Checklist

[omkarakatta]

• Evaluate empirical costs at all bandwidths, not just "acceptable" bandwidths
• Add sections to code
• Choose better names for output objects
• Remove unnecessary arguments
• Error checks
  • check that pre_main_subsample_size and others is not NULL before checking it is an integer
  • choose defaults for subsample size? maybe not
  • ignore *_control_subsample_size if estimator is "ba"
• Update documentation
• Create print and summary methods
  • print --- basic output
  • summary --- full output
  • summary_placebo --- summarize placebo results
  • summary_subsample --- summarize subsample results
• Create plot and table methods
  • plot_data_hist --- overlayed histograms of pre- and post-distributions
  • plot_placebo_summary --- plot summary results of placebo vs. bandwidth
  • plot_empirical --- plot empirical results vs. bandwidth
    • add option for plotting placebo as well
    • add option for plotting treated and control, not just the difference
  • plot_placebo --- plot placebo results vs. bandwidth (add option for plotting empirical as well)
  • plot_subsample --- plot histogram of subsample results
• Create retrieval methods
  • retrieve_placebo --- send placebo results
  • retrieve_subsample --- send subsample results
  • retrieve_empirical --- send empirical results

[omkarakatta]

• order bandwidth_vec; take unique values; ensure that it has non-zero length

[omkarakatta]

• Find a way to hide the vlines and hlines

[omkarakatta]

• Fix the default bandwidth selection criterion for difference-in-transports estimator (use d-d for bandwidth selection and use 2d-d for empirical computation and subsampling); allow option to use 2d-d for bandwidth selection criterion
• Go over subsampling procedure

[omkarakatta]

• Organize main and control distributions so they are in order according to the support
• Ensure that placebo is less than 0.05% (before-and-after)

[omkarakatta]

• Choose a random seed as default