Discussion of ‘Causal inference with misspecified exposure mappings: separating definitions and assumptions’
对《暴露映射误设下的因果推断：区分定义与假设》的讨论：推断所需的干扰限制AI 解读 ↓
← 返回论文列表与筛选
阅读说明
本页整理公开书目信息，不生成学术质量评分。AI 解读根据题名及所注明的摘要或原文片段生成，未完成全文核验，可能有误；具体假设、证明与结论请核对原文。摘要保留来源文本，未经模型改写。
DOI 首次登记不等于接收或发表日期。部分更正、评论等条目没有独立摘要；缺少可靠材料时保留待补状态。
引用量采用 OpenAlex 单条关联记录，不合计多个记录；关联记录可能包含正式发表版本，不等于此版本单独引用。缺失不视为零，引用指标不代表论文质量。
关注作者按书目中的完整姓名匹配，同名和姓名变体需人工辨别。相关论文先按共享细分方向和具体方法排列，再参考通用方法与发表日期。
分类依据与纠错
AI 根据所提供的论文材料归类：材料明确围绕干扰、暴露映射、潜在结局及因果估计量的一致性展开。
反馈分类问题
正在核对关联版本…
代码、数据与补充材料
正在核对对应版本的资源…
关注这篇论文的作者
对《暴露映射误设下的因果推断：区分定义与假设》的讨论：推断所需的干扰限制
- 研究问题
- 如果暴露映射仅用于定义估计目标，而不再限制潜在结局中的干扰，统计推断还需要哪些替代限制来控制单位间相关性？
- 方法与思路
- 讨论作者区分暴露映射的目标定义与干扰约束两种功能，并评述被讨论论文中 Horvitz–Thompson 估计量的一致性结果。
- 材料陈述的贡献
- 所给材料赞同分离两种功能，同时强调放弃映射约束仍需其他干扰限制；其中关于弱限制下一致性的陈述是对主论文的转述。
- 阅读关注点
- 应核对误设映射时估计目标如何改变、替代干扰限制具体为何，以及讨论观点与被讨论主论文的定义和证明边界是否一致。
- 阅读边界
- 材料为 OpenAlex 索引摘要，稿件版本未注明；属于讨论文章，未给出所转述一致性结果的完整条件，不能将其视为讨论作者的新定理。
原文摘要
We congratulate the author on an insightful paper that provides a discerning critique of the use of exposure mappings in studies with interference. The main observation is that exposure mappings serve two different functions; they define the estimand and impose structure on potential outcomes by restricting interference. The first function commonly leads researchers to select simple exposure mappings that summarize treatments assigned to units ‘near’ the ego. However, due to the second function, such choices may be restrictive, ruling out interference from further units. When potential outcomes fail to satisfy the restriction, the author refers to the exposure mapping as ‘misspecified’. We agree with the author’s suggestion that researchers should choose exposure mappings for the first function but consider not employing them for the second. However, absent the second function, researchers must consider alternative restrictions on interference in order for statistical inference to be possible since, without any restrictions, outcomes may be arbitrarily correlated across units. The author studies the typical Horvitz–Thompson estimator used in the literature, which compares the average outcomes of units under two different values of the exposure mapping. His main result provides weak restrictions on interference under which the estimator is consistent for a certain estimand when exposure mappings are misspecified.