How accurate are the memories that are recovered under hypnosis

contentserver.asp_.pdf

Home >Psychology homework help >How accurate are the memories that are recovered under hypnosis

306

NHYP0020-7144Journal of Clinical and Experimental Hypnosis, Vol. 53, No. 03, May 2005: pp. 0–0Journal of Clinical and Experimental HypnosisTHE HARVARD GROUP SCALE OF HYPNOTIC SUSCEPTIBILITY:

Accuracy of Self-Report and the Memory for Items

HGSHS AND ACCURACY OF SELF-REPORT AND MEMORY FOR ITEMSJARRED YOUNGER ET AL. JARRED YOUNGER1,2

Arizona State University, Tempe, USA

DAVID D. KEMMERER, JUSTIN D. WINKEL, AND MICHAEL R. NASH

University of Tennessee, Knoxville, USA

Abstract: Whereas early studies have found moderately high agree- ment between self- and observer-rated scores on the Harvard Group Scale of Hypnotic Susceptibility, Form A (HGSHS:A), these studies shared a common confound in that participants were aware of being directly observed. In the present study, confederates made surrepti- tious observations of group participants’ hypnotic responding. Fol- lowing the hypnotic procedure, participants indicated whether or not they remembered each item and provided self-reports of their hyp- notic response. The study assesses the accuracy of participant self- report for hypnosis items when individuals are unaware of being observed. Thirty-two percent of participants failed to recognize at least one item from the hypnosis session, suggesting that the inability to remember items is a common phenomenon. When participants reported not remembering an item, the accuracy of their self-reported response was no better than chance.

The adaptation of hypnotizability scales to group settings, as done with the Harvard Group Scale of Hypnotic Susceptibility: Form A (HGSHS:A; Shor & Orne, 1963), has provided researchers an economical method of acquiring data. The use of large group sessions has become commonplace in hypnosis research (e.g., Lavertue, Kumar, & Pekala, 2002; Sapp & Hitchcock, 2003), providing investigators with the statis- tical power to test hypotheses that would otherwise be prohibitively labor intensive if individual protocols were used. In addition to their use in large correlational designs, group scales have also proved valuable

Manuscript submitted March 20, 2004; final revision received December 17, 2004. 1We wish to thank Krista Moe for her help in editing and preparing this manuscript. 2Address correspondence to Jarred Younger, St. Luke’s Medical Center, 1492 S. Mill

Ave., Suite 305, Tempe, AZ, 85282, USA. E-mail: [email protected]

HGSHS AND ACCURACY OF SELF-REPORT AND MEMORY FOR ITEMS 307

in the screening of participants for further investigation (e.g., Bryant & Idey, 2001; Page, Handley, & Carey, 2002).

Several potential problems are introduced with group hypnotizability scales, as they rely on the use of self-reported scores, which, unlike indi- vidually administered protocols, are recorded retrospectively (after hypnosis has been terminated). In order for this retrospective, self- reported assessment to be a valid measure of true behavioral response, many assumptions must be met. First, individuals must be able to accurately gauge motor response with proprioceptive, muscular feed- back alone, without the aid of normal visual cues (e.g., to pass the hand-lowering item of the HGSHS:A, the subject must endorse or reject the following statement: “My hand lowered at least six inches . . .”). Previous research in the somatosensory field has found that the esti- mated distance of hand and foot motion is distorted by motor speed when participants are not allowed to observe their movements (Leder- man, Klatzky, Collins, & Wardell, 1987). In general, slower movements result in overestimations of distance covered (Hollins & Goble, 1988). In a hypnotic environment, this phenomenon may translate into over- estimation of hypnotic response and inflated self-rated scores, espe- cially when a hypnotic measure makes extensive use of self-assessed motor behavior, as with the HGSHS:A.

A second assumption for self-reports to be a valid measure of true response is that individuals can retain an accurate memory of their behavioral responses to suggestions for later reporting. Memory of suggestions and experiences may be affected by normal forgetting, sleeping during the session, as well as by the hypnotic procedure itself (Kunzendorf & Benoit, 1985–1986). For example, the HGSHS:A includes a suggestion for posthypnotic amnesia of hypnotic items. Although the suggested amnesia is removed before any self-reports are made, the experience of posthypnotic amnesia may, in some people, extend beyond the suggestion itself.

