A few things that need to be cleaned up as I've been thinking about the package/writing things up. Most of these are relatively minor, but non-trivial updates that should be done.
[x] Fix PlaceboTest to work with analytical standard errors.
[ ] optimizations to PlaceboTest for bootstraps
[ ] Exact matching should handle time-variant variables (Tabling this for now, will be integrated when caliper functionality is fully done)
[ ] exact matching should handle multiple variables (Tabling this for now, will be integrated when caliper functionality is fully done)
[ ] Refactor code
[x] Ensure that the results of DisplayTreatment are relatively normal/simple ggplot2 objects and can be manipulated with standard syntax. Right now the function seems complicated
[x] DisplayTreatment produces an object that is "backward". Fix it.
[x] balance_scatter function needs to be updated with the expectations of dimensions of the new get_covariate_balance results.
[x] Simplify the arguments of balance_scatter. No need to take an unrefined matched set, when we can just calculate it using get_covariate_balance from a refined set.
[x] Fix ART in moderator bug over in #92
[x] Standardize diagnostic function names/UIs. Utility/diagnostic functions should have similar formats/names. Eg. PlaceboTest and get_covariate_balance should follow similar schemes. Everything should take similar arguments and produce similar objects.
[x] Incorporate restrict.control.period option from big_data_checks
[x] update documentation/code results with resolution of ATE discussion
[x] (optional) include an additional covariate in the included Acemoglu data. This might be helpful for documenting some features
[x] implement the correct ATC
[x] remove continuous matching remnants for stable release
A few things that need to be cleaned up as I've been thinking about the package/writing things up. Most of these are relatively minor, but non-trivial updates that should be done.
DisplayTreatmentare relatively normal/simpleggplot2objects and can be manipulated with standard syntax. Right now the function seems complicatedbalance_scatterfunction needs to be updated with the expectations of dimensions of the newget_covariate_balanceresults.balance_scatter. No need to take an unrefined matched set, when we can just calculate it usingget_covariate_balancefrom a refined set.PlaceboTestandget_covariate_balanceshould follow similar schemes. Everything should take similar arguments and produce similar objects.restrict.control.periodoption frombig_data_checks