Vrije Universiteit Brussel

Psychometric validation of the Childhood Trauma Questionnaire-Short Form (CTQ-SF) in a

Danish clinical sample
Kongerslev, M.T.; Bach, B. ; Rossi, Gina; Trauelsen, M.; Ladegaard, N.; Lokkegaard, S.; Bo,
S.
Published in:
Child Abuse & Neglect

DOI:
10.1016/j.chiabu.2019.104026

Publication date:
2019

Document Version:
Accepted author manuscript

Link to publication

Citation for published version (APA):

Kongerslev, M. T., Bach, B., Rossi, G., Trauelsen, M., Ladegaard, N., Lokkegaard, S., & Bo, S. (2019).
Psychometric validation of the Childhood Trauma Questionnaire-Short Form (CTQ-SF) in a Danish clinical
sample. Child Abuse & Neglect, 94, [104026]. https://doi.org/10.1016/j.chiabu.2019.104026

Copyright
No part of this publication may be reproduced or transmitted in any form, without the prior written permission of the author(s) or other rights
holders to whom publication rights have been transferred, unless permitted by a license attached to the publication (a Creative Commons
license or other), or unless exceptions to copyright law apply.

Take down policy

If you believe that this document infringes your copyright or other rights, please contact openaccess@vub.be, with details of the nature of the
infringement. We will investigate the claim and if justified, we will take the appropriate steps.

Download date: 26. Sep. 2023

Psychometric validation of the Childhood Trauma Questionnaire-Short Form (CTQ-SF) in a

Danish clinical sample

Mickey T. Kongersleva,b,*, Bo Bacha, Gina Rossic, Anne M. Trauelsend, Nicolai Ladegaarde, Sille S.

Løkkegaardf, & Sune Boa

a
Psychiatric Research Unit, Faelledvej 6, 4200 Slagelse, Region Zealand, Denmark.
b
Department of Psychology, University of Southern Denmark, Campusvej 55, 5230 Odense M,

Denmark.
c
Faculty of Psychology & Educational Sciences, Personality and Psychopathology Research Group,

Vrije Universiteit Brussel (VUB), Pleinlaan 2, 1050 Brussels, Belgium.

d
Psychotherapeutic Ambulatory, Mental Health Center Amager, Digevej 10, Capital Region, 2300

Copenhagen, Denmark.
e
Department of Affective Disorders, Aarhus University Hospital – Psychiatry, Palle Juul-Jensens

Boulevard 175, 8200 Aarhus N, Denmark.

f
Danish National Center for Psychotraumatology, Department of Psychology, University of

Southern Denmark, Campusvej 55, 5230 Odense M, Denmark.

*
Corresponding author at: Psychiatric Research Unit, Faelledvej 6, 4200 Slagelse, Region Zealand,

Denmark.

E-mail addresses: mkon@regionsjaelland.dk (M. T. Kongerslev), bbpn@regionsjaelland.dk (B.

Bach), Gina.Rossi@vub.be (G. Rossi), marietrauelsen@gmail.com (A. M. Trauelsen),

nicolai.ladegaard@ps.rm.dk (N. Ladegaard), sschandorph@health.sdu.dk (S. S. Løkkegaard),

subh@regionsjaelland.dk (S. Bo).

1
Funding: This research did not receive any specific grant from funding agencies in the public,

commercial, or not-for-profit sector.

Summary declaration of interest statement: The authors declare they have no interests to declare.

2
Psychometric validation of the Childhood Trauma Questionnaire-Short Form (CTQ-SF) in a

Danish clinical sample

ABSTRACT

BACKGROUND

The Childhood Trauma Questionnaire – Short Form (CTQ-SF) is a widely utilized self-report

instrument in the assessment and characterization of childhood trauma. Yet, research on the

instrument’s psychometric properties in clinical samples is sparse, and the Danish version of the

CTQ-SF has not been previously evaluated in clinical samples.

OBJECTIVES

To examine the structural validity, internal consistency reliability, and multi-method convergent

validity of the CTQ-SF in a heterogenous clinical sample from Denmark.

PARTICIPANTS AND SETTING

The study was based on data from four Danish clinical samples (N = 393): 1) Outpatients diagnosed

with personality disorders, 2) Patients commencing psychiatric treatment for non-affective first-

episode psychosis, 3) Patients diagnosed with first-episode or prolonged depression recruited from

general practitioners and an outpatient mood disorder clinic, and 4) detained delinquent boys.

METHODS

Confirmatory factor analysis was used to explore structural validity. Also, we calculated internal

consistency and multi-method convergent validity with interview-based ratings of adverse

parenting.

RESULTS

Confirmatory factor analyses indicated that the five-factor structure described in CTQ-SF manual

with three error correlated items best fitted the data, as compared to various other models.

3
Coefficients of congruence also supported factorial similarity across countries (i.e. US substance

abuser and a mixed Brazilian sample). Internal consistency reliability was acceptable and

comparable to estimates previously published. Multi-method convergent validity associations

further corroborated the validity of the CTQ-SF.

CONCLUSION

These findings provide support for the reliability and validity of the Danish version of the CTQ-SF

in clinical samples.

KEYWORDS

Adverse childhood experiences; CTQ; Danish; Developmental trauma; Reliability; Validity

4
Background

According to the World Health Organization (WHO) 23% of children worldwide

report that they have suffered from physical abuse, 36% from emotional abuse, 16% from physical

neglect, and 18% of girls and 8% of boys from sexual abuse in the past year (World Health

Organization, 2017). Thus, developmental trauma in the form of child maltreatment, comprising

various forms of abuse and neglect, is widespread worldwide (Gilbert et al., 2009; Moody,

Cannings-John, Hood, Kemp, & Robling, 2018; Stoltenborgh, Bakermans-Kranenburg, Alink, &

van Ijzendoorn, 2015). It poses a serious public health problem, given the fact that childhood

trauma is a robust and powerful pluripotent risk factor for both concurrent and future detrimental

developmental outcomes, including poor mental and physical health (Cicchetti, 2016; Karterud &

Kongerslev, 2019; Kessler et al., 2010; McCrory, Gerin, & Viding, 2017; Vachon, Krueger,

Rogosch, & Cicchetti, 2015; World Health Organization, 2017; Zeanah & Humphreys, 2018).

Furthermore, almost all common mental disorders across the lifespan have been shown to be

strongly associated with various forms of child maltreatment. Childhood adversity may thus be the

single greatest known environmental predictor of transdiagnostic psychiatric problems (Green et al.,

2010; Kessler et al., 2010; Zeanah & Humphreys, 2018). The strong link betwixt child maltreatment

and mental disorders is also underscored by research documenting the negative effects of childhood

trauma on presentation, severity, course, and treatment response in adolescent and adult patients

diagnosed with for example depression (Williams, Debattista, Duchemin, Schatzberg, & Nemeroff,

2016), bipolar disorder (Cakir, Tasdelen Durak, Ozyildirim, Ince, & Sar, 2015; Etain et al., 2013),

borderline personality disorder (Bo & Kongerslev, 2017; Bo, Sharp, Fonagy, & Kongerslev, 2017;

Levey, Apter, & Harrison, 2016; Zanarini, Frankenburg, Hennen, Reich, & Silk, 2006), and

psychosis (Schäfer & Fisher, 2011). Importantly, childhood trauma also increases the risk for

5
suicide and suicide attempts considerably (Angelakis, Gillespie, & Panagioti, 2019; Gerdner &

Allgulander, 2009; Zatti et al., 2017) – for example a recent systematic review reported a two- to

threefold increased risk of suicide attempts for all types of child maltreatment (Angelakis et al.,

2019). Consequently, assessment of childhood abuse and neglect is important in routine clinical

practice, in order to inform treatment planning and assess risk of suicide and prognosis.

