Use of With-in-Primary-Sample-Unit Variances to Assess the Stability of a Standard Design-Based Variance Estimator

This article on survey methodology addresses the problem of estimating degrees of freedom values from stratified multistage designs when a small number of primary sampling units (PSUs) are selected per stratum. Due to the small number of PSUs selected, the traditional Satterthwaite-based degrees of freedom can be a severe underestimate. This paper proposes an alternative way to estimate the degrees of freedom that uses the within-PSU variances to provide additional information on the relative magnitudes of the overall stratum-level variances.