*Bolton Institute and **University of Liverpool, UK

Transcription

1

2 Contemporary Hypnosis (1997) Vol. 14, No. 3, pp MAIN PAPER AN INVESTIGATION INTO THE INFLUENCE OF HYPNOSIS ON THE CONFIDENCE AND ACCURACY OF EYEWITNESS RECALL Mark R. Kebbell* and Graham F. Wagstaff** *Bolton Institute and **University of Liverpool, UK ABSTRACT Research suggests that although confidence accuracy (C A) relationships are typically low, investigative interviewing with hypnosis may have a particularly adverse influence on C A relationships. However, it is possible that researchers may have paid insufficient attention to the issue of item difficulty. To address this issue an experiment was conducted which measured recall of information seen in a video film. Questions were constructed to range in difficulty and no misleading post-event information and/or leading questions were presented. Subjects were assigned to either a hypnosis condition or one of two control conditions. Higher C A correlations than have been usually reported were found, regardless of interview condition.also, when subjects were absolutely certain that a piece of information was correct they almost invariably were accurate, and again interview condition did not influence this. Some practical implications of these findings are discussed. INTRODUCTION Many studies have shown that eyewitnesses who are confident in their testimonies are more likely to be perceived as accurate than eyewitnesses who are hesitant in their testimonies (Cutler, Penrod & Stuve, 1988; Cutler, Penrod & Dexter, 1990; Fox & Walters, 1986; Leippe, Manion & Romanczyk, 1992; Lindsay, 1994; Lindsay, Wells & O Connor, 1989). Consequently, the confidence which eyewitnesses express in information that they provide heavily influences both the investigative process, and the credence which jurors give to eyewitness testimony. Thus, care must obviously be taken to ensure that eyewitnesses are interviewed in a way that does not adversely affect the relationship between their expressed confidence and the accuracy of their testimony. The possible use of hypnosis as an interview procedure to enhance eyewitness memory has provoked much debate (see, for example, Reiser, 1980, 1990; Wagstaff, 1993). According to some, forensic investigators should not use hypnosis because it encourages various forms of memory distortion (for reviews see, for example, Diamond, 1988; Orne, Sokis, Dinges & Orne, 1984; Sheehan, 1988; Wagstaff, 1993). Such conclusions are based on a number of investigations that have found that hypnotic subjects, 157

3 158 Kebbell and Wagstaff especially (but not always) those categorized as highly hypnotically susceptible, often show increased confabulation and a spurious confidence in inaccurate memories (see, for example, Dywan & Bowers, 1983; Rainer, 1983; Sheehan & Tilden, 1983; Sheehan, Grigg & McCann, 1984; Zelig & Beidleman, 1981); though this is not always the case (see, for example, Redson & Knox, 1983; Sanders & Simmons, 1983). On the basis of such studies, Sheehan (1988) concludes that probably the most consistent finding is that, whatever their accuracy, hypnotic subjects express high levels of confidence in information that they provide, both accurate and inaccurate. Hence, while not all the data support a weak or negative confidence accuracy (C A) relationship, the evidence suggests an absence of a large, positive C A relationship in situations in which hypnosis is used to facilitate the retrieval of eyewitness information. However, the failure to find strong C A relationships does not appear to be limited to situations in which hypnosis is used. Many studies suggest that in non-hypnotic situations, there is either no C A relationship at all, or only a small, positive C A relationship (for reviews see, Bothwell, Deffenbacher & Brigham, 1987; Deffenbacher, 1980; Fruzzetti, Tolland, Teller & Loftus, 1992; Wells & Murray, 1984). Nevertheless, Kebbell, Wagstaff and Covey (1996) have suggested that previous researchers in this area may have paid insufficient attention to the issue of item difficulty. Typically in work on C A relationships, researchers attempt to select items to avoid floor and ceiling effects; i.e. they try to avoid items that are either very easy or very hard to remember. In real life, however, forensic investigators may ask eyewitnesses some questions that are easier to answer than others. For example, for an assault, an investigator may ask an eyewitness What sex was the attacker?. Gender is among the first items noticed about an individual and is very likely to be answered accurately. Furthermore, most eyewitnesses are likely to be very confident that the identification of an individual s sex is correct. Alternatively, an investigator might ask an eyewitness, What was the colour of the attacker s eyes? This question might be more difficult to answer and eyewitnesses may be less confident about their accuracy (see Christianson & Hubinette, 1993). Consequently, for easy questions eyewitnesses might be both confident and accurate while for difficult questions they might be less confident and less accurate. In these circumstances a positive relationship between confidence and accuracy would occur. It may be the case, therefore, that in an attempt to avoid ceiling and floor effects, previous researchers may have chosen unrealistic and overly homogeneous pools of items, thus reducing the variance necessary for high correlations. Consequently, higher C A relationships might result with the use of a heterogeneous range of hard or easy questions. Another possibly important related factor is the relationship between absolutely certain responses and accuracy. Absolutely certain responses may have a particularly strong impact on the police and jurors but are unlikely to occur unless items that are easy are included. Regardless of the overall C A relationship, it could be the case that the relationship between these absolutely certain responses and accuracy remains high (Gruneberg & Sykes, 1993). Kebbell et al. (1996) addressed these issues by conducting two experiments that measured C A relationships in response to information seen in video films. In each case, questions were used that ranged in difficulty, from easy to hard. Higher C A correlations than usually reported were found in both experiments (r = 0.54 and r = 0.78). Furthermore, when subjects were absolutely certain that a piece of information was correct they almost invariably were accurate. These findings have since been replicated (Pike, Towel & Kemp, 1995; see also, Sporer, Penrod, Read & Cutler,

