01May10e C NL on Hagen on Rind cs

Dr. Margaret Hagen's critique of RBT, summarized and commented on

Chris, The Netherlands, May 2001
& Comment & discussion, Peter, Germany

The Skeptical Inquiry of January 2001 has a critique by psychologist Margaret Hagen of the Rind et al. 1998 meta-analysis. While Hagen is far from an unreasonable hysteric, she berates Rind et al. for pushing an agenda in their report, as she sees it. Hagen also says the report is weak not because it messes with figures, but because it draws overly confident conclusions from the figures. Below I encapsulate her article and provide some commentary.

(Margaret A. Hagen is a professor of psychology at Boston University and is the author of the book "Whores of the Court: The Fraud of Psychiatric Testimony and the Rape of American Justice." She can be reached at: Department of Psychology, Boston University, 64 Cummington Street, Boston, MA 02215. E-mail: hagen@bu.edu .)

Hagen criticizes victimologists’ exaggeration of harm caused by CSA. She asserts that there is "an important difference between adult males aroused by five-year-olds and men aroused by mature teens". She decries the Congressional condemnation.

Hagen questions Rind et al.’s recommendations that (lack of) willingness be taken into consideration by researchers, that researchers differentiate between young children and adolescents, and that they differentiate between children whose reactions reportedly were positive and those whose reactions reportedly were not. She finds all of these recommendations too ambitious in view of the limited evidence, which she thinks is meta-analyzed into a cluttered mess.

Hagen writes: "Of the boys, 1,957 reported that the CSA had been "both willing and unwanted," and 990 said it had been unwanted only. For the girls, 9,363 reported that their CSA had been "both willing and unwanted" and 2,268 reported that it had been unwanted only."

I’m not sure I understand the meaning of "both willing and unwanted" versus "unwanted only". Hagen’s wording makes it seem like there were groups of ambiguous persons versus groups of persons who were definitely unwilling. I am told that rather than comprising persons who were both willing and not willing to engage in sex at the same time, the "both willing and unwanted" groups comprise a number of persons who were willing and a number of persons who were not willing, but it is unknown how many were willing and how many were not willing. Rind et al. even write, "If unwanted CSA had been contrasted with willing CSA only, instead of a combination of unwanted and willing CSA, then consent would likely have moderated CSA-symptom relations more strongly."

Hagen says Rind et at. did not look at the (exact) age of the children who reportedly consented. She writes, "It should go without saying, although apparently it does not, that consent to sex with an adult is not at issue for preadolescents." She thinks preadolescents are incapable of informed consent, which she thinks is a prerequisite for sexual acts. She writes that "Rind et al. ignore the absence of reliable age data in concluding from their review that consent is a crucial variable in determining long-term effects of abuse". In the paper, Rind et al. do not address the difference between informed consent and simple consent, as they do in their 1999 talk for the Society for the Scientific Study of Sexuality and the American Association of Sex Educators, Counselors, and Therapists, http://www.humanbeing.demon.nl/ipceweb/Library/99118_rbt_defense_nov99.htm.

Hagen questions the value of the college students’ self-appraisals: "Attempting to identify harm to children from early sexual experiences by asking 18-year-olds what they think about it is hardly the most appropriate measure, and concluding that the sexual exploitation of children is, in most cases, over the long-term, harmless because some 18-year-olds do not see the harm in it is not only foolish, it is bad science." Since she has criticized the recovered memory movement, she should be wary of the notion of harm arising only after an "incubation period" during which subjects do not see as negative (but as inconsequential or positive) the experience which is the putative cause of dormant future harm. And the report itself says: "younger and older adults [under 30 and over 30, respectively] did not differ in CSA-adjustment relations in Neumann et al.’s (1996) meta-analysis." Rind et al. also explain that "college samples closely parallel national samples with regard to prevalence of CSA, types of experiences, self-perceived effects, and CSA-symptom relations". And: "mean effect sizes from college samples, as reported by Jumper, were similar to those from national samples (Rind & Tromovitch, 1997), nonclinical samples (Neumann et al., 1996), and community samples (Jumper, 1995, after corrections)." So perhaps Hagen is a little too skeptical in her assessment of the value of students’ self-reports in this context. However, I think she rightly points out the shortcomings of the self-report method – the value of which is also determined by the level of detail of the reports. Sandfort’s study (1981, 1982) was unrepresentative (which he stated clearly, but for which he was still attacked as though he had not stated it) but the Sandfort study used very detailed self-reports, as well as techniques to assess their truthfulness. These reports proved there validity in other areas.

