REGRESSION-ANALYSIS OF LOG-TRANSFORMED DATA - STATISTICAL BIAS AND ITS CORRECTION

被引：139

作者：

NEWMAN, MC

机构：

[1] University of Georgia, Savannah River Ecology Laboratory, Aiken, South Carolina, 29801, P.O. Drawer E

来源：

ENVIRONMENTAL TOXICOLOGY AND CHEMISTRY | 1993年 / 12卷 / 06期

关键词：

STATISTICS; REGRESSION; BIAS; LOG-TRANSFORMED VARIABLES;

D O I：

10.1002/etc.5620120618

中图分类号：

X [环境科学、安全科学];

学科分类号：

08 ; 0830 ;

摘要：

Power and exponential models are used frequently in environmental chemistry and toxicology. Such models can generate biased predictions if derived with least-squares, linear regression of log-transformed variables. An easily calculated but seldom used estimate of bias can enhance the accuracy of subsequent predictions. This prediction bias and means of correcting it are presented, along with several examples.

引用

页码：1129 / 1133

页数：5