Abstract
This article applies structural equation modeling techniques to explore the relationships between constructs related to family strength and cohesion, acculturation, and trauma symptoms. The study uses a purposive sample comprised of 122 immigrant men and women from the Mexican community based in a large midwestern metropolitan Area. The data were collected using standardized instruments: Family Adaptability and Cohesion Evaluation Scale (FACES IV), The Trauma Symptom Inventory (TSI), and the Short Acculturation Scale for Hispanics (SASH). A factor analysis was conducted with three scales and the modified versions were found to be a better fit for this sample. The model reporting good fit indices is presented in this article. Cohesion and anxiety were found to be significantly and inversely related (.574), in other words family cohesion explained 57% of the variance in TSI. The indicators of the latent variable FACES shared variance of more than 47% with the underlying latent factor. Although the latent factor of TSI informed the latent factor of intrusive thoughts and the indicators of anxiety from 36% to 83%, the model found a significant inverse relationship between the constructs of family strengths and trauma. Level of acculturation (as measured by the Acculturation Scale) did not report a significant relationship with the other two constructs. The implications for research, public policy, and clinical practice with Mexican immigrant families are discussed.
ACKNOWLEDGEMENTS
This study was done in collaboration with the Mexican Consulate of Chicago and the Sistema Nacional para el Desarrollo Integral de la Familia (SNDIF), the Mexican federal government's principal agency for protecting the well-being of Mexicans and providing essential human services. Some of the questionnaire items used in this study come from the national SNDIF study of Mexican Family Dynamics (Diagnostico de la Familia), conducted in Mexico in June 2005. The study was supported by a grant from the Global Initiatives Incentive Fund, Loyola University Chicago. Our appreciation extends to SNDIF, the Mexican Consulate of Chicago, Loyola University, and all those students who were helpful in completing this research.
Notes
1. The number of U. S. residents of Mexican origin in 2008 was 30.7 million, constituting 10% of the nation's total population and 66% of the Hispanic population. 11.3 million of U. S. residents of Mexican origin were foreign born. (http://mexican-american-proarchive.com/2011/02/facts-about-mexican-americans-from-the-census-bureau-2010/).
2. Family cohesion has been defined as the emotional bonding that family members have towards one another.
3. This pilot study was done in collaboration with the Sistema Nacional para el Desarrollo Integral de la Familia (SNDIF), the Mexican federal health and human service agency and the Mexican Consulate of Chicago.
4. The model fit was determined by assessment of goodness of fit. “Goodness of fit” tests determine if the model being tested should be accepted or rejected. These overall fit tests do not establish that particular paths within the model are significant. If the model is accepted, the researcher will then go on to interpret the path coefficients in the model (“significant” path coefficients in poor fit models are not meaningful). Chi Square indices are an indicator of an absolute fit, i.e. the model specified fits the variance and covariance structure observed in the data. The chi square does not have an upper bound, but a smaller chi square statistic is desirable. It is affected by sample size and so a larger sample size results in larger Chi square values. Therefore other supplemental indices of model are also reported and discussed in all Structural Equation Model analysis. All the fit indices are reported, however given the sample size (<150) and model simplicity, absolute fit index of chi-square, CMIN, and RMSEA, the baseline fit measures comprising IFI and TLI and CFI; and parsimony measures PNFI and PCFI to compare initial and final model fits, and the information theory measures including AIC and ECVI are reported. Given the RMSEA is not sensitive to smaller sample sizes, the ECVI can be a better test indices for the likelihood that the model can be found in another sample or population. Lower levels of ECVI are preferred in comparing models (CitationSchermelleh-Engel, Moosbrugger, & Muller, 2003).