Hagen writes that Rind et al. "conclude straight out that older children are much less likely to suffer later from sex abuse than younger children. How would they know that?"

I can’t find where they conclude that. They mention, "Baker and Duncan (1985) commented that girls in their national survey in Great Britain may have found their CSA experiences to be more damaging than boys did because they had more intrafamilial CSA and experienced CSA at younger ages." They also write, "Adolescents are different from children in that they are more likely to have sexual interests, to know whether they want a particular sexual encounter, and to resist an encounter that they do not want. Furthermore, unlike adult-child sex, adult-adolescent sex has been commonplace cross-culturally and historically, often in socially sanctioned forms, and may fall within the "normal" range of human sexual behaviors (Bullough, 1990; Greenberg, 1988; Okami, 1994)."

Hagen complains that the analyzed studies "disproportionately included subjects whose CSA experiences were relatively minor". But the overinclusion that the CSA concept has fallen prone to is not the fault of Rind et al.. (See also question & answer 5 at Science & Morality, a refutation of some common criticisms of the meta-analysis.) Also, Rind et al. found that studies that researched only contact CSA did not show significantly more (or less) effect than studies that researched all CSA. They also looked at a number of moderators that would indicate more serious CSA, and found that duration, frequency and penetration were not related to the effect size, whereas force and incest were.

Seemingly unaware of the investigated moderators, Hagen writes: "The failure of researchers to differentiate the type of abuse and to examine the possibility of differential effects as a function of severity essentially equates flashing with forcible sodomy. […] These numerous information gaps mean that the authors simply did not have the evidence necessary to make the farreaching and politically provocative conclusions they did."

Whether the conclusions are "politically provocative" is irrelevant to their validity and appropriateness in the report. Are they "far-reaching" in light of the evidence?

What did Rind et al. conclude? First and foremost that "CSA does not cause intense harm on a pervasive basis regardless of gender in the college population" (and they cite evidence that the college population is comparable to the population at large). Hagen claims that the evidence is not strong enough for these conclusions to be drawn. However, if there are various indications for this conclusion, and no indications to the contrary, I see little use in censuring the authors for challenging the widely-held (in academe and elsewhere) assumption that CSA causes intense harm on a pervasive basis regardless of gender!

Rind et al. also noted: "In most studies examined in the current review, CSA was defined based on legal and moral, rather than empirical and phenomenological, criteria. […] The positive reports of reactions and effects, along with normal adjustment for willing participants, are scientifically inconsistent with classifying these male students as having been abused. Their experiences were not associated with harm, and there appears to be no scientific reason to expect such an association (i.e., predicting psychologically harmful effects from events that produced positive reactions lacks face validity). […] The foregoing discussion does not imply that the construct CSA should be abandoned, but only that it should be used less indiscriminately to achieve better scientific validity. […] Overinclusive definitions of abuse that encompass both willing sexual experiences accompanied by positive reactions and coerced sexual experiences with negative reactions produce poor predictive validity."

So among the "farreaching and politically provocative conclusions" disapproved of by Hagen is the conclusion that the CSA construct is overinclusive; the exact observation used by Hagen to criticize the meta-analysis. Rind et al. address the construct’s overinclusiveness in all aspects: "Problems of scientific validity of the term CSA are perhaps most apparent when contrasting cases such as the repeated rape of a 5-year-old girl by her father and the willing sexual involvement of a mature 15-year-old adolescent boy with an unrelated adult."

