ARON tO

SAGE a PS ARON tO PSYCHOLOGICAL SCIENCE A Prospective Study of Memory for Child Sexual Abuse: New Findings Relevant to the Repressed-Memory Controversy Author(s): Gail S. Goodman, Simone Ghetti, .Todi A. Quas, Robin S. Edelstein, Kristen Weede Alexander, Allison D. Redlich, Ingrid M. Cordon and David P. H. Tones Source: Psychological Science, Vol. 14, No. 2 (Mar., 2003) pp. 113-118 Published by: Sage Publications, Inc. on behalf of the Association for Psychological Science Stable URL: http://www.jstor.orestable/40063780 Accessed: 22-04-2018 02:00 UTC 3'STOR is a not-for-profit service that helps scholars, researchers. and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about SSTOR, please contact supporUitjstor.org. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://about.jstor.org/terms JSTOR Sege Publications. Inc.. Association for Psychological Science are collaborating with 3'STOR to digitize, preserve and extend access to Psychological Science This content downloaded from 169.237.160.75 on Sun. 22 Apr 201802:00:39 INC All use subject to hup://about.protorsttelms 3534-014 Page 1 of 7 EFTA_000 10296 EFTA00159944 PSYCHOLOGICAL SCIENCE Research Article A PROSPECTIVE STUDY OF MEMORY FOR CHILD SEXUAL ABUSE: New Findings Relevant to the Repressed•Memory Controversy Gail S. Goodman,' Simona Ghetti,' Jodi A. Quas,2 Robin S. Edelstein,' Kristen Weede Alexander,' Allison D. Redlich,' Ingrid M. Cordon,' and David P.H. Jones' 'University of California. Davis: 'University of California. Irvine; 'Stanford University; and 'University of Oxford. Par* Hospital for Children. Oxford. England Abstract—Previous research indicates that many adults (nearly 40%) fail to report their own documented child sexual abuse (CSA) when asked about their childhood experiences. These controversial results could reflect lack of consciously accessible recollection, thus bolster- ing claims that traumatic memories may be repressed. In the present study. 175 individuals with documented CSA histories were inter- viewed regarding their childhood trauma. Unlike in previous studies. the majority of participants 691%1 in our study reported the docu- mented abuse. Older age when the abuse ended. maternal support following disclosure of the abuse, and more severe abuse were associ- ated with an increased likelihood of disclosure. Ethnicity and dissoci- ation also played a role. Failure to report CSA should not necessarily be interpreted as evidence that the abuse is inaccessible to memory. although inaccessibility or forgetting cannot be ruled out in a subset of cases. Central to the controversy over repressed or lost memory is whether childhood experiences of sexual abuse can become inaccessible to con- sciousness. Although core features of stressful events are often panicu- lwly well retained (Christianson. 1992), debate centers on whether, for highly traumatic events, special memory mechanisms supercede normal memory processes. Some researchers argue that traumatic and benign events are subject to identical cognitive operations (e.g., Loftus, Garry, & Feldman. 1994; Shobe & Kihlstrom• 1997). According to this view, traumatic experiences, like nontraumatic ones, can be forgotten over time, especially if they occurred at a young age; special memory mecha- nisms to explain forgetting are not needed. Others, however, contend that some events, such as child sexual abuse (CSA). may be so traumatic to the psychological self that they become consciously inaccessible for extended time periods because of special memory mechanisms, such as repression or dissociation (e.g., A. Freud, 1936/1946; S. Freud, 19151 1957; Williams. 1994b: but see Erdelyi. 2031). Proponents of this latter view have relied largely on retrospective studies of CSA (e.g.. Briere & Conte, 1993; Elliott & Briere, 1995). which are limited because partici- pants' reports cannot be verified. However, two prospective studies (Wi- dom & Monis, 1997; Williams, 1994a) revealed sizable percentages of individuals who did not report their previously documented CSA. In her seminal research, Williams (1994a) interviewed 129 women (66% of the original sample), mainly African Americans, who had been seen at a hospital in the early 1970s because of alleged CSA. When questioned 17 years later about a variety of personal topics, in- Address correspondence to Gail S. Goodman. Department of Psychology. University of California. One Shields Ave.. Davis• CA 95616. VOL. la. NO. 2. MARCH 2003 eluding CSA. 38% of the women failed to disclose the documented abuse. Of the nondisclosers, 68% described other unwanted sexual ex- periences, a percentage comparable to that among the women who disclosed the documented case. Findings also revealed that younger victims and those closely related to the perpetrator were less likely to disclose• whereas greater use of force during the abuse was positively related to disclosure. Williams concluded that forgetting of CSA