American Journal of Clinical Hypnosis

ISSN: 0002-9157 (Print) 2160-0562 (Online) Journal homepage: https://www.tandfonline.com/loi/ujhy20

The Hypnotic Induction Profile: Normative Observations, Reliability and Validity Donnel B. Stern Ph.D. , Herbert Spiegel M.D. & John C. M. Nee Ph.D. To cite this article: Donnel B. Stern Ph.D. , Herbert Spiegel M.D. & John C. M. Nee Ph.D. (1978) The Hypnotic Induction Profile: Normative Observations, Reliability and Validity, American Journal of Clinical Hypnosis, 21:2-3, 109-133, DOI: 10.1080/00029157.1978.10403967 To link to this article: https://doi.org/10.1080/00029157.1978.10403967

Published online: 22 Sep 2011.

Submit your article to this journal

Article views: 78

Citing articles: 29 View citing articles

Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=ujhy20

THEAMERICAN JOURNAL OF CLINICAL HYPNOSIS Volume 21, Number 2 & 3, October 1978 / January 1979 Printed in U . S . A .

The Hypnotic Induction Profile: Normative Observations, Reliability and Validity'y2 DONNEL B. STERN, Ph.D., HERBERT SPIEGEL, M.D. and JOHN C . M. NEE, Ph.D. College of Physicians and Siirgeons Coliimbia University This paper presents standardization data for the Hypnotic Induction Profile (HIP), a five to ten minute test of clinical hypnotizability. Normative observations are presented by age, sex, and educational level. Test-retest and inter-rater reliability of the HIP measures are assessed. The HIP scores predict clinical outcome in psychotherapy in which hypnosis is used, which is one sign of validity. Also, the HIP is useful in the assessment of the severity of psychopathology. Included are new, more explicit instructions for calculating the HIP induction score and profile configuration.

The Hypnotic Induction Profile (HIP) is an induction procedure and a test of hypnotizability which can be administered in five to ten minutes. Because of its brevity, it is most practical for the clinic and is presently used by clinicians in psychology and psychiatry, family practice, obstetrics, dentistry, surgery and orthopedics. The test has been developed in the course of clinical psychiatric practice, which means that potentially anxiety-arousing or embarrassing items (e.g., challenges, halI The research summarized here as well as the writing of this paper, was supported in part by the Charles E. Merrill Trust and the Merlin Foundation. * The authors would like to thank Drs. Barbara DeBetz, and Jerome Haber for their contributions as examiners in the two test-retest studies of profile patterns. Others whose help in the collection of the data was invaluable are: Laurie Lipman, Michelle Aronson and Brian Maruffi. Reprint requests should be addressed to Herbert Spiegel, M.D., 19 East 88th Street, New York, N.Y. 10028.

lucinations) are not included. As informative as such items can be, they carry the risk of interruption or prevention of rapport. For the practitioner, the most important use of the test of hypnotizability is as a screening procedure. If an individual shows little evidence of clinically usable hypnotic capacity, therapy using hypnosis may not be the method of choice. If there is clear evidence of hypnotizability, however, the clinician usually tries to help the individual to tap the capacity fully. This is not to deny the utility of clinical measures of more finely differentiated degrees of hypnotizability (the induction scale of the !{IP is meant to be such a measure), but to emphasize that the distinction between hypnotizable and nonhypnotizable is essential to the decision-making of practitioners who use hypnosis. In laboratory hypnosis, this dichotomy does not carry the same importance and would be disputed by the many inves-

109

110

tigators who view hypnotizability as a continuously distributed characteristic. (Of course, there is no a priori reason why clinically usable hypnotic capacity must be the same ability tapped by laboratory hypnotizability tests. Alternatively, it is possible that below a certain point on a continuum, hypnotizability is measurable but not clinically useful .) Laboratory tests of hypnotizability are also frequently used as screening procedures, but most often the paramount question to be answered by these tests is where an individual’s degree of hypnotizability falls on a continuum. The laboratory instruments must be quite sensitive to differences within the range of hypnotizability, a requirement which results in a large number of items and a relatively long administration time. Thus, the different ways in which hypnosis is used in the clinic and the laboratory play a part in determining the forms taken by tests of hypnotizability appropriate to the two settings. The HIP, it is hoped, while serving the needs of the practitioner, also incorporates enough of the features of the laboratory test to become a link between the two methods of studying hypnosis, a way in which the findings of clinicians can be quantified and thus compared. The HIP may be used for three complementary purposes in clinical work, or to put it another way, there are three general hypotheses associated with the test: ( 1 ) the HIP measures clinically usable hypnotizability; (2) the HIP can help in the identification of persons with relatively severe psychopathology; (3) from HIP performance can be inferred certain broad personality styles. In the final part of this paper, evidence for the validity of the first two of these inferences is examined. The hypotheses regarding personality are reviewed elsewhere (Spiegel, 1974, 1977; Spiegel and Spiegel, - - 1978). Prior to presentation of data on validity, HIP admihistration and

STERN, SPIEGEL AND NEE

scoring are described, and normative observations and data on reliability are presented. ADMINISTRATION The HIP is best described as an objectively scorable hypnotic interaction, part of which also serves as a hypnotic induction. In order to use the test properly, rhythm or momentum must be established and maintained during the interaction - there should be no long silences or pauses, nor should the pace be so rapid that the subject does not have a chance to attend to the experience. The HIP requires of the examiner a high degree of familiarity with the testing procedures - the examiner himself is the insri-xnent, and if he is not finely tuned the HIP will not be valid. Thus, persons new to the test should not expect to be able to master the technique immediately and should be aware that several - perhaps many practice administrations are the necessary prerequisite to valid administration. Descriptions of five of the six test items used in summary scores appear in Table 1 . Detailed instructions for administration are presented elsewhere (Spiegel and Spiegel , 1978). Scoring procedures for the sixth test item, the eye-roll sign (ER), are presented in Figures 1 and 2. To observe ER, the examiner asks the subject to roll the eyes upward as far as possible, (the upgaze) and then, while maintaining upward gaze, to close the eyelids slowly. During eye closure, the examiner scores the eye-roll (Figure 1) and the squint, if present (Figure 2). The sum of these two scores is the score of the eye-roll sign. In passing, it may be noted that ER was a serendipitous observation. Why ER is a meaningful measure is not clear. Although only six items are used in summary scores, other items are administered either for preparatory or clinical purposes. For instance-, the signalled arm ievi-

HYPNOTIC INDUCTION PROFILE

111

TABLE1 . THEITEMSOF THE HIP INDUCTIONSCALE Dissociation.

“Spontaneous,

”

uninstructed. Score positive

( 1 or 2) if subject reports that the arm used in

. Signalled Arm Levitation (Lev).

Control Differential (CD).

cu-off.

Float.

the preparatory levitation task feels “less a part” of the body than the other arm, or if that hand feels “less connected to the wrist” than the other hand. Score positive if, on the instructed signal, the arm rises to upright position. Positive scores vary 1-4 depending on the number of verbal reinforcements necessary. “Spontaneous, ” uninstructed. Score positive (1 or 2) if subject feels less control over the arm used in the Lev item. The examiner’s questions do not indicate which arm is expected to be less controllable. Score positive ( 1 or 2) if, on instructed signal, subject reports normal sensation and control return to arm used in Lev item. Score positive ( 1 or 2) if subject reports having felt the instructed floating sensation during the administration of the Lev item.

tation (Lev), requires that the subject have a ity should be noted: two of the HIP items, prior experience with arm levitation. This Dissociation (D) and Control Differential preparation (“preparatory arm levitation”), (CD), are not instructed. They occur sponis carried out in a standardized way, during taneously or not at all. The possibility of formal hypnosis, and may be scored. How- subtle demand characteristics as the basis of ever, because preparatory arm levitation these experiences and responses is not disscores correlated with both the collection of counted, but if such is the case, it is most items which constitute the HIP induction interesting that knowing a person’s rescore and the group of items which does not sponse to subtle interpersonal cues like (Table 5), this item was not included in the these can be of help in estimating hypinduction score. notizability. Subjects’ response to these The amnesia item, also scored but not items can be described as an experience of used in the induction score because of lack discovery. It seems likely that the experiof correlation with other items (Table 5), is ences in question are naturally occurring one of the items administered because of concomitants or results of the hypnotic potential clinical utility. Although an am- state. nesia for the signal which terminates the T h e tndirction Score and the ProJile Lev item is not instructed, many subjects Conjigirration spontaneously experience it. The clinician asks about amnesia in order that later, he Two summary hypnotizability scores can can use its occurrence or lack of occurrence be calculated, each of which serves a difin explanations. A skeptical subject might ferent purpose. The indirction score is a unbe reminded of the spontaneous amnesia; a ifactorial, actuarial index of degree of hypfearful subject might be reassured by refer- notizability which is the sum of the scores ence to its absence. One difference between for the five items described in Table 1 . the HIP and other measures of hypnotizabil- (These five items are those which form Fac-

