论文部分内容阅读
对偏倚与混杂的认识和定义虽然在不断深入和改变,但流行病学家们从未忽略控制偏倚和混杂的重要性。历史上,对于偏倚的定义已经包含三个主要意思:①观察者的偏见;②器械错误造成的偏差;③错误的研究设计造成的结果。20世纪70年代出现的偏倚类型学对于定义在流行病学研究设计、实施和分析过程中存在的偏离真实值的情况有很大的帮助。随后,偏倚逐渐形成了更加正式和系统的定义。现代流行病学把偏倚归纳为选择偏倚、信息偏倚和混杂偏倚。混杂可以看作是偏倚的一个特殊类型,在简明牛津英文词典中提到它是一个中世纪的拉丁文词:“con-fundere”,意思是混在一起,源自中世纪法语“confondre”。
Although awareness and definition of bias and confusion continue to deepen and change, epidemiologists have never overlooked the importance of control bias and confusion. Historically, the definition of bias has included three main meanings: ① the prejudice of the observer; ② the deviations caused by the device error; ③ the result of the wrong research design. The typology of bias that emerged in the 1970s was instrumental in defining deviations from true values in the design, implementation, and analysis of epidemiological research. Subsequently, bias gradually evolved into a more formal and systematic definition. Modern epidemiology summarizes bias as choice bias, information bias, and hybrid bias. Confusion can be seen as a special type of bias, which is mentioned in the concise Oxford English Dictionary as a medieval Latin word: “con-fundere,” meaning mixed, from the medieval French “confondre.”