Normalize and dispatch bootstrap input

Internal helper that standardizes bootstrap input supplied to calculate_bootstrap_ci() and handles early dispatch for boot objects. It contains no statistical logic itself, but controls workflow routing.

Usage

process_boot_input(
  bootstrap_results,
  grouping_var,
  type,
  conf,
  h,
  hinv,
  no_bias,
  boot_args
)

Arguments

bootstrap_results

A dataframe with bootstrap replicates, a single boot object, or a list of boot objects.

grouping_var

Optional character vector of grouping variables used to identify indicators or strata.

type

Character vector specifying the type(s) of confidence interval to compute (e.g. "perc", "bca", "norm", "basic").

conf

Numeric confidence level between 0 and 1.

h

Optional transformation function applied before CI calculation.

hinv

Optional inverse transformation function.

no_bias

Logical scalar. If TRUE, bias correction is skipped and boot objects are converted to a dataframe. If FALSE, confidence intervals are computed directly from the boot objects.

boot_args

Optional list of additional arguments passed to boot::boot.ci().

Value

A list with two elements:

• data: A dataframe of bootstrap replicates, or NULL if confidence intervals were computed directly

• result: A dataframe of confidence intervals, or NULL if the dataframe workflow should continue.

Details

The function supports three input modes:

Data frame input

Returned unchanged and passed on to the dataframe-based CI workflow.

boot input with no_bias = TRUE

boot objects are converted to a dataframe of bootstrap replicates and passed on to the dataframe-based CI workflow.

boot input with no_bias = FALSE

Confidence intervals are computed directly using boot::boot.ci() logic and returned immediately.

See also

Other indicator_uncertainty_helper: boot_list_to_dataframe(), bootstrap_cube_raw(), calculate_boot_ci_from_boot(), derive_bootstrap_method(), resolve_bootstrap_method()