Third, there is an assumption that participants’ self-ratings are not influenced by what they feel they are supposed to report. Self-reported responses might stray from actual behavior due to the demand charac- teristics of the hypnotic procedure. It is possible that, wittingly or unwittingly, the subjects might present themselves as more responsive than they actually were in order to please the experimenters. Contex- tual factors and situational demands have been found to introduce reporting biases (Murray, Cross, & Whipple, 1992; Spanos, Burgess, Cocco, & Pinch, 1993). The need to insure (a) accurate perception, (b) accurate memory, and (c) unbiased reporting is especially important in hypnotic self-report, since hypnosis has been shown to, in some cases, manipulate the accuracy of perception (Nordby, Hugdahl, Jasiukaitis, & Spiegel, 1999; Roberts, 1991; Von Kirchenheim & Persinger, 1991), accuracy of memory (Sheehan, 1988; Weekes, Lynn, Green, & Brentar,

308 JARRED YOUNGER ET AL.

1992), and reporting bias (Perlini, Haley, & Buczel, 1998; Spanos, 1992). Although the type and degree of distortion may vary with the specific procedure and suggestions used, it is important to note that qualities inherent to the hypnosis process may affect the accuracy of self- reported behavioral response.

The possibility of measurement error due to the previously stated factors was recognized by early hypnosis researchers, who compared self-reported with observer-reported hypnotic suggestibility scores. Early studies on the agreement between self- and observer-reported scores generally found correlations in the .80 to .90 range (Bentler & Hilgard, 1963; Coe, 1964; O’Connell, 1964; Shor & Orne, 1963), although many studies found relationships between .60 and .70 (Bentler & Roberts, 1963; Hartman, 1965; Levitt, Aronoff, & Morgan, 1974).

Based on these previous results, it would seem that self-reported scores are a reasonable substitute for observer-rated scores. However, all previous studies share a common confound—the use of manifest observers. Therefore, the subjects in these studies were aware that their responses were being observed and recorded. In large group hypnosis sessions, subjects likely feel more anonymous. Hence, we do not know how well the HGSHS:A self-report score indexes actual hypnotic behavior in a standard, large-group environment. Are subjects less accurate in their hypnotic response reporting under more anonymous conditions? If so, is the error random or do subjects over- or underesti- mate their actual behavior? Eisen (1996) did not observe behavior but found that telling subjects they were being observed suppressed scores relative to subjects who were told their responses were anonymous. The first objective of the present study is to assess the accuracy of participant self-report for items when individuals are unaware of being observed. This procedure more accurately approximates the conditions under which the HGSHS:A is usually administered.

The second objective of the present study is to monitor the extent to which HGSHS:A items are forgotten by subjects. There are several possible factors that could lead to failure to remember an item following a hypnotic procedure. For example, individuals may fall asleep during a session. Although such an event could be observed and rectified in an individual session, it might not be caught in a group session. Some previous research also suggests that the hypnotic procedure itself may evoke amnesia in some participants (Kunzendorf & Benoit, 1985–1986). If sleeping, spontaneous amnesia, and normal forgetting are phenomena present in hypnosis sessions, the wording of self-report packets would be troubling because it (a) strongly encourages a guess when the participant is unsure of their actual response and (b) is coded so that a participant who falls asleep could pass certain items. On the HGSHS:A, three items (arm immobilization, communication inhibi- tion, and eye catalepsy) are scored as a pass based on the total absence

HGSHS AND ACCURACY OF SELF-REPORT AND MEMORY FOR ITEMS 309

of a behavioral response. A sleeping person could, for example, honestly report that their eyes did not open during the eye catalepsy suggestion and pass that item. Therefore, this study assesses not only the extent to which subjects fail to remember items but also the impact of this nonremembering on the accuracy of self-report.

METHOD

Participants Participants were 79 students in various psychology classes at the

University of Tennessee, Knoxville. Fifty-three females and twenty-six males participated in the study. Participants responded to an announce- ment of extra credit located on a general notice board and were instructed by the announcement to meet at the session location.

Confederates Nine undergraduate research assistants served as confederates.

These assistants were trained by the primary investigators on behav- ioral scoring of the HGSHS:A. After training, confederates indepen- dently viewed and scored a videotaped hypnosis session. Cohen’s kappa was used to assess interrater agreement. Agreement was high (κ = .82), with only three discrepancies across all items and raters. These discrepancies were then reviewed and consensus achieved.

Materials In order to standardize administrations of the hypnosis sessions, a

CD version of the HGSHS:A was used. The audio presentation was played through speakers positioned at the front of the session room. To exactly replicate the standard Harvard procedure, all 12 hypnotic suggestions were given, although only 11 were considered in the present study because an observer cannot confirm the amnesia item. The standard HGSHS:A packet was modified in order to assess memory of each suggestion. The modification involved asking, prior to each item in the scoring packet, whether or not the participant remem- bered the suggestion. For example, for the hands-moving-together item, the question was worded as follows (additions in italics):

You were next told to hold your hands out in front of you about a foot apart and then told to imagine a force pulling your hands together.

