Date: December 08, 2014
Notice of Raw Links:
Gene G. Abel’s Statistical Crimes – a Persistent Blight on the Science of Sexuality
The Mark of Abel – How psychiatric malpractice has made the non-offending pedophile a hypercriminal –
Kristofor Xavier, 2012, minor updates K. Beylant 2014 with permission – article may be freely copied and distributed
In the midst of a recent crisis, a 27-year-old psychiatric study revealed itself as a major influence on the course of events. This analysis looks into that paper, and shows why it should never again be allowed to influence any social movement.
The crisis is one that slipped under the radar for most people, since journalists largely ignored it. In 2012, a subgroup of members belonging to the ‘hacktivist’ movement Anonymous attempted to force all minor-attracted (MA) persons off the internet. The vigorous quasi-military campaign, named OpPedoChat, consisted of blasting MA websites with distributed-denial-of-service ‘lazors,’ and attempting to hack directly into the websites, destroy them from within, and publish member lists and contents.
This effort soon tapered off, partly because MA web administrators were veterans of previous assaults and were able to ward off the attacks, but mainly because most of the participants soon changed their perception of the major group of target websites. They saw that the websites, rather than being centers for scheming about how to assault kids, were highly regulated forums administered and mainly populated by law-abiding people. The website admins and most of the members, however diverse their viewpoints, stood opposed to intergenerational sexual acts that would negatively impact children and teenagers.
As the OpPedoChat campaign refocused on different targets that were arguably associated with exploitation, its ad-hoc leader, usernamed TheAnon0ne (later revealed at his arrest to be Michael Crockett of Kansas City, Missouri, http://www.boychat.org/messages/1388484.htm), made one last attempt to keep our legal sites in the cross-hairs by citing a damning research statistic. He revealed that his hardline anti-pedophile attitudes were motivated, at least in part, by information he’d got from a website called bettermommies.com. He’d read their ‘questions and answers about pedophiles’ page.
The mommies reported a ‘fact’ that “the typical pedophile commits an average of 117 sexual crimes during his lifetime (280 for those molesting boys).” TheAnon0ne, a concerned young computer professional from the Nashville suburb of Fairview, TN (according to his personalized voice mail) with a young daughter and perhaps also a second child, appeared to suggest that the people at our websites shouldn’t be let off the hook so easily. If we, on average, were molesting 117 to 280 kids over the course of our lifetimes, then we were far from benign, and no one should be naïve about the threat we posed.
As someone who has molested zero children, I immediately thought ‘where in God’s name did this statistic come from that tars me with 280 sexcrimes?’ I thought over the people I’d met in the MA community. The vast majority had never been involved in sex with juveniles during their adult years, and were just doing the best they could to live with the orientation. I’d met, or seen posts by, some guys who’d had encounters with actively partner-seeking younger teens – acts that would be legal in some western-nation jurisdictions but not in others. Few if any of these men were continuing in this mode, given the current legal and social situation. And, mostly on the very diverse forum, boychat.org, I’d seen posts by a few people who admitted to doing time for past events with boys, but who were determined not to get on the wrong side of the law again.
The figure of 280 molestations per male-oriented MA person seemed to be a wild fiction. And yet, it was stated to have come from a study.
I decided to track down this study and see what it actually said.
After getting past a few misattributions to review papers that cited the study, I found the original document . It was a work from the psychiatry shop of Gene G. Abel, then from the Emory University School of Medicine at Grady Memorial Hospital in Atlanta. There were also five collaborators, three of them from Columbia University in Abel’s former home town, New York City. The whole citation is ‘Self-reported sex crimes of nonincarcerated paraphiliacs,’ by Gene G. Abel, Judith V. Beckerman, Mary Mittelman, Jerry Cunningham-Rathner, Joanne L. Rouleau, and William D. Murphy, published in the Journal of Interpersonal Violence in 1987, volume 2, pages 3 to 25. Unfortunately, the work appears not to be openly available online.
Abel is by no means a household name, but he should be. A review of his work reveals him as a surprisingly influential figure in current Western society. In particular, a paper he, Beckerman and Cunningham-Rathner wrote in 1984, ‘Complications, consent, and cognitions in sex between children and adults,’ (Int J Law Psychiatry 1984;7:89-103) has put many words onto the lips of the common man and woman. It manufactured the whole discourse on childhood consent and power that you can now hear from any children’s social worker, as well as the discourse on pedophiles’ ‘cognitive distortions’ that provides the stock courtroom response to any suggestion that a child or teen may have initiated or wanted a sexual act, or was not harmed by it. Abel is indirectly quoted by tens of thousands of professionals, editorial writers, letter-to-the-editor writers, and online commenters every year. You could almost say that he wrote the common modern civil attitude towards pedophilia.
