Psychonomic Bulletin & Review 1998.5 (2). 318-323

Recalling more childhood events leads to judgments ofpoorer : Implications for the recovered/ debate

ROBERT F. BELLI and PIOTR WINKIELMAN University ofMichigan, Ann Arbor, Michigan

J.DONREAD University ofLethbridge, Lethbridge, Alberta, Canada

NORBERT SCHWARZ University ofMichigan, Ann Arbor, Michigan and STEVEN JAY LYNN State University ofNlmJ York, Binghamton, NlmJ York

Dissociative disorders that are believed to develop from childhood sexual abuse are often con­ sidered to include amnesiafor childhood events, particularly the events thatinvolve the abuse itself. One unresolved issue is the extent to which memory recovery attempts can contribute to claims of having amnestic symptoms. Experiments with undergraduate subjects reveal that requiring more re­ ports of childhood events will increasejudgments of having poorer memory ofone's childhood. The results are consistent with the use of heuristics when one is reasoning under conditions of uncer­ tainty, as experienced difficulty in remembering more experiences is attributed to the incomplete­ ness of childhood memory. The findings challenge the validity of reports of childhood that follow memory recovery attempts.

Adults claiming to have recovered hidden dream interpretation, and other therapeutic techniques, ofbeing sexually abused as children have raised consid­ to reveal hidden memories that are repressed as a result erable controversy regarding whether any particular ofsuspected abuse (Bass & Davis, 1988; Courtois, 1991; claim is genuine or is rather the result ofsuggestions in­ Edwards, 1987). Claims regarding the relationship be­ troduced through the course ofpsychotherapy orby read­ tween amnesia and childhood sexual abuse are open to ing self-help books (Belli & Loftus, 1994, 1996; Berliner criticism, however. Although amnesia for traumatic events & Williams, 1994; Lindsay & Read, 1994). One facet of such as childhood sexual abuse clearly can occur (Briere this debate has focused on the extent to which amnesia & Conte, 1993; Freyd, 1996; Schetlin & Brown, 1996), for childhood events is a valid predictor ofone's having it is questionable whether such amnesia cannot be ac­ been sexually abused during childhood. Specialists of counted for in terms ofnormal mechanisms offorgetting dissociative identity disorder (DID), formerly referred to (see Shobe & Kihlstrom, 1997). Furthermore, the deter­ a~ multiple p~rsonality disorder (MPD), often consider a mination of the presence of amnesia does not appear to history ofchildhood sexual abuse to be a major factor in be based on any firm, objective criteria (Read, 1997; Read the onset ofthe disorder and argue that amnesia for epi­ & Lindsay, 1994). sodes ofthe childhood abuse is one likely symptom (Coons, One difficulty is that there exist no clearly established 1994; Putnam, Guroff, Silberman, Barban, & Post, 1986; norms for determining the degree of autobiographical Ross et al., 1991). Some psychotherapists encourage the memory loss that is to be expected in those who suffer am­ use of memory recovery techniques, such as hypnosis, nesia from childhood sexual abuse. There is work which indicates that memory for childhood events is impaired in persons who suffer organic amnesia from physical trau­ Portions of this work were presented at the second International mas, but these amnesic individuals remember the more Conference on Memory, held at the University ofPadua, in July 1996. remote childhood events even better than more recent We thank Chris Brewin, Ira Hyman, John Gomez, and two anonymous events, a pattern that is opposite to that ofnormal controls reviewers for their helpful comments on an earlier version of this who remember the more recent events best (Kopelman, paper. Any shortcomings are entirely our own, and we assume full re­ sponsibility for them. Correspondence regarding this article should be Wilson, & Baddeley, 1989; Zola-Morgan, Cohen, & addressed to R. F. Belli, Institute for Social Research, University of Squire, 1983). Importantly, the pattern for psychogenic Michigan, Ann Arbor, M148106-l248 (e-mail: [email protected]). amnesia that is hypothesized to result from childhood

