Package {WFC}

Convert a weighting quality report to a data frame

Description

Convert a weighting quality report to a data frame

Usage

## S3 method for class 'wf_quality_report'
as.data.frame(x, ...)

Arguments

Value

The report's primary table.

Coerce a variance result to a data frame

Description

Coerce a variance result to a data frame

Usage

## S3 method for class 'wf_variance_result'
as.data.frame(x, ...)

Arguments

Value

The result table as a data frame.

Convert WFC replicate weights to a survey replicate design

Description

Aligns recalibrated WFC replicate weights to analysis data by exact unit ID and creates a standard survey::svrepdesign() object. The returned design uses full-estimate-centered (mse = TRUE) variance so its estimates reproduce wf_variance().

Usage

as_svrepdesign(r, data, id = "id", degf = NULL, ...)

Arguments

Value

A standard svyrep.design object whose variables include the aligned .wf_weight base-weight column.

Convert WFC weights to a survey design

Description

Joins calibrated WFC weights onto an analysis data frame by exact unit ID and creates a standard survey::svydesign() object. The survey package remains a suggested dependency.

Usage

as_svydesign(
  w,
  data,
  id = "id",
  ids = ~1,
  strata = NULL,
  fpc = NULL,
  nest = FALSE,
  ...
)

Arguments

Value

A standard survey.design2 object whose variables include the aligned .wf_weight column.

Plot an automatic trim frontier

Description

Draws the worst design effect and residual margin error for every feasible finite cap.

Usage

## S3 method for class 'wf_auto_trim'
plot(x, lang = NULL, ...)

Arguments

Value

Invisibly returns x.

Plot blend lambda sensitivity

Description

Plot blend lambda sensitivity

Usage

## S3 method for class 'wf_blend_result'
plot(x, lang = NULL, ...)

Arguments

Value

Invisibly returns x.

Plot weight diagnostics

Description

Draws design effects and effective-sample-size shares by group.

Usage

## S3 method for class 'wf_diagnostics'
plot(x, lang = NULL, ...)

Arguments

Value

Invisibly returns x.

Plot propensity overlap and covariate balance

Description

Plot propensity overlap and covariate balance

Usage

## S3 method for class 'wf_propensity_weights'
plot(x, lang = NULL, ...)

Arguments

Value

Invisibly returns x.

Plot calibrated weight distributions

Description

Draws one histogram per group, including the group mean and recorded raking trim bounds when available.

Usage

## S3 method for class 'wf_weights'
plot(x, max_groups = 9, lang = NULL, ...)

Arguments

Value

Invisibly returns x.

Print attrition weights

Description

Print attrition weights

Usage

## S3 method for class 'wf_attrition_weights'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print an automatic trim recommendation

Description

Print an automatic trim recommendation

Usage

## S3 method for class 'wf_auto_trim'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print a guided weighting result

Description

Print a guided weighting result

Usage

## S3 method for class 'wf_autoweigh_result'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print a blend result.

Description

Print a blend result.

Usage

## S3 method for class 'wf_blend_result'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print an outcome-blind cell merge plan

Description

Print an outcome-blind cell merge plan

Usage

## S3 method for class 'wf_cell_merge_plan'
print(x, ...)

Arguments

Value

x, invisibly.

Print a collapse plan

Description

Print a collapse plan

Usage

## S3 method for class 'wf_collapse_plan'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print outcome-free survey design data

Description

Print outcome-free survey design data

Usage

## S3 method for class 'wf_design_data'
print(x, ...)

Arguments

Value

x, invisibly.

Print diagnostics

Description

Print diagnostics

Usage

## S3 method for class 'wf_diagnostics'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print post-lock outcome impact

Description

Print post-lock outcome impact

Usage

## S3 method for class 'wf_impact'
print(x, ...)

Arguments

Value

x, invisibly.

Print influence diagnostics

Description

Print influence diagnostics

Usage

## S3 method for class 'wf_influence'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print a collapse-ladder draft

Description

Print a collapse-ladder draft

Usage

## S3 method for class 'wf_ladder_draft'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print a pipeline specification

Description

Print a pipeline specification

Usage

## S3 method for class 'wf_pipeline'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print a post-stratification plan

Description

Print a post-stratification plan

Usage

## S3 method for class 'wf_poststrat_plan'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print a precheck result

Description

Print a precheck result

Usage

## S3 method for class 'wf_precheck'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print a weighting quality report

Description

Print a weighting quality report

Usage

## S3 method for class 'wf_quality_report'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print replicate weights

Description

Print replicate weights

Usage

## S3 method for class 'wf_replicate_weights'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print a weight-validation result

Description

Print a weight-validation result

Usage

## S3 method for class 'wf_validation'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print a variance result

Description

Print a variance result

Usage

## S3 method for class 'wf_variance_result'
print(x, ...)

Arguments

Value

Invisibly returns x.