4 1995). This presents the possibility that by using questions of varying difficulty, it may be possible to compare more sensitively and realistically differences between hypnosis and control procedures in terms of their effects on the C A relationship. The aim of the present study was to address this issue. On the basis of previous research and theorizing we hypothesized that hypnosis would have more of an adverse effect on C A relationships than in control conditions. Two control conditions were used. A major problem in research in this area is determining what features of hypnotic interviewing procedures are responsible for uniquely hypnotic effects. A whole host of factors are associated with hypnotic interviewing quite apart from the hypnotic induction procedure itself (Spanos & Chaves, 1989). These include instructions to the witness to reinstate context mentally, report everything, recall events in different orders and change perspectives (Geiselman, Fisher, MacKinnon & Holland, 1985; Gudjonsson, 1992; Hibbard & Worring, 1981; Reiser, 1980, 1990; Wagstaff, 1982, 1993). Thus, for realism, an evaluation of the influence of hypnosis on C A relationships should include these instructions together with a hypnotic induction procedure. However, such instructions themselves form the components of a procedure called the cognitive interview (Geiselman, Fisher, Firstenberg, Hutton, Sullivan, Avetissian & Prosk, 1984) that can have a positive effect on eyewitness performance (Fisher, 1995; Fisher & Geiselman, 1992; Memon & Bull, 1991). The first control group therefore was a group given identical instructions to the hypnosis group but without any induction procedure or mention of hypnosis. Nevertheless, in standard police interviews the above memory enhancement instructions are not routinely used (Fisher, Geiselman & Raymond, 1987; George, 1990). Therefore, the performance of subject-eyewitnesses in a control group using these memory enhancement instructions may overestimate that achievable using a standard police interview (Kebbell & Wagstaff, 1996). Consequently, a further standard control group was included in which subjects were given minimal instructions to enhance memory. METHOD Hypnosis and confidence 159 Subjects Fifty-three subjects (34 F; 19 M) of various backgrounds participated. Their mean age was years (range 18 52, SD = 6.19). Materials and Procedure Subjects were randomly assigned to one of three conditions; hypnosis (N = 18), cognitive interview (N = 17) or standard (N = 16). Subjects were tested either individually or in groups of up to five. They were shown a 5.5 min black and white video film that concerned the implied murder of a male by a female. After the film subjects were given a face-recognition filler task that lasted approximately 10 minutes. Once the filler task had been completed subjects received one of the three following procedures. Hypnosis condition. Subjects in the hypnosis condition were given the following information: One method that the Police use that may improve memory is hypnosis. I will now use a hypnotic procedure. These subjects were then played a 7.5 min taped hypnotic induction procedure. This was a slightly modified version of the induction

5 160 Kebbell and Wagstaff procedure provided by Barber (1969, pp ) for use with the Barber Suggestibility Scale. Subjects were required to report their state of hypnotic depth according to the modified Long Stanford Scale (LSS) (Tart, 1970). The LSS was modified to enable subjects to write down their answers (see Wagstaff & Ovenden, 1979). After the hypnotic induction subjects were given four taped instructions for memory facilitation, which took 2 minutes. These were based on the four cognitive interview mnemonics used by Geiselman et al. (1984), though as mentioned earlier, similar instructions have been used by hypno-investigators. (References to photographs in these instructions refer to the previously mentioned face-recognition filler task. None of these photographs concerned anyone shown on the film.) The instructions were as follows: Reinstate context. What I would like you to do now is to think carefully about the photographs and the film that you saw. Think about what you felt and what you thought when you looked at the photographs and the film. Think about how you were feeling at the time, and of your reactions to the photographs and to the film. Try to reinstate the context in your mind, of the physical environment in which you saw the photographs and the film, such as the location of objects and people in the room. Report everything. It is known that some witnesses hold back information because they are not sure about what they can remember or do not know if the information is relevant. However, you must try to report everything that you can. Recall the events in different orders. Most people remember details of an event in a certain order, from beginning to end. However, you should also try to recall the photographs and the film that you saw in a different order perhaps starting at the end of the period that you viewed the photographs or the end of the film and working backwards. Alternatively, you can start at a photograph or part of the film that you can remember particularly well and work either forward or backwards from there. Change perspectives. People who witness events sometimes try to remember events from somebody else s perspective, so try to remember what you would have seen if you were a different person viewing the photographs or what you would have felt and seen if you were one of the characters in the film. Subjects were then instructed to open their eyes, while remaining hypnotized, and to answer a 33-item questionnaire (as used by Kebbell et al., 1996) devised to test their recall of the film. Questions were open-ended (i.e., not multiple-choice), but subjects were required to provide an answer, even if this was only a guess. A forced-response task was to allow comparison with other work in this area (e.g., Kebbell et al., 1996; Smith, Kassin & Ellsworth, 1989). The questionnaire was devised by two experimenters who agreed on 11 questions in each of three categories of item difficulty: easy, medium or hard. To reduce the influence of being correct by chance, each question was devised such that a range of plausible answers was possible. For example, an easy question was What song was the woman singing? ; a medium difficulty question was What was on the dish next to the television set? ; and a hard question was What was behind the Tabasco sauce bottle in the kitchen?. After answering each question subjects were required to rate their confidence in their answer on a 10-point Likert scale, ranging from pure guess (1) to absolutely certain (10).

