Measuring the Impact of Media Exposure and Hospital Treatment on Patients
Alleging Satanic Ritual Abuse
by Frank Leavitt, Ph.D.
[From Treating Abuse Today 8(4), July/Aug 1998, pp. 28-29.]
Abstract
Reported incidents of satanic ritual abuse based on recovered memory have
become a source of controversy. Some scholars (Ganaway, 1989; Victor, 1995;
Weir & Wheatcroft, 1995) suggest that exposure to (1) media accounts of
satanic ritual abuse or (2) experiences in inpatient hospital settings
specializing in the care of patients reporting histories of sexual abuse
play an central role in the generation of these memories. The Word
Association Test was utilized to determine the impact of these environmental
influences on satanic word association knowledge of patients who report such
incidents. These two exposure variables were not found to be significant
predictors of satanic word associations in 43 patients reporting sexual
abuse. Paradoxically, less media exposure was associated with significantly
higher rates of satanic word associations in patients reporting satanic
ritual abuse. The production of satanic word associations was not found to
be an artifact of hospital exposure; and word association repertoires of
patients with and without histories of satanic ritual abuse were not found
to be contaminated by interactions with hospital staff or other patients.
Media and hospital exposure may allow patients to respond conversationally
about satanic abuse, but these variables do not account for unique satanic
word association knowledge found among patients reporting satanic ritual
abuse.
Reports of satanic ritual abuse (SRA) often come to clinical attention based
on testimony of patients about events remembered from childhood (Hill &
Goodwin, 1989; Young, Sachs, Braun, & Watkins, l991). These memories, often
dating to the early years of development, contain intriguing similarities
suggesting a special information base. Since little direct evidence to
corroborate the amnestic memories is often the case (Lanning, 1991),
similarity of detail and recurrence of patterns in patient accounts often
serve to weigh these accounts in favor of personal experience. The fact that
patients making the allegations come from vastly distant geographic regions,
yet share an unique knowledge base, further adds to the sense of credibility
(Van Benschoten, 1990; Young, Sachs, Braun & Watkins, 1991).
The Word Association Test (WAT: Cramer, 1968) was used to examine word
association repertoires of patients reporting SRA. The WAT is a familiar
psychological instrument for identifying meanings shared by most people for
a stimulus word. The meaning of a word is defined by responses that are
frequently associated to the word. Single words are used as the starting
stimulus for the free expression of associations that come to mind.
Non-clinical samples show predictable patterns of association emitting the
same responses with high frequency (Palmero & Jenkins, 1964). These high
frequency emissions define the consensual associations. Variations in
meaning from the consensual often yield clues that are useful in delineating
areas of conflict and exposure to different life experiences.
A word association study (Leavitt & Labott, in press) revealed another
similarity shared by SRA patients. While many of the patients’ associations
were no different than those commonly found in other samples (Postman &
Keppel, 1970), some of their word associations clustered around a domain
that was specific to SRA reporters, suggesting that a subtle information
base is broadly prevalent in this patient population. The fact that two
clusters better accounted for associations common to patients reporting SRA,
compared to a single cluster for associations common to patients reporting
sex abuse (SA), fit with the novel word association paradigm proposed in
Leavitt & Labott (in press) and served as the stimulus for this follow up
study.
Development of the word association paradigm was guided by the prospect that
two domains of experience are operating in patients reporting SRA experience
and that these domains may be obscured by treating all associations as
arising from experiences in a single domain. If what SRA patients allege is
true, then they have substantive experience in two domains: the normative
domain and a trauma domain. For example, children unfortunate enough to grow
up in an environment in their formative years where abuse was common within
circles (Hill & Goodwin, 1989) may come to associate “pain and fright” to
the stimulus word “circle” rather than “ball or round”. Our approach to the
problem was to categorize associations relevant to SRA populations
separately by double norming 16 stimulus words: sex, circle, fire, baby,
cross, triangle, blood, table, secret, black, altar, costume, eyes, skull,
knife, and ceremony. Double norming is illustrated for the stimulus word
“circle” in Table 1.
Words listed in column 1 are common associations for the normative domain.
Words listed in column 2 are common associations for the satanic domain.
Development of associational profiles in two domains has been described
elsewhere (Leavitt & Labott, in press).
