Chapter 12 Sample Surveys Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 1 Sampling Outline 1 Understand a Population i e all the units of interest 2 Sampling Biases 3 Randomized Samples Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 2 Populations The concept of a population is key to this entire course Having a well defined population helps researchers select better samples When doing statistics we typically want to understand certain aspects of a population Examples Cholesterol level of Male Native Americans Income of Banking Executives GPA of UT students that own laptops Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 3 The Population We want to learn details about the population We summarize this by saying that we want to learn about a population parameter A population parameter can be a Chapter12 Presentation 1213 Mean Variance Percentile Percent or some other function of the population distribution Copyright 2009 Pearson Education Inc 4 The Sample To learn about the population parameters we usually take a sample Given the data from a sample we can compute sample statistics Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 5 Notation We typically use Greek letters to denote parameters and Latin letters to denote statistics Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 6 How Large of a Sample How large a random sample do we need for the sample to be reasonably representative of the population It s the size of the sample not the size of the population that makes the difference in sampling Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 7 How Large of a Sample Cont To estimate the proportion of the population in Knox County that approves of Obama s health care plan with a margin of error of 3 we need a sample slightly bigger than 1000 To estimate the proportion of the population in the U S that approves of Obama s health care plan with a margin of error of 3 we need a sample slightly bigger than 1000 The fraction of the population that you ve sampled doesn t matter It s the sample size itself and how the sample is selected that s important Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 8 Collecting the Sample Sampling Frame The sampling frame is a list or collection of individuals from which the sample is drawn When the list or collection of individuals misses part of the population we say that the sampling frame is biased Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 9 Collecting the Sample We want to collect the sample that gives us an unbiased picture of the population This chapter describes sampling strategies that avoid bias and produce accurate results Good samples are called Representative First let s review what can go wrong if we don t have a representative sample Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 10 Voluntary Response Sampling A large group is invited to respond All those that respond are counted Voluntary response bias the sample only includes people with a strong interest in the topic Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 11 Ex Voluntary Response Bias Statistics 201 samples students opinions on a range of issues and uses the data for a class project If students aren t required or strongly encouraged to respond what effect might it have on the responses Who will respond How might their opinions or behaviors differ from other students in the class Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 12 Convenience Sampling Suppose that we only collect samples from population members that are convenient Ex A recent internet survey found that of the population had internet access Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 13 Sample Biases Undercoverage some portion of the population is not sampled or has smaller representation than in the original population Example US Census forms are sent to all residences in America What segment of the US population is undercovered Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 14 Sample Biases cont Nonresponse those that don t respond to the sample differ in their opinions from those that do Example Telephone surveys Have you ever hung up on a person doing a phone survey Are your opinions and others that hang up different from those that don t Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 15 Sample Biases cont Response Bias question wording interviewer behavior can affect answers Example Are you a racist will get you a no response almost 100 of the time Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 16 Sampling Bias Swine Flu The recent outbreak of Swine Flu in Mexico terrified the world because of the seemingly high death rate CNN regularly reported on the proportion of hospitalized patients that died Later research showed a much lower mortality rate Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 17 Sampling Bias Swine Flu Population Parameter of Interest Sampling Frame Sample Sampling Method Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 18 Sampling Designs We will explore each of these designs Simple Random Sample Systematic Sample Stratified Random Sample Cluster Sample Multistage Sample Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 19 Simple Random Sample Elements of the population are enumerated like a lottery where each element of the population gets a ticket with a unique number on it Each member of the population has the same probability of being selected When to use it When you have easy access to the entire population e g the registrar taking a sample of UT students Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 20 Systematic Sample Each element of the population is enumerated and every k th element of the population is sampled beginning at a randomly determined starting point among the first k elements Chapter12 Presentation 1213 Copyright 2009 Pearson Education Inc 21 Stratified Sampling The population is first sliced into homogeneous groups called strata before the sample is selected The population members are thought to be fairly homogeneous within each strata but each strata is thought to be different from one another Simple random sampling is used within each stratum before the results are combined Chapter12 Presentation 1213 Copyright 2009 Pearson Education
View Full Document
Unlocking...