Skip to content

Post-stratified Point Estimator

This function implements the standard FIA post-stratified point estimator. Astute statistical users will need to forgive a small abuse of nomenclature here. Indeed, this estimator is a weighted sum of post-stratified estimators across estimation units within an evaluation. We will prefer this simpler term at the risk of some confusion, as it is the standard term used in FIA documentation. Because the estimator is standard it requires the presence of pop_stratum and pop_estn_unit tables in the evaluation, which encode the post-strata weights among other details. With the default var_est = "auto", the standard ve_post_strat() variance estimator is selected automatically. This is equivalent to supplying ve_post_strat() explicitly and provides a convenient default for the standard non-ratio estimator.

Post-stratified point estimation, in the FIA context, computes estimates for each estimation unit using a post-stratified estimator, then sums across the set of estimation units (assuming independence) to produce a single estimate for the evaluation. For estimation unit gg we obtain

Y^g=∑hWghY^gh\hat{Y}_g = \sum_h W_{gh} \hat{Y}_{gh}

where hh indexes a post-stratum within estimation unit gg that unit, Y^gh\hat{Y}_{gh} is the stratum-specific estimate, and WghW_{gh} is its post-stratum weight. Then, the overall estimate is

Y^=∑gKgY^g\hat{Y} = \sum_g K_g \hat{Y}_g

where KgK_g is the proportion of the total evaluation area in estimation unit gg. The estimator can return either the weighted mean or the corresponding total through the output argument to estimate(), wherein each estimation unit is weighted by its size in acres, AgA_g.

pe_post_strat(var_est = "auto")
  • var_est: A variance-estimator specification, or "auto".

A PostStratifiedEstimator specification.