Using this word association paradigm to quantify associational patterns,
Leavitt and Labott (in press) found that a unique knowledge base was
operative in SRA patients that mimics the SRA history reported by patients
alleging ritual abuse. SRA patients differed from other patient groups along
3 dimensions: (1) they were more cognitively active to the test stimuli
producing, on average, an additional 14 associations; (2) despite the large
associational inequality, SRA patients produced fewer common associations in
the normative domain; and (3) they showed a strong response bias for
associations linked to the satanic domain. The mean common response in the
normative domain was 16.0 for SRA patients and 20.1 for other SA patients;
the mean response in the satanic domain was 9.8 for SRA patients and 1.3 for
SA patients.
From both a clinical and legal standpoint, whether historical or
non-historical variables account for the unique word association knowledge
shared by patients reporting SRA is a critical issue. To begin to address
this issue, two situational variables not related to historical events were
examined to determine if they set the stage for the unique sharing of word
associations. The first variable examined was media exposure. Since we live
in a multi-media environment that promotes sensational topics, publicity in
the media is an often-mentioned source of information on SRA (Ganaway, l989;
Noblitt, 1995; Putman, 1991; Victor, l995; Weir & Wheatcroft, l995).
According to Putnam (1991), patients do not have to come together to share
knowledge. “There is massive media dissemination of material on the satanic
through dramatic autobiographical accounts, sensational talk shows and news
reports of alleged cases, not to mention numerous movies and television
programs … that can account for a large degree of apparent similarity in SRA
allegations” (p. 177).
Purpose
The major purpose of the present study was to examine exposure to, and the
effects of, media exposure (movies, television shows, books, magazines and
newspapers) on the production of associations as measured by the WAT. If
satanic associations are related to media information, then groups with high
exposure to SRA media material should exhibit an increased number of satanic
associations. Conversely, groups with low media exposure should exhibit
fewer satanic associations.
A second purpose was to examine the impact of inpatient hospital stay on a
specialized dissociative disorders unit dealing with sexual trauma as
measured by word association performance. If immersion in clinical
activities of the hospital unit affects production of satanic word
associations, these associations should be more noticeable in the protocols
of patients with longer lengths of stay. Conversely, new arrivals to the
unit should show fewer satanic word associations.
Method
Participants
The participants were 86 female, psychiatric inpatients with a history of
childhood SA who were clinically referred for psychological testing. Two
groups were formed on the basis of sexual history.
The SRA group consisted of 43 patients who alleged SA in the context of
satanic cults. They were admissions to a dissociative disorders unit over a
29-month-period with histories that included the symbols, group structure
and paraphernalia of satanic cults including: ritual use of robes, chanting,
circles, altars, sex, blood, urine, feces, body parts and sacrifice (Hill &
Goodwin, l989; Smith & Pazder, 1989; Van Benschoten, l990). Their average
age was 36.7 (SD = 8.8), and they averaged 14.6 (SD = 2.3) years of
education.
The SA group consisted of the first 43 patients during the same time frame
who reported SA without allegations of SRA. They were admissions to the same
dissociative disorders unit. Their average age was 39.3 (SD = 2.3) with 13.4
(SD = 2.3) years of education. The one year mean educational difference was
significantly lower ( p < .02).
Additional criteria for inclusion in the SRA and SA groups were: (a) a
history of SA involving penetration, (b) onset of SA prior to the age of 12,
(c) exposure to multiple sexual violence for a period of at least 6 months,
(d) a period of amnesia for sexual trauma, (e) emergence of initial memories
of sexual trauma after the age of 18, (f) partial or full memory of abuse
prior to entering treatment and (g) absence of neurologic or psychotic
symptoms. Criteria (a) through (f) were based on self-report for which there
was no independent corroboration. Criteria (e) was necessary because
patients reporting SRA invariably report a period of amnesia for the memory.
Criteria (f) was employed to reduce potential treatment effects. Patients
were excluded if their history did not allow clear group placement or if
they were unable to complete the WAT.
Materials
Media Exposure Questionnaire (MEQ). Questions regarding exposure to media
material covered 5 areas (books, magazine articles, newspaper articles,
television shows and movies). Exposure was assessed using the following
question format with wording slightly modified to fit with each media. “Have
you read any books on the following topics?” The topics were listed on
separate lines with space provided for answering “yes” or “no” and
participants were asked “How many?” for answers in the affirmative. The
topics were alcohol abuse, HIV, sexual abuse, homelessness, multiple
personality, SRA, health care, and religious cults. SRA was always embedded
as the seventh topic in the sequence for each question.