is a relatively common occurrence. She inferred that the high nondisclo- sure rate for the documented cases was not due solely to reluctance to discuss sensitive topics (given frequent disclosure of other unwanted sexual experiences); rather, many of the women appeared to have for- gotten the former CSA, probably because of mechanisms specific to memory for traumatic events. In another prospective study, Widom and Morris (1997) inter- viewed 96 men and women with a history of substantiated CSA that occurred 20 years previously, between 1967 and 1971. Participants were prompted four times regarding CSA. Thirty-seven percent of the adults did not report any CSA experiences. However. Widom and Morris did not compare their participants' current reports with the official records to determine whether participants had disclosed the documented cases. Thus, their nondisclosure rate may be an underesti- mate because cases other than the documented cases may have been included among those disclosed. Although only two published studies have investigated memory for documented CSA after long delays. the results have far-reaching im- plications. For instance, findings could be used to support the concept of repression (Kandel & Kandel. 1994; Terr. 1994). as well as statutes of limitation that begin from the time alleged abuse is remembered. Still. questions remain regarding the methodology, interpretations, and replicability of the two studies. For instance, because Widom and Morris (1997) did not verify whether the documented case was dis- closed. their results concerning predictors are uninterpretabk; and be- cause Williams's (1994a) sample was restricted in ethnicity. and certain crucial variables (e.g.. abuse severity) were missing, the generalizability and interpretation of her findings are open to question. Given the pro- found implications, it is imperative to extend this body of research. THE PRESENT STUDY The present study included young adults who, more than a decade earlier, were involved in criminal prosecutions as victims of CSA and participated in a study of court testimony (Goodman et al., 1992). De- tailed documentation of the abuse was available for all cases. The re- search goals were to (a) compare disclosure rates of documented CSA in the current versus former prospective studies and (b) identify pre- dictors of disclosure. Copyright 0 2003 American Psychological Society 113 This content downloaded from 169.237.160.75 on Sun. 22 Apr 20111 02:0039 (TIC All use subject to hup://about.jstor.oegfierms 3534-014 Page 2 of 7 EFIA_000 10297 EFTA00159945 PSYCHOLOGICAL SCIENCE Remembering Abuse Several predictors of disclosure were examined. First we consid- ered age. Forgetting is more likely for events that occurred before rather than after the offset of childhood amnesia. at approximately age 3 to 5 years (Pillemer & White, 1989; Quas et at, 1999). Indeed. Wil- liam (1994a) reported that women who were abused before age 7 were significantly less likely to disclose than women who were abused at or after age 7. In contrast, Widom and Mont (1997) found no effect of age at time of abuse. However, because they did not compare cur- rent reports with the original cases, it is possible that participants who were abused at an early age, according to documentation, discussed abuse that occurred at a later age (i.e., after the offset of childhood am- nesia). thus deflating an age effect. A second potentially important predictor of CSA disclosure is gen- der. Previous research suggests that compared with females, males are less likely to disclose (e.g., DeVoe & Faller, 1999; Finkelhor, Hotel- lag. Lewis. & Smith. 1990; Widom & Morris. 1997). Williams and Banyard (1997) uncovered significant gender differences when they compared disclosure in Williams's ( 1994a) sample of women with dis- closure rates in a sample of men with documented CSA. However, in a regression analysis, only age emerged as a significant predictor of dis- closure. Widom and Morris found that men's and women's definitions of sexual abuse differed, which may have led to differential disclosure rates. Specifically, of participants whose responses indicated CSA, 64% of the women labeled their experiences as abusive, whereas only 16% of the men did so. Men's resistance to define certain sexual acts as abuse, perhaps because of shame, embarrassment, or fear of stigma- tization, may underlie gender differences in disclosure. Third. the relation between severity of abuse and disclosure is cur- rently under debate. Several researchers have speculated that memory deteriorates as severity of trauma increases (Terr, 1991; van der Kolk, 1997). Ten (1991). for instance, proposed that repression and dissoci- ation are more common coping reactions in response to long, antici- pated. and repeated traumas than in response to single, unanticipated, and short-lived traumas. To the extent that individuals push traumatic memories out of consciousness (i.e., repress the information) or disso- ciate during a trauma, memory would likely be adversely affected when the trauma is more severe. In fact. Williams (1994a) found that intrafamilial abuse, which tends to be severe, was less likely to be dis- closed than other abuse. However, nondisclosure of severe abuse or abuse perpetrated by a family member may be due to factors other than forgetting, such as reluctance or embarrassment. Nevertheless, because factors like event salience and personal rele- vance may facilitate memory retention (e.g., Baker-Ward. Hess. & Flanagan, 1990: Bower & Sivers. 