Statistics, With Rigor

For those of you who, like me, sat through Statistics 101 thinking that it would be great if there was a long needle within reach so I could stick it in my eye rather than have to sit through any more lectures, much of the discussion about the “1 in 5” rape epidemic stat that has driven the hysteria upon which evisceration of basic due process rights is grounded is problematic.

Sure, we understand basic statistical flaws, such as the lack of random sampling and itty-bitty sample sizes. Or there’s the facile interpretation of responses to conform to the outcome desired by the people conducting the study. Or the problems when definitions are fudged to the degree that “some guy stared at me at a party” becomes “stare rape,” which becomes rape.  Or rape is whatever someone decides it is the day after consensual sex.

But these are just the surface flaws.  We’re not qualified to parse the details with a statistician’s eye, so we rely on those who know what they’re really talking about.  One such person, new to the blogosphere, is Francis Walker.

Whether that’s a real name or a nym isn’t clear.* I’ve reached out to Francis Walker to get some additional information, but have yet to hear back. Maybe I will before the day is out. But what Francis produced at Data Gone Odd is so thorough, well-grounded and detailed as to stand on its own, regardless of who did the writing.

Before going to Francis’ posts, one point that demands making is that this isn’t some rhetorical challenge to statistics.  It’s not the mere utterance of “the study was flawed,” or some shallow challenge to the obvious self-selected group of respondents who took the survey on a website dedicated to eradicating sexual abuse.  This is what rigor looks like.

Francis has two posts, both of which are long, academic in style and highly detailed.  They’re entitled “How To Lie And Mislead With Rape Statistics,” parts one and two.  Despite their length and style, they are provocative in the sense that they go where the numbers lead.

As it turns out, only 7.8% of rape reports are true.

I know that may seem hard to believe, but I didn’t just make it up.  Technically, it is completely true.  It is also completely horse shit.  It is so misleading and built upon so many undisclosed caveats, that most people would consider it as good as lying if they knew how it was actually derived. The thing is, that “only 2-8% of rape allegations turn out to be false” figure you may have heard?  Not only is it just as misleading (if not more), it actually comes from the exact same data set.

In a discussion in the comments to this post, the typical empty rhetoric was propounded to perpetuate the phony “1 in 5” stat, not because the commenter believed it to be accurate but because

there are many good reasons to suspect that the .61% significantly undercounts, and that its source – the NCVS – finds an extremely low rape prevalence rate compared to almost any other study of rape prevalence.

Are there “many good reasons”?  What about “almost any other study of rape prevalence”?  Neither assertion offers anything remotely resembling a well-grounded argument. These are conclusory assertions without basis, meaningless rhetoric.  Francis goes through the studies at length.  Unlike Barry Deutsch, the commenter who offered these empty assertions, as if one could run through the faults in meaningful detail in a blog comment, Francis performs the rigorous scrutiny required.

Francis’ posts address primarily the issue of false rape allegations, that claim at issue when advocates contend that all “survivors” must be believed, and that there are minimal, if any, false claims.  Francis picks apart the studies, the methodologies, the assumptions and the biases behind the numbers.  It’s tedious work, but if we’re to work with statistics to make a point, it’s critical work since we only blowing smoke if the stats are crap.

But many of his points reflect the failure of rigor that confuses well-intended but misguided advocates in all aspects of this issue:

A multi-site, multi-year study?  Seems like we actually have some good data to work with here.  They then give a website where you can find out more information about the MAD Project, http://www.evawintl.org.  Clicking through you’ll find out that their primary source for a “methodologically rigorous” study on the number of false rape reports comes from the organization End Violence Against Women International. If someone tried to present their view that the number of people killed by cigarettes each year is actually pretty low and their primary source was a study done by Phillip Morris, how much credence would you lend that research?  What about a study on gun violence done by the NRA? And yet, for some unknown reason, the authors don’t seem to think that End Violence Against Women International might in some way be biased as to frequency of false rape reports.

When I embraced the DoJ’s Bureau of Justice Statistics report (notably not one of my personal favorite sources of information since the DoJ and I don’t tend to see things eye to eye in most instances), Deutsch ironically accused me of being guided by my political bias.

So you’re saying that all studies that don’t show the outcome you want are politically biased? And you think I’m the one making desperate wishes?

No one has ever accused me of trusting the Department of Justice too much before, or sharing a political bias with them.  Unfortunately, this has nothing to do with bias, political or otherwise, despite claims by those who would rather cling to statistics that are conclusively known to be false.

If there was, in fact, an epidemic of rape and sexual assault, as those offenses are legally defined rather than offenses based on how women feel about things the next day, then it would demand redress. We would still argue whether the evisceration of due process is the right solution, but at least it would be based on a problem that existed in fact rather than fertile imaginations.

