A synthetic dataset mimicking the NSECE 2019 survey structure for demonstration of the DER diagnostic pipeline. Contains N = 6785 observations across J = 51 states with unequal survey weights, clustered PSU structure, and three fixed-effect covariates (intercept, within-cluster poverty, between-cluster tiered reimbursement policy).
Format
A list with components:
- draws
Matrix of posterior draws (4000 x 54), columns 1:3 are fixed effects (beta), columns 4:54 are random effects (theta).
- y
Binary outcome vector (length 6785).
- X
Design matrix (6785 x 3) with columns: intercept, poverty_cwc (group-mean centered), tiered_reim (binary policy).
- group
Integer state group indicator (1 to 51).
- weights
Survey weights (positive, length 6785). Log-normal distributed, normalized within state.
- psu
PSU indicators (integer, length 6785).
- param_types
Character vector of length 3:
c("fe_between", "fe_within", "fe_between").- family
Model family:
"binomial".- sigma_theta
Random effect SD (0.66).
- N
Number of observations (6785).
- J
Number of groups (51).
- p
Number of fixed effects (3).
Examples
data(nsece_demo)
str(nsece_demo, max.level = 1)
#> List of 12
#> $ draws : num [1:4000, 1:54] 0.3321 0.0921 0.1227 -0.0035 0.2898 ...
#> ..- attr(*, "dimnames")=List of 2
#> $ y : int [1:6785] 0 0 0 0 0 1 0 1 0 0 ...
#> $ X : num [1:6785, 1:3] 1 1 1 1 1 1 1 1 1 1 ...
#> ..- attr(*, "dimnames")=List of 2
#> $ group : int [1:6785] 1 1 1 1 1 1 1 1 1 1 ...
#> $ weights : num [1:6785] 0.327 6.848 0.269 0.272 0.937 ...
#> $ psu : num [1:6785] 1 1 1 1 1 1 1 1 1 1 ...
#> $ param_types: chr [1:3] "fe_between" "fe_within" "fe_between"
#> $ family : chr "binomial"
#> $ sigma_theta: num 0.66
#> $ N : num 6785
#> $ J : int 51
#> $ p : int 3