Self and Informant Report Across the Borderline Personality Disorder Spectrum

Abstract

Individuals with Borderline Personality Disorder (BPD) features may be unaware of or unwilling to report their own personality or maladaptive behaviors, which complicates the assessment of BPD. Informants who know the individuals with BPD features may be uniquely suited to aid in the personality assessment of these individuals. The present study analyzed the comparative ability of individuals (targets) and informants to report BPD features across the continuum of BPD severity. The sample consisted of 1387 targets, ages 55 to 65 (56% women), who were recruited for an epidemiological longitudinal study examining the effects of PDs on health and social functioning. Each target nominated an informant who provided information about the target’s personality. Results indicated relatively low levels of agreement between perspectives and that informants reported BPD symptoms with more precision and at lower levels of BPD severity than targets. The benefits of including an informant perspective when measuring the BPD continuum are discussed; these benefits may include gains in reliability and improvement in the prediction of outcomes.

The DSM-5 (American Psychiatric Association, 2013) lists a series of behaviors and cognitions that define BPD. The nine core diagnostic criteria are: 1) Marked efforts to avoid real or imagined abandonment; 2) Unstable, intense relationships; 3) Persistent and significant self-image instability or unstable self-regard; 4) Potentially damaging impulsivity; 5) Suicidal behavior and self-harm; 6) Affective instability; 7) Chronic emptiness; 8) Intense anger or anger dyscontrol; and 9) Occasional paranoid ideation or dissociation due to stress. These criteria reflect multiple personality domains, including identity, cognitions, and affect. We describe how dysfunction in these areas might impair a person’s ability to report on his/her own personality, and suggest reasons why informant reports might be used to improve and complement self-reported personality assessments.

Identity.

Individuals with BPD are characterized by disturbances in their identity (per DSM-5; American Psychiatric Association, 2013). These disturbances come in several forms. One form of identity disturbance is a lack of self-knowledge. Individuals with BPD often appear to have a poor sense of themselves (Linehan, 1993) and a poor understanding of their own internal states (Bateman & Fonagy, 2003; Ebner-Priemer et al., 2008). Indeed, they show a poor awareness of a variety of factors that contribute to self-concept, such as poor awareness of personal goals and values. They also can have a difficult time predicting their own behavior (Dammann et al., 2011) and demonstrate less understanding of how they are perceived by others (Carlson & Oltmanns, 2015).

Lack of self-knowledge has implications for identifying diagnostic features of BPD via self-report. Deficits in self-knowledge impose an upper limit on how much information can be collected through self-report; individuals cannot provide information that they cannot access. Individuals with BPD who are unaware of their own impulsivity cannot report on impulsive personality features, and individuals with BPD who are unaware of inconsistency in their beliefs about others cannot report precisely on their shifting interpersonal cognitions. Because of their limited access to accurate self-knowledge, individuals with BPD may tend to provide shallow, superficial descriptions of themselves (Clarkin, Yeomans, & Kernberg, 2007). Informants (those who know individuals with BPD) may be in an advantaged position to report unique, and possibly more accurate, information about a target person’s BPD features. They may be able to gather observations and place them in a cause and effect narrative in a manner that the target cannot.

In addition to lack of self-knowledge, individuals on the BPD spectrum may have unstable identities. These also can act as a barrier to valid self-report. When an individual’s “self” is unstable (Clarkin et al., 2007), the individual’s identity fluctuations may influence how they interpret and therefore how they report their identity. Their current self-perspective becomes a distortion lens through which they perceive their behaviors and experiences (Green & Sedikides, 2001). While informants do not have direct access to the target’s fluctuating internal or self-perceptions, they can make inferences from a stable vantage point. Thus, informants may be uniquely situated to offer a clearer, more reliable perspective (Balsis, Cooper, & Oltmanns, 2015).

Cognition.

Individuals with BPD features often have distorted cognitions. Specifically, they often make simplified judgments about people and situations. This type of thought pattern is called “splitting” (Kernberg, 1967), which is the use of a dichotomized (“all or nothing”) judgment. Individuals with BPD features often show splitting in their perceptions of others by categorizing others as being all good or all bad, though this belief can shift rapidly (Siegel, 2006). This tendency to encode experiences into different schemas may affect the way individuals with BPD perceive people and situations. In this conceptualization, experiences or features that are subthreshold in intensity are likely to be minimized, dismissed, or simply go unnoticed. Conversely, experiences that meet threshold are likely to be seen as very severe. This simplification of self-perception may lead individuals on the BPD spectrum to have difficulty in producing nuanced information about themselves.

