STAT*3320: Sampling Theory with Applications
1. Introduction: Big Picture
Method of collecting information for the purpose of describing a
population, or comparing characteristics (analyzing data) between two or
more (sub-) populations,
5 Estimating a Ratio (Chapter 6 of text)
5.1 Why would we be interested in a ratio?
(1) Sometimes we are more interested in estimating a ratio R instead of a
mean, total or proportion.
Eg. The average number of hours of TV watched per week by c
3. Probability Sampling
Goal: estimate population parameters (total, mean, variance, proportion)
Q. How accurate of an estimate do we want?
= parameter of interest
B = bound on error of estimation
Error of estimation = | | < B
Pr | | B
ST3320 Ali F15
Assignment # 1
Due: Friday, October 2, 2053 in class
1. To estimate how many books in the library need rebinding, a librarian uses a
random number table to randomly select 100 locations on library shelves. He then
walks to each
4 Estimating Sample Size
At some stage of designing a sample survey, the size of the sample must be
Too large a sample
Too small a sample
Sampling theory provides a framework to think through this problem.
1. What is the population size
3.8 Inverse Weighting
Recall that in our discussion of totals in Section 3.3, we noted that when sampling
with replacement, we can re-write the total estimator as:
and this estimator is unbiased.
However, when sampling without replace