HOPWOOD AND MOREY

Borderline Personality

PSYCHOLOGICAL CONFLICT IN
BORDERLINE PERSONALITY AS
REPRESENTED BY INCONSISTENT
SELF–REPORT ITEM RESPONDING

CHRISTOPHER J. HOPWOOD AND LESLIE C. MOREY
Texas A&M University

Interpersonal theory conceptualizes variants of personality pathology as describ-
able by two dimensions: affiliation and control. Inconsistent results in placing bor-
derline personality along these dimensions have been interpreted by some as
representing internal conflict or ambivalence on these dimensions, as opposed to a
lack of relevance of these dimensions. This hypothesis was tested in a large clinical
sample using inconsistency in self–report item responding on scales measuring af-
filiation and control to operationalize psychological conflict. Individuals with bor-
derline personality features were more inconsistent in item responding to both
scales than were individuals without borderline features. Item response inconsis-
tency did not differentiate antisocial from non–antisocial participants. Results sup-
port the view that variability, as well as mean scores, on the interpersonal
dimensions may be important for the conceptualizing some disorders, such as
borderline, and offers a novel approach for representing such conflict.

Interpersonal theory (e.g., Kiesler, 1996; Leary, 1957) typically concep-
tualizes personality in terms of one’s standing on the interpersonal
circumplex (see Benjamin, 1974 for a notable exception), which is char-
acterized by two orthogonal dimensions: Control (dominant to submis-
sive) and Affiliation (warm to cold; Figure 1). Research has validated
this hypothesis with most, but not all, personality disorders, by project-

Journal of Social and Clinical Psychology, Vol. 26, No. 9, 2007, pp. 1065–1075

1065

Portions of this manuscript were presented at the Society for Interpersonal Theory and

Research (SITAR), Philadelphia, PA, May 5, 2006.

Christopher J. Hopwood, M.S., Department of Psychiatry, Massachusetts General Hos-

pital and Harvard Medical School, Department of Psychology, Texas A&M University,
Leslie C. Morey, Ph.D., Department of Psychology, Texas A&M University.

Correspondence concerning this article should be addressed to Christopher J.

Hopwood at [e-mail]

ing personality disorder scales onto circumplex space. For example, in
13 efforts to map antisocial personality onto the circumplex using vari-
ous scales, all 13 yielded significant projections, and 11 of 13 projected
onto the cold–dominant quadrant of the interpersonal circumplex
(Blackburn, 1998; DeJong, , van den Brink, Jansen, & Schippers, 1989;
Matano & Locke, 1995; Morey, 1985, 1991; Pincus & Wiggins, 1990; Sim
& Romney, 1990; Soldz, Budman, Demby, & Merry, 1993; Wiggins &
Pincus, 1989; several of these papers reported multiple studies; results
depicted in Figure 1). However, only 8 of 13 studies yielded significant
projections of borderline personality onto the circumplex, and results
were inconsistent: three projected it onto the cold–submissive quadrant,
three onto the warm–submissive quadrant, and two onto the warm–do-
minant quadrant.

Perhaps in part because of these findings, interpersonal researchers

have begun to view borderline personality as characterized by psycho-
logical conflict, particularly along the Affiliation dimension (e.g.,
Benjamin, 1993; Kiesler, 1996). Psychological conflict would help ex-
plain why the projections of borderline features onto the circumplex
have been unremarkable and inconsistent. For example, a person who
presents themselves as both extremely warm and extremely cold will
have a neutral average self–representation (Gurtman, 1994). In this case,
the consistency of self–representation would be necessary to differenti-
ate the borderline individual from a person who consistently presents
themselves as neutral with respect to Affiliation. Otherwise, a person
who is neither warm nor cold would be considered the same as a person
who is both warm and cold.

This conception is consistent with other theoretical perspectives and

empirical data on borderline personality. Recent research suggests that
individuals with borderline personality are affectively unstable (e.g.,
Cowdry et al., 1991; Koenisberg et al., 2001), and that they have unstable
self–concepts (Zeigler–Hill & Abraham, 2006). To the extent that vari-
ability in affect and self–concept influences interpersonal style, it might
be predicted that interpersonal behavior, and thus patterns of item re-
sponding on scales measuring interpersonal styles, would be more vari-
able among individuals with borderline personality. This logic is sup-
ported by research using interpersonal circumplex methods to
demonstrate that neuroticism (emotional instability) is associated with
cross–situational variability in interpersonal behavior (Moskowitz &
Zuroff, 2004). The association of borderline personality with psycholog-
ical conflict and, by extension, inconsistency in interpersonal style, is ex-
plicit in psychodynamic theory. Kernberg (1984) described borderline
personality organization in terms of identity diffusion, or the lack of a
fundamental sense of self necessary to manifest a coherent identity. He

