Skip to contents

Generate Normal Variables

Source: R/sim-generators.R
gen_normal.Rd

Create a generator specification for univariate Gaussian variables.

Usage

gen_normal(
  vars,
  level = c("single", "level2", "multilevel"),
  mean = 0,
  sd = NULL,
  fixed_intercept = NULL,
  random_cov = NULL,
  residual_var = NULL,
  scale_fixed_intercept = NULL
)

Arguments

vars

Character scalar naming the generated variable.

level

Simulation level. "single" generates one row-level value per observation, "level2" generates one group-level value and expands it to each row in the group, and "multilevel" generates row-level values from a random-intercept model.

mean, sd

Direct mean and standard deviation for "single" and "level2" generators. Values are recycled to the number of generated units.

fixed_intercept

Location intercept for multilevel generators. For normal variables this is on the response scale.

random_cov

Group-level random-intercept covariance. A scalar is accepted for univariate location-only generators; a matrix is required when location and scale random effects are modeled jointly.

residual_var

Residual variance for multilevel normal generators when no scale model is used.

scale_fixed_intercept

Optional log residual standard-deviation intercept for multilevel generators. When supplied, residual standard deviations are exp(scale_fixed_intercept + group_random_effect).

Value

An mlsim_generator_spec object for use in simulate_data().

Details

For non-multilevel normal variables, mean and sd define independent draws directly. For multilevel variables, fixed_intercept, random_cov, and either residual_var or scale_fixed_intercept define a random-intercept data-generating model.

See also

Other predictor generators: continuous-generators, count-generators, gen_categorical(), gen_custom(), gen_mvn()

Examples

sim <- simulate_data(
  n_groups = 2,
  n_per_group = 3,
  seed = 1,
  generators = list(
    x = gen_normal("x", mean = 0, sd = 1),
    u = gen_normal(
      "u",
      level = "multilevel",
      fixed_intercept = 0,
      random_cov = 0.2,
      residual_var = 1
    )
  )
)
sim$data
#>    group_id obs_id          x           u
#>       <int>  <int>      <num>       <num>
#> 1:        1      1 -0.6264538  0.79376625
#> 2:        1      2  0.1836433 -0.08740349
#> 3:        1      3 -0.8356286  1.72976607
#> 4:        2      1  1.5952808  0.72003208
#> 5:        2      2  0.3295078 -0.29105173
#> 6:        2      3 -0.8204684 -1.88451104