Print a reviewable weight plan

Description

Print a reviewable weight plan

Usage

## S3 method for class 'wf_weight_plan'
print(x, ...)

Arguments

Value

x, invisibly.

Print calibrated weights

Description

Print calibrated weights

Usage

## S3 method for class 'wf_weights'
print(x, ...)

Arguments

Value

Invisibly returns x.

Summarize a post-stratification plan

Description

Summarize a post-stratification plan

Usage

## S3 method for class 'wf_poststrat_plan'
summary(object, ...)

Arguments

Value

The plan diagnostics data frame.

Apply a category collapse plan

Description

Applies category-merge maps consistently to sample data and a target object.

Usage

wf_apply_collapse(sample, target, plan)

Arguments

Value

A list with collapsed sample and target.

Record a human attestation for a reviewed weight plan

Description

Approval records are separate from plans. AI agents may prepare plans but cannot create an approval while declaring themselves as an agent.

Usage

wf_approve_plan(plan, approver, role, note = NULL, actor_type = "human")

Arguments

Value

A wf_plan_approval attestation bound to the plan identity.

Assess descriptive outcome impact after weights are locked

Description

Compares fixed unweighted and locked-weight estimates. This function cannot accept target values, selection callbacks, or planning controls, and it never recalculates weights.

Usage

wf_assess_impact(weights, data, id, outcomes, level = 0.95)

Arguments

Value

A wf_impact object with fixed descriptive summaries.

Attach locked weights to analysis data by exact unit ID

Description

Attach locked weights to analysis data by exact unit ID

Usage

wf_attach_weights(data, weights, id, weight_name = ".weight")

Arguments

Value

data in its original row order with one locked-weight column.

Estimate panel attrition weights

Description

Models wave-to-wave retention in a panel and returns inverse-retention weights for retained units. The result inherits from wf_weights, so it can be composed with later calibration stages through wf_compose().

Usage

wf_attrition(
  panel,
  retained,
  formula,
  id = NULL,
  by = NULL,
  stabilize = TRUE,
  trim = NULL
)

Arguments

Value

A wf_attrition_weights object inheriting from wf_weights, with a ⁠$balance⁠ diagnostic table and per-group retention log.

Export a self-contained WFC audit file

Description

Writes a JSON audit record containing result provenance, pipeline metadata, optional guided-workflow decision ledger, input hashes, and user-supplied extra metadata. The writer is dependency-free and intended for machine archiving, not human formatting.

Usage

wf_audit_export(x, file, inputs = NULL, extra = NULL)

Arguments

Value

Invisibly returns file.

Recommend a weight-trimming cap

Description

Sweeps candidate upper caps by rerunning raking and reports the trade-off between design effect and residual margin error. The function recommends a cap but never applies it to a production result.

Usage

wf_auto_trim(
  design,
  target,
  caps = c(2, 3, 4, 5, 6, 8, 10, 12),
  lo = 0.05,
  max_deff = 6,
  max_residual = 0.02,
  ...
)

Arguments

Value

A wf_auto_trim object containing the candidate frontier and the recommended cap. Inf means no trimming is needed; NA means no candidate meets both criteria.

Run a guided workflow from verified design and target objects

Description

Provides the guided WFC experience while enforcing the same WFC 2.0 input boundary used by direct calibration and approved-plan execution.

Usage

wf_autoweigh(
  design,
  target,
  dims,
  method = c("auto", "raking", "poststrat", "logit"),
  ladder = NULL,
  min_cell = NULL,
  bounds = c(0.3, 3),
  trim = "auto",
  max_deff = 6,
  max_residual = 0.02,
  interactive = base::interactive(),
  lang = NULL,
  ...
)

Arguments

Value

A wf_autoweigh_result carrying verified input identities.

Blend online and offline calibrated estimates

Description

Combines two wf_weights sources at the estimator level. Each source is estimated within each fusion cell first; the source estimates are then combined using the effective lambda recorded in the result.

