A large number of children and adolescents are victims of traumatic experiences, the most common being bullying, cyberbullying, and emotional, physical, and/or sexual abuse, including that perpetrated by peers. Posttraumatic stress disorder (PTSD) is one of the most prevalent disorders among children and adolescents who have experienced a traumatic event. A recent meta- analysis showed that 16% of children and adolescents exposed to a traumatic event subsequently developed PTSD (Alisic et al., 2014). The Diagnostic and statistical manual of mental disorders- 5th edition ([DSM-V] American Psychiatric Association [APA], 2013) adds four new clusters of symptoms to this disorder: intrusion, avoidance, negative alterations in cognition and mood, and alterations in arousal and reactivity. Therefore, instruments with good psychometric proprieties to assess PTSD among adolescents are needed.

One such instrument is the Child PTSD Symptom Scale (CPSS) developed by Foa, Johnson, Feeny, and Treadwell (2001). The CPSS is brief, easy to administer, available for free, and has great potential for research and clinical applications. It comprises 17 items. The total symptom score and three symptom clusters of the CPSS demonstrated suitable internal consistency, as well as high test–retest reliability for both the total score and the three subscales (intrusion, avoidance, and arousal symptoms) (Foa et al., 2001). The CPSS has demonstrated very good convergent and some support for its discriminant validity has also been obtained (Nixon, Sterk, & Pearce, 2012; Stewart, Ebesutani, Drescher, & Young, 2015). Cross-cultural translations of the scale have shown good reliability and validity, including Hebrew (Rachamim et al., 2011) and Turkish versions (Kadak, Boysan, Ceylan, & Çeri, 2014). In spite of this, there has been limited evaluation of the instrument’s construct validity and its psychometric properties in the Spanish language, although Spanish is one of the most widely spoken language in the world.

Several studies aimed at analyzing the dimensions of PTSD have been performed and a meta-analysis demonstrated that the dysphoria model outperformed the numbing model in almost all subsamples (Yufik & Simms, 2010). However, in the study by Helpman et al. (2014) based on responses to the CPSS findings demonstrated that the model that best fit the data was the four-factor numbing model, compared to the three-factor model (intrusion, avoidance, and arousal symptoms), a three-factor dysphoria model (intrusion, dysphoria, and arousal), and a four-factor dysphoria model (intrusion, avoidance, dysphoria, and arousal).

Gudiño and Rindlaub (2014) examined the CPSS and its psychometric properties in a sample of Latino students who were exposed to chronically elevated levels of community violence. Their results supported a three-factor model (re-experiencing/ intrusion, avoidance, and arousal), but did not support the four- factor numbing or four-factor dysphoria models (King, Leskin, King, & Weathers, 1998; Simms, Watson, & Doebbeling, 2002, respectively). High internal consistence was obtained for the total scale (α = .92). Meyer, Gold, Beas, Young, and Kassam-Adams (2014) also tested the psychometric properties of the CPSS in Latino children residing in USA. Compared to the Spanish version, the English version had a better fit to the data for all the models of the PTSD symptoms structure. The four-factor numbing and dysphoria models had the best fit for both versions. Nevertheless, the fit indices were quite modest, particularly for the Spanish model (Meyer et al., 2014), which indicates that this version of the CPSS lacks construct validity.

The first aim of the present study was to provide data on the factor structure of the CPSS in a sample of Spanish adolescents who had experienced peer abuse (i.e., bullying, cyberbullying, and/or dating violence) during the previous year. Five models were examined: a single-factor model, the DSM-IV-TR three-factor model, the three-factor dysphoria model, the four-factor numbing model, and the four-factor dysphoria model corresponding to the dimensions of PTSD described in the DSM-V (APA, 2013). The second aim was to establish the internal consistency of the measure and to examine its criterion validity.

Method

Participants

The original sample comprised 699 high-school students from Málaga (Spain). 49 participants were eliminated due to incomplete responses and 274 participants were also eliminated because they did not fulfil the criterion of having been exposed to peer violence during the previous year. Therefore, the final sample comprised 339 adolescents (172 boys and 167 girls; mean age 13.95 years; SD = 1.29).