6 On completion of the questionnaire subjects were again required to attempt the face-recognition, filler task (again this was the face-recognition task not described here), this also took approximately 10 minutes. Subjects were then woken up by counting from five to one. They were then debriefed and thanked for their participation. Cognitive interview condition. Subjects in the cognitive interview condition were first given a reading filler activity designed to take the same time as the hypnotic induction procedure. The filler activity consisted of two reasonably interesting magazine articles. They were then given the following information, One method that the Police use that may improve memory is a cognitive interview. I will now use a cognitive interview. Subjects were then played the tape recording of the four memory mnemonics used for the hypnosis group. This was so that hypnosis per se would not be confounded with differences in memory facilitation instructions between the two groups. After these instructions subjects were required to complete the 33-item questionnaire. On completion of the questionnaire subjects were again required to attempt the face-recognition, filler task. Control condition. The procedure for the control condition was identical to that of the cognitive interview condition but without the mnemonic instructions. The reading filler activity was increased in length so that there was a similar delay between stimuli presentation and memory testing as in the other groups. After the filler activity subjects were instructed as follows, One method that the Police use that may improve memory is asking witnesses to try hard. Please try hard to remember the film and the photographs. After this they were tested in the same way as the cognitive interview condition. RESULTS Hypnosis and confidence 161 The mean hypnotic depth score for the hypnosis condition was 3.06 (SD = 2.15), approximating to a report of a state slightly deeper than lightly hypnotized. The range of these scores, however, was from 0 to 6, ranging from not at all hypnotized, to quite deeply and strongly hypnotized (Tart, 1970). The questionnaire was analysed in a similar manner to that used by Kebbell et al. (1996). A one-way ANOVA (control/cognitive interview/hypnosis) was conducted for correct answers, this was not significant. The means and standard deviations were as follows: control condition M = (SD = 2.55); cognitive interview condition M = (SD = 2.78); and hypnosis condition M = (SD = 2.70). C A correlations were calculated for each subject across the 33 questions, producing 51 within-subjects correlations (for a discussion of the relative importance of the various correlations that follow see Kebbell et al., 1996; and, Smith et al., 1989). As with previous work in this area, a Fisher s z score was calculated for each subject s correlation. Average within-subjects C A correlations were calculated for the different interview conditions and these were also transformed into z scores and the average z score tested against zero. The averages of these correlations and their levels of significance (based on the converted z scores) are shown in Table 1. As can be seen in Table 1, all the within-subjects C A correlations are significant. To assess the effects of interview condition a one-way ANOVA was conducted on the within-subjects z scores across interview conditions (control/cognitive interview/hypnosis). This was not significant.

7 162 Kebbell and Wagstaff Each subject s average accuracy was correlated with his or her average confidence rating. These average correlations are also shown in Table 1. All three correlations are significant. Table 1. Within-subjects C A correlations, each subject s average confidence correlated with his or her average accuracy, between-subjects C A correlations and the average confidence expressed in a question correlated with average accuracy for that question for interview condition Interview condition Control Cognitive interview Hypnosis Within-subjects 0.84** 0.83** 0.79* (0.8) (0.6) (0.08) N = 16 N = 17 N = 18 Within-subjects (average) 0.69** 0.78** 0.63** N = 16 N = 17 N = 18 Between-subjects (0.39) (0.44) (0.36) N = 20 N = 18 N = 24 Between-subjects (average) 0.97*** 0.96*** 0.96*** N = 33 (items) N = 33 (items) N = 33 (items) Standard deviations are in brackets. *p < 0.05, **p < 0.01, ***p < A C A relationship was also then calculated across-subjects ( between-subjects ) for each of the 33 questions. These correlations were again averaged, converted to z scores and tested for significance against zero for each interview condition. These correlations and significance levels for the z scores are shown in Table 1. None was significant. The average number of correct answers and average confidence score for each question was correlated for the three interview conditions; the results are displayed in Table 1. All three correlations are significant. Average confidence in correct answers was compared with average confidence in incorrect answers with a 3 2 ANOVA (Control/Cognitive interview/hypnosis Confidence in correct/incorrect answers) with repeated measures on the second factor. Means and standard deviations are shown in Table 2. This analysis indicated that subjects showed higher confidence in correct answers than incorrect answers, F(1,47) = , p < There was no effect of interview condition and no interaction between confidence in correct/incorrect answers and interview condition. Absolutely certain responses were considered independently. In the control condition subjects made 141 absolutely certain responses of which % were correct; for the cognitive interview condition 175 absolutely certain responses were made of which % were correct; and, for the hypnosis conditions 162 absolutely certain responses were made of which % were correct. Clearly there were no obvious differences between the conditions in this respect.

