The Estimation Of Disability Rate Of TianJin Based On Small Area Model

The disabled person is one of the mankind's social members, they require social concerns and helps. For many years, Chinese government devotes to the establishment and consummates of safeguard system of disabled person's human rights gradually, and helps the disabled person to participate in the social life equally.This article uses data from the national sample survey of the disabled(Tianjin volume), then carries out a statistical analysis of the Tianjin disabled. Results show that, the disabled person who belongs to disadvantaged groups, still faces many difficulties on education, marriage as well as life guarantee etc, and will encounter discrimination and the prejudice frequently, so the government and the community should give them the universal care. So before the implementation of policies, all levels of the governments must understand the situations of their own jurisdiction clearly, which leads to the problem of multistage inferences, which is also called small area estimation.Small area estimation is one of the important frontier directions in the field of sample survey, this article summarizes the theory of small area estimation, in which small area model is the hot spot. The small area model estimation is that according to the distribution of the dependent variable, and establish appropriate statistical model by using the subsidiary variables, then carry on the inference using the model to the target variable of every small area.This article estimates the disability rate of every district of Tianjin using some methods of small area estimation, and outlines the performance of each of methods and discusses their advantages and disadvantages. For direct estimation, the estimation of disability rate is relative stabilization in the district with large population. The traditional indirect estimation methods include synthetic regression estimation with no auxiliary data and with auxiliary data (age/sex). The difference of the results of these two methods is produced by the age/sex. Poisson model and Logistic model are two estimations based on small area model, and Poisson model utilizes auxiliary data highly related to disability rate, drawing a very good conclusion. By using Logistic model, this article makes the relationship between disability rate and age.