Assessing posttraumatic stress disorder with or without reference to a single, worst traumatic event: Examining differences in factor structure.

Abstract

The authors examined the effects of a methodological manipulation on the Posttraumatic Stress Disorder (PTSD) Checklist's factor structure: specifically, whether respondents were instructed to reference a single worst traumatic event when rating PTSD symptoms. Nonclinical, trauma-exposed participants were randomly assigned to 1 of 2 PTSD assessment conditions: referencing PTSD symptoms to their worst trauma (trauma-specific group, n = 218) or to their overall trauma history in general (trauma-general group, n = 234). A 3rd group of non-trauma-exposed participants (n = 464) rated PTSD symptoms globally from any stressful event. Using confirmatory factor analysis, the authors show that the 4-factor PTSD model proposed by D. W. King, G. A. Leskin, L. A. King, and F. W. Weathers (1998; separating effortful avoidance and emotional numbing) demonstrated the best model fit for trauma-general and non-trauma-exposed participants. The 4-factor PTSD model proposed by L. J. Simms, D. Watson, and B. N. Doebbeling (2002; emphasizing a general dysphoria factor) demonstrated the best model fit for trauma-specific participants. Measurement invariance testing revealed that non-trauma-exposed participants were different from both trauma-exposed groups on factor structure parameters, but trauma groups were not substantially different from each other.