Borderline Versus Other

Personality Disorders in the

Eating Disorders: Clinical
Description

Stephen A. Wonderlich, Ph.D.

William J. Swift, M.D.
(Accepted 15 September 1989)

The authors compared eating-disordered patients with borderline personality disor-

der to those patients with other personality disorders or none at all. Borderline sub-
jects differed from other personality disorders in several ways; most notably their
histories of sexual abuse, self-mutilation, suicide gestures, and perceived hostile pa-
rental relationships. Borderline subjects failed to differ from other personality-disor-
dered subjects in eating attitudes and showed only limited differences in eating
symptoms. The results are discussed also in terms of their support for Benjamin’s
theory of borderline personality.

Several recent studies suggest that borderline personality is quite common in

bulimic samples, although the prevalence rate across studies has been variable,
ranging from 0% to 41% for outpatients and from 29% to 40% for inpatients
(Levin & Hyler, 1986; Yates, Sieleni & Reich, 1988; Frankel, Sadik, Dantzier,
Charles, Roose, and Walsh, 1988; Johnson, Tobin, & Enright, 1989). Similarly,
studies of the comorbidity of borderline personality and anorexia nervosa have
revealed moderately high prevalence rates for both bulimic (Gartner, Marcus,
Halmi, & Loranger, 1990; Piran, Lerner, Garfinke1,Kennedy& Brouillette, 1988;
Wonderlich ,Swift, Slotnick & Goodman, 1990) and restricting (Piran et al.,
1988; Wonderlich et al., 1990) subtypes. However, Pope and his colleagues
(Pope, Frankenburg, Hudson, Jonas, & Yergulun-Todd, 1987, Pope & Hudson,
1989) found that only 2% of outpatient bulimics evidenced borderline person-

Stephen A. Wonderlich, Ph.D. is Assistant Professor, Department of Neuroscience, University of North

Dakota School of Medicine and is also Co-Director of the Fargo Clinic Eating Disorders Program.
William j. Swift, M.D. is Associate Professor of Psychiatry and Acting Chief of the Division of Child and
Adolescent Psychiatry at the University of Wisconsin School of Medicine. A preliminary version of this
paper was presented at the meeting of the American Psychological Association, New Orleans, Louisana.
August 10-1 5, 1989. Address correspondence and reprint requests to the first author at the Department
of Neuroscience, UND School of Medicine, 1919 North Elm Street, Fargo, North Dakora 58702, USA.

International journal of Eating Disorders, Vol. 9, No. 6, 629-638 (1990)

0 1990 by John Wiley & Sons, Inc. CCC 0276-3478/90/060629-10$04.00
630 Wonderlich and Swift

ality when a highly specific measure of borderline personality was employed.

They concluded that primary disturbances in mood most likely result in spuri-
ous borderline personality diagnoses that inflate the prevalence estimates of
borderline personality in this population. Although this is certainly one expla-
nation, the findings of Pope et al. (1987) also support the possibility that there
is a high rate of comorbidity between depression and borderline personality in
an eating disorders population (Johnson et al., 1989; Wonderlich et al., 1990).
Clearly, conceptual controversy regarding the borderline-personality-disorder
construct continues to be a prominent empirical issue that contributes to the
differing prevalence estimates in eating disorder samples.
Johnson et al., (1989) provide the first effort to compare the clinical character-
istics of eating disorder patients who display borderline personality to those
who are not borderline. The borderline patients differed minimally from non-
borderline patients in eating symptoms but they showed greater disturbance in
eating attitudes, emotional distress, psychosocial adaptation, family environ-
ment, as well as more self-destructive behavior, and a more complicated treat-
ment history. Even though these findings provide important information
regarding the clinical characteristics of eating-disordered individuals with bor-
derline personality disorder, it remains unclear if these characteristics are spe-
cific to borderline personality disorder or may also be seen in other DSM-III-R
personality disorders. The present study addressed this uncertainty by com-
paring clinical characteristics, psychiatric symptomatology, and family vari-
ables in four groups: eating-disordered subjects with borderline personality
disorder, eating-disordered subjects with a nonborderline personality disorder,
eating-disordered subjects without personality disorder and normal controls.
A second focus for this study was to test empirically Benjamin’s (1987; sub-
mitted) interpersonal model of borderline personality disorder using her Struc-
tural Analysis of Social Behavior (SASB) (Benjamin, 1974, 1984). Based on
Benjamin’s theory and associated circumplex methodology, several predictions
were generated. The first prediction was that borderline subjects would per-
ceive more attack and hostile disengagement in their parental relationships
than would those subjects in other groups. This is based on Benjamin’s notion
that fighting, interpersonal attack and hostile disengagement are common and
etiologically significant in the families of borderlines. Second, it was predicted
that borderlines would perceive their parents as exerting less control and influ-
ence over them than would those in the other groups. This reflects Benjamin’s
emphasis on the considerable parental neglect of the borderline within the
family. Finally, it was predicted that borderlines would be less likely to adopt a
submissive position in relation to their parents. Although Benjamin does not
specifically address submission in her theory of borderline personality, it was
thought that given the high level of neglect and hostile disengagement in the
families, it was unlikely that the borderline subject would submit to her parents.

