Studies Blame TV Networks for Election ‘Debacle’

WASHINGTON —  Two reports issued Friday on the election night “debacle”--in which television networks prematurely called the winner of the presidential election--blamed both the networks and their source of polling data for playing fast and loose with the figures.

A study commissioned by Cable News Network took all the networks to task for “recklessly endangering the electoral process” in their competition to be first with election results.

And a second report commissioned by Voter News Service, which provided the networks with their exit polling data, cited VNS for statistical errors and polling practices ill-suited for close elections. The study found no political bias behind the mistakes.

The two studies examined the factors that led to an error-riddled evening of news coverage. Most dramatically, the network’s early-hours declaration of Texas Gov. George W. Bush as the next president prompted Vice President Al Gore to place a concession call that he retracted just moments before he was to concede publicly.

What went wrong during the networks’ election night coverage will be the focus of congressional hearings scheduled in Washington for Feb. 14.

However, the independent report issued by the Research Triangle Institute, a North Carolina-based not-for-profit research firm, found that VNS’ performance on election night was largely positive. VNS is funded by a media consortium made up of Associated Press and five networks--ABC, CBS, CNN, Fox News and NBC. The purpose of the 36-year-old organization is to help its members project the outcome of elections.

Since election night, many of those same news organizations have placed much of the blame for the mistakes on the VNS data. The outside report was commissioned by VNS to address the calls for an in-depth look at the methodology behind the numbers. It found errors in several key areas, including:

* An exit poll sample size in Florida that was too small to produce accurate results.

* An undersampling of absentee and early voters that skewed exit polling results.

* A format for providing data that made the information “prone to misinterpretation.”

* A failure to clearly warn their members of possible inaccuracies in the data.

The report recommends significantly improving the statistical models used--including better methodology for estimating absentee votes and a more accurate system for determining how many votes a candidate needs to win.

All six organizations prematurely projected that Gore had won Florida, most before polls in the Panhandle had closed. Those projections became a major point of contention for Republicans during the heated 37-day legal battle over the state’s recount.

The networks declared Bush the winner within minutes of each other, following the lead of Fox News--where Bush’s first cousin John Ellis worked as an analyst. Associated Press, whose independent reports showed Florida tallies different from those of VNS, did not call the election for Bush.

CNN on Friday--as a result of its analysis--announced substantive changes in how it will report future elections.

Among other steps, the network vowed not to use exit polls to call close races and not to project winners until all polls in a state are closed.

Network executives also said that they would stay with VNS “if--and only if--significant changes are made to assure that the errors that plagued election night coverage in 2000 do not recur.”