Self-fulfilling prophecy and hypnotic response are not the same thing

Journal of Personality and Social Psychology Copyright 1998 by the American Psychological Association, Inc. 1998, Vol. 75, No. 6, 1604 1613 0022-3514/98/$3.00

Self-Fulfilling Prophecy and Hypnotic Response Are Not the Same Thing

Grant Benham, Shannon Bowers, Michael Nash, and Robert Muenchen University of Tennessee, Knoxville

The authors used the "bogus-item" methodology originally used by C. Wickless and I. Kirsch (1989) to examine the effects of response expectancy manipulations on subjects' subsequently measured hypnotizability scores. The results of the first experiment failed to replicate Wickless and Kirsch's (1989) findings that surreptitious confirmation of suggested items (the bogus-item manipulation) leads to higher scores on subsequently measured hypnotic responsiveness. The second experiment tested if, in fact, response expectancies were enhanced by the bogus-item manipulation, and the manipulation's effect on behavioral and subjective responsiveness to hypnosis was reexamined. A significant increase was found in response expectancies as a result of the bogus item expectancy manipulation, but again no evidence that this manipulation led to increased hypnotic responsiveness was found. Hypnotic responsiveness may not be as reactive to expectancy manipulations as previously claimed, and it is suggested that the notion of hypnotic responsiveness as a traitlike ability is viable.

Broadly sketched, social learning and contextualist theories minimize the importance of traitlike predispositions and cham- pion the influence of response expectations and reciprocal role interactions. Even when stability in performance obtains and change across contexts is quite minimal, it is difficult to empirically test to what extent this stability is due to powerfully enduring social influences or traitlike characteristics. The experimenter is faced with two challenges: (a) identifying a behavior that is both stable across contexts and at the same time novel enough to be relatively unburdened by preexisting expectational s e t s - - i f this condition is met, it becomes possible to test if proximal manipulation of response expectations affects the level of performance and establishes stability across time; and (b) identifying a behavior that is both stable across contexts and has two distinct theories accounting for this stability, one trait- based and the other based on social learning. If this condition is met, empirical work can plausibly begin to unravel the relative merit of each theory regarding the source of the stability. Individ- ual differences in hypnotic responsiveness are well suited on both counts.

Most researchers, regardless of theoretical perspective, agree that the degree to which a person responds to hypnosis is re- markably stable over time. Perhaps the most compelling illustra- tion of this stability is the long-term follow-up study in which the Stanford Hypnotic Susceptibility Scale (SHSS), Form A, was readministered to participants 10 to 25 years after it was first administered to them (Piccione, Hilgard, & Zimbardo, 1989). Overall, the test-retest reliability for hypnotic susceptibility was .71. This kind of stability compares favorably with test-retest reliabilities for IQ tests over similar periods of time. 1

Grant Benham, Shannon Bowers, and Michael Nash, Department of Psychology, University of Tennessee, Knoxville; Robert Muenchen, Sta- tistical Consulting Center, University of Tennessee, Knoxville.

Correspondence concerning this article should be addressed to Mi- chael Nash, Department of Psychology, University of Tennessee, Knox- ville, Tennessee 37996. Electronic mail may be sent to [email protected].

There remains, however, considerable disagreement about why this stability obtains: Does this stability reflect an enduring cognitive ability or personality trait of "hypnotic susceptibility," or is it a product of response expectations that consolidate at first testing? According to the trait view, people vary (roughly normally) in their capacity for experiencing hypnosis; individual differences in response to hypnosis are then simply a reflection of people's differing on a cognitive processing ability we call hypnotic susceptibility (e.g., Hilgard, 1965; Piccione et al., 1989). The nature of this cognitive ability is elusive, but it may have its foundation in genetics (heritability index h 2 of .64; Morgan, 1973, p. 58), cognitive processing (Tellegen & Atkin- son, 1974), and brain physiology (Graffin, Ray, & Lundy, 1995). In contrast, response expectancy theory (Kirsch, 1985) proposes that individual differences in hypnotizability are primarily the result of differences in the participant's response expectations, and that the temporal stability of hypnotic performance reflects enduring expectancies rather than a fixed personality trait. Proponents of the response expectancy theory contend that a person's belief about his or her hypnotic ability is the primary determinant of how many suggestions he or she responds to successfully. According to this theory, if naive participants are exposed to procedures that increase their expectations of how well they will do in hypnosis, they will in fact score higher than controls will on a measure of hypnotic susceptibility when subsequently tested. Whether individual differences in hypnotizability reflect the operation of a stable trait or are instead a function of response expectancies also has important clinical implications. As Kirsch states, "I t is possible that, with sufficiently strong response expectancies, all individuals would show high levels of hypnotic response" (1985, p. 1196).

The question of whether manipulation of proximal response expectancies might affect hypnotic responsiveness was originally studied by Wickless (1987) and was later reported by

1 A 13-year test-retest reliability for the Wechsler Adult Intelligence Scale IQ test is reported as .73 (Kangas & Bradway, 1971).

1604

SELF-FULFILLING PROPHECY AND HYPNOTIC RESPONSE 1605

Wickless and Kirsch (1989) in an influential article indexing the effects of various expectancy manipulations on subsequent hypnotic susceptibility scores. In a study extending an early study not involving hypnosis (Wilson, 1967), Wickless and Kirsch used a 2 x 2 design (Bogus Experiential Manipulation × Verbal Manipulation), randomly assigning 60 participants to one of four cells (15 participants in each cell). The bogus experiential manipulation and the verbal manipulation were de- signed to fool participants into believing they were good hypnotic responders. The four groups were as follows: (a) The verbal- manipulation group received a verbal expectancy manipulation prior to hypnotic induction, where false feedback from three filler personality tests was used, and the participants were told that on the basis of their test results, they had been identified as highly hypnotizable; (b) the bogus-item group received an experiential expectancy manipulation where, following the hypnotic induction, they were given various bogus suggestions for visual phenomena (such as the room taking on a rosy glow, becoming greener and greener, or becoming dark) while at the same time these suggestions were surreptitiously confirmed by means of a hidden lighting arrangement (i.e., the room was faintly illuminated with the suggested color); (c) the combined group was exposed to both the verbal and the bogus item expectancy manipulations; and (d) the control group received neither expectancy manipulation.

Following the manipulations, all of the participants in the Wickless and Kitsch (1989) study were assessed for hypnotic responsiveness using the SHSS, Form C (SHSS:C; Weitzenhof- fer & Hilgard, 1962) for the first time (Time 1 ). In an attempt to demonstrate that the enhancement of hypnotic susceptibility would not deteriorate over time, Wickless and Kirsch first debriefed the participants in the bogus-item group and the combined group regarding the manipulations (i.e., they were told about the bogus items), and then they readministered the standard SHSS:C (without any manipulation) during a second testing session (Time 2) approximately 2 weeks later.