Word Association Test (WAT). Word association performance was assessed with
a revised 29-item WAT containing 16 SRA relevant words randomly arranged
among 13 buffer words. Details of the revision are described in Leavitt &
Labott (in press).
Procedure. Participants were administered the WAT individually in the first
session of a two-part psychological evaluation. The continuous word
association method was employed; i.e., each stimulus word was read aloud and
patients were encouraged to report all words brought to mind by the stimulus
word. Sixty seconds were allowed for the production of associations to each
stimulus word. Indication by subjects that they were finished was also
permitted.
Scoring was limited to a maximum of 8 associations per stimulus word and
repeated associations were not scored. One point was scored for each
association that was present among the common associations in the normative
and satanic lists for a given stimulus word (See Table 1). By way of
illustration, one patient gave 8 associations to the stimulus word “circle”:
ball-ring-fear-night-tornado-altar-blood-unbreakable. Two (ball, ring) are
common associations in the normative domain; three (altar, blood,
unbreakable) are common associations in the satanic domain. Thus, 2 points
are scored for the normative domain and 3 points are scored for the satanic
domain for the stimulus word "circle." Associations for the other 15
stimulus words were scored in the same way. Three summary scores could be
tallied for each protocol: (1) total number of associations, (2) total
number of normative associations and (3) total number of satanic
associations. However, since the study focus was on inflation of SRA
associations by media and hospital treatment sources, only (1) total number
of associations and (3) total number of satanic associations were tallied.
The MEQ was filled out by the patient after the first session ended, and
returned at the start of the second test session to ensure that the
investigator was blind to the patient’s media exposure at the time of WAT
examination and during the scoring.
Results
A comprehensive measure of media exposure was created by summing exposure to
the 5 media sources into one composite score. Composite exposure scores
ranged from 0 to 101 for SRA patients, and 0 to 56 for SA patients. The
frequency distribution of composite scores for the two groups was reasonably
symmetrical, with 35% (n=15) of each group reporting no exposure to satanic
information from media sources (see Table 2).
Visual inspection of the distribution of composite scores was used to divide
patients into two groups with different amounts of media exposure. Patients
were assigned to the Low Media Exposure group when composite exposure to the
5 media sources was 4 or less. Patients assigned to High Media Exposure
group had composite scores of 7 or more. Table 3 shows exposure to 5 media
sources in the High and Low Media Exposure groups.
Means and medians are both presented because some data are highly skewed.
High Media Exposure groups in both the SRA and SA population are quite
comparable using medians as the preferred measure of central tendency.
These data indicate that the High Media Exposure groups as defined by
composite scores are sufficiently different from Low Media Exposure groups
(median = 0 for all 5 media sources) in information received from media
sources to permit meaningful evaluation of the impact of media influence on
SRA word association performance.
As a next step, a t-test was run to determine whether patients in the SRA
and SA groups differed in total number of word associations. A significant
result was obtained (t(84) = 3.07, p < .003). The mean number associations
produced by SRA patients was 63.7 (SD = 19.4); for SA patients, the mean was
52.1 (SD = 15.6). As a result, total number of associations was a covariate
in the subsequent analysis of group effects.
The time interval from admission to testing was used as the measure of
clinical exposure to the hospital milieu. Cases were reduced to 3 time
intervals based on a consideration of the frequency distribution and the
goal of creating groups with different exposure times. The intervals chosen
to operationalize clinical exposure to the inpatient facility were 0 to 1
days (0 indicating examination on day of admission), 2 to 6 days, and 7 or
more days. The range in the SRA group was 0 to 20; in the SA group, the
range was 0 to 21. Table 4 provides means and standard deviations of the
satanic word association scores of the SRA and the SA patients with minimal,
moderate and high clinical exposure.
An analysis of variance was run using total number of associations as a
covariate to assess potential differences in the number of satanic word
associations produced by the two groups across two media exposure conditions
(High vs. Low) and 3 treatment exposures (minimal, moderate and high).
Results indicated significant main effects of both group (SRA > SA) and
exposure (Low > High); there was no main effect for time in hospital
treatment. The only significant interaction was a group x media exposure
interaction (F(1,74) =15.2, p < .001). This interaction is presented in
Figure 1. Here it can be clearly seen that media exposure makes no
difference on the number of satanic word associations of patients in the SA
group. However, individuals in the SRA group with low media exposure
produced a greater number of satanic associations than did individuals
reporting high media exposure to satanic ritual abuse content. T-tests
yielded significant differences between SRA-low exposure (M = 10.5) and
SRA-high exposure (M = 4.6) groups, t(41)=3.66, p < .001. SRA-high exposure
subjects also had significantly more satanic abuse responses than did the
SA-high exposure subjects (M = 1.2), t(30)=2.74, p < .01.
