When a random effects meta analysis has been used, this function visualizes how the between-study standard deviation depends on the explanatory variable.
Arguments
- x
Object returned from
metagam
.
Examples
library("mgcv")
set.seed(1233)
shifts <- c(0, .5, 1, 0, -1)
datasets <- lapply(shifts, function(x) {
## Simulate data
dat <- gamSim(scale = .1, verbose = FALSE)
## Add a shift
dat$y <- dat$y + x * dat$x2^2
## Return data
dat
})
models <- lapply(datasets, function(dat){
b <- gam(y ~ s(x2, bs = "cr"), data = dat)
strip_rawdata(b)
})
meta_analysis <- metagam(models, method = "REML")
plot_between_study_sd(meta_analysis)