[Back to The Rind et al. files]
Suffer the children: Long-term effects of sexual abuse
David Spiegel, Society, 05/01/2000, 37 -4, 18-20
The "Psychological Bulletin" ignited a storm of controversy by publishing an article by Bruce Rind, Philip Tromovitch and Robert Bauserman that appeared to offer evidence that there are few if any lasting ill effects of child sexual and physical abuse. Spiegel suggests that the study is seriously flawed in its assumptions, methods, and conclusions.
The prestigious journal Psychological Bulletin ignited a storm of controversy by publishing in their July, 1998 issue an article that would, on first blush, appear to be a beacon of hope. The authors conducted a so-called "meta-analysis," a statistical review of the literature on the long-term effects of sexual and physical abuse in childhood, and reached the surprising conclusion that there are few if any lasting ill effects. Here, it appeared, was evidence of the resiliency of the human spirit, a chance that the past would not be prologue.
However, the result has been a discounting of the seriousness of the long-term consequences of the abuse of children, and ammunition for those who do not take seriously those who have been abused and those who treat and advocate for them. Is this a case of the triumph of passion over logic, or mistaken conclusions masquerading as reason? As both a clinician and a researcher, I can understand the heat on both sides of this issue. There are many times when systematic research has challenged prevailing wisdom, to our ultimate benefit, no matter how uncomfortable the process of dispatching cherished beliefs.
If it were the case that the ill effects of being sexually abused by a parent or stranger in childhood did not last, and we finally realized it, that would be one thing. At the same time, as a clinician I see and treat individuals who suffer the effects of an abusive childhood: their depression, inappropriate guilt about the abuse, their erosion of self-esteem, mistrust in relationships, and difficulty in enjoying their sexuality. In this case, however, reason does not need to override intuition, because the Rind et al. study is seriously flawed in its assumptions, methods, and conclusions.
Rind et al. stacked the deck, slanting their methods in the direction of their conclusions. To begin with, they included in their review only studies of college students. They rationalize this rather odd choice with data purporting to show that the frequency of abuse is similar in non-college student populations. Even if this were the case, the severity could be different, and the consequences are undoubtedly different. Those they chose to study were least likely to be seriously affected by abusive experiences, because they had made it through high school to college.
By design Rind et al. ignored those so mired in drug abuse, criminal activity, prostitution, or financial and educational hardship, that they could not make it to college. This is a systematic bias in favor of their conclusion, and it results in not looking at those most likely to suffer adverse consequences from sexual abuse, making it easier to conclude that there are not any even before the studies have been analyzed.
The type of review technique they employed, meta-analysis, compares different studies statistically by weighting them, taking into account the "effect size," the magnitude of the effect observed, and also the number of subjects involved. The reason for this is that it makes sense to weigh more heavily an observation made among thousands of people than one based upon a study of 25. Yet it just so happens that some of the larger studies included by Rind et al. in their analysis involved very mild sexual trauma, for example situations in which teenagers successfully fended off an unwanted sexual advance, rather than being overpowered and physically victimized.
These studies were weighed heavily because they had a large sample size. Yet it makes sense that such situations would result in relatively mild long-term effects. They should not be combined in an analysis that also examines the effects of repeated physical and sexual abuse of helpless children.
Common and Serious Symptoms
The Rind et al. paper does not include an examination of some of the most common and serious symptoms which arise from sexual abuse. An entire two-column densely typed page in the paper is devoted to the measures they analyzed as evidence of long-term ill effects. It is thus remarkable that the extensive list does not include the symptoms of post-traumatic stress disorder, the primary syndrome seen after rape and assault. The authors thus make it look as though they are being comprehensive, but there is no reference to the most salient symptoms seen after traumatic stress.
Patterns of Abuse
The manner in which the analysis was conducted seemed to examine all the possibilities but actually minimized the power to detect an effect of sexual abuse on subsequent adjustment. Rind and colleagues conducted their analysis in such a way that they ask a large number of individual questionsdoes abuse produce depression, anxiety, etc., but they reduced the possibility of finding long-term ill effects by taking each possible adverse outcome one at a time, rather than looking for patterns. Where they did look at patterns, they did so to minimize the effect of abuse per se. For example, they entered family dysfunction as a variable into their analysis, using it to prove that it is the general atmosphere of family problems rather than abuse that accounts for later distress.