METHOD

Subjects
Young women (n = 75) between 18 and 34 years of age took part in the
study. Of these subjects, 46 were consecutive adult admissions or patients al-
Personality Disorders 63 1

ready receiving treatment at the University of Wisconsin Hospital and Clinic‘s

Eating Disorders Program. Ten subjects met DSM-111-R criteria for anorexia
nervosa (restricting type), 26 fulfilled criteria for bulimia nervosa (normal
weight), and 10 were diagnosed as bulimic anorexic, fully meeting criteria for
both bulimia and anorexia. This sample is described more fully in a previous
report (Wonderlich et al., 1990).
Patients were interviewed by the first author with the Structured Clinical In-
terview for DSM-111-R Personality Disorders (SCID-11; Spitzer & Williams,
1986). The SCID-I1is a semistructured interview including specific probes assess-
ing criteria for each of the DSM-111-R personality disorders. Subjects may receive
more than one personality disorder diagnosis. The interviewer was aware that
the subjects were eating-disordered, but unaware of the subject’s specific eating-
disorder subtype diagnosis at the time of administration of the SCID-11.
This procedure resulted in 11 subjects meeting DSM-111-R criteria for border-
line personality disorder (BPD), 22 subjects meeting criteria for other personal-
ity disorders (OPD),’ and 13 subjects without personality disorder (NPD). In
addition, 29 control subjects with no history of documented eating or other
psychiatric disorder were recruited from a University of Wisconsin Educational
Psychology course. Eating-disordered subjects were paid $30.00 and controls,
$10.00, for their participation. In order to assess interrater reliability of the
SCID-11, the second author listened to 14 (30%)audiotaped interviews com-
pleted by the first author. Interrater agreement was adequate with complete
agreement on the presence of personality disorder in 12 of 14 interviews. Un-
weighted Kappas ranged from 0.56 to 0.77 for the personality disorders rated
five or more times in the study, which included: Obsessive-Compulsive (K =
0.77), Histrionic (K = 0.75), Borderline (K = 0.74), Dependent (K = 0.66), and
Avoidant (K = 0.56). Disagreements were resolved through concensus decision
following discussion by the two interviewers.

Measures

Diagnostic Survey of Eating Disorders (DSED)

The DSED (Johnson, 1985) is a self-report questionnaire that provides de-
tailed information on demographics, family history, psychiatric history and
eating behaviors.

Medical History Questionnaire (MHQ)

The MHQ (Klein, 1987) is a self-report questionnaire designed to pro-
vide information on the history of medical illnesses, abuse and trau-
ma, and behavioral problems during childhood, adolescence, and adult-
hood.

‘Specific personality disorders in the OPD group are documented more completely in Wonder-
lich et al., (1990). Personality disorders in the OPD group rated five or more times included
obsessive-compulsive, histrionic, dependent, and avoidant. Other personality disorders encoun-
tered less frequently in the OPD group included passive-aggressive, schizotypal, and self-defeat-
ing.
632 Wonderlich and Swift

Symptom Checklist 90-Revised (SCL-90-R)

The SCL-90-R (Derogatis, 1983) is a 90-item self-report measure of psychiatric
symptomatology. It yields eleven scales, reflecting various aspects of psychiat-
ric disturbance.

Eating Disorders Inventory (EDI)

The ED1 (Garner, Olmsted, & Polivy, 1983) is a 64-item self-report measure
that yields eight scales measuring eating attitudes relevant to the eating disor-
ders.

