In order to best estimate the number of students who bring their lunch to school every day, the sample should be representative of the entire student population at the school. Among the four options provided, the best sample would be:
80 students selected at random during lunch.
This option is likely to give a diverse representation of the student body, as it includes students from different grades, genders, and backgrounds, and is directly focused on the behavior of bringing lunch, as you'll be capturing students during a time when lunch decisions are relevant.
The other options are limited in scope and may not adequately represent the overall population, such as focusing only on members of a specific club or on a specific gender or year.