1998; Christianson, 1992). severe abuse, which typically includes invasive sexual contact, force, or both. would be more likely to be remembered than less severe abuse. Fur- ther, given that severe abuse typically involves recurring assault (Goodman et al., 1992). it is possible that enduring memories are fash- ioned through script formation (e.g.. Hudson, Fivush. & Kuebli, 1992; Nelson, 1986). That is. children who experience repeated abuse by the same perpetrator may develop a script of what usually occurs, which would enhance long-term memory of the gist of the abuse and de- crease the likelihood of complete forgetting. A fourth potential predictor of CSA disclosure is the extent of legal involvement children experience. Within the legal system, child vic- tims are often required to talk repeatedly (e.g.. during forensic inter- views. while testifying in court) about abuse incidents. Recounting abuse multiple times may result in elaborative rehearsal, strengthening children's event representations and enhancing long-term memory 114 (e.g.. Bahrick, 2020). In both Williams's (1994a) and Widom and Morris's (1997) samples, the CSA came to authorities' attention. sug- gesting that some of the children were involved in subsequent legal cases. Yet neither study addressed the relation between disclosure and extent of children's legal involvement. Given the importance of elabo- rative rehearsal for memory, it is of interest to examine whether legal involvement, including being questioned by authorities and testifying. increases the likelihood of adult memory and disclosure of CSA. Familial charactetistics, particularly maternal support, may influ- ence long-term memory for abuse. Maternal support and communica- tion are positively related to children's emotional and psychological adjustment following CSA disclosure (Everson. Hunter. Runyan, Edelson's. & Coulter, 1989; Sas, 1993), to their memory for stressful experiences (Goodman. Quas, Batterman-Faunce, Riddlesberger. & Kuhn. 1994), and to their maintenance of allegations across interviews (Bradley & Wood. 1996; Rieser, 1991; Sorenson & Snow. 1991). To the extent that supportive mothers legitimate their children's allega- tions or talk more often with their children about the abuse, maternal support may enhance long-term memory for, or at least individuals willingness to discuss, the abuse. Ethnicity could also play a role in disclosure. CSA sequelae. in- cluding disclosure, may differ according to ethnicity (e.g., Kenny & McEachem. 2000). Moreover, ethnicity may be correlated with other factors that could affect memory for CSA (e.g., additional life trau- mas, number of CSA experiences). Finally, our longitudinal study afforded the opportunity to examine several additional potential predictors of long-term memory for abuse that arc of substantial theoretical importance. These included mental health (e.g.. current dissociative tendencies), relationship betrayal. self-blame for the abuse, and number of other traumas or abuse experi- ences. Dissociation is believed to create holes in autobiographical memory (Bride & Conte, 1993); betrayal of attachment relations is said to underlie lost memory for abuse (Freyd. 1996): self-blame should contribute to repression (S. Freud, 1915/1957): and other trau- mas or abuse experiences could interfere with memory access to the documented (target) case. The degree to which these factors contribute to loss of CSA memory provides insight into whether special memory mechanisms underlie nonreponing by adults of childhood trauma. METHOD Participants and Sample Characteristics Between 1985 and 1987, 217 children (51 male, 166 female), ages 4 to 17 years. participated in the original study.' At that time, detailed information was collected from multiple sources (i.e., prosecutor files. nonoffending caregivers. child victims) regarding characteristics of the abuse (e.g.. perpetrator's identity, sexual acts, abuse duration) and legal case (e.g.. number of times the child testified). None of the cases involved sensational allegations of day-care or satanic ritual abuse. Approximately 13 years later (original study to current interview. M 13.09 years, range: 10.66-16.58; end of abuse to current interview. I. In the questionnaires for the current study. CSA was defined based on state law, specifying a 4-year child-perpetrator age difference. One participant from the original study is not included in the present report because the pape- tralor was not 4 years older. On that basis, the court dismissed the case. This content downloaded from 169.237.160.75 on Sun. 22 Apr 2015 020039 OTC All use subject to hop://aboutjstor.orgAerms VOL. 14, NO. 2, MARCH 2003 3534-014 Page 3 of 7 EFIA_000 10298 EFTA00159946 PSYCHOLOGICAL SCIENCE G.S. Goodman et al. Al = 1186 years. range: 11.50-19.00). the former participants were located and invited to take part in a study of legal attitudes and experi- ences. Of the 217 original participants. 