The best sample for estimating the number of students who bring their lunch to school every day would be a random sampling of students from the school population. This would ensure that a representative sample of students is included in the study and provide a more accurate estimation of the number of students who bring their lunch to school each day.