Parent Ratings (Intrex)

Intrex (Benjamin, 1983) is a 144-item relationship rating scale based on Ben-
jamin’s Structural Analysis of Social Behavior (Benjamin, 1974, 1982). Each
item on Intrex corresponds to a point on the full SASB model and depicts one
aspect of interpersonal behavior and is rated on a scale ranging from 0 (not at
all characteristic of the rated relationship) to 100 (perfectly characteristic of the
rated relationship). Based on the pattern of a subject‘s ratings for a particular
relationship, four parameters of interpersonal functioning were derived: attack,
hostile withdrawal, control, and submission. Parameters ranged from - 1.OO
(the relationship is characterized by the opposite of the parameter in question)
to +l.OO (the relationship is characterized by an extremely high degree of the
parameter in question).
The four pattern-coefficients were calculated for subjects’ ratings of the
target’s behavior and again for the subjects’ ratings of their own behavior in
relationship to the target, resulting in eight coefficients for each relationship
rated (e.g., mother attacks me, I attack mother, mother hostilely withdraws
from me, I hostilely withdraw from mother, mother controls me, I control
mother, mother submits to me, I submit to mother). These eight coefficients
served as dependent measures in statistical analyses of each relationship
rated. (No theoretical predictions had been made for parental submission
or the subjects control of parents. These were included for heuristic pur-
poses.)

RESULTS

Nominal level variables were analyzed with a Chi-square procedure and in-
dividual interval level variables were analyzed with Analysis of Variance. The
EDI, SCL-90-R, and Intrex data were analyzed with Multivariate analyses of
variance (MANOVA). Significant multivariate findings were followed by
univariate F tests with Tukey pairwise comparison tests to identify specific
mean differences. Given the large number of tests in each MANOVA, alpha
was conservatively set at .O1 for univariate Fs to reduce the possibility of Type
1 errors.

Demographic Variables
There were no significant differences between the four groups in marital sta-
tus, ~ ’ ( 6 )= 10.5, p < .11, or education level, F(3,71) = 2.41, p < .07 (Ms range
Personalily Disorders 633

from 13.5 to 14.4 years). There was a significant group difference in age,
F(3,71) = 5.27, p < .003, with the BPD subjects slightly older than the OPD and
the control groups (p’s < .05) (mean ages: BPD, 25.9 years (SD = 4.9); OPD,
21.9 years (SD = 4.6); NPD, 21.9 years (SD = 3.2); controls, 20.7 years (SD =
2.7).

Eating-Related Behavior and Treatment History

Comparison of the three eating-disordered groups revealed a significant dif-
ference between groups in length of disorder, F(2,43) = 6.32, p < .004, with
BPD subjects experiencing eating problems longer than the OPD or NPD eat-
ing-disordered groups. There was no difference between the groups in age at
onset of the eating problems: F < 1, weight as a percentage of ideal body
weight; F < 1, binging frequency; F < 1, and frequency of laxative use x2 (4) =
3.8, p < .43. There was a significant group effect on the number of vomiting
episodes per week, F(2,43) = 5.36, p < ,008, with the BPD group engaging in
more vomiting (M = 21.5) than did the NPD group (M = 4.1) but not the OPD
( M = 10.5) group.
The groups did not differ in terms of alcohol consumption, F < 1, or in fre-
quency of previous alcohol treatments x2(2) = 2.49, p < .28. The groups were
equally likely to be hospitalized currently for an eating disorder x2(2) = 2.98,
p < .24, but the BPD and OPD groups were more likely to have been hospital-
ized repeatedly than the NPD group, ~ ‘ ( 2 )= 9.00, p < .01. BPD subjects were
also most likely to have been treated with psychotropic medications, x2(2) =
6.4, p < .04.

