Pre-publication copy of the paper eventually published as Whitehouse, W. G., Dinges, D. F., Orne, E. C., & Orne, M. T. Hypnotic hypermnesia: Enhanced memory accessibility or report bias? Journal of Abnormal Psychology, 1988, 97, 289-295.


Hypnotic Hypermnesia: Enhanced Memory Accessibility or Report Bias?

Wayne G. Whitehouse, David F. Dinges, Emily Carota Orne, and Martin T. Orne

The Institute of Pennsylvania Hospital and University of Pennsylvania



Abstract: Laboratory studies of hypnotic hypermnesia have yielded inconsistent evidence of memory enhancement, and the processes responsible for the occasional positive findings have eluded identification. The present experiment assessed delayed recall for filmed material under conditions in which subjects were required to answer all questions, by guessing if necessary. They also rated confidence in the accuracy of each response. After an initial wake-baseline forced-recall test, subjects were randomly assigned to hypnosis or waking conditions for a second forced-recall test. Both groups of subjects recalled additional correct details on the second test, but the magnitude of this hypermnesia was no greater for subjects exposed to the hypnosis treatment. Hypnotized subjects did, however, exhibit a significantly greater increase in confidence for responses designated as "guesses" on the prior waking test -- a finding consistent with the view that hypnosis engenders a shift in the subjective criterion for what constitutes a "memory." Implications of these findings for the use of hypnosis in forensic situations are discussed.



Anecdotal evidence from clinical and forensic sources suggests that hypnosis can be useful for the restoration of memory. Unfortunately, despite many positive claims, the value of such evidence in establishing the reality of memory enhancement (hypermnesia) by hypnosis is weak at best. A major limitation is that the correspondence between the hypnotically aided recollections and the events originally witnessed by the individual is, as a rule, not subject to independent verification; hence, there is no way to distinguish memories from pseudomemories. Controlled laboratory studies provide the means to determine the accuracy of hypnotic memory reports and to delineate the circumstances under which hypermnesia may occur. Accordingly, research on the topic has proliferated in recent years (for a review, see Orne, Whitehouse, Dinges, & Orne, 1988). These efforts have succeeded in identifying some of the conditions that determine whether hypnosis will increase recall as well as several consequences of the use of hypnosis to enhance recall, but they have yet to resolve the fundamental reality of hypnotic hypermnesia.

There is virtual consensus among studies that hypnosis will fail to enhance recall if the to-be-remembered (TBR) material is devoid of either intrinsic, contextual, or personal meaning (e.g., Baker, Haynes, &Patrick, 1983; Barber & Calverley, 1966; Dhanens & Lundy, 1975); this is consistent with the view that stimulus meaningfulness may be a prerequisite for hypermnesia in general (e.g., Belmore, 1981; Erdelyi, Buschke, & Finkelstein, 1977). On the other hand, the use of meaningful TBR stimuli, such as pictures, stories, films, and so on, appears not to be sufficient to ensure the superiority of hypnotic over waking memory. In some instances (Dhanens & Lundy, 1975; Dywan & Bowers, 1983; Stager & Lundy, 1985), hypnosis resulted in better recall only among highly responsive subjects, whereas in other cases no differentiation of hypnotic and waking memory was observed (McConkey & Nogrady, 1984; Nogrady, McConkey, & Perry, 1985; Putnam, 1979; Sheehan, in press; Yuille & McEwan, 1985). A lack of advantage to hypnosis may be partic-



ularly evident when contrasted with the effects of motivating instructions (Cooper & London, 1973) or the use of other cognitive retrieval strategies (Geiselman, Fisher, MacKinnon, & Holland, 1985; Mingay, 1986).

It has been suggested that the evidence for recall superiority may not depend on the hypnotic augmentation of memory retrieval processes, but instead, it may result from a loosening of the subject's criterion for reporting information as memories (Dywan & Bowers, 1983; Klatzky & Erdelyi, 1985; Orne, 1979; Smith, 1983). This could be due, in part, to the subject's prior beliefs about the efficacy of hypnosis in facilitating recall (i.e., an implicit prehypnotic suggestion), to a decrease in critical judgment, to altered perceptions of the consequences of recall, or to an increased motivation to cooperate. Regardless of the underlying impetus, lowering one's report criterion will often achieve increased recall productivity, consisting of both correct and incorrect novel information being reported.

Dywan and Bowers (1983) reported findings that are consistent with the report-criterion shift hypothesis. Following repeated recall trials over a 1-week interval, high- and low-hypnotizable subjects were exposed to a hypnotic induction or to task-motivating instructions and received suggestions for enhanced recall. Subsequently, hypnotizable subjects in the hypnosis treatment recalled significantly more new correct items than any other group of subjects, but they also significantly increased the number of new incorrect items reported. Thus, subjects' overall productivity was enhanced as a consequence of the hypnosis treatment, possibly because they became less critical of information about which they had been uncertain previously.

