Four years ago, voters went to the polls to cast their votes for the presidential election and later sat glued to their televisions to follow the endless hours of poll coverage in hopes of learning which candidate won the election.
In order for the news programs to report a winner, polling information must be collected from each state to determine which candidate won the majority of votes. \There's a consortium that's formed by all the different media organizations that purchase exit polling data from a particular group of opinion researchers who have, essentially, the ability to predict,"" said UW-Madison political science Professor Dhaven Shah.
Although the polling process was the same as in previous years, election night of 2000 proved to be an embarrassment for the broadcast news media. The country turned on their televisions expecting to get clear results from the news channels they trusted but instead got unclear and sometimes false poll results.
Shah attributes the problems experienced on election night to the fact that the media acted too hastily in an attempt to be the first to break the news. ""I think they called it when it was too close. They felt the pressure to call it and they did,"" he said. ""I think they fulfilled the assumption that people had about how the news was supposed to act that night.""
Behind the scenes, polling experts, including UW-Madison political science Professors Ken Goldstein and Charles Franklin, worked vigorously to collect and analyze each state's voting data, putting the poll results through numerous statistical assessments and, upon achieving statistical confidence in the numbers, reported the results to their affiliated news headquarters.
Continuing their job this year, both professors will be in New York tonight working for ABC News headquarters where several important changes have been made to ensure more accurate poll reporting.
""The system is completely new, completely rewritten, using new software, new hardware, a new kind of presentation which is, we certainly all hope, going to take care of some of the technical shortcomings that were problematic in 2000,"" Franklin said. ""The trouble was just that Florida was one error in a very very visible election.""
Even with new technology, calling an election will never be a flawless process. Franklin continues, ""no statistical system can call a race that's separated by 500 votes like Florida ... I guess the main thing I would stress is that we really do have a good model for this stuff and it doesn't mean it's infallible and it sure doesn't mean that it's something that anybody takes lightly, but I think, I assume, that all of the networks are going to show more caution, simply because it's so obvious what the stakes are.""