These simplified cognitions and schemas may impair a person’s ability to provide accurate information via self-report. PD features may be particularly likely to be overlooked or distorted (Kernberg, 1984). For example, when asked whether one has outbursts of intense anger, an individual with BPD may overlook certain angry outbursts, viewing them as insufficiently intense to warrant endorsing “intense outbursts of anger.” Anger is often egosyntonic (Howells, 1998) and this may skew perceptions of intensity. “Reasonable” anger, for example anger at a perceived insult, may not be viewed as an “intense outburst”, but rather as an appropriate response to the situation. In this way, egosyntonic features may be overlooked at lower levels of intensity.

Informants may be able to produce more nuanced reports about targets. This is because informants may have perceptions that allow them to identify positive, negative, and mixed features in the target. As a result, informants may be particularly well suited to identifying lower intensity BPD features. At the very least, informants may provide an alternative perspective that can be contrasted against the interpretations of the target. Agreement in self and informant reports could be used to gather information about shared perspectives that are not directly influenced by the cognitive distortions specific to the disorder.

Affect.

BPD is exemplified by distortions in emotional states (Linehan, 1993). One such disruption is a heightened affective intensity. The intense emotions experienced by individuals with BPD features may interfere with their ability to accurately report on their thoughts and situations. For example, if an individual is experiencing extreme feelings of anger, they may find it difficult to report on emotional states that are less prominent than their current feelings, such as feelings of sorrow or anxiety. Informants, comparatively, may be less influenced by emotions in this way because they are more relatively likely to resemble the general population and thus should be less prone to intense affect compared to individuals with BPD features (Santangelo, Bohus, & Ebner-Priemer, 2014).

In addition to increased affective intensity, individuals on the BPD spectrum have affective instability (Koeningsber et al., 2002; Trull et al., 2008), which can interfere with self-report a number of ways. First, affective instability makes any distortion due to extreme emotion unpredictable over even relatively brief periods. Indeed, Thomas (1996) found that self-report measures of personality and psychopathology often significantly varied alongside fluctuating moods in BPD patients. Informants may act as comparatively stable landmarks in the context of assessment because informants are less likely to experience the same degree of emotional turbulence as individuals with BPD (Santangelo et al., 2014).

The reactive component of affective instability in BPD may further complicate assessment via self-report for BPD individuals because these individuals can have significant affective reactions to item content (Sansone & Sansone, 2010). Individuals with BPD related affective instability may be distressed if assessment involves unpleasant cognitions or emotions (Cheavens et al., 2005; Gratz et al., 2006; Rosenthal et al., 2005) such as those associated with stigmatized BPD personality features or unpleasant memories (e.g. trauma, relationship issues). This sensitivity to item content may then skew perceptions or motives during self-report. Informants may not share the same reactivity in affect as BPD individuals for a variety of reasons. One, compared to BPD individuals, informants may not be as reactive to item content in general (Sansone & Sansone, 2010). Two, informants are not being asked to report on their own negative features, thus the specific item content is at least one step removed from personal experience. Even informants with close relationships to targets appear to show less of a tendency to enhance or minimize target personality features compared to the self-enhancement and self-diminishment effects shown by targets (John & Robins, 1993).

Agreement.

There are many challenges to the assessment of BPD features. Perhaps because of the reasons we have outlined here, self- and informant report, have shown positive yet very low agreement in the literature (Carlson, Vazire, & Oltmanns, 2013; Clifton, Turkheimer, & Oltmanns, 2004, 2005; Klonsky, Oltmanns, & Turkheimer, 2002; Lawton, Shields, & Oltmanns, 2011; Miller, Pilkonis, & Clifton, 2005; Samuel & Widiger, 2010; Hyler, Rieder, Williams, Spitzer, Lyons, & Hendler, 1989). Indeed there are even examples in which different conclusions about BPD concepts, such as the prevalence across gender or change over time, can be drawn based on the perspective one considers (Busch, Balsis, Morey, & Oltmanns, 2016; Cooper, Balsis, & Oltmanns, 2014). Thus far, it is unclear exactly what factors contribute to the differences across perspective.

