Interpret Qini Results for Both Binary and Multi-Arm Treatments

This function interprets Qini results for both binary and multi-arm treatments. It detects the treatment type based on the input data structure and calls the appropriate sub-function.

Usage

margot_interpret_qini(
  multi_batch,
  label_mapping = NULL,
  alpha = 0.05,
  decimal_places = 2
)

Arguments

multi_batch

List output from margot_batch_policy() with diff_gain_summaries.

label_mapping

Optional named list mapping model names to readable labels.

alpha

Significance level for confidence intervals. Default 0.05.

decimal_places

Number of decimal places for rounding. Default 2.

Value

List with summary tables and explanations. Format varies for binary and multi-arm treatments.

References

Sverdrup, E., Wu, H., Athey, S., & Wager, S. (2024). Qini Curves for Multi-Armed Treatment Rules. arXiv preprint arXiv:2306.11979.

Examples

if (FALSE) { # \dontrun{
label_mapping <- list(
  "t2_env_not_climate_chg_concern_z" = "Deny Climate Change Concern",
  "t2_env_not_climate_chg_cause_z" = "Deny Humans Cause Climate Change"
)

binary_results <- margot_interpret_qini(binary_multi_batch, label_mapping = label_mapping)
print(binary_results$summary_table)

multi_arm_results <- margot_interpret_qini(multi_arm_multi_batch, label_mapping = label_mapping)
print(multi_arm_results$arm1$summary_table)
} # }