[ Pobierz całość w formacie PDF ]
Psychological Trauma: Theory, Research,
Practice, and Policy
Betrayal Trauma and Borderline Personality
Characteristics: Gender Differences
Laura A. Kaehler and Jennifer J. Freyd
Online First Publication, August 15, 2011. doi: 10.1037/a0024928
CITATION
Kaehler, L. A., & Freyd, J. J. (2011, August 15). Betrayal Trauma and Borderline Personality
Characteristics: Gender Differences.
Psychological Trauma: Theory, Research, Practice, and
Policy
. Advance online publication. doi: 10.1037/a0024928
Psychological Trauma: Theory, Research, Practice, and Policy
© 2011 American Psychological Association
2011, Vol.
●●
, No.
●
, 000 – 000
1942-9681/11/$12.00
DOI: 10.1037/a0024928
Betrayal Trauma and Borderline Personality Characteristics:
Gender Differences
Laura A. Kaehler and Jennifer J. Freyd
University of Oregon
Borderline Personality Disorder has been linked to both trauma and insecure attachment styles. Betrayal
Trauma Theory proposes those who have experienced interpersonal trauma may remain unaware of
betrayal in order to maintain a necessary attachment. This study attempts to replicate the association
between self-reported betrayal trauma experiences and borderline personality characteristics found by
Kaehler and Freyd (2009); however, this project includes participants from a community, rather than a
college, sample. Using multiple regression, all three levels of betrayal (high, medium, and low) and
gender were significant predictors of borderline personality characteristics. Separate regression analyses
were conducted for both genders to explore the associations of betrayal trauma on these traits. For men,
all three levels of betrayal trauma were significant predictors; for women, only high and medium betrayal
traumas were significant. These findings suggest trauma may be a key factor of borderline personality
disorder, with differential effects for betrayal and gender.
Keywords:
borderline personality disorder, betrayal trauma, gender
The defining characteristics of Borderline Personality Disorder
(BPD) are volatile interpersonal relationships, identity confusion,
pronounced emotional lability, and poor impulse control. Preva-
lence rates for this serious mental disorder are approximately 2%
of the general population, 10% of psychiatric outpatients, and
15–20% of clinical inpatients (
Diagnostic and statistical manual of
mental disorders, 4th ed-text revision;
APA, 2000). Yet, an inter-
esting gender effect has been observed with this disorder: 75% of
those diagnosed are women (
DSM–IV–TR
) and approximately
80% of clients receiving treatment for BPD are women (Skodol et
al., 2002). Due to both the large numbers of people affected by this
disorder and the sociocultural factors associated with BPD, many
theories have been suggested regarding the causal factors of this
disorder, including the roles of trauma and attachment.
One consistent finding regarding BPD and attachment is the
association between BPD and insecure attachment styles (Levy,
2005). In fact, some have suggested a disorganization of the
attachment system is a key contributor of BPD features (Gunder-
son & Lyons-Ruth, 2008). A review of studies exploring this
association by Agrawal, Gunderson, Holmes, and Lyons-Ruth
(2004) revealed that adults with BPD most frequently display
either fearful or unresolved (with a secondary classification of
preoccupied) attachment styles. A fearful attachment style is char-
acterized by a desire for intimacy, while simultaneously fearing
hurt or rejection by the partner. Like the fearful attachment style,
a person with an unresolved/preoccupied attachment style also
seeks an intimate relationship, but is instead sensitive to a per-
ceived dependency on the partner. A link between infant insecure
attachment and subsequent development of BPD symptoms has
also been demonstrated (Lyons-Ruth, Yellin, Melnick, & Atwood,
2005; Rogosch & Cicchetti, 2005). Recent findings from a pro-
spective longitudinal study (Carlson, Egeland, & Sroufe, 2009)
showed that a disorganized infant attachment style, which includes
a sequential or simultaneous display of contradictory approach/
avoid behaviors, predicted adult borderline symptoms. A hallmark
feature of relationships with persons diagnosed with BPD is in-
consistencies in thoughts and actions (e.g., “push-pull dynamics”),
similar to the descriptions of attachment styles mentioned above.