One potential explanation for the discrepancies between perspectives is that each has different advantages and limitations. Self-report questionnaires provide a unique perspective that includes the individual’s thoughts, intentions, feelings, and motivations (Mount, Barrick, & Strauss, 1994). However, those with PDs may have an especially difficult time observing the ways in which their maladaptive personality features affect those around them (John & Robbins, 1994; Oltmanns, Turkheimer, & Strauss, 1998). This limited insight may cause individuals to have trouble reporting accurately on their personality features. There is also the potential for individuals to be motivated to present themselves in an overly positive or negative light. Conversely, informants may be more likely to provide a less biased view of the individual. Informants may also be in an excellent position to report accurately on how the target’s behaviors affect those around them. However, informant reports are not without their own limitations. Just as selves can lack insight, informants may not know the individual well enough (in general and/or in certain contexts) to accurately report on certain symptoms. The different advantages and disadvantages offered by each perspective may contribute to the low agreement.

Recent research has shown, however, that both self-reports and informant reports provide a unique and at least partially valid perspective for measuring BPD (Vazire & Mehl, 2008). Each perspective contains a certain degree of validity, and as a result one might expect to find higher agreement than we observe in the literature. Given that each perspective is at least somewhat valid, the low agreement between perspectives remains to be explained. In the current study, we will analyze the low agreement, and compare the ability of each perspective, self and informant to indicate the same latent dimension. So even when the two perspectives are indicating the same exact construct, do they indicate that construct in an equivalent way? Are they perhaps indicating different levels of the same construct, for example? Perhaps informants are sensitive to the BPD continuum in a way that selves are not. If so, then this result may help to explain a certain amount of the limited agreement.

To explore this possibility, it may be beneficial to consider the latent BPD continuum within a framework consistent with dimensional constructs. Recall that in the DSM, each of the BPD features serves as an indicator of the target’s overall BPD intensity. An approach that can empirically examine and quantify the variations in the degree to which each of the features indicates BPD severity across self- and informant reports could reveal key aspects of their disagreement. Multidimensional Item Response Theory (MIRT; Reckase, 2009) in particular provides a method of examining the ability of an informant and a self to indicate the BPD continuum, while controlling for influences unique to each perspective. That is, MIRT offers insight into how specific criteria differentially contribute to information regarding the general latent BPD continuum, while capturing the unique contributions of the self and informant report separately. This approach clarifies each item’s ability to indicate a relative portion of the BPD latent continuum, in this case BPD intensity, and explores each item’s strength of relationship to that general continuum and the specific unique subfactors as depicted in Figure 1. With this model, we are able to capture the ability of each item to indicate the latent BPD continuum when used within a certain perspective (informant vs. self-report) and make valid comparisons across the perspectives.

Figure 1.

Structural equation bifactor measurement model of self- and informant reported DSM-5 BPD criteria fits the data well

Current Study.

The assessment of BPD typically relies primarily on self-report, which may be limited particularly for identifying this construct. Informants have shown promise as a way to bypass measurement challenges that often accompany BPD, such as deficits in an individual’s self-knowledge and motivational barriers to forthright disclosure. Thus, we hypothesize that self-reported and informant reported BPD features share at best modest levels of agreement. This hypothesis is consistent with the notion that self-and informant reports on PDs provide somewhat non-overlapping information, and that informant data adds incremental validity to information acquired via self-report in the assessment of various personality disorders (Oltmanns & Turkheimer, 2003; South, Oltmanns, & Turkheimer, 2003). We also will begin to analyze what underlies the disagreement between selves and informants. We will address this issue within a MIRT framework, and analyze the degree to which informants and selves provide information about the BPD continuum. Specifically, we will determine whether informants provide more information about the construct than selves, and whether informants can provide information about the construct even in its relatively subtle form (at relatively low levels of BPD intensity).

Method

Participants and Recruitment.

Participants were recruited from the city of St. Louis and its surrounding suburbs. Participants that met recruitment criteria were offered $60 to complete a 3 hour assessment. St. Louis’ ethnic and racial diversity allowed for a wide range of demographic representation within the participant pool. Within the recruitment region, 30% of the population is African American and 60% is Caucasian, though only 2% of this population is Hispanic. It should be noted here that this research was performed as part of the St. Louis Personality and Aging Network (SPAN) study (please see Oltmanns, Rodrigues, Weinstein, & Gleason, 2014, for a detailed description). The study received IRB approval.