186 (85.7%) were located and 175 (80.6%) were interviewed. Of those located but not interviewed. 10 (4.6% of the original sample) declined to participate, and 1 was deceased. Of the 175 participants interviewed. 141 (80.6%) were female. Partici- pants' age at time of interview ranged from 16.7 to 30.3 years = 23.0). Sixty-two percent of the sample was Caucasian non-Hispanic. 11% was African American, 16% was Hispanic. and 11% indicated that they had "other" racial-ethnic backgrounds. Participants' age when the docu- mented abuse began ranged from 2 to 16 years (Al = 8.7 years), and age when the abuse ended ranged from 3 to 16 years (M = 9.2). Mean age at the start of the legal case (when involvement in the original studY began) was 9.5 years. The alleged perpetrator was classified as a parent (24%). person in mist position (Lt. teacher, relative: 40%). acquaintance (29%). or stranger (7.4%). Of the cases, 89% involved genital contact and 42% involved penetration. The results of : tests comparing participants who did and did not take part in the current study failed to reveal any significant differences in age when abuse began or ended, severity of abuse, and ex- tent of legal imolvement, ts(199 to 216) = —1.08 to 1.60. ps > .11. al- though females were more likely than males to participate in the current researdt x(1. N = 217) = 8.1, p < .01. Procedure The study was conducted in three phases: (a) a phone interview (n = 175); (b) mailed questionnaires Or - 129). including psychometri- cally sound. standardized questionnaires of psychopathology, such as the Dissociative Experiences Scale (Bernstein & Putnam. 1986) and the Posttraumatic Diagnostic Scale (Foa. Cashman. laycox. & Perry. 1997); and (c) an in-person interview (n w 107).' For scientific and ethical reasons, participants were never told that we knew of their past victimization, legal involvement, or participation in the original study. The current research was described as concerning legal attitudes and experiences. In all three phases. information about mental health, vic- timization, legal experiences, and attitudes toward the legal system was elicited. For the present report, we examined only data relevant to CSA disclosure, which consisted primarily of data collected during the initial interview. Questions about CSA were interspersed through- out the interview and concerned legal involvement as a result of al- leged CSA. as well as experiences of CSA whether or not they were reported to the authorities. CSA was defined as exhibitionism, sexual touching, kissing, rape, oral sex. intercourse, or any type of completed or attempted CSA that occurred when the participant was under age 18 and with a person at least 4 years older. Upon disclosure, additional questions (e.g.. "What was the person's name?" "What was his/her re- lationship to your) were asked to identify the documented case. Target-Case Identification Several criteria were used to determine whether participants reported the target case (the one documented in the original study). The primary 2. The number of respondents varies across the study's phases because some participants who had completed the first phase declined to participate in the second and third phases (5% and 2%, respectively) or could not be re- located (21% and 19%. respectively). VOL. 14. NO. 2. MARCH 2003 criteria were the name or identity of the alleged perpetrator and the par- ticipant's age at the time of the legal case. Secondary criteria included type of sexual activity, frequency and duration of the abuse, and out- come of the legal case. Participants' responses were classified, by two raters, into one of the following categories: (a) clear disclosure of target case; (b) no disclosure of target case but clear disclosure of another CSA experience: (c) disclosure of CSA, but unclear whether the disclosure matched the target case: or (d) no disclosure of CSA. RESULTS AND DISCUSSION Disclosure Rate During the study's first phase. 