首页 | 本学科首页   官方微博 | 高级检索  
     


Missing regressor values under conditions of multicollinearity
Authors:R. Carter Hill  Rod F. Ziemer
Affiliation:1. Economics Department , University of Georgia , Athens, Georgia, 30602;2. Agricultural Economics Department , Texas AAmp;3. M University , College Station, Texas, 77843
Abstract:
The performance of some common procedures for replacing missing regressor values is investigated under varying conditions of multicollinearity" Analytical and numerical results indicate that the popular method of substituting sample means for missing values appears preferrable to other more sophisticated procedures given ill-conditioned designs In addition, results indicated that incomplete sample observations should not be thrown away under conditions of extreme multicollinearity.
Keywords:ill-conditioned designs  incomplete sample observations  missing regressor values
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号