You want to estimate the number of students in your middle school school who buy school lunch. which sample is unbiased?

A biased example: Asking students who are in line to buy lunch 

An unbiased example: Asking students who are leaving/going to lunch(NOT buying lunch).

But in this case, the answer choices can be... confusing.
Don't panic! You're given numbers and, of course, your use of logic.

Answer choice A: 100 students grades 6-8
Answer choice B: 20-30 students any one grade
Answer choice C: 5 students
Answer choice D: 50 students grade 8

An unbiased example would be to choose students from any grade. So we can eliminate choices B and D.

Now, the question wants to estimate how many people at your middle school buy lunch. This includes the whole entire school, and if you are going to be asking people, you aren't just going to assume that if 5 people out of 5 people you asked bought lunch, the whole school buys lunch.

So, to eliminate all bias and/or error by prediction, answer choice A, the most number of students, is your answer.