April 26, 2001
Vol. 20 No. 15

current issue
archive / search

    NORC creating nation’s best archive of votes made during presidential election in Florida

    By William Harms
    News Office

    County courthouses and other county offices throughout Florida became outposts of the National Opinion Research Center during the past three months, as teams of researchers looked over disputed ballots from the 2000 presidential election.

    The teams, who work largely in groups of three with a supervisor, compiled what will become the nation’s best archive of information on the Florida presidential election, one of the most disputed in the nation’s history.

    Rather than interviewing people, as is the usual method of gathering information for NORC surveys, the researchers, or coders, “interviewed” pieces of paper by recording descriptions of about 180,000 disputed ballots. They were the ballots on which more than one presidential candidate was selected (an overvote) or ones on which no presidential candidate was selected (an undervote).

    The clients are the nation’s leading news organizations, each of which will analyze the data and publish their own observations. Included in the group are The New York Times, The Wall Street Journal, The Washington Post, Tribune Publishing (which includes the Chicago Tribune, Los Angeles Times and a number of other newspapers), CNN, the Associated Press, the St. Petersburg Times and the Palm Beach Post.

    The analyses will look at a number of issues, such as the difficulty people had in voting, particularly in counties where punch cards, or votamatic ballots, were used.

    A variety of problems led to invalid individual votes. Some voters, for instance, did not insert their punch cards into the correct slot, but kept the ballot in hand and tried to line up the holes with the election choices listed in a ballot that opened in a booklet fashion. Some people did not insert the cards completely, and their votes were recorded in areas outside the specified holes. In some cases, the voters did not push the stylus all the way through the card, and the paper did not punch out completely, leaving either a hanging piece of paper (a chad) or only an impression made on the chad (a dimple).

    In order to describe what they saw on the disputed ballots, coders were given descriptions of what to look for and light boxes to view the ballots to see if light was showing through the dimples.

    As they went about their task, they could not touch the ballots, which were held for their examination by an election worker. Coders also were told not to talk among themselves about what they viewed. NORC imposed this quality control procedure to ensure impartiality of the ballot examination.

    Because three people looked at each ballot, NORC will be able to determine how much variability there is among different perceptions of what was on each ballot. Understanding this phenomenon is important to determine to what extent differing perspectives influence the outcome of elections in which recounts are done by hand.

    “We also had vision exams for each of the coders and made sure they were trained carefully before they began work,” said Kirk Wolter, Senior Vice President for Statistics and Methodology at NORC and Professor in Statistics.

    Although the votamatic ballots account for more than 80 percent of the votes cast in the state, Florida also uses optical scan ballots and datavote. NORC teams also carefully scrutinized these systems.

    In the optical scan ballots, voters are instructed to fill in circles, much as students fill in circles on standardized tests or college entrance examinations. Errors occur when people check the ballots rather than fill in the circles or when they fill in more than one circle for each election race.

    The datavote system consists of a lever on a machine that punches out votes after a person has lined up a ballot. That system can go astray if a voter has not correctly lined up the card.

    When NORC codes and assembles the information, it will then be provided to the news outlets for analysis. Their reports in turn will help people understand better how improvements might be made to the electoral system, not only in Florida, but also elsewhere in the country.