Parental abuse and neglect interferes with the development of a
secure attachment style (Baer & Martinez, 2006; Lamb, Gaens-
bauer, Malkin, & Schultz, 1985). For example, Minzenberg, Poole,
and Vinogradov (2006) found all types of childhood maltreatment
to be significantly associated with attachment-avoidance, with
childhood sexual abuse (CSA) also related to attachment-anxiety.
These are classic indicators of insecure attachment. Rates of mal-
treatment as high as approximately 90% have been found in BPD
patients (Zanarini et al., 1997). Looking at CSA specifically, the
prevalence rate of this trauma has been estimated to be as high as
75% in individuals with BPD, including both inpatient and outpa-
tient samples (Battle et al., 2004). It should be noted that children
who have experienced CSA are also at increased risk for experi-
encing other forms of childhood interpersonal violence, for exam-
ple, domestic violence (Bowen, 2000) and physical abuse
(Zanarini, 2000). In fact, physical abuse, emotional abuse, and
neglect are frequently observed and are all associated with BPD
(Battle et al., 2004; Carlson et al., 2009; Herman, Perry, & van der
Kolk, 1989; Johnson, Smailes, Cohen, Brown, & Bernstein, 2000;
Paris, 1997; Trull, 2001). Given the demonstrated links among
trauma, insecure attachment, and BPD, a parsimonious model in
which to explore BPD would incorporate both attachment and
trauma. Betrayal Trauma Theory (BTT; Freyd, 1996) is a concep-
Laura A. Kaehler and Jennifer J. Freyd, Department of Psychology,
University of Oregon.
Correspondence concerning this article should be addressed to Laura A.
Kaehler and Jennifer J. Freyd, Department of Psychology, University of
Oregon, Eugene, OR 97403-1227. E-mail: jjf@uoregon.edu
1
2
KAEHLER AND FREYD
tual framework that posits attachment as central to the impact of
trauma.
Bowlby (1988, p. 121) suggests the attachment relationship “has
a key survival function of its own, namely protection.” BTT
utilizes that premise to propose individuals may isolate knowledge
about betrayals, such as those that occur during maltreatment, in
order to maintain a relationship necessary for survival (Freyd,
1996). Typically, it would be advantageous to be able to detect
betrayal in order to prevent a future violation; however, one may
find it more adaptive to remain unaware of that violation if that
knowledge threatens a person’s more immediate viability (Freyd,
1996). Specifically focusing on the parent– child relationship for a
moment, the parent is solely responsible for ensuring the physical
and psychological needs of the child are met. Thus, if the parent
betrays the child—for example, via physical abuse—the child must
still remain attached to the caregiver (and ensure the caregiver is
attached to the child) in order to survive. Although a natural
response to betrayal is to withdraw, if the child would react in that
manner, his or her life would be in peril. Therefore, the child may
remain unaware of that betrayal in an effort to maintain that
necessary connection to the parent. Freyd (1996) has suggested
one mechanism by which this knowledge isolation may occur is
dissociation, defined by Bernstein and Putnam (1986, p. 727) as “a
lack of normal integration of thoughts, feelings, and experiences
into the stream of consciousness and memory.” Interestingly,
severe dissociation is one of the nine diagnostic criteria for BPD
and some have suggested it is a key, distinguishing component of
BPD (Ross, 2007; Skodol et al., 2002; Wildgoose, Waller, Clarke,
& Reid, 2000; Zweig-Frank, Paris, & Guzder, 1994).
Kaehler and Freyd (2009) explored the association between
borderline personality characteristics and betrayal trauma experi-
ences within a college population. Using the Brief Betrayal
Trauma Survey (BBTS; Goldberg & Freyd, 2006) and a modified
version of the Borderline Personality Inventory (BPI; Leichsen-
ring, 1999), results showed high- and medium-betrayal traumas
significantly predicted borderline personality characteristics, while
low-betrayal traumas did not. Interestingly, these results were
found while controlling for gender, which was not associated with
borderline personality characteristics.
