因果關係的進一步研究

Abstract

本論文的第一個目的是研究有沒有可能在X並非Y的成因(cause)，但統計分析資料卻發現它們有統計相關。這就是擾亂(confounding)效果。第二個目的在於研究變數X為因對Y的效果之理論研究並進行模擬分析。最後我們將提出一個實際資料分析的結果。

The first aim of this thesis is to study confounding effect in causal inference. It is to see when variable X is not a cause of response Y but we see the presence of statistical association between X and Y. The second aim is to study the theory of causal effect of X as the cause of response variable Y and conduct a simulation study of this causal effect. Finally we want to conduct a data analysis on causal effect.