Construct validity and factor structure of sense of coherence (SoC-13) scale as a measure of resilience in Eritrean refugees living in Ethiopia

Abstract

Background: There is a scarcity of adapted measures to study resilience and mental health of people in humanitarian settings in Africa. The aim of this study was to identify the factor structure and other psychometric properties of the Sense of Coherence (SoC-13) scale in Eritrean refugees living in Ethiopia. Methods: In a cross-sectional survey, 562 adults were selected randomly from Eritrean refugees living in Mai Aini camp, Ethiopia. The SoC-13, the Center for Epidemiologic Studies Depression scale (CES-D), the Pre and Post-Migration Living Difficulties checklist, the Primary Care Post-Traumatic Stress Disorder screener (PC-PTSD), the Oslo Social Support Scale (OSS-3), the Coping Style scale and the Fast Alcohol Screening Test (FAST) were administered concurrently. Confirmatory Factor Analysis (CFA) was used to investigate the factor structure of the SoC-13 using IBM SPSS Amos, version 21. Result: A one factor model of the SoC with twelve items had the best fit to the current data (CFA = 0.982, RMSEA = 0.035 [90%CI = 0.018, 0.050]), with good internal consistency (Cronbach's Alpha = 0.75). When all 13 items of the Tigrigna version were considered, there was an inverse association of SoC-13 with PC-PTSD(r = - 0.433, p < 0.001), CES-D (r = - 0.597, p < 0.001), Pre and post-migration living difficulties checklist (r = - 0.265, p < 0.001 respectively), and FAST (r = - 0.105, p < 0.001), providing support for the divergent validity of the scale. The SoC-13 was associated positively with the Oslo Social Support scale (OSS-3)(r = 0.363 p < 0.001) and task-oriented coping (r = 0.089, p < 0.001), demonstrating convergent validity. The four items, specifically item-1, item-2, item-3 and item-12 have shown relatively weaker item loadings (β<0.40); but item-2 demonstrated non-significant loading (β = 0.06, p>0.05) in a one factor model of SoC-13. Conclusions: Although the 13-items of the Tigrigna version of the SoC scale loaded significantly onto their respective factors in the three factor model, only 12 items loaded significantly onto the one factor model, which demonstrated superior fit to the current data. Keeping in mind that future research should examine the conceptualizations of the four items demonstrating poor convergent validity in this Eritrean sample, the reduced Tigrigna version of SoC-12 is a reasonable measure of sense of coherence in this community.