Participants were asked to identify individuals who knew them well and were capable of providing accurate descriptions of their personality traits, preferably informants that cohabited with the participants. When cohabitants were unavailable, participants were asked to identify “the person who knows you best.” Additionally, eligibility of informants for inclusion required that, historically, the informant and the participant talked at least monthly and interacted face-to-face at least yearly. Within the recruited informant pool, informants had known participants for an average of approximately 30 years. Spouses or romantic partners composed about half of the informant pool. The remainder of the informant pool was composed of the target’s other family members (e.g., an adult child of the participant) or close friends.

The final sample used in the current analyses consisted of 1387 participants (55% female) living in the St. Louis area. Participants were between the ages of 55 and 64 when they first entered the study (M = 59.5, SD = 2.7 years). The participants consisted of 68% Caucasian, 30% African American, and 2% other ethnicities (e.g., American Indian). Twenty-five participants self-identified as Hispanic or Latino, under the predicted 2% of the sample.

Materials.

Target participants and informants completed the Multisource Assessment of Personality Pathology (MAPP; Oltmanns & Turkheimer, 2006). This questionnaire includes 80 items, based on 79 diagnostic features of the 10 PD types from the DSM-IV (one item regarding narcissistic PD was divided into two items regarding jealousy of self and jealousy of others). Each criterion was translated into layman’s terminology. The MAPP includes a self-report version that asks individuals to rate the degree to which they endorse the relevant PD traits as well as an informant version in which a close other rated the target person’s personality features. The order of the items is randomized and ratings are made on a five point Likert scale ranging from “0,” which indicates that the participant is “never like this,” to “4,” which indicates that the participant is “always like this.”

Procedure and analyses.

Please see Oltmanns, Rodrigues, Weinstein, & Gleason, 2014, for a detailed description of the entire protocol. The present study compared self- and informant report in the endorsement of BPD features across degrees of disorder severity, making use of the demographic information and data from the MAPP.

The specific data analytic challenge at hand is how to use an IRT approach to compare parameters derived from this type of within-subjects variable (self vs. informant), that contains multiple dimensions of data, including two dimensions of measurement artifact. The solution we found was to use a MIRT model, which ensures that the parameters of the general dimension are validly comparable across groups (see Figure 1). Some might recognize this model as a bifactor SEM model. MIRT analyses followed a 2 parameter logistic multidimensional model as described in Reckase (2009). All analyses were conducted with flexMIRT software (Cai, 2013). The analyses produce parameters: Discrimination parameters (a) are provided for each dimension and threshold (b) parameters are calculated via procedures provided in the user manual. As part of our inspection of the results, we plotted item functions across the general BPD dimension, holding the intercepts constant and statistically controlling for the self-and informant-reported measurement artifact factors.

Results

Self-informant reliability, comparing self and informant composites of all BPD features was modest at best. As reported elsewhere, the Pearson correlation between self and informant MAPP scores for BPD was .26 (Oltmanns et al., 2014), explaining 7% of the variance (93% unexplained). At the level of individual diagnostic features, agreement was even lower. Item-level kappas computed dichotomously (feature not present, score of 0, 1, or 2 vs. feature present, score of 3 or 4) ranged from 0.07 to 0.16 (Table 1), displaying poor though statistically significant (p < 0.05) levels of agreement.

Table 1.

Item parameters across the self- and Informant-reported BPD continuum

	Self			Informant
	BPD	Artifact		BPD	Artifact
DSM-5 BPD criteria as defined by MAPP items	a	a	b	a	a	b	Kappa
1) I will do almost anything to keep those that I love from leaving me	0.10	1.05	0.69	0.39	0.43	0.44	0.08*
2) In close relationships (with friends and family members), I often switch back and forth between loving a person and hating him or her	0.94	1.64	1.89	1.34	2.71	1.36	0.09*
3) Compared to others, my opinions and preferences change more frequently	0.58	1.06	1.19	0.73	1.28	1.37	0.07*
4) I am impulsive and have done things that could be dangerous to me	0.55	1.41	1.55	1.13	1.64	1.33	0.09*
5) I have threatened to hurt, or kill myself	1.66	1.41	2.29	2.57	2.10	2.05	0.16*
6) I have strong mood swings in response to events; I have frequent periods of intense sadness, irritation or anxiety	2.01	1.64	1.31	1.80	1.84	0.83	0.15*
7) I feel emotionally unfulfilled or that life is meaningless	1.78	0.64	1.54	1.60	1.26	1.13	0.15*
8) I have sudden, intense outbursts of anger	1.33	2.28	1.75	1.31	2.18	1.24	0.10*
9) When I am under stress, I may become paranoid or suspicious of people I usually trust, or have other strange experiences that are hard to explain	1.13	2.18	1.65	1.30	1.72	1.23	0.10*