8 Hypnosis and confidence 163 Table 2. Average confidence expressed in correct and incorrect answers (on a 10- point Likert scale) with respect to interview condition and question difficulty Correct/incorrect Interview condition Control Cognitive interview Hypnosis Correct (0.60) (0.88) (1.27) N = 16 N = 17 N = 18 Incorrect (0.83) (0.60) (0.55) N = 16 N = 17 N = 18 Standard deviations are in brackets For the hypnosis group correlations were calculated between hypnotic depth and the number of correct answers, number of incorrect answers, confidence in correct answers, confidence in incorrect answers and within-subjects C A correlations. None was significant (p > 0.1 in all cases). DISCUSSION No significant differences were found between interview conditions in terms of correct answers (the effects for incorrect answers mirror those for correct answers because the paradigm was forced-choice). The failure to find any differences in correct recall between groups fits in with much previous work on hypnosis (Sheehan, 1988; Wagstaff, 1993), but appears to conflict with previous results regarding the cognitive interview; as most studies using cognitive interview procedures have shown memory enhancements (for a review see Fisher, 1995; Fisher & Geiselman, 1992; Memon & Bull, 1991). This discrepancy may be due to the use of a questionnaire and group testing, although previous experiments have shown the cognitive interview to be effective in such situations (Geiselman et al., 1984). However, the experiment presented here differed from previous studies in not physically changing the context, which might enable the context reinstatement mnemonic of the cognitive interview to show its effectiveness. When C A relationships were considered, however, considerable support was found for the view that, when questions that vary in difficulty are used, and thereby the probabilities of producing absolutely certain and pure guess responses are maximized, C A relationships are considerably higher than have been previously reported (e.g., Smith et al., 1989; Perfect, Watson & Wagstaff, 1993). These findings replicate those of by Kebbell et al. (1996) and indicate that, in general: Subjects are more confident about their correct answers than their incorrect answers. Subjects who express higher confidence in their answers are more likely to be accurate than subjects who expressed lower confidence. Questions that are given high confidence ratings are more likely to be answered accurately. When subjects are absolutely certain that an answer is correct they are almost always accurate.

9 164 Kebbell and Wagstaff Predictably also, non-significant correlations were found between subjects accuracy scores and confidence ratings for individual questions (with this particular analysis, floor and ceiling effects drastically reduce the number of correlations it is possible to calculate and decrease the homogeneity of the responses). However, the three interview conditions had virtually no effect whatsoever on these various C A measures. Thus, the adverse influence of hypnosis on C A relationships that might have been expected (e.g., Orne et al., 1984; Sheehan, 1988; Wagstaff, 1993) was not shown here. It should be pointed out, nevertheless, that no attempt was made in the present experiment to use deliberately misleading information, or to assess confabulation during free recall; that is, manipulations that might have maximized hypnotic memory distortion effects. Also, although the range of hypnotic depth scores was reasonable, none of the subjects produced hypnotic virtuoso scores on the LSS (Tart, 1970) (though none of the correlations between hypnotic depth and the other variables was significant). Also the subject-witnesses did not take part in one-to-one interactive interviews because we wished to ensure rigorous control of variables. The social situation created in interactive interviews may influence the recall of eyewitnesses and make them less or more confident in parts of their statements, depending the kind of pressure applied, and longer retention intervals may also change the C A relationship. The social situation may be particularly important for both hypnosis and the cognitive interview; indeed, recent versions of the latter emphasize manipulating the social situation to enhance recall (Fisher & Geiselman, 1992). Thus, future work might usefully consider the relationship between interview technique, C A relationships, question-difficulty and absence/presence of leading questions and/or misleading postevent information in more naturalistic settings. It should be pointed out, nevertheless, that, typically, studies that have been used to support the view that the addition of a hypnotic induction technique to an interviewing procedure renders that procedure unsafe, have not used naturalistic settings either. Thus although the present results cannot be assumed to support the view that hypnotic forensic interviewing is safe ; they do, however, suggest that it may be premature to assume that hypnosis will routinely have adverse effects on C A relationships when no attempts are made to deliberately mislead the subject. This may be important given that, in more naturalistic field studies conducted in both the USA and the UK, using police interviewers, there is little evidence that interviewers attempt to mislead the witness when their interviews are recorded (Fisher et al., 1989). ACKNOWLEDGEMENTS The authors would like to thank the Economic and Social Research Council for funding this project (Award No. R ). REFERENCES Barber, T.X. (1969). Hypnosis: A Scientific Approach. New York: Pergamon. Bothwell, R.C., Deffenbacher, K.A. & Brigham, J.C. (1987). Correlation of eyewitness accuracy and confidence: Optimality hypothesis revisited. Journal of Applied Psychology 72, Christianson, S.A. & Hubinette, B. (1993). Hands up! A study of witnesses emotional reactions and memories associated with bank robberies. Applied Cognitive Psychology 7, Cutler, B.L., Penrod, S.D. & Dexter, H.R. (1990). Juror sensitivity to eyewitness identification evidence. Law and Human Behavior 14, Cutler, B.L., Penrod, S.D. & Stuve, T.E. (1988). Juror decision making in eyewitness identification cases. Law and Human Behavior 12,