Copyright 1998 Psychonomic Society, Inc. 318 JUDGMENTS OF POORER MEMORY 319

trauma has never been fully investigated, with any dis­ ber of childhood events result in judgments of poorer tinctions regarding the extent ofautobiographical mem­ childhood memory than do attempts to fewer ory content between persons hypothesized to suffer from events. Hence, even when individuals are able to recol­ childhood trauma induced amnesia and normal controls lect many childhood events, the effort expended by re­ yet to be drawn. Further, the clinical determination of membering the past may have the non-obvious conse­ is typically not based on an attempt quence ofdecreasing one's beliefin the completeness of to systematically measure the degree to which events from one's memory. childhood can be remembered, but rather based on clin­ A number of studies have shown relationships be­ ical judgments whose criteria ofassessment are less than tween the subjective experience ofrecall and knowledge clear (Read, 1997), and for some clinicians, suchjudg­ judgments about the environment and the self(see Jacoby, ments may depend on patients' self-reports of having a Kelley, & Dywan, 1989; Schwarz & Clore, 1996; Tversky poor memory for childhood. In addition, checklists in the & Kahneman, 1973, for reviews). For example, Kelley self-help literature, which consider the inability to re­ and Lindsay (1993) found that subjects confidence in an­ member childhood events to be one ofthe symptoms as­ swers to general knowledge questions increased when sociated with childhood sexual abuse, encourage indi­ the answers, regardless ofbeing correct or incorrect, had viduals to rely on their own self-assessments of the been made easier to retrieve by being unobtrusively pre­ completeness of their childhood memory to determine sented in a previous task (see also Lindsay & Kelley, the likelihood oftheir victimization (see, e.g., Blume, 1990, 1996). In related work, Schwarz et al. (1991) found that p. xix; Fredrickson, 1992, p. 51). people form judgments of their own attributes on the Clinical scales ofdissociative disorders which include basis ofthe ease or difficulty with which relevant infor­ an indication of self-reported amnesia have also been mation comes to mind. Subjects were asked to report ei­ somewhat crude in the determination ofthe amnesia. For ther 6 or 12 examples of situations in which they be­ example, Ross's (1989, p. 325) Dissociative Disorders In­ haved assertively, and following these reports, subjects terview Schedule (DDIS) has one question dealing with had to evaluate their own assertiveness. Schwarz et al. amnesia for childhood events (answered "yes," "no," or found that subjects reported higher assertiveness after "unsure"): "Are there large parts ofyour childhood after recalling 6 rather than 12 examples, in contrast to what age 5 which you can't remember?" Interestingly, the re­ one would expect ifsubjects were basing their self- judg­ sponses to this question have been used to confirm am­ ments on the number ofevents recalled. In support ofthe nesia with MPD patients in certain studies (Anderson, conclusion that experiencing the difficulty of bringing Yasenik, & Ross, 1993; Ross eta!., 1991; Ross etal., 1990), more examples to mind reduced judgments ofassertive­ although these studies lacked a normal control group for ness, Schwarz et a!. found that subjects did rate them­ comparison. selves as more assertive in the 12-events condition than This paper questions the validity with which individ­ in the 6-events condition when given information de­ uals can assess the completeness oftheir childhood mem­ signed to discount the diagnosticity of their subjective ory, by demonstrating that the self-reported determina­ experiences. tion of amnesia can be dramatically influenced simply Similarly, the subjective experience of recalling ­ by the act of searching for childhood memories. The hood events may influencejudgments concerning the com­ search for childhood memories is one of the milder as­ pleteness ofchildhood memory. Here we report on two ex­ pects ofmemory recovery techniques that have often been periments in which we introduced an