The multidimensional structure of the Life Satisfaction Index A and its application in research on aging
Three sections, in the form of complete papers for professional publication, form the core of this dissertation. Paper 1 used data from a national survey to compare three second-order factor models of the Life Satisfaction Index A (LSIA) through confirmatory factor analyses. Support was found for all three models, with the model suggested by Liang fitting slightly better than the other two;Paper 2 examined the stability of three LSIA dimensions, and their relationships to predictor variables, across age-groups, cohorts, and time. Stacked-group LISREL models were fit to data on 58 to 85 year-olds from two national surveys. Factor loadings and causal paths were found to be invariant across age and time for persons 65 years of age, and older. Factor loadings were invariant for those 58 to 64; however, causal paths varied;Paper 3 used a multidimensional LSIA model in research on economic stress and psychological well-being. Data were collected from rural Iowans during a period of rural economic stress. Age-specific regressions on the individual LSIA dimensions showed patterns of relationships that would not have otherwise been evident;Collectively, the three papers show a multidimensional model of the LSIA to be stable when applied to older adults. There may be differences in the factor structure of the LSIA for younger persons. There is also evidence that the determinants of life satisfaction vary across age-groups. It is strongly recommended that the LSIA dimensions be analyzed individually on methodological and substantive grounds. Results from previous research, using the LSIA as a unidimensional measure, should be viewed with caution.