When the study variable y in a population of N units is
more or less proportional to a size variable x, one may use
the ratio estimator from a simple random sample of size n
without replacement (SRS). An alternative estimator in such
a situation is the Horvitz-Thompson (HT) estimator in
combination with a systematic probability proportional to
size sample from a randomly ordered list, henceforth called
a randomized PPS sample