摘 要 為研究兩種處理效應的異同，可藉由處理平均數比值及其信賴區間加以衡量。本文則進ㄧ步研究，當兩個處理組的反應變數均受另一個共變數影響時，如何衡量此兩種處理效應之異同。本文考慮針對兩個處理組的資料，分別配適簡單線性迴歸模式，然後根據所估計之迴歸直線建構兩條迴歸直線的近似條件信賴區間及信賴束。文中除模擬研究外，也以一組資料說明其應用。 Abstract To study the difference of two treatment effects, one can measure the ratio of the two treatment means together with its confidence interval. This article investigates the problem of comparing two treatment effects when the responses are affected by one covariate. The two groups of data are fitted separately by two simple linear regression models. Based on the fitted liner regression lines, the conditional confidence intervals and confidence band are constructed. This article presents the simulation results and demonstrates the use of the proposed procedures.
