set_diagnosands {DeclareDesign}

Set the diagnosands for a design

Description

A researcher often has a set of diagnosands in mind to appropriately assess the quality of a design. set_diagnosands sets the default diagnosands for a design, so that later readers can assess the design on the same terms as the original author. Readers can also use diagnose_design to diagnose the design using any other set of diagnosands.

Usage

set_diagnosands(x, diagnosands = default_diagnosands)

Arguments

x

A design typically created using the + operator, or a simulations data.frame created by simulate_design.

diagnosands

A set of diagnosands created by declare_diagnosands

Value

a design object with a diagnosand attribute

Examples

# Two-arm randomized experiment
design <-
  declare_model(
    N = 500,
    gender = rbinom(N, 1, 0.5),
    X = rep(c(0, 1), each = N / 2),
    U = rnorm(N, sd = 0.25),
    potential_outcomes(Y ~ 0.2 * Z + X + 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)) +
  declare_estimator(Y ~ Z, inquiry = "ATE")

# You can choose your own diagnosands instead of the defaults:

my_diagnosands <-
  declare_diagnosands(median_bias = median(estimate - estimand))

## Not run: 
## You can set diagnosands with set_diagnosands 
design <- set_diagnosands(design, diagnosands = my_diagnosands)
diagnosis <- diagnose_design(design)
diagnosis

## Using set_diagnosands to diagnose simulated data
simulations_df <- simulate_design(design)

simulations_df <- set_diagnosands(simulations_df, my_diagnosands)

diagnose_design(simulations_df)

# If you do not specify diagnosands in diagnose_design, 
#   the function default_diagnosands() is used, 
#   which is reproduced below.

alpha <- 0.05

default_diagnosands <- 
  declare_diagnosands(
    mean_estimand = mean(estimand),
    mean_estimate = mean(estimate),
    bias = mean(estimate - estimand),
    sd_estimate = sqrt(pop.var(estimate)),
    rmse = sqrt(mean((estimate - estimand) ^ 2)),
    power = mean(p.value <= alpha),
    coverage = mean(estimand <= conf.high & estimand >= conf.low)
  )
  
diagnose_design(
  simulations_df,
  diagnosands = default_diagnosands
)

# A longer list of potentially useful diagnosands might include:

extended_diagnosands <- 
  declare_diagnosands(
    mean_estimand = mean(estimand),
    mean_estimate = mean(estimate),
    bias = mean(estimate - estimand),
    sd_estimate = sd(estimate),
    rmse = sqrt(mean((estimate - estimand) ^ 2)),
    power = mean(p.value <= alpha),
    coverage = mean(estimand <= conf.high & estimand >= conf.low),
    mean_se = mean(std.error),
    type_s_rate = mean((sign(estimate) != sign(estimand))[p.value <= alpha]),
    exaggeration_ratio = mean((estimate/estimand)[p.value <= alpha]),
    var_estimate = pop.var(estimate),
    mean_var_hat = mean(std.error^2),
    prop_pos_sig = mean(estimate > 0 & p.value <= alpha),
    mean_ci_length = mean(conf.high - conf.low)
  )
  
diagnose_design(
  simulations_df,
  diagnosands = extended_diagnosands
)


## End(Not run)