Developmental Trauma and Impulsive Behavior

The three eating disordered groups showed no differences in the presence of
physical or sexual abuse before the age of 12. However, 72% ( n = 8) of the BPD
group reported sexual abuse during adolescence, whereas the rates were sig-
nificantly lower for the OPD (27%, n = 6), and NPD (n = 0) groups, ~ ’ ( 2 )=
15.08, p < .001. The groups did not differ in impulsive behaviors during ado-
lescence such as shoplifting, or excessive alcohol use, but more BPD subjects
(63%, n = 7) engaged in self-harmful activity during adolescence than did the
OPD (23%, n = 5) or NPD (23%, n = 3) subjects, x2(2) = 6.33, p < .05. Of the
BPD subjects 72% ( n = 8) had made a suicide attempt, whereas 46% ( n = 10) of
the OPD and none of the NPD subjects had made a suicide attempt,
) 94, p < .001.
~ ~ ( =2 13

Eating Attitudes
Results of a one-way (Diagnostic Group: BPD versus OPD versus NPD
versus controls) MANOVA on ED1 variables were significant, F(24,188) = 6.02,
p < .0001. Univariate analyses revealed significant group differences on all of
the ED1 subscales, Fs(3,71) = 6.4 to 37.1, p < .0007 to .0001. As Table 1
reveals, the BPD and OPD groups scored significantly higher on all of the ED1
scales than did controls ( p < .05), but failed to differ from each other on any
scale. The BPD and OPD groups generally scored higher than the NPD group,
634 Wonderlich and Swift

Table 1. Comparison of group means on the ED1

and SCL-90-R.
Group

Measure BPD OPD NPD Control

Eating Disorder Inventory

Drive-for-thinness 16.2" 15.6" 9.7b 4.9c
Bulimia 8.7" 7.5" 4.2"b O.gb
Body dissatisfaction 21.3"b 20.6" 13.2b' 11.1'
Ineffectiveness 18.2" 13.0" 6.Bb 1.9b
Perfectionism 11.6" 10.7" 7.3ab 5.2b
Interpersonal distrust 9.2" 7.8" 2.Bb 2.7b
Interoceptive awareness 18.1" 14.1" 5.1b 2.1b
Maturity fears 6.2" 5.5" 1.9"b 1.6b
SCL-90-R
Somatization 1.7" 1.2"b 0.7b' 0.4'
Obsessive-compulsive 2.5" 1.9" 1.4bc 0.7'
Interpersonal sensitivity 2.8" 2.P 1.4b 0.6'
Depression 3.1" 2.2b 1.4' 0.i"
Anxiety 2.5" 1.7b 0.8' 0.4'
Hostility 1.3" 1.0"b 0.6b' 0.3'
Phobic anxiety 1.6" 0.9"b 0.3bC 0.1"
Paranoia 2.0" 1.3ab 0.6b' 0.5'
Psychoticism 1.8" 1.2b 0.4' 0.4'
Global severity index 2.2" 1.6" O.gb 0.5b
Positive symptom distress 2.7" 2.3" 1.7b 1.4b
level
Note: Dissimilar superscripts across columns indicate significant pair
wise comparisons at the .05level.

but this reached significance only on the drive-for-thinness, ineffectiveness,

perfectionism and interoceptive awareness scales ( p < .05). Except for the
drive-for-thinness subscale, the NPD group did not differ from controls.

Psychological Distress
A one-way (Diagnostic Group: BPD versus OPD versus NPD versus con-
trols) MANOVA on the SCL-90-R subscales was significant, F(33,176) = 4.60,
p < .0001) and univariate analyses on the subscales were all significant,
Fs(3,73) = 9.2 to 27.1, all p < .0001. As Table 1 indicates, the BPD and OPD
groups scored significantly higher on all subscales than did controls ( p < .05),
and the NPD subjects failed to differ from the control group on any scale, ex-
cept interpersonal sensitivity. The BPD subjects distinguished themselves by
reporting more depression, anxiety, and psychotic thinking than did either the
OPD or NPD groups ( p < .05).

Ratings of Parents
Separate one-way (Diagnostic Group: BPD versus OPD versus NPD versus
controls) MANOVAs were conducted for ratings of maternal and paternal rela-
tionships. Multivariate results were significant for ratings of the mother
F(24,182) = 3.06, p < .0001 and father F(24,173) = 2.30, p < .001.
Personality Disorders 635

Mother
Univariate analyses of the mother ratings revealed significant group differ-
ences on maternal attack, F(3,69) = 10.21, p < .0001, and hostile withdrawal,
F(3,69) = 12.43, p < .0001, as well as the subject's attack of mother F(3,69) =
5.19, p < .003, and hostile withdrawal from mother F(3,69) = 15.32, p < .0001.
As Table 2 indicates, the BPD group generally perceived more attack and hos-
tile withdrawal in their relationships to their mother than any other group ( p <
.05). The OPD group also saw their mothers ( p < .05) and themselves ( p < .05)
as being more hostilely withdrawn than did control subjects, but failed to differ
on any measure from the NPD subjects. The NPD subjects did not differ from
controls on any maternal relationship rating.
There were also group differences on perception of mother's submissiveness
F(3,69) = 4.34, p < .007, and the subject's submission to mother, F(3,69) =
4.57, p < .006. Only the BPD subjects saw their mothers as less submissive to
them than did controls ( p < .05). Both the BPD and NPD subjects saw them-
selves as less submissive to their mother than did the OPD or control subjects
( p < .05).

