標題: 共變異數分析下處理差異檢定之樣本數計算
Sample Size Calculation for Treatment Difference in the Analysis of Covariance
作者: 邵逸文
謝國文
管理科學系所
關鍵字: 共變異數分析;樣本數;統計檢定力;共變項;Analysis of Covariance;Sample Size;Power;Covariate
公開日期: 2013
摘要: 共變異數分析是各門學科經常使用的統計方法,用以排除與實驗研究無關因素的影響,本論文探討在共變異數分析下,估計處理差異檢定的最適樣本數。以往統計學教科書及文獻並無明確且詳細建構出一套針對檢定共變異數分析下處理差異的樣本數估計公式,而是利用近似的方法計算樣本數,但此樣本數在進行檢定時,無法達到預設的統計檢定力,導致研究結果產生偏誤。因此本論文將直接利用共變項的分配性質來取代近似的方式,進行樣本數估計公式的推導,並探導此樣本數估計公式的正確性,以及此樣本數公式是否適合應用於實際研究。
Analysis of covariance, a common statistical method, is used to reduce the impact of factors related to the experiments. For the purpose of computing the optimal sample sizes for testing the difference between two treatment means, exact procedure is presented to improve the existing approximate methods by accommodating the underlying distribution of related covariate. Supplemental program is provided to facilitate the application of the suggested sample size procedure.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT070153134
http://hdl.handle.net/11536/74372
顯示於類別:畢業論文