Answer: C) Survey a random sample of persons within each neighborhood of the city.
==============================================================
Explanation:
I'll go through each answer choice to say why it is either the answer or not the answer.
A. This is not the answer because teachers are over-represented while non-teachers are under-represented. The teachers will most likely respond favorably to the survey, skewing the results. The other townspeople's opinions would be left out.
B. This is not the answer because people living far away from the city in question should not have sway over what happens in the particular city. It makes no sense to have someone in San Francisco determine what happens in LA schools. It's better to ask people who live in the city. Perhaps you could extend the net to nearby surrounding cities assuming those parents commute into the city to drop off their kids at those schools.
C. This is the answer. It's best to ask city residents because it is a local city issue. I would put every resident's name or address into a computer database, then have a random number generator select those individuals.
D. This is not the answer. People who do not visit city hall will not be represented at all, which skews the results of what the people of the city want overall.