10 Hypnosis and confidence 165 Deffenbacher, K.A. (1980). Eyewitness accuracy and confidence. Law and Human Behavior 4, Diamond, B.L. (1988). Inherent problems in the use of pretrial hypnosis on a prospective witness. California Law Review 68, Dywan, J. & Bowers, K. (1983). The use of hypnosis to enhance recall. Science 22, Fisher, R.P. (1995). Interviewing victims and witnesses of crime. Psychology, Public Policy, and Law 1, Fisher, R. P. & Geiselman, R. E. (1992). Memory-Enhancing Techniques for Investigative Interviewing. Springfield: Charles C. Thomas. Fisher, R.P., Geiselman, R.E. & Amador, M. (1989). Field test of the cognitive interview: Enhancing the recollection of actual victims and witnesses of crime. Journal of Applied Psychology 74, Fisher, R.P., Geiselman, R.E. & Raymond, D.S. (1987). Critical analysis of police interview techniques. Journal of Police Science & Administration 15, Fox, S.G. & Walters, H.A. (1986). The impact of general versus specific expert testimony and eyewitness confidence upon mock juror judgement. Law and Human Behavior 10, Fruzzetti, A.E., Tolland, K., Teller, S.A. & Loftus, E.F. (1992). Memory and eyewitness testimony. In M.M. Gruneberg & P.E. Morris (Eds) Aspects of Memory: The Practical Aspects. London: Academic Press, pp Geiselman, R.E., Fisher, R.P., Firstenberg, I., Hutton, L.A., Sullivan, S.J., Avetissian, I.V. & Prosk, A.L. (1984). Enhancement of eyewitness memory: An empirical evaluation of the cognitive Interview. Journal of Police Science and Administration 12, 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, George, R. (1990). A field and experimental evaluation of three methods of interviewing witnesses and victims of crime. Unpublished Masters Thesis, Polytechnic of East London, London, UK. Gudjonsson, G. (1992). The Psychology of Interrogations, Confessions and Testimony. Chichester: Wiley. Gruneberg, M.M. & Sykes, R.N. (1993). The generalisability of confidence accuracy studies in eyewitnessing. Memory 1, Hibbard, W.S. & Worring, R.W. (1981). Forensic Hypnosis: The Practical Application of Hypnosis in Criminal Investigation. Springfield, IL: C.C. Thomas. Kebbell, M.R. & Wagstaff, G.F. (1996). Enhancing the practicality of the cognitive interview in forensic situations. Psycholoquy 7(6), witness-memory.3.kebbell. Kebbell, M.R., Wagstaff, G.F. & Covey, J.A. (1996). The influence of item difficulty on the relationship between eyewitness confidence and accuracy. British Journal of Psychology 87, Leippe, M.R., Manion, A.P. & Romanczyk, A. (1992). Eyewitness persuasion: How and how well do fact finders judge the accuracy of adults and children s memory reports? Journal of Personality and Social Psychology 63, Lindsay, R.C.L. (1994). Expectations of eyewitness performance: Jurors verdicts do not follow from their beliefs. In D.F. Ross, J.D. Read & M.P. Toglia (Eds). Adult Eyewitness Testimony: Current Trends and Developments. New York: Cambridge University Press, pp Lindsay, R.C.L., Wells, G.L. & O Conner, F.J. (1989). Mock-juror belief or accurate and inaccurate eyewitnesses: A replication and extension. Law and Human Behavior 13, Luus, C.A. & Wells, G.L. (1994). The malleability of eyewitness confidence: Co-witness and perseverance effects. Journal of Applied Psychology 79, Memon, A. & Bull, R. (1991). The Cognitive Interview: its Origins, Empirical Support, Evaluation and Practical Implications. Journal of Community & Applied Social Psychology 1, Orne, M.T., Sokis, D.A., Dinges, D.F. & Orne, E.C. (1984). Hypnotically induced testimony. In G.L. Wells & E.F. Loftus (Eds), Eyewitness Testimony: Psychological Perspectives. Cambridge: Cambridge University Press, pp

11 166 Kebbell and Wagstaff Perfect, T.J., Watson, E.L. & Wagstaff, G.F. (1993). Accuracy of confidence ratings associated with general knowledge and eyewitness memory. Journal of Applied Psychology 78, Pike, G.E., Towell, N.A. & Kemp, R.I. (1995). Eyewitness testimony: Confidence of accuracy and accuracy of confidence. Paper presented to the Fifth Annual Conference of Psychology and Law (August), Budapest, Hungary. Rainer, D. (1983). Eyewitness testimony: Does hypnosis enhance accuracy, distortion and confidence? Unpublished PhD thesis, University of Wyoming, USA. Redson, M.T. & Knox, J. (1983). Is the recognition of faces enhanced by hypnosis? Paper presented to the Annual Meeting of the Society of Clinical and Experimental Hypnosis, Boston, USA. Reiser, M. (1980). Handbook of Investigative Hypnosis. Los Angeles, CA: Lehi. Reiser, M. (1990). Investigative hypnosis. In D.C. Raskin (Ed). Psychological Methods in Criminal Investigation and Evidence. New York: Springer, pp Sanders, G.S. & Simmons, W.L. (1983). Use of hypnosis to enhance eyewitness accuracy: Does it work? Journal of Applied Psychology 68, Sheehan, P.W. (1988). Confidence, memory and hypnosis. In H.M. Pettinati (Ed.), Hypnosis and Memory (pp ). New York: Guildford Press. Sheehan, P.W., Grigg, L. & McCann, T. (1984). Memory distortion following exposure to false information in hypnosis. Journal of Abnormal Psychology 93, 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, Smith, V.L., Kassin, S.M. & Ellsworth, P.C. (1989). Eyewitness accuracy and confidence: Within versus between-subjects correlations. Journal of Applied Psychology 74, Spanos, N.P. & Chaves, J.F. (1989). Hypnosis: The Cognitive Behavioural Perspective. New York: Guilford Press. Sporer, S.L., Penrod, S.D., Read, J.D. & Cutler, B.L. (1995). Choosing, confidence, and accuracy: A meta-analysis of the confidence accuracy relation in eyewitness identification studies. Psychological Bulletin 118, Tart, C.T. (1970). Self report scales of hypnotic depth. International Journal of Clinical and Experimental Hypnosis 18, Wagstaff, G.F. (1982). Recall of witnesses under hypnosis. Journal of the Forensic Science Society 22, Wagstaff, G.F. (1993). What expert witnesses can tell courts about hypnosis: A review of the association between hypnosis and the law. Expert Evidence 2, Wagstaff, G.F. & Ovenden, M. (1979). Hypnotic time distortion and free recall learning an attempted replication. Psychological Research 40, Wells, G.L. & Murray, M. (1984). Eyewitness confidence. In Wells G.L. & Loftus E.F. (Eds). Eyewitness Testimony: Psychological Perspectives. New York: Cambridge University Press, pp Zelig, M. & Beidleman, W.B. (1981). Investigative hypnosis: A word of caution. International Journal of Clinical and Experimental Hypnosis 29, Address for correspondence: Mark Kebbell, Department of Psychology, Bolton Institute, Deane Road, Bolton BL3 5AB, UK. mk1@bolton.ac.uk Received 3 October 1996; revised version accepted 9 April 1997