Although the concurrent increase in novel correct and incorrect recall obtained by Dywan and Bowers (1983) is a pattern that most clearly exemplifies the effects of lowering one's report criterion, the absence of this pattern does not imply that a criterion shift did not take place. So long as subjects are not required to disclose all the information accessible to them in a test of memory, even a



selective increase in novel correct items on a subsequent test might represent a lowering of the report threshold rather than improved access to stored memories. In this regard, a recently reported investigation of delayed recall of filmed material by Stager and Lundy (1985) observed increased correct recall for highly hypnotizable subjects exposed to a hypnotic procedure, with no commensurate increase in incorrect recall. In their study, a series of tape-recorded questions, providing many correct details to serve as retrieval cues, was asked prior to and following a hypnotic induction or a control task. More importantly, however, subjects were always free to answer or not to answer each question (permitting report criterion to vary across tests). Under these circumstances, one cannot determine how much of the newly emerged correct information reported during hypnosis actually may have been accessible previously. Accordingly, it is possible that hypnosis occasioned a shift in report criterion, which together with the availability of accurate retrieval cues embedded in the questions, created a bias toward liberating correct responses that were previously withheld because of insufficient confidence.

The pervasiveness of report-criterion changes in the current literature on hypnotic hypermnesia is unknown. Moreover, it has been argued by Klatzky and Erdelyi (1985) that no laboratory study of hypnotic recall has yet been designed to rule out the contribution of criterion shifts. According to their analysis, hypnotic hypermnesia can be assumed to have been validly demonstrated only if the increase in correct recall survives stringent control procedures that prevent variation in report criterion between hypnotic and nonhypnotic recall tests. One such procedure, developed by Erdelyi and his associates (Erdelyi, 1970; Erdelyi & Becker, 1974; Erdelyi & Kleinbard, 1978), requires subjects to provide a fixed number of responses (exceeding their free-recall level) on each recall attempt, guessing when necessary. The purpose of the technique is to prevent the occurrence of criterion shifts across repeated recall trials by setting all subjects' output at the same level.



In the present study we adapted this "forced-recall" procedure for use with the interrogatory recall format used by Stager and Lundy (1985), in an attempt to replicate their finding of hypnotic enhancement of memory under conditions in which report criterion was controlled.1 The stimulus film, recall questions, and other procedures of Stager and Lundy's study were retained in this experiment, but subjects were instructed to provide an answer to every question, even if they had to guess, and to rate their confidence in the accuracy of each response. The strategy was to assess subjects' recall prior to hypnosis under circumstances that require the reporting of information about which they may be uncertain. With such subthreshold recall already disclosed prior to hypnosis, any additionally recovered information obtained during hypnosis would imply a true hypermnesic process.



The subjects were 78 undergraduate student volunteers (33 male, 45 female) who were selected from a larger sample (n = 428) on the basis of their extreme scores (low = 0-5, high = 9-12) on the Harvard Group Scale of Hypnotic Susceptibility, Form A (HGSHS:A of Shor & Orne, 1962). Classification of subjects' hypnotic ability was supported by a subsequent individually administered Stanford Hypnotic Susceptibility Scale, Form C (SHSS:C of Weitzenhoffer & Hilgard, 1962), yielding 38 subjects who scored in the high range (8-12) and 40 who scored in the low range (0-4) of the latter scale. Subjects were paid nominally for their participation in both the initial group screening session and the individual session conducted approximately 1 week later.




Subjects were recruited through campus newspaper advertisements and letters to students who had completed volunteering cards for psychology experiments. Interested volunteers telephoned the laboratory for an appointment to take part in a group assessment of hypnotic responsivity. Sixteen group sessions were conducted, ranging in size from 9 to 42 subjects (M = 27).

Upon arrival at the laboratory for the initial group hypnotizability assessment, subjects completed an informed consent form and a number of questionnaires. Following this, the director of the laboratory established rapport by making introductory comments about hypnosis and providing an opportunity to ask questions. The tape-recorded HGSHS:A was then administered, after which subjects completed booklets assessing their behavioral and subjective responses to the hypnotic suggestions.

A brief intermission ensued, during which the HGSHS:A response booklets were collected by an independent investigator and scored unobtrusively. Subjects were then shown the stimulus film: a 15-min color presentation on poster-making techniques entitled "Posters" (distributed by AIMS Media, Van Nuys, CA). They were informed that they were about to see an interesting movie and that they should pay close attention to it. Immediately after the film, subjects completed a 20-item, multiple-choice questionnaire concerned with various topical and incidental details presented in the film, followed by several additional, but unrelated, questionnaires. A sealed take-home packet that included a stamped, self-addressed return envelope as well as a second copy of the 20-item, multiple-choice questionnaire was given to each volunteer with instructions that it should be opened and the questionnaire completed 4 days later and mailed back to the laboratory. All subjects were thanked and paid for their help in the



study, and appointments were made for those whose scores met the HGSHS:A criteria for further participation in a second session.