Rind et al. propose the introduction of the value-neutral term "adult-child sex" alongside "child sexual abuse," suggesting that (lack of) willingness be considered in future assessments. Hagen thinks Rind et al. do not have enough information to be sure that adolescents can consent (the American Psychological Association seems confident about children’s ability to give simple consent and adolescents’ ability to give informed consent in nonsexual situations, though), or that persons with a CSA experience will not be affected negatively at any point in their lives. Perhaps, but Rind et al. do not claim that henceforth researchers should speak of "beneficent adult-child sex" or "beneficent adult-adolescent sex," which would be the opposite of "CSA." Instead, based on the strong suggestions (not irrefutable proof) that there are SA persons who do not experience negative effects or who even experience positive effects, Rind et al. call for the logically more appropriate use of neutral terminology in some instances. Instances, for example, in which subjects report that they consented - which appears to matter; forget about irrefutable proof - and experienced no harm so far, with there being no evidence that harm will follow much later in life.

Lastly, Rind et al. state that adolescents should not be called children. It would appear that Hagen agrees with this view ("there is, in fact, an important difference between adult males aroused by five-year-olds and men aroused by mature teens"). Yet, by including it in a larger citation which she criticizes, she implies that this recommendation by Rind et al. is unwarranted.

Hagen criticizes the meta-analytical technique in general for comprising and mixing so much data (including incompatible and corrupt data) that its outcome is more or less meaningless.

Hagen says that society does not have to prove that every single instance of CSA causes psychological harm to the child in order to blanket condemn it. "Every night, millions of people drive while impaired by alcohol and cause no damage to anyone or anything." She sees CSA as risky behavior, and therefore finds its blanket condemnation justified. There is a long answer to that view, but such a "social policy" discussion is not immediately related to the Rind et al. report.

Hagen is concerned about the agenda of Rind et al.. From criticizing some methodological shortcomings, she now jumps to the implication that the Rind et al. report is not scientific at all; just a biased piece of unwarranted speculation. She writes, "Rind et al.’s "scientific" meta-analysis of the long-term effects of childhood sexual abuse typifies a common problem when social science research is used to drive socio-political agenda - in their case, to reformulate the definition of "child sexual abuse" to exclude putatively consensual sex between adolescents and adults. Gaping empirical lacunae are filled in with items from advocates’ wish lists". But rather than pushing an illogical agenda, Rind et al. address the empirical lacunae by advocating a more cautious scientific approach (for example, through the use of neutral terminology in some instances).

Hagen writes, "Social scientists are not in the business of policymaking and their attempts to do so through the back door of journal article writing should be thwarted. Failure to do so renders a profound disservice to all science." I believe she gets phenomena mixed up here. There is no question of Rind et al. being involved in "policymaking." Rind et al. draw conclusions and make recommendations, which are quite correct and normal scientific practices. The conclusions await falsification or affirmation. The recommendations can be considered by other scientists, which they are primarily meant for, and by actual policymakers. But Rind et al. warn that "lack of harmfulness does not imply lack of wrongfulness". They write that "the findings of the current review do not imply that moral or legal definitions of or views on behaviors currently classified as CSA should be abandoned or even altered. The current findings are relevant to moral and legal positions only to the extent that these positions are based on the presumption of psychological harm." Hagen, then, says there is no irrefutable proof that CSA does not cause psychological harm, to which Rind et al. would correctly reply that it is illogical to assume harm where (for whatever reason) there is no evidence of harm (or even the contrary - a positive testimony - is found), to which Hagen replies that the risk of harm from CSA warrants blanket prohibition. Which is extremely debatable, but in a sociolegal realm that Rind et al. do not pretend to in their report.

(Some reasons, off the cuff, why the wisdom of a blanket prohibition is debatable: it creates a culture of fear in which touching and kissing minors as well as letting them be naked is illegal; it causes children to grow up in sexual shame and ignorance and with an unnecessary distrust of adults; it brands teenage sex as criminal activity, which is as counterproductive as the Prohibition, and which can get nonabusive teenagers in trouble and turn frustrated teens into abusers; the "risk of harm" may also be cultural, i.e. determined by factors not intrinsic to the sexual relations themselves; there is a sex abuse industry that thrives on the notion of pervasive, severe harm - the recovered memory movement should be an excellent example; political and religious groups can gain popularity and power by "protecting" minors from exaggerated threats - think attempts to censor the Internet; belief in pervasive, severe harm may create iatrogenic pathology in victims who were not originally victimized by "CSA"; if harm is usually absent, then blanket prohibition is a very radical measure to reduce risk and may in fact add to harm by getting innocents in trouble and feeding moral panic.)

