Biased Estimates of Heritability Resulting from Incorrect Methods of Estimating Components of Variance

Abstract

When more than 1 record of an individual is used and full sibs are included in the data, estimates of the sire, dam, individual, and error components of variance must be provided for in the intra-class correlations method of estimating heritability. Biases in estimates of heritability were derived for 2 commonly used short-cut methods of analyses of data. Pooling sums of squares in an analysis of variance must be followed with the correct expectation of the resulting mean square if components of variance are to be derived.