analogue to the encouraged to be used to discover repressed memories of processes of remembering childhood events that would abuse (Bass & Davis, 1988; Courtois, 1991; Edwards, occur in attempts at determining whether one is a victim of 1987). Even if such techniques are not used, there is no childhood sexual abuse. Specifically, we asked subjects to doubt that people who suspect themselves to be victims report either a few or many childhood events from a rela­ ofchildhood abuse, both within and outside the context tively narrow period ofchildhood (5-1 0 years ofage). Fol­ of therapy, will engage in some search of childhood lowing this, subjects were asked to judge the completeness events. One possible factor that can influence judgments oftheir childhood memory by answering a question taken of whether one is amnesic for childhood events is the from Ross (1989): "Are there are large parts ofyour child­ feeling of effort accompanying retrieval of childhood hood after age 5 which you can't remember?" Consistent memories. Tversky and Kahneman's (1973) availability with the ease-of-retrieval hypothesis, we predicted that heuristic holds that people estimate the frequency of subjects would rate their memory as worse after success­ events on the basis of the subjectively experienced ease fully retrieving many rather than a few childhood events, with which examples can be brought to mind. Drawing and that they would not base their memory ratings on the on this research, we propose that individuals estimate the actual number ofevents that were recalled. extent to which they can remember childhood events by drawing on their subjective experience of ease or diffi­ EXPERIMENTS 1 AND 2 culty of recall. Given that recall becomes increasingly difficult as one tries to remember additional events, this To assess the role of recall experiences in judgments conjecture predicts that attempts to recall a larger num- ofchildhood memory, we asked subjects to recall 4, 8, or 320 BELLI, WINKIELMAN, READ, SCHWARZ, AND LYNN

12 childhood events. Subsequently, they rated the com­ Table 1 pleteness oftheir childhood memory in response to the Percentage ofSubjects Endorsing Each Response Category as a question cited above (Ross, 1989). In both experiments, Function ofRecall Task in Experiments 1 and 2 we employed highly similar procedures, so the two ex­ Response Category (%) periments will be presented together. No. Events Yes Unsure No N (l00%) Experiment 1 Method 4 35 33 32 66 Subjects. One hundred fifty-two University of Michigan under­ 12 51 25 25 86 graduates (55% women, 45% men; average age 19 years) participated Experiment 2 in Experiment I. One hundred fifty-nine undergraduates (66% women, 34% men; average age 19.7 years) ofthe University ofLethbridge par­ 4 26 34 40 53 ticipated in Experiment 2. 8 33 30 37 54 Procedure, The subjects in Experiment 1 responded to different ver­ 12 44 19 37 52 sions ofa memory questionnaire in individual cubicles, which ensured Note-Subjects were r~sponding to the question "Regarding child­ privacy.! The subjects in Experiment 2 responded in a classroom set­ hood memory, are there large parts ofyour childhood after age 5 which ting, prior to a lecture. The subjects in both experiments worked at their you can't remember?" own pace. The subjects in Experiment 1 were asked to report specific events that they had experienced while they were 5-7 and 8-10 years old. though they had retrieved three times as many events. They had to recall either two or six specific events for each age period (on two or six numbered lines), resulting in a total request for 4 or 12 The difference between conditions was significant for a events. comparison ofthe "yes" and the combined "unsure" and Experiment 2 followed the same procedures but included an 8­ "no" responses [X 2(1, N = 152 ) = 4.03, P = .02]. events condition (4 events each for ages 5-7 and 8-10), in addition to These findings were replicated in Experiment 2, as is a replication ofthe 4-events and 12-events conditions described above. Following the recall task, the subjects were asked, "Regarding child­ also shown in Table 1. Again, the percentage ofrespon­ hood memory, are there large