[周一午餐交流会]第十三讲：为何当自变量不被人为控制时，回归线存在偏差？

Edward A. Laws博士是美国路易斯安那州立大学的教授、路易斯安那州立大学海岸与环境学院前任院长。他本科与博士教育均在哈佛大学完成。基于杰出的教学活动与学术成就，曾被多所美国大学聘为教授，如佛罗里达州立大学、夏威夷大学等。他先后主持完成资金总额超过2亿美元的美国国家自然科学基金，目前已发表160余篇科学论文（包括Nature、Science、PNAS及其他环境科学领域顶级期刊）及3本学术专著，所有学术论文（不包括专著）至今共被引用9894次，H-index = 44，并获得多项科学荣誉。其专著《Aquatic Pollution》不仅是美国各大学经典教学参考书，且被译为多种文字在世界发行。清华大学翻译的中文版成为我国各大学的重要参考书。

讲座摘要：

Abstract: The standard model I regression method produces a biased estimate of the slope and intercept of the relationship between two variables X and Y when X is not under the control of the investigator. This talk will explain what controlling the X variable means and why lack of control over the X variable leads to bias. The extent of the bias can be estimated given knowledge of the error in X. However, it is important to remember that there are two sources of error. One is measurement error. The other is the error "inherent in the parameter being measured". The problem of bias is typically addressed by using so-called model II regression methods, but there is no completely satisfactory way to eliminate the bias when X is not controlled.