| Summary: | The basic purpose of feeding trials is to find the optimum level of feed ingredients which give the highest economical returns to the farmers. This can be achieved through estimation and comparison of means of different rations. The example we have is a study of incorporation of different levels of ensiled brewers grains in the diet of 24 hybrids weaned piglets from Landrace x Duroc x Berkshire x Large White. They were randomly divided into four groups with three replicates of two piglets per pen. They were fed 0, 10, 20, 30% incorporation of ensiled brewer's grains on dry matter basis during post-weaning period followed by 0, 30, 40 and 50% during growing period and 0, 50, 60 and 70% during finishing period. We have one explanatory variable: initial weight, and four post treatment outcome variables recorded per piglets: final weight, dry matter consumption, weight gain and index of consumption. Comparing of several multivariate treatment means model design analysis is adapted. We obtain the MANOVA (Multiple Analyse of Variance) table of each phase, where the treatment differences exist by using Wilk's lambda distribution, and we find the treatment effect by using a confidence interval method of MANOVA. This model has the advantage of computing the responses of all variables in the matrix of sum of squares and more precisely in separation of the different means percentage of Ensiled Brewer's grain.
|
|---|