Usage

wf_blend(
  online,
  offline,
  by_cell,
  lambda = c("neff", "inverse_variance", "fixed"),
  lambda_fixed = NULL,
  outcome = NULL,
  level = c("cell", "group"),
  trim_lambda = c(0.05, 0.95),
  sensitivity = TRUE
)

Arguments

Value

A wf_blend_result object.

Calibrate weights from verified design and target objects

Description

Enforces the WFC 2.0 verified-input boundary before dispatching to a calibration engine. Raw data frames and ordinary wf_target objects are not accepted.

Usage

wf_calibrate(design, target, method = "raking", ...)

Arguments

Value

A wf_weights object carrying the verified input identities.

Declare a post-stratification collapse ladder

Description

Creates an ordered, cumulative, validated category-collapse ladder for post-stratification cell resolution. Level 0 is the raw classification; each named level applies one additional set of dimension maps.

Usage

wf_collapse_ladder(dims, ...)

Arguments

Value

A wf_collapse_ladder object.

Examples

dims <- wf_dims(age = c("young", "old"), education = c("low", "high"))
wf_collapse_ladder(
  dims,
  level1 = list(age = c(young = "all", old = "all"))
)

Compose multiple weighting stages

Description

Multiplies compatible wf_weights objects into one auditable pipeline result.

Usage

wf_compose(..., id = NULL, normalize = c("none", "mean1", "sum"))

Arguments

Value

A composed wf_weights object.

Examples

stage1 <- structure(
  list(
    data = data.frame(id = c("r1", "r2"), group = "A", weight = c(2, 3)),
    log = data.frame(group = "A"),
    achieved = NULL,
    provenance = list(method = "stage1")
  ),
  class = "wf_weights"
)
stage2 <- stage1
stage2$data$weight <- c(0.5, 2)
stage2$provenance$method <- "stage2"
wf_compose(stage1, stage2)

Diagnose calibrated weights

Description

Computes per-group quality metrics for a wf_weights object.

Usage

wf_diagnose(w, target = NULL, sample = NULL, deff_ok = 3, deff_caveat = 10)

Arguments

Value

A wf_diagnostics object.

Declare calibration dimensions

Description

Declare calibration dimensions

Usage

wf_dims(..., .collapse = list())

Arguments

Value

A wf_dims object.

Examples

dims <- wf_dims(gender = c("female", "male"), age = c("young", "old"))
dims

Execute an approved plan and lock its weights

Description

Revalidates the plan, human approval, design data, target evidence, and any stored cell map before calling exactly the method recorded in the plan.

Usage

wf_execute_plan(plan, approval, design, target)

Arguments

Value

A wf_locked_weights object that also inherits from wf_weights.

Execute a guided workflow with an external human approval

Description

Execute a guided workflow with an external human approval

Usage

wf_guided_execute(workflow, approval)

Arguments

Value

A wf_locked_weights object.

Prepare a guided safe weighting workflow

Description

This practitioner-oriented entry point composes design preparation, verified target import, outcome-blind cell planning, and weight planning. It never approves a plan or computes weights.

Usage

wf_guided_plan(
  data,
  id,
  calibration,
  dims,
  target_file,
  source_file,
  source_type = c("population", "reference"),
  key_map = NULL,
  count = NULL,
  feature = NULL,
  ...
)

Arguments

Value

A reviewable wf_safe_workflow; no weights are computed.

Import a verified external reference-sample target

Description

Imports an external reference sample only when every file column has a declared design role and the companion DCF record passes the provenance and checksum checks.

Usage

wf_import_reference(
  data_file,
  source_file,
  dims,
  feature,
  by = NULL,
  production = TRUE
)

Arguments

Value

A wf_verified_target object.

Import a verified external population target

Description

Imports CSV or Excel population margins only when a companion DCF record supplies complete provenance and a matching SHA-256 checksum.

Usage

wf_import_target(
  data_file,
  source_file,
  dims,
  key_map,
  count,
  by = NULL,
  by_key = NULL,
  production = TRUE
)

Arguments

Value

A wf_verified_target object.

