inst/doc/fromscratch.R

In simr: Power Analysis for Generalised Linear Mixed Models by Simulation

## ---- message=FALSE, warning=FALSE--------------------------------------------
library(simr)

## ----options, echo=FALSE, message=FALSE---------------------------------------
simrOptions(nsim=100, progress=FALSE)

## -----------------------------------------------------------------------------
x <- 1:10
g <- letters[1:3]

X <- expand.grid(x=x, g=g)

## -----------------------------------------------------------------------------
b <- c(2, -0.1) # fixed intercept and slope
V1 <- 0.5 # random intercept variance
V2 <- matrix(c(0.5,0.05,0.05,0.1), 2) # random intercept and slope variance-covariance matrix
s <- 1 # residual standard deviation

## -----------------------------------------------------------------------------
model1 <- makeLmer(y ~ x + (1|g), fixef=b, VarCorr=V1, sigma=s, data=X)
print(model1)
model2 <- makeGlmer(z ~ x + (x|g), family="poisson", fixef=b, VarCorr=V2, data=X)
print(model2)

## -----------------------------------------------------------------------------
powerSim(model1, nsim=20)
powerSim(model2, nsim=20)