Yet the evidence says that it is both. Sexual and physical abuse often occur in families characterized by a host of other problems as well. The fact that having a bad relationship with one's parents in general leaves one miserable later on does not prove that sexual abuse did not contribute both to the bad relationship and the subsequent distress. Both conceptually and methodologically, the statistical approach employed in the article was impermissible and led to erroneous conclusions.
What Can Be Said?
Rind et al. did in fact come upon evidence of elevated levels of distress in abused samples, but they tended to minimize what they found. Furthermore, they made the false assumption that the failure to demonstrate such a relationship in their review means that there is none. Even if you cannot prove it, you have not proven the opposite-that it is not there. The most their review could prove is that we cannot be sure that there are systematic long-term ill effects of abuse. It could not prove that there are none, unless the studies showed superior mental health among abused children. That is not likely, and certainly was not the case in this study.
It was understandable that Psychological Bulletin would consider publishing an article certain to provoke outrage among clinicians and members of the public who morally deplore sexual exploitation of children. To conclude that there is no emotional damage caused by premature involuntary sexual experience easily invites the inference that is morally excusable. Rind et al. point out toward the end of their article that the absence of proven negative psychological effects does not mean that such activity is morally correct. This is true.
Yet they also recommend that a distinction be drawn between consensual and coerced sex with children, which is based upon the fallacious assumption that children can (or should be expected to) make a reasonable decision about having sex with an adult. This assumes that such a situation could conceivably be non-coercive. So while sex with children does not have to be emotionally damaging to be wrong, we had better remain clear that it is wrong. The strength of this feeling among most clinicians and the general public is underscored by the fact that the House of Representatives took the unprecedented step of passing a resolution condemning the Rind et al. article by unanimous vote.
We have statutory rape laws in all states declaring that sex with a minor amounts to rape, whether the child thought they were consenting or not. Our sexual harassment laws protect adults in similar ways, because we recognize that consent to sexual activity cannot be freely withheld in a context in which one adult (e.g. a supervisor at work) has power over another. If that is true for adults, it is surely true in spades for children, who have far less power, experience, independence, and ability to even understand what sexual activity is.
One of my patients, a successful businesswoman in her fifties, was discussing her memory of her father's first sexual advance toward her, when she was a teenager. She recounted a new discovery about an old hurl: "I realize that it wasn't just what he did to me physically. At that moment I lost my father. He was no longer a person to love and protect me. I was there to satisfy him." It is terribly damaging to experience the betrayal that comes with discovering that someone who is supposed to put you above all else is subordinating your needs to theirs.
This has been termed "betrayal trauma," by University of Oregon psychologist Jennifer Freyd, who notes that such a situation forces children to hide their natural reactions because they are by definition ambivalent. They want their parents to love them, and depend upon them for their survival. Thus they have to "get along," even when the cost of doing so is putting up with sexual and physical exploitation. This sets up relationship patterns that are laced with understandable mistrustwhy shouldn't anyone else who seems to care about me have a hidden agenda?
Furthermore, most trauma victims blame themselves inappropriately for situations over which they had no control. Oddly, it is less painful to think you brought a tragedy upon yourself than to face your vulnerability to mistreatment. Many rape victims blame themselves for not somehow having foreseen that a sudden assault would occur, even when it would not have been possible to do so. To make matters worse, children do not understand independent causation.
They are the center of their universes, and understand what happens to them as a natural result of something they have done. Abusers who call what they are doing deserved "punishment" for real or imagined misdeeds reinforce this tendency. They thus emerge not just damaged by the abuse, but feeling that they brought it upon themselves and deserved it. Their only "mistake" was choosing the wrong parents, but they see the situation as a natural product of their misdeeds. It would be stunning indeed if experiences of sexual and physical abuse by a caretaker did not produce anxiety, depression, preoccupation with traumatic experiences, and mistrust in relationships.
I believe that the Psychological Bulletin article should be dismissed not because it is distasteful or morally objectionable, but because it is wrong. It had the appearance but not the essence of good science, the goal of which is to test hypotheses in such a way that they can be proven wrong. This meta-analysis was conducted in such a way that the facts could not speak for themselves, any more than a child can when approached for sex by an adult. There is great beauty in the innocence of childhood, and it deserves our respect.
The barrage of violence and sex on television news, in movies, and in print is already invading it enough. Sex with children is morally wrong as well as emotionally and physically damaging, Rind et al. notwithstanding. Reason and common sense do not diverge here. The data could have been allowed to speak for themselves but were not. Statistical abuse has as many bad aftereffects as sexual abuse. We should not tolerate either.
David Spiegel, M.D., is professor and associate chairman, Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, California.
[Back to The Rind et al. files]