Unfortunately, however, childhood trauma oftentimes goes unrecognized in clinical

settings, due in part to the sensitivity of the subject which may make some patients or clinicians

reluctant to talk about it during a face-to-face interview (Read, Hammersley, & Rudegeair, 2018;

Zeanah & Humphreys, 2018). A brief self-report measure could therefore be useful to facilitate

systematic assessment of childhood maltreatment. Such a measure would also be time and cost

efficient making it possible to screen many patients, and it can be filled out in private, which makes

it easier for patients to disclose information on this highly sensitive topic.

Among the available instruments for the retrospective assessment of childhood

maltreatment (Dovran, Winje, Arefjord, & Haugland, 2012), the Childhood Trauma Questionnaire-

Short Form (CTQ-SF) (Bernstein & Fink, 1998; Bernstein et al., 2003) is internationally one of the

most widely used (Baker & Maiorino, 2010; Grassi-Oliveira et al., 2014). The CTQ-SF is a self-

report instrument suitable for use with adolescents and adults to screen for five types of

maltreatment: physical, emotional, and sexual abuse, and physical and emotional neglect, as well as

including a total scale score indicating the global level of childhood trauma. On the whole, the

CTQ-SF appears to meet the general requirements for a brief trauma screening instrument (Brewin,

2005), including satisfactory reliability and validity (Baker & Maiorino, 2010). Yet, results from

studies on the five-factor structure reported in the original manual (Bernstein & Fink, 1998) have

been mixed, with some studies supporting it whilst others only obtain partial support or suggest

alternative models. A number of studies, based on both clinical and non-clinical samples, and

6
including both adolescents and adults, predominantly from North America and Europe, have

obtained support for the five-factor structure reported in the CTQ-SF manual (Bernstein et al.,

2003; Dovran et al., 2013; Dudeck et al., 2015; Hernandez et al., 2013; Sacchi, Vieno, & Simonelli,

2018; Thombs, Bernstein, Lobbestael, & Arntz, 2009; Thombs, Lewis, Bernstein, Medrano, &

Hatch, 2007). Moreover, some of these studies also tested and demonstrated the structural

invariance or measurement equivalence of the original CTQ-SF five-factor structure across gender,

age and subsamples (Bernstein et al., 2003; Dovran et al., 2013; Thombs et al., 2007). Yet, other

studies, based on both clinical and non-clinical samples including adolescents and adults from for

example Sweden, Brazil, and Korea, have failed (or at least partially failed) to replicate the original

CTQ-SF five-factor structure (Gerdner & Allgulander, 2009; Grassi-Oliveira et al., 2014; Villano et

al., 2004). Particularly, the Physical Neglect scale has been found to be problematic in these studies,

and also appears more generally to have the poorest internal consistency of all the five scales in

most previous studies (Gerdner & Allgulander, 2009; Grassi-Oliveira et al., 2014). Particularly the

findings from studies in clinical samples suggest an alternative five-factor structure where items

number 2 and 26 loads onto the Emotional Neglect scale, rather than on the Physical Neglect scale

as would otherwise have been expected according to the CTQ-SF manual. This alternative CTQ-SF

five-factor was found to be the most appropriate when compared to the original factor solution in a

mixed Brazilian sample when the total CTQ-SF score was also included at the second order level

(Grassi-Oliveira et al., 2014). Another study, based on an Italian college sample, though finding that

the five-factor structure specified in the CTQ-SF manual provided best fit, also found that a four-

factor first order structure (wherein items from the Physical and Emotional neglect scales were

collapsed into one single Neglect scale) provided a good fit (Sacchi et al., 2018). Given these

indecisive research findings, more research examining the factorial structure in clinical samples is

important and needed.

7
 The Danish version of the CTQ-SF (Bernstein & Fink, 2011) has already been used in

a number of clinical studies to explore childhood trauma in relation to psychopathology (e.g., Bach

& Fjeldsted, 2017; Trauelsen et al., 2015). However, whilst the original CTQ-SF manual contains

information on the factor structure and item loadings in adolescent and adult clinical samples, the

Danish manual only contains psychometric information based on a non-clinical Danish convenience

sample. Thus, information on the psychometric performance of the Danish CTQ-SF in clinical

samples is lacking.

In summary, though internationally widely used and generally displaying adequate

psychometric properties, research on the CTQ-SF factor structure displays conflicting results.

Additionally, the instrument has never been psychometrically tested in Danish clinical sample.

Hence, the present study aimed to evaluate the psychometric properties of the Danish CTQ-SF in a

clinical sample. This involved: i) testing the instruments factor structure through comparing the fit

of the original second order five-factor model to that of competing one-, four- and five-factor

models reported in the literature as well as to a one factor model; ii) estimating the factorial

similarity for the best fitting factor model, based on factor loadings, with other international studies;

iii) assessing scale reliability; and iv) estimating convergent validity with an interview-based

measure of problematic parenting behavior, namely the Revised Childhood Experiences

Questionnaire (Zanarini, 1992).

Methods and materials

Data and procedures

Archival material from four clinical research studies in Denmark was used. Each of

these previous studies was conducted in accordance with the Declaration of Helsinki, approved by

the respective local ethics committees in Denmark, and reported to the Danish Data Protection

8
Agency. Written informed consent was obtained from all participants in the original studies. For the

present study we only had access to fully anonymized data. Apart from administration of the CTQ-

SF all participants were also tested with structured diagnostic interviews conducted by trained

clinicians. Adequate interrater-reliability for these ratings has been documented (see the original

study references below). The four clinical samples, from which our final sample was derived, can

briefly be summarized as follows:

1) Patients consecutively enrolled, as part of routine clinical assessment, from March 2012 to

June 2014 at a psychiatric outpatient clinic specialized in treatment of personality disorders

(n = 142; 68% women; Mage = 29 years, SD = 8.4). Subjects fulfilling criteria for current

psychotic disorder, current manic episode, autism or organic disorders, or organic induced

disorders were excluded. The majority of the included patients (71%) met criteria for

Borderline Personality Disorder (Bach & Sellbom, 2016). Other types of specific personality

disorders as well as various mental state disorders also occurred frequently (Bach,

Simonsen, Christoffersen, & Kriston, 2017). Moreover, in this sample, 72 randomly selected

participants were also administered an interview of parental behavior in childhood

(Zanarini, 1992). These interviews were conducted and scored by a clinical psychologist

blinded to the participants self-reported responses on the CTQ-SF. We used these data to

test the convergent validity associations between the CTQ-SF clinical scales and interview-

based ratings of childhood trauma. Moreover, 19 randomly selected interviews on parenting

behavior was also rated by another clinical psychologist to obtain estimates on interrater

reliability for this instrument.