parts ofyour childhood after age 5 which dents who reported incomplete childhood memory in­ you can't remember?" ("yes," "no," "unsure"; taken from Ross, 1989). creased with the number of childhood memories re­ In addition, the subjects in Experiment 1 rated their recall experience quested, from 26% in the 4-events condition to 33% in after completion of all other questions: "Now we want you to think back to the task where you had to write down several different child­ the 8-events condition and 44% in the 12-events condi­ hood events, How difficult was this task for you?" (l = very easy; 7 = tion. The difference in the number of"yes" responses com­ very difficult), pared with the combined "unsure" and "no" responses was significant between the 12-events and 4-events con­ Results ditions [X 2(1, N = 105) = 3.65,p < .03]. As predicted, the Manipulationchecks. Overall, compliance with instruc­ responses of subjects in the 8-events condition fell be­ tions was good. In Experiment I, all subjects assigned to tween the other two and did not significantly differ from the 4-events condition listed 4 childhood events. The the responses in the 4-events condition [X 2( 1, N = 107) = subjects assigned to the 12-events condition listed an av­ 0.61, n.s.], or in the 12-events condition [X 2(1, N= 106) erage of 11.54 events, with 86% reporting 12 events, the = 1.33, n.s.].2 remaining 14% reporting an average of8.75 events, and Analyses ofcompletions. These results are somewhat no subject retrieving less than 4 events. Moreover, sub­ ambiguous, in that the effects might have been totally jects experienced it as easier to recall 4 (M = 3.66) rather driven by subjects who did not completely recall all of than 12 events (M= 4.27) [t(148) = 2.21,p < .03]. the requested events. To rule out this possibility, analy­ In Experiment 2, the subjects in the 4-events condi­ ses were conducted on the data ofonly the subjects who tion retrieved a mean of 3.81 events, with 91 % retriev­ did complete the requested number of events. Table 2 ing all 4 events. In the 8-events condition the subjects re­ provides the frequencies for both Experiments I and 2. trieved an average of 7.76 events, with 91 % retrieving As can be seen, the trends remained consistent: More all 8 events and no subject retrieving less than 5 events. subjects said "yes" in the 12-events condition (49% in In the 12-events condition, the subjects retrieved an av­ Experiment I; 33% in Experiment 2) than in the 4-events erage of 11.13 events, with 69% retrieving all 12 events condition (35% in Experiment 1; 23% in Experiment 2). and no subject retrieving less than 6 events. Although the trend was nonsignificant with the data from Judgments of childhood memory. Because our hy­ Experiment 2, the difference in the number of"yes" re­ potheses with judgments ofchildhood memory were di­ sponses as compared with the combined "unsure" and rectional, we report one-tailed probabilities in all signif­ "no" responses was significant between the 12-events icance tests. Table 1 shows the results ofExperiment 1. As and 4-events conditions in Experiment 1 [X 2( I, N = 141) = predicted, reports ofincomplete childhood memory in­ 3.01,p < .05]. When the data from both experiments were creased with the number ofevents recalled. Specifically, combined, the significant finding ofa greater number of 35% ofthe subjects who had to recall 4 events reported "yes" responses in the 12-events than in the 4-events that there were large parts oftheir childhood after age 5 conditions persisted [X 2(1, N= 225) = 4.95,p = .02]. that they could not remember. This percentage increased Correlational analysis. Experiment 1 allows us to an­ to 51 % for the subjects who had to report 12 events, al- alyze the relationship between the judgment of child- JUDGMENTS OF POORER MEMORY 321

