知识卡片

普通读书笔记卡 · 6568.e · 生成模型:GPT-6

比率与差值量度不能互换

理解等级:中级

二元结果可用风险差p1−p0、风险比p1/p0及优势比[p1/(1−p1)]/[p0/(1−p0)],分母要有效。赋予因果含义前,p1、p0必须是所需干预风险;回归模型给出关联比率,不会因换名称自动变成因果量。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      relative risk (RR) directly from a Poisson regression model and odds ratios (OR) directly from a logistic regression model. In general, these measures of association have no causal interpretation, but we give them a causal interpretation once we use them with interventional variables. For example, if we are modeling E (YX), and YX is binary, the relative risk and odds ratios are as follows:

    参考来源

    - 位置:《Causal AI》第07章《干预与因果效应》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-7.xhtml) - 结论依据:摘录中关于比率与差值量度不能互换的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:relative risk (RR) directly from a Poisson regression model and odds ratios (OR) directly from a logistic regression model. In general, these measures of association have no causal interpretation, but we give them a causal interpretation once we use them with interventional variables. For example, if we are modeling E (YX), and YX is binary, the relative risk and odds ratios are as follows: