Big improvement on code base / functionality / testing
Deprecated mp.spatt function and replaced it with att_gt function
Calling att_gt is similar to calling mp.spatt; instead of formula for outcome of the form y~treat, now just pass the name of the outcome variable
Deprecated mp.spatt function and replaced it with conditional_did_pretest function
New est_method parameter. Can call any function for 2x2 DID in the DRDID package (default is now doubly robust estimation, but inverse probability weights and regression estimators are also supported) as well as provide custom 2x2 DID estimators
Bug fixes for including groups that are already treated in the first period
Allow for user to select control group – either never treated or not yet treated
Add functionality for uniform confidence bands for all aggregated treatment effect parameters
Introduced dynamic effects in pre-treatment periods. These allow for users to report event study plots that are common that include pre-treatment periods and are common in applied work. The event study plots in the did package are robust to selective treatment timing (unlike standard regression event study plots)
Support for using repeated cross sections data instead of panel data is much improved
Support for using sampling weights is much improved
Big improvement to website, vignettes, and code documentation
Code for dealing with unbalanced panels
Allow for event studies to be computed over subsets of event times
Allow for treatment anticipation via anticipation argument
Improved ways to summarize aggregated treatment effect parameters
Fixed bug related to needing new version of BMisc
Fixed bug related to plotting with no pre-treatment periods
Improved ways to easily plot aggregated treatment effect parameters
Added some error handling for some cases with small group sizes, and fixed some cryptic error messages
Fixes handling for data being in format besides data.frame (e.g. tibble)
Add warnings about small group sizes which are a very common source of trouble
fixed issues between BMisc and formula.tools
added point estimates for repeated cross sections data