142 participants (81.1%) disclosed the target case. Seventeen participants (9.7%) denied ever being vic- tims of CSA. Seven participants (4.0%) reported only cases other than those documented in our records. Two participants (1.1%) indicated that their parent told them that they were victims of GSA, but they had no memory of the abuse. Because of their lack of memory, the latter 2 individuals were classified as nondisclosers. Additionally, 3 individu- als (1.7%) provided only sparse and ambiguous information. and 2 other individuals (1.1%) disclosed being victims of CSA but declined to answer further questions, thus preventing us from establishing whether they were referring to the target case. One participant (0.6%) declined to answer any questions about sexual abuse. Finally, the mother of a child under 18 years old (0.6%) requested that we not ask about CSA. These latter 7 participants were excluded from subsequent analyses, leaving 168 individuals for whom it was possible to estab- lish clearly whether or not the documented case was disclosed. Of these 168 participants, only 15.5% (n = 26) did not report the target case in the phone interview.' a percentage considerably lower than that reported by both Williams (1994a) and Widom and Morris (1997). There are several potential explanations for this discrepancy. First, the age range at the time of the abuse was smaller in the former studies than in ours (Williams. 10 months-12 years; Widom & Moths, 0-11 years). When our sample was restricted to individuals abused at age 12 years and younger. however, our nondisclosure rate did not change appreciably (17%, 22 out of 131). Second. because the target case was closely followed during the prosecution (Goodman et al., 1992), extensive documentation was available to identify it. Previous studies relied on single sources of information, such as medical records, which may have led to an underestimation of disclosure because of diffi- culty matching current to past records (Loftus et al., 1994). The percent- age of individuals denying ever having been sexually abused as children is comparable in Williams's study (12%, 16 out of 129) and in the cur- rent study (10%, 17 out of 168). However, the percentage of individuals who did not disclose the target case but disclosed other cases war greater in Williams's study than the current study (26% and 4%, respec- tively). Thus, at least some of Williams's cases classified as "other may have actually been the documented case. 3. When the subset Massa (67%) with corroborative evidence (e.g.. con- fession, eyewitness, physical evidence) for the target sexual abuse was consid- ered, the disclosure rate (85.5%) was similar to that reported fee the entire sample. These findings indicate that lack of disclosure was not due to the orig- inal report of abuse being false. Three individuals now claim that the abuse al- legation was false. The basic findings do not differ whether or not these individuals are included. This content downloaded from 169.237.160.75 on Sun. 22 Apr 20111 010039 LITC All use subject to hup://about.jstor.oegrierms 115 3534-014 Page 4 of 7 EFIA_000 10299 EFTA00159947 PSYCHOLOGICAL SCIENCE Remembering Abuse Third, the majority of Williams's (1994a) sample was African American. whereas the majority of our sample was Caucasian non- Hispanic. CSA sequelae, including disclosure, may differ according to ethnicity (e.g.. Kenny & McEachem. 2000). In our study, the disclo- sure rate was higher for Caucasian non-Hispanics than African Ameri- cans: 87.6% and 70.0%. respectively. x'(I, N = 166) = 4.07. p < .05. In fact, the nondisclosure rate for African Americans (30%) ap- proached that reported by Williams. Memory interference due to a greater number of life traumas or other CSA experiences did not ex- plain the finding (rs = 1.151). Interestingly, of the 26 nondisclosers. 12 reported the target case during the second (i.e., mailed questionnaire; n = 9) or third (i.e., in- person interview; n a 3) phase of the present study. Thus, some non- disclosers may have remembered the CSA but omitted that informa- tion because of an initial disinclination to report (see della Felten*. Yeager. & Lewis. 1990). It is also possible that our questioning spurred participants to ask caretakers about past abuse or served as a retrieval cue for previously inaccessible memories. Predictors of Disclosure The study's second goal was to identify predictors of disclosure. First, to examine age effects on disclosure, we divided the sample into two groups based on whether abuse ended at age 5 or earlier (n = 28) or after age 5 (n = 140). Second. we constructed a composite measure of abuse severity consisting of abuse duration, extent of sexual contact, amount of force used, and extent of injury sustained because of the abuse (Al = 4.84, range: 2-10; higher scores indicate more severe abuse). Third. perpetrator identity was coded as a parental figure (it = 42) or not a parental figure (n =126). Fourth. participants were identi- fied as having received (n = 138) or not having received (n = 22; 8 missing) maternal support following the abuse disclosure (see Goodman et al.. 1992, for details concerning the original study's measures). Fifth. ethnicity was coded as African American (n = 20) versus