Weighted selection from a single roster


We are doing a very basic roster where we are interested in adults over 18 and selecting one for further interview. The challenge is that we want to increase the probability that individuals in a specific age range (say 18-25) are selected. Instead of a simple random sample of eligible members, we’re hoping to increase the probabilities for the 18-25-year-olds.

For example, a house with an 18-year-old and a 40-year-old there should be something like a 60/40 chance for the 18-year-old. In a house with 18, 20, and 40 there should be something like a 40/40/20 chance.

Is there some way to add some kind weight to specific selection criteria for finding the index randomly given that we don’t know how many eligible people that do and don’t fit the ideal category?