(I say ‘civil’ to distinguish the common uncivil, ape-man attitudes like ‘put them in prison forever and let big lusty Bubba have his way with them.’)
The ‘consent’ paper is a review that pulls together some data from previously published papers, but is mostly written in editorial mode. The ‘self-reported sex crimes’ paper, on the other hand, purports to be a scientific study.
It is a textbook study, all right, but it belongs in a book about the misuse of statistics. Let me show you how those of us who are MA but have molested zero children came to have 280 sexual crimes on our theoretical criminal record.
If you’ve ever taken a university statistics course, the first thing you learn is that sampling small numbers of subjects can be very powerful, but only if the sampling is truly random. Wikipedia wasn’t written when Gene Abel and friends put together their ‘self-reported crimes’ study, but the contents of the paragraphs below from the current article on biased sampling were well known:
(from a list of ‘Types of Sampling Bias’)
Self-selection bias, which is possible whenever the group of people being studied has any form of control over whether to participate. Participants’ decision to participate may be correlated with traits that affect the study, making the participants a non-representative sample. For example, people who have strong opinions or substantial knowledge may be more willing to spend time answering a survey than those who do not. Another example is online and phone-in polls, which are biased samples because the respondents are self-selected. Those individuals who are highly motivated to respond, typically individuals who have strong opinions, are overrepresented, and individuals that are indifferent or apathetic are less likely to respond. This often leads to a polarization of responses with extreme perspectives being given a disproportionate weight in the summary. As a result, these types of polls are regarded as unscientific.
Pre-screening of trial participants, or advertising for volunteers within particular groups. For example a study to “prove” that smoking does not affect fitness might recruit at the local fitness center, but advertise for smokers during the advanced aerobics class, and for non-smokers during the weight loss sessions.
Abel’s study is almost completely based on the types of sampling mentioned in these paragraphs on bias. It includes a wide range of ‘paraphiliacs,’ as defined in the then-current 3d revision of the ‘Diagnostic and Statistical Manual of Mental Disorders’ (1974), the diagnostic code of psychiatry – commonly called DSM-III. Among these paraphiliacs are groups whose presence would come as a political-incorrectness shock to most modern readers – homosexuals (but only if ‘ego-dystonic,’ i.e., unhappy about it), transsexuals and transvestites. Remember, this is being published in a journal dedicated to studies of ‘interpersonal violence.’
The paper’s statements on how the authors recruited their study population are worded in a beguiling way that makes it seem that the authors are aware of their sampling bias problems, and yet, at the same time, allows them to conceal the most serious of those problems.
This paragraph is a statistical card trick. While the reader is distracted by the true confession that the study may misrepresent the relative proportions of different kinds of paraphiliacs in the general population, Abel discreetly lets the notion slide through that the representation of the whole population within each group of paraphiliacs is okay because the ‘data are sufficiently numerous.’
That’s a serious no-no in science. No matter how large a biased sample is, it is always a biased sample. I’ll talk in a minute about what sort of a biased sample this one is, and how it obtained its bias. But first, the brazen scientific error that starts the study off needs to be addressed. What Abel and colleagues have done here, with their statement about ‘sufficiently numerous’ data, is to imply that they have interviewed so many subjects that they have overwhelmed sampling bias by, in effect, sampling the total underlying population. It is true, in statistics, that if you can collect data on an entire population, you needn’t worry about sampling bias. Your sample can’t be biased because you have polled everyone. ‘Sufficiently numerous’ is a phrase that declares that this condition has, in effect, been met. “We’ve seen so many paraphiliacs that there can’t be any significant chunks of their population that we’ve missed,” is the implication. The assertion is preposterous – the population of the Atlanta, New York and Memphis areas – the catchment areas for the authors – has by no means yielded up all or nearly all of its paraphiliacs to the study.
To be blunt, the viewpoint that larger-than-usual sample numbers obviate the need for statistical randomness is only known in one place in science: on exam papers of people who have failed the introductory statistics test.