12 Contemporary Hypnosis (1997) Vol. 14, No. 3, pp BRIEF REPORT ASSESSING HYPNOTIZABILITY AND DISSOCIATION IN DIFFERENT CONTEXTS Peter V. Butler* and Richard A. Bryant University of New South Wales, Sydney, New South Wales, Australia ABSTRACT Subjects who had completed hypnotizability screening with the Harvard Group Scale of Hypnotic Susceptibility, Form A were subsequently administered the Dissociative Experiences Scale in an independent context. Half the subjects were provided with the standard DES instructions and half were provided with information that minimized the salience of dissociative experiences. HGSHS:A scores correlated significantly with both the standard DES (r = 0.31) and the DES-Taxon scale (r = 0.35). The modified instructions did not influence correlations with HGSHS:A scores. Findings are discussed in terms of contextual influences on DES responses. INTRODUCTION The resurgence of interest in dissociation has resulted in considerable debate about the most appropriate measure of dissociation (Kihlstrom, Glisky & Angiulo, 1994). The most common measure of dissociation is the Dissociative Experiences Scale (DES; Bernstein & Putnam, 1986). The DES is widely accepted as a standard measure of spontaneous dissociation in both clinical and non-clinical samples (Waller, Putnam & Carlson, 1996). Recent revision of the DES has developed the DES-Taxon (DES-T), an 8-item abridged version of the scale that purportedly measures pathological dissociation (Waller et al., 1996). The extent to which the DES indexes dissociative processes has typically been evaluated by correlating scores on the DES with those on standardized hypnotizability tests. There is divergent opinion, however, concerning the relationship between the DES and hypnotizability. Nine experimental studies have investigated the relationship between the DES and hypnotizability. Five of these have investigated the relationship within the context of hypnotizability testing (Faith & Ray, 1994; Frischholz, Braun, Sachs, Schwartz, Lewis, Shaeffer, Westergaard & Pasquotto, 1992; Segal & Lynn, ; Smyser & Baron, 1992; Di Tomasso & Routh, 1993). These studies administered batteries of self-report inventories, including the DES, immediately prior to administering a standard test of hypnotizability (either the Harvard Group Scale of Hypnotic Susceptibility, Form A (HGSHS:A) or the Stanford Hypnotic Susceptibility Scale, Form C (SHSS:C) (Weitzenhoffer & Hilgard, 1962). Three of these studies reported *Peter V. Butler is now at the Department of Behavioural Sciences in Medicine, University of Sydney, Sydney.

13 168 Butler and Bryant significant positive correlations between the DES and hypnotizability (Frischholz et al., 1992; Smyser & Baron, 1992; Di Tomasso & Routh, 1993), with coefficient values ranging from 0.12 to These findings generally have been interpreted as evidence of the validity of the DES as a measure of dissociative tendencies. This claim is inconsistent, however, with the findings of two studies that DES and hypnotizability scores were not significantly correlated when they were tested in the same context (Faith & Ray, 1994; Segal & Lynn, ). Four studies have attempted to investigate the relationship between the DES and hypnotizability outside the context of hypnotizability testing (Nadon, Hoyt, Register & Kihlstrom, 1991; Oakman, Woody & Bowers, 1996; Silva & Kirsch, 1992; Spanos, Arango & de Groot, 1993). Each of these studies reported non-significant correlations between hypnotizability and DES scores. The discrepant correlations reported between hypnotizability and DES have resulted in much debate over the extent to which contextual factors may influence the correlations between these two measures. Different theorists have proposed that the potential influence of contextual factors on the relationship between the two measures has been attributed to an expectancy effect concerning subsequent hypnotic responding, a need for self-consistency across tests, and a priming of awareness of certain dissociative experiences (for a review, see Oakman et al., 1996). The studies conducted to date have been confounded, however, by variable sample sizes, the different number of times the measures were administered, and varying order of administration of the measures. For example, whereas Oakman et al. (1996) administered the hypnotizability test prior to the DES in one condition, they reversed this order in their second condition. Such order effects clearly have an important bearing on interpretation of the causal influence of one measure on another (Council & Kirsch, 1996). The present study aimed to investigate the relationship between the DES and hypnotizability scores by administering each only once, maintaining a consistent order of administration, and assessing each measure in different contexts. Sociocognitive theorists maintain that increasing the ambiguity about the relationship between hypnotizability and personality measures will decrease the intercorrelations between these measures (Spanos et al., 1993). Accordingly, a second aim of the present study was to investigate the influence of manipulating the instructional set of the DES by minimizing the salience of the dissociative experiences described in the DES items. Specifically, we administered one form of the DES that included the original introduction by Bernstein and Putnam (1986) and another that explicitly informed subjects that the experiences described in the DES are common experiences that may happen in partial or minimal ways. It was predicted that minimizing the salience of the dissociative nature of the DES items would minimize the correlation between the DES and hypnotizability scores. METHOD Subjects Subjects were undergraduate psychology students who participated in return for research credit. Subjects were preselected on the basis of previous participation in the group-administered, 12-item HGSHS:A (Shor & Orne, 1962). Of the 338 subjects who participated in the HGSHS:A, 50 (39 females and 11 males) subjects of mean age years (SD = 5.11) also completed the DES:A, and 44 (35 F; 9 M) subjects of mean age years (SD = 6.66) also completed the DES:B. Subjects who completed the DES did not differ from those who did not complete the DES in terms of

