Make event study or parallel trends plots — didint

didint_plot() produces either event study plots or parallel trends plots depending on what is specified via the event argument. The parallel trends plots, as well as the event study plots, are created using the means residualized by covariates under different model specifications that account for different violations of the common causal covaraites (CCC) assumptions.

Usage

didint_plot(
  outcome,
  state,
  time,
  data,
  gvar = NULL,
  treated_states = NULL,
  treatment_times = NULL,
  date_format = NULL,
  covariates = NULL,
  ref = NULL,
  ccc = "all",
  event = FALSE,
  weights = TRUE,
  ci = 0.95,
  freq = NULL,
  freq_multiplier = 1,
  start_date = NULL,
  end_date = NULL,
  hc = "hc1"
)

Arguments

outcome: A string giving the column name of the outcome variable.
state: A string giving the column identifying states. The state column should be a character column.
time: A string giving the column identifying dates.
data: A data frame containing the variables used for estimation.
gvar: String giving the column that indicates first treatment time for each state. Use either this option or the combination of treated_states and treatment_times.
treated_states: Character values specifying the treated state(s).
treatment_times: Specify the treated_states using strings, numbers, or Dates, corresponding to treated_states.
date_format: Optional string specifying the input date format when dates are supplied as character strings. Applies to start_date, end_date, treatment_times and the data in the time column if any of those are strings.
covariates: Optional string or vector of strings specifying covariates to include.
ref: Optional named list indicating the reference category for categorical covariates.
ccc: A string specifying the DID-INT specification. Any combination of "none", "hom", "time", "state", "add", and "int". Or, alternatively, "all" (default).
event: A logical value used to specify if event study plots should be made (TRUE) or if parallel trends plots should be made (FALSE).
weights: A logical value, if TRUE, estimates for the event study plot are computed as weighted averages of state level means for each period relative to their treatment period; if FALSE, uses unweighted averages.
ci: A number between 0 and 1 used to specify the size of the confidence bands.
freq: Optional string specifying the period length for staggered adoption. One of "year", "month", "week", "day".
freq_multiplier: Integer multiplier for freq. Default is 1.
start_date: Optional earliest date to retain in the data.
end_date: Optional latest date to retain in the data.
hc: Heteroskedasticity-consistent covariance matrix estimator. One of "hc0", "hc1", "hc2", "hc3", "hc4".

Value

An object of class DiDIntPlotObj, a list containing the parallel trends data or event study data (if event is set to TRUE) and the name of the outcome variable. Has an associated plot.DiDIntPlotObj method for producing event study or parallel trends plots.

Details

The arguments treated_states and treatment_times must be supplied such that their ordering corresponds with one another. That is, the first element of treated_states refers to the state treated at the date given by the first element of treatment_times, and so on.

Dates can be entered as strings, numbers, or Date objects. When character strings are supplied, the input format must be specified via the date_format argument (e.g. "yyyy-mm-dd").

Period grids are constructed automatically, based on the inputted data Otherwise, the period grid can be created manually using the arguments freq, freq_multiplier, start_date, end_date. More information on this process can be seen on the didintrjl documentation site: https://ebjamieson97.github.io/didintrjl/.

References

Karim & Webb (2025). Good Controls Gone Bad: Difference-in-Differences with Covariates. https://arxiv.org/abs/2412.14447

Examples

if (Sys.getenv("NOT_CRAN") == "true" && didintrjl_ready()) {
 file_path <- system.file("extdata", "merit.csv", package = "didintrjl")
 df <- utils::read.csv(file_path)
 res_event <- didint_plot(
   "coll", "state", "year", df, event = TRUE,
   treated_states = c(71, 58, 64, 59, 85, 57, 72, 61, 34, 88),
   treatment_times = c(1991, 1993, 1996, 1997, 1997, 1998, 1998, 1999,
                       2000, 2000),
   covariates = c("asian", "black", "male")
 )
 plot(res_event)
 DONTSHOW({
   JuliaConnectoR::stopJulia()
 })
}
#> Starting Julia ...