Personality Disorder Models and their Coverage of Interpersonal Problems

Abstract

Interpersonal dysfunction is a defining feature of personality disorders (PDs) and can serve as a criterion for comparing PD models. In this study, the interpersonal coverage of four competing PD models was examined using a sample of 628 current or recent psychiatric patients who completed the NEO Personality Inventory-3 First Half (NEO-PI-3FH; McCrae & Costa, 2007), Personality Inventory for the DSM-5 (PID-5; Krueger et al., 2012), Computerized Adaptive Test of Personality Disorder-Static Form (CAT-PD-SF; Simms et al., 2011), and Structured Clinical Interview for DSM-IV Personality Questionnaire (SCID-II PQ; First, Spitzer, Gibbon, & Williams, 1995). Participants also completed the Inventory of Interpersonal Problems-Short Circumplex (IIP-SC; Soldz, Budman, Demby, & Merry, 1995) to assess interpersonal dysfunction. Analyses compared the severity and style of interpersonal problems that characterize PD models. Previous research with DSM-5 Section II and III models was generally replicated. Extraversion and Agreeableness facets related to the most well defined interpersonal problems across normal-range and pathological traits. Pathological trait models provided more coverage of dominance problems, whereas normal-range traits covered nonassertiveness better. These results suggest that more work may be needed to reconcile descriptions of personality pathology at the level of specific constructs.

Personality Disorder Models and their Coverage of Interpersonal Problems

Personality disorders (PDs) frequently are described in terms of interpersonal dysfunction (Hopwood, Ansell, Wright, & Pincus, 2013; American Psychiatric Association [APA], 2013) that manifests quite variably (e.g., Pincus & Wiggins, 1990; Monsen, Hagtvet, Havik, & Eilertsen, 2006). For instance, Schizoid PD is defined by “detachment from social relationships,” (p. 653) whereas Histrionic PD involves “excessive emotionality and attention seeking” (APA, 2013, p. 667). Thus, in evaluating and comparing competing PD classification models, an important consideration is how each accounts for such diverse expressions of interpersonal dysfunction. Considerable research has examined Diagnostic and Statistical Manual of Mental Disorders (DSM) PDs in this regard (e.g., Soldz, Budman, Demby, & Merry, 1993); however, less is known about contemporary PD models (e.g., Widiger, 2011; Simms et al., 2011). The present study fills this gap in the literature by comparing the interpersonal impairments associated with several prominent PD models.

Models of Personality Disorder

PD models can be broadly differentiated by how they organize phenotypic variability in PD presentation. The official DSM-5 PD classification chapter (i.e., Section II; APA, 2013) organizes this variability via 10 categories supported primarily by clinical theory and expertise (e.g., Kendler, 2009). Despite this system’s endurance, it has been widely criticized as being inconsistent with research on the structure of personality pathology (e.g., Widiger & Samuel, 2005; Skodol, 2012). Moreover, these traditional categorical PDs have limited validity and usefulness due to their excessive co-occurrence, within-diagnosis heterogeneity, arbitrary cut points, and inability to describe all PD patients (e.g., Clark, 2007; Skodol, 2012).

A growing number of researchers advocate reorganizing the phenotypic variability in PD with empirically derived trait models (e.g., Krueger, Markon, Deringer, Watson, & Skodol, 2012), which consist of distinct and homogenous dimensions (Clark, 2007). Despite attempts to implement such a system (e.g., Krueger et al., 2012), disagreement remains concerning issues such as trait uni- vs bipolarity (Samuel, 2011; Krueger et al., 2011), the nature and number of lower-order PD traits (Simms et al., 2011), and whether these traits should focus on normal- or abnormal-range personality features (Widiger, 2011). In part due to these disagreements, the DSM-5 Personality and PD Work Group’s proposed trait model was placed in Section III of DSM-5 (“Emerging Measures and Models”; APA, 2013) to spur further research. As described in the first paragraph, interpersonal dysfunction is a central feature of PD; the convergence and divergence of PD models in their relations to such dysfunction can thus inform ongoing debates. A comprehensive interpersonal dysfunction model capable of making subtle differentiations in how such dysfunction manifests would facilitate this process.

Conceptualizing Interpersonal Dysfunction

One model with an internal structure and content domain appropriate for comparing PD models is the Interpersonal Circle (IPC; Leary, 1957). The IPC is a circular model defined by two orthogonal axes, which can be broadly described as agency and communion. Agency is a dimension with themes of power, mastery, and assertion on one end vs. weakness, failure, and submissiveness on the other (e.g., dominance vs. submissiveness). Communion contrasts themes of intimacy, union, and solidarity with remoteness, hostility, and disaffiliation (e.g., warmth vs. coldness; Pincus, Lukowitsky, & Wright, 2010). All points within the IPC reflect blends of these two dimensions (e.g., cold-dominance), such that external indicators (e.g., PD measures) related to agency and communion can be described precisely by their location in this space. Relations to interpersonal dysfunction may be described by their specificity (i.e., how clearly a construct relates to one region of IPC), style (i.e., which region it relates to), and severity (i.e., degree of relation to non-specific interpersonal distress; Gurtman, 1992). Describing components of a PD model with such nuance allows the IPC to provide a useful basis for comparing the amount and type of interpersonal dysfunction associated with different models for classifying and assessing personality pathology.

Interpersonal Problems and Personality Pathology

Personality pathology relates to general interpersonal distress (i.e., severity; Wright et al., 2012); however, both PDs and traits vary in the degree to which they can be characterized by a specific class of interpersonal problems (i.e., style; Soldz et al., 1993). Some expressions of personality pathology cannot be described well by one style of interpersonal problems. For instance, borderline individuals exhibit quite specific interpersonal problems at times, but as a group are heterogeneous and experience great variability in problems across time (Wright, Hallquist, Beeney, & Pilkonis, 2013; Wright, Hallquist, Morse, et al., 2013). In contrast, other forms of personality pathology relate consistently to very specific classes of interpersonal problems. Histrionic PD, for example, is related to warm-dominant interpersonal problems (i.e., being intrusive) and Avoidant PD relates to cold-submissive problems (i.e., social avoidance; Soldz et al, 1993). Consistent with this research, the traits of withdrawal and anhedonia relate to cold-submissive problems, whereas attention-seeking relates to warm-dominant problems (Hopwood, Koonce, & Morey, 2009; Wright et al., 2012). Research has identified PD constructs in all quadrants of the IPC, but the degree to which specific PD models are represented in each quadrant appears to vary considerably.

Current PD models have few components associated with warm interpersonal problems. Only two PDs (Histrionic and Dependent) consistently relate to warm problems (Pincus & Wiggins, 1990; Soldz et al., 1993; Monsen et al., 2006) and pathological trait systems do not provide additional depth in coverage (Hopwood et al., 2009; Simonsen & Simonsen, 2009; Wright et al., 2012). In contrast, normal-range personality traits may cover warm problems at the expense of cold ones (e.g., Soldz et al., 1993), although these warm problems may be less severe (Hopwood et al., 2013). In particular, Five-Factor Model Agreeableness may relate to problems with dependency that are not covered by pathological trait models (e.g., Gore & Pincus, 2013).

Notably, the PD-IIP literature exhibits limitations that may obscure links between personality pathology and interpersonal problems. For instance, studies vary in whether the specificity and severity of a PD construct’s interpersonal dysfunction are considered. In addition, faceted personality trait measures have not been examined in patient samples with the IIP, leaving open questions regarding range-restriction and generalizability. Furthermore, normal-range lower-order traits (i.e., facets) relevant to recent PD model proposals (e.g., Widiger, 2011) have not been examined with the IIP in any sample. Thus, an important step forward would be to examine how currently prominent PD models relate to the IIP in a patient sample.

The Present Study

The DSM-5 Section II PDs (APA, 2013) and three currently prominent PD trait models (Krueger et al., 2012; Simms et al., 2011; Widiger, 2011) were compared through the relations of their most commonly used measures to the IIP in a psychiatric patient sample. The aim of this research was to examine the similarities and differences in how these PD measures relate to the IIP, particularly at the level of disorders and lower-order traits. This study marks the first such comparison within a single sample and the first exploration of facet-IIP relations for two measures (Simms et al., 2011; Widiger, 2011). In addition, the measures used in this study allowed differences between normal-range and pathological traits to be more fully understood. Finally, this study used data from a patient sample, which allowed past findings to be reexamined in a more ecologically valid context.