Discussion
Using a different sample, Leavitt & Labott (in press) found that patients
who reported SA in the context of satanic rituals differed in measurable
ways from other patients groups on a WAT specifically formatted to take into
account their allegations of functioning in a trauma domain during their
childhood years. They showed an unusually strong bias for satanic word
associations. The two questions addressed in this study were whether this
novel information base was linked to (a) knowledge gained about SRA from
listening and observing media reports, or (b) knowledge gained as the result
of exposure to the environment created by treatment on an inpatient unit
specializing in sexual trauma.
The present study confirms that patients who allege SRA produce
significantly more satanic word associations. However, the broad prevalence
of satanic word associations among SRA reporters does not appear to be an
artifact of either demand characteristics of the hospital environment or
knowledge gained through media exposure. Differential responding by patients
in the SRA group can not be explained on the basis of situational demand
characteristics inherent in functioning for a period of time on a hospital
unit specializing in sexual trauma. New admissions in both the SRA and the
SA group do not differ from their counterparts who have been functioning in
the clinical environment for a week or longer, at least in terms of the
production of satanic word associations. SA patients who deny SRA experience
produce minimal satanic responses whether tested at point of admission
(0.92) or well into their hospital stay (1.83). The same pattern was found
for patients who allege SRA, except that their satanic word association
scores are uniformly higher (M = 8.69 vs. 7.45). Thus differences in demand
characteristics of the environment created by hospital treatment do not
appear causally linked to the production of satanic word associations.
The findings regarding satanic word association response and media exposure
are more complex, but again indicate that processes underlying the
production of satanic responses are not affected by knowledge gained from
media exposure. Using a mean score of 9.8 from Leavitt & Labott (in press)
as a base rate for satanic response among SRA reporters, the picture that
emerged among patients alleging SRA histories ran strikingly counter to
media influence predictions. Satanic word associations were most prominent
in the protocols of SRA patients with no or little exposure to SRA material.
They produced satanic word association responses at a level (10.5) above the
base rate score of 9.8. By contrast, the level of satanic response was much
lower in SRA patients with high media exposure. These patients produced
satanic responses at a level (4.6) well below the base rate score, and at a
rate of less than 50% of the Low Media Exposure SRA group. Thus, media
scores can be used to predict which of the SRA reporters are likely to
exhibit high numbers of satanic associations; however, for predictions to be
reliable, they must run opposite to expectation. Low exposure to media
material is the condition most likely to be associated with large numbers of
satanic word associations.
At the same time, media exposure had no affect on satanic word associations
for patients in the SA group. Like the SRA group, these patients were
divided into two groups on the basis of their composite media exposure
score. If media exposure is a mechanism mediating satanic associations, then
satanic associations should rise among patients with substantial media
exposure. The SA group may provide the cleanest test of the relationship
because if satanic associations occur, they are explainable through media
exposure rather than personal experience. Recall that patients in the SA
group denied SRA experience. Remarkably, SA patients with either High or Low
Media Exposure did not differ in number of satanic associations (1.2 vs.
1.2) despite dramatic differences in media exposure. Thus, patients who deny
SRA experience are not more likely to produce satanic responses because they
have read literature or have been exposed to movies and television shows on
this subject matter.
The composite scores of High Media Exposure patients in both the SRA and SA
groups indicate that these patients are informed about satanic ritual abuse.
At least 50% of each group had read 2 or more books, 3 or more magazine
articles, 5 or more newspaper articles; viewed 2 or more movies and watched
3 or more television shows. Yet, media-based knowledge did not serve as a
retrieval cue for vicariously acquired SRA information using the WAT as the
measure. Conversely, patients in the Low Media Exposure groups were not
SRA-knowledgeable due to media sources, based on median composite media
scores of zero. Given the evidence that WAT performance and media
information are either unrelated or related in a non-expected direction (at
least in the SRA group), it seems reasonable to conclude that information
vicariously acquired through media sources is in some fundamental way
different, or simply has a different meaning for these patients.
