Interpersonal evaluation bias in borderline personality disorder

Abstract

Background

The cognitive theory of personality disorders hypothesizes that the emotional dysregulation and interpersonal problems in individuals with borderline personality disorder (BPD) are, at least partially, caused by dysfunctional cognitive schemas. These schemas lead to biased evaluation of environmental and interpersonal stimuli.

Method

This study examined the interpersonal evaluations of individuals with BPD, depressive and healthy control participants with the thin-slice judgments paradigm. Participants were asked to evaluate six persons in six film clips, which showed these persons for 10 s, during which these persons entered a room and took a seat. Interpersonal style of the BPD group was investigated with the Inventory of Interpersonal Problems (IIP-C) questionnaire.

Results

Individuals with BPD judged the persons as being more negative and aggressive and less positive than the healthy participants, and more aggressive than the depressive individuals. In addition, individuals with BPD reported more extreme interpersonal behavior relative to the controls.

Conclusions

The findings indicate an aggressivistic evaluation bias and elevated levels of interpersonal problems in individuals with BPD as suggested in the cognitive theory.

Introduction

Borderline personality disorder (BPD) is characterized by marked problems in emotion regulation, impulsivity and poor interpersonal relationships (American Psychiatric Association, 2000). The cognitive theory of personality disorders (PDs) hypothesizes that these problems are at least partly caused by cognitive schemas leading to biased evaluation and interpretation of environmental stimuli (Beck and Freeman, 1990, Beck et al., 2004). Beliefs or schemas are thought to be basic processing units that are organized according to their functions and content. A schema is considered hypervalent when the threshold for its activation is low and when it can be triggered by remote or trivial stimuli (Beck et al., 2004). For example, the hypervalent schema “The world (and others) are dangerous” would lead to biased information processing where a person interprets neutral situations as dangerous, feels anxious, and wants to get away.

Three basic schemata are thought to play a central role in BPD: “The world is (and others are) dangerous and malevolent”, “I am powerless and vulnerable”, and “I am inherently bad and unacceptable” (Arntz, 2004, Pretzer, 1990). Several authors have tested the hypothesis that BPD is characterized by these beliefs. For example, Dreessen and Arntz (1995) developed the Personality Disorder Belief Questionnaire (PDBQ) with beliefs from the Appendix of Beck and Freeman (1990) and additional BPD specific beliefs. A set of six beliefs appeared typical for BPD. These beliefs were characterized by loneliness, unloveability, rejection and abandonment by others, viewing the self as bad, and feeling they need to be punished; themes which all correspond with Pretzer's three basic schemas for BPD. Recently, Butler, Brown, Beck, and Grisham (2002) developed a BPD beliefs scale including 14 items (e.g., “I cannot trust other people”, “People will take advantage at me if I give them the chance”, “A person whom I am close to could be disloyal or unfaithful”) that discriminate between BPD patients and persons with other PDs. Using the World Assumption Scale, Giesen-Bloo and Arntz (2005) provided additional evidence for the three dominant beliefs in individuals with BPD. They suggested that these beliefs were associated with the severity of the borderline psychopathology.

Basic schemata in BPD have also been investigated with projective approaches such as the Thematic Apperception Test (TAT), while others have relied on narrative data to assess evaluation bias in BPD patients (e.g., Nigg et al., 1992, Segal et al., 1993, Stuart et al., 1990, Westen et al., 1990, Westen et al., 1990). Some studies also asked patients with BPD to evaluate their current relationships, e.g., with relatives (Benjamin and Wonderlich, 1994, Stern et al., 1997). Other authors used standardized stimuli, but non-standardized, free response formats. For example, Arntz and Veen (2001) presented film characters in 10-minute film clips with emotional themes such as abandonment, rejection, and abuse and asked borderline and control participants for their spontaneous evaluations. Wagner and Linehan (1999) studied facial expression recognition in BPD with standardized slides of emotional faces and free responses. In general, these studies demonstrated that individuals with BPD evaluated others as being more malevolent and hostile relative to individuals with other and no mental disorders, results that are consistent with the hypotheses of the cognitive model on BPD. However, as the stimuli and/or response formats of studies were not standardized, these studies leave much room for confounding factors. On the other hand, standardized stimuli may be rather artificial. The use of schematic faces or pictures as stimuli neglects the fact that in everyday life people process a wide range of stimuli, including facial expression, behavior, and social context. Indeed, a fleeting glimpse or mere glance can lead to an instantaneous evaluative judgment.