112

STERN. SPIEGEL AND NEE

EYE-ROLL SIGN FOR HYPNOTIZABILITY UP-GAZE

ROLL

FIGURE1 tor B in Table 5 , which will be discussed clinical observations concerning its use as a below.) Because the Lev item is scored on a predictor of hypnotizability (Spiegel, 0-4 continuum (unlike the score of 0-2 1972). The second factor, Factor B, or the assigned to the other four items), the Lev induction score, is represented by control score is divided by two prior to its addition differential (CD) and signalled arm levitato the induction score. Thus the induction tion (Lev). The CD item was chosen because it showed the highest correlation with score ranges from zero to ten. The profile confipration or profile pat- Factor B. Also, prior clinical experience tern is a nominal scale which was de- had suggested that a contrast between CD veloped in order to represent the type of and ER scores was clinically significant. relationship between the two factors on the The Lev item was included because it is the HIP (Factors A and B in Table 5 ) . This HIP item which is most clearly a measure “type of relationship” was important to of the subject’s capacity to respond to hyprepresent as a score because of its notic instruction. Profile patterns fall in three general theoretical and clinical significance, facets which will be explored. The eye-roll sign categories: intact, nonintact, and zero. In was selected as the representative item from each category, subtypes are recognized. Factor A because of prior hypotheses and A . tntact projle patterns: ER and CD

HYPNOTIC INDUCTION PROFILE

113

EYE-ROLL SIGN (SQUINT)

I

2

3 FIGURE2 must be scored above zero. Complete levitation on the Lev item must have been accomplished prior to the administration of

scored zero or the fourth reinforcement has been administered on the Lev item, or both. 1. “Soft” profiles: CD is scored greater than zero. The fourth reinforcement was the foirrth reinforcement. required on the Lev item. Examples: 1. “Special intact” profiles: (Formerly termed “increment” profiles .) Lev score E R = 1 , Lev=O-1, C D = 2 ; E R = 3 . 5 , is greater than ER score by two (2.0) or Lev-0, CD- 1; ER-2, Lev=O-1-2, more units. Examples: ER= 1, Lev=3, CD-2. CD=2; ER-2, Lev-4, CD-1. 2. “Decrement” profiles: CD receives a 2. “Regular intact” profiles: Any score of zero. Lev may receive any profile which satisfies the criteria for the score. Examples: E R - 4 , Lev-0, general category of intact profiles and is CD-0; ER-1, Lev-3, CD-0. not a special intact profile falls in the C. Zero projile patterns: ER receives a broad range of the regular intact profile. score of zero. 1 . “Regular zero” profiles: CD and Lev Examples: ER-4, Lev- 1 , C D = 1; ER-2, Lev=3.5, CD-2. items are both scored zero. All profiles in this category: ER-0, L e v z l , CD-0. B . Nonintact projile patterns: ER score 2. “Special zero” profiles. CD and/or must be greater than zero. Either CD is

114

STERN, SPIEGEL AND NEE

-

suggesting little or no hypnotizability. And the zero profile seems to indicate an absence of biologically based hypnotic potential (ER is scored zero). These hypotheses are set in the context of a two-factor theory of clinical hypnotizability (Spiegel, et al., 1976) which was developed to account for some of the findings to be reviewed in the present paper. It is hypothesized that the two factors are measured by ER and the induction score. The Relationship between the Indirction Score eye-roll sign is thought to be a measure of potential hypnotic capacity. This potential and the Projile Configuration is conceived as biologically based, probaCertain relationships between the two bly inborn. The induction score is seen as HIP scores are built into the test. In prac- an indication of whether or not this potentice, the Lev item in the soft profile pattern tial can be expressed. The profile configurahas usually received a zero score. This tion is a method of representing the type of means that the induction scores of softs are relationship between potential hypnotic rarely above eight on the ten-point induc- capacity and its expression. Several combition scale. The decrement profile is partially nations are possible: (1) no potential, no defined by zero CD, which almost always expression of potential (regular zero leads to a zero score on the cutoff. Thus, profile); (2) some potential, little or no exthe induction scores of decrements do not pression of potential (soft and decrement rise above 6.0. Regular zeros (because they profiles); (3) some potential, some expresobtain zero scores on CD and Lev) also sion of potential (regular and special intact score low on the induction scale. Finally, profiles). The factors which determine since the intact profile types are partially when hypnotic potential is not expressed as defined by positive scores on CD and Lev, usable hypnotizability seem to be related to the induction scores associated with them impairments in concentration such as those usually do not fall in the lowest ranges. seen in severe psychopathology. Thus, a disproportionate number of patients who A TWO-FACTOR T H E O R YOF C L I N I C A Lscore the soft and decrement profiles seem HYPNOTIZABILITY to be characterized by relatively severe psychological disturbance. This clinical obThere are hypotheses associated with servation has been evaluated by an investimost profile configurations. In general, ingation (Spiegel, Fleiss, Bridger and Arontact profiles seem to indicate that the biologically based hypnotic potential, rep- son, 1975) which will be summarized in resented by positive ER score, has been ex- the section on the validity of the HIP as a pressed (i.e., CD and Lev scores are non- test for severity of psychopathology. A second inference made from the HIP zero). This constitutes, it is thought, eviscores is whether or not the patient seems dence of clinical hypnotizability. The speable to benefit from psychotherapy in which cial intact profile may indicate particularly hypnosis plays a significant part. According strong motivation to benefit from hypnosis to the two-factor theory, clinically usable or to experience it. Among the nonintact hypnotizability should be present only patterns, the potential suggested by the when there is evidence of both hypnotic popositive ER scores has not been expressed, Lev item (s) is scored greater than zero. Examples: ER- 0 , Lev = 1, C D 0; ER-0, Lev-0, CD-2; ER-0, Lev-4, CD-2. In practice, the profile category is often further described by the ER score. For instance, “3-decrement” refers to a decrement profile pattern in which ER score was 3 . Similarly: 1.5 - intact (regular), 2-special intact, 3.5 - soft, etc.

HYPNOTIC INDUCTION PROFILE

tential and its expression. Hypnotic potential without the expression of capacity (soft and decrement profiles) should, when psychotherapy includes hypnotic strategies, predict poor clinical outcome, while a zero score on the measure of potential (zero profile) should predict poor outcome no matter how the patient scores on the measure of usable capacity. Both the nonintact and zero profile types should predict poorer treatment response than the intact profile (i.e., those profiles showing evidence of potential and capacity). These hypotheses, too, will be examined below in the light of preliminary data. Note that the theory (as well as the validity of ER) hinges on the zero profile. That is, although there is no restriction that says the induction scores of these people cannot be high, they should not respond to treatment as well as intacts - because of the zero ER. And although zero types can and often do fail the CD and Lev items (like the soft and decrement types), the zero profile should have less value in the identification of severe psychopathology than the nonintact profiles - again, because of the zero ER. Those who have some prior experience with the HIP will notice some changes in the categories of the profile configuration and in the way the configuration is conceptualized (as a nominal scale vs. its prior presentation as an ordinal scale). However, all data presented in this paper were derived by applying the scoring criteria exactly as described here and in the HIP manual (Spiegel and Spiegel 1978; Spiegel, in press). Earlier normative observations and statistical evaluations of reliability of the profile pattern (Spiegel, et al., 1976) are no longer applicable and should be replaced by those to be described here. The earlier standardization data on the induction score and ER, presented in the same paper, may still be used.

115

NORMATIVE OBSERVATIONS All data in this section are based on the HIP scores of 4,621 private psychiatric patients seen and tested by H. Spiegel between 1969 and 1976. Approximately 500 of the patients seen in 1969 and 1970 received the items of the profile scale but were not administered one of the induction scale items, “Float,” which had yet to be devised. Other than these five hundred patients, who are not included, the normative sample is a consecutive series. Most of the normative observations and reliability data have been replicated by another examiner and will soon be available (DeBetz and Stern, in preparation). Four demographic variables were included in the early data analyses: sex, age, educational level, and marital status. The only effect (or interaction) for marital status was a mild depressing effect on the scores of widowed persons, who make up a very small proportion of the sample. It was decided that the small clinical gain that might be involved in presenting these data was not worth the complication. It was soon clear that age was by far the most important of the three remaining demographic variables. However, because there were some smaller significant effects for educational level, sex, and some of the two-way interactions between the three factors, these data are also presented here. Evaluations of interactions required analyses of measures of central tendency: therefore, the data had to be presented in this form. But we also wanted to present the distributions of the HIP scores in a more direct way. The solution we arrived at was to present the data both ways, and to present the distributions of scores by age level. The combined frequency distributions of age, sex, and educational level in the normative clinical sample appear in Table 2 . The sample is obviously quite heterogene-

116

STERN, SPIEGEL AND NEE OF THE NORMATIVE CLINICAL SAMPLE TABLE2. DISTRIBUTION B Y AGE,S E X ,AND EDUCATIONAL LEVEL.

