Reports of Recovered Memories in Therapy,Informed Consent,and Generalizability: Response to Commentaries

The various commentators on our Patihis and Pendergrast (2019) article raised some interesting, counterbalancing, and valuable ideas. Loftus and Teitcher (2019) noted that they are concerned by the suggestion that clients be given informed consent about the hazards of repressed memories. Cannell, Hudson, and Pope (2001), arguing in favor of informed consent before therapy, suggested that therapists “inform patients concerning the risk of recovering false memories” (p. 138). We understand that Loftus and Teitcher (2019) are concerned specifically that the trust in the therapy process might be undermined by such warnings. Nevertheless, if the informed consent is written in the correct way, we think the knowledge that memory is reconstructive in nature and malleable would help prevent great harm. Clients should be aware that the concept of repressed memory/dissociation is hotly debated within the psychology profession. People who are aware of the malleability of memory would still be able to benefit from evidence-based therapy techniques. Memory distortion should be avoided in therapy because, in most cases, it is morally questionable. Except under unusual circumstances, experimental, unproven techniques should not be allowed if they have the potential to fragment families and produce illusory memories. By way of comparison between the fields of psychology and medicine, consider whether unproven, controversial brain surgery would be permitted. It is advisable that clients entering therapy that involves memory recall should be informed of the potential hazards of false memory production. Such memory distortions could disrupt their own and others’ lives, sometimes leading to suicide (e.g., Pendergrast, 2017, pp. 9, 140, 300, 408). We believe that such informed consent would not “deter patients from seeking treatment” (Loftus & Teitcher, 2019, p. 26) but would instead make them aware of the controversy surrounding repressed memories and let them ask questions about it if the subject arose.

There are some limited exceptions where memory alteration may be accepted: for example, techniques that help reduce strong emotional reactions to remembering trauma (without changing factual elements of the events, for example, in scientific approaches to exposure therapy). We believe that there are plenty of therapy options that do not involve potentially unethical memory distortion, such as cognitive behavioral therapy, although we acknowledge that evidence shows that practitioners who claim to do cognitive behavioral therapy may sometimes stray outside typical techniques (Hipol & Deacon, 2013; Patihis & Pendergrast, 2019).

We agree with Loftus and Teitcher (2019) that informed consent would have to be worded very carefully so as not to undermine good therapy techniques, while still informing people that memory is malleable and that people are unlikely to completely repress years of traumatic childhood events. Indeed, those concerned about such informed consent, such as Loftus and Teitcher, should be consulted when informed consent wording is drafted. In response to Loftus and Teitcher’s concern that people would not read informed consent pamphlets, perhaps the therapist should read part of the information out loud to the client at the beginning of therapy. Regardless, a pamphlet explaining the controversy should be available in waiting rooms, whether clients read it or not. Fries and Loftus (1979; Loftus & Fries, 2008) were concerned about informed consent introducing suggestion of symptoms, and we would agree that care should be taken to avoid such wording that may lead to false reporting of such symptoms. Nevertheless, we ask the commentators and readers whether they would be happy if they or a loved one entered therapy blind to the knowledge that memory is malleable, subject to suggestion and distortion, and that the concept of repressed memories is debatable. We note that Lynn, Merckelbach, and Polizzi (2018) expressed some support for informed consent too, so perhaps a discussion could involve these two groups of commentators and others.

Loftus and Teitcher (2019) also referred to Loftus’s (1997) investigation exposing the objectively poor outcomes of many patients who made repressed memory claims. Our short survey did not assess such potential iatrogenic effects, but a future longer survey could include follow-up questions on symptoms such as those mentioned in the work by Loftus (1997; such as suicidal thoughts, hospitalization, employment, child custody, self-mutilation, and further assessment of family estrangement). The comments from participants in our study certainly offer evidence of estrangement and anecdotal mention of suicide (e.g., Patihis & Pendergrast, 2019, p. 12).

Lynn and colleagues (2018) share our concern that therapists who believe in repressed memories may instill this belief in clients and that potential iatrogenic effects may result. Future research could delve deeper into this issue, while taking into account that clients can also learn about repressed memory theory from books, the Internet, television, and other media. In some cases, we think that some clients can be self-educated in repressed memory theory and lead the way to memory recovery in or outside of therapy. Whether this is therapist led or client led, we are concerned about memory distortion and the potential of family estrangement based on untruths. We agree with Lynn et al. that the central claim in dispute—in modern parlance—is whether trauma leads to selective dissociative amnesia (which we argue is a similar concept to repressed memories). Even if trauma is shown to correlate with dissociative experiences, such as feelings of detachment or depersonalization, it cannot be emphasized enough that this does not provide evidence that memories are blocked in dissociative amnesia. We are grateful for the suggestion by Lynn et al. that future research could parse how many recovered memories of abuse involve reinterpretation of remembered events as being abusive. We add that we may also parse how many meant suppression of memory and how many meant full recovery of repressed memories (despite our key question including the phrase “that you did not even know happened before therapy,” which implies repression rather than suppression).

