Posthypnotic Amnesia in Hypnotizability Assessment: Validation of a New Scoring System for the Hypnotic Induction Profile

Abstract

The Hypnotic Induction Profile (HIP) is a standardized assessment of hypnotizability featuring a validated 0-10 scoring system, that does not factor in posthypnotic amnesia. Using confirmatory factor analyses (CFA), we compared the 10-point scoring system with a new 12-point system that includes the posthypnotic amnesia item in independent samples of individuals with fibromyalgia (n = 98) and healthy adults (n = 97). Additionally, we explored associations of the two scoring systems with measures of hypnotic phenomena. CFA results indicate that the 12-point scoring system is a good fit for the 1-factor model of hypnotizability. Posthypnotic amnesia loaded highly on the model in the fibromyalgia sample, and moderately on the model in healthy adults. Furthermore, the 12-point scoring system correlated significantly with measures of hypnotic phenomena. We conclude that the 12-point scoring system is psychometrically equivalent yet conceptually more comprehensive than the 10-point scoring system.

The Hypnotic Induction Profile (HIP; Spiegel & Bridger, 1970) is a standardized assessment of trait hypnotizability and requires approximately 5 to 10 minutes to administer. The psychometric properties of the HIP have been thoroughly validated by different groups (e.g., Debetz & Stern, 1979; Spiegel et al., 1976), and differences in HIP scores have also been shown to correspond with functional connectivity changes in fMRI (Hoeft et al., 2012). Several studies indicated that HIP scores correlate with the other measures of hypnotic susceptibility, such as the Stanford Hypnotic Susceptibility Scales (r values of .34 to .66; Frischholz et al., 1981; Gritzalis et al., 2009; Orne et al., 1979; Spiegel, 1977; Spiegel et al., 1976) and with scales of hypnotic phenomena, such as the Tellegen Absorption Scale (r values of .33 to.53; Frischholz et al., 1987).

While the HIP includes 10 individually scored items, the scoring systems that were originally proposed supported the hypothesis that hypnotizability may be unidimensional and, therefore, can be captured by factoring by specific items (Spiegel et al., 1976). In the effort of empirically finding which items of the HIP are best representative of the tendency to successfully follow suggestions and maintain trance, Spiegel et al. (1976) conducted rotated factor analyses and revealed five items that load best on the Hypnotic Induction Factor (HIF; i.e., the Induction Score; IND): dissociation (Di; ranged 0-2); signaled arm levitation (L; ranged 0-4 then divided by 2), control differential (CD; ranged 0-2), cut-off (CO; ranged 0-2), and float (F; ranged 0-2). Together, the IND is ranged from 0 to 10 (low to high hypnotizability). The other factor, the Structural Baseline Factor (SBL), conceptually represented the potential success in the initiation of the hypnotic state (Spiegel et al., 1976) and comprised the up-gaze, eye-roll (ER), and initial arm levitation. An additional scoring system, the Profile Grade (PG), follows the transition from a state of preinductive habitual awareness into and through the hypnotic state (Spiegel et al., 1976). To do that, the PG draws information from three items of both HIF and SBL: ER, L, and CD. Across these separate scores one can extract from the HIP, the IND is the most commonly used to embody hypnotizability (Spiegel, 1974; Spiegel & Spiegel, 2008).

According to the Revised APA Division 30 Definition of Hypnosis (Elkins et al., 2015), the hypnotic state is one of focused attention that is characterized by reduced peripheral awareness and increased capacity for following suggestions, while hypnotizability is one’s ability to experience changes in behavior, physiology, and thoughts following suggestions. The difficulty of successfully recalling information about events that occurred during the hypnotic process, also known posthypnotic amnesia, may stem from the limited access of information into conscious awareness during the hypnotic state (e.g., Bowers & Woody, 1996) and may, therefore, be increased in highly hypnotizable individuals.