Table 2 disorders, suicidal or self-destructive thoughts, depression. sexual dys­ Percentages for Only Those Subjects Recalling All Events in Their function, and a complex and confusing symptom picture associated Respective Conditions for Experiments 1 and 2 with the diagnosis of DID. If therapists or clients assume that such Response Category (%) problems are associated with childhood amnesia of trauma, and that memories of sexual abuse, for example, must be retrieved in psycho­ No. Events Yes Unsure No N (100%) therapy in order to effect a lasting "cure," then repeated recall attempts Panel A: Experiment I and the apparent difficulty associated with an extensive memory 4 35 33 32 66 search may reinforce the perception that amnesia is present and that 12 49 25 25 75 the suspected abuse did in fact occur. Ifsuggestive procedures are used and/or clients simply "recover memories" as a function of what they Panel B: Experiment 2 believe might have happened in a manner consistent with their ex­ 4 23 33 44 48 pectancies about the link between their symptoms and a traumatic past, 8 31 31 39 49 then such recovered memories may constitute an "experience proof" 12 33 22 44 36 that childhood amnesia was indeed present and that additional memo­ Note-Subjects were responding to the question "Regarding child­ ries must be recovered before therapy is truly complete. Because the hood memory, are there large parts ofyour childhood after age 5 which criteria for amnesia are not well specified, considerable latitude in in­ you can't remember?" terpretation regarding what constitutes "amnesia" and when memory recovery is "complete" may render attributions ofamnesia particularly vulnerable to suggestive influences and to labeling recall failures as "amnesia" by an authority figure such as a therapist. In an effort to re­ hood memory and reported ease ofretrieval. In order to cover additional memories, therapists might be tempted to use in­ conduct these correlational analyses, the responses to the creasingly specialized techniques such as hypnosis, self-help books, and guided imagery, with all the pitfalls for suggestibility associated memory question were recoded (yes = I, unsure == 0, no == with such recovery attempts (Belli & Loftus, 1994, 1996; Ceci & Lof­ -1). As was expected, there was a direct relationship be­ tus, 1994; Garry, Manning, Loftus, & Sherman, 1996; Hyman, Hus­ tween judgments of task difficulty and judgments of band, & Billings, 1995; Lindsay & Read, 1994; Lynn, Lock, Myers, & childhood memory [r(l48) = .42, P < .001]. This corre­ Payne, 1997).3 lation remained of a similar order and was significant Such a recursive cycle of suggestive influences and experiential confirmation (see Lynn & McConkey, 1998; Lynn, Stafford, Mali­ when computed separately for each experimental condi­ noski, & Pintar, 1997; Spanos, 1996), abetted by the difficulty of re­ tion [4-event r(63) = .35,p == .005; 12-event r(83) = 044, trieving childhood memories, may be associated with instances ofthe P < .001] and when the number of retrieved events was iatrogenic creation of DID as well as false memories of a traumatic partialed out [r(l45) == AI, P < .00 I]. past more generally. Survey research indicates that a sizable minority of therapists (Poole, Lindsay, Memon, & Bull, 1995) use specialized memory recovery procedures, so that the impact ofsubjective experi­ DISCUSSION ences of retrieval processes, in concert with social influence and ex­ pectation effects, may not be trivial. Experiments I and 2 highlight the malleability of people's judg­ Ofcourse, this process need not be exclusive to persons who have ments concerning the completeness of their childhood memory. Sub­ no initial memory of childhood abuse prior to the start of therapy. jects who had to recall more events from their childhood actually rated Rather, we can foresee cases in which individuals remember certain in­ their memory ofchildhood as less complete than did subjects asked to stances ofsexual abuse and, because ofengaging in extensive retrieval recall fewer memories. Although some ofthe effect was due to the per­ attempts, also come to believe they are amnesic for childhood events. formance of subjects who were unable to provide the requested num­ Such judgments may confirm the determination that amnesia is a ber of childhood events, which was more pronounced among subjects symptom of a traumatic past even among persons in whom abusive who were asked to provide a larger number ofchildhood events, the ef­ events during childhood had been remembered