Diagnostic Skill Application II

Original Article

Development of a Short Form of the Personality Styles and Disorder

Inventory (PSDI-6) Initial Validation in a Sample of Pregnant Women

Sarah Hain,1 Karin Schermelleh-Engel,1 Christine Freitag,2 Frank Louwen,3

and Silvia Oddo2,3

1Goethe University, Institute of Psychology, Department of Psychological Research Methods and Evaluation, Frankfurt/Main, Germany, 2Department of Child and Adolescent Psychiatry, Psychosomatic Medicine and Psychotherapy, University Hospital Frankfurt/Main, Germany,

3Division Obstetrics and Fetomaternal Medicine, University Hospital Frankfurt/Main, Germany

Abstract. The original Personality Styles and Disorder Inventory (PSDI) consists of 140 items capturing the non-pathological equivalents of 14 personality disorders (PDs) listed in DSM-III-R, DSM-IV-TR, and ICD-10. Because the DSM-5 proposes only six PDs in its research section and completing the PSDI is very time-consuming, a questionnaire capturing these six PDs with a smaller number of items would be desirable. Therefore, the short version PSDI-6 measuring each of the six PDs with six items was developed. Methods: Items were selected by means of exploratory and confirmatory factor analysis (CFA), using a calibration sample of nonclinical pregnant women. Measurement invariance of the PSDI-6 was demonstrated by CFA of multisample models in two independent samples. The PSDI-6 demonstrated acceptable internal consistency, factorial validity, and factorial invariance. Correlations with perfectionism, anxiety, and depression provided preliminary evidence of convergent and discriminant validity. Because our sample was limited to pregnant women, further studies should investigate the PSDI-6 in other clinical contexts.

Keywords: personality disorders, PSDI, short form, DSM-5, confirmatory factor analysis

Some Personality Styles (PSs) are known to be significant risk factors for mental illness (e.g., Caspi, Moffitt, Newman, & Silva, 1996; Skodol et al., 2005). PSs can be defined as accentuations of personality traits, which do not fulfill criteria of personality disorders (PDs; Kuhl & Kazén, 2009). PSs are often used in the field of differential psychology, whereas PDs are more restricted in clinical areas. The description of personality dimensions by means of PSs and PDs as in the Personality Styles and Disorders Inventory (PSDI; Kuhl & Kazén, 2009) allows for a sophis- ticated characterization of personality on a continuum from non-pathological personality traits/styles to clinical PDs.

The possibility to address also non-pathological person- ality characteristics facilitates the examination of risk fac- tors. A dimensional system provides greater benefit by enabling clinicians to differentiate between patients with severe PDs and those with less severe PDs or PSs and con- sequently to apply their interventions more effectively (Crawford, Koldobsky, Mulder, & Tyrer, 2011).

During the development process of the DSM-5 (American Psychiatric Association, 2013), the PDs Work Group proposed a hybrid dimensional-categorical model,

which retains six PD types. As the American Psychiatric Association (APA) evidenced excessive complexity for clinical practice, this new model was included in Section III of DSM-5, to encourage research that might support this model in the diagnosis and care of patients. The aim of the DSM-5 approach is to allow for a description of person- ality characteristics from the range of clinical PDs to per- sonality trait profiles of all patients (in contrast to the categorical DSM-IV system). Similarly, the ICD-11 (to be released in 2017; World Health Organization, 2015) Work- ing Group for the Revision of Classification of PDs devel- oped a two-step approach with the severity of personality disturbance as the first element and five personality domains as the second element (Tyrer, Crawford, & Mulder, 2011).

The PSDI is based on the dimensional approach of PDs, that is, the assumption that there is an analog PS for each clinical PD. It is a self-report inventory for the assessment of 14 PSs and PDs: paranoid, schizoid, schizotypal, border- line, histrionic, narcissistic, avoidant, dependent, obsessive- compulsive, negativistic, depressive, altruistic, rhapsodic, and antisocial.

� 2015 Hogrefe Publishing European Journal of Psychological Assessment 2016; Vol. 32(4):283–290 DOI: 10.1027/1015-5759/a000260

Th is

d oc

um en

t i s c

op yr

ig ht

ed b

y th

e A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. Th

is a

rti cl

e is

in te

nd ed

so le

ly fo

r t he

p er

so na

l u se

o f t

he in

di vi

du al

u se

r a nd

is n

ot to

b e

di ss

em in

at ed

b ro

ad ly

.

It consists of 140 items, each rated on a 4-point Likert- type rating scale ranging from 0 (= does not apply at all) to 3 (= fully applies). Kuhl and Kazén (2009) report satisfac- tory to good internal consistency, with Cronbach’s a ranging from .73 (negativistic PD) to .85 (schizotypal, borderline, obsessive-compulsive, and antisocial PD). Support for con- struct validity was demonstrated by correlations with the Big Five factors (Costa & McCrae, 1992), that is, signifi- cant correlations between neuroticism and PSDI borderline (.66), avoidant (.67), and depressive (.73) PD, as well as between extraversion and PSDI schizoid PD (�.57).

