Abstract

Published prediction equations for fat-free lean mass are widely used by producers for carcass evaluation. These regression equations are commonly derived under the assumption that the predictors are measured without error. In practice, however, it is known that some predictors, such as backfat and loin muscle depth, are measured imperfectly with variance that is proportional to the mean. Failure to account for these measurement errors will cause bias in the estimated equation. In this paper, we describe an empirical Bayes approach, using technical replicates, to accurately estimate the regression relationship in the presence of proportional measurement error. We demonstrate, via simulation studies, that this Bayesian approach dramatically improves the accuracy of the estimated equation in comparison to the fit from Ordinary Least Squares regression.

Keywords

Carcass composition, Proportional measurement error, Empirical Bayes

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Share

COinS
 
May 1st, 7:00 AM

DEVELOPING PREDICTION EQUATIONS FOR FAT FREE LEAN IN THE PRESENCE OF AN UNKNOWN AMOUNT OF PROPORTIONAL MEASUREMENT ERROR

Published prediction equations for fat-free lean mass are widely used by producers for carcass evaluation. These regression equations are commonly derived under the assumption that the predictors are measured without error. In practice, however, it is known that some predictors, such as backfat and loin muscle depth, are measured imperfectly with variance that is proportional to the mean. Failure to account for these measurement errors will cause bias in the estimated equation. In this paper, we describe an empirical Bayes approach, using technical replicates, to accurately estimate the regression relationship in the presence of proportional measurement error. We demonstrate, via simulation studies, that this Bayesian approach dramatically improves the accuracy of the estimated equation in comparison to the fit from Ordinary Least Squares regression.