How do I solve this?

You want to estimate what percentage of people in your town own a BMW. You call 100 people at random and ask them whether they own a BMW. Even assuming that everyone answers truthfully, why will you still not know accurately what percentage of people own BMWs?

A. A random sample of 100 may very well ﬁnd no BMW owners at all.

B. The results might accidently find several BMW owners, when in reality only a tiny fraction of people own BMWs.

C. Although people may not lie, the answers could vary unexpectedly. For example, for those people leasing their BMW, some may answer “yes” and some may answer “no.”

D. All of the above.

Question

How do I solve this?

You want to estimate what percentage of people in your town own a BMW. You call 100 people at random and ask them whether they own a BMW. Even assuming that everyone answers truthfully, why will you still not know accurately what percentage of people own BMWs?

A. A random sample of 100 may very well ﬁnd no BMW owners at all.
		
B. The results might accidently find several BMW owners, when in reality only a tiny fraction of people own BMWs.
		
C. Although people may not lie, the answers could vary unexpectedly. For example, for those people leasing their BMW, some may answer “yes” and some may answer “no.”
	
D. All of the above.

agent0smith · Answer

I think all of them can have an effect. As the sample gets larger, A and B become less of an issue.