A 2 x 2 analysis of variance (ANOVA) was carried out on the SHSS:C scores from Time 1 (the first testing session). There was a significant main effect for the bogus experiential manipulation, with those who received the bogus experiential manipulation scoring higher than those who did not. There was not a significant main effect for verbal manipulation, nor was there a significant interaction (Wickless, 1987, p. 26). Paren- thetically, for the bogus-item group, debriefing immediately before Time 2 "produced" (Wickless & Kirsch, 1989, p. 766) a significant increase in hypnotizability scores at Time 2 testing. That is, debriefed subjects actually scored higher on the second administration of the scale after having been informed that they had been fooled about their level of hypnotizability during the first administration. Thus, not only did Time 1 scores not deteriorate, they increased for this group.

Citing the main effect of bogus experiential manipulation for Time 1 hypnotic susceptibility scores, Wickless and Kirsch (1989) interpreted their findings as supporting the notion that hypnotic susceptibility is primarily determined by response expectancy prior to and during the initial hypnosis experience.

The Wickless and Kirsch (1989) study has played a pivotal role in challenging trait conceptions of hypnotic responsiveness, supporting instead a sociocognitive explanation of why hypnotic

susceptibility is so stable over time. Dropping the verbal manipulation and using only the bogus experiential manipulation, Wickless, Kirsch, and Moffitt (in press) essentially replicated these findings. However, both studies involve important conceptual and methodological issues that have yet to be fully addressed. We present two experiments that further investigate the role of expectations in determining hypnotic responsiveness. The first is a replication and extension of the Wickless and Kirsch (1989) study on the effects of the bogus-item manipulation. In the second experiment we tested if, in fact, response expectancies were enhanced by the bogus-item manipulation and reexamined the manipulation's effect on behavioral and subjective responsiveness to hypnosis.

Throughout Experiments 1 and 2 we used alpha levels and analyses that increased the chances of inferring a difference when such a difference really existed. First, we used one-tailed tests of significance when examining a priori hypotheses derived from response expectancy theory. We did this for four related reasons: (a) The inferences derived from the theory are in fact unidirectional (e.g., higher expectations of success in hypnosis lead to higher levels of hypnotic response), (b) the results of the Wickless and Kirsch (1989) study indeed demonstrated this directionality predicted by theory, (c) the theory would be dis- confirmed by differences in the opposite direction, and (d) we wished to give the theory as generous a test as possible by reducing the chances of Type II error. Further, we relied on standard parametric statistics (ANOVA) for three reasons: First, because this was the a priori method of comparison in the original study (Wickless, 1987); second, because this was the method used in the Wickless, Kirsch, and Moffitt replication (in press); and third, because an ANOVA is indeed the most powerful statistic under these circumstances and as such opti- mally reduces the chances of Type II error.

Finally, we investigated the effects of the bogus experiential manipulation and not the verbal manipulation for three reasons: (a) No significant main effect or interaction was obtained in the original study (Wickless, 1987; Wickless & Kirsch 1989), (b) in the replication (Wickless et al., in press) the authors themselves deleted the verbal manipulation, and (c) previous work on false verbal feedback is equivocal at best (Goebel & Stewart, 1971; Gregory & Diamond, 1973; Saavedra & Miller, 1983).

Experiment 1

Our first hypothesis, then, was that participants who received the bogus experiential manipulation in our study would have higher hypnotizability scores at Time 1 than would the controls who experienced no manipulation.

We found it rather perplexing that the bogus-item group in the Wickless and Kirsch (1989) study showed an increase in SHSS:C scores from Time 1 to Time 2, a finding consistent with neither trait nor response expectancy theory. For this reason, we also chose to investigate whether the debriefing itself had any effect on hypnotizability scores. We wondered if the increase in hypnotizability scores from Time 1 to Time 2 for these participants was a reaction to having been explicitly told that they were fooled by the experimenters. Perhaps the participants be- lieved that it was not because the experimenters had fooled them

1606 BENHAM, BOWERS, NASH, AND MUENCHEN

that they scored highly, but it was because they really were highly hypnotizable. Obtaining a lower score at Time 2 would create a conflict between their belief about being highly hypnotizable and their overt behavior. One way to avoid this conflict, as suggested by cognitive dissonance theory (Festinger, 1957), would be to behave in a manner that was consistent with one 's beliefs (i.e., to score high on the second administration of the SHSS:C) . In doing so, participants may have overcompensated, resulting in the increased hypnotic susceptibility scores at Time 2.

To test this idea in the current study, we used two bogus- item groups. The first group (the bogus debriefed group) was debriefed immediately after Time 1. 2 The second group (the bogus nondebriefed group) was not debriefed until after Time 2. Our second hypothesis, then, was that debriefed participants would evidence a significant increase in hypnotic responsiveness from Time 1 to Time 2, and that this increase would not be evident among nondebriefed participants.

Method

Participants

Participants were 10 male and 37 female undergraduate students from the University of Tennessee, Knoxville. Participant numbers were targeted around those of the Wickless and Kirsch (1989) study, which had 15 participants in each group. Approximately two thirds of the participants were enrolled in an introductory psychology course, and the re- mainder were enrolled in an abnormal psychology class. All participants received course credit in exchange for their participation. Only those reporting no previous hypnotic experience were accepted for this study. Participants were randomly assigned to one of three experimental conditions: (1) bogus debriefed, (2) bogus nondebriefed, and (3) control.

Measures

Three dependent measures were used for all participants at both Time 1 and Time 2.

Behavioral measure of hypnotic responsiveness. A behavioral index of hypnotic susceptibility, the SHSS:C (Weitzenhoffer & Hilgard, 1962), was used as per Wickless and Kitsch (1989).

A subjective measure of hypnotic responsiveness. A subjective rating of the vividness of each hypnotic suggestion (vividness index) was obtained from participants following termination of the hypnotic session. The vividness index, developed by Wickless and Kirsch (1989), is a Likert scale for each of the 5 bogus items and each of the 12 SHSS:C test items. For each item, the scale ranged from 1 to 5: A rating of 1 represented not experiencing the suggested effect at all, and 5 represented experiencing the effect as vividly as if it were actually happening. Participants in the control group were instructed to ignore the first five items on the scale (these items related to the bogus suggestions, which controls did not receive).

Subjective depth. A subjective measure of hypnotic depth was used to assess how deeply hypnotized participants felt themselves to be at various times during the hypnotic session. Participants were instructed to report a number between 1 and 10 to indicate how deeply hypnotized they felt themselves to be. A rating of 1 corresponded with feeling wide awake, and a rating of 10 corresponded with feeling very deeply hypnotized. The subjective depth probe was given at three points during the administration of the bogus items and at five points during the administration of the SHSS:C (at both Time 1 and Time 2). Such subjective measures of hypnotic depth typically correlate about .76 with behavioral measures of hypnotizability (e.g., Tart, 1979).

