Rape statistics

Due to the (in my impression so far) absurd arrest of Julian Assange, I have seen a number of recent threads around the topic of rape. There are several oft repeated faulty claims made that I wish to address here, based on a stereotypical comment. Unfortunately, the post and the comment thread^e appear to have been deleted in the mean time, but to summarize from memory:

1. Only 2 % of all rape charges are false. [Very, very incorrect.]

2. Only one in five of all rapes are reported. [Impossible to know, but very likely an exaggerated claim.]

3. Only 10 % of all reports lead to a conviction. [Semi-true, but highly misleading.]

4. Rape accusers are treated worse than the alleged rapist, have problems through not being believed, or similar. [Very incorrect and/or misleading.]

My original reply (translated to English):

The “2 %” claim is no longer taken seriously and there is reason to believe that it simply is made up (cf. the links below). Even the claim that only one in five rapes is reported is very far from being a consensus number, and could very well be something that feminists babble about for rhetorical purposes.

A quick search for “false rape charges” gave, among others, the following links:

http://www.falserape.net/false-rape.htm^e

http://www.falserape.net/falserapeafa.htm^e

http://www.foxnews.com/story/0,2933,194032,00.html^e

http://lawprofessors.typepad.com/crimprof_blog/2004/12/2_false_rape_st.html^e

http://www.billoblog.com/?p=134^e

(In addition, I recommend http://falserapearchives.blogspot.com/2010/01/false-rape-primer.html^e as a source for further links.)

To dig a little deeper:

1. Claims about false reports and under-reporting:

   The true rate of false reports is at a minimum 20 % and may be as high as 60–70 % based on the above links. To note, however, is that a false accusation is not always made deliberately, but (depending on the definitions used in any particular piece of research) can include mistakes of identity made in good faith. This distinction is of low relevance when it comes to “innocent until proved guilty”, but is important to bear in mind in other contexts—e.g. when some feminist commenter starts a rant about how misogynistic it would be to claim that half of all rape accusers are liars.

   The “one in five” is possibly caused by feminist mis-definitions of “rape” to include things that the law, a sensible person, and the alleged victim herself, do not consider anything of the kind. Notably, it is not uncommon for such mis-definitions to artificially inflate the number of claimed rapes to several times its true size.

   An interesting perspective is provided by a post that I encountered a few months ago, which contained roughly the following line of discussion (I apologize for not being able to give due credit):

   If the probability of a man committing a rape is p-rape and the likelihood that a woman will raise a false accusation is p-false, and further assuming that the two factoids of 2 % and one in five are true, then the number of reported rapes per male (or female) citizen is roughly 0.2 * p-rape + p-false. Further, by assumption, the quotient (0.2 * p-rape) / p-false would be 0.98 / 0.02 = 49. By implication, p-rape = 245 * p-false. In other words, men would have to be 245 (!!!) times more likely to commit rapes than women to make a false accusation—a claim that is so patently absurd and misandristic that the mind boggles.

   Even running through this calculation with a non-false report rate of 50 % (instead of 20 %) and a false report rate of 20 % (the minimum from serious investigations, instead of 2 %), we land at p-rape = 8 * f-false. In other words, men would still need to be eight times more prone to commit rapes than women to commit false accusations. Based on my experiences with men and women to date, I find this extremely hard to believe, and am correspondingly inclined to assume that the rate of false accusations is noticeably higher than 20 % (and possibly that the report rate is noticeably above 50 %—under no reasonable circumstances can it be the lowly 20 % claimed by some feminist propagandists). Here, however, it can make a difference whether deliberate false accusation or false accusations in general are discussed (cf. above).

2. Conviction rates:

   I have been unable to find Swedish statistics (the alleged 10 % was with reference to Sweden) on short notice; however, I did find a very interesting article on the British situation^e, where alleged numbers of just 10 % and 6.5 % are discussed. In a nut-shell: The true rate of conviction, after removing e.g. instances ruled as “no crime”, is roughly 50 %. The 10 % number is here referred to as “rate of attrition”, to which the article gives the following numbers for murder, rape, and “violence against the person”: 14 %, 6.5 %, and 4 %. Correspondingly, the rate is by no means remarkable. When considering the higher rate of false accusations for rape, the greater practical problems to provide proof that a crime has at all taken place, the possibility that a conviction is made for a lesser crime (cf. snoozeofreason’s comment to the article), whatnot, my subjective impression is that the rate for rape is higher than it should be when using other crimes as a baseline. I strongly recommend reading the linked-to article.

   The greatest error here, however, is to make the a priori assumption that almost all accused are guilty and to claim that a conviction/attrition rate of 10 % would imply that the justice system is defect. Notably, it is equally possible to turn the situation around and see the low conviction/attrition rate as a proof of many false accusations.

   More generally, a high conviction/attrition rate is only good when “false positives” can be kept down: Achieving a “perfect” conviction rate would not be hard, but doing so would fill the prisons with innocent people. Making convictions is not an end in it self—what matters is sentencing so many guilty as possible without putting innocent people in prison.

3. Treatment of the accusers and the accused:

   Frankly, this sounds mostly like yet another case of feminists claiming the exact opposite of the truth with the philosophy that if a lie is repeated often enough, it is eventually taken to be the truth. (For a discussion of some other common examples, see an earlier post. Note also my recent post on Reversing the accusation, which deals with a similar subject.)

   There may, obviously, be great variations from country to country, but in Sweden and the US (cf. e.g. the Duke Lacrosse case^e) the opposite problem of presumption of guilt, lack of due process, findings of guilt based solely on the accusers claims, and similar appear to be more common. Notably, in some countries, rape-shield laws and similar mechanisms can even make an inequality in front of the law (to the disadvantage of men) near unavoidable.

   It is important to note that the alleged “poor treatment” is often nothing more than normal investigative techniques used against anyone raising an accusation of crime: The police, the DA, the defense, must all be entitled to ask questions to probe for contradictions and implausible statements in order to get to the truth—and this must be so for all crimes or the justice system will fail as innocent people are jailed and deliberate false accusations rise. Rape cannot and must not be an exception to this.

   (Indeed, I have even repeatedly heard complaints that use of the word “alleged”, non-use of the words “victim” and “offender”, and similar, would constitute mistreatment—claims that themselves are a horrifying neglect of the legal principle of “innocent until proved guilty”, intellectual honesty, and, frankly, basic human decency.)

   The fact that (in Sweden) 38 % of all judges, 48 % of all prosecutors, and two thirds of all legal students are women makes the claim that rape accusers would be poorly treated even harder to defend—unless we assume that this mistreatement would largely stem from other women… (Numbers from 2008^e; the current numbers are likely to be even higher.)

Remark: Note that it is very important to be cautious when interpreting various surveys, statistics, etc., concerning crimes. Not only do we have the problem of feminist distortions, but also one of different criteria and definitions, uncritical handling of sources, and similar. A particular important factor is time: The number (both absolute and relative) of horse and cattle thieves in the “Wild West” is likely to be considerably lower today than two hundred years ago. Similarly, any other statistic that is not reasonably recent must be re-investigated before being brought in as a hard fact—including crime rates, report rates, speculations on unreported crime, etc., from even just a few decades ago.