頁籤選單縮合
題名 | 二次抽樣的成效評估: 以「國人血中鉛濃度調查」為例=Efficacy of Double Sampling Method in a Study of Blood Lead Levels of Taiwanese General Population |
---|---|
作者 | 劉紹興; 洪永泰; 吳德敏; 吳聰能; 江宏哲; 楊冠洋; 柯桂女; | 書刊名 | 中華職業醫學雜誌 |
卷期 | 5:3 1998.07[民87.07] |
頁次 | 頁165-177 |
分類號 | 412.78 |
關鍵詞 | 二次抽樣; 無反應者; 代表性; Double sampling; Nonrespondent; Representative; |
語文 | 中文(Chinese) |
中文摘要 | 研究者常利用抽樣方法以求在人力物力的限制下,能在短時間內得到研究結果 。本研究在進行國人血中鉛濃度調查時,收集相關資料以探討樣本的代表性,並對無反應 者的來源進行分析,以提供其他類似研究的參考。利用行政院主計處勞動力調查的樣本戶 進行二次抽樣,本研究發現空戶率是9.2%,遷移率是7.1%。因此,根據戶籍地址來抽樣的 無反應率約為16.3%,本研究亦發現有6.9%的拒訪率。無反應率和拒訪率均因縣市別和都 市化程度會有所差異。本研究利用電話訪問來求取問卷資料正確性的評估,320份抽樣問 卷中以檳榔史之錯誤率較高(17.8%),另一個錯誤率較高者是抽煙史(7.8%),其它問卷內 容的錯誤率均低於5%。 另外,本研究也探討研究對象是否具有代表性,利用問卷收集的特性與全國人口特性 分佈來比較,其中性別、抽煙和族氏的分佈均與全國人口的分佈非常相似;但年齡和職業 的分佈則有明顯差異。本研究對象在40歲以上的人口較多,而15歲至29歲的人口較少;本 研究對象中的從商者比率較全國人口少。本研究亦發現戶外替換樣本的性別和抽煙史有較 大的偏差,而戶中替換列在性別上有偏差。 |
英文摘要 | Sampling methods areusually applied in the epidemiological study to save time and cost. This article presented the gain and efficacy of double sampling method applied in a study of blood lead levels of Taiwanese general population. The vacancy rate of residential household found in this study was 9.2%. The migration rate was 7.1 %. The total non-response rate was 16.3%. The nonresponse rate varied among districts of residence and urbanization levels. In addition, there were 6.9% of sampled households refused to be visited. The accuracy of a random sample of 320 questionnaires completed by Public Health nurses were evaluated by telephone interview. The error was highest for the item of consumption of betel quit (17.8%), followed by the smoking status (7.8%). The error for other items were less than 5%. Representatives of the study population was evaluated by comparison with the demographic distribution of general population collected from governmental statistical publications. This study found that the distribution of gender, smoking status and ethnic group in study population was similar to those of general population. However, the study population was overrepresentative in elder age group and underrepresentative in commercial occupation and high education level. This study also found that substitution of sampled population by volunteers biased the distribution of gender and smoking status, while household member substitution biased the distribution of gender. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。