PD measures were characterized by the severity and specificity of their components’ relations to interpersonal problems, as well as by the predominate style of problems to which their components related. Following from this, several predictions were made. Our first hypothesis (H1) was that past findings relating traditional categorical PDs (i.e., DSM-5 Section II) and the emerging trait model (i.e., Section III) to interpersonal problems would be replicated. Second (H2), as Extraversion and Agreeableness are considered alternative interpretations of agency and communion (e.g., McCrae & Costa, 1989), we predicted that facets from these domains would relate to the IIP with the most specificity. Finally, at the facet-level, some differences among all trait measures were expected; however, major differences were hypothesized (H3) to revolve around range of measurement. Pathological traits were predicted to relate more strongly to problem severity (H3a) and cold interpersonal problems than were normal-range traits (H3b), which were expected to cover warm interpersonal problems most effectively of all models (H3c).

Method

Participants and Procedures

As part of a larger study (Simms et al., 2011), 695 current or recent psychiatric patients were recruited using flyers distributed across Western New York mental health clinics. Each participant attended a four-hour session, receiving $50 plus transportation reimbursement. Participants completed structured interviews and computerized self-report measures. After removing data influenced by inconsistent responding, non-random missingness, or questionable in-session behaviors (e.g., acute substance use), 628 participants remained. Due to questionnaire battery length and time constraints, some participants did not complete all measures, limiting the sample size for each analysis.

In the present study, participants who did not complete the IIP-SC and at least one personality measure were excluded from analyses. The remaining participants (N = 438) averaged 42 years of age, were 67% female, and identified primarily as Caucasian (69%) or African American (29%). In addition, 2% of the sample identified as American Indian or Alaskan Native and < 1% as Asian or Native Hawaiian or Other Pacific Islander. The majority of participants were either currently in treatment (21%) or had been in the past 1–2 years (70%). Based on structured clinical interviews, 32% of participants met criteria for Major Depressive Disorder, 27% for Bipolar I, 2% for Bipolar II, 11% for Dysthymia, 10% for Agoraphobia, 17% for Social Anxiety Disorder, 23% for Obsessive-Compulsive Disorder, 20% for Post-Traumatic Stress Disorder, 44% for Generalized Anxiety Disorder, 22% for Alcohol Use Disorder, 18% for Substance Use Disorder, and 3% for Schizophrenia. Relative to the 190 participants who were excluded, included participants were more likely to be female (67% vs. 56%; χ²[1, N = 627] = 7.25, p = .007), on average were 4.7 years younger (i.e., t[626] = 4.33, p < .001), and differed significantly in terms of race (χ²[4, N = 624] = 24.71, p < .001). Specifically, African American participants made up a smaller proportion of the included group (29%) relative to excluded group (47%; χ²[1, N = 624] = 19.16, p < .001). The included participants also reported significantly less personality pathology (t[599] = 3.09, p = .002), as indicated by the sum of the Structured Clinical Interview for DSM-IV Personality Questionnaire (SCID-II PQ); however, the difference between participants was small (i.e., d = .27) and thus unlikely to affect the generalizability of our findings.

Measures

Inventory of Interpersonal Problems-Short Circumplex (IIP-SC)

The IIP-SC (Soldz et al., 1995) is a 32-item measure of interpersonal behaviors that individuals perform excessively (e.g., “I argue with other people too much”) or find difficult to do (e.g., “It is hard for me to feel close to other people”). Items form 8 four-item scales (Domineering, Vindictive, Cold, Socially Avoidant, Nonassertive, Exploitable, Overly Nurturant, and Intrusive) that correspond to equally spaced octants of the IPC. Use in previous patient samples (Soldz et al., 1995) indicates adequate internal consistency and adherence to circumplex structure (Alden, Wiggins, & Pincus, 1990; Gurtman, 1992; Monsen et al., 2006). As described above, all participants (438; 100%) in this subsample completed the IIP-SC. The median alpha coefficient was .78 (Range = .73–.90)

Structured Clinical Interview for DSM-IV Personality Questionnaire (SCID-II PQ)

The SCID-II PQ is a self-report screening questionnaire used for the SCID-II interview (First, Spitzer, Gibbon, & Williams, 1995) for PDs that consists of 121 dichotomous (i.e., yes/no) items that are related to DSM PD criteria (e.g., APA, 2013). One limitation of this measure is that self-reported Antisocial PD (ASPD) symptoms cannot be scored, although conduct disorder can be. Past work supports the SCID-II PQ as a screening measure (e.g., Ulrich et al., 2008) for PDs and as an indicator of personality pathology on its own (e.g., Ryder, Costa, & Bagby, 2007). In the present sample 337 (77%) participants completed the SCID-II PQ. Internal consistency varied by disorder (Mdn = .72, Range = .51–.86)

Personality Inventory for DSM-5 (PID-5)

The PID-5 (Krueger et al., 2012) is a 220-item questionnaire with 25 lower-order scales, one for each of the DSM-5 Section III pathological personality traits. These lower-order facet scales can be organized into five domains (Negative Affectivity, Antagonism, Detachment, Disinhibition, and Psychoticism) that have been described as conceptually similar to the Personality Psychopathology-5 model (PSY-5; Watson, Stasik, Ro, & Clark, 2013). Each facet scale is measured by 4 to 14 items measured on a four-point scale of 0 (very false or often false) to 3 (very true or often true). In the present sample, 435 (99%) participants completed the PID-5. The median facet alpha coefficient was .87 (Range = .75–.96).

NEO Personality Inventory-3 First Half (NEO PI-3FH)

The NEO-PI-3 (McCrae et al., 2005) is a revision of the NEO-PI-R (Costa & McCrae, 1992) intended to improve readability and item-total correlations. Research examining problematic items yielded 37 replacement items, improving the measure’s psychometric properties without altering its structure (McCrae et al., 2005). The full measure consists of 240 statements that are rated on a five-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). These items form 30 facet scales, with six facets for each of the higher-order domains (Neuroticism, Extraversion, Openness, Agreeableness, and Conscientiousness). The NEO-PI-3FH consists of the first 120 items of the NEO-PI-R, with four items for each of the 30 facets. Despite being half the length of the original measure, relations to full-length scales are strong (Mdn = .91) and the overall-structure of the measure is preserved (McCrae & Costa, 2007). A total of 381 (87%) participants completed the NEO-PI-3FH. For facets, the median alpha was .63 (Range = .40–.80).

Computerized Adaptive Test of Personality Disorder-Static Form (CAT-PD-SF)

The CAT-PD-SF is a brief measure drawn from the full CAT-PD item pool. The CAT-PD item pool was generated as part of project to develop a comprehensive model and efficient measure of PD traits, guided by a PSY-5 framework (Simms et al., 2011). The project has resulted in a pool of 1,366 items that form scales for 33 lower-order PD traits. Due to the number of items, a balanced incomplete block design (BIBD) incorporating “planned-missingness” was used to reduce the number of items a single participant must complete (for details see Simms et al., 2011; Wright & Simms, 2014). This procedure resulted in a median sample size of 195 (range = 183–218; 42–50% of sample) for individual traits for the present study. These participants all rated a subset of 216 items that have been used to define a static form (CAT-PD-SF) of each of the 33 CAT-PD traits. A five-point scale ranging from 1 (very untrue of me) to 5 (very true of me) is used. In the current study the median coefficient alpha was .85 (Range = .80–.89). Preliminary research suggests that the CAT-PD-SF scales relate strongly to the intended a priori PSY-5 structure (Wright & Simms, 2014).

Data Analyses

All trait and PD scales were first correlated with the octant scales of the IIP-SC; however, the IIP-SC’s structure allows more detailed analyses. IIP-SC octant scale interrelations are dictated by the measure’s specific geometric structure (Gurtman, 1992), such that adjacent scales (e.g., Vindictive and Cold) are more closely related than scales located further apart (e.g., Intrusive and Nonassertive). The specificity of these interrelations allows the discriminant validity of constructs (e.g., PD traits) to be examined via their patterns of correlations across IIP-SC scales.

The Structural Summary Method (SSM) for circumplex data (e.g., Wright, Pincus, Conroy, & Hilsenroth, 2009) quantifies these patterns. The SSM models correlations between the IIP-SC and external constructs as a cosine curve (see Figure 1) that is described by elevation, amplitude, R², and angular displacement parameters. Elevation is the average correlation across octants for a scale and is an indicator of general interpersonal distress (i.e., severity). Amplitude compliments elevation; it is the difference between the midpoint (i.e., mean correlation) and peak of the curve, with large values indicating that a scale relates to specific problematic behaviors. The degree to which a scale’s observed pattern of correlations matches a prototypical cosine curve is quantified by the R² statistic. Given that a trait has a prototypical profile of correlations and shows specific relations with the IIP-SC, angular displacement will indicate the area of the IPC that is most descriptive of the trait (i.e., style of interpersonal problems).