But it’s people like Francis who do the yeoman’s work of parsing the numbers, the studies, the details, to arrive at a founded conclusion as to whether a problem exists or we’re being whipped into phony hysteria.  It’s not about anyone’s political bias, but merely accuracy and sound statistics.  The outcome goes wherever the numbers lead, and there is no hyperbole or hysterical crying that changes the value of the statistics.

* I’ve since heard from Francis Walker, who has provided me with his bona fides, but informed me that he would prefer to remain anon so that the work, and not the person, does the talking. I’m satisfied that Francis Walker is the real deal, and will honor his decision to go by a nym.

20 thoughts on “Statistics, With Rigor

  1. Barry Deutsch

    I don’t think you’re interested in productive, non-insulting discussion.

    It’s obvious that you’re either dishonest or ignorant (for instance, contrary to what you claimed twice here, the CSA does use random sampling, not self-selected respondents.) I trued text-searching the CSA for the term “stare,” both alone and with “rape,” and found no results; I don’t believe any such claim was made in the CSA.

    But in any case, you do make an excellent point, which is that your comments section is not a good home for a detailed, fact-based discussion of why the CVRS, excellent as it is, is not adequate as a measure of rape prevalence. I will take my time and write a post on my own blog on the subject, with as much rigor as I can manage.

    I thank you for your insight and hosting.

    Best of luck to you.

    1. SHG Post author

      I originally responded to this comment without realizing that you posted it here, and thought it was a continuation of your earlier comments. When I realized my error, I deleted my reply so that I could write a more fitting one.

      You write:

      It’s obvious that you’re either dishonest or ignorant (for instance, contrary to what you claimed twice here, the CSA does use random sampling, not self-selected respondents.) I trued text-searching the CSA for the term “stare,” both alone and with “rape,” and found no results; I don’t believe any such claim was made in the CSA.

      Except nowhere do I say the claim was made in the CSA. You’ve chosen to take my broader statements and falsely apply them to a specific instance, which is tantamount to a lie. So yeah, somebody is being either dishonest or ignorant, but it’s not me, pal. You’ve completely blown whatever cred you might have had by lying about what I said. You should be ashamed of yourself, but we both know you won’t be.

      You have the links here to Francis Walker’s posts, which provide all you need to understand the fallaciousness of your arguments, but you obviously haven’t read his posts and clearly have no interest in having your bubble burst. So now you’re not only a liar, but a lazy one.

      As I said earlier, that’s fine. This is America and you’re entitled to be as stupid as you want to be. But don’t play the passive-aggressive game, or make demands that I do your work to disprove your vapid arguments.

      You claimed to be open to being proven wrong, yet you can’t be bothered to read posts that don’t conform to your desired outcome. So be it. That’s your choice. Just don’t accuse of me to absolve yourself of hiding behind your political beliefs.

      I hope you enjoy writing a post that ignores the numbers and supports your politics. It won’t matter to me as I won’t read it. I deal in reality, wherever the numbers take it. But it really is a shame you couldn’t be bothered to read Francis Walker’s posts. At least then you would have some small grasp of how shallow your arguments are.

      1. Andrew

        You may not realize this, but Barry Deutsch’s blog, amptoons, is deep into extreme feminism. That explains a lot about why he’s so dense and dishonest about the statistics. His claim that he’s open minded is total bullshit, and he’s got a definite agenda.

        1. SHG Post author

          I suspect I’ve been far too kind and given Barry way too much of my time by trying to address his comments. In retrospect, he’s proven himself remarkably dishonest. As you wrote in your other comment on the other post, he’s just a feminist troll.

          1. Not Jim Ardis

            I’ve had experience with Ampersand (Barry’s nom de guerre) elsewhere.

            I am pleased that his talent for dishonest discourse has remained – comfort in that which is immutable is a failing of mine.

            1. Not Jim Ardis

              His claim to neutrality was an abject lie. He is willfully ignorant, utterly dishonest, and blindingly partisan.

              It is my opinion that he views anything that furthers his preferred politics is valid regardless of the tactic used or the dishonesty required. If his side ends up winning, then it is justified.

              I’d tell you where I used to interact with him, but you’d only spit.

  2. Anon

    I was warned not to comment here, because you’re a prick who tolerates no disagreement. But I have to ask one question. How can you understand social injustice when it’s racially motivated, and yet not care at all when its gender based? It’s like you’re two totally different people.

    1. SHG Post author

      While I am a prick, the problem isn’t disagreement. That’s the excuse made by those who don’t receive the respect and adoration to which they believe they’re entitled. Rather, I don’t tolerate people who lie, deceive, mislead, confuse and make other people stupider.

      Ideological people tend to do a lot of that.

      There are plenty of people here who disagree with me, but they do so from a position of knowledge and integrity. Bullshit does not cut it, and people who bullshit get called out. There are no tummy rubs here. People who spew nonsense get smacked, whether they agree with me or disagree with me. It makes no difference.

      As to your question, this is a law blog, not a political blog. People who adhere to a political ideology can’t conceive of what that means; they assume everyone is political just like them, and if someone opines in a way that disagrees with their orthodoxy, they must be the enemy. That’s not how it works here.