1066

HOPWOOD AND MOREY

further hypothesized that the fundamental conflict associated with bor-
derline personality involves “splitting,” or ambivalence regarding at-
tachment (Affiliation), in which the individual with borderline person-
ality vacillates between idealization (i.e., warmth) and paranoia (i.e.,
coldness). Similarly, attachment researchers conceptualize individuals
with borderline personality as having conflicting attachment goals that
create “an intense, confusing, and dissatisfying rollercoaster of
vacillating approach and avoidance efforts” (Meyer & Pilkonis, 2005, p.
253).

BORDERLINE PERSONALITY

1067

FIGURE 1. Approximate Placement of Borderline and Antisocial Personality
on the Interpersonal Circumplex Across 13 Studies.

Note. Antisocial personality represented by normal font, borderline personal-
ity is bold and underlined. Tic marks represent r = .30, perimeter represents r
= .60. Studies represented by first authors’ last names; superscripts indicate
two studies in the same publications except for Morey, which represents two
publications (Morey

1

1985, Morey

2

1991). Morey

2

represents the same data

used in the current study.

MEASURING PSYCHOLOGICAL CONFLICT

Psychological conflict has been studied using several methodologies.
Bartholomew and colleagues (Bartholomew & Horowitz, 1991; Griffin &
Bartholomew, 1994) have demonstrated that different attachment cate-
gories were differentially related to interpersonal profile complexity.
Gurtman and Balakrishan (1998) described a coefficient that represents
the statistical fit of an observed profile with a profile that conforms to the
ideal theoretical characteristics of an interpersonal problems
circumplex. They noted that profiles in which this index was low may
represent psychological conflict, as this would indicate relative eleva-
tions on non–adjacent octants, inconsistent with the assumption of
circumplexity. However, using the Inventory of Interpersonal Prob-
lems—Circumplex (Alden, Wiggins, & Pincus, 1990), which includes
eight scales representative of each interpersonal circumplex octant, Has-
lam and Gurtman (1999) found that a large majority of interpersonal
problem profiles were not characterized by complex (e.g., conflictual)
patterns, which they operationalized as elevations on opposite (psycho-
logically conflicting; e.g., warm problems and cold problems) scales.
Furthermore, complex patterns were equally common in clinical and
non–clinical samples, further supporting the authors’ contention that
complex patterns primarily resulted from measurement error, and sim-
ple patterns reflecting a single, prevailing interpersonal theme were
likely to be more clinically useful. However, Haslam and Gurtman also
noted that their results may have related to their method of
operationalizing psychological conflict.

Another approach involves investigating the patterns of specific

cross–situational behaviors. One set of behaviors that can be sampled
with relative ease involves the pattern of item responses to psychometric
instruments. Interitem variability (IIV) is an index of the variability of
item responses around an ipsative mean of item scores. It indicates, on
average, how variable a person is in responding to a set of items measur-
ing the same trait. IIV has been used in previous research to
operationalize traitedness, or the extent to which a psychological trait is
salient for an individual. Research has used several indices in addition
to IIV to represent cross–situational consistency. However, results in
this area have been inconsistent, whether IIV or other methods were
used (Tellegen, 1988).

Assuming respondents attended to item content, psychological con-

flict thus represents an alternative to the traitedness hypotheses for in-
consistency in item responding. The logic of traitedness is that individu-
als for whom a trait is not salient (i.e., the trait would not predict their
behavior reliably) will answer items on that scale in an inconsistent fash-

1068

HOPWOOD AND MOREY

ion. The alternative perspective assumes that the endorsement of con-
ceptually opposing items occurs when both are true to the respondent
because item content corresponds to both sides of a psychological con-
flict. For example, an individual with borderline personality features
may endorse items on an affiliation scale reflecting his or her desire for
closeness (i.e., “I am warm because I want others to be warm to me”) and
also be interpersonally guarded and anticipate rejection (i.e., “I am cold
to others because I expect them to be cold to me”). Such a pattern would
be consistent with theoretical articulations of borderline
phenomenology (Kernberg, 1984).