Abel and colleagues are so fond of this error, though, that they actually repeat it in their discussion of their results. Again, some verbal sleight-of-hand is used to make it appear that the biased sample problem is being addressed. Then, suddenly, the large numbers involved are again stated to negate the problem:
The ‘large numbers,’ Abel says, license the use of an intrinsically biased sample. He thus allows himself to draw conclusions about whole populations of, for example, transsexuals, even though he is using people obtained “within the context of an evaluation and treatment program for sex offenders voluntarily seeking assessment and/or treatment in a psychiatric setting.”
In the year 2014, everyone can see that this is outrageous. Why would a representative group of transsexuals be motivated to sign up, ‘in the context of a treatment program for sex offenders,’ for one to five hours of ‘structured clinical interview’ that would make them list their entire lifetime history of deviant sexual events, and count the complete number of partners involved?
This question becomes especially poignant when you look at the list of places subjects were recruited:
The study used ads in newspapers, radio and television, but less than one tenth of the participants were brought in this way. Well over half the participants in Abel and colleagues’ study came from two types of sources: lawyers, and professionals associated with mental health care. A large minority were referred by parole officers.
Abel does state that in almost all cases, the subject was asked what had ‘motivated him to seek help.’ The answers are tabulated as follows (in percentages):
Contrasting the above two lists with one another, you can correctly infer that the majority of those claiming self-motivation for joining the study were nonetheless referred by their lawyer, or by a health care professional, or by a parole officer.
The nature of the main source of bias inherent in Abel and colleagues’ sample population can be specified in one word: alarm. The subjects of Abel’s study are people who may not be incarcerated at the moment, but they have nonetheless caused alarm. They have caused alarm to professionals in law, medicine or social work, or they have become alarming to their families, or they have become alarmed at themselves. The sample population misses out on any paraphiliacs who are keeping cool, being well adjusted, or generally being reticent. It is pitched to collect the most extreme, alarming cases of non-incarcerated paraphiliacs.
Let’s go back to something Abel said in one of the quotes above. “Since paraphiliacs frequently attempt to conceal their arousal from others, it would be extremely difficult to obtain a true random sample of nonincarcerated paraphiliacs.” This is a very true statement – so true, in fact, that it technically negates much of the existing research on the topics in question, most notably any studies purporting to make general statements about pedophilia.
When you are faced with a problem of a deeply frightened or highly prudent population that deliberately frustrates being randomly sampled, you can do one of two things. One approach is to throw up your hands and say, ‘well, since we can’t get a real sample, my biased sample can be considered good enough for practical purposes, especially if it is large.’ That’s not science – but it works as a sort of political stand-in for science. Governments pay professionals a lot of money to deal with sex deviants, and the professionals can’t just come back and say, ‘we couldn’t actually get a valid sample.’
The alternative approach, which is politically unsatisfying but scientifically sound, is to accept your sample as biased and to attempt to define it exactly as it is. If you’re working with offenders incarcerated for sexual acts with children, simply say so and don’t extend your conclusions to ‘pedophiles’ as a whole. You’ve missed, among others, every pedophile who has never committed a sex act with a child, so naturally you have no basis to speak about that group of people. If you are working with a voluntary sample of people who have caused so much concern that someone has wanted an extensive psychiatric interview done about aspects of their deviant behavior, then define your population as ‘paraphiliacs whose behaviors have caused significant concern to themselves or others.’ Don’t fall off the path of science into the abyss of pseudoscience by pretending that your results apply to the whole population of paraphiliacs, all those people who are known to be ‘concealing their arousal from others.’ You have no way of knowing whether or not your biased sample results may coincidentally truly represent the reality of one or more subgroups of paraphiliacs. General knowledge of human behavior should be sufficient to warn you that in any situation that favors concealment, some people will be more concealed than others. Especially in the context of a clinical study with a particular approach.
The acts of statistical abuse in Abel’s paper, however, don’t stop here.
Further reading of the ‘self reported sex crimes’ study shows that its statistical problems are only beginning when it accepts a biased sample as representative. Let’s look at the data for non-incestuous pedophile acts. Abel gives two statistical numbers for people who have committed legally undetected sexual crimes with girls or boys – the average number of sexual assaults per person, and the median. If you’ve taken that basic statistics course, you’ll see that each of these numbers casts light on the significance of the other. If you studied ten pedophiles, and nine had committed no sexual assaults while the tenth had committed 1000, then the average number of sexual assaults per pedophile would be 100. The median, however, is the number you get if you line all the subjects up from the lowest number to the highest number, and take the number that is in the middle of the line-up. In that case, in our ten pedophiles, the median number of sexual assaults per pedophile would be zero. The hyperactive pedophile who has committed a thousand sex acts has radically skewed the average, but hasn’t affected the median. That’s because he’s a statistical outlier, an extreme case.
