A Note on the Bias due to Omitted Confounders
研究了观察性研究中遗漏未知混杂因素导致的偏倚，提出了近似公式估计渐近偏倚，并通过逻辑回归和泊松回归模型验证了其准确性，最后用低出生体重与吸烟关系的数据集进行说明。
In observational studies some confounders may be unknown and therefore omitted from the analysis while others are adjusted for. Approximations to the functions defining the relationship between the parameters in the full and reduced models are proposed leading to asymptotic bias estimates. Numerical calculations for logistic and Poisson regression models show good agreement between asymptotic and simulation bias. A data set containing the relationship between low birth weight and smoking (Hosmer & Lemeshow, 1989) is used as an illustration.