本篇使用了信賴區間評估準則來評估兩個近期推出的隨機樣本數之信賴區間程序。從使用電腦模擬開始，學者將統計方法應用在模擬出的結果。也因為如此，學者紛紛開始研究新的統計方法，從一開始的固定樣本數信賴區間程序到後來的隨機樣本數信賴區間程序。而有了這麼多的方法，也需要一套標準來評估他們的優劣。傳統的評估標準由許多面向來評估一個信賴區間程序的好壞，但多面向也帶來了不少缺點，像是各面向之間互相矛盾，這也讓信賴區間程序的發展受到了影響，因為無法設計出同時滿足各面向評估標準的信賴區間程序。後來 Yeh 和 Schmeiser 提出了單一準則的評估方法 VAMP1RE，讓評估信賴區間程序較為簡單，一開始 VAMP1RE 僅能使用在固定樣本數之信賴區間程序，後來隨著新的信賴區間程序發展，套用了序列停止條件，修改為隨機樣本數也可以使用的一般化 VAMP1RE。有了一般化的 VAMP1RE 之後，即可對近期的 ASAP3 以及 Skart 兩個信賴區間程序做評估比較。 本篇將先介紹信賴區間程序，從基本的 t 分配信賴區間到近期的 ASAP3 與 Skart，了解信賴區間程序是如何發展之後接著介紹信賴區間評估標準，從過去多面向的評估標準到後來單一準則的 VAMP1RE 以及一般化 VAMP1RE 最後使用一般化 VAMP1RE 來評估 ASAP3 與 Skart。藉由評估的結果判斷後期推出的信賴區間程序是否真的比其前面的信賴區間又更加準確。同時也藉由傳統評估方法與 VAMP1RE 比較，了解新的評估準則在評估結果是否會更方便判斷，提高判斷的效率。;This paper uses the confidence interval procedure criteria to evaluate and compare the two recently introduced random sample size confidence interval procedures. Beginning with the use of computer simulations, scholars apply statistical methods to the simulated results. Because of this, scholars have begun to study new statistical methods. From the beginning of the fixed sample size confidence interval procedure to the random sample size confidence interval procedure. With so many methods, a set of criteria is needed to assess their quality. The traditional criteria are used to evaluate the quality of a confidence interval procedure by many aspects, but many aspects also bring a lot of shortcomings, such as the contradictions between the various aspects. Later, Yeh and Schmeiser proposed a single criterion, VAMP1RE. At the beginning, VAMP1RE can only use the confidence interval procedure in the fixed sample size. With the development of the new confidence interval procedure, the sequential stopping rule is applied and the generalized VAMP1RE can be used to evaluate the random sample size confidence interval procedure. With the generalized VAMP1RE, we can evaluate the recent ASAP3 and Skart two confidence interval procedures. This paper will first introduce the confidence interval procedure, from the basic t- distribution interval to the recent ASAP3 and Skart. Understand how the confidence interval procedure develops and then introduce the confidence interval criterion. Finally, the generalized VAMP1RE was used to evaluate ASAP3 and Skart. It is judged by the result of the evaluation whether the later-suggested confidence interval program is actually more accurate than its previous confidence interval.
