Psychology of Women Quarterly, 32 (2008), 290–302. Wiley Periodicals, Inc. Printed in the USA.
Copyright

C

2008 Division 35, American Psychological Association. 0361-6843/08

BORDERLINE PERSONALITY TRAITS AND INTIMATE

PARTNER AGGRESSION: AN INTERNATIONAL MULTISITE,

CROSS-GENDER ANALYSIS

Denise A. Hines

Clark University

Although research has consistently shown that men and women use intimate partner aggression (IPA) at approximately
equal rates, there is little empirical research on whether the predictors of IPA are the same for men and women. The
current study investigated whether borderline personality (BP) differentially predicted the use of IPA for men and
women across 67 university sites from around the world. Results showed that BP predicted several forms of IPA and
that gender did not moderate the association between IPA and BP. Although the strength of the association between
IPA and BP significantly differed across sites, it was in the predicted direction for the overwhelming majority of sites.
Implications of these results and suggestions for future research are discussed.

Intimate partner aggression (IPA) is typically considered to
consist of physical, emotional, and/or sexual aggression—
any one of these in and of themselves can constitute IPA,
as can any combination of two or all three. Since the ad-
vent of research into IPA in the early 1970s, there has been
evidence that both women and men use physical and psy-
chological IPA at rates that are approximately equal (e.g.,
Gelles, 1974), whereas women’s use of sexual IPA is typi-
cally less than men’s (e.g., Cantos, Neidig, & O’Leary, 1994;
Hines & Saudino, 2003; Swan & Snow, 2002). Moreover,
several studies have shown that women and men initiate
physical IPA at equal rates and that, in 25% of violent re-
lationships, only the women had been physically aggressive
within the previous year. In 25% of violent relationships,
only the man had been physically aggressive within the pre-
vious year, and in 50% of violent relationships, both part-
ners were physically aggressive (Carrado, George, Loxam,
Jones, & Templar, 1996; Demaris, 1992; Hines & Saudino,
2003; Morse, 1995; Stets & Straus, 1990). The equivalent
rates of physical IPA by men and women has recently been
confirmed by a meta-analysis of dozens of studies of IPA

Denise A. Hines, Department of Psychology, Clark University.

This work was supported by the National Institute of Mental

Health grant T32MH15161 and by the University of New Hamp-
shire. This article is part of the International Dating Violence
Study. Other papers from that study can be downloaded from
http://pubpages.unh.edu/

∼mas2.

Address correspondence and reprint requests to: Denise A.

Hines, Clark University, Department of Psychology, 950 Main
St., Worcester, MA 01610. E-mail: [e-mail]

over a 30-year period (Archer, 2000) and is further con-
firmed by the dramatic rise in women’s arrests and pros-
ecution for domestic violence in states across the nation
since the inception of mandatory arrest policies (Buzawa
& Buzawa, 2002). Although there is an abundance of re-
search showing that men and women use approximately
equal rates of physical IPA, there is considerably less re-
search on whether the prediction of IPA is equivalent for
men and women.

IPA and Borderline Personality

Few researchers have investigated whether there are gen-
der differences in the extent to which personality dysfunc-
tion predicts various forms of IPA, even though there is a
growing body of research showing that criminal behavior
and physical IPA perpetration are manifestations of per-
sonality traits (e.g., Krueger, Caspi, & Moffitt, 2000) and
dysfunction (e.g., Holtzworth-Munroe, Meehan, Herron,
Rehman, & Stuart, 2000) in men. One personality dysfunc-
tion that is related to the use of IPA in men, and may be for
women, is borderline personality (BP). The term BP will be
used here to refer to traits of Borderline Personality Disor-
der (BPD) that do not necessarily meet diagnostic criteria.
BP is somewhat synonymous with what Dutton (1995) calls
“borderline personality organization,” which is a less severe
form of BPD and is exhibited by 11–15% of the population
(in comparison to 1.5–4% who exhibit BPD). Because BPD
is three times more common among women than among
men (American Psychiatric Association, 1994), it is likely
that BP is also more common among women, and it is,

290

Borderline Personality and Aggression

291

therefore, a good candidate for being associated with IPA
in women.

According to Dutton (1995), the essential characteris-

tics of BP organization, from most important to least, are
“proclivity for intense, unstable interpersonal relationships
characterized by intermittent undermining of the signifi-
cant other, manipulation, and masked dependency; an un-
stable sense of self with intolerance of being alone and
abandonment anxiety; and intense anger, demandingness,
and impulsivity, usually tied to substance abuse or promis-
cuity” (p. 570). People with BP organization experience re-
lationships in which their emotional needs go unmet, and
they do not have the skills to assert their needs in a healthy
way. Frustrations can increase when they do not get what
they want or need in their intimate relationship. When they
perceive a possible loss of the relationship, people with BP
organization, who have an intense fear of abandonment and
loneliness, may experience intense anger or rage. Because
people with BP organization do not have the requisite skills
to appropriately handle their anger, they are at a high likeli-
hood of physically lashing out. In addition, because people
with BP organization see their significant other as either all
good or all bad, at this stage they tend to devalue their part-
ners, further increasing the likelihood of IPA. In an effort
to keep their partners from abandoning them, they may
engage in further manipulative and controlling behaviors
to keep them from leaving the relationship. Thus, people
with BP organization tend to have very intense, unstable
relationships and identities (Dutton, 1995).

Studies of men in batterer treatment programs support

the notion that BP traits are a strong predictor of their use
of physical IPA (e.g., Dutton, 1995; Dutton & Browning,
1988; Dutton, Starzomski, & Ryan, 1996; Hart, Dutton, &
Newlove, 1993; Hastings & Hamberger, 1988; Holtzworth-
Munroe, Bates, Smutzler, & Sandin, 1997; Mauricio, Tein,
& Lopez, 2007; Oldham et al., 1985) and psychological
IPA (Dutton & Starzomski, 1993; Mauricio et al., 2007;
Oldham et al., 1985). Their use of IPA is usually impul-
sive and a function of their mood rather than any external
stimulus. Men suffering from BP organization tend to be
emotionally volatile. They are angry, jealous, and depressed
men, who tend to use IPA when there is some sort of per-
ceived “socioemotional distance” between them and their
female partners (Dutton, 1995).

The majority of this research has been conducted on clin-

ical samples of male batterers, and we do not know whether
these findings would generalize to women. However, pre-
liminary evidence suggests that some women who use IPA
might be suffering from BP. For example, Margolies and
Leeder (1995) provided a clinical description of lesbian
batterers in their intervention program. This description
mirrored symptoms of BP. The women in this study “ex-
hibited a striking propensity for dichotomous thinking and
feeling. They were prone to extreme views and hyperbole,
dividing the world into black and white” (p. 144). They
were overly dependent on their partners for both attention

and emotional support, and they often tried to emotionally
merge with their partners. They used violence when they
felt that their partners were becoming emotionally distant
or when there was a threatened physical separation. They
viewed these actions as abandonment, and they evidenced
a chronic fear of abandonment. Thus, violence was used to
maintain a connection to their partners.

Stuart, Moore, Gordon, Ramsey, and Kahler (2006), in

a comparison of women arrested for IPA and mandated
into treatment with women from the general population,
found that the odds of women arrested for IPA having BPD
were 20.3 times greater than women in the general popu-
lation. Comparison studies of men and women mandated
into treatment for IPA also show evidence that women who
perpetrate IPA evidence BP, perhaps even at rates that are
higher than those of men. For example, the female batter-
ers in Henning, Jones, and Holdford’s (2003) study were
significantly more likely to evidence BP patterns than the
male batterers (11.6% versus 2.8%).

Although there is evidence among clinical samples that

women with BP might be at greater risk of using IPA,
it is unknown whether these results would generalize to
nonclinical samples. The associations between BP and IPA
in nonclinical samples are important to consider because
most IPA never comes to the attention of authorities, yet
is still detrimental to the mental and physical health of
its victims (Straus & Gelles, 1990). In addition, although
BP predicts both physical and psychological IPA by men
and physical IPA by women, it is unknown whether these
associations would generalize to sexual IPA by both genders
or psychological IPA by women.

