Egg-Faced News Divisions Stick With VNS Despite Election Fiasco
The networks and Associated Press decide to stick with Voter News Service, cited as the culprit for last November's election night screw-ups. None would pay for an alternative.
by Warren Cohen
Thursday, May 31, 2001

After the networks' election-night debacle of incorrect victory calls last November, there was an unprecedented period of self-flagellation.

All the networks commissioned internal studies about what went wrong. So did Voter News Service, the maligned polling and vote tabulation consortium funded by ABC, NBC, CBS, CNN, Fox and the Associated Press. Fox and NBC went the furthest, threatening last November not to renew their contracts with VNS unless there were reforms that would 'ensure the accuracy and integrity of the data,' according to NBC.

So much for the tough talk. According to network news executives, all of the networks are expected to remain in the VNS consortium, with an announcement expected as soon as Friday. Starting a competiting or alternative polling operation was simply too expensive to consider, several of them said.

Although many network executives accepted some measure of blame for the mishandled election results, citing competitive pressure to call the new President before rivals, much of the focus on what went wrong fell on VNS. Founded in 1990, the consortium was designed to save the networks money on the costly task of exit polling and vote gathering across the nation. But in their reports, the networks said that VNS had outmoded equipment, inaccurate statistical processes and not enough quality control to ensure an accurate count in a razor-thin race.

Most network executives say that the silver bullet solution for these problems is more money. VNS had an operating budget of $33 million for the entire election cycle split equally among the members, according to a CNN post-mortem election report. (An additional $2 million was provided to the service by other subscribers, such as local television stations.) While all of networks have publicly committed to contributing more to VNS, they have not yet decided on dollar figures. 'We're looking at the budgets and what can be done, but there are no final decisions,' said a network news source.

So far, the members have only agreed to change the staffing on the VNS board of directors. Prior to 2000, it was composed of representatives from the election units of each network and focused on day-to-day concerns. Now, network vice-presidents will sit on the board and set broad-based policies.

On the list of fixes is replacing the old VNS mainframe computer system and software with more flexible and faster contemporary equipment. Statistical improvements will include more absentee voting in the model. According to Warren Mitofsky, a polling consultant for CNN and CBS, the VNS computations were based on models devised between 1967 and 1990, which didn't account for the same levels of absentee voting. Those votes used to comprise only about 3 percent of the national tally, but have increased due to more liberal voting laws. For example, 12 percent of Florida voters submitted absentee ballots in the last election.

Other enhancements will include mechanisms designed to catch data entry blunders. In the last election, the VNS was slow to correct a 20,000 vote error in Florida's Volusia County that overstated George W. Bush's lead over Al Gore. Instead of having a race too close to call, the networks saw the inflated total and projected the state of Florida for Bush around 2:00 a.m. on election night. At 2:51 a.m., VNS corrected the error. 'It's hard to process millions of votes in precincts and towns throughout the country without an error,' said Mitofsky. 'A more sophisticated mechanism will help catch obvious data errors.'

Even though everyone agrees on the need to make these fixes, the big question is how much money the networks will actually spend. After all, VNS was founded as a cost-cutting measure. In its own internal investigation, which was obtained by the Washington Post, VNS said that budget limitations, such as not having enough to spend on telephone polling to estimate the size of the absentee vote, 'made the task of covering elections far more difficult than necessary.' The CBS report pointed to the difficulties of increasing the VNS budget. Because the costs are borne equally and the budget must pass unanimously, CBS wrote, 'the most frugal members sets the limit on how much VNS can spend. And when its budget is limited, VNS must limit what it can do.'

Right after the election, a few of the networks raised the possibility of forming an alternative polling organization to either compete or work in tandem with VNS. That would have provided another data source so all the networks wouldn't rely on the same set of numbers and could avoid group catastrophic mistakes. CNN's report said, 'It is a general convention of good journalistic practice to seek more than one source for information to be reported, when that information is controversial and open to different interpretations. We believe there should be at least one competitive source available for both exit polling and vote counting.'

That step would have been supported by a variety of public interest groups crying for post-election reforms. Some raised the possibility that the VNS agreement between the networks may violate antitrust laws. The decision to keep VNS intact has upset some critics. 'I'm not all that surprised, given that in the end it's all about money, but I'm still disappointed,' said Gary Hill, co-chair of the Society for Professional Journalists ethics committee. Hill had pushed for an alternative-polling consortium to VNS. While he admitted that funding another organization would be expensive, he suspected that the cost would be worth in it lieu of the Florida fiasco. 'What did it cost the networks in credibility?' he asked. 'If they could buy it back with a like sum of money, wouldn't they?'

But any alternative to VNS was impractical with network news budgets being cut along with the huge expense of either polling alone or forming a VNS competitor. And privately, some network executives argue that the last election was a freak occurrence. VNS representatives have publicly stated that the organization has only a 1-in-200 chance of error and until last November, the results reflected it. 'In the ten years since the pool began, there had been only been one error in calling an election,' says Mitofsky. 'VNS has done a pretty good job with elections.'