Examining the Factor Structure and Measurement Invariance of the Trauma Symptom Checklist for Children in a Diverse Sample of Trauma-Exposed Adolescents

Abstract

The Trauma Symptom Checklist for Children (TSCC) is a widely used youth assessment of broad, transdiagnostic symptomatology following trauma. However, its factor structure has not been thoroughly tested in diverse samples. Youth ( N = 738) exposed to interpersonal violence, including physical and sexual abuse, completed the TSCC. Confirmatory factor analysis was used to test one-, six-, and eight-factor models of the TSCC clinical scales, based on previous literature and the TSCC manual. We examined measurement invariance across boys and girls and Black and non-Black participants, as well as convergent and discriminant validity. An eight-factor structure, consisting of posttraumatic stress, anxiety, depression, anger, overt dissociation, fantasy dissociation, sexual preoccupation, and sexual distress, demonstrated the best fit, with two items removed. Invariance tests supported configural and metric (but not scalar) invariance. This research highlights the need for further testing before differences between gender and racial groups can be accurately compared.

Psychometric Analysis of the Dating Anxiety Scale for Adolescents in Samples of Polish and U.S. Young Adults: Examining the Factor Structure, Measurement Invariance, Item Functioning, and Convergent Validity

“…Specifically, three levels of invariance, that is, configural, metric (weak), and scalar (strong) invariance were tested. In this study, we did not test for strict invariance, which assesses invariance of item error variances and covariances, since this level of invariance is unlikely to be achieved with real data and is regarded as overly restrictive (e.g., Gere & Mac-Donald, 2012;Morelli et al, 2020). The scheme for this analysis was performed following the standard practices that assumed the examination and comparison of increasingly restrictive models in subsequent steps of analyses (e.g., Gere & MacDonald, 2013).…”

“…Specifically, three levels of invariance, that is, configural, metric (weak), and scalar (strong) invariance were tested. In this study, we did not test for strict invariance, which assesses invariance of item error variances and covariances, since this level of invariance is unlikely to be achieved with real data and is regarded as overly restrictive (e.g., Gere & MacDonald, 2012; Morelli et al, 2020).…”