nsamp = 20     ## The number of data values in each data set
nrep = 1000    ## The number of data sets to generate
pop_mean = 0   ## The population mean of each data value
pop_var = 1    ## The population variance of each data value

## Generate a nrep x nsamp array of standard normal draws.
D = rnorm(nrep*nsamp, mean=pop_mean, sd=sqrt(pop_var))
X = array(D, c(nrep,nsamp))

## Get the variance of each row of X.
Y = apply(X, 1, var)

## Calculate the sample mean of Y.
V = mean(Y)