Shaw and Vredeveldt (2019) asked the question about whether our U.S. sample findings would generalize to Europe and provided some evidence that there may be a comparable repressed memory component in some European psychotherapy practice. Indeed, it makes sense that the countries close to the birth of the concept of repressed memories and dissociative amnesia (started in Austria by Freud, 1893–1895/1953, who was influenced by the ideas of French hypnotists) still perpetuate the meme of repressed memories. Shaw and Vredeveldt discussed research showing that most cases referring to false memory organizations in Europe involved some connection to psychotherapy (e.g., Shaw, Leonte, Ball, & Felstead, 2017). They also outlined a case in the Netherlands in which a television program credulously educated the public about repressed memories, something that has been and will be a problem in many countries. Many more people watch such programs compared with how many read psychology articles (such as this). Despite the potential presence of therapies using repressed memory in the United Kingdom, the Netherlands, and France, it is heartening to know that there are researchers there communicating skepticism (e.g., United Kingdom: Ost, Wright, Easton, Hope, & French, 2013; Shaw, 2016; Netherlands: Otgaar, Muris, Howe, & Merckelbach, 2017; France: Dodier, 2018).

Goodman, Gonzalves, and Wolpe (2019) offered some valuable skepticism of our article. Their emphasis appears to be that memory for traumatic events can be accurate. Of course, this is true. Indeed, traumatic events tend to be recalled all too well, as posttraumatic stress disorder demonstrates. It is also true that false memories can be implanted or that true memories of real events can be reappraised (e.g., the reappraisal of the intention of a parent bathing his or her child). Goodman et al. (2019) stated that they are unsure that therapists should refrain from discussing “lost memories.” In contrast, we argue that therapists should not be encouraging the reconstruction of memories of anything that the client did not know happened prior to therapy. This is because of the danger of false or distorted memories of mistreatment that could both traumatize clients and unfairly lead to estrangement from their parents. Such estrangements of parents and children are highly negative for the long-term support of psychotherapy clients.

Goodman et al. (2019) appear to believe that therapy can help people recall abuse as infants, although by definition, infantile amnesia makes that unlikely. They wrote that “most adults can remember highly consequential, even traumatic, events if they happened to them back to age 3.5 years, often with considerable detail” (p. 30). They then went on to talk about “pushing it [abuse] out of their minds,” and they suggested that “some of them [memories] are not accessible without reminders or prompts” (p. 30). Goodman et al. then cited Williams (1994) in asserting that some people can recover memories as far back as 2.75 years old, and they believe that early traumatic memories might be retrievable with cues and reminders. Here, we believe they are treading on dangerous ground because such reminders could be suggestive and because all memory is reconstructive. As an example, they say that some people suffered such extreme abuse as children that they might not recall particular isolated incidents. That is true, but most severely abuse people would know that they had been severely abused throughout their childhood (after infancy), even if they did not remember every instance. And it is dangerous to assume that there must be more that must be recalled in order to get better. For the few clients who do not remember documented abuse, whether it is because of normal forgetting mechanisms, infantile amnesia, time, or a physical head trauma, in order to get to the truth, physical evidence should be prioritized over postevent memory reconstructions. Reconstructed memory that happens in the context of the client not previously knowing about the events that are recalled is not reliable. In weighing evidence, we would place always-remembered abuse as more reliable than recovered repressed memories.

Goodman et al. (2019) cited Williams (1994) as evidence that many people may forget abuse. Here is not the place to critique that study, but Loftus, Garry, and Feldman (1994) and Pendergrast (2017) have done so. The study by Williams (1994) and similar studies have indicated that people do not report abuse because of embarrassment, ordinary forgetting, memory interference of other traumas, infantile amnesia, and so on. The lack of reporting a documented trauma is not evidence of repressed memory.