Hypnotizability measures such as the Harvard Group Scale of Hypnotic Susceptibility (HGSHS; Shor & Orne, 1963), the Stanford Hypnotic Susceptibility Scale (SHSS; Weitzenhoffer & Hilgard, 1962), and the Stanford Hypnotic Clinical Scale (SHCS; Morgan & Hilgard, 1978) contain items designed to assess the presence of posthypnotic amnesia. The analyses by Spiegel et al. (1976) revealed that posthypnotic amnesia (A), an “experimental item” at the time of HIP’s initial validation (p. 303), loaded poorly (.12) on the HIF factor. However, subsequent analyses in a replication study done by Debetz and Stern (1979) found higher loadings (.35) of A on HIF. The posthypnotic amnesia captured by the A item on the HIP is particularly interesting; while common measures of hypnotizability such as HGSHS, SHSS, and SHCS do address posthypnotic amnesia, they only measure the presence or absence of amnestic phenomena based on a number-of-events-forgotten threshold (i.e., posthypnotic amnesia items are scored dichotomously). The posthypnotic item on the HIP is trichotomous and addresses the change in sense of control throughout the hypnotic induction and, therefore, can be administered (i.e., receive a score different from 0) only if such change was reported by the inductee. Once a change in the sense of control is reported, Item A is scored ordinally with either 0, 1, or 2 points based on the extent of the amnestic phenomenon and not merely its presence.

Conceptually, posthypnotic amnesia is a phenomenon related to the transition from a habitual level of awareness in and out of the hypnotic state. Historically, late 19^th- and 20^th-century scales of hypnotic phenomena considered spontaneous posthypnotic amnesia as an indicator of hypnotic depth (Cooper, 1972; Council, 2002). More recently, Woody, Barnier, and McConkey (2005) described posthypnotic amnesia as one of the four pillars of their 4-factor model of hypnotizability. As posthypnotic amnesia is often viewed and addressed as a facet of hypnotizability, the A item in the HIP is better grouped with the HIF factor than the SBL factor. Therefore, we hypothesized that a unidimensional hypnotizability score that includes posthypnotic amnesia would better represent hypnotizability as a latent construct. We conducted confirmatory factor analysis (CFA) aiming to test whether including Item A alongside the five items of the IND is a good fit with the one-factor model and whether it improves the conceptualization of IND as a hypnotizability measure.

Method

Two scoring systems of the HIP were tested: (a) the original 10-point scale, comprising the Di, ½L, CD, CO, and F items, and (b) the 12-point scale after including the A item. To better conceptualize the validity of the two scoring systems in different settings, we used two separate samples for the analyses—a sample of healthy adults and a clinical sample.

In the clinical sample, participants were people with fibromyalgia who enrolled in a study assessing the ability of transcranial magnetic stimulation (TMS) to augment hypnotizability temporarily. At the time of the current analyses, data were available for 98 participants. During their screening appointment, participants were administered the HIP (Spiegel & Spiegel, 2004) and immediately following the Sense of Agency Rating Scale (SOARS; Polito et al., 2013) assessments at baseline screening. The SOARS is a standardized measure of the subjective experience of agency during the hypnotic process, comprising 10 7-point Likert-like scale items (high scores represent a low sense of agency), derived from two conceptual components of hypnosis-related sense of agency—involuntariness and effortlessness (Polito et al., 2013). Although it focuses solely on one’s perceived experience, without factoring observable behaviors during the hypnotic state, both involuntariness and effortlessness scales of the SOARS are a good correlate of hypnotizability (e.g., correlation with the HGSHS, r values of .39 and .56, respectively; Költő & Polito, 2017; Polito et al., 2013).