The purpose of this study is to test the hypothesis that IIV on affiliation

and control scales would differentiate individuals with and without
borderline personality, but would not differentiate individuals with and
without antisocial personality, in a clinical sample. A secondary hypoth-
esis that IIV is a substantive psychological construct, rather than an indi-
cator of random error, as the traitedness perspective would predict, was
also tested by comparing the validity coefficients of the affiliation and
control scales with external criteria for a group with high vs. low scores
on the borderline scale. Thus, if (consistent with the traitedness interpre-
tation) the higher IIV in high–borderline individuals reflects random
measurement error, validity coefficients for the interpersonal scales
would be expected to decrease in this group. However, if IIV reflects a
substantive concept, i.e., psychological conflict, then the validity
coefficient for the interpersonal scales should be comparable for both
high– and low–borderline groups.

METHOD

PARTICIPANTS

Data from the Personality Assessment Inventory (PAI; Morey, 1991)
clinical standardization sample were used to test these hypotheses. Par-
ticipants were screened for invalid protocols according to guidelines
discussed by Morey (1991), resulting in 1,079 (from the original 1,246) in-
dividuals recruited from 69 diverse clinical sites. The majority of partici-
pants (86.9%) were younger than 50 and women represented 38.6% of
the sample, which was 78.8% Caucasian and 12.6% African American.

MEASURES

The PAI is a 344–item questionnaire with a 4–point response option and
scales that have no overlapping items. Eight full scales from the Person-
ality Assessment Inventory were used in the current study. The Border-

BORDERLINE PERSONALITY

1069

line Features (BOR) and Antisocial Features (ANT) scales of the PAI
were used to operationalize those disorders, with a split at 70t (i.e., two
standard deviations among the community mean) placing individuals
into disordered and not–disordered categories. This cutoff has been
used effectively to differentiate borderline from non–borderline partici-
pants in previous research (e.g., Trull, Useda, Conforti, & Doan, 1997). In
the current sample, 325 participants were above the cutoff on BOR, 147
on ANT, and 94 above the cutoff on both scales. The internal consistency
alphas in the current sample for these scales were .91 and .86, respec-
tively, and test–retest reliabilities (retest interval averaged 24 days) in a
non–clinical sample were .90 for both scales. Both scales have been vali-
dated on a number of criterion measures in the validation studies
(Morey, 1991) and subsequent research.

The PAI Warmth (WRM) and Dominance (DOM) scales were used to

operationalize affiliation and control, respectively. These 12 item scales
demonstrated coefficient alphas in the current sample of .83 and .82, re-
spectively (Morey, 1991). IIV scores were computed for WRM and DOM
that reflected the standard deviation of the PAI normative sample–stan-
dardized item scores around personal (ipsative) means for each respon-
dent (Tellegen, 1988). Given the research question, item content is perti-
nent. The WRM items with the largest item–total correlations are “I’m a
very sociable person” and “It is easy for me to make new friends,”
whereas DOM items with the largest item–total correlations are “I’m a
‘take–charge’ type of person” and “I’m a natural leader.”

Several other PAI scales served as validating variables. The Anxiety

(ANX; 24 items; alpha = .94, test–retest = .88) and Depression (DEP; 24
items; alpha = .93, test–retest = .91) scales were used to represent internal
characteristics theoretically associated with interpersonal styles, and the
Aggression (AGG; 24 items; alpha = .90, test–retest = .85) and
Nonsupport (NON; 8 items; alpha = .80, test–retest = .81) scales to repre-
sent interpersonal characteristics. Finally, the Inconsistency (ICN) scale
was used to represent generalized inconsistency. The ICN scale consists
of 10 pairs of items with highly related content; high scores are achieved
on ICN by answering similar items inconsistently. Given that the ICN
scale is not measuring a construct but a tendency respond inconsistently
that might be influenced by situational factors such as inattention and
carelessness, it is not surprising that the reliability is relatively low
(alpha = .23, test–retest = .29).