The PSDI has been utilized in diverse research contexts (e.g., Baumann, Kaschel, & Kuhl, 2007). Existing data reveal that PSs and PDs are relevant in terms of pathogen- esis, maintenance of, and therapy for mental disorders. Godemann, Linden, Neu, Heipp, and Dörr (2004) found that dependent and avoidant personality types partially explain the development of anxiety disorders following a critical life event. Delivery can be regarded as critical life event, and PSs and PDs can therefore be expected to dem- onstrate a relationship with anxiety after childbirth. Yet, the PSDI has never been used in relation to perinatal disorders, which have been shown to be influenced by personality dimensions. High perfectionism is associated with postpar- tum depression (Gelabert et al., 2012). Jones et al. (2010) observed higher levels of Neuroticism and Psychoticism (measured by Eysenck Personality Questionnaire) in 143 postnatal depressive mothers than in healthy mothers. Therefore we assume that personality dimensions might be a determinant of perinatal mental disorders. For this pur- pose, a sample of pregnant women was examined before and after childbirth.

Since the PSDI scales were developed from a phenom- enological instead of a factor-analytic approach, the facto- rial structure has not been examined systematically up to now. As Kuhl and Kazén (2009) report high correlations between the 14 PDs, the relationships between the variables may be more appropriately represented by a smaller num- ber of factors. This assumption is supported by their results of a principal component analysis, which revealed only four instead of 14 factors.

Moreover, the demand for short measurement instru- ments is growing. In clinical practice, the application of self-rating questionnaires is common practice, but only short screening instruments are practicable for routine application. Until now, there has been no economic self- rating measure available that is commonly used to screen for DSM- and ICD-PDs. For that reason, fewer items per PSDI scale would be desirable.

We had three aims in conducting the present study. First, we wanted to develop a short version of the PSDI, the PSDI-6, consisting of only six scales according to the DSM-5 research section. The PSDI-6 is supposed to mea- sure personality dimensions in their continuous manifesta- tion, which can be labeled as PDs1 in more severe forms.

Second, we aimed for a cross-validation of the factorial structure with an independent sample, and third, we intended to provide preliminary evidence of convergent and discriminant validity.

Materials and Methods

Participants

Pregnant women were recruited at the Buergerhospital and the University Hospital Frankfurt (Germany) as part of a research project on personality and postpartum depression. The questionnaires were handed out during an information evening for pregnant women and their partners. Participants were pregnant women measured at two time points, in the third trimester of pregnancy (N = 297) and again 6 weeks postpartum (N = 104). The mean ages ranged between 20 and 45 years (M = 32.35, SD = 4.46), and 88% indicated German as their mother tongue. All participants had a school-leaving qualification, 42% a university degree, and 9% a doctoral degree. Participation was voluntary. All par- ticipants gave informed consent. The Ethics Committee of the Goethe University Frankfurt am Main approved the study.

Instruments

In addition to the PSDI, the participants filled in question- naires at two time points measuring perfectionism during pregnancy (N = 297), and anxiety and depression 6 weeks postpartum (N = 104) for construct validation. We chose perfectionism, anxiety, and depression, because earlier research indicated associations of these constructs with per- sonality dimensions.

The MPS-F (Mehrdimensionale Perfektionismus-Skala von Frost et al.; Altstötter-Gleich & Bergemann, 2006) is the German version of the Multidimensional Perfectionism Scale (Frost, Marten, Lahart, & Rosenblate, 1990). With 35 items, it captures perfectionism as a multidimensional con- struct by means of six subscales (Cronbach’s a): high per- sonal standards (.84), preference for order and organization (.90), concern over making mistakes (.89), doubting of the quality of one’s actions (.70), perception of high parental expectations (.88), and perception of high parental criticism (.85).

The STADI (State-Trait Anxiety Depression Inventory; Laux, Hock, Bergner-Köther, Hodapp, & Renner, 2013) is a new development based on the STAI (State-Trait Anxiety Inventory; Spielberger, 1983). This 80-item self-report instrument aims at distinguishing between anxiety and depression, while differentiating between the temporary

1 In the following, we will use the term ‘‘PD’’ because the PSDI-6 scales shall potentially capture personality dimensions from their weak to severe forms.

284 S. Hain et al.: PSDI-6: A Short Version of the PSDI

European Journal of Psychological Assessment 2016; Vol. 32(4):283–290 � 2015 Hogrefe Publishing

Th is

d oc

um en

t i s c

op yr

ig ht

ed b

y th

e A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. Th

is a

rti cl

e is

in te

nd ed

so le

ly fo

r t he

p er

so na

l u se

o f t

he in

di vi

du al

u se

r a nd

is n

ot to

b e

di ss

em in

at ed

b ro

ad ly

.