Age in Years Elementary

Under 25 25-34 35-44 45-54 55-64 Over 64 Total Percent of Female Sample Age in Years

Under 25

25-34 35-44 45-54 5564 Over 64 Total Percent of Male Sample

5 6 4 4 10 9 38 1.5

Females Highest Educational Level Attained Some College Secondary College Degree

66 137 162 195 128 44 732 28.2

91 197 168 173 83 38 750 28.8

50 214 172 113 54 28 63 I 24.3

Males Highest Educational Level Attained Some College Elementary Secondary College Degree

3 5 5 12 13 6 44 2.2

51 51

83 125 72 21 403 19.9

ous, underlining the need for norms specific to subject characteristics. The most immediately apparent aspects of the sample are its heavy concentration of patients in young to middle age (74% were between 25 and 54 years old), the high proportion of women (56%) and the very high education level (74% had at least some college). Males had a higher education level than females, but, contrary to most large samples, males were slightly older than females. Most of the patients (60%) sought treatment for cigarette smoking. Of the others, a common presenting complaint was overweight. The rest came for a consultation regarding one of the wide range of problems seen in the practice of any private psychotherapist. The HIP was always administered in the

62 81 96 76 51 19 385 19.1

33 121 159 173 65 16 567 27.9

Graduate Training

27 154 133 87 40 4 445 17. I

Graduate Training

22 171 200 145 55 33 626 30.9

Total

239 708 639 572 315 123 2,596

-

Total

Percent of Female Sample

9.2 27.3 24.6 22.0 12.1 4.8 100.0

Percent of Male Sample

171 429 543 53 1 256 95 2,025

8.4 21.2 26.8 26.2 12.7 4.7

-

100.0

first treatment session, and most patients were seen only once. The reason for collecting this information was to facilitate the treatment (or referral) process. The idea of conducting quantitative research was conceived later; thus, the standardization studies, as well as the other investigations to be summarized in this paper, were retrospective in design. Eye-Roll Sign

Table 3 shows the distribution of ER by age. Reading across the stepwise changes in each row of the table, it is clear how large an effect age has on ER. For instance, the frequency of ER-zero increases from 1.5% in the youngest group (less than 25 years) to 16.5% in the oldest group (over 64 years); and there was no one in the oldest

117

HYPNOTIC INDUCTION PROFILE 8

TABLE3. NORMATIVE OBSERVATIONS FOR T H E EYE-ROLL SIGN: THE EFFECTS OF AGEON THE DISTRIBUTION OF S C O R E S . ~ Less Than 25 Years

25-34 Yeas

0 1 .o 1.5 2.0 2.5 3 .O 3.5 4.0

1.5 5.4 9.5 23.7 20.8 22.7 10.3 6.1

1.4 8.4 10.4 26.7 22.9 19.0 7.1 4.1

2.0 9.1 12.2 28.8 20.4 17.6 6.2 3.6

4.2 11.7 14.0 31.0 19.1 13.9 3.8 2.3

5.3 22.6 15.0 28.2 16.4 10.9

Total

100.0

100.0

100.0

100.0

100.0

Score on Eye-Roll Sign

35-44 Years

4554 Years

5564 Years

Over 64 Years

Percent of Age Group

1.1

0.7

16.9 22.6 21.6 22.9 10.1 5.0 0.9 0.0 100.0

“N = 4621. If desired, the raw data may be retrieved by multiplying the entry in this table by the appropriate marginal frequencies in Table 2.

group who scored an ER of 4.0, while 6.1% of the youngest group performed in this range. (Approximately 10% of the total sample had a positive squint score. The squint scores are not presented separately, but are incorporated in the scores for ER). The clinical use of these findings depends upon how they are viewed. The most parsimonious interpretation is that the decline in ER represents a decrease with age in the flexibility and strength of the extraocular muscles. The less simple explanation - that a decline in ER reflects a decline in the hypnotic potential that this measure is said to represent - should not be favored unless there is some reason to reject the former hypothesis. If the simpler way of understanding this finding is tentatively accepted, then the meaning of the ER score varies according to age. Very roughly speaking, the low and high ranges of ER can be defined as follows: At 44 years old or younger, a low ER score is 1.5 or less; A high score is 3.5 or 4.0. At 45-54 years old, the low range is 1.0 or less, while the high range includes some people who score 3.0. At 55-64, the only score which can with certainty be in-

cluded in the low range is zero, while the high range includes those who score 3.0. Over 64 years, the high range drops even further to include some people scoring 2.5, while the low range is restricted to zero. These remarks should only be taken as guidelines, since the data cannot be broken down in a way that assigns an equal proportion of each age group to the low and high range. Spiegel ( 1974, 1977a, 1978) suggests that the range of ER is related to certain broad personality styles which the clinician needs to identify in order to provide the most appropriate treatment strategy. In Table 4 appear expected mean ER scores for groups formed on the basis of age, sex and educational level. The entries in the table were computed according to the following equation, which resulted from a multiple regression analysis performed on the ER scores: Mean expected ER Score = 2.791-.242 (age) .046 (sex) .037 (age x sex) - .050 (education) + .020 (educa.023 (education X sex). tion X age) In this equation, and in those presented elsewhere in this paper, age, sex, and educational level are to be entered as follows:

+

+

+

I18

STERN, SPIEGEL A N D NEE Age in Years

Sex Female = 0 Male - 1

less than 25 = 1 25 - 34 = 2 35 - 44 = 3 45 - 54 = 4 55 - 64 = 5 more than 64 = 6

Thus, this is the equation as it would be calculated for a 48 year old woman with a college degree: Mean expected ER score = 2.791 - ,242 (4-age) + .046 (0 = sex) .037 (4) (0) .050 (3 = education) .020 (4) (3) .023 (3) (0) = 1.91 All the main factors and interactions in the equation reached accepted levels of statistical significance. The nonlinearity effects of age and education were also examined. Whereas the nonlinear effects of those two variables were statistically significant, they contributed very little of the variation of ER score. Therefore, they were not included in the final regression equation. The regression analysis can be interpreted as follows: As stated above, the strongest effect is age. Beyondthat,males’ERscores were .2higher than females, ’ even with age controlled. However, this effect as due largely to the age

+

THE

+

+

Edircational L e t d Elementary School = 0 Secondary school = 1 Some college = 2 College graduate = 3 Graduate or professional training = 4

X sex interaction: ignoring the effect of education, at age = 1 (less than 25 years) the male - female difference was .14, whereas at ages 2 , 3 . . . 6 , the differences were, respectively, .18, .23, .27, .3 1, and .36. In other words, the deterioration of ER due to age was faster among women than among men. At present, we can offer no satisfactory interpretation of this effect. Surprisingly, educational level had some effect on ER. The group with the highest level of education was . 11 higher than the group with the least education. Although statistically significant, this effect was weak in comparison to the education X age interaction. The interesting finding here was that ER deterioration due to the age effect was halved among those with graduate or professional level education. These data suggest that highly educated subjects aged more slowly than subjects with less education.

