A Study for Efficient Sample Design: Focusing on the National Survey of Disabled Persons

Abstract

The current environment, characterized by rapid changes in household conditions, unexpected spread of infectious diseases, and increased awareness of personal information protection, makes it hard and costly to conduct face-to-face surveys. Despite these circumstances, however, in-person survey data remain essential to policymaking. The National Survey of Disabled Persons covers more than 40,000 households. There are two ways to reduce the burden on a survey based on enumeration districts. The first is to reduce the sample size of household and discrimination surveys. The second is to expand the survey of the disabled based on data other than enumeration districts. The existing and proposed methods were examined in various ways. The existing method is a survey based on enumeration districts. The proposed method is a survey based on both enumeration districts and databases of individuals registered as disabled. The target population for both methods is composed of disabled persons. The existing method consists of a single frame and a double sampling technique. The advantage of the existing method is that it can calculate the prevalence of individuals with disabilities and produce reliable data. The disadvantage of this method is that it is costly and comes with the burden of having to cover an increasing number of households. A further study is needed to determine the effective sample size for the first phase. The proposed method employs a dual frame and a stratified multistage sampling technique. The method excels at calculating the prevalence of individuals with disabilities even with a limited budget and enables appearance rate of the disabled under the available budget and enables conduct additional surveys if necessary. The disadvantage of this method is that it requires more skilled manpower and time compared to the existing method, and the difficulty with survey management. From the findings from our simulation, we propose that the enumeration districts can be reduced by 30% compared to the National Survey of Disabled Persons (2017). A further study needs to be conducted on the effective survey sample size for unregistered disabled persons. We examined the population size of the disabled using the capture-recapture method. The results show that the proposed estimator is similar to the known population size (National Survey of Disabled Persons, 2017). The proposed estimator has a considerably smaller standard error.