台灣地區鄉鎮市區發展類型應用於大型健康調查抽樣設計之研究

Incorporating Development Stratification of Taiwan Townships into Sampling Design of Large Scale Health Interview Survey

自從羅啟宏（1992）提出「台灣省鄉鎮發展類型之研究」一文後，利用普查資料或是其他輔助資料進行台灣地區鄉鎮市區都市化發展型態之研究，逐漸受到國內各類大型調查在進行分層抽樣設計的重視與應用。為了使2005年「國民健康訪問調查」其樣本更具鄉鎮市區發展類型之代表性，本文使用台灣最近一次之「戶口普查資料檔」（2000年）與同時期之「台灣各縣市統計要覽」以及「衛生署醫療機構現況及醫療服務統計量」資料庫來進行台灣地區鄉鎮市區都市化程度分層之研究。研究變數包括人口密度（人/平方公里）、專科以上教育程度人口比率、65歲以上人口比率、農業人口比率與每10萬人西醫人數，並使用集群分析進行研究。結果將台灣地區359個鄉鎮市區分成七個都市化程度之集群：「高度都市化市鎮」、「中度都市化市鎮」、「新興市鎮」、「一般鄉鎮市區」、「高齡化市鎮」、「農業市鎮」與「偏遠鄉鎮」。本論文結果亦已被2005年「國民健康訪問調查」所採用，做為進行分層多階段、抽取率與單位大小成比例之等機率抽樣設計之參考。

Since Luo(1992) proposed using census data to investigate the developmentstratification of Taiwan townships, using census data or related auxiliary data toinvestigate the development stratification of Taiwan townships has been regarded withincreasing importance, recently. This paper used the 2000 Taiwan census data and dataof Survey of Health Maintenance Organizations’ current status and Health ServiceUtilization (conducted by Department of Health, Executive Yuan, R.O.C.) to investigatethe development stratification of Taiwan townships. The variables we chose in thispaper included: population density (people/km2)、population ratio of people withcollege or above educational levels 、 population ratio of elder people over 65 yearsold 、population ratio of people of agriculture workers and the number of physiciansper 100,000 people. We used the Cluster Analysis, with squared Euclidean distance andWald’s minimum variance method, to study the urbanization stratification and theresults showed 7 clusters were identified. The results have been adopted as the referenceframe of the multi-stage, PPS (Probability Proportional to Size) sampling design of the2005 National Health Interview Survey (NHIS) in Taiwan.