A. I remember being told to think of a force pulling my hands together. B. I do not remember being told to think of a force pulling my hands together.

Would you estimate that an onlooker would have observed that your hands were not over six inches apart (before you were told to return your hands to their resting position)?

310 JARRED YOUNGER ET AL.

A. My hands were not more than six inches apart by then. B. My hands were still more than six inches apart by then.

Finally, a manipulation check sheet was added at the end of the protocol. This sheet was used to assess whether any participants were aware of being observed during the procedure. The sheet included both specific distractors (including the target item, “people looking around the room”) and an open-ended question that asked partici- pants to guess the purpose of the study.

Procedure Ten groups were run with an average of 8 true participants in each

session. All sessions were held in the same room, which was selected for good sound isolation and control over light levels. A table was placed in the center in the room and chairs placed on both sides with numbered packets. Participants were assigned seats randomly as they arrived. Each confederate (up to 5 in any single session) came into the room in a random fashion, so as to look like a real participant. Seat arrangement was designed so that each confederate would have a clear view of two participants directly across the table from him or her. All participants were instructed to keep at least one arm’s length away from the table. Although participants were told to do so in order to allow their arms to move freely during the session, this arrangement also allowed the confederates to clearly see all behavioral responses.

To aid in the recording of behavioral observations, each confederate wore a flesh-colored bandage on the palm of their nondominant hand. Behavioral codings (pass/fail) were tallied on these bandages during the protocol and then transferred to an observation sheet following the ses- sion. In order to avoid detection, confederates followed the protocol and kept their eyes closed except for when rating participants’ responses.

When all participants and confederates had been seated, lights were dimmed by one of the primary investigators, and the CD protocol was started. To approximate a large-group session, one experimenter remained in the room throughout the session but stayed in the back of the room, so that participants would not feel they were being closely observed during the protocol.

RESULTS

Manipulation Check Key to the study was the ability of confederates to remain undetected

by the real participants. To ascertain the effectiveness of the deception, two manipulation checks were made. First, participants were asked to rate nine factors on how much they distracted from the hypnotic experience. Exam- ples include “noises outside the room” and “temperature.” Responses

HGSHS AND ACCURACY OF SELF-REPORT AND MEMORY FOR ITEMS 311

were given on a 5-point scale ranging from not distracting to very distracting. The manipulation check, “People looking around the room,” was reported as the least distracting (M = 1.09, SD = 0.40) of the nine factors, whereas “seat uncomfortableness” was rated as the most distracting (M = 2.66, SD = 1.19). A second manipulation item, “Presence of experimenter(s),” was also rated as only slightly distracting (M = 1.38, SD = 0.84).

As a further check of the manipulation, all individuals were instructed to guess the purpose of the study. Participants’ written responses were later reviewed and no participants indicated being aware of the observ- ers. Based on these data, we can be reasonably sure that the deception was successful and the hidden observations went unnoticed.

Accuracy of Self-Reported Scores The mean self-reported hypnotic suggestibility score was 5.90

(SD = 2.60), while the mean observed score was 5.23 (SD = 2.35). These scores were significantly different (t (78) = −3.0, p = .004). The correlation between self- and observer-rated scores was .68 (p < .001), demonstrating that self-reported scores predict just below 50% of the variance of observer-rated scores. As reported in previous studies, self-reported scores were typically higher than observer-reported scores. Although general agreement in the present study is lower than that reported in previous studies, this difference might be due to the use of hidden rather than manifest observers. Table 1, column A, presents accuracy of self-reported scoring by item. Agreement ranged from 64% (arm immobilization) to 82% (hand lowering, fly suggestion, and ankle

Table 1 Agreements and Discrepancies of Self-Reported and Observer-Reported Hypnotic Response

% agreement

Cohen’s kappa

false positives

false negatives

not remembered but self passed

Head falling 79 .56 19 2 2 Eye closing 78 .35 3 19 1 Hand lowering 82 .69 10 8 1 Arm

immobilization 64 .14 22 14 17

Finger lock 76 .55 9 15 4 Arm rigidity 77 .57 19 4 3 Hands moving 80 .53 13 7 4 Communication

inhibition 71 .34 21 8 5

Fly suggestion 82 .36 8 10 0 Eye catalepsy 70 .34 9 21 6 Ankle touching 82 .07 13 5 2

312 JARRED YOUNGER ET AL.

touching). A general tendency toward overestimation of behavioral response was found, with participants inaccurately reporting a pass (seven items) more often than inaccurately reporting a fail (four items). Figure 1 shows pass frequency by item as rated by both the partici- pants and their observers. Because hypnosis scoring is dichotomous, chance agreement between two raters is very high. Cohen’s kappa adjusts for chance and is a more conservative statistic of agreement. As seen in Table 1, column B, agreement between self-reported and observer-reported scores is low. Kappas range from .07 (ankle touching) to .69 (hand lowering), with an average kappa of .41. It is important to note, however, that the kappa statistic is affected by unequal distribu- tions, such as seen in the low frequency of ankle touching passes.