all along. In addition, fect persisted when we examined only the data from subjects who had such individuals would not be immune to the initiation ofthe recursive been able to provide all of the events that had been requested. In our process that we have just identified above as leading to the generation view, these data have implications regarding the determination ofthe ofadditional, but false, memories. association between child sexual abuse and being amnestic for child­ Admittedly, the simplicity of our experiments raises questions re­ hood events. garding the generalizability offindings to the clinical context. Our ef­ A common process among individuals who suspect that they have fects are based on a single question and a very simple manipulation, become victims of childhood sexual abuse is to perform an extensive and there can be no doubt that the clinical determination of psy­ search ofmemory to determine whether instances ofabuse can be re­ chogenic amnesia for childhood events is not routinely made on the covered. Our results show that extensive searches within long-term basis of such a cursory review of the past and the answers to a single memory for childhood events, and the greater degree of difficulty ex­ question. Whether more extensive clinical assessment can provide a perienced in remembering more events, will lead one to conclude that more valid indication oftrauma-induced amnesia, and whether persons there is not much there (see also Winkielman, Schwarz, & Belli, 1998). who suffer from trauma-induced amnesia can be reliably distinguished This, in turn, may lead to unwarranted conclusions that something bad, from other individuals who, for whatever reasons, are suspected by such as sexual abuse, happened during childhood and led to this paucity therapists or by themselves of having had a traumatic past, are issues ofmemories. that are yet to be resolved 4 We recognize that these search processes are not likely to occur in In conclusion, we believe that our experiments present a reasonable isolation from other social and suggestive influences that can occur in analogue to processes that may occur in persons searching for memo­ and apart from the psychotherapeutic context. For instance, self-help rial evidence ofan abusive childhood. Obviously, our findings do not books, information presented in the media, and therapists' direct and bear on whether people who experienced an abusive childhood repress subtle suggestions and interventions can all convey to individuals that related memories and are hence likely to report poor childhood mem­ there is a link between present symptoms and a repressed or forgotten ory. Our findings do, however, highlight that these reports may simply history of childhood abuse and trauma. reflect that recalling childhood events is more difficult than we think­ We can envisage situations in which clients approach therapists to leading us to infer poorer memory, the more the events that we (suc­ deal with ongoing problems that could include low self-esteem, eating cessfully) try to recall. 322 BELLI, WINKIELMAN, READ, SCHWARZ, AND LYNN

REFERENCES LYNN, S. J., & MCCONKEY, K. M. (1998). Truth in memory. New York: Guilford. ANDERSON, G., YASENIK, L., & Ross, C. A. (1993). Dissociative expe­ LYNN, S. J., STAFFORD, J., MALINOSKI, P., & PINTAR, J. (1997). Mem­ riences and disorders among women who identify themselves as sex­ ory in the hall ofmirrors: The experience of"retractors" in psycho­ ual abuse survivors. Child Abuse & Neglect, 17,677-686. therapy. Psychological Inquiry, 8, 307-312. BASS, E., & DAVIS, L. (1988). The courage to heal: A guidefor women POOLE, D. A., LINDSAY, D. S., MEMON, A., & BULL, R. (1995). Psycho­ survivors ofchild sexual abuse. New York: Harper & Row. therapy and the recovery of memories of childhood sexual abuse: BELLI, R. E, & LOFTus, E. E (1994). Recovered memories of child­ U.S. and British practitioners' beliefs, practices, and experiences. hood abuse: A source monitoring perspective. In S. 1. Lynn & 1. Rbue Journal ofConsulting & Clinical , 63, 426-437. (Eds.), Dissociation: Theory, clinical, and research perspectives PuTNAM, E W., GUROFF, J. J., SILBERMAN, E. K., BARBAN, L., & POST, (pp. 