Direct estimates and design SEs per domain, ready for small-area estimation

Small-area estimation (SAE) area-level models (Fay-Herriot) need, for each domain, the direct estimate, its design-based variance and the effective sample size. This function computes exactly those from a recipe-aware replicate object, so the design-based ingredients flow into emdi, sae or hbsae without leaving the weightflow variance machinery. It does not fit any SAE model itself.

Usage

as_sae_input(
  object,
  variable,
  by,
  type = c("mean", "total"),
  level = 0.95,
  cv_breaks = c(0.165, 0.33)
)

Arguments

object

a weightflow_boot or weightflow_jack object (from bootstrap_weights() / jackknife_weights()).

variable

name of the study variable.

by

name(s) of the domain column(s); several are crossed.

type

"mean" (default) or "total".

level

confidence level for the interval.

cv_breaks

two increasing CV cut-points for the publishability rating (default c(0.165, 0.33), common in official statistics): a CV below the first is "publishable", between the two "review", above the second "not publishable".

Value

A data frame with one row per domain: domain, n (active units), n_eff (Kish effective sample size), estimate, se, cv, ci_lower, ci_upper and rating. Pass estimate and se^2 (the sampling variance) to a Fay-Herriot model.

Note for very small domains: the domain estimates share one bootstrap, and a replicate in which any domain has no active unit is dropped for all domains (a conservative choice). With many tiny domains this wastes replicates; use more replicates in the bootstrap, or estimate sparse domains in a separate call.

Details

The domain standard error is the recipe-aware replicate SE (it re-runs the whole recipe per replicate), so it already reflects nonresponse and calibration, not just the final weights.

See also

domain_summary(), bootstrap_estimate(), design_effect()

Other cascade audit: collect_propensities(), collect_step_detail(), collect_weights(), domain_summary(), weight_factors(), weighting_alerts()

Examples

spec <- weighting_spec(sample_survey, base_weights = pw) |>
  step_calibrate(method = "raking",
                 margins = list(region = c(table(population$region))))
boot <- bootstrap_weights(spec, replicates = 50, strata = "region",
                          psu = "psu", seed = 1)
#>   bootstrap replicate 25/50
#>   bootstrap replicate 50/50
as_sae_input(boot, "responded", by = "region")
#>   domain   n n_eff  estimate         se         cv  ci_lower  ci_upper
#> 1   East  96    96 0.5416667 0.04422887 0.08165331 0.4549797 0.6283537
#> 2  North 119   119 0.6554622 0.02944014 0.04491508 0.5977606 0.7131638
#> 3  South 121   121 0.5950413 0.05318722 0.08938407 0.4907963 0.6992864
#> 4   West 131   131 0.5190840 0.04245369 0.08178579 0.4358763 0.6022917
#>        rating
#> 1 publishable
#> 2 publishable
#> 3 publishable
#> 4 publishable