't been demonstrated to be false. It also doesn't include real claims but where the identity of the perpetrator is mistaken. But someone who is mistakenly identified is just as innocent as someone who is maliciously accused.
We can't. Which is why caution beforehand is warranted.
Belz...
Fiend God
I don't think you can. Not without changing social attitudes.
I agree, we know social attitudes can change so it may be possible.
Chris_Halkides
Penultimate Amazing
- Joined
- Dec 8, 2009
- Messages
- 12,394
A difficult number to pin down
There have been several discussions of how difficult this number is to obtain. National Review's Jason Richwine here. Bloomberg's Megan McArdle here. Slate's Cathy Young here. Reason's Robby Soave here. KC Johnson and Stuart Taylor devote the second chapter of their book on Title IX and campus sexual misconduct to this topic. Francis Walker has an insightful two part series on how to mislead using statistics with respect to the percentage of false versus true accusations. Part I here. Part II here.
Last edited:
Meadmaker
Unregistered
- Joined
- Apr 27, 2004
- Messages
- 29,033
Agreed. The article from the OP made use of two numbers in its premises.
It claimed that 5% of rape reports are false.
It claimed that 90% of rapes are not reported.
Both of those claims are contentious. I would not accept either of them without some demonstration of how they were obtained.
However, regardless of how they were obtained, and regardless of whether or not they are accurate, the conclusion of the article is still complete codswallop. The conclusion does not follow from the premises.
sadhatter
Philosopher
- Joined
- Jun 4, 2009
- Messages
- 8,694
For sure, we need to stop listening to ******* idiots and **** stirring folks.
I'm sure it will be out of style in 5 or so more years. We are already seeing the tempest of ******** is causing when that outweighs the satisfaction of getting under the other sides skin , it will change.
Brainster
Penultimate Amazing
- Joined
- May 26, 2006
- Messages
- 21,874
Bleeding-heart liberal judges?
Check.
In fact if you real the articles you linked to, most of the judges were saying that the reason they didn't give more time to the perp was because they're more interested in rehabilitating the rapists, and feel they wouldn't get the psychological treatment they needed.
theprestige
Penultimate Amazing
Is it just me, or did The Cut fail to cite any of this?
Delvo
Дэлво Δε&#
A comparison of the real numbers with what people think the numbers are is the wrong way to look at it. I don't believe that people who are concerned about false accusations have that feeling because they have any particular number in mind for how common it is. They're just thinking about how we should handle it if and when that situation does arise, which is the same issue regardless of how often or rarely it does. And there are a couple of big reasons for people to have that on their minds without any concept of how common or rare it is ever entering the picture:
1. The importance of the general principle (for all crimes, not just rape) of innocence being presumed and guilt needing to be proven
2. A tendency to imagine one's self inserted into a story: "I know I wouldn't do that, so it's automatic that any accusation against me would be false"
1. The importance of the general principle (for all crimes, not just rape) of innocence being presumed and guilt needing to be proven
2. A tendency to imagine one's self inserted into a story: "I know I wouldn't do that, so it's automatic that any accusation against me would be false"
I remember those recent allegations of senior British politicians being a part of murderous homosexual pedophile groups extremely similar to the kind of satanic cult panic from the 80's and 90's.
Even in such cases the police should apparently investigate them because it's imperative to treat all sexual abuse allegations seriously and never question the alleged victims believability.
Even in such cases the police should apparently investigate them because it's imperative to treat all sexual abuse allegations seriously and never question the alleged victims believability.
Chris_Halkides
Penultimate Amazing
- Joined
- Dec 8, 2009
- Messages
- 12,394
counting false convictions
From the article cited in the OP, "I was only able to find 52 cases in 25 years where a conviction was later overturned after a wrongful conviction based on false rape allegations. In the same period, there were 790 cases where people were found to be wrongfully convicted of murder.” For what it’s worth, 790 divided by 52 is 15.2..."
I would like to know more about these numbers. Some false convictions for murder involve both rape and murder. From this link on DNA exonerations, "We classify exonerations by the 'worst crime' for which the defendant was convicted and later exonerated." And "All told, rape is an element in 78% of known DNA exonerations (329/420). In that context, the shift since 2008 mostly means that DNA exonerations are increasingly about rape-murder rather than rape alone." Within the 790 wrongful convictions for murder were there rapes or not?
Some false convictions for rape or sexual assault were based on faulty identifications of the perpetrator, even though the offense itself happened. I would be hesitant to call such a situation a false rape allegation (I would call the allegation partially true). It is unclear to me whether or not these cases were among the 52.
If memory serves me correctly the same kind of arson evidence that was sufficient to get Todd Willingham convicted was seen as faulty in a parallel case in Texas, leading to an exoneration. For this reason among others, I would urge caution when using statistics on false convictions. In other words, the number of exonerations is likely to depend on many factors, and I am not sure that the 15.2 number means very much.
From the article cited in the OP, "I was only able to find 52 cases in 25 years where a conviction was later overturned after a wrongful conviction based on false rape allegations. In the same period, there were 790 cases where people were found to be wrongfully convicted of murder.” For what it’s worth, 790 divided by 52 is 15.2..."
I would like to know more about these numbers. Some false convictions for murder involve both rape and murder. From this link on DNA exonerations, "We classify exonerations by the 'worst crime' for which the defendant was convicted and later exonerated." And "All told, rape is an element in 78% of known DNA exonerations (329/420). In that context, the shift since 2008 mostly means that DNA exonerations are increasingly about rape-murder rather than rape alone." Within the 790 wrongful convictions for murder were there rapes or not?
Some false convictions for rape or sexual assault were based on faulty identifications of the perpetrator, even though the offense itself happened. I would be hesitant to call such a situation a false rape allegation (I would call the allegation partially true). It is unclear to me whether or not these cases were among the 52.
If memory serves me correctly the same kind of arson evidence that was sufficient to get Todd Willingham convicted was seen as faulty in a parallel case in Texas, leading to an exoneration. For this reason among others, I would urge caution when using statistics on false convictions. In other words, the number of exonerations is likely to depend on many factors, and I am not sure that the 15.2 number means very much.