event_study {did2s}R Documentation

Estimate event-study coefficients using TWFE and 5 proposed improvements.

Description

Uses the estimation procedures recommended from Borusyak, Jaravel, Spiess (2021); Callaway and Sant'Anna (2020); Gardner (2021); Roth and Sant'Anna (2021); Sun and Abraham (2020)

Usage

event_study(
  data,
  yname,
  idname,
  gname,
  tname,
  xformla = NULL,
  horizon = NULL,
  weights = NULL,
  estimator = c("TWFE", "did2s", "did", "impute", "sunab", "staggered", "all")
)

plot_event_study(out, seperate = TRUE, horizon = NULL)

Arguments

data

The dataframe containing all the variables

yname

Variable name for outcome variable

idname

Variable name for unique unit id

gname

Variable name for unit-specific date of initial treatment (never-treated should be zero or NA)

tname

Variable name for calendar period

xformla

A formula for the covariates to include in the model. It should be of the form ~ X1 + X2. Default is NULL.

horizon

Numeric. Vector of length 2. First element is min and second element is max of event_time to plot

weights

Variable name for estimation weights. This is used in estimating Y(0) and also augments treatment effect weights

estimator

Estimator you would like to use. Use "all" to estimate all. Otherwise see table to know advantages and requirements for each of these.

out

Output from event_study()

seperate

Logical. Should the estimators be on seperate plots? Default is TRUE.

Value

event_study returns a data.frame of point estimates for each estimator

plot_event_study returns a ggplot object that can be fully customized

Examples


out = event_study(
  data = did2s::df_het, yname = "dep_var", idname = "unit",
  tname = "year", gname = "g"
)
plot_event_study(out)


[Package did2s version 0.6.0 Index]