(state) and longstanding (trait) conditions. Each state or trait is measured with 10 items, which are rated on a 4-point Likert-type scale. In the present study, the state scales were used. Internal consistency is high for both state anxiety (.90) and state depression (.87).

Statistical Analysis

The total sample for the first measurement point was divided into two samples of roughly the same size accord- ing to the chronological order of participation. The first sample (N1 = 149) was drawn in order to explore the statis- tical properties of the PSDI and to identify items for the PSDI-6. The second sample (N2 = 148) was used for invari- ance testing in order to cross-validate the factorial structure and parameter estimates across groups.

Item Selection

In order to assess the factorial structure of the PSDI with the calibration sample, items were first subjected to an exploratory factor analysis (EFA) with oblique rotation using the full information robust maximum likelihood (MLR) estimator of the Mplus program (Muthén & Muthén, 1998–2011). We tested whether all 60 items of the original 140-item version that were intended to measure the six PDs included in the DSM-5 research section loaded onto the respective six PD factors. For the development of the PSDI-6, we chose 36 items, which seemed to be most suitable. The selection was guided by the results of the EFA and by contentual as well as psychometric item properties. PSDI-6 Factors, original item numbering of the 140-Items PSDI, and German item content are listed in the Appendix.

Second, using CFA, we tested in the same sample whether the six-factorial structure of the PSDI-6 fitted the data well. As the distribution of the indicator variables devi- ated from multivariate normality, CFA was conducted using the MLR estimator of the Mplus program.

Multisample Analysis

For cross-validation, measurement invariance was evaluated by conducting a series of nested invariance tests using CFAs. We conducted six invariance tests that form a nested hierarchy of increasing levels of equality constraints imposed on factor structure, factor loadings, item inter- cepts, residual variances, and factor variances and covari- ances (cf. Gregorich, 2006).

Model Fit Evaluation

Model fit for EFA and CFA was evaluated using several goodness-of-fit statistics provided by the Mplus program, the likelihood-ratio v2 test and its associated p-value, the

root mean square error of approximation (RMSEA), the comparative fit index (CFI), and the standardized root mean square residual (SRMR). Good model fit was indicated by a nonsignificant v2 value, RMSEA � .06, CFI � .95, and SRMR � .08 (cf. Hu & Bentler, 1999; Schermelleh-Engel, Moosbrugger, & Müller, 2003). For nested model compar- isons, Satorra-Bentler (SB) scaled v2 difference tests were conducted (cf. Satorra & Bentler, 2001).

Missing Data

The amount of missing data was very small (< 2%), and simultaneously replaced when estimating model parame- ters, using the MLR estimator of the Mplus program.

Results

Item Selection

Exploratory Factor Analysis

An EFA using the calibration sample revealed that 16 eigenvalues were larger than 1.0. However, the scree plot suggested keeping only five factors, while the parallel anal- ysis indicated the retention of six factors. Model fit criteria pointed to conflicting conclusions about the extent to which the six-factor model matched the observed data, with an SB-v2 value of 2447.75 (df = 1,425, p < .01), RMSEA = .069, CFI = .695, and SRMR = .052.

The results of the EFA with oblique rotation revealed that only the factors reflecting schizotypal and obsessive- compulsive PDs yielded a clear simple structure. The items intended to measure antisocial PD loaded onto two separate factors. Borderline and avoidant PDs could not be distin- guished from one another, as they loaded onto a common factor, while narcissistic PD had high cross-loadings on sev- eral factors. Thus, it was not possible to distinguish clearly between borderline, avoidant, and narcissistic PD. As it was not clear whether six factors were actually an improvement over the five factor solution, we compared the model with six factors against the model with five factors using the SB-corrected v2 difference test. This test revealed a significant improvement of the six-factor solution (DSB-v2(Ddf ) = 86.03 (55 df ), p < .01). In the following, we will therefore use six factors.

Confirmatory Factor Analysis

Using again the calibration sample, six items per PD were selected by eliminating four items from each of the six factors: schizotypal, borderline, antisocial, narcissistic, avoidant, and obsessive-compulsive PD, respectively. The 36 items were chosen by simultaneously taking into account item content and factor loadings, as well as item

S. Hain et al.: PSDI-6: A Short Version of the PSDI 285

� 2015 Hogrefe Publishing European Journal of Psychological Assessment 2016; Vol. 32(4):283–290

Th is

d oc

um en

t i s c

op yr

ig ht

ed b

y th

e A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. Th

is a

rti cl

e is

in te

nd ed

so le

ly fo

r t he

p er

so na

l u se

o f t

he in

di vi

du al

u se

r a nd

is n

ot to

b e

di ss

em in

at ed

b ro

ad ly

.

intercorrelations. The results of a six-factor CFA revealed an acceptable fit (v2 = 887.55, df = 579, p < .01, RMSEA = .06, SRMR = .08, CFI = .82).2 The moderate CFI value (.82) may be explained by relatively low intercor- relations between several items. All 36 items loaded onto the respective PD factor with the expected sign. Table 1 lists the factor loadings for the 36 PSDI-6 items on their hypothesized dimensions and the original numbering of the items in the PSDI.

