Why It Is Scientifically Respectable to Believe in Repression

A recent article (Patihis, Ho, Tingen, Lilienfeld, & Loftus, 2014) reported data suggesting that among therapists, belief in repressed memory is lower today than in the 1990s and belief in the fallibility of memory, particularly when retrieved through hypnosis, is greater today than in the 1990s. Although the therapists involved at the two time points were not the same individuals, these data offer welcome reassurance that clinical psychologists have generally been responsive to warnings about the dangers of false memories being uncritically accepted as true and have adapted their practices accordingly. However, unlike researchers, clinicians and undergraduates continue to report high levels of belief in the possibility of repressed memories. Patihis et al. described this as a “science-practice” gap (p. 519), assuming that repression is an unscientific concept and concluding that clinical-psychology practitioners are insufficiently educated about memory research. We argue that these conclusions are unjustified.

It has been pointed out many times that from the outset and throughout his writing, Freud adopted two quite different meanings of the term repression, one corresponding to a fully unconscious defense and one corresponding to a conscious defensive strategy (Bowers & Farvolden, 1996; Brewin & Andrews, 1998; Erdelyi, 1990). Whereas attempts to find experimental evidence for the unconscious version have been largely unsuccessful (Holmes, 1990), the conscious version of repression corresponds to everyday strategies such as thought avoidance and thought suppression. Avoidance of thoughts and memories is accepted as ubiquitous in psychopathology, and the ability of individuals to deliberately forget unwanted material is well established in laboratory research (Anderson & Green, 2001; Anderson & Huddleston, 2012; Geiselman, Bjork, & Fishman, 1983). The implication is that investigations of beliefs about repression must specify what type of repression is meant if respondents’ answers are to be interpretable.

In Study 1, Patihis et al. asked participants to respond to items such as “Traumatic memories are often repressed (which means the person cannot remember the traumatic event due to a defense against painful content)” and “Repressed memories can be retrieved in therapy accurately,” using a 6-point Likert scale anchored with strongly disagree and strongly agree. In our view, the high percentage of undergraduates agreeing with these statements at least “slightly” is not “surprising,” as the authors claim (p. 522). Rather, it reflects the belief that deliberate attempts to forget unwanted events and the later coming to mind of forgotten events are common in everyday life (Anderson & Huddleston, 2012; Smith & Moynan, 2008).

Further findings of this study indicated that belief in repressed memories formed a factor with other beliefs that, although controversial to some degree, do possess empirical support (see Supplemental Results for Study 1 in the Supplemental Material). For example, the belief that “hypnosis can accurately retrieve memories that previously were not known to the person” is consistent with evidence that hypnosis may sometimes lead to both additional accurate recall and additional inaccurate recall (British Psychological Society, 2001). The belief that “some people have true ‘photographic memories’” is consistent with evidence for exceptional memory abilities in a very small number of individuals (LePort et al., 2012) and was endorsed by approximately 50% of researchers in Study 2 (Table S2.2 in the Supplemental Material). It does not seem unreasonable to agree at least “slightly” with these statements. Interestingly, other straightforwardly erroneous beliefs (e.g., “with effort, we can remember events back to birth”) did not load on this factor, and beliefs in memory fallibility formed a separate and orthogonal factor. We are not persuaded that this pattern of responses demonstrates evidence for widespread faulty beliefs about memory.

In Study 2, Patihis et al. examined undergraduates’ ratings of the likely accuracy of recovered memory. In both the 1990s and 2011, the undergraduates’ mean rating was at approximately the midpoint on a 10-point scale from never accurate to always accurate (Fig. 3). This rating is consistent with the overwhelming view of professional bodies and independent commentators that recovered memories may be accurate, false, or a mixture of the two (Lindsay & Read, 1995; Wright, Ost, & French, 2006). Practitioners endorsed the ideas that traumatic memories are often repressed and can be recovered accurately to a markedly higher degree than researchers did, but again, respondents were given no clear indication about whether the questions related to the unconscious or deliberate form of repression. Moreover, belief in memory fallibility was high among both researchers and practitioners (see Table S2.2). In contrast, clinical-psychologist practitioners were much less likely than alternative therapists to endorse the ideas that people’s memories can go back to birth and that hypnotically retrieved memories are reliable (Table S2.2).

In conclusion, we suggest that the data Patihis et al. reported do not identify an important discrepancy between the views of researchers and clinical psychologists, but rather point to a marked difference between clinical psychologists and alternative practitioners, with only the former showing clear evidence of having adapted their practice in accordance with changes in the evidence base. Given the abundant evidence that events such as death, murder, and sexual assault can sometimes be forgotten (Belli, 2012; Pyszora, Barker, & Kopelman, 2003), we would like to see future research with a broader focus on plausible cognitive explanations for forgetting of traumatic events. One such mechanism, for example, is dissociation (Brewin, 2012; Brewin & Mersaditabari, 2013).