Declare Data Strategy: Assignment
Usage
declare_assignment(..., handler = assignment_handler, label = NULL)
assignment_handler(data, ..., legacy = FALSE)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.
- legacy
Use the legacy randomizr functionality. This will be disabled in future; please use legacy = FALSE.
Value
A function that takes a data.frame as an argument and returns a data.frame with assignment columns appended.
Examples
# declare_assignment in use
## Two-arm randomized experiment
design <-
declare_model(
N = 500,
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")
run_design(design)
#> inquiry estimand estimator term estimate std.error statistic p.value
#> 1 ATE 0.2 estimator Z 0.09450127 0.07783236 1.214164 0.2261312
#> conf.low conf.high df outcome
#> 1 -0.05898551 0.2479881 198 Y
# Set up population to assign
model <- declare_model(
villages = add_level(
N = 30,
N_households = sample(c(50:100), N, replace = TRUE)
),
households = add_level(
N = N_households,
N_members = sample(c(1, 2, 3, 4), N,
prob = c(0.2, 0.3, 0.25, 0.25), replace = TRUE)
),
individuals = add_level(
N = N_members,
age = sample(18:90, N, replace = TRUE),
gender = rbinom(n = N, size = 1, prob = .5)
)
)
# Assignment procedures
## Complete random assignment
design <-
model +
declare_assignment(Z = complete_ra(N = N, m = 1000))
head(draw_data(design))
#> villages N_households households N_members individuals age gender Z
#> 1 01 83 0001 2 0001 66 0 0
#> 2 01 83 0001 2 0002 47 1 0
#> 3 01 83 0002 2 0003 36 1 1
#> 4 01 83 0002 2 0004 72 1 0
#> 5 01 83 0003 1 0005 64 1 0
#> 6 01 83 0004 1 0006 73 1 0
## Cluster random assignment
design <-
model +
declare_assignment(Z = cluster_ra(clusters = villages,
n = 15))
head(draw_data(design))
#> villages N_households households N_members individuals age gender Z
#> 1 01 76 0001 1 0001 74 0 T7
#> 2 01 76 0002 4 0002 33 0 T7
#> 3 01 76 0002 4 0003 52 0 T7
#> 4 01 76 0002 4 0004 72 0 T7
#> 5 01 76 0002 4 0005 72 0 T7
#> 6 01 76 0003 4 0006 49 0 T7
## Block and cluster random assignment
design <-
model +
declare_assignment(Z = block_and_cluster_ra(
blocks = villages,
clusters = households,
block_m = rep(20, 30)
))
head(draw_data(design))
#> villages N_households households N_members individuals age gender Z
#> 1 01 90 0001 4 0001 88 0 0
#> 2 01 90 0001 4 0002 61 0 0
#> 3 01 90 0001 4 0003 77 0 0
#> 4 01 90 0001 4 0004 26 0 0
#> 5 01 90 0002 4 0005 50 1 0
#> 6 01 90 0002 4 0006 56 0 0
## Block random assignment
design <-
model +
declare_assignment(Z = block_ra(blocks = gender, m = 100))
head(draw_data(design))
#> villages N_households households N_members individuals age gender Z
#> 1 01 54 0001 3 0001 67 0 0
#> 2 01 54 0001 3 0002 25 1 0
#> 3 01 54 0001 3 0003 74 1 0
#> 4 01 54 0002 1 0004 35 0 0
#> 5 01 54 0003 2 0005 55 1 0
#> 6 01 54 0003 2 0006 34 0 0
## Block random assignment using probabilities
design <-
model +
declare_assignment(Z = block_ra(blocks = gender,
block_prob = c(1 / 3, 2 / 3)))
head(draw_data(design))
#> villages N_households households N_members individuals age gender Z
#> 1 01 73 0001 2 0001 40 0 0
#> 2 01 73 0001 2 0002 37 1 0
#> 3 01 73 0002 3 0003 28 0 1
#> 4 01 73 0002 3 0004 33 0 1
#> 5 01 73 0002 3 0005 63 0 0
#> 6 01 73 0003 4 0006 30 0 1
## Factorial assignment
design <-
model +
declare_assignment(Z1 = complete_ra(N = N, m = 100),
Z2 = block_ra(blocks = Z1))
head(draw_data(design))
#> villages N_households households N_members individuals age gender Z1 Z2
#> 1 01 56 0001 2 0001 55 1 0 1
#> 2 01 56 0001 2 0002 24 1 0 0
#> 3 01 56 0002 2 0003 83 1 0 0
#> 4 01 56 0002 2 0004 36 1 0 1
#> 5 01 56 0003 1 0005 43 1 0 0
#> 6 01 56 0004 4 0006 50 0 0 1
## Assignment using functions outside of randomizr
design <-
model +
declare_assignment(Z = rbinom(n = N, size = 1, prob = 0.35))
head(draw_data(design))
#> villages N_households households N_members individuals age gender Z
#> 1 01 73 0001 3 0001 72 0 0
#> 2 01 73 0001 3 0002 22 0 0
#> 3 01 73 0001 3 0003 45 1 0
#> 4 01 73 0002 4 0004 72 1 1
#> 5 01 73 0002 4 0005 42 1 0
#> 6 01 73 0002 4 0006 27 1 0