Moreover, IPA affects men and women in every region

of the world (World Health Organization, 2005; Straus &
the International Dating Violence Research Consortium,
2004), and studies show that BPD is evident all over the
world and is one of the most common types of personal-
ity disorders worldwide (Loranger et al., 1994). However,
there is little research demonstrating a link between BP
and IPA in non-Western cultures and countries other than
the United States. This possible link in other cultures and
countries is important to investigate because it could point
toward either similar or different prevention and interven-
tion techniques in other cultures. Moreover, if there is a
difference among cultures in the association between BP
and IPA, possible reasons for these differences should be
explored. For example, it is possible that the association
between BP and IPA is stronger in cultures in which IPA is
not culturally accepted, but weaker in cultures where IPA
is more common and acceptable.

The purpose of the current study was to evaluate:

(1) Whether BP is a significant predictor of physi-

cal, psychological, and sexual IPA among men and
women in a nonclinical population. Given the pre-
vious research establishing a positive association be-
tween physical and psychological IPA and BP among

292

H

INES

men and women in clinical samples (e.g., Dutton &
Starzomski, 1993; Stuart et al., 2006), it is expected
that this association will also generalize to nonclinical
samples.

(2) Whether there are gender differences in the pre-

diction of physical, psychological, and sexual IPA
from BP. Previous research among male and fe-
male batterers who were arrested for IPA (Hen-
ning et al., 2003) suggests that BP may be more
common among female batterers than male bat-
terers. However, other empirical research of men
and women in community, nationally representa-
tive, clinical, and birth cohort samples (e.g., Busch
& Rosenberg, 2004; Capaldi & Owen, 2001; Carney
& Buttell, 2004; Giordano, Millhollin, Cernkovich,
Pugh, & Rudolph, 1999; Henning et al., 2003; Mag-
dol, Moffitt, Caspi, & Silva, 1998; Medeiros & Straus,
2006) suggests that there is gender symmetry in the
variables that predict IPA, including various person-
ality traits and dysfunctions. Therefore, it is hypoth-
esized that the association between BP and IPA will
be similar for men and women.

(3) Whether these associations hold constant across 67

sites from around the world. Because there has been
little to no cross-cultural research on the associa-
tion between BP and IPA, no specific hypotheses
were made concerning whether these associations
would be consistent across sites. However, if differ-
ences are found, reasons for such differences were
planned.

METHOD

Participants

The data for this article were from the International Dat-
ing Violence Study (IDVS). The IDVS was conducted by
members of a consortium of researchers at universities in
various regions of the world. Consortium members volun-
teered to become part of the study after hearing about the
study at international family violence conferences, because
of previous collaborations with the principal investigator, or
through word of mouth. The questionnaires were usually
administered in classes taught by members of the consor-
tium and in other classes for which they could make ar-
rangements. Thus, it was a convenience sample. Almost all
of the classes were introductory-level psychology, sociology,
and criminal justice studies courses; thus, the majority of
the sample consisted of female college students. The data
presented in the current analysis are from 67 university
sites around the world.

Questionnaires were distributed at the beginning of the

class period. The purpose of the study and that participa-
tion was entirely voluntary was explained orally and was also
written on the cover page of the questionnaire. The students
were informed that the questionnaire was about dating

relationships and that it would include sensitive questions
concerning attitudes, beliefs, and experiences in a relation-
ship, including questions on sexual behavior. They were
guaranteed anonymity, and they were told that the session
would take about an hour. Students completed the ques-
tionnaire at their own pace and deposited the completed
(or if they chose, blank) questionnaire in a box and left the
room when they finished. A debriefing form was provided
as they turned in their questionnaire. It explained the study
in more detail and provided names and telephone numbers
of local mental health services and community resources,
such as services for battered women.

The percentage of students who chose to participate and

deposited a completed questionnaire ranged from 42% to
100%, with most participation rates ranging from 85% to
95%. A detailed description of the study, including the
questionnaires and all other key documents, is available on
the study website: http://pubpages.unh.edu/

∼mas2.

The completed questionnaires were examined for ques-

tionable response patterns, such as reporting an injury from
dating violence but not reporting an assault as having oc-
curred, or reporting an attack on a partner with a knife or
gun 10 or more times in the past year. About 7.5% of the
original 15,321 cases were identified as questionable and
were removed from the sample, leaving 14,252 students. In
addition, only students who were currently or recently (i.e.,
within the previous year) involved in a romantic relationship
were instructed to complete the measure of dating aggres-
sion; students who did not complete the measure of dating
aggression (n

= 3,152) were eliminated from the analy-

ses. Because of questionable data from the Iranian sample
that could not adequately be explained, the Iranian site was
eliminated from the analysis (n

= 98). This procedure left a

sample of 14,154 students (4,054 men and 10,100 women).
A list of the sites involved is given in Table 5.

Demographic characteristics of the sample are shown in

Table 1. These characteristics are presented for the sample
overall and for each gender, and the sites that represented
the high and low scores for each variable are also presented.
As shown for the men, sample sizes ranged from 4 men at
the Calcutta, India site to 163 men at the Swedish site. The
mean length of relationship for the male sample was 12.70
months, close to 70% of the relationships were sexual, and
95.8% of the men were involved in heterosexual intimate
relationships. For women, the sample sizes ranged from
20 at the Calcutta, India site to 511 at the Swedish site.
Women reported an average relationship length of 15.08
months, sex was part of the relationship in 73.3% of the
cases, and 98% of the women were involved in heterosexual
relationships.

Measures

There was a core questionnaire that each member of
the IDVS Research Consortium translated. All consortium
members agreed to back-translate to maintain conceptual

Borderline Personality and Aggression

293

Table 1

Demographics

Total

Men

Women

t or

χ

2

n

14,154

4,054

10,100

High

674

163

511

(Gavle, Sweden)

(Gavle, Sweden)

(Gavle, Sweden)

Low

24

4

20

(Calcutta, India)

(Calcutta, India)

(Calcutta, India)

Individual Demographics
Age (M, SD)

23.12, 6.28

23.01, 5.56

23.16, 6.55

1.41

High

39.06, 9.71

38.09, 10.04

39.98, 9.14

(Liege, Belgium)

(German-speaking Switzerland)

(Liege, Belgium)

Low

18.65, 1.36

18.86, 1.60

18.45, 1.04

(St. Petersburg 2, Russia)

(St. Petersburg 2, Russia)

(St. Petersburg 2, Russia)

Social Desirability (M, SD)

2.62, 0.36

2.59, 0.35

2.63, 0.36

6.49

∗∗∗

High

2.91, 0.42

2.88

2.94, 0.40

(Merida, Venezuela)

(Cluj-Napcoa, Romania,

(Merida, Venezuela)

SD

= 0.39; Northern

Mexico, SD

= 0.35;

Merida, Venezuela,

SD

= 0.46)

Low

2.39, 0.38

2.34, 0.38

2.39, 0.36

(Mito, Japan)

(Washington, DC, USA)

(Mito, Japan)

Borderline Personality (M, SD)

1.98, 0.49

2.00, 0.45

1.97, 0.50

3.96

∗∗∗

High

2.42, 0.44

2.29, 0.44

2.48, 0.44

(Taiwan)

(Beijing, China)

(Taiwan)

Low

1.58, 0.41

1.52, 0.34

1.59, 0.42

(Leiden, Netherlands)

(Leiden, Netherlands)

(Leiden, Netherlands)

Relationship Demographics
Relationship Length

14.40, 8.92

12.70, 9.01

15.08, 8.80

14.31

∗∗∗

in months (M, SD)
High

19.46, 7.27

19.13, 8.01

19.90, 7.01

(Gavle, Sweden)

(Leiden, Netherlands)