Figure 1.

Circumplex structural summary method parameters.

Recent developments in the use of the SSM (Wright et al., 2012; Zimmerman & Wright, under revision) provide standards for evaluating parameters and inference testing. Wright et al. (2012; Zimmerman & Wright, under revision) conducted a review of studies using the SSM and concluded that elevation and amplitude values > |.10| should be considered small effects, values > |.20| be regarded as moderate, and values > |.30| be interpreted as large. In addition, they recommend only interpreting amplitude and angular displacement when R² is above .70. Beyond this, Zimmerman and Wright (under revision) present a bootstrapping method (and R package) for placing confidence intervals (CIs) around SSM parameters, permitting hypothesis testing and additional precision. Monte Carlo simulations indicated that CIs for amplitude and angular displacement are influenced by sample size, as well as parameter magnitude. Due to this, Zimmerman and Wright recommend reporting the probability of obtaining an accurate CI (from R package) and if low (i.e., < .50), that CIs not be considered interpretable.

Elevation, amplitude, R², angular displacement, and associated CIs (for all parameters except R²) were computed for all traits and DSM PDs. Models were compared in terms of the number and proportion of components above cutoffs for elevation and amplitude, reflecting the severity and specificity of interpersonal problems they relate to. Finally, the angular displacement values across models were compared to examine hypotheses regarding differences in the interpersonal coverage of PD models.

Results

Bivariate Correlations

To aid interpretation, Table 1 contains domain-level trait intercorrelations (facet-level intercorrelations are available upon request). Table 2 presents descriptive statistics and coefficient alphas for PD construct scales, as well as bivariate correlations with all octant scales of the IIP-SC. Average IIP-SC octant correlations for the PID-5 (range = .23 to .42), CAT-PD-SF (range = .18 to .37), and SCID-II PQ (range = .14 to 35) peaked at moderate strength, whereas the NEO-PI-3FH was more weakly related to the IIP-SC (range = −.13 to .02). Scale-level correlations were examined to understand variability in PD model relations to IIP-SC octants. Most CAT-PD-SF, PID-5, and SCID-II PQ scales correlated positively and significantly with at least one IIP-SC scale; however, many NEO-PI-3FH scales were best characterized by significant negative correlations with the IIP-SC.

Table 1.

Domain-Level Intercorrelations for Personality Traits Measures

	NA	DET	ANT	DIS	PSY
Neuroticism	.78	.50	.12	.60	.40
Extraversion	−.35	−.73	.09	−.29	−.20
Agreeableness	−.27	−.19	−.67	−.34	−.34
Conscientiousness	−.46	−.39	−.19	−.73	−.40
Openness	−.05	−.31	−.07	−.04	.11

Note. Pairwise N = 379. NEO-PI-3FH domain scale names are listed in the first column and abbreviated PID-5 domain scale names are listed above columns. NA = Negative Affect, DET = Detachment, ANT = Antagonism, DIS = Disinhibition, and PSY = Psychoticism. Correlations > .10 are significant at p < .05. Correlations ≥ |.50| are in bold.

Table 2.

Personality Pathology Scale Descriptive Statistics and Correlations with IIP-SC Octant Scales

