Page Banner

United States Department of Agriculture

Agricultural Research Service

Title: Comparison of the Test Interval Method with Best Prediction for Estimating Lactation Yield

Authors
item Norman, H
item Wright, Janice
item Clay, J - DAIRY REC MGT SYS, NC

Submitted to: World Congress of Genetics Applied in Livestock Production
Publication Type: Proceedings
Publication Acceptance Date: January 11, 1998
Publication Date: N/A

Interpretive Summary: Historically, most cows enrolled in a milk recording program in the United States have had milk weights recorded monthly. However, practices for collecting milk and component data have been changing rapidly, and new plans differ widely in the number of milk weights recorded and the number of component samples taken. Recently, alternative procedures to the test interval method, which has been the accepted procedure in the United States since 1969 for estimating lactation yield from test day data, have been proposed by Schaeffer and Jamrozik in Canada and by VanRaden in the United States for more accurate estimation of milk and component yields using test day data. Daily yields from 658 Canadian Holsteins and information on U.S. testing frequencies were used to determine if the best prediction method of VanRaden could estimate lactation yield more accurately than the currently used test interval method. For the traditional environment of near monthly tests, little difference was found between the two procedures in their accuracy of estimating actual milk produced. This study was the preliminary step in determining if a change in estimation procedures is warranted. Estimates from the methods should be compared for testing environments in which weights and samples are recorded at varying intervals and frequencies. Determination of appropriate statistical methodology for the use of test day data will allow the development of more accurate genetic evaluations for use by dairy producers in making breeding decisions.

Technical Abstract: In 1969, the test interval method became the accepted procedure in the United States for estimating lactation yield from test day data. Daily yields from 658 Canadian Holsteins and information on U.S. testing frequencies were used to determine if a best prediction method could estimate lactation yield more accurately than the currently used test interval method. For the traditional environment of near monthly tests, little difference was found between the two procedures in their accuracy of estimating actual milk produced. To determine if a change in estimation procedures is warranted, estimates from these methods should be compared for testing environments in which weights and samples are recorded at varying intervals and frequencies.

Last Modified: 7/24/2014
Footer Content Back to Top of Page