General categories (low, medium, high) are frequently used in hypnosis research to screen participants for further investigation. Agreement in classification based on self-reported and observer- reported scores was low (κ = .37). Table 2 crosses frequency of low (score: 0–3), medium (4–7), and high (8–11) susceptible individuals based on both self-reported and observer-reported scores. As seen in

Figure 1. Pass frequency as rated by self and observer.

100

fa ll

cl os e

lo w er

im m ob ile lo ck ri gi d

to ge th er

in hi bi t fly

ca ta le ps y

an kl e

rater self

HGSHS AND ACCURACY OF SELF-REPORT AND MEMORY FOR ITEMS 313

the table, many more individuals would be placed in the high category if using self-reports than if using observed behavior.

Impact of “Not Remembering” on the Accuracy of Self-Report Participants’ inability to recall specific hypnosis items was investigated

as a possible factor in self-report error. First, the frequency of not remem- bering was assessed. Second, the relationship between not remembering and self-report accuracy was determined.Before each item in the HGSHS:A response booklet, participants were instructed to indicate whether or not they remembered being given the particular suggestion. These responses to the memory question were summed up across items, yielding the total number of items remembered by each participant. The incidence of not remembering was high, with 32% of participants failing to recognize at least one item; 14% indicated not remembering exactly one item; 6, 3, 5, and 1% failed to remember two, three, four, and five items, respectively. Finally, 3% failed to remember six items. These results indicate that the inability to recall items at time of self-report is a frequent phenomenon.

To determine if some items were being recalled less than others, frequency of recognition was tabulated for each item. Percentage of participants remembering suggestions (items listed in order of presentation) were: head falling (98%), eyes closing (98%), hand lower- ing (98%), arm immobilization (77%), finger locking (95%), arm rigid- ity (96%), hands moving together (95%), communication inhibition (94%), fly suggestion (94%), eye catalepsy (91%), and ankle touching (86%). In general, suggestions placed later in the procedure were rec- ognized by fewer participants. These results might point to an increasing number of participants falling asleep as the procedure continues. One interesting exception to this trend was the arm immobilization item, which was recognized by only 77% of the participants. Although it is unclear why this item would be recognized by fewer participants, wording in the response booklet may have confused some respondents.

A correlation was calculated to assess the general relationship between number of items recalled and self-report accuracy (total

Table 2 Classification of Lows (score: 0–3), Mediums (4–7), and Highs (8–11) by Self- and Observer-Ratings

Observer Rating

LOW MEDIUM HIGH

Subject Rating HIGH 17 8 MEDIUM 4 33 4 LOW 9 4

314 JARRED YOUNGER ET AL.

number of rater/observer agreements). The significant relationship (r = .31, p = .006) suggests that participants are more accurate in their self-report as they remember more items. Participants who recognized fewer items were also less accurate in their overall self-report scoring.

The error rate, across all participants, for all remembered items was 22%, while the error rate for not remembered items was 40%. A chi- square (χ2 = 9.87, p = .001) revealed that these error rates were signifi- cantly different. Further, and most important, a chi-square analysis indicated that accuracy among not remembered items was no better than what would be expected by chance (χ2 = 1.21, p = .20). Therefore, the present results suggest that accuracy of self-report is no better than chance when the individual indicates that they do not remember the item. Given the number of individuals who indicated not remembering one or more items, this phenomenon may present a threat to accurate hypnotic measurement when using only self-reported scores.

To make the findings of the present study more generalizable to other group hypnosis sessions, the entire HGSHS:A protocol was used, includ- ing the posthypnotic amnesia item. We recognize, however, that the intro- duction of an amnesia item could have an effect on memory, even though the suggestion is removed before any self-reports are made. To test for possible effects of suggested amnesia on later self-reported scoring, we conducted t tests between those passing the posthypnotic amnesia item and those failing the item. Those responding to suggested amnesia were no less accurate in their self-report than those not experiencing amnesia, t(77) = −.82, p = .42. Further, those passing the amnesia item recalled just as many items in the scoring packet as did those not experiencing amne- sia, t(77) = .10, p = .92. Therefore, the effects of temporary suggested amnesia do not appear to confound later memory for items.

