Submitted to: Journal of Dairy Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: October 10, 1995
Publication Date: N/A
Interpretive Summary: Demands to lower dairy producers' milk-testing costs, to increase technicians' ability to handle more herds, and to reduce disruption of normal milking routine have spurred investigations and implementations of alternative milk-testing methods that have the credibility of official Dairy Herd Improvement record keeping plans. Yield on test day reported as a percentage of the actual milk shipped has been proposed by dairy industry groups as a statistic for monitoring the accuracy of alternative record keeping plans. This statistic was examined to determine 1) if its properties differed with herd size, production level, and record-keeping plan and 2) whether adjusting individual daily milk yields for this statistic would increase the accuracy of genetic evaluations for milk yield. Frequency distribution of the statistic revealed a normal distribution centered around 103%, which indicated that dairy producers average 3% more milk yield on test day than they normally ship. Herd size, record keeping plan, and herd production level explained little of the differences in test-day yield as a percentage of milk shipped. Monitoring test-day yield as a percentage of milk shipped should help 1) to ensure that correct milking intervals have been reported to record keeping technicians and 2) to detect previously unknown problems associated with collection of test-day milk yields. Although adjusting individual daily milk yields for the test-day yield as a percentage of milk shipped could result in better estimates of actual milk produced, such an adjustment would have little effect on accuracy of genetic evaluations.
Milk yield recorded on Dairy Herd Improvement (DHI) test day was compared with milk yield shipped from Texas and Minnesota herds in the innovative DHI test plan "Alternate AM/PM Without a Timer." Controls were test-day yields and milk yields shipped from official DHI herds in Texas, Illinois, Minnesota, and several Northeastern U.S. states. Herd test-day milk yield as a percentage of milk shipped (%MS) was considered an indicator of accuracy of the test. Mean %MS was 103 with a standard deviation of 6 for all plans and regions. If herd test-days with missing values were excluded, the percentage of herd test-days within 96 to 110 %MS was 77 and 82 for Texas and Minnesota innovative test plans, respectively, and 82, 82, 79, and 81 for Texas, Minnesota, Illinois, and Northeastern official plans. Analysis indicated that %MS was consistent across herd sizes, data sources, and milk production levels. Eight hypothetical testing plans were examined with or without adjusting lactation yields for %MS. Multitrait animal model variance component estimates of the lactation milk yields were computed and compared. Adjusting records for %MS would decrease mean milk yields by 3% and could result in better estimates of actual milk produced but would have little effect on accuracy of genetic evaluations.