Skip to contents

Select the BPP futility threshold minimising null expected sample size

select_lambda_star.Rd

Among candidate thresholds for which the lower confidence bound on power is at least power_floor under every alternative scenario, selects the one with the smallest expected sample size under the null scenario.

Usage

select_lambda_star(
  summary_by_scenario,
  power_floor,
  null_scenario,
  alt_scenarios
)

Arguments

summary_by_scenario

A named list of data frames, one per scenario, each as returned by summarize_grid_by_lambda over the same lambda_grid.

power_floor

Minimum acceptable lower confidence bound on power.

null_scenario

Name of the null scenario in summary_by_scenario.

alt_scenarios

Character vector of alternative scenario names.

Value

A list with lambda_star (the selected threshold, or NA with a warning if none is feasible) and feasibility_table.

See also

run_calibration_grid, summarize_grid_by_lambda