TABLE4 . NORMATIVE OBSERVATIONS FOR T H E EYE-ROLL SIGN: EFFECTSOF AGE,S E X ,A N D EDUCATIONAL LEVELO N MEANEXPECTED SCORE.‘”

Educational Level Elementary Secondary Some College College Graduate Graduate Training

Sex

M F M F M F M F M F

Less Than 25 2.6 2.5 2.6 2.5 2.6 2.5 2.6 ‘ 2.5 2.6 2.4

Age in Years 25-34 35-44 2.4 2.2 2.3 2.1 2.4 2.3 2.3 2.1 2.5 2.4 2.3 2. I 2.5 2.3 2.3 2.1 2.5 2.4 2.3 2.1

45-54 2.0 1.8 2. I

Entries in this table derived by computation from regression equation (see text). Standard error of estimate = .SO4 cN = 4621. For cell n , see Table 2. a

I’

1.9

2. I 1.9

2.2 1.9 2.2 2.0

55-64

I .8 I .6 1.9 I.6 2.0 1.7 2.0 1.7 2.1 I .8

Over 64 I .6 I .3 I .7 1.4 1.8 1.5 I .9 I .6 1.9 I .6

119

HYPNOTIC INDUCTION PROFILE

Induction Score

items were passed by between 83% and 85% of the sample. Thus, failing an item is The induction score, an actuarial index, more informative than passing it. The abwas constructed on the basis of the intersence of semi-difficult and challenging correlations and factor analysis of the HIP items makes it likely that the induction items presented by Spiegel, et al. (1976) score is not highly sensitive to small difand reproduced here as Table 5 . To the surferences within the range of hypnotizabilprise of the investigators who carried out ity. the analysis, hypnotizability as measured by the HIP turned out not to be a unifactoTable 6 shows that the effect of age on rial dimension. This finding led to the es- the frequency distribution of the induction tablishment of the two HIP scores and to score was quite similar to the effect of age the development of the two factor theory of on the distribution of ER (Table 3): for inclinical hypnotizability. It was the second stance, scores of zero increased from 3.4% factor in Table 5 , factor B, which suggested in the youngest group to 17.4% of the oldthat the induction score might be a useful est group: and scores above 9.0 declined and feasible way to organize the data. (The from 12.7% of the youngest group to 1.4% sample in this analysis consisted of the of the oldest group. Here, though, the in1674 consecutive patients who had been terpretation is not so muddied by nonhypadministered all induction items at the time notic considerations. Clearly, the finding should be viewed as an indication that hypof the earlier paper.) The individual items of the induction notizability declines with age. It is a conscore are all “easy.” Except for the Lev stant decline, particularly steep in later item, which was passed (score greater than years, but observable in each decade of zero) by 73% of the sample, the induction adult life. This finding closely parallel reTABLE5 . CORRELATIONS BETWEEN10 COMPONENTS OF THE HYPNOTIC INDUCTION PROFILE, A N D ROTATEDFACTOR ANALYSIS” 1 Correlations: 1. Up Gaze 2. Eye-Roll 3. Preparatory Arm Levitation 4. Tingle 5. Dissociation 6. Signalled Levitation 7. Control Differential 8. cut Off 9. Amnesia 10. Float

Factor Loadings: Factor A Factor B ”N = 1674.

2

3

4

5

6

7

8

9

10

-

.77

-

.37 .03 .09

.37 .05 .I2

.I1 .21

.27

-

.I7

.21

.40

.20

.43

.06 .I0 - .05 .I4

.I1 .10

.16

.28 .21 - .02 .27

.22 .20 .02 .20

.58 .48 .03 .46

.59 .42 - .05 .48

.70 .02 .57

.92 .08

.54 .3 1

.oo

.05 .76

.22 .70

.03 .88

.90 .03

-.05

-

-

.40

-

-.01

-

.43

.07

.03 .77

-.12 .12

-

.I3 .73

120

STERN, SPIEGEL AND NEE OBSERVATIONS FOR THE HIP INDUCTIONSCORE: TABLE6. NORMATIVE THE EFFECT OF AGEON THE DISTRIBUTION OF SCORES. ' Induction Score

LessThan 25 Years

25-34 Years

35-44 Years

45-54 Years

55-64 Years

Over 64 Years

Percent of Age Group 0

3.4 I .2 1.7 2.4 2.6 2.7 6.9 14.2 27.4 24.7 12.7 100.0

.25- 1 1.25-2 2.25-3 3.25-4 4.25-5 5.25-6 6.2.5-7 7.25-8 8.25-9 9.25-10 Total

2.8 2.0 2.4 2.3 2.8 5.7 9.3 17.7 26.1 20.2 8.6 100.0

4.2 3.9 3.4 2.5 3.1 6.8 9.9 16.0 25.8 18.0 6.4 100.0

9.3 6.2 4.6 3.5 3.5 6.7 12.5 16.2 20.2 13.8 3.5 100.0

6.5 4.1 3.6 3.3 2.7 6.6 13.2 16.2 22.7 16.4 4.7 100.0

17.4 8.3 7.8 4.1 2.3 6.4 14.7 15.6 14.7 7.3 1.4 100.0

"N = 4621. If desired, the raw data can be retrieved by multiplying the entry in this table by the appropriate marginal frequencies in Table 2.

sults of several recent investigations (Berg and Melin, 1975; Morgan and Hilgard, 1973; Gordon, 1972), all of which employed the Stanford Hypnotic Susceptibility Scale (Weitzenhoffer & Hilgard, 1959). In Table 7 are presented the expected mean induction scores for all combinations

THE

of age, sex, and educational level. These figures were again computed by an equation which was the result of a multiple regression analysis: Mean expected Induction Score = 7.34 .480(age) + .060 (sex) + .098 (age x sex) + .160 (education) As in the case of ER, each of the main

TABLE7. NORMATIVE OBSERVATIONS FOR THE HIP INDUCTION SCORE: EFFECTS OF AGE,SEX,AND EDUCATIONAL LEVELON MEANEXPECTED SCORE."""^

Educational Level Elementary Secondary Some College College Graduate Graduate Training

Sex M F M F M F M F M F

Age in Years Less Than 25

25-34

35-44

45-54

55-64

Over 64

6.9 6.9 7.1 7.0 7.2 7.2 7.4 7.3 7.5 7.5

6.5 6.4 6.7 6.5 6.8 6.7 7.0 6.7 7.2 6.9

6.1 5.9 6.3 6.0 6.5 6.2 6.6 6.4 6.8 6.5

5.8 5.4 5.9 5.6 6.1 5.7 6.2 5.9 6.4 6.0

5.4 4.9 5.5 5.1 5.7 5.3 5.9 5.4 6.0 5.6

5.0 4.5 5.1 4.6 5.3 4.8 5.5 4.9 5.6 5.1

"Entries in this table derived by computation from regression equation (see text). "Standard e m r of estimate = 2.561 cN = 4621. For cell n, see Table 2.

HYPNOTIC INDUCTION PROFILE

121

factors and interactions in the equation has a large effect on the incidence of the reached statistical significance. The vari- decrement. The interpretation consistent ables were entered into the equation accord- with the two factor theory is that with ading to the same code used in the regression vancing age, not only does hypnotic potential decline, but more and more people cananalysis of the ER scores. In general, the effects of age, sex, and not express the potential they do possess. education on the induction score followed the same pattern as was observed for ER. The of the profile, howThese effects were much smaller for the ever, is not affected by age. The incidence to induction Score than for ER. Ten percent of of the special intact profile is constant the variance of ER was explained by the age 64 and then suddenly drops. As a regression of ER on the above three vari- of all these findings, the regular intact ables and their interaction terms, while 6% profile decreases with age. of the variance of the induction score was Only the age effect is presented in Table in this way* Age was the most 8. Sex had a much smaller effect, not large Predictor, 5% Of the enough for clinical significance: women obvariance. Among the increase in tained a larger proportion of decrement patage from One level to the next was as- terns than men (15.4% versus 13.3%), a sociated with .5 decline of the induction larger proportion of soft patterns (20.0% this effect was .‘4*versus 16.7%), a larger proportion of r e p “Ore. Among the generally, had higher lar zero profiles (2.2% versus 1.3%) and a scores than females7 but this effect was larger proportion of special zero profiles On age. At Age (less (1.3% versus .9%). The frequency of three than 25 Years), the difference was .07. The of these four categories increases with age differences increased with each age level to (Table 81, which that, because the with .18, .29, *40,*51, and *61. males in our sample had a higher mean age more education had higher induction than the females, the sex differences would scores. Each successive increase in educa- probably be larger if age were controlled. tional level was associated with .16 point This is a topic which obviously deserves increase of induction score. There was no future investigation. apparent education X sex interaction nor Educational level does have a small efwas there an education x age interaction. fect on profile scores, (increasing education is associated with a decline in the proporProjile Conjiguration tion of intact profiles), but this association Because the profile pattern is a nominal is also dwarfed by that between profile patscore, a regression analysis like that per- tern and age. Furthermore, the educational formed on the ER and induction scores level was severely contaminated by age: could not be carried out. However, the ef- younger patients had more years Of fects of the demographic variables could be ing. The effects Of these two reviewed in tabular form. Again, the age could be extricated from one another in the not be effect was very strong (Table 8 ) . From the regression equations, but tangled in the case of the profile configurayoungest to the oldest age group, frequency of the two “zero” categories increased tenfold. This finding is to be expected on the In practice, then, a patient’s or subject’s basis of the declines with age for ER and profile configuration should be compared to induction score already described. Age also the age norms in Table 8. 9

STERN. SPIEGEL AND NEE

122

TABLE 8. NORMATIVE OBSERVATIONS FOR THE HIP PROFILE EFFECTSOF AGE ON THE FREQUENCY OF EACHPROFILE CATEGORY .a

CONFIGURATION: THE

Profile Configuration

Less than 25Years

Intact (regular) Special Intact

Soft Decrement Zero (Regular) Special Zero Total

25-34 Years

35-44 Years

Percent of Age 68.3 64.6 4.0 3.4 16.9 19.4 9.5 11.4 .5 .4 1.o 1 .o

100.0

100.0

45-54 Years

55-64

Years

Over 64 Years

Group 62.9 3.4 18.0 13.9 1.o 1.o

56.4 4.1 20.1 15.5 2.6 1.6

52.1 4.4 18.5 19.9 2.5 2.8

42.2 1.8 14.7 24.3 11.9 5.0

100.0

100.0

100.0

100.0

aN = 4621. If desired, the raw data may be retrieved by multiplying the entry in this this table by the appropriate marginal frequencies in Table 2.