(Gavle, Sweden)

Low

6.21, 6.71

6.04, 5.80

6.38, 7.57

(St. Petersburg 2, Russia)

(St. Petersburg 2, Russia)

(St. Petersburg 2, Russia)

% Sex in Relationship

72.1%

69.2%

73.3%

24.30

∗∗∗

High

97.8%

98.2%

99.0%

(Gavle, Sweden &

(Quebec 2, Canada)

(Freiburg, Germany)

Freiburg, Germany)

Low

11.8%

5.9%

14.5%

(Varsanyi, Hungary)

(Varsanyi, Hungary)

(Varsanyi, Hungary)

% Heterosexual Relationships

97.4%

95.8%

98.0%

55.47

∗∗∗

High

99.6%

100%

100%

(Cluj-Napcoa, Romania)

(10 sites)

(9 sites)

Low

92.7%

68.2%

92.4%

(Jackson, MS, USA)

(Jackson, MS, USA)

(Washington, DC, USA)

Intimate Partner Aggression
% Any Physical Aggression

29.0%

23.9%

31.0%

70.14

∗∗∗

High

49.6%

54.7%

54.8%

(New York, NY, USA)

(New York, NY, USA)

(Vladivostok, Russia)

Low

15.4%

9.4%

11.6%

(Houston, TX, USA)

(Hamilton, Canada)

(Houston, TX, USA)

% Severe Physical Aggression

10.4%

8.0%

11.3%

34.95

∗∗∗

High

24.4%

28.6%

25.0%

(New York, NY, USA)

(Capetown, South Africa)

(Crete 2, Greece)

Low

1.8%

0%

2.0%

(Gavle, Sweden)

(Calcutta, India; Singapore;

(Amsterdam, Netherlands)

Hamilton, Canada; Malta)

(continued)

294

H

INES

Table 1 (continued)

Total

Men

Women

t or

χ

2

Types of Physical Aggression

0.69, 1.47

0.54, 1.29

0.76, 1.53

8.69

∗∗∗

(M, SD)

High

1.57, 2.22

2.29, 3.95

1.65, 2.32

(New York, NY, USA)

(Capetown, South Africa)

(New York, NY, USA)

Low

0.33

0.13, 0.42

0.23, 0.78

(Portugal, SD

= .98;

(Hamilton, Canada)

(Houston, TX, USA)

Malta, SD

= .82)

% Any Psychological Aggression

72.6%

65.7%

75.3%

133.96

∗∗∗

High

91.0%

100%

90.9%

(Washington, DC, USA)

(Calcultta, India)

(Washington, DC, USA)

Low

45.7%

36.7%

42.9%

(St. Petersburg 1, Russia)

(Shanghai)

(St. Petersburg 1, Russia)

% Severe Psychological

24.0%

20.9%

25.3%

30.47

∗∗∗

Aggression
High

57.7%

50.8%

60.3%

(Hong Kong 3)

(Hong Kong 1)

(Hong Kong 3)

Low

8.3%

0%

7.6%

(Gavle, Sweden)

(Winnipeg, Canada)

(Gavle, Sweden)

Types of Psychological

2.04, 1.82

1.72, 1.76

2.17, 1.82

13.55

∗∗∗

Aggression (M, SD)

High

3.03, 1.98

2.92, 1.51

3.07, 1.98

(New York, NY, USA)

(Washington, DC, USA)

(New York, NY, USA)

Low

1.20, 1.60

0.88, 1.58

1.25, 1.72

(St. Petersburg 1, Russia)

(Shanghai, China)

(St. Petersburg 1, Russia)

% Any Sexual Aggression

23.5%

29.3%

21.2%

105.00

∗∗∗

High

60.9%

65.3%

58.3%

(Merida, Venezuela)

(Merida, Venezuela)

(Merida, Venezuela)

Low

5.2%

4.6%

4.7%

(Leiden, Netherlands)

(Hong Kong 1)

(Leiden, Netherlands)

% Forced Sexual Aggression

1.7%

2.2%

1.5%

7.71

∗∗

High

7.2%

14.3%

10.5%

(Louisiana, USA)

(Capetown, South Africa)

(Dar Salaam, Tanzania)

Low

0%

0%

0%

(7 sites)

(25 sites)

(15 sites)

Types of Sexual Aggression

0.34, 0.73

0.47, 0.89

0.29, 0.64

12.18

∗∗∗

(M, SD)

High

0.86, 0.98

1.11, 1.15

0.93, 1.26

(Merida, Venezuela)

(New York, NY, USA)

(Dar Salaam, Tanzania)

Low

0.06, 0.25

0.07, 0.26

0.05, 0.21

(Leiden, Netherlands)

(Hong Kong 3)

(Leiden, Netherlands)

Note. Gender differences between the demographic variables were explored using chi-square or t tests, where appropriate. The Social Desirability

and Borderline Personality scales ranged from 1 to 4. The maximum number of types of physical aggression was 12; the maximum number of types of
psychological aggression was 8, and the maximum number of types of sexual aggression was 7.

∗∗

p

< .01.

∗∗∗

p

< .001.

equivalence (Straus, 1969) across the sites. This core ques-
tionnaire consisted of demographic items (e.g., gender,
gender of partner, whether sex was part of the relation-
ship, age of participant, parents’ education and income,
length of relationship), the Revised Conflict Tactics Scales
(CTS2; Straus, Hamby, Boney-McCoy, & Sugarman, 1996),
and the Personal and Relationships Profile (PRP; Straus,
Hamby, Boney-McCoy, & Sugarman, 1999). In addition,
the consortium members added questions to measure vari-
ables that were uniquely important for their site or con-

structs that were needed to test a theory of particular in-
terest. For the current study, only the demographic infor-
mation and questions pertaining to the use of IPA, BP, and
social desirability were used.

IPA. The use of IPA was measured by the physical ag-

gression, psychological aggression, and sexual aggression
scales of the CTS2. For each participant, the number of
different types of physically aggressive, psychologically ag-
gressive, and sexually aggressive acts used in the previous

Borderline Personality and Aggression

295

year was computed. Participants indicated on a scale from
0 to 6 how many times in the previous year they used the
acts listed (0

= 0 times; 1 = 1 time; 2 = 2 times; 3 = 3–5

times; 4

= 6–10 times; 5 = 11–20 times; 6 = more than

20 times). Following Moffitt et al. (1997), a count of the
different types of aggression for each scale was computed.
Moffitt et al. (1997) found that, statistically, the best way
to measure severity of IPA was to count the number of dif-
ferent types of acts used. A given participant could report
on up to 12 types of physical aggression, 8 types of psy-
chological aggression, and 7 types of sexual aggression. In
the multilevel modeling analyses, the count of the number
of different types of physical, psychological, and sexual ag-
gression was used. Because these are count variables, the
models were estimated using a procedure that allows for
the dependent variable to be of a Poisson distribution.

To establish past-year prevalence rates of each form of

IPA for descriptive purposes, participants were coded as 1
(yes) if they reported using any of the physically, psycho-
logically, or sexually aggressive acts listed in the CTS2 in
the previous year and 0 (no) if they reported using no acts
in the previous year. These acts included both minor and
severe acts of IPA. Because severe physical, psychological,
and sexual IPA are often a concern of researchers and ad-
vocates in the field of IPA, participants were also coded, as
per CTS2 instructions, according to whether they reported
using any severe physical (e.g., punching, kicking, beating
up) and psychological acts (e.g., threatening to hit or throw
something) and whether they reported forcing oral, anal,
or vaginal sex (i.e., severe sexual aggression). The Physical,
Psychological, and Sexual Aggression scales of the CTS2
have been shown to demonstrate good cross-cultural con-
struct validity and reliability, with overall alphas of .85, .74,
and .82, respectively. Typically, sites with low alpha coeffi-
cients were sites in which there was a low prevalence rate
for that form of IPA, which reduces the inter-item correla-
tion of the scale (Straus, 2004).

