assess {calidad}R Documentation

Assess the quality of mean estimations

Description

assess evaluates the quality of mean estimations using the methodology created by INE Chile, which considers sample size, degrees of freedom, and coefficient of variation.

Usage

assess(
  table,
  publish = FALSE,
  scheme = c("chile", "eclac_2020", "eclac_2023"),
  domain_info = FALSE,
  ...
)

Arguments

table

dataframe created by crear_insumos_media.

publish

boolean indicating if the evaluation of the complete table must be added. If TRUE, the function adds a new column to the dataframe.

scheme

character variable indicating the evaluation protocol to use. Options are "chile", "eclac_2020", "eclac_2023".

domain_info

Logical. If TRUE, indicates that the study domain information is available and will be used for assessment. This affects how the evaluation is conducted, leveraging specific domain-level data to refine the assessment results. When FALSE, domain-specific adjustments are omitted, and a generalized assessment is performed.

...

additional parameters for the evaluation. The complete list of parameters is: 1. General Parameters

  • df degrees of freedom. Default: 9.

  • n sample size. Default for chile scheme: 60. Default for CEPAL schemes: 100.

2. chile Parameters

  • cv_lower_ine lower limit for CV. Default: 0.15.

  • cv_upper_ine upper limit for CV. Default: 0.3.

3. CEPAL 2020 Parameters

  • cv_cepal limit for CV. Default: 0.2.

  • ess effective sample size. Default: 140.

  • unweighted unweighted count. Default: 50.

  • log_cv logarithmic coefficient of variation. Default: 0.175.

4. CEPAL 2023 Parameters

  • cv_lower_cepal lower limit for CV. Default: 0.2.

  • cv_upper_cepal upper limit for CV. Default: 0.3.

  • ess effective sample size. Default: 60.

  • cvlog_max maximum logarithmic coefficient of variation. Default: 0.175.

  • CCNP_b unweighted count before adjustment. Default: 50.

  • CCNP_a unweighted count after adjustment. Default: 30.

Value

dataframe with all the columns included in the input table, plus a new column containing a label indicating the evaluation of each estimation: reliable, bit reliable, or unreliable.

Examples

dc <- survey::svydesign(ids = ~varunit, strata = ~varstrat, data = epf_personas, weights = ~fe)
assess(create_mean("gastot_hd", domains = "zona+sexo", design = dc))

[Package calidad version 0.6.0 Index]