Skip to contents

Run a bulk batch of weighted survey statistics

Source: R/run_survey_stats.R
run_survey_stats.Rd

Drives survey_ci(), survey_chisq(), survey_ttest(), and survey_anova() over one or more datasets and writes the results to CSV. Each spec is a data frame describing the jobs for that analysis type; results from all jobs of a given type are stacked into one CSV (or merged into a single combined CSV if combined = TRUE).

Usage

run_survey_stats(
  datasets,
  ci_spec = NULL,
  chisq_spec = NULL,
  ttest_spec = NULL,
  anova_spec = NULL,
  output_dir = NULL,
  combined = FALSE
)

Arguments

datasets

Named list of data frames. Spec rows reference datasets by name via a dataset column.

ci_spec

Optional data frame of CI jobs. Required columns: dataset, variable, wt_var. Optional: group_var.

chisq_spec

Optional data frame of chi-square jobs. Required columns: dataset, var1, var2, wt_var.

ttest_spec

Optional data frame of t-test jobs. Required columns: dataset, outcome, group_var, wt_var.

anova_spec

Optional data frame of ANOVA jobs. Required columns: dataset, outcome, group_var, wt_var. Optional: pairwise (logical).

output_dir

Directory to write CSVs to. If NULL, results are not written; only returned.

combined

If TRUE, write a single survey_stats.csv with an analysis column instead of one CSV per analysis type. Pairwise ANOVA contrasts are always written separately when present.

Value

A named list of result data frames (ci, chisq, ttest, anova_overall, optionally anova_pairwise). The same data is written to disk when output_dir is supplied.

Examples

if (FALSE) { # \dontrun{
run_survey_stats(
  datasets = list(y6 = nptrends_y6_clean),
  ci_spec = data.frame(
    dataset = "y6",
    variable = c("LLAnyDisruption", "LLLost"),
    wt_var = "weight_year6plus",
    group_var = "SizeStrata"
  ),
  chisq_spec = data.frame(
    dataset = "y6",
    var1 = "SizeStrata", var2 = "LLAnyDisruption",
    wt_var = "weight_year6plus"
  ),
  output_dir = "stats_out"
)
} # }