Again, using his typical sleight of hand, Abel explicitly makes this point and then ignores it. (We follow standard scientific convention as used in cases of scientific fraud and major error by assigning him full responsibility as an author, even though mistakes may have been initiated by a co-author.)
The last statement is nonsense, since in reality, it is the raw total numbers of sexual acts committed, not their averages per person, that ‘indicate how often the various paraphiliac acts were completed.’ But, as you’ll see, Abel is fond of the averages and plans to use them extensively in discussion. Now let’s look at the figures themselves.
For 224 men who have had non-incestuous pedophilic acts with girls, the median number of sex acts committed was 1.4, but the average was 23.2. Nearly half, then, of the 224 girl-attracted pedophiles studied had been involved in a single sexual act with one girl, but still were assigned an average number of acts of 23.2. Among non-incestuous pedophiles who had carried out acts with boys (prepubescent only, as per the definition of pedophilia in DSM-III) the median number of such acts was 10.1, but the average was 281.4. Clearly, one or more persons up at the extreme high end of the spectrum had claimed a phenomenal number of sex acts with boys – but no actual numbers of acts per individual person are given in Abel’s paper. The median number of victims/partners listed per girl-oriented offender was 1.3, but the average was 19.8; for acts with boys, the median number of partners was 4.4, but the average was 150.2. And so on. The averages wildly exceeded the medians.
Now let’s see what acknowledgment Abel gives to the fact that the median numbers per offender are so much lower than the averages, meaning that the averages are being strongly skewed by a few extreme cases.
Yes, you saw that correctly. Abel has cited the averages as the true indicator of the activities of the ‘nonincarcerated child molesters.’
If you’re attracted to boys, on average, you’ve committed 281.7 sexual assaults, kindly reduced by the bettermommies website to 280. Highly probably, every single one of those assaults has been committed by someone else, some rampant serial child abuser. But the assaults can still be ascribed to you through statistical trickery. And Gene Abel is the man who has done it.
Is Abel, then, a deliberate fraudster? The evidence from the paper suggests that, in fact, he may be merely so statistically ignorant that he can’t see the simplest problems with his numbers. His Table 2, showing the ‘Number of victims/partners of paraphiliacs,’ lists ‘homosexuals’ as having an average of 0.1 partners, and a median of 0.0.
He never gives any explanation of how sex acts that by definition, in this study, are supposed to involve partners or victims, can be accomplished with an average number of partners of less than one. It’s not as if homosexuals didn’t claim to have completed sex acts. They had an average number of sex acts of 154.2, a median of 1.5, and a number of “completed acts per victim” (Table 3) of 1,850.5! So where did all the not-so-gay gays without partners come from? Did Abel include sexual fantasies or masturbation events in his tally of paraphilic acts and neglect to tell us about it? That would very much change the impact of the paper. We have no way of knowing. His Table 2 is unexplained, and apparently numerically impossible in the context of his stated methods.
Another problem with the statistics is that Abel doesn’t specify at what point in the subject’s life he started counting his paraphilic acts. Did he include sex acts committed with little boys when the subject himself was a little boy? Some young boys are remarkably sexually active; for example, here’s a quote from an email recently sent to me: “For me, my earliest memories involve what could be called ‘gay thoughts’ although some of it was probably more curiosity. However, (I went) through childhood with the belief that doing things with other boys was considered quite bad (despite doing it many times each week).” We have no way of knowing how much early childhood experimentation is included in Abel’s statistics. We do know that the interviews went back to the ‘age of onset of their paraphilic arousal pattern” and that “each subject’s history was examined year by year to determine the changes in frequency of his paraphilic behavior.” It might well be that my email correspondent would be listed as a profligate victimizer of boys, based entirely on sex acts he committed during his own prepubescent boyhood. Again, we have no way of knowing.
The paper is the most astounding piece of junk science I have ever seen.
I suspect the problem lies in Abel’s field of inquiry. Psychiatry has become associated with medicine, but it began as a branch of explanatory philosophy as patients were prompted to tell their doctors and psychoanalysts about their hidden thoughts. To a large extent, it became a godless replacement for religious or spiritual causal explanations of people’s behavioral motivations. As an offshoot of orthodox Judaism and Christianity among those who had become secularized, it maintained a religious fervor against homosexuality well beyond the point that any reasonable person could imagine was justifiable. It maintained the ancient Judaic concept that sexual typicality was valid and desirable, and all else was undesirable deviation or pseudo-love (a direct translation of ‘paraphilia’). And as a field based in philosophy and dialogue, psychiatry didn’t seek to give its members a firm grounding in statistics. I would gladly accuse Abel of fraud if I knew he’d taken a statistics course – but I doubt very much that he ever had.
