TPC Journal V7, Issue 4 - FULL ISSUE

338 The Professional Counselor | Volume 7, Issue 4 analyses of variances and regressions occurring simultaneously. Through this, researchers also can test mediated relationships between variables and examine the reliability of items to latent variables in a single test. Several fit indices are reported in SEM as a global examination of how well the collected data fit the hypothesized model (Hu & Bentler, 1998; Schreiber, 2008; Sivo, Fan, Witta, & Willse, 2006). Because of well-known problems of fit estimation using chi square (χ2) in large data sets, the Bentler’s Comparative Fit Index (CFI) was used with values at .95 or higher, indicating a good-fitting model (Hu & Bentler, 1998; O’Rourke & Hatcher, 2013; Sivo et al., 2006). An absolute index, the Standardized Root Mean Square Residual (SRMR), and a parsimony index, the Root Mean Square Error of Approximation (RMSEA), were used as indices assessing the difference between the proposed model and actual variances and covariances in the data. RMSEA values less than or equal to .06 and SRMR values less than or equal to .08 are preferred (Fabrigar, Porter, & Norris, 2010; O’Rourke & Hatcher, 2013; Sivo et al., 2006). Results Demography of Participants Participants’ ages ranged between 28 and 65 years. Many fell within the 31–40 year range (45.7%). More participants were male (70.9%) than female, and many were unemployed (63.7%). Of those who were employed, 34.0% earned less than US$720 a year. Family demographics were varied: 42.3% of participants were single, 30.0% were married, and 23.6% lived with their partners. Many participants (49.3%) had one or two children. The majority of participants (89.2%) indicated they were displaced during the Liberian civil war. There was no follow-up question to identify how some participants could have stayed in their homes despite the reported rampage that affected the whole country (Verdier et al., 2008), so I can only speculate that they either did not read that question well or had a way to stay safe during the war. Of those who were displaced, 60.4% became internally displaced, while the rest became refugees in and out of Africa. Almost half (48.4%) of the former refugees and IDPs had returned to Liberia since the cessation of the war. Additionally, to assess participants’ reaction to the Ebola epidemic and its possible ability to trigger traumatic experiences (Doucleff, 2015), participants were asked to compare the effect of the epidemic to the effect of the civil war on their mental health. The majority (65.8%) indicated that they linked the epidemic to the war several times a day, a process that could re- traumatize those who may have undisclosed PTSD symptoms. Results From Hypotheses The first hypothesis proposed that between 10–50% of participants in the study will meet the diagnostic criteria for PTSD (as measured by PCL-5), as well as co-occurring symptoms of distress and growth based on gender. Results of the descriptive statistics indicated that the majority (351; 79.1%) met the criteria for PTSD. Additionally, results from the one-way multivariate analysis of variance indicated that there was a statistically significant difference in post-trauma pathways based on gender ( F [2, 376] = 6.016, p = .005; Wilk’s λ = .972, partial η2 = .028). Gender had a statistically significant effect on PTG ( F [1, 377] = 6.354, p = .012, partial η2 = .017), but not on PTSD ( F [1, 377] = 3.039, p = .082, partial η2 = .008). The PTG mean score for females ( X̅ = 84.49, sd = 16.030, n = 109) was higher than for males ( X̅ = 79.56, sd = 17.663, n = 270). Both males and females demonstrated co- occurring PTG and PTSD. Even though there was a statistically significant difference in PTG scores based on gender, there was no such difference in PTSD scores.