Premium
A statistical bias in the derivation of hardness‐dependent metals criteria
Author(s) -
Newman Michael C.
Publication year - 1991
Publication title -
environmental toxicology and chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.1
H-Index - 171
eISSN - 1552-8618
pISSN - 0730-7268
DOI - 10.1002/etc.5620101008
Subject(s) - statistics , regression , least squares function approximation , data set , set (abstract data type) , econometrics , mathematics , regression analysis , computer science , estimator , programming language
Log‐transformed values of lethal effect and hardness are often employed to predict metal effects. For example, they are used to develop water quality criteria. A statistical bias encountered with backtransformed, least‐squares regression models can compromise the accuracy of associated predictions. A description and method of estimating this bias are discussed herein. In a selected data set, the bias was as high as 57%.