The healthy-adults sample included 97 individuals who were recruited in university settings and participated in a study exploring the functional brain activity and connectivity in hypnosis (Jiang et al., 2017). These individuals were taking no medications and were free of substance use and psychiatric or neurologic disorders. At an initial screening, 545 prospective participants were administered the HGSHS in an attempt to recruit individuals with high or low hypnotizability (i.e., HGSHS scores of 9-12 and 0-3, respectively). Based on the HGSHS screening, 97 individuals were classified as high (54%) or low (46%) hypnotizable and were administered the HIP. However, the HIP confirmed only 44% as high and 35% as low hypnotizable individuals (i.e., HIP induction scores of 7-10 and 0-3, respectively).

A CFA was conducted to test the one-factor measurement model of the 10-point and the 12-point scoring systems, in samples of healthy adults (n = 97) and fibromyalgia patients (n = 98). An a priori power analysis found a sample size of 94 to be adequate (β = .81) in detecting small (.1) effects given a single latent variable (i.e., hypnotizability) and six observed variables (i.e., Di, ½L, CD, CO, F, and A; Cohen, 1988; Soper, 2019; Westland, 2010). The CFA was assessed for fit with the weighted least square mean and variance adjusted χ² test for exact fit and the standardized root mean squared residual (SRMR) for approximate fit. Based on the guidelines by Asparouhov and Muthén (2018), an exact fit is concluded if the χ² test is nonsignificant (p > .05). If exact fit could not be established, an approximate fit is concluded if SRMR value less than or equal to .08, and most absolute standardized residual correlations are small (i.e., most residuals are less than .10; Kline, 2011). Analyses were done in Mplus 6.1 environment (Muthén & Muthén, 2010).

Convergent validity for the scoring systems was established vis-à-vis Spearman's Rho correlations with the HGSHS in the healthy-adults sample, and the SOARS in the fibromyalgia sample. Analyses were done in SPSS V.25 environment (IBM Corp, 2017).

Results

The mean, median, and mode of both HIP scoring systems as well as demographic information for both samples are available in Table 1. Distribution of posthypnotic amnesia item (A) scores in both samples, as well as mean HIP 10-point score per each Item A score, are available in Table 2. Tables 3 and 4 include factor loading values of individual HIP items on the one-factor model of the 10-point and 12-point systems, respectively, for both samples.

Table 1.

Demographics

	Fibromyalgia Group n = 98 M (SD)	Healthy Adults n = 97 M (SD)
Age	49.5 (11.1)	26.56 (12.3)
Gender	95.9% female	58.8% female
Race/Ethnicity
White / Caucasian	80.6%	53.6%
African American / Black	2%	7.2%
Asian	6.1%	19.6%
Native American / Pacific Islander	6.1%	4.2%
Hispanic	16.3%	6.2%
Other	10.2%	4.1%†
Education (highest)
Some graduate school	22.4%	11.3%
Completed college	24.5%	13.4%
Some college / Two-year college	31.6%	55.7%
Completed trade school	7.1%	*
Completed high school / GED	4.1%	2.1%
Less than high school	2%	2.1%
HIP Mean
10-point	4.5 (2.6)	6.2 (3.1)
12-point	5.7 (3.5)	7.5 (3.5)
HIP Median
10-point	5.0	7.5
12-point	6.8	9.0
HIP Mode
10-point	7.0	9.0
12-point	9.0	9.0
SOARS††
Involuntariness	13.7 (7.6)	**
Effortlessness	23.3 (6.1)	**
HGSHS	**	6.6 (4.3)

^†5 participants did not check any race/ethnicity

^††2 participants did not complete the SOARS

^*Trade school was not an option in the healthy adult demographic questionnaire

^**Was not administered in this sample

Abbreviations: GED = General Education Diploma; HGSHS = Harvard Group Scale of Susceptibility; HIP = Hypnotic Induction Profile; SOARS = Sense of Agency Rating Scale

Table 2.

