Title
Sampling and Nonsampling Errors in Exit Polling
Speaker
Fritz Scheuren, National Opinion Research Center (NORC), University of Chicago
Abstract
Polling typically has a much higher nonresponse rate than other surveys, yet is given great weight by the media as it often seems to provide plausible predictions. What is the justification for this? Are there reasons to expect robustness? In particular, what can be said about the ratio of sampling to nonsampling error.
What can be said, too, about the seemingly careless treatment of the margin of survey error in polling, notably when assessing whether one candidate has a "statistically significant" lead. Are the reporters choosing to ignore a properly calculated larger margin of error? Maybe just to have a headline?
Finally, how can the election poll results be used as a teaching tool to increase the quantitative literacy of students and illustrate the value of our profession to students who are not statistics majors.
Meet the speaker in Room 212 Cockins Hall at 4:30 p.m. Refreshments will be served.