In Abel’s study of ‘self-reported sex crimes,’ we still see the aching para-religious urge to include homosexuals among the sex criminals, even if they can only be included when they’re ‘ego-dystonic.’ Table 2 may be labelled ‘Number of victims/partners of paraphiliacs,’ but it is introduced in the text as giving “the number of victims by diagnostic category.” Yes, whatever that one-tenth of a partner for the ego-dystonic homosexuals consists of, he or it is considered to be a victim of paraphilia.
The Wikipedia article about DSM-III gives some historical perspective on this kind of thing:
From the wikipedia article about paraphilia:
Abel, at least in the 1980s, followed this party line to a very large extent. He absolutely disbelieved in sexual orientation. In the ‘consent’ paper, he gives a purely Skinnerian (behavior-modification) model of sexual development, based on the idea that sexuality is entirely based on what is left when socially inhibited behaviors have been subtracted from socially and personally reinforced behaviors:
The process of learning to inhibit one’s sexual arousal to specific stimuli is not always perfect. This is especially the case during a child’s early years when his sexual behavior (primarily masturbation) is hidden from others and the stimuli that the child attends to during masturbation is secret. The critical issue in the development of arousal patterns inconsistent with an individual’s culture appears to be the child’s use, during adolescence, of inappropriate stimuli during masturbation and subsequent orgasm (Abel and Blanchard, 1974). When stimuli inappropriate to the culture are fantasized during masturbation and orgasm, the pleasure and enjoyment from masturbatory activity and orgasm become associated with this inappropriate material, the fantasy material becomes erotic and eventually provokes sexual arousal in the child. In this fashion, early inappropriate experiences and activities might be fantasized hundreds of times by the boy and repeatedly associated with the sexual pleasure of masturbation and orgasm until inappropriate sexual stimuli produce erection responses and arousal in the individual. When there are no negative consequences for the use of such fantasies, from parents or others (if they knew), there is no inhibition of that arousal pattern and as a consequence, the child reaches adulthood with arousal to inappropriate stimuli (young children).
This viewpoint has sexual orientatation arse-backward. Instead of the child’s choice of masturbation materials being caused by prior attraction and internally generated fantasy, the attractions and fantasies are stated to be based on an unexplained, arbitrary and wrong choice of masturbation materials.
Apart from this leaving a weird causal lacuna (why choose the inappropriate materials in the first place in preference to appropriate ones?), this idea can scarcely be rationalized with most people’s personal recollections about their own sexual development. Typically we start off by finding certain people highly attractive, and we develop fantasies about them or even crushes on them. We may have deep crushes long before we ever discover masturbation. Most of us, at least, haven’t done anything to reinforce these primal attractions before we become vividly aware of them – they have simply happened to us.
In 2014, any medical professional attempting to apply this raw Skinnerian nonsense to the causality of homosexuality would be considered by his peers to be, at best, a disgraceful and offensive eccentric. However, because the crimes of the psychiatric war on homosexuality are still okay today if they are committed against pedophiles, Abel receives very little professional criticism, and is cited as a source of factual material in hundreds of studies.
In the bible, Cain, the son of Adam, killed his brother Abel and was cursed by the Lord, but the Lord ‘set a mark’ on him so that he would not be killed in vengeance.
In psychiatry, Gene Abel has set a mark upon all MA people, one that curses us and, by its very nature, calls for vengeance on us. He abuses statistics to mark each and every one of us with hundreds of crimes we have never committed.
Let this article be the beginning of a process that reverses his curse on us and has him stricken forever from the records of accepted science. I don’t know whether he is to be condemned as a member of history’s roster of brazen and shameless persecutors of fellow human beings, or pitied as someone whose education in science was so deficient that he blundered helplessly into catastrophic error. Either way, he is discredited.
The better mommies of the world, and the young computer activists too, need to know that this man and his colleagues published inexcusable scientific error, and that their encompassing profession, psychiatry, was so methodologically weak and politically and financially corrupt that it lacked the means and will to recognize and correct the errors.
But the errors will be corrected. As long as science functions as science, they will be weeded out. Decades of laziness and prejudice may intervene before the studies are assigned to their proper wastebin, but such episodes cannot enter the canon of valid research.