THE DISSOCIATIVE
EXPERIENCES SCALE:
FURTHER REPLICATION
AND VALIDATION
Edward]. Frischholz, Ph.D.,
Bennett G. Braun, M.D.,
Roberta G. Sachs, Ph.D.,
Laura Hopkins, Denise M. Shaeffer,
Jennifer Lewis, Frank Leavitt, Ph.D.,
James N. Pasquotto, M.A.
and David R. Schwartz
Edward J. Frischholz, Ph.D., is Director of Research at the
Dissociative Disorders Program, Rush-Presbyterian-St. Luke's
Medical Center and is an adjunct professor of psychology at
the University ofIllinois at Chicago. Bennett G. Braun, M.D.,
is Director of the Dissociative Disorders Program, Rush-
Presbyterian-St. Luke's Medical Center, Dissociative Disor-
ders Inpatient Unit, Rush North Shore Medical Center, and
is an adjunct professor of psychology at the University of
Illinois at Chicago. Roberta G. Sachs, Ph.D., is the Director
of Training at the Dissociative Disorders Inpatient Unit,
Rush North Shore Medical Center. Laura Hopkins, Denise
M. Shaeffer, andJennifer Lewis are affiliated with the Disso-
ciative Disorders Program, Rush-Presbyterian-St. Luke's
Medical Center, and University of Illinois at Chicago. Frank
Leavitt, Ph.D., is at the Department of Psychology Rush-
Presbyterian-St. Luke's Medical Center, Chicago, Illinois.
James N. Pasquotto, M.A., is associated with the Illinois
University Affiliated Program in Developmental Disabilities,
University of Illinois at Chicago. David R. Schwartz, B.A., is
associated with the Dissociative Disorders Inpatient Unit,
Rush North Shore Medical Center, Skokie, Illinois.
For reprints write Edward J. Frischholz, Ph.D., Director of
Research, Dissociative Disorders Program, Rush-North Shore
Medical Center, 9600 Gross Point Road, Skokie, Illinois
60076.
ABSTRACT
The purpose of the present study was to provide further evidence in
support ofthe validity ofthe Dissociative Experiences Scale [DES] as
a reliable measure of dissociative psychopathology. The DES was
administered to 259 college students, 33 patients with multiple
personality disorder (MPD), and 29 patients with a dissociative
disorder not otherwise specified (DDNOS). The inter-rater reliability
for the DES scoring procedure was excellent (coefficient of absolute
agreement =. 96; coefficient ofrelative agreement =. 99). The test retest
reliability ofDES scores (within approximately one month) was also
excellent (coefficient ofabsolute agreement =. 93; coefficient ofrelative
agreement=.96) and suggests that DES total scores are temporally
stable and similar in absolute value across testings. Finally, the
internal consistency of DES scores was also very high (alpha for
students=.93; alpha for MPD patients =.94; alpha for DDNOS
patients=.94; alpha for the combined total sample=. 95). Both MPD
(meanDESscore=55.0)andDDNOSpatients(meanDESscore=40.8)
earned significantly higherDESscores than students (mean DESscore
=23.8). In addition, MPD patients earned significantly higher DES
scores than DDNOS patients. The results of the present study also
suggest that a DES cutoffscore of45 to 55 maximizes the probability
of correctly classifYing students from dissociative disorder patients
(87%) while minimizing the rates offalse positive (2 to 6%) and
false negative errors (7 to 11 %). Suggestions for further validation
research are also made.
The Dissociative Experiences Scale (DES: Bernstein &
Putnam, 1986) is a brief, 28 item, self-report inventory of
both normal and abnormal experiences. The DES was de-
veloped "to offer a means of reliably measuring dissociation
in normal and clinical populations" (Bernstein & Putnam,
1986, p. 727). Preliminary psychometric data reported by the
authors of the scale indicated that summary scores were
temporally stable (r [after 4 to 8 weeks] =.84) and successful-
ly discriminated patients with multiple personality disorder
(MPD) from normals and other pathological groups (Bern-
stein & Putnam, 1986).
Two laboratories have since independently confirmed
the discriminant validity of the DES as a measure ofdissocia-
tive psychopathology relative to different kinds of normal
comparison groups. For example, Ross, Norton, and Ander-
son (1988) reported that the median DES score of 17 MPD
patients was 40.7. Although this score was significantly lower
than the median DES score of 57.06 originally reported by
Bernstein and Putnam (1986) for their 20 MPD subjects, it
was still significantly higher than the median score of 4.9
observed for their normal subjects (medical students). In a
later study, Ross, Ryan, Anderson, Ross, and Handy (1989)
reported that normal median DES scores could vary from a
low of 4.8 (for elderly subjects), 7.9 (for college student
subjects), to a high of 17.7 (for adolescent subjects). DES
scores were inversely related to age and no significant differ-
ences were observed for sex.
Ensink and van Otterloo (1989) reported that 7 MPD
patients earned a median DES score of55.4while 13 patients
with a diagnosis of Dissociative Disorder not otherwise spec-
ified (DDNOS) earned a median DES score of40.7. The median
DES scores for botl1 dissociative groups, however, we;e sig-
nificantly higher than the median DES score of 16.3 for 80
psychology students.