Procedure

T/me 1. One at a time, participants in the bogus debriefed group were seated in a comfortable chair facing the comer of the room, with the experimenter seated slightly behind them and to the left. The room was dimly lit by two 15-watt bulbs, had off-white walls, and was arranged so that no shadows were produced on the section of wall visible to participants. 3 A panel of lights (described in detail later) was installed in the wall behind the participant and hidden from view by a corkboard. The participant was instructed to close his or her eyes and to find a comfortable position, at which time the corkboard was silently removed by the experimenter (without the participants' knowledge) under the guise of turning on a white-noise generator. The participant was then instructed to open his or her eyes and look straight ahead, at which time the SHSS:C induction was begun.

Immediately following the induction, participants were given their first probe for hypnotic depth. Participants then received the bogus item expectancy manipulation. As per Wickless and If, inch (1989), the bogus items consisted of five suggestions for visual phenomena (seeing the color red, seeing the color green, seeing a flickering light, seeing dark, and seeing the color blue). Each suggestion was Confirmed by surreptitious alteration of the lighting conditions in the experimental room, using 25-watt red, green, and blue light bulbs and a 15-watt white bulb. All bulbs were mounted in the lighting unit behind the participant and controlled by a second experimenter in an adjacent room by means of rheostats. The second experimenter was able to listen to the ongoing procedure by means of a small microphone placed in the ceiling just behind the participant. During the suggestion to see a particular color, the intensity of the corresponding light bulb was gradually increased to a level previously determined by extensive piloting. Following Wickless and Kirsch's (1989) protocol, we used the levels that most pilot subjects reported as being perceptible but not distinguishable from imagined effects. The flickering-light effect was produced by rapidly increasing and decreasing the intensity of the white light, with maximum and mini- mum levels determined, as before, with piloting. Darkening of the room was produced by gradually reducing the intensity of the white light. 4 At three separate times during the presentation of the bogus items, participants were probed for hypnotic depth (immediately before the first and fourth bogus item and immediately following the fifth bogus item).

2In the Wicldess and Kitsch (1989) study, these participants were debriefed not after the first session (Time 1), but immediately before the second session (Time 2). Though this deviation from the original procedure has little impact on the issue of the effects of debriefing, it is worthy of note (see our General Discussion).

3In the Wickless and Kitsch (1989) study the walls were covered with black sheets to mask any shadows cast by the lights. In our pilot studies we found that the sheets absorbed so much of the light that individuals were unable to perceive any colors at all, even with the lights at maximum intensity. Because of this, the sheets were not used.

4 The procedures for the administration of these five bogus items as set forth by Wickless and Kirsch (1989) were followed as closely as possible. In their study, however, a sixth suggestion (bearing unusual music) was given as part of the bogus-item manipulation and was accom- panied by a 10-s musical excerpt, faintly audible to participants. Our piloting studies found this manipulation was unconvincing. Pilot participants reported that when the suggestion to hear music was given, they imagined one type of music and heard (however faintly) another type. This incongruity raised suspicion. In light of this, and because the senior author of the original study suggested to us that it be left out (Irving Kitsch, personal communication, 1995), the bogus item for music was dropped in our replication. In all other respects, the mechanical, pr9ce- dural, and piloting specifications described in the original study were followed meticulously.


Following the .bogus items, participants were administered the 12 SHSS:C items. During the administration of these items, participants were probed for hypnotic depth five times (after the first, third, fifth, seventh, and ninth SHSS:C items). Upon termination of the SHSS:C, participants were given the vividness index to fill out. Once completed, participants in the bogus debriefed group were debriefed using the word- ing reported in Wickless and Kirsch (1989):

We tried to help you become hypnotized by making sure that you would have the first few experiences that I suggested to you. Re- member when I told you to see colors on the wall? Whenever I said to imagine a color, we turned on a colored light that made the room look a tiny bit that color. When I told you to imagine a flickering light and the room getting dark, we used the lights to make those things happen too. But we did that only for the colored lights, the flickering, and the dark. Everything else you did entirely on your own, and you did very well.

Table 1 Means and Standard Deviations for Behavioral and Subjective Indexes of Hypnotic Responsiveness

Group

Bogus item collapsed Control (n = 32) (n = 15)

Hypnotic responsiveness M SD M SD

Behavioral (SHSS:C) 7.34 2.31 7.07 3.77 Subjective (vividness of

SHSS:C items) 3.22 0.81 3.15 0.78

Note. SHSS:C = Stanford Hypnotic Susceptibility Scale, Form C.

Arrangements were then made for the participant to return for the second testing session (Time 2) 2 weeks later.

For participants in the bogus nondebriefed group, all procedures before debriefing were identical to those of the bogus debriefed group. Upon completion of the vividness index, however, participants in the bogus nondebriefed group were not debriefed about the bogus-item manipulation. An appointment was made for the participants to return for Time 2 testing.

Participants in the control group were administered the standard SHSS:C without any expectancy manipulation. Hypnotic depth reports were probed during the 12 SHSS:C items as described above. The vividness index was completed after termination of hypnosis (control participants were instructed to ignore the first five [bogus] items on the vividness index). Appointments were then made for the Time 2 session.

lime 2. Two weeks after Time I, all groups returned for retesting. At Time 2, participants in all groups were administered the SHSS:C without any bogus-item suggestions and were probed again for their hypnotic depth five times during the SHSS:C. Following termination of hypnosis, all participants completed the vividness index and were debriefed as needed.

Results

Time 1 Analyses

The two primary dependent variables in this study were behavioral responsiveness to hypnosis (the SHSS:C score, possible range 0 - 1 2 ) and the vividness index score (the mean of the 12 vividness index scores for SHSS:C items, possible range 1 - 5 ) . The vividness index scores (for SHSS:C items) were significantly correlated with behavioral scores ( r = .53, p < .001 ). This correlation is in keeping with previous reports of the relationship between experiential and behavioral measures (e.g., Tart, 1979, p. 584, reported the correlation between SHSS:C experiential scores and SHSS:C behavioral scores as .77).

Because the Time 1 procedure was identical for both the bogus debriefed and the bogus nondebriefed participants, results from these two groups were collapsed for Time 1 analysis. Means and standard deviations for the SHSS:C behavioral scores and the mean vividness index scores (for SHSS:C items) are presented in Table 1.