2) Patients commencing psychiatric treatment for non-affective first-episode psychosis in

Region Zealand in Denmark (n = 101; 26% women; Mage = 23 years, SD = 3.4; Trauelsen et

9
 al., 2015). The patients were recruited consecutively over a two-year period from April 2011

to April 2013. Inclusion criteria for this study was meeting criteria for a diagnosis of non-

affective psychosis and being 18 to 35 years of age. Exclusion criteria were a previous

diagnosis of psychosis.

3) Patients diagnosed with first-episode or prolonged depression recruited from general

practitioners and an outpatient mood disorder clinic in the Central Denmark Region (n = 71;

78% women; Mage = 35 years, SD = 11.6; Ladegaard, Lysaker, Larsen, & Videbech, 2014).

The data was collected from December 2010 to December 2012. The first-episode depressed

patients (n = 44) met DSM-IV criteria for major depressive disorder, and were all

psychotropic drug-naive. The patients with prolonged depression (n = 27) were in- and

outpatients required to meet full DSM-IV criteria for major depressive disorder for a period

of minimum two years, and was additionally required to have failed to respond to two or

more pharmacotherapy treatments with antidepressants of different classes. All the recruited

patients were also required to have depression as their primary diagnosis, and a depressive

symptom severity of moderate to severe when enrolled. Current substance use disorder,

neurological illness, head trauma and chronic somatic disease were exclusion criteria.

4) Juvenile delinquent boys sampled from three secure institutions and a prison ward in

Denmark (n = 80; Mage = 17 years, SD = 0.8) during August 2010 to October 2011.

Inclusion criteria were male gender, age from 15 to 18 years, remanded or sentenced, and

willing and able to give informed consent. Exclusion criteria were profound mental

retardation, under the influence of alcohol or drugs or being psychotic on days of

assessments. The most common mental disorders in this sample were conduct disorder

(76%), personality disorders (65%), alcohol and substance abuse (58%), Attention-

Deficit/Hyperactivity Disorder (ADHD; 23%), anxiety disorders (18%) and mood disorders

10
 (8%; Gillespie, Kongerslev, Sharp, Bo, & Abu.Akel, 2018; Kongerslev, Moran, Bo, &

Simonsen, 2012; Kongerslev, Bo, Forth, & Simonsen, 2015).

We excluded one woman from sample 2 (i.e., patients diagnosed with first-episode psychosis)

because of completely missing data on all CTQ-SF items. The resulting combined Danish clinical

sample was diagnostically heterogeneous and comprised 393 respondents of which 45% (n = 177)

were women. Age ranged from 15 to 63 (M = 26 years, SD = 9.6). Of the 393 respondents included

in the present study, 391 had no missing data on the CTQ-SF. Of the two respondents with missing

data, one did not respond to item 5 and item 7 on the Emotional Neglect scale, and one did not

respond to item 16 on the Minimization/Denial scale.

Materials

Childhood Trauma Questionnaire-Short Form (CTQ-SF). The CTQ-SF (Bernstein &

Fink, 1998, 2011) is a 28-item retrospective self-report questionnaire assessing traumatic

experiences when growing up. Each item is scored on a 5-point Likert scale (1 = never true, 2 =

rarely true, 3 = sometimes true, 4 = often true, 5 = very often true). Three items compose the

Minimization/Denial scale designed to detect socially desirable response style (false negatives). The

other 25 items are divided into five clinical subscales, with five items each: Emotional Abuse,

Physical Abuse, Sexual Abuse, Emotional Neglect, and Physical Neglect. Each of the five clinical

subscales’ scores can range from 5 to 25. Scores on the 25 items can also be summed to produce a

total CTQ-SF score.

Revised Childhood Experiences Questionnaire (CEQ-R), Parental Behavior Module.

The CEQ-R (Zanarini, 1992) is a semi-structured interview that assesses 12 types of negative

parental/caretaker behaviors: emotional abuse, physical abuse, sexual abuse, emotional withdrawal,

11
physical neglect, inconsistent treatment, emotional denial, failure to protect, lack of real

relationship, verbal abuse, parentification of child, and malevolent parenting. In the present study,

each parental behavior was rated dichotomously as present (0) versus absent (1). Interrater

reliability for the CEQ-R, in the present study, was satisfactory: The median κ value for categorical

variables was 1 (range 0.31–1.00), whereas the median intraclass correlation coefficient value for

continuous variables was 0.88 (range 0.65–0.99).

Statistical analyses

The original CTQ-SF manual (Bernstein & Fink, 1998) reports a factor analysis of the

25-item version of the CTQ-SF (excluding the three items from the Minimization/Denial Scale). To

examine if this original five-factor structure could be reproduced in our Danish clinical sample we

performed a confirmatory factor analysis in Mplus version 8.3 (Muthén & Muthén, 2017). Due to

non-normality of the data at item-level (skewness and kurtosis statistics exceeded the critical levels

of respectively 2 and 7; Ryu, 2011) we applied a robust maximum likelihood estimation using the

Satorra-Bentler chi-square. To evaluate model fit a selection of fit indices regarded as the most

informative, according to Kline (2005), were selected: the root mean square error of approximation

(RMSEA), the comparative fit index (CFI) and the standardized root mean square residual (SRMR).

The RMSEA is a parsimony-adjusted index with a built-in correction for model complexity. The

guideline is that RMSEA values ≥ .10 suggest unacceptable fit. Values ≤ .08 suggest approximate

or good model fit. A key advantage is that a confidence interval can be calculated for the RSMEA

value, which provides more information regarding model fit than a point estimate alone. The upper

bound of this confidence interval should be ≤ .10 for acceptable model fit (Chen, Curran, Bollen,

Kirby, & Paxton, 2008). The CFI assesses the relative improvement in fit compared with the

independence model (i.e., null model which assumes unrelated variables). A rule of thumb is that

12
values ≥ .90 indicate a reasonably good fit. The SRMR is a measure of the mean absolute residual

correlation, so values close to 0 are a better result. Ideally, the value of the SRMR should be < 0.08.

In analogy to some previous studies that had to allow error estimates to covary to improve model fit

(e.g., Thombs et al., 2009; Thombs et al., 2007), we used modification indices to decide which

parameter could be set free. We selected the highest modification index (the value of this represents

the estimated decrease in chi-square if a previously fixed parameter were to be estimated),

combined with theoretical reasons: correlated error terms were only allowed in the model if they

also made substantive sense (i.e. an effect exists that relates the two variables, which was not

included in the specified CFA model). As soon as all fit indices indicated reasonably fit, no further

error terms were added. Finally, we checked if the model with correlated error terms indeed yielded

a significant improvement in fit when compared to the original model without error constraints by

performing a chi-square difference test using the Satorra-Bentler scaled chi-square.