ANALYSES

To test the hypothesis that item response inconsistency would incre-
ment interpersonal styles in understanding borderline personality, the

1070

HOPWOOD AND MOREY

mean IIV differences on WRM and DOM in borderline and non–border-
line and antisocial and non–antisocial groups were tested in four
one–way ANCOVAs. Because WRM and IIV on WRM are correlated, (r
= –.25, p

<

.001), indicating that individuals with low warmth scores are

more likely to be inconsistent in item responding than are individuals
with high warmth scores, and because both borderline and antisocial
features are associated with low WRM scores (Morey, 1991), the total
WRM score was included as a covariate in this analysis. Similarly, DOM
was included in analyses with respect to IIV on DOM, for the same rea-
sons (r = –.15, p

<

.001, antisocial features associated with high and bor-

derline features with low DOM scores). The ICN scale was also included
as a covariate, to control for a generalized propensity to respond in an in-
consistent fashion, in order to demonstrate that inconsistent responding
by borderline respondents on the interpersonal scales was above and be-
yond any generalized inconsistency observed in this group. Finally, to
account for overlap between the scales in identifying personality pathol-
ogy, as well as comorbidity, dimensional ANT scores were entered as
covariates in analyses investigating borderline vs. non–borderline
differences, and BOR scores were entered as covariates in analyses
investigating antisocial vs. non–antisocial differences.

To test the differential predictions of the traitedness and ambivalence

hypotheses, correlations between WRM and DOM and several validat-
ing scales were conducted. Four scales were chosen for this purpose; two
representing internal characteristics (anxiety and depression) and two
representing interpersonal characteristics (aggression and nonsupport).
The moderation of the interpersonal scale–validating variable (i.e.,
WRM, DOM) relationships by borderline status was investigated by
testing the interaction of the interpersonal scales and borderline status,
after accounting for their main effects, in predicting each validating
variable (Baron & Kenny, 1986).

RESULTS

Results for the first set of analyses are depicted in Table 1. The univariate
ANCOVAs that tested the relation of borderline status to IIV on the
WRM and DOM scales after controlling for ICN, ANT, and WRM/DOM
were significant (p

<

.01) and demonstrated moderate effects. Con-

versely, the analogous tests for antisocial status were not significant. As
expected, both ICN and the interpersonal scale covariates were also sta-
tistically significant across all four analyses. The BOR covariate was sta-
tistically significant (p

<

.01) in the antisocial analyses, but the ANT

covariate was not significant (p

>

.10) in the borderline analyses. These

BORDERLINE PERSONALITY

1071

results support the hypothesis that IIV is a substantive phenomenon that
differentiates individuals in groups predicted by the theoretical litera-
ture to differ in terms of psychological conflict, but does not differentiate
individuals in groups that are not predicted to differ in this regard.

Given the support for substantive meaning of IIV with respect to bor-

derline personality, the moderation of WRM/DOM–validating variable
relations by borderline status was investigated to test differential predic-
tions by the traitedness and conflict hypotheses. Table 2 depicts these re-
sults. Contrary to the traitedness hypotheses, the validity coefficients for
the WRM and DOM scales in predicting internal (depression and anxiety)
and interpersonal (nonsupport and aggression) variables did not differ in
groups with greater (borderline) vs. lesser (non–borderline) levels of IIV.
Moderation tests across all four variables were statistically non–signifi-
cant (p

>

.05) for both WRM and DOM. These results support the conten-

tion that IIV in this context represents substantive meaning that does not
detract from the meaning of the scale on which item responses were in-
consistent, and should not be used as an indication of trait relevance.

DISCUSSION

The current study tested the hypothesis that inconsistency in item re-
sponding (IIV) on interpersonal scales represents psychological conflict
associated with borderline, but not antisocial, personality features. Re-
sults demonstrate that, as predicted by attachment, interpersonal,
psychodynamic, and other clinical personality theories, individuals

1072

HOPWOOD AND MOREY

TABLE 1. Unadjusted Mean Inter–Item Variability on the PAI Warmth and Dominance

Scales for Patients with and without Borderline and Antisocial Features

Mean

SD

D

F

(1, 1075)

p

WRM

BOR+

.95

.54

.35

7.125

<

.01

BOR–

.77

.50

ANT+

.94

.58

.25

1.64

ns

ANT–

.81

.51

DOM

BOR+

1.04

.56

.38

16.18

<

.01

BOR–

.84

.50

ANT+

.98

.64

.18

.124

ns

ANT–

.89

.50

Note. F–tests after partialling out variance in IIV scores associated with PAI ICN scale and trait (WRM,
DOM) scale scores, which were significant across all models. Also covaried were PAI BOR (when ANT
was the independent variable)/ANT (when BOR was the independent variable) scores. BOR was a signifi-
cant covariate in ANT analyses, ANT was not significant in BOR analyses.

with borderline features are more psychologically conflicted regarding
affiliation and control than are individuals who do not have borderline
features, whereas psychological conflict with regard to interpersonal
style does not predict one’s standing with respect to antisocial features.