Subjects were tested individually in the second session, which took place between 4 and 12 days after the initial group session. Each subject was escorted to a quiet room, seated in a comfortable chair, and outfitted with a chest microphone. The experimenter informed the subject that he or she would be asked a series of 40 tape-recorded questions (different from the 20 questions contained in the earlier multiple-choice recognition tests) about the film. A 15-s period of silence followed each question to allow the subject time to respond. It was emphasized that the subject was expected to answer each question, by guessing if necessary, and that it was not permissible to say, "I don't know." In those rare instances in which a subject did not respond in 15 s, the tape was paused until a response was obtained. After each response, the subject rated his or her confidence in its accuracy using a 4-point scale (0 = not confident, just guessing; 1 = slightly confident, much doubt; 2 = moderately confident, but not certain; 3 = very confident, certain). All subjects underwent this initial baseline recall test (R 1) in the wake condition, but with eyes closed.

The second recall test (R2) took place 15 min after R1 under one of two treatment conditions (hypnosis or wake). Forty-one subjects (20 low and 21 high in hypnotizability) were randomly assigned to the hypnosis treatment, in which they were asked to remain seated comfortably with their eyes closed, while they listened to a 15-min tape-recorded hypnotic induction procedure (see Stager, 1974). Near the end of the hypnotic induction, a suggestion was given that the answers to the 40 questions would come more easily than before. This was reinforced by an additional suggestion administered by the experimenter, "You will remember the film very clearly now." The subject then responded to each of the 40 questions that had been asked during R1 and provided confidence ratings for every response as was done previously. At the completion of R2, the experimenter gave the suggestion that, when awakened, "You will remember all that has taken place during hypnosis, as well as everything that happened before you were hypnotized." The subject was then administered the tape-recorded instructions for arousal from hypnosis.



Those 37 subjects designated to remain in the wake condition during R2 (20 low and 17 high in hypnotizability) performed a 15-min tape-recorded auditory signal-detection task prior to being asked the 40 questions for the second time. The signal-detection task concluded with the same suggestion for facilitated recall that was given to subjects in the hypnosis condition.

Following the R2 recall test, all subjects were given a booklet containing the same 40 questions in a multiple-choice recognition/recall format. Below each question were five choices, the last of which was a blank line that subjects could select and fill in if they believed that the correct answer was not represented among the preceding four alternative choices. In fact, for 8 of the 40 questions the correct answer was not provided and thus it could only be given if subjects wrote it in on the blank line. Each question also included a space to record subjects' confidence in their choices, based on the same 4-point scale used for the previous recall tests.

When this final recognition/recall booklet was completed, subjects were introduced to another experimenter who administered the standardized SHSS:C assessment of hypnotic ability. This was followed by a postexperimental interview and debriefing, which was conducted by the second experimenter.


Pretreatment Measures

Immediate and delayed recognition. The purpose of the 20 item multiple-choice recognition tests administered immediately after subjects viewed the film and again 4 days later was to assess comparability among the treatment groups in terms of initial perceptual processing and long-



term retention of information contained in the film. Analysis of variance (ANOVA) conducted on recognition scores obtained on the immediate test revealed no differences among groups (F < 1.0). Similarly, no reliable intergroup differences could be detected on the delayed recognition test (F < 1.0). However, as might be expected, recognition accuracy declined over the 4-day retention interval for all groups, F(1, 74) = 5.75, p < .025.

Wake baseline forced recall. Analysis of the number of correct responses given on the 40-question wake forced-recall test (R 1) also failed to reveal any differences among groups (F < 1.0). Because subjects were required to answer each question, the equivalency of correct recall among the groups implies that they were also comparable in terms of erroneous responses. At the same time, although the mean confidence ratings assigned by subjects to correct responses were consistently and substantially greater than those assigned to incorrect recall, F(1, 74) = 1,043, p < .001, there were no significant differences in confidence among groups (F < 1.0).

The consistent lack of intergroup differences across all of the pretreatment measures (recognition as well as baseline recall) thus reduces the likelihood that differential effects that might occur during the treatment phase could be due to preexisting differences in retention level, confidence, or motivation.

Treatment Measures

Hypnotic versus wake hypermnesia. The mean number of correct responses given by subjects in each of the four groups during R 1 (baseline recall) and R2 (treatment recall) are presented in Table 1. These data were analyzed by a 2 (hypnotizability) by 2 (wake vs. hypnosis



treatment) by 2 (recall test) mixed-design ANOVA. The only significant effect to emerge from this analysis was an increase in correct recall from R1 to R2, F(1, 74) = 42.3, p < .001, which, because the forced-recall requirement was in effect for both tests, could not have resulted from a lowering of report criterion during R2. Thus, the increase in correct recall represents a true hypermnesic effect associated with the repeated effort to recall. It is worth noting, however, that this hypermnesia was unaffected by subjects' hypnotic ability and was no greater for subjects in the hypnosis treatment than for those in the waking condition (interaction Fs < 1.0). Similarly, a separate analysis determined that the amount of new correct information reported during R2 (i.e., not provided during Rl) also did not differ as a function of either hypnotizability, recall treatment, or their interaction (Fs < 1.0).