Next, given the indecisive research findings on the factor structure of the CTQ-SF in

previous studies, fit statistics of the original five-factor model specified in the manual were

compared with fit statistics for alternative models: a one factor model (representing only a total

CTQ-SF score), an original second order five-factor model and competing models at first and

second order, that is an alternative five-factor model (Grassi-Oliveira et al., 2014) and a four-factor

model (Sacchi et al., 2018). To compare the fit of these non-nested models we used two criteria

(Claeskens & Hjort, 2008): Akaike information criterion (AIC) and a sample size adjusted Bayesian

information criterion (BIC). AIC is an asymptotically efficient criterion for model selection, which

means that it tends to select the model that minimizes prediction error as sample size increases. BIC

originates from the Bayesian tradition in statistics and is concerned with the statistical property of

consistency, which refers to the one “true model” being selected with increasing probability as

13
sample size increases. Models with the lowest AIC and BIC values are considered to show the best

fit to the data.

To gauge cross-country construct equivalence by factorial similarity, we calculated

congruency coefficients with similar factors for a sample of American adult substance abusers

reported in the original CTQ-SF manual (Bernstein & Fink, 1998) and with the first order original

five-factor loadings reported in a Brazilian study combining clinical and non-clinical, adult and

adolescent samples (Grassi-Oliveira et al., 2014). The factorial similarity were evaluated with the

commonly used indicator for congruency, Tucker Phi (Tucker, 1951). A Tucker Phi value in the

range of .85 to .94 corresponds to fair similarity, while a value higher than .95 suggest that the two

factors being compared can be considered equal (Lorenzo-Seva & ten Berge, 2006).

Psychometric properties were further explored. To facilitate comparison with the

reported scale reliability in the manual (Bernstein & Fink, 1998), Cronbach’s coefficient alpha was

calculated to evaluate internal consistency. Because most of our data was non-parametrically

distributed, bivariate Spearman’s rho correlations were used to examine intercorrelations between

CTQ-SF scales and convergent associations of the CTQ-SF scales with the CEQ-R. The following

heuristic rules (Cohen, 1988) was used to interpret the effect size of the correlations: r = .10

indicates a small effect, r = .30 indicates a medium effect and r = .50 indicates a large effect. Given

the large number of tests, the Type I error rate was adjusted using a Bonferroni correction. The

conventional α = .05 was divided by the number of scales, yielding an adjusted alpha = .01.

Apart from the confirmatory factor analysis, conducted in Mplus, all other analyses

were performed using IBM SPSS Statics for MAC, version 25. Missing data was handled in Mplus

using full information maximum likelihood (FIML) and pairwise deletion in SPSS to maximize use

of information.

14
Results

Descriptive information

Table 1 provides means and standard deviations for the CTQ-SF scales across samples. In our

combined total sample kurtosis and skewness for the CTQ-SF total scale were 1.09 and 1.18,

respectively. For the Emotional Abuse, Physical Abuse, Sexual Abuse, Emotional Neglect, and

Physical Neglect subscales in the combined sample kurtosis and skewness values were –.47 and

0.74, 10.71 and 2.87, 8.50 and 2.91, –0.80 and 0.37, and 3.26 and 1.61, respectively.

Confirmatory factor analysis

We carried out confirmatory factor analysis to assess the structural validity of the

CTQ-SF in our Danish clinical sample. Because the original model did not reach adequate fit (see

Table 2), three pairs of error variances, that made substantive sense, were freed to covary: Item 9

and item 15, that refer to physical abuse, item 11 and item 12 that both include corporeal

punishment, and item 3 and item 11 that both refer to ’people in my family’. The goodness-of-fit

statistics of the original five-factor model, with three correlated error terms added, indicated good

model fit, chi-square = 673.437, df = 262, p < .001, RMSEA = .063, CFI = 0.909 and SRMR =

.071. This model, with correlated error terms, was also a significant improvement in fit compared to

the original model without error constraints, as demonstrated by a Satorra-Bentler scaled chi-square

difference test (scaling correction 0.3547, TRd 517.4292, ∆df = 3, p < .001). The standardized factor

loadings for this model, the original five-factor model with three error correlated items, are reported

in Table 3, and ranged from 0.34 (item 12 on the Physical Abuse scale) to 0.93 (item 20 on the

Sexual Abuse scale).

Competing factor models

15
 Next, the original model and competing models were compared in terms of fit

statistics. From Table 2 it is clear that based on AIC and BIC values the original model with three

correlated error terms outperformed all competing models in terms of fit to the data.

Factor congruency

Tucker Phi congruency coefficients corroborated factorial similarity of our best fitting

model (the original five-factor model with three pairs of error-correlated items) with a sample of

American adult substance abusers reported on in the original CTQ-SF manual (Bernstein & Fink,

1998) with congruency coefficients of 1.00 for Emotional Abuse, .97 for Physical Abuse, 1.00 for

Sexual Abuse, 1.00 for Emotional Neglect, and 1.00 for Physical Neglect, respectively.

Furthermore, factorial similarity was indicated with a five-factor solution derived from a Brazilian

sample combining clinical and non-clinical samples (Grassi-Oliveira et al., 2014) with congruency

coefficients of 1.00 for Emotional Abuse, .98 for Physical Abuse, 1.00 for Sexual Abuse, .99 for

Emotional Neglect, and .99 for Physical Neglect, respectively.

Intercorrelations

Table 4 displays intercorrelations between all the CTQ-SF scales. Correlations among

the five subscales ranged from .29 to .74 (ps < .001) indicating generally medium to large effects

(Cohen, 1988). These relatively high intercorrelations amongst the five subscales indicate modest

discriminant validity and suggest that it is feasible to extract a factor representing non-specific or

global childhood trauma, thereby supporting the use of the CTQ-SF total scale. Accordingly, we

proceeded to compute clinical subscale-total correlations (Table 4). The clinical subscale-total

correlations ranged from .52 to .90 (ps < .001.), revealing large effect sizes by conventional

standards. As would be expected the Minimization/Denial scale was negatively correlated with all

16
other CTQ-SF scales (rs ranged from –.16 to –.49, ps ranged from <. 01 to < 001., indicating small

to moderate effect sizes).

Internal consistency

Cronbach’s coefficient alpha values for the CTQ-SF total and subscales based on the

combined sample and for the respective subsamples are reported in Table 5. Alpha values for the

CTQ-SF total scale were high, ranging from .85 to .94 across all samples. For the five CTQ-SF

clinical subscales in the combined sample and stratified by gender the alpha values ranged from .70

to .93, and were remarkably similar to the median alpha values reported in the original CTQ-SF

manual for seven American samples (Bernstein & Fink, 1998).

Multi-method convergent validity

Convergent validity coefficients are reported in Table 6. As can be seen, on the whole

all clinical CTQ-SF scales were substantially correlated with the CEQ-R items indicating good

convergent validity. Of the total of 72 significant correlations (ps < .01) between the CTQ-SF scales

and CEQ-R items, 62 reached a medium to large effect size (rs ranged from .31 to .81). Four of the

five CTQ-SF clinical subscales (i.e. Emotional Abuse, Physical Abuse, Sexual Abuse, and Physical

Abuse) showed large positive correlations with their corresponding item on CEQ-R (rs ranged from

.52 to .81). The CTQ-SF Emotional Neglect scale was the only subscale without a distinct counter-

part scale on the CEQ-R. Still, the Emotional Neglect scale showed substantial correlations with

most of the CEQ-R items, notably Emotional withdrawal (r = .63), Lack of real relationship (r =

.62), and Malevolent parenting (r = .63).