Furthermore, results indicate that the meaning of the scale on which

individuals were inconsistent did not change as a function of that incon-
sistency. Thus, the assertion made by traitedness researchers that IIV, or
cross–situational inconsistency in general, identifies the salience of traits
for individuals (i.e., the extent to which those traits can be expected to
predict behavior) was refuted in these data. The correlations of interper-
sonal scales with aggression, nonsupport, depression, and anxiety, did
not differ for inconsistent (i.e., high borderline) and consistent (i.e., low
borderline) respondents. Instead, results suggest that both an individ-
ual’s overall interpersonal style, as well as their psychological conflict
regarding this style, are important elements of borderline personality.

These results have implications for both psychological measurement

and clinical personality theory. Regarding measurement, results sug-
gest that IIV, and perhaps other forms of cross–situational inconsis-
tency, do not always represent random error, and point towards future
investigations of inconsistency that treat the phenomena as a meaning-
ful variable that may represent, among other possibilities, psychological
conflict or differentiation (i.e., the healthy capacity to adopt varying
personality styles according to situations).

Regarding clinical personality theory, results are consistent with theo-

ries that have posited psychological conflict regarding interpersonal
motivations and behavior as relevant for understanding borderline per-
sonality. Interestingly, the current results are inconsistent with those of
Haslam and Gurtman (1999), who performed perhaps the most sophisti-
cated test of psychological conflict indicators from the interpersonal per-
spective. This may relate to either instrument differences (Haslam and
Gurtman tested the conflict hypotheses using an interpersonal problems

BORDERLINE PERSONALITY

1073

TABLE 2. Bivariate Correlations Between The PAI Interpersonal Scale and Selected

Validating Constructs for Borderline and Non–Borderline Groups

WRM

DOM

BOR+

BOR–

BOR+

BOR–

Nonsupport

–.42

–.44

–.15

–.19

Aggression

–.20

–.15

.32

.30

Depression

–.43

–.43

–.43

–.40

Anxiety

–.22

–.28

–.37

–.36

Note. All correlations were statistically significant, p

<

.001. Moderation tests indicated none of these cor-

relations significantly differed across groups.

circumplex, whereas interpersonal styles were tested in the current re-
port) or methodological differences. Direct tests of the relation of IIV on
interpersonal problems scales and borderline personality, and of meth-
ods used in their study and IIV are not possible given the current data,
because interpersonal styles (and not problems) are indicated by two
(and not eight) scales on the PAI. However, IIV on the interpersonal di-
mensions could be tested in studies in which octant measures of inter-
personal styles and problems and compared to results from analyses
similar to those used by Haslam and Gurtman, and this would appear to
be an important area of future research.

Another area for future research involves the relations between inter-

personal variability, neuroticism, and borderline personality features.
Moskowitz and Zuroff (2004) observed that neuroticism relates to
cross–situational variability in interpersonal behavior, although they
measured interpersonal variability across real–world interpersonal sit-
uations, rather than items on a psychological scale as was the case here.
Future research should compare cross–situational variability as mea-
sured by Moskowitz and Zuroff with IIV. This would be particularly im-
portant because one commonly proposed solution to the difficulty in
capturing the personality disorders within an interpersonal framework
is broadening the number of factors, as with the big five (e.g., Costa &
McCrae, 1990; Kiesler, 1996), and neuroticism has consistently been
found to relate to borderline personality (e.g., Morey et al., 2002). The re-
lation of variability in interpersonal behavior, measured as cross–situa-
tional variability, IIV, or by some other means, to neuroticism and the
unique vs. redundant relations of each to borderline personality are thus
important questions for future research. In any case, the current data
suggest that researchers should continue to explore the relation of intra
individual variability to borderline personality features, and that IIV
may be a promising technique for the identification of one important
aspect of such variability, psychological conflict.

REFERENCES

Bartholomew, K., & Horowitz, L.M. (1991). Attachment styles among young adults: A test

of a four–category model. Journal of Personality and Social Psychology, 61(2), 226–244.

Benjamin, L.S. (1974). Structural analysis of social behavior. Psychological Review, 81,

392–425.

Benjamin, L.S. (1993). Interpersonal Diagnosis and Treatment of Personality Disorders. New

York: Guilford.

