Appendix: Three Papers
In the most recently released study, Linda Meyer Williams surveyed 129 women who had documented histories of having been examined as children for sexual victimization at a hospital between 1973 and 1975.1 The abuse the parents of these children reported them to have suffered ranged ftom fondling to rape. The women, who were between ten months old and twelve years old at the time of their abuse, were interviewed seventeen years later. Williams found that forty-nine of these women had no memory of the event documented in the hospital records." The finding that 38 percent of the women did not tell the interviewer about the child sexual abuse which was documented in hospital records . . . is quite astonishing," she writes of her work. From her results she draws a number of broad conclusions supporting the predisposition of recovered memory therapists to hunt for abuse. "Having no memory of child sexual abuse is a common occurrence, not only among adult survivors in therapy for abuse but among community samples of women who were reported to have been sexually abused in childhood. . . . The current findings . . indicate that therapists should be open to the possibility of child sexual abuse among clients who report no memory of such abuse. "2
However, if we look at her methods and results, the picture becomes much less clear. First, the fact that thirty-three of the forty-nine women who could not remember the specific recorded abuse freely told of other incidents of childhood molestation is not mentioned prominently in the paper. This leaves sixteen of the 129 women (12 percent) who both reported no memory of the specific incident as well as no memories of the other abuse in childhood. There is no reason to believe that normal forgetting wasn't the cause of their inability to recall their experiences. No one questions the fact that almost all experiences-especially those of young children-are eventually forgotten. Looked at in this light, the most dramatic conclusion of the Williams study might be that close to 90 percent of women abused as children know as adults that they were abused, although they may have forgotten one or more specific instances.
The fact that many people who know they were sexually abused as children forget a specific instance is simply not, however, evidence for a repression mechanism but only for the fact that traumatic memories can sometimes be forgotten. Williams's own numbers bolster this conclusion. For example, the women who did not recall the abuse recorded in the hospital records were just as likely to report other instances of abuse as the women who did recall the hospital event. This contradicts the theory of Dr. Lenore Terr and others cited in Williams's paper that children who become able to repress trauma are more likely to repress subsequent abuse. The finding that the women who could not remember the target event were just as likely to remember other abuse is compelling evidence that these women lacked any special ability to force disturbing events from their consciousness.
It is troubling that Williams doesn't examine the sixteen women with no memories of abuse (neither of the specific event nor of other episode in their childhood) separately from the larger group of forty-nine who remember being abused but couldn't recall the target event. Because she doesn't tell the reader the type of abuse these women suffered (the abuse of the 129 women ranged from rape to nonviolent fondling) nor their ages at the time of their treatment, it is impossible to say conclusively why these memories were absent. There is no reason, however, to assume some poweful repression mechanism was responsible. Some of these sixteen women may simply have forgotten their experiences while others may not have wished to disclose any abuse memories to the interviewer. Since there is no control group (for instance, of children treated at the hospital for other reasons), Williams gives the reader no reason not to believe that garden-variety forgetting was responsible for those failing to remember the recorded event.
Williams also claims that her study has dramatic implications for the correlation between age and the ability to remember abuse. While the results show a predictable correspondence between age and recall (those who recalled the abuse were two years older on average than those who didn't), Williams purports to have found that nearly half of the women abused when they were under four years of age remembered their abuse seventeen years later. These findings would indeed have a dramatic effect on the assumption that very young children forget almost every experience (traumatic or not) were it not for a remarkable footnote in which Williams admits that: "Some of the 'memories' may be attributable to information they received from others later in life; however, this was not explored in detail in this interview." The fact that she made no effort to distinguish between abuse the subjects remembered and events they had learned of only through other sources throws the usefulness of the entire study into question-particularly in regard to the "memories" of those abused before the age of four. Using Williams's criteria, if these subjects had reported knowing the circumstances of their birth, she would have classified them as having "recalled" the event.
The Williams study is also interesting for some things it doesn't mention. Apparently, nowhere in the hospital records is there any indication that these young victims forgot their trauma immediately or shortly after it occurred, as many recovered memory therapists attest is likely. In addition, it appears that none of the women who remembered their trauma described retrieving their memories after decades with all the feelings and emotions of the original event (as is ofien described by recovered memory therapists). Because she does not document the recovery of any forgotten memories, Williams admits, in the end, that her study does not address the "validity of 'recovered' or recalled memories of once-forgotten abuse or the association of such memories with adult symptomatology." Williams proves only that traumatic events suffered by children can sometimes be forgotten. Despite its constant use by recovered memory therapists as proof of their assertions, this flawed study tells us little new about memory (because it doesn't distinguish memory and knowledge) and nothing about the repression mechanism as it is postulated by many of today's therapists.