Insert Table 1 about here

Because the recall test phase took place between 4 and 12 days after subjects had viewed the film, it is possible that the mean lengths of the retention interval may not have been equivalent across groups, thereby placing subjects in one or more of the treatment conditions at a disadvantage. However, a one-way ANOVA conducted on the number of intervening days between the acquisition and recall sessions failed to reveal a significant difference among the groups, F(3, 74) = 1.81, p > .10. In addition, no meaningful correlation was found between the length of the retention interval and the difference in correct recall between R1 and R2, which served as an index of hypermnesia (r = .05).

Hypnotic versus wake confidence. Table 2 presents the mean confidence ratings (range = 0-3) assigned to correct and to incorrect responses during R1 and R2 as a function of hypnotizability and treatment condition. Confidence ratings for correct responses increased across tests in all four



groups, F(1, 74) = 30.5, p < .001. Although subjects exposed to the hypnotic treatment showed a somewhat greater confidence increase than did wake subjects, this finding was not statistically reliable, F(1, 74)=2.6,p >.10.

Insert Table 2 about here

Confidence ratings assigned to incorrect responses, which are shown in the lower portion of Table 2, also increased significantly from R1I to R2, F(1, 74) = 109.6, p < .001. However, the inflation of confidence in incorrect recall was greater for high-hypnotizable subjects than for their low-hypnotizable counterparts, F(1, 74) = 8.4, p < .01, and also was significantly more characteristic of subjects exposed to the hypnotic treatment, F(1, 74) = 11.1, p < .005. In addition, the three-way interaction was also reliable, F(1, 74) = 4.4, p < .05, indicating that the greatest increase in confidence for incorrect responses occurred among high hypnotizables in the hypnosis condition.

Posttreatment recognition/recall. Performance on the hybrid multiple-choice recognition/recall task, which was assessed exclusively in the waking condition, was unaffected by subjects' hypnotic ability or by their prior recall treatment (Fs < 1.2); the mean percentage of correct responses ranged from 57% to 61 % among the four groups.

Analyses conducted on the recall option of this task found no intergroup differences in the number of correct responses supplied by subjects to the 8 questions for which the correct answer was otherwise unavailable (Fs < 1.0). For these very same questions, however, high-hypnotizable subjects, more often than low-hypnotizable subjects, answered incorrectly either by selecting one of the given distractor items or by "recalling" an erroneous response, F(1, 74) = 6.5, p <.025.

Confidence as an Index of Memory Report Criterion

The preceding analyses of recall during R1 and R2 were based on the number of correct responses irrespective of confidence. These analyses indicated no advantage in correct recall for subjects exposed to the hypnotic treatment relative to subjects in the waking condition. Nevertheless, the weighting of recall



by confidence ratings is of particular value to an assessment of the report-criterion shift hypothesis of hypnotic hypermnesia. This is because the use of the forced-recall technique identifies candidate responses that ordinarily would not be offered as memories (i.e., those given zero confidence ratings at R1), and allows a determination of any differential change in their report tendency (as indexed by confidence) during hypnosis at R2. Thus, by restricting analyses to those responses designated as guesses during R1, a criterion shift would be indicated by a tendency for the same responses to be given with greater-than-zero confidence ratings during R2. If hypnosis facilitates this process, increases in confidence for prior guesses should be greater among subjects exposed to hypnosis than for those tested in the waking condition.

Accordingly, mean confidence ratings assigned during R2 to those responses offered as guesses (i.e., zero confidence) during R1 were computed for each subject. Summary measures of these data are presented in Table 3 for total R1 guesses, correct guesses, and incorrect guesses. (The sample sizes on which analyses were based represent only those members of the parent samples whose guesses during R1 were repeated during R2.) An ANOVA conducted on the total number of R1 guesses repeated during R2 revealed significant main effects of hypnotizability, F(1, 73) = 5.6, p < .025, and of treatment, F(1, 73) = 7.5, p < .01. As seen in Table 3, high-hypnotizable subjects were more confident on R2 of their previous guesses than were low-hypnotizable subjects. Similarly, as predicted by the report-criterion shift account of hypnotic hypermnesia, subjects in the hypnotic treatment expressed significantly greater confidence in their earlier guesses than did subjects in the waking condition. The interaction of these factors was not significant (F < 1.0).