BP. BP traits were assessed using the BP scale of the

PRP. The BP scale is a 9-item scale based on BP features
derived from the Diagnostic and Statistical Manual of Men-
tal Disorders, 4th ed. (DSM-IV) (American Psychiatric As-
sociation, 1994), such as instability of self (e.g., I change
suddenly from being one kind of person to another) and re-
lationships (e.g., My relationships have big ups and downs),
emotional lability (e.g., My mood is always changing), aban-
donment anxiety (e.g., I’d do almost anything to keep peo-
ple from leaving me), and threats of self-harm (e.g., I’ve
told others I will kill myself). Participants responded by in-
dicating their level of agreement with each item using a 4-
point scale ranging from 1 (strongly disagree) to 4 (strongly
agree).

The nine items composing the scale were summed and

divided by the total number of items to obtain a mean
BP score. Previous research has shown that this scale
demonstrates excellent validity and reliability (Straus &

Mouradian, 1999). With the current sample, the internal
consistency reliability (Cronbach’s alpha) was .74, with the
majority of sites having alphas of .70 or greater. Three sites
(St. Petersburg 2, South Africa, and Tanzania) had alphas
between .51 and .57; therefore, the results from these sites
should be viewed with caution. It is important to note that,
although the questions were derived from the DSM-IV def-
inition of BPD, this scale was not designed as a diagnostic
tool. Thus, high scores on this subscale cannot necessarily
be interpreted as indicative of BPD. Table 1 presents the
average scores of the BP scale overall and for each gender,
and it also presents the sites that represent the high and
low average scores for BP.

Social desirability. Social desirability response bias has

been shown to be a stable cross-cultural phenomenon (e.g.,
Consalvi, 1972; Crowne & Marlowe, 1964; Mukherjee,
1967; Tsushima, 1969), although recent research that has
revisited this issue might be shedding some doubt on this
(Middleton & Jones, 2000). Given that there is nonetheless
a strong likelihood that, cross-culturally, social desirabil-
ity response bias could influence participants’ responses,
participants’ tendency to minimize socially undesirable be-
havior was controlled with the social desirability scale of
the PRP. This 13-item scale includes behaviors and emo-
tions that are slightly undesirable but true of most people,
such as “I sometimes try to get even rather than forgive
and forget.” Participants indicated on a 4-point scale (1

=

strongly disagree, 2

= disagree, 3 = agree, 4 = strongly

agree) the extent to which they agreed with each item. The
items were then summed and divided by the number of
items in the scale to obtain an average social desirability
score. The overall reliability of this scale was .68, and it
ranged from a low of .45 in Shanghai to a high of .80 in
Vladivostok. Preliminary cross-cultural predictive validity
evidence is available as well: As expected, social desirability
was a strong, negative predictor of the Antisocial Person-
ality scale (the cross-cultural validity of which has been
established elsewhere; Hines & Straus, 2007; Lysova &
Hines, in press) for all sites in this study except Calcutta,
India. For the remaining sites, correlations ranged from

−.38 to −.69, with an average correlation of −.52. Table 1
presents the average scores of the social desirability scale
overall and for each gender, and it also presents the sites
that represent the high and low average scores for social
desirability.

Socioeconomic status (SES). An SES variable was cre-

ated for each site using three variables: father’s education,
mother’s education, and family income. To create a variable
that measured the SES of each student that was relevant
to the SES of others at the student’s university, the SES
variables at each site were transformed into z scores. The
scale thus measures SES as the number of standard devi-
ations each student was above or below the mean at their
site.

296

H

INES

Analyses. To test the research questions, a series of

multilevel regression models were estimated. Multilevel re-
gression modeling is a technique that allows one to simul-
taneously consider both individual-level and group-level
influences on a variable of interest without violating
assumptions of independence (as would occur in the
individual-level analysis using site as an independent vari-
able) or losing valuable variability (as would occur in analy-
ses aggregating scores by sites). Multilevel modeling allows
one to examine how group influences interact with indi-
vidual characteristics by performing a series of nested lin-
ear models that take into account hierarchical structure
(Raudenbush & Bryk, 2002).

A series of models was estimated to investigate the hy-

potheses. Because the dependent variables were count vari-
ables, all of the models were estimated using a Poisson
model. Overdispersion was examined and found not to be
a problem. Control variables included one or more of the
following: gender, age, length of relationship, whether sex
was part of the relationship, sexual orientation, SES, and so-
cial desirability response bias of the participant. If a control
variable did not significantly predict IPA, it was dropped
from the model. The individual-level predictors included
the BP score of the participant and the interaction of BP
and gender. The predictor variables were allowed to vary
across sites, and possible differences in the strength of the
association between IPA and BP were examined. For pur-
poses of brevity, only the final models will be displayed in
the Results section.

Table 2

Gender Interaction Model Predicting Number of Types of Physical Intimate Partner Aggression (IPA)

as a Function of Borderline Personality

Fixed Effects

Coefficient

SE

t Ratio

Event Rate Ratio

Outcome:

Number of Types of Physical IPA (Intercept),

γ

00

−1.92

0.31

−6.30

∗∗∗

0.15

Covariates:

Gender

a

,

γ

10

0.42

0.17

2.40

1.51

Age,

γ

20

−0.03

0.01

−3.44

∗∗∗

0.97

Relationship Length,

γ

30

0.03

0.01

14.04

∗∗∗

1.04

Sex Part of the Relationship

b

,

γ

40

0.36

0.05

7.27

∗∗∗

1.43

Social Desirability,

γ

50

−0.70

0.06

−12.45

∗∗∗

0.49

Predictors:

Borderline Personality,

γ

60

0.90

0.15

6.18

∗∗∗

2.45

Borderline Personality

× Gender Interaction, γ

70

−0.09

0.08

−1.13

0.92

Random Effects

Variance

df

χ

2

Site Mean, u

0j

0.16

66

1403.67

∗∗∗

Number of Types of Physical IPA–Borderline Personality Slope, u

6j

0.17

66

261.16

∗∗∗

Number of Types of Physical IPA-Borderline

× Gender Interaction Slope, u

7j

0.02

66

208.60

∗∗∗

Note. Number of types of Physical IPA is a count of the different types of physical IPA the participants reported using and can range

from 0 to 12. Gender, Age, Relationship Length, Sex Part of the Relationship, and Social Desirability are group mean centered
and constrained to have equal variances across sites. Borderline Personality and Borderline Personality

× Gender Interaction are

group mean centered and allowed to vary across sites.

a

Gender: 1

= Male, 2 = Female

b

Sex Part of the Relationship: 1

= No, 2 = Yes.

∗∗∗

p

< .001.

In the final series of multilevel models, three Slopes-as-

Outcomes Models were estimated. Specifically, a Level 2
(i.e., site-level) predictor, the prevalence of each form of
IPA, was added to investigate possible site-level influences
on the site differences observed in the association between
BP and the three forms of IPA.

RESULTS

Prevalence and Frequency of IPA

The percentage of men and women who used physi-
cal, psychological, and sexual IPA and the mean num-
ber of types of aggression were computed for the full
data set and at each site. Summary data are presented in
Table 1.