In a paper titled "Self-Reported Amnesia for Abuse in Adults Molested as Children," researchers John Briere and Jon Conte more directly attempt to prove the existence of repression. The pair located 450 therapy patients who identified themselves as abuse victims and asked them whether there had ever been a time before their eighteenth birthday when they "could not remember" their abuse. Answering this yes or no question, 59 percent of the subjects reported that they had experienced such a time. From this result the two researchers concluded: "Amnesia for abuse ... appears to be a common phenomenon among clinical sexual abuse survivors.', As to how their empirical conclusions should be applied to the clinical setting, Briere and Conte write that "it is likely that some significant portion of psychotherapy clients who deny a history of childhood victimization are, nevertheless, suffering from sexual abuse trauma. . . . Thus the clinician who has some reason to believe that his or her client was molested as a child . . . may be well advised to continue to entertain that hypothesis during treatment, even in the absence of specific abuse memories."3
It doesn't take a social scientist to see that the survey question Briere and Conte used to draw their far-reaching conclusion is that flawed at its premise. The question ("During the period of time between when the first forced sexual experience happened and your eighteenth birthday, was there ever a time when you could not remember the forced sexual experience?") turns on the "could not." Patients are not only asked to remember a time when they did not remember something, but to have known whether, during that time, they could or could not have remembered the event if they had tried. Asking patients to remember a time when they couldn't remember something is a logical quandary. The question borders on the ridiculous, for it assumes the subject would have knowledge of the status of a memory during a period when that memory by the subject's own admission never came into consciousness.
Retrospective studies to determine past symptoms, as we have noted earlier, are difficult because of patients' propensity to redefine their experiences in terms of the questions the doctor asks. In assuming that patients would know what they could and couldn't remember over a period of many years, Briere and Conte's question takes this problem to an extreme and appears designed to maximize the chances that respondents will conform their answer to the perceived assumptions of the questioners. Because of the broad and ill-defined nature of the question, people with relatively continuous memories were likely to respond that they experienced periods when they could not remember only because they experienced periods when they did not think about the abuse.
Another obvious problem with the study is the likelihood that Briere and Conte's pool of respondents was composed in large part of recovered memory patients who, by definition, would classify their therapy-created "memories" as inaccessible during their pretherapy life. If the memories of abuse were not "discovered" in therapy but rather "created" in therapy, the patient would, of course, believe that she had been amnesiac for the material. This is to say that the survey is only valid if the recovered memories are valid. Briere and Conte offer no evidence to counter this problem, saying only that due to their "clinical experience" they "doubt that abuse confabulation is a major problem in abuse research." The validity of this supposedly empirical study, then, rests on the clinical assumption that patients are not likely to create histories of abuse in therapy. Now that there is evidence that therapists-infused with a false confidence in their ability to identify the symptoms of someone who was abused-can and often do create false memories of abuse in the minds of their patients, that assumption should no longer be allowed to stand unchallenged. That Briere and Conte further encourage therapists to hold to their abuse interpretations regardless of the patients' memories (or lack of memories) is likely only to make this problem worse.
Judith Lewis Herman and Emily Schatzow's paper entitled "Recovery and Verification of Memories of Childhood Sexual Trauma" is perhaps the most-often quoted paper in the recovered memory movement. In this study the researchers worked with fifty-three women in group-therapy settings and encouraged them to search for evidence that their memories of abuse were accurate. They conclude that "three out of four patients were able to validate their memories by obtaining corroborating evidence from other sources."4 In reporting this result, Herman and Schatzow are, in fact, being modest. Because six of these patients (11 percent) made no attempt to confirm their abuse memories, it could be said that a full 89 percent of the forty-seven women who made an effort were able to confirm their memories.
This study, published in the prestigious Journal of Psychoanalytic Psychology, has had a profound impact on the recovered memory debate. It is cited often not only within the recovered memory literature but also in newspapers and magazine articles on the subject of recovered memory therapy. Herman's credentials as a doctor, her association with Harvard University, and her previous book, Father-Daughter Incest, gained her a national reputation. Her recent book Trauma and Recovery was hailed in the New York Times Book Review as one of the "most important psychiatric works to be published since Freud."
The first and most striking problem with applying the results of this study to the recovered memory debate is that it does not focus on women who supposedly repressed and then recovered their memories. Among the fifty-three women were three groups, those with no amnesia for their abuse (38 percent), those with "moderate" amnesia (36 percent), and those with "severe memory deficits" (26 percent). The vast majority of these patients, then, had at least some previous and apparently continuous knowledge of their abuse. Only fifteen of the fifty-three women had what Herman and Schatzow refer to as severe memory deficits, qualifying for this label if "they reported recent eruptions into consciousness of memories that had been entirely repressed, or if this kind of recall occurred during the course of group treatment." In reporting their results about clients who confirmed their abuse, Herman and Schatzow made no distinction between patients with previous knowledge of abuse and those who discovered that they were abused through therapy.
According to the researchers, nearly all the women who couldn't remember abuse on entering treatment were able to find their memories during these group encounters. The authors attest that "participation in the group proved to be a poweful stimulus for recovery of memory in patients with severe amnesia." It is worth wondering why, if these women were initially amnesiac for their abuse, they signed up to participate in therapy groups specifically for "incest survivors." That is to say, how did they know they were sexually abused by family members if they had no memories of these events?
