Placebo DiD at fake treatment times

For each candidate fake time in placebo_times, redefines the post indicator and estimates a 2x2 DiD on pre-true-treatment data.

Usage

morie_did_placebo_test_time(
  data,
  outcome,
  treatment,
  time,
  true_treatment_time,
  placebo_times,
  covariates = NULL,
  cluster = NULL,
  alpha = 0.05
)

Arguments

data

Data frame.

outcome, treatment, time

Column names.

true_treatment_time

The actual treatment-onset time (data are restricted to pre-period observations).

placebo_times

Vector of candidate fake treatment times.

covariates

Optional character vector of covariate column names.

cluster

Optional cluster ID column for CR1 standard errors.

alpha

Significance level for confidence intervals (default 0.05).

Value

A data frame, one row per placebo time.

Examples

set.seed(1)
df <- data.frame(
  y = rnorm(500), d = rbinom(500, 1, 0.5),
  time = sample(1:8, 500, replace = TRUE)
)
out <- morie_did_placebo_test_time(df, "y", "d", "time",
                                   true_treatment_time = 7,
                                   placebo_times = c(3, 4, 5))
out
#>              placebo_time     estimate std_error   p_value significant
#> interaction             3  0.010094814 0.2271821 0.9645577       FALSE
#> interaction1            4 -0.002246453 0.2199366 0.9918505       FALSE
#> interaction2            5  0.025201883 0.2297443 0.9126510       FALSE