Internal consistency of the composite scores was esti- mated by McDonald’s omega (McDonald, 1999), because Cronbach’s a is based on the assumption of s-equivalent items, an assumption only rarely met in empirical data (Sijtsma, 2009). The consistency coefficients varied between .70 and .86.

Multisample Analysis

To cross-validate these findings, we tested invariance hypotheses in nested multisample analyses using both the calibration sample (N = 149) and the validation sam- ple (N = 148). The model fit criteria provided reason- ably strong evidence of dimensional invariance (Model 1), SB-v2(1,158 df) = 1,782.14, p < .01. The restrictions did not reduce the model fit, as all v2 difference tests were not significant. Thus, the strictest invariance test (structural invariance, Model 6) was confirmed. These results suggest that the same latent variable model underlies the PSDI-6 in both samples. Table 2 presents the model fit criteria for the invariance tests.

Construct Validation

In the construction strategy, we followed the ‘‘ABC of test construction’’ (Ziegler, 2014). We hypothesized that the PSDI-6 scales would positively correlate with measures of perfectionism, anxiety, and depression. Specifically, we expected significant relationships between borderline PD and avoidant PD with concern over mistakes, doubts about actions, anxiety, and depression, as well as between obses- sive-compulsive PD and organization.

The data of PSDI-6 and perfectionism were collected at the first time of measurement (N = 297); anxiety and depression were investigated using the second sample 6 weeks postpartum (N = 104) (see Table 3).

Regarding the relationships between the PSDI-6 factors, a relatively high correlation can be observed between bor- derline PD and avoidant PD (r = .68), while all other corre- lations are small or near zero. This high correlation could

be expected, as both scales of the original PSDI loaded onto a single factor in the EFA.

Correlations between the different instruments were found as follows: As expected, the most important associa- tions prevail between borderline PD and avoidant PD with the perfectionism scales concern over mistakes, doubts about actions, as well as postpartum anxiety and postpartum depression. The obsessive-compulsive PD correlated with the perfectionism scale organization. Interestingly, medium significant correlations were found between the perfection- ism scale personal standards and narcissistic PD and obses- sive-compulsive PD as well as between the perfectionism scale parental criticism and borderline PD and avoidant PD.

Discussion

The aim of our investigation was to develop a short form of the PSDI, the PSDI-6, capturing personality dimensions equivalent in their severe forms to the six PDs described in the DSM-5 (APA, 2013). Our results show that the PSDI-6 measures these personality dimensions in a highly efficient manner, while retaining sufficient reliability and validity. Albeit, PDs cannot yet be screened for with the PSDI-6. To that end, a sensitivity and specificity analysis on a clinical sample with known diagnoses of PDs would be necessary in a further study. By this means, cut-off values could be set and women at risk for perinatal mental disorders may be detected already during pregnancy.

The factorial structure of the PSDI-6 could be con- firmed for the population of pregnant women through a very strict examination based on increasingly restrictive invariance tests. Modification indices pointed to an improvement of model fit by freeing some error covari- ances. These error covariances may be explained by similar item wordings or content. For example, items ST-031 and ST-073 both use the word ‘‘sense’’ and are similar in content as they both refer to feeling something that is not present. These method effects should be investigated in more detail in future research.

Correlations between PSDI-6 scales are relatively small, with the exception of avoidant PD and borderline PD, as was expected. Their association might be explained by con- tent overlap, due to problems with self-identity, fear of criticism, and withdrawal from social relationships (cf. Zanarini et al., 2004).

Additionally, some initial indications of convergent validity were demonstrated. Borderline PD and avoidant PD correlated with all of the perfectionism scales (except organization), as well as with anxiety and depression.

2 Modification indices pointed to an improvement of model fit by freeing fixed parameters. Therefore, five error covariances, which indicated method effects because of equal item wordings or content, could be set free. As a result, the CFA with additional correlated residuals revealed an improved model fit in the calibration sample (v2 = 805.08, df = 574, p < .01, RMSEA = .05, SRMR = .08, CFI = .87) and the validation sample (v2 = 805.89, df = 574, p < .01, RMSEA = .05, SRMR = .08, CFI = .84). Yet, in the validation sample, one of the five error covariances did not reach significance.