One-way ANOVAs on these data revealed that the bogus item manipulation produced no significant effects on SHSS:C behavioral scores or vividness index scores, F ( 1, 45) = 0.097, p =

.757, and F ( 1, 45) = 0.086, p = .771, respectively. 5 As stated above, the most appropriate analysis for these data is a parametric that exploits the interval properties of the SHSS:C. Neverthe- less, we also carried out a 3 × 2 chi-square analysis with control and bogns-item participants divided into categories of low hypnotizability (SHSS:C scores of 0 - 4 ) , medium hypnotizability (SHSS:C scores of 5 - 8 ) , and high hypnotizability (SHSS:C scores of 9 - 1 2 ) . The chi-square was nonsignificant, X2(4, N = 47) = 3.57,p = .167. I tmust further be noted that two of the cells had expected counts below 5 (under these circumstances a chi-square may overestimate the degree of relationship).

Supplemental Analyses o f Time 1 Data

In addition to the primary dependent variables used in the Wickless and Kirsch (1989) study, we also examined the subjective depth score (the mean of the five probes for hypnotic depth during the SHSS:C, possible range 1 - 1 0 ) . The mean depth scores were moderately correlated with both the vividness index scores and the behavioral scores for the SHSS:C ( r = .35, p = .017, and r = .37, p = .010, respectively). A one-way ANOVA showed that the mean depth scores for the collapsed bogus-item groups (M = 6.13, SD = 1.91 ) were not significantly different from those of the control group (M = 5.85, SD = 1.76), F(1 , 45) = 0.2266, p = .636, demonstrating that the bogns-i tem manipulation had no significant effect on subjective levels of hypnotic depth.

For the two groups receiving bogus items, we also examined the average vividness index scores for bogus items. Though somewhat attenuated, the relationship between vividness reported during bogus-item administration (the mean of participants' ratings for the five bogus items, possible range 1 - 5 ) seemed related to behavioral responsiveness, subjective depth, and vividness scores obtained during the SHSS:C administration

5 We also analyzed these data more conservatively. The SHSS:C behavioral scores and vividness index scores were initially examined using a multivariate analysis of variance. This analysis failed to reach significance (p < .94), but we proceeded with one-way ANOVAs so as to enable a direct comparison with the Wickless and Kitsch (1989) results.


(r = .316, p = .078; r = .344, p = .054; and r = .290, p = .107, respectively).

Results of Retesting

At Time 1, participants in the bogus debriefed group were debriefed and invited to return for retesting. Participants in the bogus nondebriefed group were also asked to retum for retesting but were not debriefed until after Time 2. Eleven participants failed to arrive for Time 2 testing (4 controls, 4 bogus debriefed, and 3 bogus nondebriefed). 6 Means and standard deviations of the Time 1 and Time 2 SHSS:C scores of the 36 participants who were retested are presented in Table 2.

Comparisons of Time 1 behavioral and vividness means for these 36 returning participants (see Table 2) with the Time 1 means of all 47 participants tested at Time 1 (see Table 1) revealed no significant difference, suggesting that the participants who returned for retesting were representative of the larger groups from which they were drawn.

A 3 × 2 (Group x Time) repeated measures ANOVA on SHSS:C behavioral scores failed to show any significant effects of time, group, or Group x Time interaction, F ( 1, 33) = 0.90, p = .351; F(2 , 33) = 1.29, p = .290; and F(2 , 33) = 1.45, p = .249, respectively.

A 3 x 2 (Group x Time) repeated measures ANOVA on vividness index scores for SHSS:C items showed a significant time effect, F (1 , 33) = 59.63, p < .001. Further analyses (paired two-tailed t tests) revealed significant differences between Time 1 and Time 2 for all groups (control, bogus debriefed, and bogus nondebriefed), with participants scoring lower at Time 2, t (10) = 3.01, p = .013; t ( l l ) = 5.04, p < .001; and t(12) = 5.54, p < .001, respectively.

Discussion

In contrast with the results of the original study on expectancy manipulation (Wickless & Kirsch, 1989), the results of Experi-

Table 2 Means and Standard Deviations for Behavioral and Subjective Indexes of Hypnotic Responsiveness at Time 1 and Time 2

Group

Bogus Bogus debriefed nondebriefed Control (n = 12) (n = 13) (n = 11)

Hypnotic responsiveness M SD M SD M SD

Behavioral (SHSS:C) Time 1 6.67 2.64 7.92 1.94 6.09 3.94 Time 2 5.67 2.06 7 . 3 1 2.43 6.64 2.94

Subjective (vividness of SHSS:C items)

Time 1 2.87 0.67 3.38 0.76 2.9 0.75 Time 2 2.67 0.61 3.14 0.79 2.76 0,71

Note. SHSS:C = Stanford Hypnotic Susceptibility Scale, Form C.

ment 1 failed to demonstrate a significant effect on hypnotic responsiveness for the bogus-item manipulation across behavioral and subjective measures. Further, contrary to our cognitive dissonance hypothesis, there was no significant effect on hypnotic responsiveness attributable to the timing of debriefing.

E x p e r i m e n t 2

A conceptual limitation of Experiment 1, and indeed the entire line of research, is related to internal validity. Put simply, does the bogus-item procedure actually enhance participants' expectations about their own hypnotizability, as assumed by Wickless and Kirsch (1989) and ourselves? This was never directly addressed by our first experiment, Wickless and Kirsch (1989), or Wickless et al. (in press). It is not unreasonable to conclude that surreptitiously manipulated bogus items impact expectations (on the basis of the pilot studies, ratings of vividness for bogus items, and subjects' responses during debriefing), but this needs to be put to the empirical test.

Accordingly, the present study administered a measure of expectations at various times during the procedure: most impor- tantly before and after the bogus items were administered. In this way, we attempted to index any change in expectancies concerning hypnotic responsiveness attributable to the bogus items.

But a second question then emerges: If participants exposed to the bogus-item procedure indeed evidence increased expectations, and if this increment is specifically due to surreptitious confirmation, then participants who are administered the bogus items without surreptitious confirmation (no manipulation of lighting) should not evidence such an increase in expectations. Thus, including a group that receives bogus items without surreptitious confirmation enables us to further determine what, if any, component of the bogus item procedure accounts for changes in expectations and subsequent behavioral responsiveness, if any such changes indeed occur.

Method

Design and Hypotheses

Given the results of Experiment 1, we were not interested in the stability of the SHSS:C scores in this particular study. Thus, participants in Experiment 2 attended only one session. As before, participant numbers were targeted around those of the Wickless and Kitsch (1989) study, with approximately 15 participants in each of three groups. All participants were enrolled in an introductory psychology course and received course credit in exchange for their participation. Only those participants reporting no previous hypnotic experience were accepted for this study.