The case report of one of the women with "severe amnesia" answers this question and sheds a great deal of light on the work of these two therapists. As described in the paper, Doris, a thirty-seven-year-old housewife and mother of four children, went to couples therapy with her husband because of sexual problems they were having. According to Herman and Schatzow, it was the marriage counselor who suspected the sexual abuse and referred Doris to one of their incest-survivor groups. For the first five weekly sessions, Doris said almost nothing and listened to the other group members tell of their memories of abuse. In the sixth session, Herman and Schatzow write, Doris began to "moan and whimper and wring her hands. In a childlike voice she cried, 'The door is opening! The door is opening!'" Fearing that her memories were returning too quickly, the therapist running the group "instructed (her] to tell her memories to go away and not come back until she was ready to have them." Doris was reportedly able to do this by talking to her memories- telling them over and over to back into her unconscious. In the three weeks following this session Doris was "flooded with memories" of sexual abuse lasting from the age of six to the age of twelve, including rape by her father and being forced to "service" a group of her father's friends while he watched. Doris also remembered becoming pregnant by her father and being taken to an underground abortionist.
According to the paper, Doris was among the group of five women who could not fully "confirm" their memories but received evidence that indicated the "strong likelihood" that they had been abused. This evidence consisted of a sister asking Doris at a Thanksgiving dinner, "Did Daddy ever try anything funny with you?" (It would of course be tragically ironic if Doris's sister was in therapy herself and attempting to confirm her own newly visualized repressed memories.) Herman and Schatzow do not explain how or why this question should be accepted as evidence for the validity of Doris's new belief that she had suffered through years of sexual assaults, forced abortions, and gang rapes.
Doris's case is also interesting because it shows that the researchers didn't distinguish between confirming evidence that existed before the repressed memories were discovered and evidence that came to light after the discovery. The evidence for the validity of Doris's recovered memories came two years before the memories themselves. Indeed, it may have been this very incident which started Doris on the road to recovered memory therapy. Doris's sister's comment appears to have been the trigger for Doris's anxiety attacks which, in turn, were the reason she went into therapy. Because recovered memory therapy often begins with the identification of a potential abuser; it is likely that the perpetrator identified in the patient's repressed memories would be someone suspected to be a pedophile through another source-perhaps in Doris's case her sister's question. However, if a given piece of information is what started the patient on the hunt for repressed abuse, it should hardly be considered confirming.
The questionable way memories suffaced for the women with severe amnesia is intricately tied to Herman and Schatzow's conclusions on the confirmation of these memories, but it takes a little digging to understand why. Reading a paper Herman and Schatzow wrote three years earlier describing their group-therapy methods, one learns that group members were pressured to achieve a preset "goal" during the course of the meeting. As they describe it, at the beginning of the ten or twelve weekly sessions, patients were encouraged to define personal objectives for the course of treatment. For many patients described in the 1984 paper this goal was the recovery of suspected repressed memories. From Herman and Schatzow's description of those patients, it is clear that they felt considerable group pressure to find abuse in their past and in so doing affirm their group membership."Women who wished to recover memories," they wrote in 1984, "were often preoccupied with obsessive doubt about . . whether they belonged to the group at all," noting also that some women defined their goal for treatment by stating, 'I just want to be in the group and feel I belong.' "5
After the fifth session, Herman and Schatzow would remind their patients that they were reaching the midpoint of the therapy, in order to spur group members to "clarify their goals and begin taking action." This, no doubt, added to the pressure on the group members who had so far failed to find any memories or achieve their goal. (It is interesting to note that Doris began to report the suffacing of memories in the sixth session). Not surprisingly, the authors report that during this time the goals of the group members were often achieved in "chain reaction" fashion.
This sort of goal setting and group pressure has dramatic implications for the 1987 paper because it seems clear that the same sort of process was employed in encouraging patients to find "confirmation" of their abuse. "Participation in group therapy offered an opportunity for many patients to gather corroborating evidence of abuse," they wrote in the 1987 paper. "The groups were structured around the definition and achievement of a personal goal related to the abuse." The pressure to achieve the stated goal of finding confirming evidence by the end of the group sessions and the pressure to belong quite possibly had a significant impact on reports of "confirmation." It is not hard to imagine that the obligation these patients felt to achieve their goals, combined with pressure from the therapists and the group dynamics, might have affected the quality of what the patients accepted as "confirmation."
This leads to two other critical problems: Herman and Schatzow never clearly state what sort of evidence would qualify as confirmation, nor did they attempt to independently confirm the reports of the patients. They attest that 40 percent of the women obtained corroborating evidence from the perpetrator himself, from other family members, or from physical evidence, while another 34 percent confirmed their suspicions through the discovery that another child had been abused by the same perpetrator. Why did Herman and Schatzow not view any of this physical evidence for themselves or double check the stories by speaking directly to the source of the confirmation? It appears that the patient's preset goal was achieved by simply reporting to the group an account of having her memories confirmed. In addition, there is no way to tell what percentage of reports of confirmation came from the women who had continuous knowledge of their abuse over the course of their lives.
Cast differently, Herman and Schatzow's studies might be seen as illustrations of the power of group settings to elicit conformity from group members. However, in this light they would tell us nothing new. The conclusion that highly charged group settings are capable of producing conformity in belief is far from new or surprising and would add nothing to the literature on interpersonal and group influence.