We also argue that, in addition to avoiding false memories, therapists should refrain from reappraising real events in ways that are destructive to their family relationships. We are concerned that reappraisals of childhood events during therapy can lead to distortions in memory of the emotions that people once felt during such events (for the effects of changing appraisals on memories of emotions, see Levine, 1997; Levine, Lench, & Safer, 2009). For example, re-evaluating childhood events in the negative direction could lead to misremembering how angry or distressed the client was during that time. These changes in memory for emotions could also lead to family estrangement or at least to damaged family relationships. If therapists damage family relationships, they need to show evidence that in the decades that follow, the relationship with the therapist is more supportive than the relationship with the estranged family would have been. We acknowledge that some families are indeed abusive but argue that such genuinely abusive families leave continuous memories and knowledge of the trauma in the abused and others. Therapeutic relationships often are relatively short lived (compared with parental relationships), are conditional on insurance payments, and involve less kin altruism; therefore, in most cases, the therapy should not damage the relationship with parents. This is true even if the client reports being empowered by the victimhood or independence gained in therapy via memory distortions.

Goodman et al. (2019) also noted that sometimes people might categorize suppressed memories or reappraised continuous memories as being repressed memories. This is certainly possible, as we noted earlier. Nevertheless, we encourage the commentators and readers to examine the comments made by the participants in our Supplemental Material (Patihis & Pendergrast, 2019) to test the idea that participants were not talking about unconscious repressed memories that the client was not previously aware of. Goodman et al. appear to question the use of Amazon Mechanical Turk, but there are several articles that investigated the value of such participants (e.g., Buhrmester, Kwang, & Gosling, 2011; Gosling, Vazire, Srivastava, & Oliver, 2004), and in general, the conclusions are that such samples can provide high quality data. Our data set appeared to be of good quality, and for the purposes of our study, there were many advantages to being able to survey adults of all ages spread across the United States. We want to emphasize the value of asking the general public about recovered memories in therapy; it may be a way of getting to the truth of what some therapists do in therapy. Asking the therapists themselves may result in cautious reporting because they may know the litigious controversy surrounding repressed memories.

Our and Goodman et al.’s (2019) concern about generalizability should spur follow-up research with differing samples. In Patihis and Pendergrast (2019), we investigated representativeness on gender, ethnicity, socioeconomic status, and age and made adjustments where possible. Goodman et al. provided no evidence of their claim that these participants have a higher interest in psychology than the general population or any evidence that they did not know what the therapy type was. Also, the therapy modality’s ideal practice recommendations likely did not always match the actual practice, as we noted. Contrary to Goodman et al., there was no evidence either way of nonmotivation to answer honestly, and in fact, there was some evidence that many participants were happy to spend more time on the survey than was needed. For example, they could have skipped through the questionnaire by answering “no” to key questions, and they could have skipped over the open-ended questions completely, and many did not (see their comments in the Supplemental Material).

Goodman et al. (2019) wrote about the long-term negative effects of childhood trauma, and we support efforts to reduce such trauma. They wrote, “Risks associated with clinicians not asking about childhood trauma, remembered or forgotten, are arguably greater than risks of creating false memories” (p. 30). We are not, of course, suggesting that clinicians should not ask about always-recalled childhood trauma. But we strongly disagree that they should imply to clients that they may have been abused and that they repressed the memories.

We argue that addressing childhood trauma will be more effective when as few as possible memory distortions occur in therapy (or outside of therapy). The utopian goal of eliminating all child abuse by shifting criteria to accept a lower quality of evidence (such as recovered memory reports inconsistent with earlier reports and evidence) would result in many innocent people being unfairly accused. Society should enforce child abuse laws on cases in which the person remembers being abused and the evidence aligns or on cases in which the person does not remember but there is good evidence (e.g., video, photographs, messages, physical evidence). As Goodman et al. (2019) affirmed, there may be large numbers of such cases of real abuse, although we are skeptical that there are many cases of genuine abuse that are repressed. The gain to society of such enforcement is a general good, as long as it is empirically grounded and there is a reasonable attempt to minimize errors, and such enforcement should not be based on ungrounded ideology or unfalsifiable theory. Memory distortions in therapy, whether they be major distortions as in false memories or more subtle distortions of remembered emotion (via reappraisals), will undermine the resolution of real child abuse cases. The false allegations that can stem from exhuming repressed memories could dilute the real cases.

In closing, Patihis, Ho, Tingen, Lilienfeld, and Loftus (2014) established that high percentages of the public believe in repressed memories, so there may be a demand for the recall of repressed memories in some of the population who seek psychotherapy. That study also established that many therapists agreed with the theory of repressed memory, so some therapists might be willing to meet that potential demand from the public. In Patihis and Pendergrast (2019), we found that indeed a sizable percentage of the public in our survey sample (4%, which would mean millions of cases if the sample is representative) reported recovering memories in therapy that they were previously unaware of. In light of this, we encourage researchers to not assume that repressed memories in therapy are a thing of the past and to examine and expand research such as ours as well as examine recent accounts of ongoing problematic memory recovery practices (e.g., Brown, 2018; Efrati et al., 2018).