Measure/Scale Name	α	M	SD	PA	BC	DE	FG	HI	JK	LM	NO
NEO-PI-3FH
Neuroticism	.90	74.7	17.4	.39	.41	.37	.48	.43	.42	.39	.36
Anxiety	.67	14.2	3.6	.16	.23	.23	.38	.44	.42	.40	.22
Angry Hostility	.75	10.9	4.0	.54	.47	.25	.24	.04	.05	.11	.30
Depression	.77	13.8	4.3	.18	.26	.33	.41	.35	.35	.33	.18
Self-Consciousness	.62	12.2	3.7	.24	.34	.41	.59	.53	.48	.35	.23
Impulsiveness	.56	11.9	3.4	.31	.23	.15	.16	.21	.24	.22	.36
Vulnerability	.68	11.6	3.7	.33	.35	.31	.40	.41	.39	.37	.34
Extraversion	.83	75.1	15.1	−.13	−.35	−.54	−.73	−.50	−.37	−.25	.02
Warmth	.62	14.7	3.6	−.32	−.52	−.61	−.62	−.26	−.18	−.02	.04
Gregariousness	.75	10.7	4.1	−.09	−.27	−.43	−.61	−.31	−.18	−.13	.10
Assertiveness	.70	11.6	3.9	.12	−.07	−.25	−.52	−.52	−.42	−.26	.05
Activity	.41	11.9	3.2	−.04	−.15	−.27	−.38	−.29	−.23	−.21	−.07
Excitement-Seeking	.40	12.9	3.7	.02	.00	−.13	−.27	−.33	−.21	−.21	−.04
Positive Emotions	.67	13.2	3.9	−.21	−.39	−.48	−.52	−.28	−.27	−.19	−.04
Agreeableness	.82	84.9	12.9	−.52	−.53	−.26	−.07	.14	.15	.21	−.16
Trust	.75	12.3	3.8	−.33	−.53	−.38	−.31	−.07	−.11	−.04	−.05
Straightforwardness	.66	14.4	3.7	−.40	−.31	−.11	.07	.11	.09	.07	−.17
Altruism	.63	16.7	2.8	−.42	−.48	−.35	−.25	−.04	.05	.21	−.07
Compliance	.59	12.0	3.6	−.44	−.42	−.21	−.09	.20	.20	.13	−.11
Modesty	.56	13.7	3.3	−.15	−.03	.20	.38	.32	.29	.29	−.19
Tender-Mindedness	.60	15.9	3.0	−.24	−.25	−.11	−.09	.01	.06	.19	.02
Conscientiousness	.88	78.5	15.6	−.29	−.25	−.26	−.31	−.36	−.34	−.26	−.38
Competence	.56	14.0	3.1	−.31	−.32	−.34	−.36	−.37	−.38	−.30	−.31
Order	.75	13.7	3.8	−.01	.08	.04	−.07	−.23	−.21	−.13	−.22
Dutifulness	.50	12.9	3.3	−.29	−.27	−.20	−.17	−.10	−.08	−.06	−.26
Achievement Striving	.73	13.6	3.6	−.18	−.25	−.32	−.37	−.42	−.38	−.25	−.24
Self-Discipline	.80	13.2	4.1	−.23	−.17	−.23	−.33	−.42	−.35	−.29	−.32
Deliberation	.66	11.1	3.4	−.28	−.21	−.13	−.08	−.03	−.09	−.12	−.33
Openness	.80	84.0	13.1	−.21	−.29	−.24	−.20	−.05	−.03	.03	.05
Fantasy	.54	12.9	3.4	−.05	−.13	−.19	−.09	−.03	−.03	−.04	.14
Aesthetics	.66	14.4	3.8	−.16	−.26	−.17	−.18	−.08	−.06	.05	−.02
Feelings	.49	14.8	3.1	−.05	−.08	−.15	−.10	.09	.12	.20	.18
Actions	.43	12.5	2.9	−.18	−.28	−.21	−.29	−.18	−.13	−.10	−.03
Ideas	.62	14.0	3.7	−.17	−.20	−.13	−.12	−.10	−.10	−.07	−.05
Values	.40	15.5	2.8	−.23	−.22	−.12	.01	.14	.10	.11	−.02
M (facets)	.62	13.24	3.56	−.10	−.13	−.12	−.11	−.04	−.02	.02	−.01
SD (facets)	.11	1.48	.39	.24	.27	.25	.32	.28	.25	.21	.20
M (re-keyed facets)	−	−	−	.20	.25	.24	.27	.20	.17	.13	.12
SD (re-keyed facets)	−	−	−	.17	.16	.15	.20	.20	.18	.17	.15
PID-5
Negative Affect	.94	1.3	0.7	.46	.49	.40	.46	.41	.46	.47	.43
Submissiveness	.78	1.3	0.7	.12	.17	.17	.29	.56	.60	.51	.31
Separation Insecurity	.87	1.0	0.8	.39	.38	.26	.32	.25	.31	.31	.40
Anxiousness	.91	1.5	0.8	.37	.47	.43	.48	.44	.46	.45	.32
Emotional Lability	.91	1.4	0.8	.43	.42	.35	.41	.38	.42	.47	.41
Perseveration	.87	1.1	0.7	.42	.42	.40	.44	.44	.46	.45	.39
Detachment	.93	1.1	0.6	.34	.52	.74	.68	.44	.39	.29	.12
Depressivity	.94	0.9	0.7	.38	.48	.54	.54	.47	.50	.44	.33
Suspiciousness	.83	1.2	0.7	.44	.61	.48	.38	.14	.24	.25	.19
Restricted Affect	.75	0.9	0.6	.26	.39	.53	.35	.16	.15	.01	−.03
Withdrawal	.91	1.2	0.7	.30	.51	.70	.73	.39	.31	.21	−.01
Intimacy-Avoidance	.86	0.7	0.7	.19	.30	.53	.33	.23	.21	.10	.06
Anhedonia	.90	1.2	0.8	.31	.44	.55	.57	.42	.40	.37	.22
Antagonism	.92	0.7	0.5	.48	.44	.27	.04	.07	.09	.05	.33
Manipulativeness	.84	0.8	0.7	.41	.32	.19	−.04	.03	.05	.03	.28
Deceitfulness	.87	0.6	0.6	.46	.47	.28	.11	.14	.14	.07	.30
Hostility	.88	1.1	0.7	.66	.66	.46	.37	.14	.15	.13	.31
Callousness	.89	0.4	0.5	.58	.65	.49	.27	.09	.12	.03	.20
Attention Seeking	.91	0.9	0.7	.41	.21	.04	−.17	.01	.07	.08	.51
Grandiosity	.78	0.7	0.6	.38	.37	.24	.05	.02	.04	.03	.27
Disinhibition	.93	1.0	0.6	.52	.50	.46	.40	.37	.42	.35	.47
Irresponsibility	.81	0.7	0.6	.45	.44	.40	.30	.29	.32	.23	.40
Impulsivity	.86	1.0	0.7	.50	.46	.36	.25	.16	.24	.22	.41
Distractibility	.91	1.2	0.8	.38	.38	.41	.45	.47	.49	.42	.39
Rigid Perfectionism	.91	1.3	0.7	.38	.42	.37	.33	.17	.22	.24	.16
Risk Taking	.87	1.2	0.6	.26	.22	.09	−.09	−.12	−.04	−.04	.19
Psychoticism	.96	0.8	0.6	.50	.48	.46	.35	.24	.31	.30	.40
Eccentricity	.96	1.0	0.8	.45	.45	.43	.39	.28	.35	.31	.40
Perceptual-Dysregulation	.87	0.7	0.6	.50	.50	.45	.35	.26	.32	.30	.37
Unusual Beliefs	.83	0.6	0.6	.38	.34	.33	.18	.08	.14	.19	.28
M (facets)	.87	1.00	.69	.39	.42	.38	.30	.24	.27	.23	.28
SD (facets)	.05	.28	.09	.11	.12	.15	.21	.18	.17	.17	.13
CAT-PD-SF
Affective Lability	.86	2.8	1.1	.50	.56	.47	.45	.39	.40	.38	.34
Anger	.85	2.7	1.1	.54	.51	.36	.38	.00	.09	.20	.25
Anhedonia	.89	2.5	1.0	.35	.45	.49	.54	.31	.26	.29	.32
Anxiousness	.85	3.0	1.0	.26	.40	.36	.42	.40	.41	.42	.29
Callousness	.84	1.6	0.7	.54	.54	.43	.26	.14	.10	−.12	.12
Cognitive Problems	.88	2.6	1.0	.28	.34	.43	.47	.48	.51	.51	.33
Depressiveness	.88	2.7	1.0	.29	.42	.41	.48	.29	.25	.29	.30
Domineering	.84	2.2	0.9	.70	.42	.31	.09	−.04	−.08	−.02	.29
Emotional Detachment	.86	2.9	1.0	.15	.23	.55	.50	.26	.25	.11	−.24
Exhibitionism	.83	2.2	1.0	.26	.09	−.08	−.37	−.19	−.16	−.09	.37
Fantasy Proneness	.83	2.6	1.0	.41	.41	.43	.37	.30	.35	.40	.39
Grandiosity	.82	1.9	0.8	.39	.36	.16	.05	.01	−.02	−.07	.33
Health Anxiety	.84	2.6	1.0	.37	.39	.39	.37	.26	.29	.39	.31
Hostile Aggression	.87	1.5	0.7	.64	.69	.43	.28	−.03	.12	.11	.28
Irresponsibility	.86	2.0	0.8	.23	.18	.10	.15	.23	.18	.00	.22
Manipulativeness	.85	1.6	0.7	.37	.47	.36	.19	.22	.14	−.01	.16
Mistrust	.88	2.8	1.1	.47	.65	.49	.37	.19	.23	.21	.17
Non-Perseverance	.88	2.5	1.0	.42	.41	.38	.47	.42	.42	.32	.34
Non-Planfulness	.84	2.5	1.0	.36	.34	.29	.15	.16	.18	.20	.32
Norm Violation	.84	2.3	1.0	.46	.41	.19	.08	−.15	−.07	.01	.28
Peculiarity	.82	2.5	1.0	.45	.43	.44	.36	.35	.42	.43	.36
Perfectionism	.86	3.3	1.0	.08	.15	.13	.10	−.04	−.02	.14	.13
Relationship-Insecurity	.83	2.5	1.0	.45	.58	.50	.49	.34	.34	.26	.17
Rigidity	.81	2.6	0.8	.43	.34	.25	.22	.07	.06	.20	.32
Risk Taking	.84	1.8	0.9	.22	.26	.18	.04	.00	.03	−.04	.09
Romantic Disinterest	.88	2.2	1.0	.00	.10	.25	.16	.07	.03	−.12	−.09
Rudeness	.80	2.1	0.9	.63	.44	.27	.19	.06	.18	.29	.54
Self Harm	.86	1.6	0.9	.30	.23	.21	.25	.17	.25	.21	.19
Social Withdrawal	.87	2.8	1.0	.20	.39	.61	.79	.49	.37	.23	−.08
Submissiveness	.86	2.3	1.0	.20	.32	.29	.36	.56	.66	.52	.37
Unusual Beliefs	.85	1.5	0.7	.31	.27	.28	.07	.09	.14	.10	.26
Unusual Experiences	.82	1.6	0.8	.46	.46	.44	.22	.16	.23	.26	.35
Workaholism	.86	2.5	0.9	.09	.14	.18	.08	.05	.11	.13	.10
M	.85	2.33	.92	.36	.37	.33	.27	.18	.20	.19	.24
SD	.02	.48	.12	.16	.15	.15	.21	.19	.18	.18	.16
SCID-II PQ
Avoidant PD	.73	10.4	2.1	.20	.35	.49	.68	.47	.39	.31	.02
Dependent PD	.60	10.2	1.8	.28	.32	.29	.31	.30	.35	.33	.28
Obsessive Compulsive PD	.54	13.5	1.9	.27	.19	.26	.21	.10	.09	.10	.08
Paranoid PD	.78	12.0	2.5	.41	.55	.49	.41	.16	.16	.17	.19
Schizotypal PD	.72	14.9	2.5	.33	.36	.36	.33	.16	.20	.20	.20
Schizoid PD	.51	8.2	1.5	.19	.35	.46	.35	.08	.09	.01	.03
Histrionic PD	.51	8.8	1.8	.30	.16	−.02	−.17	−.06	.02	.06	.38
Narcissistic PD	.78	22.2	3.6	.46	.41	.27	.11	−.03	.02	.01	.28
Borderline PD	.86	21.6	4.2	.46	.51	.45	.42	.25	.27	.30	.31
Antisocial PD/Conduct Dis.	.81	16.2	2.6	.32	.32	.22	.10	−.03	−.04	−.01	.02
M	.68	13.8	2.45	.32	.35	.33	.28	.14	.15	.15	.18
SD	.13	5.0	.84	.10	.12	.16	.23	.17	.15	.13	.13

Note. PA = Domineering (90°), BC = Vindictive (135°), DE = Cold (180°), FG = Socially Avoidant (225°), HI = Nonassertive (270°), JK = Exploitable (315°), LM = Overly Nurturant (0/360°), and NO = Intrusive (45°; Soldz et al., 1995). Bold values represent the largest correlation in a row. For personality traits, domain-level scales are left-justified relative to facets. For all measures, r > |.15| is significant at p < .05 and r > |.25| significant at p < .01. Lowest pair-wise N by measure: CAT-PD-SF, N = 136; PID-5, N = 437; NEO-PI-3FH, N = 381; SCID-II PQ, N = 433.