Discussion

17
 The present study is the first to formally evaluate the psychometric performance of the

Danish version of the CTQ-SF in a heterogenous clinical sample. Our study provides support for

the Danish translation of the instrument in terms of structural validity, scale reliability, and multi-

method convergent validity. The original CTQ-SF five-factor model with three correlated error

terms provided the best fit to the data when compared with competing factor models. Additionally,

we could demonstrate factorial congruency of our best fitting model with factor models based on a

US and Brazilian sample.

Concerning reliability, the obtained coefficient alpha’s were very similar to those

reported in the original CTQ-SF manual, both in terms of magnitude and pattern (Bernstein & Fink,

1998), as well as to those reported in various other studies across countries and groups (e.g.,

Bernstein & Fink, 2011; Gerdner & Allgulander, 2009; Jiang et al., 2018; Thombs et al., 2009).

Consistent with previous research (Bernstein & Fink, 2011; Gerdner & Allgulander, 2009), the

subscales of Physical Abuse and Physical Neglect showed the lowest alpha coefficients, except in

the subsample of patients diagnosed with depression and in the subgroup of men in the combined

sample where the Sexual Abuse subscale was the lowest. Relatedly, the Physical Neglect scale

displayed the lowest internal consistency estimate of all the CTQ-SF clinical scales across all

samples, except among women in our combined sample.

We also obtained support for the convergent validity of the CTQ-SF clinical scales

with interview-ratings of adverse parental behaviors. This is a strong test of convergent validity,

considering the monoconstruct-heteromethod design employed, effectively minimizing the

possibility for shared method variance to inflate the estimates (Campbell & Fiske, 1959).

The applied implications of the present study are that clinicians in Denmark now have

a validated instrument to briefly screen for a wide range of childhood trauma. Effective intervention

hinges on detection. The CTQ-SF can easily be administered as part of routine assessment, to

18
facilitate recognition of various types of childhood trauma in diverse clinical groups. Such

recognition may be beneficial for patients and clinicians when making treatment plans, including

informing assessment of suicide risk, treatment needs, and prognosis. Moreover, assessment of

childhood trauma may also be valuable for formulating individualized case-formulations (Karterud

& Kongerslev, 2019) or make use of adjunct trauma informed care (Hopper, Bassuk, & Olivet,

2009) when indicated together with other forms of specialized treatment.

Some limitations should be considered when interpreting the results of this study. First

and foremost, our findings may not be generalizable to clinical groups different from those we have

investigated. This pertains especially to the convergent validity data which was only performed in a

small subsample of adult outpatients diagnosed with personality disorders. Moreover, our sample

size was too small to allow for factor analytic comparison between men and women and different

age groups. This must be addressed in future research, preferably together with potential differences

based on ethnicity/cultural background, for which we did not have data in the present study. This

way the factor invariance can be further examined across important subgroups. We provided some

preliminary evidence, in this study, for the cross-culturally factorial invariance of the five-factor

structure by calculating coefficients of congruency with a sample from US and Brazil. Still, future

studies could aim at collecting data from different countries to perform multi-group confirmatory

analyses, so that degree of measurement invariance can be directly tested by different measurement

models such as a congeneric model (same pattern of factor loadings), a tau-equivalent model (equal

factor loadings) and a parallel model (equal factor loadings and same amount of error). Also,

though a few prior studies, as mentioned in the introduction of this paper, have found evidence

indicating measurement invariance across age for the CTQ-SF there is indeed a need for more

future studies to test this – the factorial or measurement invariance of the CTQ-SF with respect to

age. Inspired by developmental research, such future studies could preferably use a longitudinal

19
study design to further explore the factorial invariance of the CTQ-SF (Widaman, Ferrer, & Conger,

2010).

Summarized, the findings from the present study provide evidence in support of the

reliability and validity of the Danish version of the CTQ-SF in clinical samples, and generally

suggest adequate psychometric properties comparable to those previously reported in the American

manual and in previous international studies.

References

Angelakis, I., Gillespie, E. L., & Panagioti, M. (2019). Childhood maltreatment and adult

suicidality: a comprehensive systematic review with meta-analysis. Psychological Medicine.

Advance online publication. doi:10.1017/S0033291718003823

Bach, B., & Fjeldsted, R. (2017). The role of DSM-5 borderline personality symptomatology and

traits in the link between childhood trauma and suicidal risk in psychiatric patients.

Borderline Personality Disorder and Emotional Dysregulation, 4, 12. doi:10.1186/s40479-

017-0063-7

Bach, B., & Sellbom, M. (2016). Continuity between DSM-5 Categorical Criteria and Traits

Criteria for Borderline Personality Disorder. Canadian Journal of Psychiatry, 61(8), 489-

494. doi:10.1177/0706743716640756

Bach, B., Simonsen, E., Christoffersen, P., & Kriston, L. (2017). The Young Schema Questionnaire

3 Short Form (YSQ-S3). European Journal of Psychological Assessment, 33(2), 134-143.

doi:10.1027/1015-5759/a000272

Baker, A. J. L., & Maiorino, E. (2010). Assessment of emotional abuse and neglect with the CTQ:

Issues and estimates. Children and Youth Services Review, 32, 740-748.

20
Bernstein, D. P., & Fink, L. (1998). Childhood Trauma Questionnaire: A Restrospective Self-

Report Manual. San Antonio: Pearson.

Bernstein, D. P., & Fink, L. (2011). Childhood Trauma Questionnaire: A Retrospective Self-Report.

Vejledning (Dansk version). Stockholm: Pearson.

Bernstein, D. P., Stein, J. A., Newcomb, M. D., Walker, E., Pogge, D., Ahluvalia, T., . . . Zule, W.

(2003). Development and validation of a brief screening version of the Childhood Trauma

Questionnaire. Child Abuse & Neglect, 27(2), 169-190. doi:https://doi.org/10.1016/S0145-

2134(02)00541-0

Bo, S., & Kongerslev, M. (2017). Self-reported patterns of impairments in mentalization,

attachment, and psychopathology among clinically referred adolescents with and without

borderline personality pathology. Borderline Personality Disorder and Emotion

Dysregulation, 4, 4. doi:10.1186/s40479-017-0055-7

Bo, S., Sharp, C., Fonagy, P., & Kongerslev, M. (2017). Hypermentalizing, attachment, and

epistemic trust in adolescent BPD: Clinical illustrations. Personality Disorders, 8(2), 172-

182. doi:10.1037/per0000161

Brewin, C. R. (2005). Systematic review of screening instruments for adults at risk of PTSD.

