Skip to contents

Runs a local IPM reproducibility workflow: validates inputs, estimates IPM KPIs, loads published IPM benchmarks, and compares local estimates to those benchmarks. It consumes already-built score and flag variables or can build flags from registered component columns. It does not implement new raw IPM component rules and does not read microdata files directly.

Usage

enemdu_run_ipm_reproducibility(
  data,
  period = "2025-12",
  survey_type = "anual",
  by = "area",
  ids = "upm",
  strata = "estrato",
  weight = "fexp",
  build_components = FALSE,
  build_flags = TRUE,
  strict = TRUE,
  tolerance_pp = 0.5,
  missing_component_policy = c("error", "complete_case"),
  ...
)

Arguments

data

A data frame.

period

Benchmark period.

survey_type

ENEMDU survey type.

by

Optional domain variable. Defaults to "area".

ids

Primary sampling unit variable.

strata

Survey strata variable.

weight

Survey expansion factor variable.

build_components

Logical. If TRUE, call enemdu_build_ipm_components() before flag validation.

build_flags

Logical. If TRUE, build row-level flags from registered components when available.

strict

Logical. If TRUE, abort on invalid reproducibility inputs.

tolerance_pp

Benchmark-comparison tolerance in percentage points.

missing_component_policy

Missing IPM evidence policy. "error" refuses incomplete registered components or score/flag inputs. "complete_case" excludes rows with incomplete registered components, or incomplete score/flag inputs when components are unavailable, before estimation.

...

Additional named arguments passed to enemdu_kpi_ipm() and, when relevant, enemdu_build_ipm_components().

Value

A structured list with preflight, validation, estimates, benchmarks, comparison, complete-case diagnostics, and non-official validation metadata.

Details

Benchmarks are published comparison values, not institutional validation.