Abstract
Official classification records (755,744) of the Holstein Association of America were used to ascertain effects of genetic merit of herdmates on sire evaluations for type. Progeny were stratified into three groups on herdmate average for final score: low (less than 79.75), middle (79.75 to 81.75), and high (more than 81.75). Predicted Differences for type were computed in each group for 670 sires with 20 or more progeny in each of the herdmate groups. Mean Predicted Differences for type for three herdmate groups were .12, −.28, and −.45, indicating that failure to adjust for non-random use of sires affected herdmate comparisons for type. The correlation between herdmate average and average Predicted Difference for type of herd mates’ sires was .48, and between herd mate average and average Cow Index Type of herdmates was .45. Two iterative procedures and a mixed model comparison were examined as alternatives to sire evaluation by herdmate comparison. All were effective in reducing differences by herdmate groups. However, proofs in the low group still were favored slightly by all three models. The mixed model comparison appeared to be the most effective in reducing effects of nonrandom usage.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have