Journal of Traumatic Stress, 18(1), 53-62. doi:10.1002/jts.20007

Cakir, S., Tasdelen Durak, R., Ozyildirim, I., Ince, E., & Sar, V. (2015). Childhood trauma and

treatment outcome in bipolar disorder. Journal of Trauma & Dissociation, 17(4), 397-409.

doi:10.1080/15299732.2015.1132489

Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validity by the multitrait-

multimethod matrix. Psychological Bulletin, 56(2), 81-105. doi: 10.1037/h0046016

21
Chen, F., Curran, P. J., Bollen, K. A., Kirby, J., & Paxton, P. (2008). An empirical evaluation of the

use of fixed cutoff points in RMSEA test statistic in structural equation models. Sociological

Methods & Research, 36(4), 462-494. doi:10.1177/0049124108314720

Cicchetti, D. (2016). Socioemotional, personality, and biological development: Illustrations from a

multilevel developmental psychopathology perspective on child maltreatment. Annual

Review of Psychology, 67, 187-211.

Claeskens, G., & Hjort, N.L. (2008). Model Selction and Model Averaging. Cambridge: Cambridge

University Press.

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. Hillsdale, N.J.: L.

Erlbaum Associates.

Dovran, A., Winje, D., Arefjord, K., & Haugland, B. S. M. (2012). Traumatic events and

posttraumatic reactions among children and adolescents in out-of-home placement: A 25-

year systematic literature review. Journal of Child & Adolescent Trauma, 5(1), 16-32. doi:

10.1080/19361521.2012.644654

Dovran, A., Winje, D., Overland, S. N., Breivik, K., Arefjord, K., Dalsbo, A. S., . . . Waage, L.

(2013). Psychometric properties of the Norwegian version of the Childhood Trauma

Questionnaire in high-risk groups. Scandinavian Journal of Psychology, 54(4), 286-291.

doi:10.1111/sjop.12052

Dudeck, M., Vasic, N., Otte, S., Streb, J., Wingenfeld, K., Grabe, H. J., . . . Spitzer, C. (2015).

Factorial validity of the Short Form of the Childhood Trauma Questionnaire (CTQ—SF) in

German psychiatric patients, Inmates, and University Students. Psychological Reports,

116(3), 685-703. doi:10.2466/16.03.PR0.116k27w5

22
Etain, B., Aas, M., Andreassen, O. A., Lorentzen, S., Dieset, I., Gard, S., . . . Henry, C. (2013).

Childhood trauma is associated with severe clinical chracteristics of bipolar disorder.

Journal of Clinical Psychiatry, 74(10), 991-998. doi: 10.4088/JPC.13m08353

Gerdner, A., & Allgulander, C. (2009). Psychometric properties of the Swedish version of the

Childhood Trauma Questionnaire-Short Form (CTQ-SF). Nordic Journal of Psychiatry,

63(2), 160-170. doi:10.1080/08039480802514366

Gilbert, R., Widom, C. S., Browne, K., Fergusson, D., Webb, E., & Janson, S. (2009). Burden and

consequences of child maltreatment in high-income countries. Lancet, 373(9657), 68-81.

doi:10.1016/S0140-6736(08)61706-7

Gillespie, S. M., Kongerslev, M. T., Sharp, C., Bo, S., & Abu.Akel, A. M. (2018). Does affective

theory of mind contribute to proactive aggression in boys with conduct problems and

psychopathic tendencies? Child Psychiatry & Human Development, 49(6), 906-916.

doi:https://doi.org/10.1007/s1057

Grassi-Oliveira, R., Cogo-Moreira, H., Salum, G. A., Brietzke, E., Viola, T. W., Manfro, G. G., . . .

Arteche, A. X. (2014). Childhood Trauma Questionnaire (CTQ) in Brazilian samples of

different age groups: findings from confirmatory factor analysis. PLoS One, 9(1), e87118.

doi:10.1371/journal.pone.0087118

Green, J. G., McLaughlin, K. A., Berglund, P. A., Gruber, M. J., Sampson, N. A., Zaslavsky, A. M.,

& Kessler, R. C. (2010). Childhood adversities and adult psychiatric disorders in the

national comorbidity survey replication I: associations with first onset of DSM-IV disorders.

Archives of General Psychiatry, 67(2), 113-123. doi: 10.1001/archgenpsychiatry.2009.186

Hernandez, A., Gallardo-Pujol, D., Pereda, N., Arntz, A., Bernstein, D. P., Gaviria, A. M., . . .

Gutierrez-Zotes, J. A. (2013). Initial validation of the Spanish Childhood Trauma

23
 Questionnaire-Short-Form: factor structure, reliability and association with parenting.

Journal of Interpersonal Violence, 28(7), 1498-1518. doi:10.1177/0886260512468240

Hopper, E. K., Bassuk, E. L., & Olivet, J. (2009). Shelter from the storm: Trauma-Informed Care in

homelessness services settings. The Open Health Services and Policy Journal, 2, 131-151.

doi:10.2174/1874924001003010080

Jiang, W. J., Zhong, B. L., Liu, L. Z., Zhou, Y. J., Hu, X. H., & Li, Y. (2018). Reliability and

validity of the Chinese version of the Childhood Trauma Questionnaire-Short Form for

inpatients with schizophrenia. PLoS One, 13(12), e0208779.

doi:10.1371/journal.pone.0208779

Karterud, S. W., & Kongerslev, M. T. (2019). A Temperament-Attachment-Mentalization-Based

(TAM) theory of personality and its Ddsorders. Frontiers in Psychology, 10(518).

doi:10.3389/fpsyg.2019.00518

Kessler, R. C., McLaughlin, K. A., Green, J. G., Gruber, M. J., Sampson, N. A., Zaslavsky, A. M., .

. . Williams, D. R. (2010). Childhood adversities and adult psychopathology in the WHO

World Mental Health Surveys. British Journal of Psychiatry, 197(5), 378-385.

doi:10.1192/bjp.bp.110.080499

Kline, R. B. (2005). Principles and Practice of Structural Equation Modeling. New York:

Guildford Press.

Kongerslev, M., Moran, P., Bo, S., & Simonsen, E. (2012). Screening for personality disorder in

incarcerated adolescent boys: preliminary validation of an adolescent version of the

Standardised Assessment of Personality – Abbreviated Scale (SAPAS-AV). BMC

Psychiatry, 12, 94. doi: 10.1186/1471-244X-10-10

Kongerslev, M. T., Bo, S., Forth, A. E., & Simonsen, E. (2015). Assessment of the affective

dimensions of psychopathy with the Danish version of the Inventory of Callous-

24
 Unemotional Traits among incarcerated boys: A study of reliability, criterion validity, and

construct validity. Scandinavian Journal of Child and Adolescent Psychiatry and

Psychology, 3(1), 80-96. doi: 10.21307/sjcapp-2015-008

Ladegaard, N., Lysaker, P. H., Larsen, E. R., & Videbech, P. (2014). A comparison of capacities for

social cognition and metacognition in first episode and prolonged depression. Psychiatry

Research, 220(3), 883-889. doi:10.1016/j.psychres.2014.10.005

Levey, E. J., Apter, G., & Harrison, A. M. (2016). The global problem of child maltreatment:

perspectives on mechanisms of influence and illness presentation. International Journal of

Culture and Mental Health, 10(1), 90-96. doi:10.1080/17542863.2016.1264440

Lorenzo-Seva, U., & ten Berge, J. M. F. (2006). Tucker's congruence coefficient as a meaningful

index of factor similarity. Methodology, 2(2), 57-64. doi:10.1027/1614-2241.2.2.57

McCrory, E. J., Gerin, M. I., & Viding, E. (2017). Annual Research Review: Childhood

maltreatment, latent vulnerability and the shift to preventative psychiatry - the contribution

of functional brain imaging. Journal of Child Psychology and Psychiatry, 58(4), 338-357.

doi:10.1111/jcpp.12713

Moody, G., Cannings-John, R., Hood, K., Kemp, A., & Robling, M. (2018). Establishing the

international prevalence of self-reported child maltreatment: a systematic review by

maltreatment type and gender. BMC Public Health, 18(1). doi:10.1186/s12889-018-6044-y

Muthén, L. K., & Muthén, B. O. (2017). Mplus User’s Guide (8th ed.). Los Angeles: Muthén &

Muthén.