Examples

# WFC ships a synthetic demo target beside its companion evidence record.
csv_file <- system.file(
  "extdata", "safe-target-example.csv",
  package = "WFC"
)
csv_source <- paste0(csv_file, ".source.dcf")

dims_safe <- wf_dims(sex = c("F", "M"), age = c("18-34", "35+"))

# `production = FALSE` is required here only because the bundled file is
# declared demo-only. Do not use it to admit an authoritative source whose
# evidence record is incomplete.
target <- wf_import_target(
  csv_file,
  csv_source,
  dims_safe,
  key_map = c(sex = "sex", age = "age"),
  count = "count",
  production = FALSE
)
target$identity

# The same demo file is refused as a production source, so a synthetic
# file cannot impersonate an authoritative population.
try(wf_import_target(
  csv_file,
  csv_source,
  dims_safe,
  key_map = c(sex = "sex", age = "age"),
  count = "count"
))

Diagnose high-influence calibrated units

Description

Reports the units that contribute most to weight instability. The core diagnostics are per-unit weight ratio to the group mean, share of the group's squared-weight mass, and leave-one-out design effect for the top units. When sample and target are supplied, WFC also reports the largest share of any target margin represented by that unit.

Usage

wf_influence(w, target = NULL, sample = NULL, id = NULL, top = 20)

Arguments

Value

A wf_influence object with a ⁠$table⁠ data frame.

Declare a production weighting pipeline

Description

Creates a serializable specification for a recurring weighting run. The specification records which verified target mode and calibration stage to use and which post-run validation thresholds to check.

Usage

wf_pipeline(target, stages, validate = NULL)

Arguments

Value

A wf_pipeline object.

Plan deterministic, outcome-blind support-cell merging

Description

Uses only declared design fields, external target margins, ordered dimension levels, and an optional explicit collapse ladder. Study outcomes and custom scoring callbacks are not accepted by this interface.

Usage

wf_plan_cells(
  design,
  target,
  dims,
  min_cell = 5,
  max_weight_ratio = 4,
  boundary = target$by,
  ladder = NULL
)

Arguments

Value

A reviewable wf_cell_merge_plan. No weights are computed.

Plan post-stratification cell resolution

Description

Resolves each population joint cell to the finest supported collapse-ladder level without computing weights. Review this plan before running wf_poststrat().

Usage

wf_plan_poststrat(
  sample,
  target,
  min_cell,
  ladder,
  granularity = c("adaptive", "province"),
  empty_cell = c("redistribute", "flag", "error"),
  id = NULL
)

Arguments

Value

A wf_poststrat_plan object.

Build a reviewable outcome-blind weight plan

Description

Performs structural checks and records approved-capable settings without calling a calibration engine or computing any weights.

Usage

wf_plan_weights(
  design,
  target,
  dims,
  method = c("raking", "logit", "poststrat"),
  bounds = c(0.3, 3),
  min_cell = 5,
  cell_plan = NULL
)

Arguments

Value

A wf_weight_plan. The weights field is always NULL.

Post-stratify verified design data to a verified joint target

Description

Enforces the WFC 2.0 verified-input boundary before cell-level post-stratification.

Usage

wf_poststrat(design, target, min_cell, ladder, ...)

Arguments

Value

A wf_weights object with cell_report and collapse_map.

Precheck sample and target compatibility

Description

Runs structural feasibility checks before weighting.

Usage

wf_precheck(
  sample,
  target,
  id = NULL,
  na = c("fractional", "drop", "error"),
  max_na_dims = 2,
  thin_min = 5,
  risk_ratio = 10
)

Arguments

Value

A wf_precheck object.

Examples

dims <- wf_dims(gender = c("female", "male"))
pop <- data.frame(gender = c("female", "male"), count = c(55, 45))
target <- wf_target_population(pop, c(gender = "gender"), "count", dims)
sample <- data.frame(id = 1:4, gender = c("female", "male", "female", "male"))
wf_precheck(sample, target, id = "id")

Prepare outcome-free survey design data

Description

Creates a strict design object containing only identifiers, calibration variables, and declared sampling-design fields. Any undeclared column blocks construction so study outcomes cannot silently enter weight planning.

Usage

wf_prepare_design(
  data,
  id,
  calibration,
  base_weight = NULL,
  strata = NULL,
  clusters = NULL,
  fpc = NULL
)

Arguments

Value

A wf_design_data object.

Examples

design_frame <- data.frame(
  id = sprintf("r%02d", 1:8),
  sex = rep(c("F", "M"), 4),
  age = rep(c("18-34", "35+"), each = 4),
  base_weight = 1,
  stringsAsFactors = FALSE
)

design <- wf_prepare_design(
  design_frame,
  id = "id",
  calibration = c("sex", "age"),
  base_weight = "base_weight"
)
design

# An outcome column has no declared design role, so it blocks construction
# instead of silently entering weight planning.
with_outcome <- cbind(
  design_frame,
  satisfaction = seq(40, 70, length.out = 8)
)
try(wf_prepare_design(
  with_outcome,
  id = "id",
  calibration = c("sex", "age"),
  base_weight = "base_weight"
))

Correct a non-probability sample by inverse-propensity pseudo-weighting.

