Deprecated. Please use the reveal_outcomes function within a declare_measurement declaration.
Usage
declare_reveal(..., handler = declare_reveal_handler, label = NULL)
declare_reveal_handler(
data = NULL,
outcome_variables = Y,
assignment_variables = Z,
attrition_variables = NULL,
...
)Arguments
- ...
arguments to be captured, and later passed to the handler
- handler
a tidy-in, tidy-out function
- label
a string describing the step
- data
A data.frame containing columns for assignment and potential outcomes.
- outcome_variables
The outcome prefix(es) of the potential outcomes.
- assignment_variables
Unquoted name(s) of the assignment variable(s).
- attrition_variables
Unquoted name of the attrition variable.
Details
Potential outcomes declarations indicate what outcomes would obtain for different possible values of assignment variables.
But realized outcomes need to be "revealed."
declare_reveal generates these realized outcomes using information on
potential outcomes (for instance generated via declare_potential_outcomes) and the relevant
assignment variables (for example created by declare_assignment).
Revelation steps are usefully included after declaration of all assignments of conditions required to determine the realized outcome.
If a revelation is not declared, DeclareDesign will try to guess appropriate revelations. Explicit revelation is recommended however.
declare_reveal declares how outcomes should be realized.
A "revelation" uses the random assignment to pluck out the correct potential outcomes (Gerber and Green 2012, Chapter 2).
Revelation requires that every named outcome variable is a function of every named assignment variable within a step. Thus if multiple outcome variables depend on different assignment variables, multiple revelations are needed.
Examples
design <-
declare_model(
N = 100,
U = rnorm(N),
Y_Z_0 = U,
Y_Z_1 = U + rnorm(N, mean = 2, sd = 2)
) +
declare_assignment(Z = complete_ra(N, m = 50)) +
declare_measurement(Y = reveal_outcomes(Y ~ Z))
head(draw_data(design))
#> ID U Y_Z_0 Y_Z_1 Z Y
#> 1 001 -0.4587637 -0.4587637 -0.7492207 1 -0.7492207
#> 2 002 -1.4140904 -1.4140904 0.4659158 0 -1.4140904
#> 3 003 1.5278572 1.5278572 5.2125382 0 1.5278572
#> 4 004 -0.7434868 -0.7434868 2.4997315 0 -0.7434868
#> 5 005 -0.1596565 -0.1596565 -1.1361638 0 -0.1596565
#> 6 006 -0.0715566 -0.0715566 2.2933125 1 2.2933125
# Declaring multiple assignment variables or multiple outcome variables
design <-
declare_model(
N = 10,
potential_outcomes(Y1 ~ Z),
potential_outcomes(Y2 ~ 1 + 2 * Z),
potential_outcomes(Y3 ~ 1 - X * Z, conditions = list(X = 0:1, Z = 0:1))
) +
declare_assignment(Z = complete_ra(N)) +
declare_assignment(X = complete_ra(N)) +
declare_measurement(Y1 = reveal_outcomes(Y1 ~ Z),
Y2 = reveal_outcomes(Y2 ~ Z),
Y3 = reveal_outcomes(Y3 ~ X + Z))
head(draw_data(design))
#> ID Y1_Z_0 Y1_Z_1 Y2_Z_0 Y2_Z_1 Y3_X_0_Z_0 Y3_X_1_Z_0 Y3_X_0_Z_1 Y3_X_1_Z_1 Z
#> 1 01 0 1 1 3 1 1 1 0 0
#> 2 02 0 1 1 3 1 1 1 0 0
#> 3 03 0 1 1 3 1 1 1 0 0
#> 4 04 0 1 1 3 1 1 1 0 1
#> 5 05 0 1 1 3 1 1 1 0 1
#> 6 06 0 1 1 3 1 1 1 0 1
#> X Y1 Y2 Y3
#> 1 0 0 1 1
#> 2 1 0 1 1
#> 3 0 0 1 1
#> 4 1 1 3 0
#> 5 0 1 3 1
#> 6 0 1 3 1
design <-
declare_model(
N = 100,
age = sample(18:95, N, replace = TRUE),
potential_outcomes(Y ~ .25 * Z + .01 * age * Z),
potential_outcomes(R ~ rbinom(n = N, size = 1, prob = pnorm(Y_Z_0)))
) +
declare_assignment(Z = complete_ra(N, m = 25))
declare_measurement(R = reveal_outcomes(R ~ Z),
Y = reveal_outcomes(Y ~ Z),
Y = ifelse(R == 1, Y, NA))
#> declare_measurement(R = reveal_outcomes(R ~ Z), Y = reveal_outcomes(Y ~
#> Z), Y = ifelse(R == 1, Y, NA))
head(draw_data(design))
#> ID age Y_Z_0 Y_Z_1 R_Z_0 R_Z_1 Z
#> 1 001 61 0 0.86 0 0 0
#> 2 002 55 0 0.80 0 0 1
#> 3 003 31 0 0.56 1 1 1
#> 4 004 59 0 0.84 1 1 0
#> 5 005 50 0 0.75 1 1 0
#> 6 006 40 0 0.65 1 0 0