wanted to know only about internet users, you could put a survey out there on the
internet, and people came to you and chose themselves to be in the sample. Since
your population if interest is internet users we do not have undercoverege.
7.
The dean of a college of business with five departments wants to estimate the student
hours lost per month due to illness of professors and compare them. Describe a sampling
plan that would be suitable for estimating the collegewide loss and for comparing
departments.
Stratified sample; take a simple random sample from each of the 5 depts. If you
just take a simple random sample from the whole college, you may miss a
department that is small, for example, and they will have no representation.
Stratified sampling forces there to be representation from each group.
8.
A politician wants to estimate the mean age of registered voters in her district.
Unfortunately, she does not have a complete list of households. Describe a sampling
plan that would be suitable for her purposes.
Many different answers possible here. One idea would be look at the list of
precincts in the district (a precinct is a specific area within a city designed for
voting purposesif you live in precinct A, you all vote in the same place and you are
all on the same list.) From there you can get a list of everyone in each precinct that
is registered to vote, and take a random sample from that. If that’s not allowed you
could divide the district into regions and use a map to randomly select households
to sample from (more difficult!) You’d have to eliminate anyone not registered to
vote.
9.
You are on the staff of a member of Congress who has 5,000 constituents (voters in her district.)
She is considering a bill that would require all employers to provide health insurance for their
employees. She sends a survey to all 5,000 constituents asking them their opinion on this bill (support,
oppose, no opinion.) Of the 1,128 surveys received, 871 of them oppose the bill (which equals 77.2%.)
a. What is the response rate for this survey?
1128/5000 = 0.2256=22.56%
2
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Documentb. Looking at these results you are not convinced that a majority ( > 50%) of all the voters in
her district necessarily oppose this bill. State briefly how you would explain this to the
congresswoman. Use an argument based on statistical ideas.
There is a very high rate of nonresponse to this survey. You don’t know how those
people would have responded to the question. You cannot assume that the
responses from the nonrespondents would be the same as for the respondents. It
could
be that they didn’t care; so if you did have the data from all 5,000 people
sampled, you wouldn’t have 77.2% opposing the bill; it would likely be lower in
that case.
10. An online dating service has a quick poll on its website showing the following results. Which of the
This is the end of the preview.
Sign up
to
access the rest of the document.
 Fall '11
 Johnson
 Statistics, researcher

Click to edit the document details