parallel.trends.data: Creates a data frame used for generating plots to diagnose...

In skranz/ParallelTrendsPlot: Experimental Package: Plots to diagnose parallel trends in DID regression with additional control variables.

Description

The resulting data frame can be directly passed to parallel.trends.plot or be used to generate a more customized manual plot.

Usage

parallel.trends.data(
  df,
  timevar = "t",
  yvar = "y",
  treatdummy = "treat",
  expdummy = "exp",
  treat_exp_dummy = NULL,
  cvars = NULL,
  extravars = NULL,
  add.no.control = TRUE,
  constant.type = c("means", "zero")[1]
)

Arguments

df	The data frame used for the DID regression
timevar	Name of the time variable, e.g. "year"
yvar	Name of the dependent variable
treatdummy	Name of the 0-1 dummy variable indicating whether the observation belongs to the treatment group
expdummy	Name of the 0-1 dummy variable indicating whether the observation is in an experimental period, i.e. a period in which the treatment group gets treatment.
treat_exp_dummy	Optional name of the interaction variable treatdummy * expdummy. If not given a column with name "treat_exp_dummy" is added to the data frame.
cvars	Names of the additional control variables used in the DID regression.
extravars	Names of variables that are not used in the regression but shall be included in the returned data frame.
add.no.control	If TRUE (default) add rows that are used for a parallel trends diagnostic plot that assumes that no control variables are added.

Value

A data frame that can be passed to parallel.trends.plot. Essential columns are .t, .y, .group, .exp, .mode. Note that the data frame only contains rows that have no NA for the specified columns.

skranz/ParallelTrendsPlot documentation built on Dec. 23, 2021, 3:23 a.m.