all others (n = 146; 2 missing). Finally, participants' legal involvement was scored as 1. child never went to courthouse (n = 58); 2, child went to courthouse but did not testify (n = 59); or 3. child went to courthouse and testified (n a 51): the mean score for this variable was 1.96. We performed a logistic regression analysis in which the dichoto- mous disclosure variable was regressed on dichotomous age at the end of the abuse. victim gender, abuse severity, perpetrator relationship, maternal support. ethnicity. and extent of legal involvement (see Table 1). Age when abuse ended, abuse severity, and maternal support sig- nificantly predicted disclosure. The results for ethnicity closely ap- 11)ble 1. Logistic regression predicting disclosure Variable B SE Wald df p Victim age 1.21 0.58 4.34 .04 Victim gender 0.12 0.60 0.04 .84 Abuse severity 0.34 0.16 4.35 .04 Maternal support 1.38 0.60 5.27 .02 Relationship with perpetrator -0.25 038 0.19 .66 Legal involvement 0.48 0.31 2.31 .13 Ethnicity —1.18 0.62 3.61 .06 Nose Nagetkerke R3 a .201.267.N = 158) 4 19.34.p <.01. 116 preached significance. Legal involvement. gender. and perpetrator relationship were not significant predictors. As in previous research (e.g., Howe. 2000). experiencing CSA in early childhood reduced the likelihood of CSA recall: 87% of the par- ticipants whose abuse ended after age 5 disclosed the documented case, compared with 69% of the participants whose abuse ended at age 5 or earlier. Contrary to the notion that abuse severity should be nega- tively associated with disclosure. victims of more severe abuse were more likely to disclose (mean severity for disclosers = 4.93) than were victims of less severe abuse (mean severity for nondisclosers = 4.37). These results appear consistent with the idea that memory for abuse, like memory for other events, is positively affected by event du- ration and salience (e.g., extent of sexual contact, level of force, and injury, which were indicators of abuse severity). Maternal support was also associated with increased disclosure rates. Eighty percent of the individuals who received maternal support disclosed, whereas 68% of those who did not receive maternal support reported the documented case. Individuals who, as children. felt be- lieved, supported. and legitimated when making their allegations may be more willing or able years later to discuss their victimization. Significant gender differences in disclosure were not found (15.2% of males and 15.6% of females failed to disclose). It is possible that, as a result of changes in societal attitudes toward sexual abuse, males who experienced CSA in the 1970s rather than 1980s. or who were in- terviewed nearly a decade before the present study (as in the former prospective research). may have been more reluctant to admit to vic- timization. Alternatively, males in the present study may be nonrepre- sentative of male CSA victims. Legal involvement did not significantly predict disclosure. al- though disclosure rates tended to be higher among individuals with more extensive legal involvement. Of the 51 participants who testified in court. 8% did not disclose. compared with 17% of the 59 partici- pants who went to the courthouse to testify but never took the stand and 21% of the 58 who never went to the courthouse in the target case Note that even for the 58 individuals with limited legal involvement, the proportion of nondisclosers was considerably smaller than that in former studies.' Finally. several additional variables were examined because of their theoretical importance. Number of other CSA experiences and other traumas might predict nondisclosure to the extent that interfer- ence erodes memory. Frequency of the target abuse should also predict lost memory, according to a psychoanalytic perspective (Tem. 1991). although a general-memory-mechanism perspective suggests that it should strengthen recall. None of these variables was significantly re- lated to disclosure. rs < I ICC ns < 168. Of clinical importance, rela- tionship betrayal (Freyd. 1996). childhood behavior problems (Child Behavior Checklist total T score; Achenbach, 1991). and current post- traumatic stress disorder symptoms (Foot et al., 1997) were not signifi- cant predictors. rs < 1.111. ns < 168. Self-blame, r(124) = .21. p < .05. was significantly correlated with disclosure, but in a direction op- posite to Freudian prediction (more self-blame, more likely to dis- close). and it was nonsignificant in logistic regression analysis. Interestingly, dissociation (Dissociative Experiences Scale) was sig- 4. The delay from abuse to interview was longer in previous studies than in ours. When delay (time from the police report to our phone interview) was also considered in the logistic regression analysis, its contribution did not reach sig- nificance (p = 14 and other relations were virtually unchanged. This content downloaded from 169.237.160.75 on Sun. 