To pool or not to pool: The discrete data
暂无分享,去创建一个
The purpose of this paper is to discuss the classical problem of pooling proportions of two independent random samples taken from two possibly identical binomial distributions. The primary interest is in the estimation of [theta]1 from population I. Three estimators, i.e. the classical estimator, pooled estimator and preliminary test estimator are proposed. Their asymptotic mean squared errors are derived and compared.
[1] Frederick Mosteller,et al. On Pooling Data , 1948 .
[2] Inference for some incompletely specified models involving normal approximations to discrete data. , 1967 .
[3] T. A. Bancroft,et al. On Biases in Estimation Due to the Use of Preliminary Tests of Significance , 1944 .
[4] T. A. Bancroft,et al. On Pooling Means when Variance is Unknown , 1968 .