The participating students reflected the general characteristics of children attending urban secondary schools in Andalusia (Spain). Most students (85.3%) were between 13 and 17 years of age. Of the total sample, 94.7% were living with their families (with both or one of their parents, and with their brothers/sisters, if applicable).

Instruments

All participants were requested to give written information regarding their age, gender, educational level, and school year.

Adolescent Victimization through Mobile Phone and Internet Scale (CIBVIC; Buelga, Cava, & Musitu, 2012). This questionnaire includes 8 items refer to mobile phone cyberbullying and 10 items to internet cyberbullying experienced over the previous year. The items are rated on a 4-point Likert-type scale. Cronbach’s alpha for this questionnaire was 0.81.

Self-reported Victimization Questionnaire (Cava, Musitu, & Murgui, 2007). The questionnaire uses a 20-item scale are rated on a 4-point Likert-type on which participants indicate how often during the last school year they have experienced 20 victimizing experiences. Ten items refer to peer overt victimization (physical and/or verbal assault), and 10 items refer to peer relational victimization (social ostracism). Cronbach’s alpha for this questionnaire was 0.90.

Conflict in Adolescent Dating Relationships Inventory (Spanish version, Fernández-Fuertes, Fuertes, & Pulido, 2006). The brief 34-item scale was used to assess five types of intimate violence in adolescent dating relationships: sexual abuse, relational abuse, verbal or emotional abuse, threatening behaviour, and physical abuse. The items are rated on a 4-point Likert-type scale. Cronbach’s alpha for this questionnaire was 0.89.

Emotional Quotient Inventory Youth Version (Spanish version, López-Zafra, Pulido, & Berrios, 2014). The EQ-i:YV is a self-report measure that assesses the level of emotional and social functioning in children and adolescents of 7 to 17 years of age. This study used the 8 items that measure general mood. The instrument uses a 5-point scale. Cronbach’s alpha for this questionnaire was 0.83.

Satisfaction with Life Scale (Spanish version, Atienza, Pons, Balaguer, & García-Merita, 2000). This questionnaire was developed as a measure of subjective global life satisfaction and well-being. It includes 5 items rated on a 5-point Likert-type scale ranging from 1 (totally disagree) to 5 (totally agree). Cronbach’s alpha for this questionnaire was 0.74.

Kessler Psychological Distress Scale (Spanish version, Vargas, Villamil, Rodríguez, Pérez, & Cortés, 2011). The questionnaire includes 10 items that refer to the level of anxiety and depressive symptoms experienced during the last month. The items are rated on a 5-point Likert-type scale ranging from 1 (never) to 5 (always). Cronbach’s alpha for this questionnaire was 0.84.

KIDSCREEN-10 Index (Spanish version, Erhart et al., 2009). It assesses perceived health-related quality of life. The 10 items are rated on a 5-point scale ranging from 1 (none of the time) to 5 (all of the time). It also includes a question on perceived health. Cronbach’s alpha for this questionnaire was 0.75.

Child PTSD Symptom Scale (CPSS; Foa, Johnson, Feeny, & Treadwell, 2001). Comprises 17 items corresponding to PTSD symptoms and is designed for use with children aged 8-18 years. Participants rate how often each symptom has occurred in the past month on a 4-point scale ranging from 0 (not at all) to 3 (5 or more times a week). The total score is calculated by summing all items (see below for further details). Both English and Spanish versions are available. The Spanish version was used in the current study, although some items were slightly reworded to facilitate understanding by Spanish populations. A list of the items in the questionnaire can be requested from the authors.

Procedure

The current study was part of a larger school-based study of traumatic stress events in adolescents, which received institutional review board approval at the University of Málaga (Spain).

Teachers, parents, and students were informed about the aim of the research before they agreed to participate. The protocol for the study was approved by the Institutional Review Board of the school. After parental consent had been obtained and prior to data collection, children voluntarily agreed to participate in the study. Assessments were conducted in small groups of 20 to 25 students and a single trained clinical psychologist controlled the procedure. All tests followed the authors’ instructions on use of the instruments and were conducted during school hours. Each participant anonymously completed a battery of instruments, which were always presented in the same order.

Data analysis

Univariate and multivariate distributions were analysed. Inspection of Mahalanobis d2 values indicated that there were no multivariate outliers in the sample. Little’s MCAR test was used and missing values were replaced by using the multiple imputation method.

