MAT 107 Lecture Notes - Lecture 3: Bradley Effect
Document Summary
When we wish to study a certain characteristic of a large population, it is not always practical to try to gather info from every member of that population (a census) It is often necessary and/or convenient to select a sample of the population to study instead. Bias -- the tendency for samples to differ from the corresponding population in some systematic way. Selection bias -- excludes some part of the population. Telephone poll (not everyone has a telephone, esp. those unemployed) Measurement or response bias -- the method of observation produced values that differ from the true value in a systematic way. The bradley effect: the appearance or behavior of the person asking the question. Dishonesty regarding illegal behavior or unpopular opinions. Nonresponse bias -- responses are not gathered from all members of the sample. Calling during 9-5 (work hours, people won"t answer their phone)