False Memories and True Memories of Childhood Trauma: Balancing the Risks

Patihis and Pendergrast’s (2019; this issue, p. 3) research raises questions about balancing risks of false memories with risks of not treating childhood trauma that may have been forgotten. Their central concern is that when clinicians ask about repressed memory, many clients will form false memories of child abuse.

They emphasize suggestibility/false memory, but their review omits important studies that moderate their concerns. For example, when Pezdek, Finger, and Hodge (1997) tried to implant a false memory in adults of receiving a childhood enema, the error rate was zero. Although some adults, including with trauma histories, agree with schema-consistent false suggestions about childhood events, when it comes to taboo acts of a sexual nature, Goldfarb, Goodman, Larson, Eisen, and Qin (in press) again found zero false memories. Many suggestibility/false memory studies use creative coding, such as when “partial false memories” (“That never happened to me, but if my mother said it did, it could have been at the mall”) are nearly buried in statistics reported.

That said, we acknowledge the reality of false abuse memories in some individuals as possibly induced or encouraged by therapists, particularly those who use hypnosis or psychotropic drugs (e.g., in combination with religious or other doctrines; Bottoms, Shaver, & Goodman, 1996). Still, it is unclear that clinicians should refrain from discussion with clients about lost memory (a term we prefer because it does not invoke “repression” mechanistically), given that therapy can help memory: Child victims who sought therapy during/soon after legal involvement (vs. did not) had more accurate long-term memory for abuse a decade later (Goodman, Goldfarb, Quas, & Lyon, 2017).

We note that Patihis and Pendergrast (2019) leave largely unaddressed that lost memories of childhood trauma can occur, as can recovery of them. In longitudinal studies of documented child sexual abuse (CSA), 15% to 38% of victims failed to recall the target case. Of those who recalled it, 15% to 16% said there were times of not remembering it (Goodman et al., 2003; Williams, 1994). More severe abuse was associated with reporting having forgotten the CSA but also actually predicted greater memory accuracy for it (Ghetti et al., 2006).

Relevant are our past findings regarding misinterpretation of research questions. Victims often misinterpreted questions about “repressed memory” and “amnesia,” wrongly assuming we were asking if there was a time when they were not consciously thinking about the CSA experience (Ghetti et al., 2006). We do not know how Mechanical Turk workers interpreted the questions asked. For example, clinicians, as mandated reporters of child maltreatment, often have clients sign an acknowledgement of such, which could lead to what participants interpret as instigating discussion of (lost) memories of abuse. Moreover, given variation in interpretations, survey respondents who reported they came to remember being abused as a child may mean they reinterpreted true childhood experiences as abusive (e.g., realizing a “whooping” that left bruises was physical abuse).

In our study of CSA victims who reported periods of complete forgetting, frequently endorsed reasons were that the CSA was so horrible and frightful that they pushed it out of their minds. These responses reflect individuals’ active attempts to avoid thinking about the trauma. Others indicated that it happened so often that they could not remember it all; they had (naturally) lost memory for parts of the repeated assaults. A few participants indicated they had forgotten about it because they did not think the CSA was important. Given these responses and the links of child maltreatment to psychopathology (Edwards, Dube, Felitti, & Anda, 2007), clinicians may need at times to ask about lost traumatic memories.

Although we cannot rule out, from the studies, unwillingness to disclose, there are reasons why lost memories might be awakened in therapy. One reason is related to infantile amnesia (i.e., the absence of explicit memory for one’s early life events). Most adults can remember highly consequential, even traumatic events if they happened to them back to age 3.5 years, often with considerable detail. However, some people accurately remember traumatic events back to 2.75 years of age (Williams, 1994); others’ autobiographical memory starts at age 6 or 7. It is an empirical question (one we are currently examining) whether these early traumatic memories may be retrievable with cues and reminders. It is well established that cues and reminders can heighten recall and that spreading activation (which may take some time) can increase memory access (Howe, 2011). Thus, retrieving long-ago events after some time, thought, and prompting fits with scientific research.

Another possible reason for lost and recovered memory relates to polyvictimization: having so many childhood traumatic events and sources of distress that some of them are not accessible without reminders or prompts. Adults with traumatic childhoods often tell us that they do not remember the target maltreatment because they had such chaotic childhoods (e.g., intra- and extrafamilial assaults, multiple foster homes, living on the streets) that it is hard to remember it all. Their backgrounds are quite different from those of participants in most undergraduate samples studied by many research psychologists.

Note that individual differences exist in coping with trauma that affect true and false memory. Although most people remember traumatic events particularly well, those with more attachment-related avoidant coping show greater loss of detail from such memories (Edelstein et al., 2005). Avoidant coping could help lead to a clinician’s couch.

Finally, in extrapolating their findings to the general public, the authors do not emphasize several important factors: for example, the Turk sample’s likely strong interest in psychology, making the survey results not readily generalizable to the U.S. population; the possibility that clients first raised the repressed memory topic themselves; respondents’ likely lack of knowledge as to what types of therapies they received; participants’ possible nonmotivation to answer honestly; and behavioral/cognitive behavioral therapies (which were especially frequent) resulting in the greatest number of individuals asked about repressed memory.

Repression as a mechanism is hard to validate. To the extent Freud meant lost memory for a childhood trauma resulting from avoidance of memory, many within experimental and clinical psychology can likely agree that such forgetting exists. But to the extent that Freud was referring to the complete loss of conscious memory for trauma that had negative implications for the self (and thus subject to repression) and was retrievable with a supportive therapist, there is still valid disagreement. However, children with insecure attachment (or with parents with avoidant attachment) benefit from a supportive interviewer for providing their memories (Chae et al., 2017; Milojevich & Quas, 2017). This is consistent with Freud’s concept that memories are sometimes accessed in a supportive, therapeutic context.

Note that mounting scientific evidence exists for pervasive, toxic, long-term effects of childhood trauma on mental and physical health (Edwards et al., 2007). Recounting memory for actual childhood maltreatment in a therapeutic context is likely an important part of the healing process. Risks associated with clinicians not asking about childhood trauma, remembered or forgotten, are arguably greater than risks of creating false memories—with the risk assessment surely guided not only by science but also by one’s values and fears. For perspective, the Innocence Project documents 358 cases of DNA exonerations since 1989, and many of the original convictions were based on faulty eyewitness memory. However, in 2016, there were 676,000 child victims of substantiated maltreatment in the United States (U.S. Department of Health and Human Services, 2018). That is almost 2,000 times more in a single year. Both figures likely represent “the tip of the iceberg”; both are extremely serious. Although research can guide us to avoid error, we urge scientists and clinicians to confront multiple risks involved, not just the risk of false memory.