DISCUSSION

In allowing participants to self-score their hypnotic susceptibility, many assumptions are made about the ability of individuals to accurately monitor, remember, and report their behavioral responses. Distortions of perception, memory, or reporting could all confound hypnosis data col- lected from participant-scored sessions. The present study was designed to examine the impact of some of these distortions on self-report accuracy. By placing hidden observers in group hypnosis sessions, it was possible to contrast self-scored behavior to observer-rated behavior, enabling us to observe the prevalence of inaccurate participant reporting.

Previous research has found that individuals have difficulty perceiving the distance of their motor behaviors when movements are slowed and visual feedback is removed (Lederman, Klatzky, Collins & Wardell, 1987). In the present study, the tendency to overestimate response when using proprioceptive feedback alone (see Hollins &

HGSHS AND ACCURACY OF SELF-REPORT AND MEMORY FOR ITEMS 315

Goble, 1988) was considered in the hypnotic environment. To the extent that this somatosensory phenomenon affects accuracy of self- report in hypnosis procedures, we would expect individuals to overes- timate their response to items involving judgments of distance covered (head falling, hand lowering, arm rigidity, and hands moving together). As seen in Table 1, columns B and C, individuals did, on average, overestimate their response on all these items, suggesting that findings reported in the somatosensory field are relevant to hypnosis sessions. However, since these items were no more likely to be inaccu- rately reported than other items (see Table 1, column A), the impact of this phenomenon on self-report accuracy is probably minor.

Memory for hypnotic suggestions, the second questioned assumption, had a profound impact on the accuracy of hypnotic self-report. Logi- cally, the inability to recall an item should interfere with the accurate reporting of one’s response to that item. Traditionally, however, partic- ipants are strongly encouraged to guess the degree of their response even if they do not remember being given the suggestion. In the HGSHS:A, for example, participants are informed, “failure to give a definite answer to every question may lead to disqualification of your record.” The assumption inherent to this practice is that a guess provides reliable data toward scoring hypnotic suggestibility, although this assumption has not been demonstrated empirically. In fact, previous research has shown that forced responses to items that are not remembered can lead to random responding (Gudjonsson, 1987). Indeed, the results of the present study argue strongly against the encouragement of respondent guessing, as participants were accurate at no better than chance when scoring a not-remembered item. Given the high number of participants (32%) who failed to recognize at least one item in the response booklet, participant guessing could introduce a significant amount of error in self-reported hypnotic measurement.

The impact of participant memory becomes especially important with items passed by the absence of movement. Three items in the HGSHS:A (eye catalepsy, communication inhibition, and arm immobilization) are scored as passes if the participant indicates making no movement. With the standard response protocol and instructions, a person who falls asleep during the protocol or is otherwise unable to recall the item is, in those cases, forced to make a report that would likely result in a positive score. If forced to respond to every item, a person could sleep through an entire protocol and score a three on the scale. Indeed, as seen in Table 1, column D, the three items involving absence of movement are also the three items most frequently containing the paradoxical “not remem- bered, but passed” score. Unfortunately, no study has provided esti- mates of the sleeping frequency in hypnosis sessions. Although the frequency of sleeping was not assessed in the present study, sleeping could explain why more individuals failed to recall items placed later in

316 JARRED YOUNGER ET AL.

the protocol. Sleeping could also help explain factor loadings of HGSHS:A items performed by other researchers. For example, items that could be passed by sleeping individuals cluster together in Sadler and Woody’s (2004) unrotated, two-factor structure.

Memory for experiences under hypnosis is potentially confounded by the final suggestion in the HGSHS:A, which is amnesia for items. Although this suggestion is removed before any self-report of hypnotic responding is made, some might suggest that memory in highly susceptible individuals is nonetheless altered beyond the removal of the suggestion, which is achieved by stating, “Now you can remember everything.” The results of the present study, however, do not suggest that amnesia remains after the suggestion. Individuals who passed the amnesia item later recognized the same number of items and were just as accurate in their self-report as those not passing the amnesia item.