Blackburn, R. (1998). Relationship of personality disorders to observer ratings of interper-

sonal style in forensic psychiatric patients. Journal of Personality Disorders, 12(1),
77–85.

1074

HOPWOOD AND MOREY

Costa, P.T., Jr., & McCrae, R.R. (1990). Personality disorders and the five–factor model of

personality. Journal of Personality Disorders, 4, 362–371.

Cowdry, R.W., Gardner, D.L., O’Leary, K.M., Leibenluft, E., & Rubinow, D.R. (1991). Mood

variability: A study of four groups. American Journal of Psychiatry, 148, 1505–1511.

DeJong, C.A.J., van den Brink, W., Jansen, J.A.M., & Schippers, G.M. (1989). Interpersonal

aspects of DSM–III Axis II: Theoretical hypotheses and empirical findings. Journal
of Personality Disorders, 3(2), 135–146.

Griffin, D., & Bartholomew, K. (1994). Models of the self and other: Fundamental dimen-

sions underlying measures of adult attachment. Journal of Personality and Social Psy-
chology, 67(3), 430–445.

Haslam, N., & Gurtman, M.B. (1999). Detecting complex patterns in interpersonal profiles.

British Journal of Medical Psychology, 72, 23–32.

Kernberg, O.F. (1984). Severe Personality Disorders. New Haven, CT: Yale University Press.
Kiesler, D.J. (1996). Contemporary Interpersonal Theory and Research. New York: Wiley.
Koenisberg, H.W., Harvey, P.D., Mitropoulou, V., New, A.S., Goodman, M., Silverman, J.,

Serby, M., Schopick, F., & Siever, L.J. (2001). Are the interpersonal and identity dis-
turbances in borderline personality disorder criteria linked to the traits of affective
instability and impulsivity? Journal of Personality Disorders, 15(4), 358–370.

Leary, T.F. (1957). Interpersonal Diagnosis of Personality. New York: Ronald.
Matano, R.A., & Locke, K.D. (1995). Personality disorder scales as predictors of interper-

sonal problems in alcoholics. Journal of Personality Disorders, 9(1), 62–67.

Meyer, B., & Pilkonis, P.A. (2005). An attachment model of personality disorders. In M.F.

Lenzenweger and J.F. Clarkin (Eds.), Major Theories of Personality Disorder, 2nd Ed.
(pp. 231–281). New York: The Guilford Press.

Morey, L.C. (1985). An empirical comparison of interpersonal and DSM–III approaches to

classifcation of personality disorders. Psychiatry, 48, 358–364.

Morey, L.C. (1991). Personality Assessment Inventory Manual. Odessa, FL: Psychological As-

sessment Resources.

Morey, L.C., Gunderson, J.G., Quigley, B.D., Shea, M.T., Skodol, A.E., McGlashan, T.H.,

Stout, R.L., & Zanarini, M.C. (2002). The representation of borderline, avoidant, ob-
sessive–compulsive, and schizotypal personality disorders by the five–factor
model. Journal of Personality Disorders, 16, 215–234.

Moskowitz, D.S., & Zuroff, D.C. (2004). Flux, Pulse, and Spin: Dynamic additions to the

personality lexicon. Journal of Personality and Social Psychology, 86(6), 880–893.

Sim, J.P., & Romney, D.M. (1990). The relationship between a circumplex model of inter-

personal behaviors and personality disorders. Journal of Personality Disorders, 4(4),
329–341.

Soldz, S., Budman, S., Demby, A., & Merry, J. (1993). Representation of personality disor-

ders in circumplex and five–factor space: Explorations within a clinical sample. Psy-
chological Assessment, 5(1), 41–52.

Tellegen, A. (1988). The analysis of consistency in personality assessment. Journal of Person-

ality, 56, 621–663.

Trull, T.J., Useda, J.D., Conforti, K., & Doan, B.T. (1997). Borderline personality disorder

features in nonclinical young adults: 2. Two–year outcome. Journal of Abnormal Psy-
chology, 106, 307–314.

Wiggins, J.S., & Pincus, A.L. (1989). Conceptions of personality disorders and dimensions

of personality. Psychological Assessment: A Journal of Consulting and Clinical Psychol-
ogy, 1(4), 305–316.

Zeigler–Hill, V., & Abraham, J. (2006). Borderline personality features: Instability of

self–esteem and affect. Journal of Social and Clinical Psychology, 25(6), 668–687.

BORDERLINE PERSONALITY

1075