Description

Fits the membership model declared in a wf_target_propensity() object and converts each online unit's fitted membership probability into a pseudo-design weight. The result is a wf_weights object suitable as an init_weight for wf_rake() / wf_poststrat() and as a stage in wf_compose().

Usage

wf_propensity(
  target,
  weight = c("ipw", "kernel", "matching"),
  stabilize = TRUE,
  trim = NULL
)

Arguments

Value

A wf_propensity_weights object inheriting from wf_weights, with ⁠$overlap⁠ and ⁠$balance⁠ diagnostics.

Rake verified design data to a verified external target

Description

Enforces the WFC 2.0 verified-input boundary before grouped raking.

Usage

wf_rake(design, target, ...)

Arguments

Value

A wf_weights object carrying verified input identities.

Generate re-calibrated replicate weights for variance estimation.

Description

Perturbs base weights by bootstrap, jackknife, or BRR multipliers and re-runs a calibration pipeline (refit) on each replicate, so the resulting variance captures calibration uncertainty. Pair with wf_variance().

Usage

wf_replicates(
  data,
  refit,
  method = c("bootstrap", "jackknife", "brr"),
  R = 500,
  strata = NULL,
  clusters = NULL,
  id = NULL,
  base_weight = NULL,
  seed = NULL,
  rho = 0,
  parallel = FALSE,
  progress = FALSE
)

Arguments

Value

A wf_replicate_weights object.

Build a weighting quality report

Description

Creates one structured dossier with manager and analyst projections. It can report calibrated/composed weights or a blend result and render the same payload as Markdown or dependency-free HTML.

Usage

wf_report(
  w,
  target = NULL,
  audience = c("manager", "analyst", "decision", "statistician"),
  lang = NULL,
  output = c("object", "markdown", "html"),
  file = NULL
)

Arguments

Value

A wf_quality_report for output = "object"; otherwise a rendered character string, or invisibly the structured report when file is used.

Examples

# Panel attrition weights are derived from the panel itself, so this
# example runs end to end. Reporting calibrated weights follows the same
# call; see vignette("safe-weighting-workflow") for that path, which
# additionally requires a verified target and a recorded approval.
panel <- data.frame(
  id = paste0("p", 1:24),
  region = rep(c("A", "B"), each = 12),
  retained = rep(c(TRUE, TRUE, TRUE, FALSE), times = 6),
  age = rep(c(25, 35, 45, 55, 65, 75), times = 4),
  sex = rep(c("female", "male"), times = 12),
  stringsAsFactors = FALSE
)
weights <- wf_attrition(
  panel,
  retained = "retained",
  formula = ~ age + sex,
  id = "id",
  by = "region"
)

wf_report(weights, audience = "decision")

analyst <- wf_report(weights, audience = "statistician")
names(analyst$sections)

Run a production weighting pipeline

Description

Executes a wf_pipeline() specification against verified design and target objects, records provenance, and evaluates declared validation thresholds.

Usage

wf_run(
  spec,
  sample,
  dims = NULL,
  population = NULL,
  reference = NULL,
  base_weight = NULL
)

Arguments

Value

A wf_weights object with pipeline provenance and optional ⁠$pipeline_validation⁠.

Suggest collapse plans from precheck findings

Description

Converts selected precheck issues into a reviewable collapse plan using collapse ladders declared in wf_dims().

Usage

wf_suggest_collapse(
  precheck,
  dims,
  checks = c("cat_infeasible", "support_thin", "risk_extreme_ratio"),
  max_steps = 1
)

Arguments

Value

A wf_collapse_plan object.

Draft a post-stratification collapse ladder

Description

Examines worst-group category support and proposes adjacent category merges in the explicit level order declared by wf_dims(). The result is a draft for review; no input is modified and no merge is applied automatically.

Usage

wf_suggest_ladder(sample, target, dims, min_cell = 5)

Arguments

Value

A wf_ladder_draft containing reviewable levels and a validated wf_collapse_ladder in ⁠$ladder⁠.