The authors suggest the gender differences observed in BPD
diagnoses may be attributable to the nature of the trauma experi-
enced, rather than differences between genders. As reported in
Goldberg and Freyd (2006), women tend to experience more
high-betrayal traumas, while men experience more low-betrayal
events. Work by Johnson and colleagues (2003), as part of the
Collaborative Longitudinal Personality Disorders Study (CLPS),
support this interpretation. These researchers compared men and
women who met diagnostic criteria for BPD and found no signif-
icant differences between the two groups for rates of childhood
sexual abuse, physical abuse, and witnessing abuse. However, the
authors do not report on the closeness of the relationship to the
perpetrator, so it is not possible to differentiate between the levels
of betrayal present. A recent study by Woodward, Taft, Gordon,
and Meis (2009) showed that clinicians, when evaluating ambig-
uous case vignettes of a person who experienced childhood sexual
abuse that included symptoms of both BPD and PTSD, were not
more likely to assign women the BPD diagnosis. Thus, clinicians
were responding to the nature of the trauma, not the client’s
gender, when interpreting the case.
In the current study, we attempt to replicate the findings of
Kaehler and Freyd (2009) using a community sample. Thus, it was
predicted that high- and medium-betrayal trauma would be asso-
ciated with borderline personality characteristics (with high-
betrayal as the stronger predictor), while low-betrayal trauma
would not be a significant predictor.
Method
Participants
Participants were members of the Eugene-Springfield Commu-
nity Sample (ESCS), who were recruited via direct mailings from
a list of homeowners in 1993. Since then, this group has been sent
questionnaires yearly; in 2000, the 16-page questionnaire included
items related to trauma history and borderline personality charac-
teristics. Of the approximately 850 participants who were admin-
istered this questionnaire, 749 participants (57% women) com-
pleted them. The mean age of the sample was 50.7 years old (
SD
12.6). The sample predominantly identified as Caucasian (96.0%)
and a majority was married (80%). This was a highly educated
(23% postgraduate degrees) sample, with 43% of participants
employed full-time (20% retired). Please see Goldberg (1999a,
1999b) and Goldberg and Freyd (2006) for additional information
regarding the ESCS.
Materials and Procedure
As in the Kaehler and Freyd (2009) study, the Brief Betrayal
Trauma Survey (BBTS; Goldberg & Freyd, 2006) and a modified
version of the Borderline Personality Inventory (BPI; Leichsen-
ring, 1999) were used. The BBTS is a 12-item, self-report measure
of major traumatic events participants may have experienced dur-
ing two time periods (before and after age 18). Each item is
classified as having one of three levels of betrayal: low, medium,
or high. Noninterpersonal traumas (e.g., motor vehicle accidents)
are viewed as a low-betrayal, while interpersonal traumas (e.g.,
being attacked) are considered a medium- or high-betrayal. To
distinguish high-betrayal items from medium-betrayal items, the
closeness of the perpetrator is assessed. An example of a low-
betrayal item is “been in a major motor vehicle accident that
resulted in significant loss of personal property, serious injury,
death, or fear of own death.” An example of a medium-betrayal
item is “you were deliberately attacked so severely as to result in
marks, bruises, blood, or broken bones by someone with whom
you were
not
close [italics added].” An example of a high-betrayal
item is “you were deliberately attacked that severely by someone
with whom you were
very
close [italics added].” Respondents
select the frequency of the event from three options (“never”, “one
or two times”, or “more than that”). The BBTS has been demon-
strated to have both good construct validity (DePrince & Freyd,
2001) and test–retest reliability (Goldberg & Freyd, 2006). The
authors were interested in first-hand trauma experiences (i.e., not
witnessing) and so used 7 of the 12 items in data analyses. Two
items were classified as low-betrayal, two as medium, and three as
high-betrayal. This version of the BBTS differed from the version
used in the Kaehler and Freyd (2009) study in two ways: the age
categories of the events and the response choices. In the previous
study, there were three age categories (“before age 12”, “age 12
BORDERLINE PERSONALITY AND BETRAYAL
3
through 17”, and “age 18 and older”) for each event, while for this
study there were only two age categories (“before age 18” and
“age 18 or older”). Respondents simply selected whether the event
occurred (“yes”) or did not (“no”) in the Kaehler and Freyd (2009)
study; in this study, participants selected the frequency it occurred
from three response choices (“never”, “one or two times”, and
“more than that”). Scores could range from 0 to 12 for high-
betrayal and 0 to 8 for low- and medium-betrayal.