^*p < 0.01

The strength of agreement between self and informant can be observed by the strength of the Kappas: <0.00 = Poor; 0.00–0.20 = Slight; 0.21–0.40 = Fair; 0.41–0.60 = Moderate; 0.61–0.80 = substantial and 0.81–1.00 = almost perfect (Cyr & Francis, 1992). MAPP = Multisource Assessment of Personality Pathology.

Given the relatively low agreement across perspective, we investigated whether that agreement could be due at least in part to the differential ability of each perspective to identify the latent BPD continuum. We approached these analyses within an MIRT framework, which depends on the bifactor model fitting the data. To evaluate fit, we performed a confirmatory factor analysis using the comparative fit index (CFI), the Tucker-Lewis index (TLI) and the root mean squared error of approximation (RMSEA). CFI/TLI values ≥ .95 and RMSEA values ≤ .06 suggest good model fit (Hu & Bender, 1999). Results of these analyses suggested sufficient fit for MIRT analyses. CFI (.99) and TLI (.99) were both greater than .95 and RMSEA (.02) was less than .06. After confirming the fit, we performed an MIRT analysis using flexMIRT software (Cai, 2013). Across all nine items, the within variable t-test was statically significant. The informant-reports had lower b values on average (M = 1.22, SD = .43) than the self-reports (M = 1.54, SD = .45), t(8) = 4.32, p = .003 (see Table 1). Notably, eight of the nine items had higher b parameters for the informants relative to the selves. Additionally, the items tended to be more strongly related to the general dimension for informants than selves, though with just nine items we only were able to show this as a strong statistical trend. Specifically, the informant-reports tended to have higher a₁ parameters on average (M = 1.12, SD =.64 than the self-reports (M = 1.35, SD = .62), t(8) = −1.92, p = .09. See Figure 2 for a representation of the relationship between the items and the main latent dimension. For those interested, Table 1 contains the parameter estimates from all dimensions, including the unique self and informant artifact dimensions.

Figure 2.

Item functions for self- and informant reported BPD diagnostic criteria

Finally, informants were more likely than targets to endorse items. This relative frequency was greater at higher levels of the scale, as reflected by the trend line in Figure 3. In other words, there was a trend in the target/informant ratio with informants increasingly likely to endorse more features than targets at higher levels of the scale.

Figure 3.

Ratios (informant/self) of raw score endorsement frequencies across the number of BPD criteria endorsed

Discussion

Our data were collected using a large community sample. As expected among older adults, relatively few individuals in this sample passed the DSM-5 threshold for a diagnosis of BPD. Nevertheless, many of the participants exhibited one or multiple features of the disorder. In fact, 27 participants were rated as showing 3 or more symptoms of BPD on the basis of a semi-structured diagnostic interview (Oltmanns et al., 2014). Perhaps most important, dimensional scores for BPD symptoms have been shown to be significantly related to impaired health and social adjustment in this sample (Gleason et al., 2014). We know that borderline personality pathology is meaningfully represented in this sample of older adults, and the MAPP is a valid measure to identify symptoms of this disorder.

The analyses reported in this paper were designed to evaluate the relative utility of self and informant reports for symptoms of BPD. As we explained in the introduction, the presence of features of BPD may cause a person to be less able or less willing to report on their own personality characteristics. Thus, we expected that informants would provide additional information about the targets’ personality—information that targets might not be able or willing to provide themselves. Indeed, our analyses showed that informants and targets provided different information about targets’ BPD features, as evidenced by relatively low BPD item level kappa values.

MIRT analyses shed further light on these differences, revealing that self- and informant reported BPD were sensitive to different parts of the underlying BPD continuum. Even at low levels of BPD, informants were more likely to endorse eight of the nine BPD criteria. One feature did not fit that general trend: Identity Problems. We can only speculate as to why targets more readily reported identity problems. It is possible that identity problems are less observable by informants. BPD related identity instability can cause an individual to change their personal values and goals to match the individuals around them (Clarkin, Yeomans, & Kernberg, 2007). Informants may be more likely to see the individual expressing similar values, beliefs, and goals to their own across multiple interactions. This is possibly because BPD related identity instability leads the target to mirror the values, beliefs, and goals of the informant. Thus, it may be difficult for any single informant to fully observe the fluctuations in identity.