Immediately after hypnotic induction the participants in all three groups reported how well they expected to do in hypnosis. Another measure of expectations was obtained after bogus-item administration. The three groups differed as to whether and how the bogus-item procedure was managed: (a) The lights group was administered bogus item suggestions with surreptitious confirmation (as per Wickless & Kirsch, 1989, and as in the bogus-item groups of Experiment 1); (b) the no- lights group was administered bogus-item suggestions but without con-

6 This no-show rate ( 11 of 47) is similar to that in the Wickless and Kirsch (1989) study (8 of 30).


flrmation of these suggestions through surreptitious manipulation of lighting; and (c) the control group received no bogus-item administration and no manipulation of lighting. Participants in all groups then completed the array of hypnotic behavioral and vividness responsiveness measures detailed in Experiment 1. Three hypotheses were tested through a priori within-subjects and between-groups comparisons.

Manipulation check on lights and no-lights conditions. Logically we would expect that participants in the lights group would report more vividly experiencing the bogus items than would the participants in the no-lights group. After all, the lighting was in fact literally changed for the lights group and not at all changed for the no-lights group. However, this must be empirically tested, as we must establish that the lighting changes in the lights group were indeed above threshold. Accordingly, we expected that the bogus-item vividness rating for the lights participants would be greater, and less variable, than that for the no-lights participants whose report would presumably rely on suggestibility alone.

Testing if the bogus item procedure increases expectations. If the bogus-item procedure (as used by Wickless & Kirsch, 1989, and ourselves) does indeed eulaance expectations regarding hypnotic responsiveness, then those subjects in the lights group should evidence an increase in expectations regarding their future hypnotic responsiveness from immediately before bogus-item administration to immediately after bogus-item administration. Further, if this effect is due to surreptitious confirmation of the bogus items (and not, i.e., simply because of more time in hypnosis), then this increase in expectations from before to after bogus-item administration should not obtain for participants in the no- lights group. 7

Testing if higher expectations lead to greater hypnotic responsiveness. If the bogus-item procedure affects expectations (see above), and especially if the expectations of lights group participants immediately prior to administration of the SHSS:C items can be shown to be greater than that for control participants immediately prior to administration of the SHSS:C items, then we would reasonably expect that the behavioral and subjective responsiveness to hypnosis for lights group participants (SHSS:C scores and vividness) should be greater than that for control group participants who received no such expectation-enhancing procedure prior to administration of the SHSS:C items.

M e a s u r e s

Three dependent measures were used for all participants. A behavioral measure of hypnotic responsiveness. The SHSS:C

(Weitzenhoffer & Hilgard, 1962) was used as per the Wickless and Kirsch (1989) study and Experiment 1.

A subjective measure of hypnotic responsiveness. A subjective rating of the vividness of each hypnotic suggestion (vividness index) was obtained from participants following termination of the hypnotic session as per the Wickless and Kitsch (1989) study and Experiment 1.

Expectations of hypnotic responsiveness. Participants were asked to estimate, if given 20 hypnotic suggestions at some future session, how many suggestions they thought they would respond to. The expectancy measure was administered four times: (a) prior to hypnotic induction, (b) immediately following the hypnotic induction (before the bogus items), (c) just after the bogus items, and (d) just prior to termination of the SHSS:C.

Procedure

Lights group. One at a time, the 15 participants in the lights group were seated in a comfortable chair facing the corner of the room, with the experimenter seated slightly behind them and to the left. The room was arranged in exactly the same manner as in the first experiment and, with the exception of the administration of hypnotic expectancy probes, the procedure for the bogus-item groups of Experiment 1 was followed

precisely. After participants had found a comfortable position, the first hypnotic expectancy probe was administered:

If at some future time we were to give you 20 suggestions, at that time (knowing what you know now) how many of those 20 suggestions do you think you would respond to?

The induction portion of the SHSS:C was then administered, after which participants received the second hypnotic expectancy probe. The second and subsequent expectancy probes differed slightly from the first:

If we were to give you 20 suggestions at some future time, how many of those do you now think you would respond to?

Participants then received the same bogus.item expectancy procedure that had been used in the first experiment. Immediately after the last bogus item, participants were administered the third expectancy probe. Participants were then administered the 12 SHSS:C items and were probed for hypnotic expectancy immediately prior to the 12th SHSS:C item. Upon termination of the SHSS:C, participants were given the vividness index to complete as a final task.

No-lights group. The procedures and measurement of dependent variables were identical to those for the lights: group with the exception that when bogus-item suggestions were made (seeing the color red, seeing the color green, seeing a flickering light, seeing dark, and seeing the color blue), there was no alteration in the lighting conditions. Sixteen participants were assigned to the no-lights group.

Control group. The 15 participants in the control group were administered the standard SHSS:C but no bogus items. Hypnotic expectancy reports were probed in the same manner described above, except for the probe following bogus items (because the control group was not exposed to the bogus suggestions). The vividness index was completed after termination of hypnosis (control participants were instructed to ignore the first five bogus items on the vividness index).

Resu l t s

Man ipu la t i on C h e c k on L igh t s and No-L igh t s Condi t ions

As per Table 3, the manipulation check on the vividness ratings of bogus items revealed that participants in the lights group reported significantly greater vividness in response to the bogus items than did participants in the no-lights group (M = 3.60 and M = 2.93, respectively), independent t for unequal variances, t (25) = -2 .97 , p = .003. Further, Levene 's test for equality of variances confirmed that the variability of vividness for participants in the lights group was significantly less than that of participants in the no-lights group (SD = 0.464, SD = 0.762, respectively; F (1 , 28) = 5.19, p = .031). It seems safe then to infer that participants in the lights group perceived the changes in lighting, and that this led to vividness reports that

7 Arguably there is a confound here. Because no-lights participants are left to their own devices regarding the five bogus items, and because such positive hallucination suggestions are exceedingly difficult to passDless than 10% of subjects do so (Weitzenhoffer & Hilgard, 1962)--these participants are potentially being exposed to five succes- sive failure experiences before they even begin the SHSS:C items proper. Although this does not compromise the group's utility as a manipulation check (vividness scores), it does compromise its usefulness related to all responses after the bogus-item procedure.


Table 3 Means, Standard Deviations, and Significant Findings for Experiment 2

Group

Lights No-lights Control

Dependent measure M SD M SD M SD

Expectations before bogus items 14.43 c 5.12 13.00 5.57 12.57 d 5.96

Expectations after bogus items 16.07 °'d 4.38 12.31 5.63 - - - -

Vividness rating of bogus items 3.60" .46 b 2.93 ! .76 b - - - -

Hypnotic responsiveness: Behavioral (SHSS:C) 7.20 2.08 6.94 2.91 6.87 2.85

Hypnotic responsiveness: Subjective (vividness of SHSS:C items) 3.18 c .64 3.06 .79 3.09 .40

Note. SHSS:C = Stanford Hypnotic Susceptibility Scale, Form C. Dashes indicate that bogus items were not administered to the control group. "Significant difference in means for vividness ratings of bogus items, p = .003. b Significant difference in variability between vividness ratings of bogus items, p = .031. c Significant increase in expectation scores following bogus items for lights group, p = .018. d Significant difference in expectation scores immediately prior to SHSS:C items, p = .04. c Significant difference in vividness scores for SHSS:C and bogus items, p = .011.

were more uniform and higher than those of participants in the no-lights group who were never exposed to lighting changes.