286 S. Hain et al.: PSDI-6: A Short Version of the PSDI

European Journal of Psychological Assessment 2016; Vol. 32(4):283–290 � 2015 Hogrefe Publishing

Th is

d oc

um en

t i s c

op yr

ig ht

ed b

y th

e A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. Th

is a

rti cl

e is

in te

nd ed

so le

ly fo

r t he

p er

so na

l u se

o f t

he in

di vi

du al

u se

r a nd

is n

ot to

b e

di ss

em in

at ed

b ro

ad ly

.

The highest correlations existed between borderline PD and avoidant PD and the MPS-F scales doubts about actions and concern over mistakes, which correspond with disturbances in self-image, aims, and internal preferences, as well as apprehension and feelings of inferiority, respectively. This can in turn result in circular thought processes such as worry and rumination, which conforms to the cor- relation between anxiety and depression. Additionally, these

correlational associations are consistent with Gelabert et al. (2012), who showed that high perfectionism is associated with postpartum depression. Moreover, our findings con- firm Godemann et al. (2004), who found that an avoidant and insecure personality is associated with the development of anxiety after a critical life event. Narcissistic PD yields positive correlations with concern over mistakes, parental expectations, and personal standards, which can be

Table 1. PSDI-6 PD-factors and item numbering, item content, means (M), standard deviations (SD), and standardized factor loadings (k) for the calibration sample (N1 = 149)

PD Item Item content M SD k

Schizotypal ST-003 I often have sudden inspirations. 0.88 0.85 .65 ST-031 I believe that others sometimes sense my feelings, even

though they are elsewhere. 0.58 0.66 .55

ST-073 Sometimes I sense the presence of a faraway person so strongly, it is as if he/she were really present.

0.64 0.81 .67

ST-101 There is something like a sixth sense. 1.52 0.88 .72 ST-115 I believe in thought transference. 0.87 0.88 .81 ST-129 There are supernatural powers. 0.88 0.86 .69

Borderline BL-004 I often feel an inner emptiness. 0.42 0.61 .57 BL-046 My self-respect can switch abruptly between very positive

and very negative. 0.98 1.03 .81

BL-060 My feelings often change abruptly and impulsively. 0.86 0.90 .74 BL-074 Sometimes I have the feeling that my life could go

completely off course. 0.37 0.70 .66

BL-088 I often have groundless fears. 0.67 0.81 .69 BL-102 I often get angry about feelings, that I have. 1.02 0.80 .60

Narcissistic NA-006 As a child, I often had the feeling of being someone special. 1.01 0.92 .27 NA-034 When I make suggestions on the job, I expect them to be

accepted without question. 0.79 0.81 .49

NA-048 I have a pronounced feeling for what is special. 1.43 0.82 .54 NA-062 The thought of being a famous person excites me. 0.49 0.68 .66 NA-104 I don’t dream of great successes. (R) 1.77 0.93 .40 NA-118 Being the center of attention really appeals to me. 0.86 0.81 .75

Avoidant AV-021 I don’t reveal all aspects of myself, because I am afraid that I would lose the sympathy of some people.

0.65 0.78 .47

AV-035 Criticism hurts me quicker than it does others. 1.07 0.90 .57 AV-063 When I become aware of a weakness, it can worry me for a

long time. 1.37 0.87 .64

AV-077 When I feel observed, I become anxious. 0.85 0.82 .53 AV-091 I can cope well with some people not liking me. (R) 1.61 0.89 .46 AV-133 I often have pangs of conscience. 0.85 0.85 .64

Obsessive-Compulsive OC-009 I am a person with fixed habits. 1.60 0.82 .47 OC-023 Even under time pressure, I cannot stop being thorough. 1.30 0.99 .56 OC-037 Consistency and firm principles rule my life. 1.64 0.74 .64 OC-107 I am a conscientious person. 2.31 0.64 .54 OC-121 Precision and order are very important to me. 1.84 0.85 .66 OC-135 I cannot stand it when others don’t do their duty. 1.88 0.74 .67

Antisocial AS-042 If someone behaves dismissively toward me, I really get back at them.

0.67 0.76 .87

AS-070 If people turn against me I really get back at them. 0.64 0.79 .84 AS-084 I prefer to attack, rather than letting others attack me. 0.85 0.85 .57 AS-098 If others cause troubles for me, I can make things very

uncomfortable for them. 1.16 0.81 .73

AS-126 I don’t put up with any nonsense from others. 1.41 0.74 .54 AS-140 People who want to harm me can count on retaliation. 0.83 0.80 .70

Notes. (R) = reverse coded. Preliminary translation of the German items.

