This preview shows page 1. Sign up to view the full content.
Unformatted text preview: eing selected. Most of our inference methods require the data be considered a … RANDOM SAMPLE. This implies that the responses are to be independent and identically distributed (iid). We will make this more formal later after probability, but here are the basic ideas between these two properties. Independent = the response you will obtain from one individual will not influence the response you will get from another individual. Identically distributed = all of the responses have the same model (use time plots). Many sampling designs are discussed in your text (SRS, stratified, cluster, etc). We will not cover the details of these various methods, nor work with a random number table. However, we will expect you to think about whether the data available can be considered a random sample, based on the fundamental rule for using data for inference from Chapter 5. We will also discuss various graphs that sometimes can be used for checking assumptions, one of which is a time plot for assessing the identically distributed property of a random sample (if the data are collected over time). 5.5 Difficulties and Disasters in Sampling This section presents some of the problems that can arise even when a sampling plan has been well designed. It talks about sampling from the wrong population, relying on volunteer response, and meaningless polls. Read over these few pages – especially Case Study 5.1 on page 166. 5.6 How to Ask Survey Questions The wording and presentation of questions can significantly influence the results of a survey. This section presents seven pitfalls that are possible sources of response bias in a survey. Good examples! Asking the Uninformed (page 168) People do not like to admit that they don’t know what you are...
View Full Document
This document was uploaded on 02/25/2014 for the course STATS 250 at University of Michigan.
- Summer '10