Furthermore, this information does not trigger the benchmark satanic
associations at the frequency expected, and thus is not of central
importance in understanding this phenomenon. From a scientific perspective,
media exposure is neither a necessary nor a sufficient condition for the
emergence of satanic associations on the WAT.
This study raises a number of important questions. If satanic responses do
not flow from actual experience, then from where do they arise? Exposure to
media materials does not appear to be a plausible explanation, nor does
exposure to a hospital setting. In a similar vein, it is difficult to
understand the reverse media effect in the SRA group. Why do only SRA
patients with minimal or no media information respond to the WAT (i.e., show
the base rate satanic associations) in a manner consistent with their report
of clinical information? Why are higher levels of media exposure among SRA
patients related to substantially lower satanic associations? Though
treatment-related variables were not directly studied, they do not appear
explanatory. These, as well as other questions, merit further research.
Several limitations of this study warrant consideration. As with any study
relying on self-report for determining trauma and media exposure history,
the standard caveats regarding accuracy of self reports apply. A second
concern is that the first author administered the WAT to all patients. While
collection of information on group status was delayed until the second
session, historical information that indicated group status was either
available or spontaneously introduced in a number of cases. A third concern
is that simple frequency counts of the time interval between admission and
the administration of the WAT may not be sufficient for accurately ordering
the patients along a dimension of exposure to the clinical environment.
Clearly, treatment events other than duration of exposure have potential for
affecting a patient’s response. A final concern is that the sample was
reasonably well educated. Thus, the findings cannot be generalized to less
educated segments of the population.
CONCLUSION
This study provides evidence that clients who report SRA exhibit a set of
associations to SRA-related words that cannot be explained by exposure to
the popular media or from inpatient treatment. The subtle set of
associations shared by those claiming SRA, are similar to those of SRA cases
reported in the literature. As such, these associations may be the indelible
imprints of past trauma. In addition, this study describes a new method for
evaluating claims of satanic ritual abuse that relies less heavily on self
reports. Putting new lenses in an old pair of word association glasses
enabled the influence of media input and hospital treatment exposure to be
evaluated with fewer confounds than encountered with more familiar dependent
variables such as counts of content extracted from descriptions of abuse.
REFERENCES
Ganaway, G.K. (1989). Historical versus narrative truth: Clarifying the role
of exogenous trauma in the etiology of MPD and its variants. Dissociation,
2, 205-220.
Hill, S. & Goodwin, J. (1989). Satanism: similarities between patient
accounts and pre-inquisition historical resources. Dissociation, 2, 39-44.
Lanning, K.V. (1991). Ritual abuse: A law enforcement view or perspective.
Child Abuse & Neglect, 15, 171-173.
Leavitt, F. & Labott, S. (in press). Revision of the Word Association Test
for assessing association of patients reporting satanic ritual abuse.
Journal of Clinical Psychology.
Noblitt, J.R. (1995). Psychometric measures of trauma among psychiatric
patients reporting ritual abuse. Psychological Reports, 77, 743-747.
Postman, L. & Keppel, G. (1970). Normal word associations. New York:
Academic Press.
Putnam, F.W. (1991). The satanic ritual abuse controversy. Child Abuse &
Neglect, 15, 175-179.
Van Benschoten, S.C. (1990). Multiple Personality Disorder and satanic
ritual abuse: The issue of credibility. Dissociation, 3, 22-33.
Victor, J.S. (1995, June). Satanic panic updates: The dangers of moral
panics. Skeptic Magazine, 3, 2-9.
Young, W.C., Sachs, R.G., Braun, B., & Watkins, R.T. (1991). Patients
reporting ritual abuse in childhood: a clinical syndrome report of 37 cases.
Child Abuse & Neglect, 15, 181-189.
Weir, K. & Wheatcroft, M.S. (1995). Allegations of children’s involvement in
ritual sexual abuse: Clinical experience of 20 cases. Child Abuse & Neglect,
19, 491-505.
AUTHOR NOTE
Frank Leavitt, Ph.D. is a board certified clinical psychologist who
specializes in dissociative disorders. He has served as Section Director of
the Department of Psychology at Rush Medical Center for the past 24 years
and is an associate professor on the faculty of Rush Medical College.
Send correspondence and reprint requests to:
Dr. Frank Leavitt
Department of Psychology
Rush Medical College, 1653 West Congress Parkway,
Chicago, Illinois 60612, U.S.A
e-mail: fleavitt@rush.edu
FAX: (312) 942-4990
