Sometimes it is cheaper to 'cluster' the sample in some way e.g. by selecting respondents from certain areas only, or certain time-periods only. (Nearly all samples are in some sense 'clustered' in time – although this is rarely taken into account in the analysis.)
oopt = ani.options(ani.height = 350, ani.width = 500, outdir = getwd(), nmax = 30, interval = 1, title = "Demonstration of the cluster sampling", description = "Once a cluster is sampled, all its elements will be chosen.") ani.start() par(mar = rep(1, 4)) sample.cluster(col = c("bisque", "white")) ani.stop() ani.options(oopt)