我們提出新的概似函數方法,來分析成對設計的資料,主要目的在比較兩個篩檢之陽性預測值或是兩個陰性預測值。直觀地,比較陽性預測值時僅需兩個篩檢結果為陽性的病患,反之,比較陰性預測值時僅需兩個篩檢結果為陰性的病患。現今大多數以多項式模型為基礎的方法,在模型中包含多餘的參數,導致推導出的統計量較為複雜。本文所介紹的新方法,只使用了最少量的感興趣參數。而我們的強韌分數檢定統計量與Kosinski (2013) 所提出的加權廣義分數統計量是完全相同的。;We propose a new likelihood approach to comparing two positive predictive values (PPVs) or two negative predictive values (NPVs) in paired designs. Intuitively, one only needs patients with two positive screening test results for PPVs comparison, and those with two negative screening test results for contrasting NPVs. evertheless, current existing methods rely on the multinomial model that includes superfluous parameters unnecessary for specific comparisons. This practice results in complex statistics formulas. We introduce a novel approach that fits the intuition by including a minimum number of parameters of interest and show that our robust test statistic is identical to the weighted generalized score test statistic proposed by Kosinski (2013).