分析具有巢狀誤差結構的迴歸模式時，惹忽略隨機誤差項之間的相關性，而採用最小平方(OLS)估計量所導出的標準 F 統計量（以 F^S表之）進行檢定，會導致過大的型 I 錯誤機率；若將隨機誤差項之間的相關性納入考量，而採用廣義最小平方(GLS)估計量所導出的 F 統計量 （以 F^GLS表之），則計算上會較為繁雜。因此我們藉由轉換方式，將模式轉換成隨機誤差項之間彼此獨立的新模式後，再以 F^S 進行檢定，其結果與直接以 F^GLS 檢定相同，且可使計算較為方便。由於模式轉換所需的轉換矩陣為母體變異數的函數，因此當母體變異數未知時，我們以 Henderson 的常數配適 （fitting-of-constants）方法來估計之。藉由模擬結果得知，若各段的觀察個數相等，則不論巢狀誤差結構為二段式(two-stage)或三段式(three-stage)，廣義最小平方估計量(GLS)均較最小平方估計量(OLS)表現穩定，且 F^GLS 在檢定力及實際顯著水準方面的表現也都比 F^S 好。 / When analyzing the regression model with nested-error structure, if the correlations between errors are ignored, and conduting the model adequacy test by the standard F statistic (F^S) led from the ordinary leastsquares estimator (OLSE) , then the type I error rate will be inflated. However, if the corrlated structure is considered and the model is tested by F^GLS led from the general least-squares estimator (GLSE) , the calculation will be more complicate. The model can be transformed to a new model with independent random errors and then, tested by F^S . The result is the same as the one by F^GLS , also it is more convenient for calculation. Since the transformation matrix is a function of variance components, we estimate variance components by Henderson's fitting-of-constants when they are unknown. Through simulation, it is concluded that if the observations in each stage of nested-error structure are the same, the GLSE is more stable than the OLSE in both two-stage and tree-stage structures. Also, the power and the sizes of F^GLS will perform better than those of F^S .

Taneichi(1993)提出一個新的適合度檢定統計量J²，具有近似卡方分配的性質。然而在小樣本的情形下，計算機模擬結果顯示，它的估計顯著水準大於期望顯著水準。所以本論文的重點之一，就是對J²進行改進，根據不同的準則，來選取一個適當的常數a。我們建議對每一觀測次數加一常數0.32，作為我們修正後的統計量，這個統計量我們記為J₁²。 　　另一探討的重點是在比較皮爾生卡方統計量X²，概似比例統計量G²，Cressie & Read統計量 I(2/3)，J²和J₁²之性質，我們想要了解在小樣本的情形之下，何者較接近於卡方分配，何者具有較強的檢定力。研究結果顯示，X²和I(2/3)較接近卡方分配，但J₁²又較G²及J²好；至於檢定力，我們發現沒有一個統計量在文中所探討的對立假設的情況下，同時都具有最大的檢定力。這些現象都可以用觀測次數對期望次數比值間的關係來解釋。 / Taneichi(1993) introduces a new goodness-of-fit statisticJ², which has an asymptotic chi-squared distribution. However, the results of simulation indicate that the levels of significance are in general bigger than the nominal levels, which prompts us to device a version of J² statistic which would perform better under small sample size situations. We suggest adding 0.32 to each observed value and find that the adjustment indeed works rearonably well. This version of J^2 statistic is denoted as J(1)^2. 　　Although Pearson chi-square statistic X², likelihood ratio statistic G², Cresse-Read statistic I(2/3), J^2 and J(1) ^2 all have asymptotic chi-squared distributions, their small sample behaviors are not expected to be the same. Comparisons based on simulation studies are then made. The conclusions are as follows : (1) In terms of levels of significance, X² and I(2/3) behave more like a chi-squared distribution. Though J(1) ^2 does not perform as good as X² and I(2/3), it does outperform G² and J². (2) In terms of powers, it does not seem that any of the test statistics has a clear advantage over the others.

微陣列晶片是發展及應用較為成熟的生物晶片技術。由於微陣列實驗程序複雜，故資料常包含多種不同來源的實驗誤差，為了適當的區分實驗中來自處理、晶片及基因的效應，我們提出混合效應變異數分析模型來調整系統誤差。針對各基因在不同實驗環境的差異性假設檢定問題，利用最小平方法推導出點估計以及對應的檢定統計量。本研究介紹多重檢定問題中的族型一誤差，並證明在此模型下，Sidak調整法為適當的多重檢定方法。在給定族型一誤差率的顯著水準，利用檢定力的公式，運算出在預設檢定力的最低水準下所需最小樣本（晶片）數。最後我們透過電腦模擬，以蒙地卡羅法來估計檢定力與族型一誤差率，由模擬結果發現，採用此最小樣本數結果，其檢定力可達到預期的水準以上，並且其族型一誤差率皆適當地控制在顯著水準以內。