Item A Distribution and Mean HIP and SOARS Scores

	Fibromyalgia Group n = 98 M (SD)	Healthy Adults n = 97 M (SD)
Mean	1.11 (0.96)	1.30 (0.83)
Score (%)
0	40.8	23.7
1	7.1	22.7
2	52	53.6
HIP 10-point mean total score per item A score
0	1.95 (1.82)	3.87 (3.76)
1	6.25 (1.45)	7.66 (2.55)
2	6.36 (1.23)	6.63 (2.51)
SOARS mean involuntariness score per item A score*
0	8.95 (7.28)	**
1	18.86 (5.93)	**
2	16.64 (6.11)	**
SOARS mean effortlessness score per item A score*
0	22.69 (7.00)	**
1	21.71 (5.56)	**
2	24.06 (5.43)	**

Abbreviations: HIP = Hypnotic Induction Profile; Item A = Posthypnotic Amnesia item; SOARS = Sense of Agency Rating Scale.

^*2 participants did not complete the SOARS

^**Was not administered in this sample

Table 3.

Confirmatory Factor Analyses (CFA) Standardized Factor Loadings for the HIP 10-Point System

	Fibromyalgia	Healthy Adults	Combined Samples
Dissociation	0.356	0.789	0.584
Signaled Arm Levitation	0.210	0.722	0.518
Control Differential	0.899	0.966	0.931
Cutoff	0.861	0.615	0.749
Float	0.389	0.711	0.561

Abbreviations: HIP = Hypnotic Induction Profile

Table 4.

Confirmatory Factor Analyses (CFA) Standardized Factor Loadings for the HIP 12-Point System

	Fibromyalgia	Healthy Adults	Combined Samples
Dissociation	0.334	0.778	0.554
Signaled Arm Levitation	0.234	0.712	0.490
Control Differential	0.929	0.982	0.970
Cutoff	0.828	0.611	0.735
Float	0.403	0.697	0.529
Amnesia	0.923	0.384	0.675

Abbreviations: HIP = Hypnotic Induction Profile

We conducted post hoc analyses of the difference in HIP 10-point and SOARS total scores per each Item A score (see Table 2 for means and standard deviations). One-way ANOVA revealed significant difference in HIP 10-point total score between Item A scores in the healthy-adults sample (F = 7.68, p = .001; Figure 1) and in the fibromyalgia sample (F = 99.85, p < .001; Figure 1). Furthermore, there were significant differences in SOARS involuntariness scores between Item A scores in the fibromyalgia sample (F = 17.19, p < .001; Figure 2) but no significant differences in SOARS effortlessness scores (p = .448; Figure 2). However, the involuntariness and effortlessness scales of the SOARS in the sample did not intercorrelate (p = .864). Moreover, there were no significant differences in HIP or SOARS scores between Item A ratings of 1 and 2 (all p values ≥ .326).

Figure 1:

Mean HIP 10-point totals score per item A score in the Fibromyalgia (FMS; full line) and healthy adults (HA; broken line) sample.

Figure 2:

Mean SOARS factor scores per item A score in the Fibromyalgia sample.

Fibromyalgia Sample

The one-factor model for both 10-point, χ²(5) = 4.660, p = .458, and 12-point, χ²(9) = 9.872, p = .361, systems had an exact fit to the data. Standardized factor pattern loadings ranged from .210 to .899 for the 10-point system and from .234 to .929 for the 12-point system. The A item loaded .923 on the model. Both 10-point and 12-point systems were significantly (p < .001) correlated with the SOARS involuntariness score (r_s = .466 and r_s = .456, respectively. However, neither 10-point nor 12-point systems were correlated with the SOARS effortlessness score (p = .203 and p = .142, respectively; the correlation analysis included available for 96 participants as 2 participants did not complete the SOARS).

Healthy-Adults Sample

The one-factor model for both 10-point, χ²(5) = 4.115, p = .533, and 12-point, χ²(9) = 10.183, p = .336, systems had an exact fit to the data. Standardized factor pattern loadings ranged from .615 to .966 for the 10-point system and from .384 to .982 for the 12-point system. The A item loaded .384 on the model. Both 10-point and 12-point systems were significantly (p < .001) correlated with the HGSHS score (r_s =.748 and r_s =.739, respectively; the correlation analysis included available for 87 participants as 10 HGSHS forms were scored incorrectly).