RELIABILITY Eye-Roll Sign When two independent examiners scored the same patient’s E R , the intraclass correlation coefficients representing the degree of relationship between the two sets of “Ores have been among a group Of 53 Psychiatric clinic outpatients (SPiegel and Haber9 1976) and .73 among a group Of 43 private psychiatric patients (Spiegel and DeBetz, 1977). Wheeler et al. (1974) reported that the product-moment between two sets of ER scores, each collected by a separate examiner, was .75*In a group Of 75 private psychiatric Office patients Seen twice by One Of the authors after intervening periods Of a few months to three years, the correlation between first and second ER scores was .90. This represents excellent test-retest reliability.

tndiiction Score Among a group of 75 private psychiatric patients, level of test-retest reliability has been reported to be .76 (Spiegel et al., 1976). Degee of inter-rater agreement for the induction score has been calculated to

be .75 for the group of 53 psychiatric clinic outpatients (Spiegel et al., 1976). These coefficients indicate adequate reliability for the induction ore.

Pro!le l n an earlier paper, (Spiegel, et a l . , 1976), degree of reliability of the profile Score was represented by a correlation coefficient. Unfortunately, the same simple strategy cannot be followed here. Because the profile Score is now considered a nominal scale, degree of agreement must be determined separately for each of its categories. The statistic used for estimation of reliability was kappa (Fleiss, 1971). K a p p a , a statistic which ranges from 0 to 1.0, indicates the degree of agreement (in two sets of nominal scores) which is above the degree of agreement expected by chance. lt is directly to the intraclass correlation coefficient, which makes it possible, in the case of the present data, to estimate on the same scale the reliability of E R , induction score, and profile configuration. In Table 9 are presented test-retest results for 167 patients seen by H.S. in private psychiatric practice. These people consti-

HYPNOTIC INDUCTION PROFILE

123

tute a consecutive series of patients who were administered a second HIP. The intervening time period ranged from a few weeks to six years. Of the 99 patients who scored a regular intact profile the first time, 90 (91%), also obtained a regular intact profile on their second testing (kappa = .70). Ten of the 14 patients (71%) who first scored the special intact pattern, obtained the same pattern on their retesting (kappa = .68). For the broader category of intact profiles, test-retest reliability is excellent: of 113 patients whose initial HIP fell in the intact range, 108, or 9696, also scored an intact profile on the second examination. The decrement pattern also shows acceptable stability over time, (kappa = .63) but two thirds of the patients who were initially softs scored a different pattern on retest (kappa = .34). The trend here is toward a regular intact profile on retest: 59% (99) of the sample scored a regular intact pattern on the initial testing, while 70% ( 1 17) scored a regular intact pattern on the second testing. This may be due to a regression to the mean phenomenon: that is, CD and Lev are so “easy” that subjects who fail them once tend to pass them on the second try. However, there is at least one reason to doubt the completeness of this interpretation. Despite its lower test-retest reliability, data to

be reviewed below indicate that the soft score predicts the same relatively poor response to treatment as does the decrement profile. (Both scores predict poorer therapeutic outcome than the intact profile.) This fact makes it seem worthwhile to retain the soft score and look for the source of its low stability in what it does measure rather than viewing the instability as a sign that the score should be discarded. Table 10 presents data from the two studies of inter-rater reliability of the profile configuration. Study 1 (Spiegel et al., 1976) was performed in the outpatient department of the Payne Whitney Clinic, Cornell Medical School by one of the authors (H.S.) and a colleague. The two testters administered the test independently and without knowledge of prior test findings. Order of the examiners was systematically varied; there was no evidence of an order effect. All the above comments also apply to Study I1 except that it was performed in a private psychiatric setting for clinical purposes, and the testers were H . S . and a second colleague. In Study I, the colleague was male, in Study I1 the colleague was female. In Study I the time intervening between testings was only a matter of hours, whereas in Study I1 the intervening period ranged from one day to several years. Since the direction of the findings of the

TABLE9. TEST-RETEST RELIABILITY OF r H E SIX HIP PROFILE CONFIGURATIONS. Profile Configuration Obtained on First Examination

Profile Configuration Obtained on Second Examination Total

Intact (Regular) Special Intact Soft Decrement Zero (Regular) Special Zero Total Kappa

99 14 21

29 1 3

167

Intact (Regular)

Special Intact

90 3 14 9 0 1 117 .70

5 10 0 0 0 0 15 .68

Soft

3 0 7

6 0 0 16 .34

Decrement 0 0 0 14 0 0 14 .63

Zero (Regular) 0 0 0 0 1 0 1

1 .oo

Special Zero 1 1 0 0 0

2 4 .57

STERN. SPIEGEL AND NEE

124

TABLE10. INTER-RATER RELIABILITY OF THE SIXHIP PROFILE CONFIGURATIONS: Two S T U D I E S . ~ First Examiner Total Intact (Regular) Speical Intact Soft Decrement Zero (regular) Special Zero Total by Tester Total kappa

37 3 15 23 0 2

Intact (Regular) 12 1 4

18 1

1 0 0

2 0 0 24

18

80

3

42 .65

Second Examiner Special Intact Soft Decrement 2 0 0 0 0 0 2 3 .32

0 1 0 0 0 0

I

1 0 5

0 0 2

3 0 1 3 0 1 4 0 0 0 0 0 0 9 2 18 11 22 .71 .50

1 0 0 3 0 0 4

Zero (Regular) 0 0 0 0 0

0

0 0 0 0 0 0

0

0

0 -

Special Zero 0 0 0 0 0 0 2

0 0 0 0 0 0 0

2 1 .o

=Datafrom Study 1 are printed in roman type. Data from Study 2 are printed in italics.

two studies is similar, they will be collapsed for the purpose of discussion. (The larger proportion of decrement profiles in Study I is to be expected. It represents the fact that the clinic sample was more severely disturbed than the sample drawn from private practice. The decrement pattern seems to predict the presence of severe psychopathology.) The results: The regular intact category shows good reliability 30/37, or 81% of those who scored a regular intact pattern on the first examination also did so on the second (kappa = .65). Also as before, the general intact category shows a high degree of intertester agreement (35/40 or 88%). However, kappa was low, only .32 for inter-rater reliability of the special intact pattern. A relatively low proportion of initially soft profiles (7/15, or 47%) remained soft on the second tester’s examination (kappa = .50). The decrement pattern showed a much higher degree of inter-rater agreement: 20/23, or 87%, remained in the decrement category on second testing (kappa = .71). Collapsing soft and decrement profiles to form the group of nonintact profile patterns, we find that 28/38, or 74% of those who were nonintact on initial exami-

nation were also nonintact on second examination. The reliability of neither of the zero patterns could be evaluated on the basis of these data. In summary, the regular intact profile and the intact profiles as a whole show a high degree of test-retest and inter-tester reliability. The nonintact patterns have a much lower degree of reliability. Of the two nonintact patterns, the decrement pattern is the more stable. This finding coincides with the observation made by one of the writers (H.S.) that the decrement profile is earned by persons who show a definite “break in the ribbon of concentration,” while the soft pattern is obtained by persons with less severe psychopathology, who present with a kind of “wavering” back and forth across the boundary between intact and mildly impaired concentration. The question of either collapsing the regular and special zero patterns or of doing away with the zero category altogether was raised by some readers of earlier drafts of this manuscript. The reasoning was that the small number of people who obtained these patterns was not enough justification to retain the scores. Several considerations lead