14 hypnotizability, age, or gender. Subjects who completed the DES but did not participate in the HGSHS:A included 34 subjects who completed the DES:A and 34 who completed the DES:B. Procedure The DES:A and DES:B were distributed to subjects during class between one and two weeks following completion of the HGSHS:A. In order to minimize any association between the DES and hypnotizability testing, the DES was packaged together with a separate questionnaire on thought intrusions. Subjects were requested to complete the questionnaires at home and return them during their next class. The instructions for the DES:A were identical to those of Bernstein and Putnam s (1986) initial version of the DES. The instructions for the DES:B contained an additional paragraph that explicitly informed subjects that dissociative experiences were normal experiences that may occur to most people to some degree. The amended instructions were: Please note that not everyone will have experiences that exactly match the ones described in the questionnaire. We are interested in how often these experiences sort of happen to you. That is, you might find that you experience some of these things in a partial way. Please think about each item and consider if this sort of thing happens to you in any way at all. Remember, it may happen intensely or it may happen in a mild sort of way. We are interested in all sorts of experiences. RESULTS Dissociation and hypnozability 169 Table 1 presents the correlation coefficients of HSHS:A scores and DES:A and DES:B. Scores on the DES:A and HGSHS:A were significantly correlated, r (50) = 0.31, p < Scores on the DES:B and HGSHS:A were not significantly correlated, r (44) = 0.24, p = A comparison of these two correlations indicated that the difference was not significant, z = 0.35, p = Table 1 also presents intercorrelations with the DES-Taxon (DES-T). The DES-T is a recent subscale of the DES that comprises eight items that purportedly measure pathological dissociation (Waller et al., 1996). Table 1 indicates that the pattern of correlations between the HGSHS:A and DES-T were similar to those between the HGSHS:A and the DES; the correlation between the DES-T subscale and the HGSHS:A was significant for the DES:A [r (50) = 0.35, p < 0.05] but not for the DES:B [r (44) = 0.27, p = 0.07]. A comparison of these two correlations indicated that the difference was not significant, z = 0.37, p = 0.34 Table 1. Correlations of DES and HGSHS:A scores Form A Form B DES DES-T DES DES-T HGSHS:A Note. Form A = DES:A, Form B = DES:B.

15 170 Butler and Bryant The mean DES score for subjects who completed the DES:A (M = 18.18, SD = 13.54) was not significantly different from the mean score of those who completed the DES:B (M = 18.57, SD = 12.75). Similarly, the DES-T scores of subjects who completed the DES:A (M = 11.75, SD = 11.18) was not significantly different from the mean score of those who completed the DES:B (M = 13.15, SD = 13.01). Carlson and Putnam (1993) suggest that a score of at least 30 on the DES is indicative of potential dissociative psychopathology. Fourteen subjects (16.7%) who completed the DES:A and 20 (25.6%) who completed the DES:B scored at least 30. A Mann-Whitney test indicated that this difference was not significant. DISCUSSION The current finding represents the first report of a significant correlation between the DES and the HGSHS:A when administered in independent contexts. Further, the strength of the correlation (0.31) was markedly higher than most previous reported correlations with the DES (Spanos et al., 1993). We recognize that subjects may have perceived a connection between DES administration and hypnotizability testing. Considering the novelty of hypnotic responses for many undergraduate students, and the apparent similarity between some items on the HGSHS:A and some items described in the DES, it is plausible that participation in hypnotic testing may have cued subjects to respond in a hypnotic-like manner on the subsequently administered DES. Council and Kirsch (1996) have argued that contextual influences can be most parsimoniously explained in terms of subjects need to maintain consistency in their self-presentation. It is possible that our subjects responded to the DES in a manner that was consistent with their previous hypnotic performance so that their presentation appeared stable. Further, it is possible that recent hypnotic testing may have had a priming effect that directed subjects attention to experiences that are associated with items on the HGSHS:A (Roche & McConkey, 1990). The minimum period between test measures in the current study (one week) was shorter than in other studies that have separated the contexts of testing (Nadon et al., 1991). It is possible that extending the inter-test interval may have decreased the likelihood that subjects would have perceived that the two measures were related, and potentially decreased the correlation between the measures. Finally, we recognize that permitting our subjects to complete the DES at home and return the questionnaire a week later may involve different demands from studies that have required completion of the DES within a specified and more formal setting. The current finding needs to be replicated in studies with substantially increased sample sizes, and with designs that ensure that subjects do not perceive that the two assessments are related. Manipulating the instructional set for the DES did not affect the correlation between the DES and HGSHS:A. This finding is inconsistent with previous studies that have manipulated instructions to the DES. For example, Spanos and colleagues found that the DES and the Carleton University Responsiveness to Suggestion Scale (CURSS; Spanos, Radtke, Hodgins, Bertrand, Stam & Dubreuil, 1983) were significantly correlated when a link between the two measures was communicated in the preliminary instructions to the DES, but not when these instructions were omitted. It is possible that our instructional manipulation was too subtle and that the instructions that preceded DES:B did not communicate to subjects a sufficiently clear message concerning the desired perception of the DES items. Similarly, manipulation of instructional set did not influence the absolute level of responding on the DES.