Means, standard deviations, and correlations were calculated for each variable. To test the factor structure of the CPSS, a confirmatory factor analysis (CFA) was performed. Maximum likelihood estimation was used for the analyses. To test model fit, a chi-squared statistic was used as an absolute index of goodness of fit. The model fit was considered to be satisfactory according to the following criteria (Ullman, 2006): a) the adjusted goodness- of-fit-index (AGFI) and the comparative fit index (CFI) with values of .90 indicating a good fit; b) the root mean square error of approximation (RMSEA) and the standardized root mean square residual (SRMR) were less than .08. In all CFA analyses, correlations between errors were constrained to zero, items were constrained to load on only one factor, and factors were free to correlate.

Comparisons between nested models were conducted using the S-B χ2 test (Satorra & Bentler, 2001). Comparisons between non-nested models were conducted by comparing the Akaike information criterion (AIC). Firstly, we assessed the model fit of each of the five models examined. We then compared the nested models followed by the non-nested models. Once the best-fitting model was found, it was modified by modelling the error terms that were correlated and the model fit was assessed again.

A series of moderated multiple regression analysis were then performed to analyse the effects of the interaction of the CPSS scores and general mood and happiness, and the interaction of the CPSS scores and life satisfaction, on psychological distress and health-related quality of life. A series of standardized product variables were then created to represent interactions between the CPSS scores and psychological distress, and between the CPSS scores and health-related quality of life. Interaction effects were only analysed in those cases in which the predictors significantly predicted the outcome variables considered in the analyses.

Internal consistency was calculated using Cronbach’s α coefficient for the total score and for the score of each factor. To assess the criterion validity of the CPSS, associations were analysed between the CPSS global score and general mood and happiness, life satisfaction, psychological distress, and health- related quality of life, while controlling for sex.

Results

Descriptive and preliminary analyses

Preliminary examinations of the data revealed that the overall level of missing data was 4.73%. Missing values were imputed after finding no statistically reliable deviation from randomness using Little’s MCAR test, χ2 (119) = 90.133, p = 0.07. The remaining analyses were conducted on the imputed data set.

The types of abuse reported by the participants were as follows: cyberbullying by mobile phone (49.4%), by internet (41.2%), or by both (32.5%), bullying (11.2%), bullying and some type of cyberbullying (28.9%), dating violence (36.3%), or all these types of abuse (10.0%). As shown in Table 1, score ranges, means and standard deviations for each variable were calculated. Table 1 also shows the partial correlations between the measures considered in the study while controlling for sex. Medium significant negative effect size correlations were found between the scores on the CPSS and general mood and happiness, life satisfaction, and health- related quality of life. A medium significant and positive effect size was found between the CPSS and global psychological distress.

Girls had higher scores than boys on the CPSS and reported greater psychological distress, and had lower scores on happiness and perceived health-related quality of life. The mean CPSS total score for the whole sample was 29.19 (SD = 9.19), and 53.4% of participants met or exceeded the clinical cutoff of 11 established by Foa et al. (2001). When a cutoff score of 15 was used, as recommended by the International Society of Traumatic Stress Studies (2012), 34.8% of participants met the criteria for a diagnosis of PTSD. When a clinical cutoff score of 16 or greater was used, recommended by Nixon et al. (2013), 30.9% of the participants were identified as having probable PTSD.

Factor structure

Model fit was assessed for each model (see Table 2). The nested models were then compared (i.e., the four-factor numbing model and the three-factor model, as well as the four-factor dysphoria model and the three-factor dysphoria model). A significant difference was found between the four-factor numbing model and the three-factor model, χ2 (3) = 35.93, p < .000. The four-factor numbing model, χ2 (113) = 317.00, provided a better fit than the three-factor model, χ2 (116) = 352.94. A significant difference was found between the four-factor dysphoria model and the three- factor dysphoria model, χ2 (3) = 42.45, p < .000. The four-factor dysphoria model, χ2 (111) = 306.43, provided a better fit than the three-factor dysphoria model, χ2 (116) = 323.86.

