Social Sciences · Political Science and International Relations
INSIDE THE WITHIN PRECINCT ERROR
See on the mapAbstract
Like any measurement, exit polls have both systematic and random error. The distinction is important for the reliability of the statistical models used to project the results of elections. We estimate both types of error for the 1996 US presidential election. Although the exit polls in 1996 were fairly accurate, individual precincts had substantial systematic errors, an indication that the overall accuracy might not be dependable for other elections. We also investigate interviewer and precinct characteristics that affect precision.