论文检索

观察性研究中处理效应估计的差异性回忆偏差AI 解读

Observational studies are frequently used to estimate the effect of an exposure or treatment on an outcome. To obtain an unbiased estimate of the treatment effect, it is crucial to measure the exposure accurately. A common type of exposure misclassification is recall bias, which occurs in retrospective cohort studies when study subjects may inaccurately recall their past exposure. Particularly challenging is differential recall bias in the context of self-reported binary…

连续暴露的局部因果效应：平均因果导数效应的匹配估计AI 解读

The estimation of causal effects is a fundamental goal in the field of causal inference. However, it is challenging for various reasons. One reason is that the exposure (or treatment) is naturally continuous in many real-world scenarios. When dealing with continuous exposure, dichotomizing the exposure variable based on a pre-defined threshold may result in a biased understanding of causal relationships. In this paper, we propose a novel causal inference framework that can…