Type of sampling that uses the differences that already exist in a population as the basis for selecting a sample; knowing the percentage of the population that falls into a particular category, the researcher then randomly selects a number of persons to be studied from each category in the same proportion as exists in the population