Prediction of IPA by BP

The multilevel models predicting the number of types of
physical, psychological, and sexual IPA (dependent vari-
ables) from BP (independent variable) are presented in
Tables 2, 3, and 4, respectively. Inspection of the models in
each of these tables indicates that, after controlling for gen-
der, age, relationship length, whether sex was part of the
relationship, and social desirability, BP was a significant
predictor of the number of types of physical, psychological,
and sexual IPA. For physical IPA (Table 2), the event rate
ratio was 2.45, which indicates that, for every one-point
increase of the four-point BP scale, the number of types

Borderline Personality and Aggression

297

Table 3

Gender Interaction Model Predicting Number of Types of Psychological Intimate Partner

Aggression (IPA) as a Function of Borderline Personality

Fixed Effects

Coefficient

SE

t Ratio

Event Rate Ratio

Outcome:

Number of Types of Psychological IPA (Intercept),

γ

00

−0.22

0.14

−1.55

0.80

Covariates:

Gender

a

,

γ

10

0.26

0.08

3.30

∗∗

1.29

Age,

γ

20

−0.01

0.01

−3.99

∗∗∗

0.99

Relationship Length,

γ

30

0.02

0.01

19.55

∗∗∗

1.03

Sex Part of the Relationship

b

,

γ

40

0.24

0.02

10.05

∗∗∗

1.27

Social Desirability,

γ

50

−0.48

0.03

−17.85

∗∗∗

0.62

Predictors:

Borderline Personality,

γ

60

0.42

0.07

5.74

∗∗∗

1.52

Borderline Personality

× Gender Interaction, γ

70

−0.04

0.04

−1.02

0.96

Random Effects

Variance df

χ

2

Site Mean, u

0j

0.03

66 763.50

∗∗∗

Number of Types of Psychological IPA-Borderline Personality Slope, u

6j

0.01

66 111.70

∗∗∗

Number of Types of Psychological IPA-Borderline

× Gender Interaction Slope, u

7j

0.00

66 121.88

∗∗∗

Note. Number of Types of Psychological IPA is a count of the different types of psychological IPA the participants reported
using and can range from 0 to 8. Gender, Age, Relationship Length, Sex Part of the Relationship, and Social Desirability
are group mean centered and constrained to have equal variances across sites. Borderline Personality and Borderline Person-
ality

× Gender Interaction are group mean centered and allowed to vary across sites.

a

Gender: 1

= Male, 2 = Female

b

Sex Part of the Relationship: 1

= No, 2 = Yes.

∗∗

p

< .01.

∗∗∗

p

< .001.

Table 4

Gender Interaction Model Predicting Number of Types of Sexual Intimate Partner Aggression

(IPA) as a Function of Borderline Personality

Fixed Effects

Coefficient

SE

t Ratio

Event Rate Ratio

Outcome:

Number of Types of Sexual IPA (Intercept),

γ

00

−1.99

0.31

−6.46

∗∗

0.14

Covariates:

Gender

a

,

γ

10

−0.55

0.13

−4.12

∗∗

0.58

Age,

γ

20

−0.01

0.01

−2.15

∗

0.99

Relationship Length,

γ

30

0.01

0.01

5.78

∗∗

1.01

Sex Part of the Relationship

b

,

γ

40

0.95

0.09

10.39

∗∗

2.58

Social Desirability,

γ

50

−0.50

0.06

−8.55

∗∗

0.61

Predictors:

Borderline Personality,

γ

60

0.45

0.11

3.94

∗∗

1.56

Borderline Personality

× Gender Interaction, γ

70

0.02

0.06

0.24

1.02

Random Effects

Variance

df

χ

2

Site Mean, u

0j

0.23

66

994.74

∗∗

Number of Types of Sexual IPA-Borderline Personality Slope, u

6j

0.19

66

185.21

∗∗

Number of Types of Sexual IPA-Borderline

× Gender Interaction Slope, u

7j

0.03

66

182.11

∗∗

Note. Number of Types of Sexual IPA is a count of the different types of sexual IPA the participants reported using and

can range from 0 to 7. Gender, Age, Relationship Length, Sex Part of the Relationship, and Social Desirability are group
mean centered and constrained to have equal variances across sites. Borderline Personality and Borderline Personality

×

Gender Interaction are group mean centered and allowed to vary across sites.

a

Gender: 1

= Male, 2 = Female

b

Sex Part of the Relationship: 1

= No, 2 = Yes.

∗

p

< .05.

∗∗

p

< .001.

298

H

INES

of physical IPA used more than doubled. The event rate
ratio for BP in the prediction of psychological IPA (Ta-
ble 3) shows that, for every one-point increase on the BP
scale, the number of types of psychological IPA used in-
creased by 52%. Finally, for every one-point increase on
the BP scale, the number of types of sexual IPA increased
by 56% (Table 4). These models also tested whether the
associations between BP and the three forms of IPA were
moderated by gender. As shown, for all three forms of IPA,
the interaction of gender and BP was not a significant pre-
dictor of the number of types of the different forms of IPA.
Thus, gender did not moderate the association between BP
and IPA.

Differences among the sites in the association between

BP and the number of types of the three forms of IPA are in-
dicated by a significant

χ

2

for the IPA–BP slope. As shown

in all three models, the

χ

2

was significant, which means

that there was significant variance in the strength of the
association between BP and IPA among the sites. Table 5
presents the slope estimates for the associations between
BP and physical, psychological, and sexual IPA. Slope esti-
mates provide an indication of the increase in the number of
types of IPA at each site for every one-point increase on the
BP scale. Slope estimates close to zero would indicate that
there was no association between IPA and BP. The larger
the slope estimate, the greater the number of types of IPA
for each one-point increase in the BP scale. As shown, slope
estimates ranged from 0.09 (Hong Kong 2) to 1.50 (Hous-
ton, TX) for physical IPA; the mean slope estimate was 0.90
(SD

= 0.35). For psychological aggression, slopes ranged

from 0.24 (Hungary) to 0.62 (Tanzania); the mean slope
estimate was 0.42 (SD

= 0.07). Finally, slopes ranged from

−0.30 (Washington, DC) to 1.32 (Leiden, Netherlands) for
sexual IPA; the mean slope estimate was 0.45 (SD

= 0.35).

The models also tested whether there were differences

among the sites in the moderating influence of gender. A
significant

χ

2

indicates that the moderating influence dif-

fered among the sites. Tables 2–4 show that, for all three
forms of IPA, the moderating influence significantly dif-
fered among the sites; however, the percent of variance
among the sites ranged from

<1% to 3%. Given that the

moderating influence overall was nonsignificant and there
was little practical variance among the sites, these slight
differences were not explored.

A final set of analyses examined whether the strength of

the association between BP and IPA was relatively weaker
at sites in which there was increased general acceptance of
each form of IPA. Because we did not have a measure that
directly tapped attitudes of acceptance of each form of IPA,
the prevalence of each form of IPA was used as a proxy for
attitudes; in other words, it was assumed that acceptance
of IPA would be greater at sites where IPA occurred more
frequently.

The prevalence of physical, psychological, and sexual

IPA at each site was used as a Level 2 predictor for the
association between BP and physical, psychological, and

sexual IPA, respectively. For the BP–physical IPA slope,
the prevalence of physical IPA was a significant predictor,

β = 0.01, SE = 0.01, T(65) = 2.65, p < .05, explaining
4% of the variance in the slopes among the sites. Thus,
the association between physical IPA and BP was stronger

Table 5

Slopes for Associations Between Intimate Partner

Aggression (IPA) and Borderline Personality at Each Site

Type of IPA

Site

Physical Psychological Sexual

Africa

South Africa, Cape Town

1.37

0.43

0.40

Tanzania, Dar Salaam

1.06

0.62

−0.08

Asia

China, Shanghai

0.37

0.47

0.37

China, Beijing

0.59

0.39

0.50

Hong Kong 1

0.48

0.37

0.90

Hong Kong 2

0.09

0.36

0.11

Hong Kong 3

1.09

0.38

0.50

India, Calcutta

0.58

0.45

0.64

India, Pune

0.49

0.40

−0.12

Japan, Mito

0.56

0.45

−0.17

Singapore

0.92

0.45

0.56

South Korea, Pusan

0.40

0.50

0.85

Taiwan

0.43

0.45

0.41

Australia/New Zealand

Australia, Adelaide

1.10

0.47

0.81

New Zealand, Christ Church

0.89

0.38

0.11

Canada

Hamilton

0.80

0.43

0.58

London

0.22

0.45

0.01

Quebec 1

1.33

0.49

0.69

Quebec 2

0.91

0.43

0.46

Toronto

1.02

0.45

0.64

Winnipeg

1.05

0.37

0.23

Eastern Europe/Russia

Hungary, Varsanyi

0.57

0.24

0.22

Lithuania, Vilnius

0.67

0.36

0.55

Romania, Cluj-Napoca

0.88

0.40

0.08

Russia, Bernaul

0.88

0.46

0.44

Russia, St. Petersburg 1

0.80

0.53

0.46

Russia, St. Petersburg 2

1.08

0.53

0.63

Russia, Vladivostok

0.46

0.38

0.42

Latin America

Brazil, Sao Paulo

0.84

0.45

0.29

Guatemala

0.33

0.35

0.59

Mexico, Northern

0.63

0.42

0.46

Venezuela, Merida

1.35

0.52

0.13

Mediterranean/Middle East

Greece, Crete 1

1.22

0.47

0.93

Greece, Crete 2

1.49

0.28

0.66

Israel, Emek Yezreel

1.35

0.45

1.06

Malta, Msida

1.42

0.51

0.08

(continued)