In this study, NEO-PI-3FH scales initially were scored in the direction suggested by their scale names (e.g., a high score on “Altruism” indicates trait altruism). However, NEO traits are considered bipolar, implying low scores are meaningful as well (e.g., Samuel, 2011). To focus on dysfunction, all NEO-PI-3FH scales with negative average correlations with the IIP-SC were reverse-scored. These re-keyed traits were used in the remaining analyses and marked with minus signs. Reverse scoring these scales altered the average correlations and standard deviations in Table 2, therefore adjusted means (range = .12 to .27) and standard deviations are reported below the original values. Although the mean values still were lower than those for the other measures, they were considerably higher than the original NEO-PI-3FH values.

SSM Parameters

SSM parameters and associated CIs are summarized in Table 3. Median elevation parameters for CAT-PD-SF facets (Mdn = .29) and PID-5 facets (Mdn = .34) indicate relations to severe dysfunction, whereas SCID-II-PQ PD scales (Mdn = .24) and NEO-PI-3FH (Mdn = .20) facets were less severe. This finding supports Hypothesis 3a that pathological traits will relate more strongly to severe problems than normal-range traits. Median amplitude values indicated that measures included a similar proportion, and considerable number, of scales related to specific interpersonal problems. To understand these relations, SSM Parameters for individual scales with amplitudes greater than or equal to .10 and R² greater than or equal to .70 are organized by IIP-SC octant and displacement degree in Table 4. The remaining scales yielded less meaningful angular displacements and thus are provided separately in Table 5.

Table 3.

Median SSM parameters by measure

Scale	Elevation	Amplitude	R2
SCID-II PQ	.24 (.09, .37)	.17 (.02, .25)	.89 (.28, .95)
CAT-PD-SF	.29 (−.02, .44)	.15 (.02, .34)	.82 (.03, .97)
PID-5 Facets	.34 (.06, .46)	.16 (.01, .31)	.83 (.47, .95)
PID-5 Domains	.43 (.22, .45)	.12 (.01, .26)	.88 (.13, .93)
NEO-PI-3FH Facets	−.11 (−.33, .40)	.16 (.03, .33)	.88 (.59, .98)
NEO-PI-3FH Domains	−.13 (−.36, .35)	.16 (.04, .37)	.87 (.31, .93)
NEO-PI-3FH Facets (re-keyed)	.20 (.03, .40)	.16 (.03, .33)	.88 (.59, .98)
NEO-PI-3FH Domains (re-keyed)	.31 (.12, .40)	.17 (.04, .37)	.87 (.31, .93)

Note. Median values for each parameter are provided, along with the range for the parameter within parentheses. Elevations and amplitudes = .10 are small, .20 moderate, .30 large. R² > .70 are acceptable.

Table 4.

SSM Parameters for PD and PD Trait Scales Related to Specific Interpersonal Problems

Angular Displacement	Measure	Scale	Elevation	Amplitude	R2
PA (67.5° – 112.5°): Domineering
73.8 (54.3, 91.4)	CAT-PD-SF	Exhibitionism	−.02 (−.13, .11)	.30 (.22, .39)	.85
80.5 (53.5, 106.7)	NEO-PI-3FH	C: -Deliberation	.16 (.09, .22)	.13 (.08, .18)	.89
85.2 (59.4, 104)	CAT-PD-SF	Rudeness	.33 (.20, .44)	.25 (.17, .34)	.95
98.4 (78.9, 115.1)	PID-5	Dis: Risk Taking	.06 (−.01, .13)	.19 (.13, .25)	.94
98.8 (82.8, 114.1)	PID-5	Ant: Manipulativeness	.16 (.09, .23)	.20 (.15, .25)	.86
101.8 (39, 143.8)	CAT-PD-SF	Non-Planfulness	.25 (.14, .35)	.11 (.04, .20)	.93
105.8 (74.8, 141.5)	CAT-PD-SF	Rigidity	.23 (.14, .33	.17 (.07, .27)	.92
106.7 (63.4, 139.4)	CAT-PD-SF	Unusual Beliefs	.19 (.00, .34)	.12 (.06, .19)	.80
106.8 (88.5, 124.3)	PID-5	Dis: Impulsivity	.33 (.27, .38)	.16 (.11, 21)	.93
107 (92.7, 119.2)	PID-5	Antagonism	.22 (.14, .29)	.23 (.18, .28)	.88
107.4 (94.7, 119.8)	NEO-PI-3FH	A: -Straightforwardness	.08 (.01, .15)	.25 (.20, .31)	.93
109.1 (84.8, 131)	CAT-PD-SF	Norm Violation	.15 (.02, .28)	.29 (.20, .39)	.97
110.1 (88.3, 130.9)	CAT-PD-SF	Grandiosity	.15 (.02, .28)	.23 (.13, .33)	.84
110.7 (93.5, 124.3)	PID-5	Ant: Grandiosity	.17 (.09, .24)	.20 (.15, .26)	.94
BC (112.5° - 157.5°): Vindictive
112.8 (91.1, 133.7)	PID-5	Psy: Unusual Beliefs and Experiences	.24 (.17, .31)	.14 (.09, .19)	.95
113.8 (97.2, 128.5)	CAT-PD-SF	Domineering	.21 (.10, .30)	.34 (.25, .43)	.89
114 (77.1, 139.9)	CAT-PD-SF	Unusual Experiences	.32 (.16, .46)	.15 (.07, .24)	.92
114.1 (98.7, 127.2)	PID-5	Ant: Deceitfulness	.25 (.17, .32)	.19 (.14, .24)	.83
115.3 (101.5, 128.8)	NEO-PI-3FH	N: Angry Hostility	.25 (.19, .31)	.23 (.18, .29)	.90
115.7 (103.2, 126.6)	SCID-II PQ	Narcissistic PD	.19 (.13, .25)	.25 (.20, .30)	.95
117.9 (80.6, 145.9)	NEO-PI-3FH	C: -Dutifulness	.18 (.11, .24)	.11 (.06, .17)	.84
120.9 (95.3, 143.8)	PID-5	Psychoticism	.38 (.31, .44)	.12 (.07, .17)	.93
123.4 (108.3, 138)	CAT-PD-SF	Hostile Aggression	.32 (.19, .42)	.32 (.24, .41)	.91
124.5 (107.2, 141.2)	NEO-PI-3FH	O: -Values	.03 (−.04, .10)	.19 (.14, .24)	.94
124.7 (115.3, 133.1)	NEO-PI-3FH	A: -Compliance	.09 (.03, .16)	.33 (.28, .39)	.96
125.1 (101.6, 148.7)	PID-5	Psy: Perceptual Dysregulation	.38 (.31, .44)	.12 (.07, .17)	.91
125.8 (96, 148.5)	CAT-PD-SF	Anger	.29 (.19, .39)	.23 (.14, .34)	.82
126.4 (118.7, 134.2)	NEO-PI-3FH	−Agreeableness	.13 (.07, .19)	.37 (.32, .43)	.93
130.1 (89.8, 170.8)	CAT-PD-SF	Risk Taking	.10 (−.01, .20)	.14 (.05, .24)	.87
131.1 (122.2, 139.5)	PID-5	Ant: Hostility	.36 (.30, .42)	.28 (.23, .33)	.92
135 (122.2, 148)	SCID-II PQ	Antisocial PD (Childhood Criter	ia) .11 (.04, .19)	.19 (.14, .25)	.90
137.4 (130.3, 144.8)	PID-5	Ant: Callousness	.30 (.23, .38)	.30 (.26, .35)	.91
143.2 (124.6, 164.3)	SCID-II PQ	Borderline PD	.37 (.31, .42)	.13 (.08, .18)	.89
145.9 (130.3, 161.4)	PID-5	Det: Suspiciousness	.34 (.28, .40)	.19 (.13, .25)	.78
146.9 (127.8, 165)	NEO-PI-3FH	A: -Tenderness	.05 (−.01, .11)	.18 (.12, .24)	.79
147.4 (132.1, 165.6)	CAT-PD-SF	Callousness	.25 (.12, .36)	.29 (.20, .37)	.84
149 (126.2, 170.8)	CAT-PD-SF	Mistrust	.35 (.25, .43)	.21 (.13, .30)	.81
149 (140.2, 158.3)	NEO-PI-3FH	A: -Altruism	.17 (.10, .24)	.31 (.25, .37)	.91
150 (127.9, 180.4)	CAT-PD-SF	Manipulativeness	.24 (.10, .35)	.18 (.09, .28)	.80
150.3 (127, 182.5)	PID-5	Dis: Rigid Perfectionism	.28 (.22, .34)	.11 (.05, .18)	.73
151.9 (123.9, 181)	SCID-II PQ	Schizotypal PD	.27 (.20, .33)	.10 (.06, .16)	.85
154 (141.1, 167.3)	SCID-II PQ	Paranoid PD	.32 (.26, .37)	.21 (.16, .26	.93
155.6 (140.4, 170.3)	NEO-PI-3FH	A: -Trust	.23 (.16, .29)	.22 (.17, .28)	.86
DE (157.5° - 202.5°): Cold
165 (145.1, 184)	NEO-PI-3FH	−Openness	.12 (.04, .19)	.17 (.11, .22)	.87
168.1 (139.3, 197.9)	NEO-PI-3FH	O: -Aesthetics	.11 (.04, .17)	.12 (.06, .19)	.84
175.5 (144.4, 205.7)	CAT-PD-SF	Relationship Insecurity	.39 (.28, .49)	.16 (.09, .24)	.76
175.7 (161.6, 192.1)	SCID-II PQ	Schizoid PD	.20 (.13, .26)	.22 (.17, .27)	.92
176.2 (139.9, 230.7)	CAT-PD-SF	Anhedonia	.38 (.27, .47)	.12 (.04, .22)	.82
180.2 (168, 193.4)	PID-5	Det: Restricted Affectivity	.23 (.15, .30)	.24 (.19, .30)	.89
187.5 (177.9, 197.7)	NEO-PI-3FH	E: -Warmth	.31 (.25, .37)	.33 (.27, .38)	.93
190.6 (152.3, 231.6)	NEO-PI-3FH	O: -Fantasy	.05 (−.01, .11)	.10 (.04, .16)	.59
191.2 (157.3, 231.9)	NEO-PI-3FH	E: -Actions	.18 (.10, .25)	.10 (.05, .17)	.76
194.4 (175.4, 212.3)	PID-5	Det: Intimacy Avoidance	.24 (.17, .31)	.18 (.13, .23)	.82
198.6 (162.6, 236.8)	CAT-PD-SF	Romantic Disinterest	.05 (−.07, .16)	.16 (.07, .25)	.92
FG (202.5° - 247.5°): Socially Avoidant
202.7 (192.9, 213.2)	PID-5	Det: Withdrawal	.39 (.34, .44)	.31 (.27, .36)	.90
203.5 (192.4, 214.6)	PID-5	Detachment	.44 (.39, .49)	.26 (.22, .30)	.89
203.6 (186.3, 221.5)	NEO-PI-3FH	E: -Positive Emotions	.30 (.24, .36)	.20 (.15, .26)	.86
214.1 (193.5, 233.2)	CAT-PD-SF	Emotional Detachment	.22 (.12, .33)	.29 (.22, .37)	.79
215.7 (192.7, 241)	PID-5	Det: Anhedonia	.41 (.36, .46)	.14 (.09, .19)	.83
217 (203.6, 229.5)	NEO-PI-3FH	E: -Gregariousness	.24 (.18, .30)	.27 (.23, .32)	.89
221.7 (206.5, 237.2)	CAT-PD-SF	Social Withdrawal	.37 (.27, .48)	.34 (.27, .42)	.91
228.7 (217.6, 239.3)	NEO-PI-3FH	−Extraversion	.36 (.30, .41)	.30 (.26, .35)	.93
231.2 (219.1, 242.8)	SCID-II PQ	Avoidant PD	.36 (.31, .42)	.25 (.21, .29)	.87
246.7 (221.2, 271.8)	NEO-PI-3FH	E: -Activity	.21 (.15, .26)	.14 (.10, .20)	.90
HI (247.5° - 292.5°): Nonassertive
252.3 (234.5, 268.8)	NEO-PI-3FH	N: Self-Consciousness	.40 (.34, .45)	.17 (.12, .22)	.96
256.8 (222.4, 286.8)	NEO-PI-3FH	N: Depression	.30 (.23, .35)	.10 (.06, .15)	.84
260.4 (226.5, 290.4)	NEO-PI-3FH	C: -Achievement-Striving	.30 (.24, .36)	.11 (.06, .16)	.30
263.5 (252.6, 273.8)	NEO-PI-3FH	E: -Assertiveness	.23 (.17, .29)	.32 (.28, .37)	.98
265 (252.1, 277)	NEO-PI-3FH	A: Modesty	.14 (.07, .21)	.27 (.23, .32)	.84
275.4 (255.2, 294.2)	NEO-PI-3FH	E: -Excitement-Seeking	.15 (.08, .21)	.17 (.11, .23)	.94
283.5 (223.4, 321.8)	CAT-PD-SF	Cognitive Problems	.42 (.33, .50)	.11 (.04, .20)	.82
291.2 (269.5, 310.9)	NEO-PI-3FH	N: Anxiety	.31 (.25, .37)	.14 (.10, .19)	.92
JK (292.5° - 337.5°): Exploitable
308.2 (286.9, 330.8)	CAT-PD-SF	Submissiveness	.41 (.30, .50)	.19 (.11, 29)	.86
311.8 (300.5, 323.3)	PID-5	Na: Submissiveness	.34 (.28, .40)	.25 (.20, .31)	.95
326.2 (301.5, 348.6)	NEO-PI-3FH	C: -Order	.09 (.03, .16)	.14 (.08, .21)	.80
LM (337.5° - 22.5°): Overly Nurturant
353.3 (335.8, 13.3)	NEO-PI-3FH	O: Feelings	.03 (−.04, .10)	.17 (.12, .23)	.92
NO (22.5° - 67.5°): Intrusive
66 (52.3, 79.9)	SCID-II PQ	Histrionic PD	.09 (.02, .15)	.23 (.19, .28)	.89
66.8 (54.5, 79.2)	PID-5	Ant: Attention-Seeking	.14 (.08, .21)	.27 (.23, .31)	.83

