Examining Factor Structure of the Impact of Event Scale-Revised in Adult Male Survivors of Childhood Sexual Abuse: Confirmatory Factor Analysis

Background & Purpose

The Impact of Event Scale Revised (IES-R; Weiss & Marmar, 1997) is a 22-item self-report (Likert-type) measure that assesses the extent of the symptomatic responses of Posttraumatic Stress Disorder (PTSD) based on the three DSM diagnostic criteria (avoidance, intrusion, and hyperarousal). As one of the most frequently used scales assessing the symptoms of PTSD, the IES-R has been translated into several languages including Chinese, French, German, Japanese, Spanish, Dutch, Italian, Norwegian, and Persian (Weiss, 2007). However, unlike the proposed three factor model  by the developers (Weis & Marmar, 1997), validation studies (including the English version and international versions of the IES-R) show inconsistencies and disagreement regarding its psychometric properties (factor structure) from one factor model to five factor model (e.g., Amdur & Liberzon, 2001; Creamer et al., 2003; King et al., 2009; Weiss & Marmar, 1997). Therefore, the purpose of this research is to evaluate the psychometric properties of the IES-R to confirm the original author’s three factor model.   

Methods

Secondary data analysis is used as a research method. The study sample includes adult male survivors of childhood sexual abuse who have received counseling services from 23 counseling services in the west region of Ontario Canada from September 2011 to July 2015 (N = 310). The data was collected through the lead organization that coordinates service provision within 23 counseling services in the west region, Ontario. The researchers used the demographics questions, Impact of Event Scale Revised (Weiss & Marmar, 1997) and Self-Report Instrument for Disorders of Extreme Stress (Spinazzola, Blaustein, Kisiel, & van der Kolk, 2001). SPSS-23 and AMOS-23 were used for reliability test, Exploratory Factor Analysis, and Confirmatory Factor Analysis.

Results

Internal consistency using Cronbach’s alpha for the three subscales ranges from “acceptable” to “good” level: Avoidance (α = .79), Intrusion (α = .89), and Hyperarousal (α = .87). However, the model fit indexes for the three factor model show “poor” except for CFI (.904): RMR (.100), GFI (.879), NFI (.862), TLI (.877), and RMSEA (.078). The further analysis confirms that the data supports the four factor model: RMR (.088), GFI (.905), NFI (.903), TLI (.935), CFI (.949), and RMESA (0.057).

Conclusions & Implications

The IES-R has good internal consistency. However, the confirmatory factor analysis shows that the current data does not support the three factor model but the four factor model. This result is consistent with several studies that do not support the authors’ proposed three factor model. However, there are variations among different studies, samples, and researchers regarding the most effective factor model and factor structure. It is evident that there are needs to be more research completed to verify the psychometric properties of the IES-R. The present study is unique and meaningful because there has been no validation study conducted to confirm the three factor model of the IES-R by using the adult survivors/victims of childhood sexual abuse since the introduction of Posttraumatic Stress Disorder (PTSD) in 1978.