The Effect of Herd Milk Production on Scoring Dairy Herd Improvement Performance Measures

Abstract Objectives of this research were to convert DHI performance measures into percentile scores based on expected values at a producer's rolling herd average and to compare performance with percentile scores based on means of all herds. Holstein herd data from the Minnesota Dairy Records Processing Center were used to derive regression coefficients to predict expected values of DHI management variables from three forms of rolling herd average 4% FCM. Variables were converted to percentiles using expected parameter value, assumption of normal distribution, and the mean square error of the prediction equation as an estimate of the variance. With some exceptions in reproductive categories, expected values for DHI performance measures at the mean rolling herd average were generally better than expected values in lower producing herds and worse than expected values in higher producing herds. Absolute values for the 75th percentile of performance measures based on the mean of all herds translated to percentiles of 54 to 82 when scoring was based on expected values at the 75th percentile of rolling herd average. Electronically importing DHI data and converting to percentiles based on rolling herd average allowed for comparison of DHI performance measures with similar rather than all herds.