R Guide

Basic Usage

library(dlm)

mod <- distributed_lags_model(
  data = outcome_data,
  exposure_data = treatment_data,
  from_rt = -3, to_rt = 3,
  outcome = "y", exposure = "treated",
  unit = "id", time = "year"
)

The R version takes separate data frames for outcomes and exposure. This is useful when the exposure is defined at a different level (e.g., pharmacy-level treatment merged onto beneficiary-level outcomes).

Data Setup

The R DLM expects two data frames:

1. data — contains unit, time, outcome, and any covariates
2. exposure_data — contains unit, time, and the exposure variable (one row per unit-time)

library(dlm)
library(dplyr)

# Example: separate outcome and exposure data
outcome_data <- full_data %>% select(id, year, y, x1, x2)
exposure_data <- full_data %>% select(id, year, treated) %>% distinct()

If your data is a single data frame, just select the relevant columns:

df <- generate_data(seed = 42)

data <- df %>% select(group, time, outcome)
exposure_data <- df %>% select(group, time, post) %>% distinct()

Examples

Basic estimation

library(dlm)
library(dplyr)

df <- generate_data(seed = 42, n_groups = 500, n_times = 20, treat_prob = 0.4)

data <- df %>% select(group, time, outcome)
exposure_data <- df %>% select(group, time, post) %>% distinct()

mod <- distributed_lags_model(
  data = data,
  exposure_data = exposure_data,
  from_rt = -3, to_rt = 3,
  outcome = "outcome", exposure = "post",
  unit = "group", time = "time"
)

mod$betas

Output:

           time_to_event        coef        se
post_lead2            -3 -0.11846256 0.4729260
post_lead1            -2 -0.06327287 0.5139965
post_lag0              0 -2.64104056 0.5438700
post_lag1              1 -2.26029265 0.5234023
post_lag2              2 -3.04210098 0.5674556
post_lag3              3 -2.61751913 0.4214689

Custom reference period

mod <- distributed_lags_model(
  data = data,
  exposure_data = exposure_data,
  from_rt = -3, to_rt = 3,
  outcome = "outcome", exposure = "post",
  unit = "group", time = "time",
  ref_period = -2
)

With covariates

data_with_covs <- df %>% select(group, time, outcome, x1, x2)

mod <- distributed_lags_model(
  data = data_with_covs,
  exposure_data = exposure_data,
  from_rt = -3, to_rt = 3,
  outcome = "outcome", exposure = "post",
  unit = "group", time = "time",
  covariates = c("x1", "x2")
)

Additional fixed effects

mod <- distributed_lags_model(
  data = data,
  exposure_data = exposure_data,
  from_rt = -3, to_rt = 3,
  outcome = "outcome", exposure = "post",
  unit = "group", time = "time",
  addl_fes = c("region")
)

Multiple outcomes

Use distributed_lags_models (plural) to estimate multiple outcomes:

mods <- distributed_lags_models(
  data = data,
  exposure_data = exposure_data,
  from_rt = -3, to_rt = 3,
  outcomes = c("outcome1", "outcome2"),
  exposure = "post",
  unit = "group", time = "time"
)

# Access each model
mods[[1]]$betas
mods[[2]]$betas

Weighted regression

mod <- distributed_lags_model(
  data = data,
  exposure_data = exposure_data,
  from_rt = -3, to_rt = 3,
  outcome = "outcome", exposure = "post",
  unit = "group", time = "time",
  weights = "pop_weight"
)

Working with Results

The returned object is a list:

names(mod)
# [1] "betas" "plot" "model" "vcov" "data_periods_included"
# [6] "fmla_str" "from_rt" "to_rt" "exposure" "outcome"

Beta coefficients

mod$betas
#            time_to_event        coef        se
# post_lead2            -3 -0.11846256 0.4729260
# post_lead1            -2 -0.06327287 0.5139965
# post_lag0              0 -2.64104056 0.5438700
# post_lag1              1 -2.26029265 0.5234023
# post_lag2              2 -3.04210098 0.5674556
# post_lag3              3 -2.61751913 0.4214689

Built-in event-study plot

mod$plot

The plot is a ggplot2 object. Customize it:

mod$plot +
  ggplot2::labs(title = "My Event Study", x = "Years to Treatment") +
  ggplot2::theme_minimal()

Access the underlying fixest model

summary(mod$model)
fixest::coeftable(mod$model)

Test Data Generator

df <- generate_data(seed = 42, n_groups = 500, n_times = 20, treat_prob = 0.4)

Argument	Default	Description
seed	1234	Random seed
n_groups	676	Number of panel units
n_times	20	Number of time periods
treat_prob	0.4	Treatment probability

Key variables: group, time, treat, treatment_time, years_to_treatment, post, outcome.