第400行: |
第400行: |
| * 封闭世界假设意味着模型考虑到了所有因变量,然而实际情况并不是这样 | | * 封闭世界假设意味着模型考虑到了所有因变量,然而实际情况并不是这样 |
| | | |
− | 6.3.3 无益代理使必要性失效 Failing Necessity via Barren Proxies 186 | + | 6.3.3 无效中介变量使必要性失效 Failing Necessity via Barren Proxies 186 |
| + | |
| + | * 无效中介变量如何使相关性准则的必要条件失效,以及提出改进的相关性准则 |
| | | |
| * | | * |
| | | |
| 6.3.4 偶然抵消使必要性失效 Failing Necessity via Incidental Cancellations 188 | | 6.3.4 偶然抵消使必要性失效 Failing Necessity via Incidental Cancellations 188 |
| + | |
| + | * 利用一个例子阐释了模型的偶然抵消 |
| | | |
| '''6.4 稳定无偏与偶然无偏 Stable versus Incidental Unbiasedness 189''' | | '''6.4 稳定无偏与偶然无偏 Stable versus Incidental Unbiasedness 189''' |
| | | |
| 6.4.1 动机 Motivation 189 | | 6.4.1 动机 Motivation 189 |
| + | |
| + | * 区分稳定无偏与偶然无偏 |
| | | |
| 6.4.2 形式化定义 Formal Definitions 191 | | 6.4.2 形式化定义 Formal Definitions 191 |
| + | |
| + | * 稳定无偏,结构稳定无混杂这些概念的定义,常识原理以及证明 |
| | | |
| 6.4.3 稳定无混杂的运算检验 Operational Test for Stable No-Confounding 192 | | 6.4.3 稳定无混杂的运算检验 Operational Test for Stable No-Confounding 192 |
| + | |
| + | * 稳定混杂准则及其证明 |
| | | |
| '''6.5 混杂,可压缩性和可交换性 Confounding, Collapsibility, and Exchangeability 193''' | | '''6.5 混杂,可压缩性和可交换性 Confounding, Collapsibility, and Exchangeability 193''' |
| | | |
| 6.5.1 混杂和可压缩性 Confounding and Collapsibility 193 | | 6.5.1 混杂和可压缩性 Confounding and Collapsibility 193 |
| + | |
| + | * 可压缩性定义,引理:稳定无混杂说明可压缩性 |
| | | |
| 6.5.2 混杂与混杂因子 Confounding versus Confounders 194 | | 6.5.2 混杂与混杂因子 Confounding versus Confounders 194 |
| + | |
| + | * 罗斯曼(Rothman)和格陵兰(Greenland)关于混杂因子的看法 |
| | | |
| 6.5.3 可交换性与混杂结构分析 Exchangeability versus Structural Analysis of Confounding 196 | | 6.5.3 可交换性与混杂结构分析 Exchangeability versus Structural Analysis of Confounding 196 |
| + | |
| + | * 可交换性和混杂结构分析间的联系,因果效应两种不同定义方式的分析 |
| | | |
| '''6.6 总结 Conclusions 199''' | | '''6.6 总结 Conclusions 199''' |
| + | |
| + | * 难以将混杂与相关性联系起来的原因,结构分析模型的优势 |
| | | |
| === 7 结构化反事实的逻辑 The Logic of Structure-Based Counterfactuals 201 === | | === 7 结构化反事实的逻辑 The Logic of Structure-Based Counterfactuals 201 === |