S. Hain et al.: PSDI-6: A Short Version of the PSDI 287

� 2015 Hogrefe Publishing European Journal of Psychological Assessment 2016; Vol. 32(4):283–290

Th is

d oc

um en

t i s c

op yr

ig ht

ed b

y th

e A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. Th

is a

rti cl

e is

in te

nd ed

so le

ly fo

r t he

p er

so na

l u se

o f t

he in

di vi

du al

u se

r a nd

is n

ot to

b e

di ss

em in

at ed

b ro

ad ly

.

explained by such characteristics as a demanding attitude, exaggerated self-esteem, and the expectation of attracting interest as someone special. Unsurprisingly, obsessive-com- pulsive PD shows meaningful correlations with the perfec- tionism scales personal standards, organization, and concern over mistakes, as perfectionism itself is one of the diagnostic criteria for this PD. Lastly, the moderate po- sitive correlation between antisocial PD and personal stan- dards contains a reference to disregard for social norms, incapacity to maintain enduring relationships, and low tol- erance of frustration. Discriminant validity was demon- strated by low or nonexistent correlations, for example, schizotypal PD with perfectionism, or narcissistic PD, obsessive-compulsive PD, and antisocial PD with anxiety and depression.

Because PSs and PDs can influence the severity of acute symptom disorders and because they affect therapeutic interactions, therapist awareness of PDs is relevant to the conduct and success of therapy. Therefore, dimensional diagnostics are important, because they enable practitioners to apply their interventions more effectively (Crawford et al., 2011). This is in accordance with the DSM-5 that shares the aim of measuring the range of clinical PDs as

well as nonclinical PDs. Most notably, the PSDI-6 may prove to be valuable, as the PDs measured with this inven- tory refer to differentiated pathological dimensions used in clinical routine that facilitate the deduction of therapeutical approaches, in contrast to the more general personality traits (e.g., the Big Five).

Conclusion

Our results show that the PSDI-6 measures the personality dimensions included in the DSM-5 in a highly efficient manner. The PSDI-6 could be particularly useful in the pre- vention and therapy of postpartum depression. Together with the application of standard screening instruments, the PSDI-6 may be helpful in identifying women already at risk during pregnancy, so that an appropriate therapy can be initiated as early as possible.

A limitation of the present study is that the generaliz- ability of our results is somewhat reduced, because the sample was limited to pregnant women. Further studies should thus investigate the PSDI-6 in other clinical contexts as well.

Table 2. Model fit summary for factorial invariance tests across two random samples

Model Invariance SB-v2 df DSB-v2 Ddf RMSEA CFI SRMR

1 Dimensional 1,782.14 1,158 .060 .805 .081 2 Configural 1,795.37 1,188 17.62ns 30 .059 .810 .082 3 Metric 1,815.17 1,218 19.09ns 30 .057 .813 .082 4 Strong factorial 1,842.06 1,254 29.06ns 36 .056 .816 .083 5 Strict factorial 1,852.98 1,260 11.01ns 6 .056 .814 .083 6 Structural 1,881.25 1,281 28.43ns 21 .056 .812 .088

Notes. SB-v2 = Satorra-Bentler scaled v2, DSB-v2 = SB-v2 difference; ns = not significant, RMSEA = root mean square error of approximation, CFI = comparative fit index, SRMR = standardized root mean square residual.

Table 3. Correlations between PSDI-6 factors and measures of perfectionism (MPS-F), anxiety, and depression (STADI), as well as internal consistencies (McDonald’s omega) of the PSDI-6 scales

1 2 3 4 5 6

PSDI-6a Schizotypal PD 1 .82 Borderline PD 2 .12* .81 Narcissistic PD 3 .16** .05 .67 Avoidant PD 4 �.003 .68** .12* .73 Obsessive-compulsive PD 5 .05 .11 .01 .23** .78 Antisocial PD 6 .07 .04 .28** .02 .15** .84

MPS-Fa Personal standards 7 �.002 .19** .34** .29** .39** .26** Organization 8 .06 �.04 �.04 .07 .66** .07 Concern over mistakes 9 �.06 .52** .14* .61** .24** .09 Doubts about actions 10 .01 .60** �.10 .53** .09 �.12* Parental expectations 11 .09 .26** .16** .27** .13* .12* Parental criticism 12 .07 .40** �.03 .32** .15* .08

STADIb Anxiety 13 �.05 .37** �.01 .32** .004 �.13 Depression 14 .13 .20* .07 .20* .11 �.10

Notes. aN = 297, bN = 104; **p < .01, *p < .05, internal consistencies along the diagonal are in italics.

288 S. Hain et al.: PSDI-6: A Short Version of the PSDI

European Journal of Psychological Assessment 2016; Vol. 32(4):283–290 � 2015 Hogrefe Publishing

Th is

d oc

um en

t i s c

op yr

ig ht

ed b

y th

e A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. Th

is a

rti cl

e is

in te

nd ed

so le

ly fo

r t he

p er

so na

l u se

o f t

he in

di vi

du al

u se

r a nd

is n

ot to

b e

di ss

em in

at ed

b ro

ad ly

.

Acknowledgment

The authors are grateful to Dr. Brian Bloch for his compre- hensive editing of the manuscript.