Collectively, the findings from three different laborato-
ries have each demonstrated that patients with dissociative
psychopathology (either MPD or DDNOS) earn significantly
higher DES scores than various normal comparison groups.
However, there was significant variation among DES scores
within both the dissociative (range ofmedian DES scores=40.7
to 57.06) and normal groups (range of median DES
scores=4.38 to 17.7). This variation makes it difficult to
151
DlSSOCIATIOK, Vol. III, ~o. 3: September 1990
I
FURTHER REPllCATION & VALIDATION . . I
TABLE 1
Parametric Data for DES Scores Among
Three Different Groups
Group N Mean (Sd) Median
MPD 33 55.0 (19.2) 53.9
DDNOS 29 40.8 (18.3) 44.8
STUDENTS 259 23.8 (14.1) 22.9
estimate an optimal cutoffscore for separating normals from
patients suffering from some type of dissociative disorder.
The purpose of the present study was fourfold:l) to
estimate the inter-rater reliability, temporal stability, and
internal consistency of DES scores in both normal and clin-
ical groups; 2) to provide additional evidence that normals
earn significantly lower DES scores than do patients with
different kinds of dissociative psychopathology; 3) to dem-
onstrate that MPD patients earn significantly higher DES
scores than do patients with other types of dissociative
psychopathology (i.e., DD OS); and 4) to examine the
consequences ofusing differen tDES cutoffscores to discrim-
inate between normals and dissociative disorder patients.
METHOD
Procedure
The DES was routinely administered on an individual
basis to 62 inpatients at the Dissociative Disorders Inpatient
Unit of Rush North Shore Medical Center by either a psy-
chologist, nurse, or mental health worker. Thirty-three of
. these patients met both DSM-III (American Psychiatric Asso-
ciation, 1980) and DSM-III-R (American Psychiatric
Association, 1987) criteria for a diagnosis of MPD
while the other 29 met these criteria for a diagnosis of
DDNOS. Ninety-five percent of this patient group
were female and the average age was 35.1 (sd=6.17).
Forty-four percent had graduated from college and/
or had earned an advanced college degree.
The 259 college students were all undergradu-
ates from the University of Illinois at Chicago who
completed the DES as partofastudyon hypnotizability
and personality in order to earn course credit. Seventy-
five percent of this group were female and their
average age was 19.8 (sd=3.6).
tive judgement on the part of the rater, the intra-class
correlation method recommended by Shrout&Fleiss (1979)
was used to estimate absolute and relative indices of inter-
rater agreement. Twenty DES protocols were independently
scored by four raters. The in tra-class correlation of absolute
agreement (called ICC[l,I] by Shrout & Fleiss, 1979 p. 423)
was .96 while the coefficient of relative agreement (called
ICC[3,1] by Shrout & Fleiss, 1979, p. 423) was .99.
The temporal stability of DES total scores was also as-
sessed using an intra-class correlation model. Thirty DD pa-
tients (15 MPD and 15 DDNOS) completed the DES twicewithin
approximately one month's time. The absolute temporal
stability coefficient was .93, while the relative temporal sta-
bility coefficient was .96. These findings indicate that DES
total scores are highly stable over time and that there are no
significant mean differences between testings.
Cronbach'salphacoefficient (Cronbach, 1951) was used
to estimate the internal consistency ofDES scores. The alpha
coefficient for the 33 MPD DES scores was .94, .94 for the 29
DDNOS DES scores, and .93 for the 259 student DES scores. In
addition, the alpha for the combined total sample ofpatients
and students (n=321) was .95.
TABLE 2
Classification Rates and Errors for Different DES Cutoff Scores
DES % % %
Cutoff Correct False Positive False Negative
Score Classification Error Error
10 34 65 01
15 45 54 01
20 54 43 02
25 63 34 03
30 70 25 05
35 76 18 06
40 82 12 06
45 87 06 07
50 87 03 10
55 87 02 11
60 85 01 14
Inter-rater Reliability, Temporal Stability, and
Internal Consistency ofDES Scores
Since the scoring of DES items involves a subjec-
Scoring of the DES
The DES was scored according to the graphic
rating method originally proposed and described by
Bernstein and Putnam (1986). The sum of the 28
items was divided by 28 to form an overall DES score
for each individual.
RESULTS
Statistical Analysis
Although Bernstein and Putnam (1986) ongl-
nally recommended using nonparametric data ana-
lytic techniques with DES scores, they now seem to
advocate the use of more powerful statistical proce-
dures based on a general linear model (Eve Carlson,
personal communication,june 29,1990). Therefore,
all of the data analyses calTied out in the present study
used an analysis of variance model (e.g., Guilford &
Fruchter, 1978).