The AIC goodness-of-fit index was then used to compare the remaining models. Table 2 shows that the four-factor dysphoria model had the lowest value and the best fit. The results of the evaluation of the final model indicated an adequate fit, χ2 (108) = 306,43, p = .000, NFI = .92, CFI = .95, RMSEA = .06. This final model included four factors: intrusion (items 1-5), avoidance (items 6-8), dysphoria (items 9-15), and arousal (items 16 and 17) (see Figure 1). Factor loadings were equal or more than .51.

Criterion validity

We analysed the effects of the interactions between the global score of the CPSS and general mood and happiness, as well as the effects of the interactions between the global score on the CPSS and life satisfaction, on psychological distress and health- related quality of life (see Table 3). Psychological distress was significantly predicted by general mood and happiness, and by the total CPSS score. The interaction between general mood and the CPSS score added significant incremental variance, 1.8%, B =.016, p = .003. General mood and CPSS score were significantly and independently associated with health-related quality of life, although no interaction effects were found. Life satisfaction and the global score on the CPSS significantly and independently predicted psychological distress, although no interaction effects were found. Furthermore, life satisfaction and CPSS scores were significantly and independently associated with health-related quality of life, but no interaction effects were found.

Internal consistency

The CPSS total symptom scale demonstrated high internal consistency with the full sample (α = .90). Internal consistency was moderate to good within the subscale symptom factors: α = .80 on the re-experiencing/intrusion subscale, α = .70 on the avoidance subscale, α = .83 on the dysphoria subscale, and α = .74 on the arousal subscale.

Discussion

Although the CPSS is a widely used measure to assess PTSD in young people, to the best of our knowledge this study is the first to examine the factor structure and psychometric properties of the CPSS in a Spanish sample. The purpose was to determine the factor structure of the Spanish version of the CPSS in a sample of adolescents who had experienced peer abuse during the previous year. In addition, the internal consistency and criterion validity of the questionnaire were examined.

The adolescents had been exposed to some type of peer abuse in a non-negligible percentage. Moreover, 30.9% of participants in this study met the criteria for a diagnosis of PTSD when considering the recommendations of the International Society of Traumatic Stress Studies (2012), increasing up to 53.4% according to the clinical cutoff established by Foa et al. (2001).

Confirmatory factor analyses suggested that the findings are best described by the four-factor dysphoria model. In fact, the models that have been most consistently replicated across studies are the four-factor models of King et al. (1998) and Simms et al. (2002). Using the CPSS, Kassam-Adams, Marsac, and Cirilli (2010) found that both the numbing four-factor model and the dysphoria four-factor model had a good fit in a sample of children and adolescents who had experienced unintentional injury. In the study by Meyer et al. (2014) the dysphoria four-factor model had a slightly better fit in the Spanish sample, which was also the case in the current study. Hukkelberg & Jensen (2011) also found that the dysphoria four-factor model had the best fit to the observed data. Overall, the results suggest that four dimensions define PTSD in youth samples, which is in line with proposals regarding adult data (i.e., King et al., 1998; Simms et al., 2002).

The Spanish version of the CPSS had sound psychometric properties with good reliability and validity. The total symptom scale exhibited high internal consistency and reliability for the subscales and was similar to the initial validation study by Foa et al. (2001) and subsequent psychometric analyses (Gillihan, Aderka, Conklin, Capaldi, & Foa, 2013; Gudiño & Rindlaub, 2014; Kadak et al., 2014; Meyer et al., 2014; Rachamim et al., 2011; Stewart et al., 2015). Regarding criterion validity, the total symptom scale correlated with all the measures considered in the study. Nevertheless, no interaction effects were found between the CPSS scores and general mood and happiness in the prediction of health-related quality of life. In addition, no interaction effects were found between the CPSS scores and life satisfaction in the prediction of psychological distress or health-related quality of life. Hence, higher scores on PTSD independently predicted psychological distress and health-related quality of life. The global score on the CPSS moderated the relationship between general mood and happiness and global psychological distress, which suggests that this negative relationship was stronger when CPSS scores were high. Thus, the results suggest that adolescents with higher scores on PTSD, as measured by the CPSS, would feel less happy, more distressed, and have a poorer quality of life.