415-433). New York: Guilford. R. M. (1986). The clinical phenomenology of multiple BELLI, R. E, & LOFTus, E. E (1996). The pliability ofautobiographi­ disorder: Review of 100 recent cases. Journal ofClinical Psychia­ cal memory: Misinformation and the false memory problem. In try, 47, 285-293. D. C. Rubin (Ed.), Remembering ourpast: Studies in autobiograph­ READ, J. D. (I997). Memory issues in the diagnosis of unreported ical memory (pp. 157-I 79). New York: Cambridge University Press. trauma. In 1. D. Read & D. S. Lindsay (Eds.), Recollections oftrauma: BERLINER, L., & WILLIAMS, L. M. (1994). Memories of child sexual Scientific evidence and clinical practice (pp. 79-100). New York: abuse: A response to Lindsay and Read. Applied Cognitive Psy­ Plenum. chology, 8, 379-387. READ, J. D., & LINDSAY, D. S. (1994). Moving toward a middle ground BRIERE, J., & CONTE, J. R. (1993). Self-reported amnesia for abuse in on the "false memory debate": Reply to commentaries on Lindsay molested as children. Journal ofTraumatic , 6, 2I-31. and Read. Applied , 8, 407-435. BLUME, E. S. (1990). Secret survivors: Uncovering incest and its after­ Ross, C. A. (1989). Multiple personality disorder: Diagnosis, clinical effects in women. New York: Ballantine. features, and treatment. New York: Wiley. CECI, S. J., & LoFTus, E. E (1994). "Memory work": A royal road to Ross, C. A., MILLER, S. D., BJORNSON, L., REAGOR, P., FRASER, G., & false memories? Applied Cognitive Psychology, 8, 35 I-364. ANDERSON, G. (1991). Abuse histories in 102 cases ofmultiple per­ COONS, P. M. (1994). Confirmation of childhood abuse in child and sonality disorder. Canadian Journal ofPsychiatry, 36, 97-101, adolescent cases of multiple personality disorder and dissociative Ross, C. A., MILLER, S. D., REAGOR, P., BJORNSON, L., FRASER, G., & disorder not otherwise specified. Journal ofNervous & Mental Dis­ ANDERSON, G. (1990). Structured interview data on 102 cases of ease, 182,461-464. multiple personality disorder from four centers. American Journal of COURTOIS, C. A. (199 I, Fall). Theory, sequencing, and strategy in Psychiatry, 147,596-601. treating survivors. New Directionsfor Mental Health Services, SCHEFLIN, A. W., & BROWN, D, (1996). or disso­ No. 51, pp. 47-60. ciative amnesia: What the science says. Journal ofPsychiatry & COURTOIS, C. A. (1997). Informed clinical practice and the standard of Law,24,143-188. care: Proposed guidelines for the treatment ofadults who report de­ SCHWARZ, N., BLESS, H., STRACK, E, KLUMPP, G., RITTENAUER­ layed memories ofchildhood trauma. In 1. D. Read & D. S. Lindsay SCHATKA, H., & SIMONS, A. (1991), Ease ofretrieval as information: (Eds.), Recollections oftrauma: Scientific evidence and clinical Another look at the availability heuristic. Journal ofPersonality & practice (pp. 377-361). New York: Plenum. Social Psychology, 61,195-202. EDWARDS, D. 1. (1987). The dream as a vehicle for the recovery of SCHWARZ, N., & CLORE, G. L. (1996). Feelings and phenomenal expe­ childhood trauma. Clinical Social Work, 15,356-360. riences. In E. T. Higgins & A. Kruglanski (Eds.), Socialpsychology: FREDRICKSON, R. (1992). Repressed memories: A journey to recovery A handbook ofbasicprinciples (pp. 433- 465). New York: Guilford. from sexual abuse. New York: Simon & Schuster. SHOBE, K. K., & KIHLSTROM, 1. E (1997), Is traumatic memory spe­ FREYD, J. J. (1996). Betrayal trauma: The logic offorgetting childhood cial? Current Directions in Psychological Science, 6, 70-74. abuse. Cambridge, MA: Harvard University Press. SPANOS, N. P. (1996). Multiple identities andfalse memories: A socio­ GARRY, M., MANNING, C. G., LOFTUS, E. E, & SHERMAN, S. J. (1996). cognitive perspective. New York: Guilford. : Imagining a childhood event inflates confi­ TVERSKY, A., & KAHNEMAN, D. (1973), Availability: A heuristic for dence that it occurred. Psychonomic Bulletin & Review, 3, 208-2 I4. judging frequency and probability. Cognitive Psychology, 5, 207­ HYMAN, I. E., HUSBAND, T. H., & BILLINGS, E J. (1995). False memo­ 232. ries ofchildhood experiences. Applied Cognitive Psychology, 9, 181­ WINKIELMAN, P., SCHWARZ, N., & BELLI, R. E (1998). The role ofease 197. of