Read, J., Hammersley, P., & Rudegeair, T. (2018). Why, when and how to ask about childhood

abuse. Advances in Psychiatric Treatment, 13(2), 101-110. doi:10.1192/apt.bp.106.002840

25
Ryu, E. (2011). Effects of skewness and kurtosis on normal-theory based maximum likelihood test

statistic in multilevel structural equation modeling. Behavior Research Methods, 43(4),

1066-1074. doi:10.3758/s13428-011-0115-7

Sacchi, C., Vieno, A., & Simonelli, A. (2018). Italian validation of the Childhood Trauma

Questionnaire-Short Form on a college group. Psychological Trauma, 10(5), 563-571.

doi:10.1037/tra0000333

Schäfer, I., & Fisher, H. L. (2011). Childhood trauma and psychosis - what is the evidence?

Dialogues in Clinical Neuroscience, 13(3), 360-365.

Stoltenborgh, M., Bakermans-Kranenburg, M. J., Alink, L. R. A., & van Ijzendoorn, M. H. (2015).

The prevalence of child maltreatment across the globe: Review of a series of meta-analyses.

Child Abuse Review, 24(1), 37-50. doi:10.1002/car.2353

Thombs, B. D., Bernstein, D. P., Lobbestael, J., & Arntz, A. (2009). A validation study of the

Dutch Childhood Trauma Questionnaire-Short Form: Factor structure, reliability, and

known-groups validity. Child Abuse & Neglect, 33(8), 518-523.

doi:10.1016/j.chiabu.2009.03.001. doi: 10.1016/j.chiabu.2009.03.001

Thombs, B. D., Lewis, C., Bernstein, D. P., Medrano, M. A., & Hatch, J. P. (2007). An evaluation

of the measurement equivalence of the Childhood Trauma Questionnaire--Short Form

across gender and race in a sample of drug-abusing adults. Journal of Psychosomatic

Research, 63(4), 391-398. doi:10.1016/j.jpsychores.2007.04.010

Trauelsen, A. M., Bendall, S., Jansen, J. E., Nielsen, H. G. L., Pedersen, M. B., Trier, C. H., . . .

Simonsen, E. (2015). Childhood adversity specificity and dose-response effect in non-

affective first-episode psychosis. Schizophrenia Research, 165(1), 52-59. doi:

10.1016/j.schres.2015.03.014

26
Tucker, L. R. (1951). A Method for the Synthesis of Factor Analytic Studies. Washington, DC:

Department of the Army.

Vachon, D. D., Krueger, R. F., Rogosch, F. A., & Cicchetti, D. (2015). Assessment of the harmful

psychiatric and behavioral effects of different forms of child maltreatment. JAMA

Psychiatry, 72(11), 1135-1142. doi:10.1001/jamapsychiatry.2015.1792

Villano, C. L., Cleland, C., Rosenblum, A., Fong, C., Nuttbrock, L., Marthol, M., & Wallace, J.

(2004). Psychometric utility of the childhood trauma questionnaire with female street-based

sex workers. Journal of Trauma & Dissociation, 5(3), 33-41. doi:10.1300/J229v05n04_03

Widaman, K. F., Ferrer, E., & Conger, R. D. (2010). Factorial Invariance within Longitudinal

Structural Equation Models: Measuring the Same Construct across Time. Child

Development Perspectives, 4(1), 10-18. doi:10.1111/j.1750-8606.2009.00110.x

Williams, L. M., Debattista, C., Duchemin, A. M., Schatzberg, A. F., & Nemeroff, C. B. (2016).

Childhood trauma predicts antidepressant response in adults with major depression: data

from the randomized international study to predict optimized treatment for depression.

Translational Psychiatry, 6(5), e799-e799. doi:10.1038/tp.2016.61

World Health Organization (2017). Child Maltreatment: Infographics. Geneva: World Health

Organization.

Zanarini, M. C. (1992). Revised Childhood Experiences Questionnaire (CEQ-R). McLean Hospital,

Harvard Medical School. Belmont, MA.

Zanarini, M. C., Frankenburg, F. R., Hennen, J., Reich, D. B., & Silk, K. R. (2006). Prediction of

the 10-year course of borderline personality disorder. American Journal of Psychiatry, 163,

827-832. doi: 10.1176/ajp.2006.163.5.827

27
Zatti, C., Rosa, V., Barros, A., Valdivia, L., Calegaro, V. C., Freitas, L. H., . . . Schuch, F. B.

(2017). Childhood trauma and suicide attempt: A meta-analysis of longitudinal studies from

the last decade. Psychiatry Research, 256, 353-358. doi:10.1016/j.psychres.2017.06.082

Zeanah, C. H., & Humphreys, K. L. (2018). Child abuse and neglect. Journal of the American

Academy of Child and Adolescent Psychiatry, 57(9), 637-644.

doi:10.1016/j.jaac.2018.06.007

28
Table 1
Medians, means and standard deviations for the Danish Childhood Trauma Questionnaire-Short Form across samples
EA PA SA EN PN Total MD
Sample Md M Md M Md M Md M Md M Md M Md M
(SD) (SD) (SD) (SD) (SD) (SD) (SD)
Personality 14 13.87 6 7.78 5 7.75 16 15.34 8 9.16 51 53.89 0 0.16
disorder (5.52) (4.21) (4.57) (5.15) (4.42) (18.51) (0.45)
Psychosis 9 10.96 5 6.41 5 6.77 12 12.15 7 8.22 41 44.36 0 0.44
(5.21) (2.52) (4.25) (5.13) (3.21) (15.29) (0.83)
Depression 7 9.31 5 5.90 5 5.70 12 12.50 6 7.17 38 40.59 0 0.51
(4.91) (2.29) (1.55) (4.81) (2.50) (12.68) (0.88)
Delinquent 7 7.33 6 6.48 5 5.08 8 8.58 7 7.36 34 34.81 0 0.85
boys (2.47) (1.63) (0.67) (3.38) (2.49) (7.89) (1.16)
Combined 9 10.98 5 6.83 5 6.59 12 12.64 7 8.20 41 45.19 0 0.43
(5.44) (3.16) (3.71) (5.36) (3.56) (16.65) (0.85)
Men 8 9.39 5 6.88 5 5.63 10 11.39 7 8.12 37 41.32 0 0.57
(4.67) (3.49) (2.15) (5.22) (3.57) (15.29) (0.96)
Women 13 12.91 5 6.76 5 7.76 14 14.15 7 8.29 47 49.88 0 0.26
(5.70) (2.72) (4.74) (5.15) (3.56) (17.06) (0.66)
Note. EA = Emotional Abuse; PA = Physical Abuse; SA = Sexual Abuse; EN = Emotional Neglect; PN = Physical Neglect; MD =
Minimization/Denial.

