mixed_model: Generate simulated mixed model data
In dustinfife/bluepill:

Description Usage Arguments Examples

View source: R/mixed_model.R

Generate simulated mixed model data

1	mixed_model(fixed, random, sigma, clusters, n_per, vars)

`fixed`	a vector equal to the number of variables + 1. Each entry indicates the (standardized) value of the fixed effect. The first entry is for the intercept.
`random`	a vector equal to the number of variables + 1. Each entry indicates the (standardized) standard deviation for the random effects. The first entry is for the intercept.
`sigma`	proportion of variance remaining unexplained at the residual level
`clusters`	number of clusters
`n_per`	a vector of length two, indicating the mean and standard deviation of the number of observations within each cluster
`vars`	a named list. The names correspond to the variable names (including the cluster names). Each entry either contains a vector of three (indicating mean, standard deviation, and # of digits to round), or a vector containing the unique values of categorical variables. Also, you must include the name of the cluster. See examples.

# simulate data where depression = stress + life_events + parental_depression + ses
fixed = c(0, .2, .5, .3, .2)
random = c(.1, .1, 0, .2, .1)
vars = list(
 depression = c(10, 3, 0),
 stress = c(22, 7, 0),
 life_events = c("no", "yes"),
 parental_depression = c("no", "mild", "moderate", "severe"),
 ses = c(55, 15, 0),
 therapist = paste0("Dr. ", LETTERS[1:15])
)
mixed_model(fixed, random, sigma = .3, clusters=15, n_per = c(11, 3), vars=vars)