22 Apr 2018 0100:39 OTC All use subject to hup://aboutjstorsegiterms VOL. It, NO. 2, MARCH 2003 3534-014 Page 5 of 7 EFIA_000 10300 EFTA00159948 PSYCHOLOGICAL SCIENCE G.S. Goodman et al. nificantly correlated with disclosure. co = .24.P < and was a significant predictor in a logistic regression analysis: More dissocia- tive individuals were less likely to disclose. However, because the measure was included only in the mailed questionnaires, and the sam- ple was thus smaller and somewhat biased toward participants who al- ready disclosed abuse, dissociation was not included in our main regression analysis.' Conclusion Results from this study indicate that forgetting of CSA may not be a common experience, at least not in a prosecution sample. Although African Americans' nondisclosure rate was 30%, lack of willingness to disclose seems the most parsimonious explanation for the differ- ence across racial groups. Abuse severity and age at which the abuse was experienced were positively related to disclosure, a result consistent with general mem- ory principles. It is well established that salient events are generally remembered over time, provided that they occur after the offset of childhood amnesia. Because severe abuse is often a salient experience. it should be recalled. Further. that maternal support predicted disclo- sure many years later highlights the importance of social factors in re- lation to CSA disclosure. These findings do not support the existence of special memory mechanisms unique to traumatic events, but instead imply that normal cognitive operations underlie long-term memory for CSA. The role of dissociation, however, deserves further study. Our results, along with findings that abuse leads to dissociative tendencies (Briere & Runtz, 1988) and that dissociation predicts false memory (Hyman & Billings. 1998). may imply an important role for dissocia- tion in understanding lost and false memory of CSA. A fundamental challenge for future research will be to integrate cognitive, social, and clinical factors into theories of memory for childhood trauma. Acknowledgments—This snick is based on work supported by the Na- tional Science Foundation under Grant 0004369. Any opinions, findings. and conclusions or recommendations expressed in this article are those of the authors and do nor necessarily reflect the views of the National Science Foundation. We thank Deena Day. Patricia England. Jenny Garfein. Jenni- fer Noll. Linda Pon. Lydia Prado. Juliana Raskauskas. Leslie Rudy. Jenni- fer Schaaf. Daniel Stroski, Elizabeth Taub. Penelope Trickett. and many undergraduate students who provided their assistance. Thanks are also given to Sally Murphy and her interview team at the National Opinion Re- search Center. We are grateful to the waft at the Adams. Arapahoe. and Denver County District Attorney Offices. Special thanks go to District At- torneys James Peters. Robert Grant. and William Ritter and to Deputy Dis- trict Attorneys John Jordan. Jill Straus. and Norman Brisker. Gina Gallo and Lois Buckman also offered valuable assistance. REFERENCES Achenbarh.T.M. 09911. Manual foe the Child Behavior Checlasthl.18 and 1991 Prof& Burlington: University of Vermin. Devine:al of Psychiatry. 5. Nondisclosers were somewhat. but not significantly. less likely than dis- closers to complete both the second and third phases of the study. Sixty-five percent of the nondisclosers completed the second phase. compared with 75% of the disclosers: 509E of the nondisclosers completed the third phase. com- pared with 65%M the disclosers. VOL. 14. NO. 2. MARCH 2003 Bt eat. H.P. (20004 Lang-lean maintenance of knawledge. In E. Talents @ F.I.M. Croak (Eels.). The Won( handbook of memory (pa 347-362). New York: Oxford Cream say Press. Baker.Ward. L. Hess. TM . & Flanagan. D.A. 09904 The effects of involvenum on children's memory for merits Cogninse Development. S. 55-69. Benntein. E W .@ Putnam. F.W. (1986). Development reliatalay. and validity as disso- ciation scale. The Monad of Nervous and Mental Manse, 174, 727-735. Bower. G.. & Severs. II. (1998). Cognitive impact of traumatic atillS DemItmettent Ponhopornology. /0.625-633. Bradley, A R . & Wood, J.M. (1996). How do children all? The discknure process in Chili KIWI abuse. Child Abase & Makes. 9.881-891. Brien. J.. @ Conte. J.R. (1993). Self.reponed amnesia for abase in adults moksicil as children Mama( of Trainman( Sons 6. 21-31. Briere.1..& RUOIL St. 09f46.Symmomatology associated with childhood imual imam in a nondirkal adult maple. nue Abuse & Neglect. 12. 51-39. Chnstiansort. 5.-A. (1992). Emotional stress and moraines may: A critical review. Psmhological Balkan. 112. 284-309. delta Ferris's. D.. WWI. CA.. & IAwiK D.O. (1990). ChM abase: Adolescent records is. With recall. Child Abase d Nitta-41d n7-2n. Detiree. E.R.. & Faller. K.C.11999). The tharactertuies of disclosure among children who may lave been sexually abseil Child Malteratmmu.4. 