부산광역시 사회조사 표본설계 개선에 관한 연구

Abstract

This study conducted the『2009 Busan Social Survey』to the samples who were chosen for a new sampling design, analyzed the result, found the derived problems, and estimated it. The purpose of the study is to prepare for a new sampling design of『2012 Busan Social Survey』which will be done by the Busan and which will be based on the 2010 Busan population and housing censes 100% enumeration district as a sampling fraction. The result of『2009 Busan Social Survey』, the sample size as 1,068 households has ±3.0% point margin of error within general reliability 95% of country statistics under confidence level. It is good for Busan to use quinquennial population and housing census enumeration district and two-stage cluster sampling according as stratification variable within gugun like National Statistical Office. The result from comparing the 3 variabels except Housing types and agriculture and forestry and fishing household proportion in 5 classification variables, it is right to use a householder level of education as the third stratification variable like sample modeling suggestion. Using one question in a household unit(a monthly average gross income) and a member of household unit(the point of view on the economy of Busan), the mean estimate was similar to a sample mean, but the standard error estimate was large. However, the result compared with efficiency, the sample modeling was right to have efficiencies up to 80%. The Busan should extract new sampling design which can reflect the suggested matters and solve the problems for the correct result.