29
Table 2

Model fit for the confirmatory factor analysis models of the Danish Childhood-Trauma Questionnaire-Short Form in a clinical sample

Sample size

Model 2 (df) RMSEA [90% CI] CFI SRMR AIC adjusted BIC

Original 815.923 (265) .073 [.067, .079] 0.878 0.091 22438.799 22506.439

Original with error terms 673.437 (262) .063 [.057, .069] 0.909 0.071 22261.224 22331.251

Original second order 859.277 (270) .075 [.069, .080] 0.870 0.096 22496.847 22560.508

1-factor model 2379.739 (279) .139 [.134, .144] 0.537 0.207 24685.341 24741.840

Alternative 5-factor model 825.405 (265) .074 [.068, .079] 0.876 0.092 22453.252 22520.892

Second order alternative 5-factor model 864.520 (270) .075 [.069, .081] 0.869 0.096 22507.826 22571.487

4-factor model 899.685 (269) .077 [.072, .083] 0.861 0.098 22554.541 22618.998

Second order 4-factor model 901.130 (271) .077 [.072, .083] 0.861 0.100 22556.733 22619.597

Note. S-B = RMSEA = root mean square error of approximation; CI = confidence interval; CFI = comparative fit index; SRMR =
standardized root mean square residual; AIC = Akaike information criterion; BIC = Bayesian information criterion.

30
Table 3
Standardized factor loadings for the Danish Childhood Trauma Questionnaire-Short Form based
on confirmatory factor analysis in a clinical sample.

Factor/item Loadings
EMOTIONAL ABUSE
3. People in my family called me things like ‘‘stupid,’’ ‘‘lazy,’’ or ‘‘ugly.’’ 0.62
8. I thought that my parents wished I had never been born. 0.78
14. People in my family said hurtful or insulting things to me. 0.83
18. I felt that someone in my family hated me. 0.84
25. I believe I was emotionally abused. 0.71
PHYSICAL ABUSE
9. I got hit so hard by someone in my family that I had to see a doctor or go to the 0.37
hospital.
11. People in my family hit me so hard that it left me with bruises or marks. 0.52
12. I was punished with a belt, a board, a cord, or some other hard object. 0.34
15. I believe that I was physically abused. 0.80
17. I got hit or beaten so badly that it was noticed by someone like a teacher,
neighbor, or doctor. 0.43
SEXUAL ABUSE
20. Someone tried to touch me in a sexual way, or tried to make me touch them. 0.93
21. Someone threatened to hurt me or tell lies about me unless I did something 0.68
sexual with them.
23. Someone tried to make me do sexual things or watch sexual things. 0.88
24. Someone molested me. 0.82
27. I believe that I was sexually abused. 0.90
EMOTIONAL NEGLECT
5. There was someone in my family who helped me feel that I was important or 0.62
special (R).
7. I felt loved (R). 0.86
13. People in my family looked out for each other (R). 0.88
19. People in my family felt close to each other (R). 0.86
28. My family was a source of strength and support (R). 0.92
PHYSICAL NEGLECT
1. I didn’t have enough to eat. 0.48
2. I knew that there was someone to take care of me and protect me (R). 0.73
4. My parents were too drunk or high to take care of the family. 0.57
6. I had to wear dirty clothes. 0.54
26. There was someone to take me to the doctor if I needed it (R). 0.61

Note. All standardized factor loadings were statistically significant at p < .001. (R) denotes a
reverse-scored item. Fit indices were as follows: root mean square error of approximation = 0.063,
comparative fit index = 0.909, and standardized root mean square residual = 0.071.

31
Table 4

Intercorrelations among the Danish Childhood Trauma Questionnaire-Short Form scales in the

combined sample.

EA PA SA EN PN MD

Total .90 .52 .57 .89 .73 –.49

EA .44 .50 .74 .53 –.44

PA .36 .30 .38 –.16

SA .39 .29 –.18

EN .62 –.48

PN –.24

Note. All values are Spearman’s rho correlations. All correlations are significant at p < .001, except

for the correlation between the PA and MD scale which is significant at p < .01. N varies from 391

to 393 due to missing data. EA = Emotional Abuse; PA = Physical Abuse; SA = Sexual Abuse; EN

= Emotional Neglect; PN = Physical Neglect; MD = Minimization/Denial.

32
Table 5

Internal consistency (Cronbach’s coefficient alpha) for the Danish Childhood Trauma

Questionnaire-Short Form (CTQ-SF) scales across samples and gender, and in comparison, to the

median US based alpha coefficients reported in the original manual.

Sample EA PA SA EN PN Total

Personality disorder .87 .86 .91 .90 .81 .94

Psychosis .88 .75 .94 .92 .67 .92

Depression .87 .82 .75 .88 .52 .91

Delinquent boys .79 .50 .97 .85 .47 .85

Combined .89 .82 .92 .92 .73 .94

Men .89 .88 .87 .92 .73 .94

Women .88 .70 .93 .90 .73 .93

Median alpha values based on seven US samples† .89 .82 .92 .89 .66 –

Note. EA = Emotional Abuse; PA = Physical Abuse; SA = Sexual Abuse; EN = Emotional Neglect;

PN = Physical Neglect. †Median Alpha values for the five CTQ-SF subscales reported in the

original English manual (Bernstein & Fink, 1998)

33
Table 6

Associations between the Danish Childhood Trauma Questionnaire-Short Form and interview-

rated caretaker behavior based on the CEQ-R in a sample of outpatients diagnosed with

personality disorders.

CEQ-R items EA PA SA EN PN CTQ Total

Emotional abuse .52*** .44*** .39** .51*** .61*** .61***

Physical abuse .38** .67*** .18ns .41*** .40** .48***

Sexual abuse .43*** .15ns .81*** .26* .31** .49***

Emotional withdrawal .52*** .38** .35** .63*** .66*** .66***

Physical neglect .42*** .37** .25* .38** .65*** .49***

Inconsistent treatment .56*** .26* .29* .51*** .47*** .56***

Emotional denial .56*** .25* .35** .52*** .58*** .60***

Failure to protect .54*** .46*** .47*** .55*** .62*** .66***

Lack of real relationship .60*** .29* .37** .62*** .52*** .66***

Verbal abuse .61*** .39** .33** .50*** .58*** .60***

Parentification of patient .42*** .12ns .26* .42*** .58*** .44***

Malevolent parenting .66*** .41*** .56*** .63*** .64*** .73***

Note. All values are Spearman’s rho correlations. N = 72. CEQ-R = Childhood Experiences

Questionnaire, Parental Behavior Module; EA = Emotional Abuse; PA = Physical Abuse; SA =

Sexual Abuse; EN = Emotional Neglect; PN = Physical Neglect. *p < .05; **p < .01; ***p < 001;
ns
non-significant.

34