Convergent and Discriminant Validity of Police-Initiated Post-Traumatic Stress Symptomatology

Background: Post-traumatic stress disorder (PTSD) refers to problems with physiological, emotional, and behavioral reactivity and avoidance following a traumatic experience. PTSD has strong negative effects on youths’ physical health and mental health that can last into adulthood. Prior research suggests that interactions with the police can be traumatic experiences for youth. However, prior research on policing and trauma are limited because they do not measure trauma symptomatology that is explicitly linked to interactions with police. The present study explores the validity of a measure of police-initiated post-traumatic stress symptomatology (PI-PTSS). This study will lay the groundwork for further exploration as police encounters as traumatic events that have a unique impact on youths’ mental health.

Method: Data for this study are drawn from the Fragile Families Child Wellbeing Study (FFCWS). The FFCWS is a longitudinal survey that follows an ethnically and racially diverse birth cohort of children, and their families, through age 18. The study sample consisted of youths who had been stopped at least one time by the police and did not have missing data on any of the analysis variables (n = 871). The majority of youth in the sample are black (54.8%, n = 477), 69.6% (n = 453) of the sample are male, and roughly half of the sample is between the ages of 16 and 19 years old (52.0%, n = 453).

PI-PTSS was measured using nine items specifically assessing how frequently youths experienced post-traumatic stress symptoms in the context of their interactions with the police. The measure taps into the constructs of persistently re-experiencing the encounter, and negative changes in cognition and mood following the encounter. Depression was measured using five items from the CES-D, and anxiety was measured using six items from the brief symptom inventory.

We first ran three exploratory factor analyses (EFA) on the measures of PI-PTSS, CES-D, and anxiety – testing two, three, and four factor solutions. We then conducted two confirmatory factor analyses (CFA) to identify the factor structures among the constructs.

Results: Findings from the EFA suggest a two factor solution. Factor 1 consisted of items from the PI-PTSS scale whereas factor 2 consisted the anxiety and CES-D items. The two factor CFA was based on the CATPCA demonstrated acceptable model fit (RMSEA = 0.065, CFI = 0.924, TLI = 0.914), as did the three factor model (RMSEA = 0.057, CFI = 0.941, TLI = 0.933). We decided to retain the three factor solution because of the consistency with prior research, and the increase in model fit.

Conclusions: Our findings support the independence of PI-PTSS as a construct. In terms of convergent validity, indicators of PI-PTSS load on the same factor in both EFA and CFA. We also established the divergent validity of PI-PTSS because the PI-PTSS was supported as a factor unique from anxiety and depression. Future research can continue to explore the potential relationships among PI-PTSS and youth outcomes such as delinquency and recidivism.