HYPNOTIC INDUCTlON PROFILE

us to keep both scores. In the first place, one of the contributions of the HIP may be that it is sensitive to two types of low hypnotizability, represented by the nonintact and zero profile patterns. To dispense with the zero category would amount to doing away with the theory which led to the development of the test. Theory by itself might not be a substantial enough reason, but there are data to support these hypotheses. The relationships of the zero pattern to severity of psychopathology and to treatment outcome are presented below in the section on the validity of the HIP. Our decision not to collapse the two zero categories at this time is more open to question, and might be revised in the future. Again, the distinction was initially based on theory: regular zeros show no hypnotic capacity (zero ER) and no expression of it (zero CD and zero LEV), while special zeros have positive scores on LEV and/or CD. Perhaps these positive scores are due to patients’ motivation to experience and benefit from hypnosis. If so, one would expect special zeros to show a higher rate of positive outcome than regular zeros. The only presently available data relevant to this question are presented in Table 11 and described in more detail below. These data indicate a slightly higher rate of success in quitting smoking among special zeros than among regular zeros. VALIDITY Validity as a Measure of Clinically Usable Hypnotizability Estimates of the degree of relationship between the HIP scores and scores on laboratory tests of hypnotizability will soon be available. It will simplify matters, of course, if the correlations are high enough to permit the conclusion that the HIP and these previously standardized measures are tests of the same p h e n ~ m e n a .However, ~ because the HIP purports to measure clini-

125

cally usable hypnotizability, the clearest test of its validity is the degree of association between its scores and outcome in psychotherapy in which hypnosis is employed. The usual problems in carrying out this kind of study include controlling for the content and length of therapy across groups and deciding on a suitable outcome measure. These problems do not exist in the case of follow-up data on the single session treatment of smoking, which is a standardized approach varying only slightly from patient to patient (Spiegel, 1970a, 1970b). The earlier follow-up reports on this technique of treating smokers (Spiegel, 1970a, 1970b) were concerned with a samResults are now available from one investigation of the relationship between SHSS and HIP (Orne, Hilgard, Spiegel, Spiegel, Spiegel, Crawford, Evans, Ome & Frischholz, in press 1.979). This study was done in two parts, one at the University of Pennsylvania (Sample 1: n = 87) and the other at Stanford University (Sample 11: n = 58). The correlations to be reported here are between HIP scores and a score calculated by halving the sum of scores on the Stanford Scales of Hypnotic Susceptibility, Forms A and C (Weitzenhoffer & Hilgard, 1959; 1963). In Sample I, the correlation between the induction score and SHSS:A+C/2 was .41 (p < ,001); in Sample 11, the correlation was .23 (p > .05). When HIP profile patterns were dichotomized as hypnotizable (intact) and nonhypnotizable (nonintacts and zeros), the mean scores on SHSS:A+C/2 were significantly different in the two groups: In Sample I hypnotizables (n=72) scored a mean of 7.29 (SD=2.66), khile nonhypnotizables (n = 15) scored a mean of 4.50 (SD = 2.96). The difference between the means was significant beyond the .005 level. In Sample 11, hypnotizables (n=39) scored a mean of 7.09 on SHSS:A+C/2 (SD=2.55), while nonhypnotizables (n = 19) scored 5.45 (SD = 2.65). The difference between these two means was also significant (P < .05). These findings indicate a mild to moderate relationship between the HIP and SHSS. The study was carried out in two parts to study effects of order of administration of the two tests. However, differences in setting, examiner, and subjects made it impossible to isolate an order effect. A replication of this investigation is planned.

STERN, SPIEGEL AND NEE

126

ple of about 600 patients. Currently, we are beginning the questionnaire follow-up of a group of 3600 smokers. At this point, information on the induction score is not available, but we do have preliminary data on the profile configuration (Table 11). The data in Table 11 are limited to tenday follow-up because, at the next followup period (six months), profile types begin to show so much difference in the “no response” column, that differences in the positive and negative outcome columns cannot be clearly interpreted. That is, the groups with the lowest success rates also have the lowest response rates, raising the possibility that profile types differ mainly in percentage of responders and not in percentage of successful outcomes. This pattern can be seen even in the ten-day data in Table 11. At ten days though, the between - group differences in rate of no-response are much smaller than the differences in success rates. Therefore, the data in Table 11 are probably a reflection of real differences in the proportion of successful outcomes associated with various profile patterns. Data for six months, one year, and two year follow-up will be presented in a future paper. For the present purpose, we followed the conservative course of comparing the proportion of successful outcomes to the proportion of patients who responded. Since it is our impression from an unsystematic canvassing of nonresponders

that most members of the group are unsuccessful, our method of analysis is counterintuitive, minimizing what we think may be larger differences. However, adopting this method of analysis buttresses the validity of any of the predicted differences which are observed. The following hypotheses regarding clinical hypnotizability and profile configuration are supported by the data in Table 11 (all differences are based on responders only and were significant beyond the .01 level): 1) The success rate of intacts (56.9%) is higher than that of nonintacts as a group (38.7%), higher than the success rate of softs (41.8%) and higher than the success rate of decrements (34.5%) . 2) The success rate of intacts is higher than that of zeros (35%). 3) The success rates of zeros and nonintacts, the two profile types hypothesized to indicate little or no hypnotizability, are both low and differ by only four percentage points. 4) The rate of success for the special intact profile (68.5%) is higher than that for the regular intact profile (56.5%), consistent with the notion that the special intact profile indicates strong motivation. These findings support the validity of interpreting the intact profiles as evidence of positive clinical hypnotizability, and the nonintact and zero profiles as evidence of

TABLE1 1 . THEHIP PROFILE CONFIGURATION AND CLINICAL OUTCOME (AT 10 DAYS POSTTREATMENT) I N THE SINGLE SESSION TREATMENT OF SMOKING.

Profile Configuration Successful N

Special Intact Regular Intact Soft Decrement Special Zero Regular Zero

78 2155 692 542 69 85

Clinical outcome by Self-report Unsuccessful No Response

(No Smoking)

(Still Smoking)

to Follow-up

Percent of Respondents Who report Success

47.4% 38.8% 27.5% 22.1% 27.5% 17.6%

21.8% 29.9% 38.3% 41.9% 40.6% 44.7%

30.8% 31.3% 34.2% 36.0% 31.9% 37.6%

68.5% 56.5% 41.8% 34.5% 40.4% 28.3%

HYPNOTIC INDUCTION PROFILE

significantly less usable hypnotizability. Because a number of nonintacts and zeros responded favorably to treatment, though, it cannot be concluded that these profile types inevitably predict absence of clinical hypnotizability. However, if a clinician has a limited amount of time, the results suggest that the best short-term response is obtained from intacts, particularly from those relatively rare people who obtain the special intact pattern. The soft profile showed poor reliability, so poor (Tables 9 and 10) that from the point of view of measurement theory, it should be discarded. Such a recommendation would be based partly on the supposition that a score with such low reliability would also have little validity. And yet Table 11 indicates that certain interpretations from the soft profile are valid. These seemingly discrepant findings might be integrated as follows: the soft and decrement profiles require failure of the CD and/or Lev item. These items were passed by, respectively, 85% and 73% of the normative clinical sample. Since it is an unusual occurrence, failing one of these items even once during two testings is a significant indication of low hypnotizability. The second possibility, which cannot be discounted by anything in the present data, is that the “validity” of the soft profile is an experimenter effect. That is, did patients respond to what may have been the tester’s expectation that soft profiles indicate poor clinical response? Neither Spiegel nor those who have observed his work can determine any difference in his treatment of patients which might lead to differential outcomes, but this issue awaits for its solution more rigorous investigation (i.e., a therapist blind to HIP score). It may be noted here that differential therapist expectations is a potential contaminant of all the data reported in this section. A second source of data concerning the validity of the HIP as a test of hypnotizabil-

127 ity is follow-up information on a group of 97 patients who consulted H . S . between 1968 and 1974 for flying phobia. Again, a paper on this data is in preparation but preliminary findings will be reported here. Fifty-two percent of the sample was female; 87% had attended at least some college; mean age was 40.5 years. The only criterion for inclusion here was that the patient had sought treatment employing hypnosis for fear of airplane travel. Follow-up data were collected by postcard. On the card, the patient was asked to indicate whether, “At this time have you successfully mastered this problem?” The patient could check Yes, N o , or Partially. Time intervening between treatment and last follow-up contact ranged from two to nine years. It was always the last follow-up information which was selected for use here, although there were only 14 patients whose reports changed over time, and of these, 10 changed Partial to Y e s , or vice versa. Only four changed from Yes or Partial to No. Treatment in all cases consisted of a single, nearly standard session. The relationship between the profile configuration and clinical outcome for these patients is presented in Table 12. The information is preserved in its original form since there is, of course, no way to be certain of exactly what each person meant by Yes, N o , or Partially. Nevertheless, making the not unreasonable assumption that Yes and N o mean what they seem to mean, the data are a second source of support for the validity of inferences made about clinical hypnotizability from the profile configurations. (Because of the low N , the special and regular intacts were combined here, the softs and decrements were combined, and the special and regular zeros.) Of all the intacts who responded, 37% (19) reported mastery of the plobia. The proportion of Yes answers among the nonintact responders was 23% (3) and among the zero responders, 25% (1). Though the direction of

