Graduation Year


Document Type




Degree Granting Department


Major Professor

Jason W. Beckstead

Co-Major Professor

Mary E. Evans


Confirmatory factor analysis, Cross cultural analysis, Eritrea, Immigrants, Structural equation modeling



Depression is one of the oldest known mental health conditions. It is acknowledged to be a global health problem that affects people from any culture or ethnic group. The prevalence of depression widely varied across countries and cultures. The cross-cultural relevance of the concept of depression, its screening or diagnosis, and cultural equivalence of items used to measure symptoms of depression has been area of research interest. Differences in prevalence rates in depression have been suggested as being due to research artifacts, such as use of instrument developed for one culture to another culture. With the current trend of globalization and increased rate of immigration, the need for measurement scales that can be used cross-culturally is becoming essential. Translation and adaptation of existing tools to different languages is time saving and cost effective than developing a new scale. The Center for Epidemiologic Studies Depression Scale [CES-D; (Radloff, 1977)] has been widely used as a screening tool for depressive symptoms in community and clinical settings. It has been widely accepted and translated to multiple languages and its measurement equivalence tested across groups. This study was designed as a mixed method study. The purpose of this study was three fold: (a) translate and adapt the CES-D scale into Tigrigna Language for use by Tigrigna speaking Eritrean immigrants in the United States using the forward backward translation and cognitive interview techniques (b) test the psychometric properties of the Tigrigna version CES-D scale using confirmatory factor analysis under the framework of structural equation modeling and (c) test measurement equivalence of the scale by comparing data collected from 253 Eritrean immigrants using the Tigrigna version CES-D scale with a secondary data collected from 1918 non Eritrean US citizens using the English version CES-D scale in a separate study. The baseline four factor CES-D scale model originally suggested for the general population fitted the data from both samples. The fit indices for the Tigrigna sample were (χ2 = 299.87, df = 164, RMSEA = .06, SRMR = .06, GFI = .89, and CFI = .98) and for the English sample (χ2 = 1496.81, df = 164, RMSEA = .07, SRMR = .04, GFI = .92, and CFI = .98). The Multi-group confirmatory factor analysis showed reasonably adequate fit (χ2 (328) = 1796.68, RMSEA= .07, SRMR = .06, GFI = .89, CFI = .98). Fourteen of the 20 CES-D items were invariant across the two samples suggesting partial metric invariance. Partial full factor invariance was also supported. In conclusion, the findings of this study provide adequate evidence in support of the applicability of the four factor CES-D scale for measuring depressive symptoms in Tigrigna speaking Eritrean immigrants/refugees in the United States.