Standard errors of nitrogen-corrected true metabolizable energy estimates: effects of pooling excreta samples and ignoring among-control bird variation.

To estimate TME(n), it appears to be advantageous to pool excreta samples prior to analysis for energy and nitrogen; accurate (nonbiased) estimates should result. With this protocol, estimates of precision are based entirely on the among-bird variance in excreta weight of the fed birds. However, because of the inconsistent relationship between excreta weight and excreta energy, the SE resulting from this pooling protocol generally are unrelated to the correct values obtained when the energy and nitrogen concentrations are measured in the excreta of each bird. Failure to account for the among-control-bird variation in excreta energy overstates the precision (i.e. underestimates the SE) of TME(n) estimates.