217-217. EThoti D.M.. & Brine. 1. (1995). Poutniumatic stress associated with delayed recall of wheal abuse: A mars population study. Journal of Tanana& Snell. R. 629-647. Erdelyi. M.H. 110314 Defense processes can be conscious or urICOOSOOM. American Psy- chologist, 56, 761-762. Eienon. M.D.. Hunter. W.M.. Runyan. D.K.. Edelsohn. G.A.. & Coulter. M.L 119891 Maternal support following disclosure of incest. ttaweican Journal o f Or:Anal:whit). ny. 59. 197-207. Endgame. D.. Holding. G.. Lats. I.A. & Smith. C. (1990). Sexual abuse in a national survey of adult men and women: Prevalence. cluracleristics. and risk factors Child Abase d Wakes. M. 19-28. Fos. EH.. Cashman, L. Jaycos. L. @ Petty. K. 119973 The validation of a selfwepon measure of pourraunutic stress decoke The Poolroom:tic Diagnostic Scale. Psi. ehigoskoltlartanent. 9.445-451. Freud. A. (1946). The ego and the mechanisms of defense (C. Baines. Trans). New York: International Crimmins Press. (Original work published 19361 Freud. S. 119371. Repression. In J. Strachey (Ed. @ Tram.i. The standard Masao of the complete psychological worts of Simnand Freud Obl. 11. pp. 146-1581. London. Hoganh (Original work published 19151 Freyd. 1.1. (19961. Dictate! mama: The ingac of forgetting childhood alma. Cambridge. MA: Harvard University Press Goodman. G.S.. Qum. J.A. Batterman.Famice. J.M.. Riddktherger. MM.. & Kuhn. 1. 119911. Predictors of accurate and inaccurate memories of traumatic oats experi- enced in childhood. CMOtiOSAIROM and Cognition. 3. 269-294. Goodman. G.S.. Tart. EP.. Jones. D.P.H., England. P. Pon. L K.. Rudy. L.. & Prado. L. (19924 Testifying in criminal court. Monograph! of the Sawn for Aryan* in CALM DeneMptiond. 5715, Serial No. 229). Have. M.L. 12003). The foie of ear& swamies: Developmental science and the reteation of childhood americans. Washington. DC: American Psychological Association. Hudson. J.A.. Fives& R . & Kuehl'. J. 115911. Sops and episodes The development of event memory. Applied Ceram" Plahology, 6.483-505. Hyman. I.E., Jr.. & Billings El. 11998). Individual differences and the creation of false child hod memories. Memory. 6. 1-20. Kander M.. & Kande'. E.119914 Flights of memory. Dimmer. 15. 31-38. Kenny. M.C.. e McEachem. AG. 120304 Racial. ethnic. and cultural factors of childhood sexual abuse: A selected review of the literature. Clinical Pnrhokm Review. 20. 905-922. Loftus. E.F. Gany.M. & Felten:kJ 119944 Foramina sexual imam*: What does nn mean when 3844 forget? Journal of Coasalring and Cluical Psychelogv. 62. 177- 1181. Nelson. K. 11986). Event Anowledgc Sesame and function in development. Hillsdale. NJ: &leant Kline& D.B.. @ White. S.H. (1989). Childhood noes recalled by children and adults. In NW. Reese (Ed.). &thwarts in child development and Maras ;Vol D. pp. 29)- 5604 New York: Academie Press. Qum. 9A.. Goodman. G.S.. Bidrose. S.. Pipe. M..E.. Craw. S.. & Ablin. D. 119994 Erne non and inemary: Cbildren's long-tenn remembering. forgetting. and suggestibility. Joan& of Exprriawmal Child Pachologr. 72. 235-270. Rimer. M. (19914 ReCantatiOn an child sesual abase caws. Child *gam 70.611-621. Sm. L. (19931. Three years after the teaks London. Oman. Canada: London Family Cent Clinic. Shit. K.K.. & Kihlumn, J.F. 119971. Is traumatic memory special? bon AMMON in Psych:dotted Seance. d. 70-74. Sorenson. T.. & Snow. 8.11991 I. /lowchildren tell: The process of disclosure in child sex- ual abase. Child Welfare. 70. 3-15. Tem. L.C. (19911. Childhood traumas: An outline and overview. America, Journal ,e Prn (Mows. 148. 10-20. Tat L.C. 11994). Unchained inernortes. New York: gees Banks. van do Ke4k. B.A. (1997). Traumatic memories. In PS. Applebaum & L.A. Unbars Usk TIMMS and metaary: Clinical and legal aanimwnie, 1pp. 243-2603 New York: Oxford University Press. This content downloaded from 169.237.160.75 on Sun. 22 Apr 20th 02:00:39 tJTC All use subject to http://abouldstorseggerms 117 3534-014 Page 6 of 7 EFFA_00010301 EFTA00159949 PSYCHOLOGICAL SCIENCE Remembering Abuse %dm. C.S.. & Moms. S. (19971 Accuracy of adult recollectionsof childhood victuniza. non. Pan 2: Otildhood sexual abuse. Prschalogical Assestestiv. 9, 34-46. Williams. LM. (1994a). Recall of childhood trauma A prospective study of stemma's memories of &Id us! bu / !of( I ng and Clinics Psychalon 62. 1167-1176. Williams. LM (1994b) What does it mean to fury, c le kl sexual abuse? A reply so Laf. 118 tut. Garry. @ Feldman (19941. Journal cd Consulting and OAR& Psychology. 62. 1182-1186. Williams. L.M.. & Banyan,. VI. 09971 Gonda and recall of child sexual abuse: A pro. secant study. In ID. Read & D.S. Lindsay (Eds.). Recollections of inalanOt Snot- le* evidence 1pip 371-377E New Yin: Plenum. (RECEIVED 1/4/02: ACCEPTED 4/2M) This content downloaded from 169.237.160.75 on Sun. 22 Ape 2016 02:0039 UTC All use subject to http://aboutjstor.oegiterms VOL. 14. NO. 2. MARCH 2003 3534-014 Page 7 of 7 EFIA_000 10302 EFTA00159950