Father
Ratings of father revealed significant group differences on paternal attack
F(3,66) = 4.84, p < .004, and hostile withdrawal F(3,66) = 4.72, p < .005, as
well as the subject's own hostile withdrawal from father F(3,66) = 9.08, p <
.0001. As Table 2 reveals, only the BPD subjects saw their fathers as more at-

Table 2. Comparison of group means on the Intrex ratings

of mother and father.
Group

Parent Rating BPD OPD NPD Control

Mother
Mother attacks me .17" -.56b - .52b -.86b
I attack mother - .43" - .73b' - .67"' -.8T
Mother hostilely .15" - .49b - .61k -.&Ic
withdraws from me
I hostilely .33" - .32b - .57b= - .82'
withdraw from mother
Mother controls me -.02 .23 .22 .38
I control mother - .21 -.19 -.16 .12
Mother submits to me - .50b - .19ab - .OSab .16"
I submit to mother -.22b .37" -.lSb .24"
Father
Father attacks me - .13" - .45"b - .67ab - .79b
I attack father - .29 - .57 - .81 - .74
Father hostilely - .37= -.53" - .5Fb -.84b
withdraws from me
I hostilely .33" - .25b - &Mb' - .75'
withdraw from father
Father controls me - .23" .30ab .2P .46b
I control father -.49 -.17 - .35 - .04
Father submits to me - .46 - .23 - .40 -.a
I submit to father - .07 .21 .05 .18
Note: Dissimilar superscripts indicate significant pairwise comparisons
at the .05 level.
636 Wonderlich and Swift

tacking than did controls ( p < .05). Both the BPD and OPD subjects saw their
fathers and themselves as more hostilely withdrawn than did controls (p <
.05), but this was most characteristic of the BPD subjects. The groups failed to
differ on measures of control or submission for their paternal relationship ex-
cept on ratings of fathers’ control of them, F(3,66) = 4.10, p < .01. Only the
BPD group saw their fathers as significantly less controlling and influencing
than did controls ( p < .05).

DISCUSSION

The present findings highlight the significance of Borderline personality dis-

order in eating-disordered individuals. Borderline subjects showed extreme
disturbances in several areas of assessment and differed from subjects with
other personality disorder diagnoses in a distinct manner. The borderline sub-
jects reported more sexual abuse and self-destructiveness during adolescence
and more suicidal gestures than other personality-disordered subjects. This is
consistent with recent findings suggesting that childhood trauma, and specifi-
cally sexual abuse, is common and etiologically significant in borderline per-
sonality disorder (e.g., Carrol, Schaffer, Spensley, & Abramowitz, 1980;
Herman, Perry, & van der Kolk, 1989). Borderline subjects were distinguished
also from other personality disorders by the extreme attack and hostile with-
drawal they perceived in their parental rela tionships, particularly with their
mothers. Borderline subjects also perceived their mothers and themselves as
much less submissive to each other than did controls, but this failed to clearly
distinguish them from subjects with other personality disorder diagnoses. Al-
though borderlines’ ratings of father were also quite hostile, they differed from
other personality disorders only as to the ratings of the fathers’ hostile with-
drawal from them. Additionally, borderlines saw their fathers as more neglect-
ful and abandoning than did controls, as revealed by their extremely low
control scores. These parental ratings are quite consistent with other studies
that highlight a family environment chardcterized by hostility, conflict, and
emotional impoverishment (e.g., Benjamin & Wonderlich, 1989; Carrol et al.,
1980; Feldman & Guttman, 1984; Frank 15% Paris, 1981; Johnson et al., 1983), and
collectively support perspectives of borderline personality that underscore the
etiological role of hostility and attachment disruption in parental relationships
(e.g., Benjamin, 1987; Feldman & Guttman, 1984).
Measures of eating attitudes failed to distinguish borderline personality dis-
order from other personality disorders. Thus, previous findings indicating that
borderlines show more eating attitude disturbance than nonborderline patients
(e.g. Johnson et al., 1989) may not be characteristic of borderlines specifically,
but that of a broader range of personality pathology, in general. On the SCL-
90-R, however, borderlines did report more depression, anxiety, and psychotic
thinking than did subjects with other personality disorders, thus supporting
descriptions that liighlight the borderline’s affective instability (Pope, Jonas,
Hudson, Cohen, & Gunderson, 1983) and tenuous cognitive skills (Gunderson
& Zanarini, 1987).
Though both personality-disordered groups displayed significant distur-
Personality Disorders 63 7