As observed in a number of studies, participants generally err in over- estimating their responses (Levitt et al., 1974). Although general soma- tosensory mechanisms described above may be partially responsible for this phenomenon, other sources may drive self- and observer-rated dis- crepancies. Two possible explanations that have been discussed in the lit- erature include distortions in experience and distortions in reporting. Given the number of studies evidencing the possibility of hypnotically induced distortions of experience and memory (Laurence & Perry, 1983; Sheehan, Green, & Truesdale, 1992; Spanos & McLean, 1986; Weekes et al., 1992), it is possible that the tendency for participants to overestimate their hypnotic response reflects a phenomenon in which individuals experience a suggestion but display no observable behavior. For exam- ple, an individual may experience his or her hand dropping (or later believe that it did drop) in response to the hand-lowering item, but no such drop is seen by observers. Although the gold standard of hypnosis remains observable, behavioral response to suggestions, and discrepan- cies between self-reports and observer-reports may delineate different types of hypnotic responding, rather than distortions in self-reporting. Because past research has reported good agreement between self- and observer-reported scores, it has been assumed that both raters are scoring the same behavior/experience. However, when using Cohen’s kappa to adjust for chance, agreement between the participants and their observ- ers was low. These findings suggest that the internal experience of the hypnotized individual may be quite different from observable behaviors.

Rather than distorting experiences, hypnosis may manipulate reporting bias (Zamansky, 1986). Although not tested in the present study, accuracy of reporting (the third assumption) may be severely disturbed by demand characteristics (Rosnow, 2002) and other sources of reporting bias. If hypnosis is purely compliance with demands, subjects may comply with overt behavior with individual administra- tions but simply give positive reports if they feel they are not being

HGSHS AND ACCURACY OF SELF-REPORT AND MEMORY FOR ITEMS 317

observed. In order to explore these possibilities further, future studies may more closely examine instances in which self-reporting differs from outside observations. Although it is not clear which factor (if not both) leads to participant response overestimation, the end result is the same: self-reported inflation results in many more individuals falling into the highly hypnotizable category. It is interesting that the discrep- ancy is not uniform across the range of hypnotizability scores but affects the high end to a greater degree than the low.

In generalizing these findings to other groups, it is important to note that the participants in the present study were particularly motivated in that they sought out an extra-credit study and attended an evening session. Failure to remember (especially when caused by sleeping or complete noninvolvement with the protocol) may increase in groups with lower intrinsic motivation (for example, when using a captive audience in a regularly scheduled class). Future studies should assess recall difficulty and self-report accuracy in these environments.

Current difficulties in studying hypnosis may be due, in part, to researchers’ willingness to accept false data over no data. The self-report measurement of hypnosis may be improved by allowing participants the option of not answering a question of which they are unsure. It is also recommended that self-report packets be modified to assess nonin- volvement with the protocol. Examples would include instructing par- ticipants to indicate if and at what point they fell asleep, if they became disengaged from the procedure, and, as suggested by Eisen (1996), if there are particular items they cannot remember. These modifications to self-report packets would likely reduce noise in data collection and increase the power of hypothesis testing, thus boosting the effectiveness and impact of hypnosis studies. The present results further caution researchers in using the HGSHS:A as their sole means of screening for individual differences in hypnotizability. These studies (e.g., Braffman & Kirsh, 1999; Sapp & Hitchcock, 2003) may be particularly susceptible to the confounding influences of perceptual, psychological, and social factors specific to the group hypnosis environment. The net effects of these factors lead to self-rated score inflation and, in many cases, an incorrect classification of highly hypnotizable, when such a categoriza- tion system is used. The advice of the original authors of the HGSHS:A is sound: group scores should be confirmed in an individual session.

REFERENCES

Bentler, P. M., & Hilgard, E. R. (1963). A comparison of group and individual induction of hypnosis with self-scoring and observer-scoring. International Journal of Clinical and Experimental Hypnosis, 11, 49–54.

Bentler, P. M., & Roberts, M. R. (1963). Hypnotic suggestibility assessed in large groups. International Journal of Clinical and Experimental Hypnosis, 11, 93–97.

318 JARRED YOUNGER ET AL.

Braffman, W., & Kirsch, I. (1999). Imaginative suggestibility and hypnotizability: An empirical analysis. Journal of Personality and Social Psychology, 77, 578–587.

Bryant, R. A., & Idey, A. (2001). Intrusive thoughts and hypnotizability. Contemporary Hypnosis, 18(1), 14–20.

Coe, W. C. (1964). Further norms on the Harvard Group Scale of Hypnotic Susceptiblity, Form A. International Journal of Clinical and Experimental Hypnosis, 12, 184–190.

Eisen, M. L. (1996). The relationship between memory, suggestibility, and hypnotic responsivity. American Journal of Clinical Hypnosis, 39, 126–137.

Gudjonsson, G. H. (1987). Historical background to suggestibility: How interrogative suggestibility differs from other types of suggestibility. Personality and Individual Differences, 8, 347–355.

Hartman, B. J. (1965). Self-scoring and observer-scoring estimates of hypnotic suscepti- bility in a group situation. Perceptual and Motor Skills, 20, 452.