Insert Table 3 about here

A breakdown of total R1 guesses was undertaken to ascertain whether the increased confidence invested in these guesses during R2 was related in any way to their accuracy. Analysis of those guesses that were in fact correct revealed only a reliable effect of hypnotizability, F(1, 34) = 7.4, p < .01, which was due to the greater tendency of high relative to low hypnotizables to become more confident of their



correct guesses. In contrast, the increase in confidence awarded to earlier incorrect guesses was comparable in high- and low-hypnotizable subjects, but was found to be exaggerated by exposure to the hypnotic treatment, F(l, 71) = 7.9, p < .01. Thus, although hypnosis augmented confidence in responses that subjects were previously uncertain about, it appears in this instance to have done so disproportionately for responses that were in error.

As a supplemental analysis to estimate the degree to which hypnosis altered the conviction that subjects placed in their previous guesses, the percentage of the total pool of such responses that were promoted from zero (guesses) during R1 to confidence ratings of 1 (slightly confident), 2 (moderately confident), and 3 (very confident) on R2 was calculated. Subjects in the hypnosis condition reclassified 34% of their previous guesses as 1 when repeated during R2, whereas wake subjects reclassified 21% of their prior guesses as 1 (z = 2.73, p < .0 1). Promotion to a confidence rating of 2 occurred for 18% of the guesses of hypnosis subjects and for 5% of the guesses of wake subjects (z = 3.82, p < .00 1). Finally, increases to a confidence rating of 3 were extremely rare, occurring for less than 2% of the prior guesses of subjects from either group.


Under experimental conditions shown previously to promote selective increases in correct recall among hypnotizable subjects exposed to a hypnotic treatment (Stager & Lundy, 1985), we found no recall advantage during hypnosis when subjects were required to answer each question. This forced-recall requirement was imposed as a means to stabilize the report criterion between wake and hypnosis recall tests. Note that the use of the forced-recall procedure in this experiment did not interfere with the



expression of hypermnesia. Indeed, all groups displayed a significant net recall gain from baseline to treatment tests; however, the magnitude of this increase was not any greater for subjects in the hypnosis condition. The finding that hypnotizability did not contribute to the effectiveness of the suggestion for hypnotic hypermnesia has also been reported by others (e.g., Nogrady et al., 1985; Register & Kihlstrom, 1987), and would appear to indicate that intertest recall increments -- not attributable to criterion changes -- represent normal, nonhypnotic hypermnesia spawned by extended retrieval effort (see Payne, 1987, for a review). By implication, therefore, the present data suggest that evidence for superior recall in hypnosis, obtained when report criterion is not controlled (e.g., DePiano & Salzberg, 1981; Dywan & Bowers, 1983; Stager & Lundy, 1985), may very well be due to an increased tendency for subjects exposed to hypnotic procedures to include previously uncertain information in their hypnotic memory reports.

This conclusion is further supported by the results of analyses that tracked the fate of zero-confidence responses across the two recall tests. Under the typical demands for accuracy that exist in most memory assessment contexts, it is reasonable to assume that subjects refrain from reporting information about which they are highly uncertain. Thus, in the absence of the forced-recall requirement to "answer each question, by guessing if necessary," our subjects very likely would not have volunteered information in which they had zero confidence. Yet, it is precisely because the forced-recall procedure disclosed this uncertain information during the baseline test that it was possible to examine the effects of hypnosis on the subjects' later willingness to report these responses as memories.

As the uniformly greater-than-zero mean values in Table 3 attest, all groups of subjects invested somewhat higher confidence in their guesses when repeated for a second time. However, both hypnotic ability and exposure to hypnosis independently increased confidence in uncertain responses.



It is not clear why hypnotizability contributed to a greater conviction over time in the accuracy of prior guesses. Perhaps having once postulated such responses serves to enhance their plausibility, thereby increasing their subsequent acceptance by hypnotically responsive subjects. In any event, the hypnotizability effect was confined to that class of prior guesses that were in fact correct.

In contrast, subjects exposed to a hypnotic induction expressed greater certainty than subjects in the waking condition concerning previous guesses that were in error, and this outcome was independent of hypnotic ability. Consequently, if level of confidence parallels the likelihood of reporting information as a product of one's memory in a free-recall situation, this finding suggests that the overall accuracy of such a free-recall memory report would have been degraded by exposure to a hypnotic induction procedure.

We assume that the increased confidence subjects place in responses previously defined as guesses is a valid index of their willingness to include these responses in their memory reports. More global changes in confidence -- not restricted to responses labeled as guesses during baseline recall -- were also evident, and these changes mirrored those discussed earlier. General confidence increases associated with hypnosis and/or hypnotizability have been reported in several studies (Dywan, 1983; Nogrady et al., 1985; Sheehan, Grigg, & McCann, 1984; Sheehan & Tilden, 1983, 1986), and they are often found to be unrelated to the accuracy of recall. It is possible, therefore, that what we describe as a lowering of the report criterion by hypnosis may in fact be accomplished by a general confidence escalation that causes an increase in the number of responses that exceed the report threshold. Nevertheless, the critical point is that because hypnosis yields information that otherwise would be withheld from the memory report, the process is one of facilitation of report tendency rather than an enhanced accessibility to stored memories.