Note. Scales with amplitudes < .10 or R² < .70 were omitted, as placing them within one octant is problematic. N = Neuroticism, E = Extraversion, O = Openness, A = Agreeableness, C = Conscientiousness, Na = Negative Affect, Det = Detachment, Ant = Antagonism, Dis = Disinhibition, and Psy = Psychoticism. Elevations and amplitudes = .10 are small, .20 moderate, .30 large. R² > .70 are acceptable. In all cases the probability of confidence interval accuracy was above .70. Scales representing higher-order domains are in bold.

Table 5.

SSM Parameters for PD and PD Trait Scales Not Related to Specific Interpersonal Problems

Measure	Scale	Elevation	Amplitude	R2	Angular Displacement	CI Accuracy
SCID-II PQ	Dependent	.31 (.24, .37)	.02 (.01, .08)	.28	294.6 (139.7, 99.6)	.11
	Obsessive-Compulsive	.16 (.10, .23)	.09 (.04, .14)	.72	115.6 (119.9, 195.3)	.89
CAT-PD-SF	Affective Lability	.44 (.35, .51)	.08 (.02, .16)	.69	151.3 (97.3, 219.9)	.32
	Anxiousness	.37 (.27, .47)	.06 (.02, .15)	.55	265.1 (150, 347)	.25
	Depressiveness	.34 (.24, .44)	.09 (.02, .20)	.72	180.4 (127.8, 270.3)	.53
	Fantasy	.38 (.27, .48)	.04 (.01, .14)	.61	108.7 (316.1, 242.3)	.12
	Health Anxiety	.35 (.24, .44)	.05 (.01, .14)	.41	125.4 (337.7, 276)	.11
	Irresponsibility	.16 (.06, .28)	.02 (.01, .14)	.03	142.5 (329.8, 303.4)	.10
	Non-Perseverance	.40 (.30, .48)	.05 (.01, .15)	.48	213.6 (91.8, 347.3)	.17
	Peculiarity	.41 (.31, .50)	.03 (.01, .13)	.27	104.1 (297, 255.8)	.08
	Perfectionism	.08 (−.05, .21)	.07 (.02, .19)	.51	108.4 (349.9, 210.1)	.26
	Self-Harm	.23 (.10, .35)	.02 (.01, .11)	.11	112.1 (308.5, 285.1)	.05
	Workaholism	.11 (0, .21)	.02 (.01, .16)	.20	126.4 (313.5, 287.9)	.11
PID-5	Na: Anxiousness	.43 (.38, .47)	.05 (.02, .10)	.48	242.7 (174.2, 302.8)	.43
	Na: Depressivity	.46 (.41, .51)	.09 (.05, .13)	.77	227.6 (191.3, 262.1)	.74
	Dis: Distractibility	.43 (.37, .48)	.05 (.02, .11)	.92	281.5 (206.1, 331.4)	.57
	Psy: Eccentricity	.38 (.32, .44)	.08 (.03, .14)	.83	127.1 (84.6, 165.2)	.75
	Na: Emotional Lability	.41 (.36, .46)	.03 (.01, .09)	.49	21.4 (285.7, 139.8)	.27
	Dis: Irresponsibility	.35 (.28, .42)	.09 (.04, .14)	.76	120.8 (88, 152.6)	.87
	Na: Perseveration	.43 (.37, .38)	.02 (.01, .08)	.47	295.5 (148.3, 80.8)	.13
	Na: Separation Insecurity	.33 (.27, .39)	.06 (.02, .12)	.73	78.7 (18.3, 124.7)	.75
	Disinhibition	.43 (.37, .49)	.07 (.03, .12)	.76	114.3 (66, 151.4)	.75
	Negative Affect	.45 (.40, .49)	.01 (.01, .07)	.13	56.1 (278.1, 210)	.06
NEO-PI-3FH	N: Impulsiveness	.23 (.17, .30)	.08 (.03, .14)	.71	49 (359.4, 93.9)	.59
	N: Vulnerability	.36 (.31, .42)	.04 (.01, .10)	.69	285.8 (171.2, 38.4)	.30
	C: -Competence	.33 (.27, .40)	.03 (.01, .09)	.78	254.3 (142.1, 354.6)	1.00
	C: -Self-Discipline	.29 (.23, .35)	.09 (.04, .15)	.78	299.2 (249.8, 332.1)	.85
	O: -Ideas	.11 (.04, .19)	.05 (.01, .12)	.61	154.7 (71.2, 239.1)	.47
	O: -Fantasy	.05 (−.01, .11)	.10 (.04, .16)	.59	190.6 (152.3, 231.6)	.92
	Neuroticism	.40 (.35, .45)	.04 (.01, .08)	.53	242.9 (137.2, 319.5)	.24
	Conscientiousness	.31 (.25, .37)	.04 (.01, .08)	.31	320.9 (182.3, 85.2)	.21