Hollins, M., & Goble, A. K. (1988). Perception of the length of voluntary movements. Somatosensory Research, 5(4), 335–348.

Kunzendorf, R. G., & Benoit, M. (1985–1986). Spontaneous post-hypnotic amnesia and spontaneous rehypnotic recovery in repressers. Imagination, Cognition and Personality, 5(4), 303–310.

Laurence, J. R., & Perry, C. (1983). Hypnotically created memory among highly hypno- tizable subjects. Science, 222, 523–524.

Lavertue, N. E., Kumar, V. K., & Pekala, R. J. (2002). The effectiveness of a hypnotic ego- strengthening procedure for improving self-esteem and depression. Australian Journal of Clinical and Experimental Hypnosis, 30(1), 1–23.

Lederman, S. J., Klatzky, R. L., Collins, A., & Wardell, J. (1987). Exploring environments by hand or foot: Time-based heuristics for encoding distance in movement space. Journal of Experimental Psychology: Learning, Memory, and Cognition, 13, 606–614.

Levitt, E. E., Aronoff, G., & Morgan, C. D. (1974). A note on possible limitations on the use of the Harvard Group Scale Of Hypnotic Suggestibility, Form A. International Journal of Clinical and Experimental Hypnosis, 22, 234–238.

Murray, G. J., Cross, H. J., & Whipple, J. (1992). Hypnotically created pseudomemories: Further investigation into the “memory distortion or response bias” question. Journal of Abnormal Psychology, 101, 75–77.

Nordby, H., Hugdahl, K., Jasiukaitis, P., & Spiegel, D. (1999). Effects of hypnotizability on performance of a Stroop task and event-related potentials. Perceptual and Motor Skills, 88, 819–830.

O’Connell, D. N. (1964). An experimental comparison of hypnotic depth measured by self-ratings and by an objective scale. International Journal of Clinical and Experimental Hypnosis, 12, 34–46.

Page, R. A., Handley, G. W., & Carey, J. C. (2002). Can devices facilitate a hypnotic induc- tion? American Journal of Clinical Hypnosis, 45(2), 137–141.

Perlini, A. H., Haley, A., & Buczel, A. (1998). Hypnosis and reporting biases: Telling the truth. Journal of Research in Personality, 32, 13–32.

Roberts, S. (1991). Some apparently non-cerebral aspects of consciousness. Journal of the Society for Psychical Research, 58, 31–38.

Rosnow, R. L. (2002). The nature and role of demand characteristics in scientific inquiry. Prevention and Treatment, 5, np.

Sadler, P., & Woody, E. Z. (2004). Four decades of group hypnosis scales: What does item-response theory tell us about what we’ve been measuring? International Journal of Clinical and Experimental Hypnosis, 53, 132–158.

Sapp, M., & Hitchcock, K. (2003). Measuring dissociation and hypnotizability with African American college students: A new dissociation scale—the General Dissociation Scale. Australian Journal of Clinical Hypnotherapy and Hypnosis, 24, 14–22.

Sheehan, P. W. (1988). Memory distortion in hypnosis. International Journal of Clinical and Experimental Hypnosis, 36, 296–311.

HGSHS AND ACCURACY OF SELF-REPORT AND MEMORY FOR ITEMS 319

Sheehan, P. W., Green, V., & Truesdale, P. (1992). Influence of rapport on hypnotically induced pseudomemory. Journal of Abnormal Psychology, 101, 690–700.

Shor, R. E., & Orne, E. C. (1963). Norms of the Harvard Group Scale of Hypnotic Suscep- tibility, Form A. International Journal of Clinical and Experimental Hypnosis, 11, 39–47.

Spanos, N. P. (1992). Compliance and reinterpretation in hypnotic responding. Contem- porary Hypnosis, 9, 7–15.

Spanos, N. P., Burgess, C. A., Cocco, L., & Pinch, N. (1993). Reporting bias and response to difficult suggestions in highly hypnotizable hypnotic subjects. Journal of Research in Personality, 27, 270–284.

Spanos, N. P., & McLean, J. (1986). Hypnotically created pseudomemories: Memory distortions or reporting biases? British Journal of Experimental and Clinical Hypnosis, 3, 155–159.

Von Kirchenheim, C., & Persinger, M.A. (1991). Time distortion: A comparison of hypnotic induction and progressive relaxation procedures: A brief communication. International Journal of Clinical and Experimental Hypnosis, 39, 63–66.

Weekes, J. R., Lynn, S. J., Green, J. P., & Brentar, J.T. (1992). Pseudomemory in hypno- tized and task-motivated subjects. Journal of Abnormal Psychology, 101, 356–360.

