I have two populations. One has approximately 500 people and the other has around 50,000. For research purposes I need to create a dataset out of the 50,000 cases in the large distribution that mirrors the smaller distribution on variable X, and then I can compare those distributions on variables Y and Z. Does anyone have an idea how I might be able to use proc surveyselect to do that, or should I attempt something else. Thank you so much for your assistance!!!