統計方法於醫護研究之應用近年來雖然大幅地成長，然而於這類研究論文內仍不難發現分析方法使用上之錯誤。其中常見之謬誤包括：(1)p-值之誤用：一提到 p-值誤用時都與樣本數有關，但有些錯誤是不了解p-值所反映的「虛無假設」所致；也有一些是雙尾與單尾p-值之誤用，及以一個p-值作數個結論之謬誤。(2)在信度評估的過程中，雖然信度含義甚廣，通常需以不同指標來顯示不同種類信度的估值，但部分研究者仍僅以其中一種指標 (最常見的是顯示內在一致性的Cronbach’s α) 來判斷整體信度，以偏概全。(3)有時分析方法選擇正確，唯使用軟體時卻發生程式編寫錯誤，沒有適當地使用「虛擬變數」便是一個常見的例子。(4)此外又常發現研究者在進行迴歸分析前，先以簡單線性迴歸分析各自變項單獨與依變項之關係，再挑選達統計意義之項目進行複迴歸分析，殊不如在簡單迴歸分析中不顯著之自變項，在複迴歸中仍可能達統計意義。(5)至於使用之方法在論文內前後不一致，互相矛盾，或統計名詞的錯誤選擇，以致於解釋上有困難等等，也是使用統計方法之學者專家在資料分析及論文撰寫時應注意的地方。

Although statistical analyses are now more widely carried out in medical research, many researchers use inappropriate statistical methods in their studies. Common mistakes found include: (1) Misuse of the p-value: apart from the problem of sample size, the meaning of the p-value could be misinterpreted if one does not understand the associated null hypothesis. It is also unacceptable to report a 2-tailed p-value for a l-sided a1ternative, or to use a single p-value for multiple conclusions. (2) Incompleteness: several statistical indexes are needed to justify reliability, as its definition is board. However, many researchers in­correctly use only a single coefficient, such as Cronbach'sα, to indicate that the entire instrument is reliable. (3) Even when appropriate statistical methods have been chosen, it is still possible that program codes are written incorrectly. (4) Some researchers are unaware of the fact that variables not significant in a simple linear regression may become significant if included in a multiple linear regression. (5) Others such as statistical methods used are inconsistent through­out the paper, and improper choice of statistical terms for interpretations. It is hoped that more attentions will be paid to the above-mentioned points during data analysis.