The present study therefore assessed evaluation bias with the “thin-slice judgments” paradigm, in which participants have to evaluate brief excerpts of behavior (Ambady, Shih, Kim, & Pittinsky, 2001): stimuli which are both standardized and ecologically valid. In our study we used silent film clips of about 10 s, which showed a person entering a room and taking a seat. Previous work has demonstrated that the thin slices method provides reliable information about a range of psychological constructs, including dispositional characteristics and social relations (for a review Ambady, Bernieri, & Richeson, 2000), possibly because the judgments are based on relatively automatic processes (Ambady and Rosenthal, 1992, Bargh, 1994, LeDoux, 1996).

Furthermore, most of the aforementioned studies did not control their results for comorbid psychopathology. Negativistic information processing biases are thought to be particularly characteristic of depression (e.g., Beck, 1976, Bower, 1981), an assumption supported by several empirical studies. For example, Gotlib et al. applied an emotion face dot-probe, emotional Stroop and recall task, and demonstrated disorder- and content-specific negative attentional and recall biases in depressed individuals (e.g., Gotlib et al., 2004, Gotlib et al., 2004). Considering the high prevalence of major depressive disorder (MDD) in BPD (e.g., lifetime diagnosis of 90%; e.g., Barnow et al., 2007), controlling for the presence of depression seems especially important in studies on evaluation bias in BPD. Studies on evaluations in BPD with non-standardized stimuli that did control for depression suggested that negative evaluation biases are even stronger in patients with BPD than in depressive (non-BPD) patients (e.g., Baker et al., 1992, Benjamin and Wonderlich, 1994, Segal et al., 1993, Stern et al., 1997, Stuart et al., 1990, Westen et al., 1990, Westen et al., 1990). One explanation for this finding would be that BPD is characterized by histories of childhood traumata such as sexual abuse, physical abuse, and neglect (e.g., Barnow et al., 2005, Zanarini, 2000, Zanarini and Frankenburg, 1997) which have been linked to malevolent views of others (e.g., Arntz et al., 1999, Beck et al., 2004). Thus, besides applying a standardized and ecologically valid assessment of evaluation bias, this study also included a (non-BPD) depressive group to control for the effect of depression on interpersonal evaluation bias in BPD.

A final limitation of the aforementioned studies is that none of these studies included an examination of interpersonal behavior of individuals with BPD, although the cognitive theory suggests a close link between evaluative processes and interpersonal behavior (Beck et al., 2004). Interpersonal problems are one of the dominant characteristics of BPD (American Psychiatric Association, 2000) and the effect of these difficulties can be profound. However, despite the centrality of interpersonal problems to clinical descriptions of BPD pathology, there is a lack of empirical research investigating the interpersonal behavior of individuals with BPD. In clinical settings, individuals with BPD often report conflicted relationships, intense outbursts of interpersonal hostility, over-control of anger, quarrelsomeness, and submissiveness (Gunderson, 1996, Gunderson, 2001, Horowitz, 2004, Linehan et al., 1994). A recent study investigating affective experience and interpersonal behavior of individuals with BPD (Russel, Moskowitz, Zuroff, Sookman, & Paris, 2007) found elevated levels of submissive and quarrelsome behavior, reduced levels of dominant behavior, and overall more extreme behavior relative to non-clinical control participants. We therefore decided to consider interpersonal behavior in our study.

In sum, this study tested three hypotheses. Firstly, we hypothesized that BPD is characterized by a negativistic interpersonal evaluation bias. We expected that individuals with BPD would view others as having less positive, more negative, and more aggressive traits than non-BPD participants with current depression and healthy control subjects. We also expected a smaller contrast with the depressive control group than with the healthy control group, as depressive individuals are characterized by negativistic biases as well. Secondly, we assumed that an aggressivistic interpersonal evaluation bias would characterize BPD in particular. We expected more negative evaluations for aggressive traits than for non-aggressive negative and positive traits in BPD group, but not in the control groups. This hypothesis is based on the association between BPD, childhood trauma and malevolent views of others. Thirdly, we hypothesized that BPD is characterized by more extreme interpersonal behavior. More specifically, we expected that the BPD group would show lower degrees of affiliation and dominance and higher degrees of hostility, social avoidance and submissiveness than the control groups.