ConversationTerm paper of the urban planning class: Narrator: Listen to a conversation between a student and a professor. Professor: OK, Karl. How could I help you? Student: It's about my term paper of the urban planning class. Instead of doing library research, I was thinking of doing a survey of my own and using the data in the report. Professor: That's an nice idea. I think doing this kind of hands-on project is great experience, especially if you decided to go to graduate school. Student: But I have been having a lot of problems designing a survey. Professor: Ah, it's not as easy as it looks, is it? Student: No. Now I am not even sure if I can stick with my plan or switch back to a more traditional paper. Professor: Well, you might as well get some experience sooner well than later. What's the topic again? Student: Nothing really complex. I just wanted to survey local people about the proposal for widening route 62 from a two-lane road to a four-lane highway. Professor: OK. Student: So do I go about selecting the people I want to survey? A lot of what I read have stress having a representative sample. But I'm not sure how to do it. Professor: Well, yes. A representative sample is crucial. You want the people you survey to be as close to an accurate reflection of the larger population you are interested in as possible. Imagine that the town as a cake with raisins in it and imagine that for some reason, most of the raisins end up on one side of the cake. If you cut a piece from the other side, would that be truly representative? Would that be an accurate reflection of the cake? Student: No. You'll be missing the raisins. Professor: Exactly! So for your project, what if you only survey people who live in houses along route 62, you think that would be useful? Student: Well, no. Because everyone there would probably be opposed to the widening. Professor: Yes. And that's why a representative sample is important. You don't want a sample that's biased, too strongly one way or the other. So, how do you do that? Student: Well, I could make sure my survey reaches people in all parts of the town. Professor: Good. So how can you do that? Student: Well, I'll go and make phone calls, just choose numbers at random from the local phone director. Instead of passing on questionnares, because people often don't ?? But I don't see how I can get a representative sample using the phone. Professor: Well. Make sure you're getting a nice balance of numbers. Phone company can tell you what numbers have been assigned to different parts of town. For example, for numbers in my neighborhood, that was a 256 and I live on the edge of town, where I guess you'd find a lot of pro-highway people. So make sure you don't call mostly 256 numbers.