Case Study: Use of statistics to quantify the magnitude of errors in the sorting of pigs for marketing in 3 finishing barns

The objectives were to use actual data to evaluate methods to quantify the amount of error in the estimation of BW for market pigs and its effect on several statistics including the SD in carcass weight (CW). Data from 3 wean to finish barns were obtained that included slaughter dates, CW, sort loss, date ractopamine feeding was initiated, and birth dates. Pigs were sold in 3 marketing cuts (MCUT) with multiple marketing days for each MCUT. The BW of each pig was estimated from CW. The BW data were fitted to a generalized Michaelis-Menten equation with the values for K and C fixed at 191.5 d and 2.221, respectively. The BW for each pig at each marketing day was estimated using the generalized Michaelis-Menten equation including the random effect of each pig. Three marketing error rates for BW were evaluated: with no error and actual and random sorting. Statistics were calculated including mean and SD for BW and CW, sort loss, and percentage of pigs sold correctly, for each MCUT. The barn means for CW were 95.7, 100.9, and 91.2 and SD of 7.6, 8.8, and 7.4 kg. With no error, the estimated means were 95.5, 100.7, and 91.0 kg with SD of 4.6, 4.6, and 4.3 kg. Percentages of pigs sold correctly for the first 2 MCUT were 56, 48, and 52%, and differences in sort loss (estimated versus actual $/pig) were 1.22, 5.74, and 0.56 for each barn. The accuracy in which pigs were sorted for market can be evaluated from currently available data.