This study has several limitations. Firstly, due to the logistic difficulties arising from this study being part of a larger project, the diagnostic utility of the CPSS was not assessed because of the lack of a diagnostic interview measure. The properties of the Spanish version of the CPSS should be examined in clinical samples. Secondly, measurement invariance tests were not conducted between the boys and girls because of the relatively small sample size. Future studies should include a larger sample to examine whether the CPSS items are invariant across populations. Thirdly, data were collected by self-report measures alone, which likely introduced some shared method variance across all the assessment measures. Finally, the test-retest reliability of the CPSS was not analysed, which is an important aspect for future research on its use in Spanish youth.

Despite these limitations, this Spanish version of the CPSS provides clinicians and researchers with a valid and reliable measure of PTSD among Spanish adolescents. Moreover, it integrates the four-factor structure corresponding to the dimensions of PTSD described in the DSM-V (APA, 2013).

Acknowledgements

We thank Antonio J. Lechuga for his assistance in data collection. The first author received a grant from the Spanish Ministry of Education, Culture, and Sports (FPU13/04928).

References

Alisic, E., Zalta, A. K., van Wesel, F., Larsen, S. E., Hafstad, G. S., Hassanpour, K., & Smid, G. E. (2014). Rates of post-traumatic stress disorder in trauma-exposed children and adolescents: Meta-analysis. The British Journal of Psychiatry, 204, 335-340. doi:10.1192/bjp.bp.113.131227

American Psychiatric Association (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA: Author.

Atienza, F. L., Pons, D., Balaguer, I., & García-Merita, M. L. (2000). Propiedades psicométricas de la Escala de satisfacción con la vida en adolescentes [Psychometric properties of the satisfaction with life scale in adolescents.]. Psicothema, 12(2), 314-319.

Buelga, S., Cava, M. J., & Musitu, G. (2012). Validación de la Escala de victimización entre adolescentes a través del teléfono móvil y de internet [Validation of the Scale of victimization among adolescents through the mobile phone and the Internet]. Revista Panameña de Salud Pública, 32, 36-42. doi:10.1590/S1020-49892012000700006

Cava, M. J., Musitu, G., & Murgui, S. (2007). Individual and social risk factors related to overt victimization in a sample of Spanish adolescents. Psychological Reports, 101, 275-290. doi:10.2466/pr0.101.1.275-290

Erhart, M., Ottova, V., Gaspar, T., Nickel, N., & Ravens-Sieberer, U. (2009). The HBSC Positive Health Focus Group Measuring mental health and well-being of school-children in 15 European countries: Results from the KIDSCREEN-10 Index. International Journal of Public Health, 54, 160-166. doi:10.1007/s00038-009-5407-7

Fernández-Fuertes, A. A., Fuertes, A., & Pulido, R. F. (2006). Evaluación de la violencia en las relaciones de pareja adolescentes. Validación del Conflict in Adolescent Dating Relationships Inventory (CADRI)– versión española [Assessment of violence in adolescent dating relationships. Validation of the Spanish version of the Conflict in Adolescent Dating Relationships Inventory (CADRI)]. International Journal of Clinical and Health Psychology, 6(2), 339-358.

Foa, E. B., Johnson, K. M., Feeny, N. C., & Treadwell, K. R. (2001). The Child PTSD Symptom Scale: A preliminary examination of its psychometric properties. Journal of Clinical Child Psychology, 30, 376-384. doi:10.1207/S15374424JCCP3003_9

Gudiño, O. G., & Rindlaub, L. A. (2014). Psychometric Properties of the Child PTSD Symptom Scale in Latino Children. Journal of Traumatic Stress, 27, 27-34. doi:10.1002/jts.21884

Gillihan, S. J., Aderka, I. M., Conklin, P. H., Capaldi, S., & Foa, E. B. (2013). The Child PTSD Symptom Scale: Psychometric properties in female adolescent sexual assault survivors. Psychological Assessment, 25, 23-31. doi:10.1037/a0029553

Helpman. L., Rachamim, L., Aderka, I. M., Gabai-Daie. A., Schindel-Allon, I., & Gilboa-Schechtman, E. (2015). Posttraumatic symptom structure across age groups. Journal of Clinical Children and Adolescent Psychology, 44, 630-639. doi:10.1080/15374416.2014.883928

Hukkelberg, S. S., & Jensen, T. K. (2011). The Dimensionality of Posttraumatic Stress Symptoms and their relationship to depression in children and adolescents. Journal of Traumatic Stress, 24, 326-333. doi:10.1002/jts.20637