Examples

data(wfc_example)
target <- wf_target_population(
  wfc_example$population,
  c(gender = "gender", age = "age"),
  "count",
  wfc_example$dims,
  by = "province"
)
wf_suggest_ladder(wfc_example$sample, target, wfc_example$dims, min_cell = 25)

Target from external population data

Description

Converts arbitrary population data to a canonical wf_target object.

Usage

wf_target_population(
  pop,
  key_map,
  count,
  dims,
  by = NULL,
  by_key = NULL,
  scale = c("population", "sample", "custom"),
  sample = NULL,
  totals = NULL,
  keep_joint = FALSE
)

Arguments

Value

A wf_target object.

Examples

dims <- wf_dims(gender = c("female", "male"))
pop <- data.frame(gender = c("female", "male"), count = c(55, 45))
wf_target_population(pop, c(gender = "gender"), "count", dims)

Build a propensity target: stacked reference frame and membership model spec.

Description

Stacks a self-selected online sample and a probability reference sample into one frame with a membership indicator, so the online sample's selection propensity can be modelled. No model is fit here; execution happens in wf_propensity().

Usage

wf_target_propensity(
  online,
  reference,
  formula,
  method = c("logit", "rf", "gbm"),
  by = NULL,
  id = NULL
)

Arguments

Value

A wf_target_propensity object.

Target from a weighted reference sample

Description

Builds a canonical wf_target from a reference sample where feature is the reciprocal of the design weight.

Usage

wf_target_reference(
  ref,
  feature,
  dims,
  by = NULL,
  feature_na = c("error", "drop"),
  feature_gt1 = c("warn", "allow")
)

Arguments

Value

A wf_target object.

Examples

dims <- wf_dims(gender = c("female", "male"))
ref <- data.frame(gender = c("female", "male"), feature = c(0.5, 0.25))
wf_target_reference(ref, "feature", dims)

Create a safe external-target import template

Description

Writes a CSV or Excel data template and a separate companion DCF source record. The source filename appends .source.dcf to the complete data filename so CSV and Excel files cannot accidentally share a checksum.

Usage

wf_target_template(file, dims, by = NULL, example = FALSE)

Arguments

Value

Invisibly, a list with data_file and source_file paths.

Compare calibrated weights against a reference release

Description

Detects drift in recurring production weights by comparing group coverage, design effect, effective sample size, total weight, optional margin residuals, and per-unit weight ratios for matching IDs.

Usage

wf_validate(
  new,
  reference,
  target = NULL,
  max_deff_delta = 1,
  max_ess_loss = 0.2,
  max_total_shift = 0.05,
  max_margin_delta = 0.01,
  max_ratio_p99 = 2,
  on_issue = c("warn", "error", "none")
)

Arguments

Value

A wf_validation object.

Combine replicate weights and an estimator into a variance and CI.

Description

Applies the unified replication rule ⁠Var = scale * sum_r rscales_r * (theta_r - theta)^2⁠ to any estimator, using the ⁠(scale, rscales)⁠ stored by wf_replicates().

Usage

wf_variance(
  replicates,
  estimator,
  data,
  level = 0.95,
  ci = c("normal", "percentile")
)

Arguments

Value

A wf_variance_result object.

Broom-style projections for WFC results

Description

These S3 methods return base data frames and are registered for the generic names supplied by the suggested generics package. They do not require or return tibbles, and all programmatic column names remain stable English keys.

Usage

## S3 method for class 'wf_weights'
tidy(x, ...)

## S3 method for class 'wf_diagnostics'
tidy(x, ...)

## S3 method for class 'wf_blend_result'
tidy(x, ...)

## S3 method for class 'wf_variance_result'
tidy(x, ...)

## S3 method for class 'wf_weights'
glance(x, ...)

## S3 method for class 'wf_diagnostics'
glance(x, ...)

## S3 method for class 'wf_blend_result'
glance(x, ...)

## S3 method for class 'wf_variance_result'
glance(x, ...)

## S3 method for class 'wf_weights'
augment(x, data, id = "id", ...)

Arguments

Value

tidy() returns an object's primary detail table. glance() returns a one-row summary. augment() returns data with .weight and .feature appended in data-row order.

Simulated survey weighting example data

Description

A small, fully simulated dataset for examples and tests. It contains no private source records.

Usage

wfc_example

Format

A list with three elements:

sample

Simulated respondent-level sample data.

population

Simulated population cell counts.

dims

A wf_dims object for gender and age.

Source

Generated by data-raw/make-wfc-example.R.