References

Altstötter-Gleich, C., & Bergemann, N. (2006). Testgüte einer deutschsprachigen Version der Mehrdimensionalen Perfek- tionismus Skala von Frost, Marten, Lahart und Rosenblate (MPS-F) [The statistics and validity of German version of a Multidimensional Perfectionism Scale of Frost, Marten, Lahart, and Rosenblate (MPS-F)]. Diagnostica, 52, 105–118.

American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (5th ed.). Washington, DC: Author.

Baumann, N., Kaschel, R., & Kuhl, J. (2007). Affect sensitivity and affect regulation in dealing with positive and negative affect. Journal of Research in Personality, 41, 239–248.

Caspi, A., Moffitt, T. E., Newman, D. L., & Silva, P. A. (1996). Behavioral observations at age 3 years predict adult psychi- atric disorders: Longitudinal evidence from a birth cohort. Archives of General Psychiatry, 53, 1033–1039.

Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory (NEO-PI-R) and NEO Five-Factor Inventory (NEO-FFI): Professional manual. Odessa, FL: Psycholog- ical Assessment Resources.

Crawford, M. J., Koldobsky, N., Mulder, R., & Tyrer, P. (2011). Classifying personality disorder according to severity. Jour- nal of Personality Disorders, 25, 321–330.

Frost, R. O., Marten, P., Lahart, C., & Rosenblate, R. (1990). The dimensions of perfectionism. Cognitive Therapy and Research, 14, 449–468.

Gelabert, E., Subirà, S., García-Esteve, L., Navarro, P., Plaza, A., Cuyàs, E., . . . Martín-Santos, R. (2012). Perfectionism dimensions in major postpartum depression. Journal of Affective Disorders, 136, 17–25.

Godemann, F., Linden, M., Neu, P., Heipp, E., & Dörr, P. (2004). A prospective study on the course of anxiety after vestibular neuronitis. Journal of Psychosomatic Research, 56, 351–354.

Gregorich, S. E. (2006). Do self-report instruments allow meaningful comparisons across diverse population groups? Testing measurement invariance using the confirmatory factor analysis framework. Medical Care, 44, 78–94.

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multi- disciplinary Journal, 6, 1–55.

Jones, L., Scott, J., Cooper, C., Forty, L., Smith, K., Sham, P., . . . Jones, I. (2010). Cognitive style, personality and vulnera- bility to postnatal depression. The British Journal of Psychiatry, 196, 200–205.

Kuhl, J., & Kazén, M. (2009). Persönlichkeits-Stil- und Störungs-Inventar (PSSI) [Personality Styles and Disorder Inventory (PSDI)]. Göttingen, Germany: Hogrefe.

Laux, L., Hock, M., Bergner-Köther, R., Hodapp, V., & Renner, K.-H. (2013). Das State-Trait-Angst-Depressions-Inventar (STADI) [The State-Trait Anxiety Depression Inventory (STADI)]. Göttingen, Germany: Hogrefe.

McDonald, R. P. (1999). Test theory: A unified treatment. Mahwah, NJ: Erlbaum.

Muthén, L. K., & Muthén, B. O. (1998–2011). Mplus User’s Guide (6th ed.). Los Angeles, CA: Muthén & Muthén.

Satorra, A., & Bentler, P. M. (2001). A scaled difference chi- square test statistic for moment structure analysis. Psycho- metrika, 66, 507–514.

Schermelleh-Engel, K., Moosbrugger, H., & Müller, H. (2003). Evaluating the fit of structural equation models: Tests of significance and descriptive goodness-of-fit measures. Methods of Psychological Research Online, 8, 23–74.

Sijtsma, K. (2009). On the use, the misuse, and the very limited usefulness of Cronbach’s alpha. Psychometrika, 74, 107–120.

Skodol, A. E., Gunderson, J. G., Shea, M. T., McGlashan, T. H., Morey, L. C., Sanislow, C. A., . . . Stout, R. L. (2005). The Collaborative Longitudinal Personality Disorders Study (CLPS): Overview and implications. Journal of Personality Disorders, 19, 487–504.

Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory STAI (Form Y). Palo Alto, CA: Mind Garden.

Tyrer, P., Crawford, M., & Mulder, R. (2011). Reclassifying personality disorders. The Lancet, 377, 1814–1815.

World Health Organization. (2015). The International Classifi- cation of Diseases 11th Revision is due by 2017. Retrieved from http://www.who.int/classifications/icd/revision/en/

Zanarini, M. C., Frankenburg, F. R., Vujanovic, A. A., Hennen, J., Reich, D. B., & Silk, K. R. (2004). Axis II comorbidity of borderline personality disorder: Description of 6-year course and prediction to time-to-remission. Acta Psychiatrica Scandinavica, 110, 416–420.

