Ｙａｔｅｓ校正应用于四格表 χ2 检验是合理的 ,因而也是必要的 ,但也因“过分保守”或“矫枉过正”(指校正后概率已过大以致影响统计推断 ) ,校正结果“反常”(校正后概率不变或反而变小 )而受到怀疑和指责〔1,2〕。本文试对这一使人困惑的问题作出解释 ,以供参考。一、校正 It is reasonable that Yates correction is applied to the square table χ2 test, so it is also necessary, but also because of “overconservation” or “overcorrection” (referring to the fact that the probability after correction is so large that it affects statistical inference), and the correction result is “abnormal” (correction) After the probability is unchanged or becomes smaller), it is suspected and accused [1, 2]. This article tries to explain this puzzling issue for reference. One, correction