16 Previous studies of contextual influences on DES responding have reported inconsistent effects on absolute DES scores. Whereas Nadon and colleagues (Nadon et al., 1991) reported differences in DES scores when tested in and out of the hypnotic context, others have found no differences (Oakman et al., 1996; Spanos et al., 1993). These patterns of findings suggest that contextual variables interact with a number of relevant factors to mediate the relationship between the DES and hypnotizability. The absence of an effect by our instructions suggests that future studies need to provide clear and directive communications to explore the impact of the instructional set on DES responding. In summary, the present finding provides initial evidence of a relationship between DES responses and hypnotizability when tested in independent contexts. We suggest that this finding be interpreted cautiously because the possibility of unexpected contextual influences mediating our subjects responses cannot be ruled out. There is a need to more rigorously investigate the relationship between these measures in ways that ensure that overt and covert links between measurement procedures are comprehensively masked, and that the contexts are truly independent. ACKNOWLEDGEMENTS Dissociation and hypnozability 171 This study was supported by an Australian Research Council Small Grant. We are grateful to Gladdis Warda, Fiona MacCallum, and Maree Abbott for assistance in data collection. REFERENCES Bernstein, E.M. & Putnam, F.W. (1986). Development, reliability, and validity of a dissociation scale. Journal of Nervous and Mental Disease 174, Carlson, E.B. & Putnam, J. (1993). An update on the Dissociative Experiences Scale. Dissociation 6, Council, J.R. & Kirsch, I. (1996). Explaining context effects: Expectancy and consistency. Contemporary Hypnosis 13, Di Tomasso, M.J. & Routh, D.K. (1993). Recall of abuse in childhood and three measures of dissociation. Child Abuse & Neglect 17, Faith, M. & Ray, W.J. (1994). hypnotizability and dissociation in a college age population: Orthogonal individual differences. Personality and Individual Differences 17, Frischholz, E.J., Braun, B.G., Sachs, R.G., Schwartz, D.R., Lewis, J., Shaeffer, D., Westergaard, C. & Pasquotto, M.A. (1992). Construct validity of the Dissociative Experiences Scale: II. Its relationship to hypnotizability. American Journal of Clinical Hypnosis 35, Kihlstrom, J.F., Glisky, M.L. & Angiulo, M.J. (1994). Dissociative tendencies and dissociative disorders. Journal of Abnormal Psychology 103, Nadon, R., Hoyt, I.P., Register, P.A. & Kihlstrom, J.F. (1991). Absorption and hypnotizability: Context effects reexamined. Journal of Personality and Social Psychology 60, Oakman, J.M., Woody, E.Z. & Bowers, K.S. (1996). Contextual influences on the relationship between absorption and hypnotic ability. Contemporary Hypnosis 13, Roche, S. & McConkey, K.M. (1990). Absorption: Nature, assessment, and correlates. Journal of Personality and Social Psychology 59, Segal, D. & Lynn, S.J. ( ). Predicting dissociative experiences: Imagination, hypnotizability, psychopathology, and alcohol use. Imagination, Cognition and Personality 12, Shor, R.E. & Orne, M.T. (1962). Harvard Group Scale of Hypnotic Susceptibility, Form A. Palo Alto, CA: Consulting Psychologists Press.

17 172 Butler and Bryant Silva, C.E. & Kirsch, I. (1992). Interpretative sets, expectancy, fantasy proneness, and dissociation as predictors of hypnotic response. Journal of Personality and Social Psychology 63, Smyser, C.H. & Baron, D.A. (1993). hypnotizability, absorption, and subscales of the Dissociative Experiences Scale in a nonclinical population. Dissociation 6, Spanos, N.P., Arango, M. & de Groot, H.P. (1993). Context as a moderator in relationships between attribute variables and hypnotizability. Personality and Social Psychology Bulletin 19, Spanos, N.P., Radtke, H.L., Hodgins, D.C., Bertrand, L.D., Stam, H.J. & Dubreuil, D.L. (1983). The Carleton University Responsiveness to Suggestion Scale: Stability, reliability, and relationships with expectancy and hypnotic experiences. Psychological Reports 53, Waller, N.G., Putnam, F.W. & Carlson, E.B. (1996). Types of dissociation and dissociative types: A taxometric analysis of dissociative experiences. Psychological Methods 1, Weitzenhoffer, A.M. & Hilgard, E.R. (1962). Stanford hypnotic susceptibility scale: Form C. Palo Alto, CA: Consulting Psychologists Press. Address for correspondence: Richard A. Bryant, School of Psychology, University of New South Wales, NSW 2052, Australia. . r.bryant@unsw.edu.au Received 16 July 1996; revised version accepted 13 March 1997