Abstract:
|
The management and assessement of new health care initiatives, such as the State Child Health Insurance Program (SCHIP), have created the need for estimates of the uninsured at the county and subcounty levels. Our ultimate goal is to develop a methodology that satisfies state and local needs for estimates at the county level that can be implemented within their level of expertise, budget, and other constraints.
The focus of this paper is to explore the incorporation of additional covariates in the synthetic estimates system that has been developed and tested for validity and reliability using logistic regression and goodness-of-fit tests (see, for example, Sigmund, Judson, and Popoff, 1999; and Judson, Popoff, and Fadali, 2000, 2001). The literature suggests that firm size, the rate of unemployment, and other worker characteristics, such as part-time or seasonal jobs, versus full-time jobs, may co-vary with the proportion of the population that is uninsured. We develop a synthetic system that incorporates one or more of these variables and test the model in the same manner as before using logistic regression. We also test for across-state differences using hierarchical linear modeli.
|