Note.All scales had amplitudes < .10 and thus angular displacement values should be interpreted with caution. Personality trait domains are emboldened. Confidence intervals should only be interpreted when the probability of their accuracy is > .50.

Examining Table 4 informs the prediction (H1) that previous findings with the DSM-5 PD models would replicate. Eight (out of 10; 83%) SCID-II-PQ scales related to the IIP-SC with specificity (amplitude > .10); the highest amplitudes PDs related to Vindictive (Narcissistic, Antisocial, Paranoid PDs), Cold (Schizoid PD), and Socially Avoidant (Avoidant PD) interpersonal problems. Histrionic PD was the only PD related to warm interpersonal problems (i.e., Intrusive); Dependent PD (DPD) did not show its predicted relation to warm problems. The 17 (out of 25; 68%) PID-5 facets with specific relations to the IIP-SC mostly related to Domineering (e.g., Manipulativeness), Vindictive (e.g., Callousness), and Cold (e.g., Restricted Affectivity) interpersonal problems; few traits related to warmer or purely submissive problems.

To test the hypothesis (H2) that Extraversion and Agreeableness would relate to interpersonal problems most specifically, facets of the PID-5 and NEO-PI-3FH were examined based on domain. As Tables 3 and 4 indicate, NEO Extraversion and Agreeableness facets all related to specific problems, whereas the other domains lacked such uniformity. PID-5 antagonism and detachment facets (variants of Extraversion and Agreeableness) were similarly specific; all antagonism facets and all but one detachment facet related to specific problems.

Predictions regarding interpersonal coverage differences between trait models (H3b & H3c) were examined by analyzing Table 4 on a measure-by-measure basis. As noted above, the PID-5 facets were related mostly to Domineering, Vindictive, and Cold problems. The CAT-PD-SF traits showed a similar pattern of relations with IIP-SC, as the 22 (out of 33; 67%) facets with specific relations to the IIP-SC were in the Domineering (e.g., Norm Violation), Vindictive (e.g., Hostile Aggression), and Cold (e.g., Relationship Insecurity) problems octants. The CAT-PD-SF provided some additional coverage of warm problems, as Rudeness related to somewhat warm problems (i.e., angular displacement = 85.2°). The CAT-PD-SF also provided some additional coverage of the Socially Avoidant octant, with two traits having large amplitudes (Emotional Detachment and Social Withdrawal), relative to the one PID-5 facet (Withdrawal) with very specific relations. However, overall, the PID-5 and CAT-PD-SF traits converged quite strongly in their relations to the IIP-SC. This is in line with the prediction (H3) that more dramatic differences would exist between models of differing measurement range.

Finally, the NEO-PI-3FH showed a different pattern of relations to the IIP-SC. Of the 24 facets (out of 30; 80%) with specific relations to the IIP-SC, the most specific were located within the Nonassertive (e.g., Modesty), Socially Avoidant (e.g., -Gregariousness), Cold (e.g., -Warmth) and Vindictive (e.g., -Altruism) octants. Only one facet (-Straightforwardness) related to Domineering problems with moderate specificity (amplitude > .20), whereas several NEO-PI-3FH traits related to pure submissiveness (e.g., Nonassertive problems). Similar to all measures, however, the NEO-PI-3FH did not relate strongly or specifically to warm problems. The NEO-PI-3FH facet Feelings (amplitude = .17) was the most specific; however, it did not relate to interpersonal distress (elevation = .03 [−.04, .10]). The trait of -Order was less specific in its relation to warm problems (amplitude = .14), though its elevation was slightly higher (.09). Finally, -Deliberation and Anxiety also related to warm problems, but were more strongly characterized by submissiveness and dominance than warmth.

Discussion

The present study compared four PD measures by analyzing their relations with interpersonal dysfunction in a psychiatric sample. In doing so, previous findings relating the DSM-5 PD measures to interpersonal problems largely were replicated. Across trait measures, Agreeableness and Extraversion facets related most specifically to interpersonal problems; however, PID-5 and CAT-PD traits both related to generally greater problem severity than did NEO-PI-3FH traits. In addition, pathological traits more strongly related to dominance problems than did normal-range traits, which more strongly covered submissive problems.

DSM-5 Section II and III PD Measures

Recent statistical advances (e.g., Zimmerman & Wright, in revision) and a patient sample were used to re-examine previous research relating DSM-5 PD models to the IIP (e.g., Soldz et al., 1993; Wright et al., 2012). Previous findings with DSM-5 traditional categorical PDs (i.e., Section II) generally were replicated, the exception being that dependent PD (DPD) did not relate to specific (i.e., warm) interpersonal problems, as indicated by DPD’s low amplitude value (i.e., < .01). The low amplitude value was driven by DPD’s correlations with Domineering, Vindictive, and Cold problems, which were higher than expected. Despite the low amplitude value, DPD did relate to severe interpersonal distress, as indicated by the elevation parameter (i.e., .31). Possible explanations for these findings include (a) differences between the SCID-II PQ and previously used measures, (b) failure of previous studies to quantify specificity (i.e., amplitude), and (c) random sampling error. Disentangling the influences of these factors is difficult; however, the low internal consistency of SCID-II PQ DPD scale relative to measures used in previous studies (e.g., Pincus & Wiggins, 1990) is a plausible explanation for this divergent finding.

In general, the present findings for the emerging trait model (i.e., Section III) mirror those from Wright et al.’s (2012) student sample. Exceptions were Rigid Perfectionism and Suspiciousness relating to more specific problems and Anhedonia and Irresponsibility relating to less specific ones (i.e., differences in amplitude). Notably these facet-level differences were small; in contrast, domain-level discrepancies were larger. Disinhibition and Negative Affect related to relatively non-specific problems, in contrast with their specificity in Wright et al. Differences in scoring methods may have produced this finding. The present study averaged three traits within each domain and specifically excluded interstitial traits of potential interpersonal relevance (e.g., submissiveness; APA, 2013), whereas Wright et al. used factor scores that were based on all traits. Overall, Wright et al.’s (2012) results generally replicated in our patient sample, but more work is needed to replicate these findings given the variety of PID-5 domain scoring methods used in the literature.

Extraversion and Agreeableness

As predicted from previous research (e.g., McCrae & Costa, 1989), Extraversion and Agreeableness facets related to the most specific interpersonal problems. Interestingly, NEO-PI-3FH Angry Hostility (Neuroticism) and PID-5 Submissiveness (Negative Affect) also related to quite specific problems (amplitude > .20). Weak loadings on posited domains or sizable loadings on other domains may explain this finding (e.g., McCrae, et al., 2005; Wright & Simms, 2014). Several NEO Neuroticism facets also had somewhat specific (amplitude > .10) relations to Nonassertive problems. This may be the result of the NEO Neuroticism domain containing most of the measure’s negatively valenced content, relative to pathological trait models (e.g., Wright & Simms, 2014). Despite some informative discrepancies, the confirmation of this hypothesis highlights a point of agreement between normal-range and pathological trait models.