The BPI is a 53-item self-report measure assessing characteris-
tics typical of those diagnosed with BPD. Results may be inter-
preted with either a dimensional or categorical approach. Sample
items include: “my feelings toward other people quickly change
into opposite extremes (e.g., from love and admiration to hate and
disappointment); if a relationship gets close, I feel trapped; and I
enjoy having control over someone”). The BPI has good internal
consistency and test–retest reliability (Leichsenring, 1999). The
possible responses were revised from the original scale; instead of
marking a statement as true/false, participants were given a 5-point
Likert-type scale, where 1
very inaccurate
,2
moderately
inaccurate
,3
neither accurate nor inaccurate
,4
moderately
accurate
, and 5
very accurate
. This revision was made in an
effort to increase the variability of participant’s scores. Further-
more, only 47 of the original 53 items were administered, omitting
items that would require follow-up questions.
lated with medium-betrayal (
r
.282,
p
.01) and medium-
betrayal with low-betrayal (
r
.201,
p
.01). Gender was
significantly associated with high-betrayal (
r
.263,
p
.01) and
low-betrayal (
r
.102,
p
.01), but not with medium-
betrayal (
r
.029,
p
NS). Women (
M
1.94,
SD
2.56) had
higher scores for high-betrayal experiences than men (
M
.751,
SD
1.55); however, men had higher values (
M
.787,
SD
1.14) for low-betrayal events compared to women (
M
.421,
SD
.768) (see Figure 1).
Borderline Personality Traits
The mean BPI total score was 80.6 (
SD
18.8), median was
77.0, with a range of 48 to 162 (possible score range of 47 to 235).
There was significant positive skew (skewness
1.09,
SE
.089). Thirty of the items had 50% or more respondents endorsing
that item as
very inaccurate
. Twenty of the items had less than 1%
of participants indicating that the statement was
very accurate
. The
scale had good reliability, Cronbach’s alpha
.892. Ethnicity (
r
.022,
p
NS), gender (
r
.065,
p
NS), employment status (
r
.056,
p
NS), and marital status (
r
.043,
p
NS) were not
significantly correlated with BPI score. All three levels of betrayal
events were significantly correlated with BPI score: low (
r
.177,
p
.01), medium (
r
.241,
p
.01), and high (
r
.249,
p
.01).
Procedure
Multiple Regression
As previously mentioned, participants completed both instru-
ments by mail as part of a larger, longitudinal data collection
project. Participants are compensated for their time. While com-
pleting questionnaires, respondents may decline to answer any
item. After completing the survey, each questionnaire was marked
with an identification number, which was used during data entry so
responses are anonymous when analyzing data. IRB approval was
obtained prior to the mailing of the questionnaires as well as for
this specific series of data analyses.
To test the hypotheses of interest, a multiple regression analysis
was conducted. The model, which included gender and the three
levels of betrayal trauma, did significantly predict a portion of the
variance in borderline traits,
F
(4, 581)
21.2,
p
.001, adjusted
r
2
.122. Results of the regression analysis are shown in Table 1.
As predicted, high-betrayal trauma was the largest predictor of
borderline personality features,
.201,
t
(581)
4.74,
p
.001.
Medium-betrayal trauma was also a significant predictor (
.176,
t
(581)
4.24,
p
.001). Contrary to predictions, both
gender (
.099,
t
(581)
2.43,
p
.05) and low-betrayal
(
.126,
t
(581)
3.15,
p
.01) did significantly predict
variance in BPI scores.
Separate regression models were conducted for men and women
to examine how betrayal trauma may differentially associate with
borderline personality characteristics based on gender. For men,
the model, including all three levels of betrayal trauma, did sig-
Results
Trauma History
Fifty-one percent of the sample indicated they experienced at
least one type of first-hand trauma and 35.8% experienced at least
one high-betrayal trauma. High-betrayal was significantly corre-
Figure 1.
Frequency of betrayal traumas by gender.
4
KAEHLER AND FREYD
Table 1
Multiple Regression for Betrayal Trauma Predicting BPI
(N
749)
Variable
B
SE
B
Gender
3.71
1.53
0.99
High BT
1.74
0.37
0.20
Medium BT
4.01
0.95
0.18
Low BT
2.24
0.71
0.13
Note. R
2
.128, adjusted
R
2
.122.
p
.05.
p
.01.
p
.001.
nificantly predict a portion of the variance in borderline traits,
F
(3,
264)
14.4,
p
.001, adjusted
r
2
Figure 2.
