A Rasch Analysis of the Secondary Traumatic Stress Scale with Child Welfare Supervisors and Caseworkers

Background:Secondary traumatic stress (STS) refers to the symptoms and reactions (similar to PTSD) that result from secondary exposure to traumatic material and experiences. Due to daily exposure to children and families who have endured traumatic experiences, child welfare workers and supervisors are at high risk for experiencing STS and the often negative effects. The Secondary Traumatic Stress Scale (STSS) has been used extensively over the past decade to identify STS in human service and medical professionals as well as the military and first responders.

The STSS is a 17-item measure of the frequency and extent to which a person experiences secondary traumatic stress symptoms such as feeling jumpy or annoyed, or having trouble sleeping or concentrating. The scale has demonstrated evidence of reliability with a Cronbach’s alpha = .93 (Bride, Robinson, Yegidis & Figley, 2004). However, the STSS has not undergone empirical testing using item response theory (IRT). IRT, in our case, the Rasch model, is a more sophisticated statistical analysis than classical test theory and provides a more detailed scoring of constructs and persons. The purpose of this study was to use Rasch analysis to examine the structure of the STSS and assess whether the STSS works similarly for both child welfare caseworkers and supervisors.

Method: Study participants included 1,045 child welfare supervisors and caseworkers from the Midwest who received the STSS as part of a larger survey, the Comprehensive Organizational Health Assessment (COHA). Participants were recruited via email invitation to complete the survey online as part of agency-wide participation in the COHA. Using Winsteps software, we assessed the overall structure of the STSS, running multiple rating scale analyses, and including a split samples analysis and differential item functioning to assess model invariance amongst different person variables.

Results:The STSS met expectations for unidimensionality, indicating that the items measure one construct, explaining 56.8% of the variance. The analysis demonstrated satisfactory item and person fit values, appropriate use of the response scale (all response options were endorsed), and no differential item functioning across person class variables, including: participants with extreme scores on the STSS, years of child welfare experience, previous personal trauma history, number of families on caseload, employment position, or educational degree. This indicates invariance, or the ability of the scale to measure distinct groups similarly.

Conclusion: Effectively measuring secondary traumatic stress symptoms across diverse populations in the child welfare workforce will help inform prevention and treatment efforts impacting employee well-being, job satisfaction and retention, quality services, and ultimately, child and family outcomes. The results from this study reinforced the already established psychometric properties of the STSS and newly illustrated its ability to provide a similar measure of STS across different child welfare staff. Discussion will focus on implications for the utility of this tool in professions that deal with traumatized clients and the value of the Rasch model to analyze measures which are used in diverse populations.