Borderline Personality and Aggression

299

Table 5 (continued)

Type of IPA

Site

Physical Psychological Sexual

USA

Illinois

0.92

0.40

0.42

IN, Terre Haute

1.30

0.48

−0.04

Louisiana

0.80

0.40

0.02

Mississippi

0.56

0.35

0.36

MS, Jackson

0.90

0.32

0.18

NH, Durham 1

0.37

0.36

0.51

NH, Durham 2

1.01

0.36

0.51

NY, New York

0.70

0.32

0.25

OH, Cincinnati

1.15

0.39

0.47

Pennsylvania

0.75

0.30

0.15

TN, Knoxville

1.07

0.41

0.61

TX, El Paso

0.58

0.36

−0.08

TX, Houston

1.50

0.49

0.52

TX, Lubbock

0.72

0.46

0.63

TX, Mexican-American

0.88

0.39

0.36

TX, Non-Mexican-American

0.83

0.44

0.57

TX, Nacogdoches

1.34

0.41

0.45

UT, Logan

1.30

0.46

0.16

Washington, DC

0.73

0.35

−0.30

Western Europe

Belgium, Flemish-speaking

1.01

0.40

1.31

Belgium, Liege

1.39

0.48

1.07

England, Leicester

0.80

0.28

0.20

Germany, Freiburg

1.44

0.46

0.78

Germany, Moechengladbach

0.59

0.38

−0.15

Netherlands, Amsterdam

1.32

0.49

0.93

Netherlands, Leiden

0.91

0.48

1.32

Portugal, Braga

1.39

0.54

0.92

Scotland, Glasgow

0.74

0.34

0.44

Sweden, Gavle

1.48

0.55

0.96

Switzerland, French-speaking

0.93

0.40

0.33

Switzerland, German-speaking

0.91

0.39

0.63

at sites with higher levels of physical IPA, although there
was still significant variance (13.3%) left to be explained,

χ

2

(65)

= 381.75, p < .001. The prevalence of psychological

IPA at a site was also a significant, but negative, predictor
of the BP–psychological aggression slope,

β = −0.01, SE =

0.01, T(65)

= −2.63, p < .05, explaining an additional 0.4%

of the variance in the slopes, and there was still significant
variance (0.6%) among the sites,

χ

2

(65)

= 112.80, p < .001.

Finally, the prevalence of sexual IPA was not a significant
predictor of the BP–sexual IPA slope,

β = 0.004, SE =

0.003, T(65)

= 1.27, ns.

DISCUSSION

The results of this study support the hypothesis that BP
positively predicts physical, psychological, and sexual IPA
for both men and women in nonclinical samples. Further,
the lack of moderation by gender shows that there is gender

symmetry (Straus, 2006) in the prediction of physical, psy-
chological, and sexual IPA by BP. Given the large sample
size of the IDVS, if a gender difference existed, there was
sufficient power to detect it. Thus, it seems that, for both
women and men, personality features that are consistent
with BP, such as instability of self and relationships, ma-
nipulation, self-harming behavior, fear of abandonment,
anger, jealousy, impulsivity, and emotional volatility, are
risk factors for their perpetration of IPA. Nonetheless, it is
important to consider that the current sample was a conve-
nience sample of college students, and therefore, a gender
difference may yet be found in clinical samples of men and
women. Given the research thus far with clinical samples
(e.g., Henning et al., 2003), it is likely that, if a gender dif-
ference is found, women who engage in IPA would be more
likely to evidence BP than men who engage in IPA.

The current study also showed that BP predicts all

three forms of IPA—physical, psychological, and sexual—
for both men and women. The majority of the research
thus far has concentrated mostly on male perpetrators of
physical IPA, and although associations have been found
between psychological IPA and BP in men (e.g., Dutton,
1995), and physical IPA and BP in women (e.g., Stuart
et al., 2006), the current study extends our knowledge
by showing that, among nonclinical samples of men and
women, BP predicts the three forms of IPA that are labeled
IPA.

Although BP predicted all forms of IPA for both men

and women, there were differences in the strength of these
associations across the 67 sites in this study. Nonetheless,
although the strength of the associations differed across
sites, for the overwhelming majority of the sites, the slopes
for the three forms of IPA were in the predicted direction
and were greater than zero. In addition, although the vari-
ance for the BP–psychological IPA slope was significantly
different from zero, there was only about 1% variance in
this slope among the sites. Also of note, there were no sites
in which the slope of IPA and BP was low for all three forms
of IPA. However, there were some sites in which the slopes
for one or two forms of IPA were close to or less than zero,
and thus, the reason for these differences in the strength
and direction of the associations was explored.

We found little support for the hypothesis that the asso-

ciation between BP and IPA would be stronger in sites with
lower acceptance of IPA, using prevalence of each form of
IPA as a proxy measure of acceptance. For physical IPA, the
opposite was found; sites with higher levels of IPA showed a
stronger association between physical IPA and BP. For psy-
chological IPA, the hypothesis was supported, but although
the BP–psychological aggression slope did significantly dif-
fer among the sites, the amount of between-site variance
(1%) was small to begin, and therefore, adding this predic-
tor to the model increased the variance explained by only
0.6%. For sexual IPA, the hypothesis was not supported.
The reasons for these weak and contradictory findings may
be because prevalence of IPA at a site is not a good proxy

300

H

INES

for societal attitudes about IPA. Therefore, a more direct
measure of attitudes should be used in future research.

Alternatively, the differences in the BP–IPA slopes could

be due to different manifestations of BP across cultures
(Chavira et al., 2003) and that the measure of BP did not
fully capture the BP symptoms of some participants in some
cultures. Thus far, however, studies show that BPD is found
in every region of the world (Loranger et al., 1994), and the
few studies that have empirically evaluated possible cultural
differences in the manifestations of BPD have found few
differences (Ikuta et al., 1994; Serman, Johnson, Geller,
Kanost, & Zacharapoulou, 2002). Nonetheless, it is possi-
ble that nonclinical levels of BP could manifest themselves
differently across cultures.

Future Research

In addition to exploring the reasons for cultural differences
in the prediction of IPA from BP, future research should
address some of the limitations of this study. The current
study used college students as its sample, and college stu-
dents may not be representative of the population in gen-
eral, in that people who evidence high rates of either BP
or IPA would be less likely to succeed in the college en-
vironment. Moreover, the 67 sites that participated in the
current study may not be representative of the countries
and cities in which they are located, nor are they exhaustive
of all possible cultures from around the world. Thus, future
studies should strive to obtain more representative data. It
is also important to stress that future studies assess the pre-
diction of IPA from BP longitudinally. Because the current
study was correlational, it is unknown whether BP caused
IPA, IPA caused BP, or some third variable (e.g., stress)
caused both. Given the fact that BP tends to manifest itself
in adolescence or early adulthood and is relatively long-
lasting and stable thereafter (American Psychiatric Associ-
ation, 1994), it is likely that BP predicts IPA. Nonetheless,
given the relatively young age of this sample, it could be
the case that IPA influenced the development and course
of BP.

