Diagnose and compare designs.

compare_diagnoses(
design1,
design2,
sims = 500,
bootstrap_sims = 100,
merge_by_estimator = TRUE,
alpha = 0.05
)

## Arguments

design1 A design or a diagnosis. A design or a diagnosis. The number of simulations, defaulting to 1000. sims may also be a vector indicating the number of simulations for each step in a design, as described for simulate_design. Used for both designs. Number of bootstrap replicates for the diagnosands to obtain the standard errors of the diagnosands, defaulting to 1000. Set to FALSE to turn off bootstrapping. Used for both designs. Must be greater or equal to 100. A logical. Whether to include estimator_label in the set of columns used for merging. Defaults to TRUE. The significance level, 0.05 by default.

## Value

A list with a data.frame of compared diagnoses and both diagnoses.

## Details

The function compare_diagnoses runs a many-to-many merge matching by estimand_label and term (if present). If merge_by_estimator equals TRUE, estimator_label is also included in the merging condition. Any diagnosand that is not included in both designs will be dropped from the merge.

## Examples

design_a <- declare_population(N = 100, u = rnorm(N)) +
declare_potential_outcomes(
Y_Z_0 = u,
Y_Z_1 = u + rnorm(N, mean = 2, sd = 2)) +
declare_assignment(prob = 0.5) +
declare_estimand(ATE = mean(Y_Z_1 - Y_Z_0)) +
reveal_outcomes() +
declare_estimator(Y ~ Z, estimand = "ATE")

design_b <- replace_step(design_a, step = "assignment", declare_assignment(prob = 0.3) )

comparison <- compare_diagnoses(design_a, design_b, sims = 40)