It is also possible that the wording of the MAPP item assessing Identity Problems (“Compared to others, my opinions and preferences change more frequently”) does not accurately describe or capture the sort of identity disturbances that are presumably associated with BPD. Other items in the MAPP may function more effectively for the related BPD features. Instead, this item may reflect relatively normal fluctuations in opinions and preferences more than it measures the general instability and lack of development of opinions and preferences that are commonly displayed by individuals with BPD.

We also found that compared to targets, informants tended to provide ratings that were more strongly related to the common BPD dimension. This result may have occurred because informants are motivated to report on certain features of the target or because informants have a better awareness and understanding of the target’s BPD features. Indeed, both explanations may be possible and may stem in part from features of BPD itself. Problems with trust and self-knowledge may hamper the target’s ability to self-report. In addition, affective instability, distorted cognitions, and impulsivity may reduce reporting in BPD individuals. At the same time, informants may be unaffected by these dysfunctional personality features and may be better able to report on the target’s personality. Informants also may be more willing to do so, in part because reporting on certain personality features does not directly affect them, and they may even find reporting on these features to be cathartic, depending on how disruptive, annoying, or harmful the target’s BPD personality features have been in their life. Future studies will need to pinpoint the particular factors responsible for the difference in overall reporting of BPD features by informants and targets.

There are at least three potential limitations to the analyses reported in this paper. First, we employed data from only one measure for borderline personality pathology, the MAPP. Additionally, the use of an epidemiological sample provides a great degree of utility, such as the ability to assess a wider range of BPD intensity that permitted us to analyze lower levels of BPD severity. However, the use of a clinical sample could provide meaningful information regarding the more severe ranges of BPD symptomatology and therefore should be considered for future research to complement these results. Finally, the utility of informant report in our approach was based on an assumption that informant reports are at least somewhat valid. This assumption is based on a well-established body of literature on the implementation and usefulness of informant reports in personality research (e.g. Connelly & Ones, 2010; Vazire & Mehl, 2008). It is important to note, however, that informant reports are subject to their own biases and limitations that affect their validity. Informants provide a unique and useful perspective, but they should not be considered to be more privileged with respect to truth than self-reports or clinician ratings (Galione & Oltmanns, 2014).

As research continues on both self- and informant report it will be important to consider how to optimally integrate them. Assessments of personality pathology constructs, such as the BPD continuum, may require the use of both perspectives to reliably and precisely determine an individual’s standing along the latent variable. How these two perspectives can increment and complement each other is an open empirical question. Here, we’ve endeavored to take a serious first step analyzing the information provided by each perspective and show that informants provide more total information about the BPD continuum, particularly at relatively low levels along the BPD continuum. Looking forward, a next step will be to examine how self- and informant report in combination can best predict meaningful real life outcomes and outcomes of clinical significance. Additionally, it is important to acknowledge that any factor or a test, including its scoring algorithm, scale anchors, or time constraints may affect the results.

Beyond these practical questions, many other theoretical questions are raised. BPD by definition is a disorder characterized in large part by a severely distorted view of self and other. How can we rely on these individuals to accurately report on their histories, their personalities, or other people in their lives? At the same time, to the extent that the arguments raised in our introduction are true, how might one explain the many useful and informative self-report studies? The current results suggest that selves are able to report on their personality to a certain extent, and perhaps they are well-positioned to report on unique aspects of their personality dysfunctions. At the same time, informants might provide a complementary and unique perspective as well. Researchers revisiting previous findings or theories or clinicians looking for more comprehensive and nuanced assessment of their clients might find value in both perspectives, which may lead to a deeper understanding of BPD pathology.

Lots of work remains, including understanding the influence of relationship factors on information reported by informants. Additionally, future work also may focus on other assessment approaches to BPD, including concepts introduced in section III of DSM-5. Perhaps the use of MIRT and the complementary bifactor SEM model will spur additional research and understanding in this area, as far as we know this is the first time these types of models have been used to ask these types of questions across self- and informant report.