http://washingtonpost.com/ac2/wp-dyn/A49376-2001May5?language=printer Florida Ballot Examination (Cont'd) Media Groups' Attempts to Answer 'What Ifs' Take Longer Than Expected By Peter Slevin Washington Post Staff Writer Sunday, May 6, 2001; Page A03 George W. Bush just celebrated his presidency's first 100 days and Al Gore is teaching journalism students at Columbia. But in the far corners of Florida, tens of thousands of ballots from the November election are still being examined by media groups that aim to describe what happened and answer questions of What if? What if the U.S. Supreme Court had not stepped in and abruptly halted the recounts that were underway on Dec. 9? What if canvassing boards had inspected ballots that registered no clear presidential vote on Election Day? What if Florida had better ballots and sharper technology? For months, news organizations have been dispatching teams with pens and coding sheets to Florida's 67 counties to document every last hanging chad and stray mark on roughly 180,000 ballots that were rejected by counting machines on Nov. 7. They have pressed election supervisors to sort countless ballots and they have struggled to reconcile numbers that often defy simple math. Results are expected within the next two months from the two most ambitious projects -- one led by the Miami Herald and the other a joint effort of The Washington Post and seven other news organizations. The goal is a comprehensive record that will help explain the confusion that contributed to 35 days of stalemate before Florida -- and the election -- could be called for Bush. The unprecedented effort has taken far longer than the media groups expected because of the huge logistical problems involved in winning access to the ballots, identifying the correct ones and then examining them. Each group has spent more than $500,000, in addition to untold hours planning, reporting and presenting the results. As the effort continues, reaction has followed party lines. Democratic activists are reiterating their "count every vote" mantra of last fall, hoping for evidence that Bush grabbed the state's decisive 25 electoral votes unfairly. Republicans determined to protect the legitimacy of Bush's electoral college victory have questioned the value of the projects and the validity of the varied outcomes. Florida Republicans have been particularly pointed, attending many recount sessions and firing darts at the news organizations involved in them. "Frustrated Media Gives Up Ballot Review, Begins Inventing Votes," said the headline on one of the Florida party's daily press releases challenging the media inspections. The release said news organizations intended to manufacture a hypothetical Gore victory and had begun "creating phantom votes" for the defeated former vice president. Leaders of the media groups insist they are not trying to decide who should have become president, merely who might have become president under a variety of circumstances and counting standards. Indeed, the group that includes The Post and the New York Times calls its hired ballot examiners coders, not counters. "We're not setting out to recount the election. The election's over," said Leonard Downie Jr., executive editor of The Post. "Who would have gotten how many votes depending on what standard you set for a ballot is only part of what we think will be many different kinds of stories. Our primary focus is on: Did the system work or not -- it certainly appears it did not -- and why not and what can be done to improve it?" The National Opinion Research Center at the University of Chicago is conducting the ballot review for the group that includes The Post, the New York Times, the Wall Street Journal, Tribune Newspapers (including the Los Angeles Times), the Palm Beach Post, the Associated Press, CNN and the St. Petersburg Times. Kirk Wolter, who is overseeing the project for NORC, said the coders are "merely trying to discern what is on the ballot in terms of marks or punches or partial punches and dimples and so on." Making sense of the varied media inspections requires patience and a deep appreciation for caveats. For one thing, no two projects have covered exactly the same ground. For another, each story about results includes a bushel of premises about the mechanics of the count and the analytical criteria. "There's not going to be one answer," said University of Florida political science professor Richard K. Scher. The media teams organizing the ballot studies acknowledge this point. The most determined examiner cannot know when a canvassing board in Wakulla or Glades or Miami-Dade County would have decided that a dimple meant a vote on a punch card or a faint check was intended as a vote on an optical scan ballot. But the journalists certainly have been trying, in projects large and small. "Clearly the election left a lot of people with a lot of questions, with the big one being who really won, in their minds. What we're trying to do is answer the questions," said Martin Baron, executive editor of the Miami Herald, which has already published a study of the state's undervotes -- ballots that were not counted in the presidential race because there was no discernible vote. • The Orlando Sentinel and two other newspapers studied 15,596 ballots in the 15 Florida counties with the highest rate of rejected votes. They found 1,700 ballots where they believed the voter's presidential preference was clear. When they totaled the votes, Gore gained 366 votes even