bances, it was noteworthy that eating-disordered subjects without personality

disorder appeared much less disturbed. In spite of their significant eating-dis-
order symptoms, these subjects denied sexual abuse or self-destructive behav-
ior, and also differed minimally from controls on eating attitudes, emotional
distress, and parental ratings. These findings support the contention that a
subset of eating-disordered individuals are not personality-disordered and gen-
erally show minimal levels of overall disturbance (Johnson et al., 1989).
Finally, the present study provided considerable support for Benjamin’s the-
ory of borderline personality disorder. The first prediction that borderlines
would be characterized by high levels of perceived attack and hostile with-
drawal in their parental relationships was basically supported. The perceptions
of their mothers as being attacking and that of themselves .as being hostilely
withdrawn from both parents were most specific to the borderline group.
These findings support Benjamin’s (1987) idea that for borderlines, ”chaos and
fighting comprise the family norm.” Furthermore, the high rate of sexual abuse
among borderlines found in the present study is consistent with Benjamin’s
hypothesis that borderlines are overtly abused by a parent or relative. The sec-
ond prediction, that borderlines would see their parents as neglectful was up-
held in ratings of father, but mother ratings produced only a trend in the
predicted direction. Borderlines viewed their fathers as less controlling (i.e.,
neglectful) than did controls, but were not distinguished from other personal-
ity disorders in this regard. Thus, even though this is consistent with Ben-
jamin’s notion that the borderline is neglected in the family, this finding is not
specific to borderline personality and is also characteristic of other personality
disorders. Finally, the third prediction, that borderlines would view them-
selves as less submissive to their parents than would the other diagnostic
groups, was supported partially. Borderlines saw themselves as less submis-
sive to their mothers than did controls, but again this was not clearly specific to
borderlines. Taken together, the present findings suggest that the absence of
paternal control (i.e., neglect) and the borderlines refusal to submit to their
mothers are significant features of borderline personality, but not specific,
whereas hostility in the family of the borderline is more markedly distinctive of
borderlines than of other personality disorders.
In summary, the present study suggests that in an eating-disordered
population, patients with borderline personality disorder differ from other
personality-disordered patients in their histories of sexual abuse, self-destruc-
tiveness, and perceptions of hostility in their parental relationships (particu-
larly mother). The study is limited by its assessment of personality disorder
with an instrument of unknown specificity during a symptomatic period,
possibly allowing for state effects. Also, the fact that subjects were in
treatment could have influenced responding. However, the consistency of the
data and its concordance with theories and previous studies of borderline
personality provide some degree of construct validation for the SCID-I1
diagnoses of borderline personality disorder. The data also suggest that future
examinations of the interrelationships of borderline personality and the eating
disorders may profitably explore the role of family relationships and early
traumatic experience. Furthermore, the reasonable support for Benjamin’s
theory of borderline personality suggests that it deserves further attention in
future empirical studies.
638 Wonderlich and Swift

This research was supported by NIMH Mental Health Clinical Services Research Grant
MI317139 that supported S.A.W. as a Postdoctoral Research Fellow, Department of Psy-
chiatry. University of Wisconsin.

