No because even though you aren't actually picking the exact people yo do and don't wanna survey but your only surveying those who buy hot lunches when in order to get accurate data, you have to survey everyone who bought hot and cold lunches from the school cafeteria.