The Journal of Heredity 1982:73(2):139-141
© 1982 The American Genetic Association 73:139-141
other |
Bias in estimation of the sampling variance of an estimate of heritability
Department of Mathematics, and the Department of Dairy Science, University of Illinois at Urbana-Champaign Urbana, IL 81801
Please address reprint requests to Dr. Grossman
Abstract
Bias in estimating the sampling variance of heritability estimated from full- and hall-sib analysis is evaluated. The asymptotic variance of a ratio of random variables, as well as the asymptotic covariance of ratios of random variables are compared to first approximation sampling variances and covarlance using six combinations of heritability (h2 = 0.2 and 0.4) and total sample size (N = 120, 600, and 1200). Although approximate estimates of sampling variances and covartance of heritabilities are easier to compute than asymptotic sampling variances and covartance. the biases may be important depending on sample size, heritability, methed of estimation of heritability, and the use to be made of the estimate of heritability.