Cal Trans wants to use a survey to determine the extent of public approval for a project to construct a new freeway through a particular neighborhood. Which of the following procedures would be most appropriate for obtaining a statistically unbiased sample?

A. Go door to door in the affected neighborhood, soliciting opinions.

B. Hold a public forum in the affected neighborhood, and distribute questionnaires there.

C. Place advertisements in newspapers in the affected neighborhood, asking readers to mail in their opinions.

D. Survey a selection of people whose names are randomly chosen from the city telephone directory.