Future studies should also use other measures of BP and

IPA and multiple informants of IPA on a target individual
to further validate the present results. Because both IPA
and BP were assessed through self-report, it could result in
inflated associations because response sets could influence
self-reports of personality and IPA (Cooper, 2002). Thus,
multiple informants (e.g., intimate partner) should be used.
This is especially important for IPA because a meta-analysis
of 43 studies on physical IPA provided evidence that per-
petrators of both genders tend to underestimate the extent
to which they use violence in their relationships (Archer,
1999). In addition, because our BP measure could not be
used to infer a clinical diagnosis of BPD, future studies
should explore how clinical manifestations of BPD might
differentially predict IPA for men and women and whether
differences might occur across cultures.

It is important to consider also the ways in which BP

predicts coercive control in intimate partner relationships.
Although the CTS2 is typically considered to be the best
measure of IPA, it does not contain a scale to measure co-
ercive control, a set of tactics that many perpetrators of
IPA use, and some researchers argue is a critical element
of IPA (e.g., Swan & Snow, 2002). Given the constellation
of symptoms of people with BP (e.g., anger, jealousy, ma-
nipulation, self-harming behaviors, fear of abandonment),
it is likely that BP would also predict coercive controlling
tactics in intimate relationships.

Caveats

The current study investigated whether there was gender
symmetry in the prediction of IPA. However, it is impor-
tant to note that gender symmetry in the prediction of IPA
does not mean that there is gender symmetry in the rate or
consequences of IPA. For example, this and other studies
(e.g., Cantos et al., 1994; Hines & Saudino, 2003; Swan &
Snow, 2002) show that men use significantly more sexual
IPA than women, although women who perpetrate phys-
ical IPA do sexually assault their male partners at rates
that are surprisingly high (Swan & Snow, 2002); in the cur-
rent study, women used significantly more types of physical
and psychological IPA than did men, and their past-year
prevalence rates of these two forms of IPA were also sig-
nificantly higher. Even so, there are reports that women
are more fearful, feel more threatened, are injured more,
and seek medical assistance at higher rates than men (e.g.,
Langhinrichsen-Rohling, Neidig, & Thorn, 1995; Laroche,
2005; Morse, 1995), and even though men do report fear
and injury as a result of IPA (e.g., Hines, Brown, & Dun-
ning, 2007), there probably is less gender symmetry in the
consequences of IPA (Stets & Straus, 1990) than there is in
the prediction of IPA.

Implications

The current study provides evidence that BP predicts phys-
ical, psychological, and sexual IPA among both men and
women across university campuses worldwide. Although
the current study included only college students, it pro-
vided a test of these hypotheses cross-culturally, which few,
if any, previous studies on IPA have done. In addition, be-
cause this was a large sample of both male and female
college students, possible gender differences in these as-
sociations could be investigated. The fact that none were
found has important implications for the treatment of men
and women who use IPA. Male and female perpetrators of
IPA should be assessed for evidence of BP and BPD, and
should be provided treatment that addresses these issues
where appropriate. Although traditionally BP is considered
to be a stable characteristic that has a poor prognosis af-
ter diagnosis (American Psychiatric Association, 1994), re-
cent evidence shows that certain types of therapy, such as
“Conversational Models” of psychotherapy that meet twice

Borderline Personality and Aggression

301

a week for 1 year (Stevenson, Meares, & D’Angelo, 2005),
have shown success and long-term stability of therapy gains.
Thus, male and female perpetrators of IPA who evidence
BP should undergo lengthy and intensive treatment for
their BP. However, because there is no evidence as of yet
that successfully treating BP will result in a decrease in
IPA, such treatment should be given alongside interven-
tions that address other etiological influences on IPA (e.g.,
alcohol abuse, experience of IPA in the family of origin;
Mauricio et al., 2007).

Initial submission: August 28, 2007
Initial acceptance: March 4, 2008
Final acceptance: May 6, 2008

REFERENCES

American Psychiatric Association. (1994). Diagnostic and statisti-

cal manual of mental disorders (4th ed.). Washington, DC:
Author.

Archer, J. (1999). Assessment of the reliability of the Conflict Tac-

tics Scales: A meta-analytic review. Journal of Interpersonal
Violence, 14, 1263–1289.

Archer, J. (2000). Sex differences in aggression between hetero-

sexual couples: A meta-analytic review. Psychological Bul-
letin, 126, 651–680.

Busch, A., & Rosenberg, M. (2004). Comparing women and men

arrested for domestic violence: A preliminary report. Jour-
nal of Family Violence, 19, 49–58.

Buzawa, E. S., & Buzawa, C. G. (2002). Domestic violence: The

criminal justice response. Thousand Oaks, CA: Sage.

Cantos, A. L., Neidig, P. H., & O’Leary, K. D. (1994). Injuries

of women and men in a treatment program for domestic
violence. Journal of Family Violence, 9, 113–123.

Capaldi, D. M., & Owen, L. D. (2001). Physical aggression in a

community sample of at-risk young couples: Gender com-
parisons for high frequency, injury, and fear. Journal of
Family Psychology, 15, 425–440.

Carney, M. M., & Buttell, F. P. (2004). A multidimensional eval-

uation of a treatment program for female batterers: A pilot
study. Research on Social Work Practice, 14, 249–258.

Carrado, M., George, M., Loxam, F., Jones, L., & Templar, D.

(1996). Aggression in British heterosexual relationships: A
descriptive analysis. Aggressive Behavior, 22, 401–415.

Chavira, D. A., Grilo, C. M., Shea, M. T., Yen, S., Gunderson,

J. G., Morey, L. C., et al. (2003). Ethnicity and four person-
ality disorders. Comprehensive Psychiatry, 44, 483–491.

Consalvi, C. (1972). An item and factor analysis of Danish,

Lebanese and United States college students’ responses
to the Marlow-Crowne Social Desirability Scale. Journal of
Cross-Cultural Psychology, 3, 361–372.

Cooper, M. L. (2002). Personality and close relationships: Embed-

ding people in important social contexts. Journal of Person-
ality, 70, 757–782.

Crowne, D. P., & Marlowe, D. (1964). The approval motive. New

York: Wiley.

Demaris, A. (1992). Male versus female initiation of aggression:

The case of courtship violence. In E. C. Viano (Ed.), In-
timate violence: Interdisciplinary perspectives (pp. 111–
120). Bristol, PA: Taylor and Francis.

Dutton, D. G. (1995). Male abusiveness in intimate relationships.

Clinical Psychology Review, 15, 567–581.

Dutton, D. G., & Browning, J. J. (1988). Power struggles and in-

timacy anxieties as causative factors of violence in intimate
relationships. In G. Russell (Ed.), Violence in intimate rela-
tionships (pp. 163–175). Great Neck, NY: PMA Publishing.

Dutton, D. G., & Starzomski, A. J. (1993). Perpetrator character-

istics associated with women’s reports of psychological and
physical abuse. Violence and Victims, 8, 327–337.

Dutton, D. G., Starzomski, A. J., & Ryan, L. (1996). Antecedents

of abusive personality and abusive behavior in wife as-
saulters. Journal of Family Violence, 11, 113–132.

Gelles, R. J. (1974). The violent home: A study of physical ag-

gression between husbands and wives. Beverly Hills, CA:
Sage.

Giordano, P. C., Millhollin, T. J., Cernkovich, S. A., Pugh, M. D., &

Rudolph, J. L. (1999). Delinquency, identity, and women’s
involvement in relationship violence. Criminology, 37, 17–
40.

Hart, S. D., Dutton, D. G., & Newlove, T. (1993). The prevalence

of personality disorder amongst wife assaulters. Journal of
Personality Disorders, 7, 328–340.

Hastings, J. E., & Hamberger, L. K. (1988). Personality character-

istics of spouse abusers: A controlled comparison. Violence
and Victims, 3, 31–48.