Hagen’s main point seems to be that science knows too little about CSA. It cannot be denied that the tradition of biased and sloppy research in the field (let alone the pressure to produce politically correct findings) is partly to blame for this. In my view, Hagen is blaming Rind et al. for this shortcoming of the entire field which Rind et al. even addressed. Hagen accuses Rind et al. of pushing an agenda, while their conclusions and recommendations (for instance the suggestion that consent be considered a valid construct since it seems to affect the outcome) strike me as reasonable. So although I do not see unreasonable agenda-pushing at odds with the facts on the part of Rind et al., I will include this remark from an observer: "The only reason why they have been attacked is that they push a different agenda from the majority of scientists in the field."

Hagen ends on a positive note: "If it happens that future research does in fact support the conclusions of Rind, Bauserman, and Tromovitch - and there are certainly hints in current work that that may well be so - then all citizens concerned with the pursuit of truth, not just the scientists, must shout, not whisper, at our legislators, "You may not want to hear it or even know it, but Earth does move around the Sun.""

Chris, Netherlands

Peter responds:

[Chris wrote:] Hagen criticizes victimologists exaggeration of harm caused by CSA. She asserts that there is an important difference between adult males aroused by five-year-olds and men aroused by mature teens. She decries the Congressional condemnation.

Does she? And below she apparently says the opposite.

Hagen questions Rind et al.’s recommendations that (lack of) willingness be taken into consideration by researchers, that researchers differentiate between young children and adolescents, and that they differentiate between children whose reactions reportedly were positive and those whose reactions reportedly were not. She finds all of these recommendations too ambitious in view of the limited evidence, which she thinks is meta-analyzed into a cluttered mess.

These are rather disingenuous remarks. The factual base for these recommendations is the best available, and rather much better than the individual studies that RTB analysed. (Usually one does not hear similar critiques raised w.r.t. those studies.) The last sentence betrays an amateurish view of statistics, which is after all an applied science (applied mathematics, to be precise).

Meta-analysis was introduced into the social sciences because it is generally not possible to draw valid conclusions from a collection of original studies without mathematical methods, because they easily appear to contradict each other, or indicate many spurious relationships. This results from the research techniques used and the usually very small sample sizes. (In ‘Meta-Analysis’ (1982), Hunter /Schmidt/Jackson give an instructive example.)

In their unusually long paper (too long for the critics to read, it appears) RTB give a detailed account of their work. Their result is precisely that what a meta-analysis is made for: dissolving the mist of spurious results, that is caused by sampling errors (and generations of teachers have not succeeded in instilling in their students a measure of common sense with regard to so called significant results), and improving the confidence into the real facts. -- RTB have been meticulous. If one look on the last page, one will see that it took about one and a half year to peer-review and revise the paper...

Chris is correct. RTB discuss studies part of which define CSA as coerced (more or less) and others that include consensual activities. These are *levels* of the factor consent, and the one

includes the other, so the effect of this factor is supposedly not as visible. Thirty-five studies include willing CSA, 14 only unwilling. (Table 4 on page 33). (This shows Hagen being somewhat sloppy regarding the statistical methods used.)

Their result is a little triumph of the scientific method, isn’t it? Would you have been able to tell from the raw data (appendix, p. 52-53)? With any confidence?

"Informed consent" is not a fact, a quantity or a psychic state or something else that could be observed or assessed in an individual, but simply means that some formalisms have been obeyed. If the critics were sincere they would look for parental approval (that would ordinarily be asked in cases where a child would be deemed incapable of consent, although there are exceptions).

Regarding ‘victim’s’ age, and this shows again the generally poor quality of this criticism, one should take note that "[s]imilarly, maximum age of the ‘child’ in the study’s definition of CSA was not related to the effect sizes, r(44) = .05, p > .70. two-tailed" (p.34)

If one would for the moment accept RTB’s proposal and restrict the notion of CSA to unwanted sex only, or to sex with prepubertal (how-ever defined) children, how much would it buy?

