Date Published: October 8, 2018
Publisher: Public Library of Science
Author(s): Susan Sierau, Lars Otto White, Annette Maria Klein, Jody Todd Manly, Kai von Klitzing, Philipp Yorck Herzberg, Mary C. Smith Fawzi.
Child victims’ reports of psychological and physical abuse by caregivers are a fundamental source of information beyond official records and caregiver reports. However, few or no sensitive and age-appropriate child-report instruments exist that have undergone in-depth validity and reliability testing across a broad age-range. Our study addresses this gap by examining psychometric properties of a picture-based, modularized version of the Parent-Child Conflict Tactics Scale (CTSPC-R), encompassing the maltreatment subtypes of psychological and physical abuse. A sample of 904 children and adolescents aged 4–16 years from the community (n = 568), child psychiatric services (n = 159), and from Child Protective Services (CPS; n = 177) completed the CTSPC-R. Measures to test convergent (maltreatment in parent interviews and CPS records) and concurrent validity (psychiatric symptoms) were collected. The CTSPC-R comprises 22 items, arranged in three severity modules by increasing level of psychological and physical abuse by caregivers. Companion picture cards were provided for children aged 4 and 8 years. The best fit to the data was attained with a second-order factor model, assuming three inter-correlated factors corresponding to the three severity modules, and a latent second-order factor representing combined physical and psychological abuse. The three factors showed good internal consistencies. Supporting convergent validity at the global and subtype-level of maltreatment, the CTSPC-R severity scale was associated with lifetime CPS-contact, presence of caregiver-reported emotional maltreatment and physical abuse, and dimensions of chronicity and severity. Discriminant validity was supported by non-significant correlations with caregiver-reported lack of supervision, failure to provide, and sexual abuse. Bolstering concurrent validity, moderate and severe physical abuse predicted caregiver-reported internalizing and externalizing symptoms. These effects were independent of child age, gender or community vs. non-community samples. Our study supports the CTSPC-R as a scientifically and clinically sound tool for ascertaining the child’s own perspective on psychological and physical abuse from an early age onwards.
Childhood exposure to psychological and physical abuse represents a major risk for poor physical and mental health throughout the life course [1–4]. Yet despite UNICEF’s call for the “child’s right to be heard” in cases of violence , very few, if any, instruments currently exist that have proven valid and reliable in sampling the child’s perspective on maltreatment. This study aimed to fill this crucial gap by adapting an existing instrument for use across a wider age-range from preschool-age to adolescence as well as across low-risk community and high-risk child psychiatric/CPS settings.
Our study provides a psychometric evaluation of a child-report measure of psychological and physical abuse appropriate for a broad age-range from 4 to 16 years in a heterogeneous sample of community and at-risk families. Our study therefore broadens our ability to assess children’s appraisal of physical abuse within the family context. For assessing current incidences of psychological and physical assault in a broad age range from early childhood (starting from age 4) to adolescence, we a) developed child-appropriate picture drawings to scaffold a better understanding of item content for young children, and b) implemented a sequential step-by-step, hierarchical presentation of items, arranged in three ascending levels of maltreatment severity. A confirmatory factor analysis supported the three-factor solution of the assumed three modules and revealed a latent second-order factor explaining the correlation between the factors. The second-order latent factor represents the global concept of psychological and physical abuse, and the co-occurrence of maltreatment subtypes, which is common in the literature. Yet, the inter-correlation between the three factors was moderate, indicating that these scales represent sufficiently distinct entities. Children, on the one hand, may have difficulties distinguishing between different subtypes of maltreatment, but, on the other hand, seem to be able to rate the severity of incidences. Further psychometric properties of the factors revealed good internal consistencies (Cronbach’s α between .66 and .75; McDonalds’ Ω between .75 and .78).