128

STERN, SPIEGEL AND NEE

TABLE 12. HIP PROFILECONFIGURATION A N D CLINICAL OUTCOME (2-9 YEARS POSTTREATMENT) IN THE SINGLE SESSIONTREATMENT OF FLYING PHOBIA. Profile Configuration

Intact (Regular and Special Types)

“At this time, have you successfully mastered this problem?” No response Yes Partially No to FoIIow-UP

Total

19

12

20

16

67

Nonintact (Soft and Decrement)

3

1

9

10

23

Zero (Regular and Special types)

1

0

3

3

7

23

13

32

29

97 = N

Tbtal

difference followed the prediction, the magnitude of the difference between the intact and the nonintact success rates, and between the intact and zero success rates was not statistically significant. If a Partial response is taken to mean that the patient derived some benefit from treatment, then the proportion of responders who felt helped by the contact (Yes responses plus Partial responses) was 6 1% (3 1) among the intacts, 31% (4) among the nonintacts, 25% (1) among the zeros. Analyzing the data on this basis leads to a significant difference between the success rates of intacts and nonintacts (X‘ = 3 . 8 df = 1, p < .05). The number of zeros was too small to yield a significant difference between their rate of success and the rate of intacts, though the direction of difference was as predicted:? As this paper was finished, findings concerning the relationship of demographics, the profile pattern, and clinical outcome of the treatment of flying phobia became available. Briefly, these findings are as follows: Patients with high education (graduate or professional school) and/or relatively high age (over 54 years) showed a higher incidence of positive treatment outcomes (20/35 or 57%) and a higher proportion of intact profiles (27/35, or 77%) than did patients with lower levels of age and education (l6/62, or 26% positive outcomes; 40/62, or 64% intact profiles). These findings raised the possibility that the association depicted in Table 12 between treatment

Is there an advantage to retaining two HIP scores? Does the induction score add anything to the predictive accuracy of the profile configuration? In reference to the flying phobia data, the answer to this question seems to be “no.” That is, among the intacts, there was little difference between the mean induction scores of those who responded Yes (n = 15, % = 7.79), Partial (n outcome and the profile pattern could be equally well predicted by age and education. However, among the group with high education and/or high age, the intact profile was still a much better predictor of reports of positive outcome (19/ 28, or 88%) than the nonintact and zero profiles ( 1 / 8 or 12.5%). Among the group with less education and lower age, there was a much smaller association between the profile pattern and treatment outcome (29%, or 12/42 reports of positive response from intacts; 18% or 4/22 reports of positive outcome from nonintacts and zeros). Sex had no effect. Also note that the incidence of intact profiles rises with age in this sample which is contrary to the normative data in Table 8. These findings will be explored, of course, in much more detail in the forthcoming paper, which will also present the standardized treatment method used for these patients. Finally, the rates of positive outcome reported here are based on the assumption that the Yes and Partially post card responses (see text) indicated that the patient felt some gain from treatment; and the total groups include those who were treated but did not return follow-up information.

HYPNOTIC INDUCTION PROFILE = 10, X = 8.25), andNo ( n = 15, X 7.92). (The number of patients with induction scores was slightly lower than the number with profile patterns because the induction score had not been devised in the early stages of this study.) And, in fact, all except four of the nonintacts and zeros had scores of 7.0 or below. With this much overlap, why not retain the induction score and jettison the more complicated profile pattern? The only reason not to take this step would be that the two types of profiles which seem to indicate low hypnotizability - nonintacts and zeros - are different according to some nonhypnotic criterion. According to the two factor theory, the nonintact profile should predict the presence of relatively severe psychopathology, whereas the zero profile should not. This question is taken up below.

V a l i d i t y a s A Measrire of S e v e r i t y of Psychopathology On the basis of the clinical observation that patients with severe psychopathology seemed to obtain low (0-6) induction scores and nonintact profile patterns, it has appeared that the HIP can also be used to identify these relatively disturbed patients. A preliminary study (Spiegel, et al., 1975) was undertaken to evaluate more formally the validity of this inference. Subjects in the study were 105 of 110 consecutive private psychiatric patients who had been administered both t h e HIP and a battery of psychological tests, including the Wechsler Adult Intelligence Scale and a number of projectives. (The other five patients did not have complete records.) The HIP and the psycho-diagnostic examination were always part of the patients’ treatment: neither the patient, the therapist, nor the examining psychologist were aware that the test results would be used for research purposes. The psychologist was also unaware of the patient’s HIP scores.

129

The 105 patients were assigned to one of two groups on the basis of an independent psychiatrist’s rating of the psychologists’ test report. The rater was also blind to HIP scores. The first group (N-56) included those whose psychopathology was rated as mild or moderate neurosis. The second group (N-49) consisted of patients with psychopathology rated as more severe severe neurosis, probable psychosis and obvious psychosis. The groups performed on the HIP in the predicted ways, and all predicted differences reached statistical significance: the proportion of the intacts with severe psychopathology was 29% (20/69); the proportion among nonintacts was 8 1%(29/36). Low induction scores (0-6.0) were earned by only 16% (9) of the group with mild to moderate psychopathology, but comprised 49% (24) of the patients whose psychopathology was more severe. Of the 16 high induction scores (9.25-lO.O), 15 (94%) belonged to subjects in the mild to moderate group. These findings provided clear support for the validity of the HIP as a test of severity of psychopathology. It should be noted that, given the low test-retest reliability of the soft score (Table lo), an interesting finding was that soft and decrement profiles were equally accurate in predicting degree of psychopathology. There was also an opportunity in the reanalysis of the data from this study to evaluate the meaningfulness of ER. If it is specifically the contrast between positive ER and zero scores on other items which predicts the presence of severe psychopathology, then - if ER is meaningful profile patterns which are the same as nonintact ones in all ways except for zero ER should not show an association with severe psychopathology. This hypothesis was borne out: only 25% (2) of the patients selected for this analysis were rated as showing evidence of severe psychopathology. The difference between this pro-

130

portion and the 81% of nonintacts rated as relatively severe reached statistical significance (p < .05). Thus, a zero score on CD and/or Lev predicts severe psychopathology only in the presence of positive ER. In conjunction with the rest of the HIP, the eye-roll sign appears to be useful in the prediction of seventy of psychopathology. Bringing together the evidence concerning the HIP as a test of hypnotizability and as a test of severity of psychopathology, the following conclusions can be drawn. From the profile pattern, inferences can be made concerning both clinical outcome and the presence or absence of severe psychopathology. These inferences are valid in a probabilistic sense for groups, but not for individuals. Those familiar with statistical analysis will of course know this, but it should be pointed out to other users of the test that the evidence for validity that has been presented does not necessarily mean that any one individual’s nonintact profile pattern, for instance, indicates the presence of severe psychopathology. The fact that only one of the two types of low hypnotizability - the nonintact profile - predicts severe psychopathology suggests that the profile configuration, for all its complexity, should be retained. It offers something new to the field of hypnosis, the capacity to differentiate two independent sources of low hypnotizability. The little data available to this point do not indicate that the induction score adds to the predictive power of the profile pattern. The evidence is not strong enough to support the decision to discard it, though. Also, the fact that it is an ordinal scale makes the induction score amenable to parametric statistical analysis, a potential advantage to researchers. Therefore, for the time being, both the profile configuration and the induction score will be retained.