Types of betrayal trauma and standardized male BPI score
using standardized
.141. Results of the
regression analysis are shown in Table 2. Medium-betrayal trauma
was the largest predictor of borderline personality features,
.201,
t
(264)
3.27,
p
.001. Both high-betrayal (
.154,
t
(264)
2.55,
p
.05) and low-betrayal (
.169,
t
(264)
2.83,
p
.01) also significantly predicted variance in BPI scores (see
Figure 2).
As with men, the model including the three levels of betrayal
trauma did significantly predict a portion of the variance in bor-
derline traits,
F
(3, 314)
13.0,
p
.001, adjusted
r
2
.
found by Kaehler and Freyd (2009). However, this pattern was
only found for women. Thus, for men, the experience of any type
trauma results in dysfunctional responses, while for women, only
interpersonal traumas do.
DePrince and Freyd (2002) discuss how those in a less powerful
position may perceive and respond to betrayal violations more so
than those in a more powerful position at the time of the event.
Thus, for men, who typically hold positions of power, experienc-
ing any type of trauma may be their first exposure to feelings of
disempowerment. A corollary of this is the “just world” construct,
the belief the world is “just” and so people are rewarded or
punished based on their actions (Janoff-Bulman, 1989; Lerner &
Miller, 1978). Men (O’Connor, Morrison, McLeod, & Anderson,
1996; Swickert, DeRoma, & Saylor, 2004), older Americans (Cal-
houn, Cann, Tedeschi, & McMillan, 1998), and European Amer-
icans (Calhoun, & Cann, 1994) tend to have higher just world
beliefs than women, the younger generation, and African Ameri-
cans, respectively. Janoff-Bulman (1989), in her work with trauma
survivors, found that most people usually operate on the basic
belief of invulnerability (including the idea of a just world), which
is changed after experiencing a traumatic event. Thus, perhaps for
people who have stronger beliefs in a just world (e.g., men),
experiencing any type of event that threatens this belief results in
maladaptive coping. Recent work by Giesen-Bloo and Arntz
(2005) found that BPD patients hold significantly lower beliefs of
benevolence of people and benevolence of the world compared to
patients with either a Cluster C Personality Disorder, Axis I
Disorder, or member of the control group, and that these world
assumptions are more attributable to the borderline psychopathol-
ogy of the person, rather than the experience of trauma. Differ-
.111.
Results of the regression analysis are shown in Table 3. High-
betrayal trauma was the largest predictor of borderline personality
features,
.229,
t
(314)
4.02,
p
.001. Medium-betrayal
trauma was also a significant predictor (
.144,
t
(314)
2.52,
p
.05). However, low-betrayal (
.081,
t
(314)
1.50,
p
.135) did not significantly predict variance in BPI scores (see
Figure 3).
Discussion
These findings support an association between betrayal and
borderline personality traits. As predicted, high-betrayal traumas
were the largest contributor to explained variance of borderline
characteristics, and medium-betrayal traumas also significantly
predicted borderline features. However, unexpectedly, experienc-
ing traumas with low-betrayal accounted for borderline variance,
as did gender. When exploring the relationship between betrayal
trauma and borderline personality separately for men and women,
interesting results are found. For men, all three types of betrayal
traumas predict borderline personality characteristics; however, for
women, only experiencing high- and medium-betrayal events pre-
dict those characteristics. It was hypothesized that high-and
medium-betrayal trauma would be associated with borderline per-
sonality characteristics, while low-betrayal would not be, as was
Table 2
Multiple Regression With Betrayal Trauma Predicting BPI for
Men (N
317)
Table 3
Multiple Regression With Betrayal Trauma Predicting BPI for
Women (N
432)
Variable
B
SE
B
Variable
B
SE
B
High BT
1.92
0.76
0.15
High BT
1.73
0.43
0.23
Medium BT
4.15
1.27
0.20
Medium BT
3.63
1.44
0.14
Low BT
2.75
0.97
0.17
Low BT
1.58
1.05
0.08
p
.05.
p
.01.
p
.001.
p
.05.
p
.001.
[ Pobierz całość w formacie PDF ]

Wątki

Search