Testing if the Bogus-Item Procedure Increases Expectations

Participants in the lights group evidenced a significant increase in expectation scores following bogus-item manipulation (see Table 3; before bogus items, M = 14.43; after bogus items, M = 16.07), paired one-tailed t (14) = -2 .32 , p = .018. Mean expectation scores of the no-lights group dropped following the bogus-item suggestions but not significantly so (before bogus items, M = 13.00; after bogus items, M = 12.31), paired two- tailed t (15) = 0.67, p = .515. It appears then that the bogus- item procedure does in fact fool participants into believing they will do better in hypnosis than they previously expected, and that this effect has something to do with the manipulation of lighting.

Testing if Higher Expectations Lead to Greater Hypnotic Responsiveness

As per our hypothesis regarding the impact of higher expectations on subsequent hypnotic responsiveness, we first tested to determine if the expectations of participants in the lights group immediately prior to administration of the SHSS:C items was significantly higher than those of control subjects immediately prior to administration of SHSS:C items. This was indeed the case, t (28) = 1.81, p = .04. With this difference secured, we then compared the behavioral hypnotic responsiveness scores (SHSS:C) and vividness scores of lights and control participants (see Table 3 ).

One-way ANOVAs on these data revealed that the bogus-item procedure produced no significant effects on SHSS:C behavioral scores or vividness index scores, F(2 , 43) = 0.066, p = .936, and F(2 , 42) = 0.147, p = .863, respectively) We also carried out a 3 × 2 chi-square analysis with control and bogus- item participants categorized into low-hypnotizability, medium- hypnotizability, and high-hypnotizability groups. The chi-square was nonsignificant, X2(2, N = 46) = 2.164, p = .339. It must further be noted that four of the cells had expected counts below 5 (under these circumstances a chi-square may overestimate the degree of relationship). In sum, we could detect no effect for the increased expectations of lights participants on subsequent hypnotic responsiveness across behavioral and subjective indexes.

Finally, we combined behavioral and subjective scores across both experiments, comparing scores for controls (n = 30) with those of participants who received the bogus-item manipulation (n = 47). Means on the SHSS:C were 6.97 and 7.30, respectively, a nonsignificant difference, F ( 1, 75) = 0.280, p = .599. Because this test failed to match Wickless and Kirsch's findings, we performed a power analysis to determine the likelihood of replication, given our sample size. Using their obtained means and standard deviations (experimental M = 7.8, SD = 2.48; control M = 5.47, SD = 2.29) along with our sample size yielded a power of 98%. Thus, given our design, we are very unlikely to have missed detecting an effect of this magnitude if it were there. Means on the vividness index were 3.12 and 3.21, a nonsignificant difference, F ( 1, 75) = 0.300, p = .585. The 3

8 The correlation of SHSS:C behavioral scores with vividness scores was .653, p < .001.


× 2 chi-square analysis with control and bogus-item participants categorized into low-, medium-, and high-hypnotizability groups was also nonsignificant.

Discussion

The Experiment 2 findings suggest that the lighting procedure used was above threshold for lights participants (vividness differences), but at the same time it was sufficiently subtle to fool these participants into believing that they would be more hypnotically responsive than they originally predicted. Wickless and Kitsch had posited that the bogus-item procedure is an effective means of increasing expectations of success in hypnosis, and on testing this claim empirically we found it to be sound. However, for a second time we failed to detect an effect for these increased expectations on actual hypnotic responsiveness across both behavioral and subjective indexes.

General Discussion

When stability across time and situations is noted for some individual attribute, the cause of the invariance is a natural point of interest for those who are curious about human nature. Researchers and theorists of all persuasions generally agree that an individual's hypnotic responsiveness is indeed relatively stable, but they disagree on the stability's cause. Hilgard and others (Hilgard, 1965; Kihlstrom, 1985) argue that this stability is primarily a product of a traitlike cognitive ability that varies roughly normally across the population. From this point of view, the hypnotic responsiveness of novice individuals should be relatively unaffected by temporally proximal manipulation of response expectations. In contrast, Kirsch and others (Kirsch & Council, 1992; Lynn & Sivec, 1992) posit that hypnotic responsiveness is primarily a product of culturally mediated attitudes and expectations (both proximal and distal) that render the individual either more or less likely to enact the role of hypnotic participant. Here stability obtains because the individual's first exposure to hypnosis further consolidates preexisting response expectations and attitudes and leads the individual to respond to the same degree time and time again. From this point of view, the hypnotic responsiveness of novice participants should indeed be affected by proximal manipulation of response expectations. Those who are thereby convinced that they will be good hypnotic participants should evidence more hypnotic responsiveness.

Thus, the Wickless and Kirsch (1989) methodology enables researchers to weigh the relative merit of trait and response expectancy theories, which in the case of proximal manipulation of expectation predict different outcomes. This is why the Wick- less and Kirsch (1989) study deserves careful examination and why it has been so highly cited in the literature. The Wickless and Kirsch (1989) findings indeed appear to suggest that temporally proximal manipulation of response expectancy yields a significant increase in hypnotic responsiveness, both behavioral and subjective. This lends support to the response expectancy explanation of hypnotic susceptibility, especially because the gains in hypnotic responsiveness were achieved with such little effort (i.e., a few minutes of manipulating the environment just before testing). However, the results of our two studies are at variance with these findings. Although we were able to docu-

ment that the bogus-item procedure did in fac t increase response expectancy and that this increase is arguably due to the surreptitious lighting arrangement, we found no effect on actual hypnotic behavior or subjective experience. In other words, across two studies, the manipulation of response expectancy had no measurable effect on hypnotic responsiveness. Even though our lights participants were successfully fooled into believing that they were good hypnotic subjects, their behavior did not fulfill the prophecy. Why the difference in findings? Before moving to conceptual implications of our findings, it is important to consider what, if any, procedural or measurement differences might explain the disparity between our results and those of Wickless and Kirsch. There seem to be two general possibilities.