As Klatzky and Erdelyi (1985) have argued, there is a need to reevaluate various sources of evidence concerning the effects of hypnosis on memory, using methodology capable of differentiating bona fide



memory effects from report biases. The present experiment represents such an effort, which was undertaken to clarify the processes responsible for the apparent hypnotic hypermnesia observed by Stager and Lundy (1985); our findings did not support an interpretation based on enhanced retrieval. Further research is required along these lines not only to determine the basis for other examples of the apparent enhancement of memory by hypnosis, but also to identify the source of vulnerability to memory distortion effects (e.g., Putnam, 1979; Sheehan, in press).

Finally, the principal focus of the current study concerned the ability of hypnosis to engender an increase in recall. Although we also assessed memory with a recognition test following the termination of hypnosis, no differences in recognition performance were found among groups. This is consistent with the findings of many other studies that suggest that recognition memory may not be amenable to enhancement by hypnosis (see Orne et al., 1988) or by waking hypermnesic procedures (Erdelyi, 1984; Erdelyi & Stein, 1981; Payne & Roediger, 1987). However, a recent study by Shields and Knox (1986, Experiment 2) found evidence of a modest hypnotic enhancement of recognition for words presented in an orienting task designed to foster deep perceptual processing, but not for words involved in a more superficial processing task. Because recognition judgments were required for all test items, and because the advantage for hypnosis subjects in the number of correctly recognized words was not accompanied by an increase in the number of falsely recognized words, it appears that this finding of Shields and Knox may provide an instance of a true increment in accessible memory produced by hypnosis. Nevertheless, as these authors noted, the lack of a prehypnotic memory test in their study raises questions about the mechanisms responsible for their results. In view of the potential theoretical significance of this issue, replication and further experimental work along these lines are needed.

Implications for the Forensic Use of Hypnosis

The findings of the present study clearly indicate that hypnosis does not facilitate the retrieval from



memory of information concerning meaningful material presented several days earlier. It does appear to exaggerate the "creation" of confidence for already accessible but uncertain information, thereby increasing the probability that it will be reported subsequently with some degree of conviction.

In light of these findings, one might suppose that the use of hypnosis with a witness or victim of a crime who has incomplete recollection of the events in question may be of value despite the lack of effect on memory retrieval per se. Thus, it is conceivable that the confidence-bolstering effect of hypnosis could be advantageously applied to cause an individual to share uncertain recollections, which in turn may provide useful leads in an investigation. Unfortunately, our data indicate that hypnosis differentially increased confidence only for erroneous uncertain responses, and not for uncertain information that happened to be correct. Nonetheless, given both unlimited resources and recognition by the investigating officers that the hypnotized individual's conviction is not a reliable guide to the accuracy of recall, hypnosis or any other procedure that liberalizes report criterion may yield additional leads.

In the more usual situation, however, there is a considerable risk that the inherent unreliability of information confidently provided by a hypnotized witness may actually be detrimental to the truth-seeking process. Taken further, if hypnosis is permitted to form the basis of testimony in court, the confident "recollections" of a previously hypnotized witness could very well create a serious miscarriage of justice.

The present findings challenge the rationale for using hypnosis in the forensic context. Hypnosis produced no greater increase in correct recall than did simply attempting to recall a second time in the waking condition. Unlike hypnosis, however, the latter procedure did not so markedly escalate confidence in erroneous responses. Thus, even if further research were ultimately to confirm the efficacy of hypnosis under limited circumstances, such as, for example, in overcoming memory



blocks associated with traumatic or violent emotional experiences, its value would need to be weighed carefully against the increased potential for distortions of self-perceived memory accuracy.

Finally, no effort was made in this study to create circumstances analogous to those that confound many forensic situations where, for example, the hypnotist may have preconceptions about the events to be remembered and unwittingly lead the witness's recall. It is striking, therefore, that in the "sterile" laboratory setting it is still possible to demonstrate one of the mechanisms responsible for some of the problems associated with the use of hypnosis to aid recall.




Baker, R. A., Haynes, B., & Patrick, B. S. (1983). Hypnosis, memory, and incidental memory. American Journal of Clinical Hypnosis, 25, 253-262.

Barber, T X., & Calverley, D. S. (1966). Effects on recall of hypnotic induction, motivational suggestions, and suggested regression: A methodological and experimental analysis. Journal of Abnormal Psychology, 71, 169-180.

Belmore, S. M. (1981). Imagery and semantic elaboration in hypermnesia for words. Journal of Experimental Psychology: Human Learning and Memory, 7, 191-203.

Cooper, L. M., & London, P (1973). Reactivation of memory by hypnosis and suggestion. International Journal of Clinical and Experimental Hypnosis, 21, 312-323.

DePiano, F.A., & Salzberg, H.C. (1981). Hypnosis as an aid to recall of meaningful information presented under three types of arousal. International Journal of Clinical and Experimental Hypnosis, 29, 383-400.