That consent matters has been demonstrated. The effects are a bit higher, but not much. The age of the child doesn’t matter, as far as they could check it, and looking at Ford and Beach or similar sources, I’d doubt that it matters much. (There is from the developmental viewpoint no reason to assume a susceptibility to harm that is a monotonically decreasing function of time, b.t.w. It could be greater in periods of faster development, e.g.)

One does not only use self-appraisals, although one cannot do without them, in this case. The studies used also common psychological tests. Sandfort, b.t.w., whom Chris cites, used also objective methods, albeit, if I remember correctly, not as simplistic ones as the Americans like. (See Bauserman’s rejoinder to the criticism of Sandfort by various luminaries of the abuse industry in J.Homosex, special issue, Brongersma, Sandfort and N.N eds.)

That is not the case, as RTB did look into this question (vide supra). On the other hand she now proposes as scientifically warranted a position that simply ignores the fact that adolescents are sexually active all the time?

And consider that first she says that the little children are (only) affected and the older ones dilute the effects until they disappear (which is in itself possible), and now the contrary...

On a second thought, there appear also to be ways in which younger children would be more resilient. Certain emotional risks that threaten adolescents are far from the younger children’s world.

[...]

Hagen complains that the analyzed studies "disproportionately included subjects whose CSA experiences were relatively minor". But the overinclusion that the CSA concept has fallen prone to is not the fault of Rind et al.. (See also question & answer 5 at Science & Morality, a refutation of some common criticisms of the meta-analysis.) Also, Rind et al. found that studies that researched only contact CSA did not show significantly more (or less) effect than studies that researched all CSA. They also looked at a number of moderators that would indicate more serious CSA, and found that duration, frequency and penetration were not related to the effect size, whereas force and incest were.

"Disproportionately" w.r.t. what? Sure, the samples were mostly convenience samples, I suppose, but they did not differ much from the national probability samples used in the first meta-analysis, did they?

Whom does she address? RTB found that most of these factors were irrelevant, so they were looked at, supposedly. As Hagen apparently protests against the recommendation by the authors to restrict the notion of abuse to cases more probably harmful, as indicated by the evidence, (not only to cases where harm did occur), it can be said that she has a problem with the logic as well.

[...]

Rind et al. propose the introduction of the value-neutral term "adult-child sex" alongside "child sexual abuse," suggesting that (lack of) willingness be considered in future assessments. Hagen thinks Rind et al. do not have enough information to be sure that adolescents can consent (the American Psychological Association seems confident about children’s ability to give simple consent and adolescents’ ability to give informed consent in nonsexual situations, though), or that persons with a CSA experience will not be affected negatively at any point in their lives. Perhaps, but Rind et al. do not claim that henceforth researchers should speak of "beneficent adult-child sex" or "beneficent adult-adolescent sex," which would be the opposite of "CSA." Instead, based on the strong suggestions (not irrefutable proof) that there are SA persons who do not experience negative effects or who even experience positive effects, Rind et al. call for the logically more appropriate use of neutral terminology in some instances. Instances, for example, in which subjects report that they consented - which appears to matter; forget about irrefutable proof - and experienced no harm so far, with there being no evidence that harm will follow much later in life.

That the CSA construct is over-inclusive is quite evident even without the meta-analysis, and from the beginning it was designed as such. In fact this is more a sociological problem.

Only if CSA is as pervasive and harmful as it is portrayed by the child-savers, does it work for them. One must consider that they are not interested in the children, else they would have proposed a more precise and professional useful definition. (CSA is not a medical condition, and it is highly unprofessional and unethical to treat someone for it. Same for psychology.)

This works also in the direction of effects: not only is a wonder disease that causes whatever you want "from acne to death" (J. Money) research-efficient -- of 100 symptom-associations five are on the average significant so CSA research is never in vain -- it also provides great latitude for all involved. Scientists’ Marlboro Country (‘Freiheit und Abenteuer’ as German movie-goers know).

