Chapter 9 Measurement Bias

文章目录

9.1 Measurement Error
The structure of measurement error
9.3 Mismeasured confounders
9.4 Intention-to-treat effect: the effect of a misclassified treatment
9.5 Per-protocol effect
Fine Point
- The strength and direction of measurement bias
- Per-protocol analyses
- Pseudo-intention-to-treat analyses
- Effectiveness versus efficacy
Technical Point
- Independence and nondifferentiality
- The exclusion restriction

HernKaTeX parse error: Can't use function '\'' in math mode at position 1: \̲'̲{a}n M. and Robins J. Causal Inference: What If.

已经介绍过两个bias: confounding和selection, 这里介绍第三个, measurement bias.
这个measurement bias 不是指样本数目过少导致的误差, 用记录误差更为准确, 比如明明是A=1A=1A=1, 但是记录成了A=0A=0A=0.

9.1 Measurement Error

如上图所示, 我们所希望的实际上只有A→YA \rightarrow YA→Y这一条先, 但是由于measurement error UAU_AUA导致我们最后得到的只有A∗A^*A∗, 此时想要计算AAA关于YYY的causal effect就比较麻烦了.

不光是intervention, outcome YYY实际上也可能会有measurement error.

The structure of measurement error

measurement error 可以分成四种:

independent nondifferential: Figure 9.2;
dependent nondifferential: Figure 9.3;
independent differential: Figure 9.4;
dependent differential: Figure 9.6.

9.3 Mismeasured confounders

这里介绍了另外一种特殊的情况, A,YA, YA,Y都没有测量误差, 但是LLL是存在测量误差的, 此时因为缺乏LLL, 我们没法依赖其获得想要的条件可交换性.

同样的, 类似的情况也可能导致selection bias.

9.4 Intention-to-treat effect: the effect of a misclassified treatment

这一节讨论的是, 预先打算的策略ZZZ对YYY的causal effect和实际进行的AAA关于YYY的causal effect之间的差别.

9.5 Per-protocol effect

主要讲了什么时候Intention-to-treat effect会是一个不错的选择.

Fine Point

The strength and direction of measurement bias

Per-protocol analyses

as treated: 只管AAA, 不管ZZZ(所以是不考虑ZZZ的存在?).
per protocol: 只在A=ZA=ZA=Z的子集上讨论, 所以这种很有可能会导致selection bias. 如Figure 9.14 所示.

Pseudo-intention-to-treat analyses

Effectiveness versus efficacy

Technical Point

Independence and nondifferentiality

The exclusion restriction

指的是没有直接的箭头由assign treatment ZZZ 指向 outcome YYY.