Inadequate Application of the Bogus-ltem Procedure

It might be argued that we failed to adhere to the subtleties of the lighting situation. After all, the entire procedure is predi- cated on not being too obvious about the light change, but at the same time it requires that the lights be detectable. Wickless and Kirsch (1989) used a black cloth to eliminate shadows, which apparently were problematic in their lab room. In pilot work, we managed this problem by removing objects that would cast shadows and found no discernible shadows in our trial runs. We find that explaining different outcomes as being due to procedural differences exceedingly unlikely for two reasons. First, we earnestly followed the extensive mechanical, piloting, and equipment specifications detailed in the dissertation upon which the Wickless and Kirsch study was based (Wickless, 1987). More important, the lights-no-lights and lights-control differences in Experiment 2 bear out our contention that man- agement of the lights was perceptible but not obvious, and that these surreptitious lighting changes were successful in increasing expectations.

Measurement of Expectations

In Experiment 2 we carried out the first fully reported empirical test of whether the bogus-item procedure actually does affect expectations of success in hypnosis as inferred, but not tested, in the Wickless and Kirsch (1989) study. Our measure of expectations was novel (i.e., asking participants to project themselves into the future and guess how responsive they might be), and therefore Its psychometric and structural properties are relatively unknown. On the basis of the failure of their verbal manipulation procedure and the success of the experiential bogus-item manipulation, Wickless and Kirsch speculated that it is the experiential nature of the bogus item manipulation that increases both expec 7 tations and confidence in expectations. We did not measure confidence and must accept what appears to be a sound theory- based inference by Wickless and Kirsch regarding the role of confidence. However, in both of our studies we did use the response expectancy manipulation Wickless and Kirsch en- dorsed as most likely to increase both expectancy and confidence (i.e., the bogus-item manipulation), and in Experiment 2 we documented that participants who receive this manipulation do indeed expect to do better than they had originally thought.


Subject Contamination by Debriefing

One difference between the Wickless and Kirsch (1989) study and our first experiment was that we debriefed participants immediately after their first hypnosis session, whereas Wickless and Kirsch debriefed their participants before the second session and after all participants had completed the first session. It could be that we inadvertently provided an opportunity for participants to communicate with one another about the study and thus render the bogus-item manipulation not credible. We believe this sce- nario is quite unlikely for two reasons. First, we, like Wickless and Kirsch, emphasized the importance of discretion at debriefing; second, two thirds (20) of the 30 participants exposed to bogus items were sampled from an introductory psychology course with over 1,500 students and six sections and instructors. The remaining 10 participants were sampled from an abnormal psychology class of 120 students. More important, such a sce- nario is not consistent with the procedure or results of Experi- ment 2. Here our debriefing protocol was timed exactly as in Wickless et al.'s replication study (in press). Thus, there is no reason, on the basis of procedure, to believe that our Experiment 2 participants were any more (or less) vulnerable to contamination than those of Wickless et al. (in press). Beyond this, the results of Experiment 2 are fully inconsistent with the notion of broad-spectrum contamination: Subjects exposed to the bogus-item manipulation did in fact report an increase in expectations, suggesting that they were indeed fooled by the manipulation and that our secret was reasonably secure.

Measurement of Hypnotizability

Beyond the procedures, examination of means on the SHSS:C across laboratories is instructive. In many respects the SHSS:C scores for our participants exposed to bogus items are similar to those of Wickless and Kirsch: 7.80 (Wickless and Kirsch, 1989) versus 7.34 and 7.20 (Experiments 1 and 2). However, the Wickless and Kirsch (1989) mean for control participants (5.47) is substantially lower than the means obtained for con- trois in Experiments 1 and 2 (7.07 and 6.87). Given that the original standardization for the SHSS:C yielded a mean of 5.07 for Stanford University undergraduates (Weitzenhoffer & Hil- gard, 1962), could our control scores be unusually high? To the contrary, there is good reason to believe that the control means in Experiments 1 and 2 are typical of current studies. The SHSS:C was normed in the late 1950s and early 1960s (Weit- zenhoffer & Hilgard, 1962). More contemporary studies of undergraduates over the past 20 years obtain much higher scores on the SHSS:C for naive and unselected subjects, with means typically in the 6 to 7 range and some higher (Bowers, 1993; Connors & Sheehan, 1978; Crawford, Hilgard, & Macdonald, 1982; Kearns & Zamansky, 1984; Matthews & Isenberg, 1995; Springer, Sachs, & Morrow, 1977). Just as important, two inves- tigations carried out in the same laboratory as the Wickless and Kirsch (1989) study yielded SHSS:C scores similar to our own: Council, Kirsch, Vickery, and Carlson (1983) sampled un- screened subjects, eliminated 2 of the 12 SHSS:C items, and still obtained a mean of 6.60 on the resulting 10-item version of the SHSS:C. Further, in their replication study, Wickless et al. (in press) report a mean of 6.73 for controls on the full 12-

item SHSS:C. In the same study, another, arguably inactive, control group obtained a mean of 7.97. In sum, the control means in Experiments 1 and 2 are comparable to those secured by other laboratories in recent years and are within the range of reported means obtained previously in the Wickless and Kirsch laboratory itself. 9

Taken together, our studies suggest that this type of minimal, and temporally proximal, manipulation of expectation does not meaningfully alter hypnotic responsiveness. In this sense, then, the trait theory of hypnotic susceptibility seems to hold up reasonably well. This is not to say that expectancies play no role in determining the quality of hypnotic response (Orne, 1959), that clinical patients might not benefit from a preliminary training or instruction procedure (Frankel & Orne, 1976), or that the essential cognitive skill can not be somewhat modified with training (Spanos, 1986). Further, a trait of responding well to externally generated expectations might actually be related to hypnotic susceptibility (but see Moore, 1964). However, the idea that individual differences in hypnotic responsiveness reflect, at least in part, essential differences in hypnotic ability remains a viable (and disconfirmable) position with methodological implications for basic research in cognitive and social psychology, as well as clinical implications for treatment planning.

A strong response expectancy position is not necessarily com- promised by our findings. It may still be true as Kirsch would have it that "with sufficient strong response expectancies, all individuals would show high levels of hypnotic response" (Kirsch, 1985, p. 1196). This to, is a viable and ( if tested in terms of proximal manipulations) disconfirmable proposition. Though our two studies suggest that the bogus-item manipulation is not sufficiently strong to alter hypnotic responsiveness, there may be other procedures that are. For instance, the strength and durability of a response expectancy manipulation may be enhanced with a procedure that is more aggressive, of longer duration, and interspersed throughout the initial hypnotic experience. Or perhaps more subtle and psychometrically sophisticated measurement of expectations, confidence, and hypnotic responsiveness may reveal the extent to which expectations affect (or do not affect) aspects of responsiveness.