If symptom-associations fail to materialise, it is said one should look for symptom-combinations; if this fails, too (RTB checked this), then one is told it is some symptom not checked for, or even one only to be found after psychiatric examination (Spiegel, Teicher).

This is possible, just not probable. -- This all boils down to a sort of nursery-science: there are innumerable children severely harmed, and if they do not appear in the statistics, it’s because they were abducted by aliens. Any questions?

(Almost all factors believed to be aggravating the CSA experience fail to do so, compare penetration, frequency, duration; i.e. the whole idea is flawed.)

What I always found unspeakably silly is the idea that harmfulness were a Heaviside function of time, 100% until 18 (14 in Germany) and then 0%. Scientifically sound is only the idea of a positive probability, which would never be zero, and as we may now conclude, which is not so great at all.

However, what we find mostly is more of a cross-fertilisation. The newly prominent social problem CSA adds some spice to an otherwise dull professional life. After she has felt the thrill of drawing an audience by speaking about CSA a pediatric will easily forget about the dubious sources of her knowledge; public acclaim filling the holes in her arguments.

What I, however, do not understand at all, is the loosening of standards that has occurred. So the pediatric practitioner who after having spent a week with an American psycho-sect diagnosed one fifth of her pre-school boy patients as having been raped could place their hallucinations in as reputed a journal as the Lancet, although this finding is hardly compatible with the received wisdom. So someone without any prior research experience can publish something that would contradict almost anything one knows about age preferences, sexual practices etc. and must not even bring some corroborative evidence?

That people as fanatic and dimwitted as the CSA sects would be exploited by police, politicians and similar scum goes without saying, those have their interests to guide them, but for the reasons that guide professionals, editors and the like, I’m at a loss.

Katharina Rutschky considered this a special case of anti-intellectualism, which is attractive because it relieves one of the rigour and the anxiety at the same time.

Hagen has no idea what she is writing about. The opposite is the case. Meta-analysis does not help against systematic errors but then no statistical method can do anything about it, e.g. using non-representative samples.

To imprison someone without reason is a crime. I do not see that in Germany there is anyone prepared to say that it should be different. If one would accept the results of RTB, the law could stand no longer as it is.

It is obvious that the legal profession feels the same. The most outlandish reasons are proffered, e.g. the law on child-pornography is considered as part of a sort of economic warfare: we starve the street-children of the third world into modesty, so to say. If it is foreign policy, it is best, as there is no judicial review in this area, an idea from the ancien régime that comes handy. If it is economic crime it also makes a big difference because your business can be much more easily controlled and searched than your private life. So hapless citizens appear in court as international porn traders and have to endure the most exacting economic analyses from the bench.

A police search is an armed robbery, if successful, and in principle there is no reason not to expect persecution of police and the abetting ministeresses once C.P. is no longer seen as a question of child saving but as a grave insult to civil liberties. Compare the Sharpe case in Canada. It is mostly a question of perspective, less so of principle.

So again one sees how important the CSA construct is. -- One critic writes RTB are so antediluvian that they believe in objective facts and the like, whilst contemporary social science has long started to deconstruct everything. The most vitriolic critic was however sparked by exactly that what was correctly understood as the beginning of the deconstruction of CSA, and rightly so. As soon as one starts discussing it, all will fall apart.

(B.T.W. Victimless crime is such an abomination for liberal thought because it is precisely the rooting in the society, that it lacks. It is a product of ideology and propaganda; note how, if any victim at all is presented, care is taken that he be well prepared. So these are not moral crimes, in the sense that they violate moral norms, understood as the result of the society’s life and tradition. They exist in an amoral society, and they destroy the moral fibre of a society.)

But we come to the heart of the matter:

Such a proof is impossible, of course, so this whole argument is void. The basic argumentative structure of all criticism of RTB’s work is the same: after some methodological points, not necessarily as confused as in this case, one come to the real argument, invariably the same, viz. how dare they discuss something that we don’t want to discuss. But they dare; such is life.

But surely enough to found a politics of terror on it? --

[...]

So there is a little dark cloud on the horizon, and we have to wait for the thunderstorm.

Regards, Peter