though Bush had carried all but one of the counties on Election Day. One reason was the large numbers of voters in counties with optical scan ballots -- which function like answer sheets on standardized tests -- who voted for their candidate and then also wrote in his name. The Sentinel found that 962 people voted that way, with Gore supporters making the mistake 194 more times than Bush voters. The Orlando reporters also found 275 clear votes where voters made marks in the wrong place or used ink that the counting machines did not recognize. In Charlotte County, Bush lost 12 votes and Gore two votes because a machine registered a crease in absentee ballots as a second vote. Among the published reports so far: • The Palm Beach Post reported that Gore likely lost thousands of votes because of the notorious butterfly ballot in Palm Beach County, where more than 8,000 voters punched a hole next to Gore's name and another next to the candidate immediately above or below him. Reporters discovered that 2,908 voters punched holes for Gore and the candidate beneath him on the ballot, obscure Socialist Party nominee David McReynolds, who received 622 votes statewide. An additional 5,330 voters chose Gore and the candidate just above him, Patrick J. Buchanan, the former Republican firebrand who ran on the Reform Party ticket. A far smaller number voted for Bush and another candidate. • The Miami Herald, Knight Ridder and USA Today declared in April that Bush would have tripled his 537-vote margin of victory if the Supreme Court had not stopped the December recount and if a lenient ballot standard had been used. But if the count had gone forward and county canvassers had used a stricter standard for determining a voter's intent, the organizations said, Gore might have won by three votes out of 6.1 million cast. If the undervotes had been counted from scratch using the most lenient standard, the newspapers concluded that Gore would have won by 393 votes. The most lenient standard would have counted as a vote every punch-card ballot with a dimple, pinprick or hanging chad next to a candidate's name. The strictest standard would have counted only those ballots where the chad was punched clean. The Herald survey was based on an examination of 64,248 undervotes conducted by the BDO Seidman LLP accounting firm. The company is now inspecting the overvotes -- the ballots rejected when voters cast votes for more than one presidential candidate. Results of the full study are expected in "days." Baron said the results "may get at the broader issue of who may have been preferred by most people." An attempt to organize a single media study foundered in December over issues about control and methodology. The goals of the two competing teams are similar, although there are differences in their approach to the ballots. The Herald project, for example, assigned one accountant and one reporter to examine each ballot, with only the accountant's description included in the published results. The consortium, by contrast, arranged for the National Opinion Research Center to dispatch three staff members or temporary workers to study each undervote independently. The nonprofit social science research organization aims to produce a reliable record of ballots in which the observations of each team member can be studied in a public database. Leaders of both projects have discovered the impossibility of replicating Election Day conditions. In the published Herald study, the newspapers relied on Florida counties to produce the exact undervotes from November, but hundreds of ballots were missing, either because chads had fallen out or machines read the ballots differently when they were separated for the recount. In Palm Beach, the Herald reporters and BDO Seidman accountants were not shown undervotes from all precincts, according to the Palm Beach Post, which was simultaneously viewing the same ballots. In another case, they were shown portions of the same precinct twice without being told -- due to a bomb scare that caused a brief evacuation, the Palm Beach Post reported. "We only had eight counties that could come up with the same number of ballots as they did on election night," said Mark Seibel, managing editor of the Herald. "In the grand scheme of things, I'm not sure it would change the result, but it's troubling to be missing that many." An important purpose of the ballot reviews, according to executives of the news organizations that are paying for them, is to understand how so many voters could spoil their ballots unintentionally -- and which design or technology might improve voting in the future. The 15-county Orlando Sentinel examination found 781 people who voted in one column for everyone but Bush and 197 for everyone but Gore, invalidating their ballots. Similarly, on ballots where the list of 10 presidential candidates stretched into a second column, countless voters chose one candidate from the first column and another from the second column. "To me in the long run," Downie said, "it is more important to know why people were unable to get their votes to count -- which is crucial to our democracy -- than whether or not one of the other candidates would have picked up votes if votes had been counted." © 2001 The Washington Post Company ''' (0 0) ----oOO----(_)---------- | the geek shall | | inherit the earth | -----------------oOO---- |__|__| || || ooO Ooo ------------------------------------------------------------ FREE EMAIL from AUSI at http://ausi.com