It is important to note here that Kirsch and other social learning theorists have not "closed the door" on the possibility that "some capacity or ability factor has an impact on hypnotic response that is not mediated by expectancy" (Kirsch, 1991, p. 445). Although we contend that the data support trait theory as much more than a hypothetical possibility, we concur with the spirit of this message. Arguably then, researchers who wish to test the relative contribution of essential ability and social learning to hypnotic responsiveness might begin to focus more carefully on their interaction.

9 The obtained correlations between behavioral and experiential scores were r = .53, p < .001, and r = .65, p < .001, for Experiments 1 and 2. These are somewhat lower than the corresponding correlation obtained in the Wickless and Kirsch study (1989; r = .80). This difference, though marginal, underscores the importance of the separate analyses carried out for indexes of behavioral and vividness responsiveness in Experiments 1 and 2.


Refe rences

Bowers, K.S. (1993). The Waterloo-Stanford Group C (WGSGC) scale of hypnotic susceptibility: Normative and comparative data. In- ternational Journal of Clinical and Experimental Hypnosis, 41, 35- 46.

Connors, J. R., & Sheehan, P. W. (1978). The influence of control comparison tasks and between versus within-subjects effects in hypnotic responsivity. International Journal of Clinical and Experimental Hyp- nosis, 26, 104-122.

Council, J.R., Kitsch, I., Vickery, A.R., & Carlson, D. (1983). "Trance" versus "skill" hypnotic inductions: The effects of credibil- ity, expectancy, and experimenter modeling. Journal of Consulting and Clinical Psychology, 51, 104-122.

Crawford, H.J., Hilgard, J.R., & Macdonald, H. (1982). Transient experiences following hypnotic testing and special termination procedures. International Journal of Clinical and Experimental Hypnosis, 30, 116-126.

Festinger, L. (1957). A theory of cognitive dissonance. Evanston, IL: Row, Peterson.

Frankel, E H., & Orne, M. T. (1976). Hypnotizability and phobic behavior. Archives of General Psychiatry, 33, 1259-1261.

Goebel, R. A., & Stewart, C. G. (1971). Effects of experimenter bias and induced expectancy on hypnotic susceptibility. Journal of Person- ality and Social Psychology, 18, 263-272.

Graffin, N. E, Ray, W. J., & Lundy, R. (1995). EEG concomitants of hypnosis and hypnotic susceptibility. Journal of Abnormal Psychol- ogy, 104, 123-131.

Gregory, J., & Diamond, M. J. (1973). Increasing hypnotic susceptibility by means of positive expectancies and written instructions. Journal of Abnormal Psychology, 82, 363-367.

Hilgard, E.R. (1965). Hypnotic susceptibility. New York: Harcourt, Brace & World.

Kangas, J., & Bradway, K. ( 1971 ). Intelligence at middle age: A thirty- eight-year follow-up. Developmental Psychology, 5, 333-337.

Keams, J. S., & Zamansky, H. S. (1984). Synthetic versus analytic im- aging ability as correlates of hypnotizability. International Journal of Clinical and Experimental Hypnosis, 32, 41-50.

Kihlstrom, J. E (1985). Hypnosis. Annual Review of Psychology, 36, 385-418.

Kirsch, I. (1985). Response expectancy as a determinant of experience and behavior. American Psychologist, 40, 1189-1202.

Kirsch, I. ( 1991 ). The social learning theory of hypnosis. In S. J. Lynn & J. W. Rhue (Eds.), Theories of hypnosis: Current models and perspec- tives (pp. 439-465). New York: Guilford Press.

Kirsch, I., & Council, J. R. (1992). Situational and personality correlates of hypnotic responsiveness. In E. Fromm & M. R. Nash (Eds.), Con- temporary hypnosis research (pp. 267-291). New York: Guilford Press.

Lynn, S. J., & Sivec, H. (1992). The hypnotizable subject as creative

problem-solving agent. In E. Fromm & M. R. Nash (Eds.), Contempo- rary hypnosis research (pp. 292-333). New York: Guilford Press.

Matthews, W. J., & Isenberg, G. L. (1995). A comparison of the hypnotic experience between signing deaf and hearing participants. Interna- tional Journal of Clinical and Experimental Hypnosis, 43, 375-385.

Moore, R. K. (1964). Susceptibility to hypnosis and susceptibility to social influence. Journal of Abnormal and Social Psychology, 68, 282-294.

Morgan, A.H. (1973). The heritability of hypnotic susceptibility in twins. Journal of Abnormal Psychology, 82, 55-61.

Orne, M. T. (1959). Hypnosis: Artifact and essence. Journal of Abnor- real and Social Psychology, 58, 277-299.

Piccione, C., Hilgard, E. R., & Zimbardo, P.G. (1989). On the degree of stability of measured hypnotizability over a 25-year period. Journal of Personality and Social Psychology, 56, 289-295.

Saavedra, R. L., & Miller, R. J. (1983). The influence of experimentally induced expectations on responses to the Harvard Group Scale of Hypnotic Susceptibility. International Journal of Clinical and Experi- mental Hypnosis, 31, 37-46.

Spanos, N. P. (1986), Hypnosis and the modification of hypnotic susceptibility: A social psychological perspective. In P. L. N. Naish (Ed.), What is hypnosis? Current theories and research (pp. 85-120). Phil- adelphia: Open University Press.

Springer, C.J., Sachs, L. B., & Morrow, J. E. (1977). Group methods of increasing hypnotic susceptibility. International Journal of Clinical and Experimental Hypnosis, 25, 184-191.

Tart, C.T. (1979). Measuring hypnotic depth. In E. Fromm & R.E. Shor (Eds.), Hypnosis: Developments in research and new perspec- tives (2nd ed., pp. 567-601 ). New York: Aldine de Gruyter.

TeUegen, A., & Atkinson, G. (1974). Openness to absorbing and self- altering experiences ("absorption"), a trait related to hypnotic susceptibility. Journal of Abnormal Psychology, 83, 268-277.

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

Wickless, C. (1987). The role of confirmation of expectancy in the modification of hypnotic susceptibility. (Doctoral dissertation, Univer- sity of Connecticut, 1990). Dissertation Abstracts International, 48- lOB, 3126.

Wickless, C., & Kirsch, I. (1989). Effects of verbal and experiential expectancy manipulations on hypnotic susceptibility. Journal of Per- sonality and Social Psychology, 57, 762-768.

Wickless, C., Kirsch, I., & Moffitt, K. (in press). Expectancy and suggestibility: Are the effects of environmental enhancement due to detec- tion? International Journal of Clinical and Experimental Hypnosis.

Wilson, D. L. (1967). The role of confirmation of expectancies in hypnotic induction. Dissertation Abstracts International, 28, 4787B. (University Microfilms No. 68-6781 )

Received January 14, 1997 Revision received June 15, 1998

Accepted June 29, 1998 •

Self-fulfilling prophecy and hypnotic response are not the same thing

Documents