REFERENCES
Benjamin, L. S. (1974). Structural analysis of social behavior. Psychological Review, 81, 392-425.
Benjamin, L. S. (1984). Principles of prediction using Structural Analysis of Social Behavior (SASB).
In R. A. Zucker, J. Aronoff, & A. J. Rabin (Eds.), Personality and prediction of behavior, (pp.
121-174). New York Academic Press.
Benjamin, L. S. (1983). The ZNTREX users manual, Parts I and 11. Madison, Wisconsin: Intrex Inter-
personal Institute.
Benjamin, L. S. (1987). An interpersonal approach. (Commentary on a borderline self-mutilator)
Journal of Personality Disorders, I (4), 317-324, 338.
Benjamin, L. S. & Wonderlich, S. A. (1989). Social perceptions and depression in borderline, uni-
polar and bipolar disorders. (Unpublished manuscript). University of Wisconsin.
Benjamin, L. S. Interpersonal diagnosis and treatment: The SASB Approach. (Submitted) New York
Guilford.
Carrol, J., Schaffer, C., Spensley, J. & Abramowitz, S. I. (1980). Family experiences of self-mutilat-
ing patients. American Journal of Psychiatry, 137, 852-853.
Derogatis, L. R. (1983). The SCL-90-R: Administration, scoring and procedures manual 11. Balti-
more: Baltimore Clinical Psychometric Research.
Feldman, R. B., & Guttmann, H. A. (1984). Families of borderline patients: Literal-minded parents,
borderline parents, and parental protectiveness. American Journal of Psychiatry, 141, 1392- 1396.
Frank, H. & Paris, J. (1981). Recollections of family experience in borderline patients. Archives of
General Psychiatry, 38, 1031- 1034.
Frankel, J. S., Sadik, C., Dantzic, S., Charles, E., Roose, S. P. & Walsh, B. T. (1988). The system-
atic study of personality disorders in bulimic patients. Proceedings of the American Psychiatric As-
sociation Annual Meeting, Montreal, Canada, May, 1988.
Gartner, A,, Marcus, R., Halmi, K., & Loranger, A. (1990). DSM-111-R personality disorders in eat-
ing disorder patients. American Journal of Psychiatry, 746, 1585-1591.
Gunderson, J. G. & Zanarini, M. C. (1987).Current overview of the borderline diagnosis, Journal of
Clinical Psychiatry, 48 (supplement), 5- 11.
Herman, J. L., Perry, J. C. & van der Kolk, B. A. (1989). Childhood trauma in borderline person-
ality disorder. American journal of Psychiatry, 146, 490-495.
Johnson, C. (1985). The initial consultation for patients with bulimia and anorexia nervosa. In
D. M. Garner & P. E. Garfinkel (Eds.), Handbook of psychotherapy for anorexia nervosa and bulimia,
New York: Guilford.
Johnson, C., Tobin, D., Enright, A. (1989). Prevalence and clinical characteristics of borderline pa-
tients in an eating disordered population. journal of Clinical Psychiatry, 50, 9-15.
Klein, M. (1987). The medical history questionnaire. (Unpublished document). University of Wis-
consin.
Levin, A. P. & Hyler, S. E. (1986). DSM-111 personality diagnosis in bulimia. Comprehensive Psychi-
atry, 27, 47-53.
Piran, N., Lerner, P., Garfinkel, P. E. Kennedy, S. H. & Brouillette, C. (1988). Personality disor-
ders in anorexic patients. International Journal of Eating Disorders, 7, 589-599.
Pope, H. G., Frankenburg, F. R., & Hudson, J. I., Jonas, J. M. & Yergulun-Todd, D. (1987). Is
bulimia associated associated with borderline personality disorders: a controlled study. journal
of Clinical Psychiatry, 48, 181-184.
Pope, H. G. & Hudson, J. I. (1989). Are eating disorders associated with borderline personality? A
critical review. International Journal of Eating Disorders, 8, 1-9.
Pope, H. G., Jonas, J. M., Hudson, J. I., Cohen, B. M., & Gunderson, J. G. (1983).The validity of
DSM-111 borderline personality disorder. Archives of General Psychiatry, 40, 23-30.
Spitzer, R. L. & Williams, J. B., (1986). Structured clinical interview for DSM-Zll-Personality Disorders
(SCID-ZZj. New York: New York State Psychiatric Institute.
Wonderlich, 5. A., Swift, W. J., Slotnick, H. B., Goodman, S. (1990). DSM-111-R personality disor-
ders in eating disorder subtypes. International Journal of Eating Disorders, 9, 607-616.
Yates, W. R., Sielini, B. & Reich, J. (1988). DSM-111 personality disorders among bulimics. Proceed-
ings of the American Psychiatric Association Annual Meeting, Montreal, Canada, May, 1988.