Dhanens, T P., & Lundy, R. M. (1975). Hypnotic and waking suggestions and recall. International Journal of Clinical and Experimental Hypnosis, 23, 68-79.

Dywan, J. (1983). Hypermnesia, hypnosis and memory: Implications for forensic investigation. Unpublished doctoral dissertation, University of Waterloo, Ontario, Canada.

Dywan, J., & Bowers, K. S. (1983). The use of hypnosis to enhance recall. Science, 222, 184-185.

Erdelyi, M. H. (1970). Recovery of unavailable perceptual input. Cognitive Psychology, 1, 99-113.



Erdelyi, M. H. (1984). The recovery of unconscious (inaccessible) memories: Laboratory studies of hypermnesia. In G. H. Bower (Ed.), The psychology of learning and motivation: Advances in research and theory (Vol. 18, pp. 95-127). New York: Academic Press.

Erdelyi, M. H., & Becker, J. (1974). Hypermnesia for pictures: Incremental memory for pictures but not words in multiple recall trials. Cognitive Psychology, 6, 159-171.

Erdelyi, M. H., Buschke, H., & Finkelstein, S. (1977). Hypermnesia for Socratic stimuli: The growth of recall for an internally generated memory list abstracted from a series of riddles. Memory & Cognition, 5, 283-286.

Erdelyi, M. H., & Kleinbard, J. (1978). Has Ebbinghaus decayed with time?: The growth of recall (hypermnesia) over days. Journal of Experimental Psychology: Human Learning and Memory, 4, 275-289.

Erdelyi, M. H., & Stein, J. B. (1981). Recognition hypermnesia: The growth of recognition memory (d') over time with repeated testing. Cognition, 9, 23-33.

Geiselman, R. E., Fisher, R. P., MacKinnon, D. P., & Holland, H. L. (1985). Eyewitness memory enhancement in the police interview: Cognitive retrieval mnemonics versus hypnosis. Journal of Applied Psychology, 70, 401-412.

Klatzky, R. L., & Erdelyi, M. H. (1985). The response criterion problem in tests of hypnosis and memory. International Journal of Clinical and Experimental Hypnosis, 33, 246-257.

McConkey, K. M., & Nogrady, H. (1984). Hypnosis, hypnotizability and story recall. Australian Journal of Clinical and Experimental Hypnosis, 12, 93-98.



Mingay, D. J. (1986). Hypnosis and memory for incidentally learned scenes. British Journal of Experimental and Clinical Hypnosis, 3, 173-183.

Nogrady, H., McConkey, K. M., & Perry, C. (1985). Enhancing visual memory: Trying hypnosis, trying imagination, and trying again. Journal of Abnormal Psychology, 94, 195-204.

Orne, M. T (1979). The use and misuse of hypnosis in court. International Journal of Clinical and Experimental Hypnosis, 27, 311-341.

Orne, M. T, Whitehouse, W G., Dinges, D. F., & Orne, E. C. (1988). Reconstructing memory through hypnosis: Forensic and clinical implications. In H. M. Pettinati (Ed.), Hypnosis and memory (pp. 21-63). New York: Guilford Press.

Payne, D. G. (1987). Hypermnesia and reminiscence in recall: A historical and empirical review. Psychological Bulletin, 101, 5-28.

Payne, D. G., & Roediger, H. L. III. (1987). Hypermnesia occurs in recall but not in recognition. American Journal of Psychology, 100, 145-165.

Putnam, W. H. (1979). Hypnosis and distortions in eyewitness memory. International Journal of Clinical and Experimental Hypnosis, 27, 437-448.

Register, P. A., & Kihlstrom, J. F. (1987). Hypnotic effects on hypermnesia. International Journal of Clinical and Experimental Hypnosis, 35,155-170.

Sheehan, P. W. (in press). Memory distortion in hypnosis. International Journal of Clinical and Experimental Hypnosis.

Sheehan, P. W., Grigg, L., & McCann, T. (1984). Memory distortion following exposure to false information in hypnosis. Journal of Abnormal Psychology, 93,259-265.

Sheehan, P. W., & Tilden, J. (1983). Effects of suggestibility and hypnosis on accurate and distorted retrieval from memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 9, 283-293.

Sheehan, P. W., & Tilden, J. (1986). The consistency of occurrences of memory distortion following hypnotic induction. International Journal of Clinical and Experimental Hypnosis, 34, 122-137.



Shields, I. W., & Knox, V. J. (1986). Level of processing as a determinant of hypnotic hypermnesia. Journal of Abnormal Psychology, 95, 358-364.

Shor, R. E., & Orne, E. C. (1962). Harvard Group Scale of Hypnotic Susceptibility, Form A. Palo Alto, CA: Consulting Psychologists Press.