Zamansky, H. S. (1986). Hypnotically created pseudomemories: Memory distortions or report- ing biases? Comment. British Journal of Experimental and Clinical Hypnosis, 3, 160–161.

Die Harvard Group Scale of Hypnotic Susceptibility: Genauigkeit des Selbstberichts und Gedächtnis für die Items

Jarred Younger, David D. Kemmerer, Justin D. Winkel und Michael R. Nash Zusammenfassung: Frühe Studien berichten eine mäßig hohe Übereinstimmung zwischen Selbstberichts- und Fremdberichts-Ratings der Harvard Group Scale of Hypnotic Susceptibility, Form A (HGSHS:A). Bei all diesen Studien besteht jedoch das Problem, dass die Teilnehmer wussten, dass sie direkt beobachtet wurden. In der vorliegenden Studie beobachteten eingeweihte Komplizen heimlich die hypnotischen Reaktionen der übrigen Teilnehmer einer Gruppe. Im Anschluss an die Hypnosesitzung gaben die Teilnehmer an, ob sie sich an die einzelnen Items erinnerten und gaben Selbstberichte ihrer Hypnoseerfahrungen ab. In der Studie wurde die Genauigkeit des Selbstberichts der Teilnehmer für die Hypnoseitems erhoben, während die Teilnehmer nicht wussten, dass sie beobachtet wurden. Zweiunddreißig Prozent der Teilnehmer konnten mindestens ein Item der Hypnosesitzung nicht mehr wiedererkennen, was darauf hindeutet, dass das Nichterinnern von Items ein verbreitetes Phänomen darstellt. Wenn die Teilnehmer angaben, dass sie sich nicht an ein Item erinnern konnten, lag die Genauigkeit ihres Selbstberichts auf Zufallsniveau.

RALF SCHMAELZLE University of Konstanz, Konstanz, Germany

L’échelle de groupe de susceptibilité hypnotique d’Harvard : exactitude de l’auto-évaluation et de la mémoire des items.

Jarred Younger, David D. Kemmerer, Justin D. Winkel, et Michael R. Nash Résumé : alors que des études précédentes montraient un accord modérement élevé entre les résultats ‘auto-évalués’ et ‘évalués par un

320 JARRED YOUNGER ET AL.

observateur’ sur l’échelle de groupe de susceptibilité hypnotique d’Harvard, formulaire A (HGSHS :A), ces études avaient en commun un biais de recherche, à savoir que les participants savaient qu’ils étaient observés. Dans cette nouvelle étude, des complices ont effectué des observations discrètes d’un groupe de participants sous hypnose. Suite à la procédure hypnotique, les participants devaient indiquer s’ils se souvenaient ou non de chaque items et devaient effectuer un auto-compte-rendu de leur réponse hypnotique. L’étude a évalué l’exactitude de l’auto compte-rendu de participants quand ceux-ci n’étaient pas conscients d’être observés. 32% des participants n’arrivaient pas à se souvenir d’au moins un item de la séance d’hypnoe, suggérant que l’inabilité à se souvenir d’un item est un phénomène commun. Lorsque les participant rapportaient ne pas se souvenir d’un item, l’exactitude de l’auto évaluation était ni plus ni moins due au hasard.

VICTOR SIMON Psychosomatic Medicine & Clinical Hypnosis Institute, Lille, France

La Escala Grupal de Susceptibilidad Hipnótica de Harvard: La precisión del auto-informe y el recuerdo de los reactivos

Jarred Younger, David D. Kemmerer, Justin D. Winkel, y Michael R. Nash Resumen: Mientras que los primeros estudios encontraron un acuerdo relativamente alto entre las puntuaciones de auto informes y de observadores en la Escala Grupal de Susceptibilidad Hipnótica de Harvard, Forma A (HGSHS:A), estos estudios compartieron el problema de que los participantes sabían que los estaban observando. En este estudio, los confederados del experimentador realizaron observaciones surrepticias de las respuestas hipnóticas de los participantes. Después del procedimiento hipnótico, los participantes indicaron si recordaban cada reactivo y proporcionaron auto-informes de su respuesta hipnótica. El estudio valoró la precisión del auto-informe de los participante sobre reactivos de hipnosis cuando los individuos no sabían que se les observaba. El 32% de los participantes no pudo reconocer por lo menos un reactivo de la sesión de hipnosis, sugiriendo que la incapacidad para recordar reactivos es un fenómeno común. Cuándo los participantes mencionaron no recordar un reactivo, la precisión de su respuesta fue al azar.

ETZEL CARDEÑA University of Texas, Pan American, Edinburg, Texas, USA