STAT 100 Chapter Notes - Chapter 1: Response Bias
Document Summary
Subjects are asked for a response: observational study. Subjects behavior is observed: controlled experiment. All subjects in the population are studied. Portion or group of the subjects from the population. Good sample must be representative of the total population. Size of sample, not population is important. ** a census is the whole and a sample is just a small part** 4 types of bias: voluntary response bias. Volunteers for samples tend to have strong opinions (primarily negative: undercoverage bias. Non representative of the group, a whole portion is not represented or they are not represented equally: nonresponse bias. Those who do respond differ from those who do not: response bias. Anything in the design of the survey that may influence responses. ** randomizing is the only way to reduce bias, bias cannot be eliminated completely, bias prevents us from drawing conclusions about populations** Posting an ad in a newspaper asking people to respond on a topic.