A Bayesian approach to retransformation bias in transformed regression.

Ecological data analysis often involves fitting linear or nonlinear equations to data after transforming either the response variable, the right side of the equation, or both, so that the standard suite of regression assumptions are more closely met. However, inference is usually done in the natural metric and it is well known that retransforming back to the original metric provides a biased estimator for the mean of the response variable. For the normal linear model, fit under a log-transformation, correction factors are available to reduce this bias, but these factors may not be generally applicable to all model forms or other transformations. We demonstrate that this problem is handled in a straightforward manner using a Bayesian approach, which is general for linear and nonlinear models and other transformations and model error structures. The Bayesian framework provides a predictive distribution for the response variable so that inference can be made at the mean, or over the entire distribution to incorporate the predictive uncertainty.

[1]  Yair Mundlak,et al.  Estimation in Lognormal Linear Models , 1970 .

[2]  E. Crow,et al.  Lognormal Distributions: Theory and Applications , 1987 .

[3]  Marty D. Matlock,et al.  A Watershed-Level Ecological Risk Assessment Methodology , 1996 .

[4]  Vic Barnett,et al.  Comparative Statistical Inference , 1974, Technometrics.

[5]  Robert M. Hirsch,et al.  Estimating constituent loads , 1989 .

[6]  M. C. Newman,et al.  Regression analysis of log‐transformed data: Statistical bias and its correction , 1993 .

[7]  Kenneth H. Reckhow,et al.  Empirical models for trophic state in southeastern U.S. lakes and reservoirs , 1988 .

[8]  D. Canfield,et al.  Prediction of Total Phosphorus Concentrations, Chlorophyll a, and Secchi Depths in Natural and Artificial Lakes , 1981 .

[9]  Richard J. Smith.  Logarithmic transformation bias in allometry , 1993 .

[10]  Timothy A. Cohn,et al.  Load Estimator (LOADEST): A FORTRAN Program for Estimating Constituent Loads in Streams and Rivers , 2004 .

[11]  Don M. Miller,et al.  Reducing Transformation Bias in Curve Fitting , 1984 .

[12]  F. Ramsey,et al.  The Statistical Sleuth , 1996 .

[13]  Craig A. Stow,et al.  Estimator bias in a lake phosphorus model with observation error , 1996 .

[14]  Roy W. Koch,et al.  Bias in Hydrologic Prediction Using Log-Transformed Regression Models , 1986 .

[15]  Walter R. Gilks,et al.  A Language and Program for Complex Bayesian Modelling , 1994 .

[16]  D. Sprugel,et al.  Correcting for Bias in Log‐Transformed Allometric Equations , 1983 .