retrieval and attribution in memory judgments: Judging your JACOBY, L. L., KELLEY, C. M., & DYWAN, J. (1989). Memory attribu­ memory as worse despite recalling more events. Psychological Sci­ tions. In H. L. Roediger III & F.I. M. Craik (Eds.), Varieties ofmem­ ence,9,124-126, ory and : Essays in honour ofEndel Tulving (pp. 391­ ZOLA-MoRGAN, S., COHEN, N. 1., & SQUIRE, L. R. (1983). Recall ofre­ 422). Hillsdale, NJ: Erlbaum. mote in amnesia. Neuropsychologia, 21, 487-500. KELLEY, C. M., & LINDSAY, D. S. (1993). Remembering mistaken for knowing: Ease of retrieval as a basis for confidence in answers to NOTES general knowledge questions. Journal ofMemory & Language, 32, 1-24. I. In each session, a group of participants was tested at one time, KOPELMAN, M. D., WILSON, B. A., & BADDELEY, A. D. (1989). The auto­ and subjects filled out the questionnaire in individual cubicles. Since biographical memory interview: A new assessment of autobio­ subjects would know when others had finished, we randomly assigned graphical and personal in amnesic patients. Jour­ the treatment condition to sessions so that subjects within a session re­ nal ofClinical & Experimental Neuropsychology, 11, 724-744. ceived the same treatment and would finish at approximately the same LINDSAY, D. S., & KELLEY, C. M. (1996). Creating illusions of famil­ time. Our concern was that the experience of seeing someone finish iarity in a cued recall remember/know paradigm. Journal ofMem­ very quickly could the performances of the subjects in the 12­ ory & Language, 35, 197-211. events condition. Since the sessions varied in size in unpredictable LINDSAY, D. S., & READ, J. D. (1994). Psychotherapy and memories of ways, owing to scheduled subjects' not showing up for the experiment, childhood sexual abuse: A cognitive perspective. Applied Cognitive this randomization procedure led to unequal numbers of subjects as­ Psychology, 8, 281-338. signed to the conditions. LYNN, S. J., LOCK, T., MYERS, B., & PAYNE, D. G. (1997). Recalling the 2. We also analyzed the results for both experiments by testing for unrecallable: Should hypnosis be used for memory recovery in psy­ potential effects of gender. Our primary interest was to examine chotherapy? Current Directions in Psychological Science, 6, 79-83. whether there was an interaction between gender and condition, which JUDGMENTS OF POORER MEMORY 323

could suggest that the manipulation affected only one gender. Separate 4. A related issue is whether the subjects who recalled more events analyses of variance for each experiment were conducted on recoded were generally more accurate in their judgments concerning the rela­ responses ("yes" = I, "unsure" = 0, "no" =-I). There was no evidence tive incompleteness of their childhood memories in comparison with ofan interaction between condition and gender [Experiment I F(1,148) = the more optimistic judgments of the subjects who retrieved fewer 0.06; Experiment 2 F(2, 152) = 0.53]. Experiment 2 did reveal a main events. Since in these experiments we did not employ an objective met­ effect of gender [F(I,152) = 5.32, p = .02], with men (M = 0.17) re­ ric with which to determine the completeness of childhood memory, porting worse memory than did women (M= -0.14). Gender was not this issue cannot be resolved. However, the important point is that the significant for Experiment I, however [F(I,148) = 0.29]. experiential difficulties associated with retrieving many childhood 3. Some therapists are becoming increasingly aware that memory re­ events is a normal consequence ofengaging in many retrieval attempts covery techniques can promote false memories, as is evidenced by the and is not an indication that something traumatic is responsible for recent work ofCourtois (1997) that espouses guidelines that explicitly such difficulties. recognize the unreliability ofmemory. Although we support these ef­ forts and are cautiously hopeful that memory recovery practices are changing for the better, we await the presentation offirm data that doc­ ument changes in clinical practice and the effectiveness of these (Manuscript received June II, 1997; changes. revision accepted for publication October 28, 1997.)