The Science of Polling (broadcast Friday, February 1st, 2008)

Voting in community election. Greenbelt, Maryland, 1937. Photo by Arthur Rothstein. Library of Congress image.

Would you say that polls are very reliable, somewhat reliable, somewhat unreliable, or very unreliable?

Sample sizes, the way those samples are selected, and the precise wording of questions can all influence the outcome of a poll. In this segment, Ira talks with polling experts about what makes a reliable poll and how polls should be conducted. Why is it that sometimes election exit polls seem to be right on the money, while other times election results can surprise even expert analysts? Teachers, find more information about using Science Friday as a classroom resource in the Kids' Connection.

Guests

Andrew Kohut
President, Pew Research Center
Director, Pew Research Center for the People and the Press
Director, Pew Global Attitudes Project
Washington, DC

Karlyn Bowman
Senior Fellow
American Enterprise Institute for Public Policy Research
Washington, DC

Charles Franklin
Professor, Political Science
University of Wisconsin, Madison
Co-Developer, www.pollster.com
Madison, Wisconsin

Related Links

• Polling Report
• Roper Center: Polling 101
• American Association for Public Opinion Research: Polling FAQs
• Pollster.com FAQs

Segment produced by:Annette Heist

Elsewhere on Sciencefriday.com

Explore More

• Math
• Science And Society
• Science Policy and Funding