Accuracy of Test Interval Method of Calculating Dairy Herd Improvement Association Records

Abstract

The test interval method (TIM) of testing, modified to adjust production in the beginning and last test periods, and modified to adjust production in the beginning test periods only were simulated on individual daily milk weights of 403 Holstein lactations from the Cornell University dairy herd. The test interval method overestimated average lactation production by 57.1, 91.4, and 114.7kg of milk; whereas, correcting production in the beginning and last periods overestimated actual lactation production by 9.1, 9.0, and −2.0kg of milk for monthly, bimonthly, and trimonthly testing, respectively. Corrections for only the beginning test periods were moderately successful in reducing the bias for short test intervals; whereas, the trimonthly testing actually increased the average bias by 49.3% over the bias of the unadjusted test interval method. The variances of deviations from actual yield were reduced to 87.7, 83.1, and 78.2% of the variances in the unadjusted monthly, bimonthly, and trimonthly test interval method. The centering date method (CDM) estimated actual production more closely than the test interval method. But, if beginning and last period adjustments are made to test interval method, it errs less than the unadjusted centering date method.Correcting production in the beginning and last test periods merits consideration in all forms of production testing and especially in bimonthly and trimonthly testing.