      My perspective is that of the criminal defense lawyer. I’m not MRA nor feminist, but concerned with the law, constitutional rights, and due process for people accused of crimes. Let me put the question back on you: how do you explain being so concerned about the rights of black defendants, but couldn’t care less about them when the charge is rape?

      So I’m neither liberal nor conservantive (nor libertarian for that matter). I’m a criminal defense lawyer, and I don’t care whether the charge is murder, public urination or rape. Gender doesn’t change my perspective on the Constitution, and I am no more tolerant of bullshit coming from one side than the other. Oh yeah, and I am a prick. If that’s too unpleasant for you to tolerate, then whoever told you not to comment here gave you sound advice.

      Edit: Just noticed that this is your second comment here, not your first. Nabbed.

  3. The Real Peterman

    “its source – the NCVS – finds an extremely low rape prevalence rate compared to almost any other study of rape prevalence”

    That’s true. On the other hand, rates of all types of violent crime have been absolutely nosediving for decades now. So the idea that any particular group of women has a 20-25% likelihood of being raped defies what we know of our society as a whole. As one author put it, do we really think a woman ar Harvard is ten times as likely to be a victim of rape as a resident of Detroit is to be the victim of *any* crime?

    1. Anne Krone

      Of course it does. Provided you define rape as an unattractive man asking a woman out for coffee.

    2. David M.

      But we have to have some reason for screwing up the lives of poor, black people. Imagine if they went around attending college without being harassed. And the guys with Asperger’s are even less understandable.

  4. Sgt. Schultz

    What’s really a shame is that this was a post about Francis Walker’s truly great job of breaking down the phony false rape stats, yet this flaming narcissist, Barry, made it all about him, his tiny penis and fragile fem ally ego. Instead of a discussion about things that are real and matter, this thread (and the other one) became all about some prolix douche trying to dig his way out of a huge hole of stupid.

    Francis Walker’s work was truly spectacular, thorough and thoughtful. At the very least, the contrast between this lunatic Barry and Francis Walker makes it conclusively clear that the ideologues will spew their bullshit as long and hard as possible for the sole purpose of deceiving and misleading, hoping to pick off the most ignorant with their nonsense.

    What a shame to have this happen to Francis Walker. What a shame that Barry is nothing but wasted, pathetic noise.

    1. SHG Post author

      You’re right, it is a shame that Barry tried to hijack this post to turn it into an “all-about-Barry” discussion. That said, I too think the juxtaposition of sound statistical analysis like Francis Walker’s and the lunacy of Barry Deutsch served to make an important point about how worthwhile good information is, and how worthless the ideological crap Barry spews is.

      So the upshot is that Barry’s effort merely proved the value of Francis Walker’s efforts.

  5. EH

    The issue of truth or falsity is obviously dependent on whether you are taking an accuser or defendant perspective in the initial definitions.

    From the perspective of an accuser who who believes she was raped, an accusation is “true” provided that she satisfies both of two conditions:
    1) she believes it to be true, which is to say that she believes the encounter legally constituted rape; and
    2) she does her best to relay facts as accurately as she can (which, based on some reports of rape trauma and outcomes, may not be with 100% accuracy.)

    From the perspective of a defendant who has been accused of rape, an accusation is “false” if it satisfies either of two competing conditions:
    1) The facts as described do not meet the legal criteria for rape; or
    2) Any of the facts–any at all–are inaccurate in a way that favors the defendant over the accused.

    If you look at those definitions and the overlap between them, it should be apparent that you will end up with a large set of rapes in which the classification of the accusation as “true” or “false” is entirely dependent on which definition you select.

    It is truly a pity that such an important issue has been constantly hamstrung by the semantic arguments over basic terminology.

    1. SHG Post author

      While you’ve grossly oversimplified what’s involved, your conclusion is way off base. This isn’t a semantic argument over basic terminology, but a political argument over what constitutes rape and sexual assault.

      You use the “legal definitions,” which is problematic since different jurisdictions have different definitions. So that is a facial problem at the outset. Since there is no fixed “legal definition,” and even if there was, the people conducting the survey aren’t lawyers and so aren’t qualified to determine whether conduct meets a legal definition in any event, they instead employ whatever definition they choose.

      Moreover, respondent narratives rarely match up well with efforts at definition. Women who believe they were raped or sexually assaulted tell narratives to support their belief. Since there is no corresponding survey of their alleged rapists, there is no “other side” to balance the narrative.

      As the words “rape” and “sexual assault” have long become untethered from definitions, particularly on college campuses where most of these surveys are conducted, and as respondents, who make an active choice to respond usually because they want to express their feelings on the subject, obtaining a valid result has so many moving parts, so many political influences, so many definitional problems, so many interpretative problems, that it’s extremely difficult to conduct a valid survey and obtain reliable results. To reduce it to “semantic arguments” is grossly simplistic.

Comments are closed.