A new generalized confidence interval for the among-group variance in the heteroscedastic one-way random effects model

Abstract

Based on the generalized inference idea, a new kind of generalized confidence intervals are derived for the among-group variance component in the heteroscedastic one-way random effects model. We construct structure equations of all variance components in the model based on their minimal sufficient statistics, meanwhile, the fiducial generalized pivotal quantity can be obtained through solving an implicit equation of the parameter of interest. Then the confidence interval is derived naturally from the fiducial generalized pivotal quantity. Simulation results demonstrate that the new procedure performs very well in terms of both empirical coverage probability and average interval length.

Estimation in meta‐analyses of mean difference and standardized mean difference

“…Arendacká and Liu et al propose new confidence intervals for τ 2 in the one‐way heteroscedastic random‐effects model. These intervals can be used directly in meta‐analysis of means in noncomparative studies.…”