STERN, SPIEGEL AND NEE

DISCUSSION The reliability indices for the induction score, ER, and for the intact profile patterns are high enough to fall within most test users’ limits of acceptability. Reliability of the zero profiles has not been adequately tested yet. Reliability of the decrement profile, and particularly the soft profile, is probably not high enough, from a psychometric point of view, to merit their use. However, there is a mitigating circumstance: these two profile patterns are valuable in identifying patients whose psychopathology can be characterized as relatively severe. The profile configuration and the induction score predict equally well clinical outcome in therapy in which hypnosis is used. The profile pattern offers the added advantage of identifying two types of low hypnotizability, one associated with relatively severe psychopathology, as stated above, and one associated with zero ER or, according to the hypothesis, low inborn potential for hypnosis. This finding is evidence for the meaningfulness of ER in the assessment of hypnotizability. The relationship between ER and useful hypnotizability is not direct - even the relationship between ER and the HIP induction score is low (Table 5 ) . Rather, it seems likely that the relationship between scores on tests of hypnotizability and clinical outcome is mediated by ER. This, at least, is the case for the HIP positive CD and/or Lev scores are better predictors of positive outcome when ER is positive than when it is scored zero. This indirect relationship may be part of the reason for the negligible correlation recorded between ER and laboratory measures of hypnotizability (Eliseo, 1974; Switras, 1974; Wheeler, Rice, Wolff, Grupsmith and Mordkoff, 1974). On the other hand, a high correlation between ER and laboratory tests is not to be expected. Spiegel, (1972) in introducing ER described it as a “clinical

131

HYPNOTIC INDUCTION PROFILE

‘soft focus’ observation. ” Given the data collected to date, that seems to have been an accurate way to put it. It is the distinction between zero ER and nonzero ER which appears to be most meaningful. However, we decided to test the possibility that the lack of correlations between ER and other measures might have been partly due to not taking into account the indirectness of the relationship. This test could be accomplished by comparing the correlations between ER and induction score in groups with and without severe psychopathology. The correlation should be low or negligible among the “severe” group because these people seem to have such difficulty in expressing whatever hypnotic potential they have. That is, induction scores should be low no matter what the score for ER. However, among the group without severe psychopathology, there should be less interference with expression of potential and thus more relationship between ER and induction score. Of course, the finding that nonintact profiles are more common among the people with severe psychopathology led us to expect that our hypothesis would be confirmed. What we did not know was what the magnitude of difference would be. The result was a highly significant association ( r = . 5 2 , p < . O O l ) among the “healthy” group ( n = 5 6 ) and a nonsignificant correlation (r = .15 p >. 10) among the more disturbed group (n=49). The difference between the correlations was significant ( z = 2 . 3 , ~ < . 0 5 ) .The samples were the same ones described by Spiegel et al., ( 1975). The normative observations contain a number of interesting effects and interactions of age, sex, and education. Interpretation of these effects awaits future investigation. The age effect, though, consistent across all three HIP measures, is notable. What makes this consistency more intrigu-

ing is that all of these findings were wholly unexpected. The data collected to date suggest that the HIP is a valid and reliable instrument uniquely suited to clinical use. It provides a way to make clinical data comparable from one investigator to another. In the past, a problem has been the absence of easily available and explicit instructions for the administration and scoring of the test. At about the same time the present paper appears, a newly revised and more thorough HIP manual (Spiegel, 1978: Spiegel and Spiegel, 1978) will become available. This new presentation should make it considerably easier for examiners unfamiliar with the test to learn how to use it. 19 E. 88th St. New York, N.Y. 10028

REFERENCES BERG,S., & MELIN,E. Hypnotic susceptibility in old age - some data from residential homes for old people. International Joirrnal of Clinical and Experimental Hypnosis, 1975,23, 184189. DEBETZ,B., & STERN,D. B. Normative observations and reliability of the Hypnotic Induction Profile: A replication study. In preparation. ELISEO,T. S . The Hypnotic Induction Profile and hypnotic susceptibility. International Journal of Clinical arid Experimental Hypnosis, 1974 2 2 , 32C326. FLEISS,J. L. Measuring nominal scale agreement among many raters. Psychological Birlletin , 1971, 7 6 , 378-382. GORDON, M. C . Age and performance differences of male patients on modified SHSS. Inrernarional Joirrnal of Clinical and Experimental Hypnosis, 1972,20, 152-155. MORGAN, A., & HILGARD, E. R. Age differences in susceptibility to hypnosis. International Joi4rnal o j Clinical and Experimental Hypnosis, 1973, 2 1 , 78-85. SPIEGEL,H. Termination of smoking by a single treatment. Archives of Environmental Health, 1970a, 2 0 , 736742.

132

STERN, SPIEGEL AND NEE

O R N E ,M . T . , H I L G A R DE. , R . . S P I E G E LH, . , S P I E G E LD, . , CRAWFORD. H. J., EVANS,F. J., ORNE.E. C., & FRI~CHHOLL, E. J. The Hypnotic Induction Profile: Internal characteristics and relation to Stanford Scales of Hypnotic Susceptibility, Forms A and C . Interncitionul Joiirnirl of Clinicnl und E.rperimenta1 Hypnosis. in press. SPIEGEL, H . A single treatment method to stop smoking using ancillary self-hypnosis. Intrrricifioncil Jorirntrl oj’ C l i n i i d und E.ywrimi~nttil Hypiosis 1970b. 18. 2 3 5 2 5 0 . SPIEGEL, H. An eye-roll test for hypnotizability. American Jorrrnul o/’ C l i n i i d Hylmisis 1972. 15. 2 5 2 8 . SPIEGEL, H . The grade 5 syndrome: the highly hypnotizable person. Intcvwcrtioriul Joronirl o/‘ C l i n i c u l a n d E.r~writnentiil Hypno.si.s, 1974. 2 2 , 303-3 19. SPIEGEL, H . The Hypnotic Induction Profile: A review of its development. In Conceptrid cintl 1ni.e.stigutiw Approaches to Hypnosis iind Hypnotic Phenomenu . Annuls o j ’ tliij Neicv York Acudivny o/’ Sciences, Volume 296, 1977. SPIEGEL. H . Manrial / o r the Hypnotic Inclrrc~tion Profile. 1st Edition. 4th Rei.ision. New York: Basic Books, 1978. S P I E G E LH.. , A R O N S O NM , . , F L E I S S ,J . L . , & H A B E RH, . Psychometric analysis of the Hypnotic

.

.

o/ C ‘ l i i i i c d Induction Profile. l r i t i ~ r n d o t i Jorrrntrl d E.vperimiwii1 Hxptiosis, 1976. 24 , 300-3 15. SI~IEGI;I., H . , & DtBI.I/.. B . Unpublished data, 1977. S P I ~ G E H.. L . F I . E I M .J . L . , BHINI:R. A . A , . & ARONSON. M . Hypnotizability and mental health. In S . Arieti & G . Chnanowski ( E d s . )N N tlitnciisions in iiwrld p.sy?.c.ltitrtry. New York: Wiley. 1975. Si’iECiEt., H . , & HABER,H. Unpublished data. 1976. SPIEGEI., H., & SPIEGEI.. D. Triiricx, tititl Trccrtiniwt. New York: Basic Books, 1978. Swi.rRAs, J . E . A comparison of the eye-roll test t o r hypnotizability and the Stanford Hypnotic Susceptibility Scale: Form A. Atnericun Journal of Clinical Hypnosis, 1974. 1 7 . 54-55. W E I T Z E N I ~ O F FA. E RM . . . & H I L G A R DE.. R . Skin/iird Hypnotic Srrsi~i~ptihilit~ Sinlc: f‘orrns A iinil A . Palo Alto. C A . : Consulting Psychologists Press. 1959. WEITZENHOFFER, A. M. & HILGARD, E. R. StunJiord Hypriotic Sccile. Fortir C. Palo Alto. CA: Consulting Psychologists Press, Inc., 1962.

crrid

WHEtLER. L., REIS, H. T . , Woi.tt. E . , GRI1i’SMITti, E. & MORDKOFF. A. M . Eye-roll and hypnotic susceptibility. Internutionul Jorrrriul of Clinicd u n d Experimentul Hypnosis. 1 9 7 4 . 2 , 327-344.

Donne1 B. Stem obtained his B.A. from the University of California, Riverside and his Ph.D. in Clinical Psychology from Michigan State University in 1975. He did his postdoctoral year in the Laboratory of Neuropsychology with Dr. Arthur Benton at the University of Iowa. He is currently a candidate for a certificate in Psychoanalysis at the William Alanson White Institute, doing private practice, and engaged in research through the Department of Psychology, College of Physicians and Surgeons, Columbia University.

DONNEL B. STERN, Ph.D.

133

HYPNOTIC INDUCTION PROFILE

HERBERT SPIEGEL, M.D. Herbert Spiegel received his M.D. at the University of Maryland i n 1939. After psychiatric training at St. Elizabeths Hospital, Washington, D.C., he served in the U.S. Army during World War I1 in North Africa and later as an instructor at the School of Military Neuropsychiatry. In 1947, he completed psychoanalytic training at William Alanson White Institute of Psychiatry. In 1961, he joined the faculty of the College of Physicians and Surgeons at Columbia University where he is currently Clinical Professor of Psychiatry and directs the postgraduate courses in hypnosis. Recently published: Triirice titiil Trctiti t w r i t : Tlie Cliriictil Uses of' Hyposis (with Spiegel, D.) Basic Books. New York, July 1978.

JOHN C . M. NEE, Ph.D., M.P.H. John C . M. Nee is a Senior Research Scientist at the New York Psychiatric Institute. He earned his doctoral degree from the Pennsylvania State University in 1973 and a M.P.H. degree in biostatistics from Columbia University in 1977.