152
DISSOCIATION, Vol. III, No.3: September 1990
-FRISCHHOLZ/BRAUN/SACHS/HOPKINS/SHAEFFER/LEWIS/LEAVITT/PASQUOTTO/SCHWARTZ
Discriminant Validit~ ofDES Scores
The means, standard deviations, and median DES scores
for each patient group and the student sample are presented
in Table 1. MPD patients earned a mean DES score of 55.03
(sd=19.2). DDNOS patients earned a mean DES score of 40.8
(sd=18.3), and the students earned amean DES score of23.8
(sd=14.1). A one-way analysis ofvariance using group as the
independent variable and DES scores as the dependent vari-
able indicated that there were significant differences be-
tween groups (F[2,318]=73.0, p <.00001). Two a priori
planned contrasts were then carried out to ascertain the
nature of these differences. The first con trast indicated that
the mean DES scores for the combined dissociative patient
group was significantly higher than the mean DES score for
the studentgroup (1[318]=11.3, p<.OOI, one tailed test). The
second contrast indicated that the mean DES score for the
MPD patients was significantly higher than the mean DES
score for DDNOS patients (1[318]=3.68, p <.001, one tailed
test). Thus, the present findings again confirm the discrim-
inant validity of DES scores. Patients with dissociative psy-
chopathology earn significantly higher DES scores than do
normals and MPD patients earn significantly higher DES
scores than do patients with other types of dissociative
psychopathology.
Developing DES CutoffScores
In order to assess the consequences of using different
DES cutoff scores to discriminate between patients with
dissociative psychopathology and normals, one must consid-
er three different values: 1) the percentage correctly classi-
fied; 2) the percentage of false positive error; and 3) the
percentage of false negative error. Table 2 presents these
values for various DES cutoff scores to discriminate between
the students and the combined patient group.
Examination ofTable 2 suggests that a DES cutoff score
of 45 to 55 maximizes the percentage of correctly classified
(87%) while minimizing the false positive (2 to 6%) and false
negative (7 to 11 %) error rates.
DISCUSSION
Collectively, the results of the present study provide
further support for the utility of the DES as a measure of
dissociative psychopathology. The inter-rater reliability for
the DES scoring procedure is excellent (absolute agreement
=.96; relative agreement=.99) and DES scores were found to
be highly internally consistent (.94 in the MPD group; .94 in
the DDNOS group; .93 in the student group; and .95 in the
combined total sample). As reported earlier, patients with
dissociative disorders earn significantlyhigher DES scores than
do normals (Bernstein & Putnam, 1986; Ensink & van Otter-
10,1989; Ross, Norton, & Anderson, 1988) and MPD patients
earn significantly higher DES scores than do patients with
other kinds of dissociative psychopathology (Enskink & van
Otterlo, 1989). The findings also suggest that a DES cutoff
score of 45 to 55 will optimize the correct classification of
patients from normals while minimizing the rates of false
positive and false negative elTors. In this regard, one must
also consider the possibility that some of the student group
may actually have had some type of DSM-III-R dissociative
disorder. However, the present authors consider this possi-
bility to be extremely unlikely for two reasons. First, the
incidence ofdissociative disorders in the general population
is very rare. Second, the first author worked in the same
psychology department clinic over a period of four years
where the present student sample was drawn. During this
time period, not one student was ever diagnosed as suffering
from any type of dissociative disorder.
The results of the present study are encouraging and
suggest that the DES may be a valuable instrument in the
differential diagnosis ofdissociative disorders from normals.
However, additional research is still necessary to further
substantiate the utility of the DES in discriminating disso-
ciative disorder patients from patients with other kinds of
psychopathology.
The authors would like to gratefully acknowledge the
assistance of Eve Carlson, Ph. D., who served as one of the
four raters in the present study. •
REFERENCES
American Psychiatric Association (1980). Diagnostic and statistical
manual ofrnental disorders (3rd ed). Washington, DC: Author.
American Psychiatric Association. (1987). Diagnostic and statistical
manual ofmental disorders (3rd. - revised). Washington, DC: Author.
Bernstein, E.M., & Putnam, F.W. (1986). Development, reliability,
andvalidityofa dissociation scale.joumalofNervousandMentaLDisease,
174, 727-734.
Cronbach, LJ. (1951). Coefficientalpha and the internal structure of
tests. Psychometrika, 16, 297-334.
Ensink, BJ., &van Otterloo, D. (1989). Avalidation ofthe dissociative
experiences scale in the Netherlands. DISSOCIATION, 2,221-224.
Guilford,]. P., &Fruchter,B. (1978) .Fundamentalstatisticsinpsychology
and education. New York: McGraw Hill.
Ross, CA, Norton, G.R., & Anderson, G. (1988). The dissociative
experiences scale: a replication study. DISSOCIATION, 1 (3),21-22.
Ross, C. A., Ryan, L., Anderson, G., Ross, D. , & Hardy, L. (1989).
Dissociative experiences in adolescents and college students.
DISSOCIATION, 2, 240-243.
Shrout, P.E., & Fleiss,]. L. (1979). Intraclass correlations: uses in
assessing rater reliability. Psychological Bulletin, 86, 420-428.
153
DISSOCIATION, Vol.llI, ~o. 3: September 1990