Differences Between Normal-Range and Pathological Measures

Hypotheses regarding measurement-range were broadly supported, in that all pathological measures related to somewhat different and more severe problems (i.e., higher elevation) than the NEO-PI-3FH. Notably, these conclusions held regardless of whether the NEO-PI-3FH traits were keyed normally or keyed to maximize relations with the IIP-SC. However, the prediction that normal-range and pathological traits would be differentiated by coverage of warm vs. cold problems was not supported. Evaluating this hypothesis is complex, in that NEO-PI-3FH traits were re-keyed. Correlational results suggest that many of these scales, if not re-keyed, would have positive amplitudes, large negative elevations, and angular displacements indicating warmth. However, large negative elevations imply that these scales relate to warm interpersonal behaviors rather than problems (e.g., Hopwood et al., 2013). Re-keyed NEO facets did not cover warm problems better than pathological traits, instead relating to cold problems with similar strength as pathological traits. In this sense, pathological and normal-range traits converged in their (lacking) coverage of warm interpersonal problems.

It is notable that the present study provided representation of all currently prominent PD models and thus a broad array of PD constructs, yet still found limited coverage of warm problems. Several possible explanations for this limited coverage exist. First, warm behavior may largely be adaptive rather than impairing (e.g., Baumeister & Leary, 1995); it might be uncommon for warm behaviors to lead to problems. If few people have warm problems, then fewer relations to common personality pathology might be expected. Second, linguistic constraints could explain the relative dearth of warm pathology across PD models, as there are fewer negatively valenced descriptors for warm behaviors (Leary, 1957; Coker, Samuel, & Widiger, 2002). Thus it might be the case that either (a) PD trait models fail to represent reality because they were founded on lexical work (e.g., Block, 1995), or (b) the IIP-SC overemphasizes warm problems. However, linguistic underrepresentation or cultural minimization of warm pathology does not equate to it being nonexistent or clinically irrelevant (e.g., Block, 1995). Finally, people may be less aware of their warm interpersonal problems. Supporting this possibility is research suggesting that individuals with PD bother others most when they attempt to engage in warm, affiliative behaviors (e.g., Williams et al., in press).

More vexing is the finding that more normal-range traits (NEO-PI-3FH) related to submissive problems than did pathological traits, which related in greater number to dominant problems. If the facets within these measures are taken to represent distinct patterns of thoughts, feelings, and behaviors that lead to or include interpersonal problems, then this finding may be interpreted as measures differing in their ability to describe diverse pathways to, or contexts for, dominant vs. submissive problems. In other words, substantive differences in the constructs represented by these measures may explain this finding. For instance, no NEO trait appears to correspond to grandiosity in interpersonal content (i.e., domineering), whereas grandiosity is among a number of CAT-PD and PID-5 traits related to domineering problems. In contrast, the NEO traits better cover specifically nonassertive problems; Modesty and -Assertiveness both related to problems within this octant and lacked CAT-PD or PID-5 counterparts.

Implications

The CAT-PD, PID-5, and NEO-PI-3FH trait measures of PD correspond relatively well at the domain level (e.g., Wright & Simms, 2014); however, the present study suggests potential facet-level differences in constructs. Such discrepancies in PD models could impede the ability to translate research findings across models. In addition, these models may yield nonequivalent clinical descriptions. For instance, in the PID-5 and CAT-PD trait measures, there are numerous ways to describe problematically dominant individuals, whereas the NEO facets provide fewer options. Conversely, the pathological trait models do not describe clients that struggle to assert themselves as well as the NEO.

These differences have considerable clinical importance when one considers traits in terms of social-cognitive processes (e.g., Fleeson & Jayawickreme, 2015). From this perspective, pathological personality traits describe interrelated, clinically relevant social-cognitive processes. Interpersonal theory provides a framework for connecting traits to processes by coordinating multiple levels of interpersonal phenomena (e.g., problematic behaviors, desired outcomes, sensitivities; Hopwood et al., 2011) across time (e.g., Locke & Sadler, 2007). For instance, grandiose individuals may be prone to interpret dominant behavior in others as cold or threatening, relative to less grandiose individuals (Roche, Pincus, Hyde, Conroy, & Ram, 2013). Explicating personality pathology in such terms may inform a variety of therapy approaches by suggesting interpersonal schemas (e.g., Safran, 1990), object-relations (e.g., Lukowitsky & Pincus, 2011), and avoidant coping strategies (e.g., Antony & Roemer, 2003) that may serve as treatment targets. The present results suggest that prominent PD models differ in their ability to provide such interpersonally relevant information about individuals. When choosing a PD measure or explaining clinical phenomena, clinicians should consider the extent to which PD models provide diverse ways to account for these problems.

It is then comforting that the PID-5 and CAT-PD facets converged in their relations with interpersonal problems. These models were constructed from the bottom-up by independent teams (Simms et al., 2011; Krueger et al., 2012); thus this agreement lends confidence to the mapping of lower-order PD traits. Despite this, some CAT-PD traits may present unique contributions to PD diagnosis. For instance, CAT-PD Emotional Detachment provides an additional way to describe individuals with Socially Avoidant problems and the traits Norm Violation and Rudeness do so similarly for Domineering problems. Beyond this, neither model covered warm or purely submissive problems very well. Future work may consider additional constructs (e.g., Self-Effacing; Gore, Presnall, Lynam, Miller & Widiger, 2012) to expand their interpersonal coverage.

Limitations and Future Directions

The present study benefitted from a large patient sample, broad representation of PD models, and a nuanced interpersonal model. An additional strength was the disaggregation of dysfunction severity from style. Despite these strengths, the present results should be evaluated in the context their limitations. First, generalizability may be limited to White and African American populations, as other ethnic groups largely were absent from our sample; however, personality-IIP relations generally have replicated across cultures in previous studies (Monsen et al., 2006; Simonsen & Simonsen, 2009). Additionally, generalizability may be slightly impacted given that male, African American participants with slightly more severe personality pathology were less likely to have completed all questions and thus were excluded. Although it is possible that that these excluded individuals would show different effects, we know of no research supporting such a concern. Second, all data were self-reported, whereas previous research has indicated the value of informant reports (e.g., Clifton, Turkheimer, & Oltmanns, 2005). In particular, some of the traits and problems investigated are highly “evaluative,” or undesirable, meaning that individuals may struggle to admit to or lack knowledge of these characteristics (e.g., Carlson, Vazire, & Oltmanns, 2013). Future work in this domain would benefit from examining the interpersonal impact of these PD models from the perspective of individuals who know the patients well.

There also were several measure-related limitations. In the present study, ASPD could not be examined in full, although conduct disorder symptoms related to interpersonal problems in a manner similar to previous findings with ASPD (e.g., Pincus & Wiggins, 1990). In addition, we used the NEO-PI-3FH to assess FFM facets. Some of these scales have poor internal consistency, potentially attenuating correlations and explaining the lower elevation parameters for NEO facets. Full NEO-PI-3 facet scales have better internal consistency; however, even the full NEO scales have generally lower alphas than PID-5 and CAT-PD-SF scales (McCrae, Costa, & Martin, 2005). An additional concern is that the NEO may be an inadequate representation of the FFM of PD (e.g., Widiger, 2011). Recently developed PD-specific faceted FFM measures assess maladaptive variants of FFM traits (e.g., Gore et al., 2012) and may provide broader coverage of personality pathology than the NEO; however, further work is needed to integrate the scales from these individual models into a broader descriptive model. For instance, a number of these measures contain pathological versions of the same facet, but with different names (Lynam, 2012). Future research is needed to integrate across these new measures and demonstrate their relations to interpersonal problems.

Conclusion

Past findings relating DSM-5 PD models to interpersonal problems were replicated and competing PD trait models converged in relations to interpersonal problems at the domain-level, as previous research would predict (e.g., Wright & Simms, 2014). Beyond this, the finding that pathological and normal-range trait models differ in their ability to describe individuals with domineering vs. nonassertive problems at the facet-level adds to our understanding of the lower-order structure of PD trait models. Importantly, these facet-level findings suggest that these models may have non-trivial differences in the constructs they include and, more generally, point to the importance of comparing models at the facet-level. Finally, across all models, few traits or PDs related to warm interpersonal problems, suggesting they are under-emphasized or less central to common PD features. Future work is needed to replicate the present findings and identify a comprehensive set of PD traits that reflect the full range of interpersonal problems seen in clinical practice.