Henning, K., Jones, A., & Holdford, R. (2003). Treatment needs

to women arrested for domestic violence: A comparison
with male offenders. Journal of Interpersonal Violence, 18,
839–856.

Hines, D. A., Brown, J., & Dunning, E. (2007). Characteristics of

callers to the Domestic Abuse Helpline for Men. Journal
of Family Violence, 22, 63–72.

Hines, D. A., & Saudino, K. J. (2003). Gender differences in psy-

chological, physical, and sexual aggression among college
students using the Revised Conflict Tactics Scales. Violence
and Victims, 18, 197–218.

Hines, D. A., & Straus, M. A. (2007). Binge drinking and violence

against dating partners: The mediating effect of antisocial
traits and behaviors in a multi-national perspective. Aggres-
sive Behavior, 33, 441–457.

Holtzworth-Munroe, A., Bates, L., Smutzler, N., & Sandin, E.

(1997). A brief review of the research on husband violence.
Part I: Maritally violent versus nonviolent men. Aggression
and Violent Behavior, 2, 65–99.

Holtzworth-Munroe, A., Meehan, J. C., Herron, K., Rehman, U.,

& Stuart, G. L. (2000). Testing the Holtzworth-Munroe
and Stuart (1994) batterer typology. Journal of Consulting
and Clinical Psychology, 69, 1000–1019.

Ikuta, N., Zanarini, M. C., Minakawa, K., Miyake, Y., Moriya,

N., & Nishizono-Maher, A. (1994). Comparison of Amer-
ican and Japanese outpatients with borderline per-
sonality disorder. Comprehensive Psychiatry, 35, 382–
385.

Krueger, R. F., Caspi, A., & Moffitt, T. E. (2000). Epidemiological

personology: The unifying role of personality in population-
based research on problem behaviors. Journal of Personal-
ity, 68, 967–998.

Langhinrichsen-Rohling, J., Neidig, P. H., & Thorn, G. (1995).

Violent marriages: Gender differences in levels of current
violence and past abuse. Journal of Family Violence, 10,
159–176.

302

H

INES

Laroche, D. (2005). Aspects of the context and consequences of

domestic violence–situational couple violence and intimate
terrorism in Canada in 1999. Quebec, Canada: Institut de
la statistique du Quebec.

Loranger, A. W., Sartorius, N., Andreoli, A., Berger, P.

Buchheim, P., Channabasavanna, S. M. et al. (1994).
The International Personality Disorder Examination: The
World Health Organization/Alcohol, Drug Abuse, and
Mental Health Administration international pilot study of
personality. Archives of General Psychiatry, 51, 215–224.

Lysova, A. V., & Hines, D. A. (in press). Binge drinking and

violence against intimate partners in Russia. Aggressive Be-
havior.

Magdol, L., Moffitt, T. E., Caspi, A., & Silva, P. A. (1998). De-

velopmental antecedents of partner abuse: A prospective-
longitudinal study. Journal of Abnormal Psychology, 107,
375–389.

Margolies, L., & Leeder, E. (1995). Violence at the door: Treat-

ment of lesbian batterers. Violence Against Women, 1, 139–
157.

Mauricio, A. M., Tein, J.-Y., & Lopez, F. G. (2007). Borderline and

antisocial personality scores as mediators between attach-
ment and intimate partner violence. Violence and Victims,
22, 139–157.

Medeiros, R. A., & Straus, M. A. (2006). Risk factors for physical

violence between dating partners: Implications for gender-
inclusive prevention and treatment of family violence. In
J. Hamel & T. Nicholls (Eds.), Family interventions in do-
mestic violence: A handbook of gender-inclusive theory and
treatment (pp. 59–85). New York: Springer.

Middleton, K. L., & Jones, J. L. (2000). Socially desirable re-

sponse sets: The impact of country culture. Psychology &
Marketing, 17, 149–163.

Moffitt, T. E., Caspi, A., Krueger, R. F., Magdol, L., Margolin,

G., Silva, P. A., et al. (1997). Do partners agree about abuse
in their relationship?: A psychometric evaluation of inter-
partner agreement. Psychological Assessment, 9, 47–56.

Morse, B. J. (1995). Beyond the Confict Tactics Scales: Assess-

ing gender differences in partner violence. Violence and
Victims, 10, 251–272.

Mukherjee, B. N. (1967). A cross-validation of the Marlowe-

Crowne social desirability scale on an Indian sample. Jour-
nal of Social Psychology, 72, 299–300.

Oldham, J., Clarkin, J., Appelbaum, A., Carr, A., Kernberg, P.,

Lotterman, A., et al. (1985). A self-report instrument for
borderline personality organization. In T. H. Mcglashan
(Ed.), The borderline: Current empirical research (pp. 1–
18). Washington, DC: American Psychiatric Press.

Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear

models: Applications and data analysis methods (2nd ed.).
Thousand Oaks, CA: Sage.

Serman, N., Johnson, J. G., Geller, P. A., Kanost, R. E., & Zachara-

poulou, H. (2002). Personality disorders associated with
substance use among American and Greek adolescents.
Adolescence, 37, 841–854.

Stets, J. E., & Straus, M. A. (1990). Gender differences in report-

ing marital violence and its medical and psychological con-
sequences. In M. A. Straus & R. J. Gelles (Eds.), Physical
violence in American families: Risk factors and adaptation
to violence in 8,145 families (pp. 151–166). New Brunswick,
NJ: Transaction.

Stevenson, J., Meares, R., & D’Angelo, R. (2005). Five-year out-

come of outpatient psychotherapy with borderline patients.
Psychological Medicine, 35, 79–87.

Straus, M. A. (1969). Phenomenal identity and conceptual equiv-

alence of measurement in cross-national comparative re-
search. Journal of Marriage and the Family, 31, 233–
239.

Straus, M. A. (2004). Cross cultural reliability and validity of the

Revised Conflict Tactics Scales: A study of university dating
couples in 17 nations. Cross-Cultural Research, 38, 407–
432.

Straus, M. A. (2006). Future research on gender symmetry in

physical assaults on partners. Violence Against Women, 12,
1086–1097.

Straus, M. A., & the International Dating Violence Research Con-

sortium. (2004). Prevalence of violence against dating part-
ners by male and female university students worldwide.
Violence Against Women, 10, 790–811.

Straus, M. A., & Gelles, R. J. (Eds.). (1990). Physical violence in

American families. New Brunswick, NJ: Transaction.

Straus, M. A., Hamby, S. L., Boney-McCoy, S., & Sugarman,

D. (1996). The Revised Conflict Tactics Scales (CTS-2):
Development and preliminary psychometric data. Journal
of Family Issues, 17, 283–316.

Straus, M. A., Hamby, S. L., Boney-McCoy, S., & Sugar-

man, D. (1999). The Personal and Relationships Profile
(PRP). Retrieved June 15, 2007, from http://pubpages.
unh.edu/

∼mas2/.

Straus, M. A., & Mouradian, V. E. (1999). Preliminary psycho-

metric data from the Personal and Relationships Profile
(PRP): A multi-scale tool for clinical screening and re-
search on partner violence. Retrieved June 15, 2007, from
http://pubpages.unh.edu/

∼mas2/

Stuart, G. L., Moore, T. M., Gordon, K. C., Ramsey, S. E., &

Kahler, C. W. (2006). Psychopathology in women arrested
for domestic violence. Journal of Interpersonal Violence,
21, 376–389.

Swan, S. C., & Snow, D. L. (2002). A typology of women’s use of vi-

olence in intimate relationships. Violence Against Women,
8, 286–319.

Tsushima, W. T. (1969). Responses of Irish and Italians of

two social classes on the Marlowe-Crowne Social De-
sirability scale. Journal of Social Psychology, 77, 215–
219.

World Health Organization. (2005). WHO multi-country

study on women’s health and domestic violence against
women: Initial results on prevalence, health out-
comes and women’s responses. Geneva, Switzerland:
Author.