A Sorting Method to Reduce Variation in Starting Pen Weights in Poultry Research Studies Utilizing Microsoft Excel

Abstract

Variability within nutrition experiments can be problematic, especially when one is trying to pick up small differences between experimental treatments. A variety of methods have been used over the years to reduce starting animal weight such as weighing animals into different weight groupings followed by selection of animals from the different group in an attempt to obtain smaller differences in starting weights. This paper describes a computer based sorting method that will substantially reduce variation of starting animal weights. In the example used, 700 turkey poults were sorted in three ways at one week of age into 70 pens of 10 poults per pen. Initially 800 poults were weighed and banded with top and bottom weight poults removed to the 700 poults needed. Random assignment of poults resulted in pen weights that ranged from 122.12 to 151.94 g/bird, a difference of 24.4%. Using a 3 weight grouping resulted in an average pen weight range of 130.22 to 140.4 g/bird, a difference of 10.18 g/pen or 7.8%. Using the sorting program resulted in average pen weights ranging from 134.88 to 135.56 g/bird, a difference of 0.68 g or 0.5%, substantially lower than the group sorting. It is believed that this type of sorting of animals at the beginning of an experiment will result in substantially reduced variation.