Discussion

Our results indicate that the 12-point HIP scoring system has a good fit for the unidimensional model of hypnotizability in both clinical and healthy samples. Psychometrically, both scoring systems have a good fit for the model, suggesting that the use of both scoring systems is scientifically valid. Our findings also suggest that including the posthypnotic amnesia item in the total score can elaborate on facets of hypnotizability in people who already experience hypnotic phenomena. Healthy adults who experienced partial amnesia (i.e., A score of 1) apparently had a higher hypnotizability mean score than those who experienced complete amnesia (i.e., A score of 2). However, this difference was nonsignificant and is unlikely to represent a true difference in hypnotizability between such individuals. Due to the conceptual relevance and measurable contribution of posthypnotic amnesia to hypnotizability, we conclude that the 12-point system is more comprehensive than the 10-point system in representing the construct of hypnotizability and, therefore, warranted for use.

Notably, while posthypnotic amnesia loaded better on the model in the fibromyalgia sample than in healthy adults (standardized loadings of .923 versus .384, respectively), loadings in both samples were high enough to include the item in the model. Furthermore, additional differences in item loadings have been observed in the results; Di, L, and F, while showing moderately high loadings on the model in healthy adults (.778, .712, and .697, respectively), loaded lower in the fibromyalgia sample (.334, .234, and .403, respectively). Conversely, the CO item loaded more in the fibromyalgia sample than in healthy adults (.828 versus .611, respectively). CD was the only item to load highly on both the fibromyalgia and healthy-adults samples (.929 and .982, respectively).

Our results also indicated that people who demonstrated posthypnotic amnesia had higher overall hypnotizability scores, and they experienced less agency during hypnosis than those who did not demonstrate posthypnotic amnesia. Posthypnotic amnesia is likely a polychotomous phenotype (i.e., has a distinguishable phenotypic order or “severity”) and, therefore, there is a conceptual benefit in a polychotomous over a dichotomous assessment of posthypnotic amnesia (Corbett et al., 2004). However, in our samples there were no significant differences in hypnotizability or sense if agency between the two levels of posthypnotic amnesia (i.e., Item A scores of 1 and 2). Future examinations of this issue in should be tested in bigger samples.

Data were collected from the samples across two different studies, with substantial differences between the samples (e.g., gender, age, and clinical diagnosis differences). Furthermore, in the heathy-adults samples, the majority of participants had high or low hypnotizability (an inclusion criterion for the study for which they were originally recruited). Therefore, the results should not be interpreted as a comparison of dominant factors of hypnotizability between individuals with fibromyalgia and healthy adults. Rather, the results evidence that including the A item in the HIP global score is beneficial for the measurement of hypnotizability, as different populations may differ in the way hypnotizability is manifested throughout the standardized induction process of the HIP. The main limitations of the current study stem from the limitations of the samples used; almost all participants in the fibromyalgia sample were women, who are predominantly White. Interestingly, in the fibromyalgia sample, HIP scores correlated well with the involuntariness factor of the SOARS but not the effortlessness factor. Similarly, involuntariness and effortlessness did not corelate. As this is a novel finding in the relationship between hypnotizability and sense of agency in fibromyalgia, replication of such a relationship should be done by future studies before a claim can be made on either the HIP or people with fibromyalgia and hypnosis-driven alterations of agency. Furthermore, although SOARS scores are a good correlate of hypnotizability, it is possible that alterations in subjective perception that are unrelated to hypnotizability influence the sense of agency (Költő & Polito, 2017). Further research with larger, more diverse samples can advance our understanding of the mechanism of hypnotizability in posthypnotic amnesia and vice versa.