A school district is trying to estimate the proportion of students at the high school who would eat a school breakfast if the cafeteria were opened to students at 6 am. A new employer opened up within the district and that caused an abnormal spike in the size of the freshman class this year, so there are a much small number of sophomores, juniors, and seniors than there are freshmen. The school district would like to capture the effect of this lopsided class structure in their sample. What technique would be the best for this?