International Society of Traumatic Stress Studies (2012). Child PTSD Symptom Scale. Retrieved from www.istss.org/assessing-trauma/child-ptsd-symptom-scale.aspx

Kadak, M. T., Boysan, M., Ceylan, N., & Çeri, V. (2014). Psychometric properties of the Turkish version of the Child PTSD Symptom Scale. Comprehensive Psychiatry, 55, 1435-1441. doi:10.1016/j.comppsych.2014.05.001

Kassam-Adams, N., Marsac, M. L., & Cirilli, C. (2010). PTSD symptom structure in injured children: Relationships with functional impairment and depression symptoms. Journal of the American Academy of Child and Adolescent Psychiatry, 49, 616-625. doi:10.1016/j.jaac.2010.02.011

King, D. W., Leskin, G. A., King, L. A., & Weathers, F. W. (1998). Confirmatory factor analysis of the clinician-administered PTSD scale: Evidence for the dimensionality of posttraumatic stress disorder. Psychological Assessment, 10, 90-96. doi:10.1037/1040-3590.10.2.90

López-Zafra, E., Pulido, M., & Berrios, P. (2014). EQI-versión corta (EQI-C). Adaptación y validación al español del EQ-i en universitarios [EQI- short form (EQI-C). Spanish adaptation and validation of the EQ-i in university students]. Boletín de Psicología, 110, 21-36.

Meyer, R. M. L., Gold, J. I., Beas, V. N., Young, C. M., & Kassam-Adams, N. (2014). Psychometric evaluation of the Child PTSD Symptom Scale in Spanish and English. Child Psychiatry & Human Development, 46, 438-444. doi:10.1007/s10578-014-0482-2

Nixon, R. D., Meiser-Stedman, R., Dalgleish, T., Yule, W., Clark, D. M., Perrin, S., & Smith, P. (2013). The Child PTSD Symptom Scale: An update and replication of its psychometric properties. Psychological Assessment, 25, 1025-1031. doi:10.1037/a0033324

Nixon, R. D., Sterk, J., & Pearce, A. (2012). A randomized trial of cognitive behaviour therapy and cognitive therapy for children with posttraumatic stress disorder following single-incident trauma. Journal of Abnormal Child Psychology, 40, 327-337. doi:10.1007/s10802-011-9566-7

Rachamim, L., Helpman, L., Foa, E. B., Aderka, I. M., & Gilboa- Schechtman, E. (2011). Validation of the Child Posttraumatic Symptom Scale in a sample of treatment-seeking Israeli youth. Journal of Traumatic Stress, 24, 356-360. doi:10.1002/jts.20639

Satorra, A., & Bentler, P. M. (2001). A scaled difference chi-square test statistic for moment structure analysis. Psychometrika, 66, 507-514. doi:10.1007/BF02296192

Simms, L. J., Watson, D., & Doebbeling, B. N. (2002). Confirmatory factor analyses of posttraumatic stress symptoms in deployed and nondeployed veterans of the Gulf War. Journal of Abnormal Psychology, 111, 637-647. doi:10.1037//0021-843X.111.4.637

Stewart, R.W., Ebesutani, C., Drescher, C. F., & Young, J. (2015). The Child PTSD Symptom Scale: An investigation of Its psychometric properties. Journal of Interpersonal Violence, 12, 1-20. doi:10.1177/0886260515596536

Ullman, J. B. (2006). Structural equation modeling: Reviewing the basics and moving forward. Journal of Personality Assessment, 87, 35-50. doi:10.1207/s15327752jpa8701_03

Vargas, B. E., Villamil, V., Rodríguez, C., Pérez, J., & Cortés, J. (2011). Validación de la escala Kessler 10 (K-10) en la detección de depresión y ansiedad en el primer nivel de atención. Propiedades psicométricas [Validation of the Kessler 10 Scale (K-10) for the detection of depression and anxiety in primary care. Psychometric properties]. Salud Mental, 34(4), 323-331.

Yufik, T., & Simms, L. J. (2010). A meta-analytic investigation of the structure of posttraumatic stress disorder symptoms. Journal of Abnormal Psychology, 119, 764-776. doi:10.1037/a0020981