Ziegler, M. (2014). Stop and state your intentions! Let’s not forget the ABC of test construction. European Journal of Psychological Assessment, 30, 239–242. doi: 10.1027/1015- 5759/a000228

Date of acceptance: December 18, 2014 Published online: June 26, 2015

Sarah Hain

Goethe University Institute of Psychology Department of Psychological Research Methods and Evaluation PEG, P.O. Box 75 Theodor-W.-Adorno-Platz 6 60629 Frankfurt/Main Germany Tel. +19 177 738-1404 E-mail [email protected]

S. Hain et al.: PSDI-6: A Short Version of the PSDI 289

� 2015 Hogrefe Publishing European Journal of Psychological Assessment 2016; Vol. 32(4):283–290

Th is

d oc

um en

t i s c

op yr

ig ht

ed b

y th

e A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. Th

is a

rti cl

e is

in te

nd ed

so le

ly fo

r t he

p er

so na

l u se

o f t

he in

di vi

du al

u se

r a nd

is n

ot to

b e

di ss

em in

at ed

b ro

ad ly

.

Appendix

PSDI-6 Factors, Original Item Numbering of the 140-Items PSDI, and German Item Content

PD Item German item content

Schizotypal 3 Ich habe schon öfter Eingebungen gehabt. 31 Ich glaube, dass andere manchmal meine Gefühle spüren, auch wenn sie

sich anderswo aufhalten. 73 Manchmal spüre ich die Anwesenheit einer fernen Person so stark, als

wäre sie wirklich da. 101 Es gibt so etwas wie den 6. Sinn. 115 Ich glaube an Gedankenübertragung. 129 Es gibt übernatürliche Kräfte.

Borderline 4 Ich spüre oft eine innerliche Leere. 46 Meine Selbstachtung kann abrupt zwischen sehr positiven und sehr

negativen Empfindungen wechseln. 60 Meine Gefühle wechseln oft abrupt und impulsiv. 74 Manchmal habe ich das Gefühl, als könnte mein Leben völlig aus der Bahn

geraten. 88 Ich habe oft ohne Grund ängstliche Gefühle.

102 Ich ärgere mich oft über Gefühle, die in mir aufkommen. Narcissistic 6 Ich habe als Kind oft das Gefühl gehabt, etwas Besonderes zu sein.

34 Wenn ich im Beruf Ideen einbringe, erwarte ich, dass sie von anderen anstandslos akzeptiert werden.

48 Ich habe ein ausgeprägtes Gefühl für das Besondere. 62 Der Gedanke, eine berühmte Persönlichkeit zu sein, reizt mich.

104 Ich träume nicht von großen Erfolgen. (R) 118 Im Mittelpunkt zu stehen, hat für mich einen besonderen Reiz.

Avoidant 21 Viele Seiten von mir zeige ich nicht, weil ich befürchte, dass ich die Sympathie mancher Menschen verlieren würde.

35 Kritik tut mir schneller weh als anderen. 63 Wenn mir eine Schwäche bewusst wird, kann mich das eine ganze Zeit

belasten. 77 Wenn ich mich beobachtet fühle, werde ich ängstlich. 91 Ich kann es gut verkraften, dass manche Menschen mich nicht mögen. (R)

133 Ich habe oft Gewissensbisse. Obsessive-Compulsive 9 Ich bin ein Mensch mit festen Gewohnheiten.

23 Meine Gründlichkeit kann ich auch dann nicht ablegen, wenn ich unter Zeitdruck stehe.

37 Beständigkeit und feste Grundsätze bestimmen mein Leben. 107 Ich bin ein gewissenhafter Mensch. 121 Genauigkeit und Ordnung sind mir sehr wichtig. 135 Ich kann es nicht leiden, wenn andere ihren Pflichten nicht nachkommen.

Antisocial 42 Wenn sich jemand mir gegenüber ablehnend verhält, kann ich ihn ohne weiteres fertigmachen.

70 Wenn Leute sich gegen mich wenden, kann ich sie fertigmachen. 84 Ich greife lieber an, als mich angreifen zu lassen. 98 Wenn andere mir Schwierigkeiten machen, kann ich sehr ungemütlich

werden. 126 Ich lasse mir von anderen nichts gefallen. 140 Wer mir schaden will, muss mit einer Vergeltung rechnen.

Note. (R) = reverse coded.

290 S. Hain et al.: PSDI-6: A Short Version of the PSDI

European Journal of Psychological Assessment 2016; Vol. 32(4):283–290 � 2015 Hogrefe Publishing

Th is

d oc

um en

t i s c

op yr

ig ht

ed b

y th

e A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

tio n

or o

ne o

f i ts

a lli

ed p

ub lis

he rs

. Th

is a

rti cl

e is

in te

nd ed

so le

ly fo

r t he

p er

so na

l u se

o f t

he in

di vi

du al

u se

r a nd

is n

ot to

b e

di ss

em in

at ed

b ro

ad ly

.