Smith, M. C. (1983). Hypnotic memory enhancement of witnesses: Does it work? Psychological Bulletin, 94, 387-407.

Stager, G. L. (1974). The effect of hypnosis on the learning and recall of visually presented material. Dissertation Abstracts International, 35, 3075B. (University Microfilms No. 74-28,985)

Stager, G. L., & Lundy, R. M. (1985). Hypnosis and the learning and recall of visually presented material. International Journal of Clinical and Experimental Hypnosis, 33, 27-39.

Weitzenhoffer, A. M., & Hilgard, E. R. (1962). Stanford Hypnotic Susceptibility Scale, Form C. Palo Alto, CA: Consulting Psychologists Press.

Yuille, J. C., & McEwan, H. (1985). Use of hypnosis as an aid to eyewitness memory. Journal of Applied Psychology, 70, 389-400.



Author Notes

This research was supported in part by Grant 82-IJ-CX-0007 from the National Institute of Justice, U.S. Department of Justice, in part by Grant MH-19156 from the National Institute of Mental Health, U.S. Public Health Service, and in part by a grant from the Institute for Experimental Psychiatry Research Foundation.

We are indebted to Matthew H. Erdelyi for illuminating discussions of the conceptual issues addressed in this article and for suggesting a paradigm that we were able to adapt for use in this work. We are also grateful to Campbell W. Perry for his helpful criticisms of the paper and to Mary F Auxier, Barbara R. Barras, Lindsay S. Carota, Noel F Carota, Jennifer L. Donohue, Stephen R. Fairbrother, Jean A. Friday, Sheila R. Holtz, John W. Powell IV, Wayne H. Rambo, Ann M. Whitehouse, and Elizabeth M. Whitehouse for their various contributions to the execution and scoring phases of the study.

Correspondence concerning this article should be addressed to Wayne G. Whitehouse, Unit for Experimental Psychiatry, The Institute of Pennsylvania Hospital, 111 North 49th Street, Philadelphia, Pennsylvania 19139-2798.




1 We wish to express our appreciation to Richard M. Lundy for facilitating this study by generously providing the relevant tape recordings and transcripts of recall questions.



Table 1

Mean Number of Correct Responses on the 40-Item Recall Tests Administered During Baseline (R1) and Treatment (R2) Phases

Hypnotizability/treatment condition

R 1   R2   Recall gain
  M SD M SD  
Low/wake 12.77 7.71 14.68 5.29 1.91
Low/hypnosis 13.68 3.43 15.44 3.68 1.76
High/wake 11.89 4.30 13.90 4.46 2.01
High/hypnosis 12.98 4.41 14.60 4.25 1.62



Table 2

Mean Confidence Rating (Range = 0-3) for Correct and Incorrect Responses to the 40-Item Recall Tests Administered During Baseline (R1) and Treatment (R2) Phases

Hypnotizability/treatment condition R 1   R2    
          Confidence increase
  M SD M SD  
    Correct Recall    
Low/wake 1.94 0.51 2.08 0.49 0.14
Low/hypnosis 1.99 0.39 2.17 0.39 0.18
High/wake 1.92 0.47 2.03 0.59 0.11
High/hypnosis 2.01 0.43 2.30 0.44 0.29
    Incorrect Recall    
Low/wake 0.68 0.32 0.86 0.39 0.18
Low/hypnosis 0.70 0.24 0.96 0.37 0.26
High/wake 0.69 0.31 0.92 0.48 0.23
High/hypnosis 0.66 0.19 1.21 0.35 0.55



Table 3

Mean Confidence Ratings (Range = 0-3) for Responses Given During Treatment Recall (R2) That Were Rated Previously as Guesses (Zero) During Baseline (R1)

          Confidence during R2        
Hypnotizability/treatment condition R1 Total Guesses R1 Correct Guesses R1 Incorrect Guesses
  M SD n M SD n M SD n
Low/wake 0.44 0.37 19 0.27 0.41 11 0.44 0.38 19
Low/hypnosis 0.64 0.56 20 0.48 0.53 8 0.64 0.56 20
High/wake 0.60 0.53 17 0.75 0.85 8 0.49 0.40 15
High/hypnosis 1.04 0.56 21 1.12 0.68 11 0.95 0.61 21

Note. Sample size (n) differs across the three dependent variables because not all subjects in the original treatment conditions provided at
least one instance of the class of guesses under consideration and/or repeated the response during R2.

The preceding paper is a reproduction of the pre-publication submission to the Journal of Abnormal Psychology (Whitehouse, W.G., Dinges, D.F., Orne, E.C., & Orne, M.T. Hypnotic hypermnesia: Enhanced memory accessibility or report bias?). It was eventually published in Journal of Abnormal Psychology, 1988, 97, 289-295. American Psychological Association © 1988. It is reproduced here with the kind permission of the American Psychological Association. No further reproduction or distribution of this artiocle is permitted without written permission of the publisher. Please consult the published versoin to cite quotations.