declare_estimator {DeclareDesign}R Documentation

Declare estimator


Declares an estimator which generates estimates and associated statistics.

Use of declare_test is identical to use of declare_estimator. Use declare_test for hypothesis testing with no specific inquiry in mind; use declare_estimator for hypothesis testing when you can link each estimate to an inquiry. For example, declare_test could be used for a K-S test of distributional equality and declare_estimator for a difference-in-means estimate of an average treatment effect.


  handler = label_estimator(method_handler),
  label = "estimator"

  handler = label_estimator(method_handler),
  label = "estimator"


  .method = estimatr::lm_robust,
  .summary = tidy_try,
  term = FALSE



arguments to be captured, and later passed to the handler


a tidy-in, tidy-out function


a string describing the step


A function that takes a data.frame as an argument and returns a data.frame with the estimates, summary statistics (i.e., standard error, p-value, and confidence interval), and a term column for labeling coefficient estimates.


a data.frame


A method function, e.g. lm or glm. By default, the method is the lm_robust function from the estimatr package, which fits OLS regression and calculates robust and cluster-robust standard errors.


A method-in data-out function to extract coefficient estimates or method summary statistics, such as tidy or glance. By default, the DeclareDesign method summary function tidy_try is used, which first attempts to use the available tidy method for the method object sent to method, then if not attempts to summarize coefficients using the coef(summary()) and confint methods. If these do not exist for the method object, it fails.


Deprecated argument. Use .method instead.


Deprecated argument. Use .summary instead.


Symbols or literal character vector of term that represent quantities of interest, i.e. Z. If FALSE, return the first non-intercept term; if TRUE return all term. To escape non-standard-evaluation use !!.


declare_estimator is designed to handle two main ways of generating parameter estimates from data.

In declare_estimator, you can optionally provide the name of an inquiry or an objected created by declare_inquiry to connect your estimate(s) to inquiry(s).

The first is through label_estimator(method_handler), which is the default value of the handler argument. Users can use standard method functions like lm, glm, or iv_robust. The methods are summarized using the function passed to the summary argument. This will usually be a "tidier" like broom::tidy. The default summary function is tidy_try, which applies a tidy method if available, and if not, tries to make one on the fly.

An example of this approach is:

declare_estimator(Y ~ Z + X, .method = lm_robust, .summary = tidy, term = "Z", inquiry = "ATE")

The second approach is using a custom data-in, data-out function, usually first passed to label_estimator. The reason to pass the custom function to label_estimator first is to enable clean labeling and linking to inquiries.

An example of this approach is:

my_fun <- function(data){ with(data, median(Y[Z == 1]) - median(Y[Z == 0])) }

declare_estimator(handler = label_estimator(my_fun), inquiry = "ATE")

label_estimator takes a data-in-data out function to fn, and returns a data-in-data-out function that first runs the provided estimation function fn and then appends a label for the estimator and, if an inquiry is provided, a label for the inquiry.


A function that accepts a data.frame as an argument and returns a data.frame containing the value of the estimator and associated statistics.


# Setup for examples
design <-
    N = 500,
    gender = rbinom(N, 1, 0.5),
    U = rnorm(N, sd = 0.25),
    potential_outcomes(Y ~ rbinom(
      N, 1, prob = pnorm(0.2 * Z + 0.2 * gender + 0.1 * Z * gender + U)
  ) +
  declare_inquiry(ATE = mean(Y_Z_1 - Y_Z_0)) +
  declare_sampling(S = complete_rs(N = N, n = 200)) +
  declare_assignment(Z = complete_ra(N = N, m = 100)) +
  declare_measurement(Y = reveal_outcomes(Y ~ Z))

# default estimator is lm_robust with tidy summary
design_0 <-
  design +
  declare_estimator(Y ~ Z, inquiry = "ATE")


# Linear regression using lm_robust and tidy summary
design_1 <-
  design +
    formula = Y ~ Z,
    .method = lm_robust,
    .summary = tidy,
    term = "Z",
    inquiry = "ATE",
    label = "lm_no_controls"


# Use glance summary function to view model fit statistics
design_2 <-
  design +
  declare_estimator(.method = lm_robust,
                    formula = Y ~ Z,
                    .summary = glance)


# Use declare_estimator to implement custom answer strategies
my_estimator <- function(data) {
  data.frame(estimate = mean(data$Y))

design_3 <-
  design +
  declare_inquiry(Y_bar = mean(Y)) +
  declare_estimator(handler = label_estimator(my_estimator),
                    label = "mean",
                    inquiry = "Y_bar")


# Use `term` to select particular coefficients
design_4 <-
  design +
  declare_inquiry(difference_in_cates = mean(Y_Z_1[gender == 1] - Y_Z_0[gender == 1]) -
                    mean(Y_Z_1[gender == 0] - Y_Z_0[gender == 0])) +
  declare_estimator(Y ~ Z * gender,
                    term = "Z:gender",
                    inquiry = "difference_in_cates",
                    .method = lm_robust)


if(require("broom")) {

  # Use glm from base R
  design_5 <-
    design +
    declare_estimator(Y ~ Z + gender,
                      family = "gaussian",
                      inquiry = "ATE",
                      .method = glm)
  # If we use logit, we'll need to estimate the average marginal effect with
  # marginaleffects::avg_slopes. We wrap this up in a function we'll pass to 
  # .summary.
  if(require("marginaleffects")) {
    library(marginaleffects) # for predictions
    library(broom) # for tidy
    tidy_avg_slopes <- function(x) {
    design_6 <-
      design +
        Y ~ Z + gender,
        .method = glm,
        family = binomial("logit"),
        .summary = tidy_avg_slopes,
        term = "Z"
    # Multiple estimators for one inquiry
    design_7 <-
      design +
      declare_estimator(Y ~ Z,
                        .method = lm_robust,
                        inquiry = "ATE",
                        label = "OLS") +
        Y ~ Z + gender,
        .method = glm,
        family = binomial("logit